The physiological role of transferrin (Tf) receptor 2 (TfR2), a homolog of the well-characterized TfR1, is unclear. Mutations in TfR2 result in hemochromatosis, indicating that this receptor has a unique role in iron metabolism. We report that HepG2 cells, which endogenously express TfR2, display a biphasic pattern of Tf uptake when presented with ligand concentrations up to 2 μM. The apparently nonsaturating pathway of Tf endocytosis resembles TfR1-independent Tf uptake, a process previously characterized in some liver cell types. Exogenous expression of TfR2 but not TfR1 induces a similar biphasic pattern of Tf uptake in HeLa cells, supporting a role for TfR2 in this process. Immunoelectron microscopy reveals that while Tf, TfR1, and TfR2 are localized in the plasma membrane and tubulovesicular endosomes, TfR2 expression is associated with the additional appearance of Tf in multivesicular bodies. These combined results imply that unlike TfR1, which recycles apo-Tf back to the cell surface after the release of iron, TfR2 promotes the intracellular deposition of ligand. Tf delivered by TfR2 does not appear to be degraded, which suggests that its delivery to this organelle may be functionally relevant to the storage of iron in overloaded states.
- iron transport
- HepG2 cells
transferrin (tf) receptor 1 (TfR1) is a well-characterized membrane protein that facilitates Tf-mediated iron uptake (for review, see Ref. 31). A homolog of TfR1, transferrin receptor 2 (TfR2), has been identified as a 105-kDa glycoprotein of unknown function (18, 19). Exogenous expression of human TfR2 in Chinese hamster ovary TfR variant clone b (CHO-TRVb) cells increases Tf uptake, suggesting that like TfR1, TfR2 participates in Tf-mediated iron uptake (19). These cells also have increased cell surface Tf binding, further supporting a direct role for TfR2 in Tf endocytosis (19). However, TfR2 does not appear to compensate for TfR1 function, because TfR1-knockout mice display embryonic lethality (25). Furthermore, TfR1+/− mice have lower hepatic iron levels relative to wild-type mice (25), in contrast to the hepatic iron overload associated with human or murine mutations in TfR2 (4, 9). Although its exact role remains unclear, TfR2's physiological importance is well documented by clinical reports of patients with hemochromatosis who had different mutations in this receptor (4, 14, 26, 32). Thus comparative studies of TfR2 with the well-characterized TfR1 are important in identifying functional distinctions between the two receptors.
TfR2 expression is distinct from classic TfR1 expression. It is highly expressed in the liver, with more modest mRNA levels found in some other tissues (19). The expression of TfR2 in murine liver increases during development, while TfR1 decreases (17). In contrast, TfR2 levels in the spleen are constant throughout development, while TfR1 levels increase (17). The murine TfR2 promoter can be regulated by the erythroid/liver-related transcription factors GATA, CCAAT/enhancer binding protein (C/EBP), erythroid Kruppel-like factor (EKLF), and friend of GATA-1 (FOG-1), implying tissue-specific regulation (17). The differential tissue expression of TfR1 and TfR2 during murine development and the high level of TfR2 expression in the liver suggest a tissue-specific function for TfR2 that may be distinct from TfR1 function.
Although the expression patterns of these two proteins appear to be quite different, Tf-related activities studied to date seem to be quite similar. Both bind Tf and facilitate Tf-mediated iron uptake (18, 19). However, a role for TfR2 in a process that has been referred to as “TfR1-independent Tf uptake” (15, 24, 27, 36, 37) has not been fully examined. This second Tf uptake mechanism is characterized by a pattern of biphasic internalization kinetics observed in some cell types, including liver HuH-7 cells and rat hepatocytes. While steady-state internalization of Tf typically saturates at low concentrations (<0.5 μM) (see, e.g., Refs. 1, 5), this atypical pathway displays a linear phase of uptake at ligand concentrations >0.5 μM, which appears to be independent of TfR1, because it continues when TfR1 expression is suppressed by antisense strategies or TfR1 activity is blocked by anti-TfR1 antibodies (15, 24, 36, 37). The TfR1-independent pathway in HuH-7 cells displays a slower rate of Tf recycling, an increased level of Tf degradation and a potentially unique mechanism of iron acquisition from Tf (15). Because the TfR1-independent pathway has been observed primarily in hepatic cells that express TfR2 (24, 38, 42), we considered the possibility that this receptor might be involved. Our results show that TfR2 functions in the Tf endocytic cycle are quite different from those of TfR1. Gain-of-function experiments revealed that exogenous expression of TfR2 resulted in a biphasic pattern of Tf uptake, directly demonstrating its role in a pathway previously described only in biochemical studies. Furthermore, the subcellular distribution of internalized ligand indicated that this receptor delivered Tf to the late endocytic pathway, where it accumulated in multivesicular bodies (MVB) devoid of receptors and deficient in lysosome-associated membrane protein 1 (LAMP1). Because 125I-labeled Tf did not appear to be degraded in this pathway, these coupled observations support a model in which the apparently nonsaturable linear phase of TfR1-independent Tf uptake can be explained by intracellular accrual of the ligand. The liver serves as a major depot of iron in the body, and we hypothesize that the unique mechanism of Tf import by TfR2 plays a critical role in the regulation of iron release and/or storage by this tissue.
MATERIALS AND METHODS
Cellular binding and uptake of 125I-Tf.
Radioiodinated Tf was prepared as described previously (39), except that labeled protein was separated from free 125I by using Micro Bio Spin-6 columns (Bio-Rad, Hercules, CA). HepG2 cells were grown in DMEM containing 50 U/ml penicillin, 50 μg/ml streptomycin, and 10% FBS. Before experiments, subconfluent cells in six-well plates were washed twice and subsequently incubated at 37°C for at least 15 min in serum-free medium containing 20 mM HEPES. Cells were then incubated with serum-free DMEM containing 20 mM HEPES, pH 7.4, and 2 mg/ml ovalbumin or bovine serum albumin (BSA), and concentrations of 125I-labeled Tf as indicated in the figure legends. After incubation at 37°C for 1 h, cells were put on ice and washed four times in ice-cold wash buffer (in mM: 150 NaCl, 1 CaCl2, 5 KCl, and 20 HEPES, pH 7.4). To measure surface binding, cells were kept on ice in the same buffer containing at least 2 mg/ml ovalbumin, and a duplicate set of six-well plates were prepared containing 100 μM unlabeled Tf. After addition of 125I-labeled Tf, binding assays were performed for 70 min at 4°C, and cells were subsequently washed as described above. At the end of either assay, cells were lysed in solubilization buffer (0.1% Triton X-100, 0.1–0.4% NaOH), 125I-Tf was measured using a gamma counter, and protein levels were assessed using the Bradford assay (2) to determine 125I-Tf concentration in picomoles per milligram of protein. To calculate specific cell surface binding at 4°C, the amount of radioactivity associated with cells in the presence of unlabeled ligand was subtracted from values obtained in the absence of cold Tf.
Transfection of HeLa cells.
To transiently overexpress TfR1 in HeLa cells, TfR1 cDNA was directionally subcloned into pcDNA 3.1+hygro (Invitrogen, Carlsbad, CA) using EcoRV/XbaI restriction sites (pcDNA3.1-TfR). TfR2 was exogenously expressed in HeLa cells as a FLAG-tagged fusion protein using a vector (pcDNA3-FLAG-TfR2) kindly provided by Drs. H. Kawabata and H. P. Koeffler (Division of Hematology/Oncology, Department of Medicine, University of California at Los Angeles School of Medicine, Los Angeles, CA). Briefly, subconfluent HeLa cells were washed twice in serum-free medium. Plasmids—control (pcDNA3.1), pcDNA3-FLAG-TfR2, or pcDNA3.1-TfR1—were combined with Lipofectamine (GIBCO-BRL, Grand Island, NY) in serum-free medium, and cells were subsequently transfected according to the manufacturer's instructions. Cells were trypsinized and replated onto six-well plates. 125I-Tf uptake experiments were performed 1–2 days posttransfection as described above. Western blot analysis confirmed the absence or presence of exogenous FLAG-TfR2 expression or TfR1 overexpression (data not shown). To study the degradation of internalized ligand, cells transfected to express TfR1 or TfR2 were incubated with 1.5 μM 125I-Tf at 37°C for up to 6 h. Aliquots (2 μl) of the cell-conditioned medium were removed and mixed with 1 ml of 20% trichloroacetic acid (TCA)-2% casein hydrolysate for 1 h on ice. After microcentrifugation at 14,000 rpm for 5 min at 4°C, TCA-soluble radioactivity was determined using a gamma counter.
After the incubations of transfected HeLa cells were performed as described in the figure legends, cells were washed four times with (in mM) 150 NaCl, 1 CaCl2, 5 KCl, and 20 HEPES, pH 7.4, and then fixed with paraformaldehyde (4% wt/vol) and glutaraldehye (0.1% wt/vol). HepG2 cells were washed twice and incubated for 1.5 h at 37°C to deplete cells of Tf in serum-free medium containing 20 mM HEPES, pH 7.4. HepG2 cells were then incubated for 1 h at 37°C with 2 μM holotransferrin in serum-free medium containing 20 mM HEPES, pH 7.4, and 2 mg/ml BSA; washed four times in PBS; and fixed as described above. Processing for cryosectioning and immunolabeling was performed as described by Griffiths (13). Labeling for TfR2 was performed using a 1:200 dilution of M2-monoclonal anti-FLAG antisera (Sigma Chemical, St. Louis, MO) while TfR1 was detected with a 1:50 dilution of mouse anti-TfR1 (Zymed Laboratories, South San Francisco, CA). Tf was detected using 1:50 (HeLa) and 1:200 (HepG2) dilutions of rabbit anti-Tf (Rockland Immunochemicals, Gilbertsville, PA). A 1:60 dilution for Tf and a 1:100 dilution for LAMP1 (mouse anti-human LAMP1; Pharmingen, San Diego, CA) were used in colocalization studies of HeLa cells transfected with TfR2. After incubation with appropriate secondary antibody if necessary, 5 or 10 nm of protein A-gold (Dr. Jan Slot, Department of Cell Biology, University of Utrecht Medical School, Utrecht, the Netherlands) was used to detect immunoreactivity. Microscopy was performed using a Jeol 1200EX transmission electron microscope (Jeol, Peabody, MA) at 80 kV with primary magnification of ×15–30,000.
Tf uptake by HepG2 cells is biphasic.
HuH-7 cells and rat hepatocytes are known to express both TfR1 and TfR2 (24, 38, 42), and a biphasic mechanism for Tf uptake has been documented at higher ligand concentrations for these liver-derived cells (15, 24, 36, 37). In HuH-7 cells, this mechanism correlates with a TfR1-independent pathway of endocytosis and iron delivery (15, 24, 37); therefore, we speculated that TfR2 might be involved. However, previous investigations of human hepatoma HepG2 cells, which also express both TfR1 and TfR2 (3, 7, 19, 38), did not reveal a biphasic pattern for ligand uptake (6, 34, 35). To our knowledge, these studies did not measure binding and uptake at higher ligand concentrations at which the linear phase of internalization associated with TfR1-independent uptake might have been observed, and thus we first sought to characterize Tf uptake by HepG2 cells more fully. To characterize ligand interactions, surface binding of 125I-Tf was measured at 4°C for 1 h. Under these conditions, ligand binding was saturable with respect to Tf concentrations up to 2 μM (Fig. 1A). To examine whether uptake of ligand was saturable, similar experiments were performed with cell incubations at 37°C. The association of 125I-Tf with HepG2 cells (binding and uptake) was curvilinear up to 0.2–0.5 μM but then increased linearly at concentrations up to 2 μM (Fig. 1B). The latter was not due to a kinetic effect of ligand interactions with the receptors, because control experiments confirmed that uptake in the presence of 1.5–3 μM 125I-Tf achieved steady-state levels within 1 h, the time period used in our study (data not shown). The observed biphasic pattern of ligand binding and uptake closely resembles the characteristics of HuH-7 cells and rat hepatocytes (15, 24, 36, 37), which supports the hypothesis that TfR2 may be involved. Because serum Tf concentrations range from 20 to 90 μM (40), it is possible that the second “linear” phase of uptake may saturate at even higher and more physiological concentrations of the plasma protein.
Tf uptake by HeLa cells exogenously expressing TfR2 is biphasic.
To compare Tf uptake in cells known to display saturable steady-state ligand internalization kinetics, similar assays were performed using the HeLa cell line. As previously reported (23, 29), uptake of Tf at 37°C becomes saturated at ∼0.5 μM ligand, with no further increase at higher Tf concentrations that would reflect biphasic uptake as observed for HepG2 cells (Fig. 2A). A key difference between HeLa and HepG2 cells is that while the latter express TfR2 (3, 7, 19, 38), expression of TfR2 cannot be detected in HeLa cells by performing Northern or Western blot analysis (Robalinho-Teixeira R, Robb A, and Wessling-Resnick M, unpublished observations). Therefore, to directly test whether TfR2 expression promotes the observed linear increase in uptake at higher Tf concentrations, HeLa cells were transiently transfected to express TfR2 and the amount of cell-associated Tf was determined as a function of ligand concentration as described above. Figure 2B demonstrates that uptake was curvilinear at lower Tf concentrations (<0.5 μM) and became linear at concentrations up to 2 μM, a pattern similar to that observed for HepG2 cells (Fig. 1B). Control experiments confirmed that HeLa cells transfected with empty vector or with a vector to overexpress TfR1 displayed the simple curvilinear pattern of nontransfected cells (data not shown; see Fig. 2A). Combined, these findings strongly support a functional role for TfR2 in the biphasic mechanism of Tf uptake.
TfR1 and TfR2 colocalize to plasma membrane and endosomes.
The differences in steady-state Tf uptake observed in TfR2-expressing cells prompted us to compare the subcellular distribution of TfR1 and TfR2 by performing immunoelectron microscopy of transfected HeLa cells. Double-labeling experiments (Fig. 3) demonstrated that these receptors colocalize and are found predominantly on the plasma membrane and tubulovesicular endosomes. In agreement with these data, previous immunofluorescence studies of endogenous TfR2 distribution in K562 cells also showed colocalization with TfR1 (38). The pattern of colocalization between TfR1 and TfR2 indicates that both transit the same compartments of the early endocytic pathway.
Tf localizes to MVB in TfR2- but not TfR1-expressing cells.
To compare the distribution of Tf internalized by the receptors, HeLa cells transfected with either TfR1 or TfR2 were incubated with 2 μM ligand for 1 h, then fixed for immunostaining with anti-Tf antibodies (Fig. 4). When only TfR1 was expressed, the ligand was found exclusively localized to plasma membrane and tubulovesicular endosomes (Fig. 4, left). However, in cells expressing TfR2, Tf was also observed in MVB (Fig. 4, right). This information is difficult to quantify because not every MVB contained Tf immunoreactivity; however, in 49 clear profiles of MVB that were counted, 190 protein A-gold particles were observed. In contrast, Tf labeling of MVB in the absence of TfR2 expression was very rare (Fig. 4, lower left). In 21 MVB that were counted in HeLa cells transfected to overexpress TfR1, only 14 protein A-gold particles were observed. Because receptor immunoreactivity was not detected in MVB, these combined observations suggest that Tf must be released from TfR2 as it transits the early endosomal system. The intracellular accrual of ligand that is implied by the results of these experiments is consistent with the linear phase of Tf uptake observed in the TfR2-expressing HeLa cells (Fig. 2B).
The distribution of Tf was also studied in HepG2 cells by immunoelectron microscopy. In addition to plasma membrane and endosomal compartments, Tf was detected in MVB in these cells as well (Fig. 5A). While these observations support the idea that Tf is delivered to MVB via TfR2-mediated endocytosis, a caveat is that HepG2 cells synthesize and secrete Tf (21). Indeed, immunoreactivity was detected in elements of the secretory pathway, including the endoplasmic reticulum (ER) and the Golgi apparatus (Fig. 5B). Thus the possibility exists that newly synthesized ligand might be sorted to this compartment directly from the biosynthetic pathway. However, evidence to support this idea is lacking, and the presence of the ligand in HepG2 cell MVB is entirely consistent with its localization after internalization by TfR2-expressing HeLa cells, which do not express Tf (8, 28).
Delivery of Tf to MVB is not associated with its degradation.
Proteins delivered to MVB are often destined for lysosomal degradation. To examine whether Tf staining could be detected in lysosomal (LAMP1 positive) compartments, double-labeling experiments were performed (Fig. 6A). The results indicated that MVB containing Tf are markedly deficient in LAMP1. However, because proteolysis of internalized Tf would result in loss of immunoreactivity, the lack of labeling in LAMP1-positive lysosomes did not rigorously address whether the ligand was ultimately degraded. To directly follow the fate of internalized ligand, HeLa cells transfected to express either TfR1 or TfR2 were incubated with 1.5 μM 125I-Tf, and the release of TCA-soluble radioactivity was subsequently followed. No significant degradation of 125I-Tf was observed up to 6 h postinternalization in both TfR1- and TfR2-expressing HeLa cells (Fig. 6B). On the basis of these combined data, it seems likely that Tf delivered to MVB via TfR2 does not traffic to late endosomes/lysosomes, avoiding proteolytic degradation.
While steady-state internalization of Tf saturates at low concentrations (<0.5 μM) in most cell types, a biphasic mechanism for Tf uptake has been documented at higher ligand concentrations for liver-derived HuH-7 cells and rat hepatocytes (15, 24, 36, 37), and also, as shown by the present study, human hepatoma HepG2 cells. A common feature of the liver-derived cell lines that display this atypical uptake mechanism is that they all express TfR2 (24, 38, 42). It is interesting to note that human erythroleukemia K562 cells also express TfR2 (19), and some studies have shown a similar pattern of biphasic kinetics (33). In HuH-7 cells, the mechanism of biphasic Tf uptake has been shown to correlate with a TfR1-independent pathway of endocytosis and iron delivery (15, 24, 37), although details about the molecular basis for these effects were unknown (36, 37). The fact that steady-state internalization of Tf by HeLa cells becomes biphasic when TfR2 but not TfR1 is exogenously expressed directly implies a role for TfR2 in this mechanism. Thus our data help to shed significant light on the molecular basis for the “TfR1-independent” pathway that had been described previously only in biochemical studies.
Consistent with functional differences in the kinetics of ligand uptake, the intracellular movement of Tf taken up by TfR1 and TfR2 is also strikingly different. Our data reveal that intracellular trafficking of Tf in HeLa cells is altered by exogenous expression of TfR2 such that the ligand is directed into MVB. The fact that TfR2 does not appear in MVB suggests that Tf is released during the internalization pathway mediated by this receptor. Similar to its association with TfR1, apo-Tf is reported to strongly bind TfR2 at pH <6.5, but published studies by Kawabata et al. (19) suggest that holotransferrin dissociates from TfR2 at pH values between 6.5 and 7.0, while the iron-loaded ligand remains bound to TfR1 under these conditions. Thus it seems probable that holotransferrin could dissociate from TfR2 in early endocytic compartments to be sorted for delivery to MVB. Delivery of Tf to MVB provides an explanation for the linear phase of the biphasic uptake pattern due to the intracellular accrual of the ligand.
MVB are late endosomal compartments typically associated with the progression of membrane traffic toward the lysosome, resulting in constitutive degradation of the organelle contents (16). Perhaps best characterized for their role in EGF-EGF receptor membrane traffic, MVB lack TfR and often remain negative for LAMP1, a lysosome-associated marker, while they do contain EGF-EGF receptor complexes. Kinetic studies have shown that acid-soluble 125I release due to degradation of 125I-EGF can be detected and measured in counts per minute (cpm) as early as 30 min after internalization and transit through MVB (10). In contrast, our studies show that 125I-Tf endocytosed to MVB via TfR2 is not degraded for at least 6 h. This observation prompts the speculation that Tf is delivered to MVB as a result of its association with a saturable binding component that directs sorting of the ligand to this compartment but prevents further trafficking to lysosomes. Because of experimental limitations, we were able to study 125I-labeled Tf uptake only at concentrations up to 2 μM; therefore, it is possible that saturation of such a binding component would be achieved at even higher concentrations. In HuH7 cells, the TfR1-independent pathway for Tf uptake was recently reported to saturate to ∼10 μM concentration (24). In preliminary studies, we have found that recycling kinetics of 125I-labeled Tf are not altered in TfR2-expressing HeLa cells despite the accrual of ligand (Robb A and Wessling-Resnick M, unpublished observations), and this result is also consistent with saturable association of internalized ligand with some intracellular component. Besides TfR1 and TfR2, only the cubilin-megalin complex is known to bind Tf (22), so it is clear that further work is necessary to explore the molecular basis for the localization of Tf to MVB and its resistance to lysosomal proteolysis. Because TfR2 is important for iron homeostasis, we hypothesize that the differential localization of Tf in this organelle might facilitate downstream effects, perhaps via interactions with the putative MVB-targeted receptor. Interestingly, in some cell types, MVB function in regulated exocytosis such that their contents are released (20); they may also function in the biogenesis of secretory organelles that subsequently undergo regulated secretion to export their cargo (41). Tf is known to be secreted in bile, so one hypothesis is that its appearance in hepatocyte MVB might be coupled to transcytotic clearance from circulation.
Regardless of the precise mechanism, downstream consequences of the differential intracellular targeting of Tf to MVB most likely involve modulation of serum Tf concentration and/or decreased Tf stabilization or recycling. Any of these factors would contribute to the pathology of iron overload observed in patients with hemochromatosis who have mutations in TfR2 (4, 14, 26, 32). Clinical information about iron loading associated with TFR2 mutations indicates that the metal accumulates in periportal hepatocytes (30). On the basis of these observations and the pattern of Tf traffic reported in the present study, it seems rather unlikely that hepatic iron import is the major role of TfR2. In this respect, it is also interesting to note that hepatic expression of TfR2 and a newly identified iron-regulatory factor, hepcidin, are significantly correlated in a manner that is independent of body iron status (12). Hepcidin appears to be a negative regulator of iron release (11). It is possible that the delivery of Tf to MVB is associated with the regulation of hepcidin expression such that disruption of this trafficking pathway because of loss of TfR2 function leads to alterations in tissue iron storage. It is clear that further characterization of the Tf-TfR2 trafficking pathway is necessary to elucidate key elements involved in the regulation of iron homeostasis.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants P01 DK-55495 (Project 3) and R01 DK-56160 (to M. Wessling-Resnick).
We gratefully acknowledge Ted Galanopoulos for constructing pcDNA3.1-TfR1 and Racquel Robalinho-Teixeira for performing Northern blot analysis of TfR2 mRNA expression in HeLa cells.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society