Transferrin-dependent uptake and distribution of iron in osteoclast-like cells

Abstract

Iron is an essential micro component and is involved in numerous critical cellular processes and in energy production. While its roles in oxygen transport and in oxidative phosphorylation are well documented, it remains to be elucidated, whether iron modulates specific cellular processes in different organs. Iron deficiency has been found to lead to a decrease in the capacity of osteoclasts to dissolve amorphous calcium phosphate. Furthermore, levels of iron in the cellular environment led to significant changes in the levels of transcripts encoding iron transport proteins. Within the present study, the uptake of iron by osteoclasts and the kinetics of intracellular transport were analyzed. For this purpose, M-CSF (Macrophage-Colony Stimulating Factor) dependent non-adherent osteoclast progenitor cells were differentiated to osteoclasts in media containing M-CSF and RANKL (Receptor Activator of NF-κB Ligand). For the upregulation of iron transport capacity, media were supplemented with Deferoxamine, an iron complexor, rendering the cultures virtually iron-free. To analyze iron uptake by osteoclast like cells, holo-transferrin, loaded with ⁵⁵Fe was added to the cells and iron uptake was quantitated in whole cell lysates and in fractionated cells. The data demonstrates that Deferoxamine-treated osteoclasts absorb higher quantities of iron as compared to untreated control cells. By density gradient centrifugation, cell associated iron can be separated into two major pools. Pool I represents non-transferrin associated iron in cytoplasmic fractions, while pool II contains transferrin/ transferrin receptor associated iron. Within 4 h of incubation in iron-deficient medium, pool I disappears, as does transferrin. Pool II iron and transferrin receptor, however, remain detectable. Furthermore, iron peaks did not associate with ferritin nor with mitochondria, demonstrating that these two mechanisms of iron storage did not become activated during the course of the study. The data thus demonstrates that iron uptake by osteoclasts can be modulated by exogenous iron and that cell associated iron forms either a labile iron pool of free iron that is lost within a short period of time or a vesicular pool of non-transferrin bound iron that remains stable over the experimental period. No further trafficking of iron into ferritin particles or mitochondria was detected.

Highlights

• •Exogenous iron levels modulate levels of transferrin receptors in osteoclasts.
• •Iron-starved osteoclasts are more efficient in taking up iron than saturated osteoclasts.
• •Iron taken up by osteoclasts through a transferrin-mediated mechanism remains in the cells.
• •No iron trafficking to mitochondria or ferritin particles was observed in osteoclasts.

1. Introduction

Iron is an essential microelement, playing a role in numerous vital functions. Iron is present in heme groups and in iron‑sulfur (Fe—S) cluster-containing proteins (Muckenthaler et al., 2017; Altamura et al., 2020; Koleini et al., 2021). As a cofactor, iron is required in mitochondrial respiration and DNA synthesis (Johnson et al., 2005). Despite its essential roles in these processes, free iron is toxic and an equilibrium is maintained between protein bound ferric (Fe³⁺) and free ferrous (Fe²⁺) forms (Goswami et al., 2002) During the Fenton reaction free ionized iron will give rise to reactive oxygen species, inducing oxidative stress and subsequently causing deleterious effects on proteins, lipids and DNA which will ultimately lead to cellular malfunction (Toyokuni, 2002). Therefore, due to the beneficial and deleterious properties of the metal, iron homeostasis is under firm control to prevent deficiency and anemia or overload and organ damage. Bone is one of the organs susceptible to changes in iron homeostasis, and patients with hereditary hemochromatosis, which is characterized by iron overload and ferroptosis (Ru et al., 2023), develop premature osteoporosis (Weinberg, 2008; Tsay et al., 2010; Ledesma-Colunga et al., 2021). As of now, however, the mechanisms causing these effects on the skeletal system are not yet identified, direct vs. indirect mechanisms and involvement of energy metabolism or inflammatory reactions are possibilities to be elucidated.

Bone is a metabolically active tissue that is continuously turned over to allow for adaptation to changing mechanical needs, for repair and for mineral homeostasis. Imbalance between bone formation and resorption, as is the case in malfunctioning iron homeostasis, leads to changes in bone mass and microarchitecture and eventually to a failure of the mechanical stability of skeletal elements (Parfitt, 1982; Eriksen, 1986). The correct development and activation of osteoclast (OC) lineage cells are crucial for adequate bone homeostasis. Two growth factors, macrophage colony-stimulating factor (M-CSF) and receptor activator of NF-κB ligand (RANKL) are essential for osteoclastogenesis, their absence causing an osteopetrotic phenotype due to a lack of OC differentiation (Felix et al., 1990; Kong et al., 1999a). There are numerous further components of the microenvironment, i.e. growth factors, cytokines, and trace elements, among them iron, however, contributing to the regulation of OC development and bone resorption (Kong et al., 1999a; Kong et al., 1999b; Balga et al., 2006).

Cellular iron concentrations are regulated by transferrin (TF) – transferrin receptor (TFRC) mediated iron uptake and by ferroportin (FPN, Slc40A1) mediated iron release (Donovan et al., 2000; Nemeth and Ganz, 2021). Macrophages, which take up large amounts of iron by phagocytosing dying erythrocytes express high levels of FPN1 to allow for efficient extrusion of iron (Lu et al., 2020). Osteoclasts, as another differentiated cell type of the monocyte/ macrophage lineage, express low levels of FPN1, while expressing high levels of TFRC (Ishii et al., 2009; Xie et al., 2016). Therefore, it can be suggested that during the development of OC from monocytic progenitor cells, they change from an “iron export” to an “iron uptake” phenotype. During resorption, osteoclasts require high amounts of energy for the maintenance of an acidic pH in the resorption lacuna to dissolve the bone's hydroxyapatite. Protons are transported into the resorption lacuna against a concentration gradient of ∼1000, with an intracellular pH of ∼7.4 and a lacunar pH of ∼4.5. The transport is accomplished by a vacuolar type H⁺-ATPase (V-ATPase) proton pump, requiring the continuous availability of sufficient ATP as the source of energy for H⁺-transport (Baron et al., 1986). To ensure the necessary levels of the trace element iron, OC lineage cells increase the expression of the TFRC during differentiation and react to iron deficiency with an efficient upregulation of TFRC, TF binding and iron uptake. Expression of FPN1, on the other hand, decreases during the development of OC since regulated iron release may be of lesser relevance for the cells (Xie et al., 2016).

Within the present study, iron uptake by and intracellular iron trafficking within in vitro generated osteoclast-lke cells (OCL) was analyzed. We found TF mediated iron uptake to be modulated by extracellular iron levels. Binding of TF to the surface of the OCL, however, seemed not to depend on the loading of TF with iron and intracellularly, iron and TF fractions that are not associated with TFRC were detected. After iron uptake, all cellular TF was removed from the cells within 4 h. The data demonstrates the dependence of iron uptake in vitro by OCL on TF, but the independence of TF binding to the cell surface on TF iron load.

2. Material & methods

2.1. Animals

This study was approved by the Local Committee for Animal Experimentation (Bern Committee for the Control of Animal Experimentation, Bern, Switzerland, permit number BE19/19 to WH). According to local regulations, animals were kept in groups in the specific pathogen free (SPF) facility of the Medical Faculty of the University of Bern following FELASA regulations.

2.2. Cell cultures

Osteoclast-like cells were grown in vitro from M-CSF-dependent osteoclast progenitor cells (OPC) as described previously (Xie et al., 2016; Ruef et al., 2017). Briefly, bones (femora, tibiae, humeri) from 8 to 10 week old female mice were dissected and bone marrow cells were flushed out with Hanks' Balanced Salt Solution (HBSS), supplemented with 100 U/ml penicillin and 100 U/ml streptomycin. Cells were cultured overnight in αMEM supplemented with 10% heat inactivated Fetal Bovine Serum (FBS) and 30 ng/ml M-CSF (Chiron, Emeryville, CA). After 24 h, non-adherent cells were sedimented by centrifugation for 7 min / 250 xg at 4 °C, resuspended in αMEM/ 10% FBS, counted and plated at a density of 4 × 10⁵ cells/ml. Osteoclast-like cells were grown in media supplemented with M-CSF (30 ng/ml) and RANKL (20 ng/ml, PreProTech, LubioScience, Switzerland) for 5 days with a medium change after 3 days. For the experiments, OCL were released from the culture dishes by incubating the cells for 15 min at 37 °C with PBS/ 0.02% EDTA. Cells were pelleted by centrifugation for 5 min/ 200 xg at 4 °C and diluted as necessary. If it was necessary to render the cultures iron free, the media were supplemented with 10 μM Deferoxamine (DFO, Calbiochem/ Sigma-Aldrich, Buchs, CH), a chelator of ferric (Fe³⁺) iron, for days 4 and 5.

2.3. Iron uptake

To assess the capacity of the cells for iron uptake, OCL were incubated for up to 6 h with holo-transferrin (holoTF) that was in vitro loaded with ⁵⁵Fe, with apo-transferrin (apoTF; Sigma-Aldrich, Buchs, CH) to assess binding of apoTF to the cells, and with ⁵⁵Fe alone to control for non-specific binding und uptake. To load apoTF with ⁵⁵Fe, a master mix of Buffered Salt Solution (BSS; 135 mM NaCl/50 mM KCl/10 mM CaCl₂/10 mM MgCl₂/180 mM HEPES)/ 10 mM NaHCO₃/ 10 μM FeCl₃ (⁵⁵FeCl₃ spec. Activity 20′000 Bq/ 100 μl; Perkin Elmer, Schwerzenbach CH)/ 5 μM apoTF was incubated overnight at room temperature. For iron uptake, osteoclasts were seeded into wells of 48-well plates and incubated for the required time in 200 μl αMEM/1% Pen/Strep/0.4% BSA and 200 μl holoTF (or, if required, apoTF or no TF). After incubation, the cells were washed 3 x with BSS, lysed with 1% TX-100 in water and stored overnight at −20 °C. Iron was counted in a TRI-CARB Liquid Scintillation Analyzer 2200CA (Packard, Ramsey, MN) after mixing 180 μl of the cell lysate with 4 ml of Ultima Gold Scintillator (supplier).

2.4. RNA isolation

Total RNA was isolated using the NucleoSpin® RNA Plus Kit (Macherey-Nagel AG, Oensingen, CH) following the instructions of the manufacturer. RNA quantity and quality were determined using a NanoDrop™ Spectrophotometer (Thermo Fisher, Reinach, CH). 500 ng of RNA were reverse transcribed. Quantitative RT-PCR was performed on an ABI7500 (Life Technologies, Dübendorf, CH), using the respective Assays on Demand (AoD, Thermo Fisher/ Life Technologies, Zug, CH). The AoD used in this study are listed in Table 1.

Table 1.

Assays on Demand (AoD) for quantitative RT-PCR.

Target transcripts	AoD number
F4/80	Mm00802529_m1
TRAP	Mm00475698_m1
TFRC	Mm00441941_m1
DMT1 (Slc11a2)	Mm00435363_m1
FPN (Slc40a1)	Mm00489837_m1
GUSB	Mm01197698_m1

2.5. Cell fractionation

Cultures of OCL were grown in 20 cm² plates. After cooling on ice and collection with a rubber policeman, the cells were swollen for 10 min at 4 °C in hypotonic ReSuspension Buffer (RSB) (10 mM NaCl, 10 mM Tris/HCl, 15 mM MgCl₂, ph 7.4), supplemented with a protease inhibitor cocktail (Sigma-Aldrich, Buchs, CH). The cells were transferred into an icecold Dounce Homogenizer (clearance 13–64 μm) and disrupted manually with 40 strokes. Thereafter, 1.5 volumes of icecold mannitol/sucrose (MS) buffer (525 mM mannitol, 175 mM sucrose, 12.5 mM Tris/HCl, 2.5 mM EDTA, ph 7.5) was added and the samples were transferred into 2 ml Eppendorf tubes and centrifuged for 10 min at 4 °C and 1′000 x g. For density based cell fractionation, 1 ml of the sample was loaded onto a 0–60% sucrose gradient and centrifuged for 18 h at 4 °C and 200′000 xg (Sorvall Centrifuge, Rotor Thermo Scientific TH-641). After centrifugation, fractions of 0.75 ml were removed manually from the top, precipitated with trichloroacetic acid and redissolved in Laemmli Sample Buffer (BioRad, Cressier, CH). Cell preparations containing ⁵⁵Fe were fractionated as described, the fractions were mixed with 4 ml Ultima Gold Scintillator and counted in a TRI-CARB Liquid Scintillation Analyzer 2200CA.

2.6. Western blot

Fifteen μl of the prepared samples were applied onto pre-fabricated 4–15% gradient polyacrylamide gels (BioRad, Cressier, CH). Gels were run for 65 min at RT and 150 V and subsequently, proteins were transferred onto Nitrocellulose membranes (BioRad, Cressier, CH). The membranes were saturated with 5% milk powder in TBS/0.1% Tween for 3 h at RT. After washing with TBS/0.1% Tween, primary antibodies were diluted in TBS/0.1% Tween and incubated overnight at 4 °C. For a list of antibodies and respective dilutions see Table 2. First antibodies were used to detect ßActin as a loading standard, TFRC, TF, FT (ferritin light chain) to assess storage of intracellular iron, CD51 (Integrin αV, a membrane protein to detect membrane vesicles, EEA1 (Early Endosome Antigen 1), a marker for early endosomes and HFE (hemochromatosis protein), a regulator of TF mediated cellular iron uptake. After incubation with the first antibody, membranes were washed with TBS/Tween 0.2% and incubated with the secondary antibody, each at a dilution of 1:10′000 in TBS/Tween 0.1%. The membranes were measured using an Odyssey® Sa Infrared Imaging System (LI-COR Biosciences, Lincoln, NE, USA) and quantitated with the Image Studio 5.2 software (LI-COR Biosciences, Lincoln, NE, USA).

Table 2.

Antibodies for western blots. All antibodies were purchased from Invitrogen, Life Technologies, Zug, CH).

Protein	Antibody	Dilution
Ferritin (FN)	rabbit mAb (IgG)	1:1′000
Transferrin (TF)	rabbit pAb (IgG)	1:1′000
CD51	rabbit pAB (IgG)	0.388888889
Early Endosome Antigen 1 (EAA1)	rabbit mAb (IgG)	1:1′000
Transferin Receptor (TFRC)	mouse mAb (IgG1)	1:2′000
ß-Actin	rabbit mAb (IgG)	1:5′000
Golgin	mouse mAB (IgG1)	1:1′000
Goat anti-Rabbit IgG IRDye800CW		1:10′000
Goat anti-Mouse IgG IRDye680LT		1:10′000

2.7. Statistical analysis

Differences in transcript levels and cellular iron uptake were evaluated by one-way ANOVA, with Bonferroni's post tests using GraphPad Prism Version 10 for Windows (GraphPad Software, San Diego, CA, USA). P-values >0.05 were considered significant.

3. Results

3.1. Exogenous iron levels modulate levels of transcripts encoding iron transport molecules

Mature OCL were grown from OPC for five days with M-CSF/RANKL. Three days into the culture, the conditions were rendered iron-deficient with DFO (10 μM). To assess the modulation of gene expression by DFO, total RNA was prepared at day 5 and transcripts encoding F4/80, a murine pan-macrophage marker and TRAP (Tartrate-Resistant Acid Phosphatase), a marker for osteoclasts, as well as TFRC (transferrin receptor 1), DMT1 (Divalent Metal Transporter 1, Slc11A2), required to release ferrous iron from lysosomal vesicles into the cytoplasm, and Ferroportin (FPN, Slc40A1), the only known transport protein for the release of cytoplasmic ferrous iron from the cells, were determined (Fig. 1). F4/80 and TRAP mRNA levels were not affected by the lack of iron in the culture media. Iron deficiency caused an increase in the levels of transcripts encoding TFRC (6820 ± 190 vs. 590 ± 67; p < 0.001) and DMT1/Slc11A2 (197 ± 14 vs. 129 ± 7; p > 0.1) and a decrease in the levels of transcripts encoding FPN/Slc40A1 (126 ± 11 vs. 154 ± 11; p > 0.05) in OCL. All transcripts were normalized for the levels of transcripts encoding β-Glucuronidase (GUSB). The data demonstrates that differentiation of macrophages (F4/80⁺) and OC (TRAP⁺) was not affected by iron levels in the culture, but iron deficiency caused an increase of levels of transcripts encoding TFRC and DMT1/Slc11A2.

Fig. 1.

Modulation of gene expression by exogenous iron in cultures of osteoclast progenitor cells (OPC). OPC were grown with (black bars) or without (white bars) RANKL in media rendered iron deficient (+DFO) or control media (ØDFO) for the last 48 h of the culture. Gene expression was characterized after 5 d of culture by quantitative RT-PCR and normalized against levels of transcripts encoding ß-glucuronidase. Levels of transcripts encoding the pan-MΦ marker F4/80 were reduced in osteoclastogenic cultures and transcript levels for the OC marker TRAP were detected only in cells grown with RANKL. Levels of transcripts encoding TFRC were low in MΦ and and significantly elevated in OCL in control media, and were further increased when OCL were grown in conditions of iron deficiency. Similarly, levels of mRNA encoding DMT1 were increased during development of OCL as compared to MΦ, and were significantly elevated in OCL grown without iron. Levels of mRNA encoding FPN were lower in OCL as compared to MΦ. The data shown represent the mean of triplicates (±SD) of one out of 3–4 biological replicates; * p < 0.05.

3.2. Iron uptake by osteoclasts is increased in iron deficient conditions

Cultures of OCL were treated from day 3 with and without DFO. At day 5, after transfer of the cells, the OCL were incubated for 6 h with different preparations of ⁵⁵Fe holoTF. The cellular uptake of ⁵⁵Fe was measured every hour (Fig. 2). Treatment of the cells for 48 h with DFO before the uptake phase caused a three-fold increase in iron uptake (6468 ± 1328 cpm vs. 17,630 ± 2809 cpm), while non-specific binding remained low (189 ± 23 cpm) (Fig. 2A). To further characterize the uptake of iron by OCL, cells were grown in media supplemented with DFO. During the uptake period, culture media were supplemented with or without DFO, ⁵⁵Fe was preincubated with DFO before addition of apoTF, and apoTF was preincubated with DFO before addition of ⁵⁵Fe (Fig. 2B). Addition of DFO to holoTF during the uptake period affected iron uptake after 6 h only slightly and non-significantly (17,630 ± 2809 cpm without DFO vs. 21,840 ± 2020 cpm with DFO). Preincubation of apoTF with DFO for 2 h and subsequent addition of ⁵⁵Fe caused a decrease in iron uptake by 80% (3774 ± 516 cpm), while preincubation of ⁵⁵Fe and DFO for 2 h before addition of apoTF blocked ⁵⁵Fe uptake completely (75 ± 6 cpm). In conclusion, iron uptake by OCL is increased when the cells are grown in iron deprived conditions, which is correlated to the observed elevated levels of transcripts encoding TFRC.

Fig. 2.

Uptake of iron by OCL in vitro. Osteoclast-like cells were pre-cultured for 48 h with or without DFO to render conditions iron-free. Thereafter holoTF was added for up to 6 h and iron uptake was measured (A). Upon pre-treatment of OCL with DFO (⯁-⯁), iron uptake was increased by a factor of 3 as compared to cultures without DFO (⬤-⬤). Non-specific binding of ⁵⁵Fe was virtually zero (■-■). To assess the affinity of DFO and TF to iron, iron uptake was analyzed in OCL grown for 48 h in iron depleted media and DFO was added at different time points during the preparation of holoTF and iron uptake (B). No significant differences in iron uptake were detected, when DFO was added during the process of iron uptake (▲-▲, with DFO; ⯁-⯁, without DFO). Pre-incubation of TF with DFO and subsequent addition of ⁵⁵Fe caused a decrease in iron uptake by 80% (⬤-⬤), while pre-incubation of iron with DFO prevented subsequent iron uptake (■-■). The data points represent the mean (±SD) of triplicates of one of 4–5 biological replicates; * p > 0.05, ** p > 0.01 compared to unspecific binding.

3.3. Transferrin binds to the cell surface independent of its iron load

To analyze cellular binding of TF, OCL were grown for 48 h in iron depleted medium. After harvesting, OCL were incubated with holoTF and apoTF for 2 h and subsequently incubated for another 4 h in culture medium devoid of TF and DFO. After lysis and gel electrophoresis, the blots were probed with antibodies against βactin as the loading control, and against TFRC and TF (Fig. 3A, B). Levels of TF (Fig. 3C) and of TFRC (Fig. 3D) were similar in all conditions, independent of the protocol used, whether holoTF or apoTF was added to the cells, or whether either Fe or TF were preincubated with DFO. After the 2 h uptake period, levels of TFRC were reduced by approx. 20% in comparison to control cells not exposed to TF. After another 4 h incubation, levels of TFRC were reduced by another 50%. Transferrin, either loaded or not loaded with Fe was bound to the cells after the 2 h period. Virtually no TF signal was detectable after another 4 h of incubation. Quantitative values have been normalized to βactin. The data demonstrates binding of TF to the cell surface, independently of its iron load.

Fig. 3.

Binding of TF to OCL. To assess the specificity of binding of holoTF to OCL, different preparations of holoTF, apoTF, and conditions as described in Fig. 2, were added to OCL and analyzed by western blot (A). Binding of TF was detected in all conditions after 2 h of incubation (t = 0 h), while the TF signal was missing after another 4 h (t = 4 h) of incubation in the absence of TF. Levels of TFRC were similar in all conditions, at both time points. The western blot was quantified and normalized to actin (B). Levels of TF were similar in all cell cultures after 2 h, and below detection after another 4 h (C). Levels of TFRC were decreased after 4 h by approx. 50% (D). Analysis was performed at least 4 times with similar results.

3.4. Osteoclast-like cells store intracellular iron in membrane vesicles

After the demonstration of iron uptake by OCL in vitro, its intracellular distribution was analyzed. For this purpose, cell lysates were fractionated on a sucrose gradient (0% - 60%) by ultracentrifugation and the gradient was divided into 15 fractions (Fig. 4). After 2 h of iron uptake, the ⁵⁵Fe was divided into two pools (Fig. 4C). Pool I was located in the low density fractions (F1-F4) and Pool II was located in fractions of higher density (F7-F12) of the gradient. In these same fractions (F7-F12), TFRC and CD51 were detected. After another 4 h of incubation of the cells, iron Pool I was no longer detectable, all cellular iron was found in Pool II. At this time point, cellular TF was virtually undetectable, while the locations of TFRC, EEA1, CD51 and FT remained the same for both time points (Fig. 4A, B). The data demonstrates that TF is lost during the 4 h incubation, while levels of CD51 remained constant. After 4 h, also the iron pool associated with the low density fractions had disappeared while the vesicle associated fraction was maintained, even in the absence of TF. Iron did not accumulate within ferritin vesicles, even though levels of ferritin light chain were increased.

Fig. 4.

Fractionation of OCL and distribution of iron and proteins. Osteoclast-like cells were incubated with holoTF for 2 h/0 h and for 2 h/4 h and were fractionated thereafter. Western blots were performed to detect TF, TFRC, ßactin, CD51, FT, and EEA1 on fractionated cells after 2 h of incubation (t = 0 h) (A) and after another 4 h of incubation without TF (t = 4 h) (B). In parallel, cells that were allowed to take up ⁵⁵Fe, were fractionated on a sucrose gradient, and Fe distribution within the fractions was assessed (C). Analysis was performed at least 4 times with similar results.

3.5. Intracellular iron is not transferred to ferritin particles or mitochondria in osteoclasts

To assess the distribution of proteins participating in cellular iron uptake and intracellular trafficking, OCL were grown for 48 h in culture medium supplemented with DFO. Thereafter, the cells were incubated for 2 h with hTF to allow for iron uptake. After the 2 h incubation, half of the cultures were prepared for cell fractionation, while for the other half of the cells, the period of iron uptake was followed by 4 h incubation in medium supplemented with DFO to allow for iron trafficking. The cells were solubilized and a fraction containing nuclei and cellular fragments was separated from the cytoplasmic fraction by a short centrifugation step. The distribution of proteins under study between the cytoplasmic and the pelleted fractions was assessed by western blot. Proteins were found to distribute unevenly between the pellet and the soluble fractions. Approx. 90% of the mitochondrial marker ATP5A1 was localized in the pellet fraction. Of actin, TF and TFR, 40% were detected in the pellet, for the Golgi Marker Golgin97 the number was 50%. The other proteins of this study (EEA1, early endosomes; CD51 cell membrane; DMT1/Slc11A2; HFE, hemochromatosis gene; Ft, ferritin light chain) localized predominantly in the soluble fraction (Table 3). The distribution of these proteins was very similar after 2 h and 2 + 4 h, respectively.

Table 3.

Cell fractionation – distribution of proteins in supernatants (solubilized proteins) and pellets (non-solubilized proteins).

	Protein	SN (%)	Pellet (%)
2 h/0 h	Actin	65	35
	Transferrin	60	40
	Transferrin Receptor	65	35
	EEA1	90	10
	ATP5A1	10	90
	CD51	85	15
	DMT1	100	0
	HFE	100	0
	Golgin 97	50	50
	Ferritin	95	5
	55Fe	83	17
2 h/4 h	Actin	60	40
	Transferrina	60	40
	Transferrin Receptor	60	40
	EEA1	80	20
	ATP5A1	10	90
	CD51	95	5
	DMT1	100	0
	HFE	100	0
	Golgin 97	50	50
	Ferritin	90	10
	55Fe	88	12

^a2 h/4 h < 10% 2 h/0 h.

Immediately after the 2 h iron uptake period, TF was detected in two not well separated pools, Pool I in fractions F3-F6 and Pool II in F9-F12 (Fig. 5). TFRC was most enriched in fractions F8-F12 and co-localized with the membrane marker CD51, which was detected in fractions F8-F11 (Fig. 4, 5). The mitochondrial marker ATP5 A1 was localized in F12, FT in F7-F9. Actin, which was used as a cytoplasmic marker, was most prominent in F2-F5, but was detected at low levels in all fractions of the gradient (Fig. 5). EEA1 was detected in F3 and F4 (Fig. 5), as was DMT1/Slc11A2 and HFE (Fig. 5). Ferritin was found to be localized in F7-F9.

Fig. 5.

Protein distribution in cell fractions. Osteoclasts were incubated for 2 h/0 h (black bars) and 2 h/4 h (white bars). Western blot signals were quantified for ßactin, TFRC, TF, FT, DMT1/ SLC11A2, and CD51. Analysis was performed at least 4 times with similar results.

After 2 h of iron uptake and a further 4 h of incubation in the presence of DFO, the levels of cellular TF were decreased by more than 90% and TFRC levels remained constant if normalized against total actin. The abundancy of FT increased by a factor of approx. 8, while the levels of the other proteins tested in these experiments were not significantly changed.

4. Discussion

Iron homeostasis is a crucial physiological process, regulating iron at the levels of the organism, the organs and the cells. To prevent iron storage or deficiency diseases, cellular iron uptake, trafficking, storage and elimination need to be tightly regulated. Within the present study some aspects of the uptake of iron by in vitro generated OCL and of intracellular trafficking were investigated.

Previous data suggests that during their development from monocyte/macrophage lineage cells to OC, the cells switch from an iron exporting phenotype in macrophages (MΦ) to an iron uptake phenotype in OC (Ishii et al., 2009; Xie et al., 2016). The differences in the cellular regulation of iron uptake reflects the different physiological roles of these cells in iron metabolism, MΦ recycling heme iron for further use (Vogt et al., 2021), OC using iron to generate the necessary energy for mineral dissolution and bone resorption (Das et al., 2022).

Cellular iron uptake mediated by TF, whose binding sites for ferric iron are occupied only 25% - 50% (Ritchie et al., 2002a; Ritchie et al., 2002b), is known to follow a defined route, with holoTF, if fully loaded with 2 ferric iron ions, binding to the TFRC. The iron-TF-TFRC complex is internalized within clathrin-coated pits, pH is lowered in the maturing endosomes, Fe³⁺ is released from TF and reduced to ferrous iron, which in turn is transferred into the cytoplasm by the endosomal divalent metal transporter 1 (DMT1/SLC11A2). The empty TF-TFRC complex subsequently is recycled to the cell membrane (Luck and Mason, 2012).

The data presented in this study do suggest that other mechanisms may regulate iron uptake and storage in osteoclast lineage cells. Growing OCL in media supplemented with the iron chelator DFO leads to an upregulation of TFRC (Ishii et al., 2009; Xie et al., 2016; Zhang et al., 2019) and thus, as shown here, to an increase in iron uptake. The data also shows that iron uptake is not hampered when holoTF is pre-incubated with DFO for 2 h before addition to the cell cultures, demonstrating that bound iron is not released from TF-Fe in the presence of DFO, even though the affinity of DFO to ferric iron exceeds the one of TF (DFO log K 30–31 vs. TF log K 21–23) (Steere et al., 2012; Pawlaczyk and Schroeder, 2021). Pre-incubation of apoTF or ⁵⁵Fe with DFO for 2 h, due to ferric iron – DFO complex formation, leads to a decrease in binding of iron to TF and as a consequence to a decrease in cellular iron uptake, demonstrating the formation of the DFO-Fe complex to be faster than the formation of the TF-Fe complex. Furthermore, if one of these complexes is formed, virtually no further exchange of iron between TF and DFO will take place. In all these conditions, TF binding to osteoclast lineage cells, however, is not changed. Previously it has been shown that also apoTF, though with lower affinity than holoTF, binds to TFRC (Xiu-Lian et al., 2004). This study also describes for apoTF a similar processing pathway as is observed for holoTF. Furthermore, more recently, apoTF was described to bind both to TFRC and TFR2 (Kleven et al., 2018). While the main pathway of iron uptake makes use of TFRC, in OC, TFR2 was demonstrated to play a pivotal role in controlling bone mass, deficiency in TFR2 leading to an increase in bone mass through a BMP2 dependent pathway (Rauner et al., 2019; Losser et al., 2024). In OC, levels of transcripts encoding TFRC, however, are much higher than levels of transcripts encoding TFR2. Even though it can be assumed that apoTF bound to TFRs will be competed off by holoTF, this mechanism may not be relevant in the experiments within the present study, since apoTF is added to the cell cultures without holoTF.

Upon separating cell organelles on a sucrose gradient, two cell-associated populations of iron can be distinguished after the 4 h uptake period. The first iron population (Pool I) is localized in low-density fractions (fractions 1–4) the second population (Pool II) is localized in medium to medium-high density fractions (fractions 8–12). Iron Pool I does not seem to be associated with any cell organelles or proteins, since even actin, as a cytoplasmic marker, is detected at highest levels only in fractions 2–4, while the highest levels of iron are already present in fraction 1. After another 4 h, iron Pool I has completely vanished, while iron Pool II is virtually unchanged in size and location on the gradient. Interestingly enough, immediately after iron uptake, TF also is separated into two pools on the density gradient, approximately equal levels of TF being detected in fractions 3–6 and in fractions 8–11. While TF in peak I is not associated with iron Pool I, peak II of TF and iron Pool II coincide. It might be tempting to speculate that either iron Pool I or the peak I of TF is associated with early endosomes, characterized by the Early Endosome Antigen 1, EEA1 (Simonsen et al., 1998), containing DMT1/SLC11A2 for iron extrusion. Two arguments speak against this interpretation, first, iron Pool I and the first TF peak are not superimposed and in the fractions containing early endosomes no TFRC was detected. Therefore it is suggested that the first TF peak and iron Pool I are derived from cell surface holoTF binding, which has not yet been internalized and which has been released during the procedure of cell fractionation. The parallel appearance of early endosome markers and TF peak I are rather a coincidence and do not demonstrate TF being part of early endosomes. The data also demonstrates that during gradient centrifugation, the presence of a marker such as EEA1 demonstrates the presence of early endosomes, but by no means demonstrates the exclusive presence of early endosomes in these fractions.

In contrast to iron Pool I, iron Pool II in fractions 8–12 after the 2 h uptake period is associated with TF, TFRC and the OC surface marker CD51 (Integrin αV), which is part of the vitronectin receptor (Horton et al., 1991). After another 4 h, TF can no longer be detected in these fractions, in contrast to iron, TFRC and CD51. Two sources may contribute to CD51/TFRC positive vesicles, (i) membrane vesicles generated during disruption of the cell membrane and (ii) internalized vesicles upon TF binding to TFRC. The latter are recycled, as described above to release TF and to reexpose TFRC on the cell surface, the former are being formed, independent of TF binding. After the uptake period, fractions 8–12 contain a mixture of (i) and (ii), while after 4 h of further incubation, only membrane vesicles are present.

This leaves the question of the nature of the iron storing organelles after the recycling of the TFRC/TF complex. Of particular interest seems the fact that at both time points investigated, the TFRC/TF complex is never associated with DMT1/SLC11A2, required to transport endosomal ferrous iron into the cytoplasm. Furthermore, cellular iron is not associated with ferritin particles, the ferritin light chain being upregulated during the 4 h incubation period after the uptake phase, nor with mitochondria, which are predominantly pelleted during the cell disruption process and no iron accumulation in the pellet was detected.

In summary, this study demonstrates that iron uptake by OCL generated in vitro is dependent on TFRC levels, which in turn are modulated by the availability of exogenous iron. Western blot data suggest that both, apoTF and holoTF bind to the cell surface. This study, however, is not able to demonstrate whether apoTF is internalized after cell surface binding as is the case for holoTF and ⁵⁵Fe. The TF-TFRC complex subsequently is recycled to the cell surface, while iron stays within a vesicular fraction captured in the cells. In this study, it was not possible to track the further destiny of the cellular iron, but it could be excluded that within the investigated time frame, iron was stored in ferritin particles or in mitochondria. It will take further studies to elucidate the function and roles of iron in mature OC, in particular, whether iron transport to be activated requires OC actively resorbing bone.

CRediT authorship contribution statement

Silvia Dolder: Methodology, Investigation, Formal analysis, Data curation. Romina Cabra: Writing – review & editing, Validation, Investigation, Formal analysis, Data curation. Jonas Zaugg: Validation, Methodology, Investigation, Data curation. Giuseppe Albano: Validation, Methodology, Data curation. Daniel G. Fuster: Writing – review & editing, Supervision, Formal analysis, Conceptualization. Christiane Albrecht: Writing – review & editing, Writing – original draft, Validation, Supervision, Investigation, Formal analysis, Conceptualization. Willy Hofstetter: Writing – review & editing, Writing – original draft, Validation, Supervision, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.