Ferritin ferroxidase activity: A potent inhibitor of osteogenesis

Abstract

Hemochromatosis is a known cause of osteoporosis, and iron overload has deleterious effects on bone. Although iron overload and its association with osteoporosis has long been recognized, the pathogenesis and exact role of iron have been undefined. Bone is an active tissue with constant remodeling capacity. Osteoblast (OB) development and maturation are under the influence of core binding factor α-1 (CBF-α1), which induces expression of OB-specific genes, including alkaline phosphatase, an important enzyme in early osteogenesis, and osteocalcin, a noncollagenous protein deposited within the osteoid. This study investigates the mechanism by which iron inhibits human OB activity, which in vivo may lead to decreased mineralization, osteopenia, and osteoporosis. We demonstrate that iron-provoked inhibition of OB activity is mediated by ferritin and its ferroxidase activity. We confirm this notion by using purified ferritin H-chain and ceruloplasmin, both known to possess ferroxidase activity that inhibited calcification, whereas a site-directed mutant of ferritin H-chain lacking ferroxidase activity failed to provide any inhibition. Furthermore, we are reporting that such suppression is not restricted to inhibition of calcification, but OB-specific genes such as alkaline phosphatase, osteocalcin, and CBF-α1 are all downregulated by ferritin in a dose-responsive manner. This study corroborates that iron decreases mineralization and demonstrates that this suppression is provided by iron-induced upregulation of ferritin. In addition, we conclude that inhibition of OB activity, mineralization, and specific gene expression is attributed to the ferroxidase activity of ferritin. © 2010 American Society for Bone and Mineral Research

Introduction

Bone is a specialized connective tissue composed of cells and extracellular matrix. Its distinctive feature is the mineralization of the matrix, which produces a hard tissue capable of providing support and protection. Like its close relatives, the fibroblast and chondroblast, the osteoblast (OB) is a versatile secretory cell that retains the ability to divide and proliferate.

The accumulation of excessive amounts of iron in the body can damage various organ systems, particularly the liver, the pancreas and other endocrine organs, and the heart. The term hemochromatosis refers to all forms of iron storage disease. Some of these occur without any other underlying disorder and therefore are designated as primary. Others are secondary to other diseases, particularly those associated with ineffective erythropoiesis. Although traditionally lower levels of sex steroid hormones were suggested to be the major mediator of bone loss in these patients, many lines of evidence now indicate that iron and its overload have direct deleterious effects on bone, causing both osteoporosis and osteopenia.1–4 These studies (reviewed in ref. 4) show that iron loading in primary and secondary hemochromatosis has a direct effect on bone and OB activity. They also conclude that the effect of iron is to decrease bone deposition rather than to increase bone resorption. Despite numerous in vitro and in vivo data, the precise mechanism by which iron can cause a decrease in bone mass and its mineralization has not been thoroughly studied yet.

Iron is essential for many important cellular functions, and its potential toxicity is modulated mainly by the ferritins. These are iron storage molecules, tightly regulated by iron, that have a highly conserved three-dimensional structure. It consists in a hollow protein shell (outside diameter 12 to 13 nm, inside 7 to 8 nm, M_r ≈ 500 kDa) that permits storage of up to 4500 Fe³⁺ atoms in a safe and bioavailable form.5 Each apoferritin (iron-free ferritin) shell is made up of 24 polypeptide chains of two types.6, 7 The H subunit (21 kDa) has a relatively acidic electrophoretic mobility and carries a ferroxidase activity that promotes iron incorporation and oxidizes Fe²⁺ into the safer Fe³⁺ form.6 The L subunit (19 kDa) is associated with iron nucleation, mineralization, and long-term iron storage.5, 8 The ratio between H and L subunits in a ferritin shell varies widely in different tissues, and the expression of ferritin is under delicate control at both the transcriptional and posttranscriptional levels.9, 10

Osteoporosis and osteopenia that accompany iron overload prompted us to closely examine the process by which iron can cause decreased bone mass. Iron is a strong inducer of ferritin; hence we investigated whether iron is solely responsible for such inhibition or whether it is rather the upregulation of ferritin caused by iron administration that can provide such inhibition. Since ferritin sequesters iron and possesses ferroxidase activity, we have examined whether one or both of these features can have any impact on the level of OB activity. In order to confirm the role of ferroxidase activity, we investigated the role of ceruloplasmin, which is well known to have ferroxidase activity.

Materials and Methods

Cell culture and reagents

Human osteoblasts were purchased from Cambrex, and human osteosarcoma 143-B cells were obtained from the American Type Culture Collection and fetal bovine serum (FBS) from GIBCO. Unless otherwise mentioned, all other reagents were obtained from Sigma. Cell cultures were maintained in growth medium (GM) DMEM (high glucose) containing 15% FBS, 100 U/mL penicillin, 100 µg/mL streptomycin and neomycin, and 1 mM sodium pyruvate. Cells were grown to confluence and used from passages 3 to 7. Iron was introduced as ammonium ferric citrate dissolved in deionized water. Apoferritin and ceruloplasmin were from Sigma. Human recombinant wild-type H-chain ferritins and the H-chain mutant 222 deleted ferroxidase activity were expressed in Escherichia coli and purified as described previously.11 Final concentration of ferritins was 2 mg/mL, and ceruloplasmin was 3 mg/mL.

Induction of calcification

At confluence, OBs were switched to calcification medium, which was prepared by adding 2.5 mmol/L of inorganic phosphate to the growth medium. 143-B cells calcification medium contained 4 mmol/L of inorganic phosphate. Both growth medium and calcification medium were changed every 2 days. For time-course experiments, the first day of culture in calcification medium was defined as day 0.

Quantification of calcium deposition

Cells grown on 48-well plates were washed twice with phosphate-buffered saline (PBS) and decalcified with 0.6 mol/L HCl for 24 hours. Calcium content of the supernatants was determined by the QuantiChrome Calcium Assay Kit (Gentaur). After decalcification, cells were solubilized with a solution of NaOH 0.1 mol/L and SDS 0.1%, and protein content of samples was measured with a BCA protein assay kit (Pierce). Calcium content of the cells was normalized to protein content and expressed as µg/mg protein. Mineralization also was determined by alizarin red staining.

Alkaline phosphatase activity assay

Cells grown on 6-well plates were washed with PBS twice, solubilized with 1% Triton X-100 in 0.9% NaCl, and assayed for ALP activity. Briefly, 130 µL of Alkaline Phosphatase Yellow Liquid Substrate (Sigma) was combined with 5 µg protein samples, and then the kinetics of p-nitrophenol formation were followed for 30 minutes at 405 nm at 37°C. Maximum slope of the kinetic curves was used for calculation.

Western blot to detect ALP, CBF-α1, and ferritin H-chain

To evaluate ALP and CBF-α1 protein expression, cell lysate was electrophoresed in 12.5% SDS-PAGE. For ferritin H-chain detection, cell lysate was subjected to 8% nondenaturing PAGE. Western Blotting was performed with a polyclonal anti-ALP antibody at 1:2000 dilution (Calbiochem), polyclonal anti-CBF-α1 antibody at dilution 1:200 (Santa Cruz) or with mouse anti-human ferritin H-chain antibodies (from P. Arosio) at 1:1000 dilution followed by HRP-labeled antimouse IgG antibody. Antigen-antibody complexes were visualized with the horseradish peroxidase chemiluminescence system (Amersham Biosciences). After detection, membranes were stripped and reprobed for glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Quantification of chemiluminescence was done by using Alpha DigiDoc RT quantification software.

Quantification of ferritin and osteocalcin

Ferritin content of cell lysate was measured with the IMx ferritin enzyme immunoassay (Abbott Laboratories). For osteocalcin detection, extracellular matrix of cells grown on 6-well plates was dissolved in 300 µL of EDTA (0.5 mol/L, pH 6.9). Osteocalcin content of the EDTA-solubilized extracellular matrix samples was quantified by an ELISA (Bender MedSystems).

Phosphate measurement

P_i content of cell lysate was determined by the QuantiChrome Phosphate Assay Kit (Gentaur). After 24 hours of incubation, cells were washed twice with PBS and solubilized with 1% Triton, and the cell lysates were assayed for P_i. Phosphate content of the cells was normalized to protein content and expressed as µmol/L/mg cell protein.

Statistical analysis

Data are shown as mean ± SD. Statistical analysis was performed by ANOVA test followed by a post hoc Newmann-Keuls test for multiple comparisons. A value of p < .05 was considered significant and marked with the letter a, and p < .01 was considered highly significant and marked with the letter b.

Results

Iron inhibits calcification in a dose-responsive manner

To develop an in vitro model of osteoblastic activity and mineralization, we cultured human OBs in calcification medium that was prepared by addition of 2.5 mmol/L of P_i to the growth medium (GM). Granular deposits developed in OBs grown in calcification medium for 14 days but not in the control culture grown in normal GM (Fig. 1A), as demonstrated by alizarin red staining. We found that addition of iron to the calcification medium suppresses granular deposit development (see Fig. 1A) and extracellular calcium deposition in a dose-responsive manner, causing highly significant inhibition at a concentration of 25 µmol/L and a complete inhibition at 50 µmol/L (see Fig. 1B). Since iron is a very potent inducer of ferritin (see Fig. 2A, B), next we tested whether the observed inhibitory effect of iron on calcium deposition is mimicked by apoferritin. We found that iron-free apoferritin at a dose of 2 mg/mL abolishes granule formation (see Fig. 1A) and dose dependently inhibits calcium deposition, causing complete inhibition when applied at a concentration of 2 mg/ml (see Fig. 1C). Then we asked if decreasing the level of available iron with the iron chelator deferoxamine (DFO), which, in turn, leads to the posttranscriptional downregulation of ferritin synthesis (see Fig. 2A–C), could alter the level of calcification. Indeed, addition of DFO (10 µmol/L) to the calcification medium increased the levels of calcium deposition by approximately 20% (see Fig. 1A, panel v, and Fig. 1D). In fact, both endogenous upregulation and downregulation of ferritin H and L chains with Fe or DFO, respectively, and exogenous administration of apoferritin caused alterations in calcium deposition of OBs. We found a strong negative correlation between the ferritin levels of the cells and the observed calcium depositions (see Fig. 2D).

Figure 1. Iron inhibits extracellular calcium deposition in a dose-responsive manner. (A) Human osteoblasts were cultured in growth medium and calcification medium supplemented with iron or apoferritin for 14 days. In vitro OB calcification was determined by alizarin red staining. Representative images of stained plates (upper panel) and microscopic views (×100, lower panel) from three independent experiments are shown. (B–D) OBs were cultured in GM or calcification medium alone or supplemented with 10, 25, or 50 µmol/L or iron (B); 0.1, 0.5, 1, and 2 mg/mL of apoferritin (C); or 1, 5, and 10 µmol/L of DFO (D). Calcium content of cells was measured after 14 days' incubation, and the level was normalized by protein content of cells. Data are presented as mean ± SD of three independent experiments performed in duplicates. ^ap < .05. ^bp < .01.

Figure 2. Iron inhibits OB mineralization via induction of ferritin. OBs were cultured in GM or calcification medium alone or in the presence of iron (50 µmol/L), DFO (10 µmol/L), iron + DFO (50 µmol/L each), and apoferritin (2 mg/mL) for 14 days. (A) Representative Western blot shows expression of ferritin H (Ft-H). Same cell lysates were electrophoresed in 12.5% SDS-PAGE to detect GAPDH and show equal loading of protein. (B) Densitometric measurement of the band intensities for ferritin H was normalized to GAPDH and is representative of three independent experiments. (C) Cell lysate was used to quantify ferritin expression by an immunoassay as described in the Methods section. Graph shows mean ± SD of three separate experiments performed in duplicate. (D) Strong negative correlation between ferritin levels of cells and calcium deposition in OBs cultured in calcification medium alone (♦) or supplemented with iron (▴), DFO (●), iron + DFO (▪), or apoferritin (▾). ^bp < .01.

Ferroxidase activity is responsible for inhibition of calcification

To investigate whether the inhibitory effect of ferritin is due to its iron sequestration capacity or its ferroxidase activity or both, we tested the effect of ceruloplasmin, a protein that possesses ferroxidase activity but not iron sequestration capacity. In fact, as indicated in Fig. 3A, exogenous ceruloplasmin dose-dependently inhibited mineralization, causing abolishment of calcium deposition at a dose of 3 mg/mL. Experiments using recombinant H-ferritin and the H-mutant 222 ferritin that lacks both ferroxidase activity and iron-storing capability provided further evidence of the role of ferroxidase activity in the observed inhibitory effect; whereas H-ferritin attenuated calcium deposition at a dose of 2 mg/mL, the H-mutant 222 ferritin did not alter calcification at all when applied at the same dose (see Fig. 3B).

Figure 3. Ferritin ferroxidase activity is responsible for inhibition of calcification. OBs were cultured in GM or calcification medium for 14 days, and the calcification medium were supplemented with 0.5, 1, 2, and 3 mg/mL ceruloplasmin (CP) (A) and H-ferritin or mutant 222 ferritin (Mutant Ft-H) 2 mg/mL. (B) Calcium deposition was measured after 14 days and normalized by cellular protein. Data are presented as mean ± SD of five independent experiments performed in duplicates. ^ap < 0.05. ^bp < .01.

The ferroxidase activity of apoferritin and H-chain ferritin was shown to transiently generate hydrogen peroxide from the reaction between ferrous iron and dioxygen,12 and hydrogen peroxide was shown to inhibit osteoblastic differentiation of bone cells.13 Therefore, we tested whether addition of catalase (degrades hydrogen peroxide) abrogates the inhibition of osteogenic mineralization provided by apoferritin and H-chain ferritin. Catalase added to OBs at a dose of 200 U/mL failed to increase calcification when applied with apoferritin or H-ferritin (5.85 ± 1.21 and 5.39 ± 1.6 versus 6.75 ± 1.52 and 3.85 ± 0.3 µg/mg protein, respectively).

Ferritin ferroxidase activity downregulates expression and subsequent activity of ALP

To investigate whether the observed effects are only restricted to reduced calcification or influence OB gene expression as well, we examined the role that ferritin may play regarding OB-specific genes. First, we examined the ALP gene because its activity is a good marker of osteoblastic activity. ALP activity is crucial in early osteogenesis by increasing local levels of P_i and therefore facilitating the formation of hydroxyapatite crystals.

Basal ALP activity of OBs was quite high when they were cultured in a normal GM, and addition of P_i to the GM caused a minor but significant increase in ALP activity (Fig. 4A). However, addition of iron to the high P_i-containing calcification medium led to a significant decrease of ALP activity, which became much lower than the basal ALP activity of OBs cultured in GM. Accordingly, this inhibitory effect was attenuated when using equimolar concentrations of iron and DFO together. Addition of 10 µmol/L of DFO alone caused a minor but significant increase in the level of ALP activity (see Fig. 4A). Supplementation of the calcification medium with apoferritin caused a dose-responsive inhibition of ALP expression and activity (see Fig. 4B–D). The role of ferroxidase activity behind the observed inhibitory effect of iron and ferritin also was assessed by using wild-type and mutant H-ferritin and ceruloplasmin. Whereas addition of both H-ferritin and ceruloplasmin to the calcification medium strongly downregulated ALP activity, H-mutant 222 ferritin did not influence ALP activity at all (see Fig. 4D), confirming ferroxidase activity as the central element of this inhibition.

Figure 4. Suppression of ALP expression and activity by ferritin ferroxidase. (A) OBs were cultured in GM or calcification medium alone or in the presence of iron (50 µmol/L), DFO (10 µmol/L), or iron + DFO (50 µmol/L each) and apoferritin (2 mg/mL) for 14 days. ALP activity of cells was measured as described in Methods. Data are expressed as means ± SD of five independent experiments each performed in duplicates. ^bp < .01. ^ap < .05. (B) OBs were cultured in GM or calcification medium alone or supplemented with 0.1, 0.5, 1, and 2 mg/mL of apoferritin for 14 days. Representative of three Western blots shows ALP expression. Membrane was reprobed for GAPDH to show equal loading of proteins. (C) Densitometric measurement of the band intensities for ALP was normalized to GAPDH and presents measurements from three independent experiments. (D) OBs were cultured as in (B), and ALP activity was measured. Data show average ± SD of three independent assays performed in triplicate. (E) OBs were cultured in GM or calcification medium alone or supplemented with recombinant ferritin H-chain wild-type or mutant 222 (2 mg/mL) or ceruloplasmin (3 mg/mL) for 14 days. ALP activity was measured and is shown as mean ± SD of three independent experiments done in duplicates.

Ferritin ferroxidase activity suppresses the exclusive OB product: Osteocalcin

While maintaining OBs in calcification medium for 14 days results in an approximately 13-fold increase in the amount of osteocalcin deposition in the newly synthesized extracellular matrix, addition of iron decreased the level of osteocalcin dose-responsively and caused very significant inhibition at doses of 25 µmol/L and almost complete inhibition at a dose of 50 µmol/L (Fig. 5A). Accordingly, supplementation with apoferritin also caused a dose-dependent inhibition in the amount of osteocalcin, providing very significant inhibition at 0.5 mg/mL, and complete inhibition was resulted at a dose of 2 mg/mL (see Fig. 5B). To further confirm that such inhibition is provided mainly by ferroxidase activity, ceruloplasmin and H-ferritin (wild-type and mutant 222) were added to the calcification medium. Whereas ceruloplasmin and H-ferritin abolished osteocalcin production, H-mutant 222 ferritin did not cause any significant inhibition regarding the amount of osteocalcin deposition (see Fig. 5C).

Figure 5. Ferritin attenuates upregulation of osteocalcin. OBs were cultured in GM or calcification medium alone or supplemented with iron (50 µmol/L) or DFO (10 µmol/L) (A), apoferritin (0.1, 0.5, 1, or 2 mg/mL) (B), and recombinant ferritin H-chain or mutant 222 at a dose of 2 mg/mL or ceruloplasmin 3 mg/mL (C) for 14 days. The extracellular matrix was dissolved, and osteocalcin deposition was quantified as described in methods. Data derived from three separate experiments performed in triplicates and shown as mean ± SD. ^bp < 0.01. ^ap < 0.05.

Intracellular P_i concentrations are not affected by apoferritin or ceruloplasmin

Because ferric iron is known to bind phosphate and to examine the mechanism by which both mineralization and OB gene expression are downregulated, we measured P_i uptake of OBs after 24 hours of incubation in calcification medium. Our results indicate (Fig. 6A) that iron caused a mild but significant decrease in the level of intracellular P_i. This is probably due to the phosphate-binding capacity of ferric iron. On the contrary, neither apoferritin nor ceruloplasmin altered intracellular P_i concentrations 24 hours after incubation. Hence iron is capable of binding to phosphate and hence causing a minor decrease in the level of extracellular P_i, but this seems to contribute only in minor proportion to the mechanism of action of iron on bone deposition. This notion is firmly confirmed by the inhibition caused by apoferritin, H-ferritin, and ceruloplasmin, which are all free of iron.

Figure 6. Intracellular P_i concentrations are not affected by apoferritin or ceruloplasmin. (A) OBs were cultured in GM or in calcification medium alone or supplemented with iron (50 µmol/L), apoferritin (2 mg/mL), or ceruloplasmin (3 mg/mL) for 24 hours. Cell lysates were used to measure P_i levels, as described in the methods. Results are presented as mean ± SD of three independent experiments performed in duplicates. CBF-α1 levels are suppressed in a dose-responsive manner by apoferritin supplementation. (B) After 14 days culturing of OBs in GM or calcification medium alone or supplemented with 0.1, 0.5, 1, and 2 mg/mL apoferritin, cells were lysed, and Western blot was performed as described in Methods. After detection of CBF-α1, membrane was stripped and reprobed for GAPDH to prove equal loading. (C) Densitometric measurement of the band intensities for CBF-α1 was normalized to GAPDH. Both the Western blot and densitometric measurements are representative of three separate experiments. ^bp < .01. ^ap < .05.

Ferritin inhibits OB-specific transcription factor CBF-α1

It is well recognized that OB maturation and differentiation are powerfully influenced (if not dictated) by the transcription factor CBF-α1. Hence we investigated if ferritin could downregulate expression of CBF-α1 and therefore put forward a possible explanation for the observed inhibitory effects of ferritin on OB activity. Indeed, supplementation of the calcification medium with apoferritin caused a dose-dependent downregulation of CBF-α1 protein expression (see Fig. 6B).

Ferritin prevents calcification of human 143-B cells

In order to confirm the inhibitory effect of ferritin and ferroxidase activity on mineralization, we employed another human osteoblastic cell line: 143-B osteosarcoma cells. Granular deposits developed in 143-B cells grown in calcification medium for 7 days (Fig. 7A) but not in the control culture grown in normal GM, as demonstrated by alizarin red staining. We found that addition of iron to the calcification medium prevented granular deposit development and extracellular calcium deposition at a concentration of 50 µmol/L (see Fig. 7A, B). Furthermore, exposure of cells to apoferritin also abolishes granule formation and completely inhibits calcium deposition when applied at a concentration of 2 mg/mL (see Fig. 7A, B). Importantly, cells treated with ceruloplasmin fail to exhibit granule formation and extracellular calcium accumulation at a concentration of 3 mg/mL. These results confirm that ferroxidase activity acts as an inhibitor in OB mineralization.

Figure 7. Ferritin prevents calcification of human 143-B cells. (A) 143-B cells were culture in growth medium and calcification medium alone or supplemented with iron (50 µmol/L), iron with DFO (both are 50 µmol/L), apoferritin (2 mg/mL), or ceruloplasmin (3 mg/mL). Granular deposits were determined by alizarin red staining. Representative images of stained plates (upper panel) and microscopic views (×100, lower panel) from three independent experiments are shown. (B) After 7 days of culturing of 143-B cells, extracellular calcium deposition was determined and was normalized to protein level. Data are presented as mean ± SD of three independent experiments performed in duplicates. ^bp < .01.

Discussion

In this study we examined the process of osteogenesis in vitro by growing human OBs in a phosphate-rich medium, and in particular, we studied the effect of iron. Our findings strongly suggest that the inhibition of OB activity that occurs with iron supplementation is mainly due to the iron-mediated induction of H-ferritin and its ferroxidase activity. By using DFO, we demonstrated that the iron sequestration capability of ferritin is not responsible for the observed effects. The inhibition occurs after the addition of recombinant H-ferritin and ceruloplasmin, proteins with ferroxidase activity, but not by H-ferritin mutant with inactivated ferroxidase activity. Here we show that ferritin ferroxidase activity plays a crucial role in downregulation of mineralization and the underlying gene expression.

Bone is a dynamic, highly vascularized tissue with a unique internal repair capacity to heal and remodel without scarring.14, 15 Several studies have suggested a link between excessive and misplaced iron and secondary decreased bone mass. In a study of iliac crest biopsies from 21 individuals with severe osteoporosis (at least one vertebral fracture), iron bone concentration (cortical and trabecular) was evaluated using inductively coupled plasma optical emission spectrometry. A significant increase in iron content in cortical bone was found in osteoporotic patients versus 12 controls.16 There is a growing body of evidence that iron can play a deleterious role in bone. In genetic hemochromatosis, previous human studies have found Perl's Prussian blue staining (which unambiguously identifies iron) in bone trabeculae of patients.17, 18 Moreover, bone mineral density is decreased in patients with genetic hemochromatosis and severe iron overload19 and negatively correlates with hepatic iron concentration (a good index of total-body iron overload) and bone mineral density at the femoral neck.2 High serum ferritin also has been suggested as a reliable indicator of iron stores in chronic hemodialysis patients.20

A study of iron overload with intramuscular iron dextran was conducted in pigs over 36 days.21 The main effect was a decrease in bone formation without significant changes in bone resorption. Similarly, osteopenia was induced in Sprague-Dawley rats fed a diet containing iron lactate (5%) for 2 or 4 weeks.22 Other lines of evidence firmly confirm the close relationship between iron metabolism and osteogenesis. For instance, lactoferrin, an iron-binding glycoprotein present in epithelial secretions such as milk and in the secondary granules of neutrophils, is suggested as a potent regulator of bone cell activity, and it was reported to increases bone formation in vivo.23, 24 Additionally, there is evidence that a green tea iron chelator, epigallocatechin-3-gallate, stimulates mineralization of murine bone marrow mesenchymal stem cells.25

Clinical reports also show that several hemolytic anemias, including thalassemia and sickle-cell anemia, eventually develop iron overload. There is strong relationship between such elevated iron levels and the occurrence of osteopenia and osteoporosis in these patients.26–29 DFO is a powerful chelator of iron, zinc, cobalt, and copper, and it is used commonly to prevent iron overload. Although this may be considered to increase osteogenesis by decreasing iron and subsequent ferritin upregulation, some studies suggest that in addition to the growth retardation owing to untreated thalassemia, DFO produces a further negative effect on growth velocity by causing bone dysplasia. The decrease in growth velocity mainly affects the long bones, in particular, the distal femoral physis, which normally accounts for 70% of femoral growth. It should be noted that these trace metals are more likely to be chelated in the presence of reduced iron levels. Actually, serum zinc levels were below normal limits in 37% of chelated patients in one study.30 Zinc deficiency is associated with delayed skeletal maturation and a reduction in growth, as well as bone matrix and collagen synthesis. In fact, iron chelation with DFO and simultaneous zinc supplementation have been shown to increase growth in some thalassaemic patients. This further verifies that decreased iron levels can have a beneficial effect on OB activity.

Our study has focused on shedding new light on better understanding the mechanism by which iron causes repression of OB activity and indicates that such derangement is largely due to ferritin and its ferroxidase activity. In agreement with previous studies, we confirm that iron inhibits OB activity in a dose-responsive manner in vitro. However, our investigations clearly indicate that such inhibition is mainly due to iron-induced upregulation of ferritin ferroxidase activity. This notion is supported by the fact that apoferritin, which contains very little iron, if any, causes a dose-responsive decrease of OB gene expression and subsequent calcification. We have demonstrated dose-responsive uptake of apoferritin previously,31 and in agreement with previous studies, our Fig. 3 shows such uptake. Inhibitory effects of apoferritin can be mimicked with ceruloplasmin. This is an enzyme synthesized in the liver containing six atoms of copper in its structure that carries approximately 90% of the copper in plasma. However, we used this protein in our study because of its well-known ferroxidase activity. The importance of ferroxidase activity in the process was further confirmed by the finding that while recombinant H-ferritin had the same effects observed with apoferritin and ceruloplasmin, a structurally analogous molecule to H-ferritin, namely, the recombinant H-ferritin mutant 222, which lacks ferroxidase activity and iron-storage capability, was completely ineffectual in inhibiting calcification and OB gene expression. Our results also demonstrate that the observed effects of ferritin are not due to alterations of P_i uptake, and the decrease in the level of intracellular P_i that results from supplementation with ferric iron can be explained by its P_i-binding capacity.

Several studies have reported localization of ferritin into the nucleus, which traditionally was recognized as a cytoplasmic protein.32–34 These studies demonstrate that H-ferritin is the major form found in nucleus and that it might be protective against oxidative damage or regulate certain genes.35 In this study we demonstrate that apoferritin dose-dependently decreases the expression of CBF-α1. However, it must be noted that this may not be a direct and sole explanation for the observed effects, and there may be other mechanisms responsible for such downregulation that must be elucidated by future studies.

In the iron oxidation catalyzed by the ferritin ferroxidase center, ferrous iron reacts with dioxygen to generate hydrogen peroxide and ferric iron.12 The hydrogen peroxide is produced inside the ferritin shell, and most of it is used by the ferroxidase center to oxidize iron.36 However, some could escape signal directly via oxidative regulation of regulatory thiols in signal-transducing phosphatases.37 Indeed, depending on cell types, exogenous hydrogen peroxide or lipid hydroperoxides can either increase (vascular smooth muscle cells)13, 38, 39 or decrease (osteoblast)13, 38 the elaboration of osteogenic gene regulatory programs. The hydrogen peroxide generated by the ferroxidase activity is one of the candidates that might contribute to the suppression of mineralization and OB maturation provided by ferritin. The lack of any effect of catalase on the OB calcification inhibition by ferritin and the inhibitory effect of ferritin on calcification and osteoblastic differentiation of smooth muscle cells40 suggest that hydrogen peroxide is not a major mediator in suppression of mineralization and OB maturation provided by ferritin.

This study focused mainly on providing more insight into iron overload and its association with decreasing OB activity, bone deposition, and subsequent osteopenia and osteoporosis. Our findings for the first time explain that iron as a risk factor for decreased osteogenesis mainly exerts its inhibitory actions via upregulation of ferritin, and we suggest that ferroxidase activity of ferritin is crucial in suppressing mineralization and OB maturation. It should be noted that future in vivo experiments must be carried out to validate these current in vitro findings. These findings will expand the knowledge of both osteogenesis and pathogenesis of iron caused bone defects, and by exposing ferritin as the cause of decreased OB maturation may offer new understanding into better planning strategies to prevent or reverse iron-induced osteoporosis and osteopenia.

Disclosures

The authors state that they have no conflicts of interest.

Acknowledgements

We thank Dr. Anupam Agarwal for helpful discussions and critical reading of the manuscript.

This work was supported by Hungarian Government Grants OTKA-K61546, OTKA-K75883, ETT-337/2006, RET-06/2004, and MTA-DE-11003; Paolo Arosio is supported by MIUR-PRiN-06; and Viktória Jeney is supported by the European Commission's 7th Framework Marie Curie Grant GasMalaria.

The first two authors contributed equally to these studies. The last two authors contributed equally to these studies.