|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 NONINVASIVE CHARACTERIZATION OF SKELETAL MUSCLE DAMAGE AND REPAIR IN MURINE MODELS OF MUSCULAR By NATHAN DAVID BRYANT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010
2 2010 Nathan David Bryant
3 To my friends and family who carried me through the most difficult times and to my m entors that showed me the path. I could not have done this without you.
4 ACKNOWLEDGMENTS I would like to thank the NIH T 32 Training program in Neuromuscular Plasticity in Rehabilitation (T32H D043730) for giving me the opportunity to work in this fiel d, providing guidance in career planning and introducing me to the world of clinical research and practice. I would also like to acknowledge Dr. Elizabeth McNally for providing the sarcoglycan/ knockout mice used in this study and Dr. Philip Miles for his invaluable mentoring in the areas of scientific writing and presentation. Funding for this research was provided by the Muscular Dystrophy Association (MDA) and the National Institutes of Health (NIH) R01AR47292T, R01HL78670, and T32H D043730.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 9 LIST OF FIGURES ........................................................................................................ 10 LIST OF ABBREVIATIONS ........................................................................................... 13 ABSTRACT ................................................................................................................... 15 CHAPTER 1 INTRODUCTION AND BACKGROUND ................................................................. 17 Skeletal Muscle ....................................................................................................... 18 Growth and Repair of Skeletal Muscle ............................................................. 19 Structural Or ganization Within the Muscle Fiber .............................................. 21 Dystrophin and the Sarcolemma ...................................................................... 23 Muscular Dystrophy ................................................................................................ 23 Clinical Impact and Symptoms ......................................................................... 23 Pathology ......................................................................................................... 26 Laboratory Models of Muscular Dystrophy ....................................................... 27 Animal models of dystrophinopathies. ....................................................... 27 Animal models of sarcanopathies. ............................................................. 29 Approaches to treatment of muscular dystrophy. ....................................... 30 Monitoring disease progression. ................................................................ 34 Magnetic Resonance Imaging (MRI) ...................................................................... 35 Basic Components Required for MR or NMR. .................................................. 36 The Origin of the Signal. ................................................................................... 36 The Larmor frequency. ............................................................................... 37 The 90 RF pulse and the B1 field. ............................................................. 38 The free induction decay (FID). .................................................................. 38 Basic Parameters of MR. .................................................................................. 39 Longitudinal relaxation (T1) and pulse repetition time (TR). ....................... 39 Transverse relaxation (T2) time and echo time (TE). ................................. 40 Spatial localization of the signal. ................................................................ 42 MRI Contrast Methods. ..................................................................................... 44 MRI of Skeletal Muscle. .................................................................................... 48 MRI of muscular dystrophy. ....................................................................... 51 Summary .................................................................................................... 53
6 2 OUTLINE OF EXPERIMENTS ................................................................................ 62 Overview ................................................................................................................. 62 Experiment 1: T2 as a Marke r for Sarcolemmal Damage in Dystrophy Muscle. ...... 63 Hypothesis 1. .................................................................................................... 64 Specific Aim 1. .................................................................................................. 64 Experiment 2: Diffusion Imaging to Monitor Recovery From Damage. ................... 64 Hypothesis 2. .................................................................................................... 65 Specific Aim 2. .................................................................................................. 65 Experiment 3: Magnetization Transfer Imaging to Assess Fibrosis. ....................... 65 Hypothesis 3. .................................................................................................... 65 Specific Aim 3. .................................................................................................. 66 3 METHODOLOGY ................................................................................................... 68 Animals ................................................................................................................... 68 Rodent Handling and Care. .............................................................................. 68 Mouse Strains .................................................................................................. 68 Gene Delivery to Young LGMD Mice. .............................................................. 70 Exercise Induced Eccentric Damage: Downhill Treadmill Running. ................. 71 Magnetic Resonance Imaging ................................................................................ 72 Determinatio n of Muscle T2. ............................................................................. 72 Determination of Mean ADC at 4.7T. ............................................................... 73 Determination of ADC and FA at 11T. .............................................................. 75 Determination of MTR: ..................................................................................... 76 Measurement of Longitudinal Relaxation (T1). ................................................. 77 Histology ................................................................................................................. 78 Histological Verification of Sarcolemmal Damage. ........................................... 78 Histological Quantification of Fibrosis. .............................................................. 79 Verification of transgene expression in gene delivery experiments. ................. 79 4 EXPERIMENT 1: DETECTION OF SARCOLEMMAL DAMAGE IN DYSTROPHIC MUSCLE WITH T2 WEIGHTED MRI .............................................. 89 Abstract ................................................................................................................... 89 Introduction ............................................................................................................. 89 Dystrophic Muscle Undergoes Bouts of Acute Damage. .................................. 89 T2 Relaxation as a Measure of Muscle Damage. ............................................. 90 Multi Exponential T2 Decay in Skeletal Muscle. ............................................... 91 Proposed Origins of T2 Components in Skeletal Muscle. ................................. 92 Purpose and Summary. .................................................................................... 93 Me thods .................................................................................................................. 94 Animals. ............................................................................................................ 94 Gene Correction. .............................................................................................. 94 Histological Verification of Sarcolemmal Damage. ........................................... 94 Non Negative Least Squares (NNLS) Analysis of Multicomponent T2.............. 95
7 Results .................................................................................................................... 96 Assessment of Acute Damage in Dystrophic Muscle Using T2weighted MRI. .............................................................................................................. 96 T2 MRI: Detection of Gene Correction .............................................................. 97 NNLS Analysis of Multicomponent T2 ............................................................... 97 Discussion .............................................................................................................. 99 5 EXPERIMENT 2: IMAGING REGENERATION IN DYSTRO PHIC MUSCLE USING T2 AND DIFFUSION MRI .......................................................................... 111 Abstract ................................................................................................................. 111 Introduction ........................................................................................................... 112 Detecting Changes in Microstructure with Diffusion MR. ................................ 113 Diffusion Tensor Imaging of Skeletal Muscle. ................................................. 114 Experimental Design ............................................................................................. 117 Results .................................................................................................................. 119 Diffusion weighted imaging of recovery from acute damage in healthy control C57BL/10 mice. ............................................................................... 119 Diffusion weighting imaging of damaged and unaffected muscle in mdx mice. ........................................................................................................... 120 Diffusion weighted imaging of rAAV correct sgca/ mice. ............................... 121 Diffusion tensor imaging (DTI) in rAAV corrected sgcg/ mice. ....................... 122 Recovery from eccentric damage. .................................................................. 122 Discussion ............................................................................................................ 123 6 EXPERIMENT 3: ASSESSMENT OF DAMAGE AND FIBROSIS IN DYSTROPHIC MUSCLE USING MAGNETIZATION TRANSFER MRI ................ 136 Abstract ................................................................................................................. 136 Introduction ........................................................................................................... 136 Detection of Fibrosis with Magnetization Transfer Imaging. ........................... 137 MT Imaging of Dystrophic Muscle. ................................................................. 138 MT Imaging of Damaged and Fibrotic Muscle. ............................................... 139 Methods ................................................................................................................ 140 Animals ........................................................................................................... 140 MR Imaging Data Acquisition ......................................................................... 140 MR Image Analysis ........................................................................................ 142 Histological Measurements. ........................................................................... 143 Postnatal Gene Delivery ................................................................................. 143 Statistical analysis .......................................................................................... 144 Results .................................................................................................................. 144 Optimization of Magnetization Transfer Contrast. .......................................... 144 T2 and MTR in Gastrocnemius Muscles From Young and Old Control and Dystrophic Mice. .......................................................................................... 145 Histological Assessment of Fibrosis in Dystrophic Muscle. ............................ 147 Use of T2 and MTR in Monitoring Therapeutic Intervention. ........................... 148 Discussion ............................................................................................................ 148
8 7 CONCLUSION ...................................................................................................... 161 REFERENCES ............................................................................................................ 164 BIOGRAPHICAL SKETCH .......................................................................................... 178
9 LIST OF T ABLES Table page 6 1 Comparison of T2 and MTR for TA and GAS muscles by age and strain ......... 160 6 2 T2 and MTR for unaffecte d regions (T2 < 29 ms) of TA and GAS muscles ...... 160
10 LIST OF FIGURES Figure page 1 1 Bout of damage and repair drive disease progression in muscular d ystrophy. .. 54 1 2 Mouse models of Duchenne (DMD)/Beckers muscular dystrophy and limbgirdle muscular dystrophy (LGMD) compared to a health C57/B10 mouse. ....... 55 1 3 Limitations of classical invasive measures ......................................................... 56 1 4 Spin echo sequence. .......................................................................................... 57 1 5 Longitudinal relaxation: T1 recovery. .................................................................. 58 1 6 Transverse relaxation ......................................................................................... 59 2 1 An overview of dystrophic disease progression and the specific aims of this research. ............................................................................................................ 67 3 1 Mouse age cohorts and experimental design ..................................................... 82 3 2 Gene correction model ....................................................................................... 83 3 3 Radio frequency coil and sample positioning for MRI ......................................... 84 3 4 Diffusion in the plane perpendicular to net fiber orientation (axial) was determined by fitting the exponential diffusion dependant loss of signal with increasing diffusion gradient strength (reflected in the increased b value) in vivo at 11T .......................................................................................................... 85 3 5 MR cradle mounted heating block. ..................................................................... 86 3 6 T1 was measured in the hindlimb muscles of aged C57BL/10 and mdx mice. .... 87 3 7 sarcoglycan expression in r sg/ mice .... 88 4 1 T2 weight accessment of damage in young and old mdx muscle ................... 103 4 2 Histological verification of intensity and distribution of muscle damage in murine models of muscular dystrophy .............................................................. 104 4 3 sarcoglycan in muscle of the untreated .... 105 4 4 sg/ sg/ mice was easily observed in T2 weighted MRI ............................................................................ 106 4 5 Dystrophic muscle pathology is characterized by an increased heterogeneity in T2 .................................................................................................................. 107
11 4 6 Multiple component T2 analysis of computer simulated T2 relaxation times ..... 108 4 7 Multiple component T2 analysis in copper sulfate (CuSO4) phantoms .............. 109 4 8 Multiple component T2 analysis in dystrophic muscle ....................................... 110 5 1 The diffusion of water molecules can be restricted by physical barriers in biological tissues .............................................................................................. 127 5 2 Muscle fibers are normally rather anisotropic and in our experiments the len gth of the muscle fibers run approximately parallel to the Z axis, while the X axis and Y axis are perpendicular to the fiber ............................................... 128 5 3 Muscle fibers of the TA muscle relative to the diffusion s ensitive pulsed field gradients used in the DWI sequences at 4.7T .................................................. 129 5 4 Diffusion weighted imaging of gene corrected LGMD mice .............................. 130 5 5 Directions and diffusion weighting in diffusion tensor imaging. ......................... 131 5 6 sg/ mice .............................. 132 5 7 Schematic diagram illustrating the downhill running protoco l ........................... 133 5 8 Control C57 and mdx hindlimb muscle post downhill treadmill running. ........... 134 5 9 T2 and diffusion indices during recovery from eccentric contraction induced muscle damage. ............................................................................................... 135 6 1 Typical T2 and magnetization transfer contrast and the Z Spectra from damaged and undamaged muscle. ................................................................. 153 6 2 Z spectra from the gastr ocnemius muscles of old C57BL10 control and old mdx mice .......................................................................................................... 154 6 3 Lesions in young dystrophic mice are distinctly visible in magnetization transfer (MT) contrast images and have a similar spat ial distribution as damage observable in T2 weighted images ...................................................... 155 6 4 T2 and MTR in unaffected (T2 < 29 ms) GAS muscle. ..................................... 156 6 5 E xtensive fibrosis was seen in the muscles of older (>1.5yr) dystrophic mice and is not observed in control C57BL10 mice .................................................. 157 6 6 Extensive fibrosis was observed in the old mdx muscle as compare d to the C57 control ....................................................................................................... 158
12 6 7 MT contrast was able to detect genetic correction of dystrophic muscle tissue 159 7 1 A model of multiple methods of MRI contrast at various stages of damage and disease progression is show above ........................................................... 163
13 LIST OF ABBREVIATION S AAV adenoassociated virus B0 magnitude of static magnetic field B1 magnitude of excitatory radi ofrequency field BMD Becker muscular dystrophy. A form of muscular dystrophy with partial expression of the protein dystrophin. DAG Complex dystrophin associated glycoprotein complex. A transmembrane glycoprotein complex that structurally links the cytosol ic protein dystrophin to the extracellular matrix via the protein laminin. This complex includes as subunits the sarcoglycans which are involved in limb girdle muscular dystrophy. DMD Duchenne muscular dystrophy. The most common and severe form of muscular dystrophy due to a lack of the protein Dystrophin. DTI diffusion tensor imaging DWI diffusion weighted imaging ECM extracellular matrix FID free induction decay. FOV field of view GAS Gastrocnemieus muscle LGMD limb girdle muscular dystrophy. This includes several forms of muscular dystrophy. The causes include absence of the subunits of the dystrophinassociated glycoprotein complex, including the sarcoglycans. mdx muscular dystrophy X linked. A genetic mouse model of Duchenne and Becker Muscular Dystrophy due to its dystrophin expression being little to none. MRI magnetic resonance imaging. MTC magnetization transfer contrast MTR magnetization transfer ratio NMR nuclear magnetic resonance.
14 RF radio frequency sgca/ an sarcoglycan deficient knock out mo use that is a model of limbgirdle muscular dystrophy. sgcg/ an sarcoglycan deficient knock out mouse that is a model of limbgirdle muscular dystrophy. SNR Signal to noise ratio SOL Soleus muscle T1 longitudinal relaxation rate constant T2 transverse relaxation rate constant T2* apparent transverse relaxation time TA Tibialis Anterior muscle TE echo time TR pulse repetition time
15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillme nt of the Requirements for the Degree of Doctor of Philosophy NONINVASIVE CHARACTERIZATION OF SKELETAL MUSCLE DAMAGE AND REPAIR IN MURINE MODELS OF MUSCULAR DYSTROPHY By Nathan David Bryant May 2010 Chair: G lenn A. Walter Major: Medical Sciences Physiology and Pharmacology The muscular dystrophies represent a class of devastating neuromuscular diseases that lead to the rapid wasting of muscle. Currently, there is no cure for muscular dystrophy. Duchenne muscular dystrophy is one of the most devastating forms of muscular dystrophy resulting rapid progression and loss of ambulation in the early teenage years. Eventually the disease will prove to be fatal, due to both respiratory and cardiac muscle failure. C urrent invasive assays, such as muscle biopsy, are limited in their ability to accurately represent the entire tissue. The heterogeneous distribution, both spatially and temporally, of damaged and regenerating muscle fibers in the presence of fibrosis and fatty tissue infiltration makes biopsy much less reliable until later stages when the entire muscle is uniformly involved. In contrast, noninvasive imaging methods, such as magnetic resonance imaging (MRI), have the advantage of noninvasive detection of ti ssue pathology in three dimensions. A wide range of endogenous MRI contrast methods provides tools for researchers and clinicians to locate pathological features like muscle damage, edema, atrophy, and fat infiltration but most of these have been used in a purely qualitative manner. With this in mind, I set out to establish several working
16 models of muscle damage, repair, and therapeutic correction in animal models of muscular dystrophy in order to develop effective monitoring and quantitative assessment p rotocols using MRI. I established that transverse relaxation (T2) is a consistent indicator of sarcolemmal damage in dystrophic skeletal muscle in both Duchenne and limbgirdle models of muscular dystrophy. I also explored the multi exponential nature of m uscle tissue T2 and its relationship to muscle damage. Taking it a step further, I studied changes in myofiber structure noninvasively using diffusion weighted imaging follow eccentric damage from downhill running. The use of diffusion tensor imaging (DTI) provided detailed insight into the underlying changes occurring over the throughout a two week period post injury. Ultimately, the muscle exhausts it ability to undergo this regenerative phase and its contractile tissue is replaced by fibrofatty tissue. Unfortunately, fibrosis is extremely difficult to directly imag e, due to it s extremely short transverse relaxation times. Therefore, I investigated an alternative approach to image tightly bound water molecules associated with tissue fibrosis. M agnetizat ion transfer (MT ) allows for the indirect imaging of fibrosis by looking at the proton MT between a free liquid pool and a macromolecular bound pool, of which, collagen is a major component in fibrosis. Combining these different basic MR imaging properties of muscle, I developed a multimodal approach to gain new prospectives in monitoring disease progression and detecting efficacy of potential treatments in animal models of muscular dystrophy.
17 CHAPTER 1 INTRODUCTION AND BACKGROUND The ultimate goal in mus cular dystrophy research is to find cures for the various forms of the disease. While many promising preclinical trials are under way, there still largely remains the lack of a noninvasive monitoring protocol that provides a detailed profile of a muscles pathological condition in vivo Non invasive measures offer a significant advantage over traditional histological methods on several fronts. For one, they allow repeated measures on the same subject, in longitudinal studies, without destroying the tissue in question. Another benefit is that the researcher is able to quickly collect data representing large areas of the body simultaneously. This provides data that reflects the entire muscle in question, as opposed to being confined to the histology observabl e via a localized muscle biopsy. For instance it has been shown that an accurate measure of muscle fiber type distribution with in the healthy human quadriceps muscle requires at least three biopsies (1). Skeletal muscle and its patholog ies have been studied with noninvasive modalities such as computed tomography (CT; (2)), ultrasonograph y US; (3, 4) positron emission tomography (PET; (5)), and magnetic resonance imaging (MRI; (6)) and spectroscopy (MRS; (7)). Each modality has its unique advantages. For instance, ultrasound provides a cheap and fast measurement of fascicle structure (8), but lacks metabolic information. PET on the other hand provides extremely sensitive measures of small amounts of target tracers and glucose metabolism, but lacks anatomical information. MRI can provide both highresolution anatomical and metabolic information. This is a great advantage, as regi stration between spatial data of two unrelated noninvasive techniques can be very difficult; for example overlaying PET data over high
18 resolution anatomical CT data. By using MRS and MRI, all of the metabolic or chemical composition data can be collected w ith the same experimental set up as the high resolution imaging. In addition, depending on what nuclei the instrument coils are tuned to, a wide variety of tissue composition and relative metabolic intermediate changes can be recorded. For example, localiz ed 1H MRS can be used in conjunction with numerous two and threedimensional imaging protocols using the same 1H tuned coil. This type of experiment is especially useful in looking at water to fat ratios in various biological samples and with the ability t o precisely defining specific regions from which the spectra is collected (9) 1H MRI can also be used along side 31P MRS and has proven to be valuable in healthy and diseased muscle research (10) Such studies in skeletal muscle are often interested in following changes in phosphorylated metabolic intermediates, includi ng phosphocreatine (PCr) and the individual phosphate groups the various phosphorylated forms of the predominant energy shuttle and regulatory nucleoside adenosine; ie ATP, ADP, and AMP (11) In these 31P experiments, environmental changes in muscle tissue can also be monitored; such as changes in intracellular pH (12) For those researching diseases in humans and in animal models, the vast and powerful applications of NMR ar e commonly available in both the clinical and research laboratory setting. As such, MR is a rapidly growing field in noninvasive medical imaging that we feel is perfectly suited for the study of disease, atrophy, damage and repair in skeletal muscle. Skeletal Muscle Skeletal muscle is a unique tissue and displays impressive levels of specialization making it possible for us to move about our environment, speak, and draw each vital breath of air. Skeletal muscle exhibits incredible plasticity (13) making it sensitive to
19 disuse atrophy, but also allowing for drastic adaptation with rehabilitation (14) Muscle also has a tremendous regenerative capacity capable of completely regenerating itself within mere days of an injury (15) In a healthy individual, muscle injury is a common event and plays a role in every day hypertrophy and in the extreme cases of muscle building in weight bearing exercise. Normally, when minor damage to a muscle fiber occurs, growth factors and cytokines are released and populations of pluripotent stem cells and muscle satellite cells are activated and will home toward the source of the signal. The muscle progenitor cells either fuse with existing damaged fibers or they differentiate and fuse with each other to form new fibers. Overall this results in a very dynamic process resulting continual muscle growth, remodeling and repair. For all these reasons, it makes it all more devastating when this regenerative capacity is exhausted in the event of disease. This is dramatically illustrated in children with Duchenne muscular dystrophy (DMD) that have normal to enlarged muscles prior to clinical manifestations (Figure 11) due to the inability to match tissue regeneration with the damage associated with normal ambulation, breathing, or cardiac function in the face of a structural muscle deficiency. Growth and Repair of Skeletal Muscle Skeletal muscle tissue is made up of bundles individual muscle fibers. Each of these muscle fibers is a multinucleate d single cell. The distance between these myonuclei is heavily regulated, forming myonuclear domains (16) This plays an important role in understanding repair of damage and recovery from atrophy. As a myofiber decreases in diameter, these myonuclei are brought closer together. Once the minimal spatial barriers of the myonuclear domain are violated the myonuclei will become an apoptotic nucleus and it will be lost by the muscle fiber. Although, it has
20 been debated as to which event triggers the apoptosis in the nuclei, Gundersen et al. observed no myonuclear loss in the first weeks of atrophy but did see a decrease in fiber diameter (17) This may suggest that there is a range of acceptable fiber diameters and that there is a lower size threshold that is associated with the loss of nuclei. Mature myonuclei do not divide and the muscle fibers are limited in their diameter by the number of their nuclei; i.e. there seems to also be an upper threshold that limits fiber diameter when there ar e too few nuclei in that region. This then affects the ability of muscle tissue to recover from necrotic damage or atrophy; where many myonuclei were lost. In order for the fibers to increase in diameter, new myonuclei must be introduced to the fiber. Thes e nascent nuclei come from satellite cells (18) and muscle progenitor cells. During muscle repair, on a cellular level, many events are similar to embryonic development (19) A reserve of these satellite cells reside largely in the basement membrane, between the plasmalemma and the basil lamina, surrounding the myofibers and are generally in a quiescent state in adult muscle tissue (20) These populations of satellite cells have been shown to be heterogeneous in nature (21) but as a group they are often characterized as being positive for M cadherin, Pax7, Myf5, and neural cell adhesion molecule1 (22) During development, several transcription factors are expressed that regulate early genes in muscle cell differentiation and myofiber maturation. Among these transcription factors are Pax3 and Pax7. Lepper et al. in 2009 demonstrated that while Pax3 and Pax7 had a role in embryonic development, only Pax7 was important to mechanisms of growth and repair i n the postnatal animal (23) In fact, due to its roles in survival, proliferation and commitment to a muscle lineage of differentiation, Pax7 has become a recognized cellular marker for tracking muscle
21 satellite cel l activation during these growth and repair events in the laboratory. In addition to Pax7, several other markers of muscle cell lineage and differentiation are used to study these processes. Members of a Myc c related superfamily are closely monitored in t he laboratory and include MyoD1, myogenin, myf 5 and MRF4 (24) In general, these regulatory proteins activate other member s in the family resulting in a cascade that results in myoblast and fiber maturation (24) When new muscle growth or repair is needed in adult skeletal muscle these satellite cells (Pax7+/Pax3+/ -/Myf5+/ -) are activated and proliferate (25) A small fraction of these new satellite cells will not express MyoD, and they will presumably return to the quiescent pool to replenish the reserve f or future repair and growth (25) The MyoD+ cells become mononucleated myoblasts (Pax7+/Myf5+/MyoD+) (26) and migrate to the area of tissue that needs repair. As these myoblasts fuse, they will start to express myogenin and later the nucleus within this domain will start to express embryonic myosin. Thus, embryonic myosin is a marker of new or recently repaired myofibers. These nascent muscle fibers will also have centrally located myonuclei, providing a histological marker of new fiber growth and repair, which move to the periphery upon maturation (marker of healthy adult skeletal muscle). Structural Organization Within the Muscle Fiber Within each muscle fiber, are dense arrays of myofibrils that are made up of repeating sarcomeres. Each sarcomere is in turn made up of overlapping thick and thin filaments (of myosin and actin/tropomyosin/troponin respectively) that are interdigitated and able to slide past each other during contraction. At each end of the sarcomere, is a Z disk. The Z disks organize the ends of the filaments and in striated skeletal muscle, are bound to neighboring Z disks, giving the tissue its recognizable striped or striated
22 appearance, and add to the organized structure of the myofibrils. The Z disks are tethered to the center of each sarcomere by the massive protein titin. Titin is the largest known protein in humans and each individual molecule spans half of a s arcomere. One terminal domain of titin has an elastic property, while the other largely acts to regulate the length of extended sarcomeres. The thin filaments are bound to the protein nebulin, which performs a similar function for regulating the length of the thin actin filaments (27) During contraction it is the heads of the myosin proteins making up the thick filaments that bind to active sites on the actin filament and then follow through with a power stroke, pulling the Z disks one increment closer together. The myosin head then releases, triggered by Ca++ influx to the cytopl asm from the sacroplasmic reticulum, and re cocks as ATP is hydrolysed by the protein; thus it actually requires energy to relax the muscle as opposed to contract it. This can be observed in the muscle stiffening that occurs following death; rigor mortis The myosin head alternate relaxing, binding the muscle contracts as each sarcomere simultaneously shortens. During eccentric contractions the muscle is forced to lengthen during this myosin head cycling event and damage is done to the fibers at the level o f the z disk (28) and muscle membrane (29) This will be an important concept later in this dissertation. The Z disks are rich in the protein desmin, which is also found in costameres (30) Desmin is solely found in muscle tissue and is another biological marker for muscle lineage differentia tion (both myoblasts and fibers are positive for desmin). Costameres are complexes associated with desmin and other cytoskeletal proteins and are found in the outer membrane of the cell and serve to link the contractile machinery and f actin cytoskeleton t o the
23 extracellular matrix (31) One such costamere that is particularly import in muscular dystrophy is the dystrophinassociated glycoprotein complex or DAG complex, because this is where the gene mutation in many dystrophies aff ects the stability of the cell membrane making it susceptible to damage. Dystrophin and the Sarcolemma In muscle fibers, the cellular plasma membrane is called the sarcolemma. The f actin cytoskeleton is tethered to the sarcolemma through the structural pr otein dystrophin. The dystrophin proteins are anchored to the inner surface of the sarcolemma through the dystrophin associated glycoprotein (DAG) complex (10) costamere. On the extracellular side, the complex is then attached to laminin, which in turn is bound to the extracellular matrix (ECM). The ECM is largely made up of the protein collagen and proteoglycans. The various forms of muscular dystrophy arise when gene mutations prevent any of the proteins in this network that stabilizes the sarcolemma during muscle contraction. Not all of the DAG complexes functions are structural. The complex is large and has many subunits. Some of these bind to cellular proteins that are important in cellular signaling and regulation metabolism. One such protein that forms a part of the complex is neural nitric oxide synthase (nNO S). nNOS is required in skeletal muscle to maintain activity following exercise (32) This lead to the observation that mice with disr upted DAG complexes were more susceptible to fatigue (32) further debilitating the individual. Muscular Dystrophy Clinical Impact and Symptoms The muscular dystrophies are a group of genetic diseases exhibiting progressive skeletal muscle wasting due to a structural weakness in the sarcolemma. The most
24 prevalent and severe form, Duchenne muscular dystrophy (DMD), arises from a mutation in the 2.4Mb gene coding the protein dystrophin located on the X chromosome and its phenotype is recessive. In DMD, this mutation is generally a nonsense or frameshift mutation that results in the complete absence of the functional protein. In the less se vere Becker muscular dystrophy (BMD), some level of dystrophin is present, though it is often structurally truncated or has a reduced expression level. DMD/BMD occurs in one out of 3,500 to 5,000 male children, giving rise to an estimated 400 to 600 new ca ses in the United States each year (CDC, SGDD 2006). Onset of symptoms of DMD can often be seen between 2 to 6 years of age and manifest as general weakness of voluntary skeletal muscles, especially in the hips, pelvis, thighs and shoulders. Impaired ability to rise to a standing position from a prone position is common and the use of the arms to walk up the body until upright, also know as the Gowers sign, is suggestive of a dystrophic diagnosis. Enlarged calf muscles are also typical and are considered to be the result of ps eu dohypertrophy (33 36) In DMD, the loss of effective ambulation usually occurs around the age of 12, but could occur as young as 7 years old. As the disease progresses, there is increasing i nvolvement of the muscles of respiration (diaphragm) and of the heart. This leads to further complications, often in the late teenage years to the late twenties, where weakened respiratory function, cardiac failure, or an increased susceptibility to infect ion may prove to be fatal. In BMD similar trends are observed, but the time of onset is shifted later in life. The symptoms are often milder than in DMD, yet the path of progression is much less predictable. In addition to the dystrophies that lack dystrophin itself, there are numerous other forms that are caused by mutations that result in the disruption of the dystrophin associated glycoprotein
25 (DAG) complex. Among these, are the limbgirdle muscular dystrophies (LGMD), of which at least 15 forms of LGMD are currently recognized. The different forms of LGMD arise from autosomal mutations and both dominant and recessive mutant alleles have been observed. Some types of LGMD are caused by the lack of sarcoglycan, transmembrane glycoproteins in the sarcolemma. These proteins are integral components of the DAG complex that acts as a structural tether between dystrophin and the extracellular matrix. Onset in LGMD is first seen in muscles around the hips and shoulders, with slower progression than in DMD, and ther e is a concern for cardiopulmonary complications later in life. Overall, LGMD is less severe and less prevalent than DMD and BMD, but since it is autosomal, females are equally affected. Beyond these examples, it should also be noted that there are a wide spectrum of muscular dystrophies that have been identified. And they do not all result in the same symptoms or severity due to a vast range of underlying causes. Several years ago it was noted that there were 34 clinical disorders that were currently under the umbrella of the dystrophies (37 ) There are the somewhat less severe distal muscular dystrophies like Miyoshi myopathy due to a mutation of dysferlin (38) and the tibial muscular dystrophy which has a mutation in titin (39) There are als o congenital muscular seizures, speech problems, and contractures (37) Fukuyama CMD effects primarily populations in Japan and causes neural malformations. Still other forms like Emery Dreifuss (EDMD) arise from deficiencies of emerin and laminin and lead to cardiomyopathy and contractures early in life. Finally I will mention myotonic (DM) that
26 has defects in myotinin protein kinase and ZNF9. Myotonic dystrophy has several different forms and is the most co mmon dystrophy observed in adults (37 ) Pathology Common to many of the dystrophies associated with dystrophin (dystrophinopathies) and the sarcoglycans (sarcoglycanopathies) is an increased susceptibility to contraction induced muscle damage due to a compromised sarcolemmal integrity (Fig. 1b). As the dystrophic muscle contrac ts, the internal cytoskeleton (f acti n) is not structurally anchored to the muscle plasma membrane and the extracellular matrix. This structural and functional disconnect, leads to a tearing of the sarcolemma during what would normally be withstood by normal muscle (see Fig 11). As the damaged muscle fibers are repaired, they still lack the missing structural component and will inevitably be torn again by future muscle activity and the process will be repeated. This res ults in an ongoing cyc le of muscle damage and repair. This is observed clinically as a rise of muscle enzymes and proteins in the blood (i.e. the m isozyme of creatine kinase) levels, indica tive of damaged muscle tissue. The tears in the membrane are large enough to allow large serum proteins such as albumin and immunoglobulin into the fiber, indicating that even large proteins are able to freely transverse the normally impermeable membrane (40) When looking at hist ological samples signs of tissue remodeling and regeneration are observed simultaneously with a heterogenious distribution. Regions of necrosis are also seen with vast areas of edema and immune cell infiltration. While such damage can be observed in stained tissue sections from muscle biopsies, it has been difficult to observe this increase in membrane permeability in vivo Furthermore, due to the sporadic spatial and temporal distribution of muscle damage, histological lab results may not well represent t he extent
27 of damage to the entire muscle. When it is also taken into consideration that not all muscles are affected equally, the need for a noninvasive in vivo detection method becomes further evident. Laboratory Models of Muscular Dystrophy In the labor atory setting researchers are now fortunate to have several animal models of the various forms of muscular dystrophy. These genetic models span a wide range of forms of the disease and have been developed in many different background species. The most comm only used models are murine and canine. The in vivo work carried out in the studies described in this dissertation was carried out using the mouse models described below. Animal models of dystrophinopathies The mdx mouse has been used extensively as a model of Duchenne and Beckers muscular dystrophy. The name mdx refers to the mouse having a X linked muscular dystrophy phenotype ( see Figure 12) that is due to specific point mutation resulting in premature termination of translation of the protein dystro phin (41) The mdx mouse has a low level of revertant fibers that express low levels of dystrophin in an otherwise dystrophin null animal. This contributes to the mild dystrophic phenotype in the strain and is the factor that allows it to also serve as a model for BMD. One sizable difference between the mdx and human DMD/BM D phenotype is the lack of severe fat deposition in the mouse model. While this does affect the ability to directly translate some of the data from noninvasive imaging techniques described later in this chapter, this does provide a simplified model where r esearchers can concentrate solely on the muscle fibers themselves giving a clear view of the source of the underlying problem. Like human individuals with an X linked mutation in the gene for dystrophin (Figure 11), mdx
28 mice are susceptible to contractioninduced injury during normal muscle use and also possess many of the histological changes described in humans. Mice with mutations in the dystrophin gene have also been described in the mdx 2 5CV mice, which display an increased variation in muscle fiber diameter, and in mdx 52 mice, which have large deletions in the dystrophin gene, similar to more than 65% of human patients (42) Since a mutation in exon 52 was induced, this generates a slicing error between exons 51 and 53 and as such this mouse model also lacks the shorter dystrophin isoforms Dp140 and Dp160, along with dystrophin itself (43) While 90% of the muscle fiber in mdx 52 mice had centralized nuclei, the muscle did still have a small amount of revertant fibers that immunolabeled posit ive for dystrophin; similar to mdx mice (43) In addition to these mdx variations, there are also double mutants, like the dystrophin/utrophin null mouse ( mdx ; utnr / -), that show a much more severe phenotype. One possible explanation for the mild mdx phenotype is the role of utrophin compensating for the loss of dystrophin in these animals. In the mdx ; utnr / double mutant, both of these proteins are m issing and a model with a much more severe pathology was created, replicating many more classical symptoms of the human disease. While the mdx mice have bouts of damage that occur early in age and then lie dormant and then gradual weakening is observed lat er in life, the mdx ; utnr / double mutant mice have a severe phenotype early on and progressively get worse. During this course they develop necrotic lesions, muscle weakening, kyphosis, contractures and fibrosis. It has therefore been argued that double mutants like this serve as a more accurate model for preclinical trials (42) Other double mutant s include mdx ; adbn/ dystrobrevin, mdx ; / which appears to be deficient in its muscles
29 to regenerate, mdx ; MyoD/ which also have a reduced capacity to regenerate and are a model of cardiomyopathy related to DMD, and finally the md x; PV/ double mutant that lack parvalbumin (PV), a calcium binding protein, also displays a slightly increased severity in the dystrophic phenotype (42) While the murine models are widely used, canine models of X linked muscular dystrophy (CXMD) are more similar to the human disease (44 46) The most establis hed of these is the golden retriever muscular dystrophy (GRMD) model. Though the model is much more relevant than the mice, the dogs display a higher amount of phenotypic variation (even within the same litter) and require and extensive amount of care in order to maintain (46) Early studies demonstrated that the phenotypes were much less severe in females and in smaller breeds of dog (45) but the beagle also remains an attractive model animal due to its smaller size making it easier to m aintain in laboratory kennels (47) In Japan (2003), a beagle model (CXMDJ) was produced as a backcross from the GRMD model via artificial insemination (47) It was argued that the beagles body size was more appropriate for a dystrophy model a nd that the milder phenotype also made it easier to care for them, as well as increasing their life span making longer longitudinal studies possible (47) Later it was reported that the symptoms in the now well established CXMDJ colony are nearly identical to those seen in GRMD, but with the added benefit of a much smaller animal (48) Finally, in 2006 Yugeta repor ted cardiac involvement in the CXMDJ model (49) further making it one of the more attractive animal models in use to date. Animal models of sarcanopathies. The BIO 14.6 cardiomyopathic Syrian hamster (50, 51) became the first animal model of human sarcoglycanopathy (specifically for LGMD2F) after it was reco gnized
30 sarcoglycan gene was the underlying cause of the pathological phenotype (46) Mouse models of limb girdle ( LGMD ) have been created in which specific dystrophinassociated glycoprotein null mice, such as sarcoglycan null ( sg/ -) or sarcoglycan null ( sg / ), have been made using knockout technology (12, 13) These mouse models serve as useful preclinical tools to test various therapeutic interventions aimed at curing or halting disease progression in human. They have been successfully used to show the potential of cell therapy (14 16) gene therapy (17 19) and pharmacological interventions (20, 21) all as potential candidate treatments for muscular dystrophy. Approaches to treatment of muscular dystrophy Gene therapy. In the majority of the for ms of muscular dystrophy, the disease is caused by a single gene mutation. This makes the dystrophies prime targets for gene therapy strategies. There are several approaches that are under development. One approach is gene delivery. There are many strategi es in the gene delivery mechanism itself. One of the most direct approaches is the direct injection of plasmids containing the full length gene into the muscle, followed by electroporation (52) In the case of DMD, it was observed that plasmid expressed full length dystrophin may have been protected for eliciting an immune response in mdx mice due to the life long presence of revertant fibers in the mouse model (52) While the se studies were valuable proof of concept experiments, long term expression and systemic delivery to all of the bodys skeletal muscle would not be possible using this delivery method. Another promising approach is viral mediated gene therapy vectors. One limitation encountered when
31 dealing with the dystrophin gene is its length, spanning over 2,300 kilobases and requires 16 hours to transcribe (53) One approach to ameliorating dystrophin deficiency using smaller Parvovirus vectors is to use modified, truncated genes that, the produc t of which, would still functionally add stability to the sarcolemma. Examples of these include minidystrophin and microdystrophin. Adenoassociated viruses (AAV) has been successfully utilized as a vector for minidystrophin delivery in mdx mice and was sh own to produce long term expression at least up to one year (54) Adeno associated viruses (AAV) have also been found to be an extremely efficient tool to deliver other foreign, booster or other missing genes to specific dystrophic muscles (55 57) For instance, our group has shown that in LGMDIIa skeletal muscle can be rescued following neonatal gene transfer of human sarcoglycan in mice (56) and clinical trials have recently shown AAV can be a safe and effective vector for delivering the sarcoglycan to human muscle (58) An indirect approach using gene therapy, is the targeted over expression of booster genes. The expression of utrophin, an autos omal homlogue of dystrophin, could be an alternative approach to rescuing dystrophic muscle when the host might have an immune response to foreign proteins (59) Furthermore, it has long been suggested that expression of proteins such as agrin may also partially rescue dystrophic muscle, largely by further promoting utrophin expression (60) And finally, other studies take still another approach, such as the use of transgenic knock in mice ( mdx : mIgf+/+) that demonstrate local over expression of insulinl ike growth factor 1 (IGF 1) increases the regenerative capacity and decreases necrosis in mdx skeletal muscle (61, 62) ; suggesting IGF 1 would also be a promising booster gene for treatment of MD. The use
32 of systemi cally delivered recombinant human IGF 1 bound to an IGF 1 binding protein 3 (rhIGF 1:rhIGFBP3; marketed as IPLEX) is currently in clinical trials for myotonic dystrophy (63) Pharmacological therapy. Since the underlying deficiency in MD is attributed to gene mutations, many pharmacological therapies are carried out in hopes of lessening the symptoms, but fail to address the primary cause. Unfortunately, these are the most commonly applied therapies in the clinical setting. By far the most common treatment is the prescription of steroids; specifically Deflazacort and Prednisolone (64 66) In the preclinical arena and in clinical trials, several other approaches are showing promise. One area is the idea that pharmacologically reduced protein degradation would result in less muscle loss upon injury. In such a strategy, protease inhibitors are administered to promote the maintenance of muscle mass. Specifically, the BowmanBirk inhibitor (BBIC), derived from soybean extract, has shown promise in preventing muscle atrophy after disuse (67) in healthy mice. Another pharmaceutical approach is to apply agents that promote read through of early stop codons. These compounds act similar to the antimicrobial gentamycin, but with much less cytotoxicity. One such compound that is currently in early clinical tri als is PTC124. The compound PTC124 (commercially known as ataluren) is a 284 Da, achiral, 1,2,4oxadiazole linked to flourobenzoic rings (68) which can be orally administered and promotes the expression of a slightly modified, but functional dytrophin protein via the readthrough of premature stop codons that would other wise sabotage the functional expression of dystrophin. Recent studies o f PTC124 in mdx mice show very promising results (69) as well as clinical trials for the treatment of cystic fibrosis and DMD (68, 69)
33 Finally, other approaches, that lie in between pharmacological and gene therapies involve exon skipping with the introduction of antisense oligonucleotides (70, 71) These methods utilize morpholinos to bind around and slice out the region where the m utation occurs in the mRNA transcript for the dystrophin gene, resulting in a truncated, but functional form of the protein to be expressed (71) There appears to be a higher rate of mutation in exons 44 55 of the dystrophin gene, which codes for the structurally spanning rod region of the protein (68) The rational for this approach to treatment comes from observations that inframe deletions in this region leads to the less severe clinical phenotype seen in Becker mus cular dystrophy (BMD) (68) By using antisense oligonucleotides (AONs), or morpholino, to target RNA splicing, the rescued mRNA is shortened but remains in the correct reading frame. Optimizing the delivery of these morpholinos to the myonuclei has been the focus of much research, including the addition of copolymers (68, 7274) similar to DNA transfect ion agents, and recombinant adenoassociated virus (rAVV) vectors (75) Clinical trials have shown promising proof of concept results in humans (76) The wide range of mutati ons in DMD patients may either require extremely personalized treatments (68) or we may find that morpholino cocktails may make a treatment that h as a broad application (77) In order for these treatments to be clinically effective, researcher must overcome inefficiencies in delivery and in restoration to dystrophin expression (68) ; it is estimated that just 30% of wild type levels of expression may be sufficient to achieve this goal (78) Cell therapy. Finally, it is worth briefly noting that there is a growing wealth of literature on studies of dystrophic tissue correction via the introduction to either gene corrected i ndigenous muscle progenitor cells or through delivery of exogenous cells
34 from a healthy donor, but an in depth discussion of this area is beyond the scope of this dissertation. Cell therapy has the benefit of potentially restoring the target tissue via de novo fiber formation, while gene therapies and traditional pharmaceuticals generally require the target tissue to be responsive in order for the treatment to be effective. Studies in mdx mice have been promising, but early trials in human patients revealed that many of the implanted cells die off rapidly (79) More recently, pluripotent cells extracted from the vasculature, dubbed mesoangioblasts, have shown great promise in engrafting and repairing dystrophic muscle in the GRMD dog (80, 81) To date, human trials have shown promise, but systemic delivery and cell survival are major hurdles that will need to be overcome to make cell transplant a viable therapy for muscular dystrophy (25) These recent developments have lead to great hope within the hearts and minds of patients with neuromuscular dis orders and their families. Unfortunately this also opens the door to those in desperate circumstances to seek out unproven experimental treatments before they have been proven safe or effective. MacReady, in 2009, discussed the ethically grey areas of an emerging trend of stem cell tourism, where families will travel to regions that lack strict medical oversight or to cities where clinics provide blatantly illegal procedures (82) Ultimately, it is thought that a combined approach using aspects of gene transfer and regulation, pharmaceutical agents and stem cells may be required for complete amelioration of the pathological manifestations of the disease. Monitoring disease progression. In evaluating natural disease progression and determining effectiveness of treatments in preclinical trials, the methodology can be categorized as being in one of two broad groups. The classical histology studies have been considered the gold
35 standard for many years and would be considered invasive procedures. In an animal model as small as a mouse, muscle histology is often a terminal end point and this limits our ability to perform truly longitudinal studies of a single muscles recovery from damage and thus, it increases the number of animals that need to be sacrificed at each time point in such studies. In addition biopsies have a rather limited scope and often may not represent the tissue very well in dystrophic muscle (see Figure 13). More recently developed methods are noninvasive and they provide many advantages. These include experimental design with repeated measures on the same animal to get a true assessment of disease. Numerous functional and behavioral tests have been described and are suitable for collecting noninvasive data (83) but these do not give the precise special and st ructural information that can be extracted from noninvasive imaging data. As mentioned at the beginning of this chapter, these methods include computed tomography (CT; (84) ), ultrasonography US; (4, 85) positron emission tomography ( PET; (5)), and magnetic resonance imaging (MRI; (6)) and spectroscopy (MRS; (7)). In our laboratorys work, we have found magnetic resonance imaging (MRI) and spectroscopy (MRS) to be particularly well suited for the threedimensional study of soft tissues such as skeletal muscle. Magnetic Resonance Imaging (MRI) The major advantage of noninvasive measures li ke MRI is that it allows the acquisition of three dimensional images at the resolution of a muscle cell, visualizing all of the muscles of the entire mouse limb, in real time and in the living animal. At its core, MRI is a spectroscopic technique and a bri ef overview of a few fundamental concepts will enrich the discussion presented in this dissertation for those unfamiliar with this methodology.
36 Basic C omponents R equired for MR or NMR. The modern designs for both clinical and experimental instruments share common elements. These include a static magnetic field, electromagnetic gradient coils used for spatial encoding of the signal, a radio frequency (RF) coil to used excite and receive the NMR the signal, RF signal amplifiers used in the transmission and reception of RF pulses and the incoming signal, and a computer workstation for the operator to control the system and to transform the acquired data. The static magnetic field is sustained by a liquid helium cooled, super conducting electromagnet. The field strength is the strongest at the convergence of the magnetic field lines, which in the case of these cylindrical magnets, is in the middle of the bore. The field strength (B) is measured in the unit Tesla (T), which is equivalent to 10,000 Gauss (G). The static field of the magnet is termed B0 and ranges between 0.1T and 21.1T for imaging purposes. The larger B0 is the larger the induced nuclear polarization and resulting signal strength will be. The majority of the work presented in this dissertation was collected at either 4.7T or 11.1T. The O rigin of the S ignal. In the presence of a static magnetic field, the nuclear spins of specific elements will be polarized in ratios that will enable us to manipulate their orientation and acquire a time dependent change in induced signal as they relax back to equilibrium. The work presented here primarily relies on hydrogen nuclei for the source of our signal. Since the hydrogen nucleus consists of a single proton, 1H MRS and MRI are often also referred to as proto n spectroscopy or proton imaging, respectively. In this simple spin system, the axis of the nuclear spins align either parallel or anti parallel to the static
37 field. The relative number of spins in either orientation is given by the Boltzmann distribution: [Equation 11] where N+ spins in the lower quantum energy level are more frequent than the number of Nspins in the higher energy state. This small excess of spins in the lower energy state is described by the exponential function of the ratio between the energy difference between the two states (complex term: 2 ( k = 1.3805 x 1023 J/Kelvin) and the temperature in Kelvin (T). The energy difference between the two energy states is equal to twice the product of the field strength B and us. At equilibrium, there is a very small number of spins aligned parallel (N-) to the Z axis of the magnet. Thus, at the fully relaxed state, the net magnetization vector that lies on the Z axis or the direction of the applied magnetic field and is referr ed to as Mz. At room temperature this holds as a good approximation for spin energy level distribution and is therefore relevant for biological samples. The Larmor frequency. The precession of the nuclear spin around this axis is at a rotational frequency that is determined by the Larmor equation: = B [Equation 12] where is the precession frequency in megahertz (MHz), is the gyromagnetic ratio of a specific atomic nucleus in megahertz per Tesla (MHz/T), and B is the magnetic field strength in Tesla (T). The gyromagnetic ratio for the hydrogen nucleus is 42.576 MHz/T.
38 Thus at a field strength of 4.7T, the precessional frequency of the 1H nucleus is 200.107 MHz about the Z axis. The 90 RF pulse and the B1 field. If the researcher induces a RF pulse, transmitted perpendicular to the Z axis, at the Larmor or resonant frequency, energy is absorbed by these specific nuclei and the net magnetization vector will be tipped into the x y plane. The RF field results in a second magnetic field and is term ed the B1 field. There are many RF pulse sequence strategies available, but in the most elementary NMR experiment the result of an effectively applied RF pulse would tip the net magnetization vector off of the Z axis completely into the X Y plane. Since the nuclei have angular momentum about the Z axis, as soon as they are tipped into the X Y plane they continue to precess about the Z axis. Since the orientation of the magnetization vector is brought from being parallel to the Z axis to being in the X Y pla ne, this is referred to as a 90 RF pulse. While tip angles other than 90 are commonly used for preparing the spins for various spectroscopic and imaging applications, it is always the net magnetization vector in the X Y plane (MXY) that is the source of the signal detected by the RF coil. The free induction decay (FID). In this simple experiment, the net magnetization vector is at equilibrium, precessing around the Z axis and Mz is at its maximum (M0); so initially Mz = M0. The M0 is a function of the proton density (PD) of the sample and the polarization of nuclear spin states increasing with magnetic field strengths (B0). A fter a 90 RF pulse plays out, the Mz is reduced to null and MXY is at its greatest. As MXY precesses freely around the Z axis, the associated changing electromagnetic field (EMF) induces current in a RF receiver coil, according to Faradays law of electromagnetic induction. If this vector were
39 to remain the same magnitude and within the X Y plane the plot of the induced current would be a sine wave. In the laboratory however there are multiple processes that result in the dampened sinusoidal decay in the magnitude of the MXY vector over time. This signal is termed the free induction decay (FID), which contains the frequency, amplitude and relaxation information and is the phenomenon that makes MRS and MRI contrast possible. Basic P arameters of MR. In three dimensions a plot of the net magnetization vector, following an initial 90 RF pulse, would resemble an inverted funnel rising out of the X Y plane, spiraling upward as it precesses increasingly tighter around the Z axis until it reaches equilibrium once again fully aligned with the Z axis. As described earlier, when the Mz is equal to M0, this spin system is at equilibrium. Longitudinal relaxation ( T1) and pulse repetition time (TR) Following a 90 RF pulse (Figure 14), Mz will be approximately zero. Over time the Mz magnetization will recover exponentially. The time constant for this recovery of the Mz is characterized by longit udinal relaxation and is called T1. The T1 is a time constant that is specific to a given sample or body tissue at a specific magnetic field strength. The T1 time constant is due to spinlattice interactions and is also referred to as the longitudinal rela xation. In general, samples that are more solid in nature have a shorter T1 than samples that are more fluid in nature due to these spinlattice interactions. T1 tends to lengthen for samples as the B0 field increases in strength, which can make measurements at higher field strengths require a greater interval of time in between signal acquisitions in order to achieve the same initial M0. This interval is termed the pulse repetition time or TR and should exceed 5 to 6 times the T1 of the
40 sample if no T1 wei ghting is desired. Otherwise, if the TR < T1, the Mz will not be allowed to fully relax to M0 and the net magnetization vector will have less magnitude to be tipped into the X Y plane for the subsequent signal acquisitions (Figure 15). Thus the reduction in signal will not be due to a lack of proton density but due to the TR being too short for complete longitudinal relaxation of the sample, resulting in signal saturation. In such a case the data would be said to be T1 weighted. Transverse relaxation (T2) time and echo time (TE). In most practical MR imaging applications there are many reasons why the FID is not directly acquired. Instead, an echo of the FID is the source of the collected data (Figure 1 4). This is partially due to limitations of common MR hardware, including the speed of signal digitization, but this can be an advantage as many useful contrast and signal selectivity methods arise by playing out other RF pulses and electromagnetic (EMF) gradients before and following the 90 pulse. There is an exponential decay of the MXY vector due to spinspin interactions, leading to the gradual loss of signal intensity, which is also referred to as the transverse relaxation. The time constant for this reduction in MXY magnitude is called T2. While the permanent loss of MXY following the initial 90 pulse is due to T2 spin spin interactions, a more rapid apparent signal loss is observed. Although the B0 field is shimmed or optimized for homogeneity for each sample, local inhomogeneities will cause the observed time consent of the decay of the FID to be shorter than the actual T2 time. This observed time constant is referred to as T2* ( T2 star ) and is commonly measured using gradient echo imaging. As was mentioned above in the Larmor equation, the freq uency of precession is due to the external magnetic field strength experienced by each nucleus and the magnet. As the protons interact with their
41 environment, including B0 field inhomogeneities and temporal interactions with neighboring molecules, some nuc lei speed up while others are slowed down. The result is a dephasing of the individual spins, which leads to an accelerated decay of the original FID and each echo, this is due to T2*. A sample with a short T2 may have magnetization in the X Y plane for a long time if it has a long T1, but the signal will die out quickly due to the rapid dephasing of T2* mechanisms. Immediately following the 90 RF pulse, the net magnetization vector in the X Y plane is at its greatest. If a 180 pulse is applied, the i nvestigator is able to refocus the coherent signal, producing an echo of the original FID. The period of time allotted to refocus the transverse magnetization is termed the echo time (TE). If a curve were plotted at the peak amplitudes of several echoes ov er time, the slope of the logarithmic function of this curve would be equal to the T2 relaxation time of the sample. It is possible to produce a series of echoes using a train of spaced 180 pulses; such an experiment would be referred to as being multi ec ho. The TE in a simple spinecho experiment is equal to twice the time interval between the 90 pulse and the 180 pulse (Figure 16). The longer the TE, the greater T2 weighting the data will have. A simple approach to determining the T2 of a sample is t o collect series of data with a sufficiently long TR and a range of TE values (Figure 16). Like T1, samples that are more fluid and aqueous in nature will generally have a longer T2. Biologically, this leads to the observation that areas of fluid accumulation will be conspicuous in T2 weighted images, having higher relative signal intensity at longer echo times than unaffected tissue surrounding them. Carr and Purcell demonstrated that if multiple 180 pulses were applied a train of echoes could be collect ed in a serial
42 manner (86) If the rf pulses applied for the 90 and 180 tip angles are not precise on multiecho experiments, stimulated echos will occur. This can lead to and artificially high value for the calculate d T2 time constant. To correct for this, Meiboom and Gill added an additional inversion pulse to the beginning of the Carr Purcell sequence to produce what is now commonly called the Carr Purcell Meiboom Gill (CPMG) sequence (87) As such the sequence plays out as follows. First a 90 rf pulse is applied and is followed by a time tau, then a pulse less than 180 (ie 175) is applied and then a time twice that of tau is waiting before applying a second 175 rf pulse. This serves to put in place an internal correction for what would likely be an imperfect single 180 pulse and the magnetization vector is rotated exactly into the X Y plane. Such a sequence in spectroscopic studies can be a valuable method to verify the accuracy of in vivo imaging methods of T2 calculation and water/fat content ratios. Spatial localization of the signal. In basic NMR studies, the signal comes from the bulk of the sample that is in range of the RF coil and has no spatial information. The ability to produce and locate spatially unique MR signals is essential to our ability to generate images. This is done through the use of electromagnetic gradients that are installed within the bore of the stat ic magnet. During the setup procedure, the B0 field is carefully shimmed to ensure that the sample experiences a homogeneous magnetic field. Since, according to the Larmor equation, the resonance frequency for all equivalent species of nuclei, for example the protons from the hydrogen atoms of water molecules, will all be the same. Using electromagnetic gradient coils, the local field strength is altered along a known direction using different strength gradients. This in turn creates a range of Larmor frequ encies for a given species of nuclei that are arranged based on their spatial
43 location. These gradients are often played out along the X, Y, and Z axis of the static magnet or laboratory frame. In advanced applications, time varying combinations of these g radients give us multiple axis gradient directions. Slice selection. Any directional gradient may be used as a slice selection gradient. For example, in a clinical body MRI this would be in the direction of head to feet. A range of resonance frequencies corresponding to the location on the sample or patient you wish to get signal from can then be selected by a frequency selective 90 RF pulse in the presence of a gradient, effectively only tipping the spins in the band of sample you are interested in. Short ly after the 90 RF pulse is over, the slice select gradient is turned off and the system returns to the base field strength and corresponding precession frequency of the homogenous B0 field. Phase and frequency encoding. One manner of spatial encoding of the MR signal is by shifting the phase of the of signal contribution of groups of spins in a spatially dependent manner. By applying an electromagnetic gradient at a time after the 90 RF pulse, when MXY is greater than zero, frequency of precession will v ary based on the hydrogen nucleis position. Thus while the phase gradient is on, spins will either wind up faster (therefore a positive phase shift) or they will slow down (a negative phase shift). The amount that they shift depends on how steep of a gradient is being used. Many gradient strength increments are played out to get a single image. Once the phase encoding gradient is turn off, the spins return to precession at the base frequency dictated by B0, but they are no longer in phase with each other. The last dimension needed for making an image is often called the read gradient. This is aptly named since it is turned on during data acquisition. This has the same basic effect, of slowing one
44 side down and speeding the other side up, but this time data are acquired while the gradient is on (spatially encoding this direction with varying frequency). The cumulative effect of applying both of these gradients (phase, frequency) successively is that signal from any point in the plane of those two perpendicular gradient directions is unique; each having a phase and frequency that indentifies it as a unique point in data space. The data space is in the time domain and is called k space. The Fourier transform is then applied and the processed data is transformed from the time domain to spatial encoded information as an MR image. Since each pixel in the calculated image represents a three dimensional space, they are considered pixels with volume and are termed voxels. The contrast of these images can vary greatly based on how the proton spins net magnetization is manipulated prior to and during data acquisition. This is done with extra RF pulses resulting in spin gymnastics and encoding electromagnetic gradients. These basic concepts form the basis of the imaging experiments described in the chapters that follow possible. MRI Contrast Methods. T1 and T2 contrast. T1 contrast can be generated by collecting MR data using relatively short repetition times between RF pulses that would place Z magnetization into the XY plane (typically 90 pulses). By not allowing magnetization to fully recover in the Z axis, tissues with a shorter T1 relaxation will appear brighter (less saturated). Therefore, tissues like adipose will then have a greater signal intensity due to their T1 than regions of edema (long T1). Another contrast method based on fundamental relaxivity properties of the sample is T2 weighting. In biological samples, high fluid retention or high lipid density will both
45 increase the proton density and tend to elongate the T2 relaxation. In those cases, such areas would appear brighter in signal as compared to surrounding areas in T2 weight images (ie TE>T2). This is in contrast to solid tissues, like bone and tendon, which have a much shorter T2 and appear darker than surrounding soft tissues (T2<
46 diffusion gradient executed, but having all other parameters equal, a level of mean diffusion for a given area in a specified direction can be determined. This measure is then repeated in multiple directions to assess natural barriers to diffusion within the sample. These individual images would be said to be diffusing weighted because the contrast will be biased toward levels of diffusion in the sample. Paramet ers that are important in generating this type of contrast are the duration of the diffusion gradient ( ) and the spacing between the dephasing and rephasing gradient ( ), and the slope of those gradients. The duration and spacing of the diffusion gradients are important in determining the diffusion distances that can be detected. The gradient amplitude level and rate of change in the diffusion gradients will be important in dictating the amount of diffusion weighting that occurs in the resulting image. T he power levels of the diffusion gradients, in combination with the duration, are discussed in terms of a bvalue, which has the units of mm/s2. As the b value increases, there is an exponential loss of signal that corresponds to diffusion in a given direc tion. In order to compare samples, several indices of diffusion are calculated. The mean diffusion in all directions is estimated by an apparent diffusion coefficient (ADC). In nonhomogenous samples such as biological tissues this is considered and appar ent diffusion measurement since macromolecules and complex macrostructures hinder truly free diffusion. If the mean diffusion along the direction of the X axis (Dx) and that of DY and DZ (along the Y and Z axis respectively) is calculated, then the ADC is equal to the mean of DX, DY, and DZ. This is shown in the equation below: ADC = ( DX + DY + DZ ) / 3 [Equation 13]
47 In addition to considering the average amount of diffusion in a given region of tissue, researchers can also evaluate the shape of the s pace which allows free diffusion of these molecules. One such index gives us an idea as to how spherical (isotropic) or cylindrical (anisotropic) the compartments containing these diffusing molecules are. Specifically, the term fractional anisotropy (FA) i s elevated if one direction of measured diffusion greatly outweighs the other two. By this definition, molecules in a spherical compartment would have a low FA. Since healthy muscle fibers are cylindrical, it would be expected to have a reasonably elevated FA as compared to many other tissue types. Taking the elements of diffusion imaging one step further, diffusion tensor imaging can be utilized. By collecting diffusion measurements in at least six different directions, and with the option of including mul tiple b values, and an image the has no diffusion weighting (A0), a diffusion tensor and its eigenvectors (L1, L2, and L3) and corresponding eigenvalues ( 1, 2, and 3) can be calculated. In this case the ADC is equivalent to the tensor trace. The orient ation of the eigenvectors for each voxel indicate the dominant direction of diffusion in that region. Magnetization transfer contrast. MRI has been shown to be a valuable tool in the study of these neuromuscular diseases, yet current methodology falls shor t of direc tly measuring tissue fibrosis. This difficulty arises from the extremely short T2 (88 91) associated with collagen. To overcome this investigators have utilized a technique which images the magnetization t ransfer of signal from a bound pool of water to the hi ghly mobile bulk pool of water (see Figure 1 8) (92) Tissues containing high concentrations of hydrated macromolecules with short T2s, such as muscle and cartilage, experience the largest magnetization transfer effect, resulting in larger
48 decreases in signal intensity compared to other tissues (ie fat, blood, water). Guo et al. (2003) suggested that magnetization transfer contrast MRI was sensitive to an increase in tissue fibrosis in a murine model of liver disease (93) They demonstrated that the amount of magnetization transfer (4.7T) was related to the degree of liver fibrosis and the hydroxyproline content. Hydroxyproline is a modified amino acid that is almost exclusively found in collagen, making it quantitative biochemical meas ure of tissue collagen content (94, 95) Magnetization transfer (MT) was first measured in skeletal muscle tissue by W olff and Balaban (1989) and it was estimated to be three times more efficient than MT in the kidney (93) MRI of Skeletal Muscle. Magnetic resonance is particularly suited for the study of dystrophic muscle due to its high soft tissue contrast in imaging combined with the ability to collect biochemical data from localized regions (10) ; thus giving detailed information about metabolism (96, 97) and tissue composition (98, 99) Many MR parameters have been investigated in various models of muscle damage and disease, including muscular dystrophy (100, 101) MRI of muscle tissue was described in early publications relied heavily on morphological differences in images to distinguish between healthy and diseased muscles (102) Later it became routine to evaluate inflammatory myopathies with contrast methods such as T1weighting or T2weighting, with or without short tau inversion (STIR) sequences to suppress the signal derived from fat. These methods have more recently become common practice for the evaluation of numerous inherited neuromuscular disorders (101) During the 1980s, following studies in rat muscle by Le Rumeur et al. (103) Fleckenstein et al. demonstrated an increase in T2 relaxation time was associated with
49 specific muscle activation and exercise after having human subjects perform submaximal contractions of their forearm muscles (10 0) Since then many reports have been made suggesting the T2 weighted imaging is sensitive to activation and edema in skeletal muscle tissue. More recently, Frimel et al. demonstrated a correlation between the amount of Evans blue dye positive staining m uscle fibers and an increased T2 relaxation in mouse soleus muscles after a weight reloading induced injury (15) thus providing strong evidence for a connection between increased T2 relaxation and compromised sarcolemmal integrity. In proton MRI, the predominant source of the induced signal is from tissue water and lipid molecules. Due to the structural compartmentalization of water in different macro and micro scale features, its is understandable that water in these different environments may have different MR relaxation properties, as well as other parameters discussed later, such as restric tion to free diffusion and the transfer of magnetization. While it is commonly recognized that T2 relaxation times increase after injury and/or activation, the exact mechanisms remain unclear. Although the transverse relaxation times of healthy muscle tiss ue are often reported as being from a single source (monoexponential), sharing a common decay rate, it has been demonstrated that the signal decay curve (signal intensity vs. echo time) can be deconvolved and is better fit to a multiexponential (104, 105) ; assuming the original data is high enough quality (106) It is also likely that with different types of damage and differing pathologies associated with the various myopathies, that contributio ns of these different water compartments maybe be in unique proportions depending on the specific state of the muscle tissue. Effort has been undertaken within the research community to indentify the anatomical or
50 physiological origins of these specific co mponents (discussed in detail in Chapter 4), but such assignments have proven difficult, due to the complexity of the tissue and numerous approaches to modeling the components, the exact number of components still remain somewhat controversial (105, 107) Despite the difficulty in these assignments, multi component analysis of muscle T2 relaxation still may provide valuable insight to the plasticity of muscle tissue following injury or during disease. Another growing area of study in muscle MRI utilizes diffusion weighted imaging to study the organization, orientation, and other physical attributes, such as fiber diameter. This type of data has an increasing set of applications that range from describing the overall order of the spacing and orientation of myofibers, as well as complex uses like three dimensional fiber tracking. Several studies have reported the mean diffusivity (ADC) and fractional anisotropy (FA) as indicators of damage and repair in skeletal muscle (108) Three dimensional tracking of the myoarchecture in complex regions of the body, such as the tongue (109 111) and uterine wall (112) has proven to be a valuable tool in better understanding fiber arrangement and function in these tissues. Numerous other studies have applied diffusion imaging to cardiac muscle to look at both structure in both damaged (113) and healthy (114, 115) conditions. In addition to resolving structures of complex organized tissue, like those examples above, diffusion tractography has also been used to investigate the skeletal muscles of the limbs in humans and in animal models (116 118) An additional line of research has shown magnetization transfer contrast to be a sensitive tool in studies of skeletal muscle. Decreased muscle MT has previously been shown to occur under conditions of large fluid shifts. Yoshioka et al. 1994 showed an
51 inverse correlation of MT in muscle with free water content in exercised muscle (119) Also Mattila et al. (1995) found that following muscle damage in a rodent model that there are large acute changes in MT that could not be associated with fatty tissue infiltration over the time scale studied (120) Acute muscle damage is associated with muscle fiber swelling, edema, cell necrosis and regeneration. It has also been suggested that other factors such as destruction of the large protein complexes, damaged cell membranes, and infiltrati on of inflammat ory cells could also reduce MT (121) Vahlensieck et al. (1999) used MT imaging to look at intramuscular tumors and observed that muscular scar tissue had a lower MTR than healthy muscle, while both had MTR values that were significantly higher than that of the tumors (122) This supports the possibility that despite collagen having a high MTR relative to most body tissues and a increase in its concentration with progressive fibrosis, that due to the higher MTR of healthy muscle, an increase in fibrosis in musc ular dystrophy could result in a decrease in the MTR that is independent of an increase in water content (edema) or lipid deposition. MRI of muscular dystrophy. Most MR studies of dystrophic muscle have centered on the fundamental MR parameters of muscle tissue and lipid, specifically proton longitudinal (T1) (123, 124) and transverse (T2) relaxation (125) or morphological observations in specific case studies. Currently, much of the MR imaging of dystrophic tissue in vivo is carried out based on T1 contrast. This method is particularly useful in assessing volumetric information and resolving areas of high lipid infiltration. Unfortunately, acute dystrophic lesions, due to muscle damage prior to fatty tissue infiltration, fail to be visible on these T1 weighted i mages at higher field strengths (126, 127) On the other hand, acute
52 muscle damage (128) and inflammation (129, 130) can be easily detected based on T2 values up to 17.6T. McIntosh et al. noted that dystrophic lesions are easily identifiable on T2weighted images, but they did not look at quantitative measures of the transverse relaxation (125) While several publications describe diffusion imaging of healthy and diseased muscle, none that we are aware of have addressed diffusion parameters in dystrophic muscl e. All of the current articles on diffusion MRI of subjects with muscular dystrophy were focused on imaging the brain and were cases of myotonic or congenital MD (131) A few studies have looked at magnetization tr ansfer in dystrophic muscle. Fatty tissue deposition has complicated the interpretation of MT results in human subjects with muscular dystrophy. For instance, Schick, 1996 et al. found that when using a water selective imaging sequence that that there was no difference in MT between affected and unaffected muscles in three patients with Erb muscular dystrophy (132) In contrast, McDaniel et al. 1999, not only found that in subjects with LGMD that MT was were dramati cally reduced in muscles with gross fatty infiltration but MT also was also reduced in muscle tissues without visual evidence of fatty infiltration (133) Overall fatty tissue infiltration is anticipated to be lesser a problem in murine models of dystrophy. Mouse models of muscular dystrophy reflect many of the hallmarks of human dystrophies with muscle fiber damage and regeneration, yet the amount of fatty tissue infiltration is not as severe (134) Thus we intend ed to use MT imaging as a metho d of detecting progressive fibrosis and muscle tissue damage in mouse models of muscular dystrophy.
53 Summary The muscular dystrophies represent a devastating group of neuromuscular disorders that greatly needs clinical solutions. Research on putative treat ments is moving ahead using traditional techniques, but there is a tremendous demand for more efficiency in the process of evaluation of potential treatments. Noninvasive imaging modalities like MRI can provide this much needed rapid feedback and in the process give a much clearer picture of the overall tissue with specific biochemical and anatomical data. While parameters such as T2, indices of diffusion, and magnetization transfer have all been demonstrated to reliably detect global changes in muscle com position and integrity, there is much room for refining the application and analysis of these imaging methodologies in order to produce a detailed assessment of the underlying pathological source o f these observable changes in MR parameters A multi parameter approach, in which a series of such MR imaging parameters are observed within one session, would allow the consolidation of several biochemically and structurally relevant measures into a unique profile that would characterize the status of the muscle tissue. Therefore I will use a combination of multi component T2, DTI and MT imaging to study dystrophic muscle.
54 Figure 11 Bout of damage and repair drive disease progression in muscular dystrophy. In Duchenne muscular dystrophy, a) the absence of of dystrophin results in a loss of sarcolemmal integrity. As a result, b) these muscle fibers will be damaged as the sarcolemma is compromised. The damaged cell will signal for the recruitment of satellite cells, which will home in on these chemical signal s and then repair the damaged cell or they will c) form new myofibers. Since dystrophin is absent in these fibers, the newly repaired myo fiber s are likely to undergo damage again. T his process continues to repeat until the tissues ability to completely repair i tself is d) eventually lost and the functional muscle fibers are replaced by fat and scar tissue.
55 Figure 12 Mouse models of Duchenne (DMD)/Beckers muscular dystrophy and limbgirdle muscular dystrophy (LGMD) compared to a health C57/B10 mouse.
56 Figure 13 Limitations of classical invasive measures. The golden standard in pathology of dystrophic muscle has been invasive methods such as biopsy and histology of excised muscle from animal models. This image depicts typical damage observed on a dys trophic GAS muscle in a mouse model of LMGD ( sg/ -) and it is very similar to what is seen in mdx mice. Note the difference in disease assessment between two hypothetical biopsy locations A) a seemingly unaffected region and B) within a dystrophic lesion. This example shows the extensive muscle dam age and fatty tissue deposition sg/ mouse.
57 Figure 14 Spin echo sequence. Above is a simplified diagram of a spin echo sequence illustrating the spacing of the 90 and 180 radio frequency pulses, the free induction decay (FID) and the echo. The echo time (TE) is the time between the 90 pulse and the echo and the repetition time (TR) is the time interval in between successive 90 pulses. In a spin echo experiment, the echo of the original FID is the signal that is ac quired and processed.
58 Figure 15 Longitudinal relaxation: T1 recovery. Following a 90 rf pulse the Mz magnetization is reduced to zero and recovers at an exponential rate determined by the T1 of the sample. As the time between successive 90 pulses, the repetition time (TR), is increased more Z magnetization is allowed to recover. This results in a predictive increase in signal intensity as the TR increases in length. In order to avoid T1 weighting in collected data, a TR of 5 to 6 times the length of the samples T1 should be chosen. If the experimenters aim is to achieve T1 weighting, shorter TR times are utilized.
59 Figure 16 Transverse relaxation: T2. Following a 90 rf pulse a coherent signal is produced from the procession of magnetization i n the XY (or transverse) plane. Interactions with neighboring nuclear spins results in the loss of the ensemble giving rise to processing MXY vector. These spinspin interactions lead to an exponential decay of the signal acquired. As the amount of time between the 90 rf pulse and the time of acquisition of an echo, (or echo time; TE) increases, the loss in signal intensity can be seen in the plot above. The rate of this exponential decay of signal is determined by the T2 time constant of the sample bein g measured.
60 Figure 17 A diffusion weighted spin echo sequence.
61 Figure 18 A schematic of a simplified magnetization transfer spin echo sequence.
62 CHAPTER 2 OUTLINE OF EXPERIMEN TS Overview The ultimate objective of this work was to establish a non invasive multimodal MR imaging protocol to monitor disease progression and therapeutic correction in dystrophic muscle tissue using magnetic resonance (MR) imaging. The ability to detect dystrophic pathology and its correction was investigated using t hr ee MRI contrasts methods The first aim was to assess sarcolemmal integrity (Figure 2 1) by measuring the T2 relaxation times of the muscle tissue. The second aim was to detect structural remodeling of dystrophic muscle using diffusion weighted MR im aging (DW I and DTI). The third aim was to determine the magnetization transfer contrast (MT C) and its relationship to fibrosis and damage in dystrophic muscle. The ability of all three MR modalities to monitor disease progression was performed by stratifying mic e into both young and old age cohorts. Several mouse models of muscular dystrophy (LGMD and DMD) were compared to healthy control mice. In the dystrophic mouse models, widespread acute damage was expected in the younger animals (42) while the chronic effects of advanced disease (i.e. increased fibrosis and fat infiltration) were not yet expected. The repair pathways in these younger (4 8 month old) mice is in constant flux and recovery of damage is similar to what has been observed in acute models of damage in heathy control animals (135) In the aged dystrophic mice ( 1 2 years old) the acute lesions are less common, but extensive tissue remodeling and histopathology was expected (136) In order to assess the ability of T2, DTI, and MT to de tect therapeutic correction mouse model s of limb girdle muscular dystrophy ( sgcg/ and sgca/ -), were treated with a recombinant adenoassociated vir us ( r AAV) based gene
63 therapy vector s which lead to mus cle specific expression of sg sg and were imaged using the three methods under investigation. Global h ypothesis. Changes in the biophysical properties of bulk water can be used to monitor changes in muscle fiber integrity in dystrophic skeletal muscle. Experiment 1: T2 as a Marker for Sarcolemmal Damage in Dystrophy Muscle. T2 weighted MR images of skeletal muscle have been shown to be sensitive to acute injury (15) and exercise induced contrast enhancement (100, 137) Elevated T2 values may be due to an increase in tissue fluid content or decreased muscle integrity (15, 137) To investigate the changes in global T2 as a function of longterm dystrophi c disease progression, I determine d the age dependent changes in T2 of lower hind limbs in two models of muscul ar dystrophy and in healthy age matched controls. In order to explore the temporal evolution of changes in T2 following injury, I utilized an eccentric damage protocol that involved downhill treadmill running. This had the effect of causing a synchronized bout of damage and provided a comparison of T2 at various stages of injury and repair. In addition to the natural history, I compared T2 of corrected and uncorrected dystrophic muscle with histological indices (83, 1 05) of mus cle damage and gene expression. It is has also been well established that the MR signal from muscle consists of multiple T2 values (105) that can be ascribed to different cellular compartments. I set out to determine whether the heterogeneity in T2 observed within in vivo normal, dystrophic and corrected dystrophic muscle correlates with histological indices of muscle damage, permeability, susce ptibility to injury, and edema. Finally I used nonnegative least squares (NNLS) analysis (105) of multiple spin echo imaging
64 data collected on in vivo muscle to stu dy changes of these multiple T2 components in muscle damaged in a physiologically relevant manner Hypothesis 1. T2 weighted magnetic resonance imaging (MRI) detects muscle damage and gene correction in dystrophic muscle. Specific Aim 1 The primary aim of the first set of experiments was to determine the ability of T2 imaging to detect muscle damage and monitor gene correction in dystrophic muscle. In testing this aim I specifically sought out to determine if: a) changes in muscle T2 are related to muscl e damage in muscular dystrophy, if (b) d amaged muscle can be further characterized by multiple T2 components, and if (c) changes in muscle T2 during the time course directly following injury, reflect structural remodeling and repair of muscle tissue. Exper iment 2: Diffusion Imaging to Monitor Recovery From Damage. Histological comparisons of dystrophic and healthy muscle reveal that there is a much greater variability in fiber size and that the presence fiber splitting (136) is found in the muscles with dystrophy. Pathological alterations such as these could potentially be detected by utilizing diffusion weighted imaging (DWI) (138) From these experiments I calculate d local tissue diffusivity and fractional anisotropy Tissue regions containing f ibers with sarcolemmal damage, as well as edematous tissue, were distinguishable from lipid depo sition in areas with elongated T2 relaxation times, by compar ison of the apparent diffusion coeffic ients (ADC). The ADC data combined with T2 and MTR later provided a multi modal approach to noninvasively characterize and track the stages of disease progression and therapeutic intervention.
65 Hypothesis 2 The diffusion of tissue water and T2 is altered in dystrophic muscle and related to structural changes following acute damage. Specific Aim 2. In this second set of experiments, I set out to d etermine if the diffusion of tissue water and T2 relaxation is altered in dystrophic muscle and related to structural changes following acute damage. In exploring this aim I: a) c ompared the apparent diffusion coefficient (ADC) of healthy and dystrophic muscle, before, after, and during recovery from acute damage and (b) followed c hanges in ADC T2, FA and primary eigenvectors to determine if they reflect structural remodeling and repair of muscle tissue. Experiment 3: Magnetization Transfer Imaging to Assess Fibrosis. The build up of collagenous scar tissue between myofibers is a significant contributor to the progressive debilitating effects of muscular dystrophy (139) Not only does this fibrosis replace functional muscle tissue, but it also creates an environment that diverts muscle progenitor cells away from myofiber formation (140) Unfortunately, fibrosis can be very difficult to directly MR image due to its extremely short T2 relaxation times (88) It has been shown that magnetization transfer contrast (MTC) allows the indirect imaging of fibrosis by looking at the proton magnetization transfer (MT) between a free l iquid pool and a macromolecular bound pool (90, 92) of which collagen is a major component in fibrosis (93) In the current study, the effects of age and disease progression in the lower limb muscles of healthy dystrophic and therapeutically corrected dystrophic muscle on magnetization transfer were deter mined. Hypothesis 3. Magnetization transfer is altered in dystrophic muscle and related to disease progression.
66 Specific Aim 3 In the third set of experiments, I set out to d etermine the ability of magnetization transfer imaging to detect muscle damage, fibrosis and gene correction in dystrophic muscle. To investigate this I: a) initially determined the effect of age and disease progression on the magnetization transfer ratio (MTR) and (b) determine the correlation between fibrosis, muscle damage and changes in MTR in muscle tissue.
67 Figure 21 An overview of dystrophic disease progression and the specific aims of this researc h.
68 CHAPTER 3 METHODOLOGY Animals Rodent Handling and Care. This study was conducted with the approval from the University of Florida institutional animal care and use committee. Mice were fed ad libitum and were housed in an AALAC accredited animal facility in a temperature (221C), humidity (5010%), and light (12 h light/dark cycle) controlled room. For age dependent studies, mice were stratified into two cohorts by age and by strain ( Figure 31). Mouse strains included healthy control C57BL/10 (C57BL/10ScN J) mdx a model strain for DMD (C57BL/10ScSnDmdmdx/J), a model strain for LGMD 2D ( sgca/ -) and a model strain for LGMD 2C (sgc g/ -) and are described below Mouse Strains Control C57BL/10 mice. The healthy control mice used in this study were from the strain C57BL/10, specifically C57BL/10ScN J and were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). These mice hav e a deletion leading to the loss of a lipopolysaccharide (LPS) response ( Tir4lps del), but this phenotype has no effect on the health of skeletal muscle and is therefore not relevant to this work. C57BL/10 mice served as healthy subjects and generally live d to 2 years of age if they were not sacrificed at an earlier time point. mdx mice. The mdx (C57BL/10ScSnDmdmdx/J), a Duchenne/Becker muscular dystrophy model, mice were initially obtained from The Jackson Laboratory (Bar Harbor, ME, USA) and were breed i n house to maintain the colony. The background strain for mdx mice are the C57/BL10ScSn colony, which is also currently maintained and
69 supplied by The Jackson Laboratory The mdx strain resulted from selective breeding of a spontaneous mutation that occurr ed in a C57BL/10ScSn colony and was first described as a model of DMD and BMD in 1984 (41) The mdx mouse has a mutation in exon 23 of the dystrophin gene. The dystrophic phenotype is considered mild compared to the human disease and these mice often live to 2 years of age, but exhibit limited ambulation and kyphosis in lat e stages. In early stages the mice look healthy, but the muscle tissue shows signs of widespread muscle damage due to the lack of dystrophin. Sgca/ mice. sarcoglycan null mice ( sgca/ -) are a model of the human disease, limb girdle muscular dystr ophy type 2D (LGMD2D). The limbgirdle muscular dystrophies arise from mutations in the various sarcoglycan genes that encode proteins that function as subunits in the dystrophinassociated glycoprotein complex. Even with dystrophin present, without one of these subunits, the remaining complex often disassociates and the function is lost. This leads to a generally more mild form of muscular dystrophy that is autosomal recessive. The sgca/ mice are derived from the background strain C57BL/6 and display a dystrophic phenotype early in life. While there are wide spread lesions in skeletal muscle, the mice are still able to live long lives; up to approximately 2 years of age. Sgcg/ mice. sarcoglycan null mice ( sgcg/ -) are a model of the human disease, limb girdle muscular dystrophy type 2C (LGMD2C). The sgcg/ -mice are similar to the sgcg/ in severity and there longevity. The sgcg/ strain has also been shown to develop cardiomyopathy and extensive muscle tissue fibrosis. The sgcg/ mice are chimer as that were derived genetically modified RW4 ES, which were injected into C57BL/6 blastocysts, and were then implanted into a pseudopregnant mother. The
70 RW4 ES cells were originally from the 129/SvJ strain of mouse, which have light colored coats. Thus m ale chimeras could be distinguished by their light color. These mice were reportedly (141) back crossed to the C57BL/6 strain and the phenotype is typically monitored by the presence of a light coat color. The sgcg/ strain was a gift to our laboratory from Elisabeth McNally and Lee Sweeney and our colony was maintained within our house breeding protocol. Gene D elivery to Y oung LGMD Mice. Recombinant adenoassociated viral (rAAV) vectors bearing corrective genes wer e delivered to the lower limb muscles of both sarcoglycan and sarcoglycan null mice as previously reported (55, 56) Gene delivery to s gca/ mice. In the case of treatment of sarcoglycan null mice, oneday o ld sgca neonates ( n = 6) were injected with 1x1011 vg of rAAV2/1 tMCK sgca 11 vg of rAAV2/1 tMCK LacZ in the contralateral hindlimb (single injection per leg) (56) As the vector was injected, the needle was carefully backed out of the port of entry, from under the knee toward the ankle, in order to evenly bathe the muscles and insure maximum delivery. Initially there was concern over the possibility that the injection alone would elicit a damage response, but this did not seam to be the case. In a previous study, when the same rAAV2/1 tMCK LacZ injected sgca/ muscles were compared with uninjected, untreated sgca muscles, no difference was observed in the two negative control groups (56) Gene delivery to s gcg/ mice. Gene delivery to the right leg of sgcg/ mice was achieved by IM injecting 3 week old mice (see Figure 32) with a muscle specific
71 recombinant adenoassociate d virus (1x1010 vg of rAAV2/8 Desmin sgcg ; the generous gift of Elisabeth Barton; University of Pennsylvania) which expresses the human form of the missing sarcoglycan ( sgcg ). The contralateral limb was not treated and served as a control. Vector diluted in phosphatebuffered saline (total volume of 100 l per injection) was directly injected into the lower hindlimb using a 29 1/2 G tuberc ulin syringe. The needle was inserted near the distal Achilles tendon and pointed upwards into the posterior compartme nt ( GAS and SOL ) V irus solution was injected while withdrawing the needle to maximize volume distribution across the lower limb. Expression of sarcoglycan was confirmed using immunohistological techniques including Western blotting and immunofluorescent microscopy (described below) MR imaging of both sgcg/ hindlimbs was performed at 7 weeks of age (4 weeks post injection) and one year post injection. Exercise Induced Eccentric Damage: Downhill Treadmill Running. Some MRI methods, such as diffusionw eighted imaging, can be very sensitive to the time course of injury and recovery in skeletal muscle (78) Since the l esions in dystrophic muscle are occurring at random times, it was important to develop a protocol that would induce mild damage to the dystrophic muscle synchronized to a known time course and in a physiologically relevant way. Susceptibility to eccentric contraction induced damage has been demonstrated in both murine (142) and canine (143) models of muscular dystrophy. Since distal limb muscles of humans with muscular dystrophy are also at risk for such injuries (144) results from studies of animal models acquiring and recovering from eccentric induced damage may prove to be more relevant than other damage models. Therefore 6 mdx and 6 C57BL/10 mice were exercise on a
72 treadmill that was positioned with a declination angle of 14 ( 24.9% grade), which was slightly less step than similar protocols reported in the literature (mean declination angle of 16 with a standard deviation of 1 in studies with a similar speed) (145147) The mice were run in individual lanes, were supervised and were encouraged, with gentle manual guidance, to run until exhaustion. At a speed of 8 10 m/min, this generally took 20 30 minutes, but a few ran as short as 10 min. A fter leaving the treadmill the mdx mice displayed outward signs of fatigue (heavy breathing and limited mobility) or until they ran for a duration of 45 minutes. For this study, imaging data were collected before running and then subsequently at time 0, day 5, day and day 10 post treadmill running. The protocol was stressful enough to elicit mild damage in the mdx mice, but did not damage the control C57 lower limb muscles. Magnetic Resonance Imaging Determination of Muscle T2. Q uantitative T2 imaging was i mplement ed to monitor muscle injury/regeneration. During all in vivo MR experiments animals were anesthetized using gaseous isoflurane (3% induction, 0.5 2.5% maintenance). Both hindlimbs were will imaged simultaneously using a custom built five turn, 1.5 cm i.d. single tuned 1H solenoid coil (200MHz) on a 4.7T Bruker Avance (Rheinstetten, Germany) horizontal bore spectrometer (Figure 3 3) To determine transverse relaxation rates (T2), multiple slice, single spin echo, diffusioncontrolled images were acqu ired with the following parameters: FOV=1x1 cm2, matrix=256x128, slices=12, slice thickness=1mm, slice gap=1mm, diffusion weighting=3 mm2/s, NEX=2 and TR=2 s. To avoid the contribution of stimulated echoes to the T2 measurement, a Hahn spin echo MR image sequence in which two separate acquisitions were acquired at echo times of 14 ms and 40 ms was implemented. All T2
73 images were analyzed using inhouse software running in the Interactive Data Language programming environment (IDL version 7.0, ITT Corp. ; CO ). Initially no attempt was made to determine the multicomponent characteristics of muscle T2. The T2 values were calculated assuming a single exponential decay with respect to TE as previously described using the equation below (41) [Equation 31] A pixel by pixel T2 map was calculated for each axial MRI Pixels with T2 values two standard deviations above the average T2 value of control (C57BL/10) muscles were considered to be significantly elevated and affected by dystrophi c lesions Subsequently the mean T2 value of the muscles of interest ( TA and GAS) was determined, as well as the mean T2 of affec ted verses unaffected regions. M uscle groups and affected/ unaffected muscle regions were manually outlined in five image slices and the mean muscle T2 values and the total number of pixels that were elevated were recorded. De termination of Mean ADC at 4.7T. D iffusion weighted imaging (DWI) was implemented to detect changes in muscle structure. As in the previous experiments, duri ng all in vivo MR experiments, animals were anesthetized using gaseous isoflurane (3% induction, 0.5 2.5% maintenance). Both hindlimbs were imaged simultaneously using a fiveturn, 1.5 cm single tuned 1H solenoid coil (200MHz) and a 4.7T Bruker Avance (Rheinstetten, Germany) horizontal bore spectrometer. Data reflecting mea n diffusion in the longitudinal direction (parallel to net muscle fiber orientation) was acquired with a low and then a high diffusion gradient
74 strength ( b value). This was followed by a separate acquisition of diffusion weighting in the axial direction to net fiber orientation (coinciding the direction perpendicular to mean fiber orientation). To determine an apparent diffusion coefficient (ADC) in the axial (with respect to the limb) and longitudinal (along the length of the limb) direction, multi slice, single spin echo, diffusionweighted images were acquired with the following parameters: FOV=1.6x1.6 cm2, matrix=256x128, slices=12, slice thickness=1mm, slice gap=1 mm, s, NEX=3 and TE/TR = 22/ 2 000 ms. In the direction of diffusion approximately perpendicular to muscle fiber orientation, the diffusion weighting (b) for the low diffusion weighted images was b=252.964 s mm 2, while it was b=912.497 smm 2 for the high diffusion weighted images. In the direction of diffusion approximately parallel to muscle fiber orientation, the diffusion weighting for the low diffusion weighted images was b=260.917 s mm 2 and b=927.679 s mm 2 for the high diffusion weighted images. T he set of low and high bvalues was different in the two directions due to differences in the strength of gradient coils in those directions. The low b value were 50.90% of the maximal gradient amplitude, while the high bvalues were 97.184% of maximal gradient amplitude. All DWI data was analyzed using inhouse software running in the Interactive Data Language programming environment (IDL version 7.0, ITT Corp.). An ADC, pertaining to a single direction, was calculated assuming a single exponential deca y o f signal with respect to the b value of the diffusion gradient pulse sequence and represent ed the linear spatial displacement of water molecules within the diff created based on pixel changes in MR signal intensity The mean ADCSlice and ADCRead values of the muscles of interest ( TA and GAS) was determined, as well as the mean ADCSlice/Read of affected
75 verses unaffected regions. Muscle groups and affected/ u naffected muscle regions were outlined in five image slices and the mean A DCSlice/Read was recorded. Dete rmination of ADC and FA at 11T. Within 72 hours of collection of DWI at 4.7T, a DTI data set was collected at 11.1 T using diffusion gradients capable of larger diffusion weighting (b) and shorter echo times The while the in crease in field strength at 11.1T slightly shortens T2, the large increase in SNR was a welcome benefit for the DTI data. But more important was the ability to use the S057 (300G/cm) gradients. The high power gradient coils allowed the application of a steeper diffusion sensitive gradient which made it possible to minimize the TE to 12.8ms; reducing unwanted T2 additional weighting (loss of signal from muscle). During all in vivo MR experiments animals were anesthetized using gaseous isoflurane (3% inductio n, 0.52.5% maintenance). Hindlimbs were imaged using a custom built 1 cm i.d. single tuned 1H loop gap coil (470MHz) and an 11 .1 T Magnex superconduction magnet coupled with a Bruker Avance (Rheinstetten, Germany) horizontal bore spectrometer (PV3) Optima were determined empirically starting with values published in the current literature (see Figure 34) These data were collected with diffusion gradients applied with low b values ( b = 1 00 smm2) and with high b valu es ( b =900 smm2) in 6 directions and one scan with no diffusion weighting (A0). In pilot studies, a large range of b values were measured and 900 smm2 was chosen as the upper level of diffusion weighting as it gave best tissue contrast for damaged versu s healthy muscle while retaining an adequate SNR. An example of a diffusion gradient power curve is shown in Figure 34, along with sample diffusion weighted images. The lower b value of 1 00 smm2 was chosen based on values from the literature (148) The se data were proce ssed using a
76 software program running in IDL (IDL version 7.0, ITT Corp.). From the diagonal diffusion elements (Dxx, Dyy, Dzz)), the eigenvectors (e1, e2, e3) and the corresponding 123) were calculated. The indices ADC and FA were determined, from which image maps were generated allowing measurement of specific regions of interest (TA and GAS ) using the Paravision JIVE (Bruker Avance) software package. Statistical analys es of mean ADC were carried out to determine differences due to age and disease progression. Great lengths were taken to ensure that there was no significant fluctuation in body temperature once the mice were in the bore of the magnet. Several approaches were tested and the most consistent solution was to circulate heated water through a custom built heating block that was installed within the cradle that holds the animal and RF coil (Figure 3 5). The core temperature of the animal was monitored using a magnet safe SAII temperature probe and the SA Monitoring system (S mall Animal Instruments. Inc, Stony Brook, NY, USA). Temperature of the water bath was manually controlled so that the animals were maintained at 35C to avoid possible changes in diffusion due to temperature and changes in perfusion due to lowered body te mperature. Determination of MTR: MT MRI was acquired with a similar Hahn spinecho sequence as described in the T2 imaging experiments described in aim 1, but only consist ed of a single slice. Specifically images were acquired with the following parameter s: one set with a fully attenuated MT pulse (150 dB) and one with an MT pulse reference gain of 40dB (roughly twice the power of the 90 pulse) MT frequency o ff set equal to 10 kHz, TE/TR=11.3/2000ms, Matrix=256x128, FOV=1.6x1.6cm, and NEX= 1. The power level of MT RF pulse at 40dB attenuation was chosen because this power level gave optimal
77 contrast between damaged and healthy muscle tissue and was well within the SAR limits that we defined. To define the pulse power ranges that were within our SAR limitat ions I prepared a 3% agarose (in PBS) phantom in a 1.5 conical eppendorf tube and placed an optical temperature probe in the center of the phantom. The phantom was then used to simulate the size and proton density of the animals hindlimb and a series of MT C images were collected with increasing power until the phantoms temperature was increased by 1C; the result was a block 32 pulse with 33 dB attenuation of a 1 kW transmitter. Then one decibel of attenuation higher was defined as the maximum limit that could be applied to an in vivo sample of a similar size and load on the coil. A series of MTC images were then collected on healthy control mice and 40 dB attenuation was found to provide optimal image contrast in the muscle tissue and was well below half that power of our maximum SAR limit. Magnetization transfer ratios (MTR) were calculated using Equation 3 2 with M0 representing the scan with no MT pulse and MSat representing the image with the MT pulse on. [Equation 32 ] First, a MTC ma p was created based on individual pixel MSat / M0 values. Subsequently, the mean MTR value was determined from the same ROI as in the T2 maps. Mean MTR values from regions of interest in healthy, dystrophy and corrected dystrophic muscle w ere then compared using statistical analysis. Measurement of L ongitudinal R elaxation (T1). Since T1 relaxation can be a dominant term in quantitative measures of MTR the l ongitudinal relaxation rates ( T1) of dys trophic and control muscles were determined by
78 progressive s aturation. Single s lice, trans axial images were acquired with a variable TR sequence with TE=7ms, FOV=2.4x 1.8cm2 matrix=128x64, slice thickness=3mm, NEX=1 and TRs=6, 3, 1.5, 0.75, 0.325, and 0.2s. T1 images were analyzed using P aravision 3.0.2 (Bruker Xtip) and T1 values were calculated assuming a single exponential function. Specifically, the mean T1 of the TA and GAS was calculated by fitting the ROI signal intensity of each muscle as a function of TR (see Figure 36) Histology Histological V erification of Sarcolemmal Damage. Following completion of the MR experiments, all animals were euthanized and muscle tissue was collected for histological verification of the extent and regions of damage. For the detection of muscle fibers that have sarcolemmal damage Evans Blue Dye (EDB ; Sigma) was systemically delivered to the mice following their last MRI and 24hrs prior to tissue harvest. The EBD was dissolved in phosphate buffered saline (0.15M NaCl, 10mM phosphate buffer, pH 7.4) at a concentration o f 1 m g/0.1 ml/10 g body weight (41, 91) The EBD solution was then filter ; Nalgene) prior to intraperitoneal injection. Within 24 hours the animal was sacrificed and the TA and GAS muscles were dissected from both hindlimbs and visualized for macroscopic accumulation of EBD in fibers and muscles. The muscles were gently stretched to resting length by dissection pins supported by dense styrofoam coated in O.C.T. gel (Tissue Tech), rapidly frozen in melting isopentane and stored at -80C. Frozen muscles were subsequently cut in half and thin sections were taken at the belly of the muscle. Frozen sections (10 m) were stai ned with Hematoxylin & Eosin and Oil Red O. Slide mounted sections were visualized and digitized under
79 brightfield at 5x and 20x magnification on a DM LB microscope (Lei ca Microsystems, Solms, Germany). Histological Quantification of Fibrosis. Following completion of the MR experiments, all animals were euthanized. The tibialis anterior and gastrocnemius muscles were dissected from both hindlimbs. The muscles were set at resting length using pins coated in O.C.T. gel (TissueTech) and rapidly frozen in melting isopentane and stored at -80C. Frozen muscles were subsequently cut in half and sectioned at the belly of the muscle. Frozen sections (10 m) were either stained using Masson's Trichrome Stain Kit (RichardAllan Scientific, procedure number 010, catalogue number 87010) or Hematox ylin and Eosin. Sections were visualized and digitized under brightfield at 5x and 20x magnification on a DM LB microscope (Leica Microsyst ems, Solms, Germany). Digital micrograph images of the trichrome stained tissue were analyzed using the software program ImageJ ( http://rsbweb.nih.gov/ij/ ) ROIs, containing the ent ire muscle cross section, wer e manually selected and the percentage of posi tive collagen stai ning tissue (based on area) wer e calculated using a HueSaturation Intensity color model for prean post thresholding measures (H:146206, S:0 255, I:0 255; Pass Filter). In this assay, 6 muscles from each group were used and the results were then statistically analyzed using a T test with P>0.05. Verification of transgene expression in gene delivery experiments. In order determine if the rAAV vectors were effective immunohistochemistry assays were conducted on excised muscle tissue. The restoration of the missing protein by viral gene delivery was assessed both by Western Blotting and immunofluorescence
80 microscopy of fixed tissue thin sections. The methods and validation results are reported below. Immunofluorescence. Immunofluorescence ( Figure 37 A) of rAAV hSG can be seen panels to the far right (treated animals), while it is absent in the center column (untreated sgcg/ -); wt healthy control muscle from a C57BL/10 mouse is shown in the left column of Figure 3 7 A. The transgenic human sarcoglycan is labeled green. There appears to be a slight amount of nonspecific binding around areas of excessive extracellular matrix on the sgcg/ tissue, but this is far less than was observed in the treated and positive control animals. The primary sarcoglycan was generated in a mouse (and was the generous gift of Dr. Elizabeth BartonDavis, Pennsylvania, USA) and the secondary was goat anti mouse IgG Alexafluor488. In order to label the peripheral boundry of the fibers, the muscl e thin sections were co labeled with rabbit anti laminin primary antibodies and goat anti rabbit IgG Rhodamine. Thus laminin is displayed in red in the top row of merged images. The labeling was so strong that a second series of merged image without lamini n is presented in the lower row so that the distrubution of sarcoglycan can be seen more clearly. Finally, DAPI was mixed into the mounting media to label nuclei, which are labeled blue. Western Blot. For the Western Blot (Figure 3 7 B) 15ug of protein per lane (using 1ug/ul, a volume 15ul of sample per lane) w as run denatured on a SDS PAGE apparatus and then subsequently blot transferred onto a nitrocellulose membrane. SG is 35 kDa and the GAPDH (used as an internal control protein concentration) is 36 kDa, the membrane was cut in half to use two differ ent primary SG primary and 1:5,000 of the GAPDH primary was
81 used. A horseradish peroxidase chemiluminescence kit (BioRad, Hercules, CA, USA) was used to autoexpose x ray film as a method of band detection. Note that t SG (35 kDa) is missing in the sgcg/ lane.
82 Figure 31 Mouse age cohorts and experimental design. Group selection for ( in vivo ) age and strain dependant studies of global changes in MR sensitive parameters (T2, ADC, and MTC). The lower hind limb muscles of six young (age: 25months) and six old (age: 1824 months) control C57BL/10SnJ, mdx and LGMD IId ( sgcg/ -) mice will be imaged using MRI.
83 Figure 32. Gene correction model. Gene delivery to the right leg of sgcg/ mice will be achieved by IM injecting 3 week old mice with a muscle specific recombinant adenoassociated virus (1x1010vg of rAAV2/8 Desmin sgcg ) which expresses the human form of the missing sarcoglycan (sgcg), The contralateral limb will not be treated and served as a control. Vector diluted in phosphatebuffered saline (total volume of 100 l per injection) will be directly injected into the lower hindlimb using a 29 1/2 G tuberculin syringe. The needle will be inserted near the distal Achilles tendon and pointed upwards int o the posterior compartment (gastrocnemius and soleus) V irus solution will be injected while withdrawing the needle to maximize volume distribution across the lower limb. Expression of sarcoglycan will be confirmed using immunohistological techniques. M R imaging of both sgcg/ hindlimbs will be performed at 7 weeks of age (4weeks post injection).
84 Figure 33. Radio frequency coil and sample positioning for MRI. A) shows the typical arrangement for in vivo imaging of mouse hind limbs. Both hind limbs w ere placed through a four turn, singletuned 1H solenoid coil. The respiration of each animal was monitored by placing a pneumatic respiratory pad underneath the thoracic cavity. Anesthesia was administered through a face mask. B) Depicted in the photograph is the coil used during the experiments at 4.7T. Here a glass tube filled with capillaries containing a series of copper sulfate solutions is position for imaging. The capacitors to the left were used to tune the coil to the resonance frequency appropri ate for B0 and the 1H nucleus. In the case where B0 = 4.7T, this frequency was 200 MHz. Using a network analyzer, the match capacitors were adjusted to optimize the coil after the sample was secured in to position and reflectance was minimized in order to maximize signal in the MR experiments that would follow immediately.
85 Figure 34. Diffusion in the plane perpendicular to net fiber orientation (axial) was determined by fitting the exponential diffusion dependant loss of signal with increasing diffusio n gradient strength (reflected in the increased b value) in vivo at 11T. We observed an increase in SNR at this increased field strength and more powerful gradient set. Different muscles were seen to have different ADC in this direction (possibly due to fi ber orientation) and lipid rich tissue was observed to have a low er ADC as we would expec t (as seen in the bone marrow), reflected in the lower figure as an elongated exponential curve of signal decay with increasing b value.
86 Figure 35 MR cradle mounted heating block. The temperature of the mice were maintained at 35C for the duration of the DTI data collection to avoid affects of dropping temperature would have on the diffusion of water in the tissue. The system was used to circulate heated water underneath the plexi glass plate that the animal rested on and was able to maintain the temperature within 1C for more than the total MR acquisition time of 90 minutes.
87 Figure 36 T1 was measured in the hindlimb muscles of aged C57BL/10 and mdx mice. In order to determine whether there was a significant difference in T1 between dystrophic and healthy mice progressive saturation experiments were carried out on the 4.7T magnet. A series of images were acquired each having a shorter TR. There was no signi ficant difference in the old mdx vs. control GAS muscles. And while there was more variation in T1 observed in the mdx than the control mice in the TA, this was less concerning since preliminary data showed that the TA was relatively protected from fibrosi s and did not have a significant change in MTR in the old animal. Again this showed that the changes observed in the old GAS muscle was not likely due to T1 differences.
88 Figure 37 sarcoglycan expression in rAAV treated sg/ mice Immunofluorescence (A) of rAAV hSG can be seen panels to the far right (treated animals), while it is absent in the center (untreated sgcg/ -); wt healthy control muscle from a C57BL/10 mouse is shown in the left column. The sarcoglycan is labeled green. The primary antibody sarcoglycan was generated in a mouse and the secondary was goat anti mouse IgG Alexafluor488. Sections were co labeled with rabbit anti laminin primary antibodies and goat anti rabbit IgG Rhodamine. Thus laminin sarcoglycan can be seen more clearly (A lower panel) without the overlay of the rhodamine labeled l aminin. Finally, DAPI was mixed into the mounting media to label nuclei, which are labeled blue. For the Western Blot (B) 15ug of protein per lane (using 1ug/ul, a volume 15ul of sample per lane) was run denatured on a SDS PAGE apparatus and then subsequently blot transferred onto a nitrocellulose membrane. Since g SG is 35KDa and the GAPDH (used as an internal control protein concentration) is 36KDa, the membrane was cut in half to use two different primary antibodies. A ratio of 1:400 of the g SG primary and 1:5,000 of the GAPDH primary antibodies were used. A horseradish peroxidase chemiluminescence kit (BioRad, Hercules, CA, USA) was used to autoexpose x ray film as a method of band detection. Note that SG (35KDa) is missing in the sgcg/ lane
89 CHAPTER 4 EXPERIMENT 1: DETECT ION OF SARCOLEMMAL D AMAGE IN DYSTROPHIC MUSCLE WITH T2 WEIGHTED MRI Abstract The muscular dystrophies (MD) are a group of genetic disease s, which lead to progressiv e weakening of skeletal muscle. This weakening is due to the l ack of structural proteins that protect the sarcolemma from damage during normal muscle activity. It has been established that the transverse relaxation rate constant T2 correlates to histological measures of sarcolemmal permeability in healthy control mus cle. Our first objective was to establish whether or not T2 is also sensitive to damage in dystrophic skeletal muscle. Secondly, I explored the multi exponential nature of the T2 data from muscle tissue and sought to find indicators of pathology at baselin e in dystrophy and within dystrophic lesions. Our results confirmed the utility of T2 in dystrophic muscle as being a solid indicator of damage. This however can be complicated by the presents of edema and fat tissue. Multi exponential analysis showed a novel T2 distribution appear in dystrophic lesions, suggesting that this could be a fruitful direction of future research in trying to further characterize the anatomical or physiological source of the unique distribution peak. Introduction Dystrophic Muscl e Undergoes Bouts of Acute Damage. One characteristic that makes dystrophic muscle damage difficult to study is that it is constantly in the process of recovering from injury. As described in Chapter 1, the dystrophic muscle fibers are susceptible to contr action induced damage, in the case of DMD, due to mutations resulting in the absence of the structural protein dystrophin. Although initially, the regenerative capacity of the dystrophic tissue is fully in tact, the
90 body is ultimately over whelmed and func tional muscle mass is lost. So, even though the muscle fibers are fully repaired in the early stages of the disease, they still lack dystrophin, and will ultimately be damaged again, with what would other wise be normal usage of the muscle. Beyond this tem poral variability, there is also great spatial variance in the location of the dystrophic lesions. Seemingly random lesions can appear over night and at any given time it is difficult to determine exactly what stage a lesion is at until a robust model is established. T2 Relaxation as a Measure of Muscle D amage. The T2 relaxation time of muscle often lengthens with exercise and is sensitive to perfusion, muscle damage, and ischemic conditions (149) Using MRI, acute muscle damage, including that of nondystrophic muscle, is highly conspicuous on T2 weighted images. Frimel et al. (2 005) provided histological confi rmation that significantly increased T2 relaxation rates directly correl ated with compromised sarcolemmal integrity, via Evans blue dye (EBD) permeability assays in injured control muscle (15) EBD tightly binds serum albumin in the blood and its accumulation in fibers indicates membranes with areas with large permeability. These findings are in agreement with our observed results concerning T2 elevation and EBD p ermeability in dystrophic mouse muscles (56, 150) Furthermore, the uptake of EBD can be prevented by the replacement of the missing sarcolemmal protein using gene transfer in mouse models of limb girdle muscular dystrophy. These studies of T2 relaxation in dystrophic muscle have proven valuable in detecting global changes in tissue damage and can consistently discriminate healthy from dystrophic muscle and corrected muscle from uncorrected, yet further refinement o f scanning techniques and data analysis may yield a wealth of additional useful information about the state of the damaged tissue. For
91 instance, was the change in global T2, due to extracellular edema or increased membrane permeability? Edema was expected to be a secondary event to the loss of me mbrane integrity, and the focus of this subaim was to determine whether it would be possible to detect the loss of sarcolemmal integrity in the presence of increased extracellular water. Therefore one of the prim ary goals was to determine if changes in the underlying NMR properties of water reflect changes in muscle membrane integrity. Multi E xponential T2 Decay in Skeletal M uscle. S uggestions of multiple components of muscle water signal in NMR experiments w ere first made over forty years ago with the observation of multiple water line w idths during NMR experiments on excised muscle (7, 151, 152) These signal components were later shown to also have unique T2 relaxation t imes (153) The precise number of these components and their exact origin remains elusive and is the topic of some debate. Many researchers feel that there are at least three characteristic T2 components in healthy muscle in vivo with some papers reporting as many as five (105) and others as few as two (130) The number of T2 components revealed by analysis in a given experiment depends greatly on the MR acquisition parameters and on the algorithm used for the decomposition of the multi component signal. Studies that employed typical imaging sequences to assess the T2 relaxometery of skeletal muscle often resulted in wh at appeared to be monoexponential decay curves (45, 46) In general, imaging based relaxometry can be limited by the lower number of echoes that can be acquired, longer echo times required and lower signal to nois e ratios (SNR) as compared to spectroscopy based data acquisition (154) Ex vivo T2 relaxation of muscle has consistently uncovered multicomponent behavior in muscle, with 34 components using spectral MR methods such as the Carr Purcell Meiboom Gill (CPMG) sequence. The
92 relaxation times associated with these components are generally reported to be around ~5ms, 20ms, 40ms, and 115ms (105) The shortest and longest are thought to be associated with the macromolecular bound water and the extracellular water, respectively (105) There still remains much debate and s peculation over the origins of the middle components, but it is felt that they are likely related to intracellular water (105, 155) While biexponential models often fit the data well, they have the a priori assump tion that only two dominant relaxation rates are present. A less constrained approach uses Hanson and Lawsons nonnegative least squares (NNLS) algorithm to produce a spectrum of T2 components (156) NNLS is now widely use d when handling multicomponent T2 data (105, 137, 157, 158) Proposed O rigins of T2 C omponents in Skeletal M uscle. While much effort has gone into assigning each of these relaxation components to anatomical compart ments within the tissue, as described above (152) (104, 159) a few reports have presented conflicting results (103, 160) leaving the overall consensus unclear. Saab et al. (1999) addressed this by adding a projection presaturation to the base sequence (PP CPMG), which nullified signal that ori ginated from outside a cylindrical region of interest. This localized spectroscopy allowed for the collection of echoes from 1000 echo times, 1.2ms apart, with an signal to noise ratio (SNR) of 3243:1 of in vivo human skeletal muscle data and five T2 comp onents were resolved (105) Gambarota et al. (2001) conducted studies of edema in muscle tissue using CPMG imaging (99) 28) methods in vivo and were been able to detect at least two T2 components, which the authors suggest arise from the intraand extracellular compartments. In that study, only the long T2 component was shortened after a systemic delivery of the paramag netic contrast agent GdDTPA. This, once again,
93 heavily suggests that the >100ms component originates from the extracellular water (155) Some have argued that the middle components (20ms and 40ms) may not be anatom ically partitioned at all, but may be the result of multiple pools of protons undergoing proton exchange at a slow rate or magnetization transfer (105) Inter estingly, glycerin treated muscles continue to have multiple T2 components even though they lack the outer membranous boundaries; although the muscles may have been to thick to realistically expect complete penetrance of the DMSO (107, 160) Yet following maceration, only a monoexponential T2 decay is detectabl e (107) This suggests that the origin of these middle components is likely due to anatomical partitioning of water, but that the barriers are not solely the membrane structures of the muscle tis sue. Such extensive studies of muscle water compartmentalization have not been done on dystrophic muscle. The focus of this subaim was to determine if the changes in T2 components that arise due to the pathology of dystrophic muscle. Purpose and Summary In these experiments, both the age and strain dependence of mean T2 in murine muscle were determined. Verification that measurements of global T2 changes are a robust index of sarcolemmal damage and that this model was applicable in dystrophic muscle was then investigated. Further exploration in the ability to detect multi exponential components of T2 from healthy and dystrophic muscle was then tested. In order to establish confidence in the methodology, computer simulated data sets with known parameters were used to explore the limits of the protocol and analysis. Those limits were then applied and tested using phantom imaging ( in vitro ) experiments before moving onto in vivo subjects.
94 Methods Animals. A total of six young (age: 25months) and six old (ag e: 18 24 months) control C57BL/10Sn/ J, mdx and LGMD IId ( sg/ -) mice were imaged to detect global T2 differences in lower hind limb muscles using MRI (Fig 3 1). The study was conducted with the approval from the University of Florida institutional animal care and use committee. Mice were fed ad libitum and were housed in an AALAC accredited animal facility in a temperature (221C), humidity (5010%), and light (12 h light/dark cycle) controlled room. Gene Correction. In a second experiment, an addition al six 4 6wks old sg/ mice underwent muscle specific delivery of humansg int o one leg and the other leg receive d an intramuscular injection of the same virus encoding for a histological marker gene (LacZ; control limb). The virus that was used for gen e transfer had a truncated muscle specific promoter (Desmin) and the gene was packaged into an AAV pseudotype 8/2 in order to achieve widespread expression following intramuscular injection (Fig 32). The mean T2 from lower hindlimb muscles was determi ned as previously described in Chapter 3: Methodology. Gene expression was verified by immunohistochemical fluorescent labeling of thin cryosectioned muscle tissue using an anti sg Ig as a primary antibody and by Western blot. Histological Verification of S arcolemmal Damage. Mice were injected after the last imaging session with EBD and 24hrs later the lower limb muscles were extracted and examined for macroscopic accumulation of EBD
95 in muscle groups and fibers followed by detailed histological examination ( see Methodology ) to determine membrane and muscle damage. In addition to assessing EBD, thin sectioned muscle was also stained with Hemotoxilin and Eosin. Histological results were then compared with mean T2 measures determined from in vivo T2 weighted MRI data. Non Negative Least Squares (NNLS) Analysis of Multicomponent T2 Computer simulations. A series of computer simulations were implemented in order to test the theoretical limits of our implementation of the T2NNLS algorithm utilizing an in house soft ware program (a modification of BVLS; graciously distributed by M. Cappellari at http://www astro.physics.ox.ac.uk/~mxc/idl/ ), run n ing in the development environment IDL (ITT; Boulder), These simulations were used to determine parameters such as the number of points (echo times) needed, acceptable levels of signal to noise (SNR), optimal spacing (in time) of echo times, and how those parameters effect the programs ability to resolve multiple decay components. Specifically, simulate d complex decay curves hav ing components with T2 times of 5ms, 20ms, 40ms, and 100ms, and weighting factors that mad e each one similar to observed fractions in published studies of muscle tissue. Phantom imaging. In the next step, the ability to resolve 5ms, 20ms, 40ms, and 100ms components in phantoms consisting of different concentrations of CuSO4. Solutions of increasing concentration of CuSO4 are known to shorten T2 relaxation as compared to that of the pure solute, water. This provided a simplified multi component sample, sim ulating muscle tissue, with discrete T2 components, the relaxation of which that could be individually imperially measured. The tiss ue phantom was imaged at 200MHz using a single slice multiecho sequence (with an appropriate crusher sequence
96 (88) to reduce stimulated echoes) with various echo time spacing and signal averaging. These experiments were also used to explore the machine/sequence limitations of detecting multiple components. Results Assessment of Acute Damage in Dystroph ic Muscle U sing T2weighted MRI. To evaluate acute muscle damage and dystrophic pathology, the t ransverse relaxation time constants (T2) of the muscle tissue were used as a noninvasive measure of acute muscle damage. T2weighted images from hindlimb mus cles of young control and dystrophic mice are shown Figure 41A Young animals were chosen because there is a greater level of acute muscle damage in younger dystrophic mice As was expected, a homogeneous distribution of T2 contrast was observed in control hindlimb muscles. On the other hand, the muscles from mdx and sg/ mice were characterized by contrast heterogeneity of the T2weighted images. The elevated T2 seen in the younger dystrophic muscle subsides later in life, but the transverse relaxation, while not significant, show a slightly elongated trend as compared to the age matched controls (Figure 4 1B). The T2 heterogeneity seen in the muscles of the young mice, is due to the presence of dystrophic lesions and is thought to be representative of acute damage to the sarcolemma of muscle fibers and edema formati on. This damage has been verified histologically by others (56, 150, 161) and can been seen in our work in Figure 4 2, by the presence of heterogeneous fiber diameter and central nuclei on H&E stained sections and t he clumped distribution of Evans Blue dye positive fibers in the dystrophic muscles on the bottom panel of the figure. These results show that T2weighted MRI is a useful marker of acute muscle damage as has been reported by others (162) Furthermore the results of the T2weighted MRI experiments provided guidance in
97 determining regions of interest (ROI) for the magnetization transfer studies in chapter 6 and was integral to the interpretation of those results. T2 MRI: Detection of Gene Correction In the L GMD gene correction studies, histological evidence of corrective gene expression and T2 values restored to normal levels (Figures 43; and 37) was seen. In a study published by Pacak et al. in 2007 (Figure 44A) shows an typical example of the effectiveness of the viral delivered gene treatment (56) and our ability to detect the prevention of pathological lesion formation using T2 weighted MRI. The limb on the right sarcoglycan protein and was nearly completely free of dystrophic lesions, while the contralateral limb displayed lesions typical of a young dystrophic animal. Figure 44B illustrates the degree of severity of the untreated limb and emphasizes the level of rescue provided by the therapeutical gene expression by renderi ng the regions, in 3D, of each limb with elevated (twice the standard deviation of healthy muscle) T2 relaxation times. Figure 44C (56) shows quantitatively that the T2 relaxation times of muscles in both the anterior and posterior compartment were returned to values of healthy muscle in the treated limb. These results remained stable for the duration of the experiment (56) and were followed out to 13 weeks (Figure 44C). NNLS Analysis of Multicomponent T2 The ability of the non negative least squares ( NNLS ) program to resolve distinct populations with differing T2 relaxation times is dependent on the number of points sampled and on the ratio of the signal to background noise (156, 163) The increased heterogeneity of untreated dystrophic muscle T2 is clearly seen by a rightward shift (increasing signal intensity) of histogram plots of signal intensity in T2 weighted images
98 as compared to seemingly healthy, treated muscle (Figure 45). In addition to this rightward shift, the untreated muscle histogram also typically has a broadening and a right leaning shelf (elongated T2) that is suggestive of at least a bi modal distribution (Figure 4 5). Our computer simulations took into a ccount these parameters and allow ed us to define the theoretical limits o f the NNLS analysis within the constraint s of our proposed experiment. Iterative monoexponential curve fits were run and followed by NNLS analysis on simulated signal intensity (SI) data points as a function of various distributions of echo times. Emphasi s of these simulations was on the resolving power of the NNLS and the effects of various weighting of each simulated component. Three components 20 ms and 40 ms in healthy muscle, observed at 1.89T (105) and 100 ms in damaged muscle, observed at 7T (130) were selected based upon the litera ture; T2 would be expected to shorten as B0 increases. An example of such simulated data is shown in Figure 46. Given 1,000 echo times, evenly spaced, the 3 components were resolved (Figure 46; lower panel). Adjusting the weighting of each component resulted in less predictable amplitudes and widths of the distribution. These limits were then empirically verified by carryin g out NNLS analysis on MR data collected on CuSO4 phantoms of known concentrations and relaxation times. Example data from one of the CuSO4 phantom trials is presented in figure 47. CuSO4 concentration of 30 mM and 50 mM were chosen for the multiple exponent ROI since their T2 relaxations of 25 ms and 40 ms, at 4.7T were similar to those (proposed intracellular) components found in skel etal muscle (105) that would 1) be necessary to resolve while being so close together (as opposed to a 100 ms component which is easily resolved) and 2) relevant to changes in sarcolemmal integrity and muscle damage. While the NNLS analysis of
99 the multi sample ROI was able to identify both known components, the integral of the distribution of each peak did not reflect the portion of the ROI that was responsible for each T2 relaxation time (50:50). In NNLS analysis of dystrophic muscle in vivo we detected several of the T2 components that have been reported by past researchers (Figure 4 7) With the current design, it was unlikely that the <5ms component could be observed, although it is possible that it was shifted into the 10 ms. The >100ms component was likely seen in muscle that is damaged and has edema (Figure 4 6) While the mi ddle relaxation components varied greatly depending on the degree of disease they w ere possibly the most revealing of the state of the myofibers. As we would expect, a long T2 component arose in ROIs that were sampled within regions of dystrophic lesions (Figure 47C). Discussion T2 weighted MR images of skeletal muscle have been shown to be sensitive to acute injury and exercise induced contrast enhancement. Elevated T2 relaxation times may be due to an increase in tissue fluid content or decreased muscle integrity. It was important for us to establish normalative data of the changes i n global T2 as a function of dystrophic disease progression. T he age dependent changes in T2 of lower hind limbs in two models of muscular dystrophy and in h ealth controls were determined. The elongated T2 time of the skeletal muscle in the young mice was indicative of the active bouts of damage that is observed at that age in dystrophic murine models. Indeed, these increased T2 relaxation times had a high degree of agreement with classical histological measures of sarcolemmal integrity (or lack there of; presence of Evans Blue positive fibers) and pathological morphology (i.e. central nuclei, variable fiber diameter, etc.). This further supports T2 relaxation and T2 weighted imaging as a robust
100 marker of sarcolemmal damage in dystrophic muscle tissue and allowed its use as a reference to identify effected (dystrophic lesions with elevated T2) verses unaffected (seemingly health muscle) when comparing other contrast methods and MR parameters in future chapters. The observed decline in T2 with age (>1.5 years) in all animals could be due to changes in activity, declining muscle growth and/or a general increase in connective tissue and reduced hydration. In the dystrophic mice, we expected that this reduced T2 may be associated with the development of fibrosi s, a topic addressed in Chapter 6. In addition to investigating changes in T2 with age and disease progression, a compar ison of T2 relaxation of corrected and uncorrected dystrophic muscle was made with histological indices of muscle damage and gene expres sion in a model of therapeutic correction. sg restoration to their respective sarcoglycan null LGMD mouse models, confirmed that the rAAV gene delivery was extremely efficient and that the protein was properly localizing with the inner m embrane surface. In 2007, we published the use of T2 sg/ mice. The ability of T2 weighted imaging to quickly identify the lesions in the untreated limb and verify the overwhelming efficacy in the treated side is a beautiful example of the value of T2 as a marker of muscle damage in aiding the preclinical development of a therapeutic agent for MD. While T2 weighting clearly allows us to define regions of unhealthy muscle, the many possible underly ing sources of alteration to the signal decay make it difficult to specifically identify the nature of the underlying pathology. Due to inhomogeneous water
101 and lipid distribution and possibilities of differential water compartmentalization, damaged muscle tissue gives rise to an MR signal that was likely the resulting sum of multiple pools of spins having differing transverse magnetization relaxation rates. On the surface this can make the task of deciphering the precise source of a change in the calculated T2 time ambiguous. However, when a multi component analysis algorithm such as nonnegative least squares (T2NNLS) is considered (105, 130, 137) the decomposition of the multi exponential data resulted in a unique cha racterization of damage; in our case lesions had an additional long component (Figure 47C), similar to those described by Fan et al. in edemas skeletal muscle (130) While a change in the profile of T2 components may yield additional information about the nature of a lesion in the future, as we saw in the simulations and CuSO4 phantom studies, the amplitude and peak width of the T2 component on the NNLS spectrum may not reflect the volume of anatomy or group of spins that you may try to assign to each component. As mentioned earlier, Fan and Does (130) discussed the complex effects of exchange between pools and diffusion may have on the apparent T2 of each component. To resolve these questions further work should be carried out in future studies that are beyond t he constraints of this dissertation. Within the multicomponent analysis, T2 relaxation may still have a good deal of information to tell us about the dynamics of injured and healing muscle tissue. In particular it would be informative to explore changes in the T2 components following bouts of synchronized damage, as described in the next chapter, and following individual lesion over time during the course of repair. This would reduce variation in measures and provide a known time course to begin constructing a relevant model of changes in T2 during tissue remodeling. In addition, such a model could be
102 further explored by the introduction of compartmentally restricted contrast agents, using a pre/post imaging strategy. Thus the researcher may be able to manipulate the relaxation of a single component in the NNLS spectra with an agent confined to a known anatomical compartment. Changes mean global T2 have been solidly establish as robust indicator of sarcolemmal integrity and in this study, this has been extended to dystrophic skeletal muscle. Furthermore, we have established that there is a slight decline of T2 relaxation times with age, but that it is only significant when comparing young verses old dystrophic mice. This may be due to the greatly elevated T2 seen in young muscle, simply returning to a value that is closer to healthy muscle as the period of widespread recurring lesions subsides in these mouse models. Despite the reduction of T2 lesions after one year of age, histological studies suggest that other pathologies are continuing to develop as the dystrophic mouse grows older. It is likely that T2 will be best used in combination with other modalities since pathological factors such as edema and fat infiltration can remain ambiguous with just T2 measur ements alone. As such, it is import for us to explore other noninvasive parameters to monitor disease and recovery in muscle, in order to produce a more complete profile of a given muscles condition, especially for the purpose of aiding the development o f new treatments for neuromuscular diseases.
103 Figure 41. T2 weight accessment of damage in young and old mdx muscle. T2 weighed axial MRI of the lower hindlimb of young (2 5 mo) control C57BL10, mdx/ -, asg/ -, and gsg/ mice. (a) The muscles of the young dystrophic mice show regions of acute damage as bright lesions in these T2w images (TE/TR = 40ms/2s). The mean T2 relaxation times in all mice declined with age and the T2 of dystrophic muscle was generally elecvated as compared to age matched c ontrols. In the gastrocnemius muscle of young mdx mice (b), the mean T2 values were significantly longer than those of age matched controls and longer than the older mdx age group.
104 Figure 42. Histological verification of intensity and distribution of muscle damage in murine models of muscular dystrophy. Healthy control C57BL10 muscle is shown in the first column. (a) As seen in the hemotoxylin/Eosin (H&E) stained cross sections, the control muscle has myofibers of a consisted diameter, a predominance o f peripheral myonuclei, and very little space between fibers. In contrast, the dystrophic muscles show a highly heterogeneous population of fiber sizes, immune cell infiltration, an increase in extracellular matrix, and many centrally located myonuclei (inset), all of which are indicative of ongoing damage and regeneration. (b) Clusters of Evans blue dye (EBD) positive fibers were seen in the dystrophic muscle and were not observed in controls. The spatial distribution of these EBD positive fibers was in agreement with damage observed in the T2w MRI.
105 Figure 43. I sarcoglycan in muscle of the untreated (a) and treated (b) limbs of a gsg/ mouse. Verification of successful gene delivery and proper spatial distr ibution of sarcoglycan) was observed in the treated limbs, as well as it absence was confirmed in the sham treated limbs.
106 Figure 4 4 sg/ sg/ mice was easily observed in T2 weighted MRI (a) by the absence of hyper intense dystrophic lesions in treated tissue, while the limb that under went sham treatment displayed the pathological damage patterns similar to those seen in the young dystrophic mice in the sg/ shown). (b) For illustrative purposes, lesions were segmented in multi slice MR image sets and render in three dimensions using computer models The red regions of damage dominate the sham treated limb, while they were completely absent in the treated limb. (c) ROI analysis of the mean T2 of the treated and untreated TA and gastrocnemius muscles, revealed that the untreated muscles continued to have a significantly elongated T2, while the corrected muscles values were similar to those observed in healthy co ntrol animals. (d) The percentage of voxels with significantly elevated T2 (>2 standard deviations above healthy control muscle) representing damaged muscle was determined (Pacak 2007). Over a course of 13 weeks the treated/sham sg/ mice were i maged and mean muscle T2 values were calculated. The global mean T2 values of the sham treated muscles remained significantly elevated as compared to the treatment group and health control animals.
107 Figure 4 5. Dystrophic muscle pathology is characteriz ed by an increased heterogeneity in T2. The histogram of signal intensity of voxels from regions of interest sampled from rAAV treated (grey) or untreated (red) / muscle in heavily T2 weighted images. The spreading of the distribution and right ward shift are indicative of elongation of T2 in a subpopulation muscle tissue that in this case was due to presence of dystrophic lesions.
108 Figure 46. Multiple com ponent T2 analys is of computer simulated T2 relaxation times. Simulations were carried out to determine the feasability of multiple T2 component resolution using NNLS analysis with data derived from single slice multiecho imaging Here a typical result is shown were three simulated T2 components were added to the simulated signal intensity (SI) as a function of echo time (te). In this case 1,000 simulated echos were curve fit to a series of exponential decay curves and were ran2 values. The best fits were then processed with an NNLS algorithm producing a spectra of T2 components that had contributed to the original signal decays The initial mono exponential fit estimates a mean T2 of 50.53 ms, where the NNLS analysis is able to resolve the three original components.
109 Figure 47. Multiple component T2 analys is in copper sulfate (CuSO4) phantoms (a). S tudies were carried out to determine the feasability of multiple T2 component resolution using NNLS analysis with data d erived from single slice multiecho imaging ( b ) Multi echo T2 weighed images were collected (60 echoes, TE/TR = 7.24ms/6s with 7.24ms increments) at 4.7 T Multiple component analysis was carried on mean values from ROIs which were curve fit ( c, e, g ) to a 2 values. The best fits were then processed with an NNLS algorithm (d, f, h) producing a spectra of T2 components that had contributed to the original signal decays The rate ` of each decay component (T2) and its fraction of total signal contribution (integral of it histogram curve) is reflected in the resultant T2 NNLS spectra. ROIs with only 50mM CuSO4 (c/d) and 30mM CuSO4 (e/f) show true monoexponential decay, while an ROI containing both T2 rela xations (g/h) shows resolution of both T2 times after NNLS and a shifted monoexponetial fit (g).
110 Figure 48. Multiple component T2 analys is in dystrophic muscle. S tudies were carried out to determine the feasability of multiple T2 component resolutio n using NNLS analysis with data derived from single slice multi echo imaging in vivo (a) Multi echo T2 weighed images were colle cted (60 echoes, TE = 7.24 ms to 434.4 ms with 7.24ms increments and a TR = 6 s ) on an 11T Bruker system Multiple component an alysis was carried out using mean relative signal intensity values from sellected ROIs which were curve fit (b) to a series of exponential decay curves 2 values. The best fits were then processed with an NNLS algorithm, producing a spectra of T2 components that had contributed to the original nonmono exponential decay. The rate of each decay component (T2) and its fraction of total si gnal contribution (integral of it histogram curve) is reflected in the resultant T2 NNLS spectra. T he T2 spectra of damaged and seamingly unaffected regions were different, suggesting that further studies may reveal the underlying pathological changes that give rise to the changes observed in global T2 measures of dystrophic muscle.
111 CHAPTER 5 EXPERIMENT 2: IMAGIN G REGENERATION IN DY STROPHIC MUSCLE USING T2 AND DIFFUSION MRI Abstract The muscular dystrophies (MD) are a group of genetic disease s, which lead to progressiv e weakening of skeletal muscle. Although t he types of MD vary in severity, many progress due to a lost of sarcolemmal integrity due a structural defect caused by gene mutations This makes the dystrophic muscle susceptible to contractioninduced damage. Dystrophic muscle is especially prone to injury during eccentric, or lengthening, contractions. While the muscle repair mechanisms are intact, recently repaired dystrophic muscle fibers will be still be prone to repeated damage. During these cy clic bouts of damage and recovery, the muscle tissue undergoes various stages of remodeling. In order to visualize these stages noninvasively, T2 and diffusion weighted MRI was used to observe changes associated with damage and repair in muscle tissue. Ini tial observations of the process of recovery from acute injury, in healthy mice, suggested that changes in indices of diffusion were independent of changes in T2 relaxation, and may have been more sensitive to structural changes that the remodeling tissue was going through. The directions of diffusion perpendicular to the long axis of the myofibers seem to be the most sensitive to this. While there does appear to be a consistent increase in mean diffusion in early dystrophic lesson, the random spatial and t emporal appearance of the damaged areas leads to high variance and makes the measurements difficult to interpret. Experimentally synchronizing the time of incidence of the injury appears to have reduced this variation and made further analysis possible.
112 Specifically, I set out to image the time course of eccentric damage and recovery in dystrophic muscle caused by downhill ( 14) treadmill running for 15 20 minutes The mice were pr e scanned before exercising, immediately afterwards, and imaged on days 1 5, and 10 days post running At each time point T2 weigh ted images and diffusion tensor (DTI) data set s were collected at 11.1T Mean T2 in lesions rapidly increased significantly and remained elevated during most of the repair. Similar to T2, the mean diffusivity (ADC) also rapidly increases, but it tended to migrate towards the baseline at a faster rate during the time course of repair. The rapid decrease in FA often inversely mirrored the changes of the ADC and possibly reflected the loss of structure associated with the necrotic damage that soon follows the initial injury. Of the eigenvalues, the 1 and 2 responded with a similar trend to the ADC and were at their peak height on immediately after exercise. The third eigenvalue, 3,exhibited a delay i n its rise and was at its highest value at the post oneday time point. These results suggest that the observed changes in the transverse relaxation time T2 and the indices of diffusion are governed by unique underlying mechanisms. Further, as it has been suggested in the literature, the third eigenvalue and its vector may be more sensitive to structural changes in the early stages of muscle repair that are otherwise masked by edema and inflammation. Introduction Histological comparisons of dystrophic and h ealthy muscle reveal that there is a much greater variability in fiber size (83) and the presence fiber splitting (136) found in the muscl es with dystrophy. Therefore it was hypothesized that these structural
113 differences associated with dystrophic muscle could be detected utilizing diffusion tensor imaging (DTI). From these experiments sev eral indices of diffusion of water in the muscle tissue were calculated, including an apparent diffusion coefficient (ADC) and fractional anisotropy (FA), along with the eigenvectors and eigenvalues. Physical barriers to water diffusion on health muscle ar e drastically altered in damaged and diseased muscle. While isotropic diffusion is unrestricted in all directions (Figure 51A), healthy skeletal muscle is rather anisotropic (Figure 51B) and as such has a higher FA than damaged muscle, which may have had the normal physical barriers of diffusion breached. Studying the effects of these physical changes, resulting from pathological alteration to the muscle tissue microstructure, on the ADC and FA, as well as comparing them to measures of T2 in prior scans o f the same region, should allow the detection of unique regions of damage and regeneration in dystrophic muscle and allow us to employ a multimodal noninvasive imaging approach to the characterization of muscle tissue undergoing damage and repair. Detectin g Changes in Microstructure with Diffusion MR. Despite the dystrophic muscles increased susceptibility to contractioninduced damage, the muscle tissue is able to transiently repair these damaged regions, at least during the early stages of disease progr ession. Histological signs of muscle regeneration include high variability in the diameter of myofibers, centrally located myonuclei, expression of embryonic myosin and an activated pool of satellite cells (Fig 1bc). Due to high frequency of damage and repair, dystrophic muscle also shows signs of remodeling of the ultrastructure, fiber splitting and changes in fiber orientation (135) Our approach for the further resolution of these structural changes due to tissue
114 remodeli ng will be the application of diffusion weighted and tensor imaging. An apparent diffusion coefficient (ADC) can be calculated for multiple planes of the muscle tissue in vivo Comparison of the diffusion of water and metabolites in healthy and damaged mus cle provides information on the structure and water compartmentalization of the tissue. MR measures of water diffusion in mammalian skeletal muscl e were first made by Cleveland et al. in 1976 during NMR experiment s with excised rat tibialis anterior muscle s. These studies confirmed that diffusion was highly anisotropic and was less hindered in the direction parallel to the muscle fibers (164) i.e. along the length of the muscle fiber Diffusion Tensor Imaging of Skeletal Muscle. Diffusion tensor imaging (DTI) has also pr oven to provide structural information about skeletal muscle through analysis of the diffusion indices i t provides. It was demonstrated that DTI could distinguish differences in the water diffusion properties of muscles with different muscle functionality and reveal ed differences based on gender by Galbn and Ladd (165, 166) Experiments by We d e en (2001) demonstrated the ability of DTI to examine the microstructure of skeletal muscle fiber organization by comparing t he primary eigenvectors, as a measure of principal fiber direction, and angular dispersion in the bovine tongue. While it is agreed upon that the principal eigenvector relates to the direction parallel to the muscle fibers, the origin of the second and thi rd eigenvectors is still not clear (110, 165, 167) In 2004, Galbn suggested that they likely represent diffusion in the directions perpendicular to the fiber direction (165) however histological data has yet to add evidence to support this. In addition to looking at diffusion indices from individual voxels, several studies have demonstrated the ability to apply fiber tr acking algorithms to DTI data sets, allowing the reconstruction of myofiber
115 trajectories in silico based on the regional diffusion data co llected in the DTI experiment (116, 168) These experiments showed the abilit y of measuring pennation angle, by measuring the angle of the fiber tracks from their origins at the tendon, and physiological cross sectional area (PCSA) which is directly perpendicular to the fiber tracks and directly proportional to the muscle maximum force production, suggests the possibility of predictive functional models derived from MRI data Diffusion studies of damaged muscle have also been conducted to assess changes these DTI indices. Muscle s that has been subjected to ischemia (78, 108) as well as trauma (169) often have increased T2 and ADC, while having a decreased fractional anisotropy (FA). In models of muscle atro phy, the ADC was not a ffe cted, but the FA was increased (170) In contrast to acute injury models, Zhang et al. in 2008 showed that denervationinduced muscle atrophy lead to a significant increase in FA and decreases in the secondary and tertiary eigenvalues (171) The decreases in 2 and 3 were interpreted by the auth ors as reflecting decreases in fiber diameter during muscle atrophy. Interestingly, T2 was increased before the observed cross sectional area reduction in the GAS, but then remained insensitive during the remaining muscle volume loss. It was speculated tha t this initial elongation of T2 may have been due to changes in perfusion or edema that where a result of the denervation (171) This uncoupling of T2 and diffusion indices gives further argument that they are sensi tive to different underlying phenomena. The relationship between T2 and ADC is complex and it has been shown that combined T2/diffusion experiments can shed additional light on the origins of these changes in diffusion indices and transverse decay with the onset of disease ( 36) Both
116 T2 relaxation and diffusion have been demonstrated to have multiple components in healthy and damaged muscle (129) When Ababneh et al. investigated both multicomponent transverse relaxation and diffusion in acute edema of muscle in rats, they saw that both T2 and ADC curves were best fit with a bi exponential funct ion (129) While the acute edema did alter the relative signal contribution from the slower (proposed intracellular) and faster (proposed extracellular) T2 components in a way that supported those anatomical water compartmentalization assignments, the same did not hold true for the fast and slow diffusion components. This lead the authors to suggest that an alternate form of compartmentalization was at play when it came to diffusion, such as surface associated verses free liquid pool water to give rise to the fast and slow diffusing water signals (129) In 2007, Fan and Does (130) observed a monoexponential T2 relaxation in healthy rat muscle tissue and a bi exponential fit in edematous or acutely injured muscle following injection of carrageenan. They further explored the relationship between T2 and ADC by varying the echo times of their DTI sequence and found a resultant change in the calculated ADC. It was felt that the two components (short and long for T2 and fast and slow for AD C) actually may represent the intracellular and extracellular compartments but that effects of exchange between the two pools of spins could be in flux and altered by the edema. In addition there could carrageenan that was used to induce the injury or inflammation and the fact that there always is some degree of T2 weighting, that could further cause a shift from a simple calculated expectation of what these values should be (130)
117 Thus, it is important to first make simple measurements of parameters of diffusion in healthy and dystrophic mouse skeletal muscle, in order establish the limits of our model and to optimize our imaging protocol. Following these pilot studies, we set out to measure T2 relaxation, along side diffusion parameters such as ADC and FA, in various mouse models of muscular dystrophy and compare them to healthy controls. Fur ther we evaluated the ability of these diffusion parameters to detect and monitor therapeutic correction in mouse models of limb girdle muscular dystrophy. Ultimately, we explore a synchronized damage model in mdx mice, via treadmill running, that allows u s to determine changes in the eigenvalues and eigenvectors, of the calculated diffusion tensor, during recovery from muscle injury and compare them to the trends of T2, ADC and FA. With this model, we hope to better understand the process of lesion repair in dystrophic muscle and to provide researchers with a tool that provides insight into structural remodeling. Experimental Design Specific protocols are described in Chapter 3, the following describes the overall design of the experiments relevant to the r esults presented in this chapter. The initial results presented in this chapter, related to myotoxin injected control mice, are from the work of kerstedt and Yap (Masters Thesis, 2004) from our laboratory and are presented here for the sake of discussion and their underlying importance to the development of the rational behind my current work. These data consist of two experiments, first measuring T2 and diffusion after notexin, a myotoxin derived from the venom of the Austrian Tiger Snake, Notechis scutat us scutatus (172) injection into the TA muscle in (n=13) healthy C57BL mice, and a second set of measures in mdx mice, comparing damaged and unaffected muscle.
118 The sa me protocols were then used to obtain T2 and diffusion weighted images to evaluate the sensitivity of these contrast methods to detect structural changes and prevention of disease in viral gene therapy treated LGMD II e ( sgca/ -) mice. In these studies, dif fusion was measured along 3 orthogonal laboratory frame axiss (X, Y and Z), with the mean fiber orientaion of the majority of lower limb muscle being parallel to the Z axis (Figure 52). Following those studies, muscle water diffusion was further explored by implementing diffusion tensor imaging (DTI) in healthy and dystrophic mice. A total of six young (age: 25months) and six old (age: 1824 months) control C57BL/10SnJ, mdx and LGMD IId ( sgcg/ -) mice were imaged using directional diffusion weighted imaging (DWI) at 4.7 T and DTI at 11 .1 T to investigate differences in the indices of diffusion in lower hind limb muscles of dystrophic mice throughout disease progression. In a parallel experiment an additional six 46 week old LGMD IId mice underwent muscle specific del High resolution, short TE, and high b value diffusion tensor imaging DTI data w as collected at 11.1 T (Figure 5 1) A fiber orientation independent ADC and fractional anisotropy (FA) was calculated using Paravisi on JIVE (PV4; Bruker Avance) software. Analysis of the DTI data provided additional parameters that gave information about the direction and magnitude of apparent water diffusion in each voxel represented by a set of eigenvectors and eigenvalues. Close ins pection of the corresponding eigenvectors and eigenvalues, along with FA, has aided us in our interpretation of the ADC results and help ed to refine a working model of muscle damage and water diffusion. T2 weight images were also
119 acquired at 11.1 T and prov ided the opportunity to directly compare changes in global T2 and changes in mean ADC between healthy and damaged muscle tissue. Treadmill model of eccentric damage. Mice were allowed to run on a downhill treadmill at an angle of 14 ( 24.9% grade) for 15 to 30 min at a speed of 8 to 10 m/min (Figure 5 2). Exercise was terminated when they could no longer run. This protocol was developed based on reporting the literature and our own experience. This work was approved by the IACUC at the University of Flori da. The mice were first imaged before running and at 0, 1, 2, 5, and 10 days post exercise. Results The goal of this study was to determine the sensitivity of diffusion and T2 weighted MRI to study structural changes during pathogenesis and repair following acute damage in dystrophic skeletal muscle. Diffusion weighted imaging of recovery from acute damage in healthy control C57BL/10 mice. In order to determine the sensitivity of diffusion weighted imaging (DWI) to track the structural changes in murine ske letal muscle following acute injury, kerstedt and and Yap (Masters Thesis, 2004) investigated the TA muscles from a set of healthy C57 control mice (n=13) that were injected with notexin, a myotoxin derived from snake venom, or saline solution (serving as a negative control) by collecting diffusion and T2 weighted MRI. To evaluated these changes over the time during the healing process, the MRI data was collected at 48 hours and 96 hours post injury and at 48 hours alone for mock injury (Figure 53). Both limbs were imaged simultaneously and diffusion weighted images were collected with the diffusion sensitive pulsed field gradients along
120 the three base axis's of the magnet (the slice selective direction, Z; the phase encoded direction, Y; and the frequency encoded direction, X). The fibers of the TA muscle were nearly parallel to the Z axis, while the X and Y axis's were approximately perpendicular to the long axis of the myofibers. At both post injection time points (48 hr and 96 hr) the T2 relaxation times of the notexin injected muscles was significantly lengthened, while it was unchanged in the saline injected muscles (as compared to mean T2 of healthy unjected data). While T2 was significantly increased at both 48 and 96 hours, there was no difference between these two post injected time points. The diffusion in all three directions (Dx, Dy, Dz) was also significantly increased at 48 hours post injection, while diffusion the saline control injected mice remained normal. The diffusion in the slice directi on that is parallel to the fiber length increased during the first two day but failed to be significantly different by day four. In contrast, the diffusion in the perpendicular directions (x and Y) to the myofibers had significantly decreased by 96 hours, as compared to the same measurements in the saline injected control muscles. So while T2 was sensitive to detecting damage and inflammation, it was not sensitive to structural or environmental changes taking place in the recovering tissue. The observations made in notexin injured muscle led to the hypothesis that bulk water diffusion could provide additional information not available with T2 measures. The dynamic elevation followed by a significant depression of the diffusion in directions perpendicular the fiber length suggests that they are the most sensitive to the structural changes during muscle repair. Diffusion weighting imaging of damaged and unaffected muscle in mdx mice. To investigate the ability of diffusion weighted imaging to detect pathologic al changes in dystrophic skeletal muscle (Figure 53), again kerstedt and and Yap
121 (Masters Thesis, 2004) investigated young mdx mice (n=7) by MR imaging them at an age known to display an abundance of dystrophic lesions (3 6 months old). Regions of inte rest were manually determined using manual tracing in custom software; dystrophic lesions (designated mdx d for damage) and unaffected regions ( mdx u) of muscle tissue were outlined for measurements. The T2 relaxation in the unaffected regions was not di fferent from that of control muscle at this age, while the dystrophic lesions consistently had a significantly lengthened T2. Like the control C57 limbs 48 hours post injected with notexin, the dystrophic lesions were determined to have increased diffusion in the axial directions, perpendicular to the muscle fibers, as compared to the unaffected tissue. Interestingly, there was no significant change in the diffusion along the long axis of the myofibers between unaffected and lesions. This again suggests tha t the long axis is less sensitive to alterations the tissue is going through during repair and damage. Diffusion weighted imaging of rAAV correct sgca/ mice. Using the same protocol established in our laboratory (described above), I set out to evaluate t he sensitivity of T2 and diffusion weighted MRI to detect the prevention of sarcoglycan null mice which were treated with a recombinant adenosarcglycan, or a mock treatment in the contra l ateral limb. At 4 weeks of age the mice were MR imaged with diffusion weighting in the directions parallel and perpendicular to the muscle fiber orientation. While there was a significant increase in the T2 relaxation in the muscles of the mock treated lim bs, the rAAV treated muscles were restored to normal levels of the T2 time constant for healthy muscle; thus successfully detecting correction of the pathological phenotype (56) Yet, the diffusion in the axial direction (perpendicular
122 to fiber length) was not found signi ficantly different from the untreated limb. (Figure 54) Nor was the diffusion along the length of the fiber. This seems to be due to several sources of variance in either the animal model itself or the imaging procedure. Diffusion tensor imaging (DTI) in rAAV corrected sgcg/ mice. To determine if the previously observed variation arose from subtle inconsistencies in the limb orientation in reference X, Y and Z axes of the laboratory frame, a rotationally invariant method of diffusion tensor imaging was utilized to evaluated a similar group of treated animals (Figure 55). In this study, sgcg/ mice were sarcoglycan. DTI studies of gene corrected of the sgcg/ mice again showed trends that would be expected with successful treatment, but the measurements had a high degree of variation and the groups were not significantly different. Since limb orientation could then be ruled out as the source of variance, we hypothesized that the random distribution of temporal stages and severity of lesions were likely contributing a large amount of variation into our statistical analysis (Figure 56) Recovery from eccentric damage. To overcome possible variations in the indices of diffusion, due to the random distri bution of occurrence of injury and severity, an attempt to synchronize the damage in a physiologically relevant manner was made. Skeletal muscle in various types of muscular dystrophy has been demonstrated to be susceptible to eccentric contraction inducted damage. As such, a model of mild damage inducted from a single bout of treadmill running, angled at a downhill angle of -14 (Figure 57), proved to be a physiologically relevant model for synchronized tissue repair in the mdx mouse than the use of a gen eral agent like a myotoxin. Using this model, the mdx mice showed
123 consistent damage in all muscle groups (predictable regions were in the medial compartment), while healthy C57 control mice did not show any damage at day 2 post exercise (Figure 58). While the mean T2 in a lesion rapidly increases and remains lengthened during most of the repair process (T2: 18.90.5 ms, 30.71.3 ms**, 27.71.7 ms**, 19.90.3ms, 17.640.2 ms; for Days Pre, 0, 1, 5, and 10. **P<0.001 vs Pre, ANOVA) there are many subtle but significant changes that can be seen in the different parameters of diffusion (Figure 59). Like the T2, the mean diffusivity (ADC) also rapidly increases, but returns to the baseline at a faster rate during the time course of repair. The rapid decrease in FA inversely mirrors the changes of the ADC and reflects the loss of structure associated with the necrotic damage that soon follows the initial injury. Of the eigenvalues, the 1 and 2 responded with a similar trend to the ADC and were at their peak hei ght immediately after exercise. The third eigenvalue, 3 exhibited a sustained significant increase for 24 hours, while 1 and 2 had become indistinguishable from unaffected muscle after the first day. The highest value for 3 was observed one day post running. Discussion These experiments were designed to establish a better understanding of the dynamic changes in T2 and the indices of diffusion following acute and chronic damage in dystrophic skeletal muscle, so that they may be monitored noninvasively using MRI. Similar histological studies rely on sacrificing representative groups of animals at various time points after the incidence of injury. While these data have high resolution and allow a vast array of biological staining and labeling procedures to report the state of the tissue, it does not give us a continuous picture of the dynamic repair process of a
124 single specific lesion. Early work presented here from kerstedt and and Yap (Masters Thesis, 2004) demonstrated that while T2 is a reliable marker for acute muscle damage, it is not very sensitive to massive structural changes going on in regions of repair. And of the directions of diffusion (relative to the laboratory frames primary axiss X, Y, and Z) the directions that were perpendicular to the long axis of the myofibers were the most sensitive to structural and environmental changes occurring at 2 and 4 days after myotoxin injection into the TA muscle of healthy C57 mice. This is in agreement with other studies of diffusion in damaged skeletal muscle (108) Similar trends were observed in dystrophic lesions as compared to seemingly unaffected regions of mdx mice. When simple diffusion weighted imaging was applied to LGMD mice treated with a recombinant ad enoassociated virus in the left hind limb and a mock treatment in the other, high levels in variation made it difficult to consistently distinguish between the treated and nontreated limb. In order to reduce variation brought on by measuring dystrophic l esions of unknown temporal existence, in various stages of repair, we utilized a downhill treadmill running protocol to synchronize the incidence of injury. By imaging the mice before and after running and specifically following lesions that had all occurr ed during the same, single bout of exercise, we then had a method to follow subtle changes in the indices of diffusion MRI and compare these observations to values of T2 as a reference of acute damage. The data acquisition was just over one hour long and g reat care was taken to maintain the mouses body temperature so that it would not confound our measures of diffusion or physiologically alter perfusion. As expected, a significant rise in T2 following exercise, which persisted for approximately one week af ter running, was observed in
125 the mdx mice and was back to normal resting values in a matter of hours in the control animals. Thus the bout of exercise was just at the threshold of damage for the dystrophic model and it was well within the range of limits f or normal use and performance of healthy muscle tissue (Figure 58). As was observed in the notexin experiments in the healthy C57 mice, immediately after injury an increase in ADC and a concomitant reduction in FA was seen in the exercised mdx muscle. Thi s was consistent with an increase in tissue swelling and loss of restriction of water, presumably mostly in the axial directions since water is already not as restricted in the direct parallel to fiber orientation. To look at these changes in diffusion closer, the eigenvalues were measured (along with noting the eigenvector orientaions). The primary and secondary eigenvalues largely followed trends similar to the T2 and the ADC, i.e. they immediately increased and slowly returned to normal values over the n ext week. Interestingly, fluctuations were also seen in seemingly unaffected muscles as well, but they were not as pronounced as the changes in the lesions themselves (Figure 5 9). The greatest difference between unaffected and affected (lesion) muscle was 3 and FA. While an anatomical assignment for the 3rd eigenvector is difficult to make, it has been suggested to be sensitive to changes such as swelling (78) and so perhaps changes in fiber diameter as well. These results suggest that the observed changes in the transverse relaxation time T2 and the indices of diffusion are governed by unique underlying mechanisms. Further, 3) and its vector may be sensitive to structural changes in the early stages of muscle repair that are otherwise masked in T2 weighted imaging by edema and inflammation. The addition of the down hill running protocol t o this
126 experiment offered a physiologically relevant method to study damage and recovery in a dystrophic animal that also reduced the variation of our measures and highlight the importance of synchronized damage in models like these. A combined analysis of T2 and diffusion parameters appears to be a promising approach for monitoring recovery from damage in longitudinal studies of dystrophic skeletal muscle. Such advances in the noninvasive characterization of repair events and markers of damage will advanc e preclinical development of viable treatments for muscular dystrophy. In addition recent reports of DTI and fiber tracking in human skeletal muscle (118) suggest that it may have applications in human MD research as well.
127 Figure 51. The diffusion of water molecules can be restri cted by physical barriers in biological tissues. If the diffusion is completely unhindered A) it is isotropic and all the probability is equal for movement of molecules traveling in all directions (A=B=C). The greater the mean diffusion in all directions i s, the greater the apparent diffusion coefficient (ADC) will be. If, on the other hand, the diffusion is limited in one or two directions B) it is anisotropic. If diffusion in one direction greatly out ways the other two, than the tissue is said to have a high fractional anisotropy (for example if A>>B fibers will tend to be rather anisotropic, due to their tubelike structure similar to cylinder depicted on the far ride side of the figure.
128 Figure 52. Muscle fibers are norm ally rather anisotropic and in our experiments the length of the muscle fibers run approximately parallel to the Z axis, while the X axis and Y axis are perpendicular to the fiber. In this cartoon of a healthy muscle fiber the multiple myonuclei are periph eral.
129 Figure 53. Muscle fibers of the TA muscle relative to the diffusion sensitive pulsed field gradients used in the DWI sequences at 4.7T. The X direction(frequency encoded) and Y direction (phase encoded) of the laboratory frame were perpendicular to mean myofiber orientation in the TA, while the Z direction (slice selection) was approximately parallel.
130 Figure 54. Diffusion weighted imaging of gene corrected LGMD mice. An AAV viral vector was used to deliver human sarcoglycan to group of sgca/ LGMD mice. Images where collected with the diffusion gradient applied either longitudinally (parallel) or axially (perpendicular) to the mean direction of muscle fibers in the limb.
131 Figure 55. Directions and diffusion weighting in diffusion tens or imaging. This series shows a typical DTI data set for a young (26 month old) sgcg/ mouse. A) The A0 image has no diffusion weighting. A dystrophic lesion can be seen on the head of the medial GAS muscle on the right limb (left side of the image). Bot h hind limbs were imaged simultaneously. Images with low and high ( b =100 and 900 s/mm2) were collected in 6 different directions. The set of images were then used to calculate the mean apparent diffusion coefficient (ADC), fractional anisotropy (FA) along with the eigenvectors and eigenvalues of the tensor. These indices of diffusion were then monitored in control and dystrophic muscle at baseline and following acute injury.
132 Figure 56. Diffusion tensor imaging of viral vector treated sg/ mice The inset image shows that the untreated left limb was continued to exhibit a dystrophic phenotype of dystrophic lesions while the right limb looks like healthy control muscle. The untreated limb had an increase in T2 and ADC and a drop in FA.
133 Figure 57. Schematic diagram illustrating the downhill running protocol. The mice ran on a treadmill with a 14 angle for 20 30 min at a speed of 810m/minute.
134 Figure 58. Control C57 and mdx hindlimb muscle post downhill treadmill running. A) Shows the pree xercise condition and B0 show 24hrs post running. Panels C and D suggest that the control C57 mice are not injured by the duration or intensity of the running protocol.
135 Figure 59. T2 and diffusion indices during recovery from eccentric contraction in duced muscle damage. The dark blue plot show the characteristics of a dystrophy lesion, while the light blue line follows a muscle that was not directly damaged by the exercise protocol. Both FA and the third eigenvalue, 3, were interesting over the cours e of recovery from damage.
136 CHAPTER 6 EXPERIMENT 3: ASSESSMENT OF DAMAGE AND F IBROSIS IN DYSTROPHIC MUSCLE USING MAGNETI ZATION TRANSFER MRI Abstract Experiments were performed to access the ability of magnetization transfer (MT) to detect tissue damage i n dystrophic muscle. Areas of muscle edema and acute damage were determined based on areas of elevated T2. The relationship between MT, age and areas of tissue damage was determined in control and mdx and LGMD mice. In addition changes in MT were visualiz ed following an acute bout of muscle damage in mdx mice. The ability of MT to detect tissue damage and fibrosis in muscular dystrophy was determined based on comparison with histology. Fibrosis was measured by staining muscle collagen, the major protein of fibrotic tissue, with Massons trichrome dye. Studies were performed on the hindlimb muscles of murine models of Duchenne ( mdx mice) and LimbGirdle ( sg/ mice) muscular dystrophies. The following results were obtained: 1) MTR is decreased in young dystrophic muscle, 2) Changes in MT contrast correspond to areas of T2 elevation in young dystrophic muscle and following acute muscle damage; 3) MT is dec reased in the presence of extensive fibrosis and decreased muscle T2 in old dystrophic mice and 4) MT was significantly increased following gene correction in LGMD. All of these data taken together suggest that the changes in MT reflect tissue damage in dy strophic muscle. The underlying mechanism resulting in MT contrast is multifactorial and results from a combination of changes in water compartmentation, fibrosis, and possibly protein structure. Introduction The major defect in DM D is the absence of dystr ophin and the primary role of dystrophin is to maintain the integrity of the muscle plasma membrane during muscle
137 contraction (173) Therefore, the sk eletal muscle cell membranes of affected individuals are susceptible to damage during contraction (174) This leads to a loss i n the integrity of muscle cells, a loss of skeletal muscle function, and eventually fibrosis (175) The mechanism involved in the muscle damage which occurs in other forms of MD, such as limb girdle MD, is similar. In these cases, there is a loss of dystrophinassociated proteins which leads to the same type of muscle damage and fibrosis (141) Throughout the progression of the dystrophic myopathies, the affected muscles undergo numer ous cycles of damage and repair. When the regenerative capacity of these muscles has been exhausted, the fibers are progressively replaced by the fat and connective tissue which is the hallmark of fibrosis (176) Fibrosis has particularly negative effects because it greatly reduces muscle function, discourages endogenous repair mechanisms (177) and interferes with many therapeutic interventions The current clinical methods for determining fibrotic changes in muscle rely heavily on histological evaluations of muscle biopsies. In addition to being invasive, this method suffers fr om the disadvantage of the inability to easily examine the entire muscle region over time. A valuable asset to the study of muscular dystrophy would be the development of a noninvasive technique to assess skeletal muscle fibrosis. Detection of Fibrosis wi th Magnetization Transfer Imaging. The repeated bouts of damage and repair that are associated with muscular dystrophy results in either an exhaustion of support cell regenerative capacity and/or a remodeling of the tissue environment that results in an ever increasing replacement of function al muscle fibers with scar tissue and fat. This build up of lipid and fibrosis both decreases the muscles elasticity and reduces the volume of tissue that will be receptive to treatment. Extensive fibrosis has been observed in animal models of muscular
138 dystrophy and progressivel y increases as the animal ages (67) MRI has been shown to be a valuable tool in the study of these neuromuscular diseases, yet current methodology falls short of directly measuring tissue fibrosis. This difficulty arises from the extremely short T2 and short diffusion times (88 91) associated with collagen. To overcome this investigators have utilized a technique which images the magnetization transfer of signal from a bound pool of water to the hi ghly mobile bulk pool of water (72) Tissues containing high concentrations of hydrated macromolecules with short T2s, such as muscle and cartilage, experienc e the largest magnetization transfer effect, resulting in larger decreases in signal intensity compared to other t issues (i.e. fat, blood, water). Guo et al. (2003) suggested that magnetization transfer contrast MRI was sensitive to an increase in tissue f ibrosis in a murine model of liver disease (73) They showed that the amount of magnetization transfer ( at 4.7T) was related to the degree of liver fibrosis and the hydroxyproline content. Hydroxyproline is a modified amino acid that is almost exclusively found in collagen, making it a quantitative biochemical meas ure of tissue collagen content ( 74, 75) Magnetization transfer (MT) was first measured in skeletal muscle tissue by Wol ff and Balaban in1989 and they estimated it to be three times more efficient than MT in the kidney (72) Based on these previous MRI studies, the initial hypothesis of this study was that the increased tissue fibrosis associated with aging dyst rophic muscle will result in an age dependent increase in muscle MT from a "bound" pool of water, associated with collagen and protoglycans, to "bulk" water. MT Imaging of Dystrophic Muscle. Fatty tissue deposition has complicated the interpretation of MT results in human su bjects with muscular dystrophy. For instance, Schick et al. found that when using a
139 water selective imaging sequence that that there was no difference in MT between affected and unaffected muscles in three patients with Erb muscular dystrophy (76) In contrast, McDaniel et al. not only found that in subjects with LGMD that MT was dramatically reduced in muscles with gross fatty infiltration but MT also was reduced in muscle tissues without visual evidence of fatty infiltration (77) Overall fatty tissue infiltration is anticipated to be lesser a problem in murine models of dystrophy. Mouse models of muscular dystrophy reflect many of the hallmarks of human dystrophies with muscle fi ber damage and regeneration, yet the amount of fatty tissue infiltration is not as severe ( 78) MT Imaging of Damaged and Fibrotic Muscle. Decreased muscle MT has also previously been shown to occur under conditions of large fluid shifts. Yoshioka et al. showed an inverse correlation of MT in muscle with free water content in exercised muscle (79) Also Mattila et al found that following muscle damage in a rodent model that there are large acute changes in MT that could not be associated with fatty tissue infiltration over the time scale studied ( 80) Acute muscle damage is associated with muscle fiber swelling, edema, cell necrosis and regeneration. It has also been suggested that other factors such as destruction of the large protein complexes, damaged cell membranes, and infiltration of inflammat ory cells could also reduce MT ( 81) Vahlensieck et al. used MT imaging to look at intramuscular tumor s and observed that muscular scar tissue had a lower MTR than healthy muscle, while both had MTR values that were significantly higher than that of the tumors (82) This supports the possibility that despite collagen having a high MTR relative to most body tissues and a increase in its concentration with progressive fibrosis, that due to the higher MTR of healthy muscle, an increase in fibrosis in muscular dystrophy could
140 result in a decrease in the MTR that is independent of an increase in water content (edema) or lipid deposition. Thus it was intended in this study to use MT imaging as a method of detecting progressive fibrosis in mouse models of muscular dystrophy. Therefore, the objective of this study was to investigate t he ability of MT to detect fibrosis and damage in animal models of muscular dystrophy which undergo extensive fibrosis with age. In addition, MT contrast was compared with T2weighted MRI, a measurement which has been used extensively to assess muscle dam age in dystrophic and injured muscle. T2 mapping was also used as guidance in determining the muscle regions of interest for the MT contrast experiments and in the interpretation of the results. Experiments were performed by using animal models of Duchenne and LimbGirdle muscular dystrophies. Methods Animals This study was conducted with approval from the University of Florida Institutional Animal Care and Use Committee (IACUC). A total of twelve C57BL/10SnJ, twelve mdx (C57BL/10ScSnDmdmdx/J), and twelve LGMD IId ( sg/ -) mice were studied. The twelve mice of each strain were further stratified into two age groups being young and old, resulting in six mice in each unique age/strain group. The animals were housed in an AAALAC accredited animal facility in a temperature (221C), humidity (5010%), and light (12 hr light/dark cycle) controlled room. MR Imaging Data A cquisition MRI was performed in the lower hindlimb muscles of control and dystrophic mice at d ifferent ages. Specifically, quantitative T2 imag ing was implemented to monitor muscle injury/regeneration and MT to assess muscle fibrosis. During all in vivo MR
141 experiments animals were anesthetized using gaseous isoflurane (3% induction, 0.52.5% maintenance). Both hindlimbs were imaged simultaneousl y using a custom built four turn, 1.5 cm single tuned 1H solenoid coil (200MHz) and a 4.7T Bruker Advance (Rheinstetten, Germany) horizontal bore spectrometer (Paravision V3/DMR) To determine transverse relaxation rates (T2), multiple slice, single spin e cho, diffusioncontrolled images were acquired with the following parameters: FOV=1cm, matrix=256x128, slices=12, slice thickness=1mm, slice gap 1 mm, diffusion weighting b = 5 mm2/s, NEX=2 and TR=2s. To avoid the contribution of stimulated echoes to the T2 measurement, I implemented a Hahn spin echo MR image sequence in which two separate acquisitions were acquired at echo times of 14 and 40 ms (15) Magnetization transfer was measured using a same spinecho sequence, except that a single transaxial image was acquired using a MT preparation pulse consisting of a total of 40, 25 msec square pulses (93) First, in a subset of dystrophic and control animals Z spectra were acquired with frequency offsets of 20, 15, 10, 5 KHz using a 1s presaturation pulse. The presaturation power level was empirically derived based on in vivo optimal MT contrast (n=6). However, measurements were only performed within SAR power levels (see methods chapter) The absence of tissue heating was determined using an agar phantom with a fiber optic thermocouple. Based on the Z spectra, consistent discrimination between control and dystrophic animals was observed at a frequency offset of 10KHz. Therefore all subsequent MT data were collected with a presaturation pulse at this frequency offset. Finally, longitudinal relaxation rates ( T1) of dystrophic and control muscles were determined by progressive saturation. Single slice, transaxial images were acquired
142 with a variable TR sequence with TE=7ms, FOV=2.4x 1.8cm, matrix=128x64, slice thickness=3mm, NEX=1 and TRs=6, 3, 1.5, 0.75, 0.325, and 0.2s. MR Image Analysis All T2 and MTR images were analyzed using inhouse software running in the Interactive Data Lang uage programming environment (IDL version 6.2, ITT Corp.). T2 was calculated assuming a single exponential decay with respect to TE as previously described (15) A T2 threshold map was created based on individual T2 values. Pixels with T2 values 2 standard deviations above the average T2 value of control (C57BL/10) muscles were considered to be affected (15) Subsequently the mean T2 value of the muscles of interest ( TA and GAS ) was determined, as well as the mean T2 of affected verses unaffected regions. Muscle groups and affected / unaffected muscle regions were outlined in five image slices and the mean T2 recorded. For the magnetization transfer experiments, the MTR ratio was calc ulated using the following equation: [ Eq uation 61 ] with M0 representing the scan with no MT pulse and MSat representing the image with the MT pulse on. First, a MTC map was created based (56) on individual pixel MSat / M0 values. Subsequently, the mean MTR value was determined from the same ROI as in the T2 maps. T1 images were analyzed using Paravision 3.0.2 (Xtip; PV3 ) and T1 was calculated assuming a single exponential function. [ Eq uation 62 ]
143 Specifically, the mean T1 of the TA and GAS was calculated by fitting the exponential increase in ROI signal intensity of each muscle as a function of TR. Histological Measurements. Following completion of the MR experiments, all animals were euthanized. The tibialis anterior and gastrocnemius muscles were disse cted from both hindlimbs. The muscles were fixed at resting length, coated in O.C.T. gel (TissueTech) and rapidly frozen in melting isopentane and stored at 80C. Frozen muscles were subsequently cut in half and sectioned at the belly of the muscle (the region of muscle with the maximum cross sectional area) Frozen sections (10 m) were either stained using Masson s Trichrome Stain Kit ( Richard Allan Scientific ) or Hematoxylin and Eosin Sections were visualized and digitized under brightfield at 5x and 20x magnification on a DM LB microscope (Leica Microsystems, Solms, Germany). Digital micrograph images of the trichrome stained tissue were analyzed using ImageJ ( http://rsbweb.nih.gov/ij/ ) ROIs, containing the entire muscle cross section, were manually selected and the percentage of positive collagen staining tissue (based on area) was calculated using a HueSaturationIntensity color model for prean post thresholding measures (H:146206, S:0 255, I:0 255; Pass Filter). Postnatal Gene Delivery Gene delivery to the left leg of sgc g/ mice was achieved by IM injection of 3 week old mice with a muscle specific recombinant adenoassociated virus which expresses the human form of the missing sarcoglycan (sgc g ), The contralateral limb received mock treatm ent Briefly, 3 week old sgc g/ mice were anesthetized by induced hypothermia. Vector diluted in phosphatebuffered saline (total volume of 35 l per injection) was directly injected into the lower hindlimb using a 29 gauge tuberculin
144 syringe. The needle was inserted near the distal tibialis anterior tendon and pointed upwards along the tibia. V irus solution was injected while withdrawing the needle to maximize volume distribution across the lower limb (56) MR imaging of both sgc g/ hindlimbs was performed at 8 weeks and 1 year of age. Statistical Analysis All data are expressed as meanstandard errors. T tests were performed to compare mean T2 values of the TA and GAS muscles in e xperiment 1. For experiment 2, one way ANOVA was used to analyze the dependence of T2 a nd MTR on age and mouse strain. Paired t tests were used to evaluate differences between mean T2 and MTR values of the treated and untreated legs of the sgca / mice in experiment 3. Statistical significance was set at P <0.05. Results Optimization of M ag net ization Transfer Contrast. In order to optimize the scan protocol for detection of damage and fibrosis in murine models of muscular dystrophy, magnetization transfer ratios (MTR) were compared over a range of presaturation offset frequencies in aged (>72 wks) healthy C57BL10 and dystrophic mdx mice. Aged mice were chosen for optimization due to the advanced pathology and fibrosis found in the older dystrophic animals. The objective was to determine the presaturation offset frequencies which optimize the differences between control and dystrophic muscle. The relative loss in signal intensity was plotted for each corresponding presaturation offset frequency and this Z spectra was used to compare MT in control and mdx gastrocnemius muscles are shown in Figur e 61 At all offset frequencies tested ( 20KHz to +20KHz; 5KHz increments) the MT ratios were lower in dystrophic than in control muscles (Figure 6 2), as seen by magnetization
145 transfer causing a less severe drop in signal intensity (i.e. the signal intensity in a dystrophic muscle remains higher following an MT pulse than it would in health control muscle) The Z spectrum in sg/ animals was similar to that in mdx mice. Because the largest differences in magnetization transfer contrast were observed between 10 and 15KHz, all subsequent data were acquired at 10KHz. A comparison of MT contrast in the hindlimb muscles of healthy control, mdx, and sg/ mice is shown in Figure 63A These results were obtained using the optimal conditions described above. The data demonstrate differences in MT contrast between control and the two dystrophic samples. Dystrophic muscles have less MT overall than control muscles, and the lesions (T2 > 2SD) within dystrophic muscle display minimal MT. T2 and MTR in Gastrocnemius Muscles From Young and Old Control and Dystrophic M ice. Magnetization transfer ratios (MTR) and transverse relaxation rate constants (T2) were measured in the GAS muscles from young and old control C57 and dystrophic mdx and sg/ mice. The rationale for this approach is that younger dystrophic animals display more acute muscle damage while older dystrophic mice display more fibrosis (177) Our hypothesis is that T2 is the better technique to assess acute muscle damage while MTR is a technique that may be useful to assess fibrosis. T2 and MTR in GAS muscles from young control and dystrophic mice are shown in Figure 41B (T2) and Figure 63B (MTR). These measures were made including the anatomical crossection of the muscle and included regions that had dystrophic lesions. A complete listing of these finding is presented in Table 61 for the TA and GAS muscle of young and old mice from healthy and dystrophic mice. While these data made it clear that MT was able to distinguish healthy from dystrophic muscle, areas of acute damage always had the
146 opposite trend as compared to T2. This was likely due to dominating effects of bulk water influx associated with inflammation in the early months of murine models of dystrophy. Because our primary aim was to assess the ability of MTR to detect fibrosis, areas of acute muscle damage with dystrophic lesions were excluded from the MTR measurements In order to exclude these areas, T2 maps were calculated and areas that had a T2 value of greater than 29 ms (two times the standard error over the mean of healthy muscle) were avoided when ROIs were defined This adjust ment allowed us to evaluate changes in MTR based on changes in the muscles while avoiding areas domi nated by pooling of fluid, i.e. areas of edema or inflammation. Even after avoiding lesions (Figure 64), t he muscles from dystrophic mdx mice displayed si gnificantly longer T2 values ( mdx GAS 280.7 ms) than those from the sg/ and the control animals ( sg/ -GAS 26.470.85 ms, C57GAS 27.381.44 ms) and the T2 values were longer in the muscles from mdx than in those from sg/ animals (Table 6 2). On the other hand, there was no significant difference in the MTR values of muscles from contro l and either dystrophic animal ( mdx GAS 0.810.02, sg/ -GAS 0.810.01, C57GAS 0.81<0.001). The control muscle had considerably less variation. The increase in T2 values suggest that young dystrophic muscle is in a state of acute d amage and that the damage is greater in mdx than in sg/ animals. The lack of a significant difference in MTR measurements between young control and young dystrophic muscle may mean that the young dystrophic mice have not yet developed significant levels of fibrosis.
147 The results of the T2 and MTR measurements in GAS muscles from old control and dystrophic mice are shown in Figure 64 There is no difference in T2 between control and dystrophic animals ( mdx GAS 26.641.34 ms, sg/ -GAS 26.941.21 ms, C57 GAS 26.480.29 ms) However, there is a significant decrease in MTR in dystrophic mice relative to control ( mdx GAS 0.760.03, sg/ -GAS 0.780.03, C57GAS 0.800.01) The decrease in MT is greater in mdx than in sg/ animals. These results suggest tha t bouts of acute damage in aged dystrophic muscle have largely subsided at this point in disease progression and are reflected by the relatively normal T2 values. Of great interest is the significant decrease (6%) in MTR in older dystrophic muscle as compa red to healthy age matched controls. This result indicates that changes in MTR are associated with the fibrosis that comes with advanced disease progression. Histological Assessment of Fibrosis in Dystrophic M uscle. In order to verify the development of fibrosis in dystrophic muscles, experiments were performed with GAS muscles from young and old control, mdx and sg/ mice. The appearance of collagen, the major protein of fibrotic tissue, was assessed by staining the muscles with Massons Trichrome. There was no apparent fibrosis in muscles f rom younger dystrophic animals. The results obtained from muscles of older animals are shown in Figure 65 and is quantified in Figure 66 using a color threshold image processing method to measure the percentage of tissue staining positive for collagen (blue in color). There is no visible fibrosis in control animals. On the other hand, there is collagen present in the muscles from both dystrophic samples. There is more apparent fibrosis in muscles from mdx than in those from sg/ animals. These results are similar
148 to those obtained with MTR; i.e. there is a larger decrease in MT signal in mdx than in sgcg/ muscles with age. Furthermore, histological analys is of the TA muscles in all groups of animals showed few signs of the development of fibrosis. This was also in agreement with MTR measures of the TA, which appears to be somewhat protected in aged dystrophic mice These histological data, in combination with our previous MR imaging results, indicate a relationsh ip between an increase in tissue fibrosis and a decrease in MTR of skeletal muscle in the hindlimb of murine models of muscular dystrophy. Use of T2 and MTR in Monitoring Therapeutic Intervention. T2 and MT contrast images of hindlimb muscles were compared from untreated sg/ mice and from mice which received intramuscular injections of a recombinate virus that expresses human alpha or gamma sarcoglycan. This type of therapeutic intervention leads to reversal of the dystrophic effects. Images were made in younger animals (3 months of age). The results are shown in Figure 67. In younger animals, there were significantly more pixels with elevated T2 in untreated than in treated mice and a significant increase in MTR between the two groups in both the TA and GAS muscles (~ 5%). This is an indication that viral gene delivery was effective in reducing the acute muscle damage which occurs in younger animals and that there was no development of fibrosis at this time. Discussion O lder dystrophic muscle wa s expect ed to be significantly more fibrotic than healthy control tissue especially as age and pathology progressed. Young dystrophic muscle show ed evidence of damage and repair, but did not show extensive overproduction of collagen in the extracellular matrix so early in the course of the disease. The trichrome
149 stain ed tissue sections were useful in assessing the spatial distribution of fibrosis in the muscle cross sections, but ex vivo histology can be difficult to register with MR data acquired in vivo The re sults of these experiments suggest that a decrease in muscle magnetization transfer (MT) is an indicator of muscle damage in dystrophic muscle. The evidence f or this includes the following: 1) histological data show no significant fibrosis in GAS muscles f rom young dystrophic mice, and there is no significant difference in MT between GAS muscles from young dystrophic animals and controls; 2) histological data show significant fibrosis in GAS muscles from older dystrophic mice, and there is a significant dec rease in MT in GAS muscles from older dystrophic animals relative to control; 3) histological data show more fibrosis in GAS muscles from mdx (Duchenne model of MD) than from sgcg/ (limb girdle model of MD) dystrophic animals, and there is a greater dec rease from control in GAS muscle MT from mdx than from sgcg/ mice; and 4) Histological data show little signs of fibrosis in TA muscles from dystrophic mice, and there was no decrease in MT. 4) sgcg/ mice which were treated with a gene transfer tech nique have no histological signs of hind limb muscle fibrosis, and there is no decrease in hind limb muscle MT relative to controls following the gene transfer treatment. All of these data taken together suggest that the decrease in muscle MT in dystrophi c mice is related to tissue damage and fibrosis. The rationale for doing this study was that most MRI techniques to image the replacement of muscle with connective tissue have not been proven to be successful For example, measurements of diffusion imagi ng and T2 were not useful in detecting fibrosis. This was due to the fact that T2 is extremely short and diffusion times are very
150 fast when imaging collagen, the major protein of fibrotic tissue. Some investigators have attempted to use magnetization transfer in order to overcome these previous difficulties (88) With this technique, the magnetization transfer of signal from a bound pool of water to a highly mobile bulk pool of water is imaged. In fact, tissues containing high concentrations of hydrated macromolecules, such as muscle, exhibit t he greatest MT effect. Therefore it was hypothesized that an increase in muscle tissue fibrosis would result in an increase in muscle MT from a bound pool of water associated with collagen to bulk water (92) In this regard, Guo et al demonstrated previously that the amount of MT is related to the degree of liver fibrosis in the Niemann Pick Type C mouse (93) Thus, in the current study, it was expected that an increase, not a decrease, i n muscle MT with fibrosis would be observed. There are at least three possible scenarios to explain the decrease, rather than an increase in MT in fibrotic muscles. First, it is possible that fatty tissue infiltration may account for some decrease in magnetization transfer in these muscles. It has been known for some time that fat tissue has an extremely low MT (178) For example, McDaniel et al showed that MT was dramatically reduced in muscles with gross fatty infiltration in subjects with limb girdle muscular dystrophy. In addition, Shick et al reported that using a water selective imaging sequence would minimize the signal from fat in patients with Erb muscular dystrophy. Although fat infiltration can be a source of decreased MT, it is probably not the case in our experiments for the following reasons: 1) I determined MT in areas in which muscle T2 was not elevated; 2) when compared to canine or human dystrophies, murine muscular dystrophy is characterized by an extremely l ow degree of fatty tissue replacement (41)
151 The second possible reaso n for a decrease in MT involves tissue water. Studies have shown that there is a decrease in magnetization transfer under conditions of an increase in tissue water content. These conditions may be associated with tissue damage as in muscle fiber swelling, edema, cell necrosis and regeneration, inflammation, and damage associated with exercise (120, 121, 179) For example, Yoshioka et al showed a correlation between a decrease in MT and the free wat er content in exercised muscle. However, it is unlikely that these areas of damage were involved in the decrease in MT in our current studies. M easurements of T2 were used to identify areas of damage (15) so that they were excluded from measurements of MT. The third possible explanation for a decrease in MT in fibrotic muscles involves the MTR of collagen and i s probably the most likely explanation. Healthy muscle tissue has a relatively high magnetization transfer ratio ; i.e., muscle is very efficient at MT (176, 180) Collagen, in the form of cartilage, also has a rela tively high MTR, but it is slightly lower than healthy muscle tissue (181) Therefore, when healthy muscle tissue is replaced by collagen in fibrosis, one may expect a decrease in MT. A similar explanation may account for the increase in MT which occurs with liver fibrosis (93) In this case, liver tissue has a relatively low magnetization transfer ratio while the MTR of collagen is greater. In liver fibrosis, healthy liver tissue with a lower MT is replaced by collagen with a greater MT which leads to an overall increase in signal. For example an increase collagen content in muscle would results in a slight decrease in MTR whereas an equal change liver would increase MTR tremendously. Therefore, all of this information taken together suggests that the most likely reason for a decrease in MT
152 with muscle fibrosis is the replacement of healthy muscle tissue with collagen which has a lower MT t han muscle tissue. Finally, the results of all these experiments suggest that the use of T2 and magnetization transfer together can be a valuable tool in MRI studies of murine models of muscular dystrophy. T2weighted MRI can be used to identify areas of edema and acute muscle damage. Therefore, these areas of damage can be identified and filtered out, or not used, for measurements of magnetization transfer. Thus, the decreases in MT in regions of muscle free from edema sugg est the development of fibrosi s.
153 Figure 61. Typical T2 and magnetization transfer contrast and the Z Spectra from damaged and undamaged muscle. (a) T2 weighted (upper) and MTC (lower) images of lower hindlimb of a young mdx mouse display the typical pathological contrast observed in youger acutely damaged dystrophic muscle. A representative Z spectrum of ROIs in mdx muscle (b) reveals differents in MT of dystrophic tissue determined (by T2 analysis; T2 > 29 ms) to be lesions (red) verses seamingly unaffected regions (blue). To deteminstrate the effect of high lipid content would have on MT an additional plot of the Z spectra of bone marrow is included (black).
154 Figure 6 2. Z spectra from the gastrocnemius muscles of old C57BL10 control and old mdx mice.The MT contrast is plotted as the mean (MSat/M0) x 102 (% signal retained following an MT presaturation pulse) as a function of offset RF 0 directly saturated). Data were collected with MT pulses at a range of offset Hz and 0 Hz). N o assymet ry was observed in the Z spectra of either mdx or control muscle. The lower mean MTR (M0 MSat/M0) in mdx muscle was significantly different from controls with MT offset frequencies beyond 10KHz, while 0KHz and 5KHz had consider able direct saturation effects on all muscles. I determined the measuring MT at 10KHz alone would be sufficient for the detection of dystrophic pathology for future experiments.
155 Figure 63. (a) Lesions in young dystrophic mice are distinctly visible in magnetization transfer (MT) contrast images and have a similar spatial distribution as damage observable in T2 weighted images. While similar in trend in contrast, the relatively inverse indices T2 and MTR did not always track with the same magnitude i n damaged and seemingly unaffected dystrophic tissue, confirming that the contrast in the methods do arise from unique endogenous properties of the tissue. (b) While the MTR of the C57BL10 control gastrocnemius muscles have a trend of slight increase with age, the dystrophic muscles showed a significant reduction. The reduction in MTR observed in the older mdx and gsg/ muscles were significant when compared to younger mice of the same strain (indicating the reduction was progressive) and were both signifi cant when compared to age matched control muscle (indicating the reduction was due to pathology).
156 Figure 64 T2 and MTR in unaffected (T2 < 29 ms) GAS muscle.
157 Figure 65 Extensive fibrosis was seen in the muscles of older (>1.5yr) dystrophic mi ce and is not observed in control C57BL10 mice. Here, Massons trichrome increased fibrotic collagen (staining blue) deposition between dystrophic muscle fibers. Severe fibrosis is commonly observed in the diaphragm muscle in muscular dystrophy and is seen in our murine models (lower panel).
158 Figure 66 Extensive fibrosis was observed in the old mdx muscle as compared to the C57 control (a and b). Quantitative image analysis, using a c olor threshold to calculate the percentage of positive collagen (blue) staining, showed that this increase in collagen was significant (c).
159 Figure 67 MT contrast was able to detect genetic correction of dystrophic muscle tissue. (a) The muscle of the treated limb showed homogeneous contrast similar to that of healthy control muscle, while the sham treatment limb displayed the patchy lesions that are commonly seen in young dystrophic tissue. (b) The MTR values of the TA and gastrocnemius muscles of the treated limb were significantly elevated as compared to the sham treatment limb and were similar to MTR values observed in healthy controls.
160 Table 61 Comparison of T2 and MTR for TA and GAS muscles by age and strain Muscle Param. Young C57 Old C57 Youn g mdx Old mdx / / TA T 2 26.340.56 25.300.36 27.480.69 27.670.67 27.540.64 27.631.18 MTR 0.788.0E 3 0.813.6E 3 0.815.3E 3 0.7867E 3 0.805.7E 3 0.788.8E 3 GAS T 2 28.280.32 26.660.29 31.290.43 27.050.31 29.330.82 28. 511.24 MTR 0.788.2E 3 0.802.3E 3 0.795.2E 3 0.765.7E 3 0.796.4E 3 0.771.7E 2 Table 62 T2 and MTR for unaffected regions (T2 < 29 ms) of TA and GAS muscles Muscle Param. Young C57 Old C57 Young mdx Old mdx / / TA T 2 24.5 51.21 24.720.37 26.980.26 26.800.37 25.930.47 25.880.35 MTR 0.815.0E 3 0.815.8E 3 0.814.8E 3 0.792.6E 3 0.805.3E 3 0.788.3E 3 GAS T 2 27.380.75 26.480.24 28.820.17 26.640.27 26.470.24 26.940.36 MTR 0.812.1E 3 0.801.4E 3 0.814.7E 3 0.766.6E 3 0.812.4E 3 0.789.0E 3
161 CHAPTER 7 CONCLUSION The muscular dystrophies are a collection of devastating genetic diseases that result in muscle wasting and currently no acceptable cure is available. Great efforts towards finding clinical solutio ns are underway and many promising avenues have been discovered. Whether it involves gene therapy, cell based treatment, pharmaceuticals or a combination of the above, advances in these fields are fueled by our collective understanding of basic muscle phys iology and growth and repair mechanisms. It is only after fully understanding these underlying components, that we may develop truly elegent ways to help the body repair itself. In this dissertation I have presented a collection of noninvasive techniques that allow us to monitor basic processes in the recovery of damage in skeletal muscle and follow disease progression in models of muscular dystrophy. Each of the MR imaging modalities discussed (T2, diffusion, and MT) have individual strengths and weakness We propose that used in concert, these imaging protocols allow the researcher to derive detailed information about spatial distribution of fiber damage, dynamic processes of remodeling following necrosis, and the pathology of chronic disease, including f ibrosis (Figure 7 1), all without altering the tissue that they are trying to learn to repair. We have extended the observation that elevated T2 corresponds to markers of fiber damage such as Evans Blue dye uptake and other histological marker of damage s uch as immune cell infiltration, central myonuclei, and heterogeneous fiber diameter. Adding to this, diffusion weighted imaging holds information about the nature of muscle fiber structure, orientation, fat infiltration, swelling and dynamic processes of repair. In a chronic disease model, the aged dystrophic mice displayed unique magnetization transfer properties that
162 repeatedly identified acute damage, like T2 weighted imaging, but also in areas where T2 relaxation analysis failed to detect any differenc es despite overwhelming histological evidence showing a marked increase in extracellular matrix proteins (fibrosis). This was an especially valuable finding since fibrosis is particularly difficult to image with MR due to greatly shortened T1 and T2 relax ation times. In addition to the imaging modalities presented in this dissertation, it cannot be stressed enough the value of a protocol such as the downhill tread mill running to illicit a physiologically relevant, eccentric contraction induced, synchroniz ed model of acute damage in skeletal muscle for the purpose of studying repair and remodeling in the tissue. This especially holds true in models of muscular dystrophy, where lesions are occurring seemly at random. The random spatial and random temporal fl ux of necrosis, inflammation and recovery introduces copious amounts of variation into such sensitive measures, such as the indices of diffusion. All together these noninvasive imaging techniques promise to guide our future work in developing effective t reatments for the numerous forms of muscular dystrophy and will strengthen the background of knowledge for those with similar aspirations in pursuit of cures and treatments for other neuromuscular disorders.
163 Figure 7 1 A model of multiple methods of MR I contrast at various stages of damage and disease progression is show above. The first four columns represent the cycles of acute damage and repair that dystrophic muscle mouse tissue undergoes early in life. The first column represents normal healthy or unaffected muscle (PRE = Preinjury), while the following show immediately after (D0= day 0), day 7 (D7), and day 10 (D10) after injury. The final column depicts the advances progression of dystrophy in murine models, exhibiting heterogeneous fiber diameter and extensive fibrosis. .
164 REFERENCES 1. Lexell J & Taylor CC (1989) Variability in muscle fibre areas in whole human quadriceps muscle. How much and why? Acta Physiol Scand 136:561568. 2. O'Doherty DS, Schellinger D, & Raptopoulos V (1977) Computed tomographic patterns of pseudohypertrophic muscular dystrophy: preliminary results. Journal of computer assisted tomography 1:482486. 3. Gersten J (1954) Ultrasonics and muscle disease. American journal of physical medicine 33:6874. 4. He ckmatt JZ, Dubowitz V, & Leeman S (1980) Detection of pathological change in dystrophic muscle with B scan ultrasound imaging. Lancet 1:1389 1390. 5. Weiss ES et al. (1976) External detection and visualization of myocardial ischemia with 11C substrates in vitro and in vivo. Circ Res 39:2432. 6. Scott JA, Rosenthal DI, & Brady TJ (1984) The evaluation of musculoskeletal disease with magnetic resonance imaging. Radiol Clin North Am 22:917924. 7. Cope FW (1969) Nuclear magnetic resonance evidence using D2O for structured water in muscle and brain. Biophys J 9:303319. 8. Hiblar T, Bolson EL, Hubka M, Sheehan FH, & Kushmerick MJ (2003) Three dimensional ultrasound analysis of fascicle orientation in human tibialis anterior muscle enables analysis of macroscopic torque at the cellular level. Adv Exp Med Biol 538:635644; discussion 645. 9. Boesch C (2007) Role of proton MR for the study of muscle lipid metabolism. 10. Lodi R et al. (1997) Correlative MR imaging and 31P MR spectroscopy study in sarcoglycan defi cient limb girdle muscular dystrophy. Neuromuscul. Disord. 7:505511. 11. Ackerman JJ, Grove TH, Wong GG, Gadian DG, & Radda GK (1980) Mapping of metabolites in whole animals by 31P NMR using surface coils. Nature 283:167170. 12. Vandenborne K, Walter G, Leigh J, & Goelman G (1993) pH heterogeneity during exercise in localized spectra from single human muscles. Am J Physiol 265:C13321339. 13. Studitsky AN (1964) Free Autoand Homografts of Muscle Tissue in Experiments on Animals. Ann N Y Acad Sci 120:789801. 14. Stevens J et al. (2004) Muscle adaptations with immobilization and rehabilitation after ankle fracture. Med Sci Sports Exerc 36:1695 1701.
165 15. Frimel T, Walter G, Gibbs J, Gaidosh G, & Vandenborne K (2005) Noninvasive monitoring of muscle damage during reloading following limb disuse. Muscle Nerve 32:605612. 16. Allen DL, Roy RR, & Edgerton VR (1999) Myonuclear domains in muscle adaptation and disease. Muscle Nerve 22:13501360. 17. Gundersen K & Bruusgaard JC (2008) Nuclear domains during muscl e atrophy: nuclei lost or paradigm lost? J Physiol 586:26752681. 18. Mauro A (1961) Satellite cell of skeletal muscle fibers. J Biophys Biochem Cytol 9:493495. 19. Partridge T (2009) Developmental biology: Skeletal muscle comes of age. Nature 460:584585 20. Schultz E (1985) Satellite cells in normal, regenerating and dystrophic muscle. Adv Exp Med Biol 182:7384. 21. Qu Petersen Z et al. (2002) Identification of a novel population of muscle stem cells in mice: potential for muscle regeneration. J. Cell Biol. 157:851 864. 22. Pault et al. (2007) Stem and Progenitor Cells in Skeletal Muscle Development, Maintenance, and Therapy. Mol. Ther. 23. Lepper C, Conway SJ, & Fan CM (2009) Adult satellite cells and embryonic muscle progenitors have distinct genet ic requirements. Nature 460:627631. 24. Funk WD, Ouellette M, & Wright WE (1991) Molecular biology of myogenic regulatory factors. Mol Biol Med 8:185195. 25. Tedesco FS, Dellavalle A, Diaz Manera J, Messina G, & Cossu G (2010) Repairing skeletal muscle: regenerative potential of skeletal muscle stem cells. J Clin Invest 120:1119. 26. Zammit PS (2008) All muscle satellite cells are equal, but are some more equal than others? J Cell Sci 121:29752982. 27. McElhinny AS, Kazmierski ST, Labeit S, & Gregorio C C (2003) Nebulin: the nebulous, multifunctional giant of striated muscle. Trends Cardiovasc Med 13:195201. 28. Friden J & Lieber RL (1992) Structural and mechanical basis of exercise induced muscle injury. Med Sci Sports Exerc 24:521 530. 29. McNeil PL & Khakee R (1992) Disruptions of muscle fiber plasma membranes. Role in exercise induced damage. Am J Pathol 140:10971109.
166 30. Costa ML, Escaleira R, Cataldo A, Oliveira F, & Mermelstein CS (2004) Desmin: molecular interactions and putative functions of the muscle intermediate filament protein. Braz J Med Biol Res 37:18191830. 31. Ervasti JM (2003) Costameres: the Achilles' heel of Herculean muscle. J Biol Chem 278:1359113594. 32. Kobayashi YM et al. (2008) Sarcolemmalocalized nNOS is required to maintai n activity after mild exercise [Supplementary Information]. Nature 456:511515. 33. Goodhart SP & Globus JH (1918) On the nature of muscular dystrophies. Neurol. Bull. 1:386. 34. Erb W (1884) ber die juvenile Form der progressiven Muskelatrophie und ihr e Beziehungen zur sogehanten Pseudohypertrophie der Muskeln. Deutsch Archiv Klin Med 34:467519. 35. Duchenne G (1868) Studies on pseudohypertrophic muscular paralysis or myosclerotic paralysis. Arch Neurol 36. Funsten R (1923) Clincal study of thirty cas es of muscular dystrophy. J Bone Joint Surg Am. 5:190203. 37. Lovering R, Porter N, & Bloch R (2005) The muscular dystrophies: from genes to therapies. Phys Ther 85:13721388. 38. Liu J et al. (1998) Dysferlin, a novel skeletal muscle gene, is mutated in Miyoshi myopathy and limb girdle muscular dystrophy. Nat Genet 20:31 36. 39. Udd B et al. (1998) Tibial muscular dystrophy -from clinical description to linkage on chromosome 2q31. Neuromuscul Disord 8:327332. 40. Straub V et al. (2000) Contrast agent enhanced magnetic resonance imaging of skeletal muscle damage in animal models of muscular dystrophy. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine 44:655 659. 41. Bulfield G, Siller W, Wight P, & Moore K (1984) X chromosomelinked muscular dystrophy (mdx) in the mouse. PNAS 81:1189 1192. 42. Willmann R, Possekel S, Dubach Powell J, Meier T, & Ruegg MA (2009) Mammalian animal models for Duchenne muscular dystrophy. Neuromuscul Disord 19:241249. 43. Araki E et al. (1997) Targeted disruption of exon 52 in the mouse dystrophin gene induced muscle degeneration similar to that observed in Duchenne muscular dystrophy. Biochem Biophys Res Commun 238:492497.
167 44. Cooper BJ & Valentine BA (1988) X linked muscular dystrophy in the dog. Trends Genet 4:30. 45. Valentine BA et al. (1992) Canine X linked muscular dystrophy as an animal model of Duchenne muscular dystrophy: a review. Am J Med Genet 42:352356. 46. Allamand V & Campbell KP (2000) Animal models for muscular dystrophy: valuable tools for the development of therapies. Hum. Mol. Genet. 9:2459 2467. 47. Shimatsu Y et al. (2003) Canine X linked muscular dystrophy in Japan (CXMDJ). Exp Anim 52:9397. 48. Shimatsu Y et al. (2005) Major clinical and histopathological characteristics of canine X linked muscular dystrophy in Japan, CXMDJ. Acta Myol 24:145154. 49. Yugeta N et al. (2006) Cardiac involvement in Beaglebased canine X linked muscular dystrophy in Japan (CXM DJ): electrocardiographic, echocardiographic, and morphologic studies. BMC Cardiovasc Disord 6:47. 50. Nigro V et al. (1997) Identification of the Syrian hamster cardiomyopathy gene. Hum Mol Genet 6:601 607. 51. Straub V et al. (1998) Molecular pathogenesis of muscle degeneration in the deltasarcoglycan deficient hamster. Am J Pathol 153:16231630. 52. Ferrer A, Foster H, Wells KE, Dickson G, & Wells DJ (2004) Long term expression of full length human dystrophin in transgenic mdx mice expressing internal ly deleted human dystrophins. Gene Ther 11:884 893. 53. Tennyson CN, Klamut HJ, & Worton RG (1995) The human dystrophin gene requires 16 hours to be transcribed and is cotranscriptionally spliced. Nat Genet 9:184190. 54. Gilbert R, et al. (2003) Prolonged dystrophin expression and functional correction of mdx mouse muscle following gene transfer with a helper dependent (gutted) adenovirus encoding murine dystrophin. Hum Mol Genet 12:12871299. 55. Cordier L, et al. (2000) Rescue of skeletal muscles of gamm a sarcoglycan deficient mice with adenoassociated virus mediated gene transfer. Mol. Ther. 1:119129. 56. Pacak CA et al. (2007) Long term skeletal muscle protection after gene transfer in a mouse model of LGMD 2D. Mol. Ther. 15:17751781. 57. Li J et a l. (1999) rAAV vector mediated sarcogylcan gene transfer in a hamster model for limb girdle muscular dystrophy. Gene Ther. 6:7482.
168 58. Mendell JR et al. (2009) Limbgirdle muscular dystrophy type 2D gene therapy restores alphasarcoglycan and associated proteins. Ann Neurol 66:290297. 59. Perkins KJ & Davies KE (2002) The role of utrophin in the potential therapy of Duchenne muscular dystrophy. Neuromuscul Disord 12 Suppl 1:S78 89. 60. Gramolini AO et al. (1998) Muscle and neural isoforms of agrin increase utrophin expression in cultured myotubes via a transcriptional regulatory mechanism. J Biol Chem 273:736743. 61. Barton ER, Morris L, Musaro A, Rosenthal N, & Sweeney HL (2002) Musclespecific expression of insulinlike growth factor I counters muscle decline in mdx mice. J. Cell Biol. 157:137148. 62. Shavlakadze T, White J, Hoh J, Rosenthal N, & Grounds M (2004) Targeted expression of insulinlike growth factor I reduces early myofiber necrosis in dystrophic mdx mice. Mol Ther 10:829 843. 63. Wheeler TM (2008) Myotonic dystrophy: therapeutic strategies for the future. Neurotherapeutics 5:592 600. 64. Fisher I et al. (2005) Prednisoloneinduced changes in dystrophic skeletal muscle. Sheng Wu Gong Cheng Xue Bao 19:834836. 65. Balaban B, Matthews DJ, C layton GH, & Carry T (2005) Corticosteroid treatment and functional improvement in Duchenne muscular dystrophy: long term effect. American journal of physical medicine & rehabilitation / Association of Academic Physiatrists 84:843 850. 66. Angelini C (2007) The role of corticosteroids in muscular dystrophy: a critical appraisal. Muscle Nerve 36:424 435. 67. Arbogast S et al. (2006) BowmanBirk inhibitor concentrate prevents atrophy, weakness, and oxidative stress in soleus muscle of hindlimbunloaded mice. J. Appl. Physiol. 102:956 964. 68. Nelson SF, Crosbie RH, Miceli MC, & Spencer MJ (2009) Emerging genetic therapies to treat Duchenne muscular dystrophy. Curr Opin Neurol 22:532 538. 69. Welch et al. (2007) PTC124 targets genetic disorders caused by nons ense mutations. Nature. 70. Lu QL, et al. (2003) Functional amounts of dystrophin produced by skipping the mutated exon in the mdx dystrophic mouse. Nat Med 9:10091014. 71. Yin H, Lu Q, & Wood M (2008) Effective Exon Skipping and Restoration of Dystrophin Expression by Peptide Nucleic Acid Antisense Oligonucleotides in mdx Mice. Mol. Ther. 16:3845.
169 72. Williams JH, Schray RC, Sirsi SR, & Lutz GJ (2008) Nanopolymers improve delivery of exon skipping oligonucleotides and concomitant dystrophin expression in skeletal muscle of mdx mice. BMC Biotechnol 8:35. 73. Sirsi SR, Schray RC, Wheatley MA, & Lutz GJ (2009) Formulation of polylactideco glycolic acid nanospheres for encapsulation and sustained release of poly(ethylene imine) poly(ethylene glycol) copolymers complexed to oligonucleotides. J Nanobiotechnolo gy 7:1. 74. Rimessi P et al. (2009) Cationic PMMA nanoparticles bind and deliver antisense oligoribonucleotides allowing restoration of dystrophin expression in the mdx mouse. Mol Ther 17:820 827. 75. Goyenvalle A, Babbs A, van Ommen GJ, Garcia L, & Davies KE (2009) Enhanced exonskipping induced by U7 snRNA carrying a splicing silencer sequence: Promising tool for DMD therapy. Mol Ther 17:12341240. 76. van Deutekom JC et al. (2007) Local dystrophin restoration with antisense oligonucleotide PRO051. N En gl J Med 357:26772686. 77. Beroud C et al. (2007) Multiexon skipping leading to an artificial DMD protein lacking amino acids from exons 45 through 55 could rescue up to 63% of patients with Duchenne muscular dystrophy. Hum Mutat 28:196202. 78. Heemsker k AM, Strijkers GJ, Drost MR, van Bochove GS, & Nicolay K (2007) Skeletal muscle degeneration and regeneration after femoral artery ligation in mice: monitoring with diffusion MR imaging. Radiology 243:413421. 79. Skuk D (2004) Myoblast transplantation for inherited myopathies: a clinical approach. Expert Opin Biol Ther 4:18711885. 80. Galvez BG et al. (2006) Complete repair of dystrophic skeletal muscle by mesoangioblasts with enhanced migration ability. J Cell Biol 174:231243. 81. Sampaolesi M et al. (2006) Mesoangioblast stem cells ameliorate muscle function in dystrophic dogs. Nature 444:574579. 82. MacReady N (2009) The murky ethics of stem cell tourism. Lancet Oncol 10:317 318. 83. Spurney CF et al. (2009) Preclinical drug trials in the mdx mous e: assessment of reliable and sensitive outcome measures. Muscle Nerve 39:591 602. 84. O'Doherty DS, Schellinger D, & Raptopoulos V (1977) Computed tomographic patterns of pseudohypertrophic muscular dystrophy: preliminary results. J Comput Assist Tomogr 1 :482 486. 85. Gersten JW (1954) Ultrasonics and muscle disease. Am J Phys Med 33:6874.
170 86. Carr HY & Purcell EM (1954) Effects of Diffusion on Free Precession in Nuclear Magnetic Resonance Experiments. Phys Rev 94:630638. 87. Meiboom S & Gill D (1958) Mo dified SpinEcho Method for Measuring Nuclear Relaxation Times. Rev Sci Instrum 29:688691. 88. Gatehouse PD & Bydder GM (2003) Magnetic resonance imaging of short T2 components in tissue. Clinical radiology 58:119. 89. Harrison R, Bronskill MJ, & Henkelm an RM (1995) Magnetization transfer and T2 relaxation components in tissue. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine 33:490 496. 90. Henkelman RM, Stanisz GJ, & Graham S (2001) Magnetization transfer in MRI: a review. NMR Biomed 14:5764. 91. Traore A, Foucat L, & Renou JP (2000) 1H nmr study of water dynamics in hydrated collagen: transverse relaxationtime and diffusion analysis. Biopolymers 53:476483. 9 2. Wolff SD & Balaban RS (1989) Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine 10:135 144. 93. Guo J, Erickson R, Trouard T, Galons J, & Gillies R (2003) Magnetization transfer contrast imaging in Niemann pick type C mouse liver. J Magn Reson Imaging 18:321327. 94. STETTEN MR (1949) Some aspects of the metabolism of hydroxyproline, studied with the aid of isotopic nitrogen. J. Biol. Chem. 181:3137. 95. GREEN NM & LOWTHER DA (1959) Formation of collagen hydroxyproline in vitro. Biochem J 71:5566. 96. Pettegrew J, Minshew N, & Feit H (1984) 31P NMR studies of normal and dystrophic c hicken muscle. Muscle Nerve 7:442446. 97. Dunn JF, Frostick S, Brown G, & Radda GK (1991) Energy status of cells lacking dystrophin: an in vivo/in vitro study of mdx mouse skeletal muscle. Biochim Biophys Acta 1096:115120. 98. Huang Y et al. (1994) Quan titative MR relaxometry study of muscle composition and function in Duchenne muscular dystrophy. Journal of magnetic resonance imaging : JMRI 4:59 64.
171 99. Gong QY et al. (2000) Estimation of body composition in muscular dystrophy by MRI and stereology. Jo urnal of magnetic resonance imaging : JMRI 12:467475. 100. Fleckenstein JL, Canby RC, Parkey RW, & Peshock RM (1988) Acute effects of exercise on MR imaging of skeletal muscle in normal volunteers. AJR American journal of roentgenology 151:231237. 101. M ercuri E et al. (2007) Muscle MRI in inherited neuromuscular disorders: past, present, and future. Journal of magnetic resonance imaging : JMRI 25:433440. 102. Murphy WA, Totty WG, & Carroll JE (1986) MRI of normal and pathologic skeletal muscle. AJR Ame rican journal of roentgenology 146:565574. 103. Le Rumeur E, De Certaines J, Toulouse P, & Rochcongar P (1987) Water phases in rat striated muscles as determined by T2 proton NMR relaxation times. Magn Reson Med 5:267272. 104. Hazlewood CF, Chang DC, Nic hols BL, & Woessner DE (1974) Nuclear magnetic resonance transverse relaxation times of water protons in skeletal muscle. Biophys J 14:583606. 105. Saab G, Thompson R, & Marsh G (1999) Multicomponent T2 relaxation of in vivo skeletal muscle. Magn Reson Me d 42:150 157. 106. Fenrich F, Beaulieu C, & Allen P (2001) Relaxation times and microstructures. NMR Biomed 14:133139. 107. Cole WC, LeBlanc AD, & Jhingran SG (1993) The origin of biexponential T2 relaxation in muscle water. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine 29:1924. 108. Heemskerk AM et al. (2006) DTI based assessment of ischemia reperfusion in mouse skeletal muscle. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine 56:272 281. 109. Gilbert RJ et al. (2006) Mapping complex myoarchitecture in the bovine tongue with diffusion spectrum magnetic resonance imaging. Biophys J 91:10141022. 110. Gilbert RJ & Napadow VJ (2005) Threedimensional muscular architecture of the human tongue determined in vivo with diffusion tensor magnetic resonance imaging. Dysphagia 20:1 7. 111. Kim S, Barnett A, Pierpaoli C, & Chi F ishman G (2007) Threedimensional mapping of lingual myoarchitecture by diffusion tensor MRI. NMR in biomedicine
172 112. Weiss S et al. (2006) Threedimensional fiber architecture of the nonpregnant human uterus determined ex vivo using magnetic resonance diffusion tensor imaging. Anat Rec A Discov Mol Cell Evol Biol 288:8490. 113. Sosnovik DE, Wang R, Dai G, Reese TG, & Wedeen VJ (2009) Diffusion MR tractography of the heart. J Cardiovasc Magn Reson 11:47. 114. Reese TG et al. (1995) Imaging myocardial fi ber architecture in vivo with magnetic resonance. Magn Reson Med 34:786 791. 115. Rohmer D, Sitek A, & Gullberg GT (2007) Reconstruction and visualization of fiber and laminar structure in the normal human heart from ex vivo diffusion tensor magnetic reson ance imaging (DTMRI) data. Invest Radiol 42:777789. 116. Damon BM, Ding Z, Anderson AW, Freyer AS, & Gore JC (2002) Validation of diffusion tensor MRI based muscle fiber tracking. Magnetic resonance in medicine : official journal of the Society of Magneti c Resonance in Medicine / Society of Magnetic Resonance in Medicine 48:97104. 117. Lansdown DA, Ding Z, Wadington M, Hornberger JL, & Damon BM (2007) Quantitative diffusion tensor MRI based fiber tracking of human skeletal muscle. J Appl Physiol 103:673 6 81. 118. Heemskerk AM, Sinha TK, Wilson KJ, Ding Z, & Damon BM (2010) Repeatability of DTI based skeletal muscle fiber tracking. NMR Biomed 119. Yoshioka H et al. (1994) Acute change of exercised muscle using magnetization transfer contrast MR imaging. M agn Reson Med 12:991 997. 120. Mattila KT et al. (1995) Magnetic resonance imaging and magnetization transfer in experimental myonecrosis in the rat. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Soci ety of Magnetic Resonance in Medicine 33:185 192. 121. Hatakenaka M (2003) Effect of muscular contraction on magnetization transfer detected at 1.5 T. Magnetic resonance in medical sciences : MRMS : an official journal of Japan Society of Magnetic Resonanc e in Medicine 2:151 154. 122. Vahlensieck M, Gieseke J, Traber F, & Schild H (1999) The quantification of the magnetization transfer contrast (MTC) effect by calculating MT quotients: does it yield additional information for the differentiation of benign and malignant diseases of the locomotor apparatus? Rofo 171:468472. 123. Matsumura K et al. (1988) Proton spinlattice relaxation time of Duchenne dystrophy skeletal muscle by magnetic resonance imaging. Muscle Nerve 11:97102.
173 124. Misra LK & Narayana PA (1989) In vivo T1 characterization of genetically induced muscle atrophy. Magn Reson Med 7:277282. 125. McIntosh LM, Baker RE, & Anderson JE (1998) Magnetic resonance imaging of regenerating and dystrophic mouse muscle. Biochem Cell Biol 76:532541. 126. Chang DC, Misra LK, Beall PT, Fanguy RC, & Hazlewood CF (1981) Nuclear magnetic resonance study of muscle water protons in muscular dystrophy of chickens. J Cell Physiol 107:139145. 127. Thibaud J L et al. (2007) Characterization of dystrophic muscle in golden retriever muscular dystrophy dogs by nuclear magnetic resonance imaging. Neuromuscul. Disord. 17:575584. 128. Foley JM, Jayaraman RC, Prior BM, Pivarnik JM, & Meyer RA (1999) MR measurements of muscle damage and adaptation after eccentric exercise J Appl Physiol 87:23112318. 129. Ababneh Z et al. (2005) Biexponential Parameterization of Diffusion and T2 Reaxalation Decay Curves in a Rat Muscle Edema Model: Decay Curve Components and Water Compartments. Magnetic Resonance in Medicine. 130. Fan R & Does M (2007) Compartmental relaxation and diffusion tensor imaging measurements in vivo in lambdacarrageenan induced edema in rat skeletal muscle. NMR in biomedicine 131. Sijens PE et al. (2007) MR spectroscopy and diffusion tensor imaging of the bra in in congenital muscular dystrophy with merosin deficiency: metabolite level decreases, fractional anisotropy decreases, and apparent diffusion coefficient increases in the white matter. Brain Dev 29:317321. 132. Schick F (1996) Pulsed magnetization transfer contrast MRI by a sequence with water selective excitation. Journal of computer assisted tomography 20:7379. 133. McDaniel JD et al. (1999) Magnetization transfer imaging of skeletal muscle in autosomal recessive limb girdle muscular dystrophy. Jour nal of computer assisted tomography 23:609 614. 134. Dunn JF & Zaim Wadghiri Y (1999) Quantitative magnetic resonance imaging of the mdx mouse model of Duchenne muscular dystrophy. Muscle Nerve 22:13671371. 135. Pastoret C & Sebille A (1995) mdx mice show progressive weakness and muscle deterioration with age. J Neurol Sci 129:97105. 136. Head SI, Williams DA, & Stephenson DG (1992) Abnormalities in structure and function of limb skeletal muscle fibres of dystrophic mdx mice. Proc Biol Sci 248:163169.
174 137. Saab G, Thompson RT, & Marsh GD (2000) Effects of exercise on muscle transverse relaxation determined by MR imaging and in vivo relaxometry. J. Appl. Physiol. 88:226 233. 138. Heemskerk AM, Strijkers GJ, Vilanova A, Drost MR, & Nicolay K (2005) Determin ation of mouse skeletal muscle architecture using threedimensional diffusion tensor imaging. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine 53:13331340. 139. Morrison J, Lu QL, Pastoret C, Partridge T, & BouGharios G (2000) T celldependent fibrosis in the mdx dystrophic mouse. Lab Invest 80:881891. 140. Vial C, et al. (2008) Skeletal muscle cells express the profibrotic cytokine connective tissue growth fact or (CTGF/CCN2), which induces their dedifferentiation. J Cell Physiol 215:410421. 141. Hack AA et al. (1998) Gammasarcoglycan deficiency leads to muscle membrane defects and apoptosis independent of dystrophin. J. Cell Biol. 142:12791287. 142. BartonD avis E (2006) Impact of sarcoglycan complex on mechanical signal transduction in murine skeletal muscle. Am J Physiol Cell Physiol 290:C411419. 143. Childers MK et al. (2002) Eccentric contraction injury in dystrophic canine muscle. Archives of physical medicine and rehabilitation 83:1572 1578. 144. Edwards RH, Newham DJ, Jones DA, & Chapman SJ (1984) Role of mechanical damage in pathogenesis of proximal myopathy in man. Lancet 1:548 552. 145. Brussee V, Tardif F, & Tremblay JP (1997) Muscle fibers of mdx mice are more vulnerable to exercise than those of normal mice. Neuromuscul Disord 7:487492. 146. Whitehead NP, Streamer M, Lusambili LI, Sachs F, & Allen DG (2006) Streptomycin reduces stretch induced membrane permeability in muscles from mdx mice. Neur omuscul Disord 16:845 854. 147. Lynch GS, Fary CJ, & Williams DA (1997) Quantitative measurement of resting skeletal muscle [Ca2+]i following acute and long term downhill running exercise in mice. Cell Calcium 22:373383. 148. Saupe N, White LM, Stainsby J Tomlinson G, & Sussman MS (2009) Diffusion tensor imaging and fiber tractography of skeletal muscle: optimization of B value for imaging at 1.5 T. AJR American journal of roentgenology 192:W282290.
175 149. Fisher MJ, Meyer RA, Adams GR, Foley JM, & Potchen EJ (1990) Direct relationship between proton T2 and exercise intensity in skeletal muscle MR images. Investigative radiology 25:480 485. 150. Walter G, Cordier L, Bloy D, & Sweeney H (2005) Noninvasive monitoring of gene correction in dystrophic muscle. M agn Reson Med 54:1369 1376. 151. Butterfield TA & Herzog W (2005) Is the force length relationship a useful indicator of contractile element damage following eccentric exercise? Journal of biomechanics 38:1932 1937. 152. Hazlewood CF, Nichols BL, & Chamber lain NF (1969) Evidence for the existence of a minimum of two phases of ordered water in skeletal muscle. Nature 222:747750. 153. Belton PS, Jackson RR, & Packer KJ (1972) Pulsed NMR studies of water in striated muscle. I. Transverse nuclear spin relaxati on times and freezing effects. Biochim Biophys Acta 286:1625. 154. Graham SJ, Stanchev PL, & Bronskill MJ (1996) Criteria for analysis of multicomponent tissue T2 relaxation data. Magnetic resonance in medicine : official journal of the Society of Magneti c Resonance in Medicine / Society of Magnetic Resonance in Medicine 35:370 378. 155. Gambarota G, Cairns BE, Berde CB, & Mulkern RV (2001) Osmotic effects on the T2 relaxation decay of in vivo muscle. Magnetic resonance in medicine : official journal of th e Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine 46:592 599. 156. Whittall KP & Mackay AL (1989) Quantitative Interpretation of Nmr Relaxation Data. Journal of Magnetic Resonance 84:134152. 157. Andrews T, Lancaster J, Dodd S, Contreras Sesvold C, & Fox P (2005) Testing the threepool white matter model adapted for use with T2 relaxometry. Magn Reson Med 54:449 454. 158. Oh J, Han E, Pelletier D, & Nelson S (2006) Measurement of in vivo multi component T2 relaxation t imes for brain tissue using multi slice T2 prep at 1.5 and 3 T. Magn Reson Med 24:3343. 159. Fullerton GD (1982) Basic concepts for nuclear magnetic resonance imaging. Magn Reson Med 1:39 55. 160. Fung BM & Puon PS (1981) Nuclear magnetic resonance transv erse relaxation in muscle water. Biophys J 33:2737. 161. Frimel T et al. (2005) A model of muscle atrophy using cast immobilization in mice. Muscle Nerve 32:672674.
176 162. Bansal D et al. (2003) Defective membrane repair in dysferlindeficient muscular d ystrophy. Nature 423:168172. 163. Dula AN, Gochberg DF, & Does MD (2009) Optimal echo spacing for multi echo imaging measurements of Bi exponential T2 relaxation. J Magn Reson 196:149156. 164. Cleveland GG, Chang DC, Hazlewood CF, & Rorschach HE (1976) N uclear magnetic resonance measurement of skeletal muscle: anisotrophy of the diffusion coefficient of the intracellular water. Biophys J 16:10431053. 165. Galbn CJ, Maderwald S, Uffmann K, de Greiff A, & Ladd ME (2004) Diffusive sensitivity to muscle arc hitecture: a magnetic resonance diffusion tensor imaging study of the human calf. Eur J Appl Physiol 93:253 262. 166. Galbn CJ, Maderwald S, Uffmann K, & Ladd ME (2005) A diffusion tensor imaging analysis of gender differences in water diffusivity within human skeletal muscle. NMR in biomedicine 18:489 498. 167. van Doorn A, Bovendeerd PH, Nicolay K, Drost MR, & Janssen JD (1996) Determination of muscle fibre orientation using DiffusionWeighted MRI. European journal of morphology 34:5 10. 168. Sinha S, Si nha U, & Edgerton VR (2006) In vivo diffusion tensor imaging of the human calf muscle. Journal of magnetic resonance imaging : JMRI 24:182190. 169. Zaraiskaya T, Kumbhare D, & Noseworthy MD (2006) Diffusion tensor imaging in evaluation of human skeletal m uscle injury. Journal of magnetic resonance imaging : JMRI 24:402 408. 170. Saotome T, Sekino M, Eto F, & Ueno S (2006) Evaluation of diffusional anisotropy and microscopic structure in skeletal muscles using magnetic resonance. Magn Reson Med 24:1925. 171. Zhang J, Zhang G, Morrison B, Mori S, & Sheikh KA (2008) Magnetic resonance imaging of mouse skeletal muscle to measure denervation atrophy. Exp Neurol 212:448457. 172. Chang LS (1996) Chemical modification of notexin from Notechis scutatus scutatus (A ustralian tiger snake) venom with pyridoxal 5' phosphate. J Protein Chem 15:473480. 173. Menke A & Jockusch H (1991) Decreased osmotic stability of dystrophinless muscle cells from the mdx mouse. Nature 349:6971. 174. Stedman HH et al. (1991) The mdx m ouse diaphragm reproduces the degenerative changes of Duchenne muscular dystrophy. Nature 352:536539.
177 175. Morizumi H, Hizawa K, Nunomura S, & Ii K (1984) Comparative study of alterations of skeletal muscle in Duchenne muscular dystrophy and polymyositis. Acta Pathol Jpn 34:12211242. 176. Blau H, Webster C, & Pavlath G (1983) Defective myoblasts identified in Duchenne muscular dystrophy. PNAS 80:48564860. 177. Alexakis, Partridge, & BouGharios (2007) Implication of the satellite cell in dystrophic muscl e fibrosis: a self perpetuating mechanism of collagen over production. Am J Physiol Cell Physiol. 178. Loesberg AC, Kormano M, & Lipton MJ (1993) Magnetization transfer imaging of normal and abnormal liver at 0.1 T. Investigative radiology 28:726 731. 179. Louie E, Gochberg D, Does M, & Damon B (2008) Transverse relaxation and magnetization transfer in skeletal muscle: Effect of pH. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Reson ance in Medicine. 180. Boss A et al. (2006) Whole body magnetization transfer contrast imaging. Journal of magnetic resonance imaging : JMRI 24:11831187. 181. Stanisz GJ et al. (2005) T1, T2 relaxation and magnetization transfer in tissue at 3T. Magneti c Resonance in Medicine 54:507 512.
178 BIOGRAPHICAL SKETCH Nathan Bryant began his academic career initially interested in the visual arts at Middle Tennessee State University. It was there that an introductory biology course opened his eyes to the prospec ts of a career in the life sciences. There he majored in general biology, with a concentration in microbiology, and minored in chemistry. He graduated with a B.S. degree in 1999. He continued there, working in the laboratory of Anthony Farone, conducting r esearch on novel Legionellalike amoebal pathogens. During this time he was a graduate teaching assistant and instructed the laboratory section of the same introductory biology course that had brought him into the field three years earlier. During this tim e he was awarded a GAANN fellowship, from the U.S. Department of Education, which he brought with him during his first year of his doctoral studies at the University of Florida. In Florida, he enrolled in the Interdisciplinary graduate program in biomedica l sciences (IDP) in the College of Medicine. He spent his first two years studying the biophysical characteristics of the capsid proteins of Parvoviruses and was interested in their role in virulence and host cell tropism. This work, in the Department of B iochemistry and Molecular Biology, was largely centered around structural studies utilizing X ray crystallography. The world of three dimensional data sets and computer modeling reunited him with his early interests in the fine arts and computer rendering. In 2006 he began working in the laboratory of Glenn A. Walter in the Department of Physiology and Functional Genomics and later received an NIH T 32 Training Fellowship in neuromuscular plasticity. His current research, which is the foundation of the diss ertation, is focused on developing noninvasive methods of monitoring damage and repair in dystrophic skeletal muscle using MRI. By delving into the science of
179 imaging, he has been able to interlock his passion for both art and science into his daily work. At present Nathan lives in Gainesville, Florida, enjoys photography, music, camping, rock climbing, Argentine tango, and the company of his friends and family.