Uptake and degradation of formaldehyde-treated 125I-labelled human serum albumin in rat liver cells in vivo and in vitro

The uptake of formaldehyde-treated 125I-labelled human serum albumin in rat hepatocytes and nonparenchymal liver cells was measured in vivo and in vitro. Isolated liver cells were prepared by treating the perfused liver with collagenase. Purified hepatocytes and nonparenchymal cells were obtained by differential centrifugation. Human serum albumin was found to be taken up exclusively or almost exclusively by nonparenchymal cells in vitro and in vivo (after intravenous injection). The maximal rate of human serum albumin-uptake in vitro was comparable to that in vivo. Nonparenchymal cells degraded human serum albumin in vitro as indicated by release of trichloroacetic acid-soluble radioactivity. Degradation started about 20-30 min after addition of human serum albumin to cells and rate of degradation was proportional to rate of uptake. Human serum albumin-degradation could be studied without interference of concurrent uptake by separating cells that had been preincubated with human serum albumin from the medium and then reincubating them with human serum albumin-free medium. The lag phase before human serum albumin-degradation starts and the inhibitory effect of chloroquine on degradation indicate that human serum albumin is degraded in lysosomes. The data obtained show that enzymatically prepared nonparenchymal liver cells retain their endocytic activity in vitro. Denatured human serum albumin should be useful both as a marker for rat liver macrophages and for the study of intracellular proteolysis in these cells.     

Uptake and degdradation of formaldehyde-treated 125I-labeled human serum albumin in rat liver cells in vivo and in vitro

The uptake of formaldehyde-treated ¹²⁵I-labelled human serum albumin in rat hepatocytes and nonparenchymal liver cells was measured in vivo and in vitro. Isolated liver cells were prepared by treating the perfused liver with collagenase. Purified hepatocytes and nonparenchymal cells were obtained by differential centrifugation. Human serum albumin was found to be taken up exclusively or almost exclusively by nonparenchymal cells in vitro and in vivo (after intravenous injection). The maximal rate of human serum albumin-uptake in vitro was comparable to that in vivo. Nonparenchymal cells degraded human serum albumin in vitro as indicated by release of trichloroacetic acid-soluble radioactivity. Degradation started about 20–30 min after addition of human serum albumin to cells and rate of degradation was proportional to rate of uptake. Human serum albumin-degradation could be studied without interference of concurrent uptake by separating cells that had been preincubated with human serum albumin from the medium and then reincubating them with human serum albumin-free medium. The lag phase before human serum albumin-degradation starts and the inhibitory effect of chloroquine on degradation indicate that human serum albumin is degraded in lysosomes. The data obtained show that enzymatically prepared nonparenchymal liver cells retain their endocytic activity in vitro. Denatured human serum albumin should be useful both as a marker for rat liver macrophages and for the study of intracellular proteolysis in these cells.     

Metabolism of asialoglycoproteins in cultured rat hepatocytes: evidence for receptor mediated uptake and degradation which is not feed-back regulated

1. Cultured rat hepatocytes took up and degraded asialofetuin by a saturable mechanism. 2. The uptake was specific for desialylated fetuin; native fetuin and was taken up to a much lesser extent. 3. Degradation was inhibited by lysosomotropic agents; both uptake and degradation were, however, unaffected by cycloheximide. 4. The uptake was also unaffected by incubating the cells for 24 hr in the presence of large extracellular concentrations of unlabelled asialofetuin. 5. The data suggest that the asialoglycoprotein receptor is recycled and not subject to down-regulation.     

Uptake of beta-hexosaminidase by nonparenchymal liver cells and peritoneal macrophages

The uptake of beta-hexosaminidase (EC 3.2.1.30) in nonparenchymal liver cells (i.e. endothelial and Kupffer's cells) and peritoneal macrophages has been determined by an enzymatic assay. A considerable uptake was noted in nonparenchymal liver cells, whereas no measurable uptake was seen in peritoneal macrophages. The endothelial cells were more active in the uptake of beta-hexosaminidase than were the Kupffer's cells. The uptake of beta-hexosaminidase by nonparenchymal liver cells showed saturation kinetics and was competitively inhibited by mannan. These findings support the concept that a cell surface receptor on nonparenchymal liver cells mediates uptake of beta-hexosaminidase and suggests a difference in the receptor mechanisms on liver and peritoneal macrophages.     

In vivo cholesteryl ester selective uptake of mildly and standardly oxidized LDL occurs by both parenchymal and nonparenchymal mouse hepatic cells but SR-BI is only responsible for standardly oxidized LDL selective uptake by nonparenchymal cells

In blood circulation, low density lipoproteins (LDL) can undergo modification, such as oxidation, and become key factors in the development of atherosclerosis. Although the liver is the major organ involved in the elimination of oxidized LDL (oxLDL), the identity of the receptor(s) involved remains to be defined. Our work aims to clarify the role of the scavenger receptor class B type I (SR-BI) in the hepatic metabolism of mildly and standardly oxLDL as well as the relative contribution of parenchymal (hepatocytes) and nonparenchymal liver cells with a special emphasis on CE-selective uptake. The association of native LDL and mildly or standardly oxLDL labeled either in proteins or in cholesteryl esters (CE) was measured on primary cultures of mouse hepatocytes from normal and SR-BI knock-out (KO) mice. These in vitro assays demonstrated that hepatocytes are able to mediate CE-selective uptake from both LDL and oxLDL and that SR-BI KO hepatocytes have a 60% reduced ability to selectively take CE from LDL but not towards mildly or standardly oxLDL. When lipoproteins were injected in the mouse inferior vena cava, parenchymal and nonparenchymal liver cells accumulated more CE than proteins from native, mildly and standardly oxLDL, indicating that selective uptake of CE from these lipoproteins occurs in vivo in these two cell types. The parenchymal cells contribute near 90% of the LDL-CE selective uptake and SR-BI for 60% of this pathway. Nonparenchymal cells capture mainly standardly oxLDL while parenchymal and nonparenchymal cells equally take up mildly oxLDL. An 82% reduction of standardly oxLDL-CE selective uptake by the nonparenchymal cells of SR-BI KO mice allowed emphasizing the contribution of SR-BI in hepatic metabolism of standardly oxLDL. However, SR-BI is not responsible for mildly oxLDL metabolism. Thus, SR-BI is involved in LDL- and standardly oxLDL-CE selective uptake in parenchymal and nonparenchymal cells, respectively.     

Comparative investigations on the uptake of phallotoxins, bile acids, bovine lactoperoxidase and horseradish peroxidase into rat hepatocytes in suspension and in cell cultures

Two alternative uptake mechanisms for phallotoxins by liver cells are debated: carrier-mediated uptake and receptor-mediated endocytosis. We have compared the properties of hepatocellular uptake of the phallotoxins, phalloidin and demethylphalloin, with the uptake of cholate as a substrate for carrier-mediated uptake and compared with iodinated bovine lactoperoxidase or iodinated horseradish peroxidase, as the latter are known to be taken up by vesicular endocytosis. Uptake of phallotoxins and [14C]cholate uptake into isolated hepatocytes is independent of extracellular calcium but inhibited by A23187 or by monensin. Uptake of bovine lactoperoxidase strictly depends on external Ca2+, was insensitive to A23197 and was not inhibited by monensin. No mutual uptake inhibition between phalloidin or cholate and peroxidases was seen, indicating independent permeation pathways in hepatocytes. However, high concentrations of cytochalasin B inhibited the uptake of either phalloidin, cholate or bovine lactoperoxidase. Horseradish peroxidase uptake, which was taken as an indicator for fluid pinocytosis, was low in isolated hepatocytes and could not account for the amount of phalloidin or cholate taken up. In cultured rat hepatocytes, uptake of phallotoxins decreased within 1 day to 10% of the uptake seen in freshly isolated hepatocytes. The results indicate different mechanisms for hepatocellular phallotoxin/bile-acid uptake and peroxidase internalization. As monolayer cultures of hepatocytes rapidly lost the carrier-mediated uptake of phallotoxins and bile acids, freshly isolated hepatocytes might be a more suitable experimental model than cultured cells for kinetic studies on this transport system.     

Influence of heat treatment on rabbit liver ferritin

In order to enhance the stability of beta-galactosidase, we conjugated the enzyme with dextran T-10 (Mr approx. 10 000). The conjugate contained 9-10 mol dextran/mol protein (beta-galactosidase, Mr 68 000), and the specific activity retained after conjugation was 90 +/- 4% (n = 3) of the initial activity. Uptake and degradation of native and conjugated beta-galactosidase in isolated hepatocytes and nonparenchymal liver cells was studied. There was a marked increase in stability against degradation in both cell types when beta-galactosidase was conjugated with Dextran. The degradation of dextran-conjugated enzyme was reduced by 35% in hepatocytes and by 43% in nonparenchymal cells, after 80 and 40 min, respectively, as compared with the free enzyme. However, there was insignificant difference between the uptake of native and conjugated enzyme into the liver cells. Upon intravenous infusion into rats, native and conjugated enzyme were cleared from plasma with only a slight difference in the clearance rate. The observed stability of dextran-conjugated beta-galactosidase towards cellular degradation was in accordance with the in vitro experiments. The conjugate showed marked thermal stability at 50 degrees C and enhanced resistance towards proteolysis by the broad specific protease subtilopeptidase A. This demonstrates that dextran conjugation may be used as a means of stabilizing lysosomal enzymes for therapeutic purposes.     

Human β-glucuronidase: In vivo clearance and in vitro uptake by a glycoprotein recognition system on reticuloendothelial cells

Previously reported studies demonstrated that infused human placental β-glucuronidase is rapidly cleared from rat plasma and localizes predominantly in rat liver. Prior treatment of the enzyme with sodium metaperiodate converted the enzyme to a very slow clearance form. This suggested that the clearance system recognized the carbohydrate structure of the glycoprotein hydrolase. This report defines the glycosyl specificity and the cell type(s) involved in the clearance process. Clearance of infused human β-glucuronidase was blocked by simultaneous infusion of glycoproteins which have mannose or N-acetylglucosamine in their exposed nonreducing position, or by some simple sugars (α-methylmannoside, mannose or L-fucose) which block clearance of these glycoproteins. Two immunohistochemical techniques demonstrated preferential localization of human β-glucuronidase in sinusoidal lining cells (nonparenchymal cells) in rat liver. Human placental β-glucuronidase was also taken up by isolated rat alveolar macrophages by the carbohydrate-mediated glycoprotein uptake system recently demonstrated in these cells. This isolated cell uptake system appears to have the same specificity as the system for plasma clearance of infused human placental β-glucuronidase in the intact rat.The combined data from in vivo clearance studies and from studies of enzyme uptake by isolated rat macrophages suggest that a mannose/N-acetylglucosamine-glycoprotein uptake system is expressed on fixed tissue macrophages in the rat, and that this system mediates plasma clearance of infused human placental β-glucuronidase.     

The effect of chemical treatments of albumin and orosomucoid on rate of clearance from the rat bloodstream and rate of pinocytic capture of rat yolk sac cultured in vitro.

Portions of a 125I-iodinated bovine serum albumin preparation were exposed to freezing, acetic acid (pH 3.5, 3.0 or 2.5), urea or formaldehyde, and the effect of these treatments on the rates of pinocytic uptake by yolk sacs from 17.5-day-pregnant rats cultured in vitro and of clearance from the rat bloodstream were studied. Uptake of albumin by the yolk sac was followed by rapid release of [125I]iodo-L-tyrosine into the culture medium. Similarly clearance of albumin in vivo was accompanied by the appearance of trichloroacetic acid-soluble radioactivity in the bloodstream. In both systems the rates of uptake of modified albumin preparations formed a series: formaldehyde or urea greater than acetic acid greater than freezing. The increased rates of uptake of modified albumin preparations could not be ascribed to the formation of aggregates nor, in the yolk-sac system, to an increase in the rate of pinosome formation. It is concluded that the various treatments to which the albumin was subjected increase to varying degrees the affinity of the albumin molecule for binding sites on that region of the plasma membrane from which pinocytic vesicles are formed. Some comparable experiments with native and desialylated human orosomucoid indicate that the rat yolk-sac epithelial cells do not possess the recognition system for uptake of asialoglycoproteins that exists on the surface of hepatic parenchymal cells.     

Clearance of circulating desialylated thyroglobulins in the rat.

Canine and rat thyroglobulins were labeled with 125I either in vitro or in vivo, and were utilized for plasma clearance studies performed with rat. The half-life of physiologically radioiodinated asialothyroglobulins was about 6 min, while that of chemically radioiodinated asialothyroglobulin was about 12 min. No marked species difference was observed in this clearance. The label which had disappeared from the blood was recovered mainly in the liver, and this uptake was blocked by the simultaneous injection of desialylated orosomucoid but not by native orosomucoid. Radiolabeled monoiodotyrosine, diiodotyrosine, triiodothyronine and thyroxine were detected in the liver 17 min after intravenous injection of asialothyroglobulin labeled with 125I in vivo, suggesting the possible production of thyroid hormones in extrathyroidal tissues.     

Endocytosis of bovine lactoperoxidase by two carbohydrate-specific receptors in rat liver

A perfused rat liver took up bovine lactoperoxidase (LPO) by a Ca2+-dependent, saturable process. This endocytosis was accomplished by both hepatocytes and Kupffer or other nonparenchymal cells (NPCs). The mediating receptors were the Gal/GalNAc lectin of hepatocytes and the Man/GlcNAc lectin of NPCs. Blocking either one of these receptors caused a large shift in distribution of accumulated LPO into the cell type whose receptor was left unblocked, but the extent of uptake was unaffected and the rate was only moderately reduced. Effective inhibition of overall uptake into the perfused organ required the presence of competitors for both receptors. Conversely, LPO was an effective competitor of other ligands (asialoorosomucoid or mannan) for either of the two receptors. The major clearance capacity for LPO was associated with hepatocytes which in suspension took it up by a process completely inhibitable by asialofetuin (ASF) and at a rate more than three times greater than for ASF. A faster cycling time for Gal/GalNAc receptors when bound to LPO is suggested. The glycoprotein selectively lost its affinity for Man/GlcNAc receptors when digested by endoglycosidase H (endo H), suggesting that LPO contains mannose-rich oligosaccharides.     

Uptake and 25-hydroxylation of vitamin D3 by isolated rat liver cells

The physiological roles played by hepatocytes and nonparenchymal cells of rat liver in the metabolism of vitamin D3 have been investigated. Tritium-labeled vitamin D3 dissolved in ethanol was administered intravenously to two rats. Isolation of the liver cells 30 and 70 min after the injection showed that vitamin D3 had been taken up both by the hepatocytes and by the nonparenchymal liver cells. The relative proportion of vitamin D3 that accumulated in the nonparenchymal cells increased with time. Perfusion of the isolated rat liver with [3H] vitamin D3 added to the perfusate confirmed the ability of both cell types to efficiently take up vitamin D3 from the circulation. By a method based on high pressure liquid chromatography and isotope dilution-mass fragmentography it was found that isolated liver cells in suspension had a considerable capacity to take up vitamin D3 from the medium. About 2.5 fmol of vitamin D3 were found to be associated with each hepatocyte or nonparenchymal cell after 1 h of incubation. 25-Hydroxylation in vitro was found to be carried out only by the hepatocytes. The rate of hydroxylation was about the same whether the cells were isolated from normal or rachitic rats (3.5 and 4 pmol of 25-hydroxyvitamin D3 formed per h per 10(6) cells, respectively). The possibility that the nonparenchymal cells might serve as a storage site for vitamin D3 in the liver is discussed.     

Enhanced stability of β-galactosidase in parenchymal and nonparenchymal liver cells by conjugation with dextran

In order to enhance the stability of β-galactosidase, we conjugated the enzyme with dextran T-10 (M_r approx. 10 000). The conjugate contained 9–10 mol dextran/mol protein (β-galactosidase, M_r 68 000), and the specific activity retained after conjugation was 90 ± 4% (n = 3) of the initial activity. Uptake and degradation of native and conjugated β-galactosidase in isolated hepatocytes and nonparenchymal liver cells was studied. There was a marked increase in stability against degradation in both cell types when β-galactosidase was conjugated with Dextran. The degradation of dextran-conjugated enzyme was reduced by 35% in hepatocytes and by 43% in nonparenchymal cells, after 80 and 40 min, respectively, as compared with the free enzyme. However, there was insignificant difference between the uptake of native and conjugated enzyme into the liver cells. Upon intravenous infusion into rats, native and conjugated enzyme were cleared from plasma with only a slight difference in the clearance rate. The observed stability of dextran-conjugated β-galactosidase towards cellular degradation was in accordance with the in vitro experiments. The conjugate showed marked thermal stability at 50°C and enhanced resistance towards proteolysis by the broad specific protease subtilopeptidase A. This demonstrates that dextran conjugation may be used as a means of stabilizing lysosomal enzymes for therapeutic purposes.     

Stimulation of growth of primary cultured adult rat hepatocytes without growth factors by coculture with nonparenchymal liver cells

DNA synthesis of adult rat parenchymal hepatocytes alone in primary culture can be stimulated only by the addition of humoral growth factors to the culture medium. However, when parenchymal hepatocytes were cocultured with nonparenchymal liver cells from adult rats, their DNA synthesis was markedly stimulated in the absence of added growth factors or calf serum. DNA synthesis of parenchymal hepatocytes was not stimulated by conditioned medium from nonparenchymal liver cells and was greatest when the parenchymal cells were plated on 24-h cultures of nonparenchymal liver cells. A dead feeder layer of nonparenchymal cells was almost as effective as a feeder layer of viable nonparenchymal cells. These results suggest that the stimulation of DNA synthesis in parenchymal hepatocytes was not due to some soluble factors secreted by nonparenchymal liver cells but to an insoluble material(s) produced by the nonparenchymal liver cells. This insoluble material(s) was collagenase- and acid-sensitive, suggesting that it was a protein containing collagen. The effect of nonparenchymal liver cells was specific: coculture with hepatoma cells, liver epithelial cells, or Swiss 3T3 cells did not stimulate DNA synthesis in parenchymal hepatocytes. Added insulin and epidermal growth factor showed additive effects with nonparenchymal cells in the cocultures. These results suggest that DNA synthesis in parenchymal hepatocytes is stimulated not only by various humoral growth factors but also by cell-cell interaction between parenchymal and nonparenchymal hepatocytes, possibly endothelial cells. This cell-cell interaction may be important in repair of liver damage and liver regeneration.     

Studies in vivo of the tissue uptake, cellular distribution and catabolic turnover of exogenous glucocerebrosidase in rat.

The kinetics of plasma clearance of highly purified human placental glucocerebrosidase in rats were biphasic with 75% of the infused dose showing a rapid clearance (t1/2 = 11 min) and the remaining 25% a considerably lower rate (t1/2 = 60 min). The majority of the enzyme (60%) was taken up by the liver. Although saturation kinetics for the clearance or uptake were not observed, the very high hepatic endocytic index (217 microliter/min) of glucocerebrosidase uptake indicated that liver uptake was mediated by an adsorptive endocytic process. Analysis of the cellular distribution of recovered glucocerebrosidase revealed predominantly parenchymal cell uptake with 38% of the exogenous enzyme in hepatocytes and only 2% in sinusoidal cells. High-mannose glycoproteins blocked hepatocyte and sinusoidal cell uptake of glucocerebrosidase equally. Kinetic experiments failed to demonstrate a transfer or shuttle of exogenous glucocerebrosidase from sinusoidal cells to hepatocytes. The possibility was raised that uptake of enzyme by the liver may be mediated by a common receptor that functions in both hepatocytes and sinusoidal cells. The catabolic turnover of exogenous glucocerebrosidase in rat liver was biphasic and the rate of decline was similar in hepatocytes and sinusoidal cells.     

Preclustered receptor arrangement is a prerequisite for galactose-specific clearance of large particulate ligands in rat liver

We had hypothesized that preclustered arrangement of galactose-specific receptor activity on rat liver macrophages enables these cells to internalize multivalent, particulate ligands in contrast to the clearance of molecules mediated by statistically distributed receptors on hepatocytes. We now took advantage of the nonclustered receptor distribution in newborn rat liver macrophages to study the in vivo clearance of particulate ligands. Gold particles 5, 17, and 50 nm in diameter (Au5, Au17, Au50), coated with lactosylated bovine serum albumin (LacBSA), were injected into the vena cava and livers were perfusion fixed after allowing for binding and uptake for 3 min. In sinusoidal cells from rats 15 days old LacBSA-Au5 and LacBSA-Au17 were taken up by endothelial cells and all sizes by liver macrophages. In newborn rat liver no LacBSA-Au50 or LacBSA-Au17 was retained in liver macrophages. Uptake of LacBSA-Au5 by sinusoidal cells was significant. LacBSA-Au17 was taken up in significant amounts by endothelial cells of newborn rats which correlates to the findings that galactose-specific binding sites on endothelial cells were found to localize as clusters over coated pits irrespective of age. These results demonstrate the crucial role of clustered receptors in binding and uptake of larger particulate ligands via this lectin-like binding activity.     

Clearance of circulating desialylated thyroglobulins in the rat

Canine and rat thyroglobulins were labeled with ¹²⁵I either in vitro or in vivo, and were utilized for plasma clearance studies performed with rat. The half-life of physiologically radioiodinated asialothyroglobulins was about 6 min, while that of chemically radioiodinated asialothyroglobulin was about 12 min. No marked species difference was observed in this clearance. The label which had disappeared from the blood was recovered mainly in the liver, and this uptake was blocked by the simultaneous injection of desialylated orosomucoid but not by native orosomucoid. Radiolabeled monoiodotyrosine, diiodotyrosine, triiodothyronine and thyroxine were detected in the liver 17 min after intravenous injection of asialothyroglobulin labeled with ¹²⁵I in vivo, suggesting the possible production of thyroid hormones in extrathyroidal tissues.     

Unaltered catabolism of desialylated low-density lipoprotein in the pig and in cultured rat hepatocytes.

Removal of the terminal sialic acid residues from many serum glycoproteins results in exposure of their penultimate galactose residues and rapid clearance from circulation by the liver. Low-density lipoprotein is a glycoprotein containing 21 galactose and 9 sialic acid residues per particle. Studies in this laboratory and others have shown that both the liver and extrahepatic tissues contribute to the degradation of low-density lipoprotein. This study was undertaken to determine whether desialylation of pig low-density lipoprotein alters its removal from circulation. Low-density lipoprotein was incubated at 37 degrees C with an agarose-bound neuraminidase, proteinase-free, from Clostridium perfringens. After 18 h at pH 5.0, 70% of the sialic acid residues were removed. The desialylated 131I-labelled and native 125I-labelled low-density lipoproteins were simultaneously injected into a pig, and their disappearance from plasma was followed for 96 h. The turnovers of the two were identical. In contrast, neuraminidase-treated fetuin was cleared about 200-fold faster than native fetuin. Studies were also performed in cultured rat hepatocytes. Rates of degradation of native and neuraminidase-treated low-density lipoprotein were similar, whereas asialo-fetuin was degraded at six to ten times the rate of native fetuin. Thus desialylation does not appear to alter low-density-lipoprotein catabolism by hepatic or extrahepatic cells.     

Studies of the pathogenesis of Gaucher's disease: tissue distribution and biliary excretion of [14C]L-glucosylceramide in rats

The time course of the clearance from the blood and the tissue localization of [14C]L-glucosylceramide, a nonmetabolizable enantiomorph of D-glucosylceramide that accumulates in Gaucher's disease, has been determined. 14C-labeled L-glucosylceramide injected intravenously in the form of micelles or liposomes is rapidly removed from the circulation. Most of this lipid is taken up by the liver where it is found in both hepatocytes and nonparenchymal cells. This sphingolipid analog is promptly cleared from hepatocytes and a significant portion is recovered in the bile. The clearance of [14C]L-glucosylceramide from Kupffer cells is greatly prolonged in comparison with its brief residence in hepatocytes. These findings have significant implications regarding the pathogenesis and treatment of Gaucher's disease.     

alpha3-galactosylated glycoproteins can bind to the hepatic asialoglycoprotein receptor.

In mammals, clearance of desialylated serum glycoproteins to the liver is mediated by a galactose-specific hepatic lectin, the 'asialoglycoprotein receptor'. In humans, serum glycoprotein glycans are usually capped with sialic acid, which protects these proteins against hepatic uptake. However, in most other species, an additional noncharged terminal element with the structure Galalpha1-->3Galbeta1-->4R is present on glycoprotein glycans. To investigate if alpha3-galactosylated glycoproteins, just like desialylated glycoproteins, could be cleared by the hepatic lectin, the affinities of alpha3-galactosylated compounds towards this lectin were determined using an in vitro inhibition assay, and were compared with those of the parent compounds terminating in Galbeta1-->4R. Diantennary, triantennary and tetraantennary oligosaccharides that form part of N-glycans were alpha3-galactosylated to completion by use of recombinant bovine alpha3-galactosyltransferase. Similarly, desialylated alpha1-acid glycoprotein (orosomucoid) was alpha3-galactosylated in vitro. The alpha3-galactosylation of a branched, Galbeta1-->4-terminated oligosaccharide lowered its affinity for the membrane-bound lectin on whole rat hepatocytes 50-250-fold, and for the detergent-solubilized hepatic lectin 7-50-fold. In contrast, alpha3-galactosylation of asialo-alpha1-acid glycoprotein caused only a minor decrease in affinity, increasing the IC50 from 5 to 15 nM. Fully alpha3-galactosylated alpha1-acid glycoprotein, intravenously injected into the mouse, was rapidly cleared from the circulation, with a clearance rate close to that of asialo-alpha1-acid glycoprotein (t1/2 of 0.42 min vs. 0.95 min). Its uptake was efficiently inhibited by pre-injection of an excess asialo-fetuin. Organ distribution analysis showed that the injected alpha1-acid glycoprotein accumulated predominantly in the liver. Taken together, these observations suggest that serum glycoproteins that are heavily alpha3-galactosylated will be rapidly cleared from the bloodstream via the hepatic lectin. It is suggested that glycosyltransferase expression in murine hepatocytes is tightly regulated in order to prevent undesired uptake of hepatocyte-derived, circulating glycoproteins.