UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

NMR imaging and spectroscopy in cyrobiology Isbell, Stephanie A. 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1994-953519.pdf [ 6.61MB ]
Metadata
JSON: 831-1.0059650.json
JSON-LD: 831-1.0059650-ld.json
RDF/XML (Pretty): 831-1.0059650-rdf.xml
RDF/JSON: 831-1.0059650-rdf.json
Turtle: 831-1.0059650-turtle.txt
N-Triples: 831-1.0059650-rdf-ntriples.txt
Original Record: 831-1.0059650-source.json
Full Text
831-1.0059650-fulltext.txt
Citation
831-1.0059650.ris

Full Text

NMR IMAGING AND SPECTROSCOPYIN CRYOBIOLOGYbySTEPHANIE A. ISBELLB.Sc., University of Montana, 1986A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF CHEMISTRYWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIASeptember 1994© Stephanie A. IsbellIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. it is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of_________________The University of British ColumbiaVancouver, CanadaDate____DE-6 (2188)iiABSTRACTMicroscopic Chemical Shift Specific Slice Selective (C45) NMR imaging isdemonstrated to be a nondestructive and noninvasive technique for monitoringthe spatial distributions of solvents and of freezing/thawing phenomena.Studies of test samples show that the C4S NMR imaging sequence can yieldquantitative maps of distributions of water., and dimethyl sulfoxide (DMSO).The unique characteristics of C4S NMR imaging are potentially very useful instudies of organ cryopreservation in which cryoprotective solvents (CPSs) likeDMSO successfully prevent freezing damage to cells in suspension but presentlycannot be used to preserve whole organs at subzero temperatures. There are noother techniques available which are capable of monitoring CPS and waterconcentration distributions as well as structural changes in organs duringcryopreservation procedures.C4S NMR imaging is shown to be useful for monitoring the diffusion ofthe cryoprotectant DMSO into rat kidney and liver tissues. A detailedinvestigation of the possibility of mapping concentration distributions ofDMSO and water at various temperatures using images with short repetitiontimes verifies that DMSO and water in tissues can be quantified from imagesignal intensities within certain concentration ranges. Diffusion rates areeasily obtained from the imaging data and are similar to those found in theliterature. A simple method for determination of effective diffusioncoefficients of DMSO in tissues is developed and shown to be more accurate andunambiguous than the commonly used chemical techniques.Aspects of the introduction of DMSO to rat kidney by perfusion throughthe vasculature are studied. C4S NMR imaging is applied to two common bloodremoval techniques and is shown to detect blood remaining in the vasculatureiiiafter these washout procedures. A new method of completely removing bloodfrom rat kidneys is shown to be more effective. Complete blood removal isdemonstrated to be essential for homogeneous equilibration of tissues with acryoprotective solution. Blood clots preventing the equilibration of DMSO ina kidney indirectly cause severe freezing damage. Equilibration of DMSO withrat kidney via perfusion with DMSO in University of Wisconsin (OW) solution isshown to cause minimal damage and proceed most quickly at 20°C. 31P NMRspectroscopy is performed on rat kidneys subjected to the new blood washoutprocedure and the relationship of the spectra to organ viability is discussed.ivTABLE OF CONTENTSABSTRACTTABLE OF CONTENTS ivLIST OF TABLES viiLIST OF FIGURES viiiLIST OF ABBREVIATIONS xixACKNOWLEDGEMENT xxCHAPTER 1: CRYOBIOLOGY AND NMR IMAGING . .1.1. CRYOBIOLOGY AND TISSUE PRESERVATION1.2. POTENTIAL USE OF NMR IMAGINGIN CRYOBIOLOGY1.4. NUCLEAR MAGNETIC RESONANCE (NMR)1.4.2. Fourier Transform NMR .1.4.3. T1, T2 and Quantitationin Pulse NMR1.4.4. Temperature Effects on T1, T21.4.5. Phosphorus-31 NMR Spectroscopy1.5. NMR IMAGING1.5.1. Frequency Selective RF Pulses.1.5.2. Linear Magnetic Field Gradients1.5.3. Slice Selection1.5.4. Spin—Warp Imaging1.5.5. Chemical Shift Selective Imaging.1.6. THESIS GOALSREFERENCESCHAPTER 2: INVESTIGATION OF THE FREEZING AND THAWINGOF LIQUIDS IN HETEROGENEOUS SYSTEMS2.1. INTRODUCTION2.1.1. Effects of Temperature on Relaxation Rates2.1.2. Effects of Temperature on the NMR Spectrum -2.2. NMR AS A TECHNIQUE FOR MONITORINGFREEZING/THAWING PHENOMENA2.2.1. Monitoring the Freezing Process with NMR Spectroscopy.2.2.2. Monitoring the Freezing Process with NMR Imaging2.3. T1 AND T2 IN THE DMSO/WATER SYSTEM2.3.1. Experimental2.3.1.1. Measurement of Relaxation Parameters.2.3.1.2. Imaging Protocol2.3.2. Results of T1 and T2 Measurements . .2.3.3. Discussion of the Proton Relaxations Times inSolutions2.4. IMAGING EXPERIMENTS: RELAXATION RATESAND CHOICE OF PARANETERS2.5. IMPLICATIONS FOR MONITORINGFREEZING/THAWING PHENOMENAREFERENCES113511- . .. 17• . . .21- -.-21-. 22- --. 23• . . . 26- -• . 27• . . - 29-. 3437- .- 3941- .- 4141- • - 45• • -47474851525253• . • 55DMSO/H2O/NaC158626971VCHAPTER 3: PROTON NMR RELAXATION TIMES OF WATER AND DMSO IN TISSUE SAMPLESAND THEIR EFFECT ON IMAGE INTENSITIES3.1. INTRODUCTION3.1.1. The Importance of Contrast in NMR Imaging.3.1.2. T1 and T2 Contrast in NonquantitativeImaging Protocols3.2. EXPERIMENTAL3.2.1. Surgical Procedures3.2.2. Measurement of T1 and T2 Relaxation Timesin Rat Organ Tissue3.3. RESULTS AND DISCUSSION3.3.1. Modelling P1 and T2 Contrast3.3.1.1. Calculation of P1 Contrast .3.3.1.2. Calculation of P2 Contrast .3.3.2. Image Contrast in Model Solutions .3.3.2.1. T1 Contrast in Short-TR Protocols3.3.2.2. P2 Contrast in Short-TR C4S NMR Images3.3.2.3. General Effects of P1 and T2 Contrast3.3.3. Image Contrast in Intact. Tissues3.3.3.1. T1 and P2 Relaxation in Rat Organ Tissues3.3.3.2. Evaluation of Contrast in Imagesof Rat Organ Tissues3.4. CONCLUSIONS: CONTRAST IN C4S NMR IMAGESOF RAT ORGAN TISSUESREFERENCESCHAPTER 4: STUDIES OF DIFFUSION OF CPS INTO RAT ORGAN TISSUE.4.1. INTRODUCTION4.1.1. Importance of Measuring CPSPenetration into Tissues4.1.2. Techniques for Measuring Diffusionof CPSs into Tissues4.1.2.1. Attempts to Estimate Diffusion Rates.4.1.2.2. NMR Imaging and Diffusionin Intact Tissues4.1.3. C4S NMR Imaging and Freezing Processes.4.2. EXPERIMENTAL4.2.1. Preparation of Bovine Tissue Samples4.2.2. Preparation of Rat Tissue Samples .4.3. C4S NMR IMAGING AS A MONITOR OF DMSO DIFFUSION.4.3.1. Preliminary Studies of Diffusionof DMSO into Bovine Tissue Samples4.3.2. Potential Complications in C4S Imagingof DMSO Diffusion4.3.2.2. Contrast in Short—TR C4S NMR Imaging Protocols4.3.2.4. Apparent Exclusion of DMSOfrom Rat Organ Tissues4.3.3. Potential and Limitations of C4S Imagingof DMSO Diffusion in Rat Tissues4.4. DIFFUSION OF DMSO IN RAT KIDNEY AND LIVERMONITORED WITH C4S NMR IMAGING4.4.1. Estimates of DMSO Diffusion Rates4.4.2. Measurements of DMSO Diffusion Coefficientsin Rat Organ Tissues4.4.2.1. Possible Errors Associated with Previous Methods4.4.2.2. Analysis of 1D Diffusion Data from NMR Images.4.4.3.3. Results of Diffusion Coefficient Measurements72727274757575• . . 76• • . 7676• . . 7779• . • 79• . . 83• . . 86• . . 8888• . 9599101102• . 102• . 102104• • 105• . 106107• . 108• . 108109• • 112112118118118122123123131131132viand Comparison with Literature Data . . . . 135142144CHAPTER 5:INTO5.1.STUDIES OF THE PERFUSION OF CRYOPROTECTIVE SOLUTIONINTACT RAT KIDNEY 146INTRODUCTION 1465.1.1. Introduction of CPS into an Intact Organ 1465.1.2. Choice of Vehicle and Cryoprotective Solutions . 1475.1.3. Perfusion Techniques 1505.1.4. Perfusion and C4S NMR Imaging 1515.2. EXPERIMENTAL 1515.2.1. Surgical Procedures and Blood Washout Techniques 1515.2.2. Perfusion Techniques 1545.2.3. Imaging of Rat Kidney 1595.2.4. 31P NMR Spectroscopy 1635.3. RESULTS AND DISCUSSION 1635.3.1. Preliminary Imaging Studies of Rat Kidneys 1635.3.2. Perfusion by Syringe 1705.3.3. Perfusion by Gravity 1765.3.3.1. Comparison of Syringe and Gravity Blood Washout ofKidney 1765.3.3.2. Study of Gravity-fed Blood Washoutsand New Surgical Procedure 1785.3.3.3. Study of Gravity—fed Perfusion at Two Temperatures1821875.3.4. Perfusion by Syringe Pump5.3.4.1. Development of a Method for theControlled Introduction and Removal of DMSO5.3.4.2. Determination of theJinount of Perfusates Required5.3.4.3. Use of Improved Perfusion Techniques inImaging Experiments at Three5.3.5. Freezing of DMSO-Perfused Kidneysand Subsequent Removal of DMSO5.3.5.1. Preliminary Imaging Studyof Freezing Kidney5.3.5.2. Freezing and DMSO Washout Experiments .5.3.5.3.1. Comparison of perfusion temperatures5.3.5.2.2. Freezing of kidneys for longer periodstime5.3.6. 31P NMR Study of Excised Rat Kidney5.4. SUMMARYREFERENCES .CHAPTER 6: CONCLUSIONS AND RECOMMENDATIONSFOR FURTHER RESEARCHREFERENCESAPPENDIX A: PARAMETERS USED FOR THE IMAGING EXPERIMENTS . . . 234APPENDIX B: PHOTOGRAPHS OF MICROSURGERY . . . 250APPENDIX C: PULSE SEQUENCES AND RELATED EQUATIONS . .CHEMICAL SHIFT SPECIFIC SLICE SELECTIVE (C4S) SEQUENCE.SHORT TIME TO ECHO (STE) SEQUENCEEQUATIONS• . . . 256257• . . . 2582594.6. SUMMARY.REFERENCES187194200205205209209of215220222224228233APPENDIX D: HISTOLOGY OF SELECTED SAMPLES 260viiLIST OF TABLESTable 2.1. T1 measurements (seconds) in DMSO/H20/NaC1 solutions at varioustemperatures. (A) T1 values of the proton resonance of water, (B) T1values of the proton resonance of DMSO. An F denotes that the linewidthof the NMR signal was too broad (>1000 Hz) and/or that the T1 could notbe confidently measured from the data. Parentheses around data indicatethat the measurement probably has error >20%, due to linewidth problemsor poor signal—to—noise ratio 57Table 2.2. T2 measurements (milliseconds) in DMSO/H20/NaCl solutions, •atvarious temperatures. (A) T2 of the proton resonance of water, (B) T2of the proton resonance of DMSO. An F indicates that the linewidth ofthe NMR signal was too large, and/or that the T2 could not beconfidently determined from the data. Parentheses around data indicatethat the error is probably >20%, dueto linewidth problems or poorsignal—to—noise ratio. . . . . 58Table 3.1. Relaxation measurements in rat organ tissues 90Table 4.1. Effective diffusion coefficients at various temperatures measuredby two methods: the “volume averaged” method is simply the rea of thesample in the image plane divided by the equilibration time; the “curvefit” method is done by fitting signal intensities from a row of pixelsthrough sequential images to Equation 4.4 (time since beginning ofexperiment in parentheses) . The first method should be comparable toliterature data from radiolabellirig studies 132Table 5.1. Solute compositions of different vehicle solutions .. . . . 150Table 5.2. List of parameters of imaging experiments monitoring perfusion162Table 5.3. List of parameters of imaging experiments: DMSO removal163viiiLIST OF FIGURESFigure 1.1. The vector j precesses about at the Larmor frequency V0..7Figure 1.2. Splitting of the nuclear spin levels for a spin 1/2 nucleus inthe presence of a magnetic field BFigure 1.3. (A) Magnetic moments of an ensemble of spins precessing about(B) The net magnetization vector M... . . . ..9Figure 1.4. (A) The effective field eff (B) Application of an RF fieldalong x’ causes the net magnetizationM to rotate about that axis. 13Figure 1.5. The magnetic moments of individual spins dephase after the RFpulse, due to the exchange of energy between nuclei.14Figure 1.6. The inversion-recovery (IR) experiment for measuring T. Thesample is excited with a 1800 RF pulse, then allowed to relax forvarying periods of time t. A 90° RF pulse then brings the magnetizationinto the x’y’ plane for observation of the FID.19Figure 1.7. The Carr—Purcell—Meiboom-Gill (CPMG) experiment for measuring T2(only one of the four phase cycles is shown). The sample is excitedwith a 90° RF pulse, then the magnetization is allowed to dephase for atime t before being inverted by a 180° RF pulse, which refocuses themagnetization at 2t, at which point the FID is recorded and then Fouriertransformed.. . . . . . . 20Figure 1.8. Pulse shapes and their Fourier transforms. The Fouriertransforms are approximately the same shape as the excitation (slice).(A) A square pulse shape (microseconds in length) yields a broadexcitation. (B) A gaussian pulse shape (milliseconds in length) yieldsa small excitation bandwidth. (C) The “sinc2” pulse shape gives anarrow, almost rectangular excitation 25Figure 1.9. A one-pulse experiment performed on a sample (two tubes ofwater) . The result is a “projection” of a cross section of the twotubes when the FID is acquired in the presence of a z field gradient:all the signal intensity at each point in the sample has been projectedonto a line parallel to the applied gradient.27Figure 1.10. Schematic representation of the slice selection protocol. Thesample, a tube of water (b), is subjected to a field gradient (a) afterexcitation. If the excitation is nonselective, the result is aprojection (C); if selective, the shape of the excited region is similarto the Fourier transform of the shape of the excitation pulse (d),corresponding to a slice through the sample perpendicular to theixdirection of the gradient 28Figure 1.11. The spin—warp imaging sequence. A selective 90° RF pulse incombination with a magnetic field gradient along the z axis is used toexcite a slice of the sample; the gradient along x is used to frequencyencode the data. A 1800 RF pulse inverts the magnetization and causesan echo at TE. The sequence is repeated at TR, changing the amplitudeof the field gradient along y for each repetition, to phase encode the data.31Figure 1.12. The relationship between imaging parameters TE and TR and theimage contrast type CT1, T2, or spin—density) 32Figure 1.13. An illustration of T1 and T2 contrast discrimination between twotissues with different relaxation times. (A) At a specific value of TR,differences in T1 cause the signal intensity from the two tissues todiffer by an amount dependent on the difference their relaxation curves(a and b). (B) Similarly, differences in T2 cause the T2 contrast tovary depending on the TE and the difference in relaxation rates Cd and c).33Figure 1.14. (A) Illustration of the results of a 90° selective RF excitationon a sample of two tubes containing water and fat, which have differentproton chemical shifts (1). The projections (2) in the presence of thegradient G2 are offset, so that the selective excitation, centered onfrequency a, selects spatially different slices of the magnetizationsfrom water and fat tubes. (B) The above explanation describes thesituation in (2); then (3) illustrates the excitation due to theselective 180° RF excitation and reversed G. (4) The resultant slicewhose magnetization will be refocussed during acquisition. . . . 35Figure 1.15. The Chemical Shift Specific Slice Selective (C4S) imagingsequence. The 180° RF pulse is made selective, and the slice selectiongradient reversed, in order to avoid refocussing the magnetization inthe off-resonance slice (which has a different chemical shift).36Figure 2.1. Relationship between correlation time and spectral density. Atlow temperatures, ; is long, and most of the spectral density is at lowfrequencies (i.e. molecular tumbling is slow) . At high temperatures, ;is short, and there are slow, medium, and fast components to themolecular motion. The area under the curve is constant for all ;.43Figure 2.2. Dependence of the relaxation times T1 and T2 on the correlationtime ç. . . . 45Figure 2.3. The coherence of transverse magnetization is affected by changesin the relative phases of individual magnetic moments due to exchange ofenergy via T2 mechanisms. Faster exchange gives a shorter T2 and abroader distribution of frequency components about o, i.e. a broaderlinewidth 46Figure 2.4. NMR spectra of water in the liquid state (A) and solid state(B) 48Figure 2.5. NMR images of a transverse slice taken through a bundle ofcapillary tubes (black circles) filled with water and surrounded byxFigure 2.5. NMR images of a transverse slice taken through a bundle ofcapillary tubes (black circles) filled with water and surrounded bywater, contained within a thin—walled outer glass tube. The temperaturedecreases from ambient (A) to —10°C (B-F) 49Figure 2.7. Surface plots of T1 and T2 in DMSO/H20/NaC1 solutions as a functionof the temperature and the DMSO concentration of the solution. A. T1 ofthe proton resonance of DMSO (s). B. T1 of the proton resonance ofwater (s). C. T2 of the proton resonance of DMSO (s). D. T2 of theproton resonance of water (s). . . . 61Figure 2.8. A. Diagram of a test sample, a 10 mm NMR tube filled with DMSO andcontaining capillary tubes filled with DMSO/H20/NaC1 solutions % watershown). The capillary tubes are surrounded by pure DMSO. B.Quantitative image of water protons in the test sample, shown as anegative (i.e. higher signal intensity is indicated by darker color).C. Rows of pixel intensities through the image. Numbers indicate areasrelative to the average areas corresponding to the two capillariescontaining 100% water. D. Stacked plot representation of the image.64Figure 2.9. Quantitative NMR images of the water protons in the DMSO/H20/NaCltest sample in Figure 2.8(A) at various temperatures. A. 0°C. B. -20°C. C. -40°C. D. -60°C. Negative images are shown above, with rowsof pixel intensities through the images shown below (a, b, C).Artifacts in the row projections are due to rotation of the images. 65Figure 2.10. Quantitative NMR images of the DMSO protons in the DMSO/H20/NaC1test sample shown in Figure 2.8(A) at various temperatures. A. 0°C. B.—20°C. C. -40°C. D. -60°C. Negative images are shown above, and rowsof pixel intensities shown below (a, b, C). Artifacts in rows are dueto rotation of images. The pure DMSO surrounding the capillary tubes inthe test sample is not observed, since it freezes at 18.3°C. . . . 66Figure 2.11. Phase diagram for the DMSO/water system, modified from Rasmussenand MacKenzie, Nature 220, 1315—1317 (1968). Area A, liquid; area B,liquid plus solid water; area C, liquid plus solid DMSO; area D(shaded), a complex set of phases not accessed in the present set ofexperiments 68Figure 3.1. Effects of T1 contrast on the calculated DMSO signal using ashort-TR protocol with TR = 1 s, compared with the model which assumesthat the contrast will reduce the signal by a scalar amount equal to theaverage contrast for all DMSO concentrations. Data shown from 20°C to—26°C 81Figure 3.2. Effects of T1 contrast on the calculated water signal, using ashort-TR imaging protocol with TR = 1 s, compared with the model whichassumes that the contrast will reduce the signal by a scalar amountequal to the average contrast for all water concentrations Data shownfrom 20°C to —26°C 82Figure 3.3. Effects of combined T1 and T2 contrast on the calculated DMSOsignal, using a short-TR imaging protocol with TR = 1 s, compared withthe model which assumes that the contrast will reduce the signal by ascalar amount equal to the average contrast for all DMSO concentrations.Data shown for —46°C and -63°C. 84xiFigure 3.4. Effects of combined T1 and T2 contrast on the calculated watersignal, using a short—TR imaging protocol with TR = 1 s, compared withthe model which assumes that the contrast will reduce the signal by ascalar amount equal to the average contrast for all waterconcentrations. Data shown for —46°C and —63°C 85Figure 3.5. Effects of combined T1 and T2 contrast on the calculated DMSOsignals using a short—TR imaging protocol with TR = 1 s, TE = 16 ms,from -63°C to 20°C. The relationship between signal intensity and DMSOconcentration is nearly linear for all temperatures. . . 87Figure 3.6. T1 of water and DMSO protons in rat kidney and liver equilibratedwith 0-50% DMSO in University of Wisconsin (UW) solution. T1 relaxationtimes are generally less in rat organ tissues than in solution, but havesimilar dependencies on the DMSO concentration 90Figure 3.7. T2 of DMSO and water protons in rat liver equilibrated with 0-50%DMSO, and in DMSO/I-{20/NaC1 solutions. T2 relaxation times for DMSOprotons are always longer than those of water protons. T2 relaxation intissues is an order of magnitude faster than that in solutions.. . 91Figure 3.8. Temperature dependence of the T1 relaxation times of the waterprotons in rat kidney and saline solution (aqueous 0.15 M NaCl). T1values decrease with temperature in both samples 92Figure 3.9. Temperature dependence of the T2 relaxation times of the waterprotons in rat kidney and saline (aqueous 0.15 M NaC1). T2 decreaseswith temperature in both samples . 93Figure 3.10. Temperature dependence of the T1 of the DMSO and water protonsin rat liver equilibrated with 50% DMSO, and in 50% DMSO in aqueous 0.15m NaC1. T1 relaxation times for DMSO protons are always longer thanthose of water protons. The largest T1 relaxation times for both waterand DMSO are at the highest temperature. . . . . 95Figure 3.11. The calculated water signal intensity in rat kidney and liverimages with combined T1 and T2 contrast (error bars) compared with themodelled signal intensity and its 5% error lines (dashed) . Data shownfor two temperatures, 20°C and 1°C. T1 and T2 contrasts in liver tissueare expected to cause the measured signal intensity to vary with waterconcentration in a nonlinear fashion 97Figure 3.12. The calculated DMSO signal intensities (error bars) in ratkidney and liver images along with modelled signal intensities and 5%error lines (dashed). The signal intensities vary with DMSOconcentration as expected from the model; contrast has simply reducedthe signal intensities at each concentration by a scalar factor. . 98Figure 4.1. Arrangement of tissue samples in the NMR tube. . . . 109Figure 4.2. An example of signal intensity data from a diffusion path throughan image. (50% DMSO in saline solution diffusing into rat liver.). 112Figure 4.3. Cross—sectional images of DMSO protons in a piece of bovine liver(3 mm x 6 mm in the imaging plane) surrounded by 30% DMSO in deionizedwater, at room temperature. The dark area in the center of the firstimage corresponds to the tissue sample. The DMSO solution surroundingit does not appear homogeneously bright due to field inhomogeneity. Thexiitime since the addition of the DMSO solution is given next to each imagein minutes. Diffusion of DMSO from outside the tissue sample (brightarea) is seen as an increase of the signal intensity inside the tissuefrom one image to the next 115Figure 4.4. Cross-sectional images of DMSO protons in a piece of bovine liversurrounded by 30% DMSO in 2H0, at room temperature. The time since theaddition of the DMSO solution is given next to each image in minutes.Diffusion of DMSO from outside the tissue (bright area) is seen as thesignal intensity inside the tissue increases from one image to the next.The gain was changed during the series of images, so the signalintensities are not directly comparable between images 116Figure 4.5. Cross—sectional images of DMSO protons in a piece of bovinekidney surrounded by 30% DMSO in 2H0, at room temperature. The timesince the addition of the DMSO solution is given next to each image inminutes. Diffusion of DMSO from outside the tissue (bright area) isseen as the signal intensity inside the tissue increases from one imageto the next. The gain was changed during the series of images, so thesignal intensities are not directly comparable between images. . . 117Figure 4.6. Images of DMSO protons in (A) rat liver and (B) rat kidneyequilibrated with 20% DMSO in UW solution. A row of pixel intensitiesthrough the image (path shown in white) is plotted beside image. Imageparameters were set so that TR = 15 s, TE = 16 ms, so that the imagingsequence was as close to quantitative as possible (it should only beaffected by T2 contrast). The apparent differences in concentrations ofDMSO between the solution and the liver and kidney tissue areapproximately 31% and 25% respectively.. . . . 120Figure 4.7. A series of images of DMSO protons in 50% DMSO in saline (aqueous0.15 M NaC1) diffusing into a sample of rat kidney at 23°C. The timesince the beginning of the experiment is shown next to each image inminutes. Equilibration has occurred after approximately 2 hours 30minutes (150 minutes) in a sample approximately 4.5 x 5 mm. Thus theDMSO is diffusing at an approximate rate of 2.9x105 cm/s in the imagingplane.. . . . . . . . . 125Figure 4.8. A series of images of DMSO protons in 50% DMSO in salinediffusing into a whole kidney with the time since the beginning of theexperiment given next to each image in minutes. The in-plane diffusionrate (half the sample width divided by the time to full equilibration)for 50% DMSO in saline in a whole kidney (with capsule intact) wasestimated to be 4.3x106 cm/s at 20°C. 126Figure 4.9. A series of images of DMSO protons showing the diffusion of 50%DMSO in saline into a 7.5 mm x 7.5 mm sample of rat liver at 25°C.Equilibration time was approximately 360 mm; thus the diffusion rate is1.7x105 cm/s 127Figure 4.10. A series of images of DMSO protons in 50% DMSO in salinediffusing into a 7 x 7 mm sample of rat kidney at 8°C. The time sincethe beginning of the experiment is given next to each image in minutes.Equilibration has occurred after approximately 5-16 hours in solution,although in this instance the sample is not homogeneous in the imagingplane. An estimate of in-plane DMSO diffusion rate in rat kidney atthis lower temperature is about 1.9x105 cm/s, although there is a largepotential error in terms of sample homogeneity 128xiiiFigure 4.11. A series of images of DMSO protons in 50% DMSO in salinediffusing into a 7 x 7 mm sample of rat liver at 10°C. The time sincethe beginning of the experiment is given next to each image in minutes.The sample is not homogeneous, so the estimated in—plane diffusion rate,5.lxlO6 cm/s. may not be as reliable as the measurements from otherimage sequences 129Figure 4.12. DMSO diffusion rates as a function of temperature. Valuesestimated from literature data are shown with values estimated from NMRimage data 130Figure 4.13. An example (same data as Figure 4.2) showing how data aremodified for fitting by eliminating data points from outside the sample(top) and how Equation 4.4 is fitted to the data (bottom). . . . 134Figure 4.14. The “volume averaged” method for measuring De(DMSO) from NMRimages compared with the more accurate method of fitting Equation 4.4 tosignal intensities from diffusion paths in the images. De(DMSO)measured by the volume averaged method is consistently higher than thatmeasured by the more accurate method. 137Figure 4.15. Measurements of De(DMSO) from image data compared with thediffusion coefficients of DMSO in solution (16) at various temperatures.Values of D(DMSO) in solution decrease with temperature according to anexponential relationship (16), and seem to show the same behavior in rattissues. The reported decrease in D(water) between solution and ratliver tissue is by a factor of approximately 3 (17). Here the decreasein D(DMSO) between solution and rat liver is by a factor ofapproximately 160 138Figure 4.16. A series of images of DMSO and water protons in a piece of ratliver equilibrated with 50% DMSO in saline (0.15 M NaC1) as temperatureis lowered. The temperature is given next to each image.. .. . . 141Figure 5.1. Schematics of the three different perfusion procedures used inthe experiments in this thesis. (A) Manually controlled perfusion with asyringe. (B) Gravity-fed perfusion, pressure equal to 100 cm water.(C) Syringe pump perfusion, highest pressure equal to 40 nun Hg.. . 157Figure 5.2. Rat kidney in a 15 mm NMR tube. Renal artery is catheterizedwith 0.28 nun I.D. polyethylene microtubing, which is threaded throughthe cap of the NMR tube and anchored to the tube with tape. Teflonmicrotubing (1 mm I.D.) is used as an aspiration line to remove excesssolution as it is perfused through the kidney 158Figure 5.3. (A) Schematic representation of the important structural featuresof the kidney. (B) Enlargement showing the geometry of the microscopicarchitecture. Groups of smaller structures including arteries, veins,and arterioles are visible in the images. Arrows show the direction ofblood flow. (C) Frame of reference for kidney slices. CD) A coronalsection (zy plane) of a rat kidney which has not been subjected to bloodwashout. The image slice is just above the renal pelvis so thisstructure is not seen. CE) A transverse section (xy plane) of a kidneywhich has been subjected to a gravity—fed blood washout. In both imagesthe cortex can easily be distinguished from the medulla. The renalpelvis is shown in the transverse image 166Figure 5.4. A series of transverse images of rat kidney held at 5°C, obtainedwith the STE spin—warp imaging sequence. TR was 0.5 s, and TE wasxivincreased from 5 to 80 ms in images A-E to demonstrate T2 contrast.Image F is a spin-density contrasted image (TE = 5 ins, TR = 12 s).Holding TE at 16 ins and increasing TR (T1 contrast) had no effect onimage contrast 167Figure 5.5. Transverse images of a rat kidney subjected to manually—controlled syringe blood washout. The dark areas correspond to thearcuate blood vessels (best seen in A and C), and the afferent (in thecortex) and efferent (in the medulla) arterioles, which still containblood despite the addition of heparin to the washout solution. . . 168Figure 5.6 Transverse images through a kidney subjected to a gravity-fedblood washout procedure. This procedure is more effective at bloodremoval than that demonstrated in the previous figure. However, smalllocalized blood clots are visible in the renal pelvis (images A-and D)and between the cortex and medulla (image B) 169Figure 5.7. Series of water and DMSO images of a rat kidney subjected tomanually-controlled syringe perfusion with 50% DMSO in saline. Each setof images was obtained after perfusioi of an aliquot of CP solution(volume shown at left of each image pair). Images show that there areblood clots left in the kidney (water image A), and that DMSO does notbecome homogeneously distributed 173Figure 5.8. DMSO and water images from kidneys subjected to manually—controlled syringe perfusion with 40 or 50% DMSO in saline after manualwashout by syringe. The first DMSO image (A) shows the “no—reflow”phenomenon, in which no perfusion of DMSO occurs. Images B and C showlarge blood clots in the arcuate arteries and arterioles whichcorrespond to incomplete DMSO distribution. Image D shows that avoidingcold storage after blood washout may improve perfusion, but this mightalso be due to somewhat more thorough blood removal in this sample.174Figure 5.9. Images of water protons during a perfusion with 20% DMSO insaline after manually-controlled blood washout. Images A and D show alarge amount of blood still left in the arcuate arteries, and B and Cshow blood clots in other blood vessels. The images do not correspondto the same place in the kidney. Each image was obtained afterperfusion with an aliquot of CP solution (volume shown below eachimage). . . . . . . . . 175Figure 5.10. A series of images of water and DMSO protons in rat kidneyduring a gravity-fed perfusion with 40% DMSO in saline. Although bloodwas washed out with the manually-controlled syringe method, it seems tohave left a smaller amount of blood in the kidney than in previousexperiments, mostly in the arterioles of the cortex. These smallblockages seem to cause a low perfusion rate (0.8 mi/hour). . . . 177Figure 5.11. Images of water protons in a series of slices from top to bottomthrough a rat kidney subjected to a gravity—fed blood washout. Darkareas at tops of the images are fat. Only very small dark areas arevisible, which seemed to correspond to cross-sections of small bloodvessels upon dissection. Less well—defined dark areas such as themedullar region in image D were associated with diffuse blood inefferent vessels. This technique seems much more effective at bloodremoval than the manually—controlled syringe method 180Figure 5.12. Images of water protons in a series of slices from top to bottomxvthrough a rat kidney subjected to a whole—body blood washout procedure.Dark areas at right in each image are fat. There is a noticeable lackof contrast in these images, due to the total removal of the blood.Subsequent histological examination showed that almost no blood cellsremained in the organ 181Figure 5.13. Pairs of images of water and DMSO protons in rat kidneysubjected to whole-body blood washout, during a gravity-fed perfusionwith 20% DMSO in saline at room temperature. Perfusion volumes andtimes are given next to each image pair. The very dark appearance ofthe cortex was unique to this rat and its sister 184Figure 5.14. An image of DMSO protons in the same kidney as the previousfigure, with no T1 contrast, after perfusion of 11 ml of CP solution.This image indicates that the blood was not removed from the corticaltissues 185Figure 5.15. Images of water and DMSO protons in rat kidney subjected to awhole—body blood washout and gravity—fed perfusion with 20% DMSO insaline at 10°C. Perfusion volumes and times are given next to eachimage pair. This image also indicates that the blood was not completelyremoved from the cortical tissues, and there are blood clots in theefferent arterioles (water images A and B) DMSO is well distributed tothe medulla 186Figure 5.16. The syringe pump used for perfusion of rat kidneys. The weighton the arm allows for release of pressure when the vascular resistancebuilds up enough to lift the weight, which corresponds to 40 mm Hg.This simulates a constant—pressure rather than a constant—rate delivery.190Figure 5.17. Water and DMSO images of rat kidney subjected to a 250 ml bloodwashout, during perfusion with 20% DMSO in saline using the syringe pumpfor delivery. Image A shows only one small blood clot remaining. ImageB shows DMSO in the renal artery (arrow). Image C shows DMSO perfusingthrough the arcuate arteries (arrow). DMSO distribution is almosthomogeneous in the last images (D) 191Figure 5.18. Images of water and DMSO protons in the rat kidney of theprevious figure during washout of DMSO with saline. In image A, theDMSO is being removed first from the arcuate arteries (compare to Figure5.17(C)). Some DMSO remains in the kidney after washout with 6 ml ofsaline . 192Figure 5.19. Quantitative proton spectra taken during the experiments inFigures 5.17 and 5.18. Each spectrum was acquired just after thecorresponding images. (Amounts of perfusate are beside each spectrum.)These spectra quantify the relative numbers of water and DMSO protonspresent in the solution and the kidney. (A) DMSO concentrationincreases as the kidney is perfused with CP solution. CB) DMSOconcentration decreases as the DMSO.is washed out with saline. . . 193Figure 5.20. Images of water and DMSO protons in rat kidney perfused with 21ml of 20% DMSO in UW solution at 22°C. DMSO concentration ishomogeneous (images A), which is emphasized by removing the surroundingfluid (DMSO image B). When fluid is returned to the NMR tube, thekidney is moved slightly. The water image B shows fat around the renalpelvis (arrow) and an air bubble (large round dark area) 197xviFigure 5.21. Images of water and DMSO protons in rat kidney during syringepump perfusion with 20% DMSO in UW solution at 10°C. The images showDMSO distributed throughout the kidney. Images taken after those shownin D indicated no further increase in the DMSO concentration. Thus, 10ml of solution seem adequate to completely equilibrate the kidneytissues with the CP solution 198Figure 5.22. Images of water and DMSO protons in the rat kidney shown in theprevious figure, during washout of DMSO with UW solution. After 20 mlof UW solution, there is still a small amount of DMSO left in thekidney. (The DMSO image in C was obtained with a very high gaiI setting,so is not directly comparable with the one before it.) 199Figure 5.23. Images of water and DMSO protons in a rat kidney duringperfusion with 10 ml of 20% DMSO in UW solution at 10°C. The large ovaldark area in the upper left of each image is a roll of Teflon tape forpositioning of the kidney, due to its small size. After 10 ml of CPsolution were perfused, the DMSO concentration was homogeneous.. . 202Figure 5.24. Images of water protons in rat kidney during a perfusion with20% DMSO in UW solution at 36°C. Image A shows some residual blood nearthe renal artery. As the perfusion continues, this blood moves to thearcuate arteries (B) and then into the arterioles (C and D). Thisperfusion was not successful, probably due to this blockage of the bloodvessels 203Figure 5.25. Images of water and DMSO protons in a rat kidney duringperfusion with 9 ml of 20% DMSO in UW solution at 36°C. There is someindication (images C and D) of blood clots moving through thevasculature (arrows) as in the previous figure, but they are muchsmaller. DMSO is successfully distributed throughout the kidney.204Figure 5.26. Images of water and DMSO protons in rat kidney equilibrated with20% DMSO. The dark oval object at top left is a roll of Teflon tapeused to position the kidney. The temperature was lowered as indicatedfor each set of images. Sheets of pure DMSO or pure water (visible tothe eye), which appear as straight dark lines in the images, begin toappear at -10°C. The loss of signal in the water images indicates thatwater protons are becoming less mobile but not necessarily that thewater has frozen 207Figure 5.27. Images of water and DMSO in the rat kidney of the previousfigure. Removal of the DMSO was attempted by perfusing with UW solutionat 10°C. The DMSO seems to be being removed by bulk diffusion to thesurrounding solution and possibly the solution in the renal artery,indicating that the vasculature may be blocked by swelling or some otherdamage 208Figure 5.28. Images of water and DMSO protons in a rat kidney subjected toperfusion with 10 ml of 20% DMSO in UW solution, then frozen at —20°Cfor one day. (A) Before perfusion with UW solution at 10°C, JDMSOdistribution is homogeneous. A swollen area (arrow) appears just belowa dark feature which could be a blood vessel in cross-section. (B) Whenperfusion begins, the dark feature is filled with solution. The swollenarea does not reperfuse well (arrow). (C) Some swelling of the kidneyhas occurred, but the DMSO has been mostly removed. The DMSO image wasobtained with a much higher gain than the previous DMSO image. . . 212xviiFigure 5.29. (A) Images of water and DMSO protons in rat kidney subjected toperfusion with 10 ml 20% DMSO in (3W solution at 0°C, and frozen for 1day at —20°C. After thawing to 10°C, DMSO distribution is homogeneous.There is a swollen area (arrow) similar to that in the kidney in theprevious figure. (B-D) Images of the kidney during washout with (3Wsolution. The swollen area does not reperfuse as well as the rest ofthe cortex, but removal is successful 213Figure 5.30. Image of water protons in rat kidney subjected to 0°C perfusionwith 10 ml 20% DMSO in (3W solution. Arrow shows an area possiblycorresponding to blood clots in the vasculature. This kidney wasseverely damaged by the perfusion procedure 214Figure 5.31. Images of water and DMSO protons in rat kidney from Figure 5.20.Kidney was frozen for 4 days at —20°C, then thawed to 10°C and imaged.The freeze/thaw process has caused swelling. The round dark features onthe surface of the kidney are air bubbles 217Figure 5.32. Images of water and DMSO protons in the kidney from Figure 5.19.This kidney was frozen for 13 days after perfusion with 18 ml 20% DMSO.The first pair of images show that the DMSO was not equilibratedthroughout the sample. This is probably due to residual blood in themedulla, seen in the first water image (arrow) . Washout was notsuccessful; the kidney did not reperfuse, though some DMSO was removedby bulk diffusion. The second water image shows that the kidney hasswelled. . . . - 218Figure 5.33. Images of water and DMSO protons in a rat kidney perfused with10 ml 20% DMSO in (3W solution at 22°C, then frozen for 13 days at —20°C,and thawed to room temperature. (A) The images show no obvious damageexcept a swollen area (arrow) - (B—D) Washout of DMSO with OW solutionwas attempted. Images of DMSO protons show that perfusion was uneven,again associating a swollen area with nonperfusion. However, theaddition of sucrose to the washout solution has prevented generalizedswelling due to osmotic imbalances 219Figure 5.34. 3’P NMR spectrum of a rat kidney which has been subjected to a250 ml blood washout and 35 minutes at 0°C, which was a typical timebetween excision and the beginning of perfusion with a CP solution inthe present work. Peaks in the spectrum correspond to (A)phosphomonoesters, (B) inorganic phosphate, (C) phosphodiesters, and ((3)NADH (reduced form of nicotinamide adenine dinucleotide) 221Figure 3.1. The dissected aorta held between two forceps, with the renalartery on the right. Black threads are silk sutures used to tie offblood vessels 250Figure B.2. A rat kidney, which has been subjected to a 250 ml gravity fedblood washout, and then catheterized. The silk sutures secure thecatheter into the renal artery 251Figure 3.3. A coronal slice through a rat kidney which has not been subjectedto a blood washout. There is still blood present (reddish color).252Figure 3.4. A coronal section of a rat kidney which has been subjected to a250 ml gravity-fed blood washout. Comparison with the previousphotograph shows that visually, as well as in the NMR images, most ofthe contrast between renal structures is due to the presence of blood.xviiiThe medullar tissues here appear virtually featureless. Structures thatare easily identifiable are the renal pelvis (center), the medulla(light colored) and the cortex (darker) 253Figure B.5. A transverse section through a rat kidney which has beensubjected to a 4 ml manually-controlled blood washout with a syringe.Notice that there is still blood present in the medulla and innercortex. The small red dots in the cortex are cross sections of bloodvessels which still contain blood This photograph is quite similar tothe NMR images 254Figure D.1. General structural features of the kidney at 25X magnification,including medulla (M) and cortex (C) with numerous glomeruli (G) whichare the blood filtration units of the nephrons, and an arcuate artery(A) and vein (V). The vein appears dilated, which suggests that theperfusate has passed through the vascular system under pressure. Thiskidney is shown in Figure 5.23, 5.26, and 5.27, and has been subjectedto 10°C perfusion with 20% DMSO in UW solution, freezing for 1 hour, anda DMSO washout with UW solution at 10°C. 261Figure D.2. Smaller structures of the kidney at 100X magnification, includingan artery (A) which branches into an arteriole above it, next to alymphatic (white). A glomerulus is visible as well. Most of the otherstructures are proximal convoluted tubules, and show pathologicalchanges consistent with acute renal ischemia (dead or dying cells aredarker) 262Figure D.3. General structure of a kidney subjected to perfusion at 34°C with20% DMSO in UW solution, shown at 25X magnification. The perfusionfailed, and the kidney was sectioned for histological examination.Comparison with Figure D.1 shows the same architecture, but indicationsthat autolysis of cells is advanced, and cells are losing their nuclei.• . . • .. . • . . ..263Figure D.4. A view of the previous slide at 100X magnification, showingsevere autolytic damage evidenced by loss of cell nuclei and disruptionof cell membranes. Ischemic changes are also present: Condensedproteinaceous material within Bowman’s space (B) of the glomerulus, anddilated tubules 264xixLIST OF ABBREVIATIONSAI4P Adenosine monophosphateADP Adenosine diphosphate -ATP Adenosine triphosphateCPMG Carr—Purcell—Meiboom--Gill NMR pulse sequence for measuringT2 relaxation timesCP Carr—Purcell NMR pulse sequence for measuring T2 relaxationtimesCP solution Cryoprotective solutionCPS Cryoprotective solventC4S Chemical shift specific slice selective NMR pulse sequencefor acquiring images of two chemical shifts simultaneouslyID Inner diameterIP Intraperitoneal (inside the abdominal cavity)IR Inversion recovery NMR pulse sequence for measuring T1relaxation timesIV IntraveinousNIH National Institute of Health (U.S.A.)NMR Nuclear magnetic resonancePCr PhosphocreatineRF RadiofrequencySTE Short time to echo NMR pulse sequence for acquiring a spinwarp image with a short TETE Time to echo, the time between the middle of the 9Q0excitation pulse and the middle of the acquisition in aspin—warp type NMP. imaging pulse sequenceTR Repetition time, the time between each of a series of pulsetrain repetitions in an imaging sequenceVT Variable temperaturexxACKNOWLEDGEMENTI am indebted to Professor Cohn A. Fyfe for his excellent insight anddirection, which was complemented by the support of his many graduate studentsand most especially by Dr. Hiltrud Grondey. The endless patience andhelpfulness of the tJBC Chemistry electronics and mechanics staff was awonderful tribute to the Canadian spirit. I would like to thank my family fortheir encouragement, and my brothers and sisters in the Richmond Church ofChrist for helping me feel less of a stranger in a strange land.Thanks to Dr. Richard xnmons for his insight into the diffusion ofcryoprotectants in tissues, and to Dr. Barry Fuller for his suggestions onpreservation solutions. Bless you, Betty Pearson, for your encouragement andmicrosurgical expertise.I thank God for teaching me to “surrender the hunger to say I must know!to have the courage to say I believe! for the power of paradox opens youreyes! and blinds those who say they can see”. I never could have imagined howwonderful it is to walk in the light of Jesus, who is indeed the Way, theTruth, and the Life.Finally I wish to thank my husband Brad, who stuck with me and made itall worthwhile.1CHAPTER 1CRYOBIOLOGY D N IMAGING1.1. CRYOBIOLOGY AND TISSUE PRESERVATIONThe study of cryobiology has very ancient roots. Probably the firstinterest in freezing of animal tissues arose from concerns about foodpreservation and prevention of frostbite. It has been known, possibly forthousands of years, that certain animals are more immune to the effects ofcold than are humans; for example, some insects, amphibians, and fish canwithstand subzero temperatures without an internal heat production mechanismby producing high concentrations of certain molecules in their blood, whichlower the freezing point of their tissues. It has been discovered that evenlarge mammals (1) have specialized ways of preventing damage to their tissuesfrom very low temperatures.The two most basic questions in cryobiology are: How do thecryoprotective designs found in nature work, and how can we imitate or modifythem for our own purposes? Modern interest in these questions was stimulatedin 1949, when Polge, Smith, and Parkes (2) accidentally discovered thataddition of glycerol to cell suspensions improved the survival of cells whenfrozen and subsequently thawed. Their chance discovery has prompted theinvestigation of many other compounds for possible cryoprotective properties.These studies have lead to the routine preservation of blood, semen, andembryos with various kinds of cryoprotectants.During the same period of time, during the 1950s and 1960s, techniques2for organ transplantation were being perfected. The most intransigent andfrustrating problem was the lack of time between retrieval of an organ and theensuing degradation of the tissues. This necessitates very quick transplantwith minimal study of the viability of the organ or whether the donor andrecipient tissue types are a good match. In fact, at present, tissue typingremains the most important and least utilized step in the transplant process,because tissue typing methods are time consuming, requiring intensive labworkup of blood samples from the donor. For this reason it is very importantto find ways of preserving organs for longer periods of time.Originally, surgeons were convinced that the reason transplanted organswere failing was due to the lack of an oxygen supply to the tissues during thetime between removal and implantation, and were trying different techniques tomaintain an oxygen supply to an explanted organ. Although the most obviousreason for the deterioration of an organ is the lack of oxygenation from ablood supply, oxygen—carrying substances have failed to preserve organs formore than a few hours at normal body temperature (3). Southard et al. (4)discovered that oxygen radicals actually cause damage in explanted organs. Analternative approach is to decrease the oxygen needs of the organ duringstorage by decreasing its metabolic rate. This can be achieved by inducinghypothermia (5)Cryobiologists recognized the analogy between preserving cells at lowtemperatures and preserving an organ between removal and reimplantation; theproblem was how to apply what had been learned about cell suspensions to themuch more complex systems of intact organs. Experiments on cooling andfreezing explanted organs were initiated, but it was soon discovered thatthere were so many variables that it was very difficult to even designexperiments. Hypothermic preservation of organs requires several steps, allof which have associated problems:1) washout of blood;2) introduction of cryoprotectant;33) cooling;4) freezing and storage;5) warming;6) reperfusion and assessment of viability.Some of these steps are actually carried out simultaneously in the clinicalsetting, but researchers have attempted to separate them experimentally inorder to understand how each procedure causes damage to tissue. These areasand their associated problems will be discussed further at appropriate pointsin this thesis.1.2. POTENTIAL USE OF NMR IMAGINGIN CRYOBIOLOGYThe study of organ preservation within the field of cryobiology hasseveral serious limitations. Since the beginning, there has not been anyreally detailed picture of why cryoprotectants work, what they do, and how touse them effectively. Despite the theoretical work of Pegg, Mazur, and others(6,7,8), most cryopreservation solutions contain a mixture of components whoseactions and interactions are not well understood. There are large numbers ofvariables in cryopreservation experiments, and the techniques themselves aredifficult, time consuming, and expensive.One of the problems with studying cryopreservation mechanisms is thatthere has been, until recently, no technique available for visualization ofthe interactions of cryoprotectant solvents and the organ tissues they arebeing used to preserve. There was no definitive way to determine thepermeability of tissues to cryoprotectants, determine bulk diffusion rates,monitor perfusion processes, and determine when a cryoprotectant hasequilibrated in an organ. It is quite possible that many failures in4cryopreservation of organ systems are due to the lack of this kind offundamental information. This thesis will demonstrate that an NMR imagingtechnique, Chemical Shift Specific Slice Selective (C4S) imaging can solve allof the above problems.Although there has been extensive development of specialized microscopesfor freezing cells and slices of tissue at specified rates, these instrumentshave the following limitations: The sample must always be destroyed; the depthof field is quite small; and it is difficult to prove that solids (icecrystals) have formed. The microscope can not reveal what fluids are leftunfrozen at specified temperatures. The C4S NMR imaging technique can detectfreezing wherever it occurs within a chosen volume of tissue. The area to beimaged can be chosen without physically disturbing the tissue, and can bechanged at will, solving the depth of field problem. C4S NMR imaging also hasthe advantage of being able to independently determine the concentrationdistributions of water and cryoprotectant present as mobile liquids at anygiven temperature.1.3. CHOICE OF A CRYOPROTECTANT SOLVENT SYSTEMFor the purposes of thi thesis, a specific cryoprotectant system waschosen to demonstrate the characteristics and abilities of the C4S NMR imagingtechnique. The cryoprotectant solvent system had to satisfy the followingrequirements:(1) The potential solvent system should lower the freezing point ofwater in the tissues, since many of the possible mechanisms forcryoprotection are related to the colligative properties of thesolvent;(2) The solvent must be able to penetrate the tissue and its constituent5cells quickly, thus avoiding the osmotic stresses which are amajor cause of cellular damage during cryoprotection procedures;(3) The solvent must be nontoxic at molar concentrations (9), since highconcentrations of solvent are needed to provide cryoprotection atvery low temperatures;(4) The solvent molecules must have protons whose chemical shift is farenough separated from that of water that the C4S imaging sequencecan be successfully implemented with the available microirnagingaccessory.Of the solvents studied, those fitting the criteria best were: Glycerol(which is a natural cryoprotectant) (2), ethylene glycol, and dimethylsulfoxide (DMSO) (10, 11). Ethylene glycol has proven to be too toxic for useas an organ cryoprotectant. DMSO is also toxic to organs at concentrationshigher than 50% and at temperatures higher than 4°C (12), but much lessconcentrated solutions have been shown to have cryoprotective properties (13).Glycerol is less toxic, but moves very slowly across cell membranes (8),causing osmotic stress to the cells. Thus DMSO is perhaps the best potentialcryoprotective solvent for organs. Fortunately DMSO protons also have achemical shift value very different from that of water protons. DMSO wastherefore chosen for investigating the potential of NMR microscopic imaging asa probe of the cryoprotection process as described in this thesis.1.4. NUCLEAR MAGNETIC RESONANCE (NMR)1.4.1. Basic NMR TheoryThe imaging technique used in this thesis is based on nuclear magnetic6resonance, so a brief introduction to the basic NMR experiment will be givento facilitate subsequent understanding of the imaging procedures.The technique of nuclear magnetic resonance has its basis in the factthat all nuclei with a nonzero nuclear spin possess magnetic moments. Themotion of the nuclear magnetic moment vector of such a nucleus in a staticapplied magnetic field is described by:[1.1]2nwhere y is the gyromagnetic ratio for the given nucleus, h is Planck’sconstant, and T is the spin angular momentum vector. The vector precessesabout at a frequency characteristic of the particular nucleus (Figure 1.1)= [1.2]° 2rrwith units of Hz (1 Hz 1 cycle/sec.),= 2rr\ = yB0 [1.3]with units of radians/second., Equation [1.2] is known as the Larmor equation,and V0 the Larmor frequency. The presence of the static magnetic fieldcauses a splitting of energy levels, such that a nucleus with spin quantumnumber I has 21+1 levels. For example, for protons, I is equal to 1/2, sothere are two energy levels (Figure 1.2). As shown in the figure, the energy7Figure 1.1. The vector j precesses about atthe Larmor frequency V0.-E—= exp(—)kTseparation of the levels iszV0fBoyhB0 [1.4]2nThe energy of thisseparation lies in theradiofrequency (RF) range,so transitions between thetwo levels can be stimulatedby RF radiation. Therelative populations of thetwo levels are governed bythe Boltzmann relation:[1.5]xy8where N and N.. are thepopulations of the lower andupper states, respectively, kis the Boltzmann constant, aridP is the temperature indegrees Kelvin. The basis ofall NMR experiments is theabsorption of RF energy by thenuclear spins, resulting in achange in the populations ofthe energy levels.Reattainment of equilibriumpopulations is observed whenthe RF field is removed. Thedetection method will bedescribed below.Figure 1.3(A) illustrates the magnetic moments of an ensemble of spins,and 1.3(B) shows the simpler picture which treats the magnetic moments as asingle composite vector, M, the net macroscopic magnetization vector. Noticethat since the x and y components of these vectors are randomly oriented, thesums M and M are zero:////=yfBB0=O B0=BFigure 1.2. Splitting of the nuclear spinlevels for a spin 1/2 nucleus in the presence ofa magnetic field= 1 +3M +M =x y z z[1.6]9xzBz4 bIBoMyFigure 1.3. (A) Magnetic moments of an ensemble of spinsprecessing about.(B) The net magnetization vector M.A tBoV0y10It is much easier to visualize the behaviour of spins and magneticmomentum vectors if the discussion is in terms of a frame of reference wherethe applied RF field appears to be static rather than rotating; i.e. if theframe of reference is rotating around the z axis at frequency /2n, thefrequency of the applied RF field. This is a mathematical translation fromthe usual frame of reference, called the “laboratory frame”, to a rotatingframe of reference. The movement of the magnetic moment can now be viewedwith respect to the fields and , both of which are now static intherotating frame of reference. The logic of this transformation is as follows.If the derivative of [1.6] is taken with respect to time, rotation of themagnetization vector about the applied magnetic field at frequency c/2n isexpressed (after some manipulation) as the cross product:dM - - -(-j>L6xMVMxBowhere the subscript L denotes the laboratory frame of reference. If the wholethree—dimensional coordinate system is rotated (mathematically) at thefrequency /2n, thendZ? a? - -(__) (—) +xM [1.8]dtL Rwhere the subscript R denotes the rotating frame. Combining these twoequations, and rearranging terms we get the equation() = x ( [1.9]0This can be rewritten as11= x [1.10]where-[1.11]V1.4.2. Fourier Transform NMRIn a pulse NMR experiment, a pulse of RF radiation of short •duration isapplied to the system, creating a field along the x axis of the rotatingframe:= + . + [1.12]eff 0 1(Figure 1.4(A)). If the pulse is applied exactly on resonance, then thefrequency of the rotating frame /2n is equal to the Larmor frequency‘,and=—- ; B0 + = 0 [1.13]‘0’ V12Then equation [1.12] becomes:Bff=[1.14]In other words, in the rotating frame, for a pulse applied exactly onresonance, the magnetic moments of the nuclei can be treated as if theyexperience only the RF field during the pulse. The net magnetization Rnow precesses about the effective field fleff’ and thus rotates in the y’z’plane around this axis (Figure 1.4(B), wher is applied along the x’ axisin a positive direction).13XIXIZIyeylFigure 1.4. (A) The effective fieldeff• (B) Applicationof an RF field along x’ causes the net magnetization Rto rotate about that axis.AIBoB0-&yBZIfBMSAL14If the field is applied for a time ti,, then during that time M movesthrough an angle 8:8 = YB1t = [1.15]where B1 is the magnitude of the applied RF field , t is the time thefield is applied, and is theangular velocity of the spinsunder the influence of the appliedRF field.After the RF field i-s turnedoff, the magnetization is againunder the influence of only thestatic field B, and so precessesaround the z’ axis (identical tothe z axis). This precession ofthe net magnetization induces avoltage in the receiver coil,which is detected and digitized asFigure 1.5. The magnetic moments of a Free Induction Decay (FID). Asindividual spins dephase after the RFpulse, due to the exchange of energy the magnetic moments of individualbetween nuclei.spins precess in the x’y’ plane,they begin to dephase (Figure 1.5) due to the exchange of energy betweennuclei. This entropy effect causes the observed signal to decay with a timeconstant T2 called the spin-spin, or transverse, relaxation constant. Inaddition to this, thermodynamic equilibrium is also reestablished by loss ofenergy to the surroundings, or “lattice”; this process is governed by the timeconstant T1, the spin—lattice or longitudinal relaxation constant. Thebehaviour of the components of the net magnetization vector along each axis ofXIyl15the coordinate system, M, M, and M, can be described by the equations:dM M dM M dM-(M -M)x__ x y_._• y z_ z 0 [116]dt T2’ dt T2’ dtThese equations are known as the simplified Bloch equations. (For themeasurements and further descriptions of T1 and T2 see Section 1.4.3.)Other processes can also affect the time course of the NMR signal. Ingeneral, if any process results in the nuclei being exposed to an RF fieldwhich is not at their Larmor frequency, the terms - ?iS/Y in equation [1.121will not cancel. Then is not equal to , and the magnitude of thefield fleff will depend on the difference between the frequency of the appliedfield and the Larmor frequency.When the difference between RF frequency and Larmor frequency is causedby inhomogeneity in the static field , the signal detected will consist ofa distribution of frequencies, which is observed experimentally as an increasein the linewidth:—I- i- + ytiB [1.17]The most important cause of shifts in resonance frequency from theexpected Larmor frequency is the “chemical shift”: the resonance frequency ofsignal from one species of nucleus (e.g. protons) changes depending on thelocal electronic (and therefore chemical) environment of each particularnucleus. These characteristic shifts in the Larmor frequency are caused bytiny induced local magnetic fields due to circulation of electrons in amolecule in the presence of the magnetic field ff. Shifts in Larmor frequencycaused by local fields can vary depending on the orientation of the molecule16with respect to the static magnetic field, and the chemical and physical stateof the lattice (the surrounding environment). This is the basis of NMR as adiagnostic technique, because nuclei that are of the same species but indifferent local magnetic fields have slightly different Larmor frequencies.It is possible to perform quantitative NMR experiments in which theintensities of the chemically shifted signals reflect the relative numbers ofnuclei in each local environment. The signal intensities are determined byapplying a Fourier transform to the time—domain signal, resulting in afrequency-domain spectrum. Thus the local magnetic fields in different sitesof a molecule and in different molecules can be compared. This type ofquantitative experiment will be discussed in the next section.The chemical shifts observed for different nuclei in molecules can berationalized, but not predicted in an absolute sense. Thus the chemical shiftis an experimental value based on measurement against a reference standard fora particular nuclear type, and is defined as:(ppm) = nucleUs — “standard 106 [1.18]standardSince the induced local magnetic fields are proportional to , this scalemakes the value of the chemical shift independent of the strength of thestatic magnetic field so data obtained at different field strengths can becompared directly. The differences in chemical shifts are small compared tothe Larmor frequency of the nucleus and are expressed in units which arefractions of the main static field. These fractions are parts per million(ppm) and the range of shifts for different proton environments is 10 ppm.171.4.3. T1, T2 and Quantitationin Pulse NMRThe ability to determine the relative numbers of nuclei in a givenpositions in a molecule, or, using a standard, to quantify the concentrationof a given nucleus present in a sample, is one of the most useful aspects ofNMR spectroscopy. The techniques involved require that the time constants T1and T2, describing the recovery to equilibrium of the longitudinal antransverse magnetization components, are known for the nucleus underinvestigation, and that the experimental parameters are carefully chosen withrespect to these time constants (which are defined in Equation 1.16). Thissection will be a brief description of techniques for the measurement of T1and T2 in samples like those studied in this thesis.The technique most commonly used to measure T1, the spin—latticerelaxation time, is called the inversion-recovery experiment (14). In thisexperiment, the sample is excited with a 1800 RF pulse, then allowed to relaxtoward for a period of time t; a 90° RF pulse then brings the partly“relaxed” magnetization back into the x’y’ plane for observation of the freeinduction decay (Figure 1.6). The experiment is repeated for various valuesof i. The FIDs are subjected to a Fourier transformation, and the amplitudesof the signals fitted to the solution of equation 1.16]. The inversion—recovery experiment depends on the sample magnetization, M being the same foreach repetition. This is the case only if the period between repetitions ofthe pulse sequence is greater than 5 times T1 so that complete recovery toequilibrium has taken place before the experiment is repeated and the nextsignal added.The method used to measure the spin-spin relaxation constant T2 iscalled the CPMG sequence (15), shown in Figure 1.7. The sample is excitedwith a 90° pulse, then the magnetization in the transverse plane is allowed to18decay for a time i before being inverted by a 1800 pulse. This pulse servesto reverse the dephasing due to chemical shift distribution or inhomogeneitiesin the magnetic field while the real entropic dephasing due to spin-spinrelaxation described by T2 (Equation 1.16) is nonreversible. Thus, after atime 2T, the signal intensity is solely determined by T2 effects. The FID(actually the latter half of a spin echo) is acquired, transformed, and thesignal amplitudes fit to the solution of equation [1.16). Notice that for theexperiment to be reliable, the value of T1 must be known in order to set therepetition time at five times T1.Once the T1 and T2 values for a signal from a specific nucleus, say theprotons of water in a sample, are known, their concentration in the sample canbe quantified using one of a number of techniques. For the purposes of thisthesis, it was not necessary to obtain the absolute quantities of the specificnucleus. Therefore, the relative amounts of each nucleus in each environmentwere measured. This can be done, in both spectroscopic and imagingexperiments, by setting the repetition time to 5 times the longest T1 value,and setting the time available for transverse relaxation during an experimentto less than 20% of the smallest T2 value in the sample. A detailedexplanation of the relationship between T1 and T2 values and the experimentalparameters needed for quantitative imaging is given in Chapter 2.x Bt2M = 0__I,’__positivesignalFigure 1.6. The inversion—recovery (IR) experiment for measuring T1. Thesample is excited with a 1800 RF pulse, then allowed to relax for varyingperiods of time t. A 90° RF pulse then brings the magnetization into the x’y’plane for observation of the FID.19zM0) y•4,1800ti<0ylXInegativeFID FT signalZIZIyIz,90x0-*goxogoxoMXIXIFID - signalM,> 0FIDyI.>MFT20Zr4,ZrxrXIyryrMZrXI XIZr ZrXI180,°y,-*180,°180,7EchomaximumZr 4rFIDFTy,-+FIDFTyr—>I.L FIDFTyr -*XIZrt3->Zr4’XIFigure 1.7. The Carr—Purcell—Meiboom—Gill (CPMG) experiment for measuring T2(only one of the four phase cycles is shown). The sample is excited with a90° RF pulse, then the magnetization is allowed to dephase for a time t beforebeing inverted by a 180° RF pulse, which refocuses the magnetization at 2t, atwhich point the FID is recorded and then Fourier transformed.211.4.4. Temperature Effects on T1, T2Relaxation of excited nuclear spins can occur by different mechanisms indifferent systems. In general, excited spins can relax if they are exposed toany local magnetic field alternating at the correct frequency, that is, theirLarmor frequency. Interactions can occur with fields being modulated at theLarmor frequency by the effects of various molecular motions, such asrotations or translations, or by internal motions of electrons within themolecule which cause its magnetic moment to fluctuate.The most important relaxation mechanism for spin 1/2 nuclei is thedipole—dipole interaction between nuclei. The frequency of modulation of thedipole fields is temperature dependent (since molecular motion is temperaturedependent), which means that the relaxation times T1 and T2 of protons in asample will change if the temperature is lowered. The implications forquantitative imaging of tissue samples at different temperatures will bediscussed in Chapters 2 and 3.1.4.5. Phosphorus-3l NMR SpectroscopyAlthough most of this thesis focuses on the use of proton NMRspectroscopy and proton NMR imaging, there is another nucleus which is usefulin the investigation of biological samples, phosphorus-31. Phosphorus plays acrucial part in the metabolism of cells, since it is a component of the so—called “energy molecules”: Adenosine triphosphate (ATP), adenosine diphosphate(ADP), aderiosine monophosphate (AMP), and phosphocreatine (PCr). Thesemolecules are the principle energy storage molecules for cells of all bodyorgans, although their relative importance depends on the specific organ. Allcellular processes such as replication, maintenance of chemical gradients,22etc., depend on the ability to build these molecules and to break them downagain for energy production.Since the basic metabolic processes depend on these molecules, the stateof health of cells (in the sense of energy availability) can be probed byphosphorus NMR spectroscopy. An NMR spectrum of the phosphorus in a tissuesample can yield information about the relative amounts of adenine nucleotidespresent in the tissue. This information may be related to the ability of thecells of an organ to maintain or resume basic functions after being removedfrom a donor’s body or after they have been subjected to a cryopreservationexperiment (see Chapter 5).1.5. NMR IMAGINGNMR imaging extends NMR spectroscopy to include three dimensionalspatial information about the distribution of nuclei within a sample. Animage consists of absolute signal intensities from each of the volume elements(voxels) within a selected volume. In principle, it is possible to sampleeach volume element sequentially (“sensitive point” imaging), but usually thesignals are encoded in a more efficient manner, where signals are collectedsimultaneously from many voxels, which allows a computer to aid in thecollection and processing of data. The two most common techniques of thistype are back projection reconstruction (16) and Fourier imaging. Backprojection is also used in other types of tomography, such as PET (PositronEmission Tomography) and CAT (Computer Aided Tomography) scans; but Fourierimaging is now the method of choice in most NMR imaging. There are manydifferent variations on this pulse sequence, developed for specificapplications or equipment. This section will describe only the generaltechnique and the two variations used in this thesis.23Two concepts are crucial to the understanding and use of NMR imaging:frequency selective RF pulses, and linear magnetic field gradients. Frequencyselective RF pulses in combination with a linear magnetic field gradient areused to select the desired slice of the object to be imaged (slice selection),and to select the specific chemical shift region of the NMR spectrum to beimaged (chemical shift selection). Additional linear magnetic field gradientsare then used to spatially encode the signal in two dimensions from the chosenslice volume by changing the frequency and/or the phase of the signal fromeach volume element (voxel). These two concepts are described more thoroughlybelow.1.5.1. Frequency Selective RF PulsesThe frequency bandwidth of an RF pulse is inversely proportional to thepulse duration, yB1 1/ti, , and is chosen for a given pulse angle by adjustingthe amplitude B1 (i.e. controlling the power from the transmitter) for thedesired duration of the pulse. The longer the pulse, the more frequencyselective it is (17); that is, the smaller the frequency range it can excite.For “hard”, or nonselective, RF pulses, the bandwidths used in the presentwork were between 200 kHz (t 5 ps) and 20 kHz (t = 50 ps). Since the rangeof chemical shifts for protons is approximately 10 ppm (4000 Hz at 9.4 Tesla),all proton resonances can be considered to be excited by these hard pulses.For “soft”, or selective, RF pulses, such as those used in the present workfor chemical shift and slice selection in imaging experiments, the bandwidthswere between approximately 450 Hz (t = 9 ms) and 2000 Hz (t = 1 ms).The pulse shape is the characteristic feature of selective pulses whicheffects their ability to excite a nearly rectangular frequency region, and24thus a slice with sharp edges through a three—dimensional sample. Althoughshaped soft pulses do not really excite nuclei in an exactly linear manner, itis a good approximation to their behaviour (18). With this approximation, thepulse shape and the excitation shape are Fourier pairs (Figure 1.8):f(t) = 1/2n F() exp (ict) d [1. 19]andF() = .9(f(t))=1(t) ezp(-iøt) dt [1.20]This is the basis for choosing pulse shapes for slice selection. Theexcitation should be homogeneous over a narrow frequency range, correspondingto a nearly rectangular slice through the sample perpendicular to the gradientwhich is applied simultaneously with the pulse (see section 1.5.4) . A “slice”is never perfectly rectangular because the shape of the RF pulse isapproximated by a series of very short pulses of varying amplitude which istruncated at a specified point. The more truncated the shape, the lessprecisely the excitation region approximates a rectangular volume. Figure 8shows some common pulse shapes and their Fourier transforms. It is generallyaccepted that the sin(x)/x, or sinc, function truncated after two lobes(“sinc2”) is the best pulse shape for imaging of biological samples (21).25I I II I I—100 0 100Time, is-4 —2 20Time, ms::40.60.20—20 0 20Frequency, kHz1.5.511Ai0.500.4.5IB00.2.5C10.1cI0-GiFrequency, kHz0.3C10.2o.i—10 —5 0 5 10Time, n—2 0 2Frequency, kHzFigure 1.8. Pulse shapes and their Fourier transforms. The Fouriertransforms are approximately the same shape as the excitation (slice). (A) Asquare pulse shape (microseconds in length) yields a broad excitation. (B) Agaussian pulse shape (milliseconds in length) yields a small excitationbandwidth. (C) The “sinc2” pulse shape gives a narrow, almost rectangularexcitation.261.5.2. Linear Magnetic Field GradientsLinear magnetic field gradients are used to destroy the homogeneity ofthe magnetic field across a sample in a precisely defined way, in order toimpose a spatial code upon the signal from each voxel in the part of thesample to be imaged. For example, Figure 1.9 shows the one—dimensional case,where a simple one—pulse NMR spectrum has been acquired in the presence of afield gradient. The spectrum is called a “projection” in this case, since allthe signal intensity at each point in the sample has been projected onto aline parallel to the applied field gradient. At each point along this line,which in Figure 1.9 is the x axis, the precession frequencies c of the nucleihave been spatially encoded:= c + yG;x [1 . 21]since each position x experiences a slightly different fieldB = B0 + G;x [1.22]The first form of imaging, back-projection, used simple projections toacquire an image (16). Projections are taken at a series of sampleorientations, then these “shadows” are projected back into the image planeusing a back-projection reconstruction algorithm. This technique is nowrarely used in clinical NMR imaging.27Figure 1.9. A one-pulse experiment performed on a sample (twotubes of water). The result is a “projection” of a cross sectionof the two tubes when the FID is acquired in the presence of a zfield gradient: all the signal intensity at each point in thesample has been projected onto a line parallel to the appliedgradient.1.5.3. Slice SelectionSlice selection is accomplished by applying a field gradientsimultaneously with a frequency selective RF pulse (Figure 1.10) . If, as inthe figure, the slice is to be taken in a plane perpendicular to the z axis,then a field gradient is applied along the z axis of the sample. If aprojection were taken, it would look like Figure 1.10(C); the signal from thesample is spread out over a frequency range Q as shown. The size of thisrange is controlled by the strength of the field gradient. If a frequencyselective RF pulse is applied in the presence of this gradient, only a0 SampleXh.____________fl_Projectionfrequency28specific subrange of 12 will be excited by the pulse, that is, the subrangecorresponding to the bandwidth of the pulse. This corresponds to excitingonly those nuclei within a specific slice through the sample, perpendicular tothe magnetic field gradient (Figure 1.10(D)). As mentioned before, the widthand sharpness of this slice is determined by the shape, amplitude, andduration of the RF pulse used.Figure 1.10. Schematic representation of the slice selection protocol. Thesample, a tube of water (b), is subjected to a field gradient (a) afterexcitation. If the excitation is nonselective, the result is a projection(c); if selective, the shape of the excited region is similar to the Fouriertransform of the shape of the excitation pulse (d), corresponding to a slicethrough the sample perpendicular to the direction of the gradient.z—Cya b C dG<Ox.291.5.4. Spin—Warp ImagingThe most common form of Fourier imaging is called spin-warp imaging(Figure 1.11). In this sequence, slice selection is accomplished by applyinga 900 frequency selective pulse simultaneously with a magnetic field gradient,shown in the figure as oriented along the z axis. Since the slice has finitethickness, the signals from different points in the slice are dephased by theapplication of the slice selection gradient. These signals are refocused byapplying a second field gradient of the opposite sign after the sliceselection. At the same time, x and y gradients are applied to encode thesignals from each voxel of the sample. A specific frequency encoding gradient(here, the x gradient) is applied for each repetition; the phase encodinggradient (here, the y gradient) is incremented in each repetition. Thisprocess yields a data matrixS(k,k) JJ p(x,y) exp[i2n(kx+ky)] dxdy [1.23)where the variables k and k are the amplitudes (in x and y directions) ofthe “reciprocal space” vectors (19)= (2nY’ y t [1.24]This data matrix is actually an average through the slice (20), which isassumed to be homogeneous in the z direction, i.e. through its thickness.Such an assumption is usually implicit in imaging, and the integral over k2 istherefore left out in Equations 1.23 and 1.25. The two—dimensional Fouriertransform of this data is the spin density function30p(x,y) =f 3 S(k,,k,) exp[-i2rI(kx+ky)] dkdk [1.25]Equation 1.25 does not take into account relaxation effects on the spindensity. This matrix instead represents the results from an ideal(“quantitative”) experiment where pulses and delays are very short, in whichcase p(x,y) contains absolute signal intensities from the voxels comprising31L IFigure 1.11. The spin-warp imaging sequence. A selective 90° RF pulse incombination with a magnetic field gradient along the z axis is used to excitea slice of the sample; the gradient along x is used to frequency encode thedata. A 1800 RF pulse inverts the magnetization and causes an echo at TE.The sequence is repeated at TR, changing the amplitude of the field gradientalong y for each repetition, to phase encode the data.the slice. The repetition time TR and the time to echo TE in this sequencecan be varied to control the contrast of the image. This relationship can beseen in Figure 1.12 as T1, T2, and p or spin density contrasts. Imagecontrast depends on the presence of nuclei (protons) in the sample which havedifferent physical environments: E.g. water in blood vessels is quite mobile,whereas water adsorbed to proteins is not. These factors cause differences inthe T1 and T2 values for protons that are in different molecules and/ordifferent locations and environments within the sample.TimeII___________H LIGII900I•IIRF1800InIIIII 900Sign1I f\II-TE1\ I ( nIII-TRI32Tspin-I densityT1+T2 T2Figure 1.12. The relationship between imagingparameters TE and TR and the image contrast type (T1,T2, or spin-density).Figure 1.13 illustrates how the proton signal from a biological sampledecays due to the T1 and T2 processes. There are ranges of relaxationconstants within the tissues, as well as differences in concentrations ofprotons in different tissue types. Depending on the TE and TR in the imagingsequence, the protons with different relaxation times will contribute more orless signal to the echo. This discrimination on the basis of relaxation timesis called T1 or T2 contrast. In analogy to an NMR spectrum, which isquantitative if the time between excitation and acquisition is short comparedto T2 and the repetition time is at least 5 times as long as T1 (see Section1.5), an image is considered quantitative if TE is less than 20% of theshortest T2 in the sample and the TR is 5 times the longest T1 in the sample.The contrast in a quantitative image is called spin density contrast, whichTRshort.* longshortTElong33means that tissues with various concentrations (densities) of spins will becontrasted.Figure 1.13. An illustration of T1 and T2 contrast discrimination between twotissues with different relaxation times. (A) At a specific value of TR,differences in T1 cause the signal intensity from the two tissues to differ byan amount dependent on the difference their relaxation curves (a and b). (B)Similarly, differences in T2 cause the T2 contrast to vary depending on the TEand the difference in relaxation rates (d and c).A T, contrast regime B T2 contrast regimebEcioEcho intensity—TR----> —TE —>341.5.5. Chemical Shift Selective ImagingThe Chemical Shift Specific Slice Selective (C4S) imaging sequence (1)is a variation of spin—warp imaging. It was originally developed to avoidlipid/water chemical shift artifacts in medical images. This artifact is dueto the protons in lipids and water molecules having different NMR chemicalshifts. When a frequency selective pulse and field gradient are applied tothis sample for slice selection, two spatially distinct slices are chosen, onefor the water and one for the fat (Figure 1.14(A), (B)). Thus signal from twoslices would be acquired simultaneously in the spin—warp sequence, and thespin—warp image would be an overlap of the two slice images with an offsetalong the frequency encoding axis equivalent to the frequency differencebetween the signals in the NMR spectrum.In the C4S imaging sequence (Figure 1.15), slices from both chemicalshifts are excited by the initial 90° RF pulse, but signal from only one ofthese slices is refocussed in the echo. This is achieved by simply reversingthe slice gradient and making the refocussing pulse frequency selective. Thustwo different slices corresponding to the off-resonance frequency are excitedby the 90° and the 180° RF pulses, respectively, and they do not contribute tothe acquired echo (Figure 1.14(B)). The other convenient feature of the C4Ssequence is that images of each chemical shift in the slice can be acquired atthe same time.fatA::Figure 1.14. (A) Illustration of the results of a 90° selective RF excitationon a sample of two tubes containing water and fat, which have different protonchemical shifts (1). The projections (2) in the presence of the gradient Gare offset, so that the selective excitation, centered on frequency a, selectsspatially different slices of the magnetizations from water and fat tubes. (B)The above explanation describes the situation in (2); then (3) illustrates theexcitation due to the selective 1800 RF excitation and reversed G. (4) isthe resultant slice whose magnetization will be refocussed during acquisition.35H20AI234-—a,I1O I fat4 ---— . -a--——I I . a - ‘.z—1GcOABzG>O• G2 ‘0 .-aslice O1=a ZH20 fatT1 2 3 436Time’:90°:1I--<-TR ->Figure 1.15. The Chemical Shift Specific Slice Selective (C4S) imagingsequence. The 1800 RF pulse is made selective, and the slice selectiongradient reversed, in order to avoid refocussing the magnetization in the off—resonance slice (which has a different chemical shift).a bLJLJn rG90° 180°R1Signalr90° 180°.I\ iLTE---)371.6. THESIS GOALSThis thesis introduces the application of an NMR imaging technique,Chemical Shift Specific Slice Selective (C4S) imaging, as a technique formonitoring bulk diffusion and perfusion of solvents into porous media and as amethod of detecting freezing/thawing phenomena, with particular emphasis onthe investigation of cryoprotectant solvent systems. The samples chosen forthe demonstration of this application, rat kidney and liver tissues, are ofparticular importance to organ transplant research, because small animals areused for most studies of immunology and surgical techniques for organtransplants.Chapter 2 describes extensive investigations of the T1 and T2 behaviourof DMSO/water mixtures as a function of temperature, along with a discussionof how these relaxation measurements can be used to determine imagingparameters, and reports a thorough demonstration of the ability of the C4S NMRimaging technique to detect freezing in test samples. Semi—quantitativeagreement between images and the DMSO/water phase diagram is found.Chapter 3 is a detailed analysis of the effects of changing the imagingparameters so that repetition time of the imaging sequence is shorter thanfive times T1. Using relaxation times measured in solutions of DMSO and waterand in rat tissues equilibrated with various concentrations of DMSO inUniversity of Wisconsin (UW) solution (see Chapter 5), it will be shown thatmeasurements of signal intensity from images with TR equal to 1 second can beused to determine concentrations of DMSO in certain situations.In Chapter 4, the use of C4S NMR imaging to visualize bulk diffusion ofcryoprotective solvent in rat liver and kidney tissues at two differenttemperatures. A new technique for measuring diffusion rates and effectivediffusion coefficients is demonstrated, and compared with existing methods.The possibility of detection of freezing within tissues, and specific demandson the imaging hardware, are discussed with respect to the relaxation times of38DMSO in rat tissues.Chapter 5 reports the results of a series of studies of perfusion ofDMSO into rat kidney. Several different preparation methods commonly used inrat kidney transplantation were compared, as well as three methods ofperfusion. The abilities of C4S NMR imaging to detect specific problemsduring perfusion procedures are discussed. Changes in the 31P spectra of ratkidneys stored at several temperatures are discussed with respect to prospectsfor the survival of kidneys subjected to the perfusion protocols developed inthis thesis.Chapter 6 gives a general discussion of the results presented in thethesis, and describes how they may be of use to transplant surgeons andcryobiologists, as well as for the investigation of freezing/thawing phenomenain other porous media. Finally, specific suggestions for further work aremade.39REFERENCES(1) Hochachka, P.W. and G.N. Somero. Biochemical Adaptation. (PrincetonUniversity Press: Princeton, N.J, 1984).(2) Polge, C., A.U. Smith, and A.S. Parkes. 1949. Revival of spermatozoa aftervitrification and dehydration at low temperatures. Nature 164, 666.(3) Keeler, R., J. Swinney, R.M.R. Taylor, and P.R. Uldall. 1966. The problemof renal preservation. Brit. J. Urol., 38, 653—656.(4) Southard, J.H., D.C. Marsh, J.F. McAnulty, and F.O. Belzer. 1987. Oxygen-derived free radical damage in organ preservation: Activity ofsuperoxide dismutase and xanthine oxidase. Surgery 101, 566-570.(5) Collins, G.M., M. Bravo-Shugarman, and P.1. Terasaki. 1969. Kidneypreservation for transportation. Initial perfusion and 30 hours’ icestorage. Lancet, 2, 1219—1222.(6) Pegg, D.E. Ice crystals in tissues and organs. D.E. Pegg and. A.M. Karow,eds. The Biophysics of Organ Preservation. (Plenum Press: NY, 1987).pp. 117—142.(7) Mazur, P. Causes of injury in frozen and thawed cells. 1965. Fed. Proc.24, 175.(8) Meryman, 1-i. T. Cryoprotective agents. 1971. Cryobiology, 8, 173—183.(9) Meryman, H.T. The exceeding of minimum tolerable cell volume in hypertonicsuspension as a cause of freezing injury. The frozen cell. G.E.Woistenholme and M. O’Connor, eds. (Churchill: London, 1970). pp.51-64(10) Lovelock, J. E., and M.W.H. Bishop. 1959. Prevention of freezing damageto living cells by Dimethyl Sulfoxide. Nature, 183, 1394—1395.(11) Lovelock, J. E. The mechanism of the protective action of glycerolagainst haemolysis by freezing and thawing. 1953. Biochem. Biophys. Acta11, 28—36.(12) Karow, A.M. Biological effects of cryoprotectarit perfusion, delivery, andremoval to nonfrozen organs. The Biophysics of Organ Preservation. D.E.Pegg and A.M. Karow, Jr. eds. NATO ASI Series A, vol. 147 (Plenum Press:New York, 1987). pp 25—41.(13) Karow, A.M., M. McDonald, T. Dendle, and R. Rao. Functional preservationof the mammalian kidney. 1986. Transplantation 41, 669—674.40(14) Canet, D., G.C. Levy, and I.R. Peat. Time saving in ‘3C spin—latticerelaxation measurements by inversion recovery. 1975. J. Magn. Res. 18,199—204.(15) Carr, H.Y., and E.M. Purcell. 1954. Effects of diffusion on freeprecession in nuclear magnetic resonance experiments. Physical Rev. 94,630—638.(16) Lauterbur, P.c. 1973. Image formation by induced local interactions:examples employing nuclear magnetic resonance. Nature (London) 242, 190—191.(17) Freeman, R. 1987. A Handbook of Nuclear Magnetic Resonance. LongmanScientific and Technical, Essex, England, pg.208.(18) Bailes, D.R. and D.J. Bryant. 1984. NMR Imaging. Contemp. Phys. 25, 441-475.(19) Mansfield, P. and P.K. Grannell. 1975. “Diffraction” and microscopy insolids and liquids by NMR. Phys. Rev. 12, 3618—3634.(20) Callaghan, P.T. Principles of Nuclear Magnetic Resonance Microscopy.(Clarendon Press: Oxford, 1991). pg 123.(21) Ibid., 107.(22) Volk, A., B. Tiffon, J. Mispelter, and J. Lhoste. 1987. chemical shiftspecific slice selection. A new method for chemical shift imaging athigh magnetic field. J. Magn. Reson. 71, 168-172.41CHAPTER 2INVESTIGATION OF THE FREEZING AD THAWINGOF LIQUIDS IN HETEROGENEOUS SYSTEMS2.1. INTRODUCTIONThis chapter will describe how NMR imaging techniques can be used as aprobe of the freezing and thawing phenomena of liquids in heterogeneouschemical and biological systems. The mixed solvent system, DMSO/water, isused to demonstrate that the Chemical Shift Specific Slice Selective (C4S)imaging technique (1) can monitor the spatial distributions of theconcentrations and mobilities of the two individual solvents in test samples.Relaxation parameters were measured in order to choose values of imagingparameters TE and TR that ensure quantitative reliability of the images.Images of a test sample at various temperatures demonstrate a semi—quantitative relationship between image intensity and the DMSO/water phasediagram.2.1.1. Effects of Temperature on Relaxation RatesAs described in section 1.4.3, the relaxation parameters associated withnuclei whose signals are being measured must be known if NMR measurements areto be quantitative. In a mixed solvent system like DMSO/water, the values ofT1 and T2 for protons are affected by several variables: The concentrations of42solvents and salts, physical environment, and temperature. In order toreliably determine relative numbers of hydrogen nuclei from each of thesignals in this system, one must know the values of the relaxation parametersas a function of these experimental variables.In the case of the solvent system DMSO/water, temperature has adramatic effect on the values of the relaxation parameters. This is due tothe fact that the main relaxation mechanism for hydrogen nuclei is via thedipole—dipole interactions. These interactions cause relaxation of excitedspins to change with temperature in a predictable way.In order for a nuclear spin to undergo a transition from an excitedstate to a lower-energy state, an emission of RF energy must be stimulated bya local magnetic field which is fluctuating at a particular frequency. Thisaffect is quite short-range, since dipole—dipole interactions occur over veryshort distances, since they vary as r3, where r is the distance between thesource of the fluctuating field and the nucleus. In dipole-dipoleinteractions, the source of most fluctuating fields is the magnetic moments ofother hydrogen nuclei. These local fields fluctuate because of the movementof molecules with respect to one another. The rate of rotation of moleculesin a sample is described by the correlation time, ;, which is the averageamount of time that a molecule remains in one motional state (the time betweenmovements, or “jumps”).43Molecular motion, and thus the correlation time, is dependent on temperature,and this is the source of the temperature dependence of relaxation ratesmentioned above.The relationship between the correlation time and T1 can be described interms of the spectral density function J(). The spectral density functiondescribes the frequency distribution of the probability of interaction betweennuclear dipoles. It is a complex function, having terms of the formFigure 2.1. Relationship between correlation time and spectral density. Atlow temperatures, ; is long, and most of the spectral density is at lowfrequencies (i.e. molecular tumbling is slow). At high temperatures, ; isshort, and there are slow, medium, and fast components to the molecularmotion. The area under the curve is constant for all ;.j> (nc) = A‘ 1 +J(co); longV0‘c shortCt) = I /t[2.1]44where A have scalar values (2). The spectral density function can begenerally described as follows: molecular tumbling motions in a sample havedifferent frequency (Fourier) components. If the average tumbling rate isvery fast (corresponding to short ta), then there are some intermediate, somefast, and some very fast components to the motion, so the frequencydistribution of J() is quite broad. On the other hand, if average tumblingis slow (corresponding to long ;, there will be few fast, or high frequency,components to the motion (3). Figure 2.1 shows this relationship betweencorrelation time and spectral density.The frequencies important for transitions leading to longitudinal (T1)relaxation are the Larmor frequency and twice the Larmor frequency, c and 2+J2(2) [2.2](4). Thus the more Fourier components of motion in the sample at and 2,the larger the relaxation rate, and the smaller T1 will be (2). Figure 2.2shows the expected relationship between t and the relaxation rate T1.The relationship between t, and T2 can be similarly expressed in terms ofthe spectral density function, but the expression contains a term which isindependent of frequency:J°(O) + JW() +J2(2c) [2.3]This “zero—frequency” term, which contains j(O), the zero-order spectraldensity term, predominates at low temperatures. Figure 2.2 shows T2relaxation time at a fixed Larmor frequency as a function of correlation time.452.1.2. Effects of Temperature on the NI’4R SpectrumFigure 2.2. Dependence oftime t.the relaxation times T1 and T2 on the correlationFigure 2.2 compares the dependencies of T1 and T2 on correlation time.Both relaxation processes have effects on the NMR spectrum. At hightemperatures, where correlation times are on the order of 1012 to lO’seconds, and i;’>> cL the T1 and T2 relaxation rates are the same. Thissituation occurs for small molecules like water and DMSO in the liquid state,and is called the “extreme narrowing” regime, because at these correlationtimes the dipolar interactions between nuclei in the fast—tumbling moleculesare weak, which allows transverse coherence of a signal to exist for arelatively long time, and thus T2 is longer (Figure 2.3). The linewidth Av112T,t= I IoG)C0.1-’CUCU0)0 T2-h--10.2 10° I 02 log (o0t)equals 1/nT2.46ZI ZIM FT‘Figure 2.3. The coherence of transverse magnetization is affected by changesin the relative phases of individual magnetic moments due to exchange ofenergy via T2 mechanisms. Faster exchange gives a shorter T2 and a broaderdistribution of frequency components about cot,, i.e. a broader linewidth.When liquids or solutions become viscous, and the correlation time is ofthe same order as the Larmor frequency (1/ta z ), the T1 and T2 relaxationtimes begin to differ, but the linewidth continues to increase as T2decreases. When the viscosity has increased to the point where the solutionbecomes a solid, the relationship between correlation time and T2 becomescompletely dominated by the zero-frequency term j(O) and1 [2.4]2(5). At this point, -r’<< c, and the linewidths become extremely large;corresponding signal intensities are very small.472.2. NMR AS A TECHNIQUE FOR MONITORINGFREEZ ING/THAWING PHENOMENA2.2.1. Monitoring the Freezing Process with NMR SpectroscopyThe ability of NMR spectroscopy to differentiate between molecules indifferent physical states, i.e. liquid or solid, on the basis of linewidth,allows monitoring of freezing and thawing processes in the following manner:When the temperature has decreased to the point where the solution becomes asolid, the linewidth increases so much that the signal can no longer bedetected. In this way, freezing effectively removes signal contributions fromthe nuclei in solids in conventional NMR experiments (Figure 2.4).When a biological sample is slowly frozen, the signal intensity isreduced as the signals from frozen water become ‘invisible’. However, it isimpossible to say out of which compartment of the sample (blood, intra— orextracellular water, etc.) the water is freezing, and in what part of thesample that compartment is located. This is due to the fact that the NMRspectrum of a sample is the total signal from all the nuclei in the samplewithin the RF coil. For example, in the case of proton NMR spectroscopy ofbiological samples, the NMR spectrum shows a single signal from water whichrepresents all water in all compartments of the sample. Although specialpulse sequences for localized spectroscopy exist, they are difficult to useand cannot give spectra from very small volumes efficiently. Thus “volumeaveraging” is a general difficulty in using NMR spectroscopy to monitorbiological processes in intact samples, and will be mentioned later in thecontext of NMR spectroscopy of diffusion processes in rat kidney.48AWATERI I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I20 10 5 0 —5PPMBICE- •1W&trt J. drf.xs?tLtjJW i I1I mlIL3—Mr1-tI- .JnjrI I I I I I I I I I I I I I I I I I I I I I I I I I I I I I20 15 10 5 0 —5PPMFigure 2.4. NMR spectra of water in the liquid state (A) and solid state (B).2.2.2. Monitoring the Freezing Process with NMR ImagingThe use of NMR in the monitoring of freezing processes can be extendedby acquiring additional information about the NMR signal. NMR imaging yieldsspatial information about the NMR signal (6), so that it is possible tomonitor the freezing process in different parts of a biological sample, and insome cases, in different compartments of the sample (7).49Figure 2.5. NMR images of a transverse slice taken through a bundle ofcapillary tubes (black circles) filled with water and surrounded by water,contained within a thin—walled outer glass tube. The temperature decreasesfrom ambient (A) to -10°C (B—F).Characterization of the freezing process by NMR imaging is illustratedin Figure 2.5(A)-(F), which show images of a transverse slice taken through abundle of capillary tubes (black circles) filled with water and surrounded bywater, all contained within a thin—walled outer glass tube. A short time toecho (STE) spin-echo sequence was used with T = 6 ms. Low signal intensityis expected from solid materials, and T, is also very short at low50temperatures (T2<<T1, Figure 2.2), so the signal will mostly decay before echoformation. The temperature was gradually changed from ambient (Figure 2.5(A))to —10°C (Figures 2.5(B)-(F)) using a stream of cooled nitrogen gas. Thefreezing of the water as a function of temperature is clearly shown by thedisappearance of image intensity progressing from the lower left portion ofthe image through the series.The same experimental protocol as outlined above can be applied tobiological systems. Figure 2.6 shows a series of STE images, TE = 6 ms,slice-selected in the xy plane, taken during the freezing of a marine musselsurrounded by water. The shell itself can be seen as the dark outline in thefigure and the progress of the freezing is, as before, reflected in theencroachment of the dark area in the image. The water within the body of theshell remains fluid after the surrounding water has frozen. Similarmeasurements can be made on more complex systems, and the phase behaviour ofthe water within the interior of the sample monitored.In order for NMR imaging to be a reliable quantitative measure of therelative amount of liquid remaining in a sample during freezing, the loss ofsignal intensity must be strictly due to the phase change from liquid tosolid. The NMR imaging techniques used in this thesis are not based on thesimple one—pulse/acquire method used in 1D NMR spectroscopy. Instead, theseimaging techniques involve spin echoes, and also use frequency selectiverather than nonselective pulses. Both of these differences must be taken intoaccount if an image is to be used to quantitatively monitor the freezingprocess.The general effect of the use of selective pulses and spin echos is thatthe time between excitation of the nuclei and acquisition of the signal (TE)is relatively long. During this time, relaxation can occur. Signal intensitycan be lost due to both T1 and T2 relaxation, but T2 losses are of moreconcern. In the DMSO/water system, T1 and T2 are both very long at roomtemperature compared with the short TE values possible with the equipment51used. However, at the low temperatures used to preserve tissues, T2 decreasesto become of the same order of magnitude as TE (see next section) . If TE isnot much less than T2, it is possible that a significant part of the signalcould be lost due to relaxation before it can be acquired, thus destroying thequantitative nature of the image.In addition to the general effects mentioned above, a specific problemwith the use of selective pulses must be addressed. This potential problem iscaused by the line broadening associated with decreases in T2 as theFigure 2.6. NMR images of a transverse slice taken through a marine musselduring freezing. The shell is seen as a dark outline. The water in thetissues of the mussel remains fluid after the surrounding water has frozen.52temperature decreases. As explained in Chapter 1, frequency selective pulseshave specific bandwidths depending on their duration. The bandwidth of aselective pulse used for chemical shift selective imaging must be small enoughto avoid acquisition of signal from another line in the NMR spectrum, yetlarger than the full linewidth of the signal of interest for maximum signalintensity. Care must be taken in setting the selective pulse durations in anexperiment where the temperature decreases, so that any line broadeningeffects do not lead to loss of selectivity and/or loss of quantitativereliability. Obviously a thorough knowledge of the temperature dependence ofT1 and T2 in the system to be studied is required.2.3. T1 AND T2 IN THE DMSO/WATER SYSTEMIn the previous section, the temperature dependence of relaxationparameters in the DMSO/water system was discussed. Another variable thataffects the values of T1 and T2, the concentration of solutes, was mentioned.In order to choose imaging parameters appropriate to the entire temperaturerange of a freezing experiment, the dependence of T1 and T2 on all thesevariables must be investigated. This section will describe measurements of T1and T2 of DMSO/H20/NaC1 mixtures, with concentrations ranging from 0% to 100%saline (water with a “physiological” concentration of NaC1, 0.15 M), over arange of temperatures.532.3.1. Experimental2.3.1.1. Measurement of Relaxation ParametersExperiments measuring the relaxation parameters of DMSO/H20/NaC1solutions were carried out using the Inversion Recovery (IR) sequence for T1and the Carr—Purcell-Meiboom-Gill (CPMG) sequence for T2. A Bruker WH 400 MHzspectrometer was used for measurements. When T2 measurements were repeated ona Bruker MSL spectrometer, a Carr—Purcell (CP) sequence was used; the IRsequence was used for T1 measurements. These pulse sequences were describedbriefly in Chapter 1.Twelve solutions of DMSO (99.9% minimum assay, BDH Inc., Toronto) anddeionized water were made up, with DMSO concentrations ranging from 0% to 100%in increments of 10%. Sodium chloride (0.15 M) was added to the water tosimulate conditions likely to be encountered in the application of thesemethods to biological samples. The solutions were well mixed and allowed toequilibrate to room temperature. An aliquot of each was then placed in aclean 5 mm NMR tube and the tubes sealed.Since DMSO/H20/NaC1 mixtures freeze at or above —73°C, imagingexperiments would thus be carried out in that temperature range andtemperatures from —63°C to 20°C were chosen for relaxation measurements. Thetemperature of the sample in he probe was controlled by passing coolednitrogen gas through a dewar surrounding the sample, and countered by athermostatically-controlled heater inserted in the dewar. The true sampletemperature was determined by stabilizing the temperature at a selected valueaccording to the variable temperature (VT) control unit. A 5 mm NMR tubecontaining methanol was then inserted into the magnet and allowed toequilibrate. The difference in the chemical shifts of the two peaks in theproton NMR spectrum of methanol was then measured, and a Varian referencechart relating methanol peak shift differences to temperature used to54determine the real temperature within the sample tube. The VT controller wasadjusted until the real temperature was as desired, always giving the methanolsample time to equilibrate after changing the controller settings.This temperature calibration procedure was carried out once using theBruker AI’4 spectrometer, and a curve drawn relating true temperature within thesample tube to the settings on the VT control unit. This curve was used toset the VT control unit for all subsequent experiments on this spectrometer.When the measurements were repeated on the Bruker MSL spectrometer, the actualtemperature was measured each time the VT. control unit’s settings werechanged. This ensured that the temperature was correct at the time ofmeasurement.2.3.1.2. Imaging ProtocolImages of water protons in Figures 2.5 and 2.6 were obtained using theShort Time to Echo (STE) imaging protocol, which is basically a simple spin—warp imaging sequence (see Figure 1.11). All images of DMSO/H20/NaC1solutions were obtained using the C4S sequence (see Figure 1.14). Imagingparameters were chosen according to two criteria: the limitations of theavailable equipment, and the requirements of the sample.The imaging probe was tested to determine the lengths of the selectiveand nonselective 900 and 180° RF pulses. Several pulse shapes were tested.Gradient strengths were measured using a sample of water, with a trace ofCuSO4 to reduce the T1, in a 10 mm NMR tube. Actual gradient strengths weredetermined for a range of settings of the DISMSL software parametersdesignated “M4”, Ml”, and “M15”, which control the x, y, and z gradientsrespectively.Both gradient strength and pulse length are affected by the mode of55connecting the hardware components. Longer cables with less shielding allowmore transmitter power to be lost before it reaches the RF coil. This is whymost of the experiments in this thesis were carried out with a 9 ms softpulse, even though this meant that the shortest possible TE for the C4Simaging protocol was 16 ms. When the cables were replaced with shieldedcables, the shortest possible soft pulse was still 7 ms long. The addition ofanother amplifier to this system might improve the images by shortening thepulses and thus the TE.The parameters for the images in Figures 2.5 and 2.6 were set bydetermining the 900 and 180° RF pulse parameters for the shortest pulse shapedas the sinc2 function, which is accepted as the best pulse shape forbiological samples (8). Other parameters were set simply to maximize thesignal—to—noise ratio.Parameters for the C4S NMR images of the test sample in Figure 2.8(A)are given in Appendix A. For C4S NMR imaging, the length of the shaped pulsewas further constrained by the requirement that its bandwidth be at least aslarge as the base width of the proton resonance of DMSO or water (whichever isbroader) in the NMR spectrum. The measurements of relaxation parameters inthe DMSO/H2O/NaCl solutions were used to select appropriate imaging parametersfor quantitative imaging of the test sample. At each temperature, the TR wasset to five times the longest T1.2.3.2. Results of T1 and T2 MeasurementsT1 and T2 relaxation times were measured by the 180-t—90 (IR) and 90-t-180 (CPMG) techniques described in Section 1.4.3, Figures 1.6 and 1.7.Results of the experiments measuring relaxation times are shown inTables 2.1 and 2.2.56A% Temperature, °CDMSO T 1—63 —46 —26—10 7.5 j 200 F F F 1.50 2.39 2.8210 (0.31) (0.17) 0.250 1.14 1.84 2.7520 (0.34) (0.18) 0.379 0.899 1.53 2.0630 (0.36) (0.16> 0.312 0.709 1.23 1.5340 (0.35) (0.17) 0.244 0.583 0.960 1.3150 0.370 0.175 0.189 0.396 0.708 0.87660 0.390 0.196 0.269 0.347 0.606 0.82170 0.384 0.207 0.169 0.313 0.543 0.71080 0.378 0.209 0.172 0.308 0.530 0.68690 F (0.25) 0.301 0.391 0.592 0.842100 ————— —B% Temperature, CDMSO—63 —46 —26 —10 7.5 200— — ————10 (0.29) (0.26) 0.428 1.79 2.62 3.0220 (0.31) (0.28) 0.412 1.61 2.39 2.7630 (0.32) (0.31) 0.673 1.32 2.09 2.3640 (0.31) (0.31) 0.609 1.17 1.80 2.1550 0.318 0.299 0.533 1.01 1.61 1.8360 0.330 0.305 0.729 0.961 1.46 1.9570 0.321 0.312 0.534 0.719 1.83 1.8080 0.313 0.308 0.523 0.962 1.51 1.8190 F (0.27) 0.400 1.24 1.84 2.30100 F F F F F 2.48Table 2.1. T1 values (s) in DMSO/H20/NaC1 solutions at various temperatures.(A) T1 values of the proton resonance of water, (B) T1 values of the protonresonance of DMSO. An F denotes that the linewidth of the NMR signal was toobroad (>1000 Hz) and/or that the T1 could not be confidently measured from thedata. Parentheses around data indicate that the measurement probably haserror >20%, due to linewidth problems or a poor signal—to—noise ratio.57A% Temperature, CDMS0—63 j —46 —26 ] —10 J 7.5 200 F F F 947 936 93210 (5.3) (27) 271 817 1000 105020 (8.0) (20) 238 633 941 78030 (8.6) (40) 265 607 823 70740 (12) (36) 226 498 739 73850 10.4 28.0 309 389 646 64360 8.2 25 182 326 — 60370 8.5 49 128 309 510 50080 9.6 102 127 306 526 55590 F (23) 103 383 589 632100 — — — — — —B% Temperature, °CDMSO—63 —46 —26 —10 7.5 200 — — — — — —10 (11) (47) 90 985 1280 136020 (15) (36) 109 961 1360 98930 (17) (74> 555 1110 1250 99840 (24) (43) 640 1010 1140 110050 22.8 66.0 789 860 1260 111060 18.6 58.0 510 765 970 111070 18.0 116 349 763 1049 115080 21.0 230 362 806 1100 115090 F (37) 100 938 1180 1020100 F F F F F 1200Table 2.2. T2 values (ms) in DMSO/H20/NaC1 solutions, at varioustemperatures. (A) T2 of the proton resonance of water, (B) T2 of the protonresonance of DMSO. An F indicates that the linewidth of the NMR signal wastoo large, and/or that the T2 could not be confidently determined from thedata. Parentheses around data indicate that the error is probably >20%, dueto linewidth problems or a poor signal-to-noise ratio.582.3.3. Discussion of the Proton Relaxation Timesin DMSO/H20/NaC1 SolutionsFigure 2.7 shows that relaxation parameters T1 and T2 for theDMSO/H20/NaC1 solutions all decrease smoothly with decreasing temperature.The following observations can be made from this data:T of water. Figure 2.7(B). T1 of water protons in DMSO/H20/NaC1 solutions islargest in the solution containing no DMSQ; at 20°C, T1 = 2.82 s. Thisdecreases to values on the order of 200 ms at —26°C and below. Valuesdecrease with increasing DMSO concentration at constant temperature,until the DMSO concentration reaches 80%. At concentrations higherthan this, the NaC1 in the saline solution is not as soluble in theDMSO/H20/NaCl mixture, so [NaC1] < 0.15 M. For this reason, the T1value increases slightly.Since the value of T1 is greater than 5 times the TR for theimaging protocols used in this thesis, the T1 of water protons will notaffect the quantitative nature of water images above —63°C, the lowesttemperature used in the imaging experiments. At the highesttemperature, however, the T1 of water will require a TR of approximately14 seconds for a quantitative image. Thus the time to acquire an imagewith TR = ST1 is very long, e.g. for a quadrature—cycled 128 X 128 pixelimage, 2 hours. If the length of the freezing process being monitoredis not on the order of many hours, most of the process will occur duringa single image acquired under these conditions.T1 of DMSO. Figure 2.7(A). The T1 values for DMSO protons show very similartrends to those of water. The longest T1 value, 3.02 s, is at 20°C in a10% DMSO solution. Values decrease to approximately 300 millisecondsbelow -26°C. The effect seen at high concentrations of DMSO, where the59NaCl in the saline is no longer completely soluble in the DMSO/I-120/NaClmixtures, is more dramatic than for water: at DMSO concentrations above80%, the T1 of DMSO increases quickly to nearly the value of T1 in the10% DMSO solution.Since the T1 values of DMSO in all solutions are greater than 5 timesthe TR values used in the imaging protocols, decreases in T1 of DMSOwill not affect the quantitative nature of the images. At the highesttemperature, however, the T1 of DMSO will require a TR of approximately15 s for a quantitative image. Thus the time to acquire an image isvery long. Again, if the length of the freezing process being monitoredis not on the order of hours, most of the process will occur during onlyone image acquired under conditions of TR = 5T1.T2 of water. Figure 2.7(D). The general trends seen in the T2 values of theprotons of water in DMSO/H20/NaC1 solutions are very similar to thetrends in the T1 values. The longest T2 value, 1 s, is in the 0%—10%DMSO solutions at 20°C. Values decrease to approximately 200 ms at—26°C, between 20 and 100 ms at —46°C, and are on the order of 10 ms at—63°C. Insolubility of NaC1 in the mixtures also increases the T2values slightly when concentrations of DMSO are at or above 80%.Since the TE values for the imaging protocols used in this thesis are15-20 ms, images of DMSO/saline solutions below —26°C will not bestrictly quantitative.T of DMSO. Figure 2.7(C). The trends seen in the T2 values of the DMSOprotons in DMSO/H0/NaCl solutions are again very similar to those seenin the T1 values. The longest T2 value is 1.3 s, for the 0% DMSOsolution at 20°C. The values decreased to 100—700 ms at —26°C, 36-230ms at —46°C, and 10—20 ins at —63°C. Increases in T2 at constanttemperature are again seen for concentrations of DMSO above 80%.60Given TE values of 15-20 ms in an imaging protocol, the T2 values below-26°C are too short for the images to be strictly quantitative.The T1 and T2 values for both DMSO and water in these mixtures followtrends with respect to one another. Both relaxation parameters for both DMSOand water protons decrease with decreasing temperature. Both parameters showincreases at constant temperatures in mixtures with DMSO concentrationsgreater than 80%. Both T1 and T2 of DMSO in all solutions are greater thanthe values of these parameters for water in those solutions.At the 1owst temperatures, relaxation parameters vary much less betweensolutions with different initial DMSO concentrations than at highertemperatures. This effect is due to the unfrozen mobile liquids achievingalmost identical concentrations of DMSO and water near the eutectictemperature, as either water or DMSO freezes out of solution.61Figure 2.7. Surface plots of T1 and T2 in DMSO/I-120/NaC solutions as a functionof the temperature and the DMSO concentration of the solution. A. T1 of theproton resonance of DMSO (s). B. T1 of the proton resonance of water (s). C.T2 of the proton resonance of DMSO (s). D. T2 of the proton resonance ofA B20rJ.6343U)I-.U,90%U)IU,3211.510.501.510.590%20-26.63Lire90%20$tUreowater (s).622.4. IMAGING EXPERIMENTS: RELAXATION RATESAND CHOICE OF PARAMETERSIn Section 2.2.2, limitations on the choice of imaging parameters withrespect to soft pulse bandwidth, pulse power, linewidth, and relaxation rateswere discussed in general. Detailed knowledge of the behaviour of T1 and T2at various temperatures now allows discussion of these limitations in terms ofthe imaging of a model DMSO/H2O/NaC1 system.A test sample (cross section shown in Figure 2.8(A)), made of a 10 mmNMR tube containing seven sealed capillaries filled with mixtures ofDMSO/H20/NaC1 (0, 20, 40, 60, 80, and 100% water with 0.15 M NaCl), andsurrounded by 100% DMSO, was used to determine whether the imaging protocolwas quantitative at higher temperatures as expected from relaxationmeasurements, and what effect the decrease of temperature would have on theimages. The test sample was imaged at several temperatures: 20, 0, —20, —40,and -60°C.Images of the test sample at room temperature should be quantitativelyreliable, according to the values of the relaxation times described in theprevious section. Figure 2.8(B) shows a C4S image of the test sample at roomtemperature. In this case the image is of the water in each capillary tubeonly, and for display purposes the image is shown as a negative (i.e. so thathigher signal intensity is indicated by darker color). Figure 2.11(D) is astacked plot of the (positive) image, which shows clearly how signal intensityvaries across the image. Figure 2.11(C) shows rows of pixel intensities takenthrough the image, and the areas corresponding to the total signal from eachcapillary tube relative to the average of the signal from the two capillariescontaining 100% water. (The bumpy artifacts are due to rotation of the imageafter its generation so that it would be in the same orientation as thediagram in Figure 2.11(A).) These measurements show that the imageintensities are quantitative with an error of 3—5%. Clearly, if the imageparameters are set so that TE is much shorter than all the T2 values in thesample, and TR is five times longer than the longest T1, the images arequantitatively reliable.6364CBFigure 2.8. A. Diagram of a test sample, a 10 mm NMR tube filled with DMSO andcontaining capillary tubes filled with DMSO/H,O/NaC1 solutions (% watershown) . The capillary tubes are surrounded by pure DMSO. B. Quantitativeimage of water protons in the test sample, shown as a negative (i.e. highersignal intensity is indicated by darker color) . C. Rows of pixel intensitiesthrough the image. Numbers indicate areas relative to the average areascorresponding to the two capillaries containing 100% water. D. Stacked plotrepresentation of the image.6511CA....a._J —BCa aJLb bJ Lr— -C D..,÷. 4’a a4- 4-b b4- 4-C4-Figure 2.9. Quantitative NMR images of the water protons in the DMSO/H20/NaC1test sample in Figure 2.8(A) at various temperatures. A. 0°C. B. —20°C. C.—40°C. D. —60°C. Negative images are shown above, with rows of pixelintensities through the images shown below (a, b, C). Artifacts in the rowprojections are due to rotation of the images.66Figure 2.10. Quantitative NMR images of the DMSO protons in the DMSO/H20/NaC1test sample shown in Figure 2.8(A) at various temperatures. A. 0°C. B. —20°C. C. -40°C. 0. —60°C. Negative images are shown above, and rows of pixelintensities shown below (a, b, C). Artifacts in rows are due to rotation ofimages. The pure DMSO surrounding the capillary tubes in the test sample isnot observed, since it freezes at 18.3°C.B••.÷2aDabAa•..caCa•..<aba67Figures 2.9 and 2.10 show a series of variable temperature ‘H NMR imagesof the same test sample as in Figure 2.8(A), with the percentage of water ineach capillary as shown in that diagram. The slice thickness and resolutionare approximately the same as for Figures 2.5 and 2.6. The images in Figures2.9 and 2.10 are taken at successively lower temperatures starting at 0°C atthe left in each figure, and progressing down in 20°C increments in eachsuccessive image. Figure 2.9 shows only the mobile water distribution,whereas Figure 2.10 presents only the mobile DMSO distribution. Grey scaleintensities are quantitative measurements of the respective solventsselectively imaged. Intensities are not comparable between images since thegain was increased as the temperature decreased; due to limitations in thesignal amplitudes which the analog-to--digital converter can accommodate, thelarge variation in signal intensity between images at high and lowtemperatures, and the Boltzman effect on S/N with decreasing temperature, itis not possible to set the RF gain to one value for all images. However, asdemonstrated in Figure 2.8, the grey scale does indicate the relative quantityof the selected nucleus (and the molecule that contains it) within each image.As can be seen, all the mixtures have lower freezing points than the purecomponents. The pure DMSO surrounding the capillary tubes is not visible,since it freezes at 18.3°C.68The data can be discussed in terms of a partial phase diagram of theDM30/water system (9,10) as shown in Figure 2.11, although the present systemdoes not correspond exactly, due to the addition of NaC1 to the solutions tosimulate physiological conditions. Close examination of the images in Figures2.9 and 2.10 shows that the freezing of the components of the solutionsFigure 2.11. Phase diagram for the DMSO/water system, modified from Rasmussenand MacKenzie, Nature 220, 1315—1317 (1968). Area A, liquid; area B, liquidplus solid water; area C, liquid plus solid DMSO; area D (shaded), a complexset of phases not accessed in the present set of experiments.200-20-40-60C.)0Dci)E0 20 40 60 80% DMSO10069(indicated by loss of intensity in the images) is qualitatively as expectedfrom the phase diagram. For example, the top left capillary tube in eachimage contains 80% water by volume. The middle lefthand tube is 60% water.Following the intensities of these two mixtures in the water images (Figure2.9) as the temperature falls, the 80% mixture loses signal intensity morequickly than the 60% mixture. This is because there is initially more waterto freeze out of the 80% mixture, as shown by the dotted line on the phasediagram. At —60°C the intensities are approximately equal as expected.The correspondence of image intensity to the phase composition of thesolution (from the phase diagram) can also be described as follows: Figure2.8(B) shows the distribution of water in the test sample at 20°C, and waterdistributions at other temperatures are shown in Figures 2.9(A)—(D). Thefurther the room temperature mixture is from the eutectic mixture of 60% DMSO,the greater the proportion of that mixture which will freeze out before theremaining liquid is at the eutectic concentration. Thus, at —60°C, theremaining liquid in all tubes is approximately 60% DMSO. The closer theoriginal mixture was to that concentration, the more relative signal it willhave at the lowest temperature. In Figure 2.9(D) (—60°C) as expected the 60%DMSO mixture has the largest intensity, followed by the 80% mixture, then the40% mixture, and lastly the 20% mixture.These two sets of images demonstrate the unique ability of NMR ChemicalShift Specific Slice Selective (C4S) imaging to quantitatively monitorfreezing of the individual components of mixed solvent systems.702.5. IMPLICATIONS FOR MONITORINGFREEZING/THAWING PHENOMENAIn the above series of experiments, the potential of the NMR microscopicimaging technique as a probe of freezing/thawing phenomena within solidmatrices is well demonstrated. The unique potential of the C4S pulse sequencein the investigation of the phase behaviour of mixed solvent systems has beenshown, and the quantitative reliability of the results have been demonstratedunder the conditions used for the imaging experiments. As will be discussedin later chapters, these techniques can be further developed to study thediffusion and perfusion of cryoprotectants in viable tissue samples, andpotentially the mode of action of cryoprotecting solvents in situ.71REFERENCES(1) Volk, A., B. Tiffon, J. Mispelter, and J. Lhoste. 1987. Chemical shift—specific slice selection. A new method for chemical shift imaging athigh magnetic field. J. Magn. Reson. 71, 168-174.(2) Ferrar, T.C., and E.D. Becker. Pulse and Fourier Transform NMR. (AcademicPress, Inc. NY: NY 1991). PP. 50—51.(3) Fukushima, E., and S.B.W. Roeder. Experimental Pulse NMR: A Nuts and BoltsApproach. (Addison-Wesley Publ. Co.: Reading, Massechusetts 1981) pp.131—133.(4) Callaghan, P.T. Principles of Nuclear Magnetic Resonance Microscopy.(Oxford University Press: NY, 1991). pp. 61.(5) Ibid., 64.(6) Lauterbur, P.C. 1973. Image formation by induced local interactions:examples employing nuclear magnetic resonance. Nature 242, 190—191.(7) Aguayo, J.B., S.J. Blackband, J. Schoemiger, M. Mattingly, and M.Hinterman. 1986. Nuclear magnetic resonance imaging of a single cell.Nature 322, 190—191.(8) Callaghan, P.T. Principles of Nuclear Magnetic Resonance Microscopy.(Oxford University Press: NY, 1991). pp. 107.(9) Rasmussen, D.H., and A.P. MacKenzie. 1968. Phase diagram for the systemwater—dimethyl sufoxide. Nature 220, 1315—1317.(10) Cocks, F.H. and W.E. Brower. 1974. Phase diagram relationships incryobiology. Cryob. 11, 340-358.72CHAPTER 3PROTON N RELAXATION TINES OF WATER AND DMSO IN TISSUE SAMPLESAND THEIR EFFECT ON IMAGE INTENSITIES3.1. INTRODUCTION3.1.1. The Importance of Contrast in NMR ImagingNonquantitative imaging protocols must often be used in situations inwhich the process being monitored occurs too quickly for the long delays ofthe quantitative imaging protocol to be used. Such situations arose duringthe study of CPS diffusion and perfusion using the C4S NMR imaging technique(see Chapters 4 and 5). In these cases, it will be critical to understand howthe imaging parameters will affect the image contrast. The C4S NMR imagingtechnique and all other spin—warp imaging techniques are based upon the spin-echo technique invented by Hahn (1). The experimental parameters control theimage contrast due to T1, T2, and spin density differences in the sample asdescribed in Chapter 2. Equation 3.1 shows the relationship between thesignal intensity of a particular resonance, the imaging parameters, and therelaxation times T1 and T2 (2) for any spin-warp type imaging sequence.S(TE,TR) = D(’H) [1 — 2e TR-TEJ2)/T1 + eT/Tl] eTtT2 [3.1)where S(TE,TR) is the calculated signal intensity, D(’H) is the spin densityof the protons contributing to that resonance, TE and TR are the time to echoand repetition time of the imaging sequence, respectively, and Ti and T273represent the spin—lattice and the spin—spin relaxation times for theresonance signal.In different imaging protocols in which the parameters are setdeliberately to achieve T1, T2, or spin density contrast, this equationsimplifies to the following three equations:S(TR) = D(1R) [1_2eTT7/2)T1 + eTITh] [3.2)S(TE =D(’R) e11’2 [3.3]S = D(1R) [3.4)D(’H) is known as the “spin density factor”, [l_eTT2Tl+eT’Tl] is the “T1factor”, and eTTz is the “P2 factor”. For Equation 3.2 to be applicable, TEis set to less than 20% of the smallest T2 in the sample to achieve only T1contrast by varying TR. To achieve T2 contrast by varying TE (Equation 3.3),TR is set to 5 times the longest T1 in the sample. To avoid T1 and T2contrast and thus achieve contrast based only on spin density (called“quantitative imaging” in Chapter 2), TE must be less than 20% of the shortestT2, and TR greater than S times the longest T1 in the sample (Equation 3.4).In practice, NMR images rarely have only T1, T2, or spin—density contrast, butdifferent kinds of contrast can be minimized (or maximized) by varying theimaging parameters knowing the T1 and T2 values in the sample.743.1.2. T1 and T2 Contrast in NonquantitativeImaging ProtocolsIn the course of this research, it was determined that quantitativeimaging was not feasible for the study of DMSO diffusion into tissue samples.The relaxation times in the mixture would require much longer imaging timesthan the time course of the diffusion allowed; thus to observe diffusion, theTR of the imaging sequence would have to be set to a value much smaller thanthat required by quantitative imaging. This constraint on TR is expected toproduce T1 contrast in the images. T contrast can either increase ordecrease the amount of total contrast in the image, depending on whether ittends to interfere constructively or destructively with the inherentspindensity contrast. In either case, the measured signal intensities in theimage will differ to some degree from those of a quantitative image,and thiscould lead to incorrect estimates of the DMSO concentration.As mentioned in Chapter 2, the equipment used in this thesis limited thesetting of TE in the imaging sequence to a minimum of 16 ms. If thesamplecontains nuclei with T2 relaxation times shorter than 5 times TE, or 80ms,then the image is expected to show T2 contrast. Since much shorter T2valuesare found in tissue samples than in DMSO/H0/NaC1 solutions (see Section3.3.3) and since T2 decreases as the temperature is lowered, problems with T2contrast in images of intact tissue samples are likely, especially at lowtemperatures. As with T1 contrast, T2 contrast can affect the relationshipbetween image signal intensities and concentrations of DMSO and water.Both T1 and T2 contrast are likely to be seen in the images ofbiological samples in this thesis, and an analysis of how this contrast willaffect the measurement of concentrations is required.753.2. EXPERIMENTAL3.2.1. Surgical ProceduresAll surgery and preparation of rat organs was done by Betty Pearson,microsurgical specialist in the research laboratories of Dr. Paul Keown at theJack Bell Research Center in Vancouver. Procedures are given in detail inChapter 5, Section 5.2.1 (blood washout procedure 3). The organs were excisedfrom the animal after blood washout, and placed in cold saline for subsequentcold transport.3.2.2. Measurement of T1 and T2 Relaxation Timesin Rat Organ TissueSolutions of 0—50% DMSO in University of Wisconsin (UW) solution (seeChapter 5) were made. Samples of organ tissue were cut and placed in thesolution at 0°C for 20—22 hours. The pieces of tissue were less than 5 mmthick to ensure that equilibration with the DMSO solution would be completewithin this time.Samples of tissue were cut with a scalpel from cold—stored organ tissuesequilibrated with DMSO solutions. The tissue was blotted dry with a Kimwipe,and inserted into a 5 mm NMR tube. Care was taken to avoid damaging thetissue further to avoid the loss of fluids.Measurements of relaxation times and control of temperature were carriedout as described in Section 2.3.1.1.763.3. RESULTS AND DISCUSSION3.3.1. Modelling T1 and T2 Contrast3.3.1.1. Calculation of T1 ContrastEquation 3.1 describes how T and T2 affect the signal measured from asample with imaging parameters TR and TE. At high temperatures (where T2>80ms) the equation simplifies to Equation 3.2, containing only the spin densityand T1 factors. Because TE<<TR in the imaging protocols in this thesis, theT1 factor in Equation 3.2 is simplified, and the expression for signalintensity becomesS(TR) D(’fl) * [1 — eTWTI] . [3.5]The intensity of signal from samples (or portions of samples) with differentconcentrations of DMSO can be calculated by multiplying the relativeconcentration of DMSO (SM = D’(H)) by the T1 factor (“TiF”; see description ofEquations 3.2—3.3); the result is SC1=SM*T1F.The signal expected from an imaging sequence with T1 contrast can beused to quantitate DMSO and water from images if:(1) the T1 relaxation times, and thus the signal reduction caused by T1,are approximately the same for all concentrations of DMSO, or(2) if the T1 relaxation times vary in a linear or near—linear fashionwith concentration such that the reduction in signal due tocontrast does not destroy the linear dependence of signalintensity on concentration.77In either of the above cases, the effect of shortening the TR of theimaging sequence can be modelled using an average T1 factor to calculate themeasured signal; the expected signal is calculated as SE1SM*T1Fvei whereTiFave is the average T1 factor over a range of concentrations. If this modelis appropriate, then T1 contrast can be assumed to cause only a scalarreduction in the amount of signal for any concentration of DMSO or water, andwill not affect measurements of concentration from image signal intensities.The model can be tested by plotting both SC1 and SE1 versus SM fordifferent concentrations of DMSO or water. The error in measuring the signalintensity from an image is approximately ±5% (see Chapter 2). The modeldescribes the data well if the T1 contrast has an effect smaller than theerror in measuring the signal intensity.3.3.1.2. Calculation of T2 ContrastT2 contrast is expected to occur in images of intact tissue samples inwhich TE is greater than 20% of the smallest T2 in the sample. Since TR wasset to less than five times T1 in images of diffusion and perfusion in thisthesis, there will be T1 contrast in these images as well. T1 and T2 contrastcan be taken into account at the same time using Equations 3.1 and 3.3. Toestimate the effect of T2 on image contrast, the T2 factor (“T2F”; seedescriptions of Equations 3.2—3.4) was calculated for each concentration ofDMSO and multiplied by the relative concentration of DMSO (SM), resulting in acalculated signal SC2=SM*T2F.As with T1 contrast, T2 contrast effects on signal intensitymeasurements can be modelled by assuming that the signal is reduced only by ascalar for all concentrations. This scalar was assumed to be the average T2factor, “T2Fave”i where T2Fave is the average T2 factor over a range of78concentrations. Values of the signal intensity expected from this model (SE2)were calculated as SE2SM*T2Fave.In order to take T1 contrast into account as well, the process describedfor calculating T1 contrast was used on the data in the following manner: Thecalculated signal intensity values for T2 contrasted signal (SC2) weremultiplied by the T1 factor, yielding a total calculated signal Sc,SC= SM* T2F* T1F [3.6]which is simply a different way of stating Equation 3.1.The model which assumes that the only effect of combined T1 and T2contrast will be a reduction of the signal by a scalar value for allconcentrations of DMSO is then calculated asSE = SMe T2F * T1F [3.71Both SC and SE can then be plotted against SM. If the model is correct, thenSC should fall within ±5% of SE for all values of SM.The error in measurement of relaxation times can also be taken intoaccount, assuming a 15% (±7.5%) measurement error in T1 and T2 (4). Theseerrors will affect SC, the calculated signal intensity at each concentration.The errors can be carried through the above calculations using Equations 3.8and 3.9:f5 [(T)2 * (3T1F)2 + (T)2 * (i5T2F)2]” [3.8]where zS is the error in the calculated signal, AT1 and AT2 are the 15%measurement errors for relaxation times, and 8T1F and öT2F are the derivativesof the T1 and T2 factors in the equation for calculated signal intensity, e.g.79&2F in Equation 3.9. The errors in calculated signal intensities for eachconcentration are shown as error bars in the following figures.5(T2F) -TE -TEöT2F= = (— ) exp(— ) [3.9](T2) T3.3.2. Image Contrast in Model Solutions3.3.2.1. T1 Contrast in Short-TR ProtocolsShort-TR imaging protocols are used in this thesis to monitor thedistribution of DMSO in tissue samples (see Chapter 4) . Since the variablecontrolling the differences in T1 values for DMSO protons at a giventemperature is the concentration of DMSO, T1 contrast will either increase ordecrease the differences in signal intensity between portions of the samplewith different DMSO concentrations.In Section 3.3.1, a model of signal intensity was described whichcompares the DMSO and water signals affected by combined T1 and T2 contrast tothose expected in the absence of these effects. This model assumed only thatwhen the TR and TE of the imaging sequence were not set to the values requiredfor quantitative imaging, the signal would decrease by a scalar, independentof the DMSO concentration. This model, where it fits the data, means thatalthough T1 and T2 values vary with concentration, the resulting differencesin contrast are not greater than the error in measuring the signal intensity,and thus can be ignored. In Chapter 2, it was noted that the error forquantitation of DMSO or water using signal intensity measurements from aquantitative image is about ±5%. This error is used in the model and is shown80as dashed “error lines” in the following figures. The error in the calculatedsignal due to measurement error in T1 and T2 is shown as error bars for eachconcentration.Figures 3.1 and 3.2 show how the differences in T1 in a samplecontaining 10—90% DMSO and 10—100% water affect the measured signal intensityfor each DMSO concentration in an experiment where TR = 1 s. The relaxationdata are from model solutions of DMSO/H20/NaCl discussed in Chapter 2 (seeTable 2.1). Since at these temperatures, all T2 relaxation times in the modelsolutions are greater than 5 times the TE of the imaging sequence, there is noT2 contrast. T1 relaxation times are all less than 20% of the TR of theimaging sequence (i.e. <0.2 s) below -26°C, so this is the lowest temperaturewhere T1 contrast will occur in images of model solutions.81Figure 3.1. Effects of T1 contrast on the calculated DMSO signal using ashort-TR protocol with TR = 1 s, compared with the model which assumes thatthe contrast will reduce the signal by a scalar amount equal to the averagecontrast for all DMSO concentrations. Data shown from 20°C to -26°C.Figure 3.1 shows that the dependence of image signal intensity on DMSOconcentration is changed by T1 contrast if TR is equal to 1 second. Atconcentrations below 45—50%, at temperatures above —26°C, T1 contrast will notaffect measurements of DMSO concentration from the signal intensities in DMSOimages. It should be possible, if necessary, to correct intensity data forexpected contrast effects below DMSO concentrations of 60-70% since in thisrange the signal intensity varies with DMSO concentration in a predictable20°C. 7.5°CRelative DMSO concentration-10°C0 0.5Relative DMSO concentration-26°C0.5 0.5Sb/1‘—V0 0.5Relative DMSO concentration0.5 -Relative DMSO concentration—Modelled DMSO signal (with ave. T factor)-— +5% error in modelled signal (with ave. T, factor)—.5% error in modelled signal (with ave. T1 factor)Z Calculated DMSO signal (with T factors for each [DMSOj)and error in calculated signal82way. At —26°C, T1 contrast affects signal intensities only slightly.Because DMSO in high concentrations is suspected of being toxic tokidney tissue, most research has been done with DMSO concentrations of 20% orless. In this case, imaging protocols with short TR should yield image signalintensities which vary in a linear fashion with respect to DMSO concentration.Signal intensities measured from images with DMSO concentrations above 45-50%will be affected by T1 contrast, and would not be useful for determining DMSOFigure 3.2. Effects of T1 contrast on the calculated water signal, using ashort-TR imaging protocol with TR = 1 s, compared with the model which assumesthat the contrast will reduce the signal by a scalar amount equal to theaverage contrast for all water concentrations. Data shown from 20°C to —26°C.7.5CC0.5 0.5CsI —- I I7 7— — —7777I I -—I I0.5Relative waler concentration-lO’C00 0.2 0.4 0.6 0.8Relative waler concentration.26eCCs 0.5U0.5I I/7=,--- —7 777/7/ 7/77I I I00 0.2 0.4 0.6Relative Water concentration0.5Relative water Concentration—Modelled water signal (with average T1 factor)— +5% enor in modelled signal (with average T1 factor)-— -5% error in modelled signal (with average T1 factor)Calculated water signal (with T1 factor for each [water]).and error in calculated signal0.8 183concentration distributions unless the signal intensities were corrected forcontrast effects.Figure 3.2 indicates that the dependence of signal intensity onconcentration in water images will also be affected by T1 contrast. Signalintensity data from water concentrations below 50% could be used forquantitation of water, but above this concentration the signal does notincrease with increased water concentration at most temperatures. This is dueto constructive interference of the T1 contrast with inherent spin densitycontrast. Thus it would be impossible to determine water concentrations above50% using the signal intensity measurements from an image with TR = 1 S.Since water concentrations in tissues are normally 70—85% (3), signalintensity data from short—TR imaging experiments will not be useful for waterquantitation in biological applications if the T1 behavior in tissues issimilar to that in the model solutions.3.3.2.2. T2 Contrast in Short-TR C4S NMR ImagesT2 contrast in C4S NMR images is not always avoidable, even ifthere are no limitations on RF pulse power, and very small linewidths(corresponding to long T2 relaxation times) in the sample. Linewidths inbiological samples are generally large, and pulse power is limited by thehardware being used, and the shape of the selective RF pulses. T2 contrastoccurs in C4S NMR images if the TE is larger than 20% of the shortest T2 inthe sample. This T2 contrast can either increase or decrease the differencesin signal intensity between portions of the sample with different DMSO orwater concentrations.T1 and T2 contrast will affect measurements of concentration fromsignal intensities if the differences in T1 and T2 for different84concentrations of DMSO are significant, and if the contrast causes signalintensity to vary with DMSO concentration in a nonlinear fashion. The modelof signal intensity described in Section 3.3.2.1 compares the effects ofcombined T1 and T2 contrast on the DMSO and water signals with the signalintensities expected if those effects were not observable. If the data fitthis model, then although the combined T1 and T2 contrast varies with DMSOconcentration, the effect is not greater than the error in measuring thesignal.Based on the relaxation times measured in model solutions, there shouldbe no T2 contrast in water or DMSO images taken at or above —26°C. However,at —46°C and below, measured T2 values fall below 80 ms. (Since TE = 16 ms,this is less than 5 times TE.) Thus T2 contrasting is possible in modelDMSO/H20/NaCl solutions at low temperatures.-46’C -63’C1 I 1 IIRelative DMSO concentration Relative DMSO concentration—Modelled DMSO signal (with ave. T1 & TL factors)— +5% error in modelled signal (with ave. T1 & T factors)-5% error in modelled signal (with ave. T1 & T factors)Calculated DMSO signal (with T & TL factors for each [DMSO])and error in calculated signalFigure 3.3. Effects of combined T1 and T2 contrast on the calculated DMSOsignal, using a short—TR imaging protocol with TR = 1 s, compared with themodel which assumes that the contrast will reduce the signal by a scalaramount equal to the average contrast for all DMSO concentrations. Data shownfor —46°C and —63°C.85Figure 3.3 shows the calculated DMSO signal affected by combined T1 andT2 contrast at —46 and —63°C. For both temperatures, below 35—40% DMSO, thecombined T and T2 effects do not change the dependence of image signalintensity on concentration enough to be distinguishable from the model. Atthese concentrations, then, DMSO concentrations could be directly calculatedfrom image signal intensities. At DMSO concentrations above 35—40%, thecombined T1 and T2 effects enhance the image contrast somewhat, interferingconstructively with inherent spin density contrast. Since measured signalintensity does not vary linearly with concentration, this would cause thelower concentrations to be underestimated. The variability of the data may be-46°C .63CC1 I I I 1 I I I05--02040.60.8Relative water concentration Relative water concentration—Modelled water signal (with average T1 & T1 factors)-— +5% error in modelled signal (with average T & T factors)-— -5% en-or in modelled signal (with average T1 & T, factors)Caictilated water signal (with T1 & T factor for each [water])and en-or in calculated signalFigure 3.4. Effects of combined T1 and T2 contrast on the calculated watersignal, using a short-TR imaging protocol with TR = 1 s, compared with themodel which assumes that the contrast will reduce the signal by a scalaramount equal to the average contrast for all water concentrations. Data shownfor -46°C and -63°C.due in part to the difficulty in measuring relaxation times at lowtemperatures; also, the estimate of 15% error in measuring relaxationparameters is conservative.86Figure 3.4 shows the calculated combined T1 and T2 effects on the watersignal at —46 and —63°C. It is possible, considering the difficultiesmeasuring the T2 relaxation times at these temperatures, that waterconcentrations less than 30% might be quantitatively calculated from imagesignal intensities. As mentioned before, this information would probably notbe very useful since water concentrations in most experiments would exceed 30%(even at very low temperatures). It is clear from the data that determinationof the water concentration above 30% is not possible with this short-TRimaging protocol.3.3.2.3. General Effects of T1 and T2 ContrastEquations 3.6 and 3.7 can be applied to the relaxation data from modelDMSO/H20/NaCl solutions to see the general effects of T1 and T2 contrast on theexpected image signal intensities. Figure 3.5 shows the calculated T1 and T2contrasted DMSO signal intensities at temperatures from —63°C to 20°C. Theeffects of relaxation on signal intensities from images with TR = 1 s and TE =16 ms with respect to concentration are similar between —10°C and 20°C. Thecurves for low temperatures are not as smooth due to the difficulties and thegreater error involved in measuring T2 at low temperatures.It is probable that parE of the reason for the “falling off” of signalat higher concentrations of water (Figure 3.2 and 3.4) at the lowesttemperatures is that the concentration of water does not actually correspondto the original concentration in the solution, since some water will havefrozen out of the solution. A similar situation can be seen in Figure 3.5,since DMSO will also freeze out of solutions with initially very high or verylow DMSO concentrations. A more detailed treatment of the data would requireknowledge of the DMSO/H20/NaC1 phase diagram (which is not available atpresent).87Figure 3.5. Effects of combined T1 and T2 contrast on the calculated DMSOsignals using a short-TR imaging protocol with TR = 1 s, TE = 16 ms, from—63°C to 20°C. The relationship between signal intensity and DMSOconcentration is nearly linear for all temperatures.These graphs indicate that the contrast behavior with respect to theconcentration of DMSO is similar for all temperatures. This observation isimportant since it could allow predictions about contrast in tissues to bemade without the extensive measurements of relaxation parameters that weredone on model solutions in the present work. Since use of NMR imaging andspectroscopy equipment is expensive, ideally researchers in cryobiology shouldbe able to make minimal measurements of relaxation times in the system to bestudied, and from these data decide how to set imaging parameters. Theability to predict image contrast from a minimal number of relaxationmeasurements will also allow the use of fewer animals.00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9Relative DMSO concentration—Calculated DMSO signal for 2(YCCalculated DMSO signal for 7.5°C-— Calculated DMSO signal for -10°C-- Calculated DMSO signal for -26°CCalculated DMSO signal for 46°CCalculated DMSO signal for -63 °C883.3.3. Image Contrast in Intact Tissues3.3.3.1. T1 and T2 Relaxation in Rat Organ TissuesSince the above analysis depends on T1 and T2 data from modelDMSO/H20/NaC1 solutions, it should be reassessed using data from actual tissuesamples. The data required for such an assessment are:(1) DMSO concentration dependence of T1 and T2 relaxation times of waterand DMSO in tissues,(2) Temperature dependence of T1 and T2 relaxation times in tissues, and(3) How these relationships compare with the relaxation data for modelsolutions.Table 3.1 and Figure 3.6 show the T1 values of the water and DMSOprotons at 20°C in rat kidney and liver equilibrated with 0-50% DMSO in UWsolution, and in 0—50% DMSO in DMSO/H20/NaC1 solutions. T1 relaxation timesin rat organ tissues are generally less than those in the DMSO/H20/NaC1solutions. T1 of water protons decreases with increasing concentration ofDMSO in the rat organ tissue samples as well as the DMSO/H20/NaC1 solutions.89Table 3.1. Relaxation times in rat organ tissues.Sample Relaxationand time Percent DZ4SO equilibrated into tissue.Temperature0% 10% 20% 30% 50%Rat kidney T1 (1120) 1.9 S 1.3 s 1.1 $ 0.95 S 0.89 S20°C T2 (HO) 30 ins 20 ins 17 ins 14 ins 13 insT1 (DMSO) — 3.2 s 2.7 s 1.8 a 1.3 ST2 (DMSO) — 120 ins 82 ins 70 ins 51 maRat kidney T1 (HO) 1.0 a 0.68 s 0.56 S 0.49 s 0.48 S1°C T2 (HO) 20 ins 17 ins 15 ins 12 ins 11 insT1 (DMSO) — 1.5 s 1.5 s 0.76 s 0.72 sT2 (OMSO) — 116 ins 92 ms 46 ins 44 insRat liver T1 (H0) — 0.98 s 0.73 s 0.67 S 0.65 S20°C T2 (HQ) — 20 ins 15 ms 13 ins 12 insT1 (DMSO) — 2.0 s 1.3 s 0.98 S 1.0 ST2 (DMSO) — 89 ins 55 ins 56 ms 56 msRat liver T1 (1120) — 0.60 s 0.52 S 0.49 $ 0.43 s1°C T2 (1120) — 20 ins 17 ins 15 ins 12 insT1 (DMSO) — 0.75 s 0.66 s 0.68 $ 0.57 sT, (DMSO) — 64 ins 45 ins 41 ins 28 ins90Figure 3.6. T1 of water and DMSO protons in rat kidney and liver equilibratedwith 0-50% DMSO in University of Wisconsin (UW) solution. T1 relaxation timesare generally less in rat organ tissues than in solution, but have similardependencies on the DMSO concentration.Table 3.1 and Figure 3.7 show the T2 values of the water and DMSOprotons in rat kidney and liver tissues equilibrated with 0—50% DMSO in UWsolution, and in 0-50% DMSQ in DMSO/1420/NaC1 solutions at 20°C. In all cases,the T2 values in rat organ tissues are an order of magnitude less than thosein the DMSO/H0/NaC1 solutions. All T2 values for water protons in rat organtissues are such that TE > 20% T2, i.e. all T2 values are greater than 80 ms.Thus T2 contrasting can be expected in all short-TR images of water in ratorgan tissues, even at temperatures above 0°C. T2 values of the DMSO protonsdecrease with increasing DMSO concentration in the rat organ tissue samples aswell as the DMSO/H20/NaC1 solutions.‘a0 5 10 15 20 25 30 35 40 45 50% DMSOT of water protons in kidney equilibrated with 0-50% DMSOT of DMSO protons in kidney equilibrated with 0-50% DMSOT of water protons in liver equilibrated with 0-50% DMSO‘“T of DMSO protons in liver equilibrated with 0-50% DMSOT of water protons in DMSO/water/NaC1 solutionsT of DMSO protons in DMSO/water/NaC1 solutions91I I I I I I I I I1000 ——0—---a--1 I I I I I I0 5 10 15 20 25 30 35 40 45 50% DMSOT2ofwater protons in kidney equilibrated with 0-50% DMSOD TofDMSO protons in kidney equilibrated with 0-50% DMSO-T. of water protons in liver equilibrated with 0-50% DMSO0Tg of DMSO protons in liver equilibrated with 0-50% DMSOT of water protons in DMSO/wateri’NaCl solutions— T of DMSO protons in DMSO/water/NaCl solutionsFigure 3.7. T2 of DMSO and water protons in rat liver equilibrated with0-50% DMSO, and in DMSO/H20/NaC1 solutions. T2 relaxation times for DMSOprotons are always longer than those of water protons. T2 relaxation intissues is an order of magnitude faster than that in solutions.92Figure 3.8. Temperature dependence of the T1 relaxation times of the waterprotons in rat kidney and saline solution (aqueous 0.15 M NaC1). T1 valuesdecrease with temperature in both samples.Since T1 and T2 relaxation times are temperature dependent, it is alsonecessary to know how they depend on DMSO concentration at each of thetemperatures that will be used in imaging experiments which monitor freezingin samples of tissue equilibrated with DMSO solution. Figure 3.8 shows thebehavior of the T1 values of the water protons at different temperatures inrat kidney tissue, and in saline solution (aqueous 0.15 M NaCl, the modelsolution for 0% DMSO in tissues). T1 shows similar behavior in both samples,decreasing with temperature. The T1 values of water protons in rat kidney arealways less than those in the model solution. The slopes of the curves aresimilar, indicating again that the relaxation of water (and by inference DMSO)in solution and in rat organ tissues is dependent on temperature in the sameway.0O—10 —5 0 5 10 15 20 25Temperature,°CT of water protons in kidneyT of water protons in aqueous 0.1 5M NaC1 (saline)93I I I I I1000— p100 ——DDD10 --1 I I I—10—5 0 5 10 15 20Temperature, CTof water protons in kidneyT of water protons in aqueous 0. 1SM NaC1 (saline)Figure 3.9. Temperature dependence of the T2 relaxation times of the waterprotons in rat kidney and saline (aqueous 0.15 M NaC1). T2 decreases withtemperature in both samples.Figure 3.9 shows the behavior of the T2 of water protons in rat kidney atdifferent temperatures compared to similar data for saline (aqueous 0.15 MNaCl). T2 in protons of water in rat kidney are less than those in the modelsolution by an order of magnitude, as noted before.The decrease in relaxation times with temperature is similar in waterand DMSO protons in tissues equilibrated with various concentrations of DMSO,as can be seen in Table 3.1. The T1 dependence on temperature is demonstratedin Figure 3.10, where the T1 values at temperatures from —50°C to 10°C for DMSOand water in rat liver equilibrated with 50% DMSO in UW solution are shownwith the comparable T1 values in 50% DMSO in saline (aqueous 0.15 M NaC1).It is clear that for both solution and liver, the largest T1 values forwater and DMSO are at the highest temperature and the lowest values at thelowest temperature. Also, values of T1 for DMSO are always larger than those94of water. To determine the correct TR for a quantitative imaging protocol,then, only a single measurement of the T1 of DMSO protons in tissue at thehighest temperature is necessary.Values of T2 for the DMSO protons at a given temperature and DMSOconcentration are always larger than those of the water protons. Thus, todetermine the correct TE for an imaging protocol where DMSO quantitation isdesired, only a single measurement of the T2 of DMSO protons in tissue at thelowest temperature is necessary.The general similarity in relaxation behaviour in the two rat organtissues studied could lead to incorrect assumptions about the image contrastin different tissues. This problem will be emphasized in the analysis ofimage contrast in the next section.95I I I I I I1.5 —10.:20Temperature, CT1 of water protons in liver equilibrated with 50% DMSOT1 of DMSO protons in liver equilibrated with 50% DMSO-* T1 of water protons in 50% DMSO in aqueous 0.15M NaClT of DMSO protons in 50% DMSO in aqueous 0. 15M NaClFigure 3.10. Temperature dependence of the T1 of the DMSO and water protonsin rat liver equilibrated with 50% DMSO, and in 50% DMSO in aqueous 0.15 mNaC1. T1 relaxation times for DMSO protons are always longer than those ofwater protons. The largest T1 relaxation times for both water and DMSO are atthe highest temperature.3.3.3.2. Evaluation of Contrast in Imagesof Rat Organ TissuesThe model describing combined T1 and T2 contrast developed in Section3.3.1 can be applied to the relaxation data for rat organ tissues presented inTable 3.1 to evaluate the accuracy of the DMSO and water concentrationsdetermined from measurements of image signal intensity. Images of rat organtissue are expected to be affected by T2 contrast far more than those of modelsolutions since the T2 relaxation times are all of the same order as the TEused in the imaging sequence (16 ms).Figure 3.11 shows the calculated water signal intensity in rat kidney96and liver images with combined T1 and T2 contrast (error bars) compared withthe modelled signal intensity and its 5% error lines (dashed). Data are shownfor two temperatures, 20°C and 1°C. T1 and T2 contrast in liver tissue isexpected to cause the measured signal intensity to vary with DMSOconcentration in a nonlinear fashion. At the water concentrations shown, 50—90%, it should be possible to use the measured signal intensities in images ofkidney (and perhaps liver) to determine water concentration, since the errorin measurement of the relaxation parameters accounts for most of the deviationfrom the model, and because the estimate of error in relaxation measurementsis quite conservative, so the error in the calculated signal may be largerthan is shown.Figure 3.12 shows the calculated DMSO signal intensities (error bars) inrat kidney and liver images together with modelled intensities (with dashed 5%error lines). The signal intensities vary with DMSO concentration as expectedfrom the model; contrast has simply reduced the signal intensities at eachconcentration by a scalar factor. Thus 0—50% DMSO concentrations should bedirectly measurable from signal intensities in DMSO images of rat kidney andliver tissues.970.40.30.20.104)— Modelled water signal (with ave. T and T factors)-— +5% error in modelled signal (ave. 1 and T factors)-— -5% error in modelled signal (with ave. T, and T factors)Z Calculated water signal, with ‘l & TL factors for each [water]and error in calculated signalFigure 3.11. The calculated water signal intensity in rat kidney and liverimages with combined T1 and T2 contrast (error bars) compared with themodelled signal intensity and its 5% error lines (dashed) . Data shown for twotemperatures, 20°C and 1°C. P1 and T2 contrasts in liver tissue are expectedto cause the measured signal intensity to vary with water concentration in anonlinear fashion.Kidney, 20°C Kidney, 1°C0.6 0.8Relative water concentration0.6 0.8 1 0.6 0.8Relative water concentration Relative water concentration98Kidney, 20’C Kidney, I ‘CI I 0.4 I0.3 — 0.30.I0.23° 0.1 0.2 0.3 OA 0.5Rlative DMSO concentration Relative DMSO concentrationLiver, 20’C Liver, I ‘C0.41 I0.402 03 04Relative DMSO concentration Relative waler concentration— Modelled DMSO signal (with ave. T and Tjactors)-— +5% efror in modelled signal (ave. T1 and T factors)-— -5% en-or in modelled signal (with ave. T1 and Tjactors)X Calculated DMSO signal (with T1 & TL factors for each [DMSO])and en-or in calculated signalFigure 3.12. The calculated DMSO signal intensities (error bars) in ratkidney and liver images along with modelled signal intensities and 5% errorlines (dashed), The signal intensities vary with DMSO concentration asexpected from the model; contrast has simply reduced the signal intensities ateach concentration by a scalar factor.993.4. CONCLUSIONS: CONTRAST IN C4S NMR IMAGESOF RAT ORGAN TISSUESThis chapter has discussed T1 and T2 data from rat organ tissuesequilibrated with various percentages of DMSO in UW solution, as compared withdata from model DMSO/H2O/NaC1 solutions. These data clarify the requirementsand the potentials of the C4S NMR imaging protocol as applied to short—TRimaging sequences applied to tissues, with special emphasis on rat liver andkidney. The main points made above will be summarized here.(1) T1 and T2 contrasts in short—TR images of DMSO in rat organ tissuesdo not significantly change the relationship betweenconcentrations of DMSO and image signal intensities in experimentsmonitoring diffusion of 50% DMSO or less; thus signal intensitiescan be used to determine DMSO concentrations in such experiments.(2) In the same type of experiment, contrast in water images of ratkidney does not cause the dependency of calculated signalintensity on concentration to vary significantly from a linearmodel. Thus water concentrations above 50% may also be determinedfrom image signal intensities.(3) Detailed relaxation measurements on an appropriate model system areuseful in a limited way for predicting the relaxation behavior intissue samples. Relaxation times in model solutions have the samegeneral dependencies on temperature and concentration as dorelaxation times in tissues, but the magnitudes of relaxationtimes (especially T2) are different.(4) Figures 3.11 and 3.12 seem to show that the calculated signalintensities have very little temperature dependence. This isbecause the influences of the T1 and T2 factors, which control the100relationship between signal intensity and temperature, interferedestructively with each other. (T1 contrast decreases withdecreasing temperature, whereas T2 contrast increases.)The above findings will be useful in interpreting the signal intensitydata from imaging experiments which monitor diffusion and perfusion (seeChapters 4 and 5).101REFERENCES(1) Hahn, E.L. Spin echos. Phys. Rev. 80, 580—594 (1950).(2) Hendrick, R.E. Image contrast and noise. Magnetic Resonance Imaging. D.D.Stark arid W.G. Bradley, eds. (The C.V. Moseby Co.: St.Louis, 1988).pp.66—83.(3) Mansfield, P. and P.G. Morris. NMR Imaging In Biomedicine. J.S. Waugh,ed. Advances in Magnetic Resonance, Supplement 2. (Academic Press, Inc.:New York, 1982). pg.13.(4) Baker, R. Controlled Release of Biologically Active Agents. (John Wileyand Sons: New York, 1987). pg. 42.102CHAPTER 4STUDIES OF DIFFUSION OF CPS INTO RAT ORGAN TISSUE4.1. INTRODUCTION4.1.1. Importance of Measuring CPSPenetration into TissuesThe use of cryoprotective solutions has lead to routine success incryopreservation of blood, embryos, semen, and other cells. This has allowedimprovements in the agriculture and fertility industries, as well as aiding inmany aspects of medical research. Similarly, cryopreservation of complextissues has the potential of revolutionizing medicine. Unfortunately, aftermore than thirty years of research, cryopreservation of intact organs is stillnot possible.It is not known why cryopreservation techniques developed for cells insuspension do not work for more complex tissues. However, there are importantdifferences between the cryopreservation experiments involving cellsuspensions and intact tissues which are dependent on the size and complexityof the sample. These differences are apparent during the followingprocedures:(1) Introduction of the CPS;(2) Freezing and thawing;(3) Assessment of viability.The difference is perhaps the most critical in the first procedure103mentioned. It is well known, as mentioned in Chapter 1, that the ability of aparticular CPS to protect a cell from freezing damage is dependent upon theCPS entering the cell. In cell suspensions, the entire surface of the cell isbathed in the CP solution, so equilibration of the CPS across the single cellmembranes is fast, especially in the case of DMSO. The equilibration of CPSwith complex tissues, however, must rely on the diffusion of the CPS throughmany layers of cells in the tissue. The fact that the success rate forcryopreservation is inversely proportional to the size of the tissue or organbeing preserved suggests a very simplistic explanation: That diffusion fromoutside a whole organ is not an efficient way to carry CPS to all the cells ina complex tissue.The development of CP solutions for the preservation of cells or tissuesis difficult. The ability of a CPS to penetrate cell membranes withoutdamaging the cells depends on the makeup of the vehicle solution, thetemperature at which the experiment is conducted, and the type of tissue beingpreserved. There is evidence that the permeability of cell membranes to DMSOchanges with temperature due to changes in the structure of the membrane (1).The complexity of these interactions means that the vehicle solution used withthe CPS must be optimized for each tissue type. Therefore it is not possibleto use the optimum solution for each tissue type in an organ.Optimization of a vehicle solution should include measuring the abilityof the CPS in that solution to penetrate tissues. If any part of the tissueis not protected by the CPS, it will be damaged by the cryopreservationprocedure. If it cannot be shown that the CPS has penetrated all the cells ofan organ, it is not reasonable to continue with a cryopreservation experiment.The techniques which are currently used to obtain information on CPSdistributions are very limited, but C4S NMR imaging can be used to obtain thisinformation much more accurately and easily.1044.1.2. Techniques for Measuring Diffusionof CPSs into TissuesMost of the techniques used to determine how quickly a CPS penetratestissues use radiolabels to trace the CPS. The measured amount of CPS in thetissue can be given a time dependence by exposing a number of identicaltissues to the CPS solution and removing them at intervals for testing. Themost commonly used of these procedures, the triple isotope method (2), wasfirst developed for studying cells in suspension, and standardizesmeasurements by assessing the amounts of protein and water as well as CPS.When this method is applied to tissue slices, however, it is assumed that thetissue samples are identical in terms of CPS penetration, and this introducessome uncertainties in interpreting the results. Other radiolabellingtechniques such as those used by Elford (3) and Harvey (4) also depend on theassumption of sample homogeneity.Chemical techniques for measuring amounts of DMSO in tissues exposed toCPS solutions have also been used. Pegg and Robinson (5) homogenized tissueslices, oxidized the CPS (glycerol) to formaldehyde, and measured a colorreaction (peak at 570 nm) with chromatotropic acid. Carpenter and Dawson (6)measured the amounts of DMSO in homogenized tissue samples using highperformance liquid chromatography.NMR spectroscopy has also been used to directly determine the totalamount of DMSO in intact rat liver (7), and to measure diffusion coefficientsof water and DMSO in rat liver tissue (16,17).The above techniques are “volume averaged”, since they all obtain thetotal CPS concentration by averaging over the whole sample volume. All volumeaveraged techniques share common problems:(1) The assumption that all tissue types within a sample have the samecapacity for the CPS,105(2) The assumption that all tissue types within a sample absorb the CPSat the same rate,(3) The inability to unequivocally determine the equilibriumconcentration of CPS in the tissue (due to problems 1 and 2),(4) The requirement that the sample be destroyed in the measurementprocess (except in the case of NMR spectroscopy).4.1.2.1. Attempts to Estimate Diffusion RatesThe problems listed above cause the results of all extantcryopreservation experiments on tissues and organs to be suspect. Despite thefact that at least one assumption underlying the use of the techniques aboveare incorrect (due to problem 1), these methods may give reasonable estimatesof diffusion rates.The earliest experiments investigating the diffusion of DMSO intotissues were performed on muscle tissues. These studies (3,8) indicated thatdiffusion was slow: Complete equilibration of 20% DMSO into a thin section ofguinea pig smooth muscle (taenia ccli) took 60 minutes at 37°C. Theseresearchers also noted that the type of solution used as a vehicle for theDMSO affected the diffusion rate.Studies of diffusion of DMSO into kidney also indicate a slow diffusionrate. Clark, Fahy, and Karow thoroughly studied the effects of CPSs indifferent vehicle solutions on diffusion rates in slices of rabbit kidneycortex (10). Kidney slices (0.5 mm) incubated at 25°C with 20% DMSOcontaining trace {‘4C)DMSO required about 20 minutes to accumulate the sameamount of labelled DMSO as the bathing solution. This study also showed thatthe diffusion rate of DMSO into the tissue sample depended somewhat on thevehicle solution used.The time required for delivery of a CPS to tissues via diffusion from106outside an intact organ is probably too great to insure the viability of thetissue after the cryopreservation procedure. In the above studies, the tissuewas cut into very small pieces, probably decreasing the diffusion barriers inthe tissue to a certain extent and disrupting the structure. However, ifthese diffusion rates are simply applied to whole organs, such as rat kidney,equilibration times on the order of hours are expected. If the intact tissueis to be preserved, it must be kept at lowered temperatures during thisprocess, further decreasing the diffusion rate.4.1.2.2. NMR Imaging and Diffusionin Intact TissuesThe limited evidence presented above has led cryobiologists to decidethat CPS must be delivered in a more efficient way; the obvious method forintact organs is to perfuse the CPS in a vehicle solution through thevasculature (blood vessels) of the organ, allowing a much larger surface areafor diffusion. This method requires that:(1) The vasculature provide surfaces which allow efficient diffusion tothe tissues;(2) The vasculature provide the same surface area for diffusion in allareas of the organ;(3) The vasculature can be kept open and accessible to CPS solutionsthroughout the procedure (11).With the available methods of measuring diffusion discussed in Section3.1.2, it is not possible to assure that the vasculature allows efficientdiffusion of the CPS through its walls. In fact, it is not practical toattempt to assess whether different tissue types within the organ (e.g. in thecase of kidney: Blood vessels, medullary and cortical tissues, etc.) absorb107CPS at the same rate or to the same extent.An ideal procedure for monitoring the penetration of a CPS into intacttissue would have the following characteristics: It would give information onwhere the CpS was located in the tissues and the concentration of CPS at eachpoint; it would discriminate between tissue types; it would discriminatebetween different CPS5; it would be non—invasive and non-destructive. C4S NMRimaging possesses all these characteristics.The measurements of diffusion rates and diffusion coefficients for DMSOin intact rat organ tissue in this chapter were not carried out in as muchdetail as the usefulness of the technique may later warrant, since visualinspection of the images indicated that the diffusion process was very slow.However, the image data was used to obtain measurements of DMSO diffusionrates and diffusion coefficients for comparison with the literature data fromthe much more time-consuming techniques described above (see Section 4.4.1).The methods for obtaining measurements of the DMSO diffusion rates anddiffusion coefficients in rat tissues from NMR image data are described andsome representative measurements made as examples.4.1.3. C4S NMR Imaging and Freezing ProcessesLittle information is available about the absorption of DMSO bydifferent tissue types within an organ. In Sections 3.1.1 and 3.1.2, it wasmentioned that techniques presently used do not give this kind of information,so cryobiologists have by necessity assumed that all tissues contain equalamounts of DMSO after equilibration. Using the C4S NMR imaging protocol withvery short TE, it is possible to independently monitor the freezing of twoliquids in a tissue sample, so freezing of water and DMSO could be monitoredsimultaneously. This technique could potentially answer the questions: Dodifferent tissue types freeze at the same rate, and how does the differential108freezing, if it exists, depend on the different DMSO concentrations found inthe tissues?Section 4.5 is a demonstration and discussion of the use of C4S NMRimaging to probe the freezing of tissues, and the requirements of thetechnique on the imaging hardware. The discussion shows clearly the effect ofthe present hardware limitation (TE 16 ins) on the ability to quantitativelymonitor DMSO and water concentrations in tissues at low temperatures.4.2. EXPERIMENTAL4.2.1. Preparation of Bovine Tissue SamplesAs an initial test of the ability of C4S NMR imaging to monitordiffusion of DMSO into a biological tissue, fresh samples of bovine liver andkidney were investigated. Pieces of these tissues were cut with a razorblade, 3 mm x 6 mm x 15 mm or 6 mm x 6 mm x 15 mm in the case of liver, and 10mm x 12 mm x 20 mm in the case of kidney. The pieces of tissue were tied atone end with thread and suspended in a 15 mm diameter NMR tube as shown inFigure 4.1.109Images of the tissue samples weretaken before addition of DMSO todetermine whether the image slice wasappropriate, that is, that there were noobvious inhomogeneities in the tissue atthat location. If inhomogeneities wereseen, the sample was moved up or down soa different slice could be imaged. Asolution of 30% DMSO in deionized waterwas then added to the NMR tube so thatit completely surrounded the tissuewhose volume was small compared to thatof the CP solution. About 5 ml of CFsolution was used in the case of theliver samples, and 8 ml in the case ofthe kidney sample.Figure 4.1. Arrangement of tissueThe NMR tube and sample were samples in the NMR tube.immediately inserted into the magnet,and the room temperature x and y shims were optimized. C4S images of waterand DMSO in the sample were taken at intervals to monitor the diffusion ofDMSO into the sample.NMR imaging parameters were set to maximize the signal intensity.Values of imaging parameters for the images in each figure are given inAppendix A.4.2.2. Preparation of Rat Tissue SamplesDetails of the surgical procedures are given in Chapter 5.For the quantitative C4S images of rat liver and kidney equilibrated110with University of Wisconsin (UW) solution, rat organs were harvested asdescribed in Section 3.2.3.1. Slices of these tissues were cut with a razorblade, and immersed in 20% DMSO in UW solution. After 24 hours the sampleswere placed in a 15 mm NMR tube surrounded by fresh 20% DMSO in tiW solution,and imaged. Imaging parameters (see Appendix A) were set so that the imagewas as quantitative as possible (TR = 5T1, and TE = 16 ras).Samples for diffusion experiments, and for the freezing experiment, weretaken by removing the rat liver or kidney from saline and cutting off anappropriately sized piece of tissue with a scalpel. The sample was thenplaced in a 10 mm or 15 mm NMR tube as described in Section 3.2.1.1 andsurrounded with 50% DMSO in physiological saline (0.15 M NaC1 in deionizedwater). The tube was immediately placed in the magnet. The probe had beenpreviously equilibrated to the required temperature, by the methods describedin Section 2.3.1.1. The tube and its contents equilibrated to thistemperature within 10 minutes. Room temperature shims were quickly optimizedduring this time and then imaging was begun.A set of experiments similar to those with bovine liver in Section 3.2.1were performed on slices of flushed cold—stored rat kidney and liver. Thesample to be imaged was hung in a 15 mm NMR tube surrounded by a solution of50% DMSO in saline. Imaging was begun after the shims had been quicklyoptimized. This experiment was done at 23°C and at 8°C for rat kidney, and at25°C and 10°C for rat liver.Diffusion of DMSO into a whole kidney was also investigated. In thiscase, a rat kidney was obtained which had been treated with heparin, butexcised without flushing with saline. The whole organ was kept on ice whilethe variable temperature unit was used to equilibrate the probe at 10°C. Theorgan was put into a 15 mm NMR tube and surrounded with 50% DMSO in saline(0.15 M NaC1). The kidney was imagedusing the C4S sequence, at intervalsbetween 30 minutes and 2 hours, using the imaging parameters given in AppendixA. A similar experiment was done with 2H0 replacing the water in the111solution to decrease the large signal from water in the CP solution.4.2.3. Obtaining Signal Intensity Datafrom NMR ImagesImages were Fourier transformed using the 1989 version of the BrukerInstruments operating system for the MSL—400 spectrometer. For quickestimates of the in—plane DMSO diffusion rate, the size of the tissue samplecross-section (cs) in the image was measured, and the time (t) to completeequilibration estimated visually from the images. The diffusion rate wasestimated as (cs/2)/t, in cm/s. When there was uncertainty about theequilibration endpoint, an average of rates was calculated using the last twoimage times in the series.Image files were also transferred to an Apple Centris 650 computer usingNMRLink software and cables from Bruker Instruments. DISMSL “headers”containing file information other than pixel intensity measurements werestripped from each file using a program written by Joseph Laughlin at BrukerInstruments, Bellarica, Massachusetts. Files of pixel intensities were thenimported into the NIH Image program, version 1.52b2, which is a public domainprogram available from the National Institute of Health, Bethesda, Maryland.The images of water from the series were examined carefully forinhomogeneities in the sample, such as different tissue types or pockets offluid. An area of the sample was chosen that had the least inhomogeneity anda path was chosen from outside the sample (in the surrounding fluid) to thecenter of the sample. The positions of the two pixels at either end of this“diffusion path” were noted.1123102•1OI - Tissue -- CPsolution D -I Da) I 00 5 10 15 20 25 30Pixel numberFigure 4.2. An example of signal intensity data from a diffusion path throughan image. (50% DMSO in saline solution diffusing into rat liver.)The DMSO images from each experiment were analyzed by exporting thepixel intensities from the diffusion path from each image in the diffusionseries to a separate file. These files then contain one—dimensional data(pixel intensity vs. position) which are essentially relative DMSOconcentration as a function of position along the diffusion path (Figure 4.2).4.3. C4S NMR IMAGING AS A MONITOR OF DMSO DIFFUSION4.3.1. Preliminary Studies of Diffusionof DMSO into Bovine Tissue SamplesQuantitative C4S NMR imaging of the diffusion process requires that thediffusion rate be slow enough that the time required for a single quantitativeimage (about one hour at room temperature) be small compared with the totaltime required for equilibration of DMSO in the tissue. If diffusion is fasterthan this, it is still possible to monitor the diffusion with NMR imaging, but113the TR must be decreased to less than 5 times the longest T1 in order todecrease the time required for a single image. According to the analysis inChapter 3, at DMSO concentrations less than 50%, this would not result in theloss of quantitative information, and would still give information on thelocation of the DMSO in the tissue at each time interval.It has been previously reported (see Section 4.1.2.1), that theequilibration time of DMSO into thin slices of rabbit kidney was of the orderof one hour. Assuming that DMSO equilibration rates are similar for all softtissues, the same estimate should hold for similar sized pieces of bovineliver and kidney. Thus in the initial diffusion studies of bovine tissue, theTR was set at as small a value as possible (1 s) while still obtaining viablesignals. This made the time for acquiring an image much less than theexpected equilibration time.Figure 4.3 shows cross—sectional images of a piece of bovine liver (3 mmx 6 mm in the imaging plane) surrounded by 30% DMSO in deionized water, atroom temperature. The time since the addition of the DMSO solution is givenunder each image in minutes. Diffusion of DMSO from outside the tissue sample(bright area) is clearly seen as an increase of signal intensity inside thetissue from one image to the next.Figure 4.4 shows cross—sectional images of a piece of bovine liversurrounded by 30% DMSO in 2H0, at room temperature. The time since theaddition of the DMSO solution is again given under each image in minutes.Diffusion of DMSO from outside the tissue (bright area) is clearly seen as thesignal intensity inside the tissue increases from one image to the next. Thegain was changed during the series of images, so the signal intensities arenot directly comparable between images.Figure 4.5 shows cross—sectional images of a piece of bovine kidneysurrounded by 30% DMSO in 2H0, at room temperature. The time since theaddition of the DMSO solution is given under each image in minutes. Diffusionof DMSO from outside the tissue (bright area) is clearly seen as the signal114intensity inside the tissue increases from one image to the next. The gainwas changed during the series of images, so again the signal intensities arenot directly comparable between images.115t=15 mmt=45 mm.4.wfr ‘4‘•—. •_f5t=30 mmt=60 mmFigure 4.3. Cross-sectional images of DMSO protons in a piece of bovine liver(3 mm x 6 mm in the imaging plane) surrounded by 30% DMSO in deionized water,at room temperature. The dark area in the center of the first imagecorresponds to the tissue sample. The DMSO solution surrounding it does notappear homogeneously bright due to field inhomogeneity. The time since theaddition of the DMSO solution is given next to each image in minutes.Diffusion of DSO from outside the tissue sample (bright area) is seen as anincrease of the signal intensity inside the tissue from one image to the next.116t=250 mmFigure 4.4. Cross—sectional images of DMSO protons in a piece of bovine liversurrounded by 30% DMSO in -HO, at room temperature. The time since theaddition of the DMSO solution is given next to each image in minutes.Diffusion of DMSO from outside the tissue (bright area) is seen as the signalintensity inside the tissue increases from one image to the next. The gainwas changed during the series of images, so the signal intensities are notdirectly comparable between images. Image quality is not as good as that ofFigure 4.3 because the images were reproduced from photographs of thermalprints (the original data was lost)t=40 mm t=150 mmt=225 mm117t=295 mmt=200 mmt=400 mmFigure 4.5. Cross-sectional images of DMSO protons in a piece of bovinekidney surrounded by 30% DMSO in 2H0, at room temperature. The time sincethe addition of the DMSO solution is given next to each image in minutes.Diffusion of DMSO from outside the tissue (brighter area) is seen as thesignal intensity inside the tissue increases from one image to the next. Thegain was changed during the series of images, so the signal intensities arenot directly comparable between images. Image quality is not as good as thatof Figure 4.3 because the images were reproduced from photographs of thermalprints (the original data was lost)t=80 mm1184.3.2. Potential Complications in C4S Imagingof DMSO Diffusion4.3.2.2. Contrast in Short-TR C4S NMR Imaging ProtocolsThe images in Figures 4.3 — 4.5 show that C4S NMR imaging can be used tomonitor diffusion of DMSO into intact tissues They also show that thediffusion process is too fast to allow quantitative imaging; that is, TRcannot be set to 5 x T1 for the longest expected T1 in the sample; thus T1contrasting is expected. In Chapter 3 it was shown that T2 contrast is alsoexpected in images of rat tissue samples below room temperature, so T2contrast is not expected in the above images.Image contrast will not affect the calculation of DMSO concentrations ifthe CP solution contains less than 50% DMSO. In the following diffusionexperiments, 50% DMSO in 0.15 M NaC1 was used as a CP solution. The lowestconcentrations may be slightly underestimated, but the error is notsignificant (see Chapter 3). Also, in the following estimation of theeffective diffusion coefficient De(DMSO) by curve fitting, this effect wassmall because the curve was fit to that part of the data where signalintensities and corresponding DMSO concentrations are highest (21).4.3.2.4. Apparent Exclusion of DMSOfrom Rat Organ TissuesFigure 4.6 shows images of slices of rat liver and kidney equilibratedwith 20% DMSO in UW solution, at 20°C. A row of pixel intensities througheach image (path shown in white) is displayed beside each image. Image119parameters were set so that TR 15 s, TE = 16 ms, so that the imagingsequence was as close to quantitative as possible. If the average DMSO signalintensities from the solution are compared with the average DMSO signalintensities in the rat organ tissues, the apparent difference inconcentrations of DMSO between the solution and the tissues are approximately25% and 31% for kidney and liver, respectively.There are several possible explanations of this discrepancy inconcentration between tissue and solution. The most obvious is, that fullequilibration has not occurred. This is not the case, however, since fromexperiments on diffusion rates in rat tissues (later in this chapter), 24hours at room temperature is more than enough time for equilibration with sucha small piece of tissue. Also, the image shows that the DMSO is evenlydistributed through the tissue; if no concentration gradient is visible,equilibration is complete.The second possible explanation is T1 or T2 contrast. Since TR = 15 5in this protocol, T1 contrast is not present. Table 3.1 gives the T2 of DMSOin rat kidney and liver tissues as 55 and 82 milliseconds, respectively, for20% DMSO at 20°C. This accounts for a 20% reduction in the amount of signalfrom the kidney sample, and 25% for liver.120Figure 4.6. Images of DMSO protons in (A) rat liver and (B) rat kidneyequilibrated with 20% DMSO in UW solution. A row of pixel intensities throughthe image (path shown in white) is plotted beside image. Image parameterswere set so that PR = 15 s, TE = 16 ms, so that the imaging sequence was asclose to quantitative as possible (it should only be affected by T, contrast).The apparent differences in concentrations of DMSO between the solution andthe liver and kidney tissue are approximately 31% and 25% respectively.Another possible explanation of the apparent differences betweensolution and tissue concentrations of DMSO in Figure 3.2 is that rat organtissues exclude Dt’4S0 to a certain extent. There is one previous mention ofconcentration differences between DMSO solutions and tissues equilibrated withthem (8), when Farrant showed that guinea pig smooth muscle actually0 20 40 60Pixel position310U)a) 4210CuC)U)a) 4. 1•10a)a::0210I5000 -a::________020 40 60Pixel positionI I121concentrated the DMSO from solution. However, in a later paper, Elford (3)disputed that finding. Other studies using radiolabelling techniques showthat the diffusion kinetics of DMSO depend on the makeup of the bathingsolution (9); since no such studies have used UW solution, there is noevidence in the literature for exclusion or inclusion of DMSO by tissues whenthe UW vehicle solution is used.If error in measurement of the signal intensities and error in thecalculated signal intensity (due to 15% measurement errors for T1 and T2) aretaken into account, the total error is approximately 7—8%. If these errorsare added to the expected reduction in signal from contrast, the imageintensities for kidney fall within the expected range. In the case of liverthe measured average intensity is 1% lower than the expected range, butprobably this is not a significant difference.Chemical techniques for determination of DMSO concentration in tissuesrequire the calculation of the “accessible volume” for diffusion. This is thevolume of the tissue which equilibrates with the DMSO, and is usuallyapproximately equal to the volume of tissue water (3) . This volume isdifferent for various tissue types, depending on their water content. Forexample, the water content of kidney is 81%; of vasculature, 78—81% (10).Thus a 20% reduction in the DMSO signal from these tissues after equilibrationwith a CP solution is expected if one assumes that the volume of the tissuenot occupied by free water will not absorb DMSO. All the possibleexplanations for the apparent difference in DMSO concentration between tissueand solution lead to the conclusion that no such differences in DMSOconcentration actually exist; however, if they did they would be quitedifficult to detect with C4S NMR imaging unless they caused at least 5—10%differences in signal intensity.From the data cited above, only a 3% difference at most in accessiblevolume is expected in the tissue types found in kidney, which should not causeany observable differences in final DMSO concentration in NMR images.122However, Figures 4.10 and 4.11 demonstrate that there are differences inequilibration rates in pieces of kidney and liver tissue. Although thedifferent tissue types in these images were not identified, it is very likelythat the areas of slowest equilibration correspond to vasculature.A more detailed study of this matter would be interesting. DMSOconcentration in the different tissue types present in rat organs couldpossibly depend on the tissue density or cell types, or simply on thedifferent equilibration rates. If these different rates are not known, andenough time is not allowed for all the tissue types to reach equilibrationwith the CP solution, then it is not possible to guarantee that the organ iscompletely protected from freezing damage. The actual concentration of DMSOin each tissue type will affect the freezing point of the tissue and thus itsability to survive freezing procedures.4.3.3. Potential and Limitations of C4S Imagingof DMSO Diffusion in Rat TissuesT1 and T2 data from rat tissues equilibrated with DMSO (Chapter 3), andpreliminary images of diffusion in bovine tissue samples (Section 4.3.1) havebeen presented. These data clarify the requirements and the potentials of theC4S NMR imaging protocol as applied to diffusion of DMSO into intact tissues,with special emphasis on rat liver and kidney. The main conclusions are:(1) The C4S NMR imaging protocol can be used to monitor diffusion ofDMSO into intact tissue.(2) Imaging protocols with TR shorter than required for quantitativeimaging must be used because the diffusion process is too fast toallow quantitative imaging.(3) Quantitative DMSO images of rat tissues equilibrated with 20% DMSO123in UW solution show that signal intensities from rat kidney andliver are apparently 15% and 31% (± 2%) lower, respectively, thanthe signal intensities from the surrounding solution in diffusionexperiments. This signal reduction is satisfactorily explained byT2 contrast and the measurement errors.4.4. DIFFUSION OF DM50 IN RAT KIDNEY AND LIVERMONITORED WITH C4S NMR IMAGING4.4.1. Estimates of DMSO Diffusion RatesFigure 4.7 shows a series of images of 50% DMSO in saline (aqueous 0.15M NaC1) diffusing into a sample of rat kidney at 23°C. The time since thebeginning of the experiment is shown next to each image in minutes.Equilibration has occurred after approximately 2 hours 30 minutes (150 miri) ina sample approximately 4.5 x 5 mm. Thus the DMSO is diffusing at anapproximate rate of 2.9x105 cm/s in the imaging plane.The in—plane diffusion rate (half the sample width divided by the timeto full equilibration) for 50% DMSO in saline in a whole kidney (with capsuleintact) was estimated to be 4.3x106 cm/s at 20°C. Figure 4.8 shows thisseries of images of DMSO diffusing into a whole kidney.Figure 4.9 shows the diffusion of 50% DMSO in saline into a 7.5 mm x 7.5mm sample of rat liver at 25°C. Equilibration time was approximately 360minutes; thus the diffusion rate is 1.7x105 cm/s.Figure 4.10 shows a series of images of 50% DMSO in saline diffusinginto a 7 x 7 mm sample of rat kidney at 8°C. Equilibration has occurred afterapproximately 5—16 hours in solution, although in this instance the sample isnot homogeneous in the imaging plane; the lower left of the sample is a124different tissue type, possibly part of the vasculature (19). An estimate ofin—plane DMSO diffusion rate in rat kidney at this lower temperature is about1.9x 10 cm/s, with a large potential error in terms of sample homogeneity.Figure 4.11 shows a series of images of 50% DMSO in saline diffusinginto a 7 x 7 mm sample of rat liver at 10°C. Again, the sample is nothomogeneous, so the estimate of in—plane diffusion rate, 5.1x106 cm/s, mustbe considered less reliable.There are only two estimates of the value of the diffusion rate of DMSOinto intact tissues in the literature (see Section 4.1). One is a diffusionrate at 37°C in guinea pig smooth muscle (taenia coli) (3), 2.8x105 cm/s.(This rate is approximately 2.8x105 cm/s, estimated by dividing half thesample thickness by the time to complete equilibration; unfortunately thesample thickness was not given in reference 3, and had to be estimated.) Theonly diffusion rate estimate for rat kidney is at 25°C, 2.1x105 cm/s,calculated from information in reference 11 which was obtained by aradiolabelling technique. All of these results suggest that the rate of DMSOdiffusion into intact tissues is very temperature dependent. However, noestimates were made at the temperatures used most commonly for theintroduction of CPS. The literature data are close to the values found in thepresent work which indicates that rough estimates of diffusion rates can bemade simply and directly using C4S NMR imaging, and that these compare wellwith the best estimates made using much more complex and hazardousradiolabelling techniques.125DMSO Water20 miii40 mm75 miii1130 miiiTE = 16 msTR = 1 5Figure 4.7. A series of images of DMSO protons in 50% DMSO in saline (aqueous0.15 N NaC1) diffusing into a sample of rat kidney at 23°C. The time sincethe beginning of the experiment is shown next to each image in minutes.Equilibration has occurred after approximately 2 hours 30 minutes (150minutes) in a sample approximately 4.5 x 5 mm. Thus the DMSO is diffusing atan approximate rate of 2.9x105 cm/s in the imaging plane.126t=120 mmflt=230 mmt=480 mmFigure 4.8. A series of images of DMSO protons in 50% DMSO in salinediffusing into a whole kidney with the time since the beginning of theexperiment given next to each image in minutes. The in-plane diffusion rate(half the sample width divided by the time to full equilibration) for 50% DMSOin saline in a whole kidney (with capsule intact) was estimated to be 4.3x106cm/s at 20°C.t=365 mm127DMSO Water30 miii160 mm220 miii780 mmiiTE = 16 msTR = 1 sFigure 4.9. A series of images of DMSO protons showing the diffusion of 50%DMSO in saline into a 7.5 nun x 7.5 mm sample of rat liver at 25°C.Equilibration time was approximately 360 mm; thus the diffusion rate is1.7x10 cm/s.128DMSO Water20 mm65 mm130 mm190 mmTE = 16 msTR = 1 sFigure 4.10. A series of images of DMSO protons in 50% DMSO in salinediffusing into a 7 x 7 mm sample of rat kidney at 8°C. The time since thebeginning of the experiment is given next to each image in minutes.Equilibration has occurred after approximately 5-16 hours in solution,although in this instance the sample is not homogeneous in the imaging plane.An estimate of in—plane DMSO diffusion rate in rat kidney at this lowertemperature is about 1.9x105 cm/s, although there is a large potential errorin terms of sample homogeneity.129DMSO Water10 miri195 mm390 mm650 mmTE = 16 msTR = 1 sFigure 4.11. A series of images of DMSO protons in 50% DMSO in salinediffusing into a 7 x 7 mm sample of rat liver at 10°C. The time since thebeginning of the experiment is given next to each image in minutes. Thesample is not homogeneous, so the estimated in—plane diffusion rate, 5.1x106cm/s, may not be as reliable as the measurements from other image sequences.130I I I I I I3- -guineapig -smooth muscle.bulk rat kidney ._—2-i- 0.5mm kidney slices. -_01 - bulkrat liver -o whole rat kidney0 I I I I I Io 5 10 15 20 25 30 35 40Temperature,6C20% DMSO diffusing into guinea pig smooth muscle.I 20% DMSO diffusing into 0.5mm kidney slices.50% DMSO diffusing into bulk rat kidney (imaging)50% DMSO difThsing into whole rat kidney (imaging)• 50% DMSO diffusing into bulk rat liver tissue (imaging)Figure 4.12. DMSO diffusion rates as a function of temperature. Valuesestimated from literature data are shown with values estimated from NMR imagedata.Figure 4.12 compares DMSO diffusion rates from all techniques as afunction of temperature. Uncertain rates from NMR image data are shown asaverages of high and low estimates (determined as described in Section 4.2.3).The diffusion rate decreases with temperature as expected. Temperatures lessthan 10°C are usually used when introducing CPSs into organs that are to becryopreserved. The reason for this is that organs survive longer without lossof viability at these lowered temperatures. Also, DMSO is more toxic attemperatures approaching 37°C (12,13,14).1314.4.2. Measurements of DMSO Diffusion Coefficientsin Rat Organ Tissues4.4.2.1. Possible Errors Associated with Previous MethodsThe “volume averaged” methods described in Section 4.3 have severalbasic flaws. The most important of these is the assumption that themeasurement reflects a single diffusion coefficient in tissue. In the studyof porous media, the measured value is actually the effective diffusivity, Deirelated to the bulk solution self—diffusivity D by Equation 4.1:DtD [4.1]where the constant t is the tortuosity, which includes effects of poregeometry. Diffusion in tissues is more likely to be accurately described bymatrices D(x,y,z), t(x,y,z), and D(x,y,z) for each volume element (x,y,z) inthe sample, since tissues with different functions and structures willprobably have different effective diffusivities. For example, in kidney, thediffusivity inside the larger blood vessels could be similar to the self—diffusivity of DMSO in water, while in the cortical tissues the diffusivitywould be much lower due to barriers such as cell membranes. It is clear thatdiffusion coefficients measured in tissue samples by radiolabelling techniquesreflect an average of the effective diffusion coefficients of the differenttissue types present within the whole volume. In addition to this averagingeffect, the self—diffusion coefficient D obtained with the pulsed fieldgradient NMR method is different from the effective diffusion coefficient Debecause it measures the microscopic diffusion of nuclear spins in a system atequilibrium, rather than the macroscopic diffusion of molecules down achemical gradient.132The data from the present studies of diffusion from NMR image dataavoids some of these problems. The tissue samples used in the present workwere thicker than those used in radiolabelling studies, and should thus beless prone to the effects of damage at the cut edges. Samples were cut so asto include only one type of tissue (e.g. kidney cortex), although thesetissues can still not be said to be homogeneous in the sense of a poroussubstance which can be modelled by Equation 4.1. The diffusion paths werechosen by inspecting the NMR images of the water protons in the sample (whichshould reveal most large differences in tissue type) for the areas with theleast tissue inhomogeneity. Paths were chosen also to correspond to theassumptions of the mathematical model used (see next section).4.4.2.2. Analysis of 1D Diffusion Data from NMR ImagesAs discussed previously, the NMR image data from the experiments shownin Section 4.2.3 can yield volume averaged data for comparison to methods inthe literature. For a measure which is analagous to that obtained byradiolabelling methods, the area of the sample in the imaging plane ismeasured, as well as how long the sample took to equilibrate with the DMSOsolution. Then Dave = sample area/time, in cm2/second.If, however, the macroscopic diffusion process can be described by aprocess resembling microscopic diffusion, such that D <r2>/4t (where <r2>is the mean square jump distance and t the correlation time of the motion),then Dave would be more accurately estimated as Daye,m (44’Dave Thisassumes the following: that the sample area can be described as itr2 for anaverage sample radius r, and that this radius is analogous to a mean jumpdistance in the microscopic sense.133The macroscopic diffusion process, in both the volume averaged methodsand the method described in the next section, is modelled by Fick’s law ofdiffusion. The CPS concentration diffusing into a tissue sample can bedescribed by a straight-line semi-infinite approximation:[4.2)atwith boundary conditionsC(O,t) C0 and C(x>O,O)O [4.3]with the restriction that the concentration gradients are far from the centerof the sample (21). C(x,t) is the concentration at point x along thediffusion path and time t, and C0 is the initial concentration of solvent inthe bathing solution. In this case the exact solution is (22):2C0 rC(x,t) — I exp( dp[4.4)2where ji is a dummy variable for numerical integration. This equation can besolved numerically by estimating values for De and optimizing the fit to thedata.Data from the “diffusion paths” described in Section 4.2.4.1 weremodified for fitting to the equation by eliminating the intensity data for thepoints outside the sample, so that the data started at the edge of the sample(Figure 4.13(A)). Only the data from the first few images (time points) wereused, to satisfy the assumption that the CP is diffusing only from the edge of134the sample where the intensity data “starts”. The intensity values for eachtime point from the modified diffusion paths were fitted separately to theequation. Best fits were chosen visually (Figure 4.13(B)> and thecorresponding De values compared. The final value of De for the sample wascalculated as the average of D values from the diffusion paths taken atdifferent diffusion times, and a standard deviation calculated (see Table4.1).Figure 4.13. An example (same data as Figure 4.2) showing how data aremodified for fitting by eliminating data points from outside the sample (top)and how Equation 4.4 is fitted to the data (bottom).CCCCCI IP DC r- ni CCC ri C C3•10Coa, 4h1Va,a,>a,0 5 10 15 20Pixel number3.4210a,x0.a,>a,0250 5 10 15 20 25Pixel number1354.4.3.3. Results of Diffusion Coefficient Measurementsand Comparison with Literature DataImages from series monitoring the diffusion of DMSO into rat liver andkidney tissues were analyzed as described in Section 4.4.2.2, and the resultsare presented in Table 4.1. The “volume averaged” method for measuring Davefrom NMR images can be compared with the more accurate method of fittingEquation 4.4 to signal intensities from diffusion paths within the images(Figure 4.14). Dave measured by the volume averaged method is consistentlyhigher than that measured by the more accurate curve fitting method; however,Davem seems to give results that are quite consistent with D, differing atmost by a factor of four. It is interesting that a microscopic model shouldgive almost the same results as the more accurate (and appropriate)macroscopic treatment.Both rat liver and kidney show an increase in the diffusion coefficientwith increased temperature, as expected. Rat kidney seems to be much morepermeable to DMSO than liver; however, the kidney was subjected to morethorough flushing procedures, and it is quite possible that the presence ofblood in the liver samples (even if uncoagulated) could slow down diffusion.136Sample Dan( cm2Is)T(°C)Dan a(cm2)s)D( cm2/s)average D( cm2/s)I_________________________________________ I________________________________I IRat 23 3.4x105 2.7x10’ 3.4x10’ (15) 2.7x10’kidney 2.0x106 (35)2.5x10 (50) (7x10’)Rat 8 2.1x105 l.7x106 4.7x10’ (15) 4.0x10’kidney 3.3x10 (40)3.7xl0 (60) (6x10e)4.2x10 (80)Rat 25 2.8x105 2.2x106 5.3x107 (40) 5.2x107liver 5.0x10’ (70)5.0x107 (115) (2x10)Rat 10 5.7x10’ 4.5x10’ 2.5x107 (65) 2.2x10’liver 2.5x10 (100)l.7x10 (195) (5xl0)Table 4.1. Average and effective diffusion coefficients at varioustemperatures measured by two methods. Dave is simply the area of the sample inthe image plane divided by the equilibration time. These values should becomparable to the literature data from radiolabelling studies. Dave,m ISDave/4ltp which treats the macroscopic diffusion as if it were a microscopicprocess. De is found by fitting signal intensities on a diffusion path fromsequential images to Equation 4.4 (time in minutes since beginning ofexperiment in parentheses) . An average value for each temperature is givenwith the standard deviation below it in parentheses.In each image series, three or four images were available during theappropriate time period. This allowed three or four separate measurements ofDc for each series, the average of which should give a stable measure. Asdemonstrated in Chapter 3, T1 and T2 image contrast does not affect thedependency of signal intensity on DMSO concentration. Thus the raw signalintensities can be used as if they were relative concentrations.Since there are no other techniques which give information on solventdistribution in tissues throughout a diffusion experiment, there are no trulycomparable measurements of De in the literature. NMP. spectroscopy can be usedto measure self—diffusion (a microscopic process) in tissues (16), andD(water) in rat liver tissues has been shown to be reduced by a factor ofapproximately 3 from that of bulk solution. It is not really appropriate tocompare the diffusion coefficients measured in bulk solution and in tissues,137since a microscopic diffusion model does not take into account the barriers ofcell membranes. However, the comparison does give an idea of the tortuosityof the sample (see description of Equation 4.1) which probably is in this casedetermined by the cell membrane barriers and the intercellular spaces orchannels.10 I-...——.o ..x1?voIume1 - ) averaged -:5‘0Temperature,CD, in rat kidney, measured by image intensity curve fitting.* D. in rat kidney, estimate by image areas.De in rat liver, measured by image intensity curve fitting.-X De in rat liver, estimate by image areas.Figure 4.14. The Dave(DMSO) from NMR images compared with the more accuratemethod of fitting Equation 4.4 to signal intensities from diffusion paths inthe images. Dav,(DMSO) measured by the volume averaged method is consistentlyhigher than that measured by the more accurate method.138100 I I10 - -_-----—---—---—--——------——° -‘-0.1 - . -.s0.01 - -0.001 - -0 I I I II I0 5 10 15 20 25 30 35 40Temperature,°CMeasurements of self-diffusion coefficients in 50% DMSOMeasurements of effective diffusion coefficients in rat liver-* Measurements of effective diffusion coefficients in rat kidneyFigure 4.15. Measurements of De(DMSO) from image data compared with the self-diffusion coefficients of DMSO in solution (16) at various temperatures.Values of D(DMSO) in solution decrease with temperature according to anexponential relationship (16), and seem to show the same behavior in rattissues. The reported decrease in D(water) between solution and rat livertissue is by a factor of approximately 3 (17). Here the difference betweenD(DMSO) in solution and D,(DMSO) in rat liver is a factor of approximately160.The measurements of De(DMSO) from image data can be compared with theknown self—diffusion coefficient D(DMSO) in solution (16). Figure 4.15 showsthe values of D(DMSO) in solution and De(DMSO) in rat tissues, with respect totemperature. Values of D(DMSO) in solution decrease exponentially withtemperature (16), and De(DMSO) seems to show the same behaviour in rattissues. The reported decrease in D(water) between solution and D(water) inrat liver tissue (measured by NMR spectroscopy) is by a factor ofapproximately 3 (17). Here the decrease between D(DMSO) in solution andDe(DMSO) in rat liver is by a factor of approximately 160. This is mostlikely due to the fact that D is a measure of microscopic motion in a systemat equilibrium, while De is a measurement of macroscopic motion of moleculesdown a chemical gradient, through multiple barriers.1394.5. C4S NMR IMAGING AS A MONITOR OF FREEZING IN TISSUESAs discussed in Section 2.2.2, NMR imaging can be used to monitorfreezing and thawing in tissue samples, using the fact that T2 of a liquiddramatically decreases when the liquid freezes, concurrent with a largeincrease in linewidth, eliminating that signal from the NMR spectrum andimage. Figure 4.16 shows a series of images of DMSO and water in a piece ofrat liver equilibrated with 50% DMSO in saline (0.15 M NaC1) as thetemperature is lowered. The temperature is given next to each image.The images look similar until the temperature is lowered to —35°C, wheresigns of freezing are observed. Since the T2 of 50% DMSO in rat liver is 28ms (<< 80 ms), however, there are effects of T2 contrast as well as freezing.At the same temperature, the surrounding solution begins to show elongatedfrozen structures, which appear in three dimensions as sheets. These areclearly visible in the NMR images as well as to the eye.As the temperature is decreased further, these structures fill more ofthe space surrounding the liver sample; the tissue itself shows no signal atall at —55°C. This could indicate, as observed in Section 4.3.2.4, that therat liver tissue is freezing at a temperature higher than expected; or, thatT2 is extremely short compared with TE. (Even with better hardware it may notbe possible to verify one of these possibilities using the C4S imagingsequence, since TE for this pulse sequence would likely have to be decreasedto around 1 ms to avoid T2 contrast. The shortest sinc2—shaped pulse with therequired selectivity of 800 Hz is 5 ms long.) At —70°C, none of the water inthe solution is left unfrozen, and only a small amount of DMSO is stillmobile.This series of images demonstrates that C4S NMR imaging is capable ofmonitoring the freezing of water and DMSO independently in tissues containingDMSO, if the T2 of DMSO in the tissue is known at each temperature, and if140hardware requirements can be met. The use of NMR spectroscopy to measure theT2 values before the imaging is done will give an idea of the “freezing point”of the tissue. C4S NMR imaging will then yield spatial information onfreezing in a chosen plane of the sample.141DMSO Water7°C-38°C-55°C-68°CTE = 16 msTR = 1 sFigure 4.16. A series of images of DMSO and water protons in a piece of ratliver equilibrated with 50% DMSO in saline (0.15 N NaC1) as temperature islowered. The temperature is given next to each image.1424.6. SUMMARYThis chapter has given a detailed analysis of the usefulness of C4S NMRimaging in the monitoring of DMSO diffusion and the measurement of DMSOdiffusion rates in intact tissues, as well as a demonstration of the use ofC4S NMR imaging to monitor freezing in a tissue sample:(1) Quantitative C4S NMR imaging as developed in Chapter 2 wasdetermined to take too long to be of use in monitoring diffusionrates.(2) “Volume—averaged” estimates of the DMSO diffusion rate in rattissues were made at two temperatures, and were shown to comparewell with rates determined by radiolabelling methods.(3) A technique was developed for manipulating data from images todetermine DMSO diffusion coefficients in rat organ tissues.(4) It was shown that freezing of both water and DMSO could be monitoredindependently in a tissue sample containing DMSO, if the ambiguitycaused by TE being of the same order as T2 at low temperatures canbe solved.(5) It was demonstrated that different tissue types probably havesimilar capacities for DMSO, but their equilibration rates withDMSO-containing CP solutions are potentially very different.This last point is extremely important, since the vascular systems of theorgans seems to be the most prone to injury by lowered temperatures. It ispossible that the reason for this is simply that the vascular system requiresa much longer time to equilibrate with CP solutions than do the other tissuesin the organ, and thus might have a lower CPS concentration than other tissueswhen the allowed equilibration time is ended. This would cause the non—equilibrated tissue to freeze at a higher temperature than the rest of theorgan. It has already been shown that the liquid in the vascular systemfreezes first in a cryopreservation experiment (20).143144REFERENCES(1) Pegg, D.E. et al. 1986. Optimization of a vehicle solution for theintroduction and removal of glycerol with rabbit kidneys. Cryob. 23, 53-64.(2) Dayian, G., S.N. Chin, and A.W. Rowe. 1971. Differences in human plateletpermeability to glycerol and DMSO. Cryob. 8, 393—394.(3) Elford, B.C. 1970. Diffusion and distribution of dimethyl sufoxide in theisolated guinea—pig taenia ccli. J. Physiol. 209, 187—208.(4) Harvey, B. and M.J. Ashwood-Smith. 1982. Cryoprotectant penetration andsupercooling in the eggs of salmonid fishes. Cryob. 19, 29-40.(5) Pegg, D.E. and S.M. Robinson. 1978. Flow distribution and cryoprotectantconcentration in rabbit kidneys perfused with glycerol solutions. Cryob.15, 609—617.(6) Carpenter, J.F. and P.E. Dawson. 1991. Quantitation of dimethyl sulfoxidein solutions and tissues by high-performance liquid chromatography.Cryob. 28, 210—215.(7) Fuller, B.J., A.L. Busza, and E. Proctor. 1989. Studies on cryoprotectantequilibration in the intact rat liver using nuclear magnetic resosnancespectroscopy: a noninvasive method to assess distribution of dimethylsulfoxide in tissues. Cryob. 26, 112—118.(8) Farrant, J. 1965. Permeability of guinea—pig smooth muscle to non—electrolytes. J. Physiol. 178, 1—13.(9) Clark, P., G.M. Fahy, and A.M. Karow, Jr. 1984. Factors influencing renalcryopreservation. I. Effects of three vehicle solutions and thepermeation kinetics of threee cryoprotectants assessed with rabbitcortical slices. Cryob. 21, 260—273.(10) Mansfield, P. and P.G. Morris. NMR Imaging in Biomedicine. Advances inMagnetic Resonance, Supplement 2, J.S. Waugh, ed. (Academic Press, Inc.:New York, 1982). pg. 13.(11) Collins, G. Organ transplantation: the role of preservation. TheBiophysics of Organ Preservation, NATO ASI Life Sciences Vol. 147, D.E.Pegg and A.M. Karow, eds. (Plenum Press: New York, 1987). pp. 3—14.(12) Fahy, G.M. 1986. The relevance of cryoprotectant “toxicity to- cryobiology. Cryob. 23, 1—13.(13) Hak, A.M., F.G.J. Offerijns, and C.C. Verheul. 1973. Toxic effects ofDMSO on cultured beating heart cells at temperatures above zero. Cryob.10, 244—250.(14) Clark, P., G.M. Fahy, and A.M. Karow, Jr. 1984. Factors influencing renal145cryopreservation. II. Toxic effects of three cryoprotectants incombination with three vehicle solutions in nonfrozen rabbit corticalslices. Cryob. 21, 274—284.(15) Hendrick, R.E. Image contrast and noise. Magnetic Resonance Imaging. D.A.Stark and W.G. Bradley, eds. (The C.V. Moseby Co.: Washington D.C.,1988). pg. 67.(16) Beall, P.T. 1983. States of water in biological systems. Cryobiology 20,32—334.(17) Beall, P.T., S.R. Imtey, and S.R. Casturii. Handbook for Biological NMR.(Pergamon Press: Elmsford, New York, 1984). pg.7.(18) Meyer, R.A. and T.R. Brown. 1988. Diffusion measurements by microscopicNMR imaging. J. Mag. Res. 76, 393—399.(19) Ahn, C.B. and Z.H. Cho. 1991. Diffusion and perfusion in high resolutionNMR imaging and microscopy. Mag. Res. Med. 19, 228-232.(20) Barner, H. B.. 1965. The vascular lesion of freezing as modified by DMSO.Cryob. 2, 55-61.(21) Dr. Richard imnons, personal communication.(22) Baker, R. Controlled Release of Biologically Active Agents. (John Wileyand Sons: New York, 1987). pg. 42.146CHAPTER 5STUDIES OF THE PERFUSION OF CRYOPROTECTIVE SOLUTIONINTO INTACT RAT KIDNEY5.1. INTRODUCTION5.1.1. Introduction of CPS into an Intact OrganThe bulk diffusion rates measured in Chapter 4 indicate that bulkdiffusion of DMSO from outside an intact rat kidney cannot deliver DMSO to allof the tissues within the kidney within a reasonable amount of time. It hasin fact been assumed since the l960s that delivery of CP solutions would bemost efficient if the vasculature of the organ was used. Since thevasculature is designed to provide adequate blood supply to all of thetissues, it is also probably appropriate for the distribution of a CPsolution. The only direct evidence for this assumption is derived from 1H NMRspectroscopy of rat liver perfused with DMSO (33). Three general points weremade about this assumption in Chapter 4 and will be repeated here:(1) The vasculature must provide surfaces which allow efficientdiffusion to the tissues;(2) The vasculature must provide the same surface area for diffusion toall areas of the organ (if distribution of CPS is to behomogeneous throughout the tissues);(3) The vasculature must be open and accessible to the CF solutionthroughout the perfusion procedure.147The first point is probably true, since CPS can be carried through thevasculature and diffuse through the walls of capillaries.The second point is not true of intact kidneys, since it is well knownthat the blood flow in the cortical tissue is greater than that of themedullary tissues due to differences in the vascularization of the tissues(2). Similar differences in vascularization should be suspected for anycomplex set of tissues.The third point is the most problematic. Perfusion of kidney causesseveral potential changes in the vasculature. Restriction of the flow throughblood vessels (vasospasm) is caused by handling of the organ (3), by °2 in theperfusate (4), by uneven cooling (5,6), and possibly by the production ofvasoconstricting chemicals by the kidney itself (7). Long exposure to DMSO(>2 hours) may also cause the vascular endothelial cells to break down,narrowing the capillaries and increasing the vascular resistance to perfusion(8). A combination of vasospasm and breakdown of endothelial cells (9) couldclose off perfusate access to certain parts of the organ during perfusion andprevent complete equilibration of CPS within the organ.5.1.2. Choice of Vehicle and Cryoprotective SolutionsIn order to introduce a CPS into an organ through the vasculature, avehicle solution must be used. The composition of such solutions has beenintensively studied. Solutions which have been most successful in preservingkidney function when used by themselves are favored since they probablycontribute less tissue damage when used with a CPS. These solutions allcontain “extracellular” concentrations of salts, but the components havechanged dramatically as research illuminates the importance of each component,as the following brief discussion will illustrate (see Table 5.1).The first solutions used to carry CPSs were simple balanced salt148solutions such as physiological saline (0.15 M NaCl). Buffers and glucosewere then added to stabilize pH and provide an osmotic balance (Krebs—Henseleit solution). Collins (10) improved the viability of kidneys byincreasing the potassium content of the solution. This solution (Collins’)has been used successfully in preserving kidneys for 24 hours in clinicalsituations (11). Adenine and glutathione were added to Collins’ in a solutioncalled RPS—2 (12). Adenine can be used by cells to produce AMP and thenceother molecules necessary to metabolic processes, and the addition ofglutathione prevents loss of function in certain renal cells (13) . Recently, anew perfusate was developed for preservation of pancreas (14), whichsubsequently was shown to provide better preservation of other organs,including kidney, than Collins’ solution. This solution, called UW after theUniversity of Wisconsin where it was developed, has also been shown topreserve function of dog (15) and rat (16) kidneys for 48—72 hours.Subsequent studies show that some of the more expensive components of UWsolution can be replaced with more commonly available equivalents (17). UWsolution is widely used in clinical settings for preservation of kidney (18),lung, liver, and multiple organs for transplant. The composition of UWsolution is given in Table 5.1, along with those of RPS—2 and Collins’solutions.Table 5.1: Solute compositions of different vehicle solutions(concentrations in inN)Solute Collins’ RPS-2 tiWK+ 115 28.5 135Na 10 10 30Mg2 28 2 5Cl- 15 34.5 —HC03 10 10 —HP04 58 7.2 25So42- 30 — 5glucose 139 180 —raffinose — — 30mannitol - — 30adenosine - 1 5glutathione — 5 3HES — - 50lacto- — — 100bionateOsmolarity, 320 300 300milliosmol/LAll solutions also contain varying amounts of insulin, heparin, and antibiotics.The CPS that a vehicle solution carries, and its concentration, willalso affect the function of the organ to be preserved. There are documentedinteractions between the vehicle and CPS. For example, DMSO causes loss ofmore potassium and magnesium from kidney cells than a vehicle solution alone(19), but this can be prevented by addition of these solutes to the vehicle.The concentration of CPS is important for two reasons: Cryoprotectivityand toxicity. For DMSO to be cryoprotective, it must be present in thetissues at concentrations of 0.5 to 5.0 M (20) . The more DMSO present, themore protective the solution; however, DMSO has been shown to have toxic1491 50effects at very high ccfncentrations (21). The osmotic imbalances caused byintroduction of high concentrations of DMSO can cause irreversible damage tocells, but this effect can be avoided by introducing DMSO at graduallyincreasing concentrations and adding impermeant solutes (solutes which cannotpenetrate the cell membranes) to the washout solution (22) or by simply usingDMSO concentrations of 20% or less (23).5.1.3. Perfusion TechniquesThere are several perfusion techniques that have been used in varioussituations. The simplest are similar to the technique for washout of blood(described in Section 5.2). The most complex methods require the simulationof physiological pressures. In general, it is necessary to wash all blood outof the organ (24) before the CPS is introduced. Pulsatile perfusion has beenused in the past (25) but has been shown not to be essential for successfulpreservation (26); in fact, constant pressure perfusion seems superior (27).The addition of oxygen to the perfusate is not necessary (28) and couldcontribute to vasospasm (4).The three perfusion techniques investigated in this thesis are:(1) Perfusion using manual pressure control;(2) Perfusion using gravity for pressure control;(3) Perfusion using a syringe pump for pressure control.The first of these methods was used frequently by the transplant researchgroup of Dr. Paul Keown at Jack Bell Research Institute, and by others (17,29, 30) to wash blood out of rat kidneys before storage. It is the methodwith the greatest likelihood of causing vascular damage since manual controlof pressure is difficult. The second method is widely used in cryobiology asa blood washout method (18, 23). The third method is a standard method of151blood washout (6) and/or perfusion of organs, either with a vehicle solutionor with vehicle and CPS (31).5.1.4. Perfusion and C4S NMR ImagingPerfusion of organs with a CP solution is a necessary step incryopreservation. The process inevitably causes damage to the organ,especially the vasculature, and this damage can affect how the CPS isdelivered and distributed in the organ. Homogeneous distribution of CPSinside the organ is necessary for complete preservation, since freezing damageto any part of an organ can contribute to its failure to function aftertransplantation. Thus it is necessary to understand how the perfusiondistributes a CPS, and what potential problems there are with CPS distributionthrough the vasculature with various perfusion techniques. Since vasculardamage can occur in the blood washout process, it is also important toinvestigate how the blood washout technique affects the later perfusion anddistribution of CPS.NMR C4S imaging is the only available technique whereby thedistributions of DMSO and water can be monitored during a perfusionexperiment. This chapter will describe a set of NMR C4S imaging experimentsdesigned to assess standard blood washout methods and techniques forcryoprotectant perfusion. This information will be discussed in relation tothe continuing failure of these techniques to preserve organs.1525.2. Experimental5.2.1. Surgical Procedures and Blood Washout TechniquesAll surgery and preparation of rat organs was done by Betty Pearson inthe research laboratories of Dr. Paul Keown, at the Jack Bell Research Centerin Vancouver. Rat organs were then transported in an ice-filled cooler to theUniversity of British Columbia. This “cold ischemia” time was between 30minutes and one hour in all cases.Adult male and female Lewis rats, mean weight 450g and 220g,respectively, were obtained from Charles River Laboratories, Montreal, Quebec.They were housed in a modern rodent facility and cared for in accordance withthe Canadian Council of Animal Care Ethics on Animal Experimentationguidelines. The rats were housed in groups of four in large polycarbonatecages containing corncob bedding. A standard laboratory diet of Purina ratchow and fresh water were freely available.All experiments were performed on kidneys obtained from anaesthetizedrats. An asceptic surgical technique was employed, utilizing an operatingmicroscope and microsurgical instruments. During surgery, body temperaturewas maintained by means of a water-filled heating pad.Two 50 ml beakers padded with gauze were filled with saline solution, toreceive the flushed kidneys and were kept at 4°C.Rats subjected to surgical techniques 1 and 2 below were anaesthetized(and subsequently euthanized) with a mixture of halothane and oxygen deliveredvia an inhalation anaesthetic machine. The dose was 4% during induction, 3%during incision and manipulation of abdominal contents, and 2% throughout theremainder of the procedure. The “toe-pinch” method of assessing depth ofanaesthesia was employed at frequent intervals, and the dose adjustedaccordingly. Rats subjected to surgical technique 3 below were anaesthetizedby an IP injection of sodium pentobarbital in a dose calculated to euthanize153the animal within minutes (100 mg/bOg weight).Three different techniques were used for removing blood from the kidney(“blood washout”):(1) The abdominal area was shaved and the rat placed supine on theheating pad. A midline abdominal incision was made from thesternum to the groin. The abdominal contents were gentlyretracted and kept warm throughout the experiment. A 5 ml syringefitted with a 16 gauge plastic cannula was filled with cold,heparinized saline (250 U.S.P. units per ml) in preparation forsubsequent flushing of kidneys. The aorta and vena cava weredissected between the liver and the bifurcation to the lowerlimbs, and all branches tied or coagulated. The renal arteriesand veins were carefully dissected, and the suprarenal vesselstied and cut. The kidneys were mobilized from surrounding fat.Three ligatures (7.0 silk) were loosely placed around theaorta. The first was positioned proximal to all the renalvessels, the second distally, near the groin, and the third midwaybetween them to secure the cannula for flushing. The first andsecond ligatures were tied and a small incision was made in theaorta just proximal to the second ligature. The cannula wasinserted and secured, care being taken that the tip did not extendas far as the left renal artery. The distal vena cava was cut topermit drainage. Four ml cold saline were delivered manuallythrough the syringe, at an approximate rate of 1 ml/min in orderto flush blood from the kidney.The left ureter was transected, and the kidney, renal arteryand renal vein were transferred to a cold saline bath. Apreviously prepared catheter was tied into the renal artery inpreparation for subsequent perfusion. The catheter was a lengthof polyethylene tubing (0.28 mm ID) fitted onto a 30 gauge needle,154connected to a 5.0 ml syringe filled with saline. The catheterwas carefully filled with saline and tiedinto the renal arteryunder fluid to ensure that no air was introduced to the kidney.(2) An identical surgical procedure was employed up to the point ofinserting the cannula into the aorta. Instead of a plasticcannula attached to a hand—held syringe, a 16 gauge steel cannulaconnected to a reservoir (drip set, see Figure 5.1(B)) of coldsaline was used. The cannula was inserted into the aorta andclamped in position with haemostats. The distal vena cava was cutto allow drainage of kidney. 20 ml of cold, heparinized salinesolution was allowed to flow through the kidney at a pressure of100 cm water. Subsequent removal and catheterization of thekidney was performed as for procedure 1 above.(3) The thoracic and abdominal areas were shaved, and, as soon as thetoe-pinch withdrawal and blinking reflexes were abolished, a fulllength ventral incision was made and the thoracic cavity wasopened. A 14 gauge steel cannula was inserted into the leftventricle of the heart while the heart was still pumping, and theright atrium was cut to permit drainage. Heparinized saline (4°C)was perfused via the cannula at a pressure of 100 cm water, insufficient quantity (about 250 ml) to ensure that all blood isflushed from the kidneys. The renal vessels were dissected, theureter transected, and the vessels and kidney transferred as inprocedure 1.The kidneys were placed in cold saline for subsequent cold transport,storage, and treatment with DMSO, if NMR imaging was to be done during theperfusion. Storage was never longer than 24 hours to ensure tissue viability.1555.2.2. Perfusion TechniquesThree techniques were used for the perfusion of DMSO and vehiclesolution into rat kidney. In all cases, the 5 ml syringe filled with salineused for the initial blood washout was removed from the microtubing catheter,and the needle was attached to another syringe filled with CP solution. Carewas taken not to introduce air bubbles into the microtubing when the syringeswere exchanged.(1) The catheter (0.2 m in length) was connected to a 5 ml syringefilled with CP solution. The solution was perfused through thekidney manually. (Figure 5.1(A).)(2) The catheter (2.5 m length) was connected to a 50 ml syringe filledwith CP solution. The syringe was hung at a height of 100 cmabove the height of the kidney for perfusion by gravity. (Figure5.1(B).)(3) The catheter (2.5 m length) was connected to a 20 ml syringe filledwith CP solution. The syringe was inserted into a syringe pumpwhich was set to deliver the solution. This setup is not constantvolume delivery but similar to constant pressure, as will beexplained in Section 5.3.4. (Figure 5.1(C).)In all of the perfusion arrangements, the kidney was placed carefully ina 15 mm NMR tube as shown in Figure 5.2. The catheter was previously threadedthrough the cap of the NMR tube. (New catheter microtubing was used for eachperfusion.) A length of Teflon microtubing was inserted into the tube to alevel 2—3 mm above the top of the kidney. This tubing was attached to avacuum source via a Teflon fitting, as shown. Teflon was used to avoidbreakdown of the tubing over several perfusions due to contact with DMSO. Itwas necessary to remove perfusate as it drained out of the kidney, to avoidoverflow into the RF coil, and to avoid the necessity of altering the room156temperature shim settings during a perfusion experiment. The details ofspecific perfusion experiments are listed in Table 5.2.Washout of the DMSO in the kidneys was attempted for several samples. Insome cases, the kidneys perfused with method 3 above were frozen at -20°Cafter perfusion with 20% DMSO in UW solution for various amounts of time. Inother cases, the washout was attempted immediately after the perfusion with CPsolution. Frozen kidneys were thawed to 10°C in the NMR probe and imaged toascertain the distribution of DMSO. In all the washout experiments, thesyringe containing the CP solution was replaced with a 20 ml syringe filledwith a washout solution. The syringe was placed in the syringe pump and thesolution flushed through the kidney. As indicated above, this technique ismore like a constant pressure delivery than a constant volume delivery (seeSection 5.3.4). Specific experiments are listed in Table 5.3.157Figure 5.1. Schematics of the three different perfusion procedures used inthe experiments in this thesis. (A) Manually controlled perfusion with asyringe. (B) Gravity-fed perfusion, pressure equal to 100 cm water. (C)Syringe pump perfusion, highest pressure equal to 40 mm Hg.A BaspirationLiReservoir(saline) —•aspiration15mmNMRtube/kidney/kidneyC aspirationfkidney15 mmNMR tubecatheter15mmNMRtubesyringe pumpwith 20 mlsyringe158Figure 5.2. Rat kidney in a 15 mm NMR tube. Renal artery is catheterizedwith 0.28 — I.D. polyethylene microtubing, which is threaded through the capof the NR tube and anchored to the tube with tape. Teflon microtubing (1 mmI.D.) is used as an aspiration line to remove excess solution as it isperfused through the kidney.1595.2.3. Imaging of Rat KidneySince NMR C4S imaging has not been used before in the investigation ofperfusion of CPSs, it was desirable to establish what the image data shouldlook like in circumstances that are common in transplant surgery: Images ofwater in rat kidney were taken without subjecting the kidney to any bloodwashout treatment or heparinization, T1 and T2 contrast in water images ofrat kidney were investigated, and images of water in rat kidney were takenafter different blood washout procedures. Imaging parameters are given inAppendix A.Images were also taken during experiments in which rat kidneys wereperfused with DMSO in a CP solution, with different perfusion techniques. Thelist of perfusion experiments and their descriptions in Table 5.2 also refersto the imaging parameters for each experiment by the figure number in AppendixA.Images of freezing of rat kidney perfused with DMSO in a CP solutionwere taken in one experiment (Experiment 7). Parameters are listed inAppendix A.Washout of DMSO from kidneys immediately after perfusion with the CPsolution or after freezing and thawing was investigated in several imagingexperiments. Table 5.2 lists these experiments and refers to the imagingparameters for each experiment which are recorded in Appendix A.In all imaging experiments, the NMR tube containing the kidney wasremoved from cold storage (ice) and inserted into the magnet after carefulpositioning of the tube to assure an image slice position near to the centerof the sample. The NMR probe had been previously equilibrated to theappropriate temperature as described in Chapter 2. The sample wasequilibrated for 10-15 minutes before imaging, during which time the roomtemperature shims were optimized.In experiments involving manual syringe perfusion of a CP solution, the160sample had to be removed from the magnet each time an aliquot (3 ml) ofsolution was perfused through the kidney. Care was taken not to change thelevel of the tube in the holder or move the kidney in the tube, either ofwhich would change the position of the image slice through the sample. In oneexperiment (#8) the microtubing catheter came loose from its anchor on the NMRtube (see Fig. 5.2), and so in this case the position of the slice in thekidney in each image is slightly different.In experiments using gravity fed or syringe pump perfusion, the samplewas not removed from the magnet during the course of the perfusion. Thus allimages can be assumed to correspond to exactly the same position in the samplein all these experiments.161Table 5.2List of parameters of imaging experiments monitoring perfusion.Expt. Blood Perfusion Temp. Vehicle %DMSO total ml total Perf. Fig.4 washout technique °C perfusate perfusion rate 4techn. delivered time ml/hr1 N/A N/A RT N/A N/A N/A N/A N/A 5.32 syringe N/A 4 N/A N/A N/A N/A N/A 5.43a syringe N/A RT N/A N/A N/A N/A N/A 5.53b gravity N/A RT N/A N/A N/A N/A N/A 5.64 syringe syringe RT saline 40% 10 ml N/A N/A 5.8.B5 syringe syringe RT saline 50% 21 ml 4 hr. N/A 5.8.A6 syringe syringe RT saline 50% 9 ml 1 hr. N/A 5.8.C7 syringe syringe 5 saline 50% 6 ml 1 hr. N/A 5.8.D8 syringe syringe 5 saline 20% 17 ml 1.5 hr. N/A 5.99 syringe gravity RT saline 40% 12 ml 15 hr. 0.8 5.1010 gravity N/A 8 N/A N/A N/A N/A N/A 5.1120 ml11 gravity gravity RT saline 20% 11 ml 8 hr. 1.4 5.1320 ml12 gravity gravity 10 saline 20% 10 ml 18 hr. 0.6 5.142Oml13 gravity gravity RT saline 20% 6 ml 4 hr. 1.5 5.17250 ml14 gravity syringe 10 UW 20% 10 ml 9 hr. 1.1 5.21250 ml pump16 gravity syringe RT UW 20% 18 ml ? 5.12250 ml pump17 gravity syringe RT UW 20% 21 ml 11 hr. 1.9 5.20250 ml pump19 gravity syringe 10 UW 20% 10 ml 18 hr. 0.5 5.23250 ml pump24 gravity syringe 34 saline 20% -— —- —— -—250 ml pump25 gravity syringe 36 UW 20% 9 ml 5 hr. 1.8 5.25250 ml pump26 gravity syringe 36 UW 0% —— —— —— 5.24250 ml pumpC’)C’)C’)I—iC’)-C—3GW EC,’ x,-t•’t;x,‘--jCDCDC‘-3cC3‘-3?I-I-’C-I-C’)I-)—‘CCCCI-CI—,)-I.-I-l-I—jI-0I-I-I-,.)).I•_)•-S..OttoC’U’-•I-.’wi-I-’HI-P0I0-0.0.--0.•-.•—.DCCD0)CDCD1<<<HC,CaE(•t•0)I-,I-I-I-.-‘I-o00000C,5‘00.CtH Co (t 0 I-tI 0) II 0) CD CD I-I Cl) 0 I-I, I-, 0)0)I-’-H 0-ICD’CD III-..I-’t-00C’)I-•I-’I-I-•I-,C’-)0 I-CD to CC’ P DiCI—’CtCI01’.)0t-tO••••I•••—..CtCtC)’tOIattoi.,,C,0H0 CC’CD CC) U (i-I 0 II CDCa,’1<•<i<<1<i<CDCDCDCDC,CC’CD00COCOCOtototom<top’to.0 0) H_ny,71U’(7’01(7’(710)tOat-.30C’)C’)at713‘0 *I—iC.’I’.)1635.2.4. 31P NMR Spectroscopy31P NMR spectroscopy was performed on a Varian 300 NMR spectrometer. Akidney was excised after a 250 ml blood washout (Procedure 3), then insertedinto a 10 mm NMR tube and placed on ice. Transport time was 35 minutes, aswith other samples used for perfusion studies. Spectra were acquired starting35 minutes after excision. Each spectrum represents 100 scans with arepetition time of 8 seconds. Sweep width was 60 ppm, with concentrated H3P04as the standard at 0 ppm. Temperature was kept at 0°C.5.3. RESULTS AND DISCUSSION5.3.1. Preliminary Imaging Studies of Rat KidneysContrast in NMR images depends on the relaxation characteristics andconcentrations of nuclei in the tissue at the molecular level, but theappearance is often so similar to the contrast seen with the eye that it ispossible to mistakenly attribute features in NMR images to physicalstructures. Although the contrast in the NMR image is usually related totissue structure, care must be taken when interpreting changes in imagesduring a process like perfusion. Since NMR C4S imaging has not been usedbefore in the investigation of perfusion of CPSs, it was necessary toestablish what the images should look like in circumstances that are common intransplant surgery.Figure 5.3 shows two NMR images, obtained with a spin—echo pulsesequence: (1) a rat kidney which has not been subjected to blood washout(coronal section, Figure 5.3(D)), and (2) a kidney subjected to a gravity-fedblood washout (see Section 5.3.3) (transverse section, Figure 5.3(E)). The164coronal image has very good contrast, and many structures can be identifieddue to the different values of NMR relaxation parameters and tissue watercontent. The image was taken in a plane just above the renal pelvis, so thisstructure is not seen. Comparison of the two images with the diagram ofkidney structure indicates that the renal cortex, medulla, and renal pelviscan be easily distinguished. Smaller structures such as arcuate bloodvessels, and groups of tubules, afferent and efferent arterioles, and nephronscan be discriminated. The latter four structures are so closely situated thatit has not been determined which of them is seen as dark streaks in theimages. However, the dark structures are not seen unless blood has not beencompletely removed from the kidney, so probably they correspond to thearterioles from which it would be most difficult to remove blood.Limitations on the repetition time of the NMR C4S pulse sequence used tomonitor diffusion and perfusion in rat kidneys was discussed in Chapter 4.The TR used in most experiments in this thesis is 1 second. TE is limited toits smallest possible value if semiquantitative data are to be obtained.However, knowledge of the kinds of contrast in rat kidney may allowconclusions to be drawn from contrast changes during an experiment. Figure5.4 shows a series of images of water in rat kidney held at 5°C, obtained withthe STE spin-warp pulse sequence, which allows smaller TE to be used. The TRwas 0.5 s and TE was increased from 5 to 80 ms in images A-E. The last imageshown is a spin-density contrasted (“quantitative”) image, which shows thatthere is little variation in the concentration of water in the kidney tissues.T1 contrasting of the images by holding TE at 16 ms and increasing TR did riotchange the image contrast at all.The first step in preparation of a kidney for transplant is the bloodwashout. The first technique investigated was manual flushing with a syringethrough a catheter in the renal artery. This technique has been usedextensively by Dr. Paul Keown’s transplant research group at Jack BellResearch Institute, and excellent survival of grafted kidneys was obtained165(32); other researchers commonly use this method as well (33, 34, 35). Thetechnique is thus assumed to cause minimal damage. Figure 5.5 shows images ofthe water in a rat kidney at different slice positions after washout with theabove technique. Visual inspection of this kidney (and many others infollowing experiments) revealed that the dark areas seen in the imagecorrespond to coagulated blood (36).The second technique used for blood washout of kidneys is the gravity-fed flush described in Section 5.2.1 (procedure 2). This technique is used bymany cryobiologists working with rat or rabbit kidneys. Figure 5.6 shows aseries of images at different slice positions of a rat kidney subjected to theabove blood washout technique. It is somewhat difficult to compare Figures5.5 and 5.6, since the resolution is not the same. However, it is obviousthat there are fewer of the dark areas indicative of coagulated blood afterthe gravity-fed blood washout.A decrease in T2 could occur if blood was not completely flushed out ofthe kidney but has coagulated in places despite heparinization. Coagulationof blood excludes water, and this solidification would cause the T2 ofremaining water protons to be much shorter. The literature does describe thepresence of blood cells forced outside the vasculature during perfusion (37),and dissection of rat kidneys after some of the experiments in this thesis hasshown areas in which blood vessels are distended and blood has collected. Inaddition, handling the organ during surgery could cause vascular spasm leadingto incomplete flushing and subsequent blood clots.166efferentarteriolearcuateveinarcuatearteryafferentarterioleFigure 5.3. (A) Schematic representation of the important structurai featuresof the kidney. (B) Enlargement showing the geometry of the microscopicarchitecture. Groups of smaller structures including arteries, veins, andarterioles are visible in the images. Arrows show the direction of bloodflow. (C) Frame of reference for kidney slices. (D) A coronal section (zyplane) of a rat kidney which has not been subjected to blood washout. Theimage slice is just above the renal pelvis so this structure is not seen. CE)A transverse section (xy plane) of a kidney which has been subjected to agravity—fed blood washout. In both images the cortex can easily bedistinguished from the medulla. The renal pelvis is shown in the transverseimage.A Cz.‘1xB DE167A DB EC FFigure 5.4. A series of transverse images of rat kidney held at 5°C, obtainedwith the STE spin—warp imaging sequence. TR was 0.5 s, and TE was increasedfrom 5 to 80 ms in images A—E to demonstrate T2 contrast. Image F is a spindensity contrasted image (TE = 5 ms, TR = 12 s) . Holding TE at 16 ms andincreasing TR (T1 contrast) had no effect on image contrast.168ADTE 16 msTR = 1 sFigure 5.5. Transverse images of a rat kidney subjected to manuallycontrolled syringe blood washout. The dark areas correspond to the arcuateblood vessels (best seen in A and C), and the afferent (in the cortex) andefferent (in the medulla) arterioles, which still contain blood despite theaddition of heparin to the washout solution.C169TE = 16 msTR = 1 sFigure 5.6. Transverse images through a kidney subjected to a gravity—fedblood washout procedure. This procedure is more effective at blood removalthan that demonstrated in the previous figure. However, small localized bloodclots are visible in the renal pelvis (images A and 0) and between the cortexand medulla (image B).1705.3.2. Perfusion by SyringeThe simplest way to introduce a CP solution is to use the same method aswas used to wash the blood out of the organ. The transplant research group atJack Bell Research Center manually flushes blood out of rat kidneys using a 5ml syringe filled with cold heparinized saline, as described in Section 5.2.1.Figure 5.5 indicates that this washout technique does not remove all the bloodfrom the vasculature. Introduction of a CP solution by this same method wouldthen be expected to cause further damage, since blood vessels may be fully orpartially blocked by clots. If blood is not completely removed from thekidney, it is probable that introduction of the CP solution would not lead tocomplete equilibration of the CPS in the tissues, since access through partsof the vasculature would be blocked.Figure 5.7 shows a series of images of DMSO in a rat kidney which wassubjected to perfusion with 50% DMSO in saline using the same manuallycontrolled syringe technique as used for blood washout (experiment #6). Theseries of image pairs correspond to successive perfusions of aliquots of CPsolution as indicated. The images show that DMSO is not evenly distributed inthe kidney, even after 21 ml of solution have been flushed through thevasculature. Dark areas in the images show where no DMSO is present and seemto correspond spatially to the arcuate blood vessels. This method ofperfusion with CP solution was attempted four times (Experiments 4-7), andeach time the vascular resistance (VR) ixmnediately became extremely high, tothe point that it was very difficult to exert enough pressure to forcesolution through the kidney. This could be due to expected VR increases withhigh concentrations of DMSO (22) but more probably is due to blockage of thevasculature.Figure 5.8 shows DMSO and water images from kidneys subjected to manualsyringe perfusion with 40 or 50% DMSO in saline after manual washout bysyringe (Experiments 4-7). The first DMSO image (Figure 5.8(A)), from171Experiment 5, shows the ‘no—reflow” phenomenon: Very little perfusion throughthe renal artery is possible after the blood washout and short cold storage.Similarly, the water image from Experiment 6 (Figure 5.8(C)) shows dark areascaused by blood washout, and the DMSO image taken after perfusion with CPsolution shows dark areas whose locations correspond to blood vessels. Imagesfrom Experiment 7 (Figure 5.8(B)) show even more blood clotting.When the same experiment was done without cold storage of the organbetween washout of blood and perfusion with 40% DMSO (Experiment *4), not asmany dark areas are seen (Figure 5.8(D)). The dark areas seen in this imageare all in the areas where arcuate vessels should appear but do not extendinto the cortex as much as they do in the other images. After 10 ml of the CPsolution was perfused, the DMSO concentration is still not as high in themedulla as the cortex. It is reasonable to assume this is because the medullais less vascularized (2). Perfusion without initial cold storage might beless damaging because cold storage can cause vascular spasm if kidneys are notsufficiently cooled by washout with cold saline (6) . This could also be asituation where blood clots were not completely blocking the vessels, allowinga certain amount of perfusion to take place. Thus, the cold storage may notbe responsible for failure of the perfusion.In order to test the hypothesis that the increased vascular resistance(VR) is due to the high concentration of DMSO in the CP solution, theexperiment was repeated with a CP solution of 20% DMSO in saline (Experiment*8). The vascular resistance was initially much lower than with the 50%concentration of DMSO, but after 3 ml were perfused, it increased greatly.Dark areas are seen in the water images after perfusion with each 3 ml aliquotof CP solution (Figure 5.9), but in different places. The explanation is thatthe dark areas seen are in slightly different planes of the sample, because inthis experiment the microtubing came loose from its anchor on the NMR tube, sothe kidney may have moved during the perfusion.Certainly the images of Figure 5.9 show blood clots in the arcuate blood172vessels as in the previous experiments. Proceeding with a perfusion despitethe presence of clots may lead to damaged blood vessels, and the damage wouldprobably be progressive with perfusion of each aliquot of CP solution. Aprogression of changes is noticeable in Figures 5.7 and 5.9: The dark areascorresponding spatially to the arcuate blood vessels become larger duringperfusion and extend into the renal cortex. Thus blood clots, or cellsdislodged from them by perfusion with the CP solution, may be moving throughthe vasculature of the kidney and causing further blockage problems.It is clear from the above experiments that manual washout and/orperfusion with a syringe causes changes in the NMR images of rat kidney, whichprobably correspond to blockage of the blood vessels by clots and ensuingstructural changes in the blood vessels. These changes seem to be associatedwith high vascular resistance independent of DMSO concentration in theperfusate. Histological examination done by Dr. John English of VancouverHospital and Health Sciences Center shows that the veins have becomedistended, “probably due to high perfusion pressures” (36). Since theperfusion pressure was deliberately kept low, this damage is probably relatedto blood clots blocking some of the vessels.Similar ballooning and subsequent collapse of small blood vessels wasnoted by Skaaring (39) after perfusion of dog kidneys with an extracellularsolution for 24 hours, using a constant pressure perfusion method whichusually allows good function of kidneys after transplant. It seems reasonablethat, if such damage occurs even when perfusion pressures are carefullymonitored, it is likely that more serious damage to the vasculature wouldresult from pressures exerted manually via syringe.173Water DMSOABC0TE = 16 msTR = 500 msFigure 5.7. Series of water and DMSO images of a rat kidney subjected tomanually—controlled syringe perfusion with 50% DMSO in saline. Each set ofimages was obtained after perfusion of an aliquot of CP solution (volume shownat left of each image pair) . Images show that there are blood clots left inthe kidney (water image A), and that DMSO does not become homogeneouslydistributed.174Water DMSOABCDTE = 16 ms(A-C)TR = 1 S(D) TR = 500 msFigure 5.8. DMSO and water images from kidneys subjected to manually-controlled syringe perfusion with 40 or 50% DMSO in saline after manualwashout by syringe. The first DMSO image (A) shows the “no—reflow”phenomenon, in which no perfusion of DMSO occurs. Images B and C show largeblood clots in the arcuate arteries and arterioles which correspond toincomplete DMSO distribution. Image D shows that avoiding cold storage afterblood washout may improve perfusion, but this might also be due to somewhatmore thorough blood removal in this sample.175ACTE = 5.1 msTR = 500 msFigure 5.9. Images of water protons during a perfusion with 20% DMSO insaline after manually—controlled blood washout. Images A and D show a largeamount of blood still left in the arcuate arteries, and B and C show bloodclots in other blood vessels. The images do not correspond to the same placein the kidney. Each image was obtained after perfusion with an aliquot of CPsolution (volume shown below each image)B01765.3.3. Perfusion by GravityThe second perfusion technique investigated was gravity—fed perfusion,as described in Section 5.2.2. This technique is commonly used in cryobiologyas a blood washout technique (18, 23).5.3.3.1. Comparison of Syringe and Gravity Blood Washout of KidneyAs a comparison to experiment #4, in which the manual blood washout andperfusion produced the fewest changes in the NMR images, experiment #9 wascarried out (Figure 5.10). The only modification of the procedure was thesubstitution of gravity-fed perfusion for manual syringe perfusion.Comparison of the last DMSO images in Figures 5.8 and 5.10 show that thegravity perfusion was successful in delivering DMSO to the kidney tissues eventhough the gravity-fed perfusion was performed on a kidney which had been coldstored for several hours. Cold storage per Se, then, probably did notexacerbate blood clotting and cause the partial failure of perfusion in theprevious experiments in which manual blood washout was done. There are fewerareas affected by blood clots in this experiment. These experiments suggestthat vascular damage due to high pressures may occur when perfusion of a CPsolution containing DMSO is attempted manually after incomplete blood removal.This is probably not dependent on the concentration of DMSO, according toexperiments #4 (40% DMSO) and #8 (20% DMSO), which show similar damage. DMSOis expected to increase vascular resistance, but the low perfusion rate inexperiment #9 (0.8 ml/hr) is probably due to partial blockage of the bloodvessels.177TE = 16 maTR = 1 5Water DMSO2 nL120 miii5 mL420 mm8 mL540 fIllJ 105 muL780 mmmlFigure 5.10. A series of images of water and DMSO protons in rat kidneyduring a gravity-fed perfusion with 40% DMSO in saline. Although blood waswashed out with the manually-controlled syringe method, it seems to have lefta smaller amount of blood in the kidney than in previous experiments, mostlyin the arterioles of the cortex. These small blockages seem to cause a lowperfusion rate (0.8 mi/hour).1785.3.3.2. Study of Gravity-fed Blood Washoutsand New Surgical ProcedureSince the previous experiments show that manually-controlled bloodwashout does not completely remove blood and manually-controlled perfusion maycause further vascular damage, the blood washout technique was also changed toa gravity—fed washout. In experiment #10, the washout technique was changedso that the vasculature of kidneys was flushed with saline at a pressure of100 cm, as described in Section 5.2.1, washout procedure 2. This gravity-fedblood washout was used to flush a kidney so NMR images of different parts ofthe kidney could be taken and examined for the kind of changes seen withsyringe perfusion. Figure 5.11 shows NMR images of water in slices throughthe kidney at different levels. There is very little contrast in theseimages, but some very small dark areas possibly corresponding to blood clotscan be seen in the medullar region near the center of the kidney.The kidney was then dissected and sketched, attempting to match thephysical slices with those in the images. The small round dark areas are mostlikely blood vessels running perpendicular to the image plane. Some of theless well defined dark areas seemed to correspond to areas in the kidney thatcontained diffuse blood. This may be due to a small amount of blood left inthe efferent vessels by incomplete washout (36) or blood forced out of thevasculature by high perfusion pressures (37).These data show that gravity fed blood washout seems to remove most ofthe blood left in the vasculature after manually—controlled syringe bloodwashout, but the blood does not seem to be completely removed after 20 ml ofheparinized saline have been flushed through the kidney. D’Silva et al.comment in their 1990 paper that a previously described method which requiredflushing 40—50 ml of saline through rat kidneys to remove blood (40) isunnecessary and used only 3—4 ml in their study. However, experiment #10179indicates that some blood may still be left in the vasculature after a 20 mlsaline washout, so it is reasonable to use a larger volume. It was decidedthat a larger volume of saline would be used in subsequent experiments.Another modification of the procedure for washout was suggested by Dr.John English of Vancouver Hospital and Health Sciences Center. In clinicalsituations, the donor is dead when organs are removed; the procedure forwashing blood out of organs before collection is similar to the washoutprocedure 3, Section 5.2.1. The method was modified to the rat by BettyPearson, microsurgical specialist at Jack Bell Research Center, Vancouver. Alarge volume of saline for washout of the kidneys can be delivered in thisfashion. The heparinized saline passes through the whole vasculature of therat before exiting the atrium. It was expected that 250 ml flowing throughthe vasculature of the whole animal should flush the kidneys with a largevolume of saline.This technique for gravity-fed blood washout was used in experiment *16.Figure 5.12 shows water images of different slice positions in the kidney.These images should be compared to those in Figure 5.11, where a low-volumegravity washout of the kidney itself was used. There is no indication in theimages of residual blood in the medulla. This technique was by far the mostsuccessful washout technique and so was used in all subsequent experiments.(Note: When this method was used to remove blood from a kidney, a manuallycomntrolled perfusion was successful in delivering DMSO to the tissues. Nodark areas corresponding to residual blood were seen, and the DMSOdistribution proceeeded as seen with gravity and syringe pump perfusions.This proves that the failure of manual perfusion seen in this section is dueto residual blood blocking the vasculature.)TE = 16 msTR = 1 SFigure 5.11. Images of water protons in a series of slices from top to bottomthrough a rat kidney subjected to a gravity—fed blood washout. Dark areas attops of the images are fat. Only very small dark areas are visible, whichseemed to correspond to cross-sections of small blood vessels upon dissection.Less well—defined dark areas such as the medullar region in image D wereassociated with diffuse blood in efferent vessels. This technique seems muchmore effective at blood removal than the manually-controlled syringe method.180DACEG181ABCTE = 16 msTR = 1 sFigure 5.12. Images of water protons in a series of slices from top to bottomthrough a rat kidney subjected to a whole-body blood washout procedure. Darkareas at right in each image are fat. There is a noticeable lack of contrastin these images, due to the total removal of the blood. Subsequenthistological examination showed that almost no blood cells remained in theorgan. The image brightness is low due to limitations of the DISMSL program(which performs the Fourier transform)E1825.3.3.3. Study of Gravity—fed Perfusion at Two TemperaturesIn order to keep the number of variables in the perfusion experiments toa minimum, it was desirable to use a single concentration of DMSO in the CPsolution. Karow et al. (23) have shown that dog kidneys function well afterperfusion with 2,8 M (20%) DMSO, but higher concentrations which wereintroduced caused damage which resulted in the failure of the organ aftertransplantation. Thus 20% DMSO was chosen for use in the rest of theperfusion experiments. This concentration is high enough to protect tissuesfrom freezing damage (20).Experiments #11 and 12 used a 250 ml saline blood washout as describedin Section 5.2.1, and a gravity—fed perfusion of 20% DMSO in saline (setupshown in Figure 5.1(B)). Figure 5.13 shows pairs of NMR images (DMSO andwater) taken at various times throughout the perfusion process at roomtemperature (22°C). The water images have quite different features thanimages of previous rat kidneys. The cortex appears very dark, and the medullais featureless. The DMSO images show that the DMSO is equilibrated in themedulla but not the cortex, which is the opposite result to that of previousexperiments and also of predictions based on the lesser vascularization of themedulla. Figure 5.14 shows an image of DMSO with no T1 contrast (which shouldbe semiquantitative) after 11 ml of CP solution had been perfused through thekidney. This image also shows some features in the medulla, so they may notbe visible in the perfusion images of Figure 5.12 due to poor resolution. Forsome reason the blood was not removed from the cortical tissue of thisparticular kidney. This image slice does not contain the arcuate bloodvessels, so it is unknown whether they contain blood clots as well.Figure 5.15 shows water and DMSO images from experiment *12, a perfusionat 10°C. In this experiment the perfusion was so slow that the DMSOconcentration was not allowed to reach equilibrium. However, in the early183stages of perfusion shown, DMSO is equilibrating more quickly in the medullathan the cortex, exactly as in experiment #11. The water images show thecortex as a much darker area, due to the presence of blood clots (also as inthe previous figure) and there are some features visible in the medulla.There are some possible explanations for the appearance of the twokidneys in Figures 5.13-5.15. The two donor rats were sisters, and thereforegenetically very similar. When structural defects in rat kidney were studied,it was found that they tend to have very high rates (up to. 70 or 80%) ofgenetic kidney defects (32). Highly inbred rats like the Lewis strain wouldbe expected to have more defects. Obviously these are not life—threateningdefects, but some may be more serious than others, and perhaps in this casethe vasculature of the rats was affected enough to cause slower perfusion tothe cortex. This would cause the blood washout to be less effective. Theother difference between these rats and the rats in previous experiments isthat these rats were younger and female. It is not known whether this affectsthe appearance and perfusion characteristics of rat kidneys. However, youngfemale rats were used in some subsequent experiments, and these odd featureswere not seen again.Despite the probably abnormal condition of the kidneys used inexperiments *11 and 12, it is clear that decreasing the temperature ofperfusion also decreases the rate of perfusion, when the perfusion pressure iskept the same. This could be due to increased vascular resistance uponchilling of the kidney, and/or the increased viscosity of the CP solution atlow temperatures. The perfusion rate at room temperature (22°C) wasapproximately 1.6 ml/hr, compared to 0.6 ml/hr at 10°C. This emphasizes theimportance of perfusing at higher temperatures to enhance the equilibration ofDMSO in the kidney tissues (23).184Water DMSOAOmio mmB2m175 mm6m225 mmilmi410 mmTE = 16 msTR = 1 sFigure 5.13. Pairs of images of water and DMSO protons in rat kidneysubjected to whole-body blood washout, during a gravity-fed perfusion with 20%DMSO in saline at room temperature. Perfusion volumes and times are givennext to each image pair. The very dark appearance of the cortex was unique tothis rat and its sister.C185TE = 16 msTR = 15 sFigure 5.14. An image of DMSO protons in the same kidney as the previousfigure, with no T1 contrast, after perfusion of 11 ml of CP solution. Thisimage indicates that the blood was not removed from the cortical tissues.186Water DMSOAOmio mEnB4m1400 mmC5m1500 mmTE = 16 msTR = 1 sFigure 5.15. Images of water and DMSO protons in rat kidney subjected to awhole—body blood washout and gravity-fed perfusion with 20% DMSO in saline at10°C. Perfusion volumes and times are given next to each image pair. Thisimage also indicates that the blood was not completely removed from thecortical tissues, and there are blood clots in the efferent arterioles (waterimages A and B). DMSO is well distributed to the medulla.1875.3.4. Perfusion by Syringe Pump5.3.4.1. Development of a Method for theControlled Introduction and Removal of DMSOThe last perfusion technique studied is a modification of the perfusioncircuit commonly used in cryobiology. A syringe pump with a 20 ml syringereplaced the perfusate reservoir in the gravity-fed experiments discussedabove. This pump was set up with a weighted arm (Figure 5.16) which releasedpressure when it built up to the point that it lifted the weight. During aperfusion, if the back pressure from the vascular resistance of the kidney istoo high to allow the CP solution to flow at the set rate, the pressure willrelease periodically. This will cause the flow to be “pulsatile” in nature,with fairly low pressure, instead of the constant delivery rate expected of asyringe pump. The pressures usually used in perfusion circuits for rat kidneyare 60 mm Hg or less; in the present work, the weight of the arm gives amaximum pressure of approximately 40 mm Hg.A pilot experiment (*13) was performed to find out whether thisperfusion circuit could introduce and remove DMSO from a kidney, how quicklysuch a perfusion could be performed, and whether changes in the vasculaturewould occur. The pump was set so that the release of pressure was infrequent,about once per minute, to imitate a constant pressure situation.Images during the perfusion of a kidney at 22°C with 20% DMSO in salineusing the syringe pump (experiment #13) are shown in Figure 5.17. The firstimage, taken before perfusion with the CP solution, shows that there is noresidual blood left by the large volume (250 ml) gravity—fed blood washout.In the next set of images, DMSO can be seen perfusing through the renal artery(arrow) into the cortex. The kidney has swollen considerably, but there donot seem to be any changes in the vasculature. In the third DMSO image (after2.1 ml of CP sol’ution have been perfused) there is a large amount of signal188from an area which likely corresponds to an arcuate blood vessel (arrow).This confirms that the flow rate is low enough that the distribution of DMSOshown in the images is accurate.The last set of images show that DMSO has been distributed into both themedulla and cortex. At this point the experiment was stopped, but from theimage data, 6 ml of 20% DMSO in saline is not enough perfusate to ensurecompletely homogeneous distribution of DMSO throughout the kidney.This perfusion technique is also used to remove CPSs from organs, so inthe next experiment (#13w) the CP solution was replaced by saline. Imagesduring this perfusion are shown in Figure 5.18. Comparison of the DMSO imagesat 2.1 ml of 20% DMSO CP solution and 0.25 ml saline perfusate (Figures5.17(C) and 5.18(A)) indicates that the area of the cortex in the left of theimage is perfusing more quickly in the area associated with the arcuate bloodvessels. The similarities between these two images emphasizes that theperfusion patterns depend on the vascular structure of the particular kidney.In this experiment, vascular structure has not been significantly alteredduring the course of the perfusion with the CP solution. As with perfusion ofthe CP solution, the washout of DMSO with saline is not complete after 6 mlhave passed through the kidney.Careful examination of the water images in Figure 5.18 reveals that theresolution decreases during perfusion with saline. Washout of DMSO is notusually attempted with unmodified saline, since it causes severe osmoticimbalances (6) and cell rupture (41). The decrease in resolution, then, ismost likely due to generalized edema from water being freed from cells. Aloss of T2 contrast is expected if the intercellular spaces of the tissues areswollen with water.Figure 5.19 shows stacked plots of quantitative proton spectra takenduring the perfusion with CF solution (Figure 5.17) and during the DMSOwashout (Figure 5.18). These spectra represent the relative numbers of waterand DMSO protons in the kidney and the surrounding solution. Signals from the189protons of water in the kidney and the solution have the same resonancefrequency, so the wide water peak from the kidney is superimposed on thesharper peak from the solution. The same is true for the DMSO peaks fromkidney and solution. This explains the narrow tops and wide bottoms of thelines in the spectra.Spectra in Figure 5.19(A) were taken just after each corresponding imagein Figure 5.17. They clearly show the DMSO concentration increasing, asexpected. Likewise, spectra in Figure 5.19(B) correspond to the images inFigure 5.18. They show how the concentration of DMSO decreases as it iswashed out of the kidney with the saline perfusion. Spectra of this type willbe useful in determining when the DMSO has been removed from the kidney. Whenperfusing with a CP solution, they can be used in combination with the DMSOimages to determine when the concentration of DMSO in the kidney has reachedequilibrium. Fuller, Busza, and Proctor (42) used quantitative protonspectroscopy to determine the equilibration rate of DMSO in rat liver. Sincethe distribution of DMSO in the organs does not seem to be dependablyhomogeneous, however, spectra alone cannot prove that equilibration iscomplete. In fact, quantitative three dimensional imagin9 would be the onlymethod which could prove equilibration of CPS within the tissues.190Figure 5.16. The syringe Pump used for Perfusi0 of rat kidneys The weighton the a allows for release of pressure when the vascular resistance buildsup enough to lift the weight, whIch correspondS to 40 mm g. This simulates aConstant_pres rather than a COnstant_rat delivery.1910=fr_Figure 5.17. Water and DMSO images of rat kidney subjected to a 250 ml bloodwashout, during perfusion with 20% DMSO in saline using the syringe pump fordelivery. Image A shows only one small blood clot remaining. Image B showsDMSO in the renal artery (arrow). Image C shows DMSO perfusing through thearcuate arteries (arrow). DMSO distribution is almost homogeneous in the lastimages (D).Water DMSOC0 mL0 miri0.8 mL64 mm2.1 mL168 mm6 mL480 mm16 ms500 ms192Water DMSOA0.25 rnL15 mnBC0.75 rnL45 mm4D6mL720 mmTE = 16 msTR = 500 msFigure 5.18. Images of water and DMSO protons in the rat kidney of theprevious figure during washout of DMSO with saline. In image A, the DMSO isbeing removed first from the arcuate arteries (compare to Figure 5.17(C)).Some DMSO remains in the kidney after washout with 6 ml of saline.193A0.25 mlB0.75 ml0 ml2.10.8 ml6m1ml6 mlFigure 5.19. Quantitative proton spectra taken durng the experiments inFigures 5.17 and 5.18. Each spectrum was acquired just after thecorresponding images. (knounts of perfusate are beside each spectrum.) Thesespectra quantify the relative numbers of water and DMSO protons present in thesolution and the kidney. (A) DMSO concentration increases as the kidney isperfused with CP solution. (B) DMSO concentration decreases as the DMSO iswashed out with saline. (See discussion on page 189).0.5 ml1945.3.4.2. Determination of themount of Perfusates RequiredIn the experiments shown in Figures 5.18 and 5.19, completeequilibration of the kidney with the CP solution and saline washout was notattempted. If the procedures were to be standardized, it was necessary toknow how much CP solution would give complete equilibration of DMSO within thekidney and how much vehicle solution was required to remove the DMSO. Also,because saline caused swelling of the organ, it was decided to switch to UWsolution as the vehicle in the CP solution and for the washout of DMSO.Experiment 417, Figure 5.20, shows a kidney perfused with 21 ml of 20%DMSO in UW solution at room temperature (22°C), over a period of 11 hours.The first water image shows less water in the cortex than the medulla, and therenal pelvis is enlarged. The enlargement is probably due to drainage of thelarge volume of perfusate. The first DMSO image indicates that the DMSOconcentration has equilibrated in the tissues. The second DMSO image wastaken after removing the solution around the kidney, emphasizing that the DMSOdistribution is homogeneous. This image is of a slightly different slicethrough the kidney due to movement of the organ. Removal of surroundingliquid causes shimming problems which adversely affect the resolution in theimages, so saline was added to the NMR tube before the second water image wastaken. This image shows an area (bottom right) which may be either a bloodclot or fat. The large round dark area is due to field inhomogeneity from atiny air bubble on the surface of the kidney.Experiment #17 clearly indicates that volume of CP solution between 6 ml(experiment *13) and 21 ml allows for complete equilibration of DMSO in thetissues. However, it is possible that the swelling of the kidney seen inFigure 5.20 is due to the very large volume of perfusate. Another concern isthat the equilibration will be slower at lower temperatures. Sinceexperiments were to be done at 10°C, it was desirable to know how much195perfusate is necessary for equilibration of DMSO at that temperature and howmuch washout solution is needed to remove it.Another experiment (#14) was done, using a temperature of 10°C andmonitoring the DMSO distribution over time, to determine what volume of CPsolution was necessary for the distribution to be homogeneous. Figure 5.21shows DMSO and water images during the perfusion process. The perfusion rateis constant during the course of the experiment, at 1.1 ml/hr, so the vascularresistance is fairly stable. Again, although the vehicle solution wasswitched to UW solution, there is evidence of swelling in the water images.The DMSO images show a typical pattern of perfusion, with DMSO entering thecortex through the afferent vessels first, then more slowly penetrating themedullary tissues. The relative concentration of DMSO in the kidney andsurrounding solution was monitored by NMR spectroscopy during the perfusionand showed no further changes in the last hour of the experiment. In the lastDMSO images, after 9 hours and 10 ml of CP solution, the DMSO concentration ishomogeneous. (Perfusion was actually continued for another two hours, but nofurther changes in the spectra or images was observed.)The C? solution was replaced with UW solution for removal of the DMSO,also at 10°C (experiment #14w) (Figure 5.22). The perfusion rate was stablethroughout the experiment at about 2.9 mi/hr. The water images show that moreswelling has occurred. DMSO images show that the DMSO is washed out in thesame pattern as it entered, flushing out of the afferent vessels quickly, thenout of the cortex, and finally out of the medulla. There is some residualDMSO in the medulla even after 20 ml of washout with UW solution, however, andimages and spectra from the last two hours (4 ml washout) did not change.Fuller, Busza, and Proctor (42) noticed this phenomenon in NMR spectra of ratliver; they estimated that approximately 2% DMSO was left in the liver afterwashout. In the case of rat kidney, this residual DMSO seems to beexclusively located in the medullary tissues.These experiments indicate that perfusion with 20% DMSO should be196complete after 10 ml of the CP solution have been flushed through the kidney.Washout with UW solution does not seem to be complete even after 20 ml ofwashout with UW solution. Furthermore, there is not much change in the DMSOconcentration after 10 ml of washout with UW solution. Evidence of DMSOreaction with proteins (34) has been found, but one would expect this to besimilar for all tissue types in the organ, and this would decrease themobility of the DMSO so that it would not be observed in the images. It seemsmore reasonable to assume that since the medullary tissue equilibrates moreslowly with the CP solution than does the cortical tissue, equilibration witha washout solution is slower in the medulla as well. Perhaps swelling in themedullary tissues and enlargement of the renal pelvis causes the vascularsystem in that area to be less efficient during the washout.197TE = 16 msTR = 1 5Water DMSOFigure 5.20. Images of water and DMSO protons in rat kidney perfused with 21ml of 20% Df1SO in (1W solution at 22°C. DP4SO concentration is homogeneous(images A), which is emphasized by removing the surrounding fluid (DMSO imageB). When fluid is returned to the NMR tube, the kidney is moved slightly.The water image B shows fat around the renal pelvis (arrow) and an air bubble(large round dark area)A198Water DMSO0 niL0 mmI 2.6 mL140 mm10 mL535 mmTE = 16 msTR = 1 sFigure 5.21. Images of water and DMSO protons in rat kidney during syringepump perfusion with 20% DMSO in UW solution at 10°C. The images show DMSOdistributed throughout the kidney. Images taken after those shown in Dindicated no further increase in the DMSO concentration. Thus, 10 ml ofsolution seem adequate to completely equilibrate the kidney tissues with theC? solution.AC199Water DMSOAB13 mL240 mmC20 mL415 mmTE = 16 msTR = 1 sFigure 5.22. Images of water and DMSO protons in the rat kidney shown in theprevious figure, during washout of DMSO with tiW solution. After 20 ml of UWsolution, there is still a small amount of DMSO left in the kidney. (The DMSOimage in C was obtained with a very high gain setting, so is not directlycomparable with the one before it.)2005.3.4.3. Use of Improved Perfusion Techniques inImaging Experiments at Three Temperatures.The experiments in the previous section demonstrate that at 10CC, with aCP solution of 20% DMSO in UW solution, equilibration of DMSO with the tissuesof a rat kidney occurs after 10 ml of perfusate have been flushed through thekidney. At higher temperatures equilibration should be faster because thebulk diffusion rate is higher. The DMSO can be removed by flushing 10-20 mlof UW solution through the vasculature, but a small percentage of DMSO remainsin the kidney after washout. This protocol was used to investigate theequilibration of DMSO in kidney at three different temperatures and also toattempt washout of DMSO after freezing the organ.Figure 5.23 shows water and DMSO images of a kidney perfused at 10°Cwith 10 ml of CP solution (experiment #19). The water images show that somechanges are occurring in the vasculature; generally it appears that some bloodvessels (probably veins, see Appendix D) are becoming distended. DMSO imagesshow the typical pattern of DMSO equilibration. DMSO concentration seemsfairly homogeneous after 10 ml of solution. The appearance of the DMSO imagesdid not change in the last 2 hours (1.5 ml) of perfusion. The perfusion ratewas approximately 0.5 mi/hr.In comparison, Experiment #17 was performed at 22°C, and perfused at arate of 1.1 mi/hr. Figure 5.20 shows that equilibration was achieved. It isclear that perfusion proceeds much more quickly at this temperature.Three attempts were made to perfuse a kidney with CP solution at atemperature close to normal body temperature for rats (37°C>. In the firstattempt, experiment #24, there was no perfusion at all due to high vascularresistance, although images showed no signs of damage. This could be due tothe use of saline as a vehicle solution or to spasm/blockage of the bloodvessels.The other failed attempt at higher temperature perfusion, experiment201#26, was done at 36°C, with UW solution as the vehicle. The perfusion wasextremely slow, 0.16 ml/hr, so that after 8 hours the DMSO images showedalmost no DMSO present in the kidney. However, water images taken during thisexperiment show interesting changes (Figure 5.24). The first image, takenjust after the perfusion started, shows dark areas indicating residual bloodin the vasculature closest to the renal artery where the pressure is probablyhighest. The following images show this blood moving into other parts of thevasculature (arrows).In one experiment (#25) at 36°C, perfusion with CP solution wassuccessful. Vascular resistance was low during perfusion of the first 1 ml ofsolution, indicated by a high perfusion rate of 4 mi/hr1 but then decreaseddramatically thereafter, the perfusion rate dropping to 1.5 mi/hr. In Figure5.25, the DMSO images show that DMSO appeared, as expected, first in thecortex, and then in the medulla. The water images show that after the first 2ml of perfusate, there were progressive changes (arrows) which are similar tothose seen in previous experiments (especially #20 and #26) associated withsmall blood clots. The overall perfusion rate was very similar to that atroom temperature (1.8 compared to 1.9 ml/hr).Judging by the above preliminary experiments, it is best to perfuse thekidneys at room temperature, since this seems to minimize damage and maximizethe perfusion rate.202Figure 5.23. Images of water and DMSO protons in a rat kidney duringperfusion with 10 ml of 20% DMSO in UW solution at 10°C. The large oval darkarea in the upper left of each image is a roll of Teflon tape for positioningof the kidney,due to its small size. After 10 ml of CP solution wereperfused, the DMSO concentration was homogeneous.Water DMSOAB.9 niL210 in33 niL36() mm/8 mL810 rmmmn10 nL1080 nmDTE = 16 insTR = 1 sAC0 mL 0mm0.9 rnL 330 mm0.6 mL 225 mm1.1 mL 375 mm203TE = 16 msTR = 500 msFigure 5.24. Images of water protons in rat kidney during a perfusion with20% DMSO in UW solution at 36°C. Image A shows some residual blood near therenal artery. As the perfusion continues, this blood moves to the arcuatearteries (B) and then into the arterioles (C and D). This perfusion was notsuccessful, probably due to this blockage of the blood vessels.II4-204Water DMSOA0 niLo nunB 2 mL40 mmC7 mL230 miriTE = 16 msTR = 1 s9 nL290 miiiFigure 5.25. Images of water and Dt4SO protons in a rat kidney duringperfusion with 9 ml of 20% DMSO in UW solution at 36°C. There is someindication (images C and D) of blood clots moving through the vasculature(arrows) as in the previous figure, but they are much smaller. DMSO issuccessfully distributed throughout the kidney.2055.3.5. Freezing of DMSO—Perfused Kidneysand Subsequent Removal of DMSO5.3.5.1. Preliminary Imaging Studyof Freezing KidneyThe next stage in cryopreservation of organs is freezing. Research hasfocussed on freezing injury mechanisms and how to avoid injury by optimizingfreezing rates for specific treatments. In this thesis the emphasis was toinvestigate how freezing of a cryoprotected organ could be monitored with C4SNMR imaging and whether any changes in the organs could be detected afterfreezing and during washout procedures.The first freezing experiment (#19w) was done after a kidney had beensatisfactorily equilibrated with 20% DMSO in UW solution. Figure 5.26 showswater and DMSO images at different temperatures. The first set of imagesshows the distribution of water and DMSO at 10°C. In the images at —10°C and—30°C, the surrounding solution contains structures corresponding to portionsof the solution (pure water or pure DM50) which have frozen out into sheets(visible to the eye), which in cross—section appear as straight dark lines.The DMSO images both show signal from the kidney tissues at —30°C. At thistemperature, the T2 has decreased so that most of the remaining signal is lostbefore it can be acquired. However, the signal visible is reasonablyhomogeneous so most of the DMSO in the tissues is still in a liquid state asexpected, and the organ has been at least partially protected from freezing.The last set of images in Figure 5.26 shows the kidney after thawing to10°C. A comparison of the first and last sets of images shows that no visiblechanges have occurred due to the freeze/thaw cycle (possibly because thereactually was very little freezing), except possibly an increase in the size ofthe very bright areas.Removal of the DMS0 in this kidney was attempted, using UW solution206(Figure 5.27). The water images show very little change during the washout.1rhe DMSO images show that the washout of DMSO does not follow the pattern seenbefore; instead, the DMSO looks as though it is removed first from the partsof the kidney proximal to the renal artery as if bulk diffusion was removingit. This could happen if the vasculature was not carrying the tiW solution tothe entire kidney efficiently, since the larger blood vessels would then actas a DMSO “sink”. There does not appear to be any change in the vasculature,though, and not much swelling. Washout was much slower than in experiment#14w (0.85 ml/hr compared to 2.9 ml/hr), which may indicate that exposure tolow temperatures has caused damage.Water DMSObcFigure 5.26. Images of water and DMSO protons in rat kidney equilibrated with20% DMSO. The dark oval object at top left is a roll of Teflon tape used toposition the kidney. The temperature was lowered as indicated for each set ofimages. Sheets of pure DMSO or pure water (visible to the eye), which appearas straight dark lines in the images, begin to appear at -10°C. The loss ofsignal in the water images indicates that water protons are becoming lessmobile but not necessarily that the water has frozen.207ABCD-i0’C-30C-r10C,;:d/TE = 16 msTR = 1 s208Water DMSOI i nL210 nmL4 mL31 nunCI 9.1 mL180 nun12 rnL840 mmnTE = 16 msTR = 1 sFigure 5.27. Images of water and DMSO in the rat kidney of the previousfigure. Removal of the DMSO was attempted by perfusing with tJW solution at10°C. The DMSO seems to be being removed by bulk diffusion to the surroundingsolution and possibly the solution in the renal artery, indicating that thevasculature may be blocked by swelling or some other damage.B2095.3.5.2. Freezing and DMSO Washout ExperimentsThe above experiment demonstrates that a kidney protected by perfusionwith 20% DMSO can be frozen and thawed without changes in the NMR images, andthe DMSO can be successfully washed out again. The next stage of this studywas to freeze cryoprotected kidneys for varying lengths of time to monitorchanges and attempt washout of DMSO. Also, a method of perfusing kidneys withCP solution on the bench was developed.5.3.5.3.1. Comparison of perfusion temperaturesafter freezing kidneys for 1 dayIn experiment 20w, the kidney was perfused on the bench at roomtemperature (23°C) with 10 ml of 20% DMSO in UW solution. This kidney wasthen frozen at -20°C for one day, thawed to 10°C in the probe, and imaged.The first set of images (Figure 5.28(A)) shows that the distribution of DMSOis homogeneous and that the kidney seems not to show signs of perfusiondamage, except possibly some swelling (arrow).Washout of DMSO with UW solution was attempted, shown in the next twosets of images (Figures 5.28(B), (C)) . The DMSO washes out of the kidney inthe expected pattern, at a very high rate (6.5 ml/hr). One part of thekidney, at right in the images, does not seem to be reperfused by the UWsolution in the 1—mi image. This could be correlated with the swollen area inthe first water image (arrow). The last DMSO image does show almost completewashout after 10 ml of UW solution. Either the area was blocked by swellingor damaged but still capable of allowing perfusion, or bulk diffusion hascarried the DMSO out of that area. At 10°C the bulk diffusion is probably tooslow to accomplish removal of DMSO in this short time. In any case, DMSO210removal was possible after the kidney was held at —20°C for one day.In the next experiment, #21w, the bench perfusion was modified so thatthe NMR tube containing the kidney was on ice. The temperature of perfusionwas then just above 0°C. After perfusion the kidney was frozen for one day at—20°C, then thawed in the probe as in experiment #20w. The first set ofimages in Figure 5.29 shows the kidney before the washout at 10°C was started.The wedge—shaped dark area at the top of the image is fat, and all images aredarkened at the bottom due to RF inhomogeneity. The image plane is slightlyabove the renal pelvis. The DMSO image shows that DMSO is equilibratedthroughout the kidney. The water image shows some signs that the bloodvessels have been distended by the perfusion (bright areas). The lowerportion of the kidney seems to be swollen (arrow), but there is no sign thatthe blood vessels there were changed by the perfusion and freezing.During the initial stages of DMSO washout (Figure 5.29(B)), the kidneyswelled, and reperfusion was not homogeneous. Areas close to the bloodvessels which were possibly distended by perfusion began to lose DMSOimmediately, but areas that seemed swollen before washout did not reperfuse aswell (arrow). The washout proceeded much more slowly than the previous one,and the areas that showed less initial reperfusion may be losing their DMSOmostly through bulk diffusion. The last DMSO image shows that DMSO has beenalmost completely removed from the kidney, so despite possible damage thewashout was successful.It is interesting to compare Figures 5.28 and 5.29, keeping in mind thatthe image plane in Figure 5.29 is slightly above the renal pelvis. Both setsof images show dark or bright areas which may indicate vascular blockage ordamage, which were then associated with uneven reperfusion. Since bothkidneys show the same lack of contrast in areas which did not reperfuse well,these may be attributable to generalized swelling due to tissue damage fromperfusion or freezing.The difference in perfusion rates can be attributed to the CP perfusion211temperature. Room temperature perfusion is definitely faster than perfusionon ice, although both methods seem to allow equilibration of DMSO in thekidney. More evidence of the low-temperature bench perfusion causing damageis seen in Figure 5.30 (experiment *23). This kidney was perfused in a salinebath surrounded by an ice/water bath, so the temperature is reliably 0°C.This kidney was black in color after the perfusion, indicating severe damageand ensuing necrosis of tissues.212Water DMSO0 niLo mm1 rnL10 mm10 rnL90 mmTE = 16 msTR = 1 SFigure 5.28. Images of water and DMSO protons in a rat kidney subjected toperfusion with 10 ml of 20% DMSO in (3W solution, then frozen at —20°C for oneday. (A) Before perfusion with (3W solution at 10°C, DMSO distribution ishomogeneous. A swollen area (arrow) appears just below a dark feature whichcould be a blood vessel in cross-section. (3) When perfusion begins, the darkfeature is filled with solution. The swollen area does not reperfuse well(arrow). (C) Some swelling of the kidney has occurred, but the DMSO has beenmostly removed. The DMSO image was obtained with a much higher gain than theprevious DMSO image.2131.8 mL60 mm2.7 mL120 ruin4.5 mL180 ruinFigure 5.29. (A) Images of water and DMSO protons in rat kidney subjected toperfusion with 10 ml 20% DMSO in OW solution at 0°C, and frozen for 1 day at—20°C. After thawing to 10°C, DMSO distribution is homogeneous. There is aswollen area (arrow) similar to that in the kidney in the previous figure.(B—D) Images of the kidney during washout with OW solution. The swollen areadoes not reperfuse as well as the rest of the cortex, but removal issuccessful.Water DMSO0 mL0 ruinTE 16 msTR = 1 s214TE = 16 msTR = 1 SFigure 5.30. Image of water protons in rat kidney subjected to 0°C perfusionwith 10 ml 20% DMSO in UW solution. Arrow shows an area possiblycorresponding to blood clots in the vasculature. This kidney was severelydamaged by the perfusion procedure.2155.3.5.2.2. Freezing of kidneys for longer periods of timeThe ultimate goal of cryoprotectiori is to achieve preservation of thestructure and function of the kidney after long periods of storage at lowtemperature. Therefore experiments with longer freezing exposure were done.A pilot experiment (#17w) was performed with the kidney from experiment *17,shown in Figure 5.20. It was frozen for 4 days, then thawed to 10°C in theprobe. Figure 5.31 shows that the DMSO concentration is homogeneous.Comparison of water images in Figure 5.20 and 5.31 shows that the kidney hasswollen, possibly from the freeze/thaw process. Round dark areas in the image(arrows) are due to tiny air bubbles, which are difficult to remove afterthawing without moving the kidney.Washout of DMSO was attempted, and failed. This is consistent with whatwas seen in Figures 5.27 and 5.28, that swelling of the tissues preventsreperfusion. In this experiment, the swelling was throughout the wholekidney, and there was no reperfusion whatsoever.Experiment #16w was performed on the kidney from experiment #16, whoseperfusion images were presented in Figure 5.19. This kidney was frozen at—20°C for 13 days after perfusion with 18 ml 20% DMSO. The first pair ofimages in Figure 5.32 show that the DMSO is not equilibrated throughout thesample. This is probably due to residual blood in the medulla, seen in thefirst water image in Figure 5.32 (arrow). Washout was unsuccessful, with thekidney not reperfusing, though some DMSO was removed by bulk diffusion. Thisis reasonable if the kidney tissues were not completely protected by DMSO.The second water image shows that the kidney has swelled, consistent withfreezing damage.The swelling which every kidney has shown during washout of DMSO isprobably caused by excessive osmotic imbalances in the tissues as DMSO isremoved. The DMSO apparently cannot leave the cells quickly enough, whichcauses the cells to swell and even burst. Fuller and Pegg (6) showed that216addition of a nonpermeant solute such as inannitol or sucrose to the washoutsolution can solve this problem by decreasing the intake of water to thecells. Thus in the next experiment the UW solution was modified by additionof 300 mM sucrose.Experiment #22w is a washout of DMSO from a kidney perfused at roomtemperature with 10 ml of 20% DMSO, and frozen at —20°C for 13 days. Thefirst DMSO image in Figure 5.33 shows that the DMSO concentration in thekidney is homogeneous, the water image shows no signs of damage except someswelling in the lower left (arrow). The second and third DMSO images showthat DMSO is removed from the kidney unevenly. There is an area ofnonreperfusion in the left cortical region. The DMSO is removed at a muchlower rate than with the kidneys which were frozen for shorter times, 0.8ml/hr compared with up to 6.5 ml/hr (experiment #20), especially since thiswashout was performed at room temperature and should be faster. The lastwater image shows no evidence of generalized swelling, so apparently theaddition of sucrose to the washout solution has prevented osmotic imbalances.This experiment demonstrates that the vasculature of a kidney which hasbeen equilibrated with 20% DMSO and frozen for 13 days is still structurallyintact enough to allow washout of DMSO. The failure of experiment #16windicates that inhomogeneous distribution of DMSO to tissues before freezingcauses damage severe enough to prevent reperfusion. C4S NMR imaging is theonly technique available which can show this kind of relationship betweenincomplete equilibration with a CPS and subsequent freezing damage.2]. 7WaterDMSOTE 16 msTR = 1 sFigure 5.31. Images of water and DM30 protons in rat kidney from Figure 5.20.Kidney was frozen for 4 days at —20C, then thawed to l0C and imaged. Thefreeze/thaw process has caused swelling. The round dark features on thesurface of the kidney are air bubbles.218TE 16 msTR = 1 5Figure 5.32. Images of water and DMSO protons in the kidney from Figure 5.19.This kidney was frozen for 13 days after perfusion with 18 ml 20% DMSO. Thefirst pair of images show that the DMSO was not equilibrated throughout thesample. This is probably due to residual blood in the medulla, seen in thefirst water image (arrow) . Washout was not successful; the kidney did notreperfuse, though some DMSO was removed by bulk diffusion. The second waterimage shows that the kidney has swelled.__:Ii;;f .--219o rnLo mm1 rnL90 mm2 mL65 mm8 mL630 mmFigure 5.33. Images of water and DMSO protons in a rat kidney perfused with10 ml 20% DMSO in UW solution at 22°C, then frozen for 13 days at -20°C, andthawed to room temperature. (A) The images show no obvious damage except aswollen area (arrow). (B—D) Washout of DMSO with UN solution was attempted.Images of DMSO protons show that perfusion was uneven, again associating aswollen area with nonperfusion. However, the addition of sucrose to thewashout solution has prevented generalized swelling due to osmotic imbalances.Water DMSOABCDTE = 16 msTR = 1 52205.3.6. 31P NMR Study of Excised Rat Kidney31P NMR spectroscopy of rat kidney was performed at 300 MHz in order tocompare with similar spectra in the literature. It has been proposed that 31PNMR could be used as a noninvasive viability test for cryopreserved organswhich was less difficult than transplantation. All cells in the body requireadenine nucleotides and other “energy molecules” (ATP, ADP, PCr, etc.) tofunction, and these molecules are depleted during the cryopreservationprocess. It was reasoned that in order for cells to maintain their osmoticbalance and proper chemical gradients during perfusion with a preservationsolution, these energy molecules must be available; if not, the cells (andthus the organ) would not be viable (44) . Studies of rat kidney showed thatwithin minutes after excision the kidney had lost appreciable amounts of itsreserve of adenirie nucleotides (45). The loss was slower at lowertemperatures, but it was not possible to maintain a normal amount of ATP andADP in the organ during a perfusion with a CP solution.Similar losses of high—energy adenine nucleotides during storage of ratliver at 0°C were noted by Busza, Fuller, and Proctor (46). However, theirstudy showed that under carefully controlled conditions, cold—stored liversreperfused with modified UW solution (17) were able to resynthesize adeninenucleotides. 31P NMR spectra of cold—stored livers before reperfusion werenearly identical to those of flushed and cold-stored kidney in the presentwork. Figure 5.34 shows a spectrum taken after 1 hour of cold storage, whichwas a typical time between excision and the beginning of perfusion with a CFsolution in the present work. Peaks in the spectrum correspond tophosphomonoesters, inorganic phosphate, phosphodiesters, and NADH.221Busza et al.’s work shows that the adenine nucleotides can beregenerated during perfusion of a cold—stored organ with a solution such as UWsolution. In the experience of Betty Pearson and others, such cold-storedkidneys are routinely transplanted successfully. Petterson et al. suggestFigure 5.34. 31P NMR spectrum of a rat kidney which has been subjected to a250 ml blood washout and 35 minutes at 0°C, which was a typical time betweenexcision and the beginning of perfusion with a CP solution in the presentwork. Peaks in the spectrum correspond to (A) phosphomonoesters, (B)inorganic phosphate, (C) phosphodiesters, and (D) NADH (reduced form ofnicotinamide adenine dinucleotide).that it is the activity of the Na-K—ATPase enzyme which must be preserved(45), not the normal levels of the adenine nucleotides themselves. Otherrecent studies have shown that the lack of adenine nucleotides is not a majorfactor in organ survival, rather, the organ’s ability to regenerate theseA0ii III I 11111 III II I 11111111 I I liii III I I 11111 I III I 11111111 I III I Il1 II30 20 10 0 —10 —20 -30PPM222molecules is important (45,47). Thus, despite the loss of adenine nucleotidesseen in Figure 5.34, the kidneys transported for perfusion in the experimentsin this thesis are probably viable before the beginning of thecryopreservation procedure.5.4. SUMMARYThis chapter has described several series of experiments which show howC4S NMR imaging can be used to monitor perfusion of a CP (DMSO) solution intorat kidney. NMR C4S imaging was demonstrated to be useful in assessing theeffectiveness of common techniques for removing blood from kidneys andtechniques for perfusing kidneys with CP solutions. Freezing of kidney wasmonitored, showing the effects of very short T2 on the images at lowtemperatures.Removal of blood from kidneys was shown to be incomplete after manually—controlled syringe washout and after gravity—fed washout. A technique whichremoves the blood from the entire animal was most successful. Very smallamounts of blood left in the kidneys can be detected as dark features in theimages.Perfusion of a CP solution into rat kidney was successful with all thecommon perfusion methods, but perfusion using a syringe pump for a “constantpressure” delivery was most efficient. Perfusion rates depend on temperatureas expected but also depend on how successful the blood washout was. The bestperfusion temperature was determined to be near 20°C. Lower temperaturesnecessitated very long perfusion times, and normal body temperatures causedperfusion failure two out of three times. Required volumes for equilibrationof a kidney with 20% DMSO in tiW solution were determined as well ascorresponding volumes of UW solution required to remove DMSO from the kidney.223Freezing of kidneys equilibrated with 20% DMSO in UW solution affectedthe subsequent removal of DMSO by causing some form of damage (probablyswelling), detectable as featureless areas in the images, which preventedperfusion of the washout solution to those areas of the kidney. Removal ofDMSO was possible in some cases, but there is probably significant freezingdamage to the kidney which would affect blood flow if the organ wastransplanted. The last experiments performed on cryoprotected kidneys whichwere frozen for 13 days indicated that DMSO could be removed from a thawedorgan if the organ had been completely equilibrated with DMSO before freezing.Thus the vasculature must be reasonably intact after the cryopreservationprocedure.The best conditions for successful removal of blood were obtained with awhole body gravity—fed blood washout procedure followed by excision andcatheterization of the kidney. Perfusion with 20% DMSO in UW solution wasmost efficiently performed with a syringe pump. Homogeneous distribution ofDMSO throughout the kidneys was most quickly and reliably achieved at roomtemperature. DMSO could in some cases be removed fron a kidney by washoutwith UW solution after freezing and thawing. Addition of sucrose to thewashout solution prevented generalized swelling of the organ.The experiments in this chapter indicate the potential usefulness of NMRimaging in the field of cryobiology. NMR imaging can be used to ensure thatblood has been removed from an organ before a CPS is introduced, therebyavoiding irihomogeneous delivery of the CPS; C4S NMR imaging monitors the CPSdistribution during and after delivery to ensure that equilibration hasoccurred throughout the tissues; it monitors freezing and thawing of the waterand CPS in the tissues; it detects freezing damage causing localized swelling;and it monitors the success of CPS removal. These abilities, combined withthe fact that the technique is nondestructive and noninvasive, are unique toC4S NMR imaging.224REFERENCES(1) Kubota, S., E.F. Graham, B.G. Crabo, et al. 1974. Influence of DMSOdistribution upon renal function following freezing and thawing. J.Surg. Res. 16, 582—591.(2) Small, A., N.J. Feduska, and R.S. Fib. 1977. Function of autotranplantedkidneys after hypothermic perfusion with dimethylsuifoxide. Cryob. 14,23—36.(3) Pegg, D.E. 1971. Vascular resistance of the isolated rabbit kidney. Cryob.8, 431—440.(4) Schoonees, R., G.S. Johnston, J.H. Groenewald, P.D.H. van Heerden, andG.P. Murphy. 1971. Radioisotope studies in perfused baboon kidneys.Cryob. 8, 134—137.(5)Marvig, B., 0. Kallskog, and B.J. Norlén. 1980. Effects of cold ischemiaon the preserved rat kidney: intrarenal distribution of perfusate.Cryob. 17, 478—485.(6) Fuller, B.J. and D.E. Pegg. 1976. The assessment of renal preservation bynormothermic bloodless perfusion. Cryob. 13, 177-184.(7) Belzer, F.0., and J.H. Southard. 1980. The future of kidney preservation.Transpl. 30, 616—615.(8) Pegg, D.E. 1972. Perfusion of rabbit kidneys with cryoprotective agents.Cryiob. 9, 411—419.(9) Jeske, A.H., M.C. Fonteles, and A.M. Karow, Jr. 1974. Functionalpreservation of the mammalian kidney. III. Ultrastructural effects ofperfusion with dimethylsulfoxide (DMSO). Cryob. 11, 170—181.(10) Collins, G.M., M. Bravo—Shugarman, and P.1. Terasaki. 1969. Kidneypreservation for transportation. Initial perfusion and 30 hours’ icestorage. Lancet 2, 1219—1222.(11) Pegg, D.E. 1978. An approach to hypothermic renal preservation.Cryobiology 15, 1-17.(12) Fahy, G.M., M. Hornblower, and H. Williams. 1979. An improved perfusatefor hypothermic renal preservation. I. Initial in vitro optimizationbased on tissue electrolyte transport. Cryob. 16, 618.(13) Leibach, F.H., M.C. Fonteles, D. Pillion, and A.M. Karow, Jr. 1974.Glutathione in the isolated perfused rabbit kidney. J.Surg. Res. 17,228—231.(14) Wahlberg, J.A., J.H. Southard, and F.0. Belzer. 1986. Development of anew cold storage solution for pancreas preservation. Cryob. 23, 477-482.225(15) Southard, J.H., M.J. Rice, and F.O. Beizer. 1985. Preservation of renalfunction by adenosine—stimulated ATP synthesis in hypothermicallyperfused dog kidneys. Cryob. 22, 237—242.(16) Biguzas, M., A.C. Thomas, K. Walls, B.O. Howden, D.F. Scott, and V.C.Marshall. 1990. Evaluation of UW solution in a rat kidney preservationmodel. Transpl. 49(5), 872—875.(17) de Mel, T., B.J. Fuller, and K.E.F. Hobbs. 1990. University of Wisconsinsolution without lactobionate and raffinose. Transpi. 50(5), 906—908.(18) Sakagami, K., S. Takasu, S. Kawamura, S. Saito, et al. 1989. A comparisonof Universtiy of Wisconsin and Euro—Collins solutions for simple coldstorage in non-heart—beating cadaveric kidney transplantation. Transpi.49(4), 824—826.(19) Keeler, R., J. Swinney, R.M.R. Taylor, and P.R. Uldall. 1966. The problemof renal preservation. Brit. J. Urol. 38, 653-656.(20) Karow, A.M., Jr. Biological effects of cryoprotectant perfusion, deliveryand removal to nonfrozen organs. The Biophysics of Organ Preservation.D.E. Pegg and A.M. Karow, Jr., eds. Nato ASI series A, vol. 147. (PlenumPress: New York, 1987).(21) Baxter, S.J. and G.H. Lathe. 1971. Biochemical effects on kidney ofexposure to high concentrations of dimethyl sulfoxide. Biochem.Pharmacol. 20, 1079—1091.(22) Barner, H.B. 1965. Mannitol as an osmotic antagonist to dimethylsulfoxide. Cryob. 1, 292—294.(23) Karow, A.M., Jr., M. McDonald, T. Dendle, and R. Rao. 1986. Functionalpreservation of the mammalian kidney. VII. Autologous transplantation ofdog kidneys after treatment with dimethyl sufoxide (2.8 and 4.2 M).Transpi. 41, 669—674.(24) Toledo-Pereyra, L.H., D.A. Gordon, G.H. MacKenzie. 1982. Organ freezing.J. Surg. Res. 32, 75—84.(25) Beizer, F.O., .S. Ashby, and J.E. Dunphy. 1967. 24-hour and 72-hourpreservation of canine kidneys. Lancet 2, 536—539.(26) Flalasz, N.A. and G.M. Collins. 1974. Simplification of perfusionpreservation methods. Transpi. 17(5), 534—536.(27) Pegg, D.E. and C.J. Green. 1973. Renal preservation by hypothermicperfusion. 1. The importance of pressure control. Cryob. 10, 56-66.(28) Pegg, D.E., C.J. Green, and J. Foreman. 1974. Renal preservation byhypothermic perfusion. 2. The influence of oxygenator design and oxygentension. Cryob. 11, 238—247.(29) van der Wijk, J., M.J.H. Sloof, B.G. Rijkmans, and G. Kootstra. 1980.Successful 96— and 144-hour experimental kidney preservation: acombination of standard machine preservation and newly developednormothermic ex vivo perfusion. Cryob. 17, 473-477.(30) Toledo-Pereyra, L.H. 1980. Factors involved in successful freezing ofkidneys for transplantation: preliminary experimental observations. J.226Surg. Res. 28, 563—570.(31) Collins, G.M. and N.A. Halasz. 1984. Studies in cryoprotection. 1. Asimple method for the controlled introduction and removal ofcryoprotective agents during organ perfusion. Cryob. 21, 1-5.(32) Pearson, B., personal communication.(33) Fisher, B. and S. Lee. 1965. Microvascular surgical techniques inresearch, with special reference to renal transplantation in the rat.Surgery 58(5), 904—914.(34) D’Silva, S., R.F. Gittes, J.L. Wolf, J. Pirenne. et al. 1990. Rat kidneytransplantation update with special reference to vesical calculi.Microsurg. 11, 169—176.(35) Zhong, R., 0. Grant, R. Black, C. Stiller, and J. Duff. 1989. Combinedsmall bowel and kidney transplant in the rat. Transpl. Proc. 21(1) 2907—2908.(36) English, J., personal communication.(37) Pegg, D.E. and C.J. Green. 1976. Renal preservation by hypothermicperfusion. 3. The lack of influence of pulsatile flow. Cryob. 13, 161—167.(38) Pegg, D.E. 1971. Vascular resistance of the isolated rabbit kidney.Cryob. 8, 431—440.(39) Skaaring, P., F. Bierring, J. Hejnal, V. Svendsen, E. Jensen, and E.Kemp. 1975. Ultrastructure of the glornurular filtration membrane ofautotranspianted canine kidneys stored for 24 hours. Cryob. 12, 224-230.(40) Korber, K.E. and B.A. Kraemer. 1988. Heterotopic renal transplantation inthe rat: an advanced microsurgical training excercise. Microsurg. 9,286—291.(41) Fuller, B.J., personal communication.(42) Fuller, B.J., A.L. Busza, and E. Proctor. 1989. Studies on cryoprotectantequilibration in the intact rat liver using nuclear magnetic resonancespectroscopy: a noninvasive method to assess distribution of dimethylsulfoxide in tissues. Cryob. 26, 112—118.(43) Isbell, S.A., C.A. Fyfe, P.A. Keown, and B. Pearson. Microscopic NMRimaging applied to a cryoprotectant system. Abstracts, 30th annualmeeting of the Society for Cryobiology, Cryob. 30, 621 (1994).(44) Pettersson, S., and T. Schersten. 1973. Sodium—, potassium—stimulatedadenosine triphosphatase (Na—K—ATPase) activity in human kidneytissue. Europ. Surg. Res. 5, 282—291.(45) Pettersson, S., G. Claes, and T. Schersteri. 1974. Correlation betweensodium- potassium stimulated ATPase activity and renal function aftertransplantation of canine kidneys. Europ. Surg. Res. 6, 239—246.(46) Busza, A.L., B.J. Fuller, and E. Proctor. 1994. Evaluation of coldreperfusion as an indicator of viability in stored organs: A 31P NMRstudy in rat liver. Cryob. 31, 26—31.227(47) Bowers, J.H., K. Teramoto, U. Khettry, and N.E. Clouse. 1992. 31P NMRassessment of orthotopic rat liver transplant viability: The effect ofwarm ischemia. Transpl. 54, 604—609.228CHAPTER 6CONCLUSIONS AND RECONDATIONSFOR FURTHER RESEARCHThis thesis has discussed the potential of C4S NMR imaging as a tool forstudying freezing/thawing phenomena. Specifically, C4S NMR imaging was usedto evaluate various techniques used in cryopreservation protocols: Bloodwashout, introduction of cryoprotectant solvent, freezing, thawing and washoutof cryoprotectant solvents (CPSs). The following discussion summarizes theresults and gives suggestions for further investigation of this application ofC4S NMR imaging.It was demonstrated in Chapter 2 that NMR spectroscopy and imaging areable to independently detect the freezing of two liquids in a mixed solventsystem. After extensive measurements of NMR relaxation parameters inDMSO/water test samples, it is possible to produce images whose signalintensities reflect the concentrations of protons of DMSO and waterrespectively, as functions of location within a test sample. Measurements ofconcentration from the signal intensities in these images are acceptablyaccurate for the purpose of detecting the amounts of a CPS in a sample, withan error of ±5%.It might be interesting to pursue the idea that quantitative C4S NMRimages of samples at different temperatures can be used to create a “phasediagram” of a CPS in a tissue to be cryopreserved: A sample can be imaged atsuccessively lower temperatures, and the amounts of CPS and water determinedat each temperature. This would provide information about how much of the229CPS (and how much water) was frozen at each temperature, and possibly whetheror not different types of tissues freeze at different temperatures. Accordingto the relaxation measurements in tissue samples in Chapter 3, this type ofstudy would at least require hardware capable of producing shorter soft pulsesthan those possible with the equipment used in this thesis, and probably anentirely different pulse sequence would have to be developed.The equipment requirements for quantitative imaging of CPS5 in tissuesamples at low temperatures may not be practical. In order to select for DMSOin the DMSO/water solvent system (at 9.4 Tesla), the pulse bandwidth must beless than 800 Hz, corresponding to a 5 ms sinc2—shaped pulse or a 2.5 ms sinc—1 shaped pulse. This would only reduce the TE of the C4S sequence to 7 ms.Because the T2 values of DMSO and water in tissues are expected to fall below7 ms before the freezing point is reached, this TE value will not allowquantitative imaging with the C4S sequence at these temperatures. A newpulse sequence, perhaps utilizing a saturation sequence for one of theresonances before imaging the other, would have to be developed.Instead of developing a new pulse sequence, it may be possible to usethe C4S sequence if extensive relaxation measurements of DMSO/water in tissuesare made, the signal intensity data from the images is corrected for theeffects of T1 and T2 contrast. In Chapter 3 a method of analyzing imageintensity data for the effects of contrast was described. Using this methodit was discovered that based on relaxation measurements in tissues, contrastdecreased the amount of signal but the signal still varied linearly withconcentration. Thus it was possible to accurately determine DMSOconcentrations of 10%—50% in rat kidney and liver tissues between 1°C and20°C, and possibly water concentrations of 50%-lOO% in the same temperaturerange, from signal intensities in images with TE and TR not set as forquantitative imaging.The most vital part of the cryoprotection technique is the equilibrationof the tissue with a CPS. Equilibration must be complete, or portions of the230tissue with less CPS may freeze more quickly, causing severe mechanical andosmotic stresses. The techniques available for studying CPS distribution intissues prior to this study have all relied on analyses of averageconcentrations of CPS in individual tissue samples after various equilibrationtimes. Values of DMSO diffusion rates and effective diffusion coefficientssimilar to those found with these “volume averaged” methods can be obtainedfrom NMR C4S images of DMSO protons in rat organ tissues, as shown in Chapter4. However, signal intensity data from NMR C4S images of DMSO protons can beused to measure effective diffusion coefficients much more accurately andunambiguously. The method described could be extended to give maps ofeffective diffusion coefficients across images of complex tissue samples. Itmight be possible to detect differences in the effective diffusioncoefficients between different tissue types. This information could becompared to the tissue types which are known to be susceptible to freezingdamage. It is possible that if a tissue type, say the vasculature, takes upDMSO less quickly, it may not have time to equilibrate with a DMSO solutionduring a cryoprotection experiment. This could explain why some tissue typesseem to be more prone to freezing damage than others (1).The discussion and images in Chapter 4 show very clearly that diffusionof DMSO into whole rat organs is too slow to be an acceptable method fordelivery of CPS to an organ for cryoprotection. The most commonly used methodfor achieving CPS equilibration in organs is perfusion of a cryoprotective(CP) solution (consisting of a vehicle solution and a CPS) through the organ’svasculature. However, it has never been proven that this technique canequilibrate a CP solution with all the different tissue types within an organ,because no appropriate method was available. C4S NMR imaging can detect bothconcentrations and locations of CPS in a sample.In Chapter 5, cryoprotective procedures used for the rat kidney wereinvestigated using C4S NMR imaging. In the rat kidney model used in manykidney transplant studies, blood is removed from the organ either by manually231forcing a solution through the vasculature with a syringe, or by using a knownconstant pressure (gravity) to force the solution through the vasculature.Both of these methods were investigated, and the effectiveness of each methodat removing blood was evaluated. The syringe method was very ineffective.The gravity—fed washout was more effective, and when it was modified to washthe blood out of the entire animal using a large volume of solution, NMRimages showed that all the blood had been removed.Perfusion of a CP solution through the vasculature of rat kidney wasstudied. C4S NMR images of perfusion of DMSO solutions into rat kidney showthat the CPS is distributed as expected: From the renal artery to the arcuatearteries, into the cortical tissues and finally into the medulla. Perfusionsuccessfully equilibrated the organ with DMSO only if the blood had beencompletely removed from the organ by the washout procedure. Perfusions werecarried out with manually controlled pressure, gravity-fed pressure, andconstant pressure from a syringe pump. All of the procedures were successful,but again only if blood was successfully removed.Since blood washout is seemingly not as critical to the survival oftransplanted pig kidneys (2), the C4S NMR imaging technique should also beapplied to other animal models, to better assess the relevance of completeblood washout to cryopreservation and transplantation of human organs.Equilibration of tissues with DMSO during a perfusion was shown to occurat different temperatures. The rate of perfusion was temperature-dependent,probably due to the changes in the viscosity of the DMSO solutions and changesin the properties of the cell membranes, etc. Organs can only survive for acertain period of time without a blood supply. This time is lengthened bylowering the temperature, since this slows the rates of biochemical reactions,thus requiring less oxygen consumption. Perfusion at 20°C was judged to bethe best compromise between the slow equilibration rate at low temperatures(10°C) and the high rate of deterioration of the tissues at high temperatures(37°C)23231P NMR spectroscopy has been used to monitor the viability of tissues(3) after their blood supply has been removed. In the research in thisthesis, more time was involved in obtaining the samples than in mosttransplantation studies, since the samples had to be transported. 31P NMRspectroscopy was used to show that spectra of the kidneys after transportationfrom the Jack Bell Research Center to the University of British Columbiacampus were similar to those of viable organs in the literature.The C4S NMR imaging technique has proven to be valuable in the study ofdiffusion and perfusion of DMSO in intact rat organ tissues. In the futurethis technique could be extended to quantify freezing in tissue samples. TheC4S NMR imaging technique is unique in its ability to quantify a CPS atdifferent locations in an intact organ, and should be used to assure thatcryoprotection protocols are achieving equilibration of CPSs and freezingprotection for the tissues. The commonly observed failures in organcryoprotection could be explained by ineffective blood washout procedureswhich prevent equilibration of CPS in nonperfused parts of the organ, andcause subsequent local freezing injury. C4S NMR imaging could be combinedwith more extensive histological examination to further assess thispossibility.233REFERENCES(1> Barner, H.B. 1965. The vascular lesion of freezing as modified by DMSO.Cryob. 2, 55—61.(2) Pearson, B., personal communication.(3) Jennings, R.B. H.K. Hawkins, J.E. Lowe, M.L. Hill, S. Kiotman and K.A.Reimer. 1978. Relation between high energy phosphate and lethal injuryin myocardial ischemia in the dog. xner. J. Path. 92, 187—196.234APPENDIX APARAMETERS USED FOR THE IMAGING EXPERIMENTSPRESENTED IN THE FIGURESFigure: 2.5Pulse sequence: STEQ.PC TE: 9 ms. TR: 1 S.SI: 512W TD: 256W NS: 4 Pulses: sinc2Ml: 1950/32 M14: 4000 DO: 1 S. D3: 200 .Ls.M4: 3950/0 M15: 250 D4: 200 ).Ls. RG: 55M7: 4000/0 M16: —330 D1O: 750 SW: 20 kHzM13: 1240 M17: N/A D14: 3 ms.Figure: 2.6Pulse sequence: STEQ.PC TE: 9 ms. TR: 1 s.SI: 512W TD: 256W NS: 4 Pulses: sinc2Ml: 1950/32 M14: 4000 DO: 1 • D3: 200 us.M4: 3950/0 M15: 250 D4: 200 us. RG: 55M7: 4000/0 M16: —330 D1O: 750 j.s. SW: 20 kHzM13: 1240 M17: N/A D14: 3 ms.235Figure: 2.8Pulse sequence: C4S TE: 16 ms. TR: 15 s.SI: 256w TD: 128w NS: 4 Pulses: siric2Ml: —2048/32 M14: 4000 DO: 12 D3: 400 is.M4: 2150/0 MiS: 250 D4: 500 Is. RG: 20M7: 2000/0 M16: —300 D1O: 500 s. SW: 15 kHzM13: 1800 M17: —250 D14: 9 ms.Figure: 2.9Pulse sequence: C4S TE: 16 ms. TR: 12.5 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 12 S. D3: 400 Is.M4: 2150/0 M15: 250 D4: 500 s. RG: 20M7: 2000/0 Ml6: —300 D10: 500 jis. SW: 15 kHzM13: 1800 M17: —250 D14: 9 ms.Figure: 2.10Pulse sequence: C4S TE: 16 ms. TR: 12.5 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Mi: —2048/32 Ml4: 4000 DO: 12 S. D3: 400 jts.M4: 2150/0 Ml5: 250 D4: 500 jis. RG: 20M7: 2000/0 M16: —300 DiD: 500.SW: 15 kHzM13: 1800 M17: —250 D14: 9 ms.236Figure: 4.3Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2368/37 M14: 3000 DO: 1 S. D3: 300 Jis.M4: 2850/0 M15: 500 D4: 1 ms. RG: 40M7: 3000/0 M16: —525 DlO: 1 ms. SW: 50 kHzM13: 1500 M17: —500 D14: 5 ms.Figure: 4.4Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2368/37 M14: 3000 DO: 1 S. D3: 300 j.ts.M4: 2850/0 M15: 500 D4: 1 ms. RG: 40M7: 3000/0 M16: —525 D1O: 1 ms. SW: 50 kHzM13: 1500 M17: —500 D14: 5 ms.Figure: 4.5Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2368/37 M14: 2000 DO: 1 5• D3: 300 .Ls.M4: 480/0 M15: 500 D4: 1 ms. RG: 10M7: 500/0 M16: —570 D1O: 1 ms. SW: 25 kHzM13: 1000 M17: —500 D14: 6 ms.237Figure: 4.6Pulse sequence: C4S TE: 16 ms. TR: 15 S.SI: 512w TD: 256w NS: 32 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 15 S. D3: 300 J.Ls.M4: 1870/0 M15: 250 D4: 1 ms. RG: 28M7: 2000/0 M16: —375 D1O: 1 ms. SW: 25 kHzM13: 1800 M17: —250 D14: 8 ms. 01: +370 (H20)—450 (DMSO)Figure: 4.7Pulse sequence: C4S TE: 16 ms. TR: 1 s.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —1024/16 M14: 4000 DO: 1 S. D3: 10 is.M4: 1950/0 M15: -250 D4: 1 ms. RG: 38M7: 2000/0 M16: —300 D10: 10 Is. SW 15 kHzM13: 1800 M17: —250 D14: 9 ms.Figure: 4.8Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2240/35 M14: 4000 DO: 1 S. D3: 400 is.M4: 2230/0 M15: 250 D4: 500 jis. RG: 38M7: 2000/0 M16: —580 DiD: 500 SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms.238Figure: 4.9Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml:—2048/32 M14: 4000 DO: 1 S. D3: 400 $Is.M4: 2230/0 MiS: 250 D4: 500 us. RG: 38M7: 2000/0 M16: —580 DlO: 500 . SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms.Figure: 4.10Pulse sequence: C4S TE: 16 ms. TR: 1 5.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2240/35 M14: 4000 DO: 1 5. D3: 400 uls.M4: 2230/0 M15: 250 D4: 500 . RG: 38M7: 2000/0 M16: —580 D10: 500 u• SW: 20 kHzMl3: 1800 M17: —250 D14: 9 ins.Figure: 4.16Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2240/35 M14: 4000 DO: 1 5. D3: 400 us.M4: 2230/0 M15: 250 D4: 500 RG: 38M7: 2000/0 M16: —580 D1O: 500 us. SW: 20 kHzM13: 1800 M17: —250 D14: 9 ins.239Figure: 5.3(a)Pulse sequence: C4S TE: 16 ms. TR: 2 s.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml:—384/3 M14: 2000 DO: 2 S. D3: 400 is.M4: 480/0 M15: 500 D4: 1 ms. RG: 38M7: 500/0 M16: —225 D1O: 1 ms. SW: 25 kHzM13: 1000 M17: —500 D14: 9 m. slice: 880 urnFigure: 5.3(b)Pulse sequence: STEQ TE: 8.4 ms. TR: 500 rns.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —8064/126 M14: 4000 DO: 1 S. D3: 970 is.M4: 6000/0 MiS: 250 D4: 50 RG: 26M7: 3000/0 M16: —160 D1O: 1 rns. SW: 100 kHzM13: 3800 M17: —250 D14: 1 ms.Figure: 5.4Pulse sequence: C4S TE: 5—80 ms. TR: 500 rns.SI: 256w TD: 128w NS: 4 Pulses: siric2Ml:—8064/126 M14: 4000 DO: 1 5. D3: 970 us.M4: 6000/0 MiS: 250 D4: 50 us. RG: 26M7: 3000/0 M16: —160 D10: 1 rns. SW: 100 kHzM13: 3800 M17: —250 D14: 1 ms.240Figure: 5.5Pulse sequence: C4S TE: 16 ms. TR: 1 s.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 • D3: 400 Is.M4: 2180/0 M15: 250 D4: 500 is. RG: 29M7: 2000/0 M16: —560 Dl0: 500 is. SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms.Figure: 5.6Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 S. D3: 400 .Ls.M4: 2180/0 Ml5: 250 D4: 500 . RG: 24M7: 2000/0 M16: 575 D1O: 500 JIs. SW: 20 kHzM13: 1800 M17: —250 Dl4: 9 ms. slice: 1.3 mmFigure: 5.7Pulse sequence: C4S TE: 16 ms. TR: 500 ms.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 S. D3: 400 J.Ls.M4: 2230/0 Ml5: 250 D4: 500 s. RG: 30 (H20)5 (DMSO)M7: 2000/0 M16: —500 D1O: 500 . SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms.241Figure: 5.8(a,b)Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 S. D3: 400 ,Is.M4: 2230/0 M15: 250 D4: 500 ,Is. RG: 30 (H20)5 (DMSO)M7: 2000/0 M16: —500 D1O: 500 SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms.Figure: 5.8(c)Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 S. D3: 400 Ls.M4: 2180/0 M15: 250 D4: 500 us. RG: 32 (H20)8 (DMSO)M7: 2000/0 M16: 575 D1O: 500 uls. SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms.Figure: 5.8(d)Pulse sequence: C4S TE: 16 ms. TR: 500 ms.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 5. D3: 400 us.M4: 2230/0 M15: 250 D4: 500 us. RG: 12 (DMSO)M7: 2000/0 M16: —500 D1O: 500 SW: 20 kHzM13: 1800 M17: —250 D14: 9 xns.242Figure: 5.9Pulse sequence: C4S TE: 5.1 ms. TR: 500 ms.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —8064/126 M14: 4000 DO: 1 S. D3: 970 Is.M4: 6000/0 M15: 250 D4: 50 RG: 26M7: 3000/0 M16: —160 Dl0: 2 ms. SW: 100 kHzM13: 3800 Ml7: —250 D14: 1 msFigure: 5.10Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 5. D3: 400 Ls.M4: 2180/0 M15: 250 D4: 500 us. RG: 32 (H20)8 (DMSO)M7: 2000/0 M16: —575 D1O: 500 u• SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms.Figure: 5.11Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 S. D3: 400 us.M4: 2180/0 M15: 250 D4: 500 us. RG: 25 (H20)M7: 2000/0 Ml6: 575 D1O: 500 SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms. slice: 1.3 mm243Figure: 5.12Pulse sequence: C4S TE: 16 ms. TR: 1 s.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 s. D3: 400 its.M4: 2180/0 M15: 250 D4: 500 .Ls. RG 25 (H20)M7: 2000/0 M16: —55 D1O: 500 ,Is. SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms. slice: 1.3 mmFigure: 5.13Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 S. D3: 400 j.is.M4: 2180/0 M15: 250 D4: 500 j.ts. RG: 15 (H20)1—10 (DMSO)M7: 2000/0 M16: 575 D10: 500 s. SW: 20 kHzM13: 1800 M17: —250 D14: 9 ins, slice: 1.3 mmFigure: 5.14Pulse sequence: C4S TE: 16 ins. TR: 15 s.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml: —2048/32 M14: 4000 DO: 1 S. D3: 400 ps.M4: 2180/0 M15: 250 D4: 500 JIs. RG: 25 (H20)M7: 2000/0 M16: 575 D1O: 500 ts. SW: 20 kHzM13: 1800 M17: —250 D14: 9 ms. slice: 1.3 mm244Figure: 5.15Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 256w TD: 128w NS: 4 Pulses: sinc2Ml:—2048/32 M14: 4000 DO: 1 S. D3: 400 I.Ls.M4: 2180/0 M15: 250 D4: 500 $.Ls. RG: 25 (H20)10 (DMSO)M7: 2000/0 M16: 575 D1O: 500 SW: 20 kI-IzM13: 1800 M17: —250 D14: 9 ms. slice: 1.3 mmFigure: 5.17Pulse sequence: C4S TE: 16 ms. TR: 500 ms.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 S. D3: 400 J.Ls.M4: 2125/0 M15: 250 D4: 500 s. RG: 40 (H20)20 (DMSO)M7: 2000/0 M16: —220 D10: 500 s. SW: 20 kHzM13: 1900 M17: —250 D14: 6 ms.Figure: 5.18Pulse sequence: C4S TE: 16 ms. TR: 500 mS.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 S. D3: 400 ILs.M4: 2125/0 M15: 250 D4: 500 RG: 40 (H20)20 (DMSO)M7: 2000/0 M16: —220 D1O: 500 Is. SW: 20 kHzM13: 1900 M17: —250 D14: 6 ms.245Figure: 5.20Pulse sequence: C4S TE: 16 ms. TR: 1 s.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 S. D3: 400 is.M4: 2125/0 M15: 250 D4: 500 jis. RG: 37 (H20)28 (DMSO)M7: 2000/0 M16: —220 Dl0: 500 SW: 20 kHzM13: 1800 M17: —250 D14: 6 ins.Figure: 5.21Pulse sequence: C4S TE: 16 ms. TR: 1 s.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 5. D3: 400 .Ls.M4: 2125/0 M15: 250 D4: 500 RG: 38 (H20)22 (DMSO)M7: 2000/0 M16: —220 D1O: 500 its. SW: 20kHzM13: 1900 M17: —250 D14: 6 ins.Figure: 5.22Pulse sequence: C4S TE: 16 ins. TR: 1 5.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 s. D3: 400 is.M4: 2125/0 MiS: 250 D4: 500 RG: 38 (2O)22 (DMSO)M7: 2000/0 M16: —220 D1O: 500 s. SW: 20 kHzM13: 1900 M17: —250 D14: 6 ins.246Figure: 5.23Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml:—2432/19 M14: 4000 DO: 1 S. D3: 400 Is.M4: 2125/0 M15: 250 D4: 500 RG: 38 (H20)15 (DMSO)M7: 2000/0 M16: —220 D1O: 500 Its. SW: 20 kHzM13: 1900 M17: —250 D14: 6 ms.Figure: 5.24Pulse sequence: C4S TE: 16 ms. TR: 500 ms.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 s. D3: 400 is.M4: 2110/0 M15: 250 D4: 500 Is. RG: 35 (H20)22 (DMSO)M7: 2000/0 M16: 240 Dl0: 500 ,.ts. SW: 20 kHzM13: 1900 M17: —250 D14: 6 ms.Figure: 5.25Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 S. D3: 400 jis.M4: 2150/0 MiS: 250 D4: 500 Its. RG: 35 (H20)M7: 2000/0 M16: —300 D1O: 500 . SW: 20 kHzM13: 1900 M17: —250 D14: 6 ms. slice: 1.3 nun247Figure: 5.26Pulse sequence: C4S TE: 16 ms. TR: 1 5.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 s. D3: 400 $Is.M4: 2125/0 M15: 250 D4: 500 is. RG: 38 (1420)15 (DMSO)M7: 2000/0 M16: —220 D1O: 500 s. SW: 20 kHzM13: 1900 M17: —250 D14: 6 ms.Figure: 5.27Pulse sequence: C4S TE: 16 ins. TR: 1 s.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 5. D3: 400 Is.M4: 2125/0 M15: 250 D4: 500 RG: 33 (H20)15 (DMSO)M7: 2000/0 M16: —220 D10: 500 . SW: 20 kHzM13: 1900 M17: —250 D14: 6 ins.Figure: 5.28Pulse sequence: C4S TE: 16 ins. TR: 1 S.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 S. D3: 400 .Ls.M4: 2125/0 MiS: 250 D4: 500 Is. RG: 33 (H2°)15 (DMSO)M7: 2000/0 M16: —220 D1O: 500 is. SW: 20 kHzM13: 1900 M17: —250 D14: 6 ins.248Figure: 5.29Pulse sequence: C4S TE: 16 ms. TR: 1 s.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 S. D3: 400 j.Ls.M4: 2125/0 M15: 250 D4: 500 Is. RG: 33 (H20)15 (DMSO)M7: 2000/0 M16: —220 D1O: 500 ps. SW: 20 kHzM13: 1900 M17: —250 D14: 6 ins.Figure: 5.30Pulse sequence: C4S TE: 16 ins. TR: 1 S.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 D3: 400 ps.M4: 2150/0 MiS: 250 D4: 500 ps. RG: 30 (H20)M7: 2000/0 M16: —300 D1O: 500 . SW: 20 kI-lzM13: 1900 M17: —250 D14: 6 ins.Figure: 5.31Pulse sequence: C4S TE: 16 ins. TR: 1 S.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 S. D3: 400 ps.M4: 2125/0 MiS: 250 D4: 500 ps. RG: 37 (H20)28 (DMSO)M7: 2000/0 M16: —220 D1O: S00 SW: 20 kHzM13: 1800 M17: —250 D14: 6 ins.249Figure: 5.32Pulse sequence: C4S TE: 16 ms. TR: 1 s.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 s. D3: 400 is.M4: 2125/0 M15: 250 D4: 500 is. RG 30 (H20)12 (DMSO)M7: 2000/0 M16: —220 D10: 500 ts. SW: 20 kHzM13: 1900 M17: —250 D14: 6 ms.Figure: 5.33Pulse sequence: C4S TE: 16 ms. TR: 1 S.SI: 512w TD: 256w NS: 4 Pulses: sinc2Ml: —2432/19 M14: 4000 DO: 1 S. D3: 400 jis.M4: 2110/0 M15: 250 D4: 500 is. RG: 28 (H20)5 (DMSO)M7: 2000/0 M16: —240 D1O: 500 ji SW: 20 kHzM13: 1900 M17: —250 D14: 6 ms.250APPENDIX BPHOTOGRAPHS OF MICROSURGERYThe following photographs were taken during surgery performed by BettyPearson, microsurgical specialist at Jack Bell Research Center, Vancouver.The camera was attached directly to the microsurgery microscope, and focussingwas accomplished by the surgeon. The photographs have varying degrees ofmagnification.251Figure B.1. The dissected aorta held between two forceps, with the renalartery on the right. Black threads are silk sutures used to tie off bloodvessels.252Figure B.2. A rat kidney, which has been subjected to a 250 ml gravity fedblood washout, and then catheterized. The silk sutures secure the catheterinto the renal artery.r253Figure B.3. A coronal slice through a rat kidney which has not been subjectedto a blood washout. There is still blood present (reddish color)254Figure B.4. A coronal section of a rat kidney which has been subjected to a250 ml gravity—fed blood washout. Comparison with the previous photographshows that visually, as well as in the NMR images, most of the contrastbetween renal structures is due to the presence of blood. The medullartissues here appear virtually featureless. Structures that are easilyidentifiable are the renal pelvis (center), the medulla (light colored) andthe cortex (darker)/9255Figure B.5. A transverse section through a rat kidney which has beensubjected to a 4 ml manually-controlled blood washout with a syringe. Noticethat there is still blood present in the medulla and inner cortex. The smallred dots in the cortex are cross sections of blood vessels which still containblood. This photograph is quite similar to the NMR images.256APPENDIX CPULSE SEQUENCES ID RELATED EQUATIONSThe following are modified pulse sequences used in this thesis for theimaging experiments. Both sequences were written for the Bruker MSL-400spectrometer, controlled by an Aspect 3000 computer.Equations used to calculate pulse bandwidth, slice thickness, and inplane resolution are also given.257CHEMICAL SHIFT SPECIFIC SLICE SELECTIVE (C4S) SEQUENCETAU=AQ/2 + D4START, LOOP Cl TIMESLOOP NS TIMESlOU [GX 0 GY 0 GZ 0]DOlOU [SYN1=@FL1]loulOU [GZ M15%]D4D14 [Fl @PLS1 ISIN SP1 M13%]lOU [GX M4% GY Ml% GZ M16%]TAUlOU [GX 0 GY 0]D1OlOU [GZ M17%]D4Dl4 [Fl @PLS2 ISIN SP1 M14%]D4lOU [GZ 0]Dl 0lOU [GX M7%]D4 [++PLS1 ++PLS2]D3 [STA]AQEND LOOP5OU [++FL1]END LOOP10 OU++M16% ++M17%GOTO STARTBEGIN LISTSPLS1, +X —Y —x +yPLS2, +X +YRLS, +X -Y -X +YFL1, (SF+Ol)(SF—Cl)END LISTS258SHORT TIME TO ECHO (STE) SEQUENCETAU=AQ/2 + D4START, LOOP NS TIMESlOU [GX 0 GY 0 GZ 0]DOlOU [SYN1=SF+O1]lOUlOU [GZ M15%]D4D14 [Fl @PLS1 ISIN SP1 M13%]lOU [GX M4% GY M1% GZ M16%]TAU [++PLS1]lOU [GX 0 GY 0 GZ 0]Dl 0lOU [GZ M15%]D4D2 [Fl @PLS2]lOU [GZ 01DiDlOU GX M7%D4 [SYN1=SF]D3 [STA)AQEND LOOP10 OU++M16%GOTO STARTBEGIN LISTSPLS1, +x —Y -x +YPLS2, +Y +XRLS, +X -Y -X +YEND LISTS259EQUATIONSPulse bandwidth of sinc(x)-shaped RE’ pulses:if =D14where n is the number of lobes of the sinc pulse shape, andD14 is the length of the pulse in ms.Slice thickness:1fSlice thickness =k Gwhere is the gyromagnetic constant in kHz/Gauss, G isthe strength of the linear magnetic field gradient along thez axis of the sample in Gauss/cm.In—plane resolution:IPR = SW/TD“‘H Gwhere SW is the sweep width, TD is the number of points inthe time domain, Yff is the gyromagnetic constant inHz/Gauss, and G is the strength of the linear magneticfield gradient along the x axis of the sample in Gauss/cm.260APPENDIX DHISTOLOGY OF SELECTED SNPLESAll histological examination and identification of kidney structures inthis appendix was performed by Dr. John English of Vancouver Hospital andHealth Sciences Center. Figures D.1—D.4 are representative of the sampleswhich were subjected to perfusion with 20% DMSO in UW solution, freezing, andDMSO washout. Photographs were chosen to show the best and worst outcomes (interms of structural preservation) of these experiments. In Dr. English’sopinion, none of the kidneys would have survived after transplantation.However, some of the structures were well preserved.261Figure D.l. General structural features of the kidney at 25X magnification,including medulla (t’4) and cortex (C) with numerous glomeruli (G) which are theblood filtration units of the nephrons, and an arcuate artery (A) and vein(V) . The vein appears dilated, which suggests that the perfusate has passedthrough the vascular system under pressure. This kidney is shown in Figure5.23, 5.26, and 5.27, and has been subjected to 10°C perfusion with 20% DMSOin UW solution, freezing for 1 hour, and a DMSO washout with OW solution at10°C.262FigureD.2. Smaller structures of the kidney at 100X magnification, includingan artery (A) which branches into an arteriole above it, next to a lymphatic(white). A glomerulus is visible as well. Most of the other structures areproximal convoluted tubules, and show pathological changes consistent withacute renal ischemia (dead or dying cells are darker)263Figure D.3. General structure of a kidney subjected to perfusion at 34°C with20% DMSO in UW solution, shown at 25X magnification. The perfusion failed,and the kidney was sectioned for histological examination. Comparison withFigure D.l shows the same architecture, but indications that autolysis ofcells is advanced, and cells are losing their nuclei.264Figure D.4. A view of the previous slide at 100X magnification, showingsevere autolytic damage evidenced by loss of cell nuclei and disruption ofcell membranes. Ischemic changes are also present: Condensed proteinaceousmaterial within Bowman’s space (B) of the glomerulus, and dilated tubules.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0059650/manifest

Comment

Related Items