ORIGINAL ARTICLE 
Annals of Nuclear Medicine Vol. 11, No.4, 291-297 1997 
Subcellular distribution of thallium: Morphological and quantitative study in rat myocardium 
Mitsutaka FUKUMOTO, Daisuke YOSHIDA and Shoji YOSHIDA 
Department of Radiology, Kochi Medical School 
The purpose of this study is to determine the subcellular distribution of thallium (SDTl) by electron microscopy and a newly designed fixation method that makes insoluble grains of Tl visible. Methods: To obtain the high dose necessary for electron microscopic visualization, we employed TlCl instead of 201TlCl. EM was performed in fixed rat myocardium resected at 20 min (early phase) and 3 hr (delay phase) after intravenous injection of TlCl. To fix TI in the cell, we used orthovanadate in our fixative. Atomic absorption spectroscopy (AAS) of Tl and quantification of subcellular distribution of 201Tl (SD201Tl) were studied to prove the propriety of our fixation. Results: AAS detected Tl in the Tl-loaded specimen but not in the control, indicating that TI was the origin of the grains observed in the former. In the early phase, numerous grains were observed in mitochondria, sarcoplasmic reticulum (SR), myofibrils, and nuclei, but no such grains were visible in controls. In the delay phase, grains were retained in mitochondria, SR and nuclei, but not in myofibrils. Electron microscopic SDTl (%) correlated with SD201Tl(%) calculated from isolated fractions. Conclusion: In both the early and delay phases, mitochondria are the major site of Tl and 201Tl uptake. 
Key words:thallium, myocardial cell, mitochondria, atomic absorption spectroscopy, electron microscopy 
INTRODUCTION 
THE ASSESSMENT of coronary blood flow and myocardial viability has become an important clinical parameter in this era of aggressive intervention and surgery. Evaluation of salvageable myocardium, region at risk, and areas of reversible dysfunction have both diagnostic and prognostic value in managing cases subjected to revascularization procedures. The increasing interest in and use of Thallium-201 (201Tl) in myocardial imaging during the last two decades have therefore prompted the undertaking of further studies to better understand the biological properties and detailed mechanism of accumulation of 201Tl. Thallium (TI and 201Tl) is similar in size to potassium and a certain proportion (approximately 50-60%) enters myocardial cells in active transport by means of an Na+-K+ ATPase dependent exchange mechanism. 1-3 Actually, 
Received April 21, 1997, revision accepted July 28, 1997. For reprint contact: Mitsutaka Fukumoto, M.D., Department of Radiology, Kochi Medical School, Kohasu, Okoh, Nankoku, Kochi 783, JAPAN. 
201Tl (or Tl) uptake is significantly high in tissues with a high concentration of Na+-K+ ATPase and mitochondria: myocardium, liver, and renal tubules. On the other hand, Tl-toxicity is characterized by mitochondrial swelling, inhibition of the mitochondrial electron transport system, and damage to respiratory enzymes.4-6 The transportation of Tl passing through the ion channel present in organelles has been examined.7-10 TI is not necessary tightly bound at the target sites in vivo because Tl has soluble and diffusible properties.11-13 Conventional fixing, dehydration and embedding methods are chemically disruptive, and cannot be applied to diffusible compounds without risk of undesirable ionic redistribution or loss of radioactivity. In some soluble materials these problems have been solved by modification and embedding techniques as well as by using dry emulsion strips instead of a liquid emulsion but because these cytochemical difficulties have not been overcome, there is no conclusive report focusing on the electron microscopic findings in subcellular Tl. This situation prompted us to devise a new cytochemical method to fix TI, with which we studied the subcellular distribution of thallium (SDTl) by electron microscopy (EM). We used atomic absorptive spectroscopy (AAS) to confirm the difference between Tl-loaded myocardium and control, namely to confirm that the grains are the result of Tl administration, and to demonstrate the retention of TI in fixed myocardium. Furthermore, to exclude the redistribution of Tl in fixed myocardium, we measured 201Tl uptake into the cellular fractions with 201Tl as the tracer. 
MATERIALS AND METHODS 
TlCl for electron microscopy 
The 201Tl content of the commercially available 201TlCl (111MBq) solution was 14.04 ng. It is not possible to theoretically and absolutely detect the localization of 201Tl with such a small amount of 201Tl taken up into the myocardial cell. To visualize the localization of Tl electron microscopically, we therefore employed non-radio-active TlCl, which is more easily used in large quantities, as a substitute for 201TlCl. 
Sample preparation for electron microscopic observation 
We used 12-week-old female, 250 g, Sprague-Dawley rats. At 3 hr and 20 min prior to killing, each rat was given a slow intravenous injection of 2 ml of TI solution ( 10 mM TlCl/0.15 M sucrose) to avoid hypervolemia. Control rats received 2 ml of 0.15 M sucrose via the tail vein. Rats anesthetized with ether were transcardially perfused with fixative-1 [2% glutaraldehyde + 10 mM EDTA/0.1 M cacodylate buffer (pH 7.2) with 5% sucrose] and with fixative-2 (1 mM orthovanadate sodium/fixative- 1). Our fixation method was devised to prevent the physiological influence of metallic vanadiuml4-16 by prior perfusion with fixative-1. During the fixation procedure, the descending aorta was clamped to maintain the perfusion pressure of fixative-1 and -2 in the coronary arteries. The hearts were then resected out and dissected into 2 mm thick sections, which were immersed and fixed in fixative-l at 4deg.C for 2 hr. Thereafter the specimens were sliced into 50 um-thick sections in EDTA solution ( 10 mM EDTA/ 0.1 M cacodylate buffer with 5% sucrose, pH 7.2) with a microslicer (DSK3000, Dosaka, Japan). The thin-sliced tissue sections were rinsed by immersion and gentle shaking in the EDTA solution as described above for at least 4 hr at room temperature. After the shaking, the tissue sections were immersed in 0.1 % lead acetate/acetic acid (pH 3.9) for 10 min. After 10 min rinsing with distilled water, the specimens were osmicated with 1%(; OsO4 in 0.1 M sodium cacodylate buffer (pH 7.2) for 10 min, and embedded in Epon 812 after dehydration in a graded series of ethanol and propylene oxide. Poststaining was carried out with uranium acetate. Electron microscopy was carried out with a HITACHI H-700. 
Qualitative analysis of the grains 
In a test tube a white-turbid precipitate was obtained by mixing 10 mM TlCl aqueous solution and 1 mM sodium orthovanadate aqueous solution. We prepared this precipitate intracellularly aiming to make it visible by aggregating lead (Pb). This mixture was then centrifuged at 3,000 rpm for 10 min. After centrifugation, the precipitate was stirred in 10 mM EDTA solution (pH 7.2) and recentrifuged under the same conditions. Finally, the Tl-V precipitate in EDTA solution was treated with 0.1% lead acetate/acetic acid (pH 3.9), and this mixture was recentrifuged under the same conditions to obtain the final precipitates made from Tl. V and Pb. The Tl content of the final precipitate was then determined by AAS. To demonstrate that the grains in the EM specimens contained Tl and were not merely non-specific Pb grains, analyses were performed by means of AAS. To show that the Tl was retained in the myocardium by our fixation, fixed and unfixed myocardial sections preserved for 7 days and respective preservative buffers ( 10 mM EDTA/ cacodylate buffer, pH 7.2) were investigated by AAS. At this time, to exclude non-specific precipitation of Pb, AAS of Pb was performed simultaneously. Flameless AAS was performed with an HITACHI Z-9000, with which atomic absorption of multiple elements can be examined simultaneously. In order to analyze Tl, each sample was treated in the following temperature program: 1) Dissolving of samples in 0.1 N HNO3; 2) Drying at 65-105deg.C for 40 sec and ashing at 900deg.C for 30 sec; 3) Atomization at 1 ,800deg.C for 8 sec; 4) Cleaning at 3,000deg.C for 3 sec; and 5) Using a wavelength of 276.8 nm. As for AAS of Pb, the program was as follows: 1) Dissolving of samples in 0.05 N HN03; 2) Drying at 70-110deg.C for 40 sec and ashing at 900deg.C for 20 sec; 3) Atomization at 1,800deg.C for 10 sec; 4) Cleaning at 3,000deg.C for 3 sec; and 5) Use of 283.3 nm wavelength. These conditions were determined by making reference to the TECHNICAL DATA AA sheet Nos.61,62 (issued by Hitachi, Ltd., Japan). 
Quantlfication of 201Tl in cellular fractions : SD201Tl (%) 
The non-physiological redistribution of subcellular Tl cannot be excluded by AAS. It was therefore necessary to demonstrate that 201Tl was fixed without redistribution by determining the 201Tl uptake ratio into the cellular fractions of fixed and unfixed 201Tl-loaded myocardium. First, fixed and unfixed myocardium from each of the rats administered 201TlCl ( 1.0 MBq) was extracted 20 min after administration (early phase) and 3 hr thereafter (delay phase). Fixation was performed in the same manner as at the time of preparation of the EM specimens. By several stages of ultracentrifugation including density gradient ultracentrifugation, the nuclei, mitochondria, sarcoplasmic reticulum (SR) and cytosol fractions of the fixed and unfixed myocardium were isolated. Cellular fractionation was performed according to the previously established methods.17-21 The 201Tl radioactivity in each fraction was measured with a well gamma counter (Aloka, ARC-300, Japan). Subcellular distribution of 201Tl in each fraction [SD201Tl (%)] was defined as follows: 
RESULTS 
Electron microscopic observations 
We observed normal myocardial cell structures and no grains in our EM study of control specimens (Fig. 1). On the other hand, numerous grains showing TI were diffusely distributed in the early fixed Tl-loaded myocardial cells (Fig. 2). The administration of Tl can be implicated in this difference between two specimens treated under the same experimental procedure. During a 3 hr observation period, grains within mitochondria, sarcoplasmic reticulum (SR), and nuclei were retained (Fig. 3). The amount of distribution in the myofibrils showed a decrease in the delay phase and the amount of grains in nuclei seemed to be comparatively small and constant during our observation. The most remarkable finding in the delay phase was the noticeable decrease in the number of grains distributed in myofibrils. To analyze these grains quantitatively, the number of grains/unit area in each organelle was counted by using the NIH image (version 1.6) in a personal computer (Power Macintosh 7300/166). Fifteen ROI of the same size were set on each organelle on the EM images, and the SDTl (%) to each organelle was determined from the distribution density of grains within the ROI. In SR, the area of which is smaller than that of other organelles, a few small ROI were totaled and the SDTl (%) determined as the number of grains/unit area. Quantitative SDTl data showed a relative increase of in the number of grains (34%-43%) within mitochondria, but no significant change in grains within nuclei (Fig. 4). A statistically significant decrease in grains (35%- 18%) was noted in myofibrils (Fig. 4). 
Atomic absorption spectroscopy 
The chemical precipitates had a high absorbance of Tl, and the concentration of Tl was 124.7 mg/g the final dried precipitates. Although the TI-loaded myocardial sections also showed a high absorbance of Tl, the preservative buffer did not (Fig. 5), but Tl which oozed from the myocardial sections was detected in preservative buffer used for unfixed setting (Fig. 5). Furthermore, the Tl-loaded myocardial specimens showed absorbance of Tl and Pb, whereas the control specimens showed absorbance of neither (Fig. 5). This clearly indicates that the numerous grains observed in Tl-loaded myocardium are not Pb-depositions. These grains can be considered to contain Tl since an injection of TlCl was the only treatment to yield such different results for Tl-loaded specimens and controls. Theoretically, although these AAS findings could be explained by the retention of Tl in tissues by our fixation, they do not prove that Tl was fixed without redistribution in the intracellular portion. 
Quantification of 201Tl in cellular fractions: SD20lTl(%) 
A noticeably different pattern distinguished the fixed- and unfixed myocardial cells (Fig. 6). In the fixed myocardium, 201Tl was mainly present in the mitochondria fraction, but in the unfixed myocardium, the highest 201Tl activity was noted in the cytosolic fraction, implying that the difference in patterns was caused by unfixed 201Tl. In the fixed tissue, 201Tl was tightly fixed with tissue by our fixative method, but in the unfixed condition a considerable amount of unfixed 201Tl shifted to the cytosolic part. The data for SD201Tl in the fixed setting correlated with those for SDTl determined by electron microscopic observations (Compare Figs.4 and 6). 
DISCUSSION 
Increased interest in the distribution and dynamics of thallium is due to the widening application of 201Tl in fields such as membrane biology in addition to fields of nuclear cardiology and oncology, but because the cytochemical fixation of TI has been difficult, no studies on the morphological distribution and localization of TI have been reported. The basic explanation why electron microscopic autoradigraphy has not been applicable to the study of the distribution of 201Tl is also related to the difficulties. With this in mind, we devised a method in which Tl is precipitated as grains within the cell and the localization of Tl made visible electron microscopically, and demonstrated the localization of Tl taken up in rat myocardial cells. Analysis of the distribution of grains clarified that 20 min after administration Tl is diffusely distributed within the cell, while after 3 hr it is retained in the nuclei, SR and mitochondria or is relatively decreased, whereas the amount of distribution in the myofibrils shows a decrease. One problem remaining to be solved in our EM study is the difficulty in morphologically determining the cytosolic part, making evaluation of Tl distribution in the cytosol imperfect. AAS can prove the retention of Tl in tissue attained with our newly developed fixation method, but cannot exclude the possibility of intracellular redistribution of Tl. Nevertheless, direct assay of 201Tl uptake into the 
cellular fractions demonstrated a similarity between distribution of 201Tl in fixed myocardium and that of Tl determined by EM, suggesting that our fixation method did not cause unphysiological redistribution of Tl. On the other hand, the high uptake of 201Tl in unfixed myocardial cytosol suggests the possibility that unfixed 201Tl shifted to the cytosol. In unfixed myocardium, the finding that the pattern of distribution of 201Tl was to a certain extent preserved is thought to be attributable fact that a certain proportion of 201Tl and Tl weakly binds to bases such as -SH, -OH or -NH3 present within the organelles even under physiological conditions.4-6 Our findings are largely consistent with those of previously reported studies on the relation between Tl and organelles. In addition to proving morphologically for the first time these previous findings, our study provides new information on the dynamics of Tl in the myocardial cell. In a study comparing Tl-loaded and potassium-loaded frog sartorial muscles, Edelmann22 used freeze substitution and low temperature embedding to show that dark-stained myosin filaments indicated the presence of Tl, but failed to adequately address the distribution of Tl in nuclei, mitochondria, SR and the cytosolic part. In good agreement with his results, we showed an abundance of grains, indicating the localization of Tl within myofibrils in the early phase, but our result clearly showed retention of Tl within nuclei, mitochondria and SR with time, as well as a marked decrease in Tl in myofibrils. Regarding the relation between Tl and SR, Fox's group7-9 also electro-physiologically ascertained that the SR channel is permeable to Tl. In our experiment as well, we could visually confirm the distribution of Tl into the SR, thus providing morphological evidence of their results. Considering that to respond to the energy requirements of the myocardium, mitochondria occupy more than 40% of the myocardial cell volume,23 it is reasonable to hypothesize that mitochondria are the main site of Tl distribution. Hitherto the following studies focusing on the relation between mitochondria and Tl/201Tl have been reported: 1) staining of living yeast mitochondria by Tl,24 2) relation between 201Tl uptake and mitochondria density in parathyroid tumor,25 and 3) transport and uptake of Tl in mitochondria.26.27 Furthermore, Uccelli's group investigated the localization in myocardial cells of 99mTc-MIBI and 99mTc-N-NOEt, and they explained discordant localization by differences between the electric charges in the two tracers, with the former a monovalent cation accumulating in mitochondria and the latter electrically neutral accumulating in the hydrophobic component (cell membrane) 28 Like 99mTc-MIBI both Tl and 201Tl are monovalent cations, and since their ionic radius is smaller than that of 99mTc-MIBI it is reasonable to speculate that Tl and 201Tl also accumulate in mitochondria. The significance of the present study lies in our demonstrating morphologically the presence of Tl within mitochondria and outlining the involvement of mitochondria in the distribution of 201Tl and Tl. Although the distribution of Tl and 201Tl in myocardial cells determined in our study and that of 201Tl in tumor cells reported by Ando's group29 differ, it is not necessary to discuss here the reasons for the discrepancy between these two sets of results. Of course, our present data cannot immediately be applied to 201Tl kinetics in tumor cells, because myocardial cells differ from tumor cells in not retaining the property of cell division. 
CONCLUSION 
To conclude, we note three important experimental findings. First, 20 min after administration, Tl and 201Tl show diffuse distribution in the nuclei, mitochondria, SR and myofibrils of rat myocardial cells. Second, after 3 hr the distribution of Tl and 201Tl in myofibrils decreased whereas that in other organelles was retained, with in particular a relative increase noted in distribution to mitochondria. Third, although the distribution of Tl and 201Tl to the cytosol cannot be disregarded, myocardial mitochondria could be said to comprise the major site of the distribution of Tl and 201Tl. 
ACKNOWLEDGMENTS 
This study was supported by grants from the Japanese Ministry of Education which made possible the achievement of this work. The authors wish to express their gratitude to the officials responsible for this valuable assistance. The authors are also grateful to KN International, Ltd. (Ohio, USA) and Mr. John Gelblum (Kanazawa) for their scientific and linguistic review of this paper. 
REFERENCES 
1. Mullins LJ. Moore RD. The movement of thallium ions in muscle. J Gen Physiol 43: 759-773, 1960.
2. Gehring PJ, Hammond PB. The uptake of thallium by rabbit erythrocytes. J Pharmacol Exp Ther 200: 215-221, 1964.
3. Britten JS, Blank M. Thallium activation of the (Na+-K+) activated ATPase of the rabbit kidney. Biochim Biophys Acta 159: 160-l66, 1968.
4. Hasan M. Chandra SV, Bajpai VK, Ali SF. Electron microscopic effects of thallium poisoning on the rat hypothalamus and hypocampus: Biochemica-changes in the cerebrum. Brain Res Bull 2: 255-261, 1977.
5. Hasan M. Electron microscopic study of the effects of thallium poisoning on the rat cerebellum. Forensic Sci Int 11: 139-146, 1978.
6. Woods JS. Fowler BA. Alteration of hepatocellular structure and function by thallium chloride: ultrastructural, morphometric and biochemical studies. Toxicol Appl Pharmacol 83: 218-229, 1986.
7. Fox J. Conductance and selectivity properties of a substate of the rabbit sarcoplasmic reticulum channel. Biophys J 47: 573-576, 1985. 
8. Fox J. Thallous ion permeation through the cation selective channel of the sarcoplasmic reticulum: Anomalous mole fraction dependence. Biochim Biophys Acta 736: 241-245, 1983.
9. Fox J. Experimental and theoretical studies on Tl interactions with the cation selective channel of the sarcoplasmic reticulum. J Membr Biol 84: 9-23, 1985.
10. Brismar T. Lefvert A, Jondal M. A method for analysis of cellular K-transport mechanisms through thallium (201Tl) uptake in human lymphocytes. Acta Physiol Scand 150: 299-303, 1994.
11. Mattew RBP. Blower PJ. Radionuclide targeting and dosimetry at the microscopic level: the role of microautoradiography. EurJ Nucl Med 21: 1347-1365, 1994.
12. Edelmann L. Subcellular distribution of potassium in striated muscles. Scanning Electron Microsc 2: 875-888, 1984.
13. Edelmann L. Freeze-substitution and the preservation of diffusible ions. J Microsc 161 : 217-228, 1991.
14. Cantley LC Jr, Josephson L, Warner R, Yanagisawa M. Vanadate is a potent (Na,K)-ATPase inhibitor found in ATP derived from muscle. J Biol Chem 252: 7421-7423, 1977.
15. Heyliger CE, Tahiliani AG, McNeill JH. Effect of vanadate on elevated blood glucose and depressed cardiac performance of diabetic rats. Science 227 (4693): 1474-1477, 1985.
16. Sakurai H. Tsuchiya K. Nukatuka M, Sofue M, Kawada J. Insulin-like effect of vanadyl ion on streptozotocin-induced diabetic rats. J Endocrinol 126: 451-459, 1990.
17. Mitchell RD, Palade P, Fleischer S. Purification of morphologically intact triad structures from skeletal muscle. J Cell Biol 96: 1008-1016, 1983.
18. Rosemblatt M. Hidalgo C, Vergara C, Ikemoto N. Immunological and biochemical properties of transverse tubule membranes isolated from rabbit skeletal muscle. J Biol Chem 256 (15): 8140-8148, 1981.
19. Mitchell RD, Volpe P, Palade P, Fleischer S. Biochemical characterization, integrity, and sidedness of purified skeletal muscle triads. J Biol Chem 258: 9867-9877, 1983.
20. Sharp AH. Imagawa T. Leung AT, Campbell KP. Identification and characterization of the dihydropyridine-binding subunit of the skeletal muscle dihydropyridine receptor. J Biol Chem 262: 12309-12315, 1987.
21. Meissner G. Adenine nucleotide stimulation of Ca2+ release in sarcoplasmic reticulum. J Biol Chem 259: 2365-2374, 1984.
22. Edelmann L. Freeze-dried embedded specimens for biologica microanalysis. Scanning Electron Microsc IV: 1337-1356, 1986.
23. Junqueira LC, Carneiro J, Kelley RO. Basic Histology, 8th ed., pp, 181-201, 1995.
24. Lindegren CC, Lindegren G. Thallium as a vital stain for yeast mitochondria. Nature 234: 297, 1971.
25. Sandrock D, Merino MJ. Norton JA. Neumann RD. Ultra-structural histology correlates with results of thallium-201/Technetium-99m parathyroid subtraction scintigraphy. J Nucl Med 34: 24-29, 1993.
26. Saris NEL, Skulskii IA, Savina MV. Glasunov VV. Mechanism of mitochondrial transport of thallous ions. J Bioenergetics Biomembr 13: 51-59, 1981.
27. Skulskii IA, Saris NEL, Savina MV. Glasunov VV. Uptake of thallous ions by mitochondria is stimulated by nonactin but not respiration alone. Eur J Biochem 120: 263-266, 1981.
28. Uccelli L, Giganti M, Duatti A, Bolzati C, Pasqualini R, Cittanti C, et al. Subcellular distribution of Technetium-99m-N-NOEt in rat myocardium. J Nucl Med 36: 2075-2079, 1995.
29. Ando A, Ando I. Katayama M, Sanada S, Hiraki T, Mori H, et al. Biodistribution of 201Tl in tumor bearing animals and inflammatory lesion induced animals. Eur J Nucl Med 12: 567-572, 1987.