Insulin increases the cell surface concentration of α2-macroglobulin receptors in 3T3-L1 adipocytes

The present study shows that insulin causes an increase in the binding of az-macroglobulin (a2M) to 3T3LI adipocytes. Scatchard analysis of the binding at 4 “C indicated an approximate 2-fold increase in the number of asM binding sites, with no change in the apparent affinity of the receptor. In addition, a 2-3fold increase in the binding of monoclonal antibody 2C6, which recognizes a component of the azM receptor, was found in cells treated at 37 ‘C with insulin and then KCN to inhibit receptor endocytosis. An increased cellular accumulation of asM was also observed in response to insulin. Interestingly, the increase in the rate of accumulation of a2M was significantly smaller than the increase in the number of azM receptors on the cell surface, suggesting that the rate of ligand internalization or subsequent processing is altered in response to insulin. Ultrastructural analysis of the internalization pathway of the azM receptor was performed using colloidal gold-coupled 2C6 monoclonal antibody. Control cells incubated for 20 min at 37 “C with the gold-conjugated antibody displayed 40% of cellular gold particles on the cell surface and 60% within intracellular structures. In insulin-treated c lls this proportion was reversed, with 64% of the particles being found on the cell surface, and only 36% within intracellular structures. Significant differences in the distribution of gold particles among intracellular structures were detected between control and insulintreated cells. Whereas in control cells, 18% of the total cellular gold particles internalized into tubulovesicles and multivesicular bodies, in insulin-treated cells only 3% of the gold particles were found within these structures. These data indicate that the movement of this receptor between endocytic compartments is altered in response to insulin, and suggest that the effect of insulin to increase the cell surface concentration of a2M receptors and the accumulation of OZM is due, at least in part, to alterations in the endocytic portion of the receptor recycling pathway.

The present study shows that insulin causes an increase in the binding of az-macroglobulin (a2M) to 3T3-LI adipocytes. Scatchard analysis of the binding at 4 "C indicated an approximate 2-fold increase in the number of asM binding sites, with no change in the apparent affinity of the receptor. In addition, a 2-3fold increase in the binding of monoclonal antibody 2C6, which recognizes a component of the azM receptor, was found in cells treated at 37 'C with insulin and then KCN to inhibit receptor endocytosis. An increased cellular accumulation of asM was also observed in response to insulin. Interestingly, the increase in the rate of accumulation of a2M was significantly smaller than the increase in the number of a z M receptors on the cell surface, suggesting that the rate of ligand internalization or subsequent processing is altered in response to insulin. Ultrastructural analysis of the internalization pathway of the azM receptor was performed using colloidal gold-coupled 2C6 monoclonal antibody. Control cells incubated for 20 min at 37 "C with the gold-conjugated antibody displayed 40% of cellular gold particles on the cell surface and 60% within intracellular structures. In insulin-treated cells this proportion was reversed, with 64% of the particles being found on the cell surface, and only 36% within intracellular structures. Significant differences in the distribution of gold particles among intracellular structures were detected between control and insulintreated cells. Whereas in control cells, 18% of the total cellular gold particles internalized into tubulovesicles and multivesicular bodies, in insulin-treated cells only 3% of the gold particles were found within these structures. These data indicate that the movement of this receptor between endocytic compartments is altered in response to insulin, and suggest that the effect of insulin to increase the cell surface concentration of a2M receptors and the accumulation of OZM is due, at least in part, to alterations in the endocytic portion of the receptor recycling pathway.
Within minutes after stimulation of cells by insulin, a 3-5fold increase in the plasma membrane concentration of glucose transporters (1, Z), transferrin receptors (3-5), and IGF-* This work was supported in part by Juvenile Diabetes Foundation Grant 5-24155 and National Institutes of Health Grant DK 19525. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Partially supported by the Charles N. Cohen Award for Diabetes Research.
II'/mannose 6-phosphate receptors (6, 7) is observed. This increase is accompanied by a stimulation in the uptake of glucose, iron, and IGF-I1 by the cell. Thus, the redistribution of specific membrane proteins in response to insulin may be important in mediating the rapid hypoglycemic actions and growth promoting effects of this hormone.
Two of the proteins that are increased on the cell surface in response to insulin, the transferrin and IGF-II/mannose 6phosphate receptors, share a 'number of common features. Both are found predominantly in intracellular vesicular structures in the perinuclear region, and in clathrin-coated pits and vesicles at or near the cell surface (8,9), and constitutively recycle between the plasma membrane and intracellular endocytic structures (8,10). This observation raised the possibility that other receptors which display similar distribution patterns, and recycle through the coated pit/coated vesicle pathway would be similarly affected in their cell surface concentration in response to insulin. We addressed this possibility by investigating the effects of insulin on the receptor for ap-macroglobulin (a2M). This receptor is present in numerous primary and cultured cell lines, and binds a2M-protease complexes with high affinity. The receptor has been shown to internalize exclusively through clathrin-coated pits (11,12) and its subcellular distribution is identical to that of the transferrin receptor in cultured fibroblasts (13).
In the present study, we show that insulin increases the binding of azM to the cell surface of 3T3-Ll adipocytes and stimulates the uptake of this ligand. Several mechanisms could be involved in eliciting the rapid increase in the cell surface concentration of receptors in response to insulin. For example, insulin could trigger the exocytosis of a latent, nonrecycling pool of intracellular receptors. Alternatively, insulin could decrease the intracellular transit time of the receptor by accelerating the recycling of internalized receptors to the cell surface, or by decreasing receptor endocytosis. To gain insight into the mechanisms involved in this effect, we have performed an ultrastructural analysis of the endocytosis of the receptor using a colloidal gold-conjugated monoclonal antibody against a component of the receptor. Our results indicate that in insulin-treated cells, a large number of gold particles remain at or near the plasma membrane, and significantly less particles are internalized into late endocytic structures, such as tubulovesicles and multivesicular bodies. Taken together, these results suggest that the insulin-mediated increase in the cell surface concentration of anM receptors is due, at least in part, to alterations in both the endocytosis of The abbreviations used are: IGF-11, insulin-like growth factor-11; azM, a*-macroglobulin; BSA, bovine serum albumin; Hepes, 4-(2-hydroxyethy1)-1-piperazineethanesulfonic acid. the receptor, and in its movement among intracellular endocytic compartments.
EXPERIMENTAL PROCEDURES az-Macroglobulin and Monoclonal Antibody 2C6--a2-Macroglobulin was isolated from outdated human plasma by the procedure of Swenson and Howard (14). This procedure involves precipitation with ammonium sulfate, shown to convert azM to the electrophoretically fast form which does not require further activation by primary amines or proteases in order for binding to high affinity receptors to occur (15). More than 85% of the preparation migrated in polyacrylamide gels as an 185-kDa polypeptide, indicating that relatively little proteolysis occurred during the isolation procedure. Purified azM was iodinated using the Enzymobeads procedure (Bio-Rad) to a specific activity of 0.8-1.2 Ci/pmol.
Monoclonal antibody 2C6 was obtained from Dr. John Hanover (National Institutes of Health) and has been reported to interact with the a2M receptor (13,16). For conjugation to colloidal gold, the antibody was extensively dialyzed against 2 mM boric acidborax buffer, pH 9.0, and added at a final concentration of 1.5 mg/100 ml to a suspension of 15-nm gold particles prepared by the method of Slot and Geuze (17). The preparation was stabilized with bovine serum albumin (BSA), and diluted in 10 mM phosphate buffer, pH 7.4, containing 0.1% BSA. The unreacted antibody was removed by column chromatography (Ultrogel AcA 44, LKB Products), and the gold conjugate was concentrated by Amicon filtration (YM100 membranes). After sterilization by filtration through 0.22-pm filters, the preparation was stored at 4 "C and used within 10 days of preparation. Experiments were performed using a final particle molarity of 10 nM.
Cell Culture"ST3-Ll cells were grown in 24-or 6-well multidishes (Nunc, Denmark) for binding experiments or studies involving electron microscopy, respectively. Fibroblasts were seeded and fed every 2 days in Dulbecco's modified Eagle's medium supplemented with 0.75 mg/l glutamine, nonessential amino acids (Gibco), 50 units/ml penicillin, 50 pg/ml streptomycin, and 10% fetal calf serum (Hyclone), and grown under 10% COz. At confluence, differentiation was started by addition of culture medium containing 0.25 p M dexamethasone (Sigma) and 0.5 mM isobutylmethylxanthine (Sigma). After 48 h, this was replaced with fresh medium, which was not removed until experiments were performed. Lipid droplets were observed in 8 0 4 5 % of the cells 4 days after initiating differentiation. Experiments were performed on the sixth day of differentiation.
Binding and Uptake of apM and Monoclonal Antibody 2C6-On the sixth day of differentiation, monolayers were washed three times with 1 ml of Krehs-Ringer/Hepes buffer supplemented with 2 mM sodium pyruvate and 2% BSA, and equilibrated for 30 min at 37 "C. Insulin (crystalline porcine, Eli Lilly) was then added at various concentrations for 5-30 min, as indicated in each experiment. The buffer was aspirated, replaced with ice-cold buffer, and dishes were placed on ice. '251-a2M was added at the indicated concentrations, and incubations were continued for 4 h. The monolayers were rapidly washed three times with 1 ml of ice-cold buffer, dissolved with 300 pl of 1 N sodium hydroxide, and the radioactivity counted in an LKB ycounter. Nonspecific binding was estimated by the addition of 5 ~L M unlabeled azM and represented approximately 15% of the total binding. For the measurement of aZM accumulation, cells were washed and equilibrated for 30 min as described above. Insulin was then added at 100 nM for 10 min, followed by 1 nM '251--a~M. At the time points indicated, the wells were washed three times with ice-cold buffer, dissolved, and radioactivity counted as described above.
For the study of antibody 2C6 binding, monolayers were washed and treated with insulin at the concentrations indicated in each experiment. The buffer was then replaced with ice-cold buffer containing the indicated concentrations of the antibody, and monolayers were placed on ice. After 2 h, the monolayers were washed with 1 ml of ice-cold buffer, and incubated for 45 min with 10 pg/ml of a mouse anti-rat immunoglobulin (Boehringer Mannheim). The wells were then washed twice with 1 ml of buffer, and incubated for 30 min in the presence of 5 pl/ml 1251-protein A (Du Pont-New England Nuclear). Alternatively, after insulin treatment KCN was added at a final concentration of 10 mM and incubations were continued for 20 min at 37 "C. Antibody binding was then measured as described above, but all procedures were conducted at room temperature, and with buffer containing 10 mM KCN. Nonspecific binding was estimated in incubations containing mouse anti-rat second antibody and 1251-protein A, and represented less than 20% of the binding in the presence of monoclonal antibody 2C6.
Analysis of Gold-2c6 Distribution-Gold-conjugated 2C6 was prepared as described above. Cells, grown in 6-well multidishes, were washed 3 times with 5 ml of buffer and equilibrated for 30 min as described above. Insulin was then added to a final concentration of 100 nM, and incubations were continued for 10 min at 37 "C. At this time, gold-2C6 was added at a final particle molarity of 10 nM, and incubations continued for 20 min with gentle shaking of the multiwell dishes. At this point, the gold suspension was aspirated, and the monolayers washed four times with 5-10 ml of ice-cold Krebs-Ringer/ Hepes without BSA. The monolayers were then fixed by addition of 2% glutaraldehyde for 1 min, scraped into microfuge tubes, and spun for 5 min at 12,000 X g. The pellets were then processed for electron microscopy as described (18) and tissue blocks sectioned and examined for gold particles. Certain definitions were adhered to in the analysis of gold particle distribution. Plasma membrane invaginations which were coated in their cytoplasmic surface by a bristle structure were defined as coated pits. Gold particles were also detected in small membrane invaginations which were smooth in their cytoplasmic surface, and were called smooth invaginations. Membrane enclosed structures within the cytoplasm were defined as vesicles. Those with a bristle coating were defined as coated vesicles. Smaller vesicles were round, had a clear lumen, and lacked a bristle coat. Tubulovesicles were slightly larger, had a clear lumen, and an elongated or irregular contour. Multivesicular bodies were large vesicles that contained small vesicular structures within their lumen, but no discernible electron-dense material. Lysosomes were defined by the presence of electron dense material within their lumen. Approximately 50 photographs were taken from each tissue block, and two to three separate tissue blocks were analyzed for each condition in each of three experiments. The gold particles found in different structures were tallied, and the results expressed as a percentage of total gold particles counted. These ranged from 400-700/condition in three different experiments, in which independent preparations of gold-2C6 were employed. Similar results were obtained in each experiment.

RESULTS
The binding of lZ5I-azM to the cell surface of 3T3-Ll cells on day 6 of differentiation was analyzed. Cell monolayers were incubated with increasing concentrations of the ligand at 4 "C, to prevent receptor endocytosis. After 4 h, the amount of a2M specifically bound to the cell monolayers was measured. The binding of a2M to control cell monolayers saturated at a concentration of approximately 0.5 nM (Fig. 1, .!eftpanel).  blasts (19) or Swiss 3T3 mouse fibroblasts (20). Treatment of cell monolayers with 100 nM insulin for 10 min prior to cooling to 5 "C resulted in an increase in '251-azM binding to the cell surface. Scatchard analysis indicated an increase in maximal binding, from 16 to 24 fmol/106 cells. No consistent difference in the Kd was observed.
The functional consequence of the increased binding of a2M to the cell surface was investigated by measuring the effect of insulin on the accumulation of a2M at 37 "C by cell monolayers. The values obtained in these experiments result from both the kinetics of binding of ~z M to its receptor, and from the kinetics of internalization of the ligand-receptor complex. Half-maximal binding of ~z M to its receptor occurs within 5 min at 22 "C (15), and more rapidly at 37 "C.' Thus the rate of accumulation of anM by cells results mainly from the rate of uptake and intracellular processing of the ligand. The rate of accumulation of azM by the cells was approximately 1 fmol/min/106 cells during the first 10 min of incubation (Fig. 2). The accumulation of a2M by cells that were pretreated for 10 min with 100 nM insulin before addition of the ligand was increased only by approximately 30%, despite the fact that under these conditions the binding of azM to the cell surface was increased by almost 2-fold (Fig. 1). After 10 min of incubation the rate of anM accumulation increased in both control and insulin-treated cells, and at 20 min, the cellassociated anM was 30 f 2 and 43 f 3 fmol/106 cells in control or insulin-treated monolayers, respectively. After 20 min at 37 "C degradation products of a2M started to appear, as detected by the presence of 20-90-kDa polypeptides in autoradiographies of 6-16% polyacrylamide gradient gels of cell extracts (data not shown). These results indicate that the increase in the number of a2M receptors on the cell surface is associated with an overall increase in the accumulation of Recently, a panel of monoclonal antibodies that recognize an 180/90-kDa glycoprotein found in coated pits and vesicles of NIH 3T3 cells has been raised (13). One of these antibodies (2C6) inhibits aZM binding and uptake when incubated with cells at 37 "C. These and other data suggest that the glyco- protein recognized by antibody 2C6 represents the azM receptor (13) . Fig. 3, left panel, shows that antibody 2C6 bound to the surface of 3T3-Ll cells, reaching saturation at approximately 10 pg/ml. The effect of insulin on the binding of this monoclonal antibody to the cell surface was compared to the effect of insulin on azM binding. Cell monolayers were preincubated for 10 min at 37 "C with increasing concentrations of insulin, and chilled to 5 "C. Half of the wells were incubated with 10 pg/ml of antibody 2C6, followed by a second antibody and '251-protein A as described under "Experimental Procedures." The other half were incubated with 1 nM 1251-azM for 4 h. Fig. 3, right panel, shows that insulin increased the cell surface binding of antibody 2C6 and anM by approximately 2-and 1.8-fold, respectively. Half-maximal stimulation of both antibody binding and aZM binding occurred at 1-5 nM insulin. These data support the notion that antibody 2C6 binds to the a2M receptor, and that the insulin mediated increase in azM binding is due to an approximately %fold increase in the number of receptors on the cell surface.
The discrepancy between the increase in cell surface receptors and the initial rate of accumulation of a2M suggests that insulin may cause a relative decrease in the rate of internalization or subsequent processing of a2M. However, the possibility remains that the comparison between the receptor numbers, measured at 4 "C, and the uptake of aZM, measured at 37 "C, may not be valid if some redistribution of the receptors occurs at 4 "C. To address this possibility, we measured the number of cell surface receptors in cells that had been treated with KCN at 37 "C. This treatment inhibits receptor internalization, and allows the measurement of cell surface receptor numbers at higher temperatures (3,6). Binding of antibody 2C6 to control or insulin-treated KCN-poisoned cells, assessed as described under "Experimental Procedures," was of 119 k 5 and 421 f 21 cpm, respectively (data are means f S.E. of 2 independent experiments performed in triplicate).
The values obtained in parallel experiments in which antibody binding was measured at 4 "C were 478 f 28 and 1013 f 78 cpm for control or insulin-treated cells, respectively (means f S.E.). These results suggest that redistribution of receptors to the cell surface indeed occurs at 4 "C, both in control and in insulin-treated cells. Nevertheless, the results indicate that the previously observed 2-fold increase in cell surface receptor numbers in response to insulin is not an artifact of this redistribution. In fact, an approximately %fold increase in cell surface receptor numbers in response to insulin was observed in KCN-treated cells.
We sought to identify the structures along the endocytic/ recycling pathway that might be regulated by insulin to bring about an increase in the cell surface concentration of a2M receptors. Control or insulin-treated cell monolayers were incubated for 20 min at 37 "C with saturating concentrations of antibody 2C6 conjugated to 15-nm colloidal gold particles. In both control or insulin-treated cells, gold particles were detected at the cell surface in undifferentiated plasma membrane, coated pits, or smooth invaginations. Intracellular gold particles were found in coated vesicles, smooth uncoated vesicles proximal to the plasma membrane, tubulovesicles, multivesicular bodies, and large vesicles containing electron dense material which resembled lysosomes (Fig. 4). Quantitative analysis of the distribution of gold particles was performed as described under "Experimental Procedures." This analysis revealed that after 20 min of incubation of control cell monolayers with gold-2C6 antibody, 60% of the gold particles were intracellular and approximately 40% were localized on the cell surface. In contrast, in insulin-treated cells this proportion was reversed, with only 36% of the gold particles found inside the cell and 64% on the cell surface. The distribution of gold particles among cell surface and intracellular structures of control or insulin-treated cells is shown in Fig. 5. The percentage of total cellular particles on the cell surface was significantly higher in insulin-treated cells. These particles did not selectively accumulate in any specific plasma membrane structure, but were distributed among membrane, coated pits, and smooth membrane invaginations in a proportion similar to that observed in control cells. In contrast, significant differences in the distribution of particles among intracellular structures could be detected between control and insulin-treated cells. Whereas in control cells 18% of the total cellular particles were found within tubulovesicles and multivesicular bodies (equivalent to 30% of intracellular particles), insulin-treated cells contained only 3% of total cellular gold particles within these structures (equivalent to 8% of intracellular particles). In addition, the number of particles associated to lysosome-like structures, which amounted to 9% of total particles in control cells decreased to 3% upon insulin treatment. These results suggest that the movement of receptors among endocytic compartments may be altered in response to insulin, and this effect may serve to shunt a large population of receptors into the recycling pathway. The effect of insulin on the distribution of aZM receptors is consistent with its effects on the binding and accumulation of a2M, and with the possibility that the insulin-mediated increase in the cell surface concentration of a2M receptors results from a decrease in the intracellular transit time of the receptor, and/or a decrease in its rate of internalization.

DISCUSSION
The main finding in this paper is that insulin increases the cell surface concentration of receptors for a2M, and stimulates the accumulation of this ligand. Together with previous data indicating that insulin increases the cell surface concentration of the receptors for transferrin (3)(4)(5) and IGF-II/Man-6-P (6, 7), these results indicate that insulin affects the distribution of numerous receptors that internalize and recycle through the clathrin-coated pit/coated vesicle pathway. This effect may be important in mediating the overall anabolic growthpromoting effects of this hormone, by promoting an increased rate of nutrient and growth factor uptake.
The physiological role of a2M is not completely understood. This protein is considered an important scavenger of circu-lating proteases, inhibiting their activity against high molecular weight polypeptides and retrieving them into the cell for degradation (reviewed in Ref. 21). Moreover, aZM has been shown to suppress protease secretion by some cell types (22). In addition, azM binds polypeptide growth factors with high affinity (23,24), and thereby produces effects on cell growth. Whether or not these properties of aZM relate to the metabolic or growth promoting effects of insulin is presently unknown.
Several mechanisms could contribute to the insulin-mediated increase in the cell surface concentration of the receptors for transferrin, IGF-II/Man-6-P, and azM. For example, the exocytosis of a latent pool of intracellular receptors could increase their concentration on the cell surface (4,5). Alternatively, the endocytic portion of the pathway could be altered by insulin, in a way that would increase the half-life of the receptor on the cell surface, and decrease its intracellular halflife. This mechanism is consistent with studies which indicate that the effects of insulin and insulin-like growth factor I to increase the cell surface concentration of transferrin receptors are at least in part due to a decrease in the rate of receptor endocytosis (4). In addition, the rate of receptor internalization, but not the rate of exocytosis, appears to determine the number of cell surface transferrin receptors expressed in murine erythroleukemic cells (25). The biochemical analysis presented in this paper shows that the initial rate of accumulation of aZM is increased only 30%, under conditions where the number of cell surface receptors is increased almost 2-fold (Figs. 1 and 2 and "Results"), suggesting that insulin causes a relative decrease in the rate of receptor endocytosis.
The internalization pathway followed by the receptor in control or insulin-treated cells was studied using a monoclonal antibody to a component of the aZM receptor. We have employed a protocol in which control or insulin-treated cells are incubated at 37 "C in the presence of high concentrations of gold-labeled antibody. This protocol differs from most ultrastructural studies of receptor endocytosis, in which cells are equilibrated at 4 "C with the labeled ligand prior to increasing the temperature to allow endocytosis to occur. This latter type of protocol was precluded by our observation that cooling cells to 4 "C elicits a redistribution of aZM receptors ("Results"). Our data are consistent with previous studies which have delineated the intracellular pathway followed by azM in 3T3-Ll cells (26), in that gold particles are found associated with coated pits, coated and uncoated vesicles, tubulovesicles, and multivesicular bodies. Few gold particles were found in lysosome-like structures, consistent with the notion that the receptor does not enter these structures under physiological conditions. Thus, although antibody 2C6 has been reported to cause down-regulation of the aZM receptor in Swiss 3T3 cells (13), no qualitative changes in the intra-ceIlular pathway of this receptor appear to occur in response to antibody binding. The ultrastructural analysis presented in this paper revealed two interesting effects of insulin. First, that the proportion of particles remaining on the cell surface is greater in insulin-treated cells compared to controls. These data are consistent with those obtained in the study of azM accumulation (Fig. Z), in that they suggest a relative decrease in the rate of receptor internalization in response to insulin. Second, whereas in control cells gold particles accumulate in tubulovesicles and multivesicular bodies, in insulin-treated cells significantly fewer particles were found within these structures. The total number of multivesicular bodies did not appear to be decreased, indicating that it is the transit of the gold-antibody to these structures which is altered by insulin treatment. Although we cannot rule out the possibility that the antibody increases the concentration of receptors that enter late endocytic compartments under basal conditions, it is apparent that under these conditions insulin inhibits the movement of the receptor into these structures. Under physiological conditions, an analogous effect could decrease the intracellular transit time of the receptors and thereby cause a steady-state increase in their cell surface concentration. Because these studies do not provide information on the intracellular pathway of aZM itself, the consequences of the changes in the transit of the receptor on aZM accumulation in multivesicular bodies or lysosomes are not known. It will be interesting to investigate whether this effect of insulin on the subcellular distribution of a2M receptors is accompanied by structural modifications of the receptor structure similar to those that have been shown to occur on the IGF-II/Man-6-P receptor (27).
In summary, it has been shown that insulin treatment of differentiated 3T3-Ll cells causes an increase in the number of cell surface receptors for azM, and a concomitant stimulation of the uptake of this ligand. Insulin appears to cause a relative decrease in the rate of azM receptor internalization. Ultrastructural analysis of the uptake and transit of the azM receptor suggests that insulin may inhibit the movement of this receptor among intracellular endocytic structures. Both of these mechanisms may operate to increase the number of azM receptors on the cell surface.