Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2014-05-31.

Permanent Link: http://ufdc.ufl.edu/UFE0042559/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2014-05-31.
Physical Description: Book
Language: english
Creator: Lee, Choong Heon
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012


Subjects / Keywords: Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre: Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Statement of Responsibility: by Choong Heon Lee.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Lin, Jenshan.
Local: Co-adviser: Blackband, Stephen J.
Electronic Access: INACCESSIBLE UNTIL 2014-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0042559:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042559/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2014-05-31.
Physical Description: Book
Language: english
Creator: Lee, Choong Heon
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012


Subjects / Keywords: Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre: Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Statement of Responsibility: by Choong Heon Lee.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Lin, Jenshan.
Local: Co-adviser: Blackband, Stephen J.
Electronic Access: INACCESSIBLE UNTIL 2014-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0042559:00001

This item has the following downloads:

Full Text




2 201 2 Choong Heon Lee


3 To my family


4 ACKNOWLEDGMENTS I would like to thank the many people who contributed directly and indirectly to this work. Most of all I have deeply indebted to my advisor, Dr. Stephen Blackband, my mentor and my friend who entrusted to me a rich vein of research and time to play with MR equipment and give me insights and lessons beyond science and laboratory experience His time of adv ising and teaching me in Blackband lab was tremendously entertaining and gave me an unlimited opportunity to work with many projects and excellent senior personnel and explore my interests as well. His ability to offer thoughtful questions and helpful sug gestions as well as easygoing nature and witty sense of humour made many casual coffee and lunch talks on any topic in any areas. Not to mention all the experiments and experiences where all the excitement was hide. Dr. John Forder, my committee member and friend, enabled me to develop and discuss ideas in a truly scientific and supportive environment. His biological expertise and insights have been an insp iration and he gave many helpful suggestions on any topic. Also I would like to thank my advisor Dr. Jenshan Lin for his invaluable guidance through my graduate studies and useful discussions. I would like to Dr. Huikai Xie for important opportunity to access the OCT equipment and insightful discussions. I would like to thank Dr. Glen n Walter for his exer tions teaching me about the diaphragm and skeletal muscle sample for the MR scan analysis. I would also like to thank Dr. Lisa Guay Woodford at the University of Alabama at Birmingham for the extraordinary generosity she has shown towards me. She has given the kidney and liver sample s welcomed my ideas with open arms, and provided the histological data for validation. I would like to thank all the former postdoc members i n Blackband lab. These include Drs. Jonathan Bui Timothy Shepherd, Peter Thelwall, S amuel Grant, and Kyle


5 Padgett. They were role models for me to think like a real scientist an d show themselves how to deal with scientific questions with patience and commit ment The es Rocca, Kelly Jenkins, and Xeve Silver also deserve my thanks for his warmness and a source of encouragement and friendship throughout my time in AMRIS. Especially I thank Dan Plant for being a source of encouragement and telling me about fishing and the of emphasizing doing well in school and setting my goals high. Jeremy Flint is my friend in the lab and he has a magical ability to explore the biological topics togeth er which I have had a difficulty in swallowing them at first. He has a very kind personality a nd is all the time pleasure with whom to work. I would like to thank for editing and advice in biological content. I am grateful to Min sik Hwang who led me to la unch my ride on the MR researching track and with his kindness and encouragement, he is all the time a good friend. I would like to thank Sharon Portnoy for helping in resolving issues in the MR machine editing my dissertation especially technical areas, and her long time job experience was also a good source for career advice I would like to thank Dr. Brian Hansen for his collaboration on the performing our MATLAB based DTI and DTT analysis. Also, I would like to offer a special thanks to Dr. Peter Veste rgaard Poulsen for his support and inspiration of my research as collaborative efforts. Especially I would like to thank my mother who is now in heaven but in my mind all the time. She believed in me and encouraged me all the time to attend graduate schoo l with countless care. I thank my brother Choong Suk for helping me through keeping


6 my role even when my mother was not in good health. Also I thank my brother Choong Yeop for being a good friend. I would thank f or my friends and colleagues at UF in particular Sew oon Choe and Hyochul Ahn, for all extra work and being on my side to help me push through when I was in difficult times.


7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 11 LIST OF FIGURES ................................ ................................ ................................ ........ 12 ABSTRACT ................................ ................................ ................................ ................... 16 CHAPTER 1 BACKGROUND AND SIGNIFICANCE ................................ ................................ ... 18 Overview of Magnetic Resonance Imaging (MRI) system ................................ ...... 18 Background and Motivation for Magnetic Resonance Microscopy .......................... 19 Types of RF Coils ................................ ................................ ................................ ... 20 Principles of Magnetic Resonance Imaging ................................ ............................ 23 Theory of SNR ................................ ................................ ................................ ........ 25 Calculation of SNR ................................ ................................ ........................... 25 Sources of Noise ................................ ................................ .............................. 26 Avenues for Noise Reduction ................................ ................................ ........... 28 High Field Effects ................................ ................................ ............................. 28 Spin Excitation, Relaxation, and Signal D etection ................................ .................. 30 Spin Excitation ................................ ................................ ................................ .. 30 Spin Relaxation ................................ ................................ ................................ 30 Signal Detecti on ................................ ................................ ............................... 31 Gradient Coil: Its role in Image Formation and the Resolution Limits ..................... 32 Considerations for High Resolution Images ................................ ..................... 32 Limitations of Resolution and Gradient Requirements ................................ ..... 34 T 2 limited resolution ................................ ................................ .................. 34 Bandwidth limited resolution ................................ ................................ ...... 34 Diffusion limited resolution ................................ ................................ ......... 35 Diffusion based Contrast Mechanism ................................ ................................ ..... 38 Diffusion Weighted Imaging ................................ ................................ .............. 38 Compartmentation Issues ................................ ................................ ................. 41 Diffusion Tensor Imag ing ................................ ................................ ........................ 42 Anisotropy Index ................................ ................................ ............................... 44 High Angular Resolution Diffusion Imaging (HARDI) ................................ ........ 45 2 MAGNETIC RESONANCE MICROSCOPY OF SUBCELLULAR STRUCTURE IN XENOPUS LAVIS OOCYTES ................................ ................................ ............ 58 Introduction ................................ ................................ ................................ ............. 58 Methods ................................ ................................ ................................ .................. 63 Sample Preparations and Protocols ................................ ................................ 63


8 Immersion fixed samples ................................ ................................ ........... 63 Immersion of fixed samples in agarose and slice preparation .................... 64 Magnetic Resonance Microscopy Imaging with Sample Retaining Apparatus ................................ ................................ ................................ ...... 64 Image Analysis and Validations ................................ ................................ ........ 65 Results ................................ ................................ ................................ .................... 66 Discussion ................................ ................................ ................................ .............. 69 Conclusions and Future Work ................................ ................................ ................. 71 3 REVEALING SUBCELLUAR ARCHITECTURE AND INTRACELLULAR DIFFUSION CONTRAST FROM L7 NEURONS OF APLYSIA CALIFORNICA ...... 85 Introduction ................................ ................................ ................................ ............. 85 Benefits of Using This Animal ................................ ................................ ........... 85 Novelty of the Current Study ................................ ................................ ............ 86 Methods ................................ ................................ ................................ .................. 89 Sample Preparation and Protocols ................................ ................................ ... 89 Sample Positioning and Magnetic Resonance M icroscopy .............................. 90 Image Analysis and Histological Comparisons ................................ ................. 91 Results ................................ ................................ ................................ .................... 92 Delineation of Subcellular Architecture in a Neural Cell of A. californica using MR Detectable Contrast Employing Novel Methodologies of Sample Preparation ................................ ................................ ...................... 92 Histological Confirmation of the MR delineated Subcellular Architecture of a Single Cell of A. californica ................................ ................................ ............ 94 Preliminary Studies to Determine MR Contrast and Intracellular Water Diffusion out of the (Sub) Cellular Arc hitecture of a Single Cell of A californica using Diffusion MRM ................................ ................................ .... 95 Preliminary Studies to Investigate the Fixation Effects on the (Sub) Cellular Architecture of a Single A. californica Neuron ................................ .............. 96 Discussion ................................ ................................ ................................ .............. 97 Novel Methodology in Sample Preparation ................................ ...................... 97 Com partmentation Issues ................................ ................................ ................. 98 Enhanced SNR and Resolution by Updating Hardware ................................ 102 ................................ ................................ ............. 103 Conclusions and Future Work ................................ ................................ ............... 103 4 MAGNETIC RESONANCE MICROSCOPY OF RAT, PIG, AND HUMAN NEURONS ................................ ................................ ................................ ............ 122 Introduction ................................ ................................ ................................ ........... 122 Methods ................................ ................................ ................................ ................ 126 Sample Preparation and MRM Imaging ................................ ......................... 126 MR Pulse Sequence ................................ ................................ ....................... 127 Image Analysis and Structural Validation ................................ ....................... 128 3D Segmentation and Reconstruction ................................ ............................ 129 Results ................................ ................................ ................................ .................. 130


9 Delineation of Cellular Architecture of Mammalian Brain Cells in the CNS Tissue of Rats using MR Microscopy using No vel Micro Surface coils and 3D Visualization ................................ ................................ .......................... 130 Confirmation of the Delineated Cellular Architecture of Neural Tissues with Corresponding Histological Validation using MR Microscopy ...................... 132 Delineation of Neural Processes in Pig Spinal Cord Employing MR Microscopy Techniques and Micro Surface coils ................................ ........ 133 Preliminary Evidenc e of Intracellular Compartmentation and How This Relates to Fast and Slow Diffusing Water Pools Inside the Human Neuronal Cell Utilizing Micro Surface coils ................................ .................. 133 Discussion ................................ ................................ ................................ ............ 134 Considerations When Imaging Human and Mammalian Cells ........................ 134 Compartmentation Issues ................................ ................................ ............... 135 Cell Swelling ................................ ................................ ................................ ... 136 Conclusions and Future Work ................................ ................................ ............... 137 5 DIFFUSION TENSOR MICROSCOPY OF THE CELLULAR STRUCTURE OF KIDNEYS AND L IVER IN THE AUTOSOMAL RECESSIVE POLYCYSTIC KIDNEY DISEASE (ARPKD) MOUSE MODEL ................................ .................... 147 Introduction ................................ ................................ ................................ ........... 147 Structure and Function of the Kidneys and Liver ................................ ............ 148 Gross Anatomy of the Kidney ................................ ................................ ......... 149 Microscopic Anatomy of the Kidney ................................ ............................... 149 Introduction to Polycystic Kidney Disease, PKD ................................ ............. 151 Rationale for Investigation of the Cellular Domain ................................ .......... 1 53 Utilization of cpk Mice ................................ ................................ ..................... 154 Contributions of MRI ................................ ................................ ....................... 154 Significance of Volumetrics ................................ ................................ ............ 156 Methods ................................ ................................ ................................ ................ 157 Wild Type and Mutant Mouse Liver and Kidney ................................ ............. 157 Volume Measurements by Water Displacemen t, MRM, and Correlative Histopathology of ex vivo Livers and Kidneys ................................ ............. 158 Micro Surface coil Imaging for Micro Structural Information and Correlation with Histopathology of ex vivo Tissues of B oth Types of Livers and Kidneys ................................ ................................ ................................ ....... 160 Results ................................ ................................ ................................ .................. 161 MRM Analysis of Wild Type and Mutant Mouse Kidneys ............................... 161 MRM Analysis and 3 D Segmentation of the Biliary Architecture in Wild Type and Mutant Mouse Livers ................................ ................................ ... 162 MRM Analysis of Fixed Kidneys Using Micro Surface Coils ........................... 163 Discussion ................................ ................................ ................................ ............ 164 Volume Measurements by Water Displacement, MRM and Correlative Histopathology of Fixed Livers and Kidne ys ................................ ................ 164 Magnetic Resonance Microscopy of Fixed Kidneys ................................ ....... 166 Conclusions and Future Work ................................ ................................ ............... 167 Diffusion Tensor Tractography of Nephron Structures ................................ ... 168


10 Volume Comparison between Normal and Mutant Kidneys and Liver at Various Stages of Disease Onset and Progress ................................ ......... 168 Observation of Collecting Duct Dilation ................................ .......................... 169 6 MR MICROSCOPY OF SKELETAL MUSCLE CELLS IN HEALTHY AND DYSTROPHIC DIAPHRAGM DIS EASE MODEL MICE ................................ ........ 180 Introduction ................................ ................................ ................................ ........... 180 Structure of Skeletal Muscle in the Diaphragm ................................ ............... 180 Duchenne Muscular Dystrophy, a Lethal Muscle Wasting Disease ................ 181 A Rat Model of the Disease, X linked Muscular Dystrophy (MDX) ................. 182 Imaging of a Single Muscle Cell ................................ ................................ ..... 183 Potential Contributions of MRM into the Monitoring and Treatment of DMD .. 184 Limitations of Current Treatment Options for DMD and Potential Contributions of MRM to Therapeutic Treatments ................................ ....... 185 Methods ................................ ................................ ................................ ................ 186 Sample and Slice Preparation ................................ ................................ ........ 186 Magnetic Resonance Microscopy ................................ ................................ ... 186 Results ................................ ................................ ................................ .................. 187 MR Microscopy of Healthy and Dystrophic Diaphragm, and Microstructural Information in the Transverse Plane. ................................ .......................... 187 MR Microscopy of Healthy and Dystrophic Diaphragm in the Coronal Plane. 189 Discussion ................................ ................................ ................................ ............ 190 MRM Imaging of the Healthy and Diseased Diaphragm ................................ 190 Conclusions ................................ ................................ ................................ .......... 193 7 CONCLUSIONS AND FUTURE WORKS ................................ ............................. 211 Conclusions ................................ ................................ ................................ .......... 211 Future Works ................................ ................................ ................................ ........ 213 LIST OF REFERENCES ................................ ................................ ............................. 215 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 237


11 LIST OF TABLES Table page 1 1 Comparison between a micro 5 gradient, and a planar gradient. ....................... 57 3 1 Osmoliarity of artificia l sea water, anaesthesia, and body fluid of A.californica 105 5 1 Volumetric analysis of wild type kidney volumes ................................ .............. 179 5 2 Volume tric analysis of mutant kidney volumes. ................................ ................ 179 5 3 Volumetric analysis of wild type liver ................................ ................................ 179 5 4 Volumetric analysis of mutant liv er ................................ ................................ ... 179


12 LIST OF FIGURES Figure page 1 1 Schematic of RF volume coils ................................ ................................ ............ 48 1 2 Schematic of RF micro coils.. ................................ ................................ ............. 49 1 3 Schematic of gradient echo pulse sequence.. ................................ .................... 50 1 4 Schematic of spi n echo pulse sequence.. ................................ .......................... 51 1 5 Schematic of a diffusion weighted pulse sequence and the corresponding MR signal characteristics. ................................ ................................ ................... 52 1 6 Schematic of transmitting and receiving the RF signal on resonance.. .............. 53 1 7 Pictures of a micro planar gradient. ................................ ................................ .... 54 1 8 Schematic of water diffusion in an isotropic and an anisotropic medium.. .......... 55 1 9 Schematic of comparison between DTI and HARDI. ................................ .......... 56 2 1 Representative p hotograph of the 500 coil.. .......................... 73 2 2 Light microscopy images of a frog ovum.. ................................ .......................... 74 2 3 Light microscopy image of nucleus of frog ovum and schematic of the subnuclear constituents.. ................................ ................................ .................... 75 2 4 Schematic of cellular structure of frog ovum and the protocol of sample preparation.. ................................ ................................ ................................ ....... 76 2 5 Direct comparison between cellular compartments in MRM and histology images. ................................ ................................ ................................ ............... 77 2 6 Direct comparison between cellular compartments in MRM and histology images ................................ ................................ ................................ ................ 78 2 7 Comparative analysis between MRM and histology images. .............................. 79 2 8 A series of T 2 wei ghted MRM. ................................ ................................ ............ 80 2 9 A series of 3D GE MRM with different resolution.. ................................ .............. 81 2 10 A series of 3D GE MRM with different resolution and TR.. ................................ 82 2 11 Comparative analysis between MRM and histology images ............................... 83


13 2 12 Representative histol ogical image of frog ovum stained with DAPI.. .................. 84 3 1 Schematic of an animal cell.. ................................ ................................ ............ 106 3 2 Schematic of an Aplys ia ganglion.. ................................ ................................ ... 107 3 3 Schematic of internal structure of ganglion. ................................ ...................... 108 3 4 Schematic of sample preparation proto col.. ................................ ..................... 109 3 5 Schematic of sample placement in the sample well embedding the RF coil. .... 110 3 6 Direct correlation of Aplysia neuron morphology in MRM and histology images.. ................................ ................................ ................................ ............ 111 3 7 Potential subnuclear structure of single neurons in MRM. ................................ 112 3 8 Representative 3D gradient echo MRM of the bisected model of the biggest single neurons.. ................................ ................................ ................................ 113 3 9 3D gradient echo MRM of bisected model of single neurons illustra ting MR detectable unknown intracellular structures under investigation. ...................... 114 3 10 Diffusion weighted images of slice of the big single neurons embedded in the agarose.. ................................ ................................ ................................ .......... 115 3 11 Direct correlation of cellular architecture of single neurons of A. californica in MRM and histology images.. ................................ ................................ ............ 116 3 12 3D reconstructed image of single neurons.. ................................ ..................... 117 3 13 Diffusion tensor images and reconstructed diffusion tensor images of single neurons ................................ ................................ ................................ ............ 118 3 14 Direct correlation of single neuron in diffusion weighted MRM images, ADC maps and histology images. ................................ ................................ ............. 119 3 15 Direct correlation of a slice of single ne urons in MRM and histology images. .. 120 3 16 Pre fixed and post fixed single neurons in MRM.. ................................ ............ 121 4 1 Schematic d iagram of sample placement in the sample well on top of RF coil.. ................................ ................................ ................................ .................. 138 4 2 MRM of rat striatal tissue. ................................ ................................ ................. 139 4 3 3D segmentation image of the striatal tissue.. ................................ .................. 140 4 4 False colored light microscopic image of rat cerebellum.. ................................ 141


14 4 5 Direct correlation of ectopic Purkinje cells of the WM of cerebellum in MRM and histology images.. ................................ ................................ ...................... 142 4 6 Ectopic Purkinje cells of the WM of cerebellum in diffusion weighted M RM images with different b values.. ................................ ................................ ........ 143 4 7 MR images of motor neuron cell bodies from human spinal cord (D, E, F) and its corresponding correlative histology (A, B, C).. ................................ ...... 144 4 8 MRM images of neuronal process in the ventral horn of the pig spinal cord in MRM and 3D visualization.. ................................ ................................ .............. 145 4 9 Dir ect correlation of ectopic Purkinje cell in the dorsal horn of the human spinal cord in MRM and histology images.. ................................ ...................... 146 5 1 2D MSME MRM of normal kidneys in axial direction.. ................................ ...... 170 5 2 3D FLASH images of a normal kidney.. ................................ ............................ 170 5 3 2D MSME MRM of a normal kidney in coronal direction. ................................ 171 5 4 2D diffusion weighted MRM of a mutant kidney in axial direction.. ................... 172 5 5 3D GE MRM of a mutant kidney.. ................................ ................................ ..... 173 5 6 Representative images of volume visualization using AMIRA. A) Wild type kidneys. B) Mutant kidney using AMIRA.. ................................ ......................... 174 5 7 Volume visualization of both types of mouse liver using AMIRA and histology images.. ................................ ................................ ................................ ............ 175 5 8 Delineation of cellular structures in a kidney slice in MRM.. ............................. 176 5 9 Direct correlation of medullar region of the kidney in MRM and histology images.. ................................ ................................ ................................ ............ 177 5 10 Direct correlation of medullar cortex regio n of the kidney in MRM and histology images.. ................................ ................................ ............................. 178 6 1 Schematic of the diaphragm.. ................................ ................................ ........... 194 6 2 Schematic of the musc le system.. ................................ ................................ .... 195 6 3 Schematic of sample placement in the sample well.. ................................ ....... 196 6 4 T 2 weighted MRM of the healthy di aphragm. ................................ ................... 197 6 5 Diffusion weighted MRM of the healthy diaphragm. ................................ ........ 198


15 6 6 T 2 weighted MRM of dystrophic diaphr agm. ................................ .................... 199 6 7 Diffusion weighted MRM of dystrophic diaphragm. ................................ .......... 200 6 8 Diffusion Tensor MRM of healthy diaphragm ................................ .................. 201 6 9 Diffusion Tensor MRM of diseased diaphragm ................................ ................ 202 6 10 Representative 2D diffusion weighted MRM along sa gittal orientation of dystrophic diaphragm. ................................ ................................ ..................... 203 6 11 Four representative 3D MRM of the healthy diaphragm. ................................ .. 204 6 12 Representative histology images and MRM of healthy diaphragm.. ................. 205 6 13 Diffusion weighted MRM of the transverse view of the dystrophic diaphragm. 206 6 14 T 2 weighted MRM of a diseased diaphragm region by region. ........................ 207 6 15 Diffusion weighted MRM of a diseased diaphragm region by region. .............. 208 6 16 b0 and three representative images out of six diffusion tensor MRM of a diseased diaphragm. ................................ ................................ ....................... 209 6 17 Com parison of diaphragm in immunohistochemistry and MRM. ...................... 210


16 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EXPLORATION OF MAGNETIC RESONANCE MICROSCOPY: FROM CELLULAR STRUCTURES TO SUBCELLULAR STRUCTURES By Choong Heon Lee May 201 2 Chair: Jenshan Lin Cochair: Stephen J. Blackband Major: Electrical and Computer Engineering Due to its a bility to visualize the biological and non biological architectures and function s, in a non invasive way, MRI has gained a lot of popularity over the last four decade s E ven further, the desire to gain information about the underlying structures inside the biological tissues pushes its ability ranging from keeping track of anatomical information by simply utilizing its inherent contrast mechanism, to visualizing the structures coming from the microscopic water movement which is essential to diagnose disease in a clinical environment. However, in spite of the merit in terms of monitoring real time tracking of biological changes inside the biological samples non invasively over other imaging modalities, relatively poor resolution put s some shackles on its strength. Mainly because of its dependence on the number of H protons in a voxel and irresistible molecular movements even in the smaller voxel it reserves much room for further enhancement to compete with other imaging modalities such as light micros copy and electron microscopy as examples.


17 With the advent of state of the art technological support, there are so many efforts to get over these limitations in the Magnetic Resonance Microscopic groups all over the world. In this dissertation, by employ ing state of the art equipment higher magnetic fields, micro surface coils, and strong, fast switching imaging gradients for higher resolution, greater sensitivity and better specificity for MRM, we investigate the MR stained cellular and subcellular str ucture s inside single animal cells such as frog ova and L7 motor neuron of sea slug, and report the first delineation of mammalian cells and even individual human cells With matching the hardware capability with the need of cellular imaging, we first empl oyed the novel methodology in sample preparation such as bisected and segmented models to maximize the sensitivity of micro surface coils. The matched capability allowed us to expand and tune the applications of MRM to other cellular structures in the othe r organs such as kidney and diaphragm in the mammalian. By identifying the MR stained microscopic components in a variety of tissues using histological correlation, we present the capacity of MRM as a potential biomarker in translating the structural infor mation into the functional understanding as well as reflecting the changes in tissue compartments as the diseases progress. Here, we e xpanded its potential applications to monitor the progression of the inherited diseases such as p olycystic kidney disease (PKD) and D uchenne muscular dystrophy (DMD). Our findings suggest that MR M is a promising and powerful noninvasive tool with the limitless applicability in investigating the micro architecture function and pathology of both small animal models and humans


18 CHAPTER 1 BACKGROUND AND SIGNI FICANCE Overview of Magnetic Resonance Imaging (MRI) system Three types of magnetic fields are required for imaging: a static uniform field (known as B 0 ), gradient fields, and radiofrequency (RF) fields (known as B 1 ) [ 1 ]. The static uniform field is generated by a large magnet and its strength ranges from 1.5 T to 3 T for the head to 9 T for whole body clinical imaging, and up to 21 T for animal imaging for research applications Three types of magnets can be used: per manent magnets, resistive magnets and superconducting magnets. The superconducting magnet is the most popular type for MRI applications and consists of a solenoid of superconducting wire wound around a cylindrical bore (into which the imaging object is pl aced). The superconductivity of the magnet is maintained by liquid helium in order to keep the temperature down to 4 .2 kelvin, where the wire has a resistance of approximately zero. The second type of magnetic field, the gradient field, is produced by grad ient coil s in three orthogonal directions x, y and z These spatially localize the induced signal from the sample The need for high resolution images means that gradient coils should produce strong magnetic fields quickly. This is accomplished through delivering stronger currents into the gradient coils and using low inductance coils for faster switching times In the meantime, in order to prevent the gradient coils from overheating their temperature should be maintained by a liquid cooling system. In this work in order to achieve cellular level resolution a different geometry of gradient coil is employed and will be presented later.


19 The third type of magnetic field, the radiofrequency field, is generated by RF coils, which are located closest to t he imaging sample. The RF field (B 1 ) excites the sample by rotating at the resonant frequency of the protons within it. The RF coil also detects the accordingly induced FID (free induction decay) signal from the sample. Depending on applications, there a re a variety of coil designs available, including volume coils and surface coils. Especially in MR microscopy research, miniaturization of coils has been used successfully for achieving high signal to noise ratio by means of accommodating smaller samples and providing a strong localized B 1 The main magnet, gradient coils and RF coil s require a lot of associated electronics for their control and maintenance. These include electrical components such as RF amplifiers, gradient amplifiers, transmitters, r eceivers and analog to digital converters. All of these components must operate in harmony for the images to be generated. Once they are generated, the images are reconstructed and shown on a computer console. Background and Motivation for Magnetic Resonan ce Microscopy The ultimate goal of magnetic resonance microscopy (MRM) is to non invasively obtain pathophysiological information at the cellular level. The term MRM refers to any MR acquisition with pixel resolution finer than 100 m [180, 245 248 ]. Al though MRM techniques have yet to be applied in the clinical imaging environment, they have been successfully used in preclinical, animal, and non biological research [43, 44, 246 ]. MR microscopy techniques have been substantially improved by enhancements in the signal to noise ratio, either through increasing the signal or lo wering the noise contribution. In addition, increased imaging resolution has been obtained through stronger gradients, such as the planar gradient system introduced here later.


20 With t he advance of RF coil technologies, which allowed for reduction of the coil size, so that it was comparable to the sample size, MRM has developed rapidly over that past ten years [13, 14, 16, 18, 20~23, 33, 41~50, 82]. In applications where the sample is volume limited, solenoid type RF coils have been exploited heavily in NMR spectroscopy and MRM research. This is because of their relatively simple design, which facilitates cost effective manufacturing, and good field homogeneity, [14, 33, 43, 44 81, 82] However, in spite of these advantage s solenoid coils must be wound around glass capillaries (where the sample is inserted) and their performance h as limited reproducibility especially as the size goes down. Compared to solenoid type coils, surface coils have strong SNR only in a localized area i.e., in close proximity with the coil surface. A pplications of surface coils have been limited to micro sized samples due to their relatively poor B1 field homogeneity. In this work, by using careful sample mani pulation and positioning, we took advantage of the high sensitivity and strong B1 field in proximity to the surface coil face. To achieve higher SNR, several methodologies were employed in th is work. O n e is high magnetic field strengths ( 14 or 17 T ) The second is reduc tion of the coil dimension and utiliz ation of the surface coil design. The third is minimiz ation of the loss coming from T 2 effect s, by utilizing a planar gradient system, which offers faste r switching and higher gradient field strengths (Fi g ure 1 7 ) Types of RF C oils Coils are composed of inductive and capacitive elements. The resonant frequency, 0 of an RF coil is determined by the inductance (L) and capacitance (C): (1 1)


21 To match the resonant frequency of the coil to that of the s ample, tuning can be performed by adjusting a variable capacitor. The sensitivity of the coil, or intensity of the received signal is described by the quality factor, Q: (1 2) A higher coil Q factor leads to increased SNR of the acquired signal, re sulting from the more sensitive detection of the time varying FID signal. Along with the Q factor, optimization of coil geometry relative to the sample size is also important to increasing SNR. When the patient or imaging object is placed inside the coil, RF eddy currents within them will produce randomly varying magnetic fields, which will induce noise voltages in the coil [2] This noise will decrease as the volume of tissue within the coil decreases. Therefore, reduction of the coil size will improve t he SNR of the resultant images, but in turn it will reduce the coverage. At higher field strengths there are additional effects, which must be considered in coil design. For instance, as the wavelength of the RF field becomes comparable to the sample dimen sion, travelling wave effects disrupt field homogeneity. However, with the small sample sizes used in microscopy applications, this is not a concern. Furthermore, eddy currents induced inside the conducting coil will push the current to the coil perimete r and make it crowded around there. This is referred to as the skin effect, which causes higher resistance in the straight wire, and is more pronounced at high field because the skin depth is inversely proportional to frequency. RF coils can be roughly d ivided into two groups: volume coils and surface coils. Generally volume coils can cover a larger volume with uniform B1 field inside the region


22 of interest. Common types of volume coils are birdcage coils [ 3 ] saddle coils [ 3 ] and transverse electromagnetic ( TEM ) resonators [ 4 ] (Fig ure 1 1) Two types of microcoils utilized in MR microscopy are solenoid coils and s urface coils (Fig ure 1 2) S olenoid coils (Fig ure 1 2 A ) are commonly used due to their highly uniform distri bution of the magnetic field inside the coils, high efficiency and low cost. The magnetic flux inside along the central axis of a solenoid coil is given by: (1 3) w here d is the coil diameter, N is number of turns, 0 is permeability of free space and h is the coil height [10] If applications require high SNR and the sample can be positioned very close to the coil sur face, surface coils are favorable (Fig ure 1 2 B and C ) Surface coils shown in Fig ure 1 2 are conventionally designed with single (Fig ure 1 2 B ) or multiple surface conductor loops (Fig ure 1 2 C ) While offering a high B 1 field strength in proximity to th e coil face and highest sensitivity, they have a weakness in spatial uniformity of the B1 field distribution. Most of the acquisitions in these works were carried out with multi looped surface coils, which have better B 1 field uniformity than single looped surface coils but their construction is much more complex. Wider coverage with better B 1 field uniformity can be obtained by a Helmholtz type coil which consists of two single loop surface coils with a space between them, into which the sample is inserte d (Fig ure 1 2 D ).


23 Principles of Magnetic Resonance Imaging T o make MR images the magnetization should be encoded spatially. This is accomplished u sing the magnetic field gradients The spatial and frequency domains are related in the following manner: (1 4) R eciprocally, (1 5 ) w here are the frequency domain coordinates (in units of m 1 ) and S( k x k y k z ) is the Fourier transform of the spatial magnetization distribution, (x ,y,z ) k is the reciprocal space vector ( ) [5 7] In order to encode multi dimensional imag es gradients fields in three orthogonal directions are required (x, y, and z, or frequency encod e phase encod e and slice select ) The Larmor frequency for spins at position r =(x,y,z), will therefore be : (1 6 ) ,where G refers t o the gradient vector, constructed from the gradient amplitudes in the three orthogonal directions: G =[G x G y G z ]. T here are many available contrast mechanism s in MRI. These include T 1 T 2 T 2 *, and diffusion. Different pulse sequences can be used to obt ain images with unique


24 contrasts. The most conventional pulse sequences are g radient e cho (GE), and s pin e cho (SE). In a g radient e cho pulse sequence as shown in Fig ure 1 3 the s p ins are dephased and rephased purely by gradients, i.e. bipolar gradients. After the excitation RF pulse, the negative gradient lobe will dephase the transverse magnetization. The positive lobe (known as the readout, or frequency encoding gradient) will then rephase the magnetization, with an echo forming at the center of the fre quency encoding period, when the total gradient area equals zero. Another commonly used pulse sequence is the s pin e cho (SE) pulse sequence [ 8 ] (Figure 1 4) which employs two RF pulses : one for excitation ( /2 radians) and one for refocusing ( The excitation pulse will rotate the magnetization vector will rotate down to the transverse (or x y ) plane. Then, due to field inhomogeneities, the transverse magnetization will dephase at a rate of for the duration TE/2 At TE/2 the refocusing will rotate the dephased magnetization about the x axis. Due to the refocusing pulse, the magnetization will begin to rephase, forming a spin echo at time TE [8] The main difference between gradient and spin echo sequences is that gradient echo signal decay is a function of ; however, spin echo signal decay is a function of A variant of the s pin e cho imaging sequence is the d iffusion w eighted i maging (DWI) s equence [9] (Fig ure 1 5) Diffusion weighted imaging is sensitive to the microscopic translational motion of water molecules within tissues and living organisms [9, 27] A spin echo sequence is modified into a DWI sequence by centering a pair of


25 equal area gradient lobes symmetrically around the refocusing pulse. The first gradient lobe will induce a phase shift, and after some time interval (the diffusion time, ) the second gradient will induce an equal phase shift in the opposite direction (due to the refocusing pulse) P rotons which did not change their location between the two gradient pulses will not accumulate any phase ; however, the protons which change d position will dephase. D ifference s in diffusivity of water mol ecules in tissue, results in high microstructural MR contrast in biological systems. T he mono exponential attenuation of MR signal in a DWI sequence was modeled by Stejskal and Tanner in 1965 [9] : (1 7 ) w here is b value or diffusion weighting factor, g is the gradient strengt h of diffusion sensitizing gradient s is the interval between two gradients is the duration of the gradients and D is the apparent diffusion coefficient (ADC) T heory of SNR Calculation of SNR The performance of an NMR system is frequently judged by the image SNR SNR is defined as ratio of the peak signal value to the rms (root mean square) of the noise voltage as shown below [10] (1 8 ) w here k 0 is a proportiona l ity constant correcting for the non uniformity of the RF field, B 1 / i is the ratio of the induced magnetic field to the applied current (known as sensitivity) Vs is the sample volume, N is number of spins per unit volume, is the gyromagnetic


26 ratio which depends on the nucleus ( for hydrogen, = 42.58 MHz/T ) is planck s constant, is Boltzmann s constant, I is the quantum number of nuclear spin ( for proton, I = ) is the angular fr equency, T is absolute temperature, and is the summation of coil and sample noise voltage. Th e e quation 1 5 suggests multiple strategies for SNR i mprovement, such as enhancement of coil sensitivity, increase in sample volume, or decrease in n oise voltage. The SNR per unit volume depends on the voltage generated by a voxel inside the sample and can be described as : (1 9 ) w here is the transverse magnetization and is the frequency. Accordingly, the SNR in MRM can be expressed as : (1 10 ) w here is a factor depending on the pulse sequence and AT is the acquisition time. In consequence, in order to increase the signal voltage, we have to increase the transverse magnetization (M xy ) voxel volume (V x ) frequency ( 0 ) or scan time (AT) Sources of Noise Traditionally, ohmic losses had been regarded as the single contributor to noise in MRI but the coupling between coil and sample should also be consider ed [11] Noise can come from the abundant charges in conducting materials such as cables and c oil wires, where electrons mov e around and collide because of th ermal energy which results in magnetic field fluctuations. Another source of charges is the ions in the


27 biological tissues, which are thermally agitated and interact with each other by Coulomb forces. Coupled into coils even through the air in between th e thermal motion of charged particles induces an electromotive force (EMF) in the coil The electromagnetic properties of biological samples are characterized by the permeability, permittivity, and conductivity. The permittivity and conductivity of a s ample can significantly contribute to the noise due to dielectric loss es, which occur when the electric field coming from the coil interact s with the sample. This is the dominant source for noise in clinical MRI. An additional charge source is d ipolar mole cules such as water molecules which are highly dielectric have an accompanying dipole field However, considering the high frequencies used in MR, this effect is not strong. I nductive losses are mostly inevitable Some r esearchers have calculated the eff ective sample resistance and concluded that it scales with where is the Larmor frequency and r is the radius of the coil [12] Therefore, a smaller coil radius would go further towards reducing the sample resistance than a higher main magnetic field The noise voltage is a consequence of components in the magnetic bore, e.g., imaging objects, cab les, and coils. This is described by the equation below : (1 1 1 ) where BW is the receiver bandwidth, Rc is coil resistance, Rs is sample resistance, Tc is coil temperature, and Ts is sample temperature. R environ and T environ are the environ mental resistance and temperature which are usually negligible but potentially significant at high field RcTc is the dominant term for the small coils and samples used in animal research whereas RsTs is dominant in clinical MRI


28 Avenues for Noise Reduct ion Therefore, to minimize the coil contribution to the noise voltage, either Tc or Ts can be reduced. However, decreasing the temperature of the sample is not easily accomplished. L owering the temperature of the coil can reduce the thermal noise contribut ion resulting in enhanced signal to noise ratio Alternatively, reduc tion of resistance can be achieved by utilizing low resistance materials for the coil wires. F or example, silver exhibits lower resistance than copper but is more expens ive [13] Otherwise, the bandwidth (BW) could be reduced at the expense of increased minimum TE. R esearch [14] on Q factor differences between sample loaded and sample unloaded micro solenoid coils, revealed virtually no change between the two conditions This led to the conclusion that at fixed temperature, the SNR per unit volume depends only on the coil especially when the diameter is less than 3 mm [10]. (1 1 2 ) In this regime, the h igher coil conductivity reduces the coil noise contribution; however, it can also give rise to an increase in the sample noise contribution. Therefore, the SNR is inherently limited by the coupling between the coil and sample [15] H igh Field Effects Advantages of employing the high field magnet include increased SNR, and higher contrast to noise ratio for many applications. When the noise from the coil is dominant the SNR in MRM is supposed to enhanced in proportion to B 0 (7/4) [10]. For example, by increasing the magnetic field from 14 .1 T to 21.1 T, the SNR is expected to be improved approximately by more than 50% with similar coils However, the


29 disadvantages include higher specific absorption rate(SAR) [11, 16] macroscopic susceptibility effects, and short RF wavelengths relative to sample size. Additionally, since MRM is carried out at the frequencies ranging up to 900 MHz, in order to calculate resistance, high fre quency effects such as the skin effect and proximity effect should be taken into account [17, 19 30 ] These are critical considerations, especially for micro coil manufacturers who favor smaller coil dimensions. B oth of these effects could increase resi stance by reducing the effective cross section of the conductor When the AC current pass es through the conduct or it induces an alternating magnetic field. The changing magnetic field will cause eddy currents in the wire resulting in current flow being limited to the periphery of the conductor Therefore, the current density within the wire decrease s exponentially with distance from the surface [1 7] The distance from the surface of wire at which the current density is reduce d by 1/e is called the skin depth which is given by the equation below [17] (1 1 3 ) where is the operating frequency and is conductivity Moving towards higher field, t he reduction of effective cross section is proportional to the square root of the f requency. E ddy current s are not only induced in the wire itself, but also in neighboring conductors. The result is that current becomes concentrated in areas of the wire which are furthest away from adjacent conductors. This is known as the proximity ef fect [ 67]


30 which leads to non uniform distribution of the current and can significantly increase resistance S pin Excitation Relaxation, and Signal Detection Spin Excitation At equilibrium the bulk magnetization vector, M 0 is aligned in the z direction parallel to the main magnetic field, Any magnetizatio n components which are not aligned parallel to B 0 will precess around at the Larmor frequency ( = B 0 ) A left circularly polarized radiofrequency field, B 1 oriented perpen dicular to B 0 and alternating at the Larmor frequency applies a torque, which rotates the magnetization vector away from z axis by an angle, The rotation angle is defined as, (1 1 4 ) w here is the RF pulse duration. 90 the magnetizat ion vector, M, will be in the transverse (x y) plane, where it precesses about B 0 The precessing magnetization vector will induce an EMF signal in the coil, which is called free induction decay (FID). Spin Relaxation After the RF pulse perturbs it, the spin system will return back to its thermal equilibrium. The transverse magnetization will decay, while the longitudinal magnetization grows back to its equilibrium state (M 0 ) The recovery of longitudinal magnetization occurs with the rate constant T 1 k nown as the spin lattice relaxation time. Longitudinal relaxation involves the exchange of energy between the nuclei and the surrounding environment, known as the lattice. T he decay of transverse magnetization occurs with the rate constant T 2 known as the spin spin or transverse


31 relaxation time. Transverse relaxation is caused by the dephasing of magnetic moments due to interactions between spins. Transverse and longitudinal relaxations are described by the Bloch equations, as shown below : (1 1 5 ) (1 1 6 ) The solutions of these two equations are: (1 1 7 ) (1 1 8 ) w here are the initial values of longitudinal and transverse magnetization, respectivel y It should be noted that in an inhomogeneous external magnetic field without t he help of refocusing pulses, transverse magnetization will decay with the rate constant T 2 which is shorter than T 2 The relationship between T 2 and T 2 is given by: (1 19 ) = B0 stands for the additional contribution of inhomogeneous static magnetic field to transverse relaxation [6] Signal Detection The NMR signal strength directly depends on the distance from the coil center and the highest sensitive regio n is inside the half sphere on the front which has the diameter of coil where the sample will be positioned. M ore coil loops provide, greater sensitivity


32 [ 39 ]. However, in the micro coil domain space is limited, which, in turn, limits the number of turns. For example, the 500 m micro surface coil, which was used in our experiment s is equipped with a sample well which is only 5 mm wide and 500 m deep By converting the time domain NMR signal to the frequency domain (through a Fourier transform) we obtain the spectrum of the NMR signal The spectrum provides important information such as chemical shifts, structural properties and environmental information about the elements under investigation. For example, the hydrogen atoms precess at different frequen cies due to environmental factors, such as shielding effects from electrons and other neighboring molecules. Gradient Coil: Its role in Image Formation and the Resolution Limits In the MRI system, gradient coils, which are located outside the RF coil, are essential for spatial encoding of the MR signal. Fundamentally, gradient coils should produce magnetic fields in three orthogonal directions, which vary as linearly in space as possible. Especially in MRM, in order to make imaging in the micro scale reali zable, several issues should be considered. These include the effects of diffusion during gradient pulses and eddy currents. In order to cover these issues, the fundamentals of gradient coils and the requirements for cellular imaging will first be discusse d. Considerations for High Resolution Images Conventionally, gradient waveform s are trapezoidal in shape, with a gradient strength G and rise time as shown in Fig ure 1 7 A The spatial resolution in the reconstructed image is inversel y proportional to the area of gradient strength and time of the gradient lobe, which is shown in an equation as below.


33 (1 20 ) To minimize susceptibility and diffusion effects, gradient pulses should be designed with maximum amplitude and minimum durati on. Short rise times are therefore important. Gradients should therefore be built with high efficiency and low coil inductance giving rise to a fast slew rate, which is defined as the rate of the change of the gradient field with time and has units of [T /m/s]. The time varying gradient magnetic fields used in MRI induce s eddy currents in neighboring conducting materials. The eddy currents create unwanted exponentially decaying magnetic fields, which distort the gradient waveforms, resulting in artifacts such as stretching or shrinking of the corresponding image. For micro imaging, shielding of gradi ent coils, where an outer coil surrounds the main gradient coil and produces the opposing magnetic field, is desirable. This prevents the magnetic field f rom spreading beyond the outer coil and inducing eddy currents in the external conducting objects [20] In applications requiring microscopic resolution any potential caus es for spatial distortion should be eliminated wherever possible In addition, since SNR is limited, strong gradients which can ramp up quickly can be used to minimize echo time and thus decrease signal losses due to T 2 or T 2 decay. This is especially he lpful in diffusion weighted imaging, where additional T 2 contrast is undesirable In this work, the benefits of a high strength gradient are realized by employing a newly designed planar gradient system by Bruker ( Z110828, B6406 ), which is capable of produ cing a maximum amplitude of 66 T/m. This is significantly higher than the conventional Micro 5 gradient system, which has a maximum strength of approximately 3 T/m (Fig ure 1 7 )


34 Limitations of Resolution and Gradient Requirements T 2 l imited r esolution Th e resolution of an MR image can be affected by T 2 relaxation during readout encoding, since the spectral linewidth is inversely proportional to T 2 (i.e. linewidth= ( T 2 ) 1 ). However, if the bandwidth per voxel (i.e. BW/number of pixels in the readout direc tion) exceeds the spectral linewidth, this effect is insignificant. When the bandwidth per pixel is smaller than the line width, the T 2 limited resolution is given by: (1 2 1 ) which is only applicable to frequency encoding direction of imaging [7, 42]. I n micro scopy images, magnetic susceptibility effects could become very prominent resulting in signal loss due to effects Bandwidth l imited r esolution T he resolution in phase encoding direction is given as (1 2 2 ) where where is the gradient pulse duration Under the assumption of the high SNR and ideal detection system, i n order to differentiate two different regions, the phase encoding is required in a way that two voxels have different phases. The phase difference between two voxels is (1 23 ) Note here is the unit of is rad/s [ 21, 23 ].


35 Considering the minimal detectable phase is for Fourier transform r econstruction, the bandwidth limited resolution is [21, 22] (1 2 4 ) By th ese two resolution limits, the acquisition time should be significantly smaller than to avoid the filtering effects from relaxation property. Therefore, the strong gradient and the short acquisition times are solutions, but under given conditions restricted by essential imaging pulse sequences, the strong gradient is a more practical solution than the shrinkage of the acquisition time. Diffusion l imited r esolution Another fundamental limitation to the image resolution is the motion of molecules within the imaging objects This is called diffusion limited resolution. It is very difficult to control because the Brownian motion of molecules in the sample [38] is due to interactions between ions and dipolar molecules I nteraction s with the local magnetic field cause image blurring or smearing resulting in line broadening effects. Assuming the random walk model, th e mean square displacement of water with diffusion coefficient, D in time, t is (1 2 5 ) Diffusion during the readout gradient causes signal attenuation and linewidth broadening, which can limit the resolution. Another calculation of diffusion re solution [21, 23] can be described with unbounded Brownian motion [38] for acquisition time, AT with self diffusion coefficient, D would be


36 (1 2 6 ) The attenua tion of the signal due to diffusion during the readout gradient is given by: (1 2 7 ) This attenuation broadens the signal by (1 2 8 ) where is the FWHM of the signal [7]. Due to this attenuation, the resolution limits by diffusion is (1 2 9 ) From above considerations, despite the controversial role of the acquisition time, the only solution to overcome the limits is the strong gradient. Then, h ow high gradient strength is required? From those calculations above, the requirement of gradient for overcoming the diffusion limited resolution essential for cellular imaging is (1 30 ) which is obtained by combining (1.21) and (1.23). Calculations:


37 In case of free water whose D is around for a 6 ms acquis ition time, the resolution is limited to 5 m. For cellular imaging where resolution should be less than or equal to 5 m, these resolution limits must be taken into account If the resolution to achieve is 5 m, then the readout duration should be about 6 ms as If the required resol ution is 1 m, then the acquisition time should be less than 227.3 s The gradient strength required for 5 m is 138 mT/m based on the calculation as : On the other hand, for 1 m resolution, the gradient strength should be larger than 172 5 0 mT/m based on the calculation : The situation becomes e ven more difficult when dealing with water diffusion inside the biological tissues In biological tissues, factors such as cell density, cell size, membrane permeability, and intra/extrace llular volume fractions can affect the diffusion coefficient. Assuming a diffusion coefficient in tissue of 1.1 x 10 9 m 2 /s the acquisition time should be no greater than 454.6 s The gradient strength should be greater than 8600 mT/m calculated by :


38 Diffusion based C ontrast M echanism A prominent advantage of using diffusion weighted imaging instead of proton density, T1 or T2 weighted imaging is the outstanding contrast offe red by DWI. DWI techniques can be used to visualize anatomical architecture, especially which causes the anisotropic movement of water, for example white matter in the central nervous system, muscle fibers, etc. The microscopic movement of water is affect ed by pathological and physiological changes in the underlying microstructure. In certain cases, such as acute stroke, the root cause of the pathological changes in diffusion properties is still under debate [ 230 242 ] By performing simple perturbations o n simple biological systems, (i.e. inducing cell swelling in the L7 neuron of Aplysia c alifornica by placement in a hypotonic solution), we can obtain information about the cellular origins of pathological changes in diffusion properties [81]. Diffusion We ighted Imaging Under the influence of a constant gradient, molecular diffusion will lead to a decrease in the amplitude of the water signal [8, 26] T he pulsed gradient spin echo (PGSE) experiment (Fig ure 1 5) make s diffusion behavior measur able by MRI [9] T he hydrogen protons will diffuse randomly through space during the diffusion time After the second diffusion gradient after will rewind the phase accumulated by those spins which did not change position H owever, s pins which displaced in the diffusion gradient direction during will accumulate phase. The loss of phase coherence amongst moving spins will result in signal decay. By comparing the signal amplitude in the presence and absence of a diff usion sensitizing gradient pair the a pparent d iffusion c oefficient (ADC) can be measured. In complex biological tissues, structural fe atures present hindr ances or restrict ions in the paths of water molecules. For instance even if


39 a spin can move 5 m in one direction in free water in a biological tissue the net displacement may be ended up with only 3 m due to reflect ion s caused by cell membranes org a nelles and the like. Therefore, the real total diffusion distance is 5 m despite the 3 m of app arent displacement When it comes to cellular imaging, this phenomenon should be dealt with carefully. D uring a magnetic field gradient, the phase accumulated by a single spin in the transverse plane is given by: (1 31) w here is the gyromagnetic rat io, is static magnetic field, G is the diffusion gradient amplitude, the gradient duration is and r is the spatial location of the spin. Following the second gradient pulse, the phase shift of a moving spin, depends on the spin displac ement in the direction of the diffusion gradients and is described as : (1 32) After the first diffusion gradient pulse, the spin phase is: (1 33) where is the initial spin location [27] The phase accumulated during the second gradient pulse, is : (1 34) where is the spin location during the second pulse.


40 With no net displacement, the phase shifts of static spins before and after the refocusing pulse will cancel However, with net displacement, the phase will not cancel resulting in net deph a sing The amount of dephasing is affe cted by parameters such as the gradient strength G, diffusion duration and diffusion time The acquired signal is the vector sum of the magnetic moments of the individual spins T he aforementioned spin dephasing will cause a reduction of transverse ma gnetization which is described as : (1 35) where is the equilibrium magnetization and is the phase of the j th spin The equation for the resultant signa l decay was derived by Stejskal and Tanner [9] and is given by: (1 36) where S is the signal strength after diffusion time and is the signal strength in the absence of diffusion gradient s It should be noted that this equation is only suitable for trapezoidal unipolar or bipolar gradients of short duration It also ignore s the in teraction between imaging, background, and diffusion gradients [9] In addition, under the narrow pulse approximation the signal attenuation depends only on the net displacement in the gradient direction and not on the actual diffusion path A popular


41 way to characterize water diffusion inside biological tissues is the measurement of the slope of the logarithm of the signal intensity known as the Apparent Diffusion Coefficient (ADC) : (1 37) Compartmentation Issues An i mportant consideration is that this diffusion imaging is inherently a one dimensio nal measurement, as it measures the signal decay due to diffusion in a single direction. However, molecular motion occurs along all three dimensions. W hen dealing with molecular motion in an anisotropic medium, a single one dimension al ADC measurement wil l not be sufficient .Things get more complicated as the diffusion weighting, or b value gets higher due to longer diffusion time s and /or higher gradient strengths Consequently, to describe biological systems which contain various barriers, the issue of m odel ing becomes important [ 25, 28] Rather than describing diffusion signal decay as mono exponential a bi exponential model with two compartments having independent ADCs, can be applied: (1 38) w here are the fraction al sizes and diffusion coefficients for eac h r espective compartment. Alternatively, there are multi compartment al models, which assume many more than two compartments: with


42 (1 39) Still evolving, th ese compartmental ideas confronted many issues for meeting the need for explaining the compli cated diffusion behavior inside the biological tissues [32, 33] For clos er to accurate explanation, the more signal should be attained by avoiding any averaging contamination or partial volume effects from neighboring regions which are getting trickier in cellular imaging. Great SNR with high resolution in short diffusion time s should be guaranteed to manage to significant quantitative analysis of diffusion cellular imaging, and in thi s works, high strength gradient using planar gradient w as employed (Fig ure 1 8 ) Diffusion Tensor Imaging Diffusion weighted imaging which measu res a single scalar diffusion coefficient would be sufficient characterize one dimensional random transport In reality, however, water can diffuse in any direction and in biological tissues, diffusion is often anisotropic. To describe the three dimensi onal diffusion process, the diffusion tensor can be used. [34 36] To describ e orientati onally dependent diffusion (Fig ure 1 8 ) DTI techniques use a symmetric 3x3 tensor as shown below. (1 40) which c onsist s of 6 independent elements T he diagonal elements are diffusion coefficients along the x, y, and z directions in the laboratory frame T he off diagonal


43 elements contain information about the rotation of the principal diffusion axes relative to the laboratory x, y, and z axes [37] This tensor representation offer s six independent terms to descr ibe the translational displacement of water molecules instead of a single ADC value as in diffusion weighted imaging [35, 36] In an anisotropic medium, the signal decay due to diffusion is related to the product of the D ma trix and the b matrix as follows : (1 41) w here is elements of diffusion te n sor, and Ideally, the b matrix should have only diagonal elements, related to the amplitudes of the three orthogonal diffusio n sensitizing gradients. However, in reality, there may be off diagonal elements or cross terms, reflecting the interactions between imaging and diffusion gradients. Experimentally, to characterize molecular motion in anisotropic medium, seven different me asurements are necessary, i.e., one without diffusion weighting and six more along six independent axes to determine the tensor elements: Once the D matrix is determined, it can be diagonalized to provide the eigenvalues, 1 2 3 and eigenvectors, 1 2 3 which characterize the diffusion ellipsoid. Following diagonalization, the diffusion tensor can be represented as follows :


44 (1 42) The three positive eigenvalues, represent the diffusion coeffici ents along the three orthonormal eigenvectors, T he primary eigenvalue, represents the diffusion coefficient along the principal axis of diffusion and re present water diffusivit i es along the two orthogonal direc tions After determining the eigenvalue s and corresponding eigenvectors from each voxel, the spatial anisotropic information can b e mapped with DEC ( Directionally Encoded Color ) [29] This information can also be used for fiber tracking. Anisotropy Index T here are several anisotropy ind ic es to show the degree of diffusion anisotropy, but two popular ones are RA (relative anisotropy) and FA (fractional anisotropy) which are dimensio nless and rotationally invariant quantities [8,10,11]. Mathematically, it is equal to the standard deviation of the principal diffusivities divided by their mean: (1 43) The fractional anisotropy is given by:


45 (1 44) Therefore, FA = 0 is in an iso tropic medium where the diffusion ellipsoid is a sphere and FA=1 in a fully anisotropic medium where diffusion occurs in only one direction. DTI based techniques, which are inherent ly noninvasive provide a better understanding of the complex, three dime nsional diffusion of water in tissues, which can be used to infer functional nerve connectivity. Furthermore they offer information on how the diffusion process is affected by pathological changes. High Angular Resolution Diffusion Imaging (HARDI) Since the inception [ 36 ], D iffusion tensor imaging (DTI) has drawn so much attention as an attractive tool in characterizing the isotropic or anisotropic diffusion by utilizing the rank 2 tensor [ 35 ] As shown in Fig ure 1 8 i n addition to presenting three diffe rent orthogonal directions through each eigenvector, the relative mobility can be described by its own eigenvalue, and the highest eigenvalue can be regarded as the dominant direction of the local fiber tracts. However, due to some main assumptions, DTI ha s its own shortcomings. Specifically, the dominant fiber direction at each voxel can be i llustrate d by only one orientation maximum indicating the direction where the highest diffusion occurs This is originated from the assumption of the underlying diffus ion displacement probability function as a single 3D Gaussian distribution whose probability distribution of carrying the molecules in the distance r at the fixed time constant t has zero mean and the covariance of 2Dt. The consequent shortcomings


46 are first, there might be possibility of inaccurate estimate t he true fiber direction s by a single fiber orientation especially in dealing with underlying multiple fiber orientations in in tissues such as brain or mus cle tissues [ 262, 263 ]. As shown in Fig ure 1 9 more than one direction of fiber cross with each other, e.g. the first eigenvector cannot correlate with any one of the actual or anatomical fiber orientations. Even with the recent hardware development, ther e is still limitation in resolving this issue along with partial volume effects when resolving the crossing of fine multiple fiber tracts in the brain. In order to make up the aforementioned drawbacks of conventional DTI model, more advanced image acquisit ion strategies and sophis ti cated reconstruction methodologies were introduced to estimate the orientation distribution function (ODF) of multiple fibers, for example, diffusion spectrum imaging [ 249, 250], Q BALL imaging [ 251 254 ] deconvolution based tech niques [ 255 257 ] Among them, high angular resolution diffusion imaging (HARDI) makes an effective presentation of the angular distribution of apparent diffusivity along many directions of diffusion gradients by sampling the shell of the fixed radius [258] and converting the estimated distributed diffusivities to elements of a spherical tensor by employing the spherical harmonic transformation [ 259 ] and later utilizing the generalized diffusion tensor model [ 260, 261 ] In our data, for the stable measuremen t of diffusion anisotropy in kidney 21 directions of diffusion gradients were considered. Then, by tracing along the preferential water diffusion in the fibrous tissues in whiter matter, muscle, or kidney tissue, the non invasive fiber tracking can be com pleted. The dominant direction is the one which has the highest eigenvalue [ 264, 265 ]. The resultant fiber pathway can be completed by


47 assembling several regularizations by stream line estimation, e.g. Fiber Assignment by Continuous Tracking (FACT) method [ 265 ]. Using the algorithm, there was the first correlation of cellular level diffusion tensor tractography with the corresponding histology of a rat spinal cord [ 266 ]. As such, researching on the more accurate and sophis ti cated description of orientation of the fibers in the complex local geometries has been a significant and still evolving technique in the medical imaging field.


48 Figure 1 1. Schematic of RF volume coils A) birdcage coil, B) saddle coil, C) Helmholtz coil D ) TEM resonator


49 Figure 1 2. Schematic of RF micro coils A ) S olenoidal micro coil B ) s ingle loop surface coil C ) multi loop surface coil D ) Helmholtz type surface coil.


50 Figure 1 3. Schem atic o f gradient echo pulse sequence. The gradient echo is formed by a pair of bipolar gradient pulses.


51 Figure 1 4. Schematic of spin echo pulse sequence. The spin echo is formed by regaining the spin coherence by a refocusin g 180 pulse.


52 Figure 1 5. Schematic of a d iffusion weighted pulse sequence and the corresponding MR signal characteristics between bounded and free water in terms of signal intensity.


53 Figure 1 6. Schematic of transmitting and receiving the RF signal on resonance A ) Illustration of signal excitation and reception by RF resonator, and the protons in the water in static magnetic field, B 0 and B ) after B1 excitation by RF coil, the proton get f lipped into the transverse plane at its resonant frequency, C ) the precession back to equilibrium and D ) the resultant signal of the FID in time domain and Lorentzian shape in frequency domain.


54 Figure 1 7 Pictures of a micro planar gradient. A ) and B ) The pictures of the p lanar g radient also named as a compact gradient is presented. C ) The picture presents the planar gradient which is mounted and plugged onto the micro 5 probe body.


55 Figure 1 8 Schematic of water diffusion in an isotropic and an anisotropic medium A ) D iffusion sphere representing average constant probability distribution of molecular displacement in an isotropic medium, and B ) diffusion ellipsoid representing constant probability distribu tion in an anisotropic medium.


56 Figure 1 9 Schematic of comparison between DTI and HARDI. A) The schematic of water molecules (yellow circle) move in the underlying fiber oriented in the horizontal direction (in orange) and crossing with another fi ber oriented in the vertical direction on the right (in blue) B) and C) Both of DTI and HARDI can present the consistency in probing a single direction (left to right) but the primary eigenvector in DTI in B fails to represent any of the true underlying f iber direction.


57 Table 1 1. Comparison between a micro 5 gradient, and a planar gradient. Gradient Type Amplitude min TE for DWI pulse sequence Micro5 Gradient 295G/cm (=2957mT/m) 15 ms P lanar Gradient 6800G/cm (=68000mT/m) 1~2 ms Table 1 2. Diffusi on coefficient values of free water and biological water, and their corresponding gradient strength required. Types of Water Resolution D value Gradient Strength Required Free Water 1 m 12 000 mT/m Biological Water 1 m 8600 mT/m


58 CHAPTER 2 MAGNETIC RESONANCE M ICROSCOPY OF SUBCELL ULAR STRUCTURE IN XENOPUS LAVIS OOCYTES Introduction A breakthrough in magnetic resonance microscopy ( MRM ) to visualize a single animal cel l occurred in 1986 when the resol ution achieved in MR images of a single cell (frog ovum, diameter ~ 1 mm ) revealed cellular compartments [41]. Such results were achieved by conducting imaging with in plane resolution on the order of tens of microns. The s tage IV ovum was visualized with hyperintense MR signal of the nucleus surrounded by hypointense signal of the cytoplasm (resolution = 10 x 13 x 250 3 ). Since then, a variety of MRM applications have been devised by various research groups for probing both biological and non biological samples [42 44] One suc h method involves a different animal model : that of the sea slug Aplysia c alifornica Its largest neuron the L7 (diameter = 300 is located inside the abdominal ganglion and, like the xenopus oocyte, is readily accessible through gross dissection. This cell was originally imaged at a resolution of 20 x 20 x 1 0 3 [20]. This is the first neuronal cell that became a wide ly used invertebrate animal model for comparative and behavioral brain research as a primitive model of the mammalian central nervous system [66, 179] Thanks to the roughly spherical geometr y, the se MRM images were carried out in solenoid coils that were relatively simple to design and implement and possessed high signal uniformity. Using MR images with in plane resolution s on the order of tens of micron s delineation of mammalian cells (diameter ~ 5 that c an not be excised like those in the aforementioned aquatic models could not be realize d However, with hardware advancement s e.g. higher magnetic field strength s stronger and faster switching


59 gradients [45] and micro surface coils whose magnetic field is non uniform but has the highest proximal sensitivity [18, 46 49] mammalian cells were finally delineated using MRM methods [50] and open ed up a new chapter in MR based c ellular imaging Since one of the major gears behind the evolution of MRI into the microscopic level has been driven by improvements in resolution along with other improvements such as sensitivity MRM was originally defined as MR I at or ution in at least two of three spatial dimensions [180] Even though resolution is not the only criteri on necessary for MRM from the view that resolution is necessary to resolve smaller objects inside a single voxel, it is the single most important criter ia for visualizing heretofore unseen structures. In the course of improving MRM, there are still barriers to overcome. First, it comes from its poor inherent sensitivity. Given the microscopic resolution employed in MRM the number of resonating spins is limited by voxel size Additionally, t he magnetic interaction between protons gets higher as they move into higher magnetic field s, therefore any inhomogeneities in the static field result in broadening of the spectral linewidth in the frequency domain. Th is can be minimized by optimizing hardware and software to obtain a uniform magnetic field and improved by develop ing RF coils whose geometry offer an optimal filling factor and improved sensitivity Among several ways to improve the signal to noise ratio the chiefly used methods are as follows. The SNR can be increased by suppressing the noise voltage through a reduction in coil temperature. Alternatively, in high resolution MRM imaging with fixed biological samples, lowering the sample temperature (prov ided it is kept above freezing) will increase SNR. When employing ever small er voxel s the temperature dependency in the temporal and spatial


60 domain becomes prominent. Some of th e protons can exit the voxel in the course of a single TR due to thermally dri ven Brownian motion, resulting in additional signal loss. This is in addition to T2 driven signal loss due to the temporal requirement implementing all imaging pulse sequence By employing a gradient strong enough to overcome the limitations caused by magn etic field inhomogeneities (T2* loss), as well as bandwidth and diffusion related signal loss Care must be taken during positioning when using micro surface coils or an unintended structure can dominate an image. Such structures include unwanted tissue tears or holes which are often caused by physical damage such as mechanical shearing force in sample preparation In this study and for the first time to our knowledge, subnuclear structure s were investigated using MRM We attempted to image nucleoli in the stage IV of frog ova using a 500 m ( Bruker Biospin Z76414) and 100 m ( Bruker Biospin Z76412) micro surface coil (Fig ure 2 1). T he potential of microscopic correlation was also investigated Background information including descriptions of known co mponents of the xenopus nucleus is revisited briefly visual prominence as a sub cellular compartment in this frog ovum comes from the inherently different refractive indices between it and the surrounding nucleoplasm This difference is disti nguishable under phase contrast microscopy (Fig ure 2 2) At stage IV, the frog ova require high protein levels for the multiple stages of cell division s associated with mitosis following fertilization. For effect ive production of protein, the cells require an elaborate assembly line of ribosome s i.e. the nucleolus The xenopus oocyte contains 1400 1600 nucleoli in a single cell which is approximately 25~30% of the entire volume of the nucleus [51, 52] Moreover, the large nucleoli


61 congregate a t the periphery of the translucent nucleus [53] and each contact s a nuclear pore to transport protein subunit s To test for MR detectability, the internal constituents of the nucle us were examined Results of these imaging experiments are descr ibed in Fig ure 2 3 The shape of the nucleoli in frog ova is primarily mono or bi spheroidal unlike figure eight shape found in animal s such as Aplysia c alifornica [54] Depending on morphological characteristics they can be classified into four types : ring shaped nucleoli nucleoli w ith nucleolonema [55] compact n ucleoli, and micronucleoli, whose overall constituents have no membrane. In a single nucleolus, the fibrillar center (FC) is surrounded by the dense fibrillar component (DF). Both of these structures are surrounded by the granular component (GC) which contains dispersed chromatin [57] T heir ultrastructure has been visualized using electron microscopy. The FC is composed of a network of numerous dense fibrils [56] The surrounding DF has a similar thickness but is sufficiently dense to be differentiated from the FC u sing electron microscop y. The GC made up of fine granule s with diameter s of approximately 15 nm is located both in side and adjacent to the FC and DF. Once the cell cycle reaches a certain stage, the size and number of GC s is reduce d as the proteins are transported from the rese rvoir to the cytoplasm. Additional subnuclear structures include vacuoles and chrom atin. The micro nucleoli contain fibrillarin and rRNA geometrically and may reflect the speed of recovery or the presence of cellular repair mechanism in damaged cells [52] In the nucleoli, vacuoles have a separate light refractive index, and a bubble like compartment filled with non nucleolar neoplasm [151].


62 The p oten tial for MRM to be used to visualize t h e s e subcellular structures is increased by the presence of metal s i.e. aluminum, zinc, iron, copper, inside the nucleoli oocytes and embryos of xenopus, quail oviduct s, and rat follicular cells [58, 59, 155, 156] Although there is controversy regarding cellular iron contrast, main ly because of its low concentration, confirmation of such contributions is not easily attained as the se levels may be insufficient to be detected through histology U tilizing an autoradiography technique, levels of the type of iron concentrated in the endoderm a form of ferritin may be assessed Ferritin is an iron storage protein which is responsible for storing intracellular iron (Fe3+) in a nontoxic form and regulating iron homeostasis [60, 61, 157] However, due to its paramagnetic properties, the existence of iron in the nucleoli establishes the possibilit y for MRM to detect them through differe nces between the nucleoli and surrounding nucleoplasm without employing exogenous agents. Apart from the physical advantage of nucleoli images taken by MRM techniques, there might be potential clinical applicati ons. Their importance in pathology is associ ated with modifications in their dynamic assembly. Such alterations result in 185]. Throughout their multiple divisions of embryonic stage cells the cells vary in size while keeping their nuclear integrity. However morphological changes in the nucleus such as alterations in shape, size, etc, can be associated with other diseases: for example, hypertrophic nuclei are a diagnostic indicator of cancer [188,189] For that reason abnormal size and aberrant number of these nucleoli can be a sign used as a common prognostic indicator. Even if in vivo MR diagnosis at the cellular level is not yet


63 practical unlike biopsy study due to the reduction of processing time and visualizati on of 3D structure s inside deep tissue without precipitating any potential ly hazardous effects from an invasive protocol, MRM might be suitable for contributing to disease prognosis by offering endogenous tissue contrast. In the treatment stage, a common c ancer treatment is chemotherapy which damages the normal cells along with targeted cancer cells due to the use of exogenous chemical ly toxic agents. However, by understanding the main mechanism for induced apoptosis or programed cell death, of cancer ce lls which is driven mainly by the nucleoli, MRM could potentially detect the death signal and thus might open up new avenues in clinical tumor therapy. Methods Sample Preparations and Protocols Immersion fixed samples O ocytes of Xenopus l avis which had be en fixed in 4% f ormaldehyde solution were isolated by gross dissection, washed in no less than 100X volume of PBS buffer (137 mM NaCl; 2.7 mM KCl; 10 mM Na 2 HPO 4 ; and 1.8 mM KH 2 PO 4 : pH 7.4 at 300 m Osm) to remove fixative, and cut into 50 m and thinner sli ces depending on experimental needs. After washing in PBS 4 to 5 ova were immersed in 2~3% agarose gels in embedding boats. Cutting neurons using a vibratome (Ted Pella, Lancer series 1000 ) yield ed slices 75 m thick. For histological validation, we teste d varying thickness es ( 25, 30, 50, 75 m ) In order to enhance the diffusion of fixative into oocytes in the fixation process we began by manually cutting oocytes in half and removing the ovum membrane and vegetal pole containing mostly yolk palate (Fig ur e 2 4) In that way, we embed ded the nucleoli in the immobilized nucleus and kept them as true to their original structure as possible.


64 Immersion of fixed samples in agarose and slice preparation The animal pole from o ocytes of Xenopus l avis were isolated and immersed in 2~3% agarose gel fixed in 4% f ormaldehyde solution, and washed overnight in no less than 100X volume of PBS buffer (137 mM NaCl; 2.7 mM KCl; 10 mM Na 2 HPO 4 ; and 1.8 mM KH 2 PO4: pH 7.4 at 300 mOsm) to remove fixative prior to MR imaging. Low melting temp (25 o C) agarose at 2.5~3% concentration was mixed in PBS buffer at a temperature of 30 C ). Six to eight oocytes were positioned at the bottom of embedding boats using a pair of micro forceps Cells were positioned with the cut surface facing up. After cooling the agarose blocks gelled they were glued to a cutting block inside a vibrat ome. Cutting media was kept ice cold to ensure integrity and firmness of the agarose embedded samples during the cutting procedure. O ocytes were sliced into 25, 35, 40, 50, 75 m slices, depending on experimental needs Cutting slices less than 25 m thick was not possible due to equipment limitations and experimental limitations in that thin slices are prone to signal to noise limitations Magnetic Resonance Microscop y Imaging with Sample Retaining Apparatus Magnetic resonance imaging was carried out usin g commercially available 500 m ( Bruker Biospin Z76414) or 100 m ( Bruker Biospin Z76412) micro surface coils provided by Bruker Biospin [146] The MRI systems used were a 600 MHz (14.1T) ( Oxford Instrument s) and 750 MHz (17.4T) ( Bruker Biospin ), then i nterfaced with Bruker consoles, vertical magnets. Highly specialized sample positioning and micro manipulation skills are required when dealing with ultra thin sliced samples S ample s less than 1 mm in diameter inside a sample well were retained to the sur face of the micro surface coil by the nylon mesh and the retention apparatus T echnical aspects about positioning are shown in Fig ure 2 4 First, in order to stabilize ultra thin slice d


65 samples at the center of the micro surface coil, they were anchored wi th purpose built equipment. These parts were delicate enough to maintain the morphological integrity of tissue specimen M icro wound 50 m mesh (Small Parts, Inc., # CMN 0053 C) was employed for this purpose Due to water surface tension, the mesh also requires additional weight to dis place air from the pores. We used PBS buffer to suspend the sample in for imaging A circular ring made ou t of polyethylene was employed to stabilize the samples on top of the coil. It was chosen for its high visibility and biocompatibility. By cutting a 1/3 open ing using a sharp razorblade, an opening was used as a route for irrigatin g the sample and a ventin g hole for air The well was closed with adhesive PCR film (ABgene, # AB 0558). Securing the PCR film by pressing the rim of the tissue well with a micro tweezer helped to prevent accidental water leakage when running long scans. P ositioning micrometer sc ale structures on the coil surface and morphological correlation work after MR scans was performed using a light microscope ( Zeiss, Axioplan 2 ). A ll experiments were performed in the AMRIS facility itute. Image Analysis and Validations Imaged tissue slices were placed in a tissue embedding boat (22 038217, Fisher) and covered in agarose compound (22 11 0 617, Fisher) After the agarose solidif ied slices were cut into 25 to 50 m sections and placed o n pre washed, poly lysine coated microscope slides (12 550 13, Fisher; P 8920, Sigma Aldrich). Tissue was allowed 24 h at ambient temperature to adhere to slides. Previously cryosectioned tissue w as immersed in a DAPI stain. Following application of DAPI slides were air dried before applying mounting medium (HS 103, National Diagnostics) and cover slip s Last,


66 stained tissue sections were photographed using a RGB Spot camera attached to a Zeiss Axioplan 2 light microscope. Visualization of n ucleoli with h i stological c o registration : Structures corresponding to nucleoli were v isualized using DAPI stain. The histology shows that structures exhibiting low T2 values in our MR micro imaging experiments were localized to areas in which DAPI staining ha d revealed the presence of nucleoli. By light microscopic imaging of intact preserv ed morphological structures inside and outside the nucleus, slices were examined thoroughly for the subsequent M R microscopic imaging in a time effective manner After collecting the MR imag es histology was performed. F or example, DAPI, which is common and very active in staining nucleic acids inside the nucleus was used. Results The representative micro surface coil which was exploited heavily in these experiments was introduced in Fig ure 2 1. The sample under research in this chapter was a frog ova, which was introduced in 1986 [41] for the first visualization of the cellular compartmentation, i.e. nucleus and cytosol, and reintroduced here for delineation of subnuclear structure in the nucleus using a dedicated hardware such as surface micro coils. The appearance of the nucleoli in the translucent nucleus from the manual dissection was scanned by the light microscopy in Fig ure 2 2. One sample light microscopic image of the internal structure of the nucleus was visualized in Fig ure 2 3. In the preliminary studies under the light microscope, a number of nucleoli were observed and their geometrical characteristics were introduced in the diagram in Fig ure 2 3 Sub nuclear structures, nuc leoli, were shown to be spherical bodies approximately 10~20 m in diameter in our histology images Most of them appear ed to be compact


67 spheres, while others are vacuolated as seen in the agarose embedded slice model of stage IV. Figures 2 2 and 2 3 illu strate the three different kinds of nucleoli and subnuclear materials visible under the light microscope. Since there are different components intermingled inside the minuscule nucleoli, depending upon the mineral or metal constituents the magnetic proper ties might generate different contrast Some of them were aggregated into groups and were targeted to have a greater chance of obtaining MR contrast (Fig ures 2 2 and 2 3). S lices were investigated to detect nucleoli seen in our light microscop y images ( F i g ure 2 5). Based on their nucleoli only occupy three voxels at the resolution in our MR images. I f there is any movement during imaging, the chance of obtaining images of subnuclear structures might be diminished The hypo intense signal regions in the nucleus of the frog ova correspond to the nucleoli (Fig ure s 2 5 through 2 1 3). The MRM images acquired with 14 T magnetic field strength (Fig ures 2 5 through 2 7) offer coarse delineation compared to 17T magnetic field strength does due to less sensitivity in det ecting the spins in the region of the interest. Furthermore, in 2D MRM images, smaller nuclei may have been impossible to visualize due to partial volume effects caused by coarse through plane resolution Nonetheless, most of the attempted correlation of n ucleoli especially by using 14T and 2D MRM imaging acquisition scheme between the two imaging modalities showed that one to one correlation of subnuclear structure at this scale was not possible (Fig ure 2 6 and 7) The delineation of nucleoli and correlatio n was attempted with a 17 T magnet and 3D imaging acquisition schemes (Fig ure 2 8 through 2 13). A number of nucleoli inside the nucleus were detected in 3D gradient echo images and compared with corresponding


68 histological images. In order to make sure th e nucleus intact as well as the to differentiate the hypointense signal areas coming from the subnuclear components with the air bubbles trapped, the T 2 weighted 2D MSME scans were collected (Fig ure 2 8). Due to slice protocol, two different in plane resol utions of 7.8 m (Fig ure 2 9 A ) and 6 m ( Figure 2 9 B ) were compared and attempted for higher delineation of the each subnuclear structure as individual as possible. By increasing repetition time, TR, from 500 ms (Fig ure 2 10 A ) to 1000 ms (Fig ure 2 10 B ) to account for increased T1 relaxation time, the signal inside the nucleus can be compensated. However, to find the optimal contrast to noise ratio with the limited signal to noise ratio, the interactions between the relaxation times and field strength neede d to be taken into account. T wo light microscopy images which present different geometric features and distribution of nucleoli (solid red arrows in Figure 2 11B and dotted red arrows in Figure 2 11C ) by varying the focal point or on the different stacks w ere aligned next to each other. These show the aggregation of nucleoli in the in plane as well as through plane dimension which sometimes makes one to one correlation between two imaging modalities non practical with the current status of MRM imaging techn ologies. While exact correlation of MR and histology images of aggregation of nucleoli in 30 m thick slices with histological images was not forthcoming, use of a thinner slice or confocal micros c opy might aid in future attempts to achieve correlative his tological conformation of subnuclear structures Histological fluorescent image employing DAPI staining for the nucleic acid show ed that the structures exhibiting low T 2 values in the nucleus in our MR microimaging experiments revealed the presence of nucl eoli (Fig ure 2 12)


69 Discussion These data represent the first instance sub nuclear structures were detected using MRI without the use of an exogenous contrast agent. In the granular material of the nucleoli, short T2 results in hypointensity against the n uclear background. Some images show basogenic material. Fig ure 2 1 shows the geometry of the Bruker surface coil. The sample in F ig ure 2 2 illustrates nucleoli resid ing in side the translucent nuclear sa ck shown as crescent shaped scaffolding Still, some r emnants attached to the nucle us are mostly cytoplasmic compartments with lipid contents from the vegetal pole. After trying to scan the intact, unsectioned cell it was realiz ed that this was not the best strategy. For that purpose, a solenoid coil was bet ter suited. A t the same time we need ed to find a way to compensate for sensitivity loss. W e chose to cut the oocyte into thin slices which turned out to be the best way to visualize and detect small subcellular structures. Most validation work was based on the MR correlation with the unstained histological images due to different light reflex indices of the nucleoli inside the translucent nucleus or germinal vesicle. However, given the need for an essential validation process of the subnuclear structures performing different staining methods for validation will comprise a large amount of future work For example, using silver stain the large nucleoli c ontaining the fibrillar region could be effectively differentiated from other subnuclear materials as m icro nucleoli and vacuoles. S ample selection before MR microscopy can be done by employ ing a higher resolution light microscope to select and position samples with optimal tissue structure Depending on the stage of oocyte development, as the oocytes get bigger in the more mature stage s the majority of mitochondria are located in the animal pole This structure pull s and indent s the membrane of the oocyte in the germinal vesicle resulting


70 in a crescent formation on the surface. Through careful s election samples with large nucleoli can be selected and micro nucleoli which have only fibrillar zone s without granular cortex can be avoided In that way, correlation work after the MR scan s could have a much higher success rate. C are should be paid in the manu al dissect io n procedure, as a second round of fixation of the nucleus was required to achieve adequate sample firmness in some cases Additionally the embedding of bisected oocytes in agarose cause d slight varia bility in slice position Consequently, the shape of the nucleus might not appear the same in different imaging The differences in chemical composition inside the two poles might affect the MR signal intensity since yolk pla t e l e t s are known to cause hindrance effects in the highly diffusion weighte d images [63] Since MRI can detect different tissue components based on inherent contrast mechanisms such as relaxation times, if any materials which have different m agnetic susceptibilities can cause sometimes unwanted image artifacts. These effects can arise from the boundary between two different media having different local magnetic field strength. The consequent decrease in T 2 causing the additional signal loss in the image is more severe in gradient echo images but less or nothing in spin echo which is immune to this de phasing [7]. For that reason the spin echo is often employed to prove that the hypo intense signal is not due to susceptibility effects from the a ir bubble. Therefore, the subnuclear components turned out not to be air bubbles (Fig ure 2 10). Given the importance of the validation process of mapping the MR visible structures with other imaging modalities, extra care dealing with this interference fro m the magnetic field inhomogeneties coming from beside the actual cellular components should be paid [7, 22, 2 43 ].


71 Last but not least, attempts were made to correlate subcellular structures seen in our MRM images with light microscopy as shown in F ig ure 2 11. Even with the SNR and CNR achievable on state of the art imaging systems, comparatively poor sensitivity and resolution of MRM still needs to be enhanced for validation of nucleoli. In Fig ure 2 11, consecutive histological images ( Figure 2 11 B and 2 11 C ) of different levels within the slice collected by varying focal points inside the 40 m slice depth is displayed. In a recent report by Brangwynne et al. (2011) they manipulated the nucleoli embedded in the elastic nucleoplasm and fused two nucleoli into a larger nucleolus [64] They report that the nucleoli possess the liquid like characteristics and can be fused after a short time (approximately in 5 min). Even though in this study the fixed cells were employed, by treating cells wi th Cyto D, a drug which destroys the f actin prior to fixation, we could induce fusion of nucleoli inside the nucleus and increase the chance of visualizing these structures by making quite a few numbers of nucleoli in one big mass. C onclusions and Future Work Throughout this preliminary investigation of the subnuclear structures, i.e. nucleoli, in Xenopus oocytes the subnuclear structures were detected with MRM and validated with other imaging modality by utilizing their different light refractive index with surrounding nucleoplasm Even if the one to one correlations with their contrast of MRM images with that of light microscopic images, the contrast from the subcellular structures at least discriminable. By having the quantitative analysis for now, we hope the future works can be designed to reveal the reasons for non mono exponential characteristics about the MR diffusion signal behavior in the intracellular compartments including the origin of non mono diffusion signal characteristics in the nucleus w hich has been


72 invisible with the conventional imaging apparatus so far. Once the reference on the MR signal characteristics out of the nucleoli in the nucleus is set up, the potential of this preliminary research is significant By having the delineating t he sub nuclear structures with MRM and the associated information about the changes in size and structure of nucleus, the cellular MRM imaging might contribute to diagnose the human diseases such as cancer [188, 189] or gene tic diseases [183, 184, 185]. O n ly if the large, ring shaped nucleoli with their fibrillar center s get stained, then as the big nucleoli reflect the rapidity of cell division this work might contribute to the tumor analysis The alternative possibility of applying MRM is far reaching y et realizable to utilize to monitor the change in the proliferation mechanism of viruses which activates after injecting their genetic the virus burst s the cell membran e and release s the proliferated material throughout the human body. Without going through any destructive protocol such as dehydration and exogenous chemical stain agents, if the inherent MR based stain works in the intact internal structures inside the de ep tissue, this preliminary research might contribute to pathodiagnose the genetic disease associated with nuclear alterations in nuclear body assembly and function such as spinal muscular atrophy or dystrophy [183] and the potential cancer diagnos e s in deep tissue. Furthermore, in conjunction with the tumor therapy with apoptosis, programmed cell death, if MRM could be utilized as a biomarker for early diagnosis or if it could access a specific signal to trigger a suici de mechanism through the nucleoli or the iron inside them, it could offer an aid to access the tuning ability of the suicide mechanisms of the cancer cells [62]


73 Figure 2 1. Representative p hotograph of the 500 coil. The 500 four turn surface micro coil which was developed by Bruker, Switzerland (Z76409) sits inside a 5 mm diameter and 500 deep tissue well


74 Figure 2 2. Light microscop y images of a frog ovu m The semi trans parent nucleus (germinal vesicle) exposed after manual dissection show s the many nucleoli appearing as compact bi spheroidal bodies with diameters of 10~20 m


75 Figure 2 3 Light microscopy image of nucleus of frog ovum and schematic of the subnuclear constituents. A ) Representative image of the slice containing the nucleoli and B ) e xpanded version of the sample nucleoli (C E) I llustration s of different types of nucleoli consisting of granular component (GC), C ) single dense and D ) s maller dispersed fibrillar component (DF), fibrillar center (FC), vacuole and chromatin dispersing irregularly inside the nucleoli, E ) M icronucleoli containing only FC surrounded by DF


76 Figure 2 4 Schematic of cellular structure of frog ovum and the protocol of sample preparation A) Oocytes were hand cut under a dissecting scope, then immersed in 4% formaldehyde solution in PBS overnight for fixation The translucent globular nucleus presents dot s of nucleoli in red animal pole surrounding nucleus in the upper hemisphere, and vegetal pole mainly made up of fatty yolk. B ) Bisected ova lie inside the agarose block and using a vibratome, cutting frog ova into slice samples whose thickness is between 25 and 75 m. C ) Inside the sample well on the micro surface coil, the slices embedded in agarose lie on top of the coil suppressed by the nylon mesh the retention ring, and sealed by the PCR fim Mesh pore diameter = 50 m.


77 Figure 2 5 Direct comparison between cellular compartments in MR M and histology image s A) The MRM image was acquired with in plane resolution of 11.7 m B) The MRM image was acquired with in plane resolution of 5.8 m acquired with 14T of magnetic fie ld strength. C) The corresponding light microscopic image which underwent the MRM acquisition MR scan parameters: 2 D T2 weighted images of 30 m slice of oocyte illustrating signal inhomogeneity in the nucleus (bright central region ) Imaging parameters: re solution = A ) 11. 7 x 11. 7 x 50 m 3 B ) 5.8 x 5.8 x 50 m 3 spectral width = 5 0 k H z TR = 300 0 msec, TE = 30 msec, Avg = A ) 30, B ) 40


78 Figure 2 6 Direct comparison between cellular compartments in MR M and histology image s A ) Representative MR M image scanned from the slice of frog ovum acquired with 17 T magnetic field strength B ) The expand ed view shows the morphology of subnuclear structures in MRM image based on different magnetic properties C) The corresponding light microscop y image presents the morphological information based on the light refractive index MR scan parameters: 3 D gradient echo images of oocyte illustrating the dark spot nucleoli in nucleus (hyperintense signal at the center). Imaging parameters: reso lution = 7. 8 x7. 8 x1 5.6 m 3 spectral width = 22058 Hz, TR = 1 00 0 msec, TE = 8 .42 msec, FOV = 1 x 1 x 1 cm, MTX = 128 x 128 x 64, Av era g e = 8, acquisition time = 18 hr 12 min.


79 Figure 2 7 Comparative analysis between MR M and histology images. A ) The r epresentative MR M image was scanned with 17T magnetic field strength and presents the MR characteristics inside the nucleus and cytoplasm of the frog ovum. B ) The comparative light microscopic image presents the characterist ics in the intracellular structures based on the light refractive index. C) The subnuclear structures such as nucleoli are well presented. MR scan parameters: 3 D gradient echo images of oocyte illustrating the dark spot nucleoli in nucleus (hyperintense si gnal at the center). Imaging parameters : resolution = 7.8 x 7. 8 x 15 m 3 spectral width = 22058 Hz, TR = 5 0 0 msec, TE = 8 .42 msec, FOV = 1 x 1 x 1, MTX = 128 x 128 x 64, Avg = 2, acquisition time = 1hr 22min.


80 Figure 2 8 A s eries of T 2 weighted MR M The images were scanned from the slice of frog ovum acquired with 17T magnetic field strength MR scan parameters: 2 D Multi Spin Multi Echo i mages of oocyte illustrating structures with hypointense signal properties in the nucleus (hyperintense signal at the center). Imaging parameters : resolution = 17 x 16 x 200 m 3 spectral width = 5 0 k Hz, FOV = 2.2 x 2.0 cm, MTX = 128 x 128, TR = 2 00 0 msec, TE = A ) 13.6 ms B ) 1 8 ms C ) 22 ms D ) 30 ms, Av era g e = 6, Acquisition Time = 25 min


81 Figure 2 9 A series of 3D GE MR M with different resolution. The 3D MR M image s wer e scanned from the slice of frog ovum acquired with 17T magnetic field strength at two different resolutions MR scan parameters: 3 D gradien t echo i mages of oocyte illustrating the dark spot s nucleus (hyperintense signal at the center). Imaging parameters : A ) R esolution = 7.8 x 7.8 x 15.6 m 3 spectral width = 22058 Hz, FOV = 1 x 1 x 1 cm, MTX = 128 x 128 x 64, TR = 5 0 0 msec, TE = 8.4 msec, Avg = 8, Acquisition Time = 9 hr 6 min, B ) R esolution = 6 x 6 x 15.6 m 3 spectral width = 22058 Hz, FOV = 1 x 1 x 1 cm, MTX = 166 x 166 x 64, TR = 5 0 0 msec, TE = 10.48 msec, Avg = 14, Acquisition Time = 20 hr 39 min


82 Figure 2 1 0 A series of 3D GE MR M with different resolution and TR. The 3D MR M image s were scanned from the slice of frog ovum acquired with 17T magnetic field strength at two different repetition times MR scan parameters: 3 D gradient echo i ma ges of oocyte illustrating the dark spot nucleoli in nucleus (hyperintense signal at the center). Imaging parameters : A ) R esolution = 7.8 x 7.8 x 15.6 m 3 spectral width = 22058 Hz, FOV = 1 x 1 x 1 cm, MTX = 128 x 128 x 64, TR = 5 0 0 ms, TE = 8.4 ms, Avg = 8, Acquisition Time = 9 hr 6 min B ) R esolution = 6 x 6 x 15.6 m 3 spectral width = 22058 Hz, FOV = 1 x 1 x 1 cm, MTX = 128 x 128 x 64, TR = 10 0 0 msec, TE = 8.4 msec, Avg = 8, Acquisition Time = 18 hr 12 min


83 Figure 2 11 Co mparative analysis between MRM and histology images A ) MR M image presents the characteristics of MR signals inside the nucleus of the frog ovum acquired with 17T magnetic field strength B ) and C) The t wo consecutive light microscopic image s along differen t focal depth present different geometrical existence of nucleoli (solid arrows in red ). C) The dotted arrows indicate the nucleoli which are out of focus MR scan parameters: 3 D gradient echo images of oocyte illustrating the dark spot nucleoli in nucleus (hyperintense signal at the center). Imaging parameters : resolution = 7. 8 x7. 8 x1 5.6 m 3 spectral width = 5 0 k Hz, TR = 5 0 0 msec, TE = 8.42 msec, Avg = 2 acquisition time = 1 hr 22 min.


84 Figure 2 12. Representative histolo gical image of frog ovum stained with DAPI. The blue, circular structures indicated by arrows (red) are DNA containing structures inside the nucleus correspond to nucleoli.


85 CHAPTER 3 REVEALING SUBCELLUAR ARCHITECTURE AND INT RACELLULAR DIFFUSION CONTRAST F ROM L7 NEURONS OF AP LYSIA CALIFORNICA Introduction Magnetic resonance microscopy ( MRM ) has recently been able to resolve rat, pig and human neurons ranging from 50 to 100 m in size Without the use of any exogenous or toxic chemical agents, MRM offer s u nique contrast thus establishing a new method for direct quantitative analysis of cellular architecture [50] MRI has been evolv ing since its first appearance in the 1970s when images were first taken at sub milimeter resolution [40] .The first practical application of MRM, a subset of MRI, was demonstrated in 1986 with imaging of sub cellular structures animal cells such as rat brain [178] frog ova [41], plant stem [177] at tens of micron s resolution A pplications were not limited to biological samples and have found alternative uses [42 44, 65] In addition to embryonic frog ova, a neuronal soma, L7, isolated from the abdominal ganglia of Aplysia c alifornica has been imaged using MR M, and ushered in MR based delineation of a more complex neural system. In addition with the embryonic frog ova, an actual neuronal soma, L7, isolated from the abdominal ganglia of Aplysia c alifornica with MR mic roscopic imaging blazed the trail for the possible delineation of the complex neural system. Despite the difference with mammalian nervous systems, i.e., non vertebrate animal and 60~100 times bigger than mammalian neurons in size, the animal ha s gained po pularity in comparative and behavioral neuroscience as a n early model of the CNS of mammalian [66] Benefits of U sing T his A nimal Among many reasons to employ this marine animal, Aplys ia c alifornica the advantages are as follows. First, from a comparative point of view it has primitive CNS


86 structures compared to the CNS structures of mammalian species. It is used in research on behavioral learning and memory, applicable to elementary memory loss from aging and to signal pathway s of the neurons. The unique feature of this animal is the giant size of specific neurons such as L7 and R2 ( F ig ure 3 2 ) and are similar to the hypertrop h y of big cells in mammalian CNS s, such as Purkinje cells [66] Second, due to easier access to its individual big neurons, its applicability to the MR receiver coil, especially a solenoidal coil type, has caught the attention of the MR community [20, 33, 41, 43, 81, 82] However, with relatively poor resolution and sensitivity, the partial volume averaging interfered with the MR signal analysis of the neuronal model. Specifically, it was noted that there was ambiguity concerning the assignment of cellular compartments [ 33 ]. That ambiguity is somewhat mitigated by improved technology, including micro surface coils and strong / fast switching imaging gradient coils developed by Bruker for much improved sensitivity and resolution. Novelty of the C urrent S tudy The novelty of this present research comes from two modifi cations of the sample preparation protocols. First, unlike a conventional protocol us ing chemically manufactured artificial seawater, the actual body fluid eject ed from the animal was employed in order to avoid any minute concentration differences and resu ltant physiological changes from the artificial sea water, such as Ca, Mg or K, protein water density in the environment around the neurons as reported by Albrecht Bethe [186] Even if the amount of liquid was not great it was sufficient to fill the samp le well of the micro surface coils for our MRM experiments, and the comparison presented in T able 3 1 shows insignificant differences in terms of osmolarity. Second, in conventional sample preparation, an enzymatic and chemical solution, e.g. collagenas e and trypsin


87 was employed to loosen the external layer of the ganglion sheath, muscle, collagen and fibroblast ( F ig ure 3 3 ) I n this study however, after learning of the possible loss of glial cells and extracellular matrix by this exogenous media [74] the surgery was carried out without the addition of any digesting media. Also additional neighboring cellular regions and structural protein in the extracellular matrix ( ECM ) offer structural support for sample positioning in the sample well of the micro surface coils. Extra care should be taken not to compress or burst the neurons although the neurons were mostly intact. By employing a primitive single neural cell body L7 of A c alifornica in our study the o rigin of the MR signal at a sub cellular level especially in the intracellular domain was investigated To our knowledge, this is the first study to delineate the cellular compartments of L7 using MRM correlated with histological data region by region. T he Aplysia neuron was prepared using several different methodologies, i.e. whole cell, bisected, and segmented as shown in Fig ure 3 4 to reveal different intracellular structures. Interestingly, the structural integrity was strong enough to hold tog ether after being fixed and most cells remained intact after fixation. Contributing to its p rominent size w as reportedly due to the secretion from the granules inside numerous glial like support cells on the periphery of the big neurons [187]. These glia t ype cells often made the invaginating contact with the nucleus known as trosphospongium [68 71] There have been historical controversies regarding its identity but in o ur slice models revealed that its different MR signal characteristics from the nucleus and direct correlation was visible by its unique MR contrast From the previous study [20], the water diffusion inside the nucleus is close to that of the free water, an d more restricted MR diffusion signal is attributed to the cytoplasm might be due to


88 structures of cellular organelles inside resulting in non single component diffusion. Even though there have been attempts to prob e the properties of organelles by varying diffusion times, the se subcellular structures other than nucleus and cytoplasm are still too small to measure ADC in [24, 73] Interestingly, t here are tiny some cells embedded in mesh like fibers that are pervasive These cells is reported to deliver nutrition in the form of glycogen to the nucleus as well as provide structural support to the cell [74, 75] When using a dehydration protocol commonly employed in electron microscopy, the glial cells read ily separate from in between the mesh like fibers [54, 76] They contain large amount s of water and glycogen, because they are involve d in homeostasis and osmotic regulation [74] These small, supportive cells and mesh like fibers outside the L7 neuron contribute to cellular communication between neural cells and can serve as an earl y indicator fo r ischemic stroke as they become sw o ll e n during this process Additionally they may be one of the primary source s of restricted diffusion in diffusion weighted imaging along with the extracellular matrix embedded in the tissue of the CNS, and sw o lle n glia l like cells might add to the fraction of compartment s contributing to the lower diffusing ADC [74, 77] Based on mapping intracellular compartments by h istology, we hope this study will contribute to knowledge about the origin of MR signal in physiology as well as pathology. In future work, by perturbing the external medium i.e. changing the tonicity of the perfusate to induc e cellular swelling with the h ypotonic medium mechanisms and dynamics behind the observed decrease in ADC as in human disease such as stroke


89 could be characterized with higher sensitivity and resolution to reveal information which can be used to improve our ability to interpret clinica l data. Methods Sample Preparation and Protocols Adult marine gastropods, A. californica were obtained from Aplysia Resources Center (University of Miami, FL) and L7 and R2 neuron s along with additional satellite cells were isolated from the abdominal ganglia by gross dissection without using a collagenase/trypsin digestive solution usually employed to disintegrate connective tissue This was to avoid any chemical interference or potential loss of glial type cells from the connective matrix in the intra cellular region after digestion of collagenase. This potential loss of glial type cells in the protocol have been reported by other research groups attempting to visualize them using different imaging modalities such as electron microscopy [74, 76 78] F urther more in order to avoid any kind of chemical contamination from artificial sea water, the body fluid of the sea slugs extracted from the abdomen through the foot by a syringe needle which has more natural protein resulting in higher viscosity and m ore visible air bubbles was used as a surgical media and imaging buffer L eaving the other cells intact and attached to the plasma membranes of the L7 and R2 neurons made positioning and manipulating the sample slices easier They were stored for a minimum of 24 hours in artificial sea water (ASW; in mM, 460 NaCl, 10.4 KCl, 55.0 MgCl 2 11.0 CaCl 2 15.0 Hepes, pH 7.85) with 4% formaldehyde as a fixative After washing in 3X serial washes in PBS a subset of neurons (usually 4 5 ) were selected and immersed in melted agarose filled embedding boats for high throughput sectioning Cutting neurons using a vibratome (Ted Pella, Lancer series 1000 ) yield ed slices from 25 m to 100 m For histological validation


90 before and after MR scan ning careful preservation of the slice sample is required. During preliminary experiments, we found that in some cases due to difference s in density between connective tissues and cellular compartments, cutting direction, and thickness, the large nucleus in the L7 neuron dislocate s ea sily from sliced samples Selection of appropriate samples i.e. those with intact nuclear compartments following slicing under a light microscope before the MR microscopy scan was required for experiments to be conducted in a n efficient manner. Special car e should was taken when positioning samples in side the MR apparatus and completing the histological staining process to ensure the tissue remained intact during these procedures Sample Positioning and Magnetic Resonance Microscop y Proper p ositioning of the circular cell slice in the micro surface coil that sits vertically inside the magnet was quite challenging. I f the sample was not settled on the coil surface in some cases, the slice itself drifted down and away from the surface coil due to the gravity. We employed two novel slicing methodologies in these experiments. The first included placing the cut surface of a bisected (Fig ure 3 4 B ), fixed L7 neuron to the surface coil thus avoiding any potential loss of cellular organelles while gain ing access sliced sections ( 25 m ), of fixed neurons for easier 2D diffusion image correlation with light microscopy. As discussed above, th e size of the L7 neuron makes them good candidates for morphological studies investigating the intracellular architecture of the neuron. Once it was confirmed under the dissecting scope that the cells plasma membrane was intact, the chosen sample underwent fixation in 4% formaldehyde overnight. To remove the


91 fixative prior to imaging, samples were washed in PBS through no less than four buffer changes. Using a razorblade under the dissecting scope, the one of the lager neurons L7 or R2 was cleaved into two pieces. In an attempt to reveal such su bcellular structures as the nucleus, cytoplasm, membrane, glial cells, and extracellular matrix (ECM), we obtained high resolution ( 10 m 3 ) 3D images The final images contained approximately 30 voxels. For the experiments conducted on bisected neurons (Fi g ure 3 4 B ), an auxiliary sample holder made was developed in house to keep the sample from experiencing mechanical motion from water surface tension during placement of the tissue and PCR film for sealing the well (Fig ure 3 5). Additionally to place th e bisected cell in the comparatively large sample well the diameter and depth of which are 5 mm and 500 m respectively, additional bag cells [66] and axons remained intact and attached to the neuron (Fig ure 3 5). These structures, while not the focus of our imaging experiments, aided in positioning of the sample and holder while secur ing them in the sample well. Measurement of Osmolarity : A rtificial sea water (ASW) has been used to match physiological measur e osmolarity for aquatic animals in convention al isolation protocol s [20] However, in order to obtain the best physiological match for our isolation medium, the body fluid [186] from A c alifornica was used. In order to extract it, a syringe (BD Syringe, 3ml 23G) was used The fluid osmolality was me asured using an osmometer (Osmette, Precision Systems, I nc .) The osmolality values are given in Table 3 1. Image Analysis and Histological Comparisons W e can utilize some features of the sample such as its large size and hardware advancement s to delineat e subcellular structures with higher sensitivity and resolution using MR M. Although it has disadvantages compared to other imaging modalities such


92 as optical or fluorescence microscopy MRM is the only method that offers MR contrast Here, for the accurate comparison between MR stained cellular compartments with other validation methods, the fixed L7 neurons were segmented after being embedded in 2~3% of agarose solution as described in Fig ure 3 4. Specifically using a histological stain ing method like Ni ssl (e.g. aniline, thionine, or cresyl violet), we can validate the location of the intracellular cytoplasm in neurons. This method highlights the cell body, specifically the endoplasmic reticulum, in a blue or purple blue hue Using this method, we could determine where the cytoplasm was located in our histology and in turn our MRM images To stain the nucle ar compartment DAPI stain (Vectashield, VectorLab H 1200) was employed It reveal ed the boundary of the nucleus under the fluorescent scope. Finally to define the boundary between the intra and extracellular space, the delineation of the plasma membrane was carried out using PKH 67 F luorescent Cell Liner Kits dye obtained from Sigma Biosciences (St. Louis, Mo.) Results Delineation of S ubcellular A rc hitecture in a N eural C ell of A c alifornica using MR D etectable C ontrast E mploying N ovel M ethodologies of S ample P reparation Only with the endogenous MR signal relaxivities, the intracellular ultrastructure of the whole cell was visualized with finer reso lution by the 3D spin echo imaging sequence ( F ig ure 3 6 ). In the F igure 3 6 there are more than two intracellular compartments visible which were not seen in earlier studies employing l ower resolution [33, 43, 81]. This also provides needed information a bout neural compartment s for tissue modeling studies When viewing the tissue first w ithout exogenous chemical stain, the presence of an unknown structure having unique light refractive indices was revealed ( F ig ure 3 6 B ). Even smaller, neighboring cells pr esent additional


93 compartments in the MR M images. S o far h owever, only histochemical or fluorescent images can be used to confirm the identity of structures in MR images When u sing the whole intact cell, Nissl stain offers limited intracellular accessibi lity due to the external sheath of ECM E ven after being stained with Nissl the individual structures could not be discerned under the light microscope due to insufficient light penetration. Interestingly, as shown in F ig s. 3 6, 3 7, and 3 8 the large s ize of the nucleus which fills more than 50% of the entire volume of the neuron the existence of subnuclear structure s, presumably nucleoli and the cellular structure termed the trosphospongium [54] The size and morphology of th is subnuclear structure possessed sufficient contrast for its d istinction from the structure of the nucleus. Even though definitive histological co rrelation is not presented here, the MR signal nature inside the nucleus from the adult animal presaged another regime of MRM. Still in the case of the bis ected cell, Nissl staining met with limited success in improving light penetration, scattering and absorption from the ECM. In qualitative MR images, validating the presumed structural elements described above was done by histology. For that purpose, ano ther methodological approach was devised: application of agarose routinely utilized in cell suspension s Once embedded in the agarose, the sample under went a cutting process using a vibratome to yield uniform ly 75 m thick slices As presented in the diffu sion weighted 2D images of F ig ure 3 10 the capabilities of DWI performed well to present the diffusion of water in the cellular architecture. The contrast of intracellular structures is not visible in a b = 0 image, but the cellular compartments were clea rly delineated in the high b value images. The se were the nucleus, cytoplasm and an unknown external layer, and for b


94 value s over 4000 [s/mm 2 ], the water signal from restricted diffusion remained visible in the external layer T he other layers as a whole bec a me hypointense. Additionally, in the light microscop y image s at 40x magnification, five different subcellular compartments were observed ( F ig ure 3 10 ). Specific f eatures found in the image included the unknown invaginated structure associated with the nuclear compartment Second, the orange field in the histology image is located inside the cytoplasm It shares common features with mammalian cells in that the color comes from ferrochrome, an iron containing enzyme, used for ATP energy production. Addit ionally, in this animal, there is another enzyme in the cytoplasm lipochondria which is activat ed by light [72] Histological C onf irmation of the MR delineated S ub cellular A rchitecture of a S ingle C ell of A c alifornica In conjunction with specific structur es individual compartments can also be identified. By uti lizing the methodological scheme described by the ex vivo slice model, the identity of the compartments was revealed. First, in order to define the intra and extracellular domain s the plasma membrane was identified. As illustrated in F ig ure 3 11 the gre en fluorescent signal illuminates the plasma membrane which is demonstrated to be located between the external layer and the intracellular compartment This reveals that the former analysis of this single cell model was carried out not only on the intracel lular components, but also included structures from the neurons exterior This may be due to SNR and sensitivity limitations, as well as lack of sufficient resolution. Our DAPI stain ing delineated the nucle ar region of cells adjacent to the large, single n eural cell as well as numerous tiny cells at the periphery of big and smaller neural cells. The last round of staining was conducted using Nissl to stain the cytoplasmic compartment T he 3D reconstructed images included unknown,


95 intracellular structures w hose identity was confirmed through histology (F ig ure 3 1 2) It shows the pervasiveness of glial cells around the large neural cell in the ECM The ECM, in turn, surrounds the intracellular structures, i.e. nucleus and cytoplasm of L7, R2 and tiny satelli te cells Preliminary S tudies to D eterm ine MR C ontrast and I ntracellular W ater D iffusion out of the ( S ub) C ellular A rchitecture of a S ingle C ell of A c alifornica using D iffusion MRM In conjunction with earlier findings, it was a natural extention to inves tigate the microstructural compositions of cells using contrast related to water diffusion such as diffusion anisotropy. B ased on the slice model, such intracellular compartments nucleus, cytoplasm and external glial cells embedded in the ECM were imaged with a b value of 1500 [s/mm 2 ] (F ig ure 3 1 3).S ix different diffusion sensitizing gradient s were used to collect diffusion tensor data in an attempt to describe the diffusion properties of the cell. A fractional anisotropy (FA) map was reconstructed by Fan dtasia : software developed by Angelos Barmpoutis at the U niversity of Florida [80] The result showed that while there was not high anisotropy both in the nucleus and t he cytoplasm Nonetheless the glia and the ECM region d isplayed a rather disoriented anisotropy due presumably to the dense fibers wrapped with glial cells. A higher number of diffusion directions and a lower diffusion time may be required to achieve a better delineat ion of anisotropy in the intra and extracellular compartments. As mentioned above, in hope s that probing diffusion on a smaller scale might reveal additional details related to anisotropy of the MR diffusion signal, we applied variable diffusion times to the slice model. In F ig ure 3 1 4 we present data from experiments conducted with two representative diffusion time scale s: 5ms and 11ms These studies were conducted on a 600 MHz spectrometer and were intended to


96 monitor the dependence on MR diffusion tim e. First, the difference in diffusion characteristics between the nucleus and trosphospongium [54] was investigated and the high diffusion restriction in the glial and ECM region was observed By sampling the diffusion time scale, signal change displaying dependency on the diffusion time was expected F ig ures 3 1 4 and 3 1 5 show data from experiments conducted on a 750 MHz spectrometer. These studies probed a wider time scale of diffusion values including 8.52, 11.3 and 19 ms D ifferent diffusion behaviors in the nucleus and cytoplasm were observed in Fig ure 3 15 B B y increasing the diffusion time, signal from the nucleus was significantly reduced in Fig ure 3 15 C T he existence of a double nuclear membrane (Fig ure 3 15 A ) is another characteristic of this la rge neural cell and might contribute to water restriction effects along or inside the dual membrane. Preliminary S tudies to I nvestigat e the F ixation E ffects on the ( S ub) C ellular A rchitecture of a S ingle A c alifornica N euron In this preliminary experime nt a live L7 neuron in vitro following a 9 hour long MR scan, was immediate ly fix ed using 4% formaldehyde. A fter washing the sample in PBS overnight, it was position ed as closely as possible to the coil and imaged. A one to one comparison between pre an d post fixation cells was carried out. In the live cell imaging, the cell was bathed in the abdominal body fluid [186], which was extracted during surgery. In the 3D gradient echo images shown in F ig ure 3 1 6 structural variation s inside the intracellular compartment were observed and displayed heterogeneous MR signal In comparison to the images conducted on fixed cell s in which the signal from the bath medium glia and the ECM region showed significant increases in signal in Fig ure 3 16 B


97 Discussion Nove l M ethodology in S ample P reparation To our knowledge, this study represents the first time in MR imaging that cellular compartments with region by region histological correlation is reported. By means of micron scale resolution, the microscopic origin of t he MR signal was elucidated. Furthermore, the microstructure of organelles in the cell were delineated: a result we hope will provide useful information to help detect the early onset of disease and shed light on resolving challenging issues in the macros copic realm using microscopic resolutions. To help understand the MR characteristics of neural circuits, the simple and large neuron of A c alifornica was employed. T he fixed neuron was utilized and, given the need for higher resolution of internal micros tructure of the cell by MRM, the cells were embedded in agarose and sliced for increased access to internal structures. Conventional surgical protocol for gross dissection of the L7 or R2 neurons require such exogenous enzymatic and chemical agents as col lagenase and trypsin to help loosen tissues surrounding the cells and allow for easier access to extraction with surgical equipment. The neurons structural integrity including the surrounding sheath filled with glial cells was carefully preserved. Having completed the extraction without proteolytic enzymes the ganglion sheath consisting of collagen, muscle, and fibroblast was largely intact. In our protocol, artificial media was eliminated so as to prevent or reduce any possible interference or physical da mage to the ECM and to preserve the overall structural integrity of the cell. Following the slice protocol, we were able to tailor the internal ganglion components for placement over the micro surface coil. By parting the ganglion sheath by gross dissectio n, whose adjacent layer is the basement


98 membrane, we preserved the microstructural morphology of the tissues in the abdominal ganglion such as large neuronal, supportive glial cells, and the extracellular matrix. When dissecting the ganglion sheath, extra attention was paid not to compress any fragile, large neurons adjacent to the L7 or R2 cell. Notably, the basement membrane, the middle layer between the outer brittle and hard ganglion sheath and cell layers, is visible. This structure can serve as an e ffective tissue structure to attach the other structures of the ganglia with minimal loss of ECM and glial cells. In previously conducted EM studies by Dr. Raymond Lasek [74] They pointed out the possibility that these embedded glial cells could occur when utilizing a collagen loosening enzyme. With our preparation protocol, successful detection of tiny glial type cells was achieved using DAPI stain, and their numbers throughout the ganglion were impressive. All sa mples were prepared and imaged less than a week after fixation because, if cells are immersed too long in fixative, the orange colored region appearing due to an enzyme named ferrochrome in mitochondrial region, can fade away: maybe due to the transition of iron from ferrous Fe(II) to ferric Fe(III) in the ferrochrome. This biochemical change occurring on the cellular level might impact natural MR relaxation parameters. Additionally, extra care was taken in preparing the slices using a vibratome. Due to t he large size of the L7 and R2 neurons and the density differences between agarose and tissue, if the slices are too thin, the nuclear material or the nucleus itself can be displaced. Therefore, slicing the samples was performed at a temperature close to f reezing. Compartmentation I ssues Regarding the issue of matching MR signal with specific tissue compartments this large sized model of a primitive animal cell has been used extensively to study the


99 complex mechanisms and dynamics behind the cell swelling associated with stroke [20, 33, 81, 82] Still, there is a formidable gap between the MR characteristics of this animal cell observed at microscopic resolution and the coarser resolution acquired in the clinical environment. First of all, fast and slow tissue compartments at a ratio of 70:30 in the CNS of mammalian and human cells are engineered to fit the signal distribution according to b values of the diffusion weighting factor. Yet, a disparity exists between the MR signal and histological data maybe because of the assumption tha t the fastest diffusion comes from extracellular water, which comprises only 20% of a histological analysis. This assumption would be valid in a cell culture or a cell suspension system where the outside of cellular area was surrounded by media similar to free water. In viable, intact tissue in addition to having cells of variable size other critical factors affecting MR diffusion behavior, such as the cell membrane and the structural protein matrix, must be factored into models describing the geometrical a nd physical footprint of the intra and extracellular compartments. In a previous study [33], analysis of 2D diffusion weighted images collected using a solenoidal radio frequency (RF) coil, delineating only the nucleus and cytoplasm. The authors report t hat the diffusion signal behavior inside the cytoplasm was multi exponential and the reasons for this have yet to be fully investigated [33]. However, through this study, elucidating the internal structure of the single cell, we positively identify in MRM images the nucleus, cytoplasm, cell membrane, and another region on the periphery of the cell where a number of glial cells were embedded in the extracellular matrix. Interestingly, this ECM area presents heavy restriction and is structured so that, even w ith a b value of diffusion weighted imaging exceeding 10,000

PAGE 100

100 [s/mm 2 ], the water signal out of glial cells and/or ECM in this area presents extremely slow water diffusion. Another paper [74] visualizes the mesh like structure revealed by SEM that addresses possible causes for this unusual restriction in the glial zone. Therefore, in studies conducted at lower resolution [33, 43, 81, 82], the partial volume averaging blurred and mixed the signal from this area with th e signal from the cytoplasm located inside the plasma membrane. In the present study, the signal from the cytoplasm possess es diffusion characteristics that are slower than the nucleus, but higher than the glial cell and ECM areas. One possible reason for the slow apparent intracellular diffusion is crowded cellular organelles that hinder the micro scale movement of water in the cytoplasm. The water in the extracellular space consists mostly of water tightly bonded with the glycogen on the glial cells and t he rest comes from the water molecules highly packed in the dense fibers surrounding the glial cells in the ECM. After removing the basement membrane, their locations and morphology were visualized. This was possible only because most of the glial cells an d ECM were preserved by the current treatment protocol. Even if the two physical and biological compartments, i.e. intracellular and extracellular compartments can be approximated by two fast and slow diffusion components represented by the MR diffusion si gnal, extra care must be taken in analyzing the shaded area such as the boundary between them and other compartments including vascular components or other compartments inside the cells. Small voxels with high resolving power can enhance the accuracy in re solving the controversy regarding the disparity between intra and extracellular volume fractions and fast and slow diffusion pools (approximately 80% and 20%, respectively). Given the importance of this issue, matching anatomical and MR assignment of int ra and

PAGE 101

101 extracellular compartments by aid of high resolution MRM imaging in this study is critical. In that regard, after identifying glial cells in the region having been previously designated as the cytoplasm in earlier studies, it was a natural extensio n to define the boundary of intra and extracellular regions by identifying the plasma membrane using histological staining. Through visualization of the plasma membrane by PKD 67 and correlation of structures seen in the MR with histology, we now know tha t extra attention should be paid, at least in this animal, when defining the intra and extracellular domains. Otherwise the agreement between the animal model and the clinical data could be misinterpreted. The role of glial cells inside the ECM have been recently revisited [83] As the name implies, glia l cells (glia = glue in Greek) play a supportive role in nourishing the neurons in a variety of ways (described previously in the introduction). They play an important role in revealing the onset of stroke as they swell earlier than neurons. This early res ponse thus serves as a means of detecting the disease [84, 85] With respect to water diffusion in the cell, another modeling approach suggested by Le Bihan presumes a plasma membrane bound water pool that restricts diffusion in close proximit y to the cell wall [86] It is an interesting perspective regarding the semi permeable cell me mbrane, but the properties described by this model have not been validated. This model might be testable on smaller cells of the ganglion, such as bag cells [179] or such sub cellular membrane encapsulated organelles as mitochondria, endoplasmic reticulum, and the like. Last, our effective bi exponential analysis, which is the summation of two exponential functions, has the potential for ambiguity when describing the diffusion behavior of cellular models. There is a problem with identifiability when present ing the basically logarithmic plot by summation of two

PAGE 102

102 fractional exponential functions, often yielding nearly same results as the tri exponential [87] This issue of non identifiability might be resolved by developing a more sophisticated model. Enhanced SNR and R esolution by U pdating H ardware In order to employ dedicated MR equipment, RF coils and gradient coils, it is essential to understand the benefits and limitations in sensitivity and specificity when imaging at micron scale resolution. By reducing the diffusion time, t diff signal loss from T2 effects can be somewhat mitigate d but there is a significant advantage in using high diffusion weighting to delineate such subc ellular or cellular structures. Additionally, thanks to advancements in resolution and SNR, in this preliminary study, research was carried out on the diffusion behav ior along variable diffusion time s to visualize any anisotropic water diffusion inside the intracellular compartments, e.g. nucleus, cytoplasm and glial cells in the ECM. Interestingly there is a noticeable change in terms of anisotropy in the intracellul ar compartments that might suggest the effects of hindrance in the cell. By utilizing this observation, we might be able to anticipate anisotropic water diffusion in perturbation studies in the future. Furthermore for wider coverage and better field homog eneity with the same sensitivity, a phased array coil [244] might yield enhanced results in microscopic intracellular studies on water diffusion. Additionally, to observe anisotropy in intracellular water diffusion, diffusion along six different axes was c onducted with diffusion tensor imaging. By investing more acquisition time and increased voxel size to obtain higher SNR, additional DTI data might offer different or enhanced perspectives on the anisotropic movement of water in the cellular domain.

PAGE 103

103 R ole in C ell S welling From a comparative pathological point of view, the L7 neuron in Aplysia californica was utilized to investigate the cell swelling event, a precursor in diseases like ischemic stroke [20, 81] Still, there is room to enhance the MRM results by employing current state of the art technologies. Now, equipped with higher resolution and SNR, this previously conducted perturbation research should be revisited. Additionally, now that the existence of tiny glial like cells around the periphery of the big neuron [187] has been validated by the histological confirmation, the role of these tiny cells in a cell swelling event might offer an infor mation for the related events in the mammalian case [84, 85] the temporal and spatial monitoring of the detailed architecture of subcellular structure of the neuron and glial zone might help us to better understand the mechanisms and dynamics behind the decrease in ADC in human stroke models. By utilizing a newly developed oscillating gradient spin echo sequence (OGSE) [190, 191], minimization of diffusion time might be realized and yield superior diffusion data in this animal species. Conclusi ons and F uture Work Along with structural validation measurement of water diffusion signals at different diffusion time s and strength s has shown intracellular diffusion is faster than extracellular diffusion. In the intracellular domain, the nucleus does not contain any barriers that significantly affect water diffusion behavior but the cytoplasm presented some hindrance effects presumably due to the existence of numerous cellular organelles such as mitochondria, the endoplasmic reticulum, and the G olgi apparatus. Finally, by way of histological validation, the glial cells and ECM region which had not been observed using MRM methods were imaged and directly validated Our diffusion data

PAGE 104

104 describe sophisticated restriction effects making diffusion signal an alysis amenable in the CNS of mammal s This study represents a significant advancement in MRM by virtue of utilizing methods for minimiz ing the discrepancy between the cell model derived from A. californica and other animal models in terms of cellular comp artmentation. Along with the technological advancement s in dedicated equipment for MRM, by delineating cellular structures with finer resolution and higher sensitivity, the unknown dynamics and mechanisms at play in the coarse resolution s employed in clini cal imaging could be better understood. This work could reveal more about MR properties of tissue at the cellular level and contribute to the elucidation of MR signal analysis U ltimately it is our hope that this knowledge will aid in improving clinical im aging analysis. In conjunction with high resolution protocols cell signaling through light activation of the lipochondria might open up a new chapter in functional studies on single cell s

PAGE 105

105 Table 3 1. Osmoliarity of artificial sea water, anaesthesia, and body fluid of A californica Medium Osmolarity [Osm] Sea water 1165, 1121, 1123 Anaethesia 1161 Body fluid 1134, 1117

PAGE 106

106 Figure 3 1. Schematic of an a nimal c ell It consist s of nucleus(N), smooth endoplasmic recticulum(SER), rough endoplasmic recticu lum(RER), mitochondria(M), peroxisome(P), lysosomes(L), Golgi apparatus(G), lysosome(L), centrioles(C).

PAGE 107

107 Figure 3 2 Schematic of an Aplysia g anglion. Oversized neurons are L7 on the left side and R2 on the right side The intracellular compartments are assigned such as nucleus (blue dots), cytoplasm (yellow areas), and dense fibers (green) around the tiny cells.

PAGE 108

108 Figure 3 3 Schematic of i nternal structure of ganglion N euronal cells and surrounding glial cells embe dded in the extracellular matrix presented as the green mesh External gang lion sheath layer, made up of mu scle (M), collagen (C) and fibroblast (F), covers the entire assemblage of cells and offers structural support and integrity ; basement membrane Base ment membrane (B) seals the area between the sheath and cell layers ; around the cells, the glial zone (G) surrounds and contacts the cells especially the big cell ; trosphospongium (T) direct ly deliver s nutrients bypassing the cytoplasm.

PAGE 109

109 Figure 3 4 Schematic of sample preparation protocol On the right geometry of micro surface coil, A) T he whole cell B) T he bisected cell to expose the intracellular organelles C) The segmented model for facilitating the histological corr elation after MR imaging The intracellular compartments are assigned such as nucleus (blue dots), cytoplasm (yellow areas), and dense fibers (green) around the tiny cells.

PAGE 110

110 Figure 3 5 Schematic of sample placement in the s ample well embedding the RF coil T he sliced fixed s ingle cell in the agarose block lies on top of the RF coil embedded in the sample well. Then, the sample will be immersed in the water by the n ylon mesh and r etention ring They will be tightly sealed by PCR film

PAGE 111

111 Figure 3 6 Direct correlation of Aplysia neuron morphology in MRM and histology images. A ) 3D spin echo image of the entire fixed big cell illustrating cellular compartments using a micro surface coil at 600 MHz NMR spectrometer B ) C orresponding light microscopic image illustrating signal heterogeneity in the outer periphery and connective tissue C ) L ower resolution image D ) additional cell image. MR M scan parameters: 3D spin echo sequence with TE/TR = 8.4/1000 ms, Resolution = A) 7.8 x 7.8 x 15 m3 and C) and D) 12 x 12 x 15 m 3 A) FOV = 1 mm x 1 mm x 1 mm, C) and D) 1.5 mm x 1.5 mm x 1 mm, Matrix = 128 x 128 x 64, NEX = A) 10, C) and D) 6, AT = A) 22 hrs 51 min, C) and D) 13 hours.

PAGE 112

112 F igure 3 7 Potential subnuclear structure of single neuron s in MRM. A ) Potential subnuclear structure, nucleolus in a single neuron L7 of A californica employed as a whole cell model B ) bisected model visualized in the 3D MR images. MR scan parameters: A ) 3D diffusion weighted spin echo sequence with TE/TR = 14.2/2000 ms, Resolution = 23 x 23 x 31 m 3 FOV = 1.5 mm x 1.5 mm x 1 mm, Matrix = 64 x 64 x 32, diffusion gradient separation ( ) = 8 B ) 3D standard s pin echo sequence with TE/TR = 4/1800 ms, Resolution = 7.8 x 7.8 x 15 m 3 FOV = 1 mm x 1 mm x 1 mm Matrix = 128 x 128 x 64, NEX = 6, AT = 24 hr s.

PAGE 113

113 Figure 3 8 Representative 3D gradient echo MRM of the bisected model of the biggest single neuron s Hypo intense signal region corresponds to the nucleus and intermediate hypo intense signal region to cytoplasm and embedded tiny cell ; ECM shows the hyper intense signal. MR M scan parameters: 3D gradient echo sequence with TE/ TR = 11.5/300 ms, Resolution = 11.7 x 11.7 x 15 m 3 FOV = 1.5 x 1.5 x 1 mm, Matrix = 128 x 128 x 64, NEX = 10, AT = 16 hr s 40 min

PAGE 114

114 Figure 3 9 3D g radient e cho MRM of bisected model of single neuron s illustrating MR detectable unknown intracellular structures under investigation suspected to be rough endoplasmic recticulum or trosphospongium by its i nvaginat ion into the nucleus. MRM scan paramet ers: 3D gradient echo with FOV = 7 x 7 x1 5 m, spectral width = 50 k Hz, TR = 300 ms, TE = 10.6 ms, A verage = 10

PAGE 115

115 Figure 3 10 Diffusion w eighted i mages of slice of the big single neuron s embedded in the agarose. MRM s can parameters: TR/TE=20/2000 ms, r esolution = 11.7 x 11.7 x 100 m 3 NEX = 2 8, Acquisition Time = 2 hours, b values = A) 0 B) 1500, C) 2500, D) 4000 (d) [s/mm 2 ], diffusion gradient separation ( diffusion gradient duration ( E) Diffusion signal changes by increasing b values and diffusion times Light microscopic images of single neuron showing five different subcellular compartments at 40 x magnification. Of particular note is t he morphological delineation of more than two intraceullular compartments.consisting of nucleus (1) prenuclear cytoplasm ic region (2) caroninoxisommes (3) enriched with mitochondria and lipochondria functioning as photoreceptors, cytoplasm (4) and possi ble trophospongium (5) F ) Light microscopic image of single neuron showing five different subcellular compartments at 20 x magnification.

PAGE 116

116 Figure 3 11 Direct correlation of c ellular architecture of s ingle neurons of A c aliforn ica in MRM and histolog y images A ) G reen fluorescent colored PKH 67 stained plasma membrane B ) 2D Diffusion weighted images (in plane resolution = 7.8 m 2 with 50 m thick slice at b = 1500 [s/mm 2 ] C ) DAPI stained nucleic acid in the big nucleus in a si ngle neuron ; smaller cells and numerous tiny glial cells around the periphery of the big neuron D ) L ight microscop y image E ) Nissl parameters: TR/TE = 20/2000 ms, r esolution = 11.7 x 11.7 x 150 m 3 NEX = 100, FOV = 0.15 cm x 0.15 cm x 150 m, a cquisition t ime = 2 hours, b values = 1500 [s/mm 2 ], diffusion gradient separatio n ( ( = 5 ms/1 ms

PAGE 117

117 Figure 3 1 2 3D reconstructed image of single neurons The 3D visualization was generated after interpolat ing multiple slices o f 3 D MRM images in Fig ure 3 6 A N ucleus in the cell (dark purple), cy toplasm (light brownish orange) and glial cells (purple) in the extracellular matrix (green).

PAGE 118

118 Figure 3 1 3 Diffusion tensor images and reconstructed diffusion tensor images of single neurons A ) 6 directions of diffusion tensor images with the slice of the single neuron of A c alifornica (b = 0 ) B ) b = 1500 [s/mm 2 ] acquired along 6 different diffusion sensitizing gradients C ) R econstructed images with a software developed by Angelos [80]. MR scan parameters : TR/TE = 20/2000 ms Resolution = 11.7x11.7x150 m 3 NEX = 100, FOV = 0.15 cm x 0.15 cm x 150 m, a cquisition t ime = 2 hours, b values = 1500 [s/mm 2 ], diffusion gradient separation ( diffusion gradient duration ( ) = 1 ms

PAGE 119

119 Figure 3 1 4 Direct correlation of single neuron in diffusion weighted MRM images, ADC maps and histology images A ) Representative light microscopic image B) 2D DWI images acquired with 600 MHz NMR spectrometer with different diffusion time of 5 ms and C) 11 m s all with the same b value of 1500 s/mm 2 D ) DAPI stained fluorescent image of a slice of L7 neuron of A c alifornica Apparent Diffusion Coefficient ( ADC ) map of the slice by different diffusion time s of E ) 5 ms and F ) 11 ms.

PAGE 120

120 Figure 3 1 5 Direct correlation of a slice of single neurons in MRM and histology images A ) L ight microscopic image of the slice of L7 neuron describes the existence of tiny speckles in the nucleus and double nuclear membrane unique in this big neuron B ) 2D DWI images of the slice of L7 neuron of A c alifornica acquired with 750 MHz NMR spectrometer with different b values of 1500 [s/mm 2 ] C ) ADC map of the slice by variable diffusion time of 8.52 ms, 11.3 ms, and 19 ms. MR scan parameters: TE/TR = 30 /3000 ms, FOV = 1.5 x 1. 5 mm, slice thickness= 0.7 mm, matrix size 2 NEX = 40, AT = 14 hr s 56 min.

PAGE 121

121 Figure 3 16. Pre fixed and post fixed single neurons in MRM. A ) 3D g radient e cho images of L7 neuron in the live state B ) the same neuron after fixation. MRM scan parameters: 3D gradient echo, resolution = 7 x 7 x 15 m 3 spectral width = 50 k Hz, TE/TR = 10.6/300 msec, AVG = 6, AT = 9 hrs

PAGE 122

122 CHAPTER 4 MAGNETIC RESONANCE M ICROSCOPY OF RAT, PIG, AND HUMAN NEURONS Introduction As a non invasive imaging modal i ty, MR I has been utilized extensively in clinical and laboratory environment s to perform non destructive image acquisition. A fter its invention in the 1970s the number of applications for bot h biological and non biological samples increased exponentially Now, as a member of the limited group of microscopic imaging modalities, the term MRM was coined early by the early developers of MRI to describe microscopy conducted by MR: this designation describes MR applications conducted in the sub mi l limeter domain [40] MRM was first demonstrated in a biological sample in 1986 and the first MRM image to visualize single animal cell was conducted at 10 x 13 m 2 of in plane resolution [41] Since then, attempts to see further and deeper into structures ha ve been conducted throughout the field starting with plants and material studies, then progressing to mammals including rats which are common models for comparative biology [43, 44, 65, 88] This advancement can be attributed to enhancement in MR resolution. The first small animal sample, a frog ov um was visualized using MR by Aguayo et al. [41] For this stud y resolution s up to 10 x 13 x 3 were applied in order to obtain MR images of the nucleus and cytoplasm inside a 1 mm diameter, stage IV frog ov um In addition to ready access to isolation and sample preparation, another large different cell type, neural cell, of biological sampl e from aquatic animal was employed. It was an L7 neuron isolated from the abdominal ganglion of Aplysia c alifornica Th is animal CNS model that possesses characteristics of m emory, exhibits hypertrophy of the large

PAGE 123

123 neuron s and was useful in genome research [66] This models neuronal structure w as delineated using a solenoidal coil which possessed high si gnal uniformity but relatively low signal sensitivity As such, previous MR microscopic images in these experiments were limited only down to 20 x 20 x 100 m domain [33, 43, 81, 82] Given the need for using basic research to answer questions relevant t o clinical practice, the ultimate attainable MR resolution needed to be increased to levels necessary to delineate typical human cells resolution could only recently be achiev ed and was accomplished primarily through recent technological update s such as stronger and fast switching imaging gradients [14] and micro surface coils [18, 46 49] Likewise, a major focus in the field of MRM is the increase of resolution, and the name MRM itself was coined to describe MR imaging below an in plane r esolution of has the capability to resolv e more than one object in a voxel, MR images with higher sensitivity can lead to the lower imaging acquisition time to get the sam e SNR T here are still challenges to overcome regarding the improvement of MRM resolution. Since MRM is a method of MRI, it shares common features with MRI, but, given the need to acquire MRM in a high magnetic field with a small volume of voxels, there a re some features unique to MRM One such feature are the line broadening effects or T 2 (T 2 ) effect s that widen the spectral line widths at higher ma gnetic fields due to interactions between neighboring proton s These effects can be alleviated by employing stronger fast switching imaging gradients and making the magnetic field as uniform as possible. Another issue to overcome is the relative lack of sensitivity Due to

PAGE 124

124 the limited volume of detection ( voxel volume ) in high resolution images fewer interact ing protons are available to be detected. When the protons fluctuate or change spatial positions by self diffusion the detectible signal becomes limited. B y employing a stronger gradient, this effect can be remedied. In the current study, by employing sta te of the art equipment for higher resolution, greater sensitivity and better specificity for MRM, MR stained cellular structure inside the mammalian and human neural tissue w as investigat ed. As a thick sample can prevent light penetration when using ligh t microscopy techniques employing a 2D projection imaging scheme with low through plane resolution result s in loss of contrast due to volume averaging in MR images 3 ) mammalian cells were successfully delineated f or the first time using MR methods [50]. I n the current study this unprecedented resolution and sensitivity was applied to the imaging of various biologica l samples including Purkinje cells in the rat cerebellum and big cell s in the ventral and dorsal horn s in rat and human spinal cord s. These tissues were used to investigate the origin of relaxation and diffusion signal in the CNS. particularly with respec t to the intr a cellular compartment. Additionally, for the first time in MR I MR signal was detected from the intracellular compartment in the bush like neuronal process es emanating from cell bodies in a porcine sample. These structures w ere visualized usin g endogenous MR contrast and identified using histologica l images of the tissue after undergoing MR. MR signal characteristics of the rat cerebellum were investigated at the cellular level. a region of a brain gov ern s many important functions such as some cognitive and movement related function. Especially, it plays a pivotal role in fine motor coordination but manifest s a

PAGE 125

125 complex structural map [181] E ven at low magnification, the collection of cell bodies in th e gray matter (GM) over underling white matter (WM) can be appreciated. These white mater pathways deliver messages between cells In the cerebellum, t he GM is com posed of three different layers. T he external layer named the granular layer is a layer of n euromodulary neurons covered with moss like fibers T he inner layer is named the molecular layer T he Purkinje cell layer ( a layer of pure output neurons) is located between the two others The granule cells project in parallel with the Purkinje cells. By employing MRM, the different anatomical and functional layers were investigated in terms of signal relaxivity and water diffusivity in the cellular compartments [182] Interestingly, in the WM of the cerebellum ectopic (= out of place) Purkinje cells wh ose diameters were approximately were observed. Their existence in the dorsal brainstem and cerebellum has been previously reported [89, 90] Due to their prominent size even when embedded in nerve fibers and neighboring fatty myelin sheath in the white matter (WM) they were visi ble at 20x magnification light microscopy in Fig ure 4 3 Compared to normal Purkinje cells, two aberrant characteristics -shape and projection -were observed in the ectopic cells : less dense dendri ti c trees and irregularly studded non planar branches. T he reason for this being was, during prenatal development of the cerebellum when the cells experienced a massive migration, they moved out of place [89, 90] Because of their atypical location between the brain and the cerebellum, ectopic cells often cause pathophysiological conditions, e.g. Chiari malformatio n and ectopic cerebellar dysplastic tissue [91] In addition to their prominent location, their irregular and peculiar size s (50~70 make them suitable for structural and functional investigation into intracellular compartmentation by MRM.

PAGE 126

126 The quantitative and qualitative analysis of MR information from deep tissue identifies the histochemical correlation and investigates the magne tic properties at the molecular and cellular levels. The 3D analysis conducted at the in plane resolution of 8 m employed here were carried out to investigate cellular micro architecture using MR. Additionally, 3D visualization as used in MRM can describe structure inside dense tissues, in locations unobtainable with light microscopy due to limitations caused by light penetration, absorption and scattering. The current study illustrates the capability of MRM to delineate micro scale maps and to play a role in demarcating cellular MR characteristics in neural tissues as a biopsy methodology. Methods Sample Preparation and MRM Imaging All the imaging work was carried out using a 600 MHz Oxford spectrometer equipped with a Bruker Biospin console providing i maging gradient strength s up to 3000 mT/m. The console itself is interfaced with micro surface coils, which were developed by Bruker Instruments [18, 45] B rain samples from the striatum as well as spinal cord (SC) slices containing motor neurons from rat s and human w ere harvested. After being wash ed in phosphate buffered saline (PBS 137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 and 1.8 mM KH 2 PO 4 : pH 7.4 at 300 mOsm ) overnight to remove any remnants of 4% formaldehyde t he striatal brain slice s w ere acquired by using a vibratome (Ted Pella, Lancer series 1000 ), and imaged using a 200 m coil (Bruker, Z76412) whereas the rat spinal cord was imaged using a 500 m micro surface coil (Bruker Biospin, B6370). The human cell was prep ared before the MR imaging with a dissecting scope (Zeiss, OPMI 1 FC) or light microscope ( Zeiss, Axioplan 2 ) (F ig ure 4 8 A ) Through a r igorous selection process, a reasonably big cell was selected and

PAGE 127

127 scaled with the inner diameter of the 50 m coil. Add itional accessories for mic r o imaging were developed custom made for the samples under investigation [50] T o place a thin slice as close as possible to the region of highest coil sensitivity a nylo n mesh screen with pore size of 50 m (Small Parts, CMN 0053 C) and a retention ring fashioned from a 3 mm NMR tube cap (Sigma Aldrich, NORS36007) were employed in Fig ure 4 1. Minute manipulations using Dumont forcepts were tracked visually with a dissecti ng scope (Zeiss, OPMI 1 FC) or light microscope ( Zeiss, Axioplan 2 ). Once all the retention parts and the sample were tightly secured in the tissue well, the sample was irrigated with PBS to avoid water loss due to evaporation during the long hour scan. Fi nally, t he last step involved using PCR film (ABgene, AB 0558) to seal the tissue well so that the tissue remained hydrated during scanning All animal procedures were conducted in accordance the Care and Use of Laboratory Animals and were approved by the University of Florida IACUC. M R Pulse Seq u ence Three dimensional MR image s of motor cells of rat brain were obtained using a conventional 3D gradient echo sequence at 4.7 m isotropic reso lution ( TE = 10 ms, TR = 150 ms, matrix size = ( 128 3 ) FOV = ( 0.6 mm 2 ) total acquisition time = 22 h rs bandwidth = 25 kHz, read and phase gradient amplitudes = 1980 and 1830 mT/m, respectively ) For visualization of the rat motor neuron s c onventional spin echo diffusion ms, slice thickness = 80 m, matrix size = 256 256, in plane resolution = 7.8 m 2 total acquisition time = 7 h 7 min, bandwidth = 50 kHz, read and phase gradient amplitudes = 591 and 650 mT/m,

PAGE 128

128 respectively) were acquired. For rat perkinje cells in the cerebellum, a 3D gradient echo ( TR/TE = 15 0 ms/ 7 ms, res = 6.25 m isotropic, Avg = 14, scan time = 63 h 43 min ) imaging protocol was employed. In order to examine diffusion behavior in the various tissues of the cerebellum, 2D diffusion weig h t ed images ( b = 4025 s/mm 2 = 6 ms, slice thickness = 80 m, matrix size = 256 256, in plane resolution = 7.8 ( m ) 2 total acquisition time = 7h 7min, bandwidth = 50 kHz, read and phase gradient amplitudes = 591 and 650 mT/m, respectively ) were employed To illustrat e pig motor neuron s and their affiliated neuronal processes in spinal cord slices (n = 3), a micro surface coil of 100 m diameter was employed (Bruker, B6372) and three dimensional spin echo images (TR/TE = 2000 ms/ 12.75 ms, isotropic resolution = (6.25 m) 3 Avg = 14, scan time = 63 h 43 min) were collected. T o investigate intracellular diffusivity 3D diffusion weighted images (TR/TE = 2000 ms/6.2 ms, res = 6.25 m isotropic, Avg = 14, scan time = 63 h 43 min) were collected at two b values (b = 300 s/m m 2 ; b = 600 s/mm 2 ) and used to map the apparent diffusion coefficient (ADC) in the sample. To visualiz e motor neurons and investigat e the intracellular diffusivity inside the se neurons, 2D d iffusion weighted scan protocols (TR/TE = 2000 ms/23.5 ms, ms, res = 7.8 m 7.8 m 50 m; b = 2000 s/mm 2 Avg = 40, scan tim e = 5 h 40 min) were employed. Image Analysis and Structural Validation For mammalian cell imaging, after MR data acquisition tissue slices were removed from the imaging coil and placed in a Nissl stain (0.5% Cresyl Violet, 0.3% glacial acetic acid in 99.2 % ddH 2 O) for one minute in the case of rat sample s and two minutes for

PAGE 129

129 human sample s Slices were then placed temporarily in a destaining bath (0.3% glacial acetic acid in ddH 2 O) to remove excess Nissl stain followed by a PBS wash before being wet mounted to pre cleaned optical microscopy slides (Fisher, 12 550 13) using Histomount mounting medium (National Diagnostics, HS 103). False color histological images were taken using a digital camera (QImaging, Retiga 4000R Fast 1394 Color) attached to a Zeiss mic roscope (Axioplan 2, Zeiss) and processed with QImaging software (QCapture Pro 6.0). Rat sample i mages acquired at 100 x magnification and human sample images acquired at 20x magnification employed an emission filter in the Texas Red range. For the rat cere bellum images bright field microscopy at 20 x magnification was utilized due to sufficient endogenous contrast from ectopic Purkinje cell s coming from the fatty myelin covered WM. 3D Segmentation and Reconstruction Three dimensional spin echo datasets inv estigating such structures as cell bodies and striat um in the rat brain and pig spinal cord tissue were reconstructed utilizing imaging software (Visage Imaging Amira 3.1.1) and segmented in order to visualize cellular structures from multiple view points. Considering the smaller size of the rat brain cell bodie s of alpha motor neurons in the striatum, the hypo intense signal was picked to avoid inevitable partial volume averaging along the boundary and interpolated with pixels from adjacent slices. By inves tigating larger cells motor neurons in the pig spinal cord, we had the opportunity to use eight voxel thick frames (50 m of resolution in the third dimension ) from the 3D dataset in the reconstruction thus eliminating areas in the dataset with uneven RF excitation on each end. First, a thresholding limit (0 8427 au : 8428 24652 au) was employed to set the range of the binary map consisting of

PAGE 130

130 voxels containing motor neuron perikarya and subsets of their processes (0 8427 au) and those containing tissue adjacent to these structures (8428 24652 au). Then the voxels residing in the noise region were manually eliminated so that our reconstruction did not contain data from background noise used to define cell regions inside ( shown as pink in images ) and surrounding the motor neurons. Results D elineat ion of C ellular A rchitecture of M ammalian B rain C ell s in the CNS T issue of R ats using MR M icroscopy using N ovel M icro S urface coils and 3D V isualization In this preliminary study, sample selection in particular its thickne ss is critical for successful MR imaging and histological validation. F igure 4 1 shows preliminary mammalian brain cell images acquired with a coil. It contains a vertical line which is present over three consecutive slices This struc ture was identified by its location and morphology as a descending WM tract This tissue is covered with a fatty myelin sheath and is surrounded by dispersed structures possessing hypointense signal characteristics. It is believed that these dark structu res correspond to the cell bodies of medium spiny neurons ; however, identification through direct histological correlation was not forthcoming Considering the small size of the structures ubiquitously dispers ed over the region of interest, a one to one correlation of MR images with the histological images was not practical at the time of MR image acquisition As shown in F ig ure 4 2, the suspected cell bodies and W M tract were visualized using a 3D visualization tool : AMIRA 3.1.1. T he non invasive nature of MRM allowed for visualization of the microstructures deep inside tissue s of the rat brain. In the 3D rendered images, the tissue represented by red transparent m esh structure is

PAGE 131

131 suggestive of the morphology of medium spiny neuron cell bodies while the pink solid line is indicative of a WM tract. Through Nissl stained histological images, the cerebellum is illustrated in F ig ure 4 3 A The g ranular layer consist s o diameter ) locat ed between the WM and the Purkinje cell layers These granule cells are surrounded by glial cells and stellate neurons. Their dendrit ic arbors are relatively short and project in parallel with those of Purkinje cells which receive signal s from climbing fibers of the inferior olive originating from the caudal brainstem. These fibers play an important role in carrying signal reception into the cerebellum and motor output to the motor areas From the false color images in F ig ure 4 3 B we can see that the WM is not translucent due to itsfatty myelin cover. The brighter areas correspond to granular cells which are contacted by the dendritic trees of the cells in the neighbor ing molecular The 3D GE sequence visualized sus pected cell bodies in the WM tissue of the cerebellum Considering the importance of landmarks to successful identification of tissue microstructure, we test ed whether MR visible structure s could be well correlated with structures identified in histologica l images Using WM tract was preferred for this purpose as it can be readily identified in both MR and histology The ectopic Purkinje cells presenting as hypointense signal in MR images were visualized along the WM tract. Due to the high strength of the i maging gradient s the intracellular compartment of Purkinje cell s appeared quickly diffusing spins in a limited volume of voxels but smaller cells could not be delineated due to volume averaging effects. The tissue types of the cerebellum w ere illustrate d using a 2D diffusion weighted spin echo sequence. These regions possess different MR signal characteristics or

PAGE 132

132 inhomogene ous signal distributions ( Fig ure 4 6) The neighboring region of the WM ( the hypointense region in ( Fig ure 4 5 C ) is composed of tight ly packed moss like fibers and neuromodulary neurons originating from pre cer e bellar ne urons in the brainstem It changes to a hyperintense signal region at b values of 3000 [s/mm 2 ] and above presumably due to the crowding effects experienced by water pro tons in the WM mesh like structure. Even though d elineation of individual tracts is not feasible at this resolution and sensitivity, the characteristics of water d iffusion in each tissue layer as measured by MRM shows that the cerebellum is a high ly organized, complex structure. Confirmation of the D elineated C ellular A rchitecture of N eural T issues with C orresponding H istological V alidation using MR M icroscop y Using 500 m micro surface coils, the cell bodies of motor neurons in th e ventral horn of the human spinal cord were scanned C orrelation work identifying structures by using histological stains w as carried out in collaboration with Dr. Jeremy Flint Inside the Nissl stained cell bodies the intensely stained area s correspond to the endoplasmic reticulum which is located inside the cytoplasm of the cells. Stil l, despite our advanced imaging hardware, substructures inside the intracellular compartment s could not be delineated ; however, the potential to visualiz e the intracellula r compartment may be appreciated from the results presented here Of note is the observation that the hypointense signal seen in high b value imaging might not correspond in 100% of cases to the cell body. As indicated by the yellow arrows in both the hist ological image and MR image, the signal void region in the MR comes from degradation of tissue integrity during the post mortem or tissue processing period.

PAGE 133

133 Delineat ion of N eural P rocesses in P ig S pinal C ord E mploying MR M icroscop y T echniques and M icro S u rface coils Three representative images from the pig spinal cord containing neural processes are presented in F ig ure 4 7 A From the 3D spin echo images taken at 6.25 m isotropic resolution, the cell bodies of motor neuron s and the ir relatively short an d thick dendrites present as an area of hypointense signal T he smaller and fine r structure of the dendrites has been blurred by partial volume effects. The continuity of the projection s w as confirmed by segmenting and reconstructing consecutive slices wit h 3D visualization software, AMIRA (F ig ure 4 7 ) Using built in threshold ing parameters the region of interest (ROI) corresponding to the visualized intracellular space (i.e. portions inside the motor neurons) was segmented out from the surrounding tissue by its contrasting signal intensity. Since the cell bodies in this dataset occupy eight consecutive slices in the through plane direction the cell diameter, 50 m, is resolved well and its bou ndaries are delineated without any noticeable distortion. In the reconstruction of the 3D segmentation images, the cell s are visualized in transparent pink and the surrounding structural protein is shown as green semi transparent mesh like lines in order t o distinguish the spatial arrangements of different tissues inside the spinal cord. Preliminary E vidence of I ntra cellular C ompartment ation and H ow T his R elates to F ast and S low D iffusi ng W ater P ools I nside the H uman N euronal C ell U tilizing M icro S urface co ils The isolat ed cell, presumably an ectopic Purkinje cell, in the dorsal horn of the human spinal cord has a high level of segregation of hemispheric cellular compartment s shown in the inset of Fig ure 4 8 A As we have observed in earlier images (Fig ure s 4 1 and 4 6 and 4 7) with high resolution 3D gradient echo by the 200 m coil and 2D

PAGE 134

134 diffusion images by the 500 m coil the subcellular compartments were not delineated; however, with the diffusion weighted imaging scheme collected at a b value of 1000 [ s/mm 2 ] with the 100 m coil the intracellular compartment s manifest in images with heterogeneous signal distribution (Fig ure 4 8) The hyperintense signal corresponds to the cytoplasm and the hypointense signal matche s the nucleus. N ote here that consider ing the limited number of voxel s used to image the human cell (50 m/7.8 m = 6 voxels) small mammalian cells or subcellular structures suffer from partial volume effects even at the resolutions employed here. Discussion Considerations W hen I maging H uman and M ammalian C ells To our knowledge, this research is the first study to report MR visualization of mammalian and human neuronal cells ex vivo Even though there is more work to be done prior to applying these methods directly to MR signal analysis in vivo this work represents a significant advance in assessing MR sig nal behavior at the cellular level using a recent technological advancement. Microscopic resolution in MR scans reveals the internal signals of cellular compartments, which are conventionally averaged out by the coarse voxel sizes used in clinical MRI. By analyzing the components inside the voxel, our study contributes new and unique insights to generating inherent contrast while interplaying with the applied magnetic field at the cellular level in a bottom up approach, and hopefully it will contribute to clinical treatments as a reference for monitoring and analyzing the pathological and anatomical information at the tissue and organ level of the human. Care must be taken in the described experiments starting with sample preparation. For the preparation o f CNS tissue, anatomical structures were well preserved by the

PAGE 135

135 perfusion fixation method in order to remove any iron containing hemoglobin in the blood [214]. However, human samples manifest loss of integrity around the cellular components originating from the inevitable post mortem interval (Fig ure 4 6). This prominent void could be easily confused with a cell body in diffusion MRM with certain scan characteristics: in this case, high through plane volume averaging. Therefore, sample thickness and resoluti on should be chosen carefully depending on the structure that is being imaged. Due to methods necessary for histological validation steps, thinner samples are preferred so that light may effectively penetrate them; however, using thinner samples results i n a higher risk of mechanical tearing as well as interference from free water signal, the last point being problematic for diffusion images taken atlow b values images. Last but not least, positioning with the aid of a light microscope prior to the long MR scan could increase the accuracy of histological validat ion, especially when dealing with the intracellular compartments Compartmentation I ssues As discussed earlier, fast diffusion outside the cell agrees with conditions in special environments such as cell culture or cell suspension where access of the media to the cells is higher and appears to behave as free water does in diffusion images. However, whether this view can be applied to in vivo biological system remains questionable since there are othe r factors involved in functional tissue constructs such as ECM, glial cells, and the like. In this study, rather than employing exogenous substances and molecules such as Manganese (Mn), Gadolinium (Gd), Cesium (Cs) which are larger than water [92, 147, 14 8], we explored the endogenous contrast of neural processes and intracellular compartments in the human cell. Neural cells in other mammalian animals were visualized and shown to exhibit higher intracellular than

PAGE 136

136 extracellular water diffusion. It is intere sting to observe that the membranes from the visualized neuronal processes do not create significant restrictive barriers to water diffusion.. As was true in the cellular model, Aplysia c alifornica in this preliminary study, the MR relaxivities of the nuc leus and cytoplasm were delineated. This was accomplished through meeting some important criteria including choosing the right sample for visualization. Factors such assize, geometrical distribution of intracellular compartments, and optimal choice of puls e sequence were all important to realizing our imaging goals. However, when interpreting the diffusion signal possessed by the cellular compartments, extra attention must be paid in order to determine if they imply a true division based on compartmentation or rather degrees of hindrance or restriction common to both compartments. Cell S welling In order to investigate the mechanism behind the decrease in ADC in the event of cell swelling followed by stroke, MR research conducted thus far has suffered from p artial volume effects due to low resolution and T2 (T2*) effects due to long diffusion times. For example, at high b value, a hypointense signal region should be validated to confirm whether these signal properties arise from T 2 effects or from a fast dif fusing water pool. Alternatively, for the hyperintense signal region, validation should be performed to determine if these signal properties originate from T 2 shine through [96] or restriction effects. In addition, with higher resolution, sensitivity, and specificity, MRM can support this validation process by reducing the diffusion time ( 1ms) with help from newly developed OGSE pulse sequences and strong, fast switching planar gradients. These methods are required to successfully probe mammalian cells in terms of issues

PAGE 137

137 related to compartmentation and the origins of non monoexponential MR diffusion signal decay. C onclusions and Future Work In this study, we examine cellular level MRM on the neural cells in mammalian CNS tissue. With dedicated hardware, the resultant resolution may finally reveal the complexity of the brain in the cellular domain using MR methods Hopefully, this investigation on the cellular origin s of MR signal, relaxation and diffusion, will contribute significantly to the analysis of physiology and detection of pathology in the clinical environment through relevant mod eling.

PAGE 138

138 Figure 4 1. Schematic diagram of sam ple placement in the sample well on top of RF co il. A) The sample well B) micro surface coil C) the sample of interest. D) N ylon mesh for sample suppression E) Retention ring. F) PCR film for s eal ing the water leak age

PAGE 139

139 Figure 4 2. MRM of rat striatal tissue MRM scan parameters: 3D gradient echo TE = 10 ms, TR = 150 ms, m atrix = 128 3 resolution = 3 FOV = 0.6 mm 2 acquisition time = 22 h rs Of particular note is that the h ypo intense dots correspond to the suspected cell bodies and the vertical tract corresponds to a WM tract.

PAGE 140

140 Figure 4 3 3D segmentation image of the striatal tis sue Images were reconstructed after interpolating the corresponding cellular structure s using AMIRA software Visualization of cell bodies (red), WM tract (pink), tightly embedded in the neighboring brain tissues including extracellular matrix and other s maller and supporting cells (beige structures in the background) are presented

PAGE 141

141 Figure 4 4 F alse colored light microscopic image of rat cerebellum. It visualiz es the inherent contrast between WM (dark) and GM (bright) consis ting of granular layer (1), Purkinje cells layer ( 2 ) molecular layer ( 3 ), and inter sulci space (4).

PAGE 142

142 Figure 4 5 Direct correlation of ectopic Purkinje c ells of the WM of cerebellum in MRM and histology images. A) and B) 3D gradient echo MRM images of cerebellum. C) L ight microscopic image of cerebellar region of interest before MR D) The tissue slice placed securely between the coil surface with a nylon mesh to suppress the tissue sample to maintain firm contact with the c oil surface. MR scan parameters: TE/TR = 10/500 ms, r esolution = 8 x 8 x 16 m 3 FOV = 1x 1 x 1 cm 3 m atrix = 128 x 128 x 64, NEX =10, AT = 11 hr 22 m in.

PAGE 143

143 Fig ure 4 6 E ctopic Purkinje c ells of the WM of cerebellum in diffusion weighted MRM images with different b values. A) and B) The hypointense region in low b values corresponds to the WM and C) through F) H yperintense signal region in the images taken at high b values corresponds to the granular layers. U nit s are [s/ mm 2 ] M R scan parameters: 2D diffusion weig h t ed images ( b = A) 0, B) 400, C) 1000, D) 3000, E) 7000, F) 12000 s/mm 2 ms, slice thickness = 80 m, matrix size = 256 256, in plane resolution = 7.8 ( m ) 2 total acquisition time = 7 h r 7 min, bandwidth = 50 kHz, read and phase gradient amplitudes = 591 and 650 mT/m, respectively

PAGE 144

144 Fig ure 4 7 MR i mages of motor neuron cell bodies from human spinal cord (D, E, F) and its corresponding correlative histology (A, B, C). Green arrow indicates the cell body and yellow arrow indicates tissue tears.

PAGE 145

145 F igure 4 8 M RM images of neuronal process in the ventral horn of the pig spinal cord in MRM and 3D visualization. A) The MRM images were acquired with 3D T 2 weighted spin echo imaging scheme B) T he corresponding 3D segmentation reconstruction was made using AMIRA sof tware. Of particular note is the v isualization of soma and dendrites (filled pink) tightly embedded in the tissues of the ventral horn (transparent green mesh).

PAGE 146

146 Figure 4 9. Direct correlation of e ctopic Purkinje cell in the dorsal horn of the human spinal cord in MRM and histology images. A) For scale purposes, the micro surface coil whose inner diameter was 50 m was aligned under the cell of interest and photographed at magnifications of 20x and (inset) 40x using light microscopy B) through F) C orresponding MRM images acquired using a 100 m micro surface coil T 2 weighted images with 100 m coil with TE = B) 2 5 ms, C) 30 ms. D iffusion weighted images taken with a 100 m coil b value of D) 0 [s/mm 2 ] E) 1000 [s/mm 2 ] F) 1700 [s/mm 2 ]. Of particular note is that there is the signal inhomogeneity in the intracellular domain at a b value of 1000 [s/mm 2 ].

PAGE 147

147 C HAPTER 5 DIFFUSION TENSOR MIC ROSCOPY OF THE CELLULAR STRUCTURE O F KIDNEYS AND LIVER IN THE AUTOSOMAL RECESS IVE POLYCYSTIC KIDNE Y DISEASE (ARPKD) MOUS E MODEL Introduction Another emerging application of MRM involves the delineation of tubular cells in the p arenchyma of the kidneys and liver. Of course like other imaging modalities, MRM c an be combined with exogenous imaging markers : for instance, employing gadolinium or manganese to facilitate T1 weighted imaging [97 100] Because gadolinium based contrast agents have been shown to result in unwanted complications, especially for patients with end stage renal disease [192, 193] applications of endogenous tissue contrasts are preferred. In that regard, one of the benefits of using MRM is that by employing inherent contrast between b iologic tissues, it c an measure both macroscopic change s in the overall ductal and tissue volumes and microscopic change s in the physiological consistency of tissue structures. Such changes include the size, number, spatial distribution, and growth rate of cysts in the kidneys and liver. Given the increas ed need for preclinical trials and therapeutic interventions, the non invasive nature of MRI has been used to great benefit as an early diagnostic indicator for specific diseases such as a utosomal r ecess ive p olycystic k idney d isease ( ARPKD ). MRM has the capability to detect early anatomical changes in organs, thus increasing the time in which to apply therapeutic treatments. importance as a diagnostic tool in clinical trials and therapeutic treatmen ts has been demonstrated and, in the current study, we intend to investigate its potential in regards to ARPKD To elucidat e a practical strategy, a brief review on ARPKD and the structure and function of the kidney is presented.

PAGE 148

148 S tructure and Function of the Kidneys and Liver Kidneys as constituents of the urinary system play a role in eliminating waste products by filtering and excret ion Most waste products are cellular wastes the end products of nutrient metabolism such as glucose, fatty acids, and a mino acids T he resulting, mostly nitrogenous wastes are transformed and filtered into urine by the kidneys. When the kidneys fail to function properly the accumulation of urea in the body manifests as daily blood urea nitrogen (BUN). As a result, BUN is used as an indicator of kidney function. Another function of the kidneys includes regulation of the constituents of the extracellular fluid: for instance, Na + Ca 2+ and c hloride. T he overall regulation of the particles is described as the regulation of o smolarity T he extracellular fluid (ECF) is regulated by circulating it in a form of plasma through the kidneys. T he kidneys also act a s an endocrine delivery system that secrete s hormones into body fluids One of the hormones is erythropoietin (erythro = red) that stimulates bone marrow to promote the formation of red blood cells ( RBCs ).Bone marrow cells require copious oxygen delivered by the RBCs for the regulation of the aforementioned constituents of the ECF. Another hormone secreted by the kidney is r enin, which activates angiotensinogen to change its form to angiotensin. The angiotensin causes generalized vaso constriction, that is it constricts the smooth muscles in the blood vessels leading to decreased blood circulation. The resulting increase of m ean systematic filling pressure makes the heart generate a greater pressure gradient to distribute the blood between arteries and veins. W hen a tumor compresses the renal arteries, this triggers the release of renin resulting in a signal that more blood is requir ed by the kidneys If this blockage continues as is observed in ARPKD the outcome is higher blood pressure and renal failure in children after neonatal stage [194] Therefore,

PAGE 149

1 49 given the weak prognosis, MRM is proposed to be a good indicator to unde rstand the clinical manifestations of ARPKD G ross Anatomy of the Kidney Kidneys are a pair of bean shaped organs that, on the ventral side ha ve a medial indentation, the hilum, which serves as the entry and exit point Their location in the abd omen is retroperitoneum retro = behind) the peritoneal membrane on both sides of the vertebral column below the 12 th rib D ue to the position of the liver on the right side of the abdomen the right kidney is lower than the left by approximately 8 .25 cm [1 95] They are embedded in the peritoneal fat which acts as a structural support much like the fat around the eyeball and unlike other body fats such as subcutaneous fat or intra musc u l ar fat that serve as energy storage. The internal structure of the kidne ys consists of many cup shaped filtering unit s called nephron s. These structures are composed of the glomerulus and tubules, surrounded by a hollow sinus that accommodates such tissues as the renal pelvis, blood vessels, and fat. Interestingly, the lobes c onsist of 5 to 11 individual lobules (seven in human s ) and are collectively called the renal medulla They are separated in the embryonic stage of development, but fuse into a pyramidal shape by adulthood. The striated renal pyramids are aligned in paralle l with tubes that widen toward the apex in the coronal section and act as a collecting duct. At the apex of the pyramid, the renal papilla finger shaped structures project into the minor calyx. Between the striated renal pyramids or medulla, there is the renal column whose tissue type is the same as the cortex. M icroscopic Anatomy of the Kidney The functional unit of any organ is defined as the smallest collection of the component parts necessary to perform the required function In the case of th e kidney,

PAGE 150

150 the functional unit is the nephron. Each kidney is composed of approximately 300,000 to 1 million nephrons on average even if the actual numbers of the nephrons may be varying on an individual basis, and each nephron is composed of renal corpuscl e, renal tubules, and peritubular capillaries [196] The renal corpuscle is composed of a internalized capillary network known as the glomerulus surrounded by a sphere shaped structure for filtrate collection known as [197] During embryon ic development, as the tubular parts expand, the end s form the spherical shape of This tissue consists of the parietal layer on its outer rim and the visceral layer on its inner rim, and serves as a filter. In the vascular system, the cap illaries play the role of a filter allowing metabolites to pass between parenchyma and blood vessels [197]. The glomerular filtration barrier covering capillaries inside consists of capillary endothelium, glomerular basement membrane, and podocyte foot processes, filter waste products such as nucleic acid, uric acid, blood plasma, and extra salt through the nephron. Wrapping around the capillaries, there are the podocytes (podo = foot) and pedicels which extend from the podocytes creating slits between them. T he blood is filtered through the se slits in what is known as the slit diaphragm [ 101] Second, the renal tubule is an extension of the and is composed of four segments: a proximal tubule, the loop of Henle, a distal tubule, and a collecting duct. The first segment the proximal tubule, is described as a convoluted tubule locate d in the cortex and the renal column, which have the same tissue composition. Each is approximately 14 mm long and 60 m in diameter, making the length roughly 250 times greater than the width T he proximal tubule where the

PAGE 151

151 maj ority of the absorption forms the filtrate is through which the plasma filtrate passes known as the loop of Henle It is a large, hair pin like dual loop locate d in the striated renal pyram id and ascend ing to the cortex. Its primary function is to concentrate the urine by absorbing water. The third segment is known as the distal tubule, and it presents as a convoluted, and twisted segment of the nephron It is locate d inside the cortex and is almost the same length as the proximal tubule. The final segment is the collecting duct, which passes through the striated region inside the renal pyramid and open s at the renal pelvis for the urine to pass through. T here is a dense vascular network ar ound the aforementioned tubular parts which is known as the peritubular capillaries. These capillaries merge into the interlobular and arcuate vein s coming back to the kidney. These four components are the main structural parts of the functional unit of th e kidney, the nephron. Using these four components, the kidney s serve to filter, reabsor b and secret e T hey are essential to surviv al and it is invaluable to understand the details regarding their structure and function. The following experiments focus o n the delineation of differences between normal and diseased (PKD) kidney structure. I ntroduction to Polycystic Kidney Disease PKD P olycystic kidney disease (PKD) is a primary cause for end stage renal failure which affect s more than 500,000 people in th e United States alone [102] It manifest s as a developmental abnormality such as defects in nephron formation, renal cystic disease, and congenital hepatic fibrosis, and is most likely, but not exclusively, caused by a hereditary genetic disorder [198, 199] Among its many symptoms the most recognizab le is the formation of cysts in the kidneys These cysts are often caused by

PAGE 152

152 aging, dialysis, drugs, and hormones [103] or as a genetic disorder. T hese genetic disorders can be autosomal dominant or recessive depending upon the mode of inheritance of PKD or nephronophthisis, medul lary cystic diseases, and the like Among the multiple types of PKD ADPKD i s the most common [200], and manifests in one of two types. Type I is caused by a mutation of the pkd 1 gene and affects up to 85 to 90% of people with the disease [149]. T ype II is due to a mutation of the pkd 2 gene and account s for the remaining 10 to 15% [104] I n spite of diff erences in the onset of symptoms and progression to renal failure b oth genes manifest clinical pathologies and share the same origin arising due to the by products, polycystin 1 or polycystin 2 formed in the renal tubular epithelia [104, 105] It is charact erized by the rapid proliferation of cysts in both kidneys, and has clinical manifestations of hematuria, polyuria, hypertension, degraded urinary concentration ability and flank pain. Along with a number of cysts in the kidneys, the ir size can create the potential for disease developments to other organs such as the liver, pancreas, and heart [106] A less common type of PKD ARPKD is cause d by a mutation in PKHD1 [107, 108] Compared to ADPKD which occurs in around 1 out of every 400 to 1000 live birth s ARPKD occurs in about 1 out of every 20,000 li ve birth s It presents as elongation, enlargement, and dilation of the collecting ducts in conjunction with intestinal fibrosis [102, 198, 199, 200]. T he parenchyma of the kidney is modified into a rough, sponge like consistency and the organs themselves e nlarge As a critical amount of enlarged ducts compress neighboring normal ly functioning tissues such as nephron s and blood vessels, BUN becomes elevated and urinary concentration is impeded. These events are followed by failure of the absorption function of the collecting ducts. At

PAGE 153

153 the fetal or neonatal stage, these developments disrupt formation of the lung s ( pulmonary hypoplasia ) and result in marked enlargement of both kidneys often necessitating kidney transplantation. Should the child surviv e it de velops renal failure and portal hypertension or hepatic fibrosis P ortal hypertension comes from interstitial fibrosis express ed by the formation of small to massive cysts originating from stretched and elongated collecting ducts in the renal system H epat ic fibrosis presents as abnormal development of the biliary ducts, and enlarged and fibrotic portal tracts [109] R ationale for Investiga tion of the Cellular Domain As discussed above, in addition to the incidence rate of two genetic disease types, ADPKD and ARPKD the survival rate is also comparable. Almost half of ADPKD patients encounter end stage renal disease by age 60 but most infan ts who surviv e the perinatal period will develop renal failure by adolescence. Both disease s terminating in renal failure arise from the cellular domain, i.e. cyst lin ed epithelial cells in tubular segments that have Na/K/ATPase in their basolateral membr ane [110, 201] As the cysts form and develop in the nephron, the cells swell and the distance between molecules and tubular cells can not communicate by widening the diffusion distance. T he re is a difference in the site of origin between the two disease types. Specifically, in ADPKD cyst complication s form in the tubular seg ment of the nephron that is, the distal tubules, proximal tubules, and the loop of Henle In ARPKD however, cysts form in the renal epithelia of the collecting ducts that regulate re absorption of water, sodium, and potassium [111] By understanding the characteristics of these disease s in the cellular domain, our diagnostics and therapeutic treatment options will improve To this end, MRM as a non invasive diagnostic tool will aid in elucidating the triggering mechani sms of cyst formation at the micron scale

PAGE 154

154 U tilization of cpk M ice H uman cystic kidney disease has gained many benefits from comparative biological research on animal models. Through comparison of biochemical, anatomical and functional characterization between human s and various murine models, the understanding and identification of pathogenic information has been elucidated with respect to PKD development and therapeutic treatments This includes knowledge of epithelial cell proliferation and differenti ation [202, 203, 204] cystic fluid accumulation [205, 206] and extracellular matrix composition [207, 208, 209, 210] Out of several mutations induced by genetic manipulation with experimental mouse models, the congenital polycystic kidney ( cpk ) mouse mo del is well characterized with respect to molecular determinant and experimental models of murine PKD [112] Even if the renal phenotype is presented in B6 cpk/cpk mice, only definite crosses permit the bil i ary phenotype to be expressed. To examine the biliary phenotype the B6 cpk strain was c rossed with the BA LB/c strain [113] In th e current study, using MR microscopy, we try to delineate and understand the characteristic morphological changes in the isolated renal systems of the cpk mouse model of ARPKD Additionally the data will be validated with the histochemi cal analysis and quantitative volumetrics of kidneys and liver s quantified in both normal and PKD mutant mice. C ontributions of MRI A variety of imaging modalities such as ultrasonogrphy, CT and MRI are applicable in clinical environments to monitor the progression of the ARPKD and the response to targeted therapy at early disease stage s There are, of course, benefits and limitations to all imaging modalities. F or example, ultrasonogr a phy has a relative ly low cost but poor accuracy and reproducibility Likewise CT has relatively higher resolution

PAGE 155

155 to measure the volume of cysts in vivo but poses potential toxicity effects from using exogenous ionizing radiation and contrast medium Exposure to these agents can result in fatal damage to patients with en d stage renal disease [192, 193] Compar ed to them, MRI can offer significant advantages, such as non invasiveness, accuracy, reproducibility, and the capacity of quantitative analysis to measure volume, structure and function al change of the renal system MRM can be used to track disease progression at the expense of long acquisition time. However, when employing such contrast medi a as gadolinium, extra attention must be paid to the potential damage to a patient in end stage kidney disease. The best strat egy to prevent any potential renal complications caused by the toxic effects above is employment of endogenous contrast rather than getting aided by any exogenous media. Furthermore, anatomical and pathophysiological changes from the early stage to the pre and post treatment stage in a non invasive manner would be more ideal for the patients who are already suffering from the disease itself In that regard, MRM with the highest possible resolution and sensitivity may open up the pathway to meet the need s of the aforementioned demands from the tissue levels out of isolated renal or hepatic organs to set up the landmark and reference to treat the human patient in the long run This deep tissue imaging using MRM can ful fill the need to visualiz e internal struct ures without concern about invasiveness from the preparation required of histology protocols, such as destruction of the cellular membrane due to dehydration MRM also does not suffer from and light scattering, absorption, and tissue auto fluorescence typi cally encountered in light microscopic techniques. Despite sacrificing some resolution on the order of 5~10 m, which is still lower than confocal or multi photon microscopy, MR microscopy has the

PAGE 156

156 capacity to quantify cystic development in the mouse models of ARPKD by using endogenous contrast inherent to the delicate soft tissues of the kidneys and liver. At the same time, to improve hands on applicability, imaging time could be saved while still keeping the efficiency and stability of the imaging protoco l with well orchestrated resolution, sensitivity and contrast. S ignificance of V olumetrics The significance of measuring the alteration of entire kidneys and liver comes from research on disease progression that can be monitored by dilation of collecti ng ducts [ 150] and the resulting growth of cysts due to the disruption of the renal cortex and medulla Additionally, samples reflect the response to treatment which is less effective at the late r stage s of the disease F urther more estimation of individu al or collective cyst volume is important to identify the early stages of this genetic disease which is important for ensuring a higher chance of therapeutic efficacy. However, there are challenges in monitoring the minute destruction of renal parenchyma a nd the variation in composition of multiple cysts that reflect different signal physiognomies. The prognostic value of volume measurements is reinforced by the fact that renal cysts are create d in utero and progress throughout the lifetime. Theref ore, in addition to the entire volume of renal systems, other renal structural parameters, such as position, size and number of cysts can also be good imaging biomarkers, closely related to other clinical manifestations of complications such as hypertens ion, pain and hematuria. Additionally, in relation to malfunction ing renal artery blood flow, several structural changes in other organs are induced such as pancreatic cysts and left ventricular hypertrophy : the latter of which is potentially correlated to chronic induced hypertension. In the current study, volumetric analysis on kidneys using voxel count

PAGE 157

157 based MRI segments displays its efficiency and capability to measure the progression of ARPKD These results are validated with conventional water displ acement measurements [114] Likewise, quantitative and 3D MR imaging with high precision contribute d to reduction in overall sample size and protocol refinement sav ed additional animal sacrifice and numerous post processing immune chemical staining works Despite many promising clinical trials, there are no proven therapeutic treatments available for slowing the dise ase progression [115] However, recent advance s i n technology and treatment give us in depth insight on the genetic causes prognosis and therapeutic interventions from the beginning stages of kidney disease to renal failure Methods Wild Type and Mutant Mouse Liver and Kidney All mice employed in th e se experiment s were maintained and administered by collaborators at the University of Alabama at Birmingham. They were bred from stock colonies of C57BL/6J +/cpk (b6 +/cpk) mice and Balb wild type (Balb WT) mice. F1 progeny heterozygous for the Cys1 cpk all ele were identified using standard PCR based genotyping techniques and intercrossed to generate F2 wild type and mutant mice. All F2 wild type and mutant mice were genotyped using standard PCR based genotyping techniques and sacrificed at 20 22 days of age For each mouse, b oth kidneys and the liver were fixed in 10% buffered formalin for two days then they were immersed in p hosphate b uffered s aline (PBS) (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 and 1.8 mM KH 2 PO 4 : pH 7.4) overnight at ambient temperature ( 22 0.5 C) to rinse any fixative from the sample This is performed in order to avoid severe chemical shift artifacts and T 2 shortening effects. PBS fill ed the space inside the biliary tree of the liver

PAGE 158

158 to resulting in better contrast in T 2 weighted imag es due to their longer T 2 relaxation times. Depending upon imaging applications, typically either PBS or FC 43 is employed to suspend samples in for imaging applications S pecifically, for this study, both liquid s were employed because they offer good con trast for volumetric imaging of whole sample s. W hen it comes to imaging the biliary trees in the liver, oil based FC 43 is very useful to eliminat e any magnetic susceptibility effects, but may have difficulties penetrating into the fine st branches of liver s In order to obtain sectioned sampl es for the surface coil imaging which offer s the highest sensitivity in proximity to the coil surface kidney and liver slices ranging from 25 m to 250 m were cut with a Lancer Vibratome (Ted Pella, series 1000, 888 5 054018). Volume Measurements by Water Displacement, MRM and C orrelative Histopathology of ex vivo L iver s and K idneys Liver and kidney volume was assessed by counting the number of voxels segmented from the each soft tissue region. For example, the voxe ls for the liver and bile ducts were counted and for both wild type (n=6) and mutant liver MR datasets (n = 6). Voxel counts of the entire kidney in wild type (n = 8) and mutant MR datasets (n = 12), were compiled and comparative analysis was carried out e x vivo [ 114 ] using water displacement to assess total organ volume The enlarged diseased kidneys presented with dilation of the collecting ducts filled with benign cysts It show ed good agreement in the resulting values between the two methodologies. The comparative data is addressed in the discussion section.

PAGE 159

159 All MRM work was performed using an Oxford 14.1 T and Bruker 17.4 T, which were equipped with 3000 mT/m imaging gradient coils Two wild type kidneys were easily accommodated in the 10 mm NMR tube s Each was wrapped in a Kimwipe (Kimberly Clark, Roswell, GA) to prevent movement during scanning. These scans were conducted in a 14.1 T magnet. However, due to the enlarge d features of diseased kidneys and the resulting fragility of the basement membran e, extra care was taken when positioning the mutant s into the 15mm NMR tubes Scanning of the diseased kidneys was conducted in a 17.4T magnet. For liver imaging, both liver sample s were tailored to fit into the 10 mm NMR tube To achieve this, the biggest lobe was selected and removed by gross dissection. Here again, due to the delicate composition of the liver tissue, extra care was taken during positioning. For deep tissue imaging of the liver and kidneys, samples were prepared as described above and th en carefully positioned in the tissue well of the surface coils. I mages were acquired with a 2D m ulti s lice m ulti e cho ( MSME ) imaging sequence. This sequence was chosen because of the high level of contrast generated between tubular and ductal structures This sequence also has the advantage of avoiding any possible magnetic susceptibility artifacts as seen in gradient echo images. After collecting pilot images to check for proper tissue positioning and RF power 2D MSME images in the coronal and sagittal planes were acquired : TE = 50 ms, TR = 2.5 s, number of averages = 10, FOV = 2 mm x 1 mm, sampling matrix (MTX) = 256 x 128, slice thickness = 200 m, in plane resolution = 78 m. I t is generally better to employ the longest axis of the sample in the read out direction as this can save significant amounts of time.

PAGE 160

160 For correlation with histology, after being embedded in paraffin all slides of kidneys and livers were micro sectioned at 5 m. All histological work, such as hematoxylin and eosin staining, an d statistical analysis was carried out at the University of Alabama, Birmingham All the protocols were approved by the Animal Care and Use Committee at UAB, which is fully accredited by the American Association of Accreditation of Laboratory Animal Care ( protocol no. 3387) Micro Surface coil Imaging for Micro Structural Information and C orrelation with H istopathology of ex vivo T issues of Both Types of L iver s and K idneys N ephrons in the kidneys and the lobules in the liver constitute the functional unit s of these organs For deep tissue imaging with micro surface geometry of RF coils, fixed kidneys were washed in PBS overnight to rinse out as much formaldehyde as possible prior to being b issected lengthwise. The cut face s were glued on a specimen block (Ted Pella, series 10076) using Loctite 404 tissue adhesive (Ted Pella, series 10035). The tissue on the block w as placed in a vibratome (Ted Pella, Lancer series 1000, 8885 05418), then bathed in PBS ( pH 7.4 300mOsm). A bag of ice was used to reduce the PBS bath temperature which improved the ease of sectioning. This extra cooling step reduces mechanical compression and tears during sectioning. Once multiple sample sections of 100 to 250 m were collected, a representative slice possessing the anatomica l features under investigation was selected and cut to fit into the coil sample well (0.5 cm diameter, 500 m depth) with the aid of a dissecting microscope (OPMI 1 FC, Carl Zeiss). T o make histological validation easier, endogenous landmarks such as cells or connective tissues and surgically crafted markers were sometimes required. As described in Chapter 2, the well was capped with PCR film During sealing, care was taken to avoid capturing any micro bubble s inside

PAGE 161

161 the tissue well MR imaging was performed with a T 2 weighted spin echo sequence with variable TEs (12.5, 20, 30, 40, 50, 60 ms), TR (2000 ms), in plane resolution (12 m), slice thickness (200 m), matrix size (128 x 128), FOV (1.5 x 1.5 mm 2 ) and a series of diffusion weighted spin echo sequences with b values (0, 500, 1000, 2000 [s/mm 2 ]), TEs (30 ms), TR (2000 ms), resolution (12 m), matrix size ( 128 x 128), FOV (1.5 x 1.5 mm). For diffusion tensor tractography, DTI tools software package (DTI Tools, Freiburg University Hospital, Germany) which utilize s the Fiber Assignment by Continuous Tracking ( FACT ) algorithm [160] The tracking was visualized across the entire FOV by diagonalized diffusion tensor with Matlab (The MathWorks Inc.). The direct co registration between MR DTI imaging data with the histological images stained with Nissl (cresyl violet) was achieved. The renal hilum, concave indentat ion of the kidney, was recognizable landmarks in both imaging datasets. Results MRM Analysis of Wild Type and Mutant Mouse Kidneys In Fig ure 5 1, t he structural inhomogeneity between medulla and cortex is visualized, and the hypointens e signal around the hilum the recessed central fissure corresponding to structural fat is also seen. The K imwipe used to secure the samples had a hypo intense signal. In Fig ure 5 2, n ormal kidneys are displayed. There are clear demarcation s between medulla and cortex regions These images were collected using FLASH ( F ast L ow A ngle S hot ) m agnetic r esonance i maging : a gradient echo based pulse sequence that rapid acquisition as well as a spin echo pulse sequence that is le ss sensitive to field inhomogene it ies ( Fig ure 5 3). I nho mogeneous signal results from the

PAGE 162

162 presence of medulla and cortex By increasing T E T2 contrast enhancement revealed the inner structures of the kidneys. I n Figs. 5 4 and 5 5, the characteristics of pkd kidneys are visualized C yst filled kidneys presenti ng with fluid filled dilated collecting ducts are visible The hyper intense signal in the images corresponds to kidney parenchyma that has been changed into deformed and tortuous scar tissue that contains numerous fibers. In diffusion weighted images, un like the free water in PBS surrounding the sample and cystic fluid inside the parenchyma, the water trapped in connective tissues generates a hyper intense signal These results illustrate the capabilit y of the MRI to monitor the disease at an early develo pmental stage (here 3 month old mouse) Up to b value s approaching to 3,000 s/mm 2 water trapped and restricted in the scar filled connective tissues displays hypo intense signal. V olume differences between the two types of the kidneys (0.5 cm vs. 1.5 cm) were quantified by visualizing wild type (n=8) and mutant (m=12) of kidneys in 3D visualization imaging softwar e, Amira version 3.1.1 (Figure 5 6). Then, the volume was calculated from the total number of pixels (Fig ure 5 6). The average volume for mutant kidneys (n=12) is about 14 times bigger than the wild type kidney (n=8) ( 1307.61181.37 mm 3 versus 92.9910.05 mm 3 Table 5 1 and 2 ). MRM Analysis and 3 D Segmentation of the Biliary Architecture in Wild Type and Mutant Mouse Livers The differences in v olume are reported in T able 5 1. L iver s from diseased and healthy mice were also analyzed quantitatively and qualitatively by MRI using AMIRA software version 3.1.1. The example data sets from both wild type (a) and PKD mutant (b) mouse livers. The segment ation works on the livers were analyzed to isolate the biliary tree. The total wild type liver volumes were larger than the mutant one by a factor

PAGE 163

163 of 1.8 (664.5161.61 mm 3 versus 374.46152.86 mm 3 ) while the mutant livers manifest the higher variability in volume. The biliary tree volumes themselves present this tendency (wild type =111.7510.85 mm 3 mutant = 44.7728.66 mm 3 ) These values were determined to be significantly different (p<0.05) The difference in volume fraction is well visualized in the ima ges in a fashion that the mutant livers manifest less dense arborization and truncation indicative of developmental abnormalities in the mutant animals. The ductal volume of mutant livers was 1.5 fold less than that of wild type livers (Table 5 1 & 2). The H&E stained histological images of (c) wild type and (d) mutant liver depict that the architecture of the portal vein (PV), bile duct (BD), and artery (A). Especially, due to the mutation the patho morphological change, e.g., enlargement of the bile duct (BD), is well presented (Fig ure 5 7). MRM Analysis of Fixed Kidneys U sing Micro Surface Coils T he medulla which has radiating pyramid like structures contains the cylindrical tubes of the nephron They are distinguishable in the images by the difference in the diameter of their cross sections ( Fig ure 5 8 ). Interestingly, with T 2 weighted images ( TE = 12.5 ms (Fig ure 5 8 A ) 16 ms (Fig ure 5 8 B ) the spin spin relaxation prop erties of the parenchyma connective tissue and hollow regions inside the kidneys a re revealed ; however, with diffusion weighted imaging the signal from areas of the tissue containing free water is suppressed ( Fig ure 5 8 C ). T ubular cells in the kidneys, are quite permeable to water due to their role in filtration, absorption and secretio n processes These features result in reduced hindrance/restriction effects unlike the elevated hindrance/restriction effects seen in the connective tissues. In Figs. 5 9 A and 5 10 A even with in plane resolution of 39 m, the gross anatomical features su ch as inner medulla, outer medulla, and cortex are highly

PAGE 164

164 resolved a nd show high structural heterogeneity. In the striat ed medulla, proximal tubules and the ascending loops of Henle are seen intermingled and interlaced with the renal vascular system Fig ur e 5 9 shows the diffusion tensor data set with in plane components of the calculated tractography. The overlaid in plane component of the primary diffusion eigenvectors correlates spatially with the radial striation of medullary rays in the histological im age. Fig ure 5 10 illustrates a slice the mouse kidney showing both regions of medulla and cortex. The DTI data set shows that the striation of medulla rays emanating from the outer medulla spread outas they approach the cortex. The relatively complex archi tecture in the cortex reflects the structural detail of the nephrons This tissue composed of glomerular bundles distal tubules and vasculature appears blurry due to volume averag ing in the MR data Discussion Volume Measurements by Water Displacement, MRM and C orrelative Histopathology of Fixed L iver s and K idneys Before expanding to in vivo and functional imaging of the kidneys, th is preliminary work with fixed samples show s the potential to us e dedicated small animal MRI scanners t o perform research to identify translation al opportunities to humans In the current study, non invasive volume measurements of normal and diseased livers were collected using MR imaging analysis Only t he largest lobe of both type s of livers, mutant and w ild type w as used so that they could be placed in side a 10mm NMR tube Samples were scanned in a 10 mm RF coil with high filling factor Interestingly, the mutant livers had lower total volume values than the wild type s. It was interesting considering th e opposite result in clinical data for human patients [186, 187] This feature develops with age probably in large since the organs of both types had been

PAGE 165

165 isolated from the in bred animals at three months of age before the second fibrosis Even though many factors affect the change of volume in the biliary system of the diseased liver characterized by portal hypertension and hepatic fibrosis, second ary hepatic fibrosis is presumed to be the primary cause of the increased volume. H ypertension and hepatic fib rosis leads to pruning of the biliary trees resulting in reductions in the volume of liver parenchyma. Within the mutant liver population, variability in the volumes of the mutant livers was observed. Such differences were odd given the presum ed uniformit y of in bred animals ; however, these differences were not as disparate when the ratio between the liver volume and duct volume was considered E ndogenous tissue contrast generated with samples immersed in either PBS or FC 43 was used in the current study. More specifically, in th e current study, both liquid s offer good contrast in scans needed for volumetric analysis of whole sample s; however, when it c ame to imaging the biliary trees PBS served our purposes better than the oil based FC 43 Fluorinert, whi ch is very useful in eliminating magnetic susceptibility effects also presents difficulties penetrating into the fine st branches of of livers. As a result of the disease ARPKD, process, prominent morphological alterations in the liver lead to volume mod ifications P runed intrahepatic biliary tree arising from dilation and enlargement of the biliary ducts due to cyst formation originating from cystic epithelial cells was observed V olumetric data analysis using histological techniques were deemed to be to o time intensive and suffered from issues such as volume changes due to dehydration during sample preparation. Comparatively, quantitative analysis of volume using 3D MR imaging analysis w as carried out in a reasonable time F ollowing the volumetric imagin g analysis, noteworthy features were found. First, the

PAGE 166

166 mutant liver is approximately 1.7 times greater than the wild type, which mirrors the conditions observed in ARPKD patients who usually have bigger livers due to fibrosis and pruned biliary trees as an inflammatory response to the disease. Second, the ductal volume of the mutant livers is highly variable mainly due to abnormality and irregular truncation of the biliary tree and hepatic ducts in the diseased condition (Table 5 1 and 5 2) Third, there wa s a slight deviation from what had been expected in both types of livers with respect to the ductal and total volume s T he water signal, the main source of the MR signal comes from water confine d in tissue. These areas include the portal vein, biliary duc ts, and sinusoids around the hepatocytes (Fig ure 5 7) The prominent morphological alterations of the disease process in the liver include truncation of the intrahepatic biliary tree and enlargement of the biliary ducts (area in 2D histological image and v olume in 3D reconstructed image) due to cyst formation arising from cystic epithelial cells. Magnetic Resonance Microscopy of F ixed K idneys From T 2 weighted and diffusion weighted images generated with a 500 m coil, to our knowledge for the first time us ing MRM the tubular structures in the medulla strips were delineated. T he nephron is revealed to be made up of tubular cells including the glomerulus collecting ducts. T he pyramidal str uctures of the medulla that radiat e from the inner to the outer kidney appear different ly in images with different slice orientations Depending on the direction of the slice images display different cross sectional segments of tubular structures starting from the thinner descending and thicker ascending loop of Henle, and progressing to the thicker proximal tubule loop of Henle and finally to the largest diameter collecting duct (Fig ure 5 8)

PAGE 167

167 The anatomical boundary between the renal medulla and renal cortex is marked by vascular structures ( actuate vessels ) convoluted tubules and renal corpuscle, which are all located in the cortex. The rays in the medulla are highly orientated, aligned in parallel, and interdigitized with epithelial cells specializ ed in absorbing water from the filtrate and regulati ng the concentration of urin e The micro architectural complexity is well visualized in MR images and reconstructed from segmentation data This well defined striation of outer medulla has attracted the a pplicability of the relatively recent diffusion related technique to trace any anisotropic water movements in the fiber bundles to visualize fiber directions in the kidneys [116, 117] The liquid bath employed in the majority of experiments was PBS but FC 43 was also be used to obtain image signal exclu sively from the tissues Using FC 43 with limited volume of samples especially in ultra high resolution MR imaging work presented difficulties regarding detection of the restricted water signal in tissues. C onclusions and Future Work I n this study, MRM w as utilized to reveal characteristics of disease progression in isolated fixed kidneys and liver of pkd mice Using MRM, this was achieved in a non invasive way in hope s of mirroring the desired imaging conditions conducted in human patient s. Compar ed wit h other imaging modalities MR M has the advantage s of non invasiveness, accuracy, reproducibility, and capacity of quantitative analysis to measure the structur al and function al changes of the renal system Such advantages, in the case of MRM, come at the expense of long acquisition time s MRM presented its capacity to visualiz e the internal structures of the renal and hepatic systems in normal and mutant mice while avoiding the effects of structural deformations from histological preparation protocols Add itionally, this deep tissue imaging by MRM offered contrast arising from

PAGE 168

168 the inherent magnetic properties of the tissues. Unlike contrast generated with light microscopic techniques including the degree of light scattering, absorption, and tissue auto flu orescence requiring exogenous agents Diffusion Tensor Tractography of N ephron S tructures As discuss ed above ARPKD is a genetic disease characterized by cyst formation in kidney collecting ducts. Compared to other segments of the kidneys such as the glom erulus, renal corpuscle, proximal tubules, distal tubules, and the loop of Henle the diameter of the collecting ducts is greatest in a normal healthy kidney. In the process of monitoring their enlargement due to cyst formation, high resolution MRM employ ing endogenous tissue contrast Thanks to its non invasivene ss, MRM can fill the needs of structur al monitoring of nephron s in kidney s to monitor for pathologies. Conventionally, biopsy specimens are usually taken for histological examination at the expens e of the invasiveness required for tissue procurement. Processing requires destruction of the epithelial membrane and lumen in the process of dehydration and slicing of the samples, which already manifest membrane abnormalit ies arising from cyst and scar t issue formation These abnormal tissue properties result from dysfunctional regulation of hydrostatic pressure between tissue colloidal osmotic pressure and blood plasma osmotic pressure. However, MRM can overcome limitation s of histological processes and further more, it can act as a reference point to future measurement of patients organs in the clinical environment. Volume C omparison between N ormal and M utant K idneys and L iver at V arious S tages of Disease Onset and Progress In this study, the reason fo r our findings that include a smaller mutant liver volume compared to wild type liver s was investigat ed Considering the age of the animals used

PAGE 169

169 3 month s old, this very young mouse model had not undergone the secondary process of fibrosis, which might hav e elevated the kidney volume significantly By widening the temporal window of monitoring pathological changes from pren atal, 3 month old, 6 month old, and beyond 6 months the threshold of chronological volume changes can be investigated in a reasonable p rocessing time. Observation of C ollecting D uct D ilation ARPKD is a single autosomal recessive disease complex with a recessive trait w hich means that once the cell lo se s its normal osmolality with normal checks and balance s it swells, divides (hypertrop hy) and starts to proliferate (hyperplasia) [211] As a consequence, the biochemical interac tions between intracellular orga nelles and the basement membrane and the malfunction of a single step in intracellular and intercellular communications giv e rise to cyst formation resulting in renal cystic diseases which is connected to the congenital hepatic fibrosis (CHF),and biliary tract truncation [212, 213 228, 229 ] Due to the impaired function between epithelial cells on the basement membrane, the water re absorption process is affected. Specifically, because of the poorly regulated membrane due to the accumulation of epithelial cells followed by endothelial retraction and enhanced capillary permeability, the leak out plasma protein pulls the water in resul ting in the increased size of collecting ducts. By monitoring abnormal water transport through collecting duct channels occurring in the abnormal nephron s, especially using DWI an MR imaging scheme which is highly sensitive to water diffusion in the micros tructure of the tissue makes it possible to track the tissue architecture using endogenous tissue water

PAGE 170

170 Figure 5 1.2D MSME MR M of normal kidneys in axial direction. The kidneys were wrapped with Kimwipe (black line). MRM scan parameters: TR = 3 s, TE = 60 ms, t hickness = 500 m, in plane r esolution = 94 m, a verage = 4, AT = 26 min Figure 5 2. 3 D FLASH images of a normal kidney MR scan parameters: TR/TE = 150 ms / A) 15 ms B ) 30 ms C ) 45 ms D ) 60 ms, In plane resolu tion = 39 m thickness = 200 m AT = 17 min.

PAGE 171

171 Figure 5 3. 2 D MSME MRM of a normal kidney in coronal direction Bright striated pattern indicates the medulla and hypo intense signal area corresponds to cortex. MR scan param eters: TR/TE = 2 s/ A ) 15 ms B ) 30 ms C ) 45 ms D ) 60 ms, resolution = 39 m thickness = 200 m AT = 17 min.

PAGE 172

172 Figure 5 4. 2D diffusion weighted MRM of a mutant kidney in axial direction. A ) Representative MSME images o f a cyst filled mutant kidney. MR scan parameters: TR = 2 s, TE = 100 ms NEX = 2, AT = 33 min; B ) E ) diffusion weighted images of mutant kidney in a 15 mm tube collected with a 20 mm RF saddle coil MR scan parameters: b values = A ) 0 [s/mm 2 ] B ) 500 [s/m m 2 ] C ) 1000 [s/mm 2 ] D ) 3000 [s/mm 2 ], NEX = 4, AT = 2 hrs 16 min.

PAGE 173

173 Figure 5 5. 3D GE MRM of a mutant kidney. A) Representative 3D gradient echo MRM of a mutant mouse kidney with large fluid filled cysts in FC 43 ; B ) through E ) diffusion weighted images in a 15 and 20 mm saddle coil with b values = B) 0 [s/mm 2 ] C) 500 [s/mm 2 ] D) 1000 [s/mm 2 ] E) 3000 [s/mm 2 ]

PAGE 174

174 Figure 5 6. Representative images of volume visualization using AMIRA. A ) W il d type kidneys. B ) M utant kidney using AMIRA. A) T wo wild type bean shaped mouse kidneys were wrapped with Kimwipes (hypointense dark lines) for MRM B) M utant type of mouse kidney with fluid filled cysts. The segmentation process was enhanced by manually outlining the kidney image, slice by slice in the multi slice data sets for each sample using the software. Images are not shown to scale.

PAGE 175

175 Figure 5 7. Volume v isualization of both types of mouse liver using AMIRA and his tology images. A) and B) The 3D volume photo illustrates the overall structure of bile ducts within the slices of 2D MR liver images. The solid yellow line corresponds to the bile ducts and the transparent yellow depicts the largest lobe of the liver. The H&E stained histological images of C) wild type and D) mutant liver. PV, portal vein; BD, bile duct; A, artery. The images are not shown to scale.

PAGE 176

176 Figure 5 8. Delineation of cellular structures in a kidney slice in MRM A) and B) T 2 weighted spin echo images show the functional units of the kidney, for example, distal tubules, proximal tubules, and the collecting ducts in order by size of diameters; C ) diffusion weighted spin echo image, MR scan parameters: Multi Spin Mul ti Echo pulse sequence with TE/TR = A) 1 2.5 ms B) 60 ms C) 2D diffusion weighted imaging with b = 2000 [s/mm 2 ], TE/TR = 30/2000 ms, resolution = 12 x 12 x 200 um 3

PAGE 177

177 Figure 5 9. Direct correlation of medullar region of the k idney in MRM and histology images. A ) Representative MR image of the striated medullar region in a normal kidney with a resolution of 39 m using a 5mm volume coil B)T he same slice with a micro surface coil, C) Nissl stained slice D ) 21 direction of DTI reconstruction using streamlines based on the FACT algorithm delineates the medullary tracts in the inner medulla area of a dorsal section of the mouse kidney corresponding to the circle on the same region of interest in C ).

PAGE 178

178 Figure 5 10. Direct correlation of medullar cortex region of the kidney in MRM and histology images. A ) Representative MR image of the transitory region from the medullar to cortex region in a normal kidney using a 5 mm volume coil, B ) one with a micro su rface coil, C ) Nissl stained slice, and D ) 21 direction of DTI reconstruction using streamlines based on the FACT algorithm depicts the medullary tracts in the outer medulla area of a dorsal section of the mouse kidney corresponding to the circle on the sa me region of interest in C ).

PAGE 179

179 Table 5 1. Volumetric analysis of wild type kidney volumes assessed with MR imaging data and segmented with AMIRA. The second column presents the total volume of the kidneys (n = 8). WILD TYPE KIDNEY TOTAL VOL [mm 3 ] MEAN 9 2.99 STD 10.05 SEM 1.26 Table 5 2. Volumetric analysis of mutant kidney volumes assessed with MR imaging data and segmented with AMIRA. The second column presents the total volume of the kidneys (n = 12). MUTANT KIDNEY TOTAL VOL [mm 3 ] MEAN 1307.61 STD 181.37 SEM 15.11 Table 5 3. Volumetric analysis of wild type liver assessed with MR imaging data and segmented with AMIRA. The first column of the table indicates the mean and standard deviation of total volume. The second column presents the volum e of the bile ducts. The third column describes the ratio of the ductal volume to the total volume (n = 6). WILD TYPE LIVER TOTAL VOL [mm 3 ] DUCT VOL [mm 3 ] RATIO [DV/TV] MEAN 519.1628 87.30204 0.168159 STD 71.57074 8.479506 0.118478 SEM 0.048368 Ta ble 5 4. Volumetric analysis of mutant liver assessed with MR imaging data and segmented with AMIRA. The first column of the table indicates the mean and standard deviation of total volume. The second column presents the volume of the bile ducts. The third column describes the ratio of the ductal volume to the total volume (n = 6). MUTANT TYPE LIVER TOTAL VOL [mm 3 ] DUCT VOL [mm 3 ] RATIO [DV/TV] MEAN 292.5488 34.97538 0.119554 STD 119.4265 22.38971 0.187477 SEM 0.076537

PAGE 180

180 CHAPTER 6 MR MICROSCOPY OF SKE LETAL MUSCLE CELLS IN HEALTHY AND DYSTR OPHIC DIAPHRAGM DISEASE MO DEL MICE Introduction Another application of MRM lies in delineation of individual muscle fiber s. Current MR microscopy methods should permit the visualization of single skeletal muscle cell s The diaphragm, a skeletal muscle plays a larger role in muscle related disease due to its ability to act as a biomarker of numerous diseases in preclinical trials and therapeutic interventions [118, 121, 123, 125, 126] In the current study, special att ention was paid to elucidate how observations of the diaphragm using MRM can contribute to the treatment of muscle pathology especially in relation to DMD (Duchenne Muscle Dystrophy) and discuss how emerging MRM technolog ies can demarcate the onset and monitor the temporal progression of disease Hopefully, these preliminary data will contribute to understanding the cellular changes associated with DMD and help lead to therapeutic treatment s targeted for th is and other disease dis ease can be detected soon after alterations to cell structure take place i.e. very early in the progression damage could be prevented before severe diaphragm dysfunction occurs First, given the significance of the diaphragm for the health of an organism we will briefly review it in terms of its anatomy and physiology We will also briefly discuss characteristics of the normal alongside the diseased diaphragm. S tructure of Skeletal Muscle in the Diaphragm One of three recognized types of musculature i.e. skeletal muscle, smooth muscle, and heart muscle, the skeletal muscle is a highly organized bundle arrangement of anisotropic muscle fibers. It shares many characteristics with skeletal muscle and governs respiration It physically separates the thoracic a nd abdominal cavity. It is

PAGE 181

181 dome shaped in the center and flat at the outer rim [ 118] F or the purpose of regulating breathing, the diaphragm creates pressure gradients by becoming shortened and flat This conformational change expand s the lower rib cage inflating both lungs. Since the diaphragm is a skeletal muscle, its components sha re anatomical features with limb muscles : both being continuous and flat. I ts muscles are separated into three different parts: crural, sterna l and left and right hem i diaphragms [215, 216] As shown in Fig ure 6 1 its structural organization is characteri zed by the convergence of skeletal muscle into the central tendon, which is the innermost portion of the diaphragm that holds tight all of the interdigitized myofibers. Like all ot her limb muscles, the mobility of the diaphragm is controlled by motor neur ons then the sensory fibers transmit the nerve pulses to the peripheral parts of the muscle from the lower sixor seven intercostal nerves [119, 215, 216] D uchenne Muscular Dystrophy, a Lethal Muscle Wasting Disease Duchen n e muscular dystrophy (DMD) is a n X chromosome linked muscle eating or muscle wasting disease which affects the diaphragm It occurs in every one out of 3000 3500 male s More specifically, the disease results from a mutation in a gene encoding the protein dystrophin [120, 121] At approximately the age of three, a sudden onset of weakness dystrophin in the skeletal muscle linking leads to weakness of the sarcolemma (Fig ure 6 2) This causes the sarcolemma to become leaky, a nd permits Ca 2+ to travel in and out of the membrane. In addition to muscle weakness, overload of Ca 2+ in the sarcoplasmic reticulum or muscle fiber cytoplasm induces necrosis [122 125] O ngoing muscle fiber necrosis from inflammation whi ch is essential to healing infiltration of T lymphocytes and replacement of muscle tissue with fibrous connective tissue accelerates the disease

PAGE 182

182 progression C linical manifestations due to subsequent biochemical imbalance include walking disabilities, we akness in the respiratory system especially the diaphragm and the cardiac muscle [152] Recently, it has been report ed that tissue destruction due to the weakness of dystrophin is not limited to the skeletal muscle but also affects the function of the cent ral nervous system ( CNS ). These changes manifest in reductions in IQ level dendritic abnormalities and destruction of neurons in the CNS [126] A R at Model of the Disease, X linked M uscular D ystrophy (MDX ) From comparative biology, research on the aforementioned disease DMD, has been carried out by employing mdx mice. These mice are genetically modified from the inbred C57BL strain established in the 1980s [127 129] Even if the detail ed characterization of the temporal progression of th is disease in mice is beyond the scope of this paper, the significance of the temporally dynamic progresses of patholog y can be appreciated and explained. In the presence of non functional dystrophin, even from the post natal day 1, the Z lines get disorga nized in a form of spreading or scattering out focal streaming of Z lines. At approximately post natal day 5, sparse necrotic muscle fibers have formed. F inally, at approximately post natal day 21, the signs of necrosis in many muscles including hind lim b are present [130] Consequently at post natal day 21, myonecrosis can be investigated [127] Based on the aforem entioned observations investigation into the dystropathology of mdx mice has gained popularity as a tool for testing therapeutic interventions of DMD disease [131 133] Genetically modified (GM) mdx mice like the ir human counterparts show impairments in short term memor y formation and passive avoidance reflexes [126] In spite of known variability due to age and sex [134] the two most popular muscle groups employed in research on muscle patholog y are the limb muscles and diaphragm.

PAGE 183

183 As discussed above the mdx mouse model is known for having same patholog ical disease characteristics as humans suffering from DMD. In the current study, we have chosen to focus our analysis on the diaphragm rather tha n limb muscles. Unlike the limb muscle the diaphragm develops more severe pathological manifestations and more closely reflect s the anatomical and physiological deterioration of the muscles as seen in` DMD [135] R ecent progress in DMD research has been made possible by the development of the diaphragm strip preparation [136] Thus, for observation of structural damage s resulting from DMD disease progression, the diaphragm is the clear choice over skeletal muscle from the limbs I maging of a Single Muscle Cell Given the need for an excellent biomarker to detect anatomical, physiological, and pathological changes associated with DMD the basic criteria should require the ability t o assess conditions of the diaphragm accurately at the cellular level As discussed above, the decrease in dystrophin levels leads to the formation of dystrophic myofibers. I mpairments of muscle sarcolemma which occur as a result of exercise resulting in r esealing or necrosis have been investigated well [137, 138] ; however, the mechanisms starting from the initial sarcolemmal damage subsequent combinations with inflammatory cells and cytokines and the resultant myofiber necrosis associated with DMD ha ve not been examined wel l. There are two approaches when examining the impairment of the muscle : either a macroscopic or microscopic analysis Macroscopic monitoring of overall muscle damage has been utilized [139, 140] and microscopic investigation was carried out by employing Evans Blue Dye (EDB) labeling of proteins. Specifically, tracing ED B allow s us to monitor the disruption occurring inside the sarcoplasm of the myofibers, or to measure the distance along the cylindrical or spindle

PAGE 184

184 shaped muscle fibers [139, 141] It also allows us to estimate the degree of damage to myofibers by measuring the blood protein levels. It is MR I that utilize s the characteristics of magnetically to determine tissue characteristics in a non invasi ve way. Unlike other methodology e.g. histology, MRI can offer quantitative and qualitativ e measurements of the whole or parts of the body and even monitor the temporal changes of individual animals in vivo without s acrific ing them, invasive disruption of tissue membrane s, or dehydration process. Alternatively, when used in conjunction with the diaphragm strip preparation, endogenous contrast based on the density or movement of the magnetically labeled spins inside the deep tissue can be employed to invest igate the degree of pathological destruction in vitro P otential Contributions of MRM into the Monitoring and Treatment of DMD MR M is distinguishable from MRI only because they possess resolution below 100 m in two of three spatial dimensions C onvention al MRI could potentially be a useful tool for compar ing normal and dystrophic muscles in mdx m ice in vivo By applying engineered contrast agents such as albumin targeted contrast agent to damaged cell s the degree of cellular damage can be successfully mea sure d [142] Furthermore, without the help of th e s e toxic contrast agents, e.g. gadolinium, is possible to attain sufficient e ndogenous contrast to track disease progression [153] However, in spite of its non invasive ness MRI is often limited in terms of resolution and sensitivity. However, with the advent of MRM employing state of the art hardware such as high field magnets an d strong/fast switching gradient coils, it has d emonstrate d its ability to attain image resolution at the micron scale Through these technical advancements, MRM is able to monitor disease progression in individual mice estimate the resources needed for

PAGE 185

185 p reclinical trials, and gauge the responses to the therapeutic treatments all without sacrificing research subjects or fundamentally altering tissue morphology through histological preparation. L imitations of Current Treatment Options for DMD and Potent ial Contributions of MRM to Therapeutic Treatments So far, there is no cure for DMD, but there are t wo treatment options to ameliorate the disease. One is to employ assistive devices such as braces and wheelchairs for walking difficulties, and ventilators for breathing and spinal fusion surgery for enhancement of quality of life through enhanced and maintained balance and correction of pelvic obliquity [217, 218] The other commonly employed option is drug treatment. F or example, corticosteroids or glucocor ticoids are often used to retard the disease progression, i.e. reducing muscle degeneration and inflammation [143] Such medications can offer relief to patients, but only alleviate the symptoms of DMD rather than cure the origin of the disease. T hese treatment options must be applied during the early stag es of DMD, and have potential long term side effects which might overshadow the potential benefits [144] In addition to aforementioned two treatment options only to decelerate the disease progression the newest and most innovative strategies to cure DMD has focused on the refurbis hment of functional dystrop h in gene and reducing inflammation associated with muscle necrosis for years In the case of gene therapy, the biggest hurdle is recognition of the newly expressed dystrop h in as foreign by the host immune system ; therefore, immun e suppression drugs are required to interrupt the immune response

PAGE 186

186 Methods Sample and Slice Preparation All mdx mice employed in this experiment were maintained and administered by a collaborator, Dr. Glenn Walter, at the University of Florida. All animal procedures were conducted in accordance for the Care and Use of Laboratory Animals and were approved by the University of Florida IACUC. The diaphragm w as washed in PBS(137mM NaCl, 2.7mM KCl, 1 0mM Na 2 HPO 4 and 1.8mM KH 2 PO 4 : pH 7.4) to remove any remaining formaldehyde and sliced into sections In order to obtain the slice s used for surface coil imaging muscle samples were collected from areas adjoining the central tendon and the coastal margin Magnetic Resonance Microscop y A single skeletal muscle cell (called as a myofiber or a muscle fiber) is a cylindrical multinucleated cell made up of inter digitized actin and myosin These units are collectively organized into functional units called sarc omere. The bundling together of these units in skeletal muscle causes the striation pattern which distinguishes this tissue from smooth muscle. For deep tissue imaging with the RF micro coils, excised tissue specimen was required. Healthy and diseased fi xed diaphragms were washed in PBS overnight to remove as much formaldehyde as possible and were cut into rectangular blocks .. The thickness of the sections varied depending upon the application from 100 to 250 m. Once the sample slides were collected, a s ample possessing salient anatomical features was selected and tailored to fit through gross dissection into the 5mm diameter sample well with the aid of a dissecting microscope (OPMI 1 FC, Carl Zeiss). T o make structural validation easier, landmarks such a s cells or connective tissues were

PAGE 187

187 sometimes required. As described in Chapter 2, to avoid leakage and sample dehydration the well was sealed with PCR film (Fig ure 6 3) For the comparative analysis between healthy and dystrophic diaphragm muscles MR ima ges in both parallel and orthogonal to the striations in the diaphragm were acquired. Results MR Microscopy of Healthy and Dystrophic Diaphragm, and Microstructural Information in the Transverse Plane. Macroscopically, the control diaphragm was observed t o be thicker and more brittle than its diseased counterpart. It displayed higher levels of bright brown pigmentation. However, following the destruction of the sarcolemma, the dystrophic diaphragm develops thinner and less compliant muscle fibers. Likewise less pigmentation is observed due to scar tissue formation and the loss of mitochondria which contain ferrochrome: an iron containing enzyme for transforming the covalent bonds of nutrients to ATP bonds. Skeletal muscle in the diaphragm whose muscle cel ls are embedded in connective tissue is shown in Figs. 6 4 and 6 5. There are packed so tightly together that even with high resolution MRM images, it is difficult to resolve individual muscle cells in the healthy animal. However, as the DMD progresses, gr owth of fibrous connective tissue by infiltration of T lymphocytes drive individual muscl e fibers further apart (Figure 6 6). DMD originated with the weakness of dystrophin protein the faulty gene product of the DMD which connects the extracellular matrix (ECM) and intracellular cytoskeleton leading to instability of the sarcolemmal membrane and muscle fiber necrosis. In terms of MR techniques which magnetically tag hydrogen protons inside the water molecules, fenestration of water molecules through the mem brane and water molecules pulled by

PAGE 188

188 proteins yield higher T 2 weighted signal rich in the connective tissue and vintage the movements of water molecules as the contrast (Fig ure 6 5). Due to pathological changes such as the weakness of skeletal muscle and dy strophic changes in the muscle cell membrane, muscle fibers in the diaphragm become interweaved, overlapped and split into branches monitored in T2 weighted and diffusion weighted images (Fig ures 6 6 and 6 7 ). Unlike the healthy tissue which possesses dens ely packed cells in a crowded environment, individual dystrophic muscle fiber was easier to visualize in the transverse plane. Above the b value of 1000 [s/mm 2 ], since the signal from free water has been killed, only the water signal trapped or restricted by T tubules or sarcoplasmic reticulum remains bright. By changing the directions of diffusion sensitizing gradients, here, 6 directions at proper b value of 600 s/mm2, the anisotropic nature of skeletal muscle architecture was examined in both healthy and dystrophic diaphragm (Fig ures 6 8 and 6 9). Using alternate sample orientations, the individual muscle fibers of dystrophic diaphragms were resolved using di ffusion weighted imaging (Fig ure 6 10 ). Here, the high isotropic resolution of MRM images, i.e. 9 .3 m were acquired with 3D gradient echo imaging sequences which are highly sensitive to suscepbility effects (Fig ure 6 11 ). The images taken with the coil perpendicuar to the muscle fibers are illustrated (Fig ure 6 1 2 ). Microscopic anatomy of the skelet al muscles in the diaphram is illustrated (Fig ure 6 12). Even when employing the shortest permissible TE to compensate for the rapid T2 signal loss, the delineation of individual muscle fibers was difficult. Such data demonstrate that the resolution is no t the only factor to consider when designing and refining an optimal pulse sequence.

PAGE 189

189 MR Microscopy of Healthy and Dystrophic Diaphragm in the Coronal Plane. As is shown in Fig ure 6 1 2 nuclei located at the periphery of skeletal muscle cells were visual ized. Interestingly, the structures presumed to be nuclei are located in the tissue regions nuclei would expect to inhabit: they are rich in areas adjoining muscle fibers and scant in the region of connective tissue scar (hyper intense signal in the middl e of the Fig ure 6 12 C and D ). In Fig ure 6 13, in addition to spin spin relaxation contrast mechanism, additional MR contrast parameter to delineating the local water environment was excellently presented. By increasing the degree of diffusion weighting abo ve b = 1000 [s/mm 2 ]) (Fig ure 6 1 3 D through 6 13G ), the signal of free bulk water exhibiting isotropic diffusion was reduced while the anisotropic nature of diffusion in the restricted environment remained hypo intense. In the dystrophic diaphragm, the dy stro pathological conditions changed its anatomical structure. Specifically, as necrosis and edema occur with the infiltration of the inflammatory T lymphocytes, the MR contrast parameters also changed dramatically. The diameter of the muscle fibers become s thinner and torn as they approach the central tendon of the diaphragm in the region 1 in Fig ure 6 14 A and 6 14B However, when the outer rim of the diaphragm was visualized, relatively intact and thicker muscle fibers were observed (Fig ure 6 14 C ) The ar ea of the diaphragm closer to the central tendon demonstrates hyper intense signal in part due to increased water restriction in its connective tissues following inflammation and scar tissue formation (Fig ure 6 15 B and C ). Contrary to the inner rim of the diaphragm, the outer rim contains more muscle fibers. In addition, remnants of thin muscle fibers are embedded in the connective tissues (Fig ure 6 15 D and 6 15E )

PAGE 190

190 The cylindrical structures of the muscle fibers embedded in connective tissue are visualize d (Fig ures 6 16 and 6 17 ). The water signal inside muscle fibers appears hyper intense in 6 direction DTI acquisitions (Fig ure 6 16) Even with pathological modifications, the size and number of muscle fibers vary region by region throughout the diaphragm. Discussion MRM Imaging of the H ealthy and D iseased D iaphragm Comparative research was carried out on normal and pathological mouse muscular dystrophy ( D M D ), which is a degenerative mutant disease originating from the loss of dystrophin. There is no known cure for DMD. This disease leads to difficulty in breathing and weakening of the heart and diaphragm. Our intent was to investigate the capability of MRM to distingui sh between healthy and pathological tissue of the skeletal muscle system: specifically, the diaphragm (Fig ure 6 1). This structure consists of bundles of myofibers. Each fiber is a single, multinucleated muscle cell. The cell membrane of these fibers is ca lled the sarcolemma. It is within this membranous structure that the pathology originates in muscular dystrophy. Examining the sarcolemma (Fig ure 6 2), we can see that, in a healthy muscle cell, there is a network of proteins that structurally tethers the internal cytoskeleton to the extracellular matrix of collagen. These structures pass through and stabilize the sarcolemmal membrane. If any one of these proteins is absent or mis folded, this complex fails to stabilize the s ripped by shear forces during normal muscle contraction. This loss of membrane integrity leads to necrosis of myofibers and subsequent rounds of damage and repair. Since the new fibers lack the functional

PAGE 191

191 complex as well, this destructive and inflammator y cycle continues. Depending on the protein missing and muscle affected, one of the various forms of muscular dystrophy develops. Using high resolution MRM with dedicated equipment, effects of the disease can be detected at the cellular level. We hope to e xtend this capability to monitor the progression of the disease from the earliest structural manifestations to the responses to treatment interventions. When viewing the healthy diaphragm in the transverse plane (Fig ure s 6 5 and 6 6 ), even with the high in plane resolution (15 m) employed, the thickness of the slice (160 m) resulted in excessive volume averaging. Even if the striation patterns were observed when viewing the tissue in the transverse plane, until viewing the tissue in the coronal plane t hese micro structures were not visible (Fig ure 6 4 and Figure 6 12). From our six direction diffusion tensor images, the complex structure of skeletal muscle system was not readily resolvable. This could potentially be due to the crowded orientation of f iber bundles and a multitude of crossing, merging, and kissing fibers (Fig ures 6 8 and 6 9). However, a s the disease progresses, protein rich scar tissue forms in individual muscle fibers (hypo intense signal). To our knowledge, this was the first time su ch structures were delineated using MR technology. This is a noteworthy accomplishment because this was achieved without employing any exogenous chemical agents. Presumably, due to its rigid structural protein and packed myofibers, the inherent T2 signal was too short to detect using a standard MR imaging pulse sequences. In the coronal view, we were able to visualize the dysfunctional muscle fibers embedded in the loose connective tissue. The nuclei located on the edges of individual muscle fibers

PAGE 192

192 were n ot seen adjoining scar tissu e (hyper intense signal). Fig ure 6 1 7 ability to see into the deep tissue layers of the diaphragm: free of any potentially Li kewise, due to its non invasive nature which allows for monitoring structural changes throughout disease progression, MRM may play an increasingly important role in monitoring the efficacy of therapeutic drug interventions and gene or stem cell therapy. The dynamics and mechanism behind muscle length and force generation could also be explored. By investigating this process using the MR contrast parameters presented above, the earliest structural effects associated with this process might be monitored in a non invasive way. Even if the etiology is not well known, scar tissue formation in DMD occurs as a result of muscle edema and compensatory hypertrophy followed by inflammatory response. During formation of the scar tissue, the diaphragm undergoes struc tural deformation due to myofiber remodeling and collagen deposition. This formation of protein rich connective tissue results in restricted water diffusion from within the fibrous areas. These barriers to the water diffusion make hinder the diffusion exch ange between cells and blood vessels necessary for tissue homeostasis resulting in delayed recovery. Such replacement of functional tissue with scar tissue ultimately leads to the weakening of respiratory system due to degradation of the diaphragm and card iac muscle. Hopefully, this research will contribute to the development of a noninvasive methodology for tracking disease progression without the need for invasive tissue biopsy.

PAGE 193

193 C onclusions In this preliminary study, by investigating the dystrophic diap hragm in a mdx mouse model, the potential capability of MRM to visualize changes associated with a devastating muscle wasting disease which effects humans: DMD. This visualization was achieved without employing exogenous contrast agents of any kind, that i s, images were generated using only endogenous contrast mechanisms. Such MRM images could clearly detect the degree of cellular and tissue damage by simple visual assessment. Still, even if the attainable resolution is much lower than imaging modalities su ch as light microscopy, by virtue of its non invasive nature, MRM is poised to become increasingly important in delineating cellular structures deep inside tissues for diagnostic purposes in the clinic.

PAGE 194

194 Figure 6 1. Schematic of the diaphragm It consist s of the central tendon(C) and surrounding skeletal muscle fibers (M).

PAGE 195

195 Figure 6 2. Schematic of the muscle system This is provided by Dr. Walter professor at the university of florida. The architecture of the skelet al muscle consisting of : A) bundles of muscle fibers or myofibers. Each fiber is a single and multinucleated muscle cell. The cell membrane of these fibers is B) the sarcolemma which is stabilized by a network of proteins including dystrophin in C) muscle fiber

PAGE 196

196 Figure 6 3 Schematic of sample placement in the sample well The diaphragm securely placed at the center of a micro surface coil consisting of central tendon (c) surrounded by muscle fibers(m), the nylon mesh keep the sample from floating on the irrigating water surface, a plastic tip of NMR tube cap as a retention ring, and the PCR film seal tight the sample well.

PAGE 197

197 Figure 6 4 T 2 weighted MRM of the healthy diaphragm MR scan parameters: TE = A) 12.1 ms B) 20 m s, C) 30 ms D) 40 ms E) 50 ms r esolution = 15 x 15 x 160 m 3 NEX/AT = 30/2hr, BW = 50 k Hz Scale bar represents 500 m.

PAGE 198

198 Figure 6 5 Diffusion weighted MRM of the healthy diaphragm. MR scan parameters: b value = A ) 0 [ s/mm 2 ] B ) 600 [s/mm 2 ] C ) 800 [s/mm 2 ] D ) 100 0 [s/mm 2 ] E ) 1300 [s/mm 2 ], r esolution = 15 x 15 x 160 m 3 NEX/AT = 30/2hr, BW = 50 k Hz Scale bar represents 500 m.

PAGE 199

199 F igure 6 6 T 2 weighted MRM of dystrophic diaphragm. MR scan parameters: TE = A) 12.1 ms B) 20 ms C) 30 ms D) 40 ms r esolution = 15 x 15 x 160 m 3 NEX/AT = 30/2hr, BW = 50 k Hz Scale bar represents 500 m.

PAGE 200

200 Figure 6 7 D iffusion weighted MRM of dystrophic diaphrag m MR para meters: b value = A ) 0 [s/mm 2 ] B ) 600 [s/mm 2 ] C ) 1000 [s/mm 2 ] D ) 1300 [s/mm 2 ], r esolution = 15 x 15 x 160 m 3 NEX/AT = 30/2hr, BW = 50 k Hz Scale bar represents 500 m.

PAGE 201

201 Figure 6 8 Diffusion Tensor MRM of healthy diaphragm. A ) b 0 image B ) through D ) R epresentative three out of six different diffusion weighted images along the directions of diffusion sensitizing gradient. Scan parameters: r esolution=15 x 15 x 100 m 3 NEX/AT = 30 / 2 hr, TR/TE = 17.5/2000 ms, b val ue = 1750 [s/mm 2 ], BW = 50 k Hz, / = 8.35/2 ms Scale bar represents 500 m.

PAGE 202

202 Figure 6 9 Diffusion Tensor MRM of di se ased diaphragm A ) b 0 image B ) through D ) R epresentative three out of s ix different directions of DTI images Scan parameters: r esolution = 15 x 15 x 100 m 3 NEX/AT = 30/2hr, TR/TE = 17.5/2000ms, b value = 600 [s/mm 2 ], BW = 50 k Hz, / = 8.35/2 ms Scale bar represents 500 m.

PAGE 203

203 F igure 6 10 R epresentative 2D d iffu sion weighted MRM along sagittal orientation of dystrophic diaphragm. MRM scan parameters: b value = 1000 [s/mm 2 ] r esolution = 15 x 15 x 160 m 3 NEX/AT = 30/2hr, BW = 50 k Hz Scale bar represents 500 m.

PAGE 204

204 Figure 6 1 1 Four r epresentative 3D MR M of the healthy diaphragm MRM scan parameters: 3D gradient ech o isotropic resolution of 9.3 m, TE/TR=2.1/300 ms, NEX/AT=16/10hr 55m Scale bar represents 500 m.

PAGE 205

205 Figure 6 1 2 Representative histology ima ges and MRM of healthy diaphragm A) and B) Histology images visualize the individual muscle fibers embedded in the connective tissue and deeply stained areas corresponding to the nuclei in the healthy model [158] C ) and D ) T 2 weighted MR images visualiz e delineation of intracellular structures in the muscle fibers, i.e., the individual muscle fibers and the nuclei (hypo intense signal) around the periphery are visualized. MR parameters: 2D MSME, TE/TR = 18/2000 ms, r esol ution = 7.8 x 7.8 x 60 m 3 NEX/AT = 92/13hr 5min BW = 101010 Hz

PAGE 206

206 Figure 6 13 Diffusion weighted MRM of the transverse view of the dystrophic diaphragm MRM scan parameters : b value = A) 0 [s/mm 2 ] B) 200 [s/mm 2 ] C) 400 [s/mm 2 ] D) 800 [s/mm 2 ] E) 1000 [s/mm 2 ] F) 1750 [s/mm 2 ] G) 3000 [s/mm 2 ], r esolution = 15 x 15 x 1 60 m 3 TE/TR = 12.4/2000 ms, NEX/AT = 30/2hr, BW = 50 k Hz, / = 4.32/1ms, NEX/AT = 30/2hr Scale bar represents 500 m.

PAGE 207

207 F igure 6 1 4 T 2 weigh t ed MRM of a diseased diaphragm region by region A) Two different regions were designated along the longitudinal MRM. Of particular note i s the different morphological manifestation between B) in proximity with the central tendon and C) distant region. MRM s can parameters: TE/TR = 18/2000 ms, In plane r es olution = 7.8 x 7.8 m 2 Slice t h ickness = 60 m NEX/AT = 92/13hr 5min BW = 101010 Hz Scale bar represents 500 m.

PAGE 208

208 F igure 6 1 5 Diffusion weighted MRM of a diseased diaphragm region by region A) Two different regions, region I and region II, were designated on the longitudinal MRM. B) and C) P roxima l r egion from the central tendon. D) and E) D istant from the central tendon MRM scan parameters: B) and D) b= 1000 [s/mm 2 ] C) and E) 1750 [s/mm 2 ] r esolution = 15 x 15 x 160 m 3 NEX/AT = 30/2hr Scale bar represents 500 m.

PAGE 209

209 Figure 6 1 6 b 0 and three representative images out of six diffusion tensor MRM of a diseased diaphragm. MR M scan parameters: r esolution = 15 x 1 5 x 160 m 3 A) NEX/AT = 30/2hr B) through D) NEX/AT= 28/1hr 59min A) b value = 0, B) through D) b value = 20 00 [s/mm 2 ] Scale bar represents 500 m.

PAGE 210

210 F igure 6 1 7 Comparison of diaphragm in i mmunohistochemistry and MRM. A) Immunohistochemistry visualizes a cross section of diaphragm muscle co reacted with antibodies to laminin (red), myosin heavy chain type I (blue) and myosin heavy chain type IIa (green). [159] B) MR stained image of dystrophic diaphragm Scale bar represents 500 m.

PAGE 211

211 CHAPTER 7 CONCLUSION S AND FUTURE WORKS C onclusions The goal of this dissertation is explorative studies about the volume and mass limited biological samples such as isolated single cells in the invertebrate animals and single cellular compartments inside the mammalian tissues using micro surface coils. To achieve th e goal, three different schemes were utilized. First, through the employment of micro surface coils which have highest sensitivity and bio specificity, the investigation on their MR characteristics regarding biological tissues was carried out. It required different methodological schemes such as bisected and s egmented model s to enhance the accessibility to the core of the deep inside the cell models e.g. motor neuron in Aplysia c alifornica and frog ova. Secondly, the MR stained deep tissue imaging was corr elated with other imaging modalities such as histology to assess its validity. The overall well known shortcomings of MR imaging are low sensitivity and resolution and long acquisition times compared to other imaging modalities such as light microscopy. On the other hand, the advantages of MRI could be its capability in generating endogenous contrast in the tissue and controlling the contrast level. Moreover, unlike optical microscopy, its requirements of invasiveness in sample preparation are drasticall y low therefore the cellular structure inside the tissue keeps its integrity. Furthermore, through its ability of deep tissue penetration, it could localize the signal in the ROI anywhere in the tissue in free of light scattering, absorption or tissue auto fluorescence which make s it a good candidate for possible collaboration with other imaging modalities.

PAGE 212

212 I n this work, while illustrating its superior capabilities to conventional volume coils such as 5 mm and 10 mm volume coils, the applications of MRM dif fusion imaging were investigated into the water diffusion in the biological systems such as nucleoli (=nucleus of nucleus) in chapter 2, single cells in different living organisms, e.g. mice, rats and human in chapter 3 and 4, striated medulla regions or proximal tubular cells in the kidney s and volumetric comparison between wild and mutant types of mice liver and kidneys in chapter 5, and single myofiber cell (=skeletal muscle fiber) whose potential anisotropy inside their architecture is worth to investi gate and anatomical changes in chapter 6. Even if the underlying causes for the restriction in the complex biological system are remained to be investigated, possible explanations are cell membranes, various organelles, water bounded to the macromolecules compartments inside the cells, and fibers, and rearrangement of protein matrix or chain of nucleic acid in the process of fixation. The challenges in operating MRM were also considered. They are mostly related to the hardware and software in terms of t he sensitivity and resolution. They are the more distinct the magnetic susceptibility between tissue components by the higher main magnetic field, the weaker signal strength due to reduction in voxel size in conjunction with higher resolution, the worse si gnal detection due to thermally driven molecular Brownian motion and bandwidth limited resolution, etc. Those issues could be maximally resolved from hardware point of view by employing fast switching and strong gradient and higher magnetic field up to 900 MHz and proper sized micro surface coils with right filling factor and Helmholtz coils for wider coverage while keeping the same

PAGE 213

213 SNR. Additionally from software point of view by developing optimized pulse sequences such as OGSE to minimize the diffusion e ffects and T 2 effects could be a solution. By opening up a new window of imaging modal i ty by combined efforts of dedicated hardware and software, MRM now could access its resolvability from body, organ, tissue down to the cellular level. By registering in dividual MR resolvable compartments with histology which have been difficult with conventional volume coils or solenoidal coils due to low resolution and sensitivity, this research could contribute to set up a highly elaborated reference in the subcellular and cellular MR registered information for more accurate interpretation of MR signals in clinical MR imaging. F uture Works For future works, in combination with RF micro surface coils with a perfusion chamber, the single Aplysia neuron could go through o smotic perturbation to monitor the physiological and pathological changes in the intracellular compartment at the high spatial and temporal resolution. By keeping the viability in a long term experiment, the triggering mechanism behind the reduction of AD C could be elucidated. As with the hypothesis behind reduction of ADC in cell swelling in the stroke, the investigation on the existence of another compartment in the intracellular domain or possible enhancement of restricted ADC in the glial cell layers c ould be monitored by this perturbation study. The fixation effects on the single cell could be estimated by comparing pre and post fixation, or by regulating doping concentration. By presence of interesting cellular organelle named liphochondria in the n euron of sea slug and known to be triggered by the light, this primitive animal might offer an hint about the temporal discrepancy between activation and diffusion of water. Alternatively, the investigation of location of osmolytes could improve more insig ht into the cellular

PAGE 214

214 swelling in affiliation with compartmentation issues. If the glial cells react first to the osmolar changes by swelling themselves, it might offer a clue to the dynamics of cell swelling in the CNS of mammalian where glial cells swell first at the onset of ischemic stroke [84, 85] Furthermore, unlike this animal since the roles and location of osmolytes and the controller were divided in a mammalian, this could be also an good initiative to the another future works on the investigation on activation of neurons in the hypothalamus named supra optic nuclei ( SON ) triggered by osmolar changes and the neurotransmitter named antidiluric hormone ( ADH ) was delivered to the kidney in the mammalian animals. Last but not least, by re gulating the calcium concentration, the single myofiber could be monitored its physiological changes or pathological manifestation in comparison with the diseased model, dystrophic diaphragm disease ( DDD ) by MRM. Throughout these studies described in this dissertation, the ultimate goal of this diagnostic tool is the aid in understanding the source of MR signal changes in pathological and physiological states and monitoring the therapeutic interventions in treating the diseases in human.

PAGE 215

215 LIST OF REFERENC ES 1. Nishimura, D. G., Principles of Magnetic Resonance Imaging 2010, Stanford Univ. 2. Redpath, T. W., Signal to noise ratio in MRI Br J. Radiol. 71(847):704 7 (1988). 3. H ayes C.E., Edelstein, W. A., Schenck, J. F., Mueller, O. M. Eash, M., An efficient, highly homogeneous radiofrequency coil for hole body NMR imaging at 1.5 T Vol. 63. 1985, Orlando, FL, ETATS UNIS: Academic Press. 4. Vaughan, J.T., Hetherington, H. P., Out, J. O., Pan, J. W., Pohost, G. M., High frequency volume coils for clinical NMR imaging and spectroscopy. Magnetic Resonance in Medicine, 1994. 32 (2): p. 206 218. 5. Mansfield, P., Grannell, P.K., NMR 'diffraction' in solids? Journal of Physics C: Solid State Physics, 1973. 6 : p. L422. 6. Liang, Z. P. and P. La uterbur, Principles of Magnetic Resonance Imaging: A Signal Processing Perspective 1999: {Wiley IEEE Press}. 7. Callaghan, Principles of Nuclear Magnetic Resonance Microscopy 1991, Claredon: Oxford. 8. Hahn, E.L., Spin Echoes. Phys. Rev., 1950. 80 : p. 58 0. 9. Stejskal, E.O., Tanner, J. E., Spin diffusion measurements: Spin echoes in the presence of a time dependent field gradient. J. Chem. Phys., 1965. 42 (1): p. 288 292. 10. Hoult, D.I. and R.E. Richards, The signal to noise ratio of the nuclear magnetic resonance experiment. Journal of Magnetic Resonance (1969), 1976. 24 (1): p. 71 85. 11. Hoult, D.I. and P.C. Lauterbur, The sensitivity of the zeugmatographic experiment involving human samples. Journal of Magnetic Resonance (1969), 1979. 34 (2): p. 425 433. 12. Chen, C. N., Hoult, D. I., Biomedical magnetic resonance technology 1989, Bristol and New York: A. Hilger. 13. McFarland, E.W. and A. Mortara, Three dimensional NMR microscopy: improving SNR with temperature and microcoils. Magn Reson Imaging, 1992. 10 (2): p. 279 88. 14. Webb, A.G., Grant, S. C., Signal to noise and magnetic susceptibility trade offs in solenoidal microcoils for NMR. Journal of Magnetic Resonance, 1996. B (113(1)): p. 83 7.

PAGE 216

216 15. Ocali, O., Atalar, E., Ultimate Intrinsic Signal to Noise Ratio in MRI. Magn Reson Med, 1998. 39 : p. 462 473. 16. Hoult, D.I., Phil, D., Sensitivity and power deposition in a high field imaging experiment. J. Magn. Reson. Imaging, 2000. 12 : p. 46 67. 17. Heaviside, O., Electromagnetic theory 1950, New York: Dove r Publications. 18. Massin, C., Boero, G., Vincent, F., Abenhaim, J., Besse, P. A., Popovic, R. S., High Q factor RF planar microcoils for micro scale NMR spectroscopy. Sensors and Actuators, 2002. A: Physical (97 98): p. 280 288. 19. Peck, T. L., Magin, R. L., Lauterbur, P. C., Design and Analysis of Microcoils for NMR Microscopy J. Magn. Reson. 108 (2), 114 124 (1995). 20. Schoeniger, J.S., Aiken, N., Hsu, E., Blackband, S. J., Relaxation time and diffusion NMR microscopy of single neurons. J Magn Reson B 1994. 103 (3): p. 261 73. 21. Cho, Z.H., Ahn, C. B., Juh, S. C., Lee, H. K., Jacobs, R. E., Lee, S., Yi, J. H., Jo, J. M., Nuclear magnetic resonance microscopy with 4 microns resolution: theoretical study and experimental results. Med Phys, 1988. 15 (6): p. 815 24. 22. Bowtell, R.W., Brown, G. D., Glover, P. M., McJury, M., Mansfield, P., Pope, J. M., Ratcliffe, R. E., Waterton, J., and Timms, W. E., Resolution of Cellular Structures by NMR Microscopy at 11.7 T [and Discussion]. Phil. Trans. R. Soc., 1990. A (333): p. 457 467. 23. Cho, Z.H., Ahn, C. B., Juh, S. C., Jo, J. M., Friedenberg, R. M., Fraser, S. E., and Jacobs, R. E., Recent Progress in NMR Microscopy towards Cellular Imaging. Phil. Trans. R. Soc., 1990. A (333): p. 469 475. 24. Norris, D.G., T. Ni endorf, and D. Leibfritz, Healthy and infarcted brain tissues studied at short diffusion times: The origins of apparent restriction and the reduction in apparent diffusion coefficient. NMR in Biomedicine, 1994. 7 (7): p. 304 310. 25. Niendorf, T., Dijkhuize n, R. M., Norris, D. G., van Lookeren Campagne, M., Nicolay, K., Biexponential diffusion attenuation in various states of brain tissue: Implications for diffusion weighted imaging. Magnetic Resonance in Medicine, 1996. 36 (6): p. 847 857. 26. Carr, H.Y. and E.M. Purcell, Effects of Diffusion on Free Precession in Nuclear Magnetic Resonance Experiments. Physical Review, 1954. 94 (3): p. 630 LP 638. 27. Le Bihan, D., Molecular diffusion nuclear magnetic resonance imaging. Magnetic R esonance Q uarterly, 1991. 7 (1): p. 1 30.

PAGE 217

217 28. Szafer, A., J. Zhong, and J.C. Gore, Theoretical model for water diffusion in tissues. Magn Reson Med, 1995. 33 (5): p. 697 712. 29. Pajevic, S. and C. Pierpaoli, Color schemes to represent the orientation of anisotropic tissues from diff usion tensor data: Application to white matter fiber tract mapping in the human brain. Magnetic Resonance in Medicine, 1999. 42 (3): p. 526 540. 30. Webb, A., Radiofrequency microcoils in magnetic resonance, Prog. Nucl. Mag. Reson. Spec., 32:1 42 (1997). 31 Callaghan, P.T., Coy, A., Forde, L. C., Rofe, C. J., Diffusive relaxation and edge enhancement in NMR microscopy Vol. 101. 1993, Orlando, FL, ETATS UNIS: Academic Press. 32. Thelwall, P.E., Shepherd, T. M., Stanisz, G. J., Blackband, S. J., Effects of t emperature and aldehyde fixation on tissue water diffusion properties, studied in an erythrocyte ghost tissue model. Magn Reson Med, 2006. 56 (2): p. 282 9. 33. Grant, S.C., Buckley, D. L., Gibbs, S., Webb, A. G., Blackband, S. J., MR microscopy of multicom ponent diffusion in single neurons. Magn Reson Med, 2001. 46 (6): p. 1107 12. 34. Pierpaoli, C., Jezzard P., Basser P.J., Chiro, G. D, Diffusion tensor MR imaging of the human brain. Radiology, 1996. 201 : p. 637 648. 35. Basser, P.J., J. Mattiello, and D. L ebihan, Estimation of the Effective Self Diffusion Tensor from the NMR Spin Echo. Journal of Magnetic Resonance, Series B, 1994. 103 (3): p. 247 254. 36. Basser, P.J., J. Mattiello, and D. LeBihan, MR diffusion tensor spectroscopy and imaging. Biophysical J ournal, 1994. 66 (1): p. 259 267. 37. Ulug, A.M. and P.C.M. van Zijl, Orientation independent diffusion imaging without tensor diagonalization: Anisotropy definitions based on physical attributes of the diffusion ellipsoid. Journal of Magnetic Resonance Ima ging, 1999. 9 (6): p. 804 813. 38. Einstein, A., On the movement of small particles suspended in a stationary liquid demanded by the molecular kinetic theory of heat Theory of the Brownian Movement. 1905, New York: Dover Publications. 1 18. 39. Eroglu, S., Gimi, B., Roman, B., Friedman, G., Magin, R. L., NMR Spiral Surface Microcoils: Design, Fabrication, and Imaging Concepts in Magnetic Resonance Part B: Magnetic Resonance Imaging, 17B (1), 1 10 (2003). 40. Mansfield, P., Morris, P. G., NMR imaging in bio medicine 1982: Academic Press.

PAGE 218

218 41. Aguayo, J.B., Blackband, S. J., Schoeniger J., Mattingly, M. A., Hintermann, M., Nuclear magnetic resonance imaging for a single cell. Nature, 1986. 322 : p. 190 191. 42. Callaghan, P.T., Forde, L. C. Rofe, C. J., Correl ated Susceptibility and Diffusion Effects in NMR Microscopy Using both Phase Frequency encoding and Phase Phase Encoding. Journal of Magnetic Resonance, 1994. B (104): p. 34 52. 43. Aiken, N.R., Hsu, EW, Blackband SJ, A review of NMR imcroimaging studies of single cells. J Magn Reson Annal, 1995. 1 : p. 41 8. 44. Ciobanu, L., D.A. Seeber, and C.H. Pennington, 3D MR microscopy with resolution 3.7 micro n by 3.3 micro n by 3.3 micro n J Magn Reson, 2002. 158 (1 2): p. 178 82. 45. Weiger, M., Schmidig, D., Denoth, S., Massin, C., Vincent, F., Schenkel, M., Fey, M., NMR microscopy with isotropic resolution of 3.0 m using dedicated hardware and optimized methods. Concepts in Magnetic Resonance Part B: Magnetic Resonance Engineering, 2008. 33B (2): p. 84 93. 46. Badili ta, V., Kratt, K., Baxan, N., Mohmmadzadeh, M., Burger, T., Weber, H., Elverfeldt, D., Hennig, J., Korvink, J. G., Wallrabe, U., On chip three dimensional microcoils for MRI at the microscale. Lab Chip, 2010. 10 (11): p. 1387 1390. 47. Ehrmann, K., Gersbach M., Pascoal, P., Vincent, F., Massin, C., Stamou, D., Besse, P. A., Vogel, H., Popovic, R. S., Sample patterning on NMR surface microcoils. Journal of Magnetic Resonance, 2006. 178 (1): p. 96 105. 48. Ehrmann, K., Saillen, N., Vincent, F., Stettler, M., J ordan, M., Wurm, F. M., Besse, P. A., Popovic, R., Microfabricated solenoids and Helmholtz coils for NMR spectroscopy of mammalian cells. Lab Chip, 2007. 7 (3): p. 373 380. 49. Weber, H., Baxan, N., Paul, D., Maclaren, J., Schmidig, D., Mohammadzadeh, M., H ennig, J., von Elverfeldt, D., Microcoil based MRI: feasibility study and cell culture applications using a conventional animal system. Magnetic Resonance Materials in Physics, Biology and Medicine, 2011. 24 (3): p. 137 145. 50. Flint, J.J., Lee, C.H., Hans en, B., Fey, M., Schmidg, D., Bui, J. D., King, M.A., Vestergaard Poulsen, P., Blackband, S. J., Magnetic resonance microscopy of mammalian neurons. NeuroImage, 2009. 46 (4): p. 1037 1040. 51. Perkowska, E., H.C. Macgregor, and M.L. Birnstiel, Gene Amplific ation in the Oocyte Nucleus of Mutant and Wild type Xenopus laevis. 1968. 217 (5129): p. 649 650. 52. Wu, Z.a. and J. Gall, "Micronucleoli" in the Xenopus germinal vesicle in Chromosoma 1997, Springer Berlin / Heidelberg. p. 438 443.

PAGE 219

219 53. Quintana, C., Cry ofixation, cryosubstitution, cryoembedding for ultrastructural, immunocytochemical and microanalytical studies. Micron, 1994. 25 (1): p. 63 99. 54. Coggeshall, R.E., A light and electron microscope study of the abdominal ganglion of Aplysia californica. J N europhysiol, 1967. 30 (6): p. 1263 87. 55. Ghosh, S., R. Lettr e and I. Ghosh, O n the composition of the nucleolus with special reference to its filamentous structure. Cell and Tissue Research, 1969. 1 01( 2): p. 254 265. 56. Olson, M.O.J., M. Dundr, and A. S zebeni, T he nucleolus: an old factory with unexpected capabilities. Trends in Cell Biology, 2000. 1 0( 5): p. 189 196. 57. Olson, M.O.J., T he Nucleolus. 2004: Springer. 58. Pantelouris, E.M., B. Knox, and H. Wallace, I ron in amphibian oocytes and embryos. Ex perimental Cell Research, 1963. 3 2( 3): p. 469 475. 59. Gautier, T., Masson, C., Quintana, C., Arnoult, J., Hernandez Verdun, D., T he ultrastructure of the chromosome periphery in human cell lines, in C hromosoma. 1992, Springer Berlin / Heidelberg. p. 502 510. 60. Gersch, M., Z. Zellforsch, Microscop. Anat., 1940. 3 0: p. 483. 61. Brown, D.D.a.C., J. D., B iochemistry of amphibian development. I. Ribosome and protein synthesis in early development of Rana Pipiens. Develop. Biol., 1962. 5 : p. 412 434. 62. Hork y, M., Kotala, V., Anton, M., Weisierska Gadek, J., N ucleolus and Apoptosis. Annals of the New York Academy of Sciences, 2002. 9 73( 1): p. 258 264. 63. Sehy, J.V., J.J.H. Ackerman, and J.J. Neil, W ater and lipid MRI of the Xenopus oocyte. Magnetic Resonance in Medicine, 2001. 4 6( 5): p. 900 906. 64. Brangwynne, C.P., Mitchison, T. J., Hyman, A. A., A ctive liquid like behavior of nucleoli determines their size and shape in Xenopus laevis oocytes. PNAS, 2011. 1 08( 11): p. 4334 4339. 65. Ciobanu, L. and C.H. Penn ington, 3 D micron scale MRI of single biological cells. Solid State Nucl Magn Reson, 2004. 2 5( 1 3): p. 138 41. 66. Moroz, L.L., Edwards, J. R.., Puthanveettil, S. V., Kohn, A. B., Ha, T., Heyland, A., Knudsen, B., Sahni, A., Yu, F., Liu, L., Jezzini, S., L ovell, P., Iannucculli, W., Chen, M., Nguyen, T., Sheng, H., Shaw, R., Kalachikov, S., Panchin, Y. V., Farmerie, W., Russo, J. J., Ju, J. Kandel, E. R., N euronal transcriptome of aplysia: neuronal compartments and circuitry. Cell, 2006. 1 27( 7): p. 1453 67.

PAGE 220

220 67. Odelblad E: Micro NMR in high permanent magnetic fields In Nordisk Forening for Obsterik och Gynekologi Edited by Karolinska Institute. Lund, Sweden; 1966. 68. Cajal, S., R., L 'appariel reticulaire de Golgi Holmgren colore par le nitrate d'argent. T rab. del Lab. de Invest. Biol., 1907. T : p. 5. 69. Bensley, R.R., O n the nature of the canalicular apparatus of animal cells. Biol. bull. Woods Hole, Mass, 1910. 1 9: p. 179 194. 70. Golgi, C., I ntorno alla struttura delle cellule nervose. Arch. ital. de Bi ol., 1898. 3 0: p. 60. 71. Kopsch, F., D ie Darstellung des Binnennetzes in spinalen Ganglienzellen und anderen Korperzellen mittels Osmiumsaure. Vol. Halbband II. 1902, Berlin: Sitzungsber. d. k. preuss. Acad. d. Wiss. 72. Baur, P.S., Jr, Brown, A. M., Roge rs, T.D. and Brower, M. E., L ipochondria and the light response of Aplysia giant neurons. J. Neurobiol., 1977. 8 ( 19 42). 73. Assaf, Y. and Y. Cohen, N on Mono Exponential Attenuation of Water andN Acetyl Aspartate Signals Due to Diffusion in Brain Tissue. J ournal of Magnetic Resonance, 1998. 1 31( 1): p. 69 85. 74. Lewis, E.R., T.E. Everhart, and Y.Y. Zeevi, S tudying neural organization in Aplysia with the scanning electron microscope. Science, 1969. 1 65( 898): p. 1140 3. 75. Ambron, R., S. Rayport, and J. Babi arz, G iant Aplysia neuron R2 has distal dendrites: evidence for protein sorting and a second spike initiation site. J. Neurosci., 1988. 8 ( 2): p. 722 731. 76. Eisenstadt, M., Goldman, J. E., Kandel, E. R., Koike, H., Koester, J., Schwartz., J. H. I ntrasoma tic injection of radioactive precursors for studying transmitter synthesis in identified neurons of Aplysia californica. Proc Natl Acad Sci U S A, 1973. 7 0( 12): p. 3371 5. 77. Ye, X., Xie, F., Romanova, E. V., Rubakhin, S. S., Sweedler, J. V., P roduction o f Nitric Oxide within the Aplysia Californica Nervous System. ACS Chem Neurosci, 2010. 1 ( 3): p. 182 193. 78. Giller, E., Jr. and J.H. Schwartz, C holine acetyltransferase in identified neurons of abdominal ganglion of Aplysia californica. J Neurophysiol, 19 71. 3 4( 1): p. 93 107. 79. Keicher, E., Maggio, K., Hernandez Nicaise, ML, Nicaise, G., T he abundance of Aplysia gliagrana depends on Ca2+ and Na+ concentrations in sea water. Glia, 1992. 5 ( 2): p. 131 8. 80. Barmpoutis, A., f anDTasia. 2010 ( http://www.cise. ufl.edu/~abarmpou/lab/fanDTasia/ )

PAGE 221

221 81. Hsu, E.W., N.R. Aiken, and S.J. Blackband, N uclear magnetic resonance microscopy of single neurons under hypotonic perturbation. Am J Physiol, 1996. 2 71( 6 Pt 1): p. C1895 900. 82. Grant, S.C., Aiken, N. R., Plant, H. D., Gibbs, S., Mareci, T. H., Webb, A. G., Blackband, S. J. N MR spectroscopy of single neurons. Magn Reson Med, 2000. 4 4( 1): p. 19 22. 83. Temburni, M.K., Jacob, M.H, N ew functions for glia in the brain. PNAS, 2001. 9 8( 7): p. 3631 3632. 84. Papadopoulos, M.C., Verkman, A. S., A quaporin 4 Gene Disruption in Mice Reduces Brain Swelling and Mortality in Pneumococcal Meningitis. Journal of Biological Chemistry, 2005. 2 80( 14): p. 13906 13912. 85. Kahle, K.T., Simard, J. M., Staley, K. J., Nahed, B. V., Jones, P S., Sun, D., M olecular Mechanisms of Ischemic Cerebral Edema: Role of Electroneutral Ion Transport. Physolology, 2009. 2 4( 4): p. 257 265. 86. Le Bihan, D., T he 'wet mind': water and functional neuroimaging. Physics in Medicine and Biology, 2007. 5 2( 7). 8 7. Rutledge, D.N., S ignal Treatment and Signal Analysis in NMR, ed. D.N. Rutledge. 1996: Elsevier. 88. Consortium, T.E.P.K.D., T he polycystic kidney disease 1 gene encodes a 14 kb transcript and lies within a duplicated region on chromosome 16. Cell, 1994. 7 7( 6): p. 881 894. 89. Scherini, E., Bernocchi, G., E ctopic Purkinje like cells are GABAergic: immunohistochemistry with an immune serum against glutamic acid decarboxylase. Cell Tissue Res, 1989. 2 58( 2): p. 437 9. 90. De Camilli, P., Miller, P. E., Levi tt, P., Walter, U., Greengard, P., A natomy of cerebellar Purkinje cells in the rat determined by a specific immunohistochemical marker. Neuroscience, 1984. 1 1( 4): p. 761 IN2. 91. Chung, C.J., Castillo, M., Fordham, L, Mukherji, S., Boydston, W., Hudgins, R ., S pinal intradural cerebellar ectopia. AJNR Am J Neuroradiol, 1998. 1 9: p. 897 9. 92. Duong, T.Q., Ackermann, J. J., Ying, H. S., Neil, J. J., E valuation of extra and intracellular apparent diffusion in normal and globally ischemic rat brain via 19F NMR Magn Reson Med, 1998. 4 0( 1): p. 1 13. 93. Narahara, H.T., Ozand, P., S tudies of Tissue Permeability: IX. The effect of insulin on the penetration of 3 methylglucose H3 in frog muscle. J. Biol. Chem., 1963. 2 38: p. 40 49.

PAGE 222

222 94. Thorne, R.G., Nicholson, C., I n vivo diffusion analysis with quantum dots and dextrans predicts the width of brain extracellular space. PNAS, 2006. 1 03( 14): p. 5567 5572. 95. Nicholson, C. and E. Sykov a Ex tracellular space structure revealed by diffusion analysis. T rends in Neuroscie nces, 1998. 21 (5 ): p. 207 215. 96. Hyslop, W.B., Lauterbur, P. C., Ef fects of Restricted Diffusion on Microscopic NMR Imaging. J Magn. Reson., 1991. 94 : p. 501 510. 97. Tzika, A.A., Zarifi, M. K., Goumnerova, L., Astrakas, L. G., Zurakowski, D., Young Pou ssaint, T., Anthony, D. C., Scott, R. M., Black, P. M., Ne uroimaging in Pediatric Brain Tumors: Gd DTPA enhanced, Hemodynamic, and Diffusion MR Imaging Compared with MR Spectroscopic Imaging. A JNR Am J Neuroradiol, 2002. 23 (2 ): p. 322 333. 98. Jackson, L.S ., Nelson, J. A., Case, T. A., Burnham, B. F. Ma nganese protoporphyrin IX. A potential intravenous paramagnetic NMR contrast agent: preliminary communication. I nvest Radiol, 1985. 20 (2 ): p. 226 9. 99. Koenig, S.H., Brown III, R. D., Goldstein, E. J., Burn ett, K. R., Wolf, G.L., Ma gnetic field dependence of proton relaxation rates in tissue with added Mn2+: rabbit liver and kidney. M agn Reson Med, 1985. 2( 2 ): p. 159 68. 100. Wolf, G.L. Co ntrast Agents in Magnetic Resonance Imaging. in Jo urnal of Computer As sisted Tomography. 1987. 101. Wickelgren, I., Fi rst Components Found for Key Kidney Filter. S cience, 1999. 28 6(5 438): p. 225 226. 102. Gabow, P.A., Au tosomal dominant polycystic kidney disease. N Engl J Med, 1993. 32 9(5 ): p. 332 42. 103. Du, J., Wilson, P. D., Ab normal polarization of EGF receptors and autocrine stimulation of cyst epithelial growth in human ADPKD. a m J Physiol Cell Physiol, 1995. 26 9(2 ): p. C487 495. 104. Mochizuki, T., Wu, G., Hayashi, T., Xenophontos, S. L., Veldhuisen, B., Saris, J. J., Reynolds, D. M., Cai, Y., Gabow, P. A., Pierides, A., Kimberling, W. J., Breuning, M. H., Deltas, C., Peters, D. J. M., Somlo, S., PK D2, a gene for polycystic kidney disease that encodes an integral membrane protein. S cience, 1996. 27 2(5 266): p. 1339 42.

PAGE 223

223 105. Burn, T.C., Connors, T. D., Dackowski, W. R., Petry, L. R., van Raay, T. J., Millholland, J. M., Venet, M., Miller, G., Hakim, R. M., Landes, G. M., Kilnger, K. W., Qian, F., Onuchic, L. F., Watnick, T., Germino, G. G., Doggett, N. A., An alysis of the genomic sequence for the autosomal dominant polycystic kidney disease (PKD1) gene predicts the presence of a leucine rich repeat. The American PKD1 Consortium (APKD1 Consortium). H um Mol Genet, 1995. 4( 4 ): p. 575 82. 106. Rinkel, G.J.E., Djibuti, M., Algr a, A., van Gijn, J., Pr evalence and Risk of Rupture of Intracranial Aneurysms: A Systematic Review. S troke, 1998(29): p. 251 256. 107. Onuchic, L., Furu, L, Nagasawa, Y, Hou, X., Eggermann, T., Ren, Z., Bergmann, C., Senderek, J., Esquivel, E., Zeltner, R. Rudnik Schoneborn, S., Mrug, M., Sweeney, W., Avner, E. D., Zerres, K., Guay Woodford, L. M., Somlo, S., Germino, G. G., PKHD1, the polycystic kidney and hepatic disease 1 gene, encodes a novel large protein containing multiple immunoglobulin like plexin transcription factor domains and parallel beta helix 1 repeats. A m J Hum Genet, 2002. 70 : p. 1305 17. 108. Ward, C., Hogan, MC, Rossetti, S, Walker, D., Sneddon, T., Wang, X., Kubly, V., Cunningham, J. M., Bacallao, R., Ishibashi, M., Milliner, D. S., Tor res, V. E., Harris, P.C., The gene mutated in autosomal recessive polycystic kidney disease encodes a large, receptor like protein N at Genet, 2002. 30 : p. 259 269. 109. Desmet, Wh at is congenital hepatic fibrosis? H istopathology, 1992. 20 : p. 465 7. 110. Wilson, P.D., Ep ithelial cell polarity and disease. A m J Physiol Renal Physiol, 1997. 27 2(4 ): p. 1466 1522. 111. Lebeau, C., Hanaoka, K, Moore Hoon, ML, and W. Guggino, Beawens, R, Devuyst, O., BS C2 is the basolateral chloride transporter in a subset of cy sts in autosomal dominant polycystic kidney disease (ADPKD). J Am SocNephrol, 1999. 10 (3 6A). 112. Guay Woodford, L., Mu rine models of polycystic kidney disease: molecular and therapeutic insights. A merican journal of physiology, 2003. 28 5(6 ): p. 1034 1049. 113. Ricker, J., Gattone, VH II, Calvet, JP, Rankin, CA, De velopment of autosomal recessive polycystic kidney disease in BALB/c cpk/cpk mice. J Am Soc Nephrol, 2000. 11 (1 0): p. 1837 1847. 114. Mrug, M., Li, R., Cui, X., Schoeb, T. R., Churchill, G. A., Gu ay Woodford, L. M. Ki nesin family member 12 is a candidate polycystic kidney disease modifier in the cpk mouse. J Am Soc Nephrol, 2005. 16 (4 ): p. 905 16. 115. Cowley, B.D., Jr., Re cent advances in understanding the pathogenesis of polycystic kidney diseas e: therapeutic implications. D rugs, 2004. 64 (1 2): p. 1285 94.

PAGE 224

224 116. Crozier, S., Doddrell, DM, In vivo localised 1Hspectroscopy of the rat eye: Preliminary studies. M AGMA, 1995. 3( 1 37). 117. Crozier, S., G. Cowin, and Z.H. Endre, MR microscopy and microspec troscopy of the intact kidney. C oncepts in Magnetic Resonance Part A, 2004. 22 A(1 ): p. 50 59. 118. Tarver, R., Conces, D J Cory, D A Vix, V. A., Im aging the diaphragm and its disorders. J Thorac Imaging, 1989. 4( 1 ): p. 1 18. 119. Standring, S., Gr ays A natomy, The Anatomical Basis of Clinical Practice. 2009, Philadelphia: Elsevier. 120. Rosenberg, R.N., DiMauro, S., Paulson, H. L, Ptacek, L. Nestler, E. J, Th e molecular and genetic basis of neurologic and psychiatric disease. 2008, Philadelphia: Wolter s Kluwer/Lippincott Williams & Wilkins. 121. Hoffman, E., Brown, RH Jr, Kunkel, LM., Dy strophin: the protein product of the Duchenne muscular dystrophy locus. C ell, 1987. 51 : p. 919 928. 122. Allen, D.G., N.P. Whitehead, and E.W. Yeung, Me chanisms of stret ch induced muscle damage in normal and dystrophic muscle: role of ionic changes. T he Journal of Physiology, 2005. 56 7(3 ): p. 723 735. 123. Fanchaouy, M., Polakova, E, Jung,C, Ogrodnik, J, Shirokova, N, Niggli, E., Pa thways of abnormal stress induced Ca2+ i nflux into dystrophic mdx cardiomyocytes. C ell Calcium, 2009. 46 (2 ): p. 114 21. 124. Gissel, H., Th e Role of Ca2+ in Muscle Cell Damage. A nnals of the New York Academy of Sciences, 2006. 10 66(1 ): p. 166 180. 125. Phillips, M.F., Quinlivan, R., Ca lcium anta gonists for Duchenne muscular dystrophy. C ochrane Neuromuscular Disease Group, 2009. 126. Anderson, J.L., Head, S. I., Rae, C., Morley, J. W., Br ain function in Duchenne muscular dystrophy. B rain, 2002. 12 5(1 ): p. 4 13. 127. Shavlakadze, T., White, J., Hoh J.F., Rosenthal, N., Grounds, M.D, Ta rgeted expression of insulin like growth factor 1 reduces the necrosis of dystrophic mdx muscle. M ol. Ther, 2004. 10 : p. 829 843. 128. Torres, L.F.B., Duchen, L.W., T he mutant mdx: inherited myopathy in the mouse: mo rphological studies of nerves, muscles and end plates. B rain, 1987. 11 0(2 ): p. 269 299. 129. Coulton, G.R., Curtin, N. A., Morgan, J. E., Partridge, T. A., T he mdx mouse skeletal muscle myopathy: II. contractile properties N europathology and Applied Neurobi ology, 1988. 14 (4 ): p. 299 314.

PAGE 225

225 130. Whitehead, N.P., E.W. Yeung, and D.G. Allen, Muscle damage in mdx (dystrophic) mice: role of calcium and reactive oxygen species C linical and Experimental Pharmacology and Physiology, 2006. 33 (7 ): p. 657 662. 131. Radl ey, H.G., Grounds, M.D., Cr omolyn administration (to block mast cell degranulation) reduces necrosis of dystrophic muscle in mdx mice. N eurobiol. Dis., 2006. 23 : p. 387 397. 132. Grounds, M.D., Torrisi, J.O., An ti TNF?(Remicade? therapy protects dystrophic skeletal muscle from necrosis. FAS EB J, 2004. 18: p. 676 682. 133. Allamand, V., Campbell, K. P., Anim al models for muscular dystrophy: valuable tools for the development of therapies. Hum Mol. Genet., 2000. 9(16 ) : p. 2459 2467. 134. Salimena, M.C., Lagr ota Candido, J., Quirico Santos, T.L., Gend er dimorphisminfluences extracellular matrix expression and regeneration of muscular tissue inmdx dystrophic mice. His tochem. Cell Biol., 2004. 122: p. 435 444. 135. Stedman, H.H., Sweeney, H. L., Shrager, J. B., Maguire, H. C., Panettieri, R. A., Petrof, B., Narusawa, M., Leferovich, J. M., Sladky, J. T., Kelly, A. M., The mdx mouse diaphragm reproduces the degenerative changes of Duchenne muscular dystrophy. 199 1. 352( 633 5): p. 536 539. 136. Faulkner, J.A., Ng., R., Davis, C. S., Li, S., Chamberlain, J. S., Diaphragm muscle strip preparation for evaluation of gene therapies in mdx mice Cli nical and Experimental Pharmacology and Physiology, 2008. 35(7 ): p. 725 729. 137. McNeil, P.L. and T. Kirchhausen, An e mergency response team for membrane repair. 200 5. 6(6) : p. 499 505. 138. Doherty, K.R. and E.M. McNally, Repa iring the tears: dysferlin in muscle membrane repair. Tre nds in Molecular Medicine, 2003. 9(8) : p. 327 330. 139. Straub, V., Rafael, J.A., Chamberl ain, J.S., Campbell, K.P., Anim al models for muscular dystrophy show different patterns of sarcolemmal disruption. J. Cell Biol., 1997. 139: p. 375 385. 140. Matsuda, R., Nishikawa, A., Tanaka, H., Visu alization of dystrophic muscle fibers in mdx mouse by vital staining with Evans blue: evidence of apoptosis in dystrophindeficient muscle. J. Biochem., 1995. 118: p. 959 964. 141. Hamer, P.W., McGeachie, J. M., Davies, M. J., Grounds, M. D., Evans Blue Dye as an in vivo marker of myofibre damage: optimising p arameters for detecting initial myofibre membrane permeability. Jou rnal of Anatomy, 2002. 200( 1): p. 69 79.

PAGE 226

226 142. Amthor, H., Egelhof, T., McKinnell, I., Ladd, M. E., Janssen, I., Weber, J., Sinn, H., Schrenk, H. H., Forsting, M., Voit, T., Straub, V., Albu min t argeting of damaged muscle fibres in the mdx mouse can be monitored by MRI. Neu romuscular Disorders, 2004. 14(1 2) : p. 791 796. 143. Hussein, M.R., Hamed, S. A., Mostafa, M. G., Abu Dief, E. E., Kamel, N. F., Kandil, M. R., The effects of glucocorticoi d therapy on the inflammatory and Dendritic cells in muscular dystrophies. Int ernational Journal of Experimental Pathology, 2006. 87(6 ): p. 451 461. 144. Alman, B., Raza, SN, Biggar, WD, Steroid treatment and the development of scoliosis in males with duch enne muscular dystrophy. J B one Joint Surg Am, 2004. 86 A (3): p. 519 524. 145. Jacobs, A. M., Ibrahim, T. S., Ouwerkerk, R ., MR Imaging: Brief Overview and Emerging Applications RadioGraphics, 2007. 27, 1213 1229 146 Bruker BioSpin (www.bruker biospin.com) 147. Neil J.J., Duong, T. Q., Ackerman, J. J., Evaluation of intracellular diffusion in normal and globally ischemic rat brain via 133Cs NMR Magn Reson Med, 1996, 35 (3):329 35 148. van der Toorn, A, Dijkhuizen, R.M., Tulleken, CAF, Nicolay K., Diffusion of metabolites in normal and ischemic rat brain measured by localized 1H MRS. Magn. Reson. Med. 1996, 36(6):914 922 149. The European Polycystic Kidney Disease Consortium, The polycystic kidney disease 1 gene encodes a 14 kb transcript and lies within a duplicated region on chromosome 16 Cell. 1994, 77(6), 881 894. 150. Sun, Y., Zhou, J., Stayner, C., Munasinghe, J., Shen, X., Beier, D. R., Albert, M. S., Magnetic Resonance Imaging Assessment of a Murine Model of recessive Polycystic Kidney Disease Comparative Medicine 2002, 52(5), 433 438. 151 Handwerger, K. E., Cordero, J. A., Gall, J. G., Cajal Bodies, Nucleoli, and Speckles in the Xenopus Oocyte Nucleus Have a Low Density, Sponge like Structure Mol Biol Cell. 2005, 16(a), 202 211 152 Qu inlan J.G., Hahn H.S., Wong B.L., Lorenz J.N., Wenisch A.S., Levin L.S. Evolution of the mdx mouse cardiomyopathy: physiological and morphological findings Neuromuscul.Disor 2004. 14, 491 496. 153 Walter, G., Cordier, L., Bloy, D., Sweeney, H. L., Noninva sive monitoring of gene correction in dystrophic muscle Magn. Res. Med. 2005, 54 (6), 1369 1376.

PAGE 227

227 154 Leijendekker, W. J., Passaquin, A. C., Metzinger, L., Ruegg, U. T., Regulation of cytosolic calcium in skeletal muscle cells of the mdx mouse under condit ions of stress Br J Pharmacol. 1996, 118(3), 611 616. 155 Nomizu, T., Falchuk, K. H., Vallee, B. L., Zinc, iron, and copper contents of Xenopus laevis oocytes and embryos Molecular Reproduction and Development. 1993, 36(4), 419 423. 156 Kandror, K. V., Stepanov, A. S., Tsuprun, V. L., The main adenosine triphosphate binding component of amphibian oocytes is ferritin Molecular Reproduction and Development 1992, 31(1), 48 54. 157 Harrison, P. M., Artymiuk, P. J., Ford, G. C., Lawson, D. M., Smith, J. M. A ., Treffy, A., White, J. L., Ferritin: Function and structural design of an iron stroage protein in Biomineralisation, Mann, S., Webb., Williams, R. J. P., Eds., VCH Publishers, Weinheim, Germany, 1989, 258. 158 Caceci, T., Smooth and skeletal muscle, 201 1, http://www.vetmed.vt.edu/education/Curriculum/VM8054/Labs/Lab10/lab10.htm 159 Lynch, G., Muscle and Hargreaves, M., Immunohistochemisty, 1994 2011, http://www.physiology.unimelb.edu.au/lab muscle/lab techniques.asp 160 Mori, S., Crain, B. J., Chacko, V. P., van Zijl, P. C., Three dimensional tracking of axonal projections in the brai n by magnetic resonance imaging Ann Neurol., 1999, 45(2): 265 9. 161 Axelson, D. E., Wooten, J. B., Bulk Analysis of Tobacco and Cigarettes by Magnetic Resonance Imaging J. Agric. Food Chem., 2000, 48 (6), 2199 2207. 162 Barreiro, P., Ortiz, C., Ruiz Alt isent, M., Ruiz Cabello, J., Fernandez Valle, M. E., Recasens, I., Asensio, M., Mealiness assessment in apples and peaches using MRI techniques Magnetic Resonance Imaging 2000, 18(9), 1175 1181. 163 Chaerle, L., Van der straiten, D., Seeing is believing: imaging techniques to monitor plant health Biochimica et Biophysica Acta (BBA) 2001, 1519(3), 153 166. 164 Cornillon, P., Salim, L. C., Characterization of water mobility and distribution in low and intermediate moisture food systems Magnetic Resonance Imaging 2000, 18(3), 335 341. 165 Almeida, G., Leclerc, S., Perre, P., NMR imaging of fluid pathways during drainage of softwood in a pressure membrane chamber International Journal of Multiphase Flow 2008, 34(3), 312 321

PAGE 228

228 166 Hernandez, R. E., Caceres, C B., Magnetic Resonance Microimaging of Liquid Water Distribution in Sugar Maple Wood Below Fiber Saturation Point Wood and Fiber Science 2010, 42(3), 259 272. 167 Maccotta, A., Fantazzini, P., Garavaglia, C., Donato, I. D., Perzia, P., Brai, M., Morreal e, F., Preliminary 1H NMR Study on archaeological Waterlogged Wood Annali di Chimica 2005, 95(3 4), 117 124. 168 Merela, M., Oven, P., Sepe, A., Sersa, I., Three dimensional in vivo magnetic resonance microscopy of beech (Fagus sylvatica L.) wood Magneti c Resonance Materials in Physics, Biology and Medicine 2005, 18(4), 171 174. 169 Morales, S., Guesalaga, A., Fernandez, M. P., Guarini, M., Irarrazaval, P., Computer reconstruction of pine growth rings using MRI Magnetic Resonance Imaging 2004, 22(3), 403 412. 170 Blumich, B., Casanova, F., Mobile NMR Modern Magnetic Resonance 2006, 373 382. 171 Blumich, B., 2000 NMR Imaging of materials Vol. 57 of Monographs on the physics and chemistry of material. Oxford Science Publications, Clarendon Press. 172 Gore J. C., Anderson, A. W., Does, M. D., Gochberg, D. F., Joers, J. M., Kennan, R. P., Parsons, E. C., Schachter, M., The relationship of problems in biomedical MRI to the study of porous media Magnetic Resonance Imaging 2001, 19(3 4), 295 300. 173 Kocken berger, W., Functional imaging of plants by magnetic resonance experiments Trends in Plant Science 2001, 6(7), 286 292. 174 Scheenen, T. W. J., Vergeldt, F. J., Heemskerk, A. M., Van As, H., Intact Plant Magnetic Resonance Imaging to Study Dynamics in Lon g Distance Sap Flow and Flow Conducting Surface Area Plant Physiology 2007, 144:1157 1165. 175 Wu, J., Yu, D., Chan, C M, Kim, J., Mai, Y W, Effect of fiber pretreatment condition on the interfacial strength and mechanical properties of wood fiber/PP co mposites, Journal of Applied Polymer Science, 76, 1000 1010 (2000). 176 Peter, G., Lee, C.H., Lundsford, K., Blackband, S., Ingram, L., Yost, R., Int181egrated Nondestructive Spatial and Chemical Analysis of Lignocellulosic Materials during Pretreatment a nd Bioconversion to Ethanol Genomic Science 2010 Awardee Workshop VIII and Knowledgebase Workshop Crystal City, Va., February 7 10, 2010. 177 Eccles, C. D. & Callaghan, P.T. High resolution imaging: the NMR microscope J. Magn. Reson. 68, 393 398(1986)

PAGE 229

229 178 Johnson, G. A., Thompson, M. B., Gewalt, S. L., & Hayes, C. E. Nuclear magnetic resonance imaging at microscopic resolution J. Magn. Reson. 68, 129 137 (1986). 179 Kaczmar ek, L.K., Finbow, M., Revel, J. P., Strumwasser, F., The Morphology and coupling of Aplysia bag cells within the abdominal ganglion and in cell culture Journal of Neurobiology, 10 (6), 535 550 (1979). 180 Glover, P., Mansfield, S. P., Limits to magnetic r esonance microscopy Rep. Prog. Phys. 65 (10), 1489 (2002). 181 Fine, E. J., Ionita, C. C., Lohr, L., The history of the development of the cerebellar examination Semin Neurol 22 (4): 375 84 (2002). 182 Voogd, J., Glickstein, M., The anatomy of the cere bellum Trends in Cognitive Sciences, 2(9), 307 313 (1998). 183 Gall J. G. Cajal bodies: The first 100 years Annu Rev Cell Dev Biol 16: 273 300 (2000). 184 Savas, J. N., Makusky, A., Ottosen, S., Baillat, D., Then, F., Krainc, D., Shiekhattar, R., Markey S. P., Tanese, N., contributes to RNA mediated gene silencing through association with Arogonaute and P bodies Proc Natl Acad Sci USA 105: 10820 10825(2008) 185 Barbato, C., Ruberti, F., Cogoni, C., Searching for MIND: Micro RNAs in Neurodegenerative Diseases Journal of Biomedicine and Biotechnology, 2009: 871313 (2009). 186 Bethe, A. The permeability of the surface of marine animals The Rockefeller University Press, 13(4), 437 444 (1930). 187 Schacher, S. M., The role of s upport cells in the growth and differentiation of neurons in the abdominal ganglion of Aplysia J. exp. Biol., 95, 205 214 (1982) 188 Webster, M., Witkin, K.L., Cohen Fix, O. Sizing up the nucleus: nuclear shape, size and nuclear envelope assembly J. Cell Sci. 122, 1477 1486(2009). 189 Zink, D., Fischer, A.H., Nickerson, J. A. Nuclear structure in cancer cells ., Nature. Rev. Cancer 4, 677 687(2004). 190 Schachter, M., Does, M. D., Anderson, A. W., Gore, J. C., Measurements of Restricted Diffusion Using an Oscillating Gradient Spin Echo Sequence Journal of Magnetic Resonance, 147(2), 232 237 (2000) 191 Does, M. D., Parson, E. C., Gore, J. C., Oscillating gradient measurements of water diffusion in normal and globally ischemic rat brain Magn. Reson. Med., 4 9, 206 215 (2003).

PAGE 230

230 192 Marckmann, P., Skov, L., Rossen, K., Dupont, A., Damholt, M. B., Heaf, J. G., Thomsen, H. S., Nephrogenic Systemic Fibrosis: Suspected Causative Role of Gadodiamide Used for Contrast Enhanced Magnetic Resonance Imaging J. Am. Soc. Nephrol. 17: 2359 2362(2006). 193 Thomsen, H. S., Nephrogenic systemic fibrosis: a serious late adverse reaction to gaodiamide 16(12), 2619 2621 (2006). 194 Zerres, K., Rudnik Schoneborn S., Deget, F., Holtkamp, U., Brodehl, J., Geisert, J., Scharer, K., Clinical course of 115 children with autosomal recessive polycystic kidney disease Acta Peadiatr 85: 437 445 (1996). 195 Chung, H. M., Chung, K. W., BRS Gross Anatomy ISBN 0781771749, Lippincott Williams & Wilkins; Sixth edition (October 12, 2007). 196 Nyengaard, J. R. & Bendtsen, T. F. Glomerular number and size in relation to age, kidney weight, and body surface in normal man Anat. Rec. 232 194 201 (1992). 197 Walter, F., PhD. Boron. Medical Physiology: A Cellular And Molecular approach Elsevier/S aunders, ISBN 1 4160 2328 3 (2004) 198 Guay Woodford, L.M. and Desmond, R. A., Autosomal recessive polycystic kidney disease: the clinical experience in North America Pediatrics 111: 1072 1080 (2003). 199 Zerres, K. Rudnik Schoneborn, S., Steinkamm, C., Becker, J., Mucher, G., Autosomal recessive polycystic kidney disease (ARPKD) J. Nephrol 16: 453 458 (2003). 200 Torres, v. E. & Harris, P.C. Mechanisms of disease: autosomal dominant and recessive polycystic kidney diseases Nature Clin. Pract. Nephrol. 2, 40 55; quiz 55 (2006). 201 Wilson, P.D., Sherwood, A. C., Palla, K., Du, J., Watson, R., Norman, J. T., Reversed polarity of Na+ K+ ATPase: mislocation to apical plasma membrane Am. J. Physiol. 260 (Renal Fluid Electrolyte Physiol. 29): F420 F430, 199 1. 202 Menezes LF, Cai Y, Nagasawa Y, Silva AM, Watkins ML, Da Silva AM, Somlo S, Guay Woodford LM, Germino GG, Onuchic LF (2004) Polyductin, the PKHD1 gene product, comprises isoforms expressed in plasma membrane, primary cilium, and cytoplasm Kidney Int 66:1345 1355. 203 Nagano J Kitamura K Hujer K M Ward C J Bram R J Hopfer U Tomita K Huang C Miller R T (2005) Fibrocystin interacts with CAML, a protein involved in Ca(2+) signaling. Biochem Biophys Res Commun 338:880 889.

PAGE 231

231 204 Zhang M Z Mai W Li C Cho S Y Hao C Moeckel G Zhao R Kim I Wang J Xiong H Wang H Sato Y Wu Y Nakanuma Y Lilova M Pei Y Harris R C Li S Coffey R J Sun L Wu D Chen X Z Breyer M D Zhao Z J McKanna J A Wu G ., PKHD1 protein encoded by the gene for autosomal recessive polycystic kidney disease associates with basal bodies and primary cilia in renal epithelial cells Proc Natl Acad Sci USA 101:2311 2316 (2004) 205 Avner, E. D., Sweeney, W. E. Jr., Nelson, W. J., Abnormal sodium pump distribution during renal tubulogenesis in congenital murine polycystic kidney disease PNAS, 89(16): 7447 7451 (1992). 206 Veizis, E. I., Carlin, C. R., Cotton, C. U., Decreased amiloride sensitive Na+ absorption in collecting duct pr incipal cells isolated from BPK ARPKD mice AJP Renal Physiol, 286(2): F244 254 (2003). 207 Gattone V H 2nd, Grantham J J ., Understanding human cystic disease through experimental models Semin Nephrol 1991; 11: 617 631. 208 Aziz N. Animal models of p olycystic kidney disease BioAssays 1995; 17:703 712. 209 Orellana S A Avner E A. Cystic maldevelopment of the kidney Semin Nephrol 1995; 15: 341 352. 210 Calvet, J. P., Polycystic kidney disease: Primary extracellular matrix abnormality or defective cellular differentiation Kidney International, 43:101 108 (1993). 211 Nauta, J., Ozawa, Y., Sweeney, W. E., Rutledge, J. C., Avner, E. D., Renal and biliary abnormalities in a new murine model of autosomal recessive polycystic kidney disease 7(2), 163 172 (1993). 212 Picolli, D. A., Witzleben, C. L. Disorders of intrahepatic bile ducts In: Walker WA, Durie PR, Hamilton JR, Walker Smith, Watkins JB (eds) Pediatric gastrointestinal disease, vol 2. Decker, Philadelphia, pp 1124 1140 (1991). 213 Kaplan, B S., Kaplan, P., Autosomal recessive polycystic kidney disease In: Spitzer A, Avner ED (eds) Inheritance of kidney and urinary tract diseases. Kluwer, Dordrecht, pp 265 276 (1990). 214 Williams, T. H., Jew, J. Y., An improved method for perfusion fixatio n of neural tissues for electron microscopy 7(3), 407 418(1975). 215 Rives, J. D., & Baker, D. D ., Anatomy of the Attachments of the Diaphragm: Their Relation to the Problems of the Surgery of Diaphragmatic Hernia Annals of Surgery, 115 (5), 745 755 (1942 ).

PAGE 232

232 216 Rives, J. D., & Baker, D. D. (1942 ). Anatomy of the Attachments of the Diaphragm: Their Relation to the Problems of the Surgery of Diaphragmatic Hernia Annals of Surgery, 115 (5), 745 755. 217 Cheuk D K Wong V Wraige E Baxter P Cole A N'Diaye T Mayowe, V., Surgery for scoliosis in Duchenne muscular dystrophy Cochrane Database Syst Rev 2007;(1):CD005375. 218 Karol L. Scoliosis in patients with Duchenne muscular dystrophy J Bone Joint Surg Am 2007;89(Suppl 1):155 162. 219 Schellinger P.D., Jansen, O., Fiebach, J.B., Pohlers, O., Ryssel, H., Heiland, S., Steiner, T., Hacke, W., Sartor, K., 2000. Feasibility and practicality of MR imaging of stroke in the management of hyperacute cerebral ischem ia. Am. J. Neuroradiol. 21 (7), 1184 1189 AUG. 220 Sotak, C.H., Nuclear magnetic resonance (NMR) measurement of the apparent diffusion coefficient (ADC) of tissue water and its relationship to cell volume changes in pathological states Neurochem. Int. 45, 569 582 ( 2004 ). 221 Van Der Toorn, A., S ykova, E., Dijkhuizen, R.M., Vorisek, I., Vargova, L., Skobisova, E., Van Lookeren Campagne, M., Reese, T., Nicolay, K., 1996. Dynamic changes in water ADC, energy metabolism, extracellular space volume, and tortuosity in neonatal rat brain during global i schemia Magn. Reson. Med. 36, 52 60. 222 Jin, J. M., Electromagnetic analysis and design in magnetic resonance imaging 1999, Boca Raton, Fla: CRC Press. 223 Slichter, C.P., Principles of magnetic resonance 3rd ed. ed. 1990, Berlin: Springer Verlag. 224 Cleveland, G.G., Chang, D. C., Hazlewood, C. F., Rorschach, H. E., Nuclear magnetic resonance measurement of skeletal muscle: anisotrophy of the diffusion coefficient of the intracellular water. Biophysical Journal, 1976. 16 (9): p. 1043 1053. 225 Cory, D.G ., Garroway, A. N., Measurement of translational displacement probabilities by NMR: An indicator of compartmentation. Magn Reson Med, 1990. 14 (3): p. 435 444. 226 Vaughan, J. T., Hetherington, H. P., Otu, J. O., Pan, J. W., Pohost, G. M., High frequency vo lume coils for clinical NMR imaging and spectroscopy Magn. Reson. Med., vol 32, 206 218(1994). 227 Baertlein, B. A., Ozbay, O., Ibrahim, T., Lee, R., Ying, Y., Kangarlu, A., Robitaille, P. M. L., Theoretical model for an MRI radio frequency resonator Bi omedical Engineering, IEEE Transactions on, 47)4_, 535 546 (2000).

PAGE 233

233 228 Everson G. T., Scherzinger, A., Berger Leff, N., Reichen, J., Lezotte, D., Manco Johnson, M., Gabow, P., Polycystic liver disease: Quantitation of parenchymal and cyst volumes from com puted tomography images and clinical correlates of hepatic cysts vol 8 (6), 1627 1634 (1988). 229 Everson, G. T., Taylor, M. R. G., Doctor, R. B., Polycystic disease of the liver vol 40 (4), 774 782 (2004). 230 Henkelman R.M., Stanisz, G. J., Kim, J. K., Bronskill, M. J., Anisotropy of NMR properties of tissues Magn Reson Med; 32: 592 601(1994). 231 Buckley D L Bui J D Phillips M I Zelles T Inglis B A Plant H D Blackband S J. The effect of ouabain on water diffusion in the rat hippocam pal slice measured by high resolution NMR imaging Magn Reson Med;41:137 142 (1999). 232 Bui J D Buckley D L Phillips M I Blackband S J ., Nuclear magnetic resonance imaging measurements of water diffusion in the perfused hippocampal slice during N methyl D aspartate induced excitotoxicity Neuroscience;93:487 490 (1999). 233 Forder J R Bui J D Buckley D L Blackband S J. MR imaging measurements of compartmental water diffusion in perfused heart slices Am J Physiol Heart Circ Physiol;281 :1280 1285 (2001). 234 Neeman M Jarrett K A Sillerud L O Freyer J P. Self diffusion of water in multicellular spheroids measured by magnetic resonance microimaging Cancer Res;51:4072 4079 (1991). 235 Pilatus U Shim H Artemov D Davis D van Zijl P Glickson J D. Intracellular volume and apparent diffusion constants of perfused cell cultures as measured by NMR Magn Reson Med;37:825 832 (1997). 236 Sotak, C.H., Nuclear magnetic resonance (NMR) measurement of the apparent diffusion c oefficient (ADC) of tissue water and its relationship to cell volume changes in pathological states Neurochem. Int. 45, 569 582 (2004). 237 Van Der Toorn, A., Sykova, E., Dijkhuizen, R.M., Vorisek, I., Vargova, L., Skobisova, E., Van Lookeren Campagne, M. Reese, T., Nicolay, K., Dynamic changes in water ADC, energy metabolism, extracellular space volume, and tortuosity in neonatal rat brain during global ischemia Magn. Reson. Med. 36, 52 60, (1996). 238 Benveniste, H., Hedlund, L.W., Johnson, G.A., Mecha nism of detection of acute cerebral ischemia in rats by diffusion weighted magnetic resonance microscopy .Stroke 23, 746 754, (1992).

PAGE 234

23 4 239 Hasegawa, Y., Latour, L.L., Formato, J.E., Sotak, C.H., Fisher, M., Spreading waves of a reduced diffusion coefficient of water in normal and ischemic Rat brain J. Cereb. Blood Flow Metab. 15, 179 187 (1995). 240 OShea, J.M.,Williams, S.R., vanBruggen, N., GardnerMedwin, A.R., Apparent diffusion coefficient and MR relaxation during osmoticmanipulation in isolated turtle c erebellum Magn. Reson. Med. 44, 427 432 (2000). 241 Dijkhuizen, R.M., deGraaf, R.A., Tulleken, K.A.F., Nicolay, K., Changes in the diffusion of water and intracellular metabolites after excitotoxic injury and global ischemia in neonatal rat brain J. Cere b. Blood Flow Metab. 19, 341 349 (1999). 242 Pfeuffer, J. Flogel, U. Dreher W., Leibfritz D. Restricted diffusion and exchange of intracellular water: theoretical modeling and diffusion time dependence of 1H NMR measurements on perfused glial cells N MR Biomed., 11(1), pp. 19 31 (1998). 243 Ludeke, K. M., Roschmann, P., Tischler, R., Susceptibility artefacts in NMR imaging Magn Res. Imaging. 3, 329 343 (1985). 244 Roemer, P.B., Edelstein, W. A., Hayes, C. E., Souza, S. P., Mueller, O. M., The NMR phase d array. Magnetic Resonance in Medicine, 1990. 16 (2): p. 192 225. 245 Blumer, P., Blumich, B., Botto, R. E., Fukushima, E.(ed), Spatially Resolved Magnetic Resonance (Weinheim: Wiley VCH) (1998). 246 Benveniste, H., blackband, S. J., Translational Neurosci ence and MR microscopy Lancet Neurol. 5(6), 536 544 (2005). 247 Callaghan, P. T., Principles of Nuclear Magnetic Resonance Microscopy Oxford University Press (1994). 248 Hoult, D.I. The NMR receiver: a description and analysis of design Progress in NMR Spectroscopy 1978; 12: 41 77. 249 Sun SW, Liang HF, Trinkaus K, Cross AH, Armstrong RC, Song SK. Noninvasive detection of cuprizone induced axonal damamge and demyelinationin the mouse corpus callosum Magn Reson Med 2006; 55: 302 308. 250 Kim JH, Bu dde MD, Liang HF, Klein RS, Russell JH, Cross AH, Song SK. Detecting axon damage in spinal cord from a mouse model of multiple sclerosis Neurobiol Dis 2006; 21: 626 632. 251 Budde MD, Kim JH, Liang HF, Russell JH, Cross AH, Song SK. Axonal inju ry detected by in vivo diffusion tensor imaging correlates with neurological disability in a mouse model of multiple sclerosis NMR Biomed 2008; 21: 589 597.

PAGE 235

235 252 Choi SJ, Lim KO, Monteiro I, Reisberg B. Diffusion tensor imaging of frontal white matter s tudy J Geriatr Psychiatry Neurol 2005; 18: 12 19. 253 Tuch DS. 2004. Q Ball Imaging Magnetic Resonance in Medicine 52: 1358 1372. 254 Hess CP, Mukherjee P, Han ET, Xu D, Vigneron DB. 2006. Q ball reconstruction of multimodal fiber orientations using the spherical harmonic basis Magnetic Resonance in Medicine 56: 104 117. 255 Tournier JD, Calamante F, Gadian DG, Connelly A. Direct estimation of the fiber orientation density function from di ffusion weighted MRI data using spherical deconvolution. Neuroimage 2004;23(3):1176 1185. 256 Anderson AW. 2005. Measurement of fiber orientation distributions using high angular resolution diffusion imaging Magnetic Resonance in Medicine 54: 1194 1206. 2 57 Kaden E, Knosche TR, Anwander A. 2007. Parametric spherical deconvolution: Inferring anatomical connectivity using diffusion MR imaging Neurolmage 37: 474 488. 258 Tuch, D.S., Reese, T.G., Wiegell, M.R., Makris, N., Belliveau, J.W., Wedeen, V.J., 2002. High angular resolution diffusion imaging reveals intravoxel white matter fiber heterogeneity Magn Reson Med 48, 577 582. 259 Frank, L.R., 2002. Characterization of anisotropy in high angular resolution diffusion weighted MRI Magn Reson Med 47, 1083 109 9. 260 Ozarslan, E., Mareci, T.H., 2003. Generalized diffusion tensor imaging and analytical relationships between diffusion tensor imaging and high angular resolution diffusion imaging Magn Reson Med 50, 955 965. 261 Ozarslan, E., Vemuri, B.C., Mareci, T .H., 2005. Generalized scalar measures for diffusion MRI using trace, variance, and entropy. Magn Reson Med 53, 866 876. 262 Alexander DC, Barker GJ, Arridge SR. Detection and modeling of non Gaussian apparent diffusion coefficient profiles in human brain data Magn Reson Med 2002; 48: 331 340. 263 Tuch DS, Reese TG, Wiegell MR, Wedeen VJ. Diffusion MRI of complex neural architecture Neuron 2003; 40: 885 895. 264 T. E. Conturo, R. C. McKinstry, E. Akbudak, and B. H. Robinson. Encoding of anisotropic diffus ion with tetrahedral gradients: A general mathematical diffusion formalism and experimental results Magn. Reson. Med., 35:399 412,1996.

PAGE 236

236 265 S. Mori, B. J. Crain, V. P. Chacko, and P. C. M. van Zijl. Three dimensional tracking of axonal projections in the brain by magnetic resonance imaging Ann. Neurol., 45:265 269, 1999. 266 Flint, J. J., Hansen, B., Fey, M., Schmidig, D., King, M. A., vestergaard Poulsen, P., Blackband, S. J., Cellular level diffusiontensor microscopy and fiber tracking in mammalian nerv ous tissue with direct histological correlation NeuroImage 52 (2010) 556 561.

PAGE 237

237 BIOGRAPHICAL SKETCH Choong Heon Lee was born to Tae Soo Lee and Jung Nim Hong in Gwangju, Republic of Korea. After graduation from Chonnam National University for his B.A. in 2 001, he attended rogram in the Department of Electrical and Computer Engineering. In the fall of 2004, he started his Ph.D. program in the Department of Electrical and Computer Engineering. After completion of his Ph .D. qualifying exam in the spring of 2004, he started MRI research in the Department of Neuroscience in the Department of Medicine in pursuit of applications of engineering to the medical research. He is a member of International Society of Magnetic Resonanc e in Medicine.