Megalin (gp330) Is an Endocytic Receptor for Thyroglobulin on Cultured Fisher Rat Thyroid Cells*

We recently reported that megalin (gp330), an endocytic receptor found on the apical surface of thyroid cells, binds thyroglobulin (Tg) with high affinity in solid phase assays. Megalin-bound Tg was releasable by heparin. Here we show that Fisher rat thyroid (FRTL-5) cells, a differentiated rat thyroid cell line, can bind and endocytose Tg via megalin. We first demonstrated that FRTL-5 cells express megalin in a thyroid-stimulating hormone-dependent manner. Evidence of Tg binding to megalin on FRTL-5 cells and on an immortalized rat renal proximal tubule cell line (IRPT cells), was obtained by incubating the cells with125I-Tg, followed by chemical cross-linking and immunoprecipitation of 125I-Tg with antibodies against megalin. To investigate cell binding further, we developed an assay in which cells were incubated with unlabeled Tg at 4 °C, followed by incubation with heparin, which released almost all of the cell-bound Tg into the medium. In solid phase experiments designed to illuminate the mechanism of heparin release, we demonstrated that Tg is a heparin-binding protein, as are several megalin ligands. The amount of Tg released by heparin from FRTL-5 and IRPT cells, measured by enzyme-linked immunosorbent assay (ELISA), was markedly reduced by two megalin competitors, receptor-associated protein (RAP) and 1H2 (monoclonal antibody against megalin), indicating that much of the Tg released by heparin had been bound to megalin (∼60–80%). The amount inhibited by RAP was considered to represent specific binding to megalin, which was saturable and of high affinity (K d ∼11.2 nm). Tg endocytosis by FRTL-5 and IRPT cells was demonstrated in experiments in which cells were incubated with unlabeled Tg at 37 °C, followed by heparin to remove cell-bound Tg. The amount of Tg internalized (measured by ELISA in the cell lysates) was reduced by RAP and 1H2, indicating that Tg endocytosis is partially mediated by megalin.

I-Tg with antibodies against megalin. To investigate cell binding further, we developed an assay in which cells were incubated with unlabeled Tg at 4°C, followed by incubation with heparin, which released almost all of the cell-bound Tg into the medium. In solid phase experiments designed to illuminate the mechanism of heparin release, we demonstrated that Tg is a heparin-binding protein, as are several megalin ligands. The amount of Tg released by heparin from FRTL-5 and IRPT cells, measured by enzyme-linked immunosorbent assay (ELISA), was markedly reduced by two megalin competitors, receptor-associated protein (RAP) and 1H2 (monoclonal antibody against megalin), indicating that much of the Tg released by heparin had been bound to megalin (ϳ60 -80%). The amount inhibited by RAP was considered to represent specific binding to megalin, which was saturable and of high affinity (K d ϳ11.2 nM). Tg endocytosis by FRTL-5 and IRPT cells was demonstrated in experiments in which cells were incubated with unlabeled Tg at 37°C, followed by heparin to remove cellbound Tg. The amount of Tg internalized (measured by ELISA in the cell lysates) was reduced by RAP and 1H2, indicating that Tg endocytosis is partially mediated by megalin.
Thyroglobulin (Tg) 1 is synthesized in thyrocytes and re-leased into the follicle lumen, where it is stored as the major component of the colloid (1,2). Post-transitional modifications of Tg that occur mainly at the cell-colloid interface lead to forms that are iodine-rich and that contain the thyroid hormones T4 and T3 (mature Tg). Hormone secretion requires uptake of Tg by thyrocytes, with transport to lysosomes, where proteolytic cleavage leads to release of the hormones from mature Tg molecules (1). Internalization of Tg may result from pseudopod ingestion under certain circumstances, such as intense, acute stimulation by the thyrotropic hormone (TSH), but micropinocytosis (vesicular internalization) is thought to be the usual route (1)(2)(3). There is evidence that micropinocytosis of Tg can take place both by nonselective fluid phase uptake and receptor-mediated endocytosis, but the relative importance of these two mechanisms is uncertain (1). Although evidence has been obtained of low affinity receptors for Tg on thyroid cells, a receptor capable of mediating Tg endocytosis has not been fully characterized (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12).
We have previously obtained evidence suggesting that megalin (gp330) may function as a receptor for Tg (13). Megalin (gp330) is a member of the LDL receptor family (14,15) and has been shown to bind multiple, unrelated ligands and to mediate endocytosis of ligands via coated pits, leading to delivery of ligands to lysosomes, where degradation occurs (16 -18). In immunohistochemical studies, megalin has been found principally on the apical surface of a restricted group of absorptive epithelial cells, including renal proximal tubule cells, epididymal cells, type II pneumocytes, and thyroid epithelial cells (19,20). Based on the assumption that physiological ligands of megalin may be identified by consideration of the composition of fluids to which it is exposed in various organs (21), we postulated that megalin on thyrocytes serves as a receptor for Tg. In support of this possibility we demonstrated in solid phase assays that purified rat megalin binds to rat Tg with high affinity (13). Binding was inhibited by several known competitors of megalin, including the receptor-associated protein (RAP), antibodies to megalin, and heparin, which in addition dissociated Tg bound to megalin (13). In the present study we have investigated FRTL-5 cells to determine whether they express megalin and if they are capable of binding and internalizing Tg via megalin. FRTL-5 cells, an established rat thyroid cell line, exhibit a number of thyroid-specific functions in a TSH-dependent manner (22,23). Here we show that FRTL-5 cells do express megalin and that they can bind and endocytose Tg via megalin.

EXPERIMENTAL PROCEDURES
Materials-Tg was purified from rat thyroids by ammonium sulfate precipitation and column fractionation, as described previously (24). * This work was supported by NIDDK, National Institutes of Health Grant 46301. 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. The Tg preparations were analyzed by Western blotting, using a rabbit anti-human Tg antibody cross-reactive with Tg from other species (Axle-Westbury, NY). Two bands were seen at about 330 and 660 kDa, corresponding to the monomeric and the dimeric forms of Tg. The RAP was used in the form of a glutathione S-transferase (GST) fusion protein. DH5␣ bacteria harboring the pGEX-RAP expression construct were kindly provided by Dr. Joachim Herz (University of Texas Southern Medical Center, Dallas, TX). The production of RAP-GST and GST was performed as described (25). Heparin (Sigma) was used because it inhibits Tg binding to megalin in solid phase assays (13), as occurs for binding of other megalin ligands (26). Furthermore, heparin dissociates Tg bound to megalin (13) and releases ligands that are bound to members of the LDL receptor family (27).
A rabbit antibody, designated A55, against immunoaffinity purified megalin was described previously (28,29). A previously described mouse anti-megalin monoclonal antibody, designated 1H2, has been shown to react with ectodomain epitopes in the second cluster of ligand binding repeats (29). A rabbit antibody against RAP-GST and a mouse antibody against the low density lipoprotein receptor-related protein (LRP) were previously described (19). The above antibodies were used as purified IgG preparations unless otherwise specified. A goat antibody against GST was purchased from Amersham Pharmacia Biotech. Alkaline phosphatase (ALP)-conjugated goat anti-rabbit IgG and horseradish peroxidase-conjugated goat anti-rabbit IgG were obtained from Bio-Rad. ALP-conjugated mouse anti-goat IgG was obtained from Axle. Horseradish peroxidase-and ALP-conjugated goat anti-mouse IgG were obtained from Sigma. Fluorescein isothiocyanate (FITC)-conjugated rabbit anti-mouse IgG and goat anti-rabbit IgG were obtained from Cappel (Durham, NC). Biotin-labeled goat anti-mouse IgG was obtained from Vector (Burlingame, CA).
Radiolabeling of Tg-200 g of rat Tg were radiolabeled with 125 I-Na (NEN Life Science Products) using IODO beads (Pierce), according to the manufacturer's instructions. The specific activity of the preparations ranged from 1,500 to 7,000 cpm/ng.
Cell Cultures-Fisher rat thyroid cells (FRTL-5) (American Type Culture Collection, Manassas, VA) were cultured as described previously (22,23) in Coon's F12 medium, supplemented with 5% fetal calf serum and with a mixture of six hormones, including TSH (10 milliunits/ml). In some experiments the medium was replaced with fresh medium lacking TSH for 24 h, after which TSH was sometimes readded. An immortalized rat renal proximal tubule cell line (IRPT), which expresses megalin but not LRP, was established as described previously (30,31). IRPT cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. CHO cells, a Chinese hamster ovary cell line that lacks megalin and expresses LRP (32), were cultured in Ham's medium supplemented with 10% fetal bovine serum.
Evaluation of Megalin and LRP Expression by Flow Cytometry (FACS Analysis)-Cells were detached from the plates, washed with Tris buffered-saline (TBS), and incubated in plastic tubes for 1 h at 4°C with the mouse monoclonal anti-megalin antibody (1H2), with the rabbit polyclonal anti-megalin antibody (A55), with the anti-LRP antibody (tissue culture supernatant), or, as controls, with purified mouse or rabbit IgG. Anti-megalin antibodies and control IgGs were used at a concentration of 20 g/ml in binding buffer (TBS, 5 mM CaCl 2 , 0.5 mM MgCl 2 , 5% fetal bovine serum). After washing with TBS, FITC-conjugated rabbit anti-mouse IgG (1:250) or goat anti-rabbit IgG (1:1500) secondary antibodies were added for 1 h at 4°C in binding buffer. Cells were then washed and analyzed by flow cytometry using a FACSCAN from Becton Dickinson (Mountain View, CA).
Detection of Megalin by Western Blotting-Cell extracts were prepared with 1% Triton X-100, 1% deoxycholate (both from Fisher) and were briefly sonicated. The extracts were then subjected to SDS-polyacrylamide gel electrophoresis under nonreducing conditions and blotted onto nitrocellulose membranes, which were incubated with tissue culture supernatant containing the mouse monoclonal anti-megalin antibody (1H2) or with the rabbit polyclonal anti-megalin antibody (A55), followed by horseradish peroxidase-conjugated goat anti-mouse or anti-rabbit IgG secondary antibodies. Bands were detected using a chemiluminescent substrate kit (Kirkegard & Perry Laboratories, Gaithersburg, MD).
Evaluation of Megalin Expression by Immunoperoxidase Staining-Cells were cultured on glass coverslips and fixed with 2% paraformaldehyde-L-lysine-sodium periodate. The preparations were blocked with avidin D (0.1 mg/ml in PBS) and incubated with tissue culture supernatant containing the mouse monoclonal anti-megalin antibody (1H2) followed by biotin-labeled goat anti-mouse IgG secondary antibody (1: 200 in PBS) or with the secondary antibody alone. Staining was detected using the peroxidase labeled biotin-avidin complex (Vector). 3,3Ј-Diaminobenzidine (Sigma) was used for the color reaction.
Cross-linking Experiments-Cells were incubated with serum-free Dulbecco's modified Eagle's medium for 1 h at 37°C and then blocked in 0.5% ovalbumin (OVA, Sigma), 25 mM Tris, 150 mM NaCl, 5 mM CaCl 2 , 0.5 mM MgCl 2, pH 8.0 , for 1 h at 4°C. Cells were then incubated with 10 nM of 125 I-Tg for 4 h at 4°C in blocking buffer, followed by cross-linking with the homobifunctional cross-linker 3,3Ј-dithiobis[sulfosuccinimidyl propionate] (Pierce) (0.5 mM in PBS) for 30 min at 4°C. After washing with TBS, cells were lysed with 1% Triton X-100, 1% deoxycholate and briefly sonicated. Cell lysates (500 g) were immunoprecipitated with the rabbit anti-megalin antibody (A55 anti-serum, 1:200) or with the rabbit anti-human Tg antibody (1:200), or, as control, with normal rabbit serum, using protein A beads (Pierce). The precipitates were separated by 4 -16% SDS-polyacrylamide gel electrophoresis under nonreducing conditions and analyzed by autoradiography. Radioactivity in the cell lysates and in the precipitates was counted with a ␥ counter.
Release of Cell Bound Exogenous Tg by Heparin-Cells were cultured in 96-well tissue culture plates until 80 -100% confluence was reached. Cells were incubated with unlabeled Tg or, as controls, with RAP-GST or GST in Coon's F12 medium containing 5 mM CaCl 2, 0.5 mM MgCl 2 , and 0.5% OVA. For inhibition experiments, FRTL-5 and IRPT cells were incubated with Tg alone, or in the presence of megalin inhibitors, namely RAP-GST (200 g/ml) or 1H2 (200 g/ml) or, as controls, with GST (200 g/ml) or normal mouse IgG (200 g/ml). After 4 h of incubation at 4°C, the cells were washed twice with ice-cold PBS to remove nonspecifically bound proteins and then incubated for 1 h at 4°C with ice-cold heparin (10 -100 units/ml in PBS), to release receptor-bound proteins from the cell surface. In certain experiments, to measure total cell-bound Tg, cell lysates were prepared by treating the cells with ice-cold H 2 O on ice, immediately after incubation with Tg at 4°C. The amounts of Tg, RAP-GST or GST in the heparin wash or in the cell lysates were measured by ELISA. For this purpose, 96-well microtiter plates were coated overnight at 4°C with the heparin wash or with the cell lysates, blocked with bovine serum albumin (Sigma) and incubated with rabbit anti-human Tg (1:500), rabbit anti-RAP (20 g/ml), or goat anti-GST (1:1000) antibodies, followed by ALP-conjugated goat antirabbit IgG (1:3000) or mouse anti-goat IgG (1:2500) secondary antibodies. After incubation with p-nitrophenyl-phosphate (Sigma), absorbance at 405 nM was determined with an El-311 ELISA microplate reader. The amount of Tg, RAP-GST, or GST was calculated using a standard obtained by coating the wells with 10 g/ml of purified Tg, RAP-GST, or GST and was normalized for the total amount of protein in the cell lysates, measured with a commercial kit (Bio-Rad).
Binding of Tg to Heparin in Solid Phase Assays-96-well microtiter plates were coated overnight at 4°C with rat Tg or, as control, with OVA, at a concentration of 100 g/ml in PBS. After blocking with bovine serum albumin, plates were incubated for 3 h at room temperature with a biotin-labeled heparin-albumin complex (Sigma) or, as control, with biotin-labeled albumin (Sigma) (in PBS, 0.05% Tween 20, 0.5% bovine serum albumin), followed by ALP-conjugated streptavidin (Vector, 1:3000) and p-nitrophenyl-phosphate. Absorbance was determined at 405 nM. For inhibition experiments, biotin-labeled heparin-albumin was added to the wells alone or in the presence of unlabeled heparin (500 units/ml).
Endocytosis Experiments-Cells were seeded in 96-well tissue culture plates and cultured until 80 -100% confluence was reached. Cells were then incubated at 37°C with unlabeled rat Tg or, as controls, with unlabeled RAP-GST or GST in Coon's F12 medium containing 5 mM CaCl 2, 0.5 mM MgCl 2 , and 0.5% OVA. For inhibition experiments, Tg (10 g/ml) was added to the plates together with one of the following: RAP-GST (200 g/ml), 1H2 (200 g/ml), or, as controls, GST (200 g/ml) or normal mouse IgG (200 g/ml). After 6 h of incubation the cells were washed twice with PBS and incubated with ice-cold heparin, as described above, to remove cell surface bound proteins, which were measured by ELISA. The cells were then lysed to measure internalized Tg, RAP-GST, or GST, using deionized H 2 O on ice. The amount of cell protein was measured in an aliquot of the cell lysate. Internalized Tg, RAP-GST or GST were measured in the cell lysates by ELISA, as described above. Fig.  1A, FRTL-5 cells were found to express megalin by FACS analysis, using either the mouse monoclonal (1H2) or the rabbit polyclonal (A55) anti-megalin antibodies. In addition, immunoperoxidase staining with 1H2 demonstrated surface megalin on FRTL-5 cells (Fig. 1B). Megalin was also detected in FRTL-5 cell extracts by Western blotting (Fig. 1C). In agreement with previous reports (30,31), IRPT cells were found to express megalin by FACS analysis, whereas megalin was not detected on CHO cells (not shown). Because LRP binds many of the same ligands as megalin, we studied its expression on FRTL-5 cells by FACS analysis and, as shown in Fig. 1D, no LRP was found. In a previous study we showed that IRPT cells do not express surface LRP (30).

Expression of Megalin by Cultured Cells-As shown in
Expression of Megalin by FRTL-5 Cells and TSH Dependence-To investigate whether megalin expression by FRTL-5 cells is regulated by TSH, cells were cultured for 24 h in medium lacking TSH, sometimes followed by 48 h of culture in fresh medium containing TSH. As shown in Fig. 2, megalin expression was reduced at 24 h after TSH deprivation, as assessed by FACS analysis. The staining for megalin ranged from 46 to 69% of the staining before TSH deprivation, depending on the antibody used. Readdition of TSH resulted in increased megalin expression, reaching 88 -90% of the levels found before TSH deprivation. The results indicate that TSH is required for high levels of megalin expression on FRTL-5 cells.
Cross-linking and Immunoprecipitation of 125 I-Tg with Antibodies against Megalin-In experiments designed to obtain evidence of Tg binding to cell surface megalin, FRTL-5 and IRPT cells were incubated with 125 I-Tg at 4°C, followed by cross-linking and immunoprecipitation with anti-Tg or antimegalin antibodies. As shown in Fig. 3A, both anti-Tg and anti-megalin antibodies produced a high molecular mass band at the same size, indicating Tg bound to megalin on the cell surface. Another band at approximately 50 kDa was produced by the anti-megalin antibody and not by the anti-Tg antibody. The identity of this product is unknown.
Binding of Tg to Megalin on Cultured Cells, Assessed by Heparin Release-We previously showed that heparin dissociates Tg from megalin in solid phase assays (13). Here we studied the ability of heparin to dissociate Tg bound to megalin on cells, to be used as a measure of cell bound Tg. Cells were incubated with unlabeled Tg at 4°C, followed by incubation with heparin. In addition, some cells were incubated with RAP (used as a RAP-GST fusion protein), as a positive megalin binding control, or with GST, as a negative control, before heparin treatment. The amounts of Tg, RAP-GST, and GST in the heparin wash were measured by ELISA.
As shown in Fig. 4A, heparin was found to release Tg and FIG. 1. A, detection of megalin on FRTL-5 cells by FACS analysis. Cells were detached from the plates and incubated at 4°C with the rabbit (A55) or with the mouse (1H2) anti-megalin antibodies, followed by secondary antibodies (FITC-conjugated goat anti-rabbit or antimouse IgG). Control, mouse IgG. The results are representative of one of three experiments performed. B, detection of megalin on FRTL-5 cells by immunoperoxidase staining. Cells were cultured on glass coverslips, fixed and blocked with avidin D, and then incubated with 1H2 followed by a biotin-labeled goat anti-mouse IgG secondary antibody, or with the secondary antibody alone (Control). Staining was detected by the peroxidase-labeled biotin-avidin complex. Surface staining is produced by 1H2 on most cells and some are intensely stained, with a punctate pattern. X100. C, expression of megalin by FRTL-5 cells, detected by Western blotting. Cell extracts were subjected to SDS-polyacrylamide gel electrophoresis under nonreducing conditions and blotted onto nitrocellulose membranes, which were incubated with 1H2, followed by horseradish peroxidase-conjugated goat anti-mouse IgG secondary antibody. The arrow indicates the only band observed, which corresponded to megalin. D, absence of LRP on FRTL-5 cells, assessed by FACS analysis. Cells were incubated with the anti-LRP antibody or, as a control, with purified mouse IgG, followed by FITC-conjugated antimouse IgG secondary antibody. RAP-GST from FRTL-5 and IRPT cells. Although FRTL-5 cells are capable of synthesizing Tg (22,23), the amount of Tg released by heparin from FRTL-5 cells incubated in medium lacking Tg was negligible. RAP-GST was also released from CHO cells, which express LRP (32) but not megalin. The amount of Tg released by heparin from CHO cells was negligible as compared with FRTL-5 and IRPT cells (Fig. 4A). No release of GST was observed from FRTL-5, IRPT, or CHO cells. The amount of Tg released from FRTL-5 cells by heparin was dependent on its concentration, with linear increase up to 100 units/ml, the highest concentration used (not shown).
Inhibition experiments showed that much of the exogenous Tg released by heparin from FRTL-5 and IRPT cells was bound to megalin on the cell surface. Thus, when the cells were incubated with Tg plus either of two megalin competitors, namely RAP-GST or 1H2 (monoclonal antibody against megalin), the amount of Tg released by heparin was markedly reduced (Fig. 4B). The mean inhibition produced by RAP-GST was 84% in FRTL-5 cells and 78% in IRPT cells. The mean inhibition produced by 1H2 was 62% in FRTL-5 cells and 73% in IRPT cells. No reduction of heparin-releasable Tg was produced by co-incubation of the cells with Tg plus GST or mouse IgG, used as negative controls.
To measure the proportion of total cell surface-bound Tg that was released by heparin, experiments were performed in which the amount of Tg released by heparin was compared with the amount of Tg found in cell lysates from FRTL-5 or IRPT cells that were not subjected to heparin treatment but immediately lysed after 4 h of incubation with Tg at 4°C. Because at this temperature no internalization occurs, the amount of Tg found in the cell lysates represents only exogenous Tg bound to the cell surface, minus a negligible amount (ϳ1%) contributed by endogenous Tg, as found in FRTL-5 cells not incubated with exogenous Tg. This amount was similar to that found in the heparin wash of FRTL-5 cells incubated with medium lacking Tg.
As shown in Fig. 5A (left bars), almost all (ϳ97%) of the total cell-bound Tg was released by heparin both from FRTL-5 and IRPT cells, indicating that heparin-releasable Tg can be considered a measure of total cell surface-bound Tg. Co-incubation of cells with Tg plus the megalin competitors RAP-GST or 1H2 resulted in reduced amounts of total-cell bound Tg (Fig. 5A,  right bars), showing that much of the Tg was bound to megalin. No reduction of total cell-bound Tg was produced by co-incubation of the cells with Tg plus GST or mouse IgG, used as negative controls. The percentage of inhibition of total cell-bound Tg obtained with RAP-GST or 1H2 was similar to the percentage of inhibition of heparin-releasable Tg produced by the same competitors (Fig. 4B). Therefore, inhibition of heparin-releasable Tg by megalin competitors can be considered as a measure of megalin-bound Tg, which ranged from ϳ60 to ϳ80%, depending on the cell type and on the competitor used.
As shown in Fig. 5B, when FRTL-5 cells were incubated with increasing concentrations of unlabeled Tg, the total amount of cell surface-bound Tg released by heparin was saturable, as is typical of receptor binding. When cells were incubated with increasing concentrations of unlabeled Tg plus a constant concentration of RAP-GST (200 g/ml), inhibition of heparin-releasable bound Tg was obtained, ranging from ϳ80% at low Tg concentrations to ϳ60% at high Tg concentrations. The amount of Tg bound in the presence of RAP-GST was considered to be unrelated to megalin, and this showed some degree of saturation, suggesting the presence of Tg receptors, in addition to megalin, on FRTL-5 cells. The amount of Tg binding that was inhibited by RAP-GST (obtained by subtracting from the total amount of heparin releasable Tg the amount released in the presence of RAP-GST) was considered to be specific binding to megalin, and this was more highly saturable. The equilibrium dissociation constant (K d ) of Tg binding to FRTL-5 cells was calculated according to the method of Furchgott (33), which is based on comparison between the binding of an unlabeled protein to a receptor in the presence or in the absence of a known competitor of the receptor, in this case RAP-GST. The calculated K d values in the three experiments performed ranged from 7.8 to 14.3 nM (mean 11.2 Ϯ 3.0 nM), indicating a high affinity interaction between megalin and Tg on FRTL-5 cells.
Binding of Heparin to Tg-To help understand the mechanism by which heparin competes with megalin for Tg binding, and in view of the knowledge that megalin does not bind to heparin (26,34), solid phase binding assays were performed to determine whether Tg is a heparin-binding protein. Tg-coated wells were incubated with biotin-labeled heparin or, as a control, with biotin-labeled albumin. As shown in Fig. 6, biotinlabeled heparin bound to Tg-coated wells (Fig. 6, line 1) but not to ovalbumin-coated wells (Fig. 6, line 4). No binding of biotinlabeled albumin was observed (Fig. 6, line 3). Binding of biotinlabeled heparin-albumin to Tg was saturable and was almost completely inhibited by co-incubation with unlabeled heparin (Fig. 6, line 2).
Uptake of Tg by Cultured Cells Is Partially Mediated by Megalin-Experiments were performed by incubating cells at 37°C with unlabeled Tg, or as positive and negative controls with RAP-GST or GST, followed by heparin treatment, to remove ligands bound to the cell surface. The amounts of Tg, RAP-GST, and GST in the cell lysates were determined by ELISA as a measure of internalization. As shown in Fig. 7A, FRTL-5 and IRPT cells were found to internalize Tg and RAP-GST, whereas no uptake of GST was observed. The amount of endogenous Tg in FRTL-5 cell lysates, estimated from cells incubated in medium lacking Tg, was only 10 -20% of that found after addition of 10 g/ml of exogenous Tg. No appreciable uptake of Tg or GST was observed by CHO cells, whereas RAP-GST was internalized by these cells.
To investigate the role of megalin in Tg internalization, experiments were performed by incubating FRTL-5 and IRPT cells with Tg alone or in the presence of RAP-GST or 1H2. The results show that megalin is partially responsible for Tg endocytosis. As shown in Fig. 7B, Tg uptake was reduced by RAP-GST (mean inhibition 46.5% in FRTL-5 cells and 71.2% in IRPT cells) and by 1H2 (mean inhibition 50.3% in FRTL-5 cells and 65.1% in IRPT cells), whereas no significant inhibition was obtained with GST or normal mouse IgG, used as negative controls (Fig. 7B).
To obtain quantitative information about Tg uptake, the ratio between the amount of internalized Tg, measured in the cell lysates, to the amount of cell surface bound Tg, measured as the amount released by heparin after incubation with Tg at 37°C, was calculated. As shown in Fig. 7C, this ratio was ϳ0.5 for both FRTL-5 and IRPT cells, and it was reduced by coincubation of the cells with Tg plus RAP-GST or 1H2, confirming a role of megalin in uptake of cell surface bound Tg. DISCUSSION In the present study we show that megalin is a receptor for Tg endocytosis on cultured thyroid cells. We first demonstrated that FRTL-5 cells, a well studied established rat thyroid cell line (22,23), express megalin when cultured in standard medium containing the thyrotropic hormone (TSH). As shown by FACS analysis and immunoperoxidase staining, megalin was present on the cell surface. FRTL-5 cells are known to maintain several specific thyroid functions, most of which are TSH-dependent (22,23). Therefore, the demonstration that megalin expression by FRTL-5 cells is TSH-dependent indirectly supports a thyroid-related function of megalin. In particular, because TSH provides a signal for Tg internalization (1-3), the finding suggests a role of megalin as an endocytic receptor.
Direct evidence of binding of Tg to megalin on FRTL-5 cells as well as on IRPT cells, a rat renal proximal tubule cell line that expresses abundant megalin (30,31), was provided by experiments in which cells were incubated with 125 I-labeled Tg followed by cross-linking and incubation of the cell extracts with antibodies against megalin, which resulted in co-immunoprecipitation of 125 I-Tg. In other experiments, binding of Tg to megalin on FRTL-5 and IRPT cells was demonstrated by incubation of cells with unlabeled Tg at 4°C, followed by treatment with heparin, which released Tg into the medium, as detected by ELISA. Because almost all (ϳ97%) of the total cell-bound Tg was released by heparin, the amount of heparinreleasable Tg can be considered as a measure of cell surface-FIG. 6. Binding of heparin to Tg. 96-well microtiter plates were coated overnight at 4°C with purified rat Tg or, as a control, with OVA and incubated with biotin-labeled heparin or, as a control, with biotinlabeled albumin, followed by ALP-conjugated streptavidin. Absorbance (OD) was determined at 405 nM. 1, binding of biotin-labeled heparin to Tg-coated wells. bound Tg.
We previously showed in solid phase experiments that heparin dissociates Tg from purified megalin (13). Goldstein and colleagues (27) had shown earlier that heparin releases the LDL from its receptor on cultured fibroblasts and used this finding to measure the amount of binding of LDL to the LDL receptor. However, it is known that heparin can release molecules not only from members of the LDL receptor family but also from heparan sulfate proteoglycans (35), which are expressed in many cell types, including FRTL-5 cells (36). Nevertheless, in our experiments with FRTL-5 cells we obtained compelling evidence that most of the Tg released by heparin had been bound to megalin. Thus, when cells were incubated with Tg plus the monoclonal anti-megalin antibody 1H2 or RAP, there was roughly a 60 -80% reduction in heparin releasable Tg as well as in total cell-bound Tg, which represents the proportion of megalin-bound Tg. 1H2 is entirely specific for megalin (29), and, although RAP binds to certain other members of the LDL receptor family, notably LRP and the very low density lipoprotein receptor (16,19,(37)(38)(39), these receptors are not expressed by thyroid epithelial cells or by renal proximal tubule cells in vivo (19,40,41). Furthermore, LRP is not expressed on FRTL-5 cells as shown here, nor as previously reported on IRPT cells (30). In addition, RAP does not bind appreciably to heparan sulfate proteoglycans nor inhibit ligand binding to heparan sulfate proteoglycans (42). Thus, in dealing with FRTL-5 or IRPT cells, inhibitory effects of RAP can be considered specific for megalin. Based on this, we showed that Tg binding to megalin on FRTL-5 cells is saturable and of high affinity (K d ϭ ϳ11.2 nM). However, the finding that binding not related to megalin (noninhibitable by RAP) also showed some degree of saturation supports the existence, in addition to megalin, of other Tg receptors on thyroid cells, as suggested in other studies (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12). This possibility is also supported by the finding that in co-immunoprecipitation experiments a higher proportion of Tg was precipitated by the anti-Tg antibody from FRTL-5 cells than from IRPT cells.
Based on the finding that heparin dissociates Tg from megalin, we postulated that Tg is a heparin-binding protein, because megalin itself does not bind to heparin (26,34) and because several megalin ligands are heparin-binding proteins (26). Indeed, this prediction was confirmed by solid phase assays, which showed specific binding of Tg to heparin. This observation suggests that regions rich in positively charged amino acid residues (arginine and lysine) in the Tg molecule may contribute to its binding to megalin, as has been demonstrated for binding of certain other megalin ligands, including aprotinin and polybasic drugs (43).
Experiments in which FRTL-5 and IRPT cells were incubated with unlabeled Tg at 37°C, followed by heparin treatment to remove cell-bound Tg, showed that megalin can mediate Tg endocytosis. The detection of Tg in FRTL-5 cell lysates clearly showed that exogenous Tg had been internalized, because Tg of endogenous origin was considerably lower, as found in lysates from cells incubated in medium lacking Tg. Furthermore, the demonstration that almost all of the cell surfacebound Tg was released by heparin provides evidence that the amount of Tg found in the cell lysates represented only Tg that had been internalized. Moreover, because Tg was measured by ELISA, the amount of Tg internalized may have been underestimated, because Tg degradation during the course of the incubation should cause some loss of immunoreactivity.
Inhibition experiments provided evidence that a certain amount of Tg uptake is mediated by megalin. Thus, internalization of Tg by FRTL-5 and IRPT cells was appreciably reduced when cells were co-incubated with exogenous Tg plus RAP or 1H2. Furthermore, we obtained evidence against the possibility that reduction of Tg uptake in FRTL-5 and IRPT cells by megalin competitors resulted merely from lowering the amount of Tg bound to the cells. Thus, the ratio of internalized Tg to cell surface bound Tg was reduced by RAP and 1H2.
As noted earlier, there is evidence that micropinocytosis of Tg by thyrocytes occurs both through fluid phase uptake and receptor-mediated endocytosis (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12). However, despite extensive investigations, a specific receptor shown to have a major role in Tg uptake by thyrocytes has not previously been characterized. Consiglio et al. (4,5) showed the existence of a specific binding site for asialogalacto-Tg in thyroid membrane preparations as well as in cultured thyrocytes, and this was confirmed by others (6). The receptor was identified in porcine FIG. 7. A, uptake of Tg and RAP-GST by FRTL-5, IRPT, and CHO cells. Cells were incubated with unlabeled Tg or RAP-GST or, as a control, with GST at a concentration of 10 g/ml for 6 h at 37°C. Heparin (100 units/ml) was then added to remove ligands bound to the cell surface. Tg or RAP-GST internalized by the cells were measured in the cell lysates by ELISA. Values were normalized for the amount of protein in the cell lysates. The amount of Tg in cell lysates from FRTL-5 cells incubated in medium lacking Tg was subtracted, as background. Results are expressed as the means Ϯ S.E. obtained in three experiments. B, inhibition of Tg uptake by megalin inhibitors. FRTL-5 and IRPT cells were incubated with 10 g/ml of unlabeled Tg, alone or in the presence of RAP-GST or mouse monoclonal anti-megalin antibody (1H2). GST and normal mouse IgG (MIgG) were used as controls. Results are expressed as the mean percentages of inhibition Ϯ S.E. obtained in three experiments. C, ratio between uptake and binding of Tg. Cells were incubated at 37°C with unlabeled Tg, alone or in the presence of RAP-GST or 1H2, followed by heparin treatment. Bound Tg was measured in the heparin wash, and internalized Tg was measured in the cell lysates. Results are expressed as the means Ϯ S.E. obtained in three experiments. thyroid membranes as a 45-kDa protein, which was suggested to be involved mainly in Tg recycling (7). More recently, Lemansky and Herzog (11) performed a study using porcine thyroid follicles, designed to investigate the role of mannose-6phosphate receptors in Tg endocytosis through interaction with mannose-6 recognition markers on Tg N-linked glycans. Although they failed to show that mannose-6-phosphate receptors are responsible for Tg endocytosis, they did obtain evidence of specific low affinity binding sites on the apical surface of thyrocytes involved in Tg endocytosis. However, the responsible receptor was not identified. In another study, Giraud et al. (10) obtained evidence of selective, moderately high affinity binding of Tg to thyrocytes in cultured "inside-out" porcine follicles, as well as to cultured CHO and Madin-Darby canine kidney cells from other tissues. The receptor was not isolated, but its pH dependence, its presence on CHO and Madin-Darby canine kidney cells, and its apparent recognition of anionic charges indicate that it is not megalin. One reason why megalin has not been found in thyroid cells in previous studies may be related to the fact that in most of those studies primary cultures of thyroid cells or tissues were used. In this regard, we have observed that primary cultures of rat renal proximal tubule cells cease to express megalin after several days (30). Furthermore, we have found that primary cultures of porcine thyroid cells or follicles cease to express megalin after a few days, although megalin is normally expressed on pig thyroid cells. 2 In summary, evidence obtained in our previous (13) and present studies supports the conclusion that megalin can function as a receptor on thyrocytes capable of mediating binding and uptake of Tg. Receptors other than megalin probably contribute to this process and fluid phase pinocytosis may play a major role (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)44). In any case, the ability of megalin to bind Tg with high affinity raises interesting questions about its function in vivo. High affinity receptors serve to mediate endocytosis of ligands that are generally present in low concentrations in extracellular fluids and thus serve to compete effectively with fluid phase uptake. However, Tg in the colloid is very highly concentrated (100 -200 mg/ml), which is consistent with the notion that nonspecific fluid phase uptake is the major mechanism for Tg endocytosis and hormone release (1). The function of a high affinity receptor for Tg on thyroid cells may therefore be to regulate the extent of endocytosis only under special circumstances. High affinity receptor binding should lead to an increase of Tg endocytosis, which would be expected to result in delivery to lysosomes, with hormone release. However, it is also possible that the receptor could divert Tg from the usual endocytic pathway, as through recycling or transcytosis. Further studies are needed to define the role of megalin in thyroid hormone release.