In vivo 31p NMR as a noninvasive tool to study rodent cerebral metabolism as a function of age

Nuclear magnetic resonance (NMR) studies were performed with a two turn copper foil surface coil employing a unique transposition design that incorporates an explicit center tap ground of a widened bottom layer. Flexibility permitted easy conformation to tissues of interest thereby enhancing coil performance. Dielectric and inductive losses were thus reduced via effective shielding of electric fields and uniform distribution of magnetic fields. Construction details of the crossover coil and performance data for coils of varied dimensions are given. To establish age variant spectral saturation, apparent 31P longitudinal relaxation times, T1, values were measured for brain phosphates in young (5-6 mo), mature (11-12 mo), and aged rats (23-24 mo). Significant differences were found for alpha-, and beta-phosphate groups of adenosine triphosphate (ATP) and bone phosphate. A large, but nonsignificant change between young and mature rats was noted for the T1 of inorganic phosphate. Age related changes in tissue composition and biochemistry are discussed as possible contributors to these results. To assess the senescent response to stress, cerebral phosphate metabolism during mild hypoxia in young, mature and old rats was studied. Young rats displayed a greater drop in phosphocreatine (PCr), an earlier and greater onset of acidosis, and a larger rise in inorganic phosphate (Pi) than either the mature or old animals. PCr and intracellular pH (pHi) levels stayed low while Pi remained elevated after normoxia was reinstated. In contrast, all metabolite levels in the mature and senescent rats return to within 10% of control levels. Thus, young animals appear to maintain adequate ATP levels via increased anaerobic glycolysis thereby developing severe acidosis, while mature and senescent animals appear to maintain ATP levels by increased oxidative respiration rates. Speculations to explain this disparity are discussed in detail. The measurement of cerebral blood flow (CBF) via deuterium washout kinetics was developed in anticipation of calculating cerebral metabolic rates of oxygen (CMR[O2). Initially investigated on a vertical NMR system, the technique was found to have less complications when measured with the animal in a horizontal position. Values of global cerebral blood flow were similar to cited literature.

IN VIVO 31p NMR AS A NONINVASIVE TOOL TO STUDY RODENT CEREBRAL METABOLISM AS A FUNCTION OF AGE Jacob Abraham Stolk A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Medicinal Chemistry The University of Utah March 1990 Copyright © Jacob Abraham Stolk 1990 All Rights Reserved I'll E U 1': I VERStT,\, OF UT.\ 1-1 CRADl: ATE SCIIOOL SUPERVISORY COrvfiVIITTEE APPROVAL of a dissertation submitted by Jacob Abraham Stalk This dissertation has been read by each member of the following Sli pervisorv committee and bv majority vott' has been found 10 be satisfactory. Z ?:S ic Stephen A. Kuby ~ Issaku Ueda THE UNIVERSITY OF UTAH GRADUATE SCHOOL FINAL READING APPROVAL To the Graduate Council of the University of Utah: I have read the dissertation of Jacob Abraham Sto 1 k in its final fonn and have found that (1) its format, citations and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate School. Dale Chair, Supervisory Cammiuee Approved for the Major Department O~!hffi~r~ Ar hu D. Broom Chair/Dean Approved for the Graduate Council B. Gale Dick Dean aCThe Graduate School ABSTRACT Nuclear magnetic resonance (NMR) studies were performed with a two turn copper foil surface coil employing a unique transposition design that incorporates an explicit center tap ground of a widened bottom layer. Flexibility permitted easy conformation to tissues of interest thereby enhancing coil performance. Dielectric and inductive losses were thus reduced via effective shielding of electric fields and uniform distribution of magnetic fields. Construction details of the crossover coil and performance data for coils of varied dimensions are given. To establish age variant spectral saturation, apparent 3Ip longitudinal relaxation times, TIl values were measured for brain phosphates in young (5-6 mo), mature 01-12 mo), and aged rats (23-24 mo). Significant differences were found for a.-, and ~-phosphate groups of adenosine triphosphate (ATP) and bone phosphate. A large, but nonsignificant change between young and mature rats was noted for the Tl of inorganic phosphate. Age related changes in tissue composition and biochemistry are discussed as possible contributors to these results. To assess the senescent response to stress, cerebral phosphate metabolism during mild hypoxia in young, mature and old rats was studied. Young rats displayed a greater drop in phosphocreatine (PCr), an earlier and greater onset of acidosis, and a larger rise in inorganic phosphate (Pi) than either the mature or old animals. PCr and intracellular pH (pHi) levels stayed low while Pi remained elevated after normoxia was reinstated. In contrast, all metabolite levels in the mature and senescent rats return to within 10% of control levels. Thus, young animals appear to maintain adequate ATP levels via increased anaerobic glycolysis thereby developing severe acidosis, while mature and senescent animals appear to maintain ATP levels by increased oxidative respiration rates. Speculations to explain this disparity are discussed in detail. The measurement of cerebral blood flow (CBF) via deuterium washout kinetics was developed in anticipation of calculating cerebral metabolic rates of oxygen (CMR02)' Initially investigated on a vertical NMR system, the technique was found to have less complications when measured with the animal in a horizontal position. Values of global cerebral blood flow were similar to cited literature. v Dedicated to Mom, Dad, and Kim. TABLE OF CONTENTS ABSTRACT ................................................................................................................ i v LIST OF TABLES ....................................................................................................... ix LIST OF FIGURES ..................................................................................................... x LIST OF ABBREVIATIONS ................................................................................... xii ACKNOWLEDGMENTS ........................................................................................ xi v CHAPTER 1. INTRODUCTION ............................................................................................... 1 Significance .................................................................................................... 1 General Perspectives ........................................................................ 1 Aging Theories .................................................................................. 3 Background .................................................................................................... 5 StructuraL ........................................................................................... 5 Metabolism ........................................................................................ 10 NMR ................................................................................................................ 19 Historical Perspective ...................................................................... 19 2. THE CROSSOVER SURFACE COIL ............................................................... 26 Theoretical Principles .................................................................................. 26 Coil Losses ...................................................................................................... 29 Sensitivity and Noise ...................................................................... 29 Intrinsic Losses .................................................................................. 31 Dielectric Losses ................................................................................ 32 Inductive Losses ................................................................................ 33 Design Objectives .......................................................................................... 34 Methods and Materials ................................................................................ 37 Coil Construction ............................................................................. 37 Quality Factor "Q" ............................................................................ 41 Field Profile ........................................................................................ 42 Coil Performance .......................................................................................... 44 Results ................................................................................................. 44 Discussion .......................................................................................... 45 3. APPARENT 31P SPIN-LATTICE RELAXATION TIMES .......................... 51 Introduction ................................................................................................... 51 Materials and Methods ................................................................................ 53 Animal Preparation ......................................................................... 53 NMR ................................................................................................... 54 Results ............................................................................................................. 57 Discussion ...................................................................................................... 61 4. CEREBRAL PHOSPHATE METABOLISM DURING HYPOXIA AS A FUCTION OF AGE ......................................................................................... 66 Introduction ................................................................................................... 66 Materials and Methods ................................................................................ 69 Results ............................................................................................................. 73 Discussion ...................................................................................................... 82 5. CEREBRAL BLOOD FLOW MEASURED BY DEUTERIUM NMR ......... 89 Introduction ................................................................................................... 89 Materials and Methods ................................................................................ 91 Results and Discussion ................................................................................ 94 6. SUMMARy .......................................................................................................... 99 APPENDIX: PHYSIOLOGIC AND METABOLIC DATA .................................. 104 REFEREN CES ............................................................................................................ 113 viii 2.1 3.1 3.2 3.3 4.1 5.1 A.l A.2 A.3 A.4 A.5 A.6 A.7 A.S LIST OF TABLES Unloaded and Loaded Q-values for the Crossover Surface Coil. .......................................................................................... 47 Apparent Tl for ~-, (I-, y-ATP, PCr, Pi, and the Broad Component as a Function of Age .................................................... 59 Saturation Factors for ~-, (I-, y-ATP, PCr, Pi, as a Function of Age ................................................................................... 62 Chemical Shifts of 13-, (I-, y-ATP as a Function of Age ................. 62 Metabolite Ratios as a Function of Age .......................................... 75 Cerebral Blood Flow ............................................................................ 96 Blood Gas Analyses as a Function of Time .................................... 105 Intracellular pH as a Function of Time During Hypoxia .................................................................................................. 106 Phosphocreatine Ratios as a Function of Time During Hypoxia .................................................................................................. 107 Inorganic Phosphate Ratios as a Function of Time During Hypoxia .................................................................................... 10S Phosphomonoester Ratios as a Function of Time During Hypoxia .................................................................................... 109 ATP Ratios as a Function of Time During Hypoxia ................... 110 ADP Ratios as a Function of Time During Hypoxia ................... 111 Phosphorylation Potential Ratios as a Function of Time During Hypoxia ......................................................................... 112 LIST OF FIGURES Figure Page 1.1 A side (lateral) view of the cerebral cortex ........................................ 8 1.2 The neuron .............................................................................................. 9 1.3 Stages in the extraction of ATP from foodstuffs .............................. 12 1.4 The five states of mitochondrial respiration .................................... 14 1.5 31p NMR visible metabolites ............................................................... 22 1.6 A typica1 31p brain spectrum ................................................................ 23 2.1 Diagrammatic representation of the magnetic field ....................... 28 2.2 Standard tuning-matching circuitry ................................................... 31 2.3 Schematic representation of dielectric losses ................................... 33 2.4 Inductive losses ....................................................................................... 34 2.5 The balance-matching scheme ............................................................ 35 2.6 Layers of the crossover coil. .................................................................. 38 2.7 The assembled crossover surface coiL ............................................... 39 2.8 The explicit balance-matching scheme ............................................. .40 2.9 Measurement of Q using reflective power ....................................... 43 2.10 Q curves of a 31p surface coil. .............................................................. .46 2.11 Bl field profiles for the 23 x 18 mm IH crossover surface coil.. .... 48 3.1 Relaxation mechanisms ........................................................................ 53 3.2 The modified pulse-burst saturation recovery sequence ............... 55 3.3 A typical inversion recovery experiment. ........................................ 58 3.4 Graphical representation of apparent Tl values .............................. 60 4.1 31p spectra before, during, and after hypoxia of young, mature and old rats ............................................................................................... 74 4.2 Arterial oxygen tension, Pa02, and pH, pHa, as a function of time ............................................................................................................ 77 4.3 Intracellular pH, pHi, and metabolic ratios as a function of time ............................................................................................................ 78 4.4 The phosphorylation potential as a function of time .................... 81 5.1 Deuterium washout curve for a "common carotid" experiment. .............................................................................................. 95 xi Abbreviation ADP AMP ATP Bl CBF CMRglu CMR02 Cr CT FAD FADH2 AD GRASS MRI NAD NADH NMR OER PaC~ Pa02 PCr LIST OF ABBREVIATIONS Adenosine diphosphate Adenosine monophosphate Adenosine triphosphate Radiofrequency magnetic induction field Cerebral blood flow Cerebral metabole rate of glucose Cerebral metabolic rate of oxygen Creatine Computerized tomography Flavin adenine dinucleotide (oxidized form) Flavin adenine dinucleotide (reduced form) Free induction decay Gradient recalled acquisition in the steady state Magnetic resonance imaging Nicotinamide adenine dinucleotide (oxidized form) Nicotinamide adenine dinucleotide (reduced form) Nuclear magnetic resonance Oxygen extraction ratio Arterial carbon dioxide tension Arterial oxygen tension Phosphocreatine PDE Phosphodiester PET Positron emission tomography pHa Arterial pH pHi Intracellular pH Pi Inorganic phosphate PME Phosphomonoester Q Quality factor of an electronic circuit RCR Respiratory control ratio TJ Longitudinal relaxation time T2 Transverse relaxation time TE Echo delay TR Repetition delay xiii ACKNOWLEDGMENTS I would like to express my deepest gratitude to Dr. Marty Schweizer for his patient guidance throughout my graduate career. His faith in my decisions and abilities has enabled me to develop a sincere appreciation of the science to come. Many thanks go out to Dr. Donald W. Alderman for his expert suggestions and guidance with the design of the crossover surface coil and to Jay I. Olsen for frequent and lengthy discussions about how NMR really works. My sincerest appreciation goes out to Mom and Dad for their never ending support, love and encouragement even though I'm sure it must have been exasperating at times. But my gratitude to Kim for her patience, love, support, friendship and encouragement, which was especially needed during the last phase of my graduate career, cannot be overstated. Her endureance for my, at times, "flakey" behavior is deeply appreciated and admirable. Finally I would like to thank all my dear friends at the University of Utah and beyond. The many evenings, BBQ's, weekends in Southern Utah, and bike rides spent together were fantastic and necessary to keep me motivated. CHAPTER 1 INTRODUCTION Significance General Perspectives Aging, from the Latin word aetas meaning age or lifetime, is not easily described in an all-encompassing definition. Timiras proposes three major definitions that reflect the primary branches of aging research.1 1) Aging as a stage of the lifespan: the sum of all changes occurring in an organism with the passage of time. 2) Aging as a deteriorative process: the sum of all changes occurring with time and leading to functional impairment and death. 3) Aging as cellular and molecular damage: changes in membranes, cytoplasm, and/or nucleus. These definitions pOint out the existence of a difference between "normal" aging and pathological aging. The latter is not so difficult to envision as being some degenerative process leading to moderate to severe malfunction of physiological and mental homeostasis. The fine line between geriatric illness and changes normal to the aging process is, however, more nebulous. Colarusso and Nemiroff offered the following definition of normal: "that mental activity and behavior which, based on current knowledge, falls within the broad range of predictable expectation, fulfills developmental potential, and is adaptive to the society in which it occurs.,,2 Thus, they consider normality within the changing biopsychosocial context of the aging process. In light of demographic studies, gerontology, the study of aging, 2 seems warranted. In 1910, 4% of the U.S. population was over 65 years of age; in 1980 this number had increased to 11 %. At present 13% of the population is older than 65,1, 3 and in 2030, it is predicted that 21 % of the population will be over 65.4 Even more rapid is the increase in the population over 75 years of age, expected to be over 8% in 2030. It is predicted that, by the year 2000 almost half of the deaths in the United States will occur after age 80.5 The increased numbers of elderly in today's society is undoubtedly due to the increase in medical knowledge and technology. This progress, however, has put an increasing financial burden on federally funded health care. A summary of a report submitted by the U.S. Senate Special Committee on Aging states the following:6 More than four out of five persons 65 and over have at least one chronic condition that needs medical attention. One out of four elderly have at least a mild degree of functional disa bili ty. Federal spending has nearly doubled since 1960. In 1986, $270 billion or 26% of the federal budget was of direct benefit to older Americans. Federal expenditures on the elderly are projected to increase from 3% of gross national product (GNP) to more than 6% in 2030. Among the many aging processes, impairment of mental faculties is perhaps one of the most debilitating. Central nervous system (CNS) disorders produce nearly half the disabilities in individuals beyond the age of 65. 1,6 Impaired memory, intellect, sensory activity, balance, and/or cognitive activity are some of the most frequent disabilities of old age. However, more and more research is pointing out that cognition, memory, 3 and learning ability may not be affected as severely in the normal aging adult as was originally thought?,8 It is becoming increasingly clear that the senescent adults can maintain mental and physical faculties if activities in these areas are kept up throughout one's lifespan. Maintenance of a healthy mind and body can not only increase longevity, but can also increase the quality of this longevity. Decline of mental and physical faculties need not be a given.9, 10 It cannot be disregarded, however, that structural and physiological changes do take place in the aging body. That the individual is nevertheless able to maintain functional integrity implies a plasticity, especially of the nervous system, that has only begun to be investigated. Consequently, it becomes important that research be conducted toward obtaining information on the normal and diseased aging brain. Neuroanatomy and neurometabolism studies are required, to address both clinical and financial problems associated with the "graying" of America. Aging Theories The study of aging can be divided into many areas of research, ranging from the molecular and cellular approach to the systemic and organismic approach. ll, 12 Molecular theories presume that there is some genetic program that determines the maximum lifespan for each species. Thus the codon restriction theory of aging is based on the hypothesis that the fidelity or accuracy of translation is impaired with aging.13 Cellular theories relate to changes that occur in structural and functional elements of cells with passage of time. They relate to the build up of damaging materials. Thus there is clear evidence of lipofuscin buildup in the cortex and hippocampus.14, 15 It has also been shown that free-radical 4 reactions are increasingly involved in aging and age related disorders.16,17 Biologic macromolecules develop cross-linkages or bonds between identical molecules with the passage of time, thereby altering their physical and chemical properties.18 The overall performance of an animal is closely related to the effectiveness of a variety of control mechanisms. Aging theories at a systemic level ascribe aging of the entire organism to decrements in the function of a key system, such as the nervous system, endocrine system, and immune system. Aging at the neuroendocrine level is primarily concerned with the maintenance of homeostasis.19 The neuroendocrine system is a complex mechanism that consists of the brain, the most important component of which (from a neuroendocrine point of view) is the hypothalamus, the pituitary, the target glands (thyroid, adrenals, gonads, etc.), tissues directly controlled by the pituitary and target glands, and other hormone producing tissues. Any adaptation to stress, either as a result of external (environmental) stimuli or internal (emotional, hormonal, immunologic, and metabolic) stimuli, depends on control mechanisms orchestrated by the nervous and endocrine systems. A decline in efficiency of this system will result in decreased function and increased pathology of most organs and tissues. The immune system protects the individual from a variety of potentially harmful substances and organisms. It has been shown that a decreased immune efficiency is primarily due to a reduced thymus and T cell function. 20 In this respect, conventional theory has predicted a decreased immune response in the elderly, and as a result an increased number of infections. However, the converse is probably more true.21 Most 5 likely, elderly persons have previously produced antibodies reactive to a sufficiently broad range of environmental pathogens to provide them with protection despite their reduced ability to produce new antigens. Increased infections seen in many nonambulatory or nonactive elderly are hence more likely due to mechanical factors rather than immunological ones. Though from the above brief overview it is apparent that all systems are affected in some way by senescence, it also may be suspected that at the heart of all normally functioning organisms lies a healthy central nervous system. It is the central nervous system that "oversees and orchestrates" all adaptive mechanisms required in response to internal and external insults in order to maintain homeostasis. Hence, this dissertation will address some anatomical, but mostly metabolic, changes associated with aging of the central nervous system and how the aging nervous system responds to stress. Background Structural Studies of structural changes in the aging brain were initially limited to the measurement of brain size. In 1860, it was reported that the human brain decreased in weight by an average of 6.6% between 20 and 80 years of age.22 A latter study found that the decrease in brain weight is linear and can be as large as 11 %.23 In contrast, a minimally significant loss in brain weight was reported between nondemented males less than 50 and males older than than 65.24 Furthermore, no significant decrease in brain weight was found for individuals older than 65, and all mean weights fell within the range of normal adult brain weights. Much of the variance is mostly due to methodology: when after death is brain weight measured; how much brain 6 tissue is removed; whether blood vessels and meninges are included in the measurement; body weight of the individual before death. These are all important yet rather variable factors. The relationship between body weight and brain weight is a well­accepted fact by neuroanatomists and neuropathologists.25 In light of this relationship it was suggested that because the mean body weight of humans has increased during the past century, cross-sectional studies show a spurious decline in brain weight with advancing age.26 Thus, younger people of today are larger and heavier than younger people of earlier generations, and one would expect that, simply based on body size, the younger, larger individuals would have a heavier brain than the average 75 year old. Brain volume in humans is traditionally assessed by measuring ventricular and sulcal dimensions in autopsy specimens. Precomputerized analyses include an increase in ventricular size with age,27 an increase in sulcal size of the frontal lobes,28 and a widening of cortical fissures and dilation of anterior horns of the lateral ventricles.29 However, these results were obtained postmortem, and accuracy may be compromised by methodological problems related to the time of measurement after death. With the advent of Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) , in vivo studies of brain volume and brain atrophy have become possible.30,31 CT studies include visual ratings by experienced observers, measurements of a variety of ventrical-to-brain indices, and volumetric pixel counting. These studies generally agree that a progressive enlargement of the ventricles and cortical sulci (liphysiologic atrophy) is a characteristic of normal aging.32 It has generally been assumed that any shrinkage of the brain with 7 increasing age is due to a loss of neurons. One estimate claims a reduction in total neuron count of 19.7% in adults older than 70.33 However, the aging process does not appear to affect all brain regions equally. In 1955, initial studies on neuron density in the cerebral cortex were performed manually.34 Across the age range of 65-95, a significant decrease in neuron density in the superior temporal gyrus, superior frontal gyrus, and the area striata (Figure 1.1) was demonstrated. However, no significant decrease in the postcentral gyrus or inferior gyrus neuron density was noted. Similarly, no significant decrease was found in any region of the brainstem, including the facial nerve nucleus, ventral cochlear nucleus, and the abducens and trochlear nuclei. A computerized counting methodology developed by Henderson et a1.35 allowed evaluation of differently sized neurons within the same cortical regions as those in the 1955 study.36 It was found in both studies that the number of "small" neurons (~12, <19 J.1m) decreases differently than the number of "large" neurons (~19 J.1m). Thus the overall decrease in small neurons was found to be 35% while the decrease in large neurons was 50%. Though decreases in the inferior temporal gyrus were least for both neuron sizes, a significant decrease, in contrast to the earlier study, was found in the postcentral gyrus. It is possible that general shrinkage of the larger neurons accounts for their greater density reduction. In other words, large neurons become small neurons with increasing age. An extensive study involving 120 brains measured cell size with a computerized digitizer to determine neuronal populations and neuronal densities.37 As well as correcting for the apparent small neuron increase, a strong emphasis was placed on correcting comparisons of neuronal densities of young versus old age groups for age-dependent differences in tissue shrinkage during processing. As a result, it was found that there is Cerebellum Central Sulcus I Precentral Gyrus = precentral motor cortex n Postcentral Gyrus = somatosensory cortex ill Superior Temporal Gyrus IV Middle Temporal Gyrus V Inferior Temporal Gyrus VI Area Striata = visual cortex VII Superior Frontal Gyrus Frontal Lobe Figure 1.1 A side (lateral) view of the cerebral cortex. Relevant anatomical features are shown and tabulated numerically. 8 9 noneuronal loss in any of the cortical areas, and in fact an increase in total neuron numbers was found in the visual cortex. It was recently pointed out that the fixing solution, methyl benzoate, is a probable cause of age­dependent differences in tissue shrinkage. Using a more traditional fixing solution, xylene, it was shown that the dependence between tissue shrinkage and age is no longer significant.38 However, it is essential to edit for small and large neuron differences. This study also concluded that the salient shrinkage of brain volume is due to a shrinkage of large neurons with consequently increasing number of small neurons. Recent studies thus appear to dispel the pessimistic conclusions that longevity can only be achieved at the expense of profound cortical neuronal depletion. In Figure 1.2 a schematic drawing is shown of a neuron and its components. A neuron consists of a cell body containing the nucleus, Dendrites Cell Body Axon Terminals Figure 1.2 The neuron. The neuron functions by the transmission of electrical impulses through the axons, from one dendrite to another. Axon terminals release neurotransmitters into the synaptic cleft where they induce an electrical pulse in the following dendrite. 10 dendrites receiving impulses, axons, transmitting impulses, and axon terminals that release neurotransmitter into the synaptic cleft. Controversy exists concerning the changes that take place at the neuron level.39 Certainly, much of the controversy is due to methodology, but also the large variability seen in different brain regions gives rise to misunderstanding. Early studies suggested that dendritic trees of surviving neurons were atrophied in normal human aging.40 However, no change was seen in synaptic density throughout adult life (ages 16-72 years) and only a slight decrease was observed in brains of older individuals.41 A substantial net dendritic growth was seen in neurons of the parahippocampal gyrus,42 a section of the cortex medial to the middle temporal gyrus. Similarly, dendrites of neurons in the dentate gyrus, a section of the hippocampus, showed growth in early old age (S;;75) but were observed to regress in late old age (>75).43 So it may be concluded that, similar to the variance in neuron density with increasing age, dendritic growth, and thus the number of synapses, can vary in different sections of the brain. A current area of interest is the theory of stimulating dendrite growth and arborization in mentally and physically active individuals. Diamond has already shown this to be the case for rats,44 and this sort of plasticity could be a plausible explanation for successful aging in mentally and physically stressful environments. Metabolism The affect of aging on metabolic processes in the brains is a valid question considering the structural degradation of the brain. Does the change in neuron number, density, and size also affect metabolism, and if so, can this be an explanation for the apparent age-related decline in 11 physiological function in nonsuccessful aging (i.e., aging associated with pathological conditions)? Many aspects of metabolism can be studied. Among these are metabolism of neurotransmittors, protein metabolism, lipid metabolism, and energy metabolism. In general it can be said that normal aging is associated with slight changes in neurotransmitter levels.45, 46 It is, however, important to note that many problems exist, making an accurate assessment of age-dependent neurotransmittor levels nontrivial. Protein synthesis was found to remain unchanged in rat brain in a recent assay using an initiating cell-free system.47 Though specific activity may be altered, rate constants and amino acid composition are not affected by increasing age.48 Overall lipid synthesis does not seem to be age-dependent. However, it is evident that membrane lipid composition changes with increasing age, thereby giving rise to changes in membrane fluidity.49 The different areas of research are extremely vast and a selection of one is necessary in order to study it adequately. It must be kept in mind however that all metabolic processes are closely linked, and in fact most rely on the synthesis of adenosine triphosphate (ATP) and reduced nicotinamide adenine dinucleotide (NADH). A change in the synthesis of the latter could have a rate limiting effect on many metabolic reactions. It is a well-known fact that the primary form of fuel in biological systems is ATP. Its extraction from the oxidation of foodstuffs follows three major stages (Figure 1.3). In the first stage, large food molecules are broken down into smaller units, producing no useful energy. In the second stage, the smaller molecules are converted into acetyl CoA units that playa central role in ATP generation. This stage is generally referred to as anaerobic Polysaccharides ~ Glucose ATP NAD+ ..... ...,....-F---- NADH FAD ..... :lII-{ Proteins Fats . ~ . Ammo aCids ~. Fatty aCids and g1 ycerol NADH ATP Stage I Stage n Stage ill 12 Figure 1.3 Stages in the extraction of ATP from foodstuffs. Stage I produces no useful energy. Stage II produces a small amount of ATP, NADH, and FADH2. Stage III, consisting of the citric acid cycle which produces much more NADH, and FADH2 used in oxidative phosphorylation to produce large quantities of A TP. 13 metabolism and is associated with the production of a small amount of ATP, NADH and reduced flavin adenine dinucleotide (FADH2)' The complete oxidation of acetylCoA to C02 in the citric acid cycle takes place in stage III, and is referred to as aerobic metabolism. NADH and FADH2, produced abundantly in this cycle, are used by the mitochondria to generate large quantities of A TP at the expense of oxidizing 02 to H20. This is known as oxidative phosphorylation. In addition to these processes, which are net energy producing processes, the brain has two additional mechanisms to help maintain ATP at constant levels. These are the creatine kinase and adenyl ate kinase reactions.50, 51 Creatine kinase catalyzes the reversible transfer of phosphate between phosphocreatine (PCr) and ATP: PCr + Mg-ADP- + H+ P Mg-ATP2- + Cr (1.1) Adenylate kinase effectively transfers the energy residing in the p­phosphate group of the original ADP molecule to produce A TP, as shown in equa tion 1.2. Mg-ADP- + ADp2- P Mg-ATP2- + AMp2- (1.2) Under steady state conditions, the reactions described above ensure that the rate of energy consumption by endergonic reactions is exactly balanced by the rate of ATP synthesis. Research on age-dependent metabolism has mostly concentrated on the changes occurring within Stage III: the citric acid cycle and oxidative phosphorylation. Besides being the most important site of energy regulation, this is also due to the relative ease of mitochondrial isolation. Based on the substrate availability in a mitochondrial suspension, five states of respiratory function can be identified (Figure 1.4).52 In the presence of oxygen but absence of substrate and adenosine diphosphate ADP Substrate 1 J ..... Time 5 State: 1 2 3 4 5 Oxygen: + + + + - Substrate: - - + + + ADP: - + + - + 14 Figure 1.4 The five states of mitochondrial respiration. The plot is of oxygen concentration of a mitochondrial suspension as a result of the addition of ADP and/or an oxidizable substrate such as succinate or glutamate. (ADP), mitochondria display state 1 respiration. The addition of ADP to the suspension results in an increase in oxygen consumption establishing state 2 respiration. The addition of both ADP and substrate yield state 3 respiration. State 4 respiration is attained when all ADP is metabolized to ATP. When all oxygen is consumed anaerobic energy synthesis is initiated and the mitochondria can no longer synthesize A TP (state 5). Dividing the rate of oxygen consumption in the presence of ADP (state 3 respiration) by the rate of oxygen consumption in the absence of ADP (state 4 respiration) yields the respiratory control ratio (RCR), a sensitive indicator of mitochondrial integrity.53 Using glutamate, malate, glutamate+malate, and pyruvate plus malate as substrates, Chiu and Richardsen have shown that the RCR decreases as a function of age in isolated brain mitochondria.54 Though state 4 respiration remained constant with increasing age, state 3 respiration 15 decreased using the named substrates. However, no change was observed with succinate as a substrate, implying that changes may be occuring in the NADH dehydrogenase complex. An age-related reduction was also observed for cytochrome oxidase activity in crude mitochondrial brain homogenates.55 Based on these results, it appears there is a definite impairment in the mitochondrial machinery. This view is supported by recent findings that some inner membrane protein fractions, corresponding to NADH dehydrogenase and cytochrome c, decrease in abundance with age.56 The authors also found an increase in other inner membranes fractions possibly corresponding to cytochrome b and subunits of cytochrome c oxidase, suggesting that there may be a decreased degradation rate as well as decreased rate of protein synthesis. Sylvia and Rosenthal concluded that in the brains of living rats, the percentage of steady state reduction of cytochrome aa3 was unaffected by aging.57 However, in a latter study these authors found an age related decrease in the capacity to respond to higher energy demands.58 The study by Syllvia and Rosenthal also illustrates the disadvantage of mitochondrial studies: cell fractions may be damaged during isolation. In contrast, whole tissue or organ preparations preserve the intracellular milieu. One such system is the isolated perfused heart. It was shown that work performance in hearts from senescent rats changed little with increasing age,59 and that ATP generation by oxidative phosphorylation was not significant! y influenced by aging. With respect to endogenous substrate availability, the primary source of acetyl-eoA is from pyruvate, a product of glycolysis. It is thus a logical query as to whether the enzymes responsible for carbohydrate metabolism are affected by age. Ulfert et. al. looked at glycolytic metabolites in brain 16 extracts and found that arterial glucose concentrations were significantly higher in 2 year old rats as compared to 6 month old rats.60 However, glucose-6-phosphate levels were unchanged as were all other glycolytic metabolites implying no glycolytic enzymatic changes with age. After extensive review of the literature prior to 1983 concerning both in vitro and in situ mitochondrial studies,61 Hansford conduded that senescent mitochondria seem to remain undamaged morphologically and maintain enzymatic entigrity, though some substrate specific changes are present. Furthermore, the discrepancies observed in the literature may very well be due to imperfect isolation procedures as well as nonphysiological environments. In fact initial studies on whole organs appear to indicate that there is no or little effect of aging on the ability to synthesize ATP under resting conditions. In the last two to three decades, much attention has been given to the in vivo study of brain metabolism in an attempt to solve the riddle of significant age variance in mitochondrial respiration seen in isolated mitochondria yet no or little variance seen in whole organ studies. Roy and Sherrington suggested that cerebral blood flow (CBF) is regulated to meet the requirements of cerebral metabolism and function.62 Since then efforts have been directed towards evaluating cerebral function by measuring CBF, cerebral metabolic rate for oxygen (CMR02> and regional cerebral metabolic rate for glucose (CMRglu) in an effort to unravel the in vitro versus in vivo mystery. The measurement of oxygen utilization was made possible by the pioneering work of Kety and Schmidt.63 Firstly, CBF is determined via the Fick Principle, which states that the quantity of any substance taken up in a given time by an organ from the blood that perfuses it is equal to the total 17 amount of the substance carried to the organ by the arterial inflow less the amount removed by the veinous drainage during the same period. After determination of CBF, the oxygen utilization is calculated by multiplying the CBF with the oxygen extraction ratio (OER), which is determined from the A-V 02 difference. Early studies employed inhaled N20 as the substance and found that CBF and CMR02 declined significantly from the first to the third decade of life followed by a more gradual decline into the tenth decade.64 However, the health status of the patients leaves some doubt as to whether the observed changes were age dependent per se, or that perhaps pathological conditions contributed to the change.65, 66 It is possible that much of the variance arises due to inconsistent criteria used for patient selection. It should be noted that no decrease in CBF or CMRo2 was found in rats using nitrous-oxide,67 or [l4C]-iodoantipyrine.68 However, one must be careful to extrapolate results obtained from rodents to humans, as the former show less age-related neuropathologies. CMRglu can be measured after a bolus injection of 2-deoxy-D-[1- 14C]glucose ([14C]DG). After initial phosphorylation by hexokinase, the [14C]DG-6-phosphate becomes a poor substrate for further enzymatic degradation and builds up in (cerebral) tissue.69 Autoradiographs of brain slices then show at what rate glucose is phosphorylated and a quantitative measurement of glucose utilization is obtained. Using this technique, it was clearly demonstrated that cerebral metabolic rates for glucose in rats change significantly during the first 12 months, yet show no significant change in the following 18 months?O, 71 Obviously, no such invasive measurements are possible on humans. Though noninvasive CBF and CMR02 measurements have been 18 possible for some time, the measurement of CMRglu and ATP metabolism has mostly been limited to animal studies. With the development of positron emission tomography (PET), and in-vivo nuclear magnetic resonance imaging and spectroscopy (MRI and MRS), real time, simultaneous, noninvasive observation of CBF, CMR02, and CMRglu, and phosphate metabolism has been realized. In addition, older techniques are again revitalized to study basal metabolic rates in the aging organism in light of the changing bias that getting older equates to reduced metabolic capacity. PET measures the emission of gamma rays that are a result of an annihilation event between an electron within the tissue and a positron originating from a positron emitting radioisotope such as 150, 11C, and 18F.72 Not only is the method noninvasive, it has the additional benefit of supplying regional metabolic information. For measurement of CBF and C M R 0 2 studies, the technique relies on the measurement of 150 concentration in regions of the brain during sequential inhalation of C150 2 and 1502.73,74 In a single experiment regional blood flow, blood volume, oxygen extraction fraction, and oxygen consumption can be determined. For the measurement of CMR02 the Sokoloff model is applied after the injection of [18F]2-fluoro-2-deoxy-D-glucose (FDG).75 Frackowiak et aL noted a decline in CBF and CMR02 in gray matter but no change in white matter in a study of healthy aging adults, screened for atherosclerosis.72 Similar results were recently obtained with 133Xe inhalation.66 With respect to glucose metabolism, PET studies have shown no or only a slight decrease in frontal and superior cortex.76, 77 A recent study, on more advanced instrumentation, found that while cortical volume was diminished, cortical glucose metabolism remained unchanged.78 19 It is still not entirely clear whether CBF, CMRo2.1 and CMRglu are age variant in the resting adult human. However, there does appear to be a trend of decreased flow, yet stable oxygen and glucose metabolic rates. It is thus necessary to continue research on normal aging in order to elucidate compensatory mechanisms by which the senescent brain can maintain the coupling between CBF and CMRglu Due to the significant role of A TP in brain function,79 much would be gained if it would be possible to study mitochondrial function based on ATP metabolism in vivo. The developments in Nuclear Magnetic Resonance have made this possible. NMR Historical Perspective Nuclear magnetic resonance spectroscopy, NMR, was first observed in the mid 1940s.80, 81 With the advent of larger magnetic fields, improved electronics, and Fourier transform methods providing increased resolution and sensitivity, it became possible to study increasingly complicated compounds. The application of NMR to macromolecular biological systems became possible with the development of sophisticated pulse sequences.82 NMR was applied to organ systems by Moon and Richards who discovered that 31 P NMR signals could be observed in intact human erythrocytes.83 Chance et al. obtained 31p NMR spectra of the head by placing an anesthetized mouse into an 18 mm glass NMR tube with a radiofrequency coil surrounding the entire tube.84 Ackerman et al. minimized the contribution of tissue outside the organ of interest with the development of the surface coil.85 Many advances have been made since these preliminary experiments. Localization techniques have been improved,86 and spectroscopic imaging is being developed that will supply 20 the researchers with metabolic maps of tissues of interest.87,88 Observation of nuclei other than 31p, such as lH, 13C, and 19F has been realized.89 Proton density has been the basis of NMR imaging and the technique has proven to be a valuable diagnostic tool for the radiologist.90 However, the same abundance of water and lipid protons makes lH spectroscopy difficult due to overlap of these resonances with metabolically relevant species. DC spectroscopy solves this problem, but is plagued by the extremely low natural isotopic abundance. Metabolically relevant data can be obtained only after the infusion of large quantities of expensive 13C labeled substrate such as 13C-glucose. 91 Because it is not involved in endogenous metabolism, the naturally 100% abundant 19F NMR lends itself extremely well to clinical research. 19F labeled therapeutic agents, such as 5- fluorouracil, 5-FU, can be followed, giving the pharmacologist a localized picture of drug uptake and metabolism. The advantage of 31p NMR is its natural abundance (100%) and its biological importance in homeostasis. Complete understanding of NMR requires quantum mechanical formulation. Fortunately, a classical picture describes NMR adequately for most applications. The magnetic moments of a spin 1=1/2 nucleus in a magnetic field will exhibit two orientations: one aligned along the magnetic field, the other against it. There will be an excess of spins aligned with the field as this is a lower, and thus favored, energetic state. Hence there will be a net magnetic moment aligned with the field. This moment is tipped after the application of a small perpendicular field. When the field is turned off, the moment relaxes to its equilibrium position, thereby inducing a signal in the receiving coil. The signal obtained after pulsing the spins is termed a free induction decay (FlD) The frequency at which a specific nucleus resonates is proportional to 21 the static magnetic field. Electrons surrounding each nucleus result in additional small local magnetic fields. As a result the resonance frequency is slightly shifted by an amount known as the chemical shift, 0, which is measured relative to the resonant frequency of a reference compound. Thus, in 31p NMR the PCr peak is the accepted reference point because its resonance frequency is stable under a variety of physiological conditions. Figure 1.4 shows the different structures visible by 31 P NMR with their corresponding chemical shifts relative to PCr. It should be noted that the three ATP peaks have another contributor. Adenosine diphosphate (ADP) and other soluble nucleotide contribute to the signaL However, this contribution is less than 10%. The phosphodiester and phosphomonoester peaks are envelopes of different species, all resonating at resonance frequencies which cannot be resolved by in vivo NMR. Only an example of each of these are indicated in Figure 1.5, and different species contribute in varying degrees. The chemical shift remains fairly constant for most metabolites. However, the Pi chemical shift changes with a changing intracellular pH, (pHi).92 The peak shifts down field, away from PCr, with increasing pH, and upfield, towards the PCr peak, with decreasing pH. When a surface coil is used for brain spectroscopy, an additional broad peak is observed. This peak is primarily bone phosphate, but has contributions from lipid and other immobile phosphates. These compounds have fast transverse relaxation rates resulting in broadening of the resonance peak. A typical spectrum obtained at 4.7T is shown in Figure 1.6. The first reported clinical application of 31p NMR spectroscopy was in a patient with muscle cramps, which biopsy showed to be due to myophosphorylase deficiency (McArdle'S syndrome).93 Arnold et. aL investigated the hypothesis that phosphocreatine resynthesis may be a Glucose-6-Phosphate a phosphomonoester (PME) 00. = 6 -7 ppm o II 1)-&-O~2~ OH OH OH Glycerolphosphorylcholine a phosphodiester (PDE) 00. = 2 - 3 ppm CH20H I CHOH 0 CH3 I II I + H2C-O-P-O-C-C-N-CH3 b H2 H2 tH3 Inorganic Phosphate (Pi) 00. = 4 - 5ppm o II -o-P-OH I a Phosphocreatine (PCr) 00. = 0 ppm a NH I II -O-P-N-C-N-CH2-Coa I H I a CH3 Adenosine Triphosphate (A TP) 00. = (-2) - (-3) ppm o~ = (-6) - (-7) ppm 0)' = (-16) - (-17) ppm :t,N t~.Jl ~ o 0 0 N N 11"( II~ IICl -o-p-o-p-o-P-O-CH2 I I I 0 a a a OHOH 22 Figure 1.5. 31 P NMR visible metabolites. Chemical shifts are referenced to phosphocreatine (Per). ADP peaks are hidden under the ATP resonances. Many species resonate at frequencies similar to the examples of phosphomonoesters (PME) and phosphodiesters (PDE). These are not resolved in in vivo NMR. PCr I PME j'-ATP I a-ATP I ~-ATP I 1"",""""']""1""1'" '''''''' ".'"'' 'I" \'" ""',' if 313 2~ 10 e -10 -20 -30 PPM 23 Figure 1.6 A typical 31p brain spectrum. Shown from left to right: PME, Pi, PDE, PCr, Y·, a-, and ~-ATP. The broad baseline hump is from bone phosphate and other immobile pho.sphates such as lipids. 24 index of mitochondrial function.94 They determined that PCr regeneration per se is not an adequate indicator of myocardial function due to the effect of lactic acid build up. However, the calculated rate of free ADP proved to be a more precise diagnostic. Studies of mitochondrial function, using exercise, have since been extended to several myocardial myopathies.95 Although 31p NMR has contributed to the understanding of muscular metabolism, much of the NMR findings had previously been shown by biopsy techniques. This is not so for brain metabolism where human biopsy is not feasible. Validation studies from several laboratories have shown that the behavior of metabolites observed by 31p NMR follow expected patterns based on conventional and model systems during hypoxia, ischemia, seizures, etc.96 Normal metabolite ratios for a variety of animal models were soon available and, under high stress conditions, ATP levels are maintained at the expense of PCr and metabolic acidosis. Human spectroscopic studies were first introduced by Bottomly et a1.97 Unfortunately, few investigations into normal aging metabolism remain possible in humans. Few studies are available and no normal statistically relevant values have been published. Often some sort of stress is required to gain insight into mechanistic aspects of homeostatic maintenance. Nevertheless, some pathological conditions have given insight into ATP metabolism in vivo as a function of age.98, 99 Elevations in both PME and PDE were observed in brains of Alzheimer's patients as compared to controls. Reduced PCr/~-ATP and increased PME/PCr ratios were also noted. Due to its capacity to directly measure oxidative phosphorylation metabolites, as well as intracellular pH, 31p NMR spectroscopy is ideally suited to investigate mitochondrial function. As it is a noninvasive 25 technique, it permits mitochondrial study under physiological conditions. The following sections will discuss the employment of in vivo NMR toward the elucidation of metabolic mechanisms as a function of age in rat brain. In Chapter 2 the development and workings of a highly shielded surface coil are discussed. Chapters 3 and 4 describe the application of this coil to the study of some metabolic aspects of aging including Tl and hypoxia studies, studied by in vivo 31p NMR. Finally, Chapter 5 covers the latest developments in the use of 2H NMR for the study of cerebral blood flow. Chapter 6 will summarize the findings of this research and describe some of the studies that may shed additional light on the metabolic mechanisms involved in the aging brain. CHAPTER 2 THE CROSSOVER SURFACE COIL 100 Theoretical Principles The application of nuclear magnetic resonance (NMR) to in vivo studies has presented the experimentalist with problems not usually encountered in high resolution NMR. The size and geometry of the coils can be quite varied depending on the particular area that is to be studied. The specific configuration that is required of such coils no longer guarantees a homogeneous field across the sample as is relied upon in high resolution systems. In other words the magnetic field experienced by a specific volume is a function of the distance it is removed from the coil as well as the coil's shape. It was furthermore recognized early in the development of in vivo NMR that losses due to conducting tissue are a significant problem in whole animal experiments.101,102 The design of a radio frequency (r.f.) coil and its magnetic properties can best be considered in terms of the theory developed by Hoult and Richards.103 Consider a single turn rJ. coil that serves both as transmitter and receiver (transceiver). The magnetic field experienced by small sample volume, BV 5, with coordinates (p, x) after application of a current through a coil is given by:85 (Bl)xy = Bx + (Bp)xy where (Bp)xy is the perpendicular component of Bp given by: (Bp)xy = Bp sin e (2.1) (2.2) 27 where e is the angle between the directions of B p and B o' These relationships can be visualized in Figure 2.1. Nuclei within the test sample will precess about the static magnetic field at their Larmor frequency, COo, given by (2.3) where 'Y is the gyromagnetic ratio of the nucleus. Following application of a current through the coil, known as an r.f. pulse, the induced magnetic field causes the nuclei to precess away from their equilibrium position at an angular rate given by COl = 'Y (B l)xy (2.4) For a constant amplitude pulse of duration, td, nuclei will process about an angle (2.5) The net magnetic moment of the nuclei, then, has a magnitude MooV s, where Mo is the equilibrium magnetization of the test sample. However, only the component of the magnetic moment in the x-y plane will induce a current in the coil upon relaxation of the nuclei to their equilibrium position. The magnitude of this moment, m, is given by m = Mo oVs sina (2.6) Because the magnetic field B1 is a function of distance from the coil, the magnetic moment that can induce a voltage is also a function of position. The voltage that is induced in the coil, or electromotive force (€), is given by Faraday's law (2.7) where <l> m is the flux in the coil originating from the rotating moment dipole m. 28 x Figure 2.1. Diagrammatic representation of the magnetic field. B1 is produced by a current passing through a single turn coil and consists of its perpendicular components Bp and Bx. Other parameters are: Bo, the static magnetic field; r, the coil radius, i, the current flowing in the coil, and 8V s, a test volume in space described by coordinates (p,x). 29 It has been shown that this electromotive force can be described by103, 104 2 N Y ){2 roo SV s (B1)xy £ = ---~=----e-t/T2 cos rot 4kTs (2.8) where N is the number of nuclei per unit volume, )( is Planck's constant divided by 21t, k is Boltzmann's constant, Ts is the sample temperature, and T2 is the transverse relaxation time. It is this voltage that produces a signal which is processed by the NMR computer to produce spectra or images. Unfortunately the signal is attenuated by noise arising from a variety of sources: Brownian motion of the electrolytes in the patient, Brownian motion of electric dipoles on the patient's skin, and the resistance of the receiving coil amd preamplifier. The former two are beyond the control of the experimentalist, and will not be addressed herein. The latter two are discussed below. Coil Losses Sensitivity and Noise The signal-to-noise ratio, SIN, is proportional to the induced voltage, £, divided by the root mean square (RMS) noise voltage, ns, induced in the coi1.105 The RMS is given by the Nyquist formula106 ns = ...J4kTs M Rs = (2.9) where r is the radius of the test sample, p is its reSistivity, and l\f is the frequency width over which the noise is measured. Rs is the total system resistance described as (2.10) where Rc is the intrinsic loss inherent to the coil, Rd the dielectric losses 30 associated with electric fields, and Rm the inductive losses due to the magnetic fields. In an LC circuit, such as is employed in in vivo NMR, Rs is defined by moL Rs = Os (2.11) where Qs is a measure of circuit resistance known as the "quality factor," and L is the coil inductance. Having expressed the total coil losses as the sum of parallel resistances in equation 10, it is also possible to break the total circuit Q down into the reciprocal sum 1 1 1 1 -Os- -Q-c+ Q-d +O-m (2.12) where the subscripts denote intrinsic, dielectric, and magnetic influences respecti vel y. The resonant frequency for the LC circuit is described as 1 roo = ...jLCs (2.13) where Cs is the total circuit capacitance. It can be readily seen from equations (2.11) and (2.13) respectively, that increasing the system resistance, Rs, by increasing any of its components, will decrease the coil Q, while increasing the total ca paci tance will decrease the Larmor resonance frequency. It has been shown that the sensitivity, 'P, of the coil can be described as 107, 108 'I' = K ~ J(BI)xy N sina dV volume (2.14) where (2.15) 31 where F has been included to account for the noise contributed by the preamplifier of the spectrometer. It can now be seen how the quality factor Q relates to the signal to noise, namely it is proportional to the square root of Q, which in turn is inversely proportional to Rs. Intrinsic Losses Intrinsic losses (Re) are due to the inherent resistance of the LC circuit consisting of the coil (L), tuning (Cl), and matching (C2) capacitors (see Figure 2.2). It can be defined as (2.15) The total circuit capacitance is now the sum of Cl and C2. The resonant frequency for the circuit can be described as 1 COo = -;.y=L(==C=l +==C==2=-) (2.16) A circuit is considered "tuned" when it resonates at the Larmor frequency of the nucleus of interest. It is considered "matched" when the circuit impedance is equal to that of the high-power, pulsed, radiofrequency transmitter. When these conditions are met, there is efficient transfer of Cl ~ 1---0-----0--..... L Figure 2.2. Standard tuning-matching circuitry. L is the coil; Re, the equivalent parallel intrinsic resistance; Cl, the tuning capacitor; and C2 the matching capacitor. 32 power from the rf generator to the probe, and pulse widths are minimized. Furthermore, matching of impedances maximizes transfer of power from the detected NMR signal to the high-gain, low noise preamplifier, assuring high sensitivity and signal-to-noise ratio.109 Apart from purchasing the highest quality capacitors, reducing the coil resistance is the only method to reduce the intrinsic noise of the coil. This can be accomplished by using wire of high conductivity and of large thickness and by properly spacing the wire of multiturn coils.lIO Although less practical, it may also be possible to reduce intrinsic noise by reducing the temperature of the coil. However, the latter has obvious complications associated with temperature maintenance of both coil and sample. The following two sections will deal with losses that can be addressed in the design of a surface coil. Dielectric Losses Dielectric losses stem from power dissipation via the distributive capacitance associated with the turns of the coil. Electric fields extend in a distributed manner from one side of the coil to the other, as is shown in Figure 2.3a. These fields interact with electric dipoles present in the membranes of surface tissues causing power dissipation and a deterioration of coil performance. While this capacitance is difficult to calculate, it may be shown that its value is approximately proportional to the coil diameter, decreases slowly with increasing coil length, and is essentially independent of the number of turns. 11 I The electronic equivalent of the dielectric effects is shown in Figure 2.3b where an additional resistance, Rd, and capacitance, Cd, is added in parallel to the basic circuit depicted in Figure 2.2. a) b) C1 ~~o---~----~----~--~ L 33 Figure 2.3. Schematic representation of dielectric losses. a) The electric fields extend from one side of the coil to the other passing through the sample contributing, b) an additional resistance, ~, and capacitance, Cd, reducing the circuit Q and resonant frequency respectively. Inductive Losses Inductive losses are a result of the interaction of magnetic fields with a conducting sample. The magnetic fields that will produce a signal via mechanisms described previously, arise from a current that is continuously turned on and off. As is the case with any conducting sample, many closed loops can be envisioned within the tissue sample that is to be studied with the surface coil. However, this continuously changing flux will also set up circulating currents within the sample, called eddy currents.1l2 It is these eddy currents that produce heat within the sample and are associated with power loss. Similar to the dielectric losses, the inductive losses add an 34 additional resistance, Ri, to the standard circuitry (Figure 2.4), thereby degrading the circuit Q. Unfortunately, NMR relies on magnetic fields for the excitation and detection of resonances and little can be done to reduce inductive losses. Design Objectives The value of optimizing the loaded coil Q with a balanced tuning circui t thereby minimizing electric fields has been recognized .113 Electric fields present an equivalent resistance given by1D1 (2.17) where t is a combined loss factor of the capacitors and the conducting sample such as the human body. To minimize dielectric losses, the ratio Rd/Rc must also be minimized.101 This ratio is proportional to the ratio Cd / C 1. In other words, dielectric losses can be minimized by ensuring that most of the dielectrical energy is stored in the tuning capacitor rather than the distributive capacitance. Murphy-Boesch described a balanced-matched tuning circuit for an in vivo system that minimized electric fields entering the sample.114 This C1 ~~--~~--~~~~ L R· 1 Figure 2.4. Inductive losses: an additional resistance. The inductive losses are represented by a parallel resistance, Ri, which further degrades the circuit Q. 35 design not only accounted for the losses associated with electric fields extending from one side of the coil to the other, coil-to-coil losses, but also considered the electric fields extending from the coil on, or within, the animal and the probe ground, coil-to-ground losses. The latter are compensated for by the addition of another matching capacitance, CIb, as shown in Figure 2.5. In addition to the balance-match circuitry, it is possible to shield the electric fields from the sample. Hoult and Lauterbur proposed the positioning of a Faraday shield between coil and sample.IOI A Faraday shield consists of of an array of wires that do not form closed loops and therefore do not link flux lines of the magnetic field. Though this can be an efficient method of reducing dielectric losses, it has some problems associated with it. Among others one must consider voltages induced in the shield when frequencies are high enough for the inductive impedance of the link to become comparable to the impedance of the coiL There may also be physical constraints that would not permit the use of a Faraday shield. Hence an alternate means to accomplish similar shielding without Cta ~~--~--~--~~ L ~~--~--~--~~ Ctb Rd,'I Figure 2.5. The balance-matching scheme. CIa and CIb are matching capacitors that compensate for coil-to-coil and coil-to-ground losses respectively. Rd,i is the combined parallel resistance contributed by dielectric and inductive effects. 36 interposing any material between the coil and sample would seem appropriate. As was already mentioned, inductive losses result from a concentration of magnetic fields close to the conductor. Since energy dissipation, and hence Q degradation, goes as a function of B12 a non­uniform field can produce hot spots close to the surface. In the case of a traditional copper wire coil, magnetic fields are highly concentrated close to surface. Most often, however, the area of interest lies below the surface, and, as a result, Q is degraded unnecessarily. The filling factor is also affected by this high concentration of magnetic fields close to the conductor but not in the region of interest. In high resolution NMR, the filling factor is loosely defined as the amount of coil volume that is taken up by sample volume, excluding air space and glass. 1 IS Thus, it is crucial to optimize this factor. Unfortunately, this definition is nonfunctional for surface coils. However, it is possible to extend the definition to surface coils by considering the filling factor as the amount of sample volume that is contained within a uniform Bl field. Because the Bl field of copper wire is not uniform close to the coil, and thus at the surface of the sample, it would be advantageous to employ a material which will produce a more uniform field throughout the sample. Copper foil can be used to optimizes both conductivity of the coil itself and uniformity of the generated Bl field. However, it has been shown that surface coils of conventional design made of copper foil are much more sensitive to loading than those made with copper wire.116 Reduction of Q values was reported to be 72% for coils constructed of copper wire and 86% for coils constructed of foil. The larger attenuation seen in foil coils was attributed to the dielectric losses arising from a larger distributed capacitance 37 due to the larger area of conductor in contact with the sample. Thus in conventional circular surface coil designs either material presents difficulties of some sort. Close conformation to the tissue of interest could partially alleviate these problems. However, coils constructed of wire thick enough to have good Q characteristics, are not flexible enough to permit close conformation to the tissue of interest. In contrast, the flexibility of copper foil allows closer conformation to the sample, thereby optimizing the filling factor and decreasing inductive losses. In summary, when designing a coil one would like to minimize dielectric losses through balancing of the circuit and by shielding electric fields from the sample. minimize inductive losses and achieve high filling factors by promoting uniform magnetic field penetration over the entire sample. maintain coil flexibility to permit close conformation to the subject, which in turn optimizes coil efficiency. Methods and Materials Coil Construction Coils were constructed, as shown in Figure 2.6, from a five layer laminate consisting of copper foil and teflon tape. If the coil was designated for monitoring of 19F nuclei, mylar tape was used. Two foil layers (B and D) were separated by tape (layer C), and insulated and held together on both sides with additional tape (layers A and E). A crossover was accomplished by connecting each half of the top foil layer (B, points x and y) to the complementary half of the bottom foil layer (D, points x' and y'). The connection of point x to point y' was accommodated through a cut (5) in the 38 A B D E Figure 2.6. Layers of the crossover coil. Two foil layers (B and D) are separated by a layer of tape (C). They are then insulated and held together by additional tape layers (A and E). The crossover is accomplished by connecting each half of the top foil layer (B, points x and y) to its complimentary halves of the bottom layer (0, points x' and y'). The crossover connection of point x to point y' is facilitated by a small tab on the bottom copper layer. middle tape layer (C). The connection of point x' to point y was facilitated by a small tab on the bottom layer (D) which folds across the top layer. The assembled coil with crossover is shown in Figure 2.7, where the tape layers have been omitted for clarity. Foil width for the top conductor was generally 1.5-3.0 mm for coils up to 2.0 cm in diameter while the bottom conductor was approximately 20%-30% wider. After assembly, the coil has three terminals, a, b, and c (Figures 2.7 and 2.8). Terminals a and b were connected to the capacitive circuit, and terminal c was connected to the ground plane of the circuit board resulting in an explicitly grounded surface coil. A variable tuning capacitor C2, and two variable balanced-matching capacitors (CIa and Clb) were connected to a PC board in a configuration similar to that of Murphy-Boesch (Figure 2.8).8 a Figure 2.7. The assembled crossover surface coil. Tape layers are omitted for clarity. Terminals a and b are connected to the tuning circuit, and terminal c is connected directly to the ground plane of the circuit board. Efficient dielectric shielding is thereby accomplished by a center tap ground. Actual length of terminals is greater than depicted to facilitate coil placement and to remove the circuit board away from the region of interest. 39 40 a ..... _ .... ..... _ .. b --- C 1a ~ ,~ ~ ~ T i ~ ~ --- II i L I * Figure 2.8. The explicit balance-matching scheme. Components consist of a variable tuning capacitor, C2, two variable matching capacitors, CIa and CIb. A chip capacitor, Cf, is often added to accomplish tuning. It is placed as close to the coil as possible. Capacitors CIa and CI b are equal in value and match the circuit to the balanced end of the 1/4 wavelength triaxial "bazooka" balanced-to-unbalanced transformer (T) constructed from Belden 9222 cable. The unbalanced coaxial end (L) presents an impedance of 50 0 to the spectrometer rf circuitry. 41 To facilitate proper tuning and to maximize power to the coil, a small fixed.chip capacitor (Cf) was added in parallel to the coil. Capacitors Cta and Clb were of equal value and adjusted proportionately to match the circuit to the balanced end of a triaxial "bazooka" 1 17 balanced-to-unbalanced transformer (T). The total length of this bazooka is c length = 0.69 4olo (2.18) where c is the speed of light, COo is the resonant frequency of the nuclei of interest, and 0.69 is a conductivity constant specific to Belden 9222 cable. The unbalanced end presents an impedance of 50 n to the spectrometer. The entire cable arrangement was constructed from a single length of Belden 9222 triaxial cable with the two shields connected together 1 14 wavelength O. = cl coo) from the coil (Figure 2.8). Both shields were connected at the spectrometer's preamplifier. The cable was taped to the magnet with copper tape in order to ground the magnet. Stray rf energy is thus minimized. Quality Factor "Q" Determining the contribution of each resistive loss is a complex issue that relies on both theory and practice. However it is possible to obtain a quantitative estimate of the quality factor Q. Q measurements of unloaded and loaded probes, that is, probes with and without the sample in place, can be used to indicate the degree of dielectric and inductive losses. Measurements were made with the coil connected to an oscilloscope as shown in Figure 2.9. The probe was connected to the sweep generator via a directional coupler. The sweep generator was connected to an oscilloscope that is in the xl y mode. The 42 unloaded coil was then tuned and matched to the frequency of interest (fo)' Q was determined by measuring the frequency width at the half­powerpoints which are calibrated by placing a 3 dB attenuator in the couple­detector connection. Q' is given by Q ' _ fo -M (2.19) where Ai is the frequency width at half power. Q' was then be multiplied by two to obtain the "free ringing" Q of the circuit. The same procedure was then repeated for a loaded coil, i.e., a coil taped to a conducting sample such as a human leg. Alternatively, the probe can be connected to the General Electric 2.0T CSI spectrometer and the Q' determined by measuring the half power width in the "tuning mode" and multiplying the ratio obtained via equation 2.18 with an appropriate scaling factor, m (2.20) where m = 4.044 10-3. Again, unloaded and loaded Q values were determined and the percent degradation was determined. The advantage of either of these procedures is the ability to observe tuning and matching characteristics as well as measure Q-values of the probe in its final configuration both on the bench and in the magnet. Field Profile To illustrate the magnetic field uniformity, imaging techniques for mapping of the Bl field were used that are based on the correspondence between signal strength, or image intensity, and the magnitude of Bl. After tuning the coil to the appropriate frequency, an image was obtained of a phantom containing a concentrated solution of the nuclei of interest using a Directional Coupler Sweep Generator and Detector (~-+"'--ef Out In ... io---.... Probe x ~ ... -M y Oscilloscope 43 Figure 2.9. Measurement of Q using reflective power. The probe is connected to the sweep generator via a directional coupler. The oscilloscope trace will touch the top line when a perfect match exists. Loading the coil broadens and degrades the match, indicated by the shaded line (f' 01 M'). 44 gradient-recalled-acquisition-in-the-steady-state (GRASS) image sequence.11B ,119 The imaging sequence utilizes a single rf pulse that rotates the net magnetization vector by an angle 9<90°. The angle usually used is small and the experiment recycling time short such that the magnetization achieves a steady state)20, 121 The relative signal intensity of each voxel was then related to the field strength within that voxel by122 S -5- = (Bl Bl max) sin(y B1 t) max " (2.21) where S is the signal intensity of any voxel and Smax the maximum signal intensity due to a perfect 90· pulse. B1 is the magnetic field experienced by the nuclei in any voxel and Bl,max the maximum magnetic field experienced by nuclei that are tipped by a 90 0 pulse. y is the gyromagnetic ratio and t the pulse width. A 20 x 18 mm IH coil was placed on the phantom, parallel to the x-z plane of the magnet. GRASS images were obtained from a CUS04 doped H20 phantom. Gradient refocused images are obtained at 1 em intervals at a repetition rate, TR, of 16.0 msec, pulse power of 17%, and a gradient reversal echo time, TE, of 9.0 msec. Total image time was =5.0 sec. Contour maps were then obtained from the relative image intensities using a look-up table generated from equation 2.21. The maximum magnetic field contour is assigned a value of 0.7 and successive contours a value of O.7n x 10 (n=2,3 ... ). Contour plots were then plotted by software supplied with the CSI spectrometer. Coil Performance Results Figure 2.10 shows the Q curves after tuning of a typica1 31p surface coil with inner axial diameters of 18xll mm. The top layer width is 2.5 mm, and 45 the bottom layer width is 4 mm. Figure 2.10a is the Q-curve of the unloaded coil while Figure 2.lOb is the Q-curve of the same coil but now loaded by placing the coil on a human leg. It can be seen from peak analyses displayed next to the curves, that there is slight degradation in height and increase in width. Unloaded Q for this coil was calculated at 174, while the loaded Q was 165 giving an overall circuit Q degradation of =5%. Table 2.1 shows additional Q values for coils of varying dimensions and resonances. Using a look-up table, GRASS images were converted to BI field maps which are shown in figure 2.11. Figure 2.11a shows a contour plot of the field produced by a 23 x 18 mm I H coil in the x-y, or axial plane, while Figure 2.11b illustrates the BI field in the x-z, or sagittal plane. The z- and y­axes represent the distance from the center of the coil to its edges while the x-axes represents the penetration depth of the magnetic field. Discussion Dielectric losses in the crossover surface coil are minimized via the construction of an explicitly balanced circuit which grounds the center of the coil. Explicit grounding reduces sensitivity to exactly equal adjustment of the two matching capacitors. This does not, however, preclude that Q degradation is minimized in a matched coil, a condition neccesary for optimum coil performance. The windings of the coil are arranged with a crossover connection at the apex of the coil (Figure 2.7) such that the portion of the winding which is closest to the ground connection (point c) is next to the surface of the subject. Those points that are driven (a and b), and are thus at an elevated electric potential, are on the side away from the sample. Electric fields produced by the top layer are thus shielded from the tissue of interest by the grounded bottom layer. Optimum shielding of the crossover a) ~ fCN' LOREN HEIGHT 799 463 T/- 7 75821 FREQ -3064 26 +1- 11.3351 WIDTH 1600.97 +1- 21.4366 I POINTS ~ 170 STD.DEV.IN Y u 32.7231 I I I I I I I I I I T-' a -5000 -10000 Hz b) ~ fCN' LOREN HEIGHT 772.757 +1- 6 53758 FREO -3118. 43 ~/- 10 4106 WIDTH 1694.33 +1- 19 ~611 I POINTS ~ 170 STDDEVIN Y ~ 28 3925 I II--r---'--r--r--,--r-r-r-'-'---' a -5000 -10000 Hz Figure 2.10 Q curves of a a 31p surface coil. Dimensions of the coil are: 18 x 11 mm inner axial diameters, 2 mm wide top layer, and 4 mm wide bottom layer. a) Q curve of unloaded coil, b) Q curve of loaded coil. Loading was accomplished by taping the coil to a human leg. Data decribing curve include height, frequency, and half height width. These are used to calculate Q as described in the text. ~ Table 2.1 Unloaded and Loaded Q-values for the Crossover Surface Coil. Coil Dimensions 1 Frequency in mm Nucleus2 (MHz) axb D 32 10 x8 H 85 23x 18 F 80 30x20 F 188 10 x8 P 34 13x 11 P 81 13x 10 103 34 60x4 85 1. a - inside long axis b - inside short axis c - width of top layer d - width of bottom layer c 1.5 4 6 1.5 3 3 5 d 2 5 8 2 4 4 7 Q-emEtl 111 190 225 160 180 105 80 130 Q-loaded 107 120 105 135 165 95 75 130 47 %Q-Ioss ~5 37 53 16 5 10 6 ~5 2. D-deuterium, F-fluorine, H-protons, and ID-inductive trap2 phosphorus­proton dual tuned coils. 3. M. D. Schnall, V. H. Subramanian, J. S. Leigh, Jr., and B. Chance, ]. Magn. Reson. 1985, 65, 122. a) -2 b) -2 o --ev .~ 1 )( Z 2 -1 -1 z-axis (em) o y-axis (em) o 48 1 2 1 2 Figure 2.11 B1 field profiles for the 23 x 18 mm 1H crossover surface coil. Profiles are calculated from the GRASS image intensities of a CUS04 doped water sample. Each contour is assigned a value of O.7T1 x 10 (n=1, 2, 3 ... , where n = 0 is Bl,max). a) B} profile of the x-z, or sagittal plane and b) B} profile of the x-y, or axial plane. The z- and y-axes display the distance from the center of the coil while the x-axis displays the penetration depth. The coil's apex is at approximately -1.15 em in the z-direction. 49 design is reflected in the small Q degradation of coils upon loading as shown in Table 2.1 Effective shielding can be further demonstrated by placing the coil "up-side-down" on human tissue and noting a 1.5 to 3.0 fold larger Q reduction. Also, effective shielding is only slightly affected if the bottom layer is made narrower than the top. For example, Q degradation increased from 5% to 10% when the bottom layer of the 20 x 18 mm 31p coil was reduced from 6 to 3 mm while maintaining the top layer at 4 mm. To illustrate that enhanced shielding of the foil coil is particular to the crossover design, an implicitly grounded two turn IH coil was constructed of 14 G copper wire with identical dimensions as the foil coil used to obtain the images in Fig. 2.11a and b. (L e. 18 x 23 mm). The Q degradation of this coil was approximately 55%, a value significantly larger than the corresponding foil coil (37%). Inductive losses are minimized by the inherent physical properties of copper foil. In contrast to wire, the magnetic fields in the sample induced by a current passing through foil are more uniformly distributed. Thus, high density magnetic fields entering the sample but not the region of interest, as is the case with wire coils, are less pronounced in a coil made of copper foil Unnecessary inductive losses are thereby reduced. This hypothesis is supported by the Bl field profiles shown in Fig. 2.11 where it can be seen that the field is relatively uniform and no extraordinary clustering of field lines close to the conductor is observed. The physical nature of the laminated foil and tape structure provides a thin flexible surface coil that easily conforms to the subject. Close sample conformation maximizes the filling factor by optimizing the distribution of uniform magnetic fields that penetrate into the region of interest. The shielded design prevents dielectric losses, normally observed when the coil 50 is in close contact with conducting tissue, from increasing any further. The field is a uniform dome with a depth of penetration of approximately 0.75 times the coil radius. It can be seen that the sagittal field profile is shorter than the axial field map despite the fact that the coil is longer in the sagittal direction. This shortening probably occurs because magnetic fields at the apex and base of the coil have a significant component parallel to the static magnetic field Bo which does not contribute to the signal. 85 In addition, the field is not concentric with the coil, but slightly shifted towards the coil's apex. This phenomenon is probably due to the fact that some of the current flowing into the coil bypasses the crossover by flowing through the capacitance between the layers setting up a countercurrent in the bottom layer at the coil's base. This reduces the total magnetic field in the base region, shifting it towards the apex. This second effect becomes greater as the ratio of the interturn capacity of the coil to the tuning capacity, C2, increases and it will thus vary with frequency and coil size. These factors are not a serious problem, but are important for proper coil placement when precise localization of the region of interest is required. In conclusion, electric fields are well-shielded from the tissue. Magnetic fields clustered close to the conductor that do not enter the region of interest are minimized, while those that do enter that region are uniformly distributed. Explicit grounding enables easy matching of the coil circuitry, while the use of foil makes the coil flexible permitting close conformation to the tissue of interest, enhancing the coils effectiveness. Thus, even though foil has traditionally been avoided in the construction of surface coils, the shielded crossover design has several advantages over wire coils, and in fact shows improved Q characteristics over a wire coil with identical dimensions. - CHAPTER 3 APPARENT 31p SPIN-LATTICE RELAXATION TIMES AS A FUNCTION OF AGE123 Introduction After pulsing the nuclear magnetic moments with a small rf field, the nuclei are tipped into a perpendicular plane. The nuclei are then allowed to relax back to their equilibrium position. Two modes of relaxation are possible (Figure 3.1). The first relaxation process is the loss of energy to the surroundings (lattice). This process is called spin-lattice, or longitudinal, relaxation and the time constant is called Tt. Tl thus describes the rate of return of the Mz magnetization from its value following excitation to its equilibrium value (Mo). The second is the interaction of neighboring spins such that they gradually get out of phase. This process is called spin-spin, or transverse, relaxation, and the time constant is called T2. Only the phase coherent part of the transverse magnetization can produce a signal. In contrast to Tt, which is an energy effect, T2 is considered an entropy effect. In any dynamic study of metabolic response to a physiological or pharmacological perturbation, it is often required to obtain measurements at short consecutive intervals. Though this is technically possible in NMR, the inherent low sensitivity of the technique requires that one requires a maximum number free induction decays (FID) within the time interval. However, the longitudinal relaxation mechanism dictates how fast one should pulse. If a second pulse is applied before the spins have returned to x Figure 3.1 Relaxation mechanisms. Following a 900 rf pulse, the net magnetization is along the y axis. Vectors fan out in the x-y plane due to spin-spin relaxation and return to the z axis due to spin-lattice relaxation. 52 their equilibrium, the detectable signal produced will be less than that produced by the first. If a series of pulses are applied in this manner one saturates the signal. Comparison of metabolic parameters from different study groups may not be valid if Tl values differ among groups. Some "external" factors that influence Tl relaxation rates are:124 • Paramagnetic ions or metals. • Organic free radicals. • Water content. • Enzyme exchange rates The dependznce of Tl on so many factors has as a result that only "apparent" longitudinal reiaxation rates can be measured in vivo. Indeed measurement of pure longitudinal relaxation rates is virtually impossible. 53 However, the rational behind this experiment is to determine to what extent observed peaks are saturated if pulsed at a certain rate. In light of the structural and biochemical changes that take place in the aging brain (see Chapter 1) one wonders whether 31p relaxation processes are age variant. If such were the case, then adjustments would be necessary to correct for differential partial saturation. Materials and Methods Animal Preparation Male Fischer 344 rats, young (5-6 mo), mature (11-12 mo) and old (24- 25 mol were grown by and obtained from commercial suppliers under the auspices of the National Institute on Aging. Rats were held in the University of Utah animal facility and exposed to standard diurnal/nocturnal cycles. Animals were given food and water ad libitum. Anesthesia was induced by placing the animals in a large jar which was flushed with 4% isoflurane in 100% 02. The animal was tracheotomized using a 16G angiocath to permit artificial respiration. Anesthesia was maintained using 1.5% isoflurane, in 95% 02 and 5% C02, administered via a small animal Harvard respirator. Respiratory stroke volume and rate were adjusted such that arterial oxygen tension, Pa02, carbon dioxide tension, PaC02, and pH, were within normal values (At 5000 ft, normal values for Pa02 PaC02, and pH are 100-130 mmHg, 35-40 mmHg, and 7.35- 7.45 respectively).125 The deep femoral artery and vein were catheterized using Intramedic PE 50 tubing. This permitted sampling for blood gas analysis which was performed at the University of Utah Medical Center Pulmonary Laboratory on a Radiometer ABL2 blood gas analyzer. In addition, arterial catheterization enabled constant monitoring of mean 54 arterial blood pressure, mabp, and pulse via a Hewlet Packard HP 7086 blood pressure monitor. Blood pressure was maintained at 100-120 mmHg throughout the acquisition period without the intervention of any cardiac stimulants. Veinous catheterization was performed to permit replenishment of volume with saline. To avoid heat loss, surgery was performed on a thermostated platform. After completion of surgery, the rat was maintained at 1.0% isoflurane, which was the minimum alveolar concentration (MAC) value as determined by the tail clamp method.126 It was then strapped to a thermostated copper litter within the 7.2 x 46 cm cylindrical aluminum probe. Body temperature was monitored with a YSI rectal probe, and maintained at 35-36 °C via circulation of warm water through the copper litter. A 1.2 x 0.9 cm crossover surface coil was taped to the rat's head and tuned and matched as described in Chapter 2. The entire assembly was then inserted into the vertically oriented magnet. These procedures conformed to guidelines and protocols were approved by the University of Utah institutional animal care and use committee (IACUC). NMR 31 P spectra were collected at 81 MHz on an IBM NR200 (AF series) spectrometer linked to a Cryomagnet Systems (Indianapolis, IN) 90-mm bore, 4.7 T magnet. Field homogeneity was optimized on brain water protons resonating at 200 MHz. Homogeneity is indicated by the half-height peak width of the water resonance. The 31P-tuned coil was adapted to 1H with a small tank circuit outside the magnet. Typical half height peak widths were 80-120 Hz. 55 Apparent Tl measurements were made using a modified pulse-burst saturation recovery technique.127 In this technique, echo formation is eliminated by the saturating pulse train consisting of ten 35 ~ pulses with progressively decreasing spacing (1-\ = 1.0 ms x 1/2n; n = 0, 1, 2, ... , 10). The pulse train is designed to produce zero z-magnetization. Following saturation, magnetization relaxes to equilibrium during time 'to The acquisition program was cycled through 10 't values ranging from 0.1 to 20 s. A sampling pulse then monitors the degree of recovery. Though faster than the inversion recovery method, the method requires accurate calibration of the 90° pulse width, which was accomplished with 1 M Na2HP04 phantoms. Pulse width for maximum mid brain excitation was 25 Ils. The program was cycled four times giving 128 scans per 't value. Spectral width was set at 10 kHz. Gaussian line broadening of 30 Hz was applied to the FlD prior to Fourier transformation to enhance signal-to-noise. 1[12 1[12 Figure 3.2 The modified pulse-burst saturation recovery sequence. 90° pulse widths, optimized on aIM Na2HP04 phantom were 25 Ils. 1-\ = 1.0 ms x 1/2n; n = 0, 1, 2, ... , 10. Saturation pulse widths were 35 Ils. The pulse sequence is cycled through 10 recovery times, 't, ranging from 0.1-20 s. 56 Peak intensities of soluble phosphates were measured above the broad component baseline and fitted to the equation I{t) = 10 - loe(-t/Tl) (3.n where I{t) is the peak intensity at time t and 10 is the equilibrium (fully relaxed) intensity. Plotting In[Io-l{t)] vs t presents a straight line with slope equal to 1 /T 1. 10 for long relaxing species was estimated by plotting arms line through the data points and extrapolating to a plateau value. Saturation factors were calculated for experiments with a repetition rate smaller than measured T1 values. Thus, for a two second repetition rate, the saturation factor was calculated from the T1 data by taking the ratio of l{oo)/I{2). The saturation factor is then equal to . 1 saturation factor = I-e-2/T1 (3.2) These saturation factor are used in all further comparisons of metabolite peak areas. One factor affecting spin-lattice relaxation rates is binding of paramagnetic or diamagnetic metals to ATP. Mg2+ is such a metal and has additional importance as a cofactor in all reactions involving ATP synthesis or usage. The relative amount of ATP-Mg{II) complexing among age groups, was evaluated by measuring the chemical shifts of ,,(-, a-, and ~-ATP relative to PCr.128, 129 A net increase in Mg-ATP complexing results in a down field shift of the ~-, and -y-ATP resonances. All data were analyzed for statistical significance via planned comparisons,130 on a Macintosh SE using Microsoft Excel. This method permits directed analysis of variance and tests the following null hypotheses: flO: Jlyoung -- Jlmature = 0 Ho: !lyoung - ~Id = 0 Ho: !lmature - !lold = 0 57 The null hypothesis was rejected when F(1,6»6.94, p=0.05, implying a statistically significant difference between group means. Results Figure 3.3 shows a typical inversion recovery experiment. It can be seen that the pulse comb resulted in complete saturation, reflected in no detectable signal in the bottom spectrum. With increasing t values, soluble metabolites relax to their equilibrium position and maximum intensity for most soluble phosphate resonances is observed at t = 20 sec. Measurement of peak areas and processing of the data via equation 3.1 gave Tl values shown in Table 3.1. Average values with standard sample deviation are shown for the three age groups. Statistical deviation from the young group is indicated by a *, while statistical deviation from the mature group is indicated by a 00. These data are plotted in Figure 3.4, giving a clear illustration of T 1 trends as a function of age. Though most Tl values do not vary significantly with age, the relaxation time of the a-phosphate decreases between young and mature rats. A further decrease is observed between the mature and old age group. The reverse is seen for the relaxation time of the ~-phosphate. Though no significant change is observed between young and mature animals for this metabolite, a significant increase in Tl was found for the old age group. No statistically significant changes in Tl values were observed for per and Pi. However, the large difference between young and mature Tl values for Pi does stand out and further investigation seems warranted. A significant increase between young and old animals was observed 58 pcr t = 20.0 s t = 14.0 s t = 12.0 s t = 10.0 s t= 6.0 s t= 4.0 s t= 2.0 s t= 1.0 s t= 0.5 s t= 0.1 s Figure 3.3 A typical inversion recovery experiment. Shown are spectra obtained at varying t values. The bottom spectrum was obtained at t = 0.1 s indicating complete saturation of the soluble phosphates. Each spectrum represents 128 scans. A Lorentzian function was applied to the Free induction decay (FlO) to enhance resolution. ~-ATP a-ATP y-ATP PCr Pi Table 3.1 Apparent Tl (s) of Rat Brain Phosphates as a Function of Age. younga maturea (5-6 mo) (10-11 mo) 1.01 ± 0.21 1.07 ± 0.09 1.85 ± 0.25 1.10 ± 0.26* 1.68± 0.25 1.73 ± 0.13 3.02± 0.35 3.05 ± 0.49 2.01 ± 0.79 3.23 ± 1.14 Broad component 14.61 ±3.05 20.54 ± 4.84 a n =4 old a (24-25 mo) 1.46 ± 0.08*,"0 0.98 ± 0.15* 1.51 ± 0.19 2.71 ± 0.32 3.00 ± 0.53 23.02 ± 2.58* * variance from the young value is statistically significant at p = 0.05 00 variance from the mature value is statistically significant at p = 0.05 59 60 Bone 30 Pi PCr "tATP III young (l-ATP II mature • old f>-ATP 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Seconds Figure 3.4 Graphical representation of apparent Tl values. The data presented in Table 3.1 are shown as bar graphs illustrating the trends of apparent Tl values as a function of age. 61 for the broad phosphate hump, which is mostly composed of calcium phosphate in bone. Though the difference between young and mature rats is not statistically significant, the data point does fit the trend. Saturation factors calculated from the Tl data are tabulated in Table 3.2. Though Tl values are less than 2 sec for the three phosphate groups of ATP, a saturation factor is still required because complete relaxation requires an interpulse delay of ~5Tl' Calculated values were used in all studies where metabolite peak areas were compared A verage chemical shifts of the pertinent phosphates are shown in Table 3.3. The data were collected in a separate study from six Fischer 344 rats per age group which were prepared analogous to the Tl study. Standard deviations are reported and planned comparisons showed no significant differences between any of the age groups. Discussion The above results clearly indicate a variable Tl as a function of age. In light of the many physiological and chemical changes that take place in the aging brain this is not too surprising. However, to assign anyone mechanism to the Tl changes would be simple minded. As already mentioned, many factors contribute to the spin lattice relaxation rate. One of the first mechanisms that could be invoked is the well­established accumulation of paramagnetic metals in older rats.131 Thus copper (Cu) and iron (Fe) increase 85% and 101 % respectively between the ages of 4 and 25 months. In contrast, manganese (Mn2+) levels decrease by 37%. One would expect, on the basis of these data, a shortening of 31p Tl'S of those phosphates that are capable of metal chelation. Binding of metals such as Mn2+ and Cu2+ to ATP, for example, has been shown to result in a ~-ATP (l-ATP y-ATP PCr Pi a n=6 ~-ATP (l-ATP y-ATP a n=6 Table 3.2 Saturation Factors for Phosphate Metabolites as a Function of Age. younga rnaturea (5-6 rno) (10-11 rno) 1.16 1.18 1.51 1.19 1.44 1.46 2.06 2.08 1.59 2.16 Table 3.3 Chemical Shifts of ATP Phosphates as a Function of Age. younga rnaturea (5-6 rno) (10-11 rno) -2.51 ± 0.12 -2.42 ± 0.01 -7.56± 0.03 -7.54± 0.02 -16.14 ± 0.05 -16.10 ± 0.01 62 olda (24-25 rno) 1.34 1.15 1.36 1.92 2.06 olda (24-25 rno) -2.44± 0.03 -7.60± 0.02 -16.10 ± 0.06 63 reduction of the T1 of primarily ~-, and y-phosphates.132 Similar results can be expected from any other paramagnetic metal chelation to ATP. However, neither ~- or y-ATP T1 values decrease significantly. In fact, the ~­resonance displays a significant increase. Opposite trends imply that the accumulation of paramagnetic metals does not appear to have a significant effect on T1'S of ATP. Diamagnetic metals, in addition to increasing relaxation rates, cause a chemical shift change for the ex and ~ resonances. It has been well-documented in vitro that binding of diamagnetic metals, such as Mg2+ to the ~- and y-phosphates results in an upfield shift of approximately 1-2 ppm.128, 133 Recently similar studies were performed in vivo.134 A significant down field shift was observed in brain tissue of Mg2+ depleted animals which the authors contributed to an altered free magnesium status. Increased Mg2+ concentrations do not appear to be a factor based on a lack of chemical shift change as is shown in Table 3.3. This is not completely unreasonable if one assumes no differential, and complete saturation of unbound ATP with Mg2+ in all age groups. A second mechanism that may be invoked is the increase in free radical concentration which has given rise to the free-radical theory of aging. 16, 17 The build up of free radicals appears to correlate with increased DNA damage, and hence cellular damage. In addition, free radicals lead to increased relaxation rates via similar mechanisms as those involved with paramagnetic relaxation. Namely, by providing a large, electron magnetic moment, local fields act as an alternative source of changing magnetic fields for molecules under observation. However, it is not likely that only a few specific species would be affected by this mechanism. Furthermore, a free radical usually carries a negative charge and would be repelled by the negatively charged phosphate groups. 64 The decrease in water concentration mentioned in Chapter 1 could be a third mechanism giving rise to altered Tl rates. The degree of hydrogen bonding to the hydration sphere surrounding molecules of interest probably increases when the total water content is decreased. This paradoxical effect is due to a concentration of water molecules about the anion. This in turn results in an increased mean correlation time, te, the average time for a molecule to progress through one radian. When te< 1 / (001 the decreased mobility results in an increased Tl. Again, one would not expect that the individual phosphates in ATP would be affected differentially. Thus, though these mechanisms are a significant factor in proton imaging,135 it is not a likely candidate to explain the Tl differences for ATP seen in this study. However, the steady increase in Tl for the broad component may be related decreased motion. Thickening of the skull is probably accompanied by an increase in the rigidity of calcium phosphate. In this semisolid matrix, the motional correlation times must be much longer than 1/(00 and further increases would decrease the rate of longitudinal relaxation, in other words relaxation times increase. A similar argument can be made for components of the broad hump, such as phospholipids and phosphorylated proteins, and the rigidification of biomembranes.49 Perhaps the most important mechanism by which apparent Tl values are affected is the influence of enzymatic exchange kinetics. The reactions that come to mind are those catalyzed by creatine kinase (1.1), adenylate kinase (1.2), and ATP phosphatase (3.1). ADP + Pi f2 ATP (3.1) The spin-lattice relaxation time in the presence of exchange, Tl,obs can be expressed asl36 _1 ___ 1_+ f Tlobs - TIA TIB + 'tB 65 (3.2) where TIA and TIB are the inherent TI values in the absence of exchangel of AI the more abundant speciesl and B. f is the ratio of B to AI and 'tB is the lifetime of the B species. In fast exchange 'tB is much smaller than TIB and therefore (3.3) Altered enzyme kinetics as a function of age are well-established in vitro. Though basic metabolic levels do not appear to be affected with agel it need not be a given that the equilibrium position of the enzymes of interest are constant with age. A change in equilibrium will affect the ratio of A to B and hence have a direct effect on the observed spin-lattice relaxation rate. In vivo studies have not established which equilibria are affected by age. Furthermorel due to the complexity of the energetic systeml determining which equilibria are most influential in the spin lattice relaxation rates of the ATP phosphates is almost impossible. The increase in TI for Pi between young and mature rats may also be attributable to a change in enzyme kinetics. ATPase is the most probable candidate exchanging ADP and Pi for ATP. Though no in vivo literature is available for this enzyme with respect to the brainl in vitro studies show that the equilibrium lies far to the right. A further shift to the right would probably cause an increase in Tl for Pi via a decrease in f in equation 3.3. This is admittedly conjecture and somewhat unlikely as it implies an increased capacity for mitochondrial respiration in the older rat. The following chapter will address this issue further. CHAPTER 4 CEREBRAL PHOSPHATE METABOLISM DURING HYPOXIA AS A FUNCTION OF AGE137 Introduction The proper functioning of brain cells relies on an abundant and continuous supply of oxygen. Even relatively moderate amounts of deprivation can lead to neuronal damage. Several factors lead to the extreme vulnerability of brain cells to hypoxic conditions. Relative to muscle tissue, neuronal elements have a higher metabolic rate, lower capillary density, essentially no oxygen stores, and meager reserves of high energy phosphates and carbohydrate substrates. Indeed, it may be said that a high degree of specialization has required the sacrifice of ruggedness seen in many other tissues. Thus while muscle tissue is capable of adequate aerobic and anaerobic metabolism, the brain is mostly dependent on aerobic metabolism. In light of the many structural changes with aging described in the introduction it would not be surprising if metabolic changes occur as well. As described, early studies on isolated rat liver mitochondria indeed indicate a reduced functional capacity,54, 55 In contrast, rat brain mitochondria had a similar capacity to carry out oxidative metabolism for both 6 month and 24 month animals.138 More recent in vivo PET studies do not indicate a decreased oxygen metabolism as a function of age in humans. However these studies were performed in resting, unstressed 67 patients.72 The metabolic changes of the normally aging brain are primarily functional and differences in oxidative metabolism become only apparent when metabolic requirements are increased.139,140 The effects of ischemia on glucose and energy metabolism as a function of age were studied in frozen brain extracts.137 Via enzymatic analysis of glycolytic and high energy metabolite concentration, an increased glycolytic flux in younger animals was noted. Succinate was found to accumulate, perhaps due to secondary reactions, and ATP appeared to be less utilized in the aged animals. Similar results were obtained after submitting rats to severe hypoxemia (Pa02=20-25 mmHg) and analyzing frozen brain tissue enzymatically. 141 However, no changes in high energy phosphate levels were observed in the latter study. 31 P NMR was employed to investigate metabolite levels in brain tissue following a 55 s period of complete anoxia, which resulted in loss of consciousness.142 After freezing, removal, and perchloric acid extraction of brain tissue, 31 P spectra were obtained. It was found that PCr and A TP drops with a concomitant rise in Pi was greater in the older rats than in the younger animals. However, no statistics were given. In a later study, synaptosomes isolated from the forebrain of rats of different ages were submitted to different degrees of hypoxia (5~Pa02~11 mmHg). The potential damage induced by the hypoxic insult which synaptosomes cannot reverse was then evaluated.143 This was accomplished by establishing the free energy difference (tltlG) between the oxidation-reduction reactions of the electron transfer chain for the two first phosphorylation sites (tlGox-red) and the phosphorylation state of the adenine nucleotide system (tlGATP). It was found that under extreme hypoxia (Pa02<35 mmHg), tltlG decreased significantly with increasing age, 68 indicating an impairment in energy transfer from the reduction of the redox pairs to the synthesis of ATP. The llGATP seemed mostly affected in this case. However, for synaptosomes submitted to moderate hypoxia (Pa02>35 mmHg), no significant changes were observed in any of the free energy parameters. Though the above studies contained in vivo components, in the sense that the hypoxic insult was accomplished in whole animals, the actual analysis was accomplished in vitro. The extent of tissue damage and phosphate hydrolysis is not known exactly, but freezing brain tissue is not an instantaneous process no matter what method is employed.144 Therefore it may not be warranted to assume that per and ATP levels remain stable after the anoxic or hypoxic period. In vivo studies on the affects of aging on the response to hypoxia are less abundant. Following exposure for four days at 10% 02 Wistar rats of 3, 20, and 27 mo did not display a significant change in spontaneous activity, food and fluid intake, or blood gas analyses.145 A follow-up study examined the learning capacity of rats following an identical regimen.146 It was found that though hypoxia reduced the learning skills of all age groups, as compared to normoxic control groups, no Significant difference was seen between young and old rats. It would seem obvious that the results obtained by in vitro analyses do not necessarily correlate with patterns obtained in vivo. Furthermore, in vitro studies are not entirely consistent and are extremely susceptible to procedural error. Thus the in vivo study of hypoxic brain metabolism as a function of age seems a warranted and necessary objective to obtain a more complete understanding of aged brain metabolism under stress. 69 Materials and Methods Male Fischer 344 rats of identical age groups as described in the previous chapter (5-6 mo; 10-12 mo; 24-25 mo: n=6), were grown under the auspices of NIA at the Harlan Laboratories. As before, rats were held in the University of Utah animal facility, given food and water ad libitum, and exposed to standard diurnal/nocturnal cycles. Following initial induction of anesthesia, each animal was tracheotomized and artificially ventilated via a Harvard rodent respirator. Anesthesia was maintained with 1-1.25% isoflurane in 36% 02, 4% C02 and 60% N2 during surgical preparation. Respiratory rate and volume were adjusted according to weight and age such that blood gas analyses were within normal range. Intramedic PE-50 tubing was used to catheterize the deep femoral vessels. One arterial1ine was introduced to allow for monitoring of blood pressure and pulse, and withdrawal of 100 - 150 III blood samples used for blood gas and blood glucose analysis. Blood sugar levels were monitored with Chemstrip® (Boehringer-Mannheim) strips. In addition, deep femoral veins in each leg were catheterized. One allowed for administration of pharmacological agents such as pancuronium bromide (Pavulon®) to prevent hyperventilation and return of volume following blood withdrawal to avoid hypovolemia. The second veinous line was inserted to accommodate slow epinepherine infusion (0.05 mg/ml @ 2 ml/hr) such that blood pressure could be maintained during the hypoxic period. After administration of Pavulon (initial bolus of 0.3 mg/kg followed by 0.05 mg/kg every half hour), the anesthetic was reduced to 0.75% ± 0.05. Body temperature was monitored via a rectal probe and maintained at 36-37 °C as before. 70 After preparation, the rat was secured to a copper block within the homebuilt aluminum probe. After proper alignment in reference to the rat's brain, the two turn copper coil (connected to an optimally tuned balance-match circuitry)100 was taped to the skull. The probe was then inserted into the NMR spectrometer. Field homogeneity was optimized by shimming on the brain water proton signal which resonates at 200 MHz. Shimming was stopped when a line width of 30-40 Hz was obtained. Following 12.5 minutes of normoxia, hypoxia was induced by decreasing the inspired 02 from 36% to 18%, then to 13% after an additional 17.5 minutes (%N2 of the inspired air mixture was increased accordingly). These procedures conformed to guidelines and protocols were approved by the University of Utah institutional animal care and use committee (IACUC). 31p spectra were obtained at 81 MHz. Using a pulse width of 25 J,.Lsec (under our conditions, the optimum 90° pulse for 31 P excitation in the cen­tral brain region, approximately 3 millimeters from the coil) and an interpulse delay of 2 seconds, 74 transients were accumulated corresponding to a 2.5 minute turnover. The consecutive time domain summations were collected and stored. Lorentzian and Gaussian linebroadening were applied for maximum resolution and signal-to-noise ratio (LB=30 Hz; GB=O.08). The individual soluble, unbound, phosphate metabolites are assigned on the basis of chemical shift and are readily seen on top of the broad hump. Individual metabolite areas were measured above the bone hump via triangulation (Area = 1/2(b x h» and then multipilied by saturation factors determined from 31p spin lattice relaxation times (T 1) to correct for differential saturation effects.123 Differential changes in Tl values due to hypoxia were assumed to be constant within each age group. 71 Contribution of scalp to the broad bone hump was investigated by collecting spectra when the scalp was removed and comparing this to spectra acquired with the coil placed directly on the bone. No difference was noted in spectral properties. Intracellular pH (pHi) was calculated from the chemical shift difference (0) of inorganic phosphate (Pi) with respect to phosphocreatine (PCr) by the following relationship.147 [ 0 - 3.29] pHi = 6.77 + log10 5.68 _ 0 (4.1) This equation is based on the inorganic orthophosphate titration curve of a phantom simulating in vivo conditions. The value 6.77 is the pKa of the phosphate acid-base pair in this particular medium. An average "control" area of the first five blocks was computed for each metabolite and then divided into the ensuing "hypoxic" areas. Ratios were averaged for each age group and plotted against time. ADP ratios were calculated by combining ATP, PCr, and intracellular pH (pHi) data, a manipulation of metabolite concentrations based on the creatine kinase equilibrium (1.1).148 From the equilibrium equation one obtains an expression for [H + ]: [A1P] [Cr] 1 [H+] = [ADP] x [OCr] x Keq Metabolite ratios during hypoxia are then compared using the ratio [H+]o [A1P]o [ADP] [PCr] [Cr]o [H+] = [A1P] x [ADP]o x [POlo x [Cr] (4.2) (4.3) where the equilibrium constant was assumed stable during hypoxia and the "0" subscript indicates control areas. Rearranging terms, and assuming that the total creatine pool (PCr + Cr) remains constant,149 an expression for ADP was obtained where -flo+ x A-T-P ADP H+ ATPo ADP = o T} Per 1/Rx Per o T} = Per 1 + 1/R- Per o and R = [PCr]o/[Cr]o is equal to 0.91 for healthy brain tissue.149 72 (4.3) (4.4) It is well known that one of the factors that governs the rate of mitochondrial respiration is the cytosolic phosphorylation state,150 defined as [ATP]/[ADP][Pj]. The change in phosphorylation state (;/;0) during the stress period was estimated by calculating the ratio of the potential during hypoxia over the potential during control. Thus Rearranging terms, ; [ATP] ([ATP]o )-1 ~ = [ADP][Pj] x [ADP]o[Pj]o ; [A TP] [ADP]o [Pi]o ~ = [ATP]o x [ADP] x [Pi] (4.5) (4.6) Statistical analysis was performed on a Macintosh ™ SE using Microsoft® Excel. Sample averages and standard deviations were calculated for each age group. Planned comparisons were performed analogous to the Tl experiments for each metabolite at each time point. The following null hypotheses were tested: Ho: Jlyoung - Jlmature = 0 Ho: Jlyoung - J.1old = 0 Ho: Jlmature - J.1old = 0 The null hypothesis was rejected when F(1,10) > 4.96, p = 0.05. 73 Results Typical 31p brain spectra before, during, and after hypoxia for different ages are shown in Figure 4.1. The soluble phosphates are clearly seen above the broad immobile phosphate hump. As would be expected, a drop in PCr, concomitant with a rise in Pi, is seen in all age groups after the hypoxic period. Recovery of metabolites appears adequate upon reinstitution of normoxia for this particular set of data, but see below for averaged data. Unexpected is the more severe drop in PCr, and thus greater rise in Pi, observed in the young age group (left trio of spectra). Initial inspection of Figure 4.1 shows minimal difference between mature (middle trio) and old adults (right trio). Relative metabolite ratio's were calculated from control spectra of each age group following correction of the areas by appropriate saturation factors. The results are shown in Table 4.1. The increase in PCr/PME ratio between young and mature rats was quite large yet statistically non­significant. This ratio decreased, significantly, between mature and old animals. A similar trend was seen for PCr /Pi ratio: an initial increase followed by a decrease, but differences were statistically nonsignificant. The trend was again repeated for the PCr/~-ATP ratio: an initial nonsignificant increase between young and mature adults was followed by a significant decrease between mature and old animals. Sample means, standard deviation and significance of physiologic and metabolic parameters, including Pa02, pHa, and pHi are tabulated in the Appendix. Statistically significant differences are as indicated It can be seen that the standard deviation was smaller for the mature adult and senescent rats than for the young rats. Furthermore, the sample error increased during the recovery period. Reasons for this are discussed below. ~ I'C, ~~l ATP H",J[(L"",~",,,,,,, '''''~'''''''''''/~''.'i~''''''\~fI'' )w~l . ~ '>J~~~ 400 20.0 0.0 -20.0 -400 -«lD PPM MATURE I'C, ~~I AT,. 7~~ ,.J./f' lA "{f",,~t~"""-- - ~,o#'''(':,V\.,It.J.r::,1r.'' )l~ "'l<t..~~'/' I \t"\ . "\,"'~ ... (\fl"f~.'(">I" .~~\N' I 40.0 20.0 0.0 -20.0 -40.0 -«lD PPM QLD fICo' ~ . )~.~ ... I \l •. WV\~r""" "'''''-V14:''''I'-o.l:i-\ ... B. }~~ l._JiJ'I\l .. ~~... /l,"'t~~ A. r '~.o'~' riD '_~'~'...oo PPM RfCO\IERY HYPOXIA NORMOXIA Figure 4.1 31p Brain spectra before, during and after hypoxia of young, mature and old rats. Each spectrum was acquired at 81 Mhz. 74 transients were collected using a 2S Jls and a 2 s interpulse delay. Sweep width was 10 kHz. Gaussian and Lorentzian functions were applied to the FlD prior to Fourier transformation to enhance resolution. From bottom to top are shown control spectra, following 17.5 min of 13% inspired 02, and 20 minutes following recovery. ~ 75 Table 4.1 Metabolite Ratios as a Function of Age younga maturea olda (5-6 mo) (10-11 mo) (24-25 mo) PCr/PME 3.59 ± 0.52 4.64 ± 1.37 3.30 ± 0.4900 PCr/Pi 4.79 ± 0.77 5.58 ±0.66 4.66 ± 1.23 PCr/p-ATP 3.26 ± 0.46 3.72±0.38 3.10 ± 0.3400 a n=4 00 variance from the mature value is statistically significant at p = 0.05 76 All data are shown graphically in Figures 4.2 through 4.4. The periods of decreased percent inspired 02 is indicated in each figure by the small vertical lines. The time of recovery, when inspired 02 was returned to 36%, is indicated by the "closed" arrow in medium and old age groups, and an open arrow in the young age group. Figure 4.2 a and b show the arterial 02 tension in mmHg (Pa02) and pH (=extracellular pH, pHa) as a function of time. It can be seen from Figure 4.2a that mature and old rats do not differ significantly in initial blood 02 levels, while the young rat had slightly higher Pa02 levels. However, both the descent into, as well as the final state of hypoxia was identical for all age groups. The arterial pH did not differ significantly between young and old animals. Both decreased in an identical fashion. Though not significant, the mature rat seemed more able to maintain pHa throughout hypoxia. N one of the rats showed complete recovery of arterial pH after the hypoxic insult. Figure 4.3 a-f depicts the metabolic ratios of pHi, PCr, Pi, PME, A TP, and ADP obtained from the spectral data. Figure 4.3a shows that, in contrast to the nonsignificant arterial pH drop, the intracellular pH, (pHi), did show a significantly greater drop in the young adults. Recovery in the young adul t is marginal, while the other age groups faired well after hypoxic exposure. Concomitant with the larger drop in intracellular pH, there is a larger drop in PCr and a greater rise in Pi (Figures 4.3b and 4.3c respectively) in the young animals, both are statistically significant. Two rats displayed a reversal of recovery in the last 10 minutes of return to normoxia. This is reflected in the rise in PCr and Pi. The marginal recovery of pHi also reflects this reversal of recovery. The other four young rats showed a continuation of the plateau levels achieved after 15 minutes of recovery. 77 a) 160 140 120 100 Pa02 80 60 40 20 II!! Young • Mature b) 7.6 • Old 7.5 7.4 7.3 pHa 7.2 7.1 7.0 6.9 36% 18% 13% 36% ~ Inspired 02 6.8 0 10 20 30 40 SO 60 70 80 time (min) Figure 4.2 Arterial oxygen tension, Pa02, and pH, pHa, as a function of time. Values are shown during normoxia (38% inspired oxygen), hypoxia (18 and 13%) and recovery. a) Pa02, b) pHa. Open arrows represent onset of recovery for young animals, while closed arrows represent recovery for mature and old rats. Blood analysis was performed on 100 JlI of blood on a Radiometer ABL blood gas analyzer. a) pHi b) PCr PCro 7.3 7.2 7.1 7.0 6.9 6.8 6.7 6.6 IS Young • Mature 1.1 • Old 1.0 0.9 0.8 0.7 36% 18% 13% 36% -E- Inspired 02 0.6 +--.-...-..,...-+......-......... ........-I-...-'r""T......-I-r..,.....,r-f-..,....,.-r-,....,.."'T"""'f'"" o ro w ~ ~ ~ ~ m ~ time (min) 78 Figure 4.3 Intracellular pH, pHi, and metabolic ratios as a function of time. a) pHi, b) PCr IPCro, c) Pi/Pio' d) PME/PMEo, e) ATP I ATPo1 f) ADP I ADP Q. pHi was calculated using the following relationship: pHi = 6.77 + log1 0[0- 3.29~ J I where 0 I.S the chem.Ical .ShIf.t dIf ference between Pi and PCr. 5.68 - u 79 c) 2.5 2.0 1.5 1.0 0.5 I!'I Young • Mature d) 1.5 • Old 1.4 1.3 1.2 PME 1.1 PMEo 1.0 0.9 0.8 18% 36% -E:- Inspired 02 0.7 0 10 20 30 40 50 60 70 80 time (min) Figure 4.3 continued. Ratio's are calculated for each metabolite by dividing the average of five control areas into the the respective hypoxic areas. Measured areas have been corrected for partial saturation. Points depict every other spectrum. Values represent the mean for each age group, n=6. Arrows represent onset of recovery as in Figure 4.2. e) 1.1 ATP ATPo f) 1.0 0.9 0.8 1.6 1.4 1.2 ADP 1.0 ADPo 0.8 0.6 0.4 36% 18% 0 10 20 80 III Young • Mature • Old 36% ~ Inspired 02 30 40 SO 60 70 80 lime (min) Figure 4.3 continued ADP / ADPo ratio's are calculated as described in the text using equations 4.3 and 4.4 81 Figure 4.4 The phosphorylation potential as a function of time. Ratios are calculated from equation 4.6 All other symbols are analogous to Figure 4.3. 82 Figure 4.3d shows the phosphomonoester, PME/PMEo, ratios through recovery. All age groups show an statistically identical rise in PME into hypoxia which falls back to control levels once inspired oxygen levels are returned to normal. Figures 4.3e illustrate ATP (measured via the ~­resonance) levels during hypoxia. Though ATP levels fluctuated throughout hypoxia, all age groups appeared to be able to maintain ATP levels at control levels at statistically identical levels. While the standard error for ADP levels is large, qualitative trends are apparent in Figure 4.3f. The two older groups displayed a slight drop followed by recovery, while the levels in the young rat seemed to fluctuate much more. An initial decrease is followed by an increase into late hypoxia in this age group. During early recovery the levels again seem to dip before returning to control levels. From Figure 4.4, it can be seen that the phosphorylation state ratio follows the trend expected from the rise in Pi. The decrease is significantly different for the young age group and was not observed for either mature or old animals. Though the later two age groups did exhibit a slight decrease in phosphorylation state, they were not significantly different from control values. Discussion In response to hypoxic hypoxia at least three mechanisms are relied upon to alleviate the reduced oxygen supply which leads to reduced mitochondrial respiration and reduced ATP availability. Firstly mitochondrial respiration itself and anaerobic glycolysis may be stimulated to produce more ATP. In other words, enzyme activities in the respective energetic stages may be enhanced in response to reduced ATP availability. 83 Secondly, ATP can be resynthesized by phosphoryl transfer to ADP from PCr by the creatine kinase (equation 1.1). Thirdly, adenylate kinase produces ATP by phosphoryl transfer to ADP from ADP (equation 1.2). It should be noted that MgATP is the true substrate for both creatine kinase and adenylate kinase and the dependence of energy state on magnesium concentration must be appreciated.151 However, in an in vivo setting, the concentration of magnesium within the cell may be assumed to be relatively constant in view of the fact that the high concentration of binding sites ensures that Mg2+ is buffered within the cell. That magnesium activities do not appear to change as a function of age was demonstrated in Chapter 3 by a lack of changes in ATP 3Ip chemical shifts. Whether or not magnesium concentrations are constant during hypoxia is not known. However, the degree of hypoxia imposed on the animals in these experiments was not severe enough that one might suspect structural damage, and the cell's capacity to maintain normal magnesium concentration may therefore be expected to be unaffected.152 As mentioned in Methods and Materials, peak areas were measured by triangulation above the broad bone hump. It was decided that this seemingly archaic method of peak integration was still superior and more consistent than computer aided integration. The reason for this was found to be due to the inability to remove the broad hump effectively and obtain a smooth baseline without affecting peak areas. A variety of techniques do exist to remove this broad signal. These include postacquisition subtraction and difference methods, selective presaturation,153 and spin echo pulse sequences.154 It was found that postacquisition techniques resulted in dis­tortion of the spectra, which affected peak shape, and thus peak area. Presaturation requires a second 31p channel, which is not available in this 84 laboratory. Spin echo sequences, and other pulse delay sequences, were attempted but the with rapid spectral tum-over required in a dynamic study, signal-to-noise ratios became inadequate. The ultimate solution to the problem would be localization employing gradient coils.1SS However, this requires the availability of gradient coils, as well as longer acquisition times. Absolute quantitation of metabolite levels in in vivo NMR is rather complicated and somewhat controversia1.1S6 Concentration measurements are extremely sensitive to coil configuration and loading. Use of external o