Using three different methods of cells fractionation, hemosiderin granules were isolated from tissues (liver and/or spleen) of three patients. The samples were obtained from a case of idiopathic hemochromatosis, a case of thalassemia major with secondary (transfusional?) hemosiderosis, and a case of transfusional hemosiderosis associated with an unclassified anemia. Iron, nitrogen, and protein content of the hemosiderin granules varied over a wide range. Electron microscopy of sectioned granules revealed aggregates of dense particles of different shapes, with diameters ranging from 10 A to about 75 A. In some of the granules dense particles corresponding to the iron hydroxide micelles of ferritin molecules were abundant. But many of the granules contained very few of these molecules. The presence of ferritin and apoferritin in the samples of hemosiderin granules was demonstrated by means of precipitin tests in agar-gel, using rabbit antiferritin sera with known antibody nitrogen concentrations. At least three antigenic components were detected in highly purified crystalline ferritin prepared from tissues of the three patients; the hemosiderin granules contained the same antigens, but probably in much smaller quantities. Both ferritin and apoferritin molecules were extracted from hemosiderin granules, and were demonstrated in the electron microscope after suitable preparation. The solubility curve of human ferritin in solutions of (NH 4 ) 2 SO 4 was investigated. The results indicate that substantial quantities of ferritin or apoferritin can be lost in saline, aqueous media during isolation of hemosiderin granules from cells. It was shown by means of electron microdiffraction on selected hemosiderin granules that the dense particles represent forms of partly hydrated α-Fe 2 O 3 . The conditions necessary for electron microdiffraction in an electron microscope precluded an exact determination of the state of hydration of the α-Fe 2 O 3 or of its structural relation to (FeOOH) micelles of pure ferritin in its undenatured state. The findings were considered in the light of evidence on the structure and disposition of hemosiderin in situ in cells, and on the structure of ferritin. Differences between endogenous hemosiderin and hemosiderin derived from injections of colloidal iron compounds were pointed out. The evidence indicates that in hemochromatosis and in secondary hemosiderosis much of the inorganic storage iron in liver and spleen is derived from degraded ferritin. The findings suggest that an abnormal cellular metabolic pathway of ferritin is implicated in the pathogenesis of hemochromatosis and transfusional hemosiderosis.

As revealed by electron microscopy and electron diffraction, the physical state of ferric hydroxide micelles contained in iron-dextran, saccharated iron oxide, and hydrous ferric oxide ("ferric hydroxide") differs notably from the state of the ferric hydroxide in ferritin or hemosiderin. By virtue of this difference one can trace the intracellular transformation of colloidal iron, administered parenterally, into ferritin and hemosiderin. One hour after intraperitoneal injection of iron-dextran or saccharated iron oxide into mice, characteristic deposits were present in splenic macrophages, in sinusoidal endothelial cells of spleen and liver, and in vascular endothelial cells of various renal capillaries. Four hours after injection, small numbers of ferritin molecules were identifiable about intracellular aggregates of injected iron compounds; and by the 6th day, ferritin was abundant in close proximity to deposits of injected iron compounds. The latter were frequently situated in cytoplasmic vesicles delimited by single membranes. These vesicles were most frequently found in tissue obtained during the first 6 days after injection; and in certain of the vesicles ferritin molecules surrounded closely packed aggregates of injected material. Much unchanged ferric hydroxide was still present in macrophages and vascular endothelial cells 3 to 4 weeks after injection. While electron microscopy left no doubt about the identity of injected ferric hydroxide on the one hand, and of ferritin or hemosiderin on the other, histochemical tests for iron failed in this respect. Precipitation of ferric hydroxide (hydrous ferric oxide) from stabilized colloidal dispersions of iron-dextran was brought about in vitro by incubation with minced mouse tissue ( e.g . liver), but not by incubation with mouse serum or blood. Subcutaneous injections of hydrous gel of ferric oxide into mice initially produced localized extracellular precipitates. Most of the material was still extracellular 16 days after injection, though part of it was phagocytized by macrophages near the site of injection; but apparently none reached the spleen in unaltered form. Five days after injection and thereafter, much ferritin was present in macrophages about the site of injection and in the spleen. The findings show that iron preparations widely used in therapy can be identified within cells, and that their intracellular disposition and fate can be followed at the molecular level. Considered in the light of previous work, they indicate that the characteristic structure of the ferric hydroxide micelles in molecules of ferritin is specific, and develops during the union of apoferritin with ferric hydroxide. Apparently this union does not depend upon specific cell components.

Hemosiderin deposits in rats and in man were studied and compared by means of electron and light microscopy. Typical, isotropic, iron-positive hemosiderin granules were found to contain innumerable, closely packed, electron-dense particles, embedded in matter that was much less dense to electrons. Similar dense particles were often scattered diffusely through the cytoplasmic matrix of cells containing hemosiderin granules. In cells of proximal convoluted tubules of rats given repeated intraperitoneal injections of hemoglobin the hemosiderin granules contained dense particles with a mean diameter of 55 A, and with a size-frequency distribution that indicated uniformity. These particles corresponded in size to the iron micelles of ferritin molecules. There was less uniformity of particles in hemosiderin granules situated in liver and reticulo-endothelial cells of rats that had been given a diet containing ethionine. The dense aggregates representing hemosiderin granules were often situated inside discrete cytoplasmic organelles that were bordered by membranes, and sometimes contained "cristae"; and often the membranous borders were markedly disrupted. The term "sidersomes" is proposed for these specialized cytoplasmic structures which may be derivatives of mitochondria, and apparently play a part in the formation of hemosiderin. Ferritin was crystallized from the livers and kidneys of the hemosiderotic rats with ease, but could not be crystallized from comparable quantities of liver and kidney tissue of untreated control rats. Specimens from the liver and spleen of a patient with advanced hemosiderosis, obtained at an operation, were also studied. In liver and reticulo-endothelial cells many particles with diameters of about 60 A were scattered through the cytoplasmic matrix. By contrast, hemosiderin granules in the same cells contained particles that varied considerably in size. In representative granules, examined at high resolution, the size-frequency distribution of particle diameters displayed a periodicity consistent with the presence of small, uniform subunits. Electron micrographs of ferritin, isolated from the spleen of the same patient, provided confirmation for the inferences that the dense particles observed inside cells are iron micelles, and that ferritin is probably a component of hemosiderin.

A marked increase of the serum beta globulins was found in rabbits developing amyloidosis as a result of prolonged treatment with ribonucleate administered by subcutaneous injections. Following cessation of treatment the beta globulin levels gradually returned to normal while the gamma globulin levels rose strikingly, the changes being accompanied by a resorption of amyloid from the spleen, and probably also from the kidneys. Electrophoretic studies provided some evidence that the increase in beta globulins which accompanied the development of amyloidosis resulted from the production of a globulin not normally present in rabbit serum. A protein or protein derivative that moved as a beta globulin when subjected to filter paper electrophoresis was excreted in substantial quantities in the urine of several amyloidotic rabbits, along with much smaller quantities of substances moving as albumin, alpha and gamma globulins. Considered as a whole, the findings indicate a causal relationship between the abnormal production of circulating beta globulins and the deposition of amyloid in rabbits treated with ribonucleate. Hence it appears that a beta globulin may be directly involved in the formation of amyloid under the conditions of the experiments here reported.

It was shown, by means of salt fractionation procedures and electrophoresis that a marked and sustained hyperglobulinemia regularly resulted when sodium ribonucleate was injected subcutaneously at frequent intervals into rabbits undergoing immunization with horse serum. The hyperglobulinemia was characterized by a large increase in the gamma globulin levels, and a slight increase in the alpha and beta globulin levels. In control experiments done concurrently, the immunization of rabbits with horse serum, accompanied by subcutaneous injections of saline instead of ribonucleate, resulted in only moderate elevations in gamma globulin levels, while injections of ribonucleate alone brought about slight elevations in all three globulin components in some of the animals. Precipitin tests showed that the rabbits immunized with horse serum and simultaneously treated with ribonucleate developed antibody titers against horse serum that were no higher than those of the immunized controls given saline instead of ribonucleate. Indeed, some of the animals treated with horse serum and ribonucleate had globulin levels that were much higher and had antibody titers that were significantly lower than were those of several rabbits receiving horse serum and saline. Injections of ribonucleate alone did not result in the formation of specific antibodies detectable by means of precipitin tests. The results made it plain that the hyperglobulinemia of the animals treated with horse serum and ribonucleate was not due to an excessive production of specific antibodies. The findings as a whole provide further evidence that nucleotides play an important role in the formation of proteins in animals, and they indicate that an abnormally increased utilization of ribonucleotides by cells capable of producing globulins may be a causative factor in the pathogenesis of hyperglobulinemia.