Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2012 Cellular Responses to Osmotic Perturbation: A High-Field ¹H and ²³Na Magnetic Resonance Microscopy Study John Walsh

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected] THE FLORIDA STATE UNIVERSITY

COLLEGE OF ENGINEERING

CELLULAR RESPONSES TO OSMOTIC PERTURBATION:

A HIGH-FIELD 1H AND 23Na MAGNETIC RESONANCE MICROSCOPY STUDY

By

JOHN WALSH

A Thesis submitted to the Department of Chemical and Biomedical Engineering in partial fulfillment of the requirements for the degree of Master of Science

Degree Awarded: Fall Semester, 2012 John Walsh defended this thesis on August 10, 2012.

The members of the supervisory committee were:

Samuel C. Grant Professor Directing Thesis

Teng Ma Committee Member

Jingjiao Guan Committee Member

The Graduate School has verified and approved the above-named committee members, and certifies that the thesis has been approved in accordance with university requirements.

ii

This thesis is dedicated to my family and friends who have continued to support me through this long process. I couldn’t have done it without you.

iii

ACKNOWLEDGEMENTS

I would first like to acknowledge the FAMU-FSU College of Engineering and the National High Magnetic Field Laboratory for use of their facilities to perform this work. Only with access to these high magnetic fields could a project of this nature have been undertaken. I would also like to acknowledge my supervising professor, Dr. Samuel C. Grant, as his knowledge of magnetic resonance imaging and RF coil design were critical in conducting the experiments and acquiring the high resolution images obtained. Additionally, my committee members Dr. Teng Ma and Dr. Jingjiao Guan provided much insight on the directions and results of this project. I would like to acknowledge Parastou Foroutan, a graduate student in Dr. Grant’s lab, who aided me with the initial sodium imaging and acquainted me with the techniques used throughout this project. Additionally, I would like to acknowledge Corey Falgas, another fellow graduate student, who aided in maintaining the animals used in this work. I would also like to thank my other colleagues Ihssan Masad, Jens Rosenberg, and Jose Muniz for their support, ideas, and help throughout the course of the project. Finally, funding for this project came from the American Heart Association (SE Division) and the FSU Department of Chemical and Biomedical Engineering. Additional funding came from the FSU Office of National Fellowships through an Undergraduate Research and Creative Activity Award. For all of these contributions, I am grateful.

iv

TABLE OF CONTENTS

List of Tables ...... vii List of Figures ...... ix Abstract ...... xiii 1. INTRODUCTION ...... 1 2. BACKGROUND AND THEORY ...... 4 2.1 Magnetic Resonance Imaging ...... 4 2.2 Cell Membranes and Osmosis ...... 12 2.3 Relation to Pathological Conditions ...... 14 2.4 Single Cell Models ...... 17 2.5 Tissue Models ...... 20 2.6 Diffusion Multi-Compartment Modeling...... 22 2.7 Sodium MRI ...... 24 3. OBJECTIVES ...... 28 3.1 High-Field MRI ...... 28 3.2 Radiofrequency (RF) Coil Design ...... 30 3.3 Neural Ganglia Model System ...... 32 3.4 Osmotic Perturbation ...... 33 3.5 Cell Viability ...... 34 4. METHODS ...... 36 4.1 Neural Ganglia Preparation ...... 36 4.2 Artificial Sea Water (ASW) Solutions ...... 38 4.3 MR Image Acquisition ...... 43 4.4 Statistical Analysis...... 50 5. RESULTS ...... 52 5.1 1H/23Na Imaging of Fixed Ganglia ...... 52 5.1.1 Isotonic Fixed Ganglia ...... 52 5.1.2 Hypertonic Fixed Ganglia ...... 60 5.1.3 Hypotonic Fixed Ganglia...... 67 5.1.4 Fixed Ganglia Summary ...... 74 5.2 1H/23Na Imaging of Viable and Nonviable Ganglia ...... 76 5.2.1 Isotonic Viable/Nonviable Ganglia ...... 76 5.2.2 Hypertonic Viable/Nonviable Ganglia ...... 85 5.2.3 Hypotonic Viable/Nonviable Ganglia ...... 93 5.2.4 Viable/Nonviable Ganglia Summary ...... 102 5.3 Comparison of Viable and Fixed Ganglia ...... 106

v

6. DISCUSSION ...... 111 7. CONCLUSIONS AND FUTURE DIRECTIONS ...... 118 References ...... 121 Biographical Sketch ...... 135

vi

LIST OF TABLES

1 Standard concentrations of key compounds in isotonic artificial sea water (ASW) ...... 38

2 Quantities of each component to make 1 L of ASW solutions of different tonicities ...... 42

1 3 Summary of MR acquisition parameters for the H spin echo T1 measurement ...... 45

1 4 Summary of MR acquisition parameters for the H spin echo T2 measurement ...... 46

1 5 Summary of MR acquisition parameters for the H gradient echo T2* measurement ...... 47

6 Summary of MR acquisition parameters for the 1H diffusion-weighted spin echo ADC measurement ...... 48

23 7 Summary of MR acquisition parameters for the Na 3D gradient echo T1 measurement ... 49

23 8 Summary of MR acquisition parameters for the Na 3D gradient echo T2* measurement ...... 50

9 Proton and sodium relaxation and diffusion measurements for the isotonic fixed ganglia .. 60

10 Proton and sodium relaxation and diffusion measurements for the hypertonic fixed ganglia ...... 67

11 Proton and sodium relaxation and diffusion measurements for the hypotonic fixed ganglia ...... 74

12 Summary of relaxation and diffusion trends in fixed ganglia at varying tonicities ...... 75

13 Proton relaxation and diffusion measurements for the isotonic viable ganglia ...... 84

14 Proton and sodium relaxation and diffusion measurements for the isotonic nonviable ganglia ...... 84

15 Proton relaxation and diffusion measurements for the hypertonic viable ganglia ...... 92

16 Proton and sodium relaxation and diffusion measurements for the hypertonic nonviable ganglia ...... 93

17 Proton relaxation and diffusion measurements for the hypotonic viable ganglia ...... 101

18 Proton and sodium relaxation and diffusion measurements for the hypertonic nonviable ganglia ...... 102

19 Summary of relaxation and diffusion trends in viable ganglia at varying tonicities ...... 104 vii

20 Summary of relaxation and diffusion trends in nonviable ganglia at varying tonicities .... 105

viii

LIST OF FIGURES

1 The orientation of magnetic moments when placed in an external magnetic field ...... 5

2 Diagram of a standard spin echo pulse sequence ...... 8

3 Representation of the net magnetization during a spin echo pulse sequence ...... 9

4 Diagram of a standard diffusion-weighted spin echo pulse sequence ...... 11

5 The fluid mosaic model of the cell membrane ...... 12

6 Mechanism of the sodium-potassium ATPase ...... 13

7 Mechanism of cell death during ischemia and stroke ...... 15

8 Diagram of the mechanism of osmotic redistributions that result due to ischemia and stroke ...... 16

9 Dissection process for isolating the abdominal ganglion from Aplysia ...... 37

10 Image of three viable isolated abdominal ganglia from Aplysia after dissection ...... 37

11 Cellular response of placing neuronal ganglia in extracellular environments of different tonicities ...... 43

12 Image of the solenoidal transmit-receive coil showing both resonant circuits for simultaneous tuning to both the proton and sodium resonant frequencies ...... 44

1 13 Representative T1-weighted H MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T ...... 54

1 14 Representative T2-weighted H MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T ...... 55

1 15 Representative T2*-weighted H MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T ...... 56

16 Representative diffusion-weighted 1H MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T ...... 57

17 Representative 23Na MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T ...... 58

18 Representative regression curve fits for the isotonic fixed ganglia ...... 59

ix

1 19 Representative T1-weighted H MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T ...... 61

1 20 Representative T2-weighted H MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T ...... 62

1 21 Representative T2*-weighted H MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T ...... 63

22 Representative diffusion-weighted 1H MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T ...... 64

23 Representative 23Na MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T ...... 65

24 Representative regression curve fits for the hypertonic fixed ganglia ...... 66

1 25 Representative T1-weighted H MR images of three fixed ganglia in hypotonic ASW acquired at 11.7 T ...... 68

1 26 Representative T2-weighted H MR images of three fixed ganglia in hypotonic ASW acquired at 11.7 T ...... 69

1 27 Representative T2*-weighted H MR images of three fixed ganglia in hypotonic ASW acquired at 11.7 T ...... 70

28 Representative diffusion-weighted 1H MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T ...... 71

29 Representative 23Na MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T ...... 72

30 Representative regression curve fits for the hypotonic fixed ganglia ...... 73

1 31 Representative T1-weighted H MR images of three viable ganglia in isotonic ASW acquired at 11.7 T ...... 78

1 32 Representative T2-weighted H MR images of three viable ganglia in isotonic ASW acquired at 11.7 T ...... 79

1 33 Representative T2*-weighted H MR images of three viable ganglia in isotonic ASW acquired at 11.7 T ...... 80

34 Representative diffusion-weighted 1H MR images of three viable ganglia in isotonic ASW acquired at 11.7 T ...... 81

x

35 Representative 23Na MR images of three nonviable ganglia in isotonic ASW acquired at 11.7 T ...... 82

36 Representative regression curve fits for the isotonic viable ganglia...... 83

1 37 Representative T1-weighted H MR images of three viable ganglia in hypertonic ASW acquired at 11.7 T ...... 86

1 38 Representative T2-weighted H MR images of three viable ganglia in hypertonic ASW acquired at 11.7 T ...... 87

1 39 Representative T2*-weighted H MR images of three viable ganglia in hypertonic ASW acquired at 11.7 T ...... 88

40 Representative diffusion-weighted 1H MR images of three viable ganglia in hypertonic ASW acquired at 11.7 T ...... 89

41 Representative 23Na MR images of three nonviable ganglia in hypertonic ASW acquired at 11.7 T ...... 90

42 Representative regression curve fits for the hypertonic viable ganglia ...... 91

1 43 Representative T1-weighted H MR images of three viable ganglia in hypotonic ASW acquired at 11.7 T ...... 95

1 44 Representative T2-weighted H MR images of three viable ganglia in hypotonic ASW acquired at 11.7 T ...... 96

1 45 Representative T2*-weighted H MR images of three viable ganglia in hypotonic ASW acquired at 11.7 T ...... 97

46 Representative diffusion-weighted 1H MR images of three viable ganglia in hypotonic ASW acquired at 11.7 T ...... 98

47 Representative 23Na MR images of three nonviable ganglia in hypotonic ASW acquired at 11.7 T ...... 99

48 Representative regression curve fits for the hypotonic viable ganglia ...... 100

49 Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 1 hypertonic extracellular tonicities for the H T1 measurement ...... 107

50 Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 1 hypertonic extracellular tonicities for the H T2 measurement ...... 107

xi

51 Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 1 hypertonic extracellular tonicities for the H T2* measurement ...... 108

52 Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and hypertonic extracellular tonicities for the 1H ADC measurement ...... 108

53 Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 23 hypertonic extracellular tonicities for the Na T1 measurement ...... 109

54 Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 23 hypertonic extracellular tonicities for the Na T2* measurement ...... 109

55 Comparisons of 23Na SNR in nonviable ganglia between hypotonic, isotonic, and hypertonic extracellular tonicities ...... 110

56 23Na images and signal intensity of a single viable ganglion in isotonic, hypertonic, and hypotonic ASW ...... 110

xii

ABSTRACT

Magnetic resonance imaging (MRI) relaxation and diffusion properties are sensitive to the physiological state of cells and tissues. In particular, cerebral ischemia and stroke is associated with a reduced apparent diffusion coefficient (ADC) and increased transverse 1 magnetization relaxation parameter (T2) in proton ( H) MRI making it useful for clinical diagnosis. Further, increases in sodium (23Na) signal intensity in 23Na MRI correlate with several neurodegenerative diseases as well as ischemia and stroke, which can be used to identify stroke lesions and predict stroke onset time. However, the contributing mechanisms underlying these 23 changes, including possible increases in intracellular sodium and/or lengthened Na T2 relaxation, are not well understood. It has been hypothesized that alterations in cell regulatory mechanisms result in cell swelling, which disrupts tissue microstructure and ionic distributions. Osmotic perturbations have been used on single neurons and neural tissue models to mimic these volume changes. Evaluated with 1H MRI, hypotonic perturbations mimic tissue ischemia and result in an increase in 1H signal intensity, a decrease in 1H ADC, and an increase 1 in H T2, while all trends are reversed for hypertonic perturbation. Single cell work has focused on the large L7 neuron from the abdominal ganglion in the sea hare Aplysia californica, which has been shown to have distinct nuclear and cytoplasmic compartments with differing relaxation and diffusion properties. Therefore, it would be useful to develop a tissue model comprised of large cells in which the intracellular contribution to the volume averaged signal could be determined. In this study, Aplysia abdominal ganglia were used due to their simple anatomy with a small collection of relatively large neurons up to 300 microns in diameter representing a simple neural tissue model system. The abdominal ganglia were dissected from the living animal, washed with isotonic artificial sea water (ASW; in mM: 460 NaCl, 10.4 KCl, 55 MgCl2, 11 CaCl2, 15 HEPES) and loaded into a 2.5-mm o.d. capillary containing isotonic, hypertonic, or hypotonic ASW. Hypertonic and hypotonic perturbations were introduced by changing the isotonic sodium chloride concentration (460 mM) to 545 mM and 345 mM, respectively. All MR imaging was performed at 11.75 T utilizing a homebuilt, double-tuned 1H/23Na solenoidal coil having a diameter of 3 mm. Imaging was performed in separate studies to quantify the proton T1,

T2, T2*, and ADC immediately following dissection on viable ganglia as well as sodium T1 and

T2* after the loss of cell viability. Changes in the MRI relaxation parameters (T1, T2 and T2*) as xiii well as the apparent diffusion coefficient (ADC) in isolated ganglia were used to assess changing cellular environments under osmotic perturbation in the context of the influence changing ionic distributions, notably sodium, have on cell swelling in disease states. With tonicity changes, the ganglionic sac remains intact although alterations in the signal intensity (particularly with respect to sodium) are evident. Proton T1 did not change significantly 1 1 with osmotic perturbation; however, H T2 and T2* decreased with increasing tonicity and H

T2* provided the most robust measurement for identifying cellular changes due to osmotic perturbation. All relaxation parameters increased with cell death and loss of viability and no changes in 1H ADC were observed. The trends observed in 1H relaxation are possibly due to an increased intracellular volume fraction while the lack of a clear trend in ADC values may be due to significant volume averaging between the nuclear, cytoplasmic, and interstitial environments and should be modeled utilizing a multi-compartmental approach. Sodium measurements reveal a general increasing trend in relaxation with osmotic perturbation but no clear conclusions could be drawn. Changes in sodium relaxation are evident in compromised ganglia; however, linearly increasing trends in sodium signal intensity are seen in all osmotic conditions and may represent changes occurring with loss of cell viability, providing evidence that the direct evaluation of ions may prove more sensitivity in assessing pathological osmotic changes. Future work with perfused ganglia as well as voxel-selective spectroscopy and contrast agents to isolate the intracellular sodium signal could be used to answer questions regarding these observed trends in sodium signal intensity and relaxation.

xiv

CHAPTER ONE

INTRODUCTION

All living cells maintain a delicate homeostatic balance of intracellular and extracellular osmolytes which drives transport across the cell membrane and is critical in regulating cell volume (Lang, et al. 1998). Ions, such as sodium, potassium, and chloride, as well as cellular proteins, maintain these osmotic gradients. In fact, a large percentage of the cell’s energy supply is expended in regulating cell volume by actively transporting sodium and other ions across the cell membrane to influence osmosis, the passive diffusion of water to regions of higher osmolyte concentration (Faller, 2008). Transport between the intracellular and extracellular environments is critical to the nourishment and health of cells. Additionally, ion distributions across the membrane create electrochemical gradients which act as the basis of signal transduction in neurons and myocytes. Often, external perturbations, both osmotic and neurotoxic, interfere with these transport systems resulting in cell volume and osmoregulation imbalances (Carpenter, et al. 1992). This loss of cellular homeostasis becomes apparent in pathological disease states such as stroke, cancer, and chronic neurodegeneration. Specifically, protein synthesis and degradation is altered in states of osmotic stress leading to protein distribution alterations that are found in many common neurodegenerative diseases such as Alzheimer’s, Parkinson’s and Huntington’s disease (Mao, et al. 2008). Further, the osmotic effects of water redistributions and changing ionic distributions are evident in states of tissue ischemia and acute stroke (Young, et al. 1987). Intricate osmoregulation pathways exist in organisms, including humans, that respond to states of osmotic changes in the extracellular environment through feedback mechanisms in an attempt to maintain cellular homeostasis (Bourque, 2008). Therefore, methods to quantify tissue microstructure, changes in cell volume, and the tissue sodium concentration in these pathological states are needed and have led to the incorporation of magnetic resonance microscopy as a tool to noninvasively evaluate changing tissue microstructure at sub-millimeter resolutions. Further, MR microscopy provides the added benefit of sensitivity to physiological function which, through different contrast mechanisms such as relaxation and diffusion weighting, can characterize the state and distribution of cellular water as well as relate restricted water diffusion to changing tissue microstructures in vivo (Beneviste & Blackband, 2002). Expanding the MR techniques to non-proton nuclei such as

1 sodium (23Na) allows the tissue sodium concentration and ionic distributions to be measured in abnormal tissues, providing another biomarker for many pathological conditions (Christensen, et al. 1996). By measuring intracellular and extracellular sodium and monitoring the rate of transport across the neural membrane through cell volume changes, a wealth of information about osmoregulation can be investigated in light of chemical and biological challenges to cell viability and function (Pasantes-Morales & Tuz, 2006). In addition, differences in microstructure and ion and protein concentration between the nucleus and cytoplasm are vital to intracellular equilibrium and provide a basis in understanding cellular responses to osmotic perturbation which may be expanded to the analysis of small collections of cells and tissues (Hsu, et al. 1996). Magnetic resonance imaging (MRI) provides a powerful tool for measuring sodium and water distributions in vivo. The correlation of MRI relaxation parameters and diffusion coefficients with the cellular response of changing intracellular volume fractions under osmotic (O’Shea, et al. 2000) and neurotoxic (Buckley, et al. 1998) perturbation allows the assessment of physiological conditions in disease states and is a direct indicator of changes in cell structure and water distribution involved in cell volume regulation. In the clinical setting, stroke and cerebral ischemia are associated with a reduced apparent diffusion coefficient (ADC) and increased 1 transverse magnetization relaxation parameter (T2) in proton ( H) MRI making it useful for diagnosis (Baird & Warach, 1998). Although usually applied to 1H imaging, the technique can be advanced to include 23Na MRI, which facilitates the direct measurement of sodium distributions in tissues, allowing the role of sodium concentration gradients and the direct maintenance of ionic sodium distributions in the regulation of cell volume to be examined in pathological conditions such as cerebral ischemia (Lin, et al. 2001). 23Na MRI can also be used to identify stroke lesions, but unlike 1H MRI, the sodium signal intensity increases in a linear fashion making it possible to predict stroke onset time which allows more appropriate treatment options to be pursued (Jones, et al. 2006). Further, increases in sodium signal intensity are evident in brain tumors and recently demonstrated in chronic neurodegeneration, opening the field of 23Na MRI into the investigation of a variety of pathologies (Thulborn, et al. 1999; Mellon, et al. 2009; Inglese, et al. 2010). MR diffusion and relaxation, as well as sodium signal intensity, measurements are aimed at examining differences in parameter values between normal and perturbed conditions in the context of understanding cells in diseased states.

2

The typical MRI relaxation parameters, T1, T2, and T2*, and the apparent diffusion coefficient (ADC) from diffusion-weighted magnetic resonance imaging are used to quantitatively assess changes in tissue microstructure and water distributions. Further, 23Na MRI provides physiological information on ionic concentrations and distributions within different tissue regions. In particular, changes in sodium signal and relaxation show a faster response to perturbation, prior to morphological change at the tissue level, indicating the benefit of this imaging modality in the early detection of pathological conditions (Schepkin, et al. 2011). Further insight can be gained through the incorporation of sodium MRI as a tissue viability marker and illustrating how these techniques can be expanded and incorporated into clinical image settings (Hussain, et al. 2009). 1 This study will be aimed at examining differences in H relaxation (T1, T2, and T2*) and 23 diffusion (ADC) as well as Na relaxation (T1 and T2*) in the abdominal neural ganglion from the sea hare Aplysia californica. As a small collection of relatively large cells, the abdominal ganglion will serve as a model system previously not characterized with MR imaging techniques. Further, as a majority of the work performed on single cells using MR involves the Aplysia L7 neuron from the abdominal ganglion, the characterization of the entire ganglion will provide insight between previous single cell studies and the more complex tissue models that have been explored. Although the measured relaxation and diffusion properties are an average of the response of the nucleus, cytoplasm, interstitial space, and extracellular medium, comparisons of cellular responses between different extracellular tonicities can provide insight into the collective ganglia response to osmotic stress and the nature of the single cell response within the tissue environment. Further, by imaging fixed, viable, and nonviable ganglia, MR changes that are characteristic of determining cell viability can be described. The results are presented in the context of understanding the mechanism of the MR changes observed in tissue ischemia and chronic neurodegeneration models.

3

CHAPTER TWO

BACKGROUND AND THEORY

2.1 Magnetic Resonance Imaging

Magnetic Resonance Imaging (MRI) is a non-ionizing medical imaging tool utilized for its tissue contrast and three-dimensional imaging capabilities. The use of MRI in the field of medical imaging was not introduced until the 1970s when it was shown that normal and abnormal tissue behaved differently when placed in a magnetic field. The increase in NMR relaxation rates (T1 and T2) in tumors as compared to normal tissues showed that this principle could be used to differentiate between tissues in disease states (Damadian, 1971). Since that time, MRI has revolutionized the medical community and become a prominent technique in the detection and diagnosis of disease. Although an insensitive imaging modality limited in spatial resolution on the order of 1-3 mm with a temporal resolution on the order of 10 minutes in clinical applications, the use of higher magnetic fields and advanced radiofrequency coil design has allowed the realm of magnetic resonance imaging to expand to resolutions on the order of tens of microns. Early from the conception of MRI, its use as a microscopic tool was recognized and currently resolutions less than 100 µm are feasible in research settings, allowing MRI to be termed magnetic resonance microscopy (Blackband, et al. 1999). The use of high-resolution imaging has allowed the cellular components of isolated tissues and single cells to be characterized through MRI relaxation and diffusion measurements. The principles of nuclear magnetism govern MRI, such that atomic nuclei with an odd atomic mass or atomic number, which results from the number and distribution of protons and neutrons in the nucleus, have a nonzero spin quantum number (denoted I). The nuclei are charged and, due to the nonzero spin, have a net angular momentum that causes them to act like magnets. Therefore, atomic nuclei which have a net nuclear spin are sensitive to magnetic resonance imaging techniques (Webb, 2003). Although clinically applied to hydrogen atoms in water molecules with a spin of 1/2 (I = 1/2), MRI can be expanded to include quadrupolar nuclei such as sodium which has a spin of 3/2 (I = 3/2). With a weaker signal emission due to its chemical nature and lower cellular concentration, sodium is much more difficult to image (Lin, et al. 2001).

4

The magnetic moment (denoted µ) of these nuclei can be related to the spin quantum number such that the magnetic moment is quantized and can only take certain discrete values as seen in equation (1) where h is Planck’s constant and γ is the gyromagnetic ratio, the ratio of the nuclear magnetic moment to the angular momentum.

(1) ⁄ | | [ ]

When placed in an external magnetic field, the nuclei align with the field; however, although the nuclei preferentially align in a low energy state parallel to the direction of the external magnetic field, there is a distribution between the low energy parallel state and the higher energy anti- parallel state as seen in Figure 1. Although illustrated as being directly aligned with the field, the magnetic moments are only partially aligned as they can only have discrete values. Also, the split into two energy levels only applies for spin 1/2 nuclei, corresponding to the nuclear magnetic quantum number which can have values of 1/2 and -1/2. For spin 3/2 nuclei, four energy levels are allowed and the magnetic moments will be distributed between the energy levels corresponding to magnetic quantum numbers of 3/2, 1/2, -1/2, and -3/2.

Figure 1. The orientation of magnetic moments when placed in an external magnetic field. Left: Without a magnetic field, the magnetic moments are random. Right: With the application of a magnetic field, the magnetic moments are aligned in a parallel or anti-parallel state resulting in a net magnetization (van Geuns, et al. 1999).

5

Each atomic nucleus has an energy related to its inherent magnet moment, gyromagnetic ratio, and the strength of the external magnetic field. For example, the proton nucleus has an energy difference between the parallel and antiparallel orientations shown in equation (2) where B0 is the strength of the externally applied magnetic field.

(2)

Based on the energy differences at a given temperature (T), there is a given distribution of spins between the parallel and antiparallel orientations as shown in equation (3) where k is the Boltzmann constant. The difference is small and represents the relative insensitivity of the MR imaging technique as it can only detect the differences between these states.

(3)

Upon being aligned with the magnetic field, each nucleus spins about the axis of the main magnetic field at a resonant frequency, often called the Larmor frequency, which is characteristic of each atomic nucleus and the external magnetic field strength as shown in equation (4). The gyromagnetic ratio (γ/2π) for the proton nucleus is 42.58 MHz/T and for the sodium nucleus is 11.26 MHz/T.

(4)

For NMR applications, the orientations of the nuclear magnetic moments are altered. Upon excitation with a radio frequency (RF) pulse at the resonant frequency, the spins absorb energy and are reversed into the high energy anti-parallel state. At the completion of the RF pulse, the net magnetization begins to return to the equilibrium state. After excitation, the nuclear spins relax to their previous thermal equilibrium state such that nuclear spins that were flipped into the anti-parallel state return to the parallel state, releasing energy that is detected as an induced voltage in a radio-frequency receiver coil. The Bloch equations represent the 6 mathematical formulation of the relaxation of the net magnetization after being altered from the equilibrium state in the x direction (5), y direction (6), and z direction (8). By definition, the z direction is aligned with the externally applied magnetic field. Before excitation, the net magnetization is only in the z direction; however, to obtain any MR signal, the net magnetization must be rotated into the x-y plane.

(5)

( )

(6)

( )

(7)

In particular, several characteristic relaxation parameters are used to assess the return of the net magnetization to equilibrium. The T1, or spin-lattice, relaxation parameter is the longitudinal relaxation in the direction of the externally applied magnetic field. The T1 is directly related to the correlation time and represents the thermal exchange of energy between the nuclear spins and the surrounding environment. The T1 relaxation parameter can be detected through a spin-echo pulse sequence that introduces a 90 degree RF pulse, rotating the net magnetization along the z axis into the x-y plane and observing the time response to the initial state prior to excitation (Kuhn, 1990). A typical spin echo pulse sequence is illustrated in Figure 2. The repetition times (TR) govern the amount of time the longitudinal magnetization has to relax back to equilibrium values before the next excitation. The repetition time, TR, or time between successive pulse sequences is varied and plotted against the mean signal detected in an exponential rise to max equation that takes the form of equation (8).

(8)

( )

7

The T1 relaxation parameter is also a measure of viscosity as more viscous substances have shorter T1 relaxation times as there is a more effective transfer of energy between nuclei and their environment.

Figure 2. Diagram of a standard spin echo pulse sequence. The echo time (TE) is the time between the 90˚ excitation pulse and the image acquisition. The repetition time (TR) is the time between successive 90˚ excitation pulses. The phase and frequency encoding gradients spatially encode the two-dimensional image within the slice encoded by the slice-selecting gradient (van Geuns, et al. 1999).

The T2, or spin-spin, relaxation parameter is the transverse relaxation in the direction perpendicular to the applied magnetic field and is due to a loss of phase coherence of oscillating nuclei. Also measured through a spin echo pulse sequence, inhomogeneities in the main magnetic field cause the transverse magnetization to disperse. Through the introduction of a 90 degree pulse followed by a 180 degree pulse, the net transverse magnetization is realigned into a characteristic echo response which can be detected as illustrated in Figure 3 (Kuhn, 1990). The echo time, TE, or time between initial excitation and signal measurement is varied and plotted against the mean signal in an exponentially decreasing function that takes the form of the equation (9) (Webb, 2003).

(9)

8

Figure 3. Representation of the net magnetization during a spin echo pulse sequence. The coordinate system shown is rotating at the resonant frequency of the nuclei. A) Application of the 90˚ pulse flips the magnetization into the transverse plane. B) The spins begin to diphase due to magnetic field inhomogeneities. C) Application of the 180˚ pulse flips the magnetization within the transverse plane and dephasing continues. D) At the echo time (TE) the spins rephase producing the spin echo (Kuhn, 1990).

Based on T2 relaxation, the T2* relaxation parameter is also the transverse relaxation but differs from T2 based on sample inhomogeneities and magnetic susceptibilities. Particularly in samples that are composed of regions that have different magnetic susceptibilities, such as in tissues, there are greater signal losses at the boundaries and is especially pronounced at interfaces with air, such as an air bubble located within the sample. Additionally, the T2* relaxation parameter is determined in a very different manner. Using a gradient echo pulse sequence, the characteristic echo response is generated not by using a 180 degree refocusing pulse, but rather through the application of a bipolar gradient that cancels phase differences that result from sample inhomogeneities and variations in the externally applied magnetic field. Again, the echo

9 time, TE, or time between initial excitation and signal measurement is varied and plotted against the mean signal in an exponentially decreasing function that takes the form of the equation (10).

(10)

The apparent diffusion coefficient (ADC) is a measure of nuclear movements within a sample. Based on the spin echo pulse sequence used for the determination of T1 and T2, an additional diffusion encoding gradient is introduced before and after the 180 degree pulse as illustrated in Figure 4. Nuclei that move within this time period are not seen in the resulting image which can be used to differentiate between diffusing and non-diffusing nuclei. The strength of the diffusion gradient, also termed the b values, is varied and an exponentially decreasing function results, as shown in equation (11) where the b values are defined in equation (12) and the ADC is the apparent diffusion coefficient (Minati &Weglarz, 2007). The strength (G), duration (δ) and delay (Δ) correspond to the pulse sequence in Figure 4.

(11)

(12)

( )

Diffusion measurements in inhomogeneous environments, such as tissues, are complicated by barriers to diffusion that impact the ADC. The theoretical models for how these factors (restrictions to due cellular membranes, diffusion anisotropy, diffusion times, and impact of specific experimental setup) contribute to diffusion measurements have been explored in some detail (Szafer et al. 1995; Yablonskiy & Sukstanskii, 2010). Factors such as a cell size distribution (Jespersen, et al. 2005) were shown to not significantly impact diffusion measurements while increases in cellularity were shown to decrease the ADC (Matsumoto, et al. 2009). Further, temperature impacts both relaxation (Kamman, et al. 1988) and diffusion (Matsumoto, et al. 2009) measurements. Frequently diffusion measurements result in biexponential signal decay resulting from two different populations of water nuclei (Minati &

10

Weglarz, 2007) such that exchange must be considered (Latt, et al. 2009). However, the biexponential behavior cannot be directly related to compartmentation found in tissues, and although the model fits experimental data well, it is not founded on physical barriers to diffusion (Kiselev & Il’yasov, 2007; Schwarz, et al. 2004; Sehy, et al. 2002). Further, the diffusion coefficient measured in MRI differs from the intrinsic dynamic ADC characterized from mean- square displacements (Grebenkov, 2010).

Figure 4. Diagram of a standard diffusion-weighted spin echo pulse sequence. The echo time (TE) remains the time between the end of the 90˚ excitation pulse and the image acquisition. The application of the first gradient encodes the spins with a phase. The second gradient rephases the spins. However, if diffusion has occurred the spins will only be partially rephased and image signal will be lost. The characteristic diffusion encoding parameters are the strength of the gradient pulse (G), duration of the pulse (δ) and diffusion time (Δ) (Sotak, 2004).

Through the variation of the given parameters (TR, TE and b) and linear regression curve fitting, the relaxation parameters and apparent diffusion coefficient can be determined through NMR techniques. In order to generate an MR image, the same principles are utilized, although applied in the presence of magnetic field gradients. Gradients are used in each of the three directions to isolate individual voxels with a given resonant frequency and phase. A slice is first selected by applying a slice-selection gradient during the initial excitation in the direction corresponding to the desired slice orientation. The remaining two directions are specified by a frequency-encoding gradient and phase-encoding gradient. The phase is encoded prior to data

11 acquisition while the frequency is encoded during data acquisition. By identifying each voxel with a specific frequency and phase, an image can be generated.

2.2 Cell Membranes and Osmosis

Cell membranes provide a selectively permeable boundary which regulates exchange between the intracellular and the extracellular environment. Modeled as a lipid bilayer, as illustrated in Figure 5, cell membranes do not facilitate the bulk transport of charged nuclei, such as sodium, or cellular proteins across the membrane. The composition and balance of intracellular and extracellular ions and osmolytes is critical in regulating osmosis (Lang, et al. 1998). Osmosis is the entropic process driven by water transport from regions where the water matrix structure is highly ordered owing to small concentrations of osmolytes to regions where this structure is more disordered due to higher concentrations of osmolytes. Gradients established by these concentration differences are the driving forces of transport across the cell membrane and the basis of signal transduction in excitable cells, such as neurons.

Figure 5. The fluid mosaic model of the cell membrane. The diagram illustrates the lipid bilayer integrated with transmembrane proteins for cellular transport (Singer, et al. 1972).

Neurons in particular are sensitive to the transport of sodium, potassium, and chloride ions across the membrane. These ions are the most abundant and play the largest role in establishing osmotic gradients within the cell. For excitable cells, the ionic distributions in the cell affect the electrochemical gradient established by the neuron, which may interfere with transport across the membrane leading to cell swelling or crenation (Pasantes-Morales & Tuz,

12

2006). Although most membranes are freely permeable to water through aquaporin channels, diffusion of sodium and other ions is limited to leakage channels found within the membrane. However, in living cells, the sodium-potassium pump actively transports sodium and potassium, maintaining osmotic gradients across the membrane and making possible the active processes of life. The sodium-potassium pump uses ATP as an energy source to actively transport three sodium ions out of the cell while simultaneously transporting two potassium ions into the intracellular environment using a mechanism illustrated in Figure 6. This transport process results in a low intracellular sodium concentration and high intracellular potassium concentration. Disruption of this mechanism through blocking the sodium potassium pump by removing all potassium from the extracellular environment (or using neurotoxins) results in drastic increases in neuronal volume in neurons from the sea slug Aplysia californica (Carpenter, et al. 1992). Similar increases in cell volume were found by changing the extracellular osmolarity, providing evidence that alterations in extracellular tonicity can be used to alter cell volume and mimic cellular changes seen in pathological conditions.

Figure 6. Mechanism of the sodium-potassium ATPase. Utilizing a majority of cellular energy in neurons through ATP and a process of dephosphorylation, the Na+/K+ pump actively transports three Na+ ions out of the cell and two K+ ions into the cell to maintain osmotic gradients, regulate cell volume, and maintain or alter the cell membrane potential to act as a + + signal transducer. The two configurations of the protein for binding Na or K are denoted E1 and E2 (Faller, 2008).

13

Although cells are usually surrounded by extracellular fluid of nearly constant osmolarity, changes in the extracellular osmolarity occur in diseased cells. The effect of differing extracellular osmolarities on mouse embryonic stem cells resulted in changes in confluence as well as cellular size (Mao et al, 2008). When isotonicity is lost, the confluence of the culture is greatly reduced as many cells lose viability. In surviving cells, cell volume changes occur as hypotonic osmotic perturbations result in cell swelling and hypertonic perturbation results in crenation. Further, changes in intracellular ion or protein concentrations can osmotically draw water into the intracellular environment, resulting in cell swelling. In molluscs, prolonged exposure to hypotonic perturbation results in decreased levels of organic osmolytes to balance the intracellular osmolarity (Silva & Wright, 1994). These regulatory mechanisms can result in either cell volume increases or decreases based on the tonicity of the extracellular environment to maintain cellular homeostasis (Strange, 2004).

2.3 Relation to Pathological Conditions

In particular, early studies have shown that water and sodium concentrations increase while potassium concentrations decrease in states of tissue ischemia with a mechanism is not fully understood (Young, et al. 1987). Assessed by atomic absorption spectroscopy, these changes are immediately evident, although continue to change with time, and could only be measured in nonliving tissues. In the mammalian brain, complex osmoregulatory circuits precisely control the intracellular and extracellular osmolarity and are maintained through feedback loops sensitive to the extracellular tonicity (Bourque, 2005). Although complex, these feedback mechanisms are in place to maintain cell volume and regulate ion and protein distributions in the intracellular and extracellular environments. However, disruptions of these transport processes are seen in pathological conditions such as tissue ischemia and stroke. In particular, ischemia leads to failure of all active transport processes as ATP becomes depleted leading to failure of the sodium-potassium pump and a corresponding increase in intracellular sodium (Lipton, 1999). Current understandings of the mechanisms responsible for cell death are summarized in Figure 7. Without ATP, the electron transport chain fails and an increase in glycolysis to produce ATP increases, leading to an increase in free radicals and changes in enzyme activity. This is further exacerbated by a decrease in protein synthesis which

14 only adds to the osmotic imbalances caused by ion transport failures. Focusing on ion transport failures with ischemia, Figure 8 illustrates the changes expected within the intracellular and extracellular environment due to ischemia. The increase in intracellular sodium depolarizes the neuron which leads to intracellular calcium accumulation (Pasantes-Morales & Tuz, 2006). As the interior of the cell becomes more positively charged and osmotic pressure increases, water and chloride ions flow into the cell causing the cells to swell. These increases in cell volume alter the tissue microstructure and can lead to cell death. Additionally, neurodegeneration is also thought to change protein synthesis and metabolic processes leading to osmotic changes in neurons and ultimately to neuronal death leading to brain atrophy.

Figure 7. Mechanism of cell death during ischemia and stroke. A lack of oxygen results in a lack of aerobic respiration leading to decreased ATP and a disruption of active processes. Notably, the sodium-potassium pump fails to maintain ionic distributions resulting in changing ionic distributions. Additional changes in protein synthesis, electron transport, and phospholipase activity disrupt many cellular physiological processes (Lipton, 1999).

15

Figure 8. Diagram of the mechanism of osmotic redistributions that result due to ischemia and stroke. Failure of the sodium-potassium pump results in increased intracellular sodium and depolarization of the cell leading to increased intracellular calcium. To preserve charge balance, chloride ions leak into the cell along with water resulting in cell swelling (Pasantes-Morales & Tuz, 2006).

Magnetic resonance imaging has been utilized to identify osmoregulation changes that occur in stroke (Baird & Warach, 1998; Merino & Warach, 2010). First studied in cats, MR images showed that ischemic stroke could be identified as early as six and a half hours after onset and had clinical implications (Buonanno, et al. 1982). Relaxation measurements, specifically increases in T2 which resulted in an increase in image signal intensity, were used as the basis for identifying ischemic stroke. However, the first diffusion-weighted MR images were shown to provide evidence of stroke at much earlier time points than conventional T2-weighted imaging in cats (Moseley, et al. 1990). One hour after the onset of ischemia, T2-weighted MR images appear normal, but significant increases in the signal intensity in diffusion-weighted MRI due to a reduced ADC were observed, indicating that diffusion-weighted imaging was more sensitive in identifying stroke. Through the combination of relaxation, diffusion, and perfusion MR measurements, tissue ischemia and stroke can be identified in rats with mismatches between the diffusion, perfusion, and relaxation infarct size (Hoehn-Berlage, et al. 1995; Calamante, et al. 1999; Beneviste, et al. 1992; van Dorsten, et al. 2002). The region of damage is underestimated

16 by relaxation measurements at early time points while perfusion imaging does not necessarily indicate salvageable tissue. Although first performed on a variety of animal models (Hoehn, et al. 2001), MR imaging is noninvasive and through relaxation and diffusion measurements, the onset time of stroke in the clinical environment can be estimated to allow better patient treatment (Petkova, et al. 2010; Jokivarsi, et al. 2010). Additionally, both cancerous tumors and neurodegeneration also exhibit drastic changes in cellular osmoregulation and cell viability. Similar MR relaxation and diffusion measurements are useful in identifying brain tumors (Naruse, et al. 1987) and identifying tumor changes with treatment (Chenevert, et al. 1997). In mouse models of Alzheimer’s disease, reductions in T2 were observed in several brain regions (Helpern, et al. 2004). Further, T2*-weighted MRI has been used to identify plaques in human tissue associated with Alzheimer’s disease (Beneviste, et al. 1999) and standard MR imaging protocols can detect reductions in hippocampal size that are indicative of neurodegeneration. Although the different pathological conditions can be identified, the cellular changes that result in different contrast in MR images are not fully understood. MRI typically looks at water in the human body and therefore changes in cellular water distributions are likely responsible; however, water distributions are heterogeneous and water diffusion is highly restricted leading to difficulties in understanding MR relaxation and diffusion measurements. However, MR diffusion decreases have been correlated with reductions in activity of the sodium-potassium pump resulting in increased intracellular sodium, decreased intracellular potassium and cell swelling (Mintorovitch, et al. 1992). Similar diffusion decreases also correlate with decreased extracellular volume fraction (increased intracellular volume fraction due to cell swelling) and increases in extracellular tortuosity (van der Toorn, et al. 1996).

2.4 Single Cell Models

As MRI hardware continues to improve and higher magnetic field strengths are achieved, higher resolution images become possible due to an improved signal-to-noise ratio. With resolutions below 100 µm, MR microscopy has been developed using small RF microcoils along with small samples to look at individual isolated cells (Ciobanu & Pennington, 2004). Initial efforts were put forth using the toad Xenopus laevis oocyte due to its large size (~1 mm) and unique regions including the nucleus, animal pole, and vegetal pole which could be clearly

17 differentiated (Aguayo, et al. 1986; Mallard, 1986). Increased signal intensity was observed in the animal pole indicating a greater concentration of free water which showed that physiological differences existed between different regions of the cell. Further, using water suppression techniques, it was shown that a majority of the lipids are isolated to the cytoplasm and the nucleus contains mostly mobile water (Posse & Aue, 1989). Since the cell is not differentiated, there are significant lipid fractions within the oocyte suitable for MR imaging (Sehy, et al. 2001) and spectroscopy (Lee, et al. 2006). Expanding on these findings, MR relaxation measurements

(T1 and T2) identified relaxation differences between these three regions. The nucleus had the longest T1 and T2 relaxation parameters at a resonant frequency of 300 MHz (T1: 1856 ± 100 ms;

T2: 29.45 ± 6.95 ms) followed by the animal pole (T1: 1207 ± 173 ms; T2: 18.81 ± 4.08 ms) and the vegetal pole (T1: 950 ± 191 ms; T2: 11.07 ± 3.05 ms) supporting the finding that the nucleus has a high proton spin density due to a large pool of free water (Pauser, et al. 1995). Relaxometry methods were used to differentiate the regions of the cell and identify the contributions of free and bound water to the MR signal. The identification of two distinct water pools was in agreement with NMR spectroscopy methods that were correlated with destructive methods to determine the water weight in entire oocytes and subcellular regions in response to osmotic perturbation using sucrose solutions (Cameron, et al. 1990). It was reported that no intracellular water pool was equivalent to bulk water or water in dilute solutions and that there were hydration layers in which the water molecules were bound and not in fast exchange. Further, the bound and free water fractions are dependent on the cell type and change in relaxation in protein concentration and other factors that may hinder water exchange. Diffusion measurements on the Xenopus oocyte were expanded to include osmotic perturbations of the extracellular environment and reveal a linear increase in ADC with increasing cell volume (Sehy, et al. 2004). Although these images were obtained on an oocyte and not a fully differentiated cell, it sparked interest in understanding the cell regulatory mechanisms that contribute to the volume- averaged tissue response seen in clinical MRI. In order to examine the effect of osmosis and ionic sodium distributions on a single neuron, many factors must be controlled in the experimental design. A common concern in large organisms is the heterogeneity of tissues and an uncontrolled cellular environment. To combat such concerns, a single isolated neuron was used for single cell studies in a controlled external environment (Grant, et al. 2000). An ideal animal model for such experimentation is the sea slug Aplysia californica which contains a large

18

L7 neuron capable of being isolated from the organism and remaining viable for extended periods of time after dissection (Schoeniger, et al. 1994; Bowtell, et al. 1995). With a diameter of 300-400 µm, the L7 neuron is much larger than typical cells and can be observed and manipulated using standard microscopic techniques. Also, neuronal cells are excitable in which sodium plays an important role in the propagation of electrical signals along the cell. Cell swelling of neurons is also of concern as damage to such cells is typically irreversible making an understanding of cell swelling due to osmosis of particular interest. In studying single neurons isolated from Aplysia californica, clear differences appear to exist between the nuclear and cytoplasmic compartments within the neuron. The nucleus and cytoplasm have a clear chemical composition of osmolytes including betaine and choline that are maintained in relatively high concentrations (hundreds of mmols) that are critical in maintaining osmotic balances in the neuron (Grant, et al. 2000). By imaging over extended time periods in upwards of eight hours, the detected signal of these osmolytes significantly decreased in both the nuclear and cytoplasmic compartments showing a loss of cell membrane integrity which leads to cell swelling and eventually cell death. The nucleus has longer T2 relaxation, resulting in an increase in signal intensity. Additionally, the ADC can be measured in both compartments (Hsu, et al. 1996), showing differences in diffusion between these intracellular regions where the biexponential nature of water diffusion is evident (Grant, et al. 2001). Since the membrane is semipermeable, it was thought that diffusion anisotropy may exist within cells as diffusion may become more prevalent in a single direction. However, upon encoding the ADC in two perpendicular directions, it was shown that tissue anisotropy is unlikely in cells as there are no clear differences in diffusion in the different direction (Hsu, et al. 1997). These results show that diffusion is likely isotropic and maintains this characteristic under osmotic perturbation. Under osmotic perturbation, diffusion in the nuclear and cytoplasmic compartments also change in different ways and can be attributed to differences in the permeability of the membranes as the nuclear membrane allows diffusion more readily. The T2 times increase under hypotonic perturbation, while the diffusion coefficient decreases in the nucleus. However, the ADC measured in the nucleus and cytoplasm remains relatively constant under osmotic perturbation (Hsu, et al. 1996). The response of osmotic perturbation of human brain slices illustrates this similar trend in an increased fast-diffusing fraction within the tissue as the tonicity is increased (Shepherd, et al. 2003). Additionally, diffusion in these human brain slices follow a

19 biexponential model, similar to that observed in the single cell analysis showing that the cellular response can be utilized in understanding how tissues respond to pathological insults and that animal models provide similar tissue environments to those found in humans.

2.5 Tissue Models

Osmotic, neurotoxic, and ischemic external perturbations of the extracellular environment interfere with cellular transport processes, resulting in cell volume and osmoregulation imbalances. Methods to quantify diffusion in these pathological states have led to the incorporation of magnetic resonance imaging (MRI), specifically diffusion-weighted imaging, as a tool to evaluate changing cellular microstructure and water distributions within cellular structures, as can be seen in an isolated neural tissue model. Using the rat hippocampus (van Pul, et al. 2005) and the turtle cerebellum (O'Shea, et al. 2000) as model tissue structures, diffusion, in the context of cell volume regulation under osmotic perturbation and oxygen and glucose deprivation, can be quantified. The apparent diffusion coefficient (ADC) is an indicator of molecular movement within a sample and can be determined using diffusion-weighted MRI. Of particular interest, the ADC decreases in states of tissue ischemia, particularly in acute stroke, with a biophysical mechanism that is not understood. O’Shea, et al. focused on the effect of osmotic perturbations on the intracellular and extracellular volume as assessed by the ADC and the MRI relaxation parameter,

T2. It was hypothesized that the ADC relates to both the intracellular and extracellular volume and that changing ADC values under osmotic perturbation leads to changing volume fractions. Similarly, van Pul, et al. tested the idea that osmotic perturbation and the resulting cell swelling only partially contributed to the decreased ADC seen in ischemic tissue, hypothesizing that oxygen and glucose deprivation also plays a role in the reduced ADC. Although, cell swelling leads to a decreased extracellular volume due to increased tortuosity, the ratio of restricted diffusion to diffusion in free water, changes in the intracellular environment and the tissue itself can lead to a reduced ADC. By separately investigating the effects due to osmotically-induced cell swelling and those due to ischemia, a new aspect of the cellular response can be quantified. As with any living in vitro tissue model, there are factors that cannot be controlled. Cellular

20 metabolism, as well as inherent repair and cell volume regulatory mechanisms, is not considered providing limitations to the applicability to in vivo model systems. To quantify these changes in the ADC, dissected tissues were perfused to maintain tissue viability during MRI analysis. Both the turtle cerebellum (O'Shea, et al. 2000) and rat cerebellum (van Pul, et al. 2005) were chosen because the characteristic intracellular and extracellular volume fractions are well defined. Additionally, comparisons of different tissue models validate the results as being applicable to different tissue types. The turtle cerebellum was held between a cotton cloth perfused with an oxygenated isotonic control solution and osmotic perturbations were introduced by changing the sodium chloride concentration. However, changing only this concentration could lead to neuronal activation which would alter the cellular response to perturbation. The rat hippocampus was held with a nylon mesh perfused with oxygenated artificial cerebrospinal fluid, however, the mesh could provide limitations to cell swelling. Osmotic perturbations were introduced by changing the amount of water added to solutions, which maintained the relative concentration of osmolytes; glucose and oxygen deprivation was introduced simply by removing glucose and bubbling the solution with nitrogen instead of oxygen. Similar diffusion-weighted MR images were acquired. The general conclusion of both studies shows that under hypertonic perturbation the ADC increased while under hypotonic perturbation the ADC decreased showing a nearly linear relationship between the ADC and the extracellular osmolarity. The average ADC was similar in both studies at 6.8 x 10-4 mm2/s compared to 4.69 x 10-4 mm2/s in the rat hippocampus and turtle cerebellum, respectively. Additionally, O’Shea et al. concluded that the T2 value decreased under hypertonic perturbation and increased under hypotonic perturbation. The diffusion coefficients are lower in the intracellular space by a factor of nearly 2, the response is much faster under hypertonic perturbation (time constant of 5 min for hypertonic and 11 min for hypotonic), and the ADC changes faster and more significantly than T2 (32-50% increase over baseline compared to 8-22% for T2). The results show comparable trends to both a fast and slow exchange model, which accounts for independent diffusion in the intracellular and extracellular compartments, showing the physical system likely obeys an intermediate combination of the two exchange models. However, when accounting for tortuosity, viscosity, and molecular interactions, these results show that the total tissue water does impact T2 and the ADC varies with the extracellular space fraction.

21

Similarly, van Pul, et al. showed little anisotropy of the diffusion within the tissue, as there is limited myelin enclosed structures and restrictive membranes. The ADC decreased by 8% under no perturbation, 83% under hypotonic osmotic perturbation, and 57% under glucose and oxygen deprivation. The decrease of the ADC under no perturbation shows there are additional cellular mechanisms accounting for ADC decrease that have not been accounted for. Additionally, oxygen deprivation is critical for the time scale of ADC changes as changes occurred much more rapidly when oxygen is removed, on the order of 3 hours to reach the final ADC deviation in comparison to 8 hours when oxygen is present. However, changes in the extracellular volume fraction are similar between perturbations. Of particular interest, osmotic perturbation resulted in little cell damage, while oxygen and glucose deprivation resulted in detrimental damage to the tissue. It follows that the ADC changes do not result exclusively from cell swelling as the intracellular ADC may be changing due to ischemic effects not related to osmotic perturbation. The ADC can be related to the intracellular and extracellular volume fractions. However, reductions in ADC are due not only to osmotically induced cell volume manipulations, but also to ischemia, which will have large impacts on future work related to the onset of ischemia and stroke in clinical situations. A method for imaging nervous tissue slices has been well developed for MR analysis (Blackband, et al. 1997). The applicability of these tissue model systems shows that both rat and human hippocampal slices exhibit similar MR relaxation and diffusion properties.

2.6 Diffusion Multi-Compartment Models

Together with other MRI studies, a multi-component diffusion model is used to quantitatively assess parameters that affect diffusion across the cell membrane through the use of diffusion weighted MRI. The model for diffusion in isolated neurons relates to the compartmentalization of water in the cell resulting in the existence of both bound and free water with slow transport properties between the intracellular and extracellular environments. A multi- compartment model of diffusion can adequately represent a tissue model of diffusion (Stanisz, 2003. Using human erythrocyte ghosts, water diffusion was found to be non-monoexponential with significant exchange between the intracellular and extracellular compartments (Thelwall, et al. 2002). This method can be used to observe how a cell responds under different osmotic

22 perturbations, the corresponding cell volume increase, and the rate of cellular osmosis. Cell membranes aid in the compartmentalization of the cell; however, the properties of the individual compartments within the cell must be known to be able to understand the heterogeneous system. A diffusion model must account for water exchange between the nucleus, the cytoplasm, and the extracellular environment, but the slow diffusion of such systems ensures that a large fraction of the water remains in the original compartment. Of concern in the diffusion weighted MRI studies is the SNR dependence on coil and voxel size becoming noticeably apparent in imaging when the resolution must be increased. In an effort to increase the resolution of such imaging sequences, the SNR must be increased significantly, often resulting in a tradeoff between SNR and resolution. The measurement of the diffusion coefficients under different osmotic perturbations, including both hypotonic and hypertonic solutions, will be a basis of understanding the mechanism of osmosis on cell volume regulation. Cellular modeling of diffusion becomes complicated by the fact that no tissue sample is homogeneous as cells, extracellular proteins, as well as the extracellular fluid can impact the rate of diffusion. Recently, efforts have been made to determine the parameters that most significantly affect the changes in ADC measured under states of perturbation by modeling the tissue as a collection of cuboidal cells. The ADC decreases by up to 50% following ischemia in tissues which can be accounted for by cell swelling under hypotonic perturbation (Harkins, et al. 2009). The intracellular diffusivity and heterogeneity of tissue structures decrease the ADC in tissues due to the presence of impermeable membranes. The intracellular diffusivity and membrane permeability most greatly effect diffusion as decreases in ADC can be accounted for by decreases in membrane permeability and diffusivity in the intracellular environment. The tortuous intracellular environment causes a faster signal decay leading to reduced ADC measurements, showing that the compartmentalization of tissues is required to accurately account for the measured ADC (Harkins, et al. 2009). It has also been shown that intracellular hydration increases under hypotonic perturbation and the diffusion coefficient decreases in states of increased protein concentration which can directly account for reductions in T2 relaxation values and ADC measurements (Kotek, et al. 2009). Changes in tissue microstructure directly impact both the intracellular and extracellular spaces and consequently alter diffusion (Norris, 2001).

23

2.7 Sodium MRI

Although sodium is much more difficult to image due to its quadrupolar nature, weak signal emission, and low cellular concentration, the use of high magnetic fields allows the imaging of sodium in tissues (Boada, et al. 2005). A quadrupolar nucleus is defined as having a quantum spin number greater than 1/2 (noting that the traditional spin 1/2 nucleus is 1H). Sodium (23Na) is a spin 3/2 nucleus and therefore has a quadrupole moment due to a higher degree of nuclear asymmetry. As opposed to spin 1/2 nuclei that split into two unique energy levels when an external magnetic field is applied, spin 3/2 nuclei split into four levels characterized by the four magnetic quantum numbers +3/2, +1/2, -1/2, and -3/2, representing four possible orientations in a static magnetic field. Therefore, the spin 3/2 nucleus has more than one allowed single quantum transition (3/2 → 1/2, 1/2 → -1/2, -1/2 → -3/2). When comparing sodium and proton MRI, it must be noted that sodium is not homogeneously distributed in tissues, but rather has an intracellular concentration of sodium of 10 mM and an extracellular concentration of 150 mM in humans. Due to poor spatial resolution, the sodium signal is volume averaged over both environments with a total sodium concentration near 50 mM in the brain, due to the large intracellular volume fraction. In comparison to proton imaging where the water concentration is frequently above 50 M, this presents a significant difference and the total sodium concentration limits the signal obtained from a given voxel. Further, several other important differences exist. 23Na is found in 100% natural abundance, similar to the near 100% natural abundance of 1H. However, protons have a gyromagnetic ratio of 42.58 MHz/T while sodium has a gyromagnetic ratio of 11.26 MHz/T, giving sodium a resonant frequency of 132 MHz at 11.75 T while protons have a resonant frequency of 500 MHz at 11.75 T. These differences give sodium a relative sensitivity of approximately 10% that of 1H. Although the sensitivity of 23Na is an order of magnitude below 1 H, its relaxation rates are two orders of magnitude faster with T1 ~ 30-40 ms and T2 ~ 30-40 ms. Factoring the low sodium concentration and reduced sensitivity, the total sensitivity of 23Na-MRI is 10,000 times lower than 1H-MRI, which limits spatial resolution and requires higher magnetic fields, small radiofrequency (RF) coils, and extensive averaging leading to longer imaging times. Due to these factors, the resolution of sodium is poor with the maximum resolution on the order of 100 μm, although maximum clinical resolutions are on the order of 1-2 mm.

24

Given the quadrupolar nature of the sodium nucleus with multiple quantum transitions, the signal obtained from each of these transitions is not uniform, but rather the relative intensities are 3:4:3. Static quadrupolar effects result due to the electric field gradient at the nucleus, while dynamic quadrupolar effects occur due to fluctuations in the electric field at the nucleus. Both longitudinal and transverse (T1 and T2) intrinsic relaxation mechanisms are affected by these interactions. Static quadrupolar effects can cause the two outer transitions (3/2 → 1/2 and -1/2 → -3/2) to experience different resonant frequencies and will be broader, distributed over the entirety of the spectral range (deGraff, 2007). In tissue samples heterogeneous broadening occurs when the tissue is not macroscopically oriented and the outer transitions become blurred together over the entire spectral frequency range. Dynamic quadrupolar effects occur due to fluctuations in the electric field at the nucleus. The extreme narrowing condition is related to the correlation time (ω0τc « 1) and is related to the resonant frequency and the rotational correlation time. If this condition is not fulfilled, biexponential behavior becomes evident. Further, the T1 and T2 relaxations of a quadrupolar nucleus are biexponential in nature. The first in vivo sodium images were produced in 1983 of a cat experiencing acute cerebral stroke (Hilal, et al. 1983). It became evident that sodium MRI was extremely sensitive to pathological conditions as one of the first changes in diseased tissue is the loss of ion homeostasis. By 1985, clinical sodium MRI of humans was feasible (Perman, et al. 1986). Although initial studies examined primarily the sodium signal changes without quantifying either the sodium relaxation parameters or the total sodium concentration, methods have been developed to address these concerns. It was observed by Christensen et al. that the signal intensity for high SNR images is expected to be linear with concentration, given that T2 weighting is minimized and the B0 and B1 fields are homogenous over the sample volume. Field homogeneity concerns can be minimized by shimming on the proton water signal using a double- tuned RF coil. The relationship between sodium signal intensity, I, and total sodium concentration, TSC, can be described in equation (13) when two sodium reference phantoms

(concentrations of C1 and C2 with corresponding image intensities of I1 and I2) are included in the imaging field of view.

25

(13)

It was shown that the total tissue sodium concentration in the rat brain using sodium MRI (45 ± 4 mM) was in excellent agreement with the 22Na radionuclide dilution technique (49 ± 6 mM), however the obvious advantage is that the sodium MRI can be used in vivo and does not require radiation (Christensen, et al. 1996). However, sodium distributions within ischemic tissue and in cancerous tumors have been shown to increase due to the perturbed states (Lin, et al. 2001; Thulburn, et al. 1999). Thulburn et al. focused on the use of sodium MRI to assess changing sodium distributions in tumors in the rat cerebellum with and without interventional treatments. MRI analysis showed a significant increase in sodium concentration in the tumor region as illustrated by an increase in signal intensity. The affected area correlated with morphological changes as seen in the proton image and with ex vivo histological staining confirming that sodium MRI was capable of identifying the affected region and monitoring changes in sodium distributions as a response to pathological conditions. The sodium concentration increased from around 40 mM to 65 mM, a significant increase, possibly illustrating that sodium has a key role in identifying diseased tissues. Although the specific treatments performed in the study did not alter the sodium concentration, this provides an initial starting point to assess therapeutic treatments for cancer. Sodium imaging was expanded by Lin et al. in comparing the sodium response to 1 ischemic conditions. Similar reductions in T2 and increase in ADC were reported for the H imaging studies which correlated with increases in 23Na signal intensity on the same animal subjects. The sodium signal intensity increased in the ischemic regions, possibly due to an increased spin density related to higher sodium concentrations. The exact relaxation mechanism, either T1 or T2, was undetermined and future studies must focus on relating these changes to the physiological conditions in perturbed tissue (Thulburn, et al. 1999). However, the sodium distributions change more rapidly and remained in an altered state for a longer period of time

26 than the corresponding proton images. Even at low resolution, quantitative analysis of diseased nervous tissue show changes in sodium distributions. Sodium imaging could appear normal, even after ischemia has set in; however, there is potential to observe the response to sodium in neurological conditions as a possible method for the early detection, prior to cell death, of neurological diseases. 23Na imaging has become more prominent due its promise of earlier identification of altered cell physiology in vivo. Similar to all MR imaging, 23Na MRI does not affect cognitive function or impose any serious health risks to patients (Atkinson, et al. 2010). Increases in 23Na signal intensity are seen in ischemia (Thulborn, et al. 2005) and tumors (Schepkin, et al. 2011). Although the increase in signal intensity may be due to increases in the total sodium concentration, it could also result from changes in the 23Na relaxation parameters. To provide 23 23 evidence of changing Na relaxation, the Na T2* relaxation parameter was mapped in the human brain and differences were seen between grey and white matter, as well as a much higher sodium content in cerebrospinal fluid (Fleysher, et al. 2009). Recently, the field of sodium MRI has expanded to look at sodium alterations that occur as a result of neurodegeneration in clinical settings. Increases in 23Na signal intensity were shown to correlate with Alzheimer’s disease (Mellon, et al. 2009) and multiple sclerosis (Inglese, et al. 2010); however contributions from 23Na relaxation were not included.

27

CHAPTER THREE

OBJECTIVES

The specific objectives of this work are based on characterizing the cellular response to osmotic perturbation, which results in cell swelling or crenation, through the MR relaxation and diffusion properties of the abdominal Aplysia californica ganglia. Both 1H and 23Na will be utilized to examine the benefit of using non-proton nuclei in assessing cell viability and cellular responses to insults of the extracellular environment. Changes in the MR relaxation parameters and diffusion properties can be used as direct indicators of change in the cell structure and water distribution involved in cell volume regulation. Therefore, MR analysis can show the impact altered extracellular tonicity has on cellular ion distributions as well as morphological change and cellular damage within the ganglia providing insight to the osmotic regulation of cell volume in isolated neural ganglia using a non-destructive technique. It is hypothesized that 23Na MRI at high magnetic field strengths will show that redistributions of ionic sodium occur prior to morphological changes in the cell or apoptotic events and are more sensitive to the physiological environment than conventional 1H MR imaging techniques. Although there are many aspects to cell volume regulation and redistributions of ionic sodium within the cell, this study will focus on altered cellular transport mechanisms due to osmotic perturbations, allowing quantitative characterizations of cell volume regulation in the model Aplysia abdominal ganglia. Specifically, 1 23 differences in H and Na magnetic resonance imaging contrast mechanisms (T1, T2, T2*, and ADC) in fixed, viable, and nonviable neural ganglia will be assessed. The project will be aimed at examining these differences in the context of understanding diseased cell states in pathological conditions.

3.1 High-Field MRI

Magnetic resonance imaging (MRI) will be used in the analysis of ion and water distributions in the isolated neural ganglia as well as in the measurement in the rate of osmosis, characterized by the apparent diffusion coefficient. MRI relies on magnetism to image atomic nuclei and both 1H and 23Na nuclei are detectable through magnetic resonance principles.

28

Although sodium is much more difficult to image, it nevertheless has critical functions in regulating cell volume and initiating both membrane depolarization and signal transduction in 1 excitable cells. Isolated Aplysia ganglia were characterized through H T1, T2, T2* and ADC 23 measurements, as well as Na T1 and T2* measurements. Changes in the relaxation and diffusion measurements were quantified and related to the extracellular osmolarity and viability of the ganglia. By imaging both 1H and 23Na, the role of sodium concentration gradients in the regulation of osmosis and the role of active transport processes on the maintenance of ionic sodium distributions in isolated neural ganglia becomes evident through examination of the MRI relaxation parameters. The MRI relaxation parameters (T1, T2 and T2*) and the apparent diffusion coefficient (ADC) from diffusion-weighted magnetic resonance imaging were used to quantitatively assess changes in tissue microstructure within the ganglia, allowing the effect of osmotic perturbations on cellular osmoregulation to be observed. Although averaged over the entirety of the ganglia, the ganglia is compartmentalized and the measurements represent the average tissue response of the nuclear and cytoplasmic regions of the neurons, as well as the interstitial environment. By quantifying changes in diffusion which is utilized as an indicator of the relative rate of osmosis, cell volume regulation through altered water distributions as well as changes in tissue microstructure due to altered barriers to diffusion are observed. The relaxation and diffusion measurements are aimed at examining changes in the context of understanding and reversing pathological insults in chronic neurodegenerative disease and tissue ischemia. It is hypothesized that 23Na MRI will permit the assessment of the osmotic regulation of cell volume in a prototypical isolated neural ganglia and that redistributions of ionic sodium will occur prior to morphological changes in the cell or apoptotic events. MR relaxation and diffusion measurements of isolated neurons from the Aplysia abdominal ganglia indicate that hypotonic perturbations result in increases in T2 and no apparent changes in ADC which serves as the foundation for work on the abdominal Aplysia ganglia (Hsu, et al. 1996). Both hypertonic and hypotonic perturbations of the extracellular environment were introduced and compared to isotonic conditions to illustrate the role changing ionic gradients have on the detection of cell swelling through MR relaxation and diffusion mechanisms.

29

3.2 Radiofrequency (RF) Coil Design

A radio frequency (RF) coil must be utilized in order to transmit RF pulses to the system and receive the MR signal. A solenoidal coil achieves the highest signal-to-noise ratio (SNR) of available coil configurations and therefore necessary to achieve the required signal to observe the cellular structure of neural ganglia samples (Glover & Mansfield, 2002). However, coil development continues to improve and a large number of coil configurations are possible (Neuberger & Webb, 2009). A tradeoff exists between spatial and temporal resolution as the resolution is directly related to the imaging time and signal obtained. For example, in imaging brain slices a resolution of 120x230x300 μm can be obtained in 4 minutes while a resolution of 15x15x300 μm necessitates a 14 hr image acquisition (Benveniste & Blackband, 2002). The resolution, however, can be improved without sacrificing imaging time through the improvement of the SNR, defined as the difference in mean signal and mean noise obtained divided by the standard deviation of the noise. A higher SNR corresponds to higher signal intensity in the acquired MR images (Minard & Wind, 2001). The SNR can be improved by utilizing higher magnetic field strengths as there is a strong dependence of SNR on the externally applied magnetic field strength, Bo, as shown in equation (14) (Peck, et al. 1995).

(14)

Further, the SNR can be improved through the use of small solenoidal coils. For a solenoidal coil of a specified length (L) and diameter (D), the SNR improves according to equation (15). It can be seen that the length to diameter ratio is important is maximizing the signal obtained (Peck, et al. 1995).

(15)

√ ⁄

√ ( ⁄ ) A radio frequency (RF) coil was constructed which could be used at multiple frequencies. The solenoidal coil was double tuned to both the resonant frequencies of 1H and 23Na. At field strength of 11.75 T, this corresponds to a coil tuned to 132 MHz and 500 MHz for 23Na and 1H, 30 respectively. The large range of required frequencies necessitates the use of a versatile coil which can be easily manipulated to tune to varying frequencies. Two separate channels were utilized on the coil and each was tuned to the operating frequency at the given magnetic field strength, which fluctuates slightly with time. Changing between the resonant frequency of the different nuclei required changing the tuning capacitor on the coil to a different capacitance, C, as the relation between the inductive and capacitive properties of the coil are shown in equation (16) where ω is the resonant frequency, L is the inductance, and C is the capacitance (Minard & Wind, 2001).

(16)

The coil was constructed by wrapping copper wire around a fused silica capillary. A capillary with a 3 mm diameter was used to facilitate the larger ganglia samples, although small microcoils (700 µm diameter) have been used to image single neurons (Grant, et al. 2001). SNR analysis will be useful in analyzing coil performance at varying field strengths (Peck, et al. 1995). Due to its low sensitivity, sodium is inherently difficult to image and requires the use of high magnetic field to achieve signal to noise ratios that yield signal strengths that can be clearly identified from the noise background. The SNR depends on the size of the RF coil, the size of the voxel which is being imaged, the static magnetic field strength, the spectral width and acquisition time (Webb, 2003). In order to increase the SNR, small coils must be used at high magnetic fields and the RF coil design is at the millimeter length scale to increase SNR (Minard & Wind, 2001). With an inherently small voxel size due to the microscopic nature of the neurons within the ganglia, these efforts must be taken to reduce the total acquisition time. This time reduction will reduce the number of averages in an effort to limit the imaging time to levels short enough to ensure cell viability. During imaging, the neural ganglia were placed in a capillary tube and centered in the RF solenoidal coil. Due to the coil configuration, 1H and 23Na imaging were performed in separate studies which did not allow for simultaneous data acquisition of both the 1H and 23Na signals. The experimental setup was simplified by not perfusing the ganglia (viability was expected to be maintained for up to seven hours after dissection) which allowed the ganglia samples to be directly inserted into the coil.

31

3.3 Neural Ganglia Model System

In understanding the role of the isolated neuron in the context of larger tissues and cellular interactions, isolated abdominal ganglia from the sea slug Aplysia californica were quantitatively described using both 1H and 23Na magnetic resonance imaging. The abdominal ganglion houses the L7 motor neuron which has been used extensively in single cell MRI studies. Expanding to the imaging of the abdominal ganglia allows comparisons of the entire ganglia to previous work using isolated neurons; cell to cell interactions and tissue homogeneity issues in the MR relaxation and diffusion properties of single neurons within a tissue environment can be explored. Further, the abdominal ganglion has been extensively characterized for cellular electrophysiology studies. It is a suitable model system due to its relatively simple anatomy, consisting of a small number of large neurons which can reach sizes more than 300 µm in diameter (Coggeshall, 1967). The general anatomy of the ganglion consists of two spherical hemispheres connected by a commissure and the entire ganglion is encapsulated in a fibrous connective tissue sheath (Frazier, et al. 1967). The sheath is not uniform, but is thickest where the nerves attach to the abdominal ganglia and thinner in other areas. Also, the ganglion is vascularized through the main dorsal aorta which branches into several arteries; however, the arteries end in the sheath surrounding the ganglia and do not penetrate into the interior of the ganglia, but rather empty the blood directly into the sheath indicating that the nervous tissue itself (consisting of the nerve and glial cells within the ganglion) are avascular and rely on diffusion for nourishment (Coggeshall, 1967). The connections of the large neurons within the abdominal ganglia have also been characterized (Kandel, et al. 1967). There is a 30-40% increase in the number of axons when large animals (200 g) are compared with juvenile animals (5 g). Further, increases in the size of Aplysia neurons can be related to temperature (Treistman & Grant, 1993) and osmoregulatory processes become slower as the animals age (Skinner &Peretz, 1989) making it important to account for factors that may alter the physiology of the ganglia. The neurons themselves are located along the periphery of the ganglion surrounding a central region containing neutrophil containing the presynaptic and postsynaptic processes of the neurons where synapses between neurons occur. Although the number of large neurons remain the same throughout development, as the animals age the total neurons increase by up to 40% mainly through additional small neurons at the boundary of the interior neutrophil. The large neurons are unipolar cells with a large nucleus that accounts for

32

30% of the cellular volume but as the cells become larger the nuclei become flatter and occupy less space. The abdominal ganglion contains between 1,000-2,000 cells and 15,000-25,000 axons making it a very simple nervous system model in comparison to higher organisms (Coggeshall, 1967). As the animals mature, several key changes occur within the neurons of the abdominal ganglion including an increase in the size of the nucleus and cell body with added irregularities in shape and granule distribution. The changes of developing neurons are evident in MR relaxation and diffusion measurements (Aiken, et al. 1996). In imaging Aplysia abdominal ganglia, temperature, age, and the general size and appearance of the ganglia were kept uniform to prevent unwanted changes in cellular physiology.

3.4 Osmotic Perturbation

By examining the cellular response to different osmotic insults of the extracellular environment, much can be learned about cell volume regulation and the influence of changing ionic concentration distributions on cell swelling and osmosis in diseased states. In states of reduced tonicity, there is a net influx of water into the cell and the cell swells, while in states of increased tonicity the cell shrinks. The morphological changes under state of osmotic stress lead to disordered cell shapes which impacts transport in cells and tissues (Chen, et al. 2000). Of the cellular ions, sodium is paramount in cell volume regulation due to the large variations of concentration in the intracellular and extracellular environments. Sodium ions are polar compounds incapable of diffusing through the highly non-polar interior region of the cell membrane in significant quantities. Relying on active transport processes, notably the sodium- potassium pump, sodium transport is directly controlled by each cell and is a regulatory mechanism which accounts for a great majority of cellular energy consumption (Lang, et al. 1998). The sodium-potassium pump is a key regulator of sodium transport across the cell 23 membrane and therefore in controlling osmosis. By quantifying the Na T1 and T2* relaxation parameters and the 23Na signal intensity, the sodium redistributions due to osmotic insult can be characterized. Through this understanding, further efforts may be undertaken to combat the effects of such insults which may be used in reversing the onset of cell death in neurological disease.

33

Osmosis is driven by osmotic pressure differences which can be determined using equation (17) where i is the dissociation factor, M is the concentration, R is the gas constant, T is the absolute temperature, and π is the osmotic pressure (Strange, 2004).

(17)

Following isotonic studies on neural tissue, the osmolarity of the perfusate will be varied by altering the ionic concentrations found in the extracellular medium. Similar MRI analysis will be performed and the parameters for the ganglia will be compared to the isotonic values initially obtained. In so doing, the effect of certain perturbations can be related to decreases in the intracellular volume fraction under states of hypertonic perturbation and increases under hypotonic perturbation. Therefore, it can be concluded that changes in the relaxation parameters are direct indicators of changes in the cell structure and water distribution involved in cell volume regulation (Hsu, et al. 1996). Although there are many aspects to cell volume regulation and redistributions of ionic sodium within the cell, studies have used MRI to examine the transport mechanisms within a cell under both isotonic conditions and with osmotic perturbations (Hsu, et al. 1996). Allowing quantitative characterizations of transport mechanisms to be developed in the model Aplysia californica isolated neuron, osmotic effects have shown to decrease the T2 relaxation time under hypotonic perturbation as is commonly encountered in tissue ischemia. Through the efforts of examining the effect of osmotic perturbation on isolated neural ganglia, the mechanism of cell volume regulation in neurons continues to be better developed to understand diseased cell states in neurodegeneration.

3.4 Cell Viability

Of concern in many single cell and tissue model systems is cell viability and how significantly physiological mechanisms are altered due to cell death. As cells undergo apoptosis, significant changes in water distributions alter MR relaxation and diffusion measurements (Kettunen & Brindle, 2005). By quantifying the proton relaxation and diffusion properties of the ganglia immediately after dissection and again after cell viability is lost, the changes associated with loss of viability can be identified. Sodium redistributions may present novel ways of

34 assessing cell viability since as cells become nonviable, active transport ceases and membrane permeability is altered resulting in significant ionic redistributions. Although the sodium relaxation parameters are only measured on nonviable ganglia, by observing the sodium signal intensity over time, insight into the changes in ionic distributions within the ganglia can be identified.

35

CHAPTER FOUR

METHODS

4.1 Neural Ganglia Preparation

Aplysia californica were acquired from the National Resource for Aplysia at the Rosenstiel School of Marine and Atmospheric Science at the University of Miami and were in the late juvenile stage of development with a weight of 46 – 75 g. The Aplysia were shipped in plastic bags filled with oxygen and sea water. Animals were maintained in a seawater aquarium at 18-20 ˚C, slightly higher than their natural water temperature, and approximate pH of 7.8-8.0. Aquarium salinity was adjusted by adding Instant Ocean sea salt to achieve a specific gravity of 1.020-1.025. The Aplysia could be maintained in this environment for several weeks by checking the salinity and feeding every three days with washed romaine lettuce. The abdominal, or visceral, neuronal ganglion from each Aplysia was obtained by first anesthetizing the slug with a magnesium chloride paralytic agent injected in three separate locations along the muscular foot of the animal. Approximately 15-20 mL of the anesthesia was injected into each slug in each of the three different injection sites in order to quickly disperse the anesthesia within the organism. Anesthesia was isotonic to artificial sea water (ASW) and contained 379 mM MgCl2 and 15 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, a common buffering agent used to maintain pH). After verifying that there were no apparent muscular responses or gill reflexes, a ventral incision was made along the muscular foot and standardized dissecting techniques using scissors and tweezers were used to locate, isolate, and remove the ganglia (Figure 9). The ganglia were easily located in each animal as they were small orange masses located in the abdominal cavity whose connections to other ganglia in the organism could be seen. No digestive enzymes were used in the isolation of the ganglia as only the large portion of connective tissue surrounding the ganglia was removed. After cutting out around the abdominal ganglia and removing the connective tissue, the abdominal ganglia exhibit a characteristic “H” shape as seen in Figure 10. The ganglia were initially washed by being placed in a Petri dish containing isotonic ASW and allowed to adjust for several minutes. The ganglia were then moved to another Petri dish containing either isotonic, 20% hypertonic, or

36

20% hypotonic ASW. Three Aplysia were dissected for each repetition and all three ganglia were loaded into a 2.5-mm outer diameter capillary, acquired from Wilmad Lab Glass, approximately 2.5-3 cm in length which contained the ASW solution and was sealed at both ends using wax.

Figure 9. Dissection process for isolating the abdominal ganglion from Aplysia. Left: Anesthetized Aplysia prior to dissection. Middle: Aplysia after the ventral incision exposing the abdominal cavity. Right: Close-up of the abdominal ganglion which is removed.

Figure 10. Image of three viable isolated abdominal ganglia from Aplysia after dissection. All three ganglia were loaded into the same capillary and then imaged.

For fixed ganglia samples, the same dissection procedure as outlined above was used. The ganglia were initially washed by being placed in a Petri dish containing isotonic ASW and allowed to adjust. The ganglia were then placed in a 4 wt% paraformaldehyde ASW solution and refrigerated for several weeks. At the time of final sample preparation, the ganglia were then

37 washed several times with isotonic ASW to remove the as much of the fixative as possible. Finally the ganglia were moved to a Petri dish containing either isotonic, 20% hypertonic, or 20% hypotonic ASW and allowed to adjust. Again, three ganglia were loaded into a 2.5 mm outer diameter capillary containing the ASW solution and sealed with wax at both ends.

4.2 Artificial Sea Water (ASW) Solutions

Aplysia californica is a marine organism whose natural environment is sea water. Therefore, to assess osmotic changes in Aplysia neurons, artificial sea water (ASW) of similar composition is prepared. ASW has standard values of several key compounds, as shown in Table

1, including sodium chloride (NaCl), potassium chloride (KCl), magnesium chloride (MgCl2), calcium chloride (CaCl2), and HEPES, a buffering agent used to maintain pH.

Table 1. Standard concentrations of key compounds in isotonic artificial sea water (ASW). The concentration, C, is given in mmol/L.

Substance C / mM NaCl 460 KCl 10.4

MgCl2 55

CaCl2 11 HEPES 15

The osmotic pressure, defined as a measure of the tendency of water to move to a more concentrated solution, of the isotonic solution was calculated using the Morse equation, shown in (18), for osmotic pressure, assuming an ideal dilute solution and full dissociation of the salt compounds.

(18)

38

In this equation i is the van’t Hoff factor for dissociation, M is the concentration of the compound in mol/L, R is the universal gas constant (0.082 Lˑatm/molˑK), T is the absolute temperature in K, and π is the osmotic pressure in atm. When multiple compounds contribute to the osmotic pressure, this equation can be expanded as shown in (19) when n species are present.

(19)

∑

Assuming the solution to be at 25 °C (293 K) and letting species 1, 2, 3, 4, and 5 be sodium chloride, potassium chloride, magnesium chloride, calcium chloride, and HEPES, respectively, the osmotic pressure can be calculated. The van’t Hoff factors for each of these species are 2, 2, 3, 3, and 1, respectively. An example calculation for the isotonic ASW is shown in (20) below.

[ (20)

( )

]

The osmotic pressure of the isotonic solution was found to be 27.7 atm. To assess the influence of changing ionic gradients, two additional solutions of ASW were made of differing osmolarities. The osmotic differences were introduced through changing the sodium chloride (NaCl) concentration in the solution. The two additional solutions were 20% hypertonic ASW and 20% hypotonic ASW and were used for the osmotic perturbation tests on both fixed and living ganglia, using isotonic ASW as a control. Using the osmotic pressure of the isotonic ASW as a basis, the hypertonic and hypotonic ASW solutions had osmotic pressures 20% lower and 20% higher than isotonic ASW, respectively. The differences in osmolarity were achieved in both cases by changing only the NaCl concentration, leaving all other ionic concentrations unaltered. The change is NaCl concentration was not expected to be high enough to induce

39 neuronal activation. The osmotic pressure of the hypertonic solution must be 20% larger than the isotonic solution. Therefore the osmotic pressure is found to be 33.3 atm, shown in (21) below.

(21)

In a similar manner, the osmotic pressure of a 20% hypotonic solution must be 20% lower than the isotonic solution. Therefore the osmotic pressure is found to be 22.2 atm, shown in (22) below.

(22)

Based on the osmotic pressures of the hypertonic and hypotonic solutions, the needed concentration of sodium chloride, C1, to achieve the desired osmotic pressure can be back calculated from the osmotic pressure. Assuming that the interior of the ganglionic sac is isotonic to ASW, an osmotic pressure gradient of 5.5 atm in the hypertonic and hypertonic solutions and no osmotic pressure gradient for the isotonic solutions is expected. A sample calculation for the sodium chloride concentration in the hypertonic ASW solution is shown in (23) below.

[ (23)

( )

]

40

In a similar manner, the sodium chloride concentration for the hypotonic solution is shown in (24) below.

( ) [ (24)

( ) ( )

( ) ( )]

From these calculations, the final concentrations of each salt needed to make the ASW solutions were determined and the amount of salt added to make 1 L of each solution is presented in Table 2. It terms of osmolarity, the isotonic solution has a total osmolarity of 1139 mOsm/L (excluding HEPES), while the hypertonic and hypotonic ASW have a total osmolarity of 1367 and 911 mOsm/L, respectively. The effect of placing neuronal ganglia in each ASW solution is illustrated in Figure 11 below, assuming the interior of the ganglia to be isotonic to ASW. When placed in a hypertonic environment, water will tend to flow out of the ganglia causing the ganglia to shrink, or crenate. Conversely, in a hypotonic environment, water will tend to flow into the ganglia causing the ganglia to swell.

41

Table 2. Quantities of each component to make 1 L of ASW solutions of different tonicities. The hypotonic solution compositions are shown in (a), isotonic in (b), and hypertonic in (c).

Substance M / (g/mol) C / mM n / mol m /g NaCl 58.44 345 0.3450 20.162 KCl 74.56 10.4 0.0104 0.775

MgCl2 203.30 55 0.0550 11.182

CaCl2 147.02 11 0.0110 1.617 HEPES 238.30 15 0.0150 3.575 (a) Substance M / (g/mol) C / mM n / mol m /g NaCl 58.44 460 0.4600 26.882 KCl 74.56 10.4 0.0104 0.775

MgCl2 203.30 55 0.0550 11.182

CaCl2 147.02 11 0.0110 1.617 HEPES 238.30 15 0.0150 3.575 (b) Substance M / (g/mol) C / mM n / mol m /g NaCl 58.44 575 0.5750 33.603 KCl 74.56 10.4 0.0104 0.775

MgCl2 203.30 55 0.0550 11.182

CaCl2 147.02 11 0.0110 1.617 HEPES 238.30 15 0.0150 3.575 (c)

42

Figure 11. Cellular response of placing fixed neural ganglia in extracellular environments of different tonicities, assuming the ganglia to be isotonic to ASW. The expected changes in cell volume due to hypotonic (A), isotonic (B), or hypertonic (C) ASW solutions are shown.

4.3 MR Image Acquisition

Initial ASW relaxivity measurements were conducted at the 11.75 T magnet at the FAMU-FSU College of Engineering and the 21.1 T ultra wide-bore magnet at the National High Magnet Field Laboratory. In order to compare relative signal intensities at varying field strengths, the solenoidal coil was single tuned to either the resonant frequency of hydrogen or sodium. At the lower filed strength of 11.75 T, this corresponds to a coil tuned to either 132 MHz or 500 MHz for sodium and hydrogen, respectively. Alternatively, at 21.1 T, these resonant frequencies are 268 MHz and 900MHz for sodium and hydrogen, respectively. Additionally, the imaging of hydrogen and imaging of sodium was performed in separate studies which did not allow for simultaneous data acquisition of both the proton and sodium signals. Changing between the resonant frequencies of the different nuclei resulted in changing the tuning capacitor of the coil. For both fixed and living imaging of Aplysia neuronal ganglia, three abdominal ganglia were isolated and loaded into a 2.5 mm diameter capillary that was around 2-3 cm long and placed at the center of the RF solenoidal coil. All imaging studies were performed at 11.75 T using a double-tuned solenoidal transmit-receive coil. The new coil was double-tuned to both resonant frequencies simultaneously, 132 MHz for 23Na imaging and 500 MHz for 1H imaging to allow both 1H and 23Na on the same ganglia. The coil was constructed by wrapping copper wire around a fused silica capillary. The capillary had an outer diameter of 3 mm (Figure 12) which allowed the capillary containing the ganglia to be inserted at the center of the coil. The coil was

43 housed in a non-protonated susceptibility-matching fluid (Fluorinert) to maximize field homogeneity at the coil boundary. Measurements of the proton T1, T2, and T2* relaxation parameters as well as the apparent diffusion coefficient (ADC) were conducted on living ganglia as well as on ganglia that had been allowed to lose viability after sitting in the magnet overnight.

Sodium T1 and T2* measurements were then made on the nonviable ganglia, as this was necessary to accommodate the long sodium imaging times.

Figure 12. Image of the solenoidal transmit-receive coil showing both resonant circuits for simultaneous tuning to both the 1H and 23Na resonant frequencies. The front of the coil is shown on the right and the back of the coil on the left.

For proton T1 relaxation time measurements, a spin echo scan sequence (dwi_se) was performed utilizing eight separate scans with differing repetition times to determine the T1 relaxation time for the ganglia samples. The echo time was 14 ms (TE = 14 ms), the repetition time varied from 650 to 7500 ms, 4 averages were acquired, and the resolution was 40 x 40 µm with a slice thickness of 0.1 mm. The tip angle was set based on finding the spin echo 180˚ tip averaged over the entire sample. The receiver gain was set for the scan with the longest repetition time (TR = 7500 ms) and that value was used for all scans. A summary of the acquisition parameters can be seen in Table 3. The scan acquisitions were acquired out of order to avoid systematic errors in signal measurement and the order of acquisition was scans with TR times of 1500, 800, 3000, 650, 2000, 7500, 5000, and 1000 ms. The excitation and refocusing pulses both had a duration of 2 ms and the spectral width was minimized to achieve the shortest echo time while still being able to maintain the specified geometry.

44

1 Table 3. Summary of MR acquisition parameters for the H spin echo T1 measurement.

Parameter Value TR [650, 800, 1000, 1500, 2000, 3000, 5000, 7500] ms TE 14 ms Averages 4 FOV 10.2 x 2.6 mm Resolution 40 x 40 µm Matrix 256 x 64 Spectral width 80 kHz (79365.1 Hz) Slice Thickness 0.1 mm (25 slices, coronal orientation, interlaced acquisition) Scan Time 1 hr 33 min

Proton T2 relaxation time measurements were also performed using a spin echo (dwi_se) scan sequence utilizing five different scans with differing echo time to determine the proton T2 relaxation time for the ganglia samples. The repetition time was 5000 ms (TR = 5000 ms), the echo time varied from 14 to 100 ms, 4 averages were acquired, and the resolution was 40 x 40 µm with a slice thickness of 0.1 mm. The tip angle was set based on finding the spin echo 180˚ tip averaged over the sample volume. The receiver gain was set for the scan with the shortest echo time (TE = 14 ms) and that value was used for all scans. A summary of the acquisition parameters can be seen in Table 4. The scan acquisitions were acquired out of order to avoid systematic errors in signal measurement and the order of acquisition was TE times of 25, 60, 14, 100, and 40 ms. Again, the excitation and refocusing pulses both had a duration of 2 ms and the spectral width was minimized to achieve the shortest echo time while still being able to maintain the specified geometry.

45

1 Table 4. Summary of MR acquisition parameters for the H spin echo T2 measurement.

Parameter Value TR 5000 ms TE [14, 25, 40, 60, 100] ms Averages 4 FOV 10.2 x 2.6 mm Resolution 40 x 40 µm Matrix 256 x 64 Spectral width 80 kHz (79365.1 Hz) Slice Thickness 0.1 mm (25 slices, coronal orientation, interlaced acquisition) Scan Time 1 hr 48 min

Proton T2* relaxation time measurements were performed using a gradient-recalled echo

(MGE) scan sequence to determine the T2* relaxation time for the ganglia samples. The repetition time was 5000 ms (TR = 5000 ms), the echo time varied from 6 to 66 ms, 8 averages were acquired, and the resolution was 40 x 40 µm with a slice thickness of 0.1 mm. The tip angle was set based on finding the 180˚ tip found using the spin echo sequence and adjusting for the 4 ms pulse duration used in this scan. The receiver gain was set for the scan with the shortest echo time (TE = 6 ms). A summary of the acquisition parameters can be seen in Table 5. The order of scan acquisitions was from low TE to high TE based on the nature of the scan sequence, as only one scan is acquired. As mentioned, the excitation pulse had a duration of 4 ms and the spectral width was minimized to achieve the shortest echo time while still being able to maintain the specified geometry and minimum echo time.

46

1 Table 5. Summary of MR acquisition parameters for the H gradient echo T2* measurement.

Parameter Value TR 5000 ms TE [6, 18, 30, 42, 54, 66] ms Averages 8 FOV 10.24 x 2.56 mm Resolution 40 x 40 µm Matrix 256 x 64 Spectral width 50 kHz (50000 Hz) Slice Thickness 0.1 mm (25 slices, coronal orientation, interlaced acquisition) Scan Time 43 min

A diffusion-weighted spin echo scan sequence (DW_SE) was performed to determine the ADC for the ganglia samples. The repetition time was 2500 ms (TR = 2500 ms), the echo time was 20.5 ms (TE = 20. 5 ms), b values ranged from 5 to 4500 s/mm2, 8 averages were acquired, and the resolution was 40 x 40 µm with a slice thickness of 0.1 mm. The excitation and refocusing pulses had a duration of 2 ms and 1.5 ms, respectively. The tip angle was set based on finding the 180˚ tip found in the spin echo sequences, using this for the excitation pulse, and adjusting for the 1.5 ms refocusing pulse duration used in this scan. A summary of the acquisition parameters can be seen in Table 6. The order of scan acquisitions was from low b to high b based on the nature of the scan sequence, as only one scan is acquired. The spectral width was minimized to achieve the shortest echo time while still being able to maintain the specified geometry and minimum echo time.

47

Table 6. Summary of MR acquisition parameters for the 1H diffusion-weighted spin echo ADC measurement.

Parameter Value TR 2500 ms TE 20.5 ms b [5.19, 523.78, 777.96, 1031.49, 1537.4, 2042.38, 3050.74, 4560.98] s/mm2 Averages 8 FOV 10.24 x 2.56 mm Resolution 40 x 40 µm Matrix 256 x 64 Spectral width 55 kHz (55147.1 Hz) Slice Thickness 0.1 mm (25 slices, coronal orientation, interlaced acquisition) Scan Time 2 hr 51 min

For sodium T1 measurements, a series of five three-dimensional gradient recalled echo scan sequences, utilizing the fast low angle shot (FLASH) technique, with differing repetition times were performed to determine the T1 for the ganglia samples. The echo time was 3.5 ms (TE = 3.5 ms) and the repetition times were 20, 50, 100, 200, and 250 ms (TR = 20, 50, 100,

200, 250 ms) to determine the T1 relaxation parameter. A total of 128 averages were acquired with a resolution of 89 x 89 x 400 µm. The excitation pulse duration was set to the 90o tip at an attenuator setting of 12 dB. The total acquisition time was 22 hours. A summary of the acquisition parameters can be seen in Table 7. The scan acquisitions were acquired out of order to avoid systematic errors in signal measurement and the order of acquisition was TR times of 50, 200, 20, 100, and 250 ms.

48

23 Table 7. Summary of MR acquisition parameters for the Na 3D gradient echo T1 measurement.

Parameter Value TR [20, 50, 100, 200, 250] ms TE 3.5 ms Averages 128 FOV 11.4 x 11.4 x 3.2 mm Resolution 89 x 89 x 400 µm Matrix 128 x 128 x 8 Slice Orientation Coronal Attenuator 0 12 dB Scan Time ~22 hr Effective Spectral Bandwidth 8928.6 Hz Echo Position 10% Duration of Read Dephase 1.0 ms Duration of Slice Rephase 1.0 ms Read Spoiler Duration 1.0 ms Read Spoiler Strength 10% Slice Spoiler Duration 1.0 ms Slice Spoiler Strength 10%

For sodium T2* measurements, a series of five three-dimensional gradient recalled echo scan sequences, utilizing the fast low angle shot (FLASH) technique, with differing echo times were performed to determine the T2* for the ganglia samples. The repetition time was 100 ms (TR = 100 ms) and the echo times were 3.5, 6, 6, 15, and 21 ms (TE = 3.5, 6, 9, 15, 21 ms) to determine the T2* relaxation parameter. A total of 128 averages were acquired with a resolution of 89 x 89 x 400 µm. The excitation pulse duration was set to the 90o tip at an attenuator setting of 12 dB. The total acquisition time was 18 hours. A summary of the acquisition parameters can be seen in Table 8. The scan acquisitions were acquired out of order to avoid systematic errors in signal measurement and the order of acquisition was TE times of 6, 15, 3.5, 9, and 21 ms.

49

23 Table 8. Summary of MR acquisition parameters for the Na 3D gradient echo T2* measurement.

Parameter Value TR 100 ms TE [3.5, 6, 9, 15, 21] ms Averages 128 FOV 11.4 x 11.4 x 3.2 mm Resolution 89 x 89 x 400 µm Matrix 128 x 128 x 8 Slice Orientation Coronal Attenuator 0 12 dB Scan Time ~18 hr Effective Spectral Bandwidth 8928.6 Hz Echo Position 10% Duration of Read Dephase 1.0 ms Duration of Slice Rephase 1.0 ms Read Spoiler Duration 1.0 ms Read Spoiler Strength 10% Slice Spoiler Duration 1.0 ms Slice Spoiler Strength 10%

4.4 Statistical Analysis

Initially, a standard difference of means t-test was used to compare differences within the viable, nonviable and fixed ganglia. MR relaxation and diffusion measurements of ganglia in ASW of different osmolarities were compared using a significance level of p<0.05. All of the proton datasets for the fixed, viable and nonviable ganglia were then grouped and a one-way ANOVA with a least significant difference (LSD) post hoc was performed to compare the differences between osmolarities within each viability condition. Additionally, comparisons were made between the viability conditions at each fixed osmolarity. Differences were deemed

50 significant at p<0.05. The LSD post hoc allows for multiple tests to be performed on a given dataset and determines statistical differences between all groups assuming a normal distribution and homogeneity. Equal variance was assumed even though in some conditions this assumption may be violated. Equal variance was difficult to achieve because frequently the differences between individual samples was greater than differences within the ganglia of a given sample. Also, because differences in the magnitude of the different MR relaxation and diffusion measurements are large and all measurements were grouped into a single analysis, equal variance may not be achieved. Further, all of the sodium datasets for the fixed and nonviable ganglia were then grouped and a similar one-way ANOVA with LSD post hoc was performed to compare the differences between osmolarities within each viability condition. Additionally, comparisons were made between the viability conditions at each fixed osmolarity. The proton and sodium datasets had to be analyzed separately because no sodium relaxation measurements were performed on viable ganglia.

51

CHAPTER FIVE

RESULTS

MR data were acquired on a total of nine fixed ganglia and 42 viable ganglia. Further, imaging on an additional three ganglia was performed to quantify 23Na signal intensity on viable ganglia. The results are divided between fixed and viable ganglia to show the effect of fixative on MR relaxation and diffusion. Further, the results are presented for each of the osmotic conditions tested showing the initial isotonic condition followed by the 20% hypertonic and 20% hypotonic osmotic perturbations. Both 1H and 23Na datasets were acquired on identical ganglia, although different ganglia were used for each osmotic condition. Viable ganglia were defined as being imaged immediately after dissection. Imaging began after scout images were acquired, shimming on the sample volume was performed, and the correct power settings to achieve the desired flip angles were determined, usually within two hours of dissection. The proton measurements lasted a total of seven hours to measure the T1, T2, T2* and ADC. Nonviable images were acquired twelve hours after the initial measurement on the viable ganglia. Identical slices and ROI size were used for both the viable and nonviable measurements on the same ganglia. Sodium imaging was also performed on the same ganglia and began immediately following the proton imaging on nonviable ganglia around 26 hours after dissection.

5.1 1H/23Na Imaging of Fixed Ganglia Fixed Aplysia ganglia were imaged in each of the osmotic conditions using the acquisition parameters described previously.

5.1.1 Isotonic Fixed Ganglia Imaging was first performed on three fixed ganglia in isotonic ASW. Fixative abruptly stops cell physiological processes and alters cellular microstructure through crosslinking proteins. Fixed cells within the ganglia loose membrane integrity. The loss of membrane integrity and even integrity of the ganglionic sac, which contains all the neural cells, was compromised as seen in the MR images as a lack of a clearly defined membrane at the periphery

52 of the ganglion. Further, no active transport processes occur in fixed tissue, and transport is through simple passive diffusion across a compromised cell membrane. A complete T1 dataset and the regions of interest (ROIs) used for analysis are presented in Figure 13. The T1 relaxation parameter was much shorter than those seen in viable tissue (1100 ms in comparison to 2000 ms). Complete T2 and T2* datasets with the ROIs outlined are shown in Figure 14 and Figure

15, respectively. T2 and T2* contrast were not developed at short TE times due to a longer T2 relaxation parameter (around 38 ms), which required the use of longer TE times for proper curve 2 fitting (R > 0.95). The T2 and T2* relaxation parameters were much longer than those seen in viable tissue (T2: 38 ms in comparison to 19 ms; T2*: 29 ms in comparison to 13 ms). A complete ADC dataset with the ROIs outlined is shown in Figure 16. Diffusion-weighted MR images at high b values directly identified the ganglia from the ASW medium, seen as a bright region (high signal intensity) due to restricted diffusion within the ganglia and a lower diffusion coefficient than the ASW medium. Sodium imaging (Figure 17) showed strong correlation between the 1H and 23Na images as the outline of the ganglionic sac remained consistent between both 1H and 23Na images, although no cellular regions could be identified on 23Na images due to the lower spatial resolution (89x89 μm in comparison to 40x40 μm). However, the 23Na signal intensity was not uniform throughout the interior of the ganglia, where the ganglia are comprised of mainly connective tissue and few neurons indicating that cellular regions have either higher 23 23 Na content or differing relaxation properties. It can be noted that the Na T1 and T2* relaxation parameters were shorter in the ganglia than the ASW medium, yet the ganglia have a lower signal intensity (appear darker in the MR images) indicating that the sodium content is lower within the ganglia than in the ASW. This finding is interesting because, even though the ganglia have lost membrane integrity due to fixation, there are 23Na differences that can be identified within the gangliar regions. All relaxation and diffusion measurements were fit to a monoexponential function as shown in Figure 18. Little or no biexponential behavior was observed as evidenced by the high R2 values indicating that the monoexponential fits were adequate in fitting the data. Although compartmentation within the ganglia is expected, volume averaging over the entire ganglia revealed that plotting average signal intensities versus the different contrast parameters (TR, TE or b) led to monoexponential curve fits. Quantitative values of the proton T1, T2, T2* and ADC, as well as the sodium T1 and T2*, for the isotonic fixed ganglia are shown in Table 9.

53

TR = 650 ms TR = 800 ms TR = 1000 ms TR = 1500 ms A

TR = 2000 ms TR = 3000 ms TR = 5000 ms TR = 7500 ms

B

1 Figure 13. Representative T1-weighted H MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with eight different TR values ranging from 650 to 1 7500 ms for the H T1 measurement. (B) Image of the isotonic fixed ganglia with a TR of 7500 ms (left) and the corresponding ROIs (right).

54

A TE = 14 ms TE = 25 ms TE = 40 ms

TE = 60 ms TE = 100 ms

B

1 Figure 14. Representative T2-weighted H MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with five different TE values ranging from 14 to 1 100 ms for the H T2 measurement. (B) Image of the isotonic fixed ganglia with a TE of 25 ms (left) and the corresponding ROIs (right).

55

A TE = 6 ms TE = 18 ms TE = 30 ms

TE = 42 ms TE = 54 ms TE = 66 ms

B

1 Figure 15. Representative T2*-weighted H MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with six different TE values ranging from 6 to 66 1 ms for the H T2* measurement. (B) Image of the isotonic fixed ganglia with a TE of 18 ms (left) and the corresponding ROIs (right).

56

2 2 2 2 A b = 5 s/mm b = 524 s/mm b = 778 s/mm b = 1031 s/mm

2 2 2 2 b = 1537 s/mm b = 2042 s/mm b = 3051 s/mm b = 4561 s/mm

B

Figure 16. Representative diffusion-weighted 1H MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with eight different b values ranging from 5 to 4561 s/mm2 for the 1H ADC measurement. (B) Image of the isotonic fixed ganglia with a b value of 524 s/mm2 (left) and the corresponding ROIs (right).

57

TR = 20 ms TR = 50 ms TR = 100 ms A

TR = 200 ms TR = 250 ms

B

C

Figure 17. Representative 23Na MR images of three fixed ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with five different TR values ranging from 20 to 250 ms for the T1 measurement and five different TE values ranging from 3.5 to 21 ms for the T2* measurement (not shown). (B) Images of the isotonic fixed ganglia with a TR/TE of 250/3.5 ms 23 for the Na T1 measurement (left) and the corresponding ROIs (right). (C) Images of the isotonic 23 fixed ganglia with a TR/TE of 100/3.5 ms for the Na T2* measurement (left) and the corresponding ROIs (right).

58

Figure 18. Representative regression curve fits for the isotonic fixed ganglia. All relaxation and diffusion curves were fit to a single exponential function. The curve fits are as follows: (A) proton T1, (B) proton T2, (C) proton T2*, (D) proton ADC, (E) sodium T1, and (F) sodium T2*.

59

Table 9. Proton and sodium relaxation and diffusion measurements for the isotonic fixed ganglia. The total sample size was three ganglia and all curve fits had an adjusted R2 >0.95. The average ± a single standard deviation are reported for all measurements.

1H ADC / 1H T / ms 1H T / ms 1H T * / ms 23Na T / ms 23Na T * / ms 1 2 2 (µm2/ms) 1 2 Ganglia 1 1198.8 41.63 31.67 1.15 35.98 9.78 Ganglia 2 1078.9 40.06 29.63 1.12 33.76 8.43 Ganglia 3 1080.5 32.94 27.79 0.97 31.42 7.68 Average 1119.4 ± 68.8 38.21 ± 4.63 29.70 ± 1.94 1.08 ± 0.10 33.72 ± 2.28 8.63 ± 1.06

5.1.2 Hypertonic Fixed Ganglia A second sample of three fixed ganglia was imaged in hypertonic ASW (575 mM NaCl). Again, these ganglia had been previously fixed and therefore membrane integrity was expected to be lost. Although no active transport occurs, an initial osmotic gradient was established with hypertonic ASW which has a higher extracellular sodium content that should equilibrate with the ganglia since membrane integrity is compromised. However, minor crenation of the ganglia is expected. A complete T1 dataset as well as the ROIs are presented in Figure 19. The hypertonic fixed ganglia had a short T1 relaxation parameter as evident by higher signal intensity within the ganglia at short repetition times (TR). However, although shorter than the viable ganglia, the T1 in the hypertonic fixed ganglia did not deviate from the isotonic fixed ganglia. A complete T2 and T2* dataset with the ROIs outlined are shown in Figure 20 and Figure 21, respectively. In the hypertonic samples, even less T2 and T2* contrast was seen at shorter TE times, indicating a longer T2 relaxation parameter. The T2 (46 ms) and T2* (35 ms) were higher than for the isotonic fixed ganglia. A complete ADC dataset with the ROIs outlined is shown in Figure 22. Diffusion- weighted imaging was again able to observe the gangliar regions and due to the orientation of the sample observe the bilobe structure of the ganglia, showing the two different lobes connected by a thin commissure in the middle. The ADC was found to be lower than for the isotonic fixed ganglia. Sodium imaging (Figure 23) showed that the 23Na signal intensity was higher than that observed in the isotonic fixed ganglia but there was little change in T1 and T2* relaxation. This indicates that there was a higher sodium content in the ganglia, consistent with a loss of membrane integrity and diffusion of the hypertonic ASW into the gangliar regions that increased the total 23Na concentration. Again, all relaxation and diffusion measurements were fit to a

60 monoexponential function as shown in Figure 24 and no biexponential behavior was observed showing that cell crenation within the ganglia did not lead to biexponential curve fits.

Quantitative values of the proton T1, T2, T2* and ADC as well as the sodium T1 and T2* for the hypertonic fixed ganglia are shown in Table 10.

A TR = 650 ms TR = 800 ms TR = 1000 ms TR = 1500 ms

TR = 2000 ms TR = 3000 ms TR = 5000 ms TR = 7500 ms

B

1 Figure 19. Representative T1-weighted H MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with eight different TR values ranging from 650 to 1 7500 ms for the H T1 measurement. (B) Image of the hypertonic fixed ganglia with a TR of 7500 ms (left) and the corresponding ROIs (right).

61

A TE = 14 ms TE = 25 ms TE = 40 ms

TE = 60 ms TE = 100 ms

B

1 Figure 20. Representative T2-weighted H MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with five different TE values ranging from 14 to 1 100 ms for the H T2 measurement. (B) Image of the hypertonic fixed ganglia with a TE of 25 ms (left) and the corresponding ROIs (right).

62

A TE = 6 ms TE = 18 ms TE = 30 ms

TE = 42 ms TE = 54 ms TE = 66 ms

B

1 Figure 21. Representative T2*-weighted H MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with six different TE values ranging from 6 1 to 66 ms for the H T2* measurement. (B) Image of the hypertonic fixed ganglia with a TE of 18 ms (left) and the corresponding ROIs (right).

63

2 2 2 2 A b = 5 s/mm b = 524 s/mm b = 778 s/mm b = 1031 s/mm

2 2 2 2 b = 1537 s/mm b = 2042 s/mm b = 3051 s/mm b = 4561 s/mm

B

Figure 22. Representative diffusion-weighted 1H MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with eight different b values ranging from 5 to 4561 s/mm2 for the 1H ADC measurement. (B) Image of the hypertonic fixed ganglia with a b value of 524 s/mm2 (left) and the corresponding ROIs (right).

64

TR = 20 ms TR = 50 ms TR = 100 ms A

TR = 200 ms TR = 250 ms

B

C

Figure 23. Representative 23Na MR images of three fixed ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with five different TR values ranging from 20 to 250 ms for the T1 measurement and five different TE values ranging from 3.5 to 21 ms for the T2* measurement (not shown). (B) Images of the hypertonic fixed ganglia with a TR/TE of 250/3.5 23 ms for the Na T1 measurement (left) and the corresponding ROIs (right). (C) Images of the 23 hypertonic fixed ganglia with a TR/TE of 100/3.5 ms for the Na T2* measurement (left) and the corresponding ROIs (right).

65

Figure 24. Representative regression curve fits for the hypertonic fixed ganglia. All relaxation and diffusion curves were fit to a single exponential function. The curve fits are as follows: (A) proton T1, (B) proton T2, (C) proton T2*, (D) proton ADC, (E) sodium T1, and (F) sodium T2*.

66

Table 10. Proton and sodium relaxation and diffusion measurements for the hypertonic fixed ganglia. The total sample size was three ganglia and all curve fits had an adjusted R2 >0.95. The average ± a single standard deviation are reported for all measurements. 1H ADC / 1H T / ms 1H T / ms 1H T * / ms 23Na T / ms 23Na T * / ms 1 2 2 (µm2/ms) 1 2 Ganglia 1 1256.6 48.01 36.90 0.76 33.08 6.64 Ganglia 2 1105.0 40.65 32.08 0.66 39.09 8.06 Ganglia 3 1473.0 49.85 35.91 0.84 48.12 13.77 Average 1278.2 ± 184.9 46.17 ± 4.87 34.96 ± 2.55 0.75 ± 0.09 40.10 ± 7.57 9.49 ± 3.77

5.1.3 Hypotonic Fixed Ganglia A third sample of three fixed ganglia was imaged in hypotonic ASW (345-mM NaCl). As a fixed tissue, no active transport occurs; however, an initial osmotic gradient was established with a lower extracellular sodium content that should equilibrate with the ganglia that results in minor swelling. A complete T1 dataset as well as the ROIs is presented in Figure 25. The hypotonic fixed ganglia T1 relaxation parameter did not change from the other tonicities in the fixed ganglia and again a high signal intensity from the ganglia was seen at short TR times indicating a much shorter T1 than the surrounding ASW. Complete T2 and T2* datasets with the ROIs outlined are shown in Figure 26 and Figure 27, respectively. Due to the washing procedure with the fixed tissue, it became difficult to avoid air bubbles during sample preparation. The air bubbles can be seen clearly in these images as large dark voids that cause magnetic field distortions. In the hypotonic samples, it was again seen that even less T2 and T2* contrast developed at shorter TE times, much like the hypertonic fixed ganglia. The T2 (45 ms) and T2* (33 ms) values were higher than for the isotonic fixed ganglia and were nearly identical to those seen in the hypertonic condition. A complete ADC dataset with the ROIs outlined is shown in Figure 28. Diffusion-weighted imaging was again able to observe the gangliar regions; however, the ADC was higher than the hypertonic fixed ganglia but was more similar to the isotonic condition. Sodium imaging (Figure 29) showed that the 23Na signal intensity was lower than that observed in the isotonic fixed ganglia, but there was little change in T1 and T2* relaxation. Due to the loss of membrane integrity, the hypotonic ASW diffused into the gangliar regions decreasing the total 23Na concentration. As seen in Figure 30, monoexponential fits were used for all relaxation and diffusion measurements showing that cell swelling did not lead to

67 biexponential curve fits. Quantitative values of the proton T1, T2, T2* and ADC as well as the sodium T1 and T2* for the hypotonic fixed ganglia are shown in Table 11.

A TR = 650 ms TR = 800 ms TR = 1000 ms TR = 1500 ms

TR = 2000 ms TR = 3000 ms TR = 5000 ms TR = 7500 ms

B

1 Figure 25. Representative T1-weighted H MR images of three fixed ganglia in hypotonic ASW acquired at 11.7 T. (A) Images were acquired with eight different TR values ranging from 650 to 1 7500 ms for the H T1 measurement. (B) Image of the hypotonic fixed ganglia with a TR of 7500 ms (left) and the corresponding ROIs (right).

68

A TE = 14 ms TE = 25 ms TE = 40 ms

TE = 60 ms TE = 100 ms

B

1 Figure 26. Representative T2-weighted H MR images of three fixed ganglia in hypotonic ASW acquired at 11.7 T. (A) Images were acquired with five different TE values ranging from 14 to 1 100 ms for the H T2 measurement. (B) Image of the hypotonic fixed ganglia with a TE of 25 ms (left) and the corresponding ROIs (right).

69

A TE = 6 ms TE = 18 ms TE = 30 ms

TE = 42 ms TE = 54 ms TE = 66 ms

B

1 Figure 27. Representative T2*-weighted H MR images of three fixed ganglia in hypotonic ASW acquired at 11.7 T. (A) Images were acquired with six different TE values ranging from 6 to 66 1 ms for the H T2* measurement. (B) Image of the hypotonic fixed ganglia with a TE of 18 ms (left) and the corresponding ROIs (right).

70

2 2 2 2 A b = 5 s/mm b = 524 s/mm b = 778 s/mm b = 1031 s/mm

2 2 2 2 b = 1537 s/mm b = 2042 s/mm b = 3051 s/mm b = 4561 s/mm

B

Figure 28. Representative diffusion-weighted 1H MR images of three fixed ganglia in hypotonic ASW acquired at 11.7 T. (A) Images were acquired with eight different b values ranging from 5 to 4561 s/mm2 for the 1H ADC measurement. (B) Image of the hypotonic fixed ganglia with a b value of 524 s/mm2 (left) and the corresponding ROIs (right).

71

TR = 20 ms TR = 50 ms TR = 100 ms A

TR = 200 ms TR = 250 ms

B

C

Figure 29. Representative 23Na MR images of three fixed ganglia in hypotonic ASW acquired at 11.7 T. (A) Images were acquired with five different TR values ranging from 20 to 250 ms for the T1 measurement and five different TE values ranging from 3.5 to 21 ms for the T2* measurement (not shown). (B) Images of the hypotonic fixed ganglia with a TR/TE of 250/3.5 23 ms for the Na T1 measurement (left) and the corresponding ROIs (right). (C) Images of the 23 hypotonic fixed ganglia with a TR/TE of 100/3.5 ms for the Na T2* measurement (left) and the corresponding ROIs (right).

72

Figure 30. Representative regression curve fits for the hypotonic fixed ganglia. All relaxation and diffusion curves were fit to a single exponential function. The curve fits are as follows: (A) proton T1, (B) proton T2, (C) proton T2*, (D) proton ADC, (E) sodium T1, and (F) sodium T2*.

73

Table 11. Proton and sodium relaxation and diffusion measurements for the hypotonic fixed ganglia. The total sample size was three ganglia and all curve fits had an adjusted R2 >0.95. The average ± a single standard deviation are reported for all measurements. 1H ADC / 1H T / ms 1H T / ms 1H T * / ms 23Na T / ms 23Na T * / ms 1 2 2 (µm2/ms) 1 2 Ganglia 1 1057.4 40.58 28.07 1.38 35.25 8.90 Ganglia 2 1074.6 44.26 31.37 0.87 32.28 9.43 Ganglia 3 1507.6 49.31 38.15 1.36 31.26 7.75 Average 1213.2 ± 255.1 44.72 ± 4.38 32.53 ± 5.14 1.20 ± 0.29 32.93 ± 2.07 8.69 ± 0.86

5.1.4 Fixed Ganglia Summary Fixed ganglia in isotonic, 20% hypertonic, and 20% hypotonic ASW were imaged, and 1H and 23Na relaxation and diffusion properties were quantified. A summary of all relaxation and diffusion measurements are shown in Table 12 along with the statistical significance of the measured parameters using a one-way ANOVA with LSD post hoc. No changes were seen between tonicities with regard to the proton T1 relaxation parameter. The T2 and T2* relaxation parameters were higher under osmotic perturbation, but only the hypertonic fixed ganglia was found to be significant (p = 0.032). A decreasing trend in ADC was observed with increasing tonicity, but only significant differences between the hypertonic and hypotonic fixed ganglia was found (p = 0.050). Increasing trends in 23Na relaxation were observed; however, no statistical significance was reached. It must be noted that these trends are based on small sample sizes (three ganglia in each osmotic condition), and the trends observed may not be completely descriptive of the cellular changes that are occurring. The data supports the finding that T1 is an insensitive parameter to cellular changes; however, T2 and T2* are much more sensitive to cell volume changes with osmotic perturbation. The T2 and T2* relaxation parameters are indicative of the interactions between nuclear spins within the tissue environment and, therefore, changes in cell volume that alter these interactions and the measured relaxation.

74

Table 12. Summary of relaxation and diffusion trends in fixed ganglia at varying tonicities. (a) Quantitative values of the 1H and 23Na relaxation and diffusion measurements (a indicates significance to the isotonic fixed ganglia, b indicates significance to the hypotonic fixed ganglia, and c indicates significance to the hypertonic fixed ganglia). (b) P values for comparison between the different tonicities using a one-way ANOVA with LSD post hoc (* indicates significance with p < 0.05).

1H ADC / 1H T / ms 1H T / ms 1H T * / ms 23Na T / ms 23Na T * / ms 1 2 2 (µm2/ms) 1 2 Hypotonic 1213.2 ± 255.1 44.72 ± 4.38 32.53 ± 5.14 1.20 ± 0.29c 32.93 ± 2.07 8.69 ± 0.86 Isotonic 1119.4 ± 68.8 38.21 ± 4.63c 29.70 ± 1.94c 1.08 ± 0.10 33.72 ± 2.28 8.63 ± 1.06 Hypertonic 1278.2 ± 184.9 46.17 ± 4.87a 34.96 ± 2.55a 0.75 ± 0.09b 40.10 ± 7.57 9.49 ± 3.77 (a)

1 1 1 1 23 23 P value H T1 H T2 H T2* H ADC Na T1 Na T2*

Iso/Hypo 0.666 0.079 0.210 0.587 0.887 0.973

Iso/Hyper 0.465 0.032* 0.021* 0.152 0.256 0.641

Hypo/Hyper 0.765 0.692 0.281 0.050* 0.203 0.665 (b)

75

5.2 1H/23Na Imaging of Viable and Nonviable Ganglia Viable Aplysia ganglia were imaged in each of the osmotic conditions using the same acquisition parameters as for the fixed ganglia. The same ganglia were then imaged again after a 12-h delay, presumably allowing cell viability to be lost (nonviable ganglia).

5.2.1 Isotonic Viable/Nonviable Ganglia Imaging was performed on twelve viable ganglia in isotonic ASW. The ganglia were imaged immediately after dissection and all cellular metabolic processes were presumed to remain unaltered. Membrane integrity and the integrity of the ganglionic sac, which contains all the neural cells, remained intact as seen in the MR images as a clearly defined membrane at the periphery of the ganglion. Further, active transport processes were unaltered and allowed the ganglia to maintain osmotic gradients between the intracellular and extracellular environment. A 1 complete and representative H T1 dataset and the regions of interest (ROIs) used for analysis are presented in Figure 31. The T1 relaxation parameter was much longer than those seen in fixed 1 tissue. A complete H T2 and T2* dataset with the ROIs outlined are shown in Figure 32 and

Figure 33, respectively. T2 and T2* contrast allowed the identification of the large neurons and contrast was developed at short TE times due to a shorter T2 and T2* relaxation parameters in comparison to fixed ganglia. A complete 1H ADC dataset with the ROIs outlined is shown in Figure 34. Diffusion-weighted MR images at high b values allowed the direct identification of the ganglia from the ASW medium, seen as a bright region (high signal intensity) due to restricted diffusion within the ganglia and a lower diffusion coefficient. However, the ADC was not significantly different from the ADC in the fixed ganglia. Sodium imaging (Figure 35) showed strong correlation between the 1H and 23Na images as the outline of the ganglionic sac remained consistent between both 1H and 23Na images, although no cellular regions could be 23 23 identified on Na images. It can be noted that the Na T1 and T2* relaxation parameters were shorter in the ganglia than the ASW medium (similar to the fixed ganglia), yet the ganglia have a lower signal intensity (appear darker in the MR images) indicating that the sodium content is lower within the ganglia than in the ASW. Further, sodium signal intensity is not uniform throughout the ganglia, but higher signal intensity is seen in the interior of the ganglia. This observation implies that the neurons around the periphery remain viable as the osmotic gradients

76 are maintained with a lower intracellular sodium concentration. All relaxation and diffusion measurements were fit to a monoexponential function as shown in Figure 36. Little or no biexponential behavior was observed, as evidenced by high R2 values indicating that the monoexponential fits were adequate in fitting the data. Although compartmentation within the ganglia is expected, volume averaging over the entire ganglia revealed that plotting average signal intensities versus the different contrast parameters (TR, TE or b) led to monoexponential curve fits. Quantitative values of the proton T1, T2, T2* and ADC as well as the sodium T1 and

T2* for the isotonic viable ganglia are shown in Table 13, and a similar quantitative analysis on the isotonic nonviable ganglia is shown in Table 14.

77

A TR = 650 ms TR = 800 ms TR = 1000 ms TR = 1500 ms

TR = 2000 ms TR = 3000 ms TR = 5000 ms TR = 7500 ms

B

C

1 Figure 31. Representative T1-weighted H MR images of three viable ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with eight different TR values ranging from 650 to 1 7500 ms for the H T1 measurement. (B) Image of the isotonic viable ganglia with a TR of 7500 ms (left) and the corresponding ROIs (right). (C). Images of the isotonic nonviable ganglia with a TR of 7500 ms (left) and the corresponding ROIs (right).

78

A TE = 14 ms TE = 25 ms TE = 40 ms

TE = 60 ms TE = 100 ms

B

C

1 Figure 32. Representative T2-weighted H MR images of three viable ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with five different TE values ranging from 14 to 1 100 ms for the H T2 measurement. (B) Images of the isotonic viable ganglia with a TE of 14 ms (left) and the corresponding ROIs (right). (C) Images of the isotonic nonviable ganglia with a TE of 14 ms (left) and the corresponding ROIs.

79

TE = 6 ms TE = 18 ms TE = 30 ms A

TE = 42 ms TE = 54 ms TE = 66 ms

B

C

1 Figure 33. Representative T2*-weighted H MR images of three viable ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with six different TE values ranging from 6 to 66 1 ms for the H T2* measurement. (B) Images of the isotonic viable ganglia with a TE of 6 ms (left) and the corresponding ROIs (right). (C) Images of the isotonic nonviable ganglia (left) and the corresponding ROIs (right).

80

2 2 2 2 A b = 5 s/mm b = 524 s/mm b = 778 s/mm b = 1031 s/mm

2 2 2 2 b = 1537 s/mm b = 2042 s/mm b = 3051 s/mm b = 4561 s/mm

B

C

Figure 34. Representative diffusion-weighted 1H MR images of three viable ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with eight different b values ranging from 5 to 4561 s/mm2 for the 1H ADC measurement. (B) Images of the isotonic viable ganglia with a b value of 5 s/mm2 (left) and the corresponding ROIs (right). (C) Images of the isotonic nonviable ganglia with a b value of 524 s/mm2 (left) and the corresponding ROIs (right).

81

A TR = 20 ms TR = 50 ms TR = 100 ms

TR = 200 ms TR = 250 ms

B

C

Figure 35. Representative 23Na MR images of three nonviable ganglia in isotonic ASW acquired at 11.7 T. (A) Images were acquired with five different TR values ranging from 20 to 250 ms for the T1 measurement and five different TE values ranging from 3.5 to 21 ms for the T2* measurement (not shown). (B) Images of the isotonic nonviable ganglia with a TR/TE of 250/3.5 23 ms for the Na T1 measurement (left) and the corresponding ROIs (right). (C) Images of the isotonic nonviable ganglia with a TR/TE of 100/3.5 ms for the T2* measurement (left) and the corresponding ROIs (right).

82

Figure 36. Representative regression curve fits for the isotonic viable ganglia. The proton measurements are on viable ganglia while the sodium measurements are on nonviable ganglia. All relaxation and diffusion curves were fit to a single exponential function. The curve fits are as follows: (A) proton T1, (B) proton T2, (C) proton T2*, (D) proton ADC, (E) sodium T1, and (F) sodium T2*.

83

Table 13. Proton relaxation and diffusion measurements for the isotonic viable ganglia. The total sample size was 12 ganglia and all curve fits had an adjusted R2 >0.95. The average ± a single standard deviation are reported for all measurements. 1 1 1 1 2 H T1 / ms H T2 / ms H T2* / ms H ADC / (µm /ms) Ganglia 1 2154.2 20.84 12.97 1.03 Ganglia 2 2315.9 18.03 12.66 1.19 Ganglia 3 2142.3 22.75 14.22 1.37 Ganglia 4 2390.6 20.41 13.52 1.42 Ganglia 5 1885.4 20.82 12.45 1.56 Ganglia 6 2369.1 19.02 12.44 1.62 Ganglia 7 1714.7 19.00 10.41 0.86 Ganglia 8 1422.5 19.45 12.83 0.93 Ganglia 9 1846.0 18.30 12.32 0.79 Ganglia 10 2287.3 21.03 12.37 1.03 Ganglia 11 1767.7 18.14 13.52 0.51 Ganglia 12 1858.1 17.10 10.89 0.85 Average 2012.8 ± 307.3 19.57 ± 1.62 12.55 ± 1.06 1.10 ± 0.34

Table 14. Proton and sodium relaxation and diffusion measurements for the isotonic nonviable ganglia. The total sample size was 12 ganglia and all proton curve fits had an adjusted R2 >0.95 (sodium curve fits adjusted R2 >0.82). The average ± a single standard deviation are reported for all measurements. The missing data point is an outlier due to volume averaging with the ASW. 1H ADC / 1H T / ms 1H T / ms 1H T * / ms 23Na T / ms 23Na T * / ms 1 2 2 (µm2/ms) 1 2 Ganglia 1 2343.6 26.41 17.92 0.71 42.00 6.98 Ganglia 2 2463.1 23.25 16.89 0.81 48.19 8.58 Ganglia 3 2359.1 26.25 17.43 1.06 49.83 8.33 Ganglia 4 2604.2 25.88 18.05 1.18 50.10 5.66 Ganglia 5 2132.2 26.36 16.73 1.30 54.85 9.79 Ganglia 6 2670.9 22.72 16.69 1.35 48.78 - Ganglia 7 3185.7 43.61 24.98 0.92 35.37 13.26 Ganglia 8 2332.6 42.09 23.76 0.61 40.10 10.36 Ganglia 9 2663.1 33.93 24.22 0.75 35.50 10.48 Ganglia 10 2764.7 34.00 18.76 0.67 39.26 9.44 Ganglia 11 2091.2 28.83 18.49 0.51 51.49 13.65 Ganglia 12 2361.3 31.88 17.18 0.75 39.87 13.17 Average 2497.6 ± 301.8 30.43 ± 6.88 19.26 ± 3.13 0.89 ± 0.28 44.61 ± 6.65 9.97 ± 2.60

84

5.2.2 Hypertonic Viable/Nonviable Ganglia Imaging was performed on fifteen viable ganglia in hypertonic ASW. The ganglia were imaged immediately after dissection and all cellular metabolic processes were presumed to remain unaltered, allowing osmotic gradients to be maintained between the intracellular and extracellular environments. The integrity of the ganglionic sac could be seen in the MR images 1 as a clearly defined membrane at the periphery of the ganglion. A complete H T1 dataset and the regions of interest (ROIs) used for analysis are presented in Figure 37. The T1 did not change significantly from the isotonic viable ganglia; however, the T1 remained significantly longer than 1 the fixed ganglia. A complete H T2 and T2* dataset with the ROIs outlined are shown in Figure

38 and Figure 39, respectively. Both the T2 and T2* increased significantly (T2: p = 0.022; T2*: p < 0.001) in comparison to the isotonic viable ganglia and the relaxation parameters were much longer than the fixed ganglia. A complete 1H ADC dataset with the ROIs outlined is shown in Figure 40. Although diffusion-weighted MR images allowed the direct identification of the ganglia from the ASW medium due to a lower diffusion coefficient, the ADC was not significantly different from the ADC in the isotonic viable or fixed ganglia. Sodium imaging (Figure 41) showed strong correlation between the 1H and 23Na images as the outline of the ganglionic sac remained consistent between both 1H and 23Na images, although no cellular 23 23 regions could be identified on Na images. It can be noted that the Na T1 and T2* relaxation parameters were shorter in the ganglia than the ASW medium (similar to the fixed ganglia), yet the ganglia have a lower signal intensity (appear darker in the MR images) indicating that the sodium content is lower within the ganglia than in the ASW. Although a general decreasing trend was seen in 23Na relaxation at lower tonicity, no significant changes are seen with regards to sodium relaxation in nonviable ganglia. All relaxation and diffusion measurements were fit to a monoexponential function as shown in Figure 42. Little or no biexponential behavior was observed as evidenced by the high R2 values indicating that the monoexponential fits were adequate in fitting the data. Although compartmentation within the ganglia is expected, volume averaging over the entire ganglia revealed that plotting average signal intensities versus the different contrast parameters (TR, TE or b) led to monoexponential curve fits. Quantitative values of the proton T1, T2, T2* and ADC as well as the sodium T1 and T2* for the hypertonic viable ganglia are shown in Table 15 and a similar quantitative analysis on the hypertonic nonviable ganglia are shown in Table 16.

85

A TR = 650 ms TR = 800 ms TR = 1000 ms TRTR == 15001500 msms

TR = 2000 ms TR = 3000 ms TR = 5000 ms TRTR == 75007500 msms

B

C

1 Figure 37. Representative T1-weighted H MR images of three viable ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with eight different TR values ranging from 1 650 to 7500 ms for the H T1 measurement. (B) Image of the isotonic viable ganglia with a TR of 7500 ms (left) and the corresponding ROIs (right). (C). Images of the isotonic nonviable ganglia with a TR of 7500 ms (left) and the corresponding ROIs (right).

86

TE = 14 ms TE = 25 ms TE = 40 ms A

TE = 60 ms TE = 100 ms

B

C

1 Figure 38. Representative T2-weighted H MR images of three viable ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with five different TE values ranging from 1 14 to 100 ms for the H T2 measurement. (B) Images of the hypertonic viable ganglia with a TE of 14 ms (left) and the corresponding ROIs (right). (C) Images of the hypertonic nonviable ganglia with a TE of 14 ms (left) and the corresponding ROIs.

87

TE = 6 ms TE = 18 ms TE = 30 ms A

TE = 42 ms TE = 54 ms TE = 66 ms

B

C

1 Figure 39. Representative T2*-weighted H MR images of three viable ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with six different TE values ranging from 6 1 to 66 ms for the H T2* measurement. (B) Images of the hypertonic viable ganglia with a TE of 6 ms (left) and the corresponding ROIs (right). (C) Images of the hypertonic nonviable ganglia (left) and the corresponding ROIs (right).

88

2 2 2 2 A b = 5 s/mm b = 524 s/mm b = 778 s/mm b = 1031 s/mm

2 2 2 2 b = 1537 s/mm b = 2042 s/mm b = 3051 s/mm b = 4561 s/mm

B

C

Figure 40. Representative diffusion-weighted 1H MR images of three viable ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with eight different b values ranging from 5 to 4561 s/mm2 for the 1H ADC measurement. (B) Images of the hypertonic viable ganglia with a b value of 5 s/mm2 (left) and the corresponding ROIs (right). (C) Images of the hypertonic nonviable ganglia with a b value of 524 s/mm2 (left) and the corresponding ROIs (right).

89

TR = 20 ms TR = 50 ms TR = 100 ms A

TR = 200 ms TR = 250 ms

B

C

Figure 41. Representative 23Na MR images of three nonviable ganglia in hypertonic ASW acquired at 11.7 T. (A) Images were acquired with five different TR values ranging from 20 to 250 ms for the T1 measurement and five different TE values ranging from 3.5 to 21 ms for the T2* measurement (not shown). (B) Images of the hypertonic nonviable ganglia with a TR/TE of 23 250/3.5 ms for the Na T1 measurement (left) and the corresponding ROIs (right). (C) Images of the hypertonic nonviable ganglia with a TR/TE of 100/3.5 ms for the T2* measurement (left) and the corresponding ROIs (right).

90

Figure 42. Representative regression curve fits for the hypertonic viable ganglia. The proton measurements are on viable ganglia while the sodium measurements are on nonviable ganglia. All relaxation and diffusion curves were fit to a single exponential function. The curve fits are as follows: (A) proton T1, (B) proton T2, (C) proton T2*, (D) proton ADC, (E) sodium T1, and (F) sodium T2*.

91

Table 15. Proton relaxation and diffusion measurements for the hypertonic viable ganglia. The total sample size was 15 ganglia and all curve fits had an adjusted R2 >0.95. The average ± a single standard deviation are reported for all measurements. 1 1 1 1 2 H T1 / ms H T2 / ms H T2* / ms H ADC / (µm /ms) Ganglia 1 1847.4 15.73 7.55 1.36 Ganglia 2 1794.7 16.53 10.42 1.47 Ganglia 3 2417.2 21.92 10.46 2.02 Ganglia 4 2060.6 15.45 9.90 0.61 Ganglia 5 1641.8 18.74 9.75 0.73 Ganglia 6 1648.5 15.82 8.94 0.74 Ganglia 7 2183.9 18.41 9.64 1.01 Ganglia 8 1813.9 21.18 9.89 1.21 Ganglia 9 2976.2 16.75 10.46 1.42 Ganglia 10 2003.6 19.50 10.13 1.19 Ganglia 11 1882.9 16.49 9.64 1.06 Ganglia 12 2223.2 18.29 9.95 1.08 Ganglia 13 1885.7 19.14 9.99 1.36 Ganglia 14 1662.8 14.84 9.80 1.12 Ganglia 15 2265.0 14.05 9.15 1.44 Average 2020.5 ± 355.6 17.52 ± 2.31 9.71 ± 0.74 1.19 ± 0.35

92

Table 16. Proton and sodium relaxation and diffusion measurements for the hypertonic nonviable ganglia. The total sample size was 15 ganglia and all proton curve fits had an adjusted R2 >0.95 (sodium curve fits adjusted R2 >0.82). The average ± a single standard deviation are reported for all measurements. The missing data points are due to gradient failures caused by overheating. 1H ADC / 1H T / ms 1H T / ms 1H T * / ms 23Na T / ms 23Na T * / ms 1 2 2 (µm2/ms) 1 2 Ganglia 1 2153.3 23.93 10.58 1.31 36.02 10.27 Ganglia 2 1984.5 22.40 13.70 1.30 44.50 10.18 Ganglia 3 2451.6 23.42 12.76 1.82 40.58 13.17 Ganglia 4 2556.9 47.10 28.77 0.84 38.46 10.68 Ganglia 5 2168.3 36.72 21.51 0.98 48.88 12.64 Ganglia 6 2279.5 38.64 22.18 0.89 48.26 11.69 Ganglia 7 2313.7 31.58 16.79 0.78 - - Ganglia 8 1935.7 30.49 16.22 0.71 - - Ganglia 9 2413.1 29.39 16.15 1.24 - - Ganglia 10 2199.7 30.71 14.16 1.03 46.95 10.33 Ganglia 11 2061.4 33.53 16.43 1.02 49.12 10.82 Ganglia 12 2202.6 30.47 16.98 0.77 45.93 7.75 Ganglia 13 2354.6 36.34 21.53 1.13 45.19 10.35 Ganglia 14 2313.7 32.85 19.98 0.92 53.88 11.63 Ganglia 15 2702.7 28.09 15.73 0.92 46.97 9.78 Average 2272.8 ± 207.7 31.71 ± 6.41 17.56 ± 4.57 1.04 ± 0.29 45.40 ± 4.97 10.77 ± 1.41

5.2.3 Hypotonic Viable/Nonviable Ganglia Imaging was performed on fifteen viable ganglia in hypotonic ASW. The ganglia were imaged immediately after dissection, and all cellular metabolic processes were presumed to remain unaltered. Membrane integrity and the integrity of the ganglionic sac remained intact as seen in the MR images as a clearly defined membrane at the periphery of the ganglion. Further, active transport processes were unaltered and allowed the ganglia to maintain osmotic gradients between the intracellular and extracellular environment. A complete and representative T1 dataset and the regions of interest (ROIs) used for analysis are presented in Figure 43. The T1 did not change significantly from the isotonic viable ganglia; however, the T1 remained significantly longer than the fixed ganglia. Complete T2 and T2* datasets with the ROIs outlined are shown in Figure 44 and Figure 45, respectively. The T2 decreased but due to larger variations in the measurement was not significant (p = 0.240) while the T2* increased significantly (p = 0.009) in comparison to the isotonic viable ganglia. Both T2 and T2* relaxation

93 parameters were much longer than the fixed ganglia. A complete ADC dataset with the ROIs outlined is shown in Figure 46. Although diffusion-weighted MR images allowed the direct identification of the ganglia from the ASW medium due to a lower diffusion coefficient, the ADC was not significantly different from the ADC in the isotonic viable or fixed ganglia. Sodium imaging (Figure 47) showed strong correlation between the 1H and 23Na images as the outline of the ganglionic sac remained consistent between both 1H and 23Na images, although no 23 23 cellular regions could be identified on Na images. It can be noted that the Na T1 and T2* relaxation parameters were shorter in the ganglia than the ASW medium (similar to the fixed ganglia), yet the ganglia have a lower signal intensity (appear darker in the MR images) indicating that the sodium content is lower within the ganglia than in the ASW. Although a general increasing trend was seen in 23Na relaxation at higher tonicity, no significant changes are seen with regards to sodium relaxation in nonviable ganglia. All relaxation and diffusion measurements were fit to a monoexponential function as shown in Figure 48. Little or no biexponential behavior was observed as evidenced by the high R2 values indicating that the monoexponential fits were adequate in fitting the data. Although compartmentation within the ganglia is expected, volume averaging over the entire ganglia revealed that plotting average signal intensities versus the different contrast parameters (TR, TE, b) led to monoexponential curve fits. Quantitative values of the proton T1, T2, T2* and ADC as well as the sodium T1 and

T2* for the hypertonic viable ganglia are shown in Table 17 and a similar quantitative analysis on the hypertonic nonviable ganglia are shown in Table 18.

94

A TR = 650 ms TR = 800 ms TR = 1000 ms TR = 1500 ms

TR = 2000 ms TR = 3000 ms TR = 5000 ms TR = 7500 ms

B

C

Figure 43. Representative 1H MR images of three viable ganglia in hypotonic ASW acquired at 1 11.7 T with eight different TR values ranging from 650 to 7500 ms for the H T1 measurement (A). Images of the viable ganglia (B) and the nonviable ganglia (C) with a TR of 7500 ms and the corresponding ROIs.

95

TE = 14 ms TE = 25 ms TE = 40 ms A

TE = 60 ms TE = 100 ms

B

C

Figure 44. Representative 1H MR images of three viable ganglia in hypotonic ASW acquired at 1 11.7 T with five different TE values ranging from 14 to 100 ms for the H T2 measurement (A). Images of the viable ganglia with a TE of 14 ms (B) and the nonviable ganglia with a TE of 25 ms (C) and the corresponding ROIs.

96

TE = 6 ms TE = 18 ms TE = 30 ms A

TE = 42 ms TE = 54 ms TE = 66 ms

B

C

Figure 45. Representative 1H MR images of three viable ganglia in hypotonic ASW acquired at 1 11.7 T with six different TE values ranging from 6 to 66 ms for the H T2* measurement (A). Images of the viable ganglia with a TE of 6 ms (B) and the nonviable ganglia with a TE of 18 ms (C) and the corresponding ROIs.

97

2 2 2 2 A b = 5 s/mm b = 524 s/mm b = 778 s/mm b = 1031 s/mm

2 2 2 2 b = 1537 s/mm b = 2042 s/mm b = 3051 s/mm b = 4561 s/mm

B

C

Figure 46. Representative 1H MR images of three viable ganglia in hypotonic ASW acquired at 11.7 T with eight different b values ranging from 5 to 4561 s/mm2 for the 1H ADC measurement (A). Images of the viable ganglia with a b value of 5 s/mm2 (B) and the nonviable ganglia with a b value of 524 s/mm2 (C) and the corresponding ROIs.

98

TR = 20 ms TR = 50 ms TR = 100 ms A

TR = 200 ms TR = 250 ms

B

C

Figure 47. Representative 23Na MR images of three nonviable ganglia in hypotonic ASW acquired at 11.7 T with five different TR values ranging from 20 to 250 ms (A). Images of the 23 nonviable ganglia with a TR/TE of 250/3.5 ms for the Na T1 measurement (B) and the 23 nonviable ganglia with a TR/TE of 100/3.5 ms for the Na T2* measurement (C) and the corresponding ROIs.

99

Figure 48. Representative regression curve fits for the hypotonic viable ganglia. The proton measurements are on viable ganglia while the sodium measurements are on nonviable ganglia. All relaxation and diffusion curves were fit to a single exponential function. The curve fits are as follows: (A) proton T1, (B) proton T2, (C) proton T2*, (D) proton ADC, (E) sodium T1, and (F) sodium T2*.

100

Table 17. Proton relaxation and diffusion measurements for the hypotonic viable ganglia. The total sample size was 15 ganglia and all curve fits had an adjusted R2 >0.95. The average ± a single standard deviation are reported for all measurements. * indicated significance (p < 0.05) in comparison to the isotonic viable ganglia using a difference of means t-test.

1 1 1 1 2 Region H T1 / ms H T2 / ms H T2* / ms H ADC / (µm /ms) Ganglia 1 1914.2 25.93 16.80 1.29 Ganglia 2 1814.6 23.52 16.86 1.00 Ganglia 3 2045.4 23.23 17.55 1.23 Ganglia 4 1991.2 22.95 16.16 1.16 Ganglia 5 1897.2 21.29 15.57 1.14 Ganglia 6 1902.2 22.90 16.73 0.86 Ganglia 7 1980.2 24.47 15.75 1.36 Ganglia 8 1945.9 24.34 17.75 1.24 Ganglia 9 1633.2 20.85 17.68 1.53 Ganglia 10 2224.2 22.71 15.68 1.17 Ganglia 11 1627.9 23.60 16.56 1.25 Ganglia 12 1933.9 24.25 18.11 1.24 Ganglia 13 2129.0 25.65 15.63 1.53 Ganglia 14 1993.2 22.75 16.66 0.99 Ganglia 15 2318.0 26.06 15.10 1.53 Average 1956.7 ± 186.1 23.63 ± 1.53* 16.57 ± 0.92* 1.23 ± 0.20

101

Table 18. Proton and sodium relaxation and diffusion measurements for the hypertonic nonviable ganglia. The total sample size was 15 ganglia and all proton curve fits had an adjusted R2 >0.95 (sodium curve fits adjusted R2 >0.90). The average ± a single standard deviation are reported for all measurements. The missing data points are due to gradient failures caused by overheating.

1H ADC / Region 1H T / ms 1H T / ms 1H T * / ms 23Na T / ms 23Na T * / ms 1 2 2 (µm2/ms) 1 2 Ganglia 1 2670.9 48.57 28.12 0.59 47.53 8.43 Ganglia 2 2204.6 39.65 24.03 0.58 50.56 9.16 Ganglia 3 2807.4 41.89 26.93 0.75 49.73 14.32 Ganglia 4 2518.9 40.37 27.53 0.87 - - Ganglia 5 2551.7 37.99 27.78 0.75 - - Ganglia 6 2532.3 43.07 31.64 0.60 - - Ganglia 7 2710.8 47.92 30.62 0.96 35.92 12.66 Ganglia 8 2354.6 47.46 33.44 1.06 33.01 12.05 Ganglia 9 2779.3 37.81 25.48 1.28 21.05 11.22 Ganglia 10 2603.5 35.83 23.60 0.82 47.24 - Ganglia 11 1872.3 32.01 22.17 0.82 45.64 - Ganglia 12 2499.4 34.71 26.43 0.87 41.27 - Ganglia 13 2524.0 38.87 25.21 1.06 - 9.03 Ganglia 14 2558.9 41.12 27.03 0.85 - 8.61 Ganglia 15 2670.2 38.57 28.26 1.26 - 5.96 Average 2523.9 ± 237.2 40.39 ± 4.82 27.22 ± 3.03 0.87 ± 0.22 41.33 ± 9.73 10.16 ± 2.59

5.2.4 Viable/Nonviable Ganglia Summary

Viable ganglia in isotonic, 20% hypertonic and 20% hypotonic ASW were imaged and 1H and 23Na relaxation and diffusion properties were quantified. A summary of all relaxation and diffusion measurements for the viable ganglia are shown in Table 19 along with the statistical significance of the measured parameters using a one-way ANOVA with LSD post hoc. No

changes were seen between tonicities with regard to the proton T1 relaxation parameter. The T2

and T2* relaxation parameters decreased as the extracellular tonicity was increased. Significance

was achieved for both perturbations in the T2* measurement, but only the hypotonic viable

ganglia were found to be significant in the T2 measurement in comparison to the isotonic viable ganglia. No clear trend in ADC was observed with increasing tonicity. The data supports the

finding that T1 is an insensitive parameter to cellular changes, similar to the fixed ganglia.

However, T2 and T2* are much more sensitive to cell volume changes with osmotic perturbation, 102 which may to be due to changing microenvironments within the ganglia as cell volumes are changing. The intracellular volume fraction and membrane permeability are factors that are altered with osmotic perturbation and may impact MR relaxation and diffusion measurements. 1 Although not measured in previous studies, the H T2* appears to be the most consistent measurement in identifying changes due to osmotic perturbation. A summary of all relaxation and diffusion measurements for the nonviable viable ganglia are shown in Table 20 along with the statistical significance of the measured parameters using a one-way ANOVA with LSD post hoc. The trends observed are similar to the viable ganglia, although the T1 relaxation parameter was also significant. All relaxation parameters increased upon loss of cell viability, and the absolute value of the relaxation parameter was dependent on how long after dissection it was measured. Relaxation measurements remained consistent for the first 7-8 hours after dissection, and then began to increase. The ADC decreased upon loss of cell viability for most ganglia, although no significance was reached. Increasing trends in 23Na relaxation were observed although was not statistically significant. However, 23Na relaxation times consistently were longer in viable ganglia in comparison to fixed ganglia and a generally increasing trend in sodium relaxation with increasing tonicity was observed.

103

Table 19. Summary of relaxation and diffusion trends in viable ganglia at varying tonicities. (a) Quantitative values of the 1H and 23Na relaxation and diffusion measurements (a indicates significance to the isotonic fixed ganglia, b indicates significance to the hypotonic fixed ganglia, and c indicates significance to the hypertonic fixed ganglia). (b) P values for comparison between the different tonicities using a one-way ANOVA with LSD post hoc (* indicates significance with p < 0.05).

1H ADC / 1H T / ms 1H T / ms 1H T * / ms 1 2 2 (µm2/ms) ac ac Hypotonic 1956.7 ± 186.1 23.63 ± 1.53 16.57 ± 0.92 1.23 ± 0.20 b bc Isotonic 2012.8 ± 307.3 19.57 ± 1.62 12.55 ± 1.06 1.10 ± 0.34 b ab Hypertonic 2020.5 ± 355.6 17.52 ± 2.31 9.71 ± 0.74 1.19 ± 0.35 (a)

1 1 1 1 P value H T1 H T2 H T2* H ADC

Iso/Hypo 0.586 0.022* <0.001* 0.202

Iso/Hyper 0.941 0.240 0.009* 0.397

Hypo/Hyper 0.511 <0.001* <0.001* 0.645 (b)

104

Table 20. Summary of relaxation and diffusion trends in nonviable ganglia at varying tonicities. (a) Quantitative values of the 1H and 23Na relaxation and diffusion measurements (a indicates significance to the isotonic fixed ganglia, b indicates significance to the hypotonic fixed ganglia, and c indicates significance to the hypertonic fixed ganglia). (b) P values for comparison between the different tonicities using a one-way ANOVA with LSD post hoc (* indicates significance with p < 0.05).

1H ADC / 1H T / ms 1H T / ms 1H T * / ms 23Na T / ms 23Na T * / ms 1 2 2 (µm2/ms) 1 2 c ac ac Hypotonic 1213.2 ± 255.1 44.72 ± 4.38 32.53 ± 5.14 1.20 ± 0.29 32.93 ± 2.07 8.69 ± 0.86 c b b Isotonic 1119.4 ± 68.8 38.21 ± 4.63 29.70 ± 1.94 1.08 ± 0.10 33.72 ± 2.28 8.63 ± 1.06 ab b b Hypertonic 1278.2 ± 184.9 46.17 ± 4.87 34.96 ± 2.55 0.75 ± 0.09 40.10 ± 7.57 9.49 ± 3.77 (a)

1 1 1 1 23 23 P value H T1 H T2 H T2* H ADC Na T1 Na T2*

Iso/Hypo 0.799 <0.001* <0.001* 0.923 0.278 0.853

Iso/Hyper 0.031* 0.464 0.115 0.142 0.778 0.397

Hypo/Hyper 0.011* <0.001* <0.001* 0.098 0.181 0.538 (b)

105

5.3 Comparison of Viable and Fixed Ganglia Through the imaging of fixed, viable, and nonviable ganglia in hypotonic, isotonic, and hypertonic extracellular environments, comparisons between the different conditions could be 1 achieved. A compete summary of the H T1 relaxation parameter trends can be seen in Figure 1 1 49. The H T2 and T2* trends can be seen in Figure 50 and Figure 51, respectively. The H ADC trends are shown in Figure 52. All relaxation parameters increased upon loss of cell viability and the absolute value of the relaxation parameter was dependent on how long after dissection it was measured. The ADC decreased upon loss of cell viability for most ganglia, although no significance was reached. Although no viable data was recorded for 23Na relaxation on viable ganglia to the extensive time it took to acquire the data, comparisons between nonviable and fixed ganglia were 23 performed. A compete summary of the Na T1 relaxation parameter trends can be seen in Figure 23 23 53 and the Na T2* trends can be seen in Figure 54. Increasing trends in Na relaxation were observed; however, no statistical significance was reached. However, 23Na relaxation was consistently longer in viable ganglia in comparison to fixed ganglia. Although not significant, the SNR of the first sodium acquisition was calculated and it was seen that the SNR was significantly higher for the hypertonic ganglia in comparison to the isotonic condition as seen in Figure 55. This indicates that at the time of sodium imaging, cell membrane integrity was most likely lost and the high sodium content of the hypertonic ASW had begun to diffuse into the ganglia, indicating that viability had already been lost. To assess the changes that are occurring in viable ganglia at short times with regard to sodium, a single ganglion was dissected and placed in ASW of varying tonicities as seen in Figure 56. Although the relaxation parameters were not measured, the signal intensity was quantified over a 48 hour period. It can be seen that the 23Na signal intensity increases linearly with time after dissection and the slope of this rise in signal intensity differs for each of the osmotic conditions. Further, the changes in 23Na signal intensity begin within the first hours after dissection indicating that although the ganglia may remain viable for a longer time period, drastic alterations in the cellular environment begin much earlier.

106

Figure 49. Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 1 hypertonic extracellular tonicities for the H T1 measurement. Significance was reached within all conditions between viable, nonviable, and fixed ganglia.

Figure 50. Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 1 hypertonic extracellular tonicities for the H T2 measurement. * indicates significance (p < 0.05) using a one-way ANOVA with LSD post hoc on the viable ganglia measurements. Significance was reached within all conditions between viable, nonviable, and fixed ganglia (except between the hypotonic fixed and nonviable ganglia).

107

Figure 51. Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 1 hypertonic extracellular tonicities for the H T2* measurement. * indicates significance (p < 0.05) using a one-way ANOVA with LSD post hoc on the viable ganglia measurements. Significance was reached within all conditions between viable, nonviable, and fixed ganglia.

Figure 52. Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and hypertonic extracellular tonicities for the 1H ADC measurement. The only significant differences were between hypotonic viable/nonviable and hypertonic viable/fixed.

108

Figure 53. Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 23 hypertonic extracellular tonicities for the Na T1 measurement. The only significant difference was between the isotonic fixed/nonviable ganglia.

Figure 54. Comparisons of fixed, viable, and nonviable ganglia between hypotonic, isotonic, and 23 hypertonic extracellular tonicities for the Na T2* measurement.

109

Figure 55. Comparisons of 23Na SNR in nonviable ganglia between hypotonic, isotonic, and hypertonic extracellular tonicities. SNR was measured in the first 23Na image in the T1 dataset (TE/TR: 3.5/50 ms) acquired 26 hours after dissection. * indicates significance (p < 0.05) using a one-way ANOVA with LSD post hoc.

Figure 56. 23Na images and signal intensity of a single viable ganglion in isotonic, hypertonic, and hypotonic ASW. Left: 23Na 3D GRE MR Images (TE/TR: 2.7/20 ms) of an abdominal ganglion in isotonic ASW 2 hours post dissection (A) and 48 hours post dissection (B). Right: Graph of 23Na signal intensity as a function of time after dissection illustrating the rise in ganglionic 23Na signal intensity (ASW control was unchanged). 23Na signal intensities from the ROIs were normalized to the initial time point.

110

CHAPTER SIX

DISCUSSION

Single cell imaging has focused on utilizing isolated neurons from the sea slug Aplysia californica as a model organism due to the large size of the individual neurons, which have a diameter of 300 μm. This size is significantly larger than average mammalian cells which have diameters on the order of 10 μm and exist below the resolution currently available to magnetic resonance techniques. However, it can be argued that the principles of cell biology measured in single cells can be used to understand and predict the tissue response to pathological conditions (Benveniste & Blackband, 2002). Similar to isolated neurons, Aplysia neural ganglia show that

T2 is more sensitive than ADC in identifying changes to osmotic perturbation. Hypotonic perturbation led to increases in T2, which can be accounted for by cell swelling and increases in the intracellular volume fraction. Even in single cells, the ADC was found to be highly variable (Hsu, et al. 1996), and no characteristic changes were seen with osmotic perturbation suggesting that modifications in isotropic diffusion (restrictions, hindrances and compartment exchange) exist even at the level of the single cell and are carried through to the entirety of the ganglia. In contrast to previous studies utilizing the single cell, no active perfusion was used between successive images of the ganglia. Although limiting changes in orientation due to perfusion that could impact relaxation and diffusion measurements, the ganglia are in a progressively worsening state of depleted oxygen and nutrients, which may impact cell viability. However, the viability of Aplysia neurons in unperfused environments have be shown previously through Trypan dye exclusion tests of cell membrane integrity to be maintained for up to seven hours 1 after dissection (Schoeniger, et al 1994). On single cells, it was shown that the H T2 and ADC remain constant throughout this time window; however, upon loss of membrane integrity, T2 increases significantly. In the MR analysis of the entire neural ganglia, 1H relaxation measurements did remain constant for the first several hours after dissection and were classified as viable. By imaging the ganglia a second time, twelve hours after the initial proton measurements and well outside the documented viability window, it was shown that in all 1 osmotic conditions H T1, T2 and T2* increased significantly, while the ADC generally decreased. These efforts show that cell viability impacts MR diffusion and relaxation measurements. Although MR relaxation and diffusion changes occur within the first 24 hours

111 after dissection, the unperfused ganglia notably maintain electrical activity for up to 48 hours after dissection (Novak and Wheeler, 1986). Based on these previous reports and the current 1H data, it would appear that the isolated ganglion maintains its integrity and viability for a time window approaching twelve hours after which 1H relaxation and diffusion changes indicate significant alteration in the cellular microenvironment. However, the current study displays significant changes is 23Na signal intensity that begin within the first two hours of dissection, indicating that cellular changes can be detected at much earlier time points than 1H MRI or even electrophysiology studies. Although 23Na relaxation did not change significantly with osmotic perturbation in nonviable ganglia, there was a generally increasing trend in relaxation with increasing extracellular tonicity. The confounding influences on measured 23Na signal intensity resulting from a signal loss associated with increasing T1 and signal gain associated with an increasing T2* are not solely responsible for the large increase in measured 23Na signal intensity. The loss of true cellular viability appears to be gradual, and although the ionic distributions of sodium may be changing and active ionic transport may be affected, the semi-impermeable nature of the cell membrane is likely maintained for longer time periods. Fixative also appears to alter MR relaxation and diffusion as fixed Aplysia ganglia consistently exhibited longer T1, T2 and T2* relaxation parameters and a generally decreasing ADC upon loss of viability. In studies of fixed tissue (Shepherd, et al. 2009) and single cells (Purea & Webb, 2006), the T1 generally decreased with fixative (even after washing) while the T2 drastically decreased with fixative but regained normal relaxation after washing. It would be expected that T2 measurements between fixed and viable tissue to be similar; however, this is not supported with the Aplysia ganglia as T2 remained significantly longer in nonviable ganglia. For ADC measurements, diffusion was assumed to be isotropic and measured at short diffusion times. Although diffusion anisotropy was not considered, in single cell studies of Aplysia neurons, no significant difference between the read and phase direction in the 1H imaging of single cells was observed (Hsu, et al. 1997). However, there were data points that were found to show statistically significant anisotropy, but this issue has not been explored further. It can be expected that diffusion anisotropy exists in any tissue environment as membranes and other cellular components provide barriers to diffusion and the distribution of transport proteins and leakage channels in the cell membrane might contribute to preferential

112 diffusion in a single direction. The compartmentalized nature of tissue model systems complicate diffusion measurements as exchange mechanisms must be considered (Stanisz, 2003). Further, intracellular and extracellular volume fractions measured through biexponential diffusion do not appear to be correlated with specific compartments within the tissue (Schwarz, et al. 2004; Sehy et al. 2002). By exploring diffusion anisotropy in the abdominal ganglia of Aplysia by diffusion tensor imaging, as well as exchange, through a more detailed approach in a more complicated environment could determine the likelihood of anisotropic diffusion in both single cell and ganglion models. Although this factor may not seem significant, if anisotropic diffusion is present, the orientation of the specimen for imaging studies would have to be maintained for consistent diffusion measurements, a practice that has not been controlled in any single cell imaging study to date. Given these concerns, clear differences do exist between the nuclear and cytoplasmic compartments in single cells. Metabolites such as betaine have different distributions in the separate regions, and the diffusion coefficients in each region are different. Utilizing a biexponential model of water diffusion in the cellular environment shows two unique fractions of water, both a bound and free pool, which exhibit slow and fast diffusion, respectively (Grant, et al. 2001). The values determined for the volume fractions of the fast and slow diffusion, do not correlate directly with the size of the nucleus and cytoplasm. It appears that a certain amount of both fast and slow diffusion occur in both the nucleus and cytoplasm. The resolution used in single cell studies is on the order of 20x20x200 μm, which was relaxed for imaging the entire ganglia to 40x40x100 μm and does require more extensive volume averaging of the nuclear, cytoplasmic and interstitial environments. Volume averaging may contribute to the large variability in diffusion measurements. However, despite the resolution, trends in 1H relaxation remain consistent. Although resolutions below 10 μm are achievable with smaller RF coils, they are not suitable for the larger size of the ganglia and would require significantly longer imaging times. The improvement in resolution may result in better resolution of each cellular compartment, which would eliminate a portion of the volume averaging and could yield diffusion fractions more representative of the cellular compartments. Although cellular resolutions have not been achieved in tissue models, the characteristic differences in diffusion coefficients likely exist in interstitial tissue environments. With increased barriers to diffusion

113 due to extracellular structures, anisotropic diffusion as well as altered volume fractions would be expected but have yet to be tested or understood. A variety of tissue models have been proposed to mimic the human tissue environment, particularly with rat and mice models. The effect of osmotic and ischemic perturbations on MR relaxation and diffusion mechanisms were previously performed using rat hippocampus (O'Shea, et al. 2000) and turtle cerebellum (van Pul, et al. 2005) tissue slice models. Both animals exhibited similar responses to osmotic perturbation through a decreased T2 and increased ADC with increasing tonicity. Shepherd et al. observed the osmotic response of osmotically perturbed human brain slices and found that the ADC did increase with increasing tonicity with results similar to those found for the turtle cerebellum and rat hippocampus, illustrating the validity of these animal models in representing pathological conditions in human brain tissue. The trends observed in the Aplysia ganglia model system exhibit similar trends in T2 relaxation and further support the use of model systems in analyzing MR diffusion and relaxation trends in human pathologies. Several factors must be carefully considered in osmotic perturbation models that eliminate unwarranted metabolic and physiological responses. Of primary concern is the freedom of the tissue to expand under hypotonic perturbation. In many studies, a nylon mesh or other restraining device is placed around the tissue slices, which may restrict free cellular expansion. Additionally, the method through which osmotic perturbation is introduced may result in unwanted cellular responses. Sodium, in particular, is responsible for initializing the depolarization of neurons. Because the tonicity of the extracellular fluid is frequently adjusted by adding additional sodium chloride, the functioning of the neurons in the tissue sample may have been affected (O’Shea et al. 2000). In imaging the Aplysia ganglia, the ganglia were suspended in a glass capillary tube in a much larger volume of ASW allowing the ganglia to remain unrestricted; however, this also allows the ganglia to move, which may result in motion artifacts, particularly during high gradient diffusion measurements. Further, although the sodium chloride concentration was altered to achieve the desired osmotic perturbations, the changes should not be significant in depolarizing the neurons or altering the physiology of the ganglia. In a study involving ischemic perturbation, artificial cerebrospinal fluid saturated deprived of oxygen and glucose was perfused through a turtle cerebellum tissue sample, limiting both the oxygen and glucose supply to the tissue (van Pul, et al. 2005). Upon creating an oxygen

114 and glucose deprived environment, the ADC rapidly decreased with the most rapid response occurring when oxygen was removed. Although, this change was attributed to changes in the intracellular environment, the experiment did not conclude with a return to baseline values. This finding may suggest that the individual cells died or were severely damaged during the experiment and were incapable of returning to baseline values. In imaging Aplysia ganglia, the ganglia remained in a sealed capillary tube in an environment where oxygen depletion and a lack of nutrients exist. Cell viability must be maintained throughout the experiment to understand the active cellular response occurring, and although the tissue was claimed to be viable, it would be illustrative if the tissue regained normal functioning after perturbation. Histological staining of the turtle cerebellum confirmed that extensive damage had occurred under oxygen-deprived conditions. Diffusion anisotropy of the tissue model systems remained unclear, again providing evidence at the tissue level confirming the analysis of single cells. The osmotic perturbation of tissues produces an expected response as the intracellular volume fraction is altered corresponding to cell swelling or crenation. However, the effects of ischemia are much more pronounced, yet are detected in a similar manner using MRI, showing that metabolic and physiological changes are occurring and are likely related to the breakdown of intracellular microstructure and the swelling of cellular organelles that would increase the intracellular tortuosity under ischemic conditions (van Pul, et al. 2005). Without the full characterization of cellular responses, the mechanisms underlying these changes in MR relaxation and diffusion could be related to both osmotic as well as ischemic challenge. Similar results can be seen for neurotoxic perturbation as the ADC decreases upon treatment with ouabain, an inhibitor of the sodium potassium pump. By perfusing a rat hippocampal tissue slice with a solution containing ouabain, a decrease in the ADC was found and accounted for by an increased in the volume fraction of the slow-diffusing compartment (Buckley, et al. 1999). Similar ADC changes are also seen with excitotoxicty in hippocampal slices (Bui, et al. 1999). This analysis shows that water is compartmentalized, and under neurotoxic perturbation water redistributions occur, altering free diffusion and increasing the water fraction confined in a more viscous intracellular environment. With this analysis, it appears that a multitude of factors can impact measurements of the ADC. It becomes difficult to distinguish the difference between osmotic perturbation, tissue ischemia and the effect of neurotoxins on the ADC. In a sealed environment, Aplysia ganglia exhibit aspects of each of

115 these contributing factors. Although the osmolarity of the extracellular environment is directly altered, a depleting supply of oxygen and buildup of cellular waste is expected which all seem to impact MR diffusion measurements. Non-proton nuclei such as 23Na are sensitive to the techniques of MRI; however, due to the quadrupolar nature of the sodium nucleus, there is an inherent loss of sensitivity, signal and resolution. Given these constraints, sodium MRI becomes practical at high magnetic field strengths or when a quantitative analysis of a large volume is needed, thereby reducing the resolution concerns. Several studies have used sodium imaging to observe changes in sodium distributions in cerebral tumors (Thulburn, et al. 1999) and cerebral ischemia (Lin, et al. 2001) in rats. It becomes evident that increasing the magnet field strength greatly enhances the resolution that can be achieved in reasonable times. However, higher field strengths are available and resolutions of 89x89x400 μm are possible at 11.75 T for the Aplysia ganglia. Sodium concentration increases in hypertonic regions, such as in cancer tumors, which relates to an increase in signal intensity. Similar increases in sodium signal are seen in states of tissue ischemia as sodium MRI appears to be adequate to differentiate diseased and normal tissue.

However, the relation of the increase in sodium signal intensity to changes in the sodium T1 and 23 T2 relaxation parameters was not previously made (Lin, et al. 2001). In the Aplysia ganglia, Na

T1 and T2* do not exhibit significant changes with osmotic perturbation, leading to the conclusion that the tissue sodium concentration is increasing. Further, the gradual and linear increase in 23Na signal intensity supports the finding that 23Na signal intensity measurements can identify the time after cellular damage occurs and as a measure of viability. Within the ganglia, inhomogeneities develop between the interior of the ganglia and the periphery where the majority of neurons are located. The cells in the interior of the ganglia first experience oxygen depletion and may lose viability at earlier time points. These observations support the findings that 23Na can be used as measure of tissue viability (Thulborn, et al. 2005) and that an inhomogeneity of 23Na distributions occur in ischemic tissue (Yushmanov, et al. 2009). Although the MRI relaxation parameters and ADC can be measured relatively easily, the physiological significance of these parameters is still not fully understood. Current efforts have used simplified model systems and numerical simulation to try and find the important physical properties responsible for these characteristic increases in ADC and reduction in T2 with increasing tonicity (Harkins et al. 2009; Kotek et al. 2009). Using a simplified red blood cell

116 system, the effect of intracellular volume fraction, extracellular proteins, and intracellular hydration on reductions in ADC under hypotonic conditions was examined (Kotek et al. 2009). It was found in this system that in cell swelling, the protein concentration and extracellular tortuosity hinder bulk movement resulting in a reduced diffusion coefficient. The identity of the extracellular components do not affect diffusion, but rather only the concentration is related to diffusion as higher concentrations of proteins reduce the diffusion coefficient. The hydration level in the intracellular space has no direct change on the diffusion coefficient, but the water shift is ultimately responsible for inducing these other changes. These changes are in agreement with direct measurements of protein levels which suggest that protein distribution is altered in states of osmotic perturbation (Mao et al. 2008). Although limited in nature due to the simplicity of the model, these measurements provide the first experimental data to identify the physical mechanisms responsible for changes in measured MRI parameters. The simplicity by which these parameters were tested suggests that similar experiments should be conducted progressively more complex samples to begin to put together an understanding of the ability of MRI to assess the cellular response to pathological conditions. The physical principles could be expanded into a numerical simulation to predict the ADC values as a function of characteristic parameters in the tissue environment. Tissue was modeled as a repeating grid of cuboidal cells undergoing isotropic, biexponential diffusion with a fast-diffusing and slow-diffusing compartment (Harkins, et al. 2009). The ADC is an average diffusion coefficient that takes into account both diffusing compartments resulting in the biexponential model of diffusion. The properties found to be important in the reduction of ADC is the intracellular diffusivity at short diffusion times and the membrane permeability at long diffusion times. In general, the ADC decreases with a decrease in the membrane permeability while the ADC and these differences can be accounted for by cell swelling (Harkins, et al. 2009). Changes occur with the cell membrane as it becomes less permeable during states of osmotic or ischemic stress. Therefore, the MRI relaxation parameters indicate the activity of membrane transport and the tonicity of the environment in which the cell is located and can be used on tissues in vivo.

117

CHAPTER SEVEN

CONCLUSION AND FUTURE DIRECTIONS

At present, a notable gap exists between the assessment of large isolated single cells and tissues using MRI, as well as the expansion of these techniques to in vivo animal and human subjects. As a majority of the single cell work has been done on the abdominal motor neuron from the sea slug Aplysia californica, expanding these studies to observe the entire ganglia using both 1H and 23Na MRI techniques provided information regarding how single cells function in the tissue environment. In particular, by using a small collection of cells, it can be seen how the cellular mechanism is responsible for the overall tissue response. Both proton and sodium MRI were used to characterize the ganglionic environment, which consists of intracellular, extracellular and interstitial environments. Of the different available relaxation mechanisms, it is evident that T2* weighting provides the greatest contrast for observing the individual cells within the neural ganglia. With tonicity changes, the ganglionic sac remains intact although alterations in the signal intensity (particularly with respect to sodium) are evident. Proton T1 did not change significantly with osmotic perturbation; however, 1 1 H T2 and T2* decreased with increasing tonicity, and H T2* provided the most robust measurement related to osmotic perturbation. All relaxation parameters increased with cell death and loss of viability, and no changes in 1H ADC were observed. The trends observed in 1H relaxation are possibly due to an increased intracellular volume fraction while the lack of a clear trend in ADC values may be due to significant volume averaging and alterations in water exchange between the nuclear, cytoplasmic and interstitial environments and should be modeled utilizing a multi-compartmental approach. Sodium measurements reveal a generally increasing trend in relaxation with osmotic perturbation. Changes in sodium relaxation are evident in compromised ganglia; however, linearly increasing trends in sodium signal intensity are seen in all osmotic conditions and may represent changes occurring with the loss of cell viability, providing evidence that the direct evaluation of ions may prove more sensitivity in assessing pathological osmotic changes. Future work will focus on the imaging of perfused viable ganglia, which would allow for different osmotic conditions to be introduced to the same ganglia. Through changing the extracellular tonicity, the duration of ganglion viability could be better assessed under

118 perturbation, while also investigating the transient impact of osmotic perturbations following multiple cycles of tonicity changes. Similar studies will be performed to quantify the proton and sodium relaxation parameters of perfused ganglia under osmotic and neurotoxic perturbation to compare these trends to those found in the unperfused and fixed ganglia. These methods will allow comparisons to be made between passive and active osmoregulation as well as between the single cell and tissue environment. Sodium distributions have been shown to be altered in states of tissue ischemia in rat tissue models, which provides speculation that similar distributions will occur in the Aplysia californica model. However, the role of T1 and T2 relaxation in these studies was not performed. By relating the sodium T1 and T2 to changes in physiological conditions, the changes in these parameters could be used to assess tonicity and sodium concentrations in perturbed tissues. This study indicates that sodium redistributions and changes in sodium signal intensity occur at much earlier time points than changes in the proton relaxation parameters and ADC. To assess 23Na relaxation in a time-sensitive manner, voxel-selective spectroscopy and contrast agents to isolate the intracellular sodium signal by chemical shift could be used to better analyze the sodium redistributions that are occurring and answer questions regarding the origin of changes in sodium signal intensity from increased intracellular sodium or altered 23Na relaxation parameters. Further, in addition to sodium, ions such as chlorine and potassium play a large role in active transport and impact the sodium-potassium pump regulating intracellular ionic concentrations. By expanding the current imaging techniques to include chlorine and potassium, a complete understanding and model of localized active transport processes in neurons could be developed. Although there exist significant sensitivity barriers that would make chlorine and potassium MRI difficult without drastic increases in SNR via much higher magnetic fields or other techniques, the feasibility of chlorine (Kirsch, et al. 2010) imaging in vivo on a rat brain and potassium imaging (Augath, et al. 2009) have been demonstrated. Recent diffusion measurements and modeling efforts can better model transport across cellular membranes and show the implications a disruption of these processes has on cellular homeostasis. Additionally, focusing on ionic MRI, a more advanced diffusion model can be applied to model the transport of ions in this neural ganglia model system, accounting for compartmentation such as the intracellular (nuclear and cytoplasmic environments) as well as the interstitial and extracellular environments. At present, the previous data from isolated single neurons has not been

119 incorporated into the current ganglionic model; however, the previously measured diffusion properties of isolated single cells can be used to quantify the contribution of individual neurons to the response of the tissue. With additional extracellular components in the ganglia model, the idea of tissue anisotropy must be further considered. Although isolated single cells were shown to exhibit isotropic diffusion, more sophisticated imaging techniques such as diffusion tensor imaging can be used to develop a more accurate representation of the properties of diffusion in single cells, particularly in their native tissue environments. However, measuring the volumes of these compartments will pose a challenge due to the complex and non-homogenous nature of the ganglion; although it has been proposed that an estimation of intracellular volume fraction is provided by paired diffusion and relaxation measurements. Through this model, exchange mechanisms could be determined and expanded to model such conditions in osmotically disrupted disease states, such as cell swelling during stroke and tissue ischemia. As with all MRI experimentation, current limitations to this technique reside in the sensitivity and resolution. Sodium, with a very low concentration, will always have a lower signal; however, high magnetic fields and the use of finely-tuned microcoils will continue to be utilized to achieve reasonable sodium resolutions. The limitation on resolution in MRI is due to the inherent diffusion of molecules and the resulting T2 relaxation where current estimates place this limit at 1 μm (Lee, et al. 2001; Tyszka, et al. 2005). With average mammalian cell sizes on the order of 10 μm, cells have the potential to be imaged with some detail of the intracellular environment as has been demonstrated (Flint, et al. 2009). However, as technologies continue to improve with higher magnetic fields becoming available on a regular basis and advances in RF coil technology, MRI continues to be expanded at the microscopic level to the point where MR microscopy may be able to assess the cellular level of human tissue under pathological conditions.

120

REFERENCES

Aguayo, J. B., Blackband, S. J., Schoeinger, J., & Mattingly, M. A. (1986). Nuclear magnetic resonance imaging of a single cell. Nature, 322, 190-191.

Aguayo, J. B., Blackband, S. J., Shoeniger, J., Mattingly, M. A., & Hintermann, M. (1986). Nuclear magnetic resonance imaging of a single cell. Nature, 322, 190-191.

Aiken, N. R., Hsu, E. W., Horsman, A., & Blackband, S. J. (1996). Maturation effects on the NMR microimaging characteristics of single neurons. American Journal of Physiology, 271, C1295-C1302.

Atkinson, I. C., Sonstegaard, R., Pliskin, N. H., & Thulborn, K. R. (2010). Vital signs and cognitive function are not affected by 23-sodium and 17-oxygen magnetic resonance imaging of hte human brain at 9.4 T. Journal of Magnetic Resonance Imaging, 32, 82-87.

Augath, M., Heiler, P., Kirsch, S., & Schad, L. R. (2009). In vivo 39K, 23Na, and 1H MR imaging using a trile resonant RF coil setup. Journal of Magnetic Resonance, 200, 134- 136.

Baird, A. E., & Warach, S. (1998). Magnetic resonance imaging of acute stroke. Journal of Cerebral Blood Flow and Metabolism, 18, 583-609.

Beck, B., Plant, D., Grant, S., Thelwall, P., Silver, X., Mareci, T., et al. (2002). Progress in high field MRI at the Univeristy of Florida. Magnetic Resonance Materials in Physics, Biology, and Medicine, 13, 152-157.

Benveniste, H., & Blackband, S. (2002). MR microscopy and high resolution small animal MRI: applications in neuroscience research. Progress in Neurobiology, 67, 393-420.

Benveniste, H., & Blackband, S. (2002). MR Microscopy and High Resolution Small Animal MRI: Applications in Neuroscience Research. Progress in Neurobiology, 67, 393-420.

Benveniste, H., Einstein, G., Kim, K. R., Huletee, C., & Johnson, G. A. (1999). Detection of neuritic plaques in Alzheimer's disease by magnetic resonance microscopy. Proceedings of the National Academy of Science, 96, 14079-14084.

Benveniste, H., Einstein, G., Kim, K. R., Hulette, C., & Johnson, G. A. (1999). Detection of neuritic plaques in Alzheimer's disease by magnetic resonance microscopy. Proceedings of the National Academy of Sciences of the United States of America, 96, 14079-14084.

Benveniste, H., Hedlund, H., & GA, J. (1992). Mechanism of detection of acute cerebral ischemia in rats by diffusion-weighted magnetic resonance microscopy. Stroke, 23, 746- 754.

121

Benveniste, H., Hedlund, L. W., & Johnson, G. A. (1992). Mechanism of detection of acute cerebral ischemia in rats by diffusion-weighted magnetic resonance microscopy. Stroke, 23, 746-754.

Blackband, S. J., Buckley, D. L., Bui, J. D., & Phillips, M. I. (1999). NMR microscopy - beginnings and new directions. Magnetic Resonance Materials in Physics, Biology, and Medicine, 9, 112-116.

Blackband, S. J., Buckley, D. L., Bui, J. D., & Phillips, M. I. (1999). NMR microscopy - beginnings and new directions. Magnetic Resonance Materials in Physics, Biology, and Medicine, 9, 112-116.

Blackband, S. J., Bui, J. D., Buckley, D. L., Zelles, T., Plant, H. D., Inglis, B. A., et al. (1997). MR microscopy of perfused brain slices. Magnetic Resonance in Medicine, 38, 1012- 1015.

Boada, F. E., LaVerde, G., Jungreis, C., Nemoto, E., Tanase, C., & Hancu, I. (2005). Loss of cell ion homeostasis: what sodium MRI can tell us. Current Topics in Developmental Biology, 70, 77-101.

Bourque, C. W. (2008). Central mechanisms of osmosensation and systemic osmoregulation. Nature Reviews Neuroscience, 9, 519-531.

Bowtell, R. W., Peters, A., Sharp, J. C., Mansfield, P., Hsu, E. W., Aiken, N., et al. (1995). NMR microscopy of single neurons using spin echo and line narrowed 2DFT imaging. Magnetic Resonance in Medicine, 33, 790-794.

Bowtell, R., GD, B., PM, G., M, M., & P, M. (1990). Resolution of cellular structures by NMR microscopy at 11.7 T. Philosophical Transactions: Physical Sciences and Engineering, 333, 456-467.

Buckley, D. L., Bui, J. D., Phillips, M. I., Zelles, T., Inglis, B. A., Plant, H. D., et al. (1999). The effect of ouabain on water diffusion in the rat hippocampal slice measured by high resolution NMR imaging. Magnetic Resonance in Medicine, 41, 137-142.

Buckley, D., Bui, J., Phillips, I., Zelles, T., Inglis, B., Plant, D., et al. (1999). The Effect of Ouabain on Water Diffusion in the Rat Hippocampal Slice Measured by High Resolution NMR Imaging. Magnetic Resonance in Medicine, 41, 137-142.

Bui, J. D., Buckley, D. L., Phillips, M. I., & Blackband, S. J. (1999). Nuclear magnetic resonance imaging measurements of water diffusion in the perfused hippocampal slice during N- methyl-D-aspartate-induced excitotoxicity. Neuroscience, 93, 487-490.

122

Buonanno, F. S., Pykett, I. L., Kistler, J. P., Vielma, J., Brady, T. J., Hinshaw, W. S., et al. (1982). Cranial anataomy and detection of ischemic stroke in the cat by nuclear magnetic resonance imaging. Nuclear Magnetic Resonance, 143, 187-193.

Burstein, D., & Mattingly, M. (1989). Magnetic resonance imaging of intracellular sodium. Journal of Magnetic Resonance, 83, 197-204.

Calamante, F., Lythgoe, M. F., Pell, G. S., Thomas, D. L., King, M. D., Busza, A. L., et al. (1999). Early changes in water diffusion, perfusion, T1 and T2 during focal cerebral ischemia in the rat studied at 8.5 T. Magnetic Resonance in Medicine, 41, 479-485.

Cameron, I. L., Merta, P., & Fullerton, G. D. (1990). Osmotic and motional properties of intracellular water as influences by osmotic swelling and shrinkage of Xenopus oocytes. Journal of Cellular Physiology, 142, 592-602.

Carpenter, D. O., Fejil, M., Ayrapetyan, S., Szarowski, D. H., & Turner, J. N. (1992). Dynamic changes in neuronal volume resulting from osmotic and sodium transport manipulations. Acta Biologica Hungarica, 43, 39-48.

Chen, K. C., & Nicholson, C. (2000). Changes in brain cell shape create residual extracellular space volume and explain tortuosity behavior during osmotic challenge. Proceedings of the National Academy of Sciences of the United States of America, 97, 8306-8311.

Chenevert, T. L., McKeever, P. E., & Ross, B. D. (1997). Monitoring early response of experimental brain tumors to therapy using diffusion magnetic resonance imaging. Clinical Cancer Research, 3, 1457-1466.

Christensen, J. D., Barrere, B. J., Boada, F. E., Vevea, J. M., & Thulborn, K. R. (1996). Quantitative tissue sodium concentration mapping of normal rat brain. Magnetic Resonance in Medicine, 38, 83-89.

Ciobanu, L., & Pennington, C. H. (2004). 3D micron-scale MRI of single biological cells. Solid State Nuclear Magnetic Resonance, 25, 138-141.

Ciobanu, L., Seeber, D. A., & Pennington, C. H. (2002). 3D MR microscopy with resolution 3.7 um by 3.3 um by 3.3 um. Journal of Magnetic Resonance, 158, 178-182.

Ciobanu, L., Webb, A. G., & Pennington, C. H. (2003). Magnetic resonance imaging of biological cells. Progress in Nuclear Magnetic Resonance Spectroscopy, 42, 69-93.

Ciobanu, L., Webb, A. G., & Pennington, C. H. (2003). Magnetic Resonance Imaging of Biological Cells. Progress in Nuclear Magnetic Resonance Spectroscopy, 42, 69-93.

Coggeshall, R. E. (1967). A light and electron microscope study of the abdominal ganglion of Aplysia californica. Journal of Neurophysiology, 30, 1263-1287.

123

Damadian, R. (1971). Tumor detection by nuclear magnetic resonance. Science, 171, 1151-1153. de Graaf, R. A. (2007). in vivo NMR spectroscopy Principles and Techniques (2nd ed.). Hoboken, New Jersey: John Wiley and Sons Ltd.

Enderle, J., Blanchard, S., & Bronzino, J. (2005). Introduction to Biomedical Engineering, 2nd ed. Burlington: Elsevier Academic Press.

Faller, L. D. (2008). Mechanistic studies of sodium pump. Archives of Biochemistry and Biophysics, 476, 12-21.

Fleysher, L., Oesingmann, N., Stoeckel, B., Grossman, R. I., & Inglese, M. (2009). Sodium long- component T2* mapping in human brain at 7 Tesla. Magnetic Resonance in Medicine, 62, 1338-1341.

Flint, J. J., Lee, C. H., Hansen, B., Fey, M., Schmidig, D., Bui, J. D., et al. (2009). Magnetic resonance microscopy of mammalian neurons. Neuroimage, 46, 1037-1040.

Flint, J., Hansen, B., Vestergaard-Poulsen, P., & Blackband, S. J. (2009). Diffusion weighted magnetic resonance imaging of neuronal activity in the hippocampal slice model. Neuroimage, 46, 411-418.

Frazier, W. T., Kandel, E. R., Kupfermann, I., Waziri, R., & Coggeshall, R. E. (1967). Morphological and functional properties of identified neurons in the abdominal ganglion of Aplysia californica. Journal of Neurophysiology, 30, 1288-1351.

Glover, P., & Mansfield, P. (2002). Limits to magnetic resonance microscopy. Reports on Progress in Physics, 65, 1489-1511.

Glover, P., & Mansfield, P. (2002). Limits to magnetic resonance microscopy. Reports on progress in physics, 65, 1489-1511.

Grant, S. C., Aiken, N. R., Plant, D., Gibbs, S., Mareci, T. H., Webb, A. G., et al. (2000). NMR spectroscopy of single neurons. Magnetic Resonance in Medicine, 44, 19-22.

Grant, S. C., Aiken, N. R., Plant, H. D., Gibbs, S., Mareci, T. H., Webb, A. G., et al. (2000). NMR Spectroscopy of Single Neurons. Magnetic Resonance in Medicine, 44, 19-22.

Grant, S. C., Buckley, D. L., Gibbs, S., Webb, A. G., & Blackband, S. J. (2001). MR microscopy of multicomponent diffusion in single neurons. Magnetic Resonance in Medicine, 46, 1107-1112.

Grant, S., Buckley, D., Gibbs, S., Webb, A., & SJ, B. (2001). MR microscopy of multicomponent diffusion in single neurons. Magnetic Resonance in Medicine, 46, 1107- 1112.

124

Grebenkov, D. S. (2010). Use, misuse, and abuse of apparent diffusion coefficients. Concepts in Magnetic Resonance Part A, 36A, 24-35.

Hancu, I., van der Maarel, J. R., & Boada, F. (2000). A model for the dynamics of spin 3/2 in biological media: signal loss during radiofrequency excitation in triple-quantum-filtered sodium MRI. Journal of Magnetic Resonance, 147, 179-191.

Harkins, K. D., Galons, J.-P., Secomb, T. W., & Trouard, T. P. (2009). Assessment of the effects of cellular tissue properties on ADC measurements by numerical simulation of water diffusion. Magnetic Resonance in Medicine, 62, 1414-1422.

Harkins, K. D., Galons, J.-P., Secomb, T. W., & Trouard, T. P. (2009). Assessment of the effects of cellular tissue properties on ADC measurements by numerical simulation of water diffusion. Magnetic Resonance in Medicine, 62, 1414-1422.

Heiler, P. M., Langhauser, F. L., Wetterling, F., Ansar, S., Grudzenski, S., Konstandin, S., et al. (2011). Chemical shift sodium imaging in a mouse model of thromboembolic stroke at 9.4 T. Journal of Magnetic Resonance Imaging, 34, 935-940.

Helpern, J. A., Lee, S.-P., Falangola, M. F., Dyakin, V. V., Bogart, A., Ardekani, B., et al. (2004). MRI assessment of neuropathology in a transgenic mouse model of Alzheimer's disease. Magnetic Resonance in Medicine, 51, 794-798.

Hilal, S. K., Maudsley, A. A., Perman, W. H., Bonn, J., Mawad, M. E., Silver, A. J., et al. (1983). In vivo NMR imaging of tissue sodium in the intact cat before and after acute cerebral stroke. American Journal of Neuroradiology, 4, 245-249.

Hoehn, M., Nicolay, K., Franke, C., & van der Sanden, B. (2001). Application of magnetic resonance to animal models of cerebral ischemia. Journal of Magnetic Resonance Imaging, 14, 491-509.

Hoehn-Berlage, M., Eis, M., Back, T., Kohno, K., & Yamashita, K. (1995). Changes of relaxation times (T1, T2) and apparent diffusion coefficient after permanent middle cerebral artery occlusion in the rat: temporal evolution, regional extent, and comparison with histology. Magnetic Resonance in Medicine, 34, 824-834.

Hsu, E. (1996). Nuclear magnetic resonance of perfused single cells and its application to the study of water diffusion changes in tissue ischemia. Ph.D. Thesis, Johns Hopkins University, Baltimore.

Hsu, E. W., Aiken, N. R., & Blackband, S. J. (1996). Nuclear magnetic resonance microscopy of single neurons under hypotonic perturbation. American Journal of Physiology, 271, C1895-C1900.

125

Hsu, E. W., Aiken, N. R., & Blackband, S. J. (1997). A study of diffusion isotropy in single neurons using NMR microscopy. Magnetic Resonance in Medicine, 37, 624-627.

Hsu, E. W., Aiken, N. R., & Blackband, S. J. (1997). A study of diffusion isotropy in single neurons using NMR microscopy. Magnetic Resonance in Medicine, 37, 624-627.

Hsu, E., Aiken, N., & Blackband, S. (1996). Nuclear magnetic resonance microscopy of single neurons under hypotonic perturbation. American Journal of Physiology-Cell Physiology, 271, 1895-1900.

Hussain, M. S., Stobbe, R. W., Bhagat, Y. A., Emery, D., Butcher, K. S., Manawadu, D., et al. (2009). Sodium imaging intensity increases with time after human ischemic stroke. Annals of Neurology, 66, 55-62.

Inglese, M., Madelin, G., Oesingmann, N., Babb, J. S., Wu, W., Stoeckel, B., et al. (2010). Brain tissue sodium concentration in multiple sclerosis: a sodium imaging study at 3 tesla. Brain, 133, 847-857.

Jespersen, S. N., Pedersen, M., & Stodkilde-Jorgensen, H. (2005). The influence of a cellular size distribution on NMR diffusion measurements. European Biophysics Journal, 34, 890-898.

Joiner, C. H., & K, L. P. (1978). The Correlation Between Ouabain Binding and Potassium Pump Inhibition in Human and Sheep Erythrocytes. Journal of Physiology, 283, 155-175.

Jokivarsi, K. T., Hiltunen, Y., Grohn, H., Tuunanen, P., Grohn, O. H., & A., K. R. (2010). Estimation of the onset time of cerebral ischemia using T1rho and T2 MRI in rats. Stroke, 41, 2335-2340.

Jones, S. C., Kharlamov, A., Yanovski, B., Kim, D. K., Easley, K. A., Yushmanov, V. E., et al. (2006). Stroke onset time using sodium MRI in rat focal cerebral ischemia. Stroke, 37, 883-888.

Kamman, R. L., Go, K. G., Brouwer, W., & Berendsen, H. J. (1988). Nuclear magnetic resonance relaxation in experimental brain edema: effects of water concentration, protein concentration, and temperature. Magnetic Resonance in Medicine, 6, 265-274.

Kandel, E. R., Frazier, W. T., Waziri, R., & Coggeshall, R. E. (1967). Direct and common connections among identified neurons in Aplysia. Journal of Neurophysiology, 30, 1352- 1376.

Kettunen, M. I., & Brindle, K. M. (2005). Apoptosis detection using magnetic resonance imaging and spectroscopy. Progress in Nuclear Magnetic Resonance Spectroscopy, 47, 175-185.

126

Kirsch, S., Augath, M., Seiffge, D., Schilling, L., & Schad, L. R. (2010). In vivo chlorine-35, sodium-23, and proton magnetic resonance imaging of the rat brain. NMR in Biomedicine, 23, 592-600.

Kiselev, V. G., & Il'yasov, K. A. (2007). Is the "biexponential diffusion" biexponential? Magnetic Resonance in Medicine, 57, 464-469.

Kotek, G., Berente, Z., Schwarcz, A., Vajda, Z., Hadjiev, J., Horvath, I., et al. (2009). Effects of intra- and extracellular space properties on diffusion and T2 relaxation in a tissue model. Magnetic Resonance Imaging, 27, 279-284.

Kotek, G., Berente, Z., Schwarcz, A., Vajda, Z., Hadjiev, J., Horvath, I., et al. (2009). Effects of intra- and extracellular space properties on diffusion and T2 relaxation in a tissue model. Magnetic Resonance Imaging, 27, 279-284.

Krizaj, D., Rice, M. E., Wardle, R. A., & Nicholson, C. (1996). Water compartmentalization and extracellular tortuosity after osmotic changes in cerebellum of Trachemys scripta. Journal of Physiology, 492, 887-896.

Kuhn, W. (1990). NMR microscopy - fundamentals, limits and possible applications. Angewandte Chemie, 29, 1-19.

Kuhn, W. (1990). NMR microscopy - fundamentals, limits, and possible applications. Angewandte Chemie, 1-19.

Lang, F., Busch, G. L., Ritter, M., Volkl, H., Waldeggar, S., Gulbins, E., et al. (1998). Functional Significance of cell volume regulatory mechanisms. Physiological Reviews, 78, 247-306.

Lang, F., Busch, G. L., Ritter, M., Volkl, H., Waldegger, S., Gulbins, E., et al. (1998). Functional significance of cell volume regulatory mechanisms. Physiological Reviews, 78, 247-306.

Latt, J., Nilsson, M., van Westen, D., Wirestam, R., Stahlberg, F., & Brockstedt, S. (2009). Diffusion-weighted MRI measurements on stroke patients reveal water-exchange mechanisms in sub-acute ischaemic lesions. NMR in Biomedicine, 22, 619-628.

LaVerde, G. C., Jungreis, C. A., Nemoto, E., & Boada, F. (2009). Sodium time course using 23Na MRI in reversible focal brain ischemia in the monkey. Journal of Magnetic Resonance Imaging, 30, 219-223.

Lee, S.-C., Cho, J.-H., Mietchen, D., Kim, Y.-S., Hong, K. S., Lee, C., et al. (2006). Subcellular in vivo 1H MR spectroscopy of Xenopus laevis oocytes. Biophysical Journal, 90, 1797- 1803.

Lee, S.-C., Kim, K., Kim, J., Lee, S., Yi, J. H., Kim, S. W., et al. (2001). One micrometer resolution NMR microscopy. Journal of Magnetic Resonance, 150, 207-213.

127

Lin, S.-P., Song, S.-K., Miller, J. P., Ackerman, J. J., & Neil, J. J. (2001). Direct, longitudinal comparison of 1H and 23Na MRI after transient focal cerebral ischemia. Stroke, 32, 925- 932.

Lin, S.-P., Song, S.-K., Miller, P., Ackerman, J., & Neill, J. (2001). Direct Longitudinal Comparison of 1H and 23Na MRI AfterTransient Focal Cerebral Ischemia. Stroke, 32, 925-932.

Lipton, P. (1999). Ischemic cell death in brain neurons. Physiological Reviews, 79, 1431-1568.

Mallard, J. (1986). First sight of a single cell. Nature, 322, 116.

Mao, L., Hartl, D., Nolden, T., Koppelstatter, A., Klose, J., Himmelbauer, H., et al. (2008). Pronounced alterations of cellular metabolism and structure due to hyper- or hypo- osmosis. Journal of Proteome Research, 7, 3968-3983.

Mao, L., Hartl, D., Nolden, T., Koppelstatter, A., Klose, J., Himmelbauer, H., et al. (2008). Pronounced alterations of cellular metabolism and structure due to hyper- or hypo- osmosis. Jounral of Proteome Research, 7, 3968-3983.

Marieb, E. N., & Hoehn, K. (2007). Human Anatomy and Physiology (7th ed ed.). San Franciso: Pearson Benjamin Cummings.

Matsumoto, Y., Kuroda, M., Matsuya, R., Kato, H., Shibuya, K., Oita, M., et al. (2009). In vitro experimental study of the relationship between the apparent diffusion coefficient and changes in cellularity and cell morphology. Oncology Reports, 22, 641-648.

Mellon, E. A., Pilkinton, D. T., Clark, C. M., Elliott, M. A., Witschey 2nd, W. R., Borthakur, A., et al. (2009). Sodium MR imaging detection of mild Alzheimer disease: preliminary study. American Journal of Neuroradiology, 30, 978-984.

Merino, J. G., & Warach, S. (2010). Imaging of acute stroke. Nature Reviews Neurology, 6, 560- 571.

Minard, K. R., & Wind, R. A. (2001). Solenoidal microcoil design - part II: optimizing winding parameters for maximum signal-to-noise performance. Concepts in Magnetic Resonance, 13, 190-210.

Minard, K. R., & Wind, R. A. (2001). Solenoidal Microcoil Design - Part II: optimizing Winding Parameters for Maximum Signal-to-Noise Performance. Concepts in Magnetic Resonance, 13(3), 190-210.

Minard, K. R., & Wind, R. A. (2001). Solenoidal microcoil design. Part I: optimizing RF homogeneity and coil dimensions. Concepts in Magnetic Resonance, 13, 128-142.

128

Minati, L., & Weglarz, W. P. (2007). Physical foundations, models, and methods of diffusion magnetic resonance imaging of the brain: a review. Concepts in Magnetic Resonance Part A, 30A, 278-307.

Mintorovitch, J., Yang, G. Y., Shimizu, H., Kucharczyk, J., Chan, P. H., & Weinstein, P. R. (1994). Diffusion-weighted magnetic resonance imaging of actue focal cerebral ischemia: comparison of signal intenisty with changes in brain water and Na+, K+-ATPase activity. Journal of Cerebral Blood Flow and Metabolism, 14, 332-336.

Moseley, M. E., Cohen, Y., Mintrovitch, J., Chileuitt, L., Shimizu, H., Kucharczyk, J., et al. (1990). Early detection of regional cerebral ischemia in cats: comparison of diffusion- and T2-weighted MRI and spectroscopy. Magnetic Resonance in Medicine, 14, 330-346.

Nagel, A. M., Bock, M., Hartmann, C., Gerigk, L., Neumann, J.-O., Weber, M.-A., et al. (2011). The potential of relaxation-weighted sodium magnetic resonance imaging as demonstrated on brain tumors. Investigative Radiology, 46, 539-547.

Naritomi, H., Kanashiro, M., Sasaki, M., Kuribayashi, Y., & Sawada, T. (1987). In vivo measurements of intra- and extracellular Na+ and water in the brain and muscle by nuclear magnetic spectroscopy with shift reagent. Biophysical Journal, 52, 611-616.

Naruse, S., Horikawa, Y., Tanaka, C., Hirakawa, K., Nishikawa, H., & Yoshizaki, K. (1986). Significance of proton relaxation time measurement in brain edema, cerebral infarction and brain tumors. Magnetic Resonance Imaging, 4, 293-304.

Neuberger, T., & Webb, A. (2009). Radiofrequency coils for magnetic resonance microscopy. NMR in Biomedicine, 22, 975-981.

Norris, D. G. (2001). The effects of microscopic tissue parameters on the diffusion weighted magnetic resonance imaging experiment. NMR in Biomedicine, 14, 77-93.

Novak, J. L., & Wheeler, B. C. (1986). Recording from the Aplysia abdominal ganglia with a planar microelectrode array. IEEE Transaction on Biomedical Engineering, BME-33, 196-202.

O'Shea, J. M., Williams, S. R., van Bruggen, N., & Gardner-Medwin, A. R. (2000). Apparent diffusion coefficient and MR relaxation during osmotic manipulation in isolated turtle cerebellum. Magnetic Resonance in Medicine, 44, 427-432.

O'Shea, J. M., Williams, S. R., vanBruggen, N., & Gardner-Medwin, A. R. (2000). Apparent diffusion coefficient and MR relaxation during osmotic manipulation in isolated turtle cerebellum. Magnetic Resonance in Medicine, 44, 427-432.

Pasantes-Morales, H., & Tuz, K. (2006). Volume changes in neurons: hyperexcitability and neuronal death. Contributions to Nephrology, 152, 221-240. 129

Pauser, S., Zschunke, A., Khuen, A., & Keller, K. (1995). Estimation of water content and water mobility in the nucleus and cytoplasm of Xenopus laevis oocytes by NMR microscopy. Magnetic Resonance Imaging, 13, 269-276.

Peck, T. L., Magin, R. L., & Lauterbur, P. C. (1995). Design and analysis of microcoils for NMR microscopy. Journal of Magnetic Resonance, Series B, 108, 114-124.

Peck, T. L., Magin, R. L., & Lauterbur, P. C. (1995). Design and Analysis of Microcoils for NMR microscopy. Journal of Magnetic Resonance, 108, 114-124.

Perman, W. H., Turski, P. A., Houston, L. W., Glover, G. H., & E., H. C. (1986). Methodology of in vivo human sodium MR imaging at 1.5 T. Radiology, 160, 811-820.

Petkova, M., Rodrigo, S., Lamy, C., Oppenheim, G., Touze, E., Mas, J.-L., et al. (2010). MR imaging helps predict time from symptom onset in patients with acute stroke: implications for patients with unkown onset time. Radiology, 257, 782-792.

Posse, S., & Aue, W. P. (1989). Spectroscopic imaging and gradient-echo microscopy on a single cell. Journal of Magnetic Resonance, 83, 620-625.

Purea, A., & Webb, A. G. (2006). Reversible and irreversible effects of chemical fixation on the NMR properties of single cells. Magnetic Resonance in Medicine, 56, 927-931.

Rooney, W. D., & Springer Jr., C. S. (1991). A comprehensie approach to the analysis and interpretation of the resonances of spins 3/2 from living systems. NMR in Biomedicine, 4, 209-226.

Rooney, W. D., & Springer Jr., C. S. (1991). The molecular environment of intracellular sodium: 23Na NMR relaxation. NMR in Biomedicine, 4, 227-245.

Roth, Y., Ocherashvilli, A., Daniels, D., Ruiz-Cabello, J., Maier, S. E., Orenstein, A., et al. (2008). Quantification of water compartmentation in cell suspensions by diffusion- weighted and T2-weighted MRI. Magnetic Resonance Imaging, 26, 88-102.

Schepkin, V. D., Bejarano, F. C., Morgan, T., Gower-Winter, S., Ozambela Jr., M., & Levenson, C. W. (2011). In vivo magnetic resonance imaging of sodium and diffusion in rat glioma at 21.1 T. Magnetic Resonance in Medicine.

Schepkin, V. D., Brey, W. W., Gor'kov, P. L., & Grant, S. C. (2010). Initial in vivo rodent sodium and proton MR imaging at 21.1 T. Magnetic Resonance Imaging, 28, 400-407.

Schoeniger, J. S., Aiken, N., Hsu, E., & Blackband, S. J. (1994). Relaxation-time and diffusion NMR microscopy of single neurons. Journal of Magnetic Resonance, Series B, 103, 261- 273.

130

Schwarcz, A., Bogner, P., Meric, P., Correze, J.-L., Berente, Z., Pal, J., et al. (2004). The existence of biexponential signal decay in magnetic resonance diffusion-weighted imaging appears to be independent of compartmentalization. Magnetic Resonance in Medicine, 51, 278-285.

Sehy, J. V., Ackerman, J. J., & Neil, J. J. (2002). Evidence that both fast and slow water ADC compnents arise from intracellular space. Magnetic Resonance in Medicine, 48, 765-770.

Sehy, J. V., Zhao, L., Xu, J., Rayala, H. J., Ackerman, J. J., & Neil, J. J. (2004). Effects of physiological challenge on the ADC of intracellular water in the Xenopus oocyte. Magnetic Resonance in Medicine, 52, 239-247.

Sehy, J., Ackerman, J. J., & Neil, J. J. (2001). Water and lipid MRI of the Xenopus oocyte. Magnetic Resonance in Medicine, 46, 900-906.

Shepard, T. M., Scheffler, B., King, M. A., Stanisz, G. J., Steindler, D. A., & Blackband, S. J. (2006). MR microscopy of rat hippocampal slice cultures: a novel method for studying cellular processes and chronic perturbations to tissue microstructure. Neuroimage, 30, 780-786.

Shepherd, T. M., Thelwall, P. E., Stanisz, G. J., & Blackband, S. J. (2009). Aldehyde fixative solutions alter the water relaxation and diffusion properties of nervous tissue. Magnetic Resonance in Medicine, 62, 26-34.

Shepherd, T. M., Wirth III, E. D., Thelwall, P. E., Chen, H.-X., Roper, S. N., & Blackband, S. J. (2003). Water diffusion measurements in perfused human hippocampal slices undergoing tonicity changes. Magnetic Resonance in Medicine, 49, 856-863.

Shepherd, T., Wirth, E., Thelwall, P., Chen, H.-X., Roper, S., & Blackband, S. (2003). Water Diffusion Measurements in Perfused Human Hippocampal Slices Undergoing Tonicity Changes. Magnetic Resonance in Medicine, 49, 856-863.

Silva, A. L., & Wright, S. H. (1994). Short-term cell volume regulation in Mytilus californianus gill. Journal of Experimental Biology, 194, 47-68.

Singer, S. J., & Nicolson, G. L. (1972). The fluid mosaic model of the structure of cell membranes. Science, 175, 720-731.

SJ, S., & Nicolson, G. L. (1972). The fluid mosaic model of the structure of cell membranes. Science, 175, 720-731.

Skinner, T. L., & Peretz, B. (1989). Age sensitivity of osmoregulation and of its neural correlates in Aplysia. American Journal of Physiology, 256, R989-R996.

131

Sotak, C. H. (2004). Nuclear magnetic resonance (NMR) measurements of the apparent diffusion coefficient (ADC) of tissue water and its relationship to cell volume changes in pathological states. Neurochemistry International, 45, 569-582.

Stanisz, G. J. (2003). Diffusion MR in biological systems: tissue compartments and exchange. Israel Journal of Chemistry, 43, 33-44.

Strange, K. (2004). Cellular volume homeostasis. Advances in Physiology Education, 28, 155- 159.

Szafer, A., Zhong, J., Anderson, A. W., & Gore, J. C. (1995). Diffusion-weighted imaging in tissues: theoretical models. NMR in Biomedicine, 8, 289-296.

Thelwall, P. E., Grant, S. C., Stanisz, G. J., & Blackband, S. J. (2002). Human erythrocyte ghosts: exploring the origins of multiexponential water diffusion in a model biological tissue with magnetic resonance. Magnetic Resonance in Medicine, 48, 649-657.

Thulborn, K. R., Davis, D., Adams, H., Gindin, T., & Zhou, J. (1999). Quantitative tissue sodium concentration mapping of the growth of focal cerebral tumors with sodium magnetic resonance imaging. Magnetic Resonance in Medicine, 41, 351-359.

Thulborn, K. R., Davis, D., Snyder, J., Yonas, H., & Kassam, A. (2005). Sodium MR imaging of acute and subacute stroke for assessment of tissue viability. Neuroimaging Clinics of North America, 15, 639-653.

Thulborn, K. R., Gindin, T. S., Davis, D., & Erb, P. (1999). Comprehensive MR imaging protocal of stroke management: tissue sodium concentration as a measure of tissue viability in nonhuman primate studies and in clinical studies. Radiology, 213, 156-166.

Thulburn, K., Deavis, D., Adams, H., Gindin, T., & Zhou, J. (1999). Quantitiative Tissue Sodium Concentration Mapping of the Growth of Cerebral Tumors with Sodium Magnetic Resonance Imaging. Magnetic Resonance in Medicine, 41, 351-359.

Treistman, S. N., & Grant, A. G. (1993). Increase in cell size underlies neuron-specific temperature acclimation in Aplysia. American Journal of Physiology, 264, C1061-C1065.

Tsang, A., Stobbe, R. W., Asdaghi, N., Hussain, M. S., Bhagat, Y. A., Beaulieu, C., et al. (2011). Relationship between sodium intensity and perfusion deficits in acute ischemic stroke. Journal of Magnetic Resonance Imaging, 33, 41-47.

Tyszka, J. M., Fraser, S. E., & Jacobs, R. E. (2005). Magnetic resonance microscopy: recent advance and applications. Current Opinion in Biotechnology, 16, 93-99.

Tyszka, J. M., Fraser, S. E., & Jacobs, R. E. (2005). Magnetic resonance microscopy: recent advances and applications. Current Opinion in Biotechnology, 16, 93-99.

132 van der Toorn, A., Sykova, E., Dijkhuizen, R. M., Vorisek, I., Vargova, L., Skobisova, E., et al. (1996). Dynamic changes in water ADC, energy metabolism, extracellular space volume, and tortuosity in neonatal rat brain during global ischemia. Magnetic Resonance in Medicine, 36, 52-60. van Dorsten, F. A., Olah, L., Schwindt, W., Grune, M., Uhlenkuken, U., Pillekamp, F., et al. (2002). Dynamic changes of ADC, perfusion, and NMR relaxation parameters in transient focal ischemia of rat brain. Magnetic Resonance in Medicine, 47, 97-104. van Pul, C., Jennekens, W., Nicolay, K., Kopinga, K., & Wijn, P. (2005). Ischemia-Induced ADC Changes Are Larger Than Osmotically-Induced ADC Changes in a Neonatal Rat Hippocampus Model. Magnetic Resonance in Medicine, 53, 348-355. van Pul, C., Jennekens, W., Nicolay, K., Kopinga, K., & Wijn, P. F. (2005). Ischemia-induced ADC changes are larger than osmotically-induced ADC changes in a neonatal rat hippocampus model. Magnetic Resonance in Medicine, 53, 348-355.

Verheul, H. B., Balazs, R., Berkelbach van der Sprenkel, J. W., Tulleken, C. A., Nicolay, K., Tamminga, K. S., et al. (1994). Comparison of diffusion-weighted MRI with changes in cell volume in a rat model of brain injury. NMR in Biomedicine, 7, 96-100.

Webb, A. (2003). Introduction to Biomedical Imaging. Hoboken, New Jersey: John Wiley and Sons, Inc.

Webb, A. (2003). Introduction to Biomedical Imaging. New Jersey: John Wiley and Sons.

Wetterling, F., Gallagher, L., Macrae, I. M., Junge, S., & Fagan, A. J. (2011). Regional and temporal variations in tissue sodium concentration during the acute stroke phase. Magnetic Resonance in Medicine.

Wetterling, F., Tabbert, M., Junge, S., Gallagher, L., Macrae, I. M., & Fagan, A. J. (2010). A double-tuned 1H/23Na dual resonator system for tissue sodium concentration measurements in the rat brain via Na-MRI. Physics in Medicine and Biology, 55, 7681- 7695.

Winkler, S. S., Thomasson, D. M., Sherwood, K., & Perman, W. H. (1989). Regional T2 and sodium concentration estimates in the normal human brain by sodium-23 MR imaging at 1.5 T. Journal of Computer Assisted Tomography, 13, 561-566.

Yablonskiy, D. A., & Sukstanskii, A. L. (2010). Theoretical models of the diffusion weighted MR signal. NMR in Biomedicine, 23, 661-681.

Young, W., Rappaport, H., Chalif, D., & Flamm, E. (1987). Regional Brain Sodium, Potassium, and Water Changes in the Rat Middle Cerebral Artery Occlusion Model of Ischemia. Stroke, 18, 751-759. 133

Young, W., Rappaport, Z. H., Chalif, D. J., & Flamm, E. S. (1987). Regional brain sodium, potassium, and water changes in the rat middle cerebral artery occlusion model of ischemia. Stroke, 18, 751-759.

Yushmanov, V. E., Kharlamov, A., Yanovski, B., LaVerde, G., Boada, F. E., & Jones, S. C. (2009). Inhomogeneous sodium accumulation in the ischemic core in rat focal cerebral ischemia by 23Na MRI. Journal of Magnetic Resonance Imaging, 30, 18-24.

Yushmanov, V. E., Yanovski, B., Kharlamov, A., LaVerde, G., Boada, F., & Jones, S. C. (2009). Sodium mapping in focal cerebral ischemia in the rat by quantitative 23Na MRI. Journal of Magnetic Resonance Imaging, 29, 962-966.

134

BIOGRAPHICAL SKETCH

John Walsh was born on June 25, 1988 in Tampa, FL and raised in Palmetto, FL. He began his undergraduate education at Florida State University in 2006 and received a B.S. in Chemical Engineering with a major in Biomedical Engineering in 2010. He continued at Florida State University, starting his M.S. in Biomedical Engineering in 2010. He plans on attending medical school after graduation to further his education in medical research.

135