Experimental Iron Overload Ultrastructural Studies

Experimental Iron Overload Ultrastructural Studies" THEODORE C. IANCU AND HANNA SHILOH Pediatric Research Unit Carmel Hospital 34362 Haifa, Israel and Faculty of Medicine Technion-Israel Institute of Technology 31096 Haifa, Israel

INTRODUCTION Neither the role of iron nor the mechanism of cellular damage has been fully elucidated in primary (idiopathic, genetic) hemochromatosis (IHC), secondary, erythropoietic hemochromatosis (EHC), or transfusional hemochromatosis. Attempts to reproduce the features of human disease in experimental animals have been hitherto only partially successful, because laboratory animals seem to be quite resistant to iron overload, at least during the relatively short period of exposure to surplus iron. Although rather static in nature, electron microscopic studies have nonetheless contributed to the understanding of cytosiderosis. This overview summarizes the information gained from previous experiments and describes ultrastructural features of the newer models.

WHAT ULTRASTRUCTURAL ABNORMALITIES SHOULD BE REPRODUCED BY THE IDEAL EXPERIMENTAL MODEL? The ideal experimental model should present clinical, biochemical, and morphological abnormalities similar to those of humans with chronic iron-loading conditions. In the absence of spontaneous overload syndromes in animals similar to human hemochromatosis, active administration of exogenous iron provided information concerning ferritin synthesis, intracellular transport, distribution of iron storage compounds in various cells and organs, mechanisms of iron elimination, as well as data on fibrogenesis and other forms of iron toxicity. In order to appreciate the value of information obtainable from experimental models, a summary of ultrastructural findings common to various human ironloading conditions may be helpful. a This work was supported in part by the Milman Fund for Pediatric Research. The model of the carbonyl-iron-fed rat was studied at the MRC-Clinical Research Centre, Harrow, U.K., in collaboration with Drs. R. J. Ward and T. J. Peters and supported by the Nuffield Foundation and the Wellcome Trust.

164

IANCU & SHILOH: ULTRASTRUCTURAL STUDIES

165

F'IGURE 1. (a) Low-magnification electron micrograph from a liver biopsy of a patient with precirrhotic IHC. A few electron-opaque granules are seen in hepatocytes (arrows).No iron accumulations are found in the sinusoidal cells (broad arrow). Unstained, magnification 4oOox. (b) In iron-injected baboons, most siderosomes are found within sinusoidal cells, with little parenchymal cell involvement, at least in the initial stage of overload. Unstained, magnification 6OOox.

166

ANNALS NEW YORK ACADEMY OF SCIENCES

Ferridn Synthesis, Transport into Lysosomes, and Hemosiderin Formation Cells exposed to an iron-rich milieu react by incorporating inorganic iron into apofemtin, resulting in synthesis of holofemtin. The core of iron oxyhydroxide located within the protein shell is electron-opaque and permits the identification of femtin particles by electron microscopy. It is generally agreed that this ironprotein assembly takes place in the cytosol, in which a plateau of femtin concentration is reached during overload, specific for various cell lines. In parallel, ironcontaining fenitin particles are transported into lysosomes (siderosomes) and sequestrated as femtin and/or hemosiderin. The latter is considered a degradation product of femtin and has a heterogeneous composition, reflected by its ultrastructural appearance.’” With ongoing iron overload, there is progressive increase in the number of siderosomes. Any successful experimental model should exhibit the features of cytosiderosis, but the amount of cytosol femtin versus iron storage compounds in siderosomes may vary. Variability in the severity of cytosiderosis among different cells and organs is also typical. Some features of iron overload as found in IHC and in an experimental model are shown in FIGURES 1 and 2.

Iron and Cellular Damage Acute ionic iron toxicity is well established and can produce cell death. In contrast, toxicity and cell damage produced by the protein-bound iron compounds femtin and hemosiderin, which accumulate during chronic overload, are not easily demonstrable. The ultrastructural features of cellular damage, as seen in humans, include changes in organelles (other than siderosomes); extreme cytosiderosis, which may displace and replace normal cellular content beyond recognition; and fibrogenesis. There are noticeable differences between various cell lines exposed to iron: macrophages have a much greater capacity to accumulate storage iron than parenchymal cells. Increased fibrogenesis appears as collagen fiber deposition, even in the early stages of ~ v e r l o a dIn . ~ humans, a critical level of liver iron concentration has been established beyond which cirrhosis develop^,^ but such a figure is not available for experimental animals.

THE EARLY EXPERIMENTAL MODELS Parenteral Iron Administration The initial electron microscopic studies used injected iron-dextran or ironsorbitol-citric-acid complex, which produced rapid, but nonlethal, liver iron overload.67 Several such investigations showed that rats react to the injected iron by increased femtin synthesis and its transport into lysosomes (siderosomes). None of these studies reported liver cirrhosis. The involvement of other organs in the process of cytosiderosis has not been studied in detail. . ~ that the intraperitoneal injection of repeated In 1979, Awai et ~ 1 reported doses of iron nitrilotriacetate (Fe-NTA) in rats and rabbits produced marked iron deposits in the pancreas and liver, mainly in parenchymal cells. This was in contrast to the finding of most of the iron in Kupffer cells in colloidal iron-injected , ~ Fe-NTA overload animals. It has also been shown, by Parmley et ~ l . that reaches parenchymal cells more than iron-dextran. However, no excessive liver

IANCU & SHILOH:ULTRASTRUCIWRAL STUDIES

167

FIGURE 2. (a) Similarly to thalassemic children, hepatocytes of patients with IHC contain femtin particles that form arrays (arrow). Such siderosomes are extremely rare in cells other than hepatocytes. Unstained, 100,OOOx magnification. (b) In baboons, interhepatocytic collagen bundles (arrows) are conspicuous after I8 months of iron-dextran injections. Unstained, magnification 8000 x .

168

ANNALS NEW YORK ACADEMY OF SCIENCES

fibrosis or cirrhosis was noted in any of these experiments. Because of the good tolerance of experimental animals to Fe-NTA and its capability to produce parenchymal overload, this model was further used to establish the importance of pathogenetic factors such as lipid peroxidation and lysosomal fragility. Adult baboons, being closer to humans, were considered better experimental models. Intramuscular administration of iron-dextran for 18 months produced elevated liver enzymes, cytosiderosis of sinusoidal and later also of parenchymal liver cells, and excessive interhepatocytic collagen deposition. However, 6 months after the cessation of iron administration, the excessive collagen had disappeared. Thus, even though conspicuous by light and electron microscopy, the fibrosis did not interrupt the liver microcirculation and cirrhosis did not develop. It appears that in this model the duration of iron overload may have been too short to reach the irreversible cirrhogenic lesion.I2 The fact that iron was in the form of injected iron-dextran and therefore accumulated in the reticuloendothelial cells before its redistribution to parenchymal organs (FIG.lb) may have also contributed to reduction in toxicity to hepatocytes and fibrogenesis. Nutritional Overload Since both in early EHC, e.g., homozygous P-thalassemia, and in IHC the excess iron originates in the gastrointestinal tract, it was only natural that efforts were directed towards producing “nutritional” iron overload. Significant liver iron overload was reported in rats when given ethionine,I3 or when fed a cholinedeficient diet.I4 In these models of liver toxicity excess iron is only an accessory agent. A more physiological approach was that of Richter,l5 who produced massive siderosis in rats by alternating periods of feeding and starvation. Nevertheless, as with other models, no liver cirrhosis was produced. Cell Cultures An additional experimental model was the use of cell cultures. Jacobs et al. studied Chang cells exposed to an iron-enriched medium and described ferritin synthesis and hemosiderin accumulation in siderosomes.16 Iron toxicity appeared to be induced within 3 weeks and was irreversible unless the cells were transferred into an iron-poor medium.I7 While important for the study of cytosiderosis and iron toxicity, the study could not mimic all the in uiuo pathogenetic steps of hemochromatosis.

THE RECENT MODELS Parenterally Administrated Iron By injecting iron-dextran or heat-denatured homologous red cells intraperitoneally, Weintraub et a1.I” documented significant increases in prolyl hydroxylase activity in the hepatic homogenates of iron-loaded rats. No light microscopic histologic evidence of cell necrosis or excessive collagen were found in the livers of the animals. However, electron microscopy revealed collagen fibrils immediately adjacent to hepatocytes in the iron-loaded rats but not in controls. These findings were similar to our observations on liver biopsies of infants with homozy-

IANCU & SHILOH: ULTRASTRUCTURALSTUDIES

169

gous p-thala~semia.~J~ They support the concept that in vivo iron overload may have a direct effect on stimulating increased prolyl hydroxylase activity and hepatocytic collagen synthesis. Cell Cultures Additional information concerning the toxicity of iron for the heart was obtained by sequentially examining cultured rat heart myocytes after exposure to a medium containing ferric ammonium These cells exhibited abnormal contractility2' and increased lipid peroxidation.22Electron microscopic examination of the iron-exposed myocytes showed features of premature aging and other

FIGURE 3. Rat heart myocytes, cultured in a medium containing ferric ammonium citrate, show lysosomal accumulation of various electron-dense particles, including typical iron-rich femtin (arrow). Unstained, magnification 88,OOOx.

signs of cellular damage, in parallel with the capacity to synthesize various uoncontaining particles, including typical iron-rich femtin, some of which formed clusters or were sequestrated in siderosomes (FIG.3).

The Carbonyl-Iron-Fed Rat Carbonyl iron is an elemental form of iron. It has good bioavailability and extremely low toxicity, due to its slow rate of solubilization and prolonged absorption.23Rats fed carbonyl iron developed liver iron o ~ e r l o a d 'and ~ - ~possibly ~ increased hepatic fibrosis.25Isolated siderosomes showed increased fragility26

170

ANNALS NEW YORK ACADEMY OF SCIENCES

similar to that found in human subjects.27 Significant changes were also found in the polyunsaturated fatty acid content of the lysosomal and other organelle membranes when compared to dextran-injected control rats, with decreases in both arachidonic and linoleic acids. These changes, characteristic of lipid peroxidation, suggested that the enhanced hepatic lysosomal fragility associated with iron overload may be caused by iron-induced membrane lipid peroxidation (Ward, personal communication). All these observations suggested that the carbonyl-fed rat may be a useful model for the study of early ultrastructural changes in organs involved in IHC and EHC. We therefore studied rats fed carbonyl iron for 4, 6, and 15 months (in collaboration with R. J. Ward and T. J. Peters).28 An overall increase in iron content of liver, spleen, intestine, and, to a lesser degree, pancreas and heart was found. Separated liver siderosomes exhibited evidence of increased membrane instability. Although detailed description of our morphological findings is beyond the scope of this review, the following sections present the main observations. Ultrastructural Features of the Liver

Typical features of iron overload were found mainly in the periportal regions (acinar zone I, according to Rappap01-t~~). These included conspicuous cytosolic ferritin and accumulation of ferritin, mainly of the iron-rich type, within siderosomes. The latter organelles had a predominantly pericanalicular location and contained also variable amounts of hemosiderin (FIG.4). As in human subjects,30 ferritin-containing microvilli were seen protruding into bile canaliculi, but only a few free particles were observed in the lumen. In contrast to baboons12 no siderosomes were present in bile canaliculi at any stage of iron overload. Sinusoidal cells (Kupffer and sinusoidal lining cells) appeared less iron-laden, especially in the rats fed carbonyl iron for only 4 months. Even in later stages of the experiment, sinusoidal cells did not exhibit the well-known hemosiderin “clumps” found in the earlier models of injected iron. Excessive collagen deposition could not be documented either by light microscopy (Masson’s trichrome stain) or by electron microscopy. The Spleen

The spleen was the only examined organ that contained conspicuous electrondense iron in control rats. It should be noted, however, that in contrast to iron-fed rats, controls showed storage iron deposition only in dendritic macrophages of the red pulp, and the overall amount was small. Macrophages of carbonyl-iron-fed rats were not essentially different from macrophages of other organs (FIG.5 ) , but endothelial cells showed a remarkable capacity of iron transfer into siderosomes, demonstrated by a nearly complete absence of ferritin particles from the cytosol. The Small Intestine

Our interest in the ultrastructural features of small intestine derived from the importance of this organ as point of entry for excess iron.The intestinal mucosa is the putative location of the primary transport defect in IHC, and probably is also involved in the iron-loading anemias. Unstained, high-magnification micrographs

IANCU & SHILOH: ULTRASTRUCI'URAL STUDIES

171

FIGURE 4. Liver of rat fed carbonyl iron for 4 months. Siderosomes (arrows)tend to cluster in the pericanalicular area of hepatocytes. Some other lysosomes contain lipofuscin (broadarrows).BC, bile canaliculus; M ,mitochondrion. Unstained, magnification 80,OOOx.

enabled the identification of two types of ferritin particles in the enterocytes of the upper third of the intestinal villus: (1) iron-poor particles, randomly dispersed in the cytosol, with the exception of microvilli, and (2) iron-rich particles, assembled in clusters within the same enterocytes (FIG.6). Within the lamina propria, the only cells containing femtin and hemosiderin were macrophages, typically located in the upper third of the villus.

172

ANNALS NEW YORK ACADEMY OF SCIENCES

The Pancreas

There is a high frequency of diabetes mellitus among patients with IHC and EHC, which justifies the interest in the earliest ultrastructural changes of the pancreas in iron overload. The postmortem appearance of the pancreas in human subjects with longstanding iron overload is severe siderosis and replacement of the parenchyma by fibrotic tissue and fat. Both acinar and islet cells display variable degrees of cytosiderosis, but much less than the interstitial tissue. These observations were made on material retrieved from patients in the “end-stage’’ of the disease, and are not very helpful in providing information concerning the early stages. In the carbonyl-iron-fed rat, the pancreatic interacinar and interlobular inter-

FIGURE 5. Macrophage from spleen of rat fed carbonyl iron for 6 months. Large, iron-rich fenitin particles are seen in a membrane-limited siderosome (arrow). Smaller, coalesced particles (divested ferritin) fill a “leaky” siderosome (broad arrow). Unstained, magniiication 86,000~.

IANCU & SHILOH: ULTRASTRUCI'URAL STUDIES

173

FIGURE 6. Small intestine of rat fed carbonyl iron for 15 months. Epithelial cells contain iron-poor fenitin in the cytosol and iron-rich fenitin in siderosomes (arrow). Unstained, magnification 125,000~.

stitial cells showed massive iron overload (FIG. 7a). In contrast, acinar and islet pcells showed only minor clusters of fenitin in the cytosol or accumulation within typical siderosomes (FIG. 7b). Rats exposed for 15 months to iron had heavier overload than those exposed for only 4 months. There was no evidence of organelle or cell damage, other than the presence of iron storage compounds. No increased fibrosis was seen in the pancreases of either experimental or control rats. The Heart

Since functional abnormalities of the iron-loaded heart are frequently the cause of death in chronic iron-loading conditions, it was important to assess its involvement in the carbonyl-iron-fed rats. Even though the total iron concentration was only moderately higher than in controls, we could identify siderosomes within myocytes with ultrastructural features similar to those seen in the acinar 8). The amount of fenitin sequestrated in siderosomes cells of the pancreas (FIG. of heart myocytes was small, and the particles had a heterogeneous appearance. The endothelial cells of capillaries contained conspicuous fenitin in the cytosol and occasional siderosomes, similar to endothelial cells of other organs. It has to be remembered that in human subjects the involvement of the pancreas and heart occurs late, possibly only when the liver and spleen are severely overladen and iron redistribution occurs.

174

ANNALS NEW YORK ACADEMY OF SCIENCES

FIGURE 7. Pancreas of rat fed carbonyl iron for 4 months. (a) Interstitial cells contain large amounts of ferritin and hemosiderin even in the early stages of overload (broad arrow). A, acinar cell. Unstained, magnification 60,OOOx. (b) Femtin particles can be identified within siderosomes of acinar cells (arrow). Z , zymogen granules. Unstained, magnification 41,OOOx.

IANCU & SHILOH: ULTRASTRUCTURAL STUDIES

175

FIGURE 8. Heart of rat fed carbonyl iron for 6 months. Part of a myocyte showing femtin particles and osmiophilic (lipid) droplets can be seen within a siderosome. Unstained, magnification 5 6 , 0 0 0 ~ .

Spontaneous Nutritional Overload

It has been known for some time that certain bird species are markedly iron overladen. Recently, captive birds were found to have been fed an iron-rich diet, and high liver iron concentrations were measured in both free and captive bird^.^' No increased levels of serum transaminases were found, indicating absence of significant parenchymal damage. We examined the morphological features of the naturally occurring avian iron overload (Iancu, Ward, and Peters, unpublished data). As in livers of other species in which excessive iron is absorbed from the gut, there was a marked predominance of parenchymal iron overload in comparison with sinusoidal cells. Avian hepatocytes contained much more cytosolic femtin than human or animal controls, and large amounts of ferritin and hemosiderin were seen sequestrated in siderosomes. Typical arrays, similar to those seen in human iron-overladen subjects, were found only in the hornbill. Even though cellular damage appeared in severely overladen areas, nodular cirrhosis similar to the human disease was encountered only in the mynah bird. Another example of spontaneous iron overload is the Svalbard reindeer studied by Schreiner and N i l ~ s e nStarving .~~ animals decrease their body weight during the winter season by 45%, with a parallel increase of liver iron concentration from 29 mg per 100 g wet weight to 300 mg per 100 g wet weight. A reversible, massive siderosis of both parenchymal and nonparenchymal liver cells can be seen in the starving animals on a high-iron diet of forage plants, without any signs of liver fibrosis or cirrhosis.

176

ANNALS NEW YORK ACADEMY OF SCIENCES

CONCLUSION AND SUMMARY The ideal experimental animal model, exhibiting all physiopathological features of the human diseases associated with iron overload, has not yet been found. To date, the various models have contributed to the understanding of some features of iron toxicity. It is still unknown why an exaggerated amount of iron is absorbed from the gut, what the exact mechanism of cell and organelle toxicity is, and why certain cells are more sensitive to iron than others. The ultrastructural studies have elucidated some of the intracellular pathways of storage iron compounds. It has been shown that most cells are capable of reacting to iron exposure by synthesis of femtin and its subsequent transfer into siderosomes and degradation to hemosiderin. A more recent experimental model, the carbonyl-iron-fed rat, has provided additional information concerning ferritin in the intestinal mucosa, as well as features of early cytosiderosis in other organs later severely involved, such as pancreas and heart. The finding of severe iron overload in pancreatic interstitial cells may be relevant to the interlobular and perilobular fibrosis found in advanced stages of hemochromatosis. The dysfunction of the endocrine pancreas can be seen as the result of disturbed islet microcirculation, secondary to the interstitial fibrosis. Similarly, the presence of storage iron in heart myocytes at such an early stage may be relevant to future severe cardiac complications of hemochromatosis, such as dysrrhythmias and heart failure. In general, experimentally or spontaneously iron-laden animals appear much less affected by iron overload than human subjects. This may be due to their excellent iron-segregating mechanisms and increased biliary excretion of iron.33 This is more likely due to the fact that the animals do not suffer from the genetic defect of patients with either IHC or EHC.The experimental animals behave like humans who have had chronic exposure to relatively innocuous exogenous iron, such as in transfusional siderosis. The fact that even severe “physiological” parenchymal iron overload, as seen in birds and reindeer, is not usually accompanied by excessive fibrosis or cirrhosis indicates that we are still missing the “cirrhogenic factor” so aggressively active in infants and children with homozygous P-thalassemia.

ACKNOWLEDGMENTS The authors wish to thank Ms. Y.Regev for photographic assistance and Dr. A. S. Luder for help with the manuscript. REFERENCES 1. RICHTER, G . W. 1978. The iron-loaded cell-the

2. 3. 4. 5.

6.

cytopathology of iron storage. A review. Am. J. Pathol. 91: 363-404. IANCU,T. C. 1983. Iron overload. Mol. Aspects Med. 6 1-100. RICHTER, G. W. 1984. Studies of iron overload. Lab. Invest. 50: 26-33. IANCU, T. C., H. B. NEUSTEIN& B. H. LANDING.1977. The liver in thalassemia major: Ultrastructural observations in iron metabolism. In Iron Metabolism. Ciba Found. Symp. 51: 293-307. & L. W. POWELL.1986. Value of hepatic iron BASSETT,M. L., J. W. HALLIDAY measurements in early hemochromatosis and determination of the critical iron level associated with fibrosis. Hepatology 6 24-29. & J. L. E. ERICSSON. 1974. Studies of iron loading of rat ARBORGH, B., H. GLAUMANN

IANCU & SHILOH: ULTRASTRUCTURAL STUDIES

7. 8.

9. 10.

1 I.

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

177

liver lysosomes: Effects on the liver and distribution and fate of iron. Lab. Invest. 30: 664-673. HULTCRANTZ, R. & B. ARBORGH.1978. Studies on the rat liver following iron overload. Acta Pathol. Microbiol. Scand. Sect. A. 86: 143-155. AWAI,M., M. NARASAKI, Y. YAMANOI & S. SENO. 1979. Induction of diabetes in animals by p a r e n t e d administration of ferric nitrilotriacetate. Am. J. Pathol. 95: 663-673. PARMLEY, R. T., M. E. MAY,S. S. SPICER,M. G. BUSE& C. J. ALVAREZ.1981. Ultrastructural distribution of inorganic iron in normal and iron-loaded hepatic cells. Lab. Invest. 44: 475-485. G. M. BRITTENHAM, C. H. PARK& R. 0. RECKNAGEL. BACON,B. R., A. S . TAVILL, 1983. Hepatic lipid peroxidation in uiuo in rats with chronic iron overload. J. Clin. Invest. 71: 429-439. PETERS,T. J., M. J. O'CONNELL& R. J. WARD.1986. Role of free-radical mediated lipid peroxidation in the pathogenesis of hepatic damage by lysosomal disruption. In Free Radicals and Liver Injury. G. Poli, K. H. Cheeseman, M. U. Dianzani & T. F. Slater, Eds.: 107-115. IRL Press. Oxford. IANCU,T. C., H. RABINOWITZ, P. BRISSOT,A. GUILLOUZO, Y. DEUGNIER & M. BOUREL.1985. Iron overload of the liver in the baboon. J. Hepatol. 1: 261-275. T. D., N. KAUFMAN & J. V. KLAVINS. 1963. Deposition of iron in association KINNEY, with a periodic acid-Schiff (PAS)-positive material in the liver of ethionine-treated rats. Lab. Invest. 12: 978-984. MACDONALD, R. A. 1960. Experimental pigment cirrhosis: Its production in rats by feeding a choline deficient diet with excess iron. Am. J. Pathol. 36: 499-520. RICHTER,G. W. 1974. Effects of cyclic starvation-feeding and splenectomy on the development of hemosiderosis in rat livers. Am. J. Pathol. 74: 481-506. & P. PERERA.1978. Iron overload in Chang cell JACOBS, A., T. G . HOY,J. HUMPHRYS cultures: Biochemical and morphological studies. Br. J. Exp. Pathol. 59: 489-498. WILLIAMS, A., T . G. HOY& A. JACOBS.1982. Cellular proliferation and susceptibility to iron toxicity in iron loaded cell cultures. Scand. J. Haematol. 28: 227-232. WEINTRAUB, L. R., A. GORAL,J. GRASSO,C. FRANZBLAU, A. SULLIVAN & S. SULLIVAN.1985. Pathogenesis of hepatic fibrosis in experimental iron overload. Br. J. Haematol. 5 9 321-331. & H. B. NEUSTEIN. 1977. Pathogenetic mechanisms in IANCU,T. C., B. H. LANDING hepatic cirrhosis of thalassemia major: Light and electron microscopic studies. Pathol. Annu. U: 171-200. IANCU,T. C., H. SHILOH,G. LINK,E. R. BAUMINGER, A. PINSON& C. HERSHKO. 1987. Ultrastructural pathology of iron-loaded rat myocardial cells in culture. Br. J . Exp. Pathol. 68: 53-65. 1983. Beating rat heart LINK,G., J. URBACH, Y. HASIN,A. PINSON& C. HERSHKO. cell cultures: An in uitro model of iron toxicity and chelating therapy (abstract). Blood 62 (Suppl. 1,9): 38a. 1985. Heart cells in culture: A model of myocarL I N K G., , A. PINSON& C. HERSHKO. dial iron overload and chelation. J. Lab. Clin. Med. 106: 147-153. HUEBERS, H. A., G. M. BRITTENHAM, E . CSIBA& C. A. FINCH. 1986. Absorption of carbonyl iron. J. Lab. Clin. Med. 108: 473-478. BACON,B. R., J. H. HEALEY,G. M. BRITTENHAM, C. H. PARK,J. NUNNARI, A. S. TAVILL& H. B. BONKOVSKY. 1986. Hepatic microsomal function in rats with chronic dietary iron overload. Gastroenterology 90. 1844- 1853. L. Lours & A. S. PARK,C. H., W. S . STASSEN,B. R. BACON,G. M. BRITTENHAM, TAVILL. 1985. Hepatic fibrosis in rats with chronic dietary iron overload (abstract). Hepatology 5: 950. LESAGE,G. D., L. J. KOST,S. S. BARHAM & N. F. LARUSSO. 1986. Biliary excretion of iron from hepatocyte lysosomes in the rat. J. Clin. Invest. 77: 90-97. SELDEN,C., M. OWEN,J. M. P. HOPKINS& T. J. PETERS.1980. Studies on the concentration and intracellular localization of iron proteins in liver biopsy specimens from patients with iron overload with special reference to their role in lysosomal disruption. Br. J. Haematol. 44: 593-603.

178

ANNALS NEW YORK ACADEMY OF SCIENCES

28. IANCU, T. C., R. J. WARD& T. J. PETERS.1987. Ultrastructural observations in the carbonyl-iron fed rat, an animal model for hemochromatosis. Virchows Arch. B. 53: 208-217. 29. RAPPAPORT, A. M. 1982. Anatomic considerations. I n Diseases of the Liver. L. Schiff & E. Schiff, Eds.: 1-57. Lippincott. Philadelphia. M. I., J. W. SINDRAM, L. H. P. M. RANDEMAKERS, F. M. J. ZUYDERHOUT. 30. CLETON, W. C. DE BRUIJN& J. J. MARX. 1986. Ultrastructural evidence for the presence of fenitin-iron in the biliary system of patients with iron overload. Hepatology 6 30-35. G. M., R. J. WARD,J. K. KIRKWOOD & T. J. PETERS.1986. A possible 31. HENDERSON, cause of hepatic haemosiderosis in birds (abstract). Presented at the European Iron Club Meeting. September 1986. Pavia, Italy. B. B. I. & K. J. NILSSEN. 1986. The Svalbard reindeer: Nature’s own 32. SCHREINER, model of liver iron overload (abstract). Presented at the European Iron Club Meeting. September 1986. Pavia, Italy. R. & H. GLAUMANN. 1982. Studies on the rat liver following iron over33. HULTCRANZ, load. Lab. Invest. 46: 383-392.