Metabolism of transferrin-bound iron by the liver: a study in vivo.

Normal hepatic iron uptake is presumed to involve active transport of iron from serum transferrin after a transferrin-hepatocyte-membrane interaction. However, the possibility remains that transferrin itself enters the cell by endocytosis, with subsequent iron transfer within the lysosome, in a process akin to that occurring in erythropoietic tissue (Hemniaplardh & Morgan, 1977). Iron is subsequently stored in the liver cell as ferritin, some of which ultimately appears in the lysosomal system. In states of iron excess, lysosomal iron and ferritin content are markedly increased, and it has been postulated as the cause of membrane damage and ultimately cellular necrosis (Peters & Seymour, 1976). The pathway of iron from transferrin to ferritin is essentially uncharted, and its further elucidation is required particularly for an understanding of the pathophysiology of iron-overload conditions. Previous studies concerning hepatic iron metabolism have used iron-carriers such as dextran and haemoglobin (e.g. Van Wyk et al., 1971), and it is likely that iron in this form has a different metabolic pathway to that of transferrin-bound iron. Equally, studies using labelled serum are open to this criticism (e.g. Zimelman et al., 1977). We have therefore studied the process of hepatic iron uptake specifically from transferrin by using the rat as an experimental model. Purified rat transferrin was prepared from male rat serum. The rivanol-precipitation method of Sutton & Karp (1965) was used, but final purification was achieved by two passages through DEAE-Sephadex A50 using a 0.05-0.5~.-Tris gradient, pH 8.5. The final transferrin solution was shown to be uncontaminated by other serum proteins by polyacrylamide-gel electrophoresis. Transferrin solution (10mg) was labelled with lZ5I by using the lactoperoxidase method (Kuenzle & Dobeli, 1973). Labelling with 59Fe to produce a final transferrin saturation of 3 0 4 0 % was carried out in the presence of carbonate ions and a IOOM excess of citrate. The double-labelled transferrin was dialysed overnight against 0 . 6 ~ NaHC03, pH7.5, to remove free iron, and the solution shown to contain only transferrin-bound iron by polyacrylamide-gel electrophoresis. Tracer doses (1 yg of transferrin per g body weight) were injected intravenously into male rats aged between 3 and 5 months. After varying periods of time the animals were killed, and the livers perfused with iso-osmotic NaCI, removed, homogenized in iso-osmotic sucrose and fractionated by differential centrifugation by a modification of the method of de Duve et al. (1955). Fractions collected after centrifugation were designated I (6OOg, IOmin), I1 (10000g, 5min), 111 (4OOOOg, IOmin), IV (78000g, 90min), V (260OOOg, 60min) and VI (post-260000g supernatant). Marker enzymes assayed were N'-acetyl-D-D-glucosaminidase, alkaline phosphatase and neutral a-glucosidase (Seymour & Peters, 1977) and glutamate dehydrogenase (Ellis & Goldberg, 1972). The hepatic concentration of '''I was maximal 5min after injection [4.7? 1 . 8 % ~ ~ . of dose (n = 6)] and subsequently declined to less than 2% by 3h after injection. The hepatic concentration of 59Fe rose to a maximum [11.7*4.6%s.~. of dose (n = 6)] at 16h, and remained relatively constant over the ensuing 6 days. Fig. 1 shows the distribution of subcellular-organelle marker enzymes after differential centrifugation of the liver homogenate. Mitochondria can be seen to predominate in fractions I and 11, lysosomes in I1 and 111, plasma membrane in I, I11 and IV and endoplasmic reticulum in 111 and IV. Fig. 1 also shows the distribution of s9Fe. At 5min after injection, the majority is seen in fraction VI, a similar distribution being found for lz5I. Both labels are predominantly associated with transferrin, and may represent its dislocation from the plasma membrane during fractionation. After 30min transferrin-associated label is still present in fraction VI; however, ferritin-associated 59Fe is also detectable in fraction VI and V. At 3h after injection ferritin-associated s9Fe is increased in fractions V and IV; 59Fe in fraction VI is markedly decreased and