Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Anthracyclines are potent antineoplastic agents used extensively in the treatment of a range of cancers (Gianni and Myers, 1992; Gerwirtz, 1999). However, their efficacy is severely hindered by the development of cardiotoxicity at high cumulative doses, resulting in heart failure (Gianni and Myers, 1992; Gerwirtz, 1999). The antineoplastic effects of anthracyclines have been well characterized, yet little is understood regarding the mechanisms responsible for their cardiotoxicity (Gianni and Myers, 1992; Gerwirtz, 1999). To enhance cancer management using anthracyclines, it is important to investigate the mechanisms of both cardiotoxicity and antitumor activity. This knowledge is required for the development of treatment regimens that will enhance the antineoplastic effects and limit the cardiotoxicity of these agents.

There is good evidence that the cardiotoxicity of anthracyclines is caused, at least in part, by their avid interaction with iron (Fe). In fact, it is well known that anthracyclines strongly bind Fe, forming metal ion complexes (Gianni and Myers, 1992). Furthermore, Fe loading has been shown to potentiate the cardiotoxicity of the anthracycline doxorubicin (DOX) (Hershko et al., 1993; Link et al., 1996). It is also of interest that the only clinical intervention for anthracycline-mediated cardiotoxicity is the Fe chelator ICRF-187 (also known as dexrazoxane; Gerwirtz, 1999). In addition, another Fe chelator, desferrioxamine (DFO), which is used clinically for the treatment of Fe overload disease, can reduce the cardiotoxic effects of DOX (Hershko et al., 1993; Link et al., 1996). Collectively, these studies convincingly suggest a role for Fe in anthracycline-mediated cardiotoxicity.

Cells obtain Fe via the binding of transferrin (Tf) to the transferrin receptor 1 (TfR1; Richardson and Ponka, 1997). The Tf-TfR1 complex is then internalized into cells by receptor-mediated endocytosis. Acidification of the endosome results in Fe release from Tf, which is then transported across the membrane by Nramp2 (also known as divalent metal ion transporter DMT1; Gunshin et al., 1997). Once Fe enters the cytosol, it becomes incorporated into a poorly characterized compartment known as the intracellular Fe pool. This pool has been suggested to contain low molecular mass Fe complexes or it could be composed of high molecular mass chaperone molecules that bind Fe (Richardson and Ponka, 1997). From this pool, Fe can be incorporated into cytochromes and [Fe-S] proteins, etc., or can be stored in ferritin (Richardson and Ponka, 1997). However, the precise pathways involved in the uptake and release of Fe from proteins such as ferritin remain unknown.

Ferritin consists of two types of subunits, heavy (H) and light (L) chains. Twenty-four subunits are organized into a symmetrical structure (with 4-, 3-, and 2-fold axes), generating a cavity within the protein for Fe storage. At the 3-fold axes of the protein structure, there are channels traversing the protein shell; these are believed to be the main entry routes for Fe(II) and its oxidation to Fe(III) (Harrison and Arosio, 1996). Although the Fe deposition pathway that leads to the incorporation of the Fe core within ferritin has been investigated, there is little information regarding the process of Fe release from ferritin under physiological conditions.

Despite this property of anthracyclines to readily bind Fe, few studies have examined the effect of these drugs on cellular Fe metabolism. The most comprehensive experiments to date have examined the effects of anthracyclines on the iron regulatory proteins (IRPs) (Minotti et al., 1998; Kotamraju et al., 2002; Kwok and Richardson, 2002). The IRPs are involved in post-translational regulation of various molecules involved in Fe metabolism, including ferritin and the TfR1 (Hentze and Kühn, 1996; Richardson and Ponka, 1997). Several studies in vitro have also assessed the ability of anthracyclines to mobilize Fe from purified ferritin (Demant, 1984; Thomas and Aust, 1986; Gianni and Myers, 1992), Tf (Demant and Norskov-Lauritsen, 1986), and microsomal membranes (Minotti, 1990; Gianni and Myers, 1992). However, it is difficult to determine whether the Fe release observed from isolated and purified ferritin is physiologically relevant.

In the present study, we investigated the effect of anthracyclines on intracellular distribution and trafficking of Fe in neoplastic cells and differentiated cardiomyocytes. These cell types were compared to understand the role of Fe in anthracycline-mediated cardiotoxicity and antitumor activity. We demonstrated in control cells that once Fe is internalized, it is mainly incorporated into ferritin, followed by redistribution to other compartments. Significantly, we showed for the first time that anthracyclines result in marked accumulation of ferritin-Fe because of inhibition of Fe mobilization from the protein. The ability of anthracyclines to inhibit cellular Fe redistribution from ferritin may affect the cytotoxicity of these antitumor agents.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Treatments and Reagents. Buthionine sulfoximine (BSO), catalase, ebselen, ferric ammonium citrate (FAC), horse spleen ferritin, menadione, N-acetylcysteine (NAC), paraquat, and superoxide dismutase (SOD) were obtained from Sigma Chemical Co. (St. Louis, MO). Cisplatin was from Pharmacia and Upjohn (Sydney, Australia). The Mn(III) tetrakis (4-benzoic acid) porphyrin complex (MnTBAP) was obtained from ICN Biomedicals (Aurora, OH). DOX, daunorubicin (DAU), and epirubicin (EPI) were obtained from Pharmacia (Sydney, Australia). DFO was obtained from Novartis Pharmaceutical Co. (Basel, Switzerland). ICRF-187 was purchased from Chiron B.V. (Paasheuvelweg, Amsterdam, The Netherlands). The Fe chelators 2-pyridylcarboxaldehyde isonicotinoyl hydrazone (PCIH) and 2-hydroxy-1-naphthylaldehyde isonicotinoyl hydrazone (also known as 311), were synthesized by standard techniques (Becker and Richardson, 1999). A polyclonal anti-ferritin antibody was obtained from Roche Diagnostics (Indianapolis, IN). All other chemicals were of analytical reagent quality.

Cell Culture. The human SK-Mel-28 melanoma, SK-N-MC neuroepithelioma, and MCF-7 breast cancer cell lines were obtained from the American Type Culture Collection (Manassas, VA). All cell lines were grown in Eagle's modified minimum essential medium (Invitrogen, Mount Waverley, Australia) containing 10% fetal calf serum (Invitrogen), 1% nonessential amino acids (Invitrogen), 100 µg/ml streptomycin (Invitrogen), 100 U/ml penicillin (Invitrogen), and 0.28 µg/ml Fungizone (Squibb Pharmaceuticals, Montréal, ON, Canada). Cells were grown in an incubator (Thermo Forma, Marietta, OH) at 37°C in a humidified atmosphere of 5% CO₂/95% air and subcultured as described previously (Richardson and Baker, 1992). Cellular growth and viability were assessed by phase contrast microscopy and Trypan blue staining.

Preparation of ⁵⁹Fe-Transferrin. Human apotransferrin (Sigma) and rat apotransferrin (kindly provided by Professor E. H. Morgan, Department of Physiology, University of Western Australia, Perth, Australia) were labeled with ⁵⁹Fe (PerkinElmer Life Sciences, Boston, MA) to produce ⁵⁹Fe₂-transferrin (⁵⁹Fe-Tf) as described previously (Richardson and Baker, 1992). Unbound ⁵⁹Fe was removed by exhaustive vacuum dialysis against 0.15 M NaCl buffered to pH 7.4 using 1.4% NaHCO₃. Fully saturated diferric Tf was used in all experiments. Human ⁵⁹Fe-Tf and rat ⁵⁹Fe-Tf were used to label human neoplastic cells and rat cardiomyocytes with ⁵⁹Fe, respectively.

Effect of Anthracyclines on ⁵⁹Fe Efflux from Prelabeled Cells. Iron efflux experiments examining the ability of various agents to mobilize ⁵⁹Fe from cells were performed using established techniques (Richardson and Milnes, 1997). Briefly, cells were prelabeled with ⁵⁹Fe-Tf ([Tf] = 0.75 µM; [Fe] = 1.5 µM) for 3 to 18 h at 37°C. This medium was aspirated and the cell monolayer washed six times with ice-cold Hanks' balanced salt solution (Invitrogen). The cells were then reincubated for 3 to 24 h at 37°C with minimum essential medium in the presence or absence of the agents to be tested. After this incubation, the overlying media containing released ⁵⁹Fe were collected in -counting tubes. The cells were removed from the Petri dishes and placed in a separate set of tubes. Radioactivity was measured in both the cell pellet and supernatant using the -scintillation counter (1282 Compugamma; Amersham Biosciences).

Effect of Anthracyclines on ⁵⁹Fe Uptake from Transferrin by Cells. The ability of various agents to affect cellular ⁵⁹Fe uptake from ⁵⁹Fe-Tf was performed using standard procedures (Richardson and Milnes, 1997). Briefly, cells were incubated with ⁵⁹Fe-Tf (0.75 µM; i.e., [Tf] = 0.75 µM; [Fe] = 1.5 µM) together with the agents of interest for 3 to 24 h at 37°C. This medium was then aspirated and the cell monolayer washed six times with ice-cold Hanks' balanced salt solution. Cells were then harvested on ice using a plastic spatula and placed in -counting tubes. As a measure of cellular density, protein concentrations were assessed using the Bio-Rad protein reagent (Bio-Rad, Hercules, CA). The data are expressed as counts per minute of ⁵⁹Fe/mg of protein. Separate experiments demonstrated that cell number was directly proportional to protein concentration.

Determination of Intracellular Iron Distribution using Native-PAGE-⁵⁹Fe-Autoradiography. Native-PAGE-⁵⁹Fe-autoradiography was performed using established techniques (Richardson et al., 1996; Richardson and Milnes, 1997). Briefly, cells were labeled with ⁵⁹Fe-Tf (0.75 µM) in the presence or absence of anthracyclines and/or other agents and then lysed using 30 µl of ice-cold 1.5% Triton X-100 containing 2 mM phenylmethylsulfonyl fluoride (Sigma), followed by one freeze-thaw cycle. Samples were then vortexed and centrifuged at 14,000 rpm for 45 min at 4°C to separate the stromal-mitochondrial membrane fraction from the cytosol. The protein concentration of the cytosol was determined using the Bio-Rad protein assay. Radioactivity was assessed using the -scintillation counter described above.

Immunoprecipitation of ⁵⁹Fe-Ferritin. To measure the amount of ⁵⁹Fe in ferritin, immunoprecipitation was performed using an anti-human ferritin antibody (Roche Diagnostics) via established procedures (Baker et al., 1985). Briefly, cells were incubated with ⁵⁹Fe-Tf (0.75 µM) for 18 h at 37°C in the presence or absence of anthracyclines (5 µM). Cell lysates were prepared as described for native-PAGE-⁵⁹Fe autoradiography. Samples were incubated with or without the anti-human ferritin antibody (Roche Diagnostics) for 24 h at 37°C, using dilutions from 1:10 to 1:50. Samples were then transferred to 4°C for 24 h to allow precipitation of the antibody-antigen complexes and centrifuged at 14,000 rpm for 1 h at 4°C. The supernatant was removed and the pellet washed twice with ice-cold phosphate-buffered saline. The radioactivity in the pellet was measured using the -scintillation counter described above.

ATP Assay. Cellular ATP levels were assessed using the Sigma diagnostic kit as per the manufacturer's instructions. In these experiments, three 75-cm² flasks of cells were treated with the agents of interest for 18 h at 37°C. Cells were then resuspended in 0.6 ml of phosphate-buffered saline and lysed in an equal volume of 12% trichloroacetic acid. The decrease in the absorbance of NADH at 340 nm was used as a measure of ATP in the sample.

Western Blot Analysis. Western blot analysis was performed essentially as described previously (Kwok and Richardson, 2002). Briefly, cells were lysed using 1.5% Triton X-100 containing complete protease inhibitor (Roche Diagnostics, Mannheim, Germany) and centrifuged at 14,000 rpm for 45 min at 4°C. Protein concentrations of cytoplasmic extracts were then determined using the Bio-Rad protein assay kit (Bio-Rad). Samples (100 µg) containing 20% -mercaptoethanol were loaded onto a SDS-polyacrylamide gel consisting of 4% stacking and 15% resolving gels. After electrophoresis, proteins were transferred onto polyvinylidene difluoride membranes (Amersham Biosciences, Piscataway, NJ) overnight at 4°C.

Glutathione Assay. Intracellular glutathione (GSH) levels were measured using 5',5'-dithiobis-(2-nitrobenzoic acid) (Sigma) as described previously (Sedlack and Lindsay, 1968). After incubation of cells for 6 to 24 h at 37°C with or without anthracyclines, cellular GSH levels were assessed. Cells were lysed and equal volumes of sample and 5% metaphosphoric acid (Merck, Darnstadt, Germany) were mixed to precipitate cellular protein before addition of 200 µM 5',5'-dithiobis-(2-nitrobenzoic acid) and incubation at 37°C for 1 to 2 h. Absorbance readings were performed at 412 nm against a GSH standard curve.

Statistical Analysis. Experimental data were compared using Student's t test. Results were considered statistically significant when p < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Anthracyclines Induce Ferritin-⁵⁹Fe Accumulation in Normal and Neoplastic Cells during ⁵⁹Fe Uptake from ⁵⁹Fe-Transferrin. Considering the high affinity of anthracyclines for Fe (Gianni and Myers, 1992; Gerwirtz, 1999), it was important to understand their effects on intracellular Fe distribution. This was particularly relevant for understanding the mechanisms involved in Fe-mediated cardiotoxicity and antitumor activity of anthracyclines. To examine this, we used the native-PAGE-⁵⁹Fe-autoradiography technique that has been used to assess the intracellular distribution of Fe and the identity of intermediates involved in Fe uptake by cells (Richardson and Milnes, 1997; Watts and Richardson, 2002). In the current study, we have concentrated our efforts on examining the effects of anthracyclines on the distribution of ⁵⁹Fe in cytosolic fraction that constituted the greatest proportion of ⁵⁹Fe in the cell. Indeed, the amount of ⁵⁹Fe present within the stromal-mitochondrial membrane fraction represented ~20 to 30% of the total cellular ⁵⁹Fe.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Anthracyclines result in ferritin-59Fe accumulation during 59Fe uptake from 59Fe- transferrin (59Fe-Tf). A, SK-Mel-28 melanoma cells were incubated with 59Fe-Tf (0.75 µM) for 3 to 24 h at 37°C in the presence or absence of 20 µM DOX. The 59Fe-PAGE autoradiography was then performed (see Materials and Methods for details). Supershift experiments were performed by incubating samples with an anti-ferritin antibody for 1 h at 25°C before 59Fe-PAGE autoradiography. The results are a typical experiment from three performed. B, densitometric analysis of the autoradiograph shown in A. C, cardiomyocyte cultures were incubated with 59Fe-Tf (0.75 µM) in the presence or absence of DOX (5 µM), DAU (5 µM), or EPI (5 µM) for 18 h at 37°C and 59Fe-PAGE autoradiography was performed. Supershift experiments were performed as in A. Results shown are a typical experiment from three performed. D, densitometric analysis of autoradiograph shown in C.

Anthracyclines Inhibit ⁵⁹Fe Mobilization from Ferritin after Prelabeling with ⁵⁹Fe-Tf. Chase experiments were designed to examine the effect of DOX on cellular ⁵⁹Fe distribution after the initial labeling of SK-Mel-28 melanoma cells with ⁵⁹Fe-Tf (0.75 µM) for 3 h at 37°C (Fig. 2). After ⁵⁹Fe labeling, the cells were then washed and reincubated for 0 to 24 h at 37°C in the presence or absence of DOX (20 µM). After 3- and 6-h reincubations of ⁵⁹Fe-prelabeled cells in control media, marked 3- and 4-fold increases in ferritin-⁵⁹Fe levels were observed, respectively, compared with the 0 h control (Fig. 2, A, compare lane 1 with lanes 2 and 4, and B). However, with longer periods of reincubation in control media, ferritin-⁵⁹Fe levels decreased and by 18 to 24 h reached levels comparable with the 0 h time point (Fig. 2, A, compare lane 1 with lanes 6 and 8, and B). In addition, after an 18- or 24-h reincubation in media alone, ⁵⁹Fe was evident in a range of very diffuse bands above ferritin, suggesting redistribution of ferritin-⁵⁹Fe to other cellular compartments in control cells (Fig. 2A, lanes 6 and 8).

View larger version (35K): [in this window] [in a new window]   Fig. 2.   DOX inhibits 59Fe mobilization from ferritin after prelabeling cells with 59Fe-Tf. A, SK-Mel-28 cells were incubated with 59Fe-Tf (0.75 µM) for 3 h at 37°C. Cells were then washed and re-incubated with control media or media containing DOX (20 µM) for 0 to 24 h at 37°C and 59Fe-PAGE autoradiography performed (see Materials and Methods for details). The results are a typical experiment from five performed. B, densitometric analysis of autoradiograph shown in A. C, SK-Mel-28 cells were prelabeled for 3 h at 37°C with 59Fe-Tf (0.75 µM). The cells were then washed and reincubated with control media or media containing DOX (1-20 µM) for 18 h at 37°C and 59Fe-PAGE autoradiography was performed. The results are a typical experiment from three performed. D, densitometric analysis of C. E, a range of neoplastic cell types, including human SK-Mel-28 melanoma cells, human SK-N-MC neuroepithelioma cells, human MCF-7 breast cancer cells, and rat cardiomyocytes were prelabeled with 59Fe-Tf (0.75 µM) for 3 h at 37°C, followed by an 18-h reincubation at 37°C in control media or DOX (5-20 µM). Results shown are a typical experiment from three performed. F, densitometry of autoradiograph shown in E.

Doxorubicin Neither Inhibits ⁵⁹Fe Uptake from Transferrin nor Mobilizes ⁵⁹Fe from Cells. It could be suggested that the ferritin-⁵⁹Fe accumulation observed when cells were coincubated with ⁵⁹Fe-Tf and DOX (Fig. 1, A-D) was caused by an increase in cellular ⁵⁹Fe uptake. To examine this possibility, SK-Mel-28 cells were labeled with ⁵⁹Fe-Tf (0.75 µM) for 3 or 24 h at 37°C in the presence or absence of DOX (5-20 µM) (Fig. 3A). Cells were subsequently washed and lysed, and total cellular ⁵⁹Fe content was assessed. The Fe chelators DFO (0.5 mM) and 311 (50 µM) were used as positive controls and significantly (p < 0.0001) decreased ⁵⁹Fe uptake into cells after both 3- and 24-h incubations with ⁵⁹Fe-Tf (Fig. 3A), as documented previously (Richardson and Milnes, 1997; Darnell and Richardson, 1999). However, in these studies, there was no increase in ⁵⁹Fe uptake from ⁵⁹Fe-Tf in the presence of DOX (5-20 µM), indicating that the accumulation of ⁵⁹Fe in ferritin was not caused by enhanced ⁵⁹Fe uptake (Fig. 3A).

View larger version (38K): [in this window] [in a new window]   Fig. 3.   DOX affects neither total cellular Fe content nor total cellular ferritin protein levels. A, SK-Mel-28 cells were incubated with 59Fe-Tf (0.75 µM) in the presence or absence of DFO (0.5 mM), 311 (50 µM), or DOX (5-20 µM) for 3 h or 24 h at 37°C. Cells were washed and total cellular 59Fe content was assessed (see Materials and Methods for details). Results are mean ± S.D. (three determinations) in a typical experiment from three experiments performed. B, SK-Mel-28 cells were incubated with 59Fe-Tf (0.75 µM) for 3 h at 37°C. Cells were then washed and reincubated with control media, DFO (0.5 mM), 311 (50 µM), or DOX (5-20 µM) for 3 or 24 h at 37°C and the release of cellular 59Fe assessed (see Materials and Methods for details). Results are mean ± S.D. (three determinations) in a typical experiment from three experiments performed. C, SK-Mel-28 cells were incubated with control media, DFO (100 µM), FAC (100 µg/ml), or DOX (5-20 µM) for 18 h at 37°C. Western blot analyses were performed as described under Materials and Methods. Results shown are a typical experiment from three performed. D, densitometry of the results in C, normalized to -actin.

Accumulation of Ferritin-⁵⁹Fe Is Not Caused by Changes in Cellular Ferritin Protein Levels. To determine whether the increase in ferritin-⁵⁹Fe levels observed upon incubation with DOX was caused by an increase in total ferritin synthesis in DOX-treated cells, Western analysis was performed (Fig. 3C). Briefly, cells were incubated with control media, DFO (100 µM), FAC (100 µg/ml), or DOX (5-20 µM) for 18 h at 37°C. The Fe chelator DFO depletes cellular Fe levels and results in a decrease in ferritin synthesis (Hentze and Kühn, 1996) and was used as a negative control. Conversely, FAC donates Fe to cells increasing the synthesis of ferritin (Wang and Pantopoulos, 2002) and was an appropriate positive control.

The Cytotoxic Agent, Cisplatin, was Far Less Effective than DOX at Increasing Ferritin-⁵⁹Fe Accumulation. The accumulation of ferritin-⁵⁹Fe induced in DOX-treated cells could be caused by an indirect effect of the cytotoxicity of this agent. To examine this possibility, the effect of anthracyclines were compared with the cytotoxic agent cisplatin (Fig. 4, A and B). The SK-Mel-28 cells were labeled with ⁵⁹Fe-Tf (0.75 µM) for 3 h at 37°C, washed, and then re-incubated for 18 h with control media, DOX, DAU, and EPI (20 µM) or cisplatin (10 or 20 µM). As expected, all three anthracyclines resulted in marked accumulation of ⁵⁹Fe-ferritin (Fig. 4, A, compare lane 1 with lanes 2 to 4, and B). In contrast, cisplatin had no effect at 10 µM (Fig. 4, A, compare lanes 1 and 5, and B) and at 20 µM only slightly increased ferritin-⁵⁹Fe incorporation (Fig. 4, A, compare lanes 1 and 6, and B).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Other cytotoxic agents are far less efficient than anthracyclines at inducing ferritin-59Fe accumulation. A, SK-Mel-28 cells were incubated with 59Fe-Tf (0.75 µM) for 3 h at 37°C. Cells were then re-incubated with control media, DOX (20 µM), DAU (20 µM), EPI (20 µM), or cisplatin (10 or 20 µM) for 18 h at 37°C and 59Fe-PAGE autoradiography was performed (see Materials and Methods for details). The results are a typical experiment from two performed. B, densitometric analysis of the autoradiograph shown in A.

Other Redox-Cycling Drugs also Caused a Marked Accumulation of ⁵⁹Fe in Ferritin. A major mechanism of anthracycline toxicity is believed to involve redox cycling and the production of free radicals (Gianni and Myers, 1992). To determine whether other redox cycling agents exerted the same effect as anthracyclines on Fe accumulation in ferritin, experiments were performed using the classic redox cycling agent menadione (Gutteridge and Halliwell, 1989). The SK-Mel-28 cells were incubated with ⁵⁹Fe-Tf (0.75 µM) for 18 h in the presence or absence of menadione (1 to 10 µM) or DOX (5 µM) and ferritin-⁵⁹Fe accumulation was assessed (Fig. 5, A and B).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   The redox-cycling agents DOX and menadione induced ferritin-59Fe accumulation. A, SK-Mel-28 cells were incubated with 59Fe-Tf (0.75 µM) for 18 h at 37°C in the presence or absence of DOX (5 µM) or menadione (1-10 µM). 59Fe-PAGE autoradiography was then performed (see Materials and Methods for details). The results shown are a typical experiment from 3 performed. B, densitometric analysis of the gel is shown in A.

View larger version (41K): [in this window] [in a new window]   Fig. 6.   The ferritin-59Fe accumulation induced by the redox-cycling agents DOX and menadione can be partially inhibited by the cell-permeable superoxide scavenger, MnTBAP. A, SK-Mel-28 cells were incubated with 59Fe-Tf (0.75 µM) for 18 h at 37°C with either control media, DOX (5 µM), or menadione (10 µM) in the presence or absence of the free radical scavengers, catalase (1000 U/ml), SOD (1000 U/ml), ebselen (15 µM), MnTBAP (200 µM), or the combination of ebselen (15 µM) and MnTBAP (200 µM) for 18 h at 37°C. 59Fe-PAGE autoradiography was then performed (see Materials and Methods for details). The results shown are a typical experiment from three performed. B, densitometric analysis of the autoradiograph shown in A.

The Effects of Chelators on Preventing Doxorubicin-Mediated Ferritin-⁵⁹Fe Accumulation by Cells during Fe Uptake. The DOX-induced accumulation of ferritin-⁵⁹Fe and the inhibition of ferritin-⁵⁹Fe mobilization to other vital Fe-containing proteins (e.g., ribonucleotide reductase) could potentially result in their decreased enzymatic activity and lead to cytostasis or cytotoxicity via Fe deprivation. To determine whether these effects of DOX could be prevented, we examined the effect of a number of Fe chelators (Fig. 7, A and B). In these studies, we assessed the clinically used Fe chelators DFO (50 µM) and ICRF-187 (500 µM) and also the active membrane-permeable aroylhydrazone chelator PCIH (50 µM), which has been demonstrated to have potential for the treatment of Fe-overload diseases (Becker and Richardson, 1999) (Fig. 7, A and B).

View larger version (28K): [in this window] [in a new window]   Fig. 7.   Some iron chelators are effective at inhibiting DOX-mediated ferritin-59Fe accumulation during cellular 59Fe uptake from 59Fe-transferrin (59Fe-Tf). A, cardiomyocyte cultures were incubated with 59Fe-Tf (0.75 µM) for 18 h at 37°C with control media, DFO (50 µM), PCIH (50 µM), or ICRF-187 (500 µM) in the presence or absence of DOX (5 µM) and 59Fe-PAGE autoradiography performed (see Materials and Methods for details). The results shown are a typical experiment from 3 performed. B, densitometric analysis of the results in shown in A.

Iron Chelators Only Partially Mobilize ⁵⁹Fe from Ferritin after its Accumulation in DOX-Treated Cells. Although DFO and, to a lesser extent, PCIH, seemed to be able to prevent DOX-mediated ferritin-⁵⁹Fe accumulation during the ⁵⁹Fe uptake process (see Fig. 7 above), it was important to examine whether these chelators were effective at mobilizing ⁵⁹Fe already bound within ferritin as a result of DOX treatment (Fig. 8).

View larger version (32K): [in this window] [in a new window]   Fig. 8.   Iron chelators can partially mobilize 59Fe from ferritin in DOX-treated cells. A, cardiomyocyte cultures were prelabeled for 18 h at 37°C with 59Fe-Tf (0.75 µM) in the presence or absence of DOX (5 µM). Cells were then washed and reincubated for 24 h at 37°C in control media or media containing the chelators DFO (50 µM), PCIH (50 µM), and ICRF-187 (500 µM), and 59Fe-PAGE was autoradiography performed (see Materials and Methods for details). The results shown are a typical experiment from three performed. B, densitometric analysis of the results in A. C, cardiomyocyte cultures were treated as described in A, and cellular 59Fe released into the overlying media was assessed after the 24 h reincubation in the presence or absence of chelators. Results are mean ± S.D. (three experiments).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Anthracyclines are potent antineoplastic agents that avidly bind Fe; this property is thought to play a role in their cardiotoxic effects (Gianni and Myers, 1992). Despite this, there have been very few studies on the effects of these drugs on Fe metabolism in neoplastic cells or cardiomyocytes. The most comprehensive experiments to date have examined the IRPs that are crucial for cellular Fe homeostasis (Minotti et al., 1998; Minotti et al., 2001; Kotamraju et al., 2002; Kwok and Richardson, 2002). In addition, very little is understood concerning the actual mechanisms involved in intracellular Fe trafficking and ferritin-Fe release and the effect of anthracyclines on these processes.

The current study is the first to investigate the effects of anthracyclines on intracellular Fe trafficking and distribution in both tumor cells and differentiated cardiomyocytes. We demonstrated that during and after cellular ⁵⁹Fe uptake from ⁵⁹Fe-Tf in control cells, ⁵⁹Fe was mainly incorporated into ferritin during the first 6 h (Figs. 1 and 2). Thereafter, ferritin-⁵⁹Fe levels dramatically decreased, the ⁵⁹Fe being redistributed to other cellular compartments (Figs. 1 and 2). However, the presence of anthracyclines inhibited the ⁵⁹Fe mobilization process from ferritin, resulting in a marked accumulation of ⁵⁹Fe within this protein (Figs. 1 and 2). Cellular ⁵⁹Fe uptake and efflux experiments demonstrated that the total cellular Fe content was not affected by anthracyclines (Fig. 3, A and B). These observations indicated that the changes in ferritin-⁵⁹Fe content were not caused by altered ⁵⁹Fe uptake or release from cells. The inhibition of ⁵⁹Fe redistribution to other cellular compartments by anthracyclines may have detrimental effects on vital Fe-dependent processes and may, at least in part, be responsible for its cytotoxicity.

Our observations are the first to demonstrate the effect of anthracyclines on inhibiting ferritin-Fe redistribution in an intact cellular system. In contrast to our data, previous studies using isolated and purified ferritin have demonstrated that anthracyclines induce Fe release from this protein (Demant, 1984; Thomas and Aust, 1986). However, it is likely that the in vivo mechanisms involved in mobilizing Fe from ferritin are quite different from that observed with the purified protein. Once isolated from its cellular environment, the structure and function of the molecule could be altered. In addition, Fe may only be mobilized within certain cellular compartments such as the lysosome (Radisky and Kaplan, 1998), and it is probable that interaction of ferritin with other molecules may be essential for Fe release. Furthermore, in studies using purified ferritin, it was difficult to determine whether Fe was being mobilized from the ferritin core or from Fe nonspecifically bound to the outer surface of the protein (Ponka and Richardson, 1997).

Although many mechanisms have been proposed for in vitro ferritin-Fe mobilization (Dognin and Crichton, 1975; Harrison and Arosio, 1996), currently there are few data describing the process of Fe release from ferritin within cells. Studies using isolated ferritins in vitro have shown that when conserved residues in the 3-fold axes are mutated, ferritin-Fe release is markedly enhanced (Takagi et al., 1998; Jin et al., 2001). Similarly, it can be speculated that residues could be altered to inhibit Fe release. Considering this, a major mechanism of anthracycline toxicity is through their ability to redox cycle, generating free radicals that result in lipid peroxidation and protein oxidation (Gianni and Myers, 1992). Oxidation of the ferritin protein could theoretically alter the Fe release process. In our current investigation, other redox-cycling drugs, including paraquat and menadione, also resulted in ferritin-Fe accumulation (Figs. 5 and 6). In addition, experiments with the intracellular superoxide scavenger, MnTBAP, suggested that superoxide may play a role in the accumulation of ferritin-Fe induced by these redox-cycling drugs (Fig. 6, A and B). Further studies are underway to determine the precise molecular mechanism involved.

At present, it is unclear whether the accumulation of ferritin-Fe after treatment with DOX is involved in its cardiotoxic or antineoplastic effects. However, the release of stored Fe from ferritin is probably critical for important physiological processes (e.g., ribonucleotide reductase activity that is necessary for DNA synthesis) and the ability of anthracyclines to prevent this may be a mechanism of cytotoxicity. Alternatively, it could also be argued that the increased storage of Fe within ferritin may be a protective cellular response that helps to prevent the interaction of Fe with anthracyclines, inhibiting its ability to redox-cycle and generate toxic free radicals.

We demonstrated in our studies that Fe chelators such as DFO and PCIH could inhibit the DOX-mediated accumulation of ferritin Fe (Fig. 7) and slightly mobilize Fe once it had been incorporated into ferritin after incubation with DOX (Fig. 8). It is important to note that DOX has been shown to be more cytotoxic in Fe-overloaded cardiomyocytes compared with controls (Hershko et al., 1993), and the ability of DOX to accumulate Fe in ferritin (Figs. 1 and 2) and prevent its redistribution may potentiate its cardiotoxicity. In addition, DFO can reduce toxicity in Fe-overloaded cardiomyocytes (Hershko et al., 1993). However, it remains to be established whether the ability of chelators to prevent ferritin-Fe accumulation (Figs. 7 and 8) is beneficial in terms of reducing DOX cardiotoxicity. It is interesting that some Fe chelators are potent antitumor agents (Richardson and Milnes, 1997) and the combination of DOX and DFO is more cytotoxic than either alone (Blatt and Huntley, 1989). Therefore, suitable chelators may both prevent DOX-mediated cardiotoxicity and potentiate its antitumor activity.

The uptake of ⁵⁹Fe into ferritin of the differentiated cardiomyocytes resulted in a diffuse band that seemed to consist of two major components (e.g., see Figs. 1C, 2E, 7A, and 8A), whereas ⁵⁹Fe uptake into ferritin by all the tumor cell lines resulted in a single well-defined band (e.g., Figs. 1A and 2, A, C, and E). This observation could be indicative of a difference in the metabolism of ferritin in the cardiomyocytes compared with the neoplastic cells (Andrews et al., 1987; Campbell et al., 1993). Although the role of ferritin as an Fe-storage molecule is well defined, there have been few studies in living cells examining the Fe uptake and release process from this protein. Further investigation is required to fully characterize the significance of the difference observed in ferritin-⁵⁹Fe incorporation between cardiomyocytes and tumor cell types.

In our previous study, we examined the effect of anthracyclines on IRP-RNA-binding activity in the same cell types assessed in this investigation, namely, SK-Mel-28 melanoma cells and cardiomyocytes (Kwok and Richardson, 2002). These experiments were important because IRPs are major regulators of intracellular Fe metabolism (Hentze and Kühn, 1996). The ferritin mRNA contains a hairpin loop structure in its 5'-untranslated region called an iron-responsive element (Hentze and Kühn, 1996). In ferritin mRNA, binding of IRP to the iron-responsive element inhibits translation, thereby decreasing Fe storage (Hentze and Kühn, 1996).

We showed in a prior investigation that the effect of anthracyclines on IRP-RNA-binding was complex, with an initial and delayed response being found (Kwok and Richardson, 2002). The initial response to DOX resulted in a decrease in IRP-RNA-binding activity observed after a 6-h incubation (Kwok and Richardson, 2002), which should theoretically increase ferritin protein synthesis. Surprisingly, this was not observed in the current study examining ferritin protein levels as a function of time. The reason for this remains unclear at present and is a subject for further investigation. The delayed effect of DOX on IRP-RNA-binding was observed after an 18- to 24-h incubation in which the initial decrease was followed by restoration of IRP-RNA-binding activity to control levels (Kwok and Richardson, 2002). This was in agreement with the observations of the present work, in which ferritin protein levels in the presence of DOX were comparable with the untreated controls (Fig. 3C). However, ⁵⁹Fe incorporation into ferritin continued to increase (Fig. 1, A and B). Again, these results suggest that the elevated ⁵⁹Fe levels in ferritin after incubation with DOX was independent of de novo ferritin synthesis.

In conclusion, this is the first study to demonstrate that anthracyclines induce ferritin-Fe accumulation by inhibiting Fe mobilization from the protein. To date, very little is understood concerning the mechanism of ferritin Fe-release, and this investigation represents the first examination of this process using intact cells and the physiological Fe donor Tf. Our experiments also showed that ferritin is a dynamic molecule that can take up and release Fe depending on cellular requirements. The inhibition of Fe redistribution to other cellular compartments by anthracyclines may have consequences in terms of the cytotoxic effects of these drugs.