SummaryThe mechanism of uptake of radio-iodinated tissue plasminogen activator (125I-t-PA) was studied in rats. When trace amounts of 125I-t-PA were injected alone, the clearance followed a biphasic pattern in which 65% and 35% were cleared with α- and β-kinetics (t1/2 (α) = 0.6 min, and t1/2 (β) = 6.4 min), respectively. Coinjection with excess unlabelled t-PA or inannan changed the uptake kinetics to the muiiupliasic β-elimination pattern. Mannosylated albumin and ovalbumin, both of which bind to the hepatic mannose receptor, reduced the proportion of t-PA cleared with t1/2 (α) to 48% and 21%, respectively. A corresponding increase in the β-elimination ot t-PA was observed. The t1/2 (α) and t1/2 (β) were unchanged. Studies on the eleaiaucc of 125I-ovalbumin also showed a biphasic elimination with an initial rapid phase, t1/2 (α), accounting for only 39% of the clearance of ovalbumin, as compared to 65% in the case of t PA. Macromolecules with affinity for the galactose-receptor only, such as asialofetuin, or galactosylated albumin, did not significantly affect the clearance kinetics at the concentrations used. Asialoorosomucoid, which also carries galactosyl residues in the terminal position, reduced somewhat (from 65% to 48%) the proportion cleared with α-kinetics. Very high concentrations of galactose and N-acetyl-galactosamine, which are also known to compete for binding to the galactose receptor, lowered the proportion of t-PA cleared in the late β-phase (reduced from 35% to 26% with galactose and to 19% with N-acetyl-galactosamine).To determine the hepatocellular site of uptake of t-PA, the protein was conjugated with 125I-labelled tyramine cellobiose (125I-TC) and injected intravenously (i.v.). This adduct is nonbiodegradable, and is trapped intralysosomally after endocytosis. I.v. injection of 125I-TC-t-PA and subsequent isolation of the liver cells showed that the Kupffer cells (KC), liver endothelial cells (LEC) and parenchymal cells (PC) contained 11%, 44% and 45%, respectively, of the radioactive label recovered in liver (hepatic uptake 80% of injected dose). The in vivo uptake per cell was about three times higher in KC and LEC than in PC. Injection of 125I-TC-t-PA together with mannose inhibited uptake in LEC and increased uptake in PC. Conversely, co-injection with galactose inhibited the uptake of 125I-TC-t-PA in PC and increased the uptake in LEC. Co-injection with excess amounts of unlabelled t-PA shifted the site of uptake from LEC to PC and changed the clearance kinetics to a monophasic β-elimination. The inhibitors used had only marginal effects on the uptake of 125I-TC-t-PA in KC. Although significant amounts of label were recovered in KC, the total size of the population of these cells is relatively small, so that the main hepatic uptake of 125I-TC-t-PA was in LEC and PC.In conclusion, the elimination of t-PA from the blood by the liver is strongly dependent on the structure of its carbohydrate side chains. The main cellular sites of clearance are LEC (via mannose receptors), and PC (via galactose receptors and an unsaturable noncarbohydrate uptake mechanism).

SummaryExperiments were carried out to charact erize the specificity of uptake of tPA in rat liver cells. Endocytosis in liver endothelial cells of the native carbohydrate variants of tissue plasminogen activator (tPA), and tPA inactivated by diisopropyl fluorophosphate was found to be competitive, suggesting that the determinant being recognized by these cells is different from the active site. Fibronectin and urokinase, which show partial homology with tPA, did not compete with tPA for uptake in liver endothelial cells. Hyaluronic acid, collagen, or IgG, which are endocytosed by specific receptors in liver endothelial cells, did not interfere with the uptake.Reduced endocytosis by liver endothelial cells was observed with tPA modified in the carbohydrate side chains, suggesting that these structures are important for uptake. Ovalbumin, mannan, mannose, fructose, and EDTA, but not galactose, effectively inhibited uptake in liver endothelial cells of both native and diisopropyl fluorophosphate-inhibited tPA, but had very little effect on the uptake of tPA modified in the carbohydrate side chains.Endocytosis of native tPA by parenchymal cells could be inhibited by galactose, ovalbumin, and EDTA, but not by mannose.These results suggest that endocytosis of tPA by liver endothelial cells and parenchymal cells is mediated by the mannose and galactose receptors, respectively.

A B S T R A C T The catabolic pathways of streptokinase, plasmin, and activator complex prepared with human plasminogen were studied in mice. '25I-streptokinase clearance occurred in the liver and was 50% complete in 15 min. Incubation with mouse plasma had no effect on the streptokinase clearance rate. Complexes of plasmin and a2-plasmin inhibitor were eliminated from the plasma by a specific and saturable pathway. Competition experiments demonstrated that this pathway is responsible for the clearance of injected plasmin. Streptokinase-plasminogen activator complex formed with either 1251-plasminogen or 125I-streptokinase cleared in the liver at a significantly faster rate than either of the uncomplexed proteins (50% clearance in <3 min). Streptokinase incubated with human plasma also demonstrated this accelerated clearance. p-Nitrophenyl-p '-guanidinobenzoate-HCl or pancreatic trypsin inhibitor-treated complex cleared slowly compared with untreated complex independent of which protein was radiolabeled. Significant competition for clearance was demonstrated between a2-macroglobulin-trypsin and activator complex only when the plasmin(ogen) was the radiolabeled moiety. Large molar excesses of a2-plasmin inhibitor-plasmin failed to retard the clearance of activator complex. Hepatic binding of streptokinase-plasmin, in liver perfusion experiments, was dependent upon prior incubation with plasma (8-10% uptake compared to a background of -2.5%). Substitution of human ca2-mac-

SummaryThe mechanism of uptake of tissue plasminogen activator (tPA) in rat liver was studied. Radio-iodinated tPA was removed from the circulation after intravenous administration in a biphasic mode. The initial half life, t1/2(α), and the terminal phase, t1/2(β), were determined to be 0.5 min and 7.5 min, resp. Separation of the liver cells by collagenase perfusion and density centrifugation, revealed that the uptake per cell was two to three times higher in the non-parenchymal cells than in the parenchymal cells.Endocytosis of fluorescein isothiocyanate-labelled or 125I-labelled tPA was studied in pure cultures of liver cells in vitro. Liver endothelial cells and parenchymal cells took up and degraded tPA. Endocytosis was more efficient in liver endothelial cells than in parenchymal cells, and was almost absent in Kupffer cells.Competitivb inhibition experiments showing that excess unlabelled tPA could compete with the uptake and degradation of 125I-tPA, suggested that liver endothelial cells and parenchymal cells interact with the activator in a specific manner. Endocytosis of trace amounts of 125I-tPA in cultures of liver endothelial cells and parenchymal cells was inhibited by 50% in the presence of 19 nM unlabelled tPA. Agents that interfere with one or several steps of the endocytic machinery inhibited uptake and degradation of 125I-tPA in both cell types.These findings suggest that 1) liver endothelial cells and parenchymal cells are responsible for the rapid hepatic clearance of intravenously administered tPA; 2) the activator is taken up in these cells by specific endocytosis, and 3) endocytosed tPA is transported to the lysosomes where it is degraded.

Electrophoretic analysis of endoglycosidase-treated tissue plasminogen activator obtained from human melanoma cells showed that the heterogeneity observed for the protein in these preparations is caused by an Nglycosidically linked N-acetyllactosamine type of carbohydrate chain which is present in about 50% of the molecules. An oligomannose type and an N-acetyllactosamine type of glycan is present in all molecules.Three glycopeptides were isolated and characterized by 'H-NMR, sugar determination, methylation analysis and amino acid determination. The exact attachment site for each of the three glycans could be deduced from the amino acid compositions of the glycopeptides. Asn-117 carries the oligomannose type of glycan, the structure of which was completely determined. Asn-184 is the site where the presence or absence of a biantennary Nacetyllactosamine type of glycan causes the size heterogeneity. The third N-glycosylation site, Asn-448, was found to carry a triantennary or tetraantennary N-acetyllactosamine type of carbohydrate chain.Tissue plasminogen activator, t-PA, is an important regulator of the fibrinolytic activity in the blood [l]. It is released from the vascular endothelium and converts plasminogen to plasmin by a specific proteolytic cleavage. Plasmin in turn efficiently degrades the fibrin network of thrombi formed in the vascular system. The activation of plasminogen by t-PA is stimulated by the presence of fibrin [2, 31, and t-PA binds to fibrin with a high affinity [4]. Thus, plasmin is selectively generated at the site of the fibrin clot which can be lysed without systemic activation of plasminogen in the circulation.The amino acid sequence of the human t-PA molecule has been deduced from isolated cDNA clones [5, 61 and from direct amino acid sequence analysis [7]. Three Asn-Xaa-Ser/ Thr sequences available for N-glycosylation were identified in the structure [7].When purified from human melanoma cells cultured in the presence of aprotinin, the activator is obtained as a singlechain protein containing a form with 530 residues as well as a proteolytically modified form with 527 residues [6,8,9]. The numbering system for the glycosylation sites discussed in this study (i.e. Asn-117, Asn-184 and Asn-448) refers to the shorter form of the single-chain activator in order to follow the nomenclature generally used.If aprotinin is omitted, or if the single-chain t-PA is treated with plasmin, a two-chain form is generated where the Nterminal half (A chain) of the parent molecule is linked to the C-terminal half (B chain) with a disulfide bond. An additional heterogeneity in the melanoma t-PA preparations is an apparent size difference of about 3 kDa [lo]. Analysis of the carbohydrate composition of the separated variants has revealed that the larger molecule is more heavily glycosylated [ 111.The effectiveness of t-PA as a thrombolytic agent has been demonstrated in animal studies [12-141 and recently clinical trials have been initiated with t-PA produced by genetic engineering of animal cells [15, 161...