Cartilage damage after intraarticular exposure to collagenase 3.

OBJECTIVE: To determine the in vivo effects of intraarticular MMP-13. METHODS: Human recombinant MMP-13 was injected intraarticularly (i.a. ) into the hamster knee joint. MMP-13 activity, collagen and proteoglycan fragments, and hyaluronan were measured in synovial fluid. Antibody 9A4 was used to localize collagen damage. Western blotting was used to determine the size of type II collagen fragments. RESULTS: MMP-13 activity measurements showed greater than 98% of MMP-13 to be cleared instantly from the joint cavity. The remainder was cleared with a t(1/2)of 2 h. Immunohistochemical staining demonstrated collagen cleavage was limited to a thin superficial band on the surface of the articular cartilage whereas collagen damage extended more deeply into the synovial capsule and the menisci. The elevation of proteoglycan and hyaluronan in synovial fluid after MMP-13 was modest. Collagen fragments appeared in synovial fluid within 15 min following MMP-13. They were cleared with a half-life of circa 1.8 h and the predominant fragment was 32 kDa. CONCLUSIONS: Activated MMP-13 leads to tissue collagen damage with the release of collagen fragments. These fragments are measurable and could provide a method for assessment of cartilage collagen damage.

Detection of collagenase-induced damage of collagen by 9A4, a monoclonal C-terminal neoepitope antibody.

To determine whether the collagen network is compromised by collagenase during acute inflammation, a monoclonal antibody (9A4) was developed with specificity for the C-terminal neoepitope sequence generated by collagenase-cleavage of type II collagen (Gly-Pro-Pro-Gly-Pro-Gln-Gly-COOH). 9A4 was shown to detect the collagen collagenase-cleavage neoepitope with a K = 1.7 x 10(-7) M (type II) and K = 2 x 10(-6) M (type I). It does not recognize uncleaved native or denatured collagen. Articular cartilage from control animals is unstained by 9A4. During acute inflammation elicited in hamsters by intra-articular LPS, positive staining for the 9A4 neoepitope indicated the collagen was damaged. Wheel running exercise was used to apply stress to control cartilage and cartilage from animals with damaged collagen. After 6 months of running, the cartilage from normal animals was unaffected. By contrast, in the group with damaged collagen, the cartilage was fibrillated in all animals and in half of those, the cartilage failed and bony eburnation resulted.

Early eosinophil infiltration into the optic nerve of mice with experimental allergic encephalomyelitis.

Experimental allergic encephalomyelitis is a murine model of preclinical autoimmune disease that has pathologic similarities to multiple sclerosis (MS). Although CD4+ T cells have been shown to play a crucial role in the development of disease, we recently demonstrated a link between the development of paralysis and eosinophil infiltration into the spinal cord. As such, CD4+ cells may initiate disease, but eosinophils may be the actual effector cells responsible for causing damage to myelin and causing paralysis. Because MS patients sometimes experience early visual problems, ie, optic neuritis, we explored whether an early eosinophil infiltrate was also observed in the optic nerves of SJL mice after the passive transfer of encephalitogenic T cells. Seven days after the passive transfer of myelin basic protein (MBP)-reactive T cell blasts, we observed a significant infiltration of eosinophils into the optic nerves of the mice. This infiltration persisted during the early phases of paralysis, then declined to baseline values by the peak of limb paralysis on Day 10, and remained at baseline during the remission phase of the disease. Remyelination of optic nerves was observed at this time. These results suggest that eosinophil infiltration into the optic nerve is one of the earliest events occurring after the passive transfer of encephalitogenic T cells in murine experimental allergic encephalomyelitis.

Inhibition of leukotriene B4-receptor interaction suppresses eosinophil infiltration and disease pathology in a murine model of experimental allergic encephalomyelitis.

Leukotriene B4 (LTB4) is a chemotactic and cell-activating factor present at inflammatory sites in a variety of autoimmune diseases including multiple sclerosis (MS). In this study, we used a murine model of MS, experimental allergic encephalomyelitis (EAE), to assess the potential role of LTB4 on cell infiltration and paralysis. Injection of encephalogenic T cells into naive animals induced paralysis and weight loss that was completely inhibited by treatment with the selective LTB4 receptor antagonist CP-105,696 (ED50= 8.6 mg/kg orally). Although migration of lymphocytes into the central nervous system was unaffected, the efficacious effects of CP-105,696 correlated with up to a 97% decrease in eosinophil infiltration into the lower spinal cord as determined by light and electron microscopy and quantitated by levels of the specific enzyme marker eosinophil peroxidase. These results demonstrate that eosinophil recruitment in EAE is dependent on LTB4 receptor ligation and further reveal a previously unrecognized role for eosinophils in the pathogenesis of this disease.

Glycemic improvement in diabetic db/db mice by overexpression of the human insulin-regulatable glucose transporter (GLUT4).

The effects of increased GLUT4 (insulin-regulatable muscle/fat glucose transporter) expression on glucose homeostasis in a genetic model of non-insulin-dependent diabetes mellitus were determined by expressing a human GLUT4 transgene (hGLUT4) in diabetic C57BL/KsJ-db/db mice. A genomic hGLUT4 construct was microinjected directly into pronuclear murine embryos of db/+ matings to maintain the inbred background. Four lines of hGLUT4 transgenic mice were bred to homozygosity at the db locus and all showed a marked reduction of both fasted and fed plasma glucose levels (to approximately 50 and 360 mg/dl, respectively) compared with age-matched nontransgenic db/db mice (approximately 215 and 550 mg/dl, respectively), as well as an enhanced disposal of an oral glucose challenge. In situ immunocytochemical localization of GLUT4 protein in muscle from hGLUT4 db/db mice showed elevated plasma membrane-associated GLUT4 protein in the basal state, which markedly increased after an insulin/glucose injection. In contrast, nontransgenic db/db mice had low levels of plasma membrane-associated GLUT4 protein in the basal state with a relatively small increase after an insulin/glucose challenge. Since the intracellular GLUT4 levels in db/db mice were similar to nontransgenic db/+ mice, the glucose transport defect in db/db mice is at the level of glucose transporter translocation. Together, these data demonstrate that GLUT4 upregulation overcomes the glucose transporter translocation defect and alleviates insulin resistance in genetically diabetic mice, thus resulting in markedly improved glycemic control.

Enhanced peripheral glucose utilization in transgenic mice expressing the human GLUT4 gene.

Human GLUT4 protein expression in muscle and adipose tissues of transgenic mice decreases plasma insulin and glucose levels and improves glucose tolerance compared with nontransgenic controls (Liu, M.-L., Gibbs, E. M., McCoid, S. C., Milici, A. J., Stukenbrok, H. A., McPherson, R. K., Treadway, J. L., and Pessin, J. E. (1993) Proc. Natl. Acad. Sci. U.S.A. 90, 11346-11350). We examined the basis of improved glycemic control in hGLUT4 transgenic mice by determining glucose homeostasis and metabolic profiles in vivo. Glucose turnover experiments indicated a 1.4-fold greater systemic glucose clearance in hGLUT4 mice relative to controls (p < 0.05), whereas hepatic glucose production was similar despite 26% lower (p < 0.05) glucose levels. Glucose infusion rate during an euglycemic-hyperinsulinemic clamp was 2-fold greater (p < 0.05) in hGLUT4 mice versus controls, and skeletal muscle and heart glycogen content were increased 3-5-fold (p < 0.05). The increased peripheral glucose clearance in hGLUT4 mice was associated with increased (25-32%) basal and insulin-stimulated glucose transport rate in soleus muscle (p < 0.01), and increased muscle plasma membrane-associated GLUT4 protein. Fed hGLUT4 mice displayed 20-30% lower plasma glucose and insulin levels (p < 0.05) and 43% elevated glucagon levels (p < 0.001) compared with controls. Triglycerides, free fatty acids, and beta-hydroxy-butyrate were elevated 43-63% (p < 0.05) in hGLUT4 mice due to hypoinsulinemia-induced lipolysis. Free fatty acids and beta-hydroxybutyrate levels in hGLUT4 mice increased further upon fasting, and skeletal muscle glycogen levels decreased markedly compared with controls. The data demonstrate that high level expression of hGLUT4 increases systemic glucose clearance and muscle glucose utilization in vivo and also results in marked compensatory lipolysis and muscle glycogenolysis during a fast.

Transgenic mice expressing the human GLUT4/muscle-fat facilitative glucose transporter protein exhibit efficient glycemic control.

To examine the physiological role of the GLUT4/muscle-fat specific facilitative glucose transporter in regulating glucose homeostasis, we have generated transgenic mice expressing high levels of this protein in an appropriate tissue-specific manner. Examination of two independent founder lines demonstrated that high-level expression of GLUT4 protein resulted in a marked reduction of fasting glucose levels (approximately 70 mg/dl) compared to wild-type mice (approximately 130 mg/dl). Surprisingly, 30 min following an oral glucose challenge the GLUT4 transgenic mice had only a slight elevation in plasma glucose levels (approximately 90 mg/dl), whereas wild-type mice displayed a typical 2- to 3-fold increase (approximately 250-300 mg/dl). In parallel to the changes in plasma glucose, insulin levels were approximately 2-fold lower in the transgenic mice compared to the wild-type mice. Furthermore, isolated adipocytes from the GLUT4 transgenic mice had increased basal glucose uptake and subcellular fractionation indicated elevated levels of cell surface-associated GLUT4 protein. Consistent with these results, in situ immunocytochemical localization of GLUT4 protein in adipocytes and cardiac myocytes indicated a marked increase in plasma membrane-associated GLUT4 protein in the basal state. Taken together these data demonstrate that increased expression of the human GLUT4 gene in vivo results in a constitutively high level of cell surface GLUT4 protein expression and more efficient metabolic control over fluctuations in plasma glucose concentrations.

Ca2+ stores in Purkinje neurons: endoplasmic reticulum subcompartments demonstrated by the heterogeneous distribution of the InsP3 receptor, Ca(2+)-ATPase, and calsequestrin.

The nature of second messenger-responsive intracellular Ca2+ stores in neurons remains open for discussion. Here, we demonstrate the existence in Purkinje cells (PCs) of endoplastic reticulum (ER) subcompartments characterized by an uneven distribution of three proteins involved in Ca2+ storage and release: the inositol 1,4,5-trisphosphate (InsP3) receptor, Ca(2+)-ATPase, and calsequestrin. Ca(2+)-ATPase and the InsP3 receptor have a widespread, although not identical, distribution throughout the ER. Calsequestrin is localized throughout the smooth ER and is particularly concentrated in pleiomorphic vesicles with a moderately electron-dense core, which appear to represent a subcompartment of the smooth ER. In double-labeling experiments many of these vesicles were unlabeled by InsP3 receptor antibodies. These results suggest a key role of the ER as an intracellular Ca2+ store and demonstrate a possible structural basis for distinct intracellular Ca2+ pools regulated by different second messengers.

Ultrastructural localization of albumin transport across the cerebral microvasculature during experimental meningitis in the rat.

Injury to the blood brain barrier (BBB) is a fundamental sequela of bacterial meningitis, yet the precise mechanism facilitating exudation of albumin across the endothelium of the cerebral microvasculature remains conjectural. After intracisternal inoculation of Escherichia coli (0111:B4) lipopolysaccharide in rats to elicit a reversible meningitis and BBB injury, we utilized in situ tracer perfusion and immunolabeling procedures to identify by transmission electron microscopy the precise topography and microvascular exit pathway(s) of bovine serum albumin (BSA). Results revealed that during meningitis there was: (a) an inducible increase in immunodetectable monomeric BSA binding to the luminal membrane of all microvascular segments in the pia-arachnoid and superficial brain cortex; (b) similar uptake of both colloidal Au-BSA (as well as monomeric BSA) by plasmalemmal vesicles but no detectable transcytosis to the abluminal side; and (c) predominant exit of both perfused Au-BSA and immunodetectable monomeric BSA through open intercellular junctions of venules in the pia-arachnoid. This was corroborated in separate experiments documenting focal pial venular leaks of in situ perfused 0.01% colloidal carbon black during experimental meningitis. These results provide precise localization of BBB injury in meningitis to meningeal venules, confirm a paracellular exit pathway of albumin via open intercellular junctions, and suggest an injury mechanism amenable to specific therapeutic intervention.

Endocytosis of simian virus 40 into the endoplasmic reticulum.

The endocytosis of SV-40 into CV-1 cells we studied using biochemical and ultrastructural techniques. The half-time of binding of [35S]methionine-radiolabeled SV-40 to CV-1 cells was 25 min. Most of the incoming virus particles remained undegraded for several hours. Electron microscopy showed that some virus entered the endosomal/lysosomal pathway via coated vesicles, while the majority were endocytosed via small uncoated vesicles. After infection at high multiplicity, one third of total cell-associated virus was observed to enter the ER, starting 1-2 h after virus application. The viruses were present in large, tubular, smooth membrane networks generated as extentions of the ER. The results describe a novel and unique membrane transport pathway that allows endocytosed viral particles to be targeted from the plasma membrane to the ER.

Synaptophysin is targeted to similar microvesicles in CHO and PC12 cells.

Synaptophysin, an integral membrane protein of small synaptic vesicles, was expressed by transfection in fibroblastic CHO-K1 cells. The properties and localization of synaptophysin were compared between transfected CHO-K1 cells and native neuroendocrine PC12 cells. Both cell types similarly glycosylate synaptophysin and sort it into indistinguishable microvesicles. These become labeled by endocytic markers and are primarily concentrated below the plasmalemma and at the area of the Golgi complex and the centrosomes. A small pool of synaptophysin is transiently found on the plasma membrane. In CHO-K1 cells synaptophysin co-localizes with transferrin that has been internalized by receptor-mediated endocytosis. These findings suggest that synaptophysin in transfected CHO-K1 cells and neuroendocrine PC12 cells is directed into a pathway of recycling microvesicles which, in CHO cells, is shown to coincide with that of the transferrin receptor. They further indicate that fibroblasts have the ability to sort a synaptic vesicle membrane protein. Our results suggest a pathway for the evolution of small synaptic vesicles from a constitutively recycling organelle which is normally present in all cells.

Interaction of angiotensin II with functional smooth muscle cells in culture.

In this study we report on the characterization of a highly enriched population of cultured vascular smooth muscle cells (SMC) prepared from collagenase-treated medial layer explant outgrowths of rabbit aortae. Studies done on cells from first passage explant outgrowths showed that the cells retain the fine structural features of vascular SMC in situ, can be immunostained with anti-smooth muscle myosin IgG, and bind [125I]angiotensin II (ANG II) in a specific and saturable manner with an apparent Kd of 1 nM. Addition of ANG II (0.1 microM) to the cultures causes obvious shape changes and retraction of cell processes. Electron microscopic autoradiography of cells labeled with [125I]ANG II show that the initial site of interaction of ANG II with the SMC is the plasma membrane. The distribution of ANG II receptors among cells in the population was studied using light microscopic autoradiography. The autoradiographical grain density varied among cells in the population ranging from cells that were heavily labeled to those that possessed virtually no label. These data imply that the expression of ANG II receptors may be limited to a certain progeny within the cell population or is a function of their stage within the cell cycle.

Transcytosis of albumin in capillary endothelium.

We have used a variety of immunocytochemical procedures to localize albumin in transit through the capillary endothelium of the murine myocardium and thereby identify endothelial cell structures involved in albumin efflux. The most informative results were obtained with a protocol that included (a) removal of endogenous albumin by perfusion of the heart with PBS supplemented with 14 mM glucose, (b) perfusion of the heart vasculature with exogenous (bovine) albumin for various short time periods, (c) fixation of the vessels by formaldehyde-glutaraldehyde mixtures, (d) processing of fixed myocardium specimens through L. R. White embedding followed by sectioning, or direct thin frozen sectioning, and (e) reacting the surface of such specimens with antialbumin antibodies followed by gold-labeled secondary antibodies. The results indicate that (a) monomeric albumin binds (with low affinity) to the luminal surface of the capillary endothelium, (b) it is restricted in transit through the endothelium to plasmalemmal vesicles, and (c) it appears in the pericapillary spaces less than 15 s after the beginning of its perfusion. No albumin concentration gradients, centered with their maxima on the exits from intercellular spaces, were detected at any time points, including the shortest ones (15 and 30 s) investigated. Additional information comparing monomeric vs. polymeric albumin transcytosis was obtained using albumin-gold complexes. The results are discussed in terms of vesicular transport of albumin across the endothelium and the relations of this type of transport to the postulated pore systems of the physiological literature.

Protein p38: an integral membrane protein specific for small vesicles of neurons and neuroendocrine cells.

An intrinsic membrane protein of brain synaptic vesicles with Mr 38,000 (p38, synaptophysin) has recently been partially characterized (Jahn, R., W. Schiebler, C. Ouimet, and P. Greengard, 1985, Proc. Natl. Acad. Sci. USA, 83:4137-4141; Wiedenmann, B., and W. W. Franke, 1985, Cell, 41:1017-1028). We have now studied the presence of p38 in a variety of tissues by light and electron microscopy immunocytochemistry and by immunochemistry. Our results indicate that, within the nervous system, p38, like the neuron-specific phosphoprotein synapsin I, is present in virtually all nerve terminals and is selectively associated with small synaptic vesicles (SSVs). No p38 was detectable on large dense-core vesicles (LDCVs). p38 and synapsin I were found to be present in similar concentrations throughout the brain. Outside the nervous system, p38 was found in a variety of neuroendocrine cells, but not in any other cell type. In neuroendocrine cells p38 was localized on a pleiomorphic population of small, smooth-surfaced vesicles, which were interspersed among secretory granules and concentrated in the Golgi area, but not on the secretory granules themselves. Immunoblot analysis of endocrine tissues and cell lines revealed a band with a mobility slightly different from that of neuronal p38. This difference was attributable to a difference in glycosylation. The finding that p38, like synapsin I, is a component of SSVs of virtually all neurons, but not of LDCVs, supports the idea that SSVs and LDCVs are organelles of two distinct pathways for regulated neuronal secretion. In addition, our results indicate the presence in a variety of neuroendocrine cells of an endomembrane system, which is related to SSVs of neurons but is distinct from secretory granules.

Epithelial polyanion (podocalyxin) is found on the sides but not the soles of the foot processes of the glomerular epithelium.

The distribution of podocalyxin was determined by several different immunoelectron microscopic methods, which included so-called diffusion methods carried out on cryostat sections and surface methods done on ultrathin sections. When localization was carried out by three different indirect immunogold methods (using ultrathin frozen sections, ultrathin sections cut from tissues embedded in Lowicryl K4M or LR White, or 20- mu cryostat sections for incubations), gold particles were found exclusively on the urinary surfaces of glomerular epithelial cells and, at a lower concentration, on the luminal surface of glomerular endothelial cells. No gold was bound to the basal surfaces (soles) of the epithelial foot processes, the filtration slit diaphragms, the glomerular basement membrane (GBM), or those regions of the endothelial plasmalemma that face the GBM. When podocalyxin was localized by an indirect immunoperoxidase method on cryostat sections, heavy deposits of reaction product were seen on the same cell surfaces as with the immunogold methods. Each of the methods used had certain advantages as well as limitations, but the collective results obtained were convergent and complimentary. It is concluded that podocalyxin is restricted in its distribution to the urinary surfaces of epithelial cells and the luminal surface of endothelial cells; it is missing or present in very low concentrations on the soles (basal surface) of the epithelial foot processes, the slit diaphragms and the basement membrane surface of the endothelium; and the podocalyxin-containing regions of the epithelial and endothelial plasmalemmae constitute microdomains of distinctive membrane protein composition on the corresponding cell surfaces.

Interaction of angiotensin II with dispersed cells from the anterior pituitary of the male rat.

Membranes from 6-week-old male rat anterior pituitaries possess saturable binding sites for angiotensin II (AII; Kd = approximately 2 X 10(-9) M). The binding is specific since it can be competed for with [Sar1,Leu8]AII and is unaffected by the presence of insulin or cholecystokinin octapeptide at nanomolar concentrations. To find out which cell types specifically interact with AII, rat anterior pituitaries were enzymatically dispersed and exposed to [125]iodo-AII (2 nM) in the absence or presence of [Sar1,Leu8]AII (400 nM). The cells were then washed free of unbound ligand and processed for light and electron microscopic autoradiography. Distribution of autoradiographic grains revealed that three cell types were specifically labeled with [125I]iodo-AII, namely mammotrophs, corticotrophs, and presumptive thyrotrophs. These cells were all labeled in the presence of [125I]iodo-AII alone (experimentals), whereas only 10-30% of them were labeled when 400 nM [Sar1,Leu8]AII was included in the binding reaction (controls). The number of grains over the labeled cells in the controls was 20% of that found in the experimental cells. These results may imply that AII can regulate anterior pituitary functions under physiological conditions by interacting directly with its secretory cells.

An enhanced incorporation of fatty acid into phosphatidyl choline that parallels histamine discharge in mast cells.

Purified rat peritoneal and pleural mast cells preincubated briefly with radioactively labeled fatty acid were treated with A23187, which bypasses primary receptors in stimulating exocytosis. An enhanced incorporation of fatty acid into phosphatidyl choline (PC) that occurred in parallel with histamine release at 24-25 degrees C was observed and was initially proportional to the total amount of histamine discharged. Enhanced PC labeling and histamine secretion were also proportional at temperatures ranging from 17-37 degrees C. Both radioactive linoleic and palmitic acids were incorporated selectively at the beta-position of the glycerol backbone of PC. PC labeling by [3H]choline was not detectably different in control and stimulated cells, and phosphatidic acid did not exhibit selectively enhanced beta-acylation. Thus, the stimulated labeling in A23187-treated cells may occur secondary to the action of a phospholipase A2 that favors PC as a substrate. Other peritoneal cell types exhibit a very similar A23187-stimulated selective labeling of PC. Therefore, autoradiography has been used to provide a direct demonstration that in purified preparations, mast cells are the principal cell type engaged in A23187-elicited incorporation of fatty acid into PC. The efficacy of this approach has relied on special procedures devised to obtain significantly different autoradiographic grain densities between control and stimulated preparations that can be attributed to differences in the level of [3H]palmitate-labeled PC. Preliminary tests using compound 48/80 as a secretory stimulus for mast cells have identified a similar selectively enhanced PC labeling. In either case, however, consideration of possible relationships between PC metabolism and the secretory process are premature since they have not been tested directly.

An electron microscope autoradiographic study of the carbohydrate recognition systems in rat liver. II. Intracellular fates of the 125I-ligands.

Electron microscope autoradiographic and biochemical methods were used to study the intracellular fates of several 125I-glycoproteins, known to be specifically bound and internalized by the different cell types in the liver. At the earliest times examined (1--2 min), 125I-glycoproteins (ASGP) were localized predominantly along the sinusoidal front of hepatocytes. Analysis of the distribution of autoradiographic grains indicated that: (a) approximately 40--60% of the 125I-ligand could be ascribed to the plasmalemma; (b) a significant fraction had already been internalized; yet (c) very little 125I-ligand was present in the lysosome-Golgi region. Between 4 and 15 min after administration of 125I-ASGPs, there was a dramatic redistribution of autoradiographic grains from regions of the plasmalemma and peripheral cytoplasm (30% decrease) to the lysosome-Golgi region (30% increase). At longer times (30 min), there was continued drainage of 125I-ASGP into this region. The grain density over secondary lysosomes was 60--90 times higher than that over recognizable Golgi elements, clearly indicating that lysosomes were the ultimate destination of the 125I-ASGP. However, no more than 60% of the total 125I-ligand could be localized to lysosome-rich regions of the hepatocyte, with the remaining 40% primarily in the intermediate cytoplasm. Biochemical evidence for proteolysis of the internalized 125I-ASGP (presumably within lysosomes) was obtained when [125I]-mono-iodotyrosine was found in the liver (i.e., hepatocytes) at times later than 15 min. The temporal redistribution observed for mannose and N-acetylglucosamine-terminated glycoproteins (ahexosamino-orosomucoid and agalacto-orosomucoid, respectively) in endothelial cells indicated that the 125I-ligands resided in macropinocytic vesicles (1--15 min) before their ultimate residence in dense bodies (15 min). The same 125I-ligands were also localized to structures resembling secondary lysosomes in Kupffer cells. The lysosomal nature of "these organelles" was implied from the appearance of [125I]mono-iodotyrosine in the liver at later times. 125I-beta-glucuronidase followed the same intracellular pathway in both cell types but was not degraded.

An electron microscope autoradiographic study of the carbohydrate recognition systems in rat liver. I. Distribution of 125I-ligands among the liver cell types.

Electron microscope autoradiography was used to study the cellular localization of seven glycoproteins rapidly cleared from the circulating plasma of rats and taken up by the liver. 1 and 15 min after intravenous administration of the 125I-glycoproteins, livers were fixed in situ by perfusion and processed for autoradiography. Autoradiographic grains in the developed sections were found to represent the intact 125I-ligand. A quantitative analysis of the distribution and concentration (density) of autoradiographic grains over the three major cell types of the liver was then performed. Three molecules, asialo-fetuin, asialo-orosomucoid, and lactosaminated RNase A dimer, the oligosaccharide chains of which terminate in galactose residues, were bound and internalized almost exclusively (greater than 90%) by hepatocytes. Conversely, four molecules, the oligosaccharide chains of which terminate in either N-acetyl-glucosamine (agalacto-orosomucoid) or mannose (ahexosamino-orosomucoid, preputial beta-glucuronidase, and mannobiosaminated RNase A dimer), were specifically bound and internalized by cells lining the blood sinusoids--that is, by Kupffer cells and endothelial cells. Endothelial cells were two to six times more active (on a cell volume basis) than were Kupffer cells in the internalization of these four 125I-ligands. Mannose and N-acetylglucosamine-terminated glycoproteins competed with each other for uptake into either endothelial cells or Kupffer cells, indicating that a single system recognized mannose or N-acetyl-glucosamine residues. Finally, agalacto-orosomucoid and ahexosamino-orosomucoid were also associated with hepatocytes, but competition experiments utilizing excess asialo-orosomucoid demonstrated that residual galactosyl residues were responsible for this association.