Biochimica et Biophysica Acta 1832 (2013) 293–301

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbad is

Insulin promotes iron uptake in human hepatic cell by regulating transferrinreceptor-1 transcription mediated by hypoxia inducible factor-1

Sudipta Biswas 1, Nisha Tapryal, Reshmi Mukherjee, Rajiv Kumar, Chinmay K. Mukhopadhyay ⁎Special Centre for Molecular Medicine, Jawaharlal Nehru University, New Delhi-110 067, India

⁎ Corresponding author. Tel.: +91 11 2674 1932; faxE-mail address: ckm2300@mail.jnu.ac.in (C.K. Mukh

1 Present address: Department of Molecular Cardiolotute, Cleveland Clinic Foundation, Ohio-44195, USA.

0925-4439/$ – see front matter © 2012 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.bbadis.2012.11.003

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 July 2012Received in revised form 24 October 2012Accepted 6 November 2012Available online 13 November 2012

Keywords:InsulinIron uptakeTransferrin receptor-1TranscriptionHypoxia inducible factor-1Hepatic iron overload

Hepatic iron is known to regulate insulin signaling pathways and to influence insulin sensitivity in insulin re-sistance (IR) patients. However, the role of insulin on hepatic iron homeostasis remains unexplored. Here, wereport that insulin promotes transferrin-bound iron uptake but shows no influence on non transferrin-boundiron uptake in human hepatic HepG2 cells. As a mechanism we detected increased transferrin receptor-1(TfR1) expression both at protein and mRNA levels. Unaltered stability of protein and transcript of TfR1suggested the regulation at transcriptional level that was confirmed by promoter activity. Involvement oftranscription factor hypoxia inducible factor-1 (HIF-1) was shown by mutational analyses of the TfR1 pro-moter region and by electrophoretic mobility shift assay. When HepG2 cells were transfected with specificsiRNA targeted to 3′UTR of HIF-1α, the regulatory subunit of HIF-1; insulin-induced TfR1 expression andiron uptake were inhibited. Transfection of cDNA expressing stable form of HIF-1α reversed the increasedTfR1 expression and iron uptake. These results suggest a novel role of insulin in hepatic iron uptake by aHIF-1 dependent transcriptional regulation of TfR1.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Iron is one of the most essential micronutrients required for thebody to maintain energy homeostasis [1]. Due to its redox nature itmay participate in many enzymic electron transfer reactions. At thesame time, it may generate potentially dangerous hydroxyl radicalin the presence of reactive oxygen species (ROS) [2]. Thus, regulationsof iron homeostasis genes are highly coordinated mostly at thepost-transcriptional level [3]. Dysregulation of iron homeostasisgenes may lead to iron deficiency or overload conditions contributingto several pathological conditions including iron overload relatedhepatic cancer [4]. Conditions like alcohol consumption, excess useof xenobiotics and hepatitis C infection contribute to hepatic ironoverload (HIO) [4]. Insulin resistance (IR) is also reported to be oneof the major contributors of HIO [4–7].

Insulin plays a critical role in energy metabolism and glucosehomeostasis [1,8]. After secretion from pancreatic beta cells into theportal circulation, insulin primes the liver for rapid alterations inhepatic carbohydrate and lipid metabolism [9]. Absence of insulinin case of insulin-dependent diabetes mellitus (IDDM) or cellularnon-responsiveness to insulin in case of non-insulin-dependentdiabetes mellitus (NIDDM) leads to severe pathophysiological condi-tions [9,10]. One of the body's responses in the early stage of insulin

: +91 11 2674 1781.opadhyay).gy, The Lerner Research Insti-

rights reserved.

resistance (IR) is to increase the synthesis and secretion of insulinfrom pancreatic beta cells causing hyperinsulinemia [9–11]. IR andcoupled hyperinsulinemia lead to a variety of abnormalities, includ-ing elevated triglycerides, low levels of HDL, enhanced secretion ofVLDL, increased vascular resistance, changes of peripheral bloodflow and weight gain [12]. In some cases, a clinical-pathological syn-drome of hepatic iron overload (HIO) in which disproportionatelyraised serum and liver iron concentrations were reported [4–7]. Inter-estingly, hepatic iron store was reported to regulate insulin signalingpathways and to influence insulin sensitivity in IR patients suggestingthe existence of a cross talk between insulin and iron metabolism[13,14]. Iron depletion by desferrioxamine (DFO) caused an increasein insulin signaling in hepatoma cells as well as in rat liver [15]. Irondepletion by phlebotomy also reported to improve insulin resistancein patients with non-alcoholic fatty liver disease (NAFLD) andhyperferritinemia [16,17]. Even insulin resistance related HIO wasimproved by venesections and iron restricted diet [18,19] thatstrengthened the concept of cross talk between insulin and iron me-tabolism. Although these findings strongly suggest the role of ironon insulin-induced cell metabolism the role of insulin on ironmetabolism is not clear so far [19].

Hepatocyte plays an important role for maintaining body ironhomeostasis and takes up both transferrin-bound iron (TBI) andnon transferrin-bound iron (NTBI) [20]. TfR1 binds iron saturatedtransferrin (Tf) for cellular iron uptake in almost all cell types includ-ing hepatocyte [21,22]; whereas, the mechanism of iron uptake fromNTBI is far less understood. Involvement of Zip 14 in iron uptake fromNTBI was reported recently [20,23]. Iron uptake from both TBI and

294 S. Biswas et al. / Biochimica et Biophysica Acta 1832 (2013) 293–301

NTBI is decreased by hereditary hemochromatosis gene product HFE[20]. Mutations in HFE affecting its function thus cause increased he-patic iron uptake as detected in hereditary hemochromatosis [19,24].However, several attempts to link HFE mutations responsible forIR-associated hepatic iron overload have failed so far [4,7,19].

Inter-organ transport and uptake of nonheme iron is largelyperformed by the Tf/TfR1 system [3,21,22]. During cellular iron deple-tion, TfR1 is regulated mainly by post-transcriptional mRNA stabilitymechanism mediated by increased binding of iron regulatory pro-teins (IRP1/IRP2) to iron response elements (IRE) present in theTfR1 3′UTR [3,21,22]. Iron depletion or hypoxia also regulates TfR1by activating hypoxia inducible factor-1 (HIF-1) that binds to hypoxiaresponsive element (HRE) present in the promoter of TfR1 [25,26].HIF-1α, the regulatory subunit of HIF-1 is stabilized during oxygenor iron depletion by a post-translational mechanism and transportedto the nucleus to bind the constitutional subunit HIF-1β to activateHIF-1. HIF-1 then binds to HRE (5′-RCGTG-3′) of targeted genes. Innormal condition HIF-1α is highly unstable due to hydroxylation oftwo proline residues (Pro402 and Pro564) by prolyl hydroxylases(PHDs) that promotes ubiquitination and subsequent degradationby proteasomal pathway [27–29]. Hypoxia leads to HIF-1α stabiliza-tion due to inhibition of prolyl hydroxylses and subsequent decreasein HIF-1α ubiquitination and degradation. Insulin induces severalgenes including glucose transporters (Glut1 and Glut-3), severalglycolytic enzymes, erythropoietin and vascular endothelial growthfactor (VEGF) [30] by activating HIF-1 but its role on regulating TfR1or iron homeostasis has not been reported so far. Insulin also inducesnuclear localization of Sp1 in the skeletal muscle cell [31]. Incidental-ly, Sp1 binds to GC-rich region of TfR1 promoter and is responsible forbasal expression of TfR1 [32]. Considering all these, we hypothesizethat insulin has a direct role in hepatic iron uptake by regulatingTfR1 expression. In the present study, we revealed a novel role ofinsulin in transferrin-bound iron uptake in human hepatic cells medi-ated by HIF-1 induced transcription of TfR1.

2. Materials and methods

2.1. Materials

Human recombinant insulin was obtained from Invitrogen. Mono-clonal antibodies to TfR1 and HIF-1α were obtained from Invitrogenand Abcam respectively. Actin antibody was from Santa Cruz Biotech-nology. All cell culture reagents and other reagents were obtainedfrom Sigma-Aldrich, if not mentioned otherwise.

2.2. Cell culture

Human hepatocarcinoma HepG2 cells (ATCC) were maintained inDulbecco's Modified Eagle's Medium (DMEM), supplemented with10% heat-inactivated fetal bovine serum (Invitrogen), 100 units/mlpenicillin, 100 μg/ml streptomycin, and 2 mM L-glutamine in a hu-midified atmosphere containing 5% CO2 at 37 °C (Forma-Scientific).Human hepatocytes were obtained from Invitrogen (HMCS 10) andcultured in Williams' Medium E as per Company's protocol. Cells at50–60% confluence were treated with appropriate concentrations ofinsulin after serum deprivation as described earlier [33,34].

2.3. Western blot analysis

After the treatment as indicated in different experiments, cellswere lysed in buffer containing 50 mM HEPES, pH 7.5, 150 mMNaCl, 1% Triton X-100, 100 mM sodium fluoride, 1 mM EGTA, 1 mMEDTA, 2 mM sodium vanadate, 10 mM sodium pyrophosphate,1 mM phenylmethylsulfonyl fluoride and a protease inhibitor mix-ture (Roche Applied Science). Then cell lysates were centrifuged at13,000 ×g for 10 min; protein (40 μg) from the cleared supernatant

was resolved by either by 7.5% SDS- polyacrylamide gel electrophore-sis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF)membrane. PVDF membranes were subjected to immunoblottingwith antibody to TfR1 (1:1000) or actin (1: 1000). For immunoblotanalysis of HIF-1α, nuclear extracts (30 μg) were run into 7.5%SDS-PAGE, transferred into PVDFmembranes and subjected to immu-noblotting with monoclonal antibody to HIF-1α (1:1000). Followingincubation with horseradish peroxidase-conjugated secondary anti-body, proteins were visualized by enhanced chemiluminescence(ECL), according to the manufacturer's protocol (Amersham Biosci-ences). To quantitate chemiluminescence signals ImageJ program(http://rsbweb.nih.gov/ij/) was used as described detail in thesoftware.

2.4. Northern blot analysis

RNA was isolated from serum deprived HepG2 cells using TriPurereagent (Roche Applied Science). Total RNA (20 μg) was denatured informamide/formaldehyde, electrophoresed through 1.2% agarose gelcontaining 6% formaldehyde, and then blotted onto nylon mem-branes (Schleicher & Schuell). After cross-linking by ultraviolet irradi-ation, membranes were hybridized to a fragment of human TfR1(generated by RT-PCR) or HIF-1α cDNA (Novus biologicals) and la-beled by random priming with [α-32P]dCTP using a New EnglandBiolab Kit. Quantitation of the blots was performed by using ImageJprogram.

2.5. Reverse transcriptase polymerase chain reaction

Total RNA (4 μg) was used for each sample to performsemi-quantitative RT-PCR by using one tube RT-PCR kit (Roche).The following primers were used for hTfR1 gene: Forward: 5′ GCTTGA AGA TCG TTA GTA TCT AAC 3′ and Reverse: 5′ CTA ACA CAGTAA AGG TCA TGC AAG 3′; human β-actin: Forward, 5′ GAC ATGGAG AAA ATC TGG CAC CAC 3′ and reverse, 5′ GAT GGA GTT GAAGGT AGT TTC GTG 3′. TfR1 mRNA stability assay was performed as de-scribed earlier [35,36]. For quantitative RT-PCR, TfR1 and controlβ-actin mRNA transcripts were measured by a two-step RT-PCRusing the same set of primers with the Applied Biosystems 7500real time PCR system. cDNA synthesis was performed with a reversetranscription reagents kit (Applied Biosystems). Transcripts were am-plified in duplicates with 100 pmol of specific sense and antisenseprimers in a 50 μl reaction mixture containing 25 μl of 2× SYBRGreen PCR Master Mix (Applied Biosystems) that underwent 95 °Cactivation for 10 min, followed by 40 repetitions of denaturation at95 °C for15 s, annealing at 56 °C for 30 s, and 1 min extension at60 °C. TfR1/β-actin ratios were calculated using 7500 system SDSSoftware (Applied Biosystems).

2.6. Construction of FLAG-tagged TfR1 cDNA and verification of TfR1protein stability by insulin

The human TfR1 cDNA was subcloned from pCD-TfR1 (ATCC) toNotI and BamHI sites of p3XFLAG-CMV-7.1 vector (Sigma). Theclone was verified by sequence analysis. The construct wastransfected in HepG2 cells along with pGL3-Control (to verify trans-fection efficiency) and after insulin treatment Western blot analyseswere performed in the cell extract using TfR1, FLAG, luciferase andactin antibodies.

2.7. Cloning of TfR1 promoter and construction of mutants

Human TfR1 promoter region (from−114 to +14 of transcriptionstart site) was amplified by PCR from genomic DNA and was clonedbetween SacI and XhoI restriction sites in pGL3 basic vector(Promega) behind the luciferase. Sequential deletion mutations

295S. Biswas et al. / Biochimica et Biophysica Acta 1832 (2013) 293–301

from 5′ end and specific response element were mutated usingsite-directed mutagenesis. The following forward primers were usedfor different deletion mutation products of TfR1 promoter: (−114to +14) — 5′ ATA CAT GAG CTC GAT CTG TCA GAG 3′; (−95 to+14) — 5′ ATA CAT GAG CTC CGA GCG TAC GTG 3′; (−82 to +14)— 5′ ATA CAT GAG CTC CTC AGG AAG TGA 3′ (underline denotesSacI). In all cases 5′ ATA TTG CTC GAG TAT CCC GAC GCT 3′ (underlinedenotes XhoI) was used as reverse primer. Site directed mutation ofHRE was performed using megaprimer method [33,34] using 5′ CCTGAG GCt ttt ACG CTC GCG 3′ as primer and −114 to +14 constructas template. This fragment was further cloned into pGL3 basic vectorbetween SacI and XhoI restriction sites.

2.8. Transient transfection of cells and reporter gene assays

To determine transcriptional efficiency of chimeric TfR1 promoterconstructs, HepG2 cells were transiently transfected for 6 h with a re-porter plasmid (1.5 μg) using Lipofectamin 2000 (Invitrogen). Tomonitor transfection efficiency, a reporter gene construct (0.25 μg)containing β-galactosidase behind an SV40 promoter was co-transfected as described earlier [34]. Transfected cells were allowedto recover for 6 h in DMEM with 10% fetal bovine serum and afterserum deprivation were incubated with insulin for 16 h. Luciferase(Promega) and β-galactosidase (Invitrogen) activities in cell extractswere determined as described in the company's protocol.

2.9. Preparation of nuclear extracts

Nuclear extract was prepared from HepG2 cells as described before[34,37]. Briefly, 1×108 cells were washed with ice-cold phosphate-buffered saline and then with a solution containing 10 mM Tris–HCl,pH 7.8, 1.5 mM MgCl2, and 10 mM KCl, supplemented with aprotease inhibitor mixture containing 0.5 mM dithiothreitol, 0.4 mMphenylmethyl-sulfonyl fluoride and 2 μg/ml each of leupeptin, pepstatinand aprotinin. After incubation on ice for 10 min, cells were lysed by 10strokes with a Dounce homogenizer and the nuclei were pelleted. Thepellet was resuspended in a solution containing 420 mM KCl, 20 mMTris–HCl, pH 7.8, 1.5 mM MgCl2 and 20% glycerol, supplemented withprotease inhibitor mixture described above and incubated at 4 °C withgentle agitation. The nuclear extract was centrifuged at 10,000×g for10 min, and the supernatant was dialyzed twice against a solution of20 mM Tris–HCl, pH 7.8, 100 mM KCl, 0.2 mM EDTA, and 20% glycerol.Protein concentration was determined using Bio-Rad reagent withbovine serum albumin as standard.

2.10. Electrophoretic mobility shift analysis (EMSA)

Sequences of sense strands for oligonucleotide probes used forEMSA were as follows: 5′ AGC GTA CGT GCC TCA GGA AGT GAC 3′(W — wild HRE) and 5′-AGC GTA AAA GCC TCA GGA AGT GAC 3′(M — mutant HRE). The sense and antisense strands were annealed,gel-purified, and end-labeled with [γ32P]ATP using T4-polynucleotidekinase (Promega). Unincorporated nucleotide was removed by gelfiltration using G-25 Sephadex columns (Quick SpinTM TE, RocheApplied Science). To measure DNA-protein interaction, 1×105 cpm ofoligonucleotide probe was incubated with 5 μg of nuclear extract and0.5 μg of sonicated, denatured salmon sperm DNA (Invitrogen) in10 mM Tris–HCl, pH 7.8, 50 mM NaCl, 1 mM MgCl2, 1 mM EDTA,5 mM dithiothreitol and 5% glycerol for 20 min at 4 °C in a total vol-ume of 20 μl. The reaction mixture was subjected to electrophoresis(200 V in 0.3× Tris-buffered EDTA solution at 4 °C) using 5%nondenaturing polyacrylamide gel. Dried gels were subjected to auto-radiography up to 24 h.

2.11. Knocking down of HIF-1α by siRNA

Custom made siRNA for HIF-1α, control siRNA and transfection re-agents were purchased (Sigma). HepG2 cells were transfected with theHIF-1α siRNA targeted to its 3′UTR. Target sequence 5′ AAT GCC AGATCA CAG CAC ATT 3′ (from 3167 to 3187 nucleotide of human HIF-1αmRNA) was designed by siRNA target finder site of Ambion (http://www.ambion.com/techlib/misc/siRNA_finder.html). HIF-1α siRNA(80 μg) was transfected in 12-well plates using N-TER™ nanoparticlesiRNA Transfection System (Sigma). In some cases, after 24 h of transfec-tion and recovery, cells were retransfected with pcDNA3-HIF-1α (P/A)or only pcDNA3 (4 μg/well) using Oligofectamine (Invitrogen). HIF-1αcDNA (P/A) (corresponding only to its open reading frame)wasmutatedat 402 and 564 from proline to alanine (kind gift from Dr. RituKulshreshtha, Indian Institute of Technology, New Delhi, India) to avoidproline hydroxylation and subsequent ubiquitination and proteasomaldegradation. After transfection serum deprived cells were treated withinsulin and Western blot analyses were performed for HIF-1α, TfR1and actin.

2.12. Iron uptake

2.12.1. 55Fe loading to Apotransferrin55FeCl3 (NEN) and nitrilotriacetate (NTA) were mixed in 1:50

molar ratio in 300 μl of carbonate buffer (10 mM NaHCO3, 250 mMTris–Cl pH 7.4). After mixing thoroughly it was kept at room temper-ature for 10–15 min. Apo-transferrin (1.5 μg dissolved in 100 μlcarbonate buffer) was incubated with 55Fe-NTA in 1:2 ratios atroom temperature for 1 h and then kept at 4 °C for overnight [38].The complex was passed through G-50 sephadex column to purifythe 55Fe-Tf complex from the unbound 55Fe-NTA.

2.12.2. 55Fe-Tf uptake assayCells were grown in 12 well plates. Sub-confluent cells were treat-

ed with insulin or desferrioxamine (DFO, 150 μM) after serum depri-vation. After requisite time, cells were incubated with 55Fe-Tf for30 min at room temperature (25 °C) in serum free media. Thencells were washed twice with wash-buffer (150 mM NaCl and100 μM EDTA) to remove the externally bound 55Fe-Tf complex.Cells were lysed with 200 μl of lysis buffer (0.1% NaOH and 0.1% Tri-ton X-100) at 4 °C for 30 min [35] and lysate was used for countingradioactivity by Liquid Scintillation Counter. Iron uptake was deter-mined as pmol/106 cell/30 min.

2.12.3. Non transferrin-bound iron (NTBI) uptake assayA solution of 55Fe-NTA was prepared by mixing 55FeCl3 (NEN)

with a 5-fold molar excess of disodium salt of NTA [39]. To measureiron uptake, 5×106 cells were incubated with 55FevNTA (1 μM) inserum free medium for 30 min at 25 °C. Cells were washed twicewith 150 mM NaCl containing 10 μM EDTA (to remove surface-bound iron) and cell-associated radioactivity was measured in celllysates (as described before) by Liquid Scintillation Counter. Ironuptake was determined as pmol/106 cell/30 min.

2.13. Statistical analysis

All experiments have been performed at least three times withsimilar results and representative experiments are shown. Densito-metric results are normalized with respect to internal controls andare expressed relative to the results in untreated control. Error barsrepresent standard deviations with a p value b0.06 was taken to bestatistically significant.

296 S. Biswas et al. / Biochimica et Biophysica Acta 1832 (2013) 293–301

3. Results

3.1. Insulin regulates transferrin-mediated iron uptake in human hepaticcells by promoting TfR1 expression

To determine the ability of insulin to promote iron uptake, HepG2cells were treated with insulin (0–30 nM) and transferrin-bound iron(55Fe-Tf) uptake was examined. A dose dependent increase in 55Fe-Tfuptake up to about 1.8-fold was detected by insulin (0–30 nM) treat-ment (Fig. 1A). Any further increase in iron uptake was not foundwith further increase in insulin concentration (data not shown).Treatment with iron chelator DFO (150 μM) increased intracellular55Fe by about 2.4-fold (Fig. 1A). Insulin had no effect on nontransferrin-bound iron (NTBI) uptake as no increase in intracellular55Fe was detected in insulin (30 nM) treated HepG2 cells after incu-bation with 55Fe-NTA (Fig. 1B). To find any appreciable iron uptakefrom 55Fe-Tf, more than 8 h of insulin treatment was required andmaximum was detected after 16 h of insulin treatment (Fig. 1C).

To find out whether insulin-induced increase in 55Fe uptake wasdue to increase in TfR1 expression, we treated HepG2 cells with in-creasing concentration of insulin and steady state TfR1 level was

Fig. 1. Role of insulin on iron uptake in HepG2 cells. A. Serum deprived cells were treat-ed with insulin (0–30 nM) or DFO (150 μM) for 16 h then incubated with 55Fe-Tf. After30 min intracellular radiolabled iron was determined as described in Materials andmethods. Results were representative of three independent experiments performedin triplicate, pb0.04. B. Serum deprived HepG2 cells were treated with insulin(30 nM) for 16 h and then incubated with 55Fe-NTA. Intracellular radiolabeled ironwas determined after 30 min. Results were representative of three independent exper-iments performed in triplicate, pb0.02. C. Similarly, cells were treated with insulin(30 nM) for different period of time or remained untreated. After each time periodcells were incubated with 55Fe-Tf. After 30 min intracellular radiolabled iron wasdetected. Results were determined from three independent experiments performedin triplicate, pb0.04.

determined in cell lysate by Western blot analysis. A concentrationdependent increase in TfR1 level was detected by insulin treatment(Fig. 2A). The maximal increase (about 2-fold) was detected by 10to 30 nM of insulin treatment. The iron chelator DFO (100 μM) wasused as a positive control (Fig. 2A). After insulin treatment morethan 8 h was required to detect any appreciable change in TfR1level and maximum increase was found at 16 h (Fig. 2B). To furtherconfirm the role of insulin on TfR1 expression, human primary hepa-tocytes were treated with insulin (30 nM) or DFO (150 μM) for 16 hand Western blot analysis was performed in cell lysates. TfR1 expres-sion was increased about 2.1-fold by insulin and about 2.6-fold byDFO treatment (Fig. 2C). Since, insulin was reported to inhibit prote-olysis [40], we examined whether insulin had any influence on TfR1stability at the protein level. To verify that cells were treated with in-sulin or remain untreated for 14 h and then protein synthesis inhibi-tor cycloheximide (CHX, 10 μg/ml) was added. TfR1 expression wasdetermined by Western blot analysis in cell lysates prepared after 0,3, 6 and 9 h of CHX addition. Reduction of 50% TfR1 expression wasdetected at about 6 h in both insulin treated and untreated cells(Fig. 2D) suggesting insulin had no role in TfR1 protein stability. Asa second approach, cells were transfected with plasmid encodingFLAG-tagged TfR1 and pGL3-control (as transfection control) andthen treated with insulin or none. After 14 h, cell lysates were pre-pared and analyzed by Western blotting. The data showed while en-dogenous TfR1 was increased (70%) but epitope-tagged TfR1(detected by FLAG antibody) remained unchanged (Fig. 2E) furthersuggesting no role of insulin on TfR1 stabilization at protein level.

3.2. Insulin regulates TfR1 synthesis by a transcriptional mechanism

To further determine the mechanism Northern blot analysis wasperformed and about 2-fold increase in TfR1 mRNA was detectedafter 16 h of insulin treatment (Fig. 3A). A similar result of about2.1-fold increase of TfR1 transcript was detected in insulin treatedcells by qPCR analysis. Since, TfR1 is mainly regulated bypost-transcriptional mRNA stability mechanism by iron depletion[3,21,22]; we determined half-life of TfR1 transcript in insulin treatedand untreated cells. Insulin treatment did not alter TfR1 mRNAhalf-life (about 2 h) as reported earlier [35,36] (Fig. 3B–C) indicatingthe regulation was primarily at the transcription level. To test thatpromoter region (−114 to +14 from transcription start site) ofTfR1 was cloned into pGL3-basic vector behind luciferase gene andtransfected into HepG2 cells along with β-galactosidase (coupledwith SV40 promoter). After 16 h of insulin treatment luciferase andgalactosidase activities were determined in cell extracts. The resultshowed about 2.1-fold increase in promoter activity (Fig. 4A)confirming TfR1 regulation by insulin was due to transcriptionalmechanism. To identify the responsible element(s) conferringinsulin-mediated increase in TfR1 expression we sequentially deletedthe promoter region from the proximal end (−95 and −82) and ex-amined the promoter activity. Insulin induced TfR1 promoter activityin−95 construct transfected cells remained unaffected, but the activ-ity was abolished in−82 construct transfected cells (Fig. 4A). This re-sult suggested that the −95 to −82 bp region of the TfR1 promoterwas critical for the transcriptional regulation by insulin. To find thepresence of any known responsive element within this region, wesearched transcription factor binding consensus sequence between−95 and −82 region upstream of the transcription start site and ahypoxia responsive element (HRE, TACGTGCC) was detected withinthat region (Fig. 4A). To examine the involvement of HRE ininsulin-induced activation of TfR1 promoter, the HRE was mutatedfrom 5′ TACGTGCC 3′ to 5′ TAAAAGCC 3′ in the −114 construct andthe mutated construct (Mut-HRE) was transfected. In response to in-sulin (30 nM) no appreciable change in luciferase activity wasdetected in Mut-HRE transfected cells, while wild-HRE transfectedcells showed more than two-fold increase in luciferase activity

Fig. 2. Role of insulin on transferrin receptor-1 expression. A. Cells were treated with insulin (0–30 nM) or desferrioxamine (100 μM) for 16 h; TfR1 (upper panel) and actin (mid-dle panel) expressions were detected in cell lysates by Western blot analyses. Relative expression was determined by densitometric analysis from three independent experiments(bottom panel), pb0.02 B. Similarly, cells were treated with insulin (30 nM) or remained untreated. TfR1 (upper panel) and actin (middle panel) expressions were determined byWestern blot analysis from cell lysates isolated after 8 h, 16 h or 24 h of insulin treatment. Relative expression was determined by densitometric analysis from three independentexperiments (bottom panel), pb0.03. C. Similarly, human primary hepatocytes were treated with either insulin (30 nM) or desferrioxamine (150 μM) or remain untreated for 16 hand Western blot analysis was performed with cell lysates for TfR1 (upper panel) and actin (bottom panel). Result represents one of the three independent experiments of similarresults. D. Cycloheximide (CHX, 10 μg/ml) was added after 14 h of insulin treated or untreated cells and cell lysates were obtained after 0, 3, 6 and 9 h of CHX treatment. Westernblot analyses were performed using TfR1 and actin antibodies in cell lysates. Relative level of TfR1 after comparing with actin (in %) was obtained from three independent exper-iments, pb0.05. Decrease of 50% TfR1 level in compare to 0 h of CHX addition has been shown by arrow. E. Cells were transfected with FLAG-tagged TfR1 and pGL3-control vector orkept untransfected (UT). Then transfected cells were treated with insulin (none or 30 nM) for 14 h. Western blot analyses were performed with cell lysates with antibodies of TfR1,FLAG, luciferase and actin. Results shown are representative of one of the three independent experiments with similar results.

297S. Biswas et al. / Biochimica et Biophysica Acta 1832 (2013) 293–301

(Fig. 4B) suggesting involvement of HRE in insulin-mediated activa-tion of TfR1 in HepG2 cells.

3.3. Identification of HIF-1 as the transcription factor responsible forinsulin-induced TfR1 regulation

When 32P labeled double stranded TfR1-HRE probe was incubatedwith nuclear extract isolated from insulin treated HepG2 cells, aninsulin-induced DNA binding activity was detected by EMSA(Fig. 5A). To find out specificity of the binding, cold competitionassay was performed. When 10× or 30× molar excess of the coldprobe was added with nuclear extract from insulin treated HepG2cell, the insulin-induced binding (shown by the solid arrowhead) as

well as non-specific (NS) binding was competed out. But when 10×or 30× molar excess of unlabeled mutant-HRE was added only thenon specific binding was competed out but insulin-induced DNAbinding remained unaltered confirming the binding was due to HRE(Fig. 5A). To confirm that insulin-induced HRE binding was actuallydue to HIF-1, supershift analysis was performed using HIF-1α andHIF-1β antibody. When nuclear extract from insulin-induced HepG2cell was incubated with either HIF-1α or HIF-1β antibody or both to-gether; in all cases the specific insulin-induced DNA binding activitywas either shifted to higher mobility or not entered into the gel prob-ably due to a larger complex formation (Fig. 5B). This experimentstrongly suggested that insulin-induced increase in TfR1 synthesiswas due to HIF-1 binding to HRE present in the TfR1 promoter.

Fig. 3. Insulin promotes TfR1 mRNA expression. A. Total RNA was isolated after 16 h ofinsulin (30 nM) treatment and abundance of TfR1 mRNA was determined after hybrid-izing with 32P labeled human TfR1 cDNA (left upper panel). The abundance of loadedRNA was determined by β-actin hybridization (left bottom panel). Relative expressionwas determined by densitometric analysis (right panel), pb0.03. B. To determine themRNA stability of TfR1, actinomycin D (7.5 μg/ml) was added to insulin (30 nM) treat-ed and untreated cells after 12 h of treatment. Total RNA was isolated at 0, 1.5, 3.0 and4.5 h of actinomycin D addition and reverse-transcriptase PCR was performed withspecific primers of human TfR1 and β-actin. C. Relative expression of TfR1 mRNA wasrepresented (in %) from three independent experiments, pb0.05. Decrease of 50%TfR1 mRNA level in compare to 0 h of actinomycin D addition has been shown byarrow.

Fig. 4. Insulin induces TfR1 promoter activity. A. Chimeric pGL3-basic vectorscontaining different regions of human TfR1 promoter were transfected in HepG2cells (with a plasmid containing β-galactosidase to correct for transfection efficiency)and then treated with insulin (30 nM) after serum deprivation. After 16 h, luciferaseactivity in cell extracts was measured by chemiluminescence and normalized forβ-galactosidase activity. Data represents means±S.D., n=3, pb0.06. B. ChimericpGL3 basic vector containing HRE (−114 to +14 segment of TfR1 promoter) eitherwild (Wild) or mutated (Mut) was transfected along with β-galactosidase plasmid.After recovery cells were serum deprived and treated with insulin (16 h, 30 nM).Luciferase activity in cell extracts was measured and normalized for β-galactosidaseactivity. Data shown are the means±S.D., n=3, pb0.04.

298 S. Biswas et al. / Biochimica et Biophysica Acta 1832 (2013) 293–301

3.4. Role of HIF-1 in insulin-induced TfR1 mediated iron uptake

To confirm the role of HIF-1 in insulin-induced TfR1-mediated ironuptake, HepG2 cells were transfected with HIF-1α siRNA targeted toits 3′UTR and also with a control siRNA. Western blot analysis showedabout 2- fold increase in HIF-1α protein level after 16 h of insulintreatment in control siRNA transfected cells (Fig. 6A). Interestingly,we detected cross reactive protein bands with similar mobility withthe HIF-1α antibody in HepG2 cells that was not affected by HIF-1αsiRNA. The same antibody did not show any cross reactivity in othercell types in normoxic condition (data not shown). Insulin-inducedHIF-1α expression was inhibited more than 90% by HIF-1α siRNAtransfected cells (Fig. 6A). To show the specificity of siRNA,HIF-1α-siRNA (targeted to its 3′UTR) transfected cells were furthertransfected with pcDNA3-HIF-1α (P/A) containing mutations at402-proline and 564-proline to alanine to overcome the effect ofprolyl hydroxylase mediated HIF-1α protein degradation. Result

showed HIF-1α expression was restored in pcDNA3-HIF-1α (P/A)transfected cells but not in pcDNA3 transfected cells (Fig. 6B). WhenHIF-1α siRNA transfected cells were treated with insulin, TfR1 ex-pression was inhibited by about 90% in compared to cells transfectedwith control siRNA (Fig. 6C). Interestingly, basal TfR1 level was not af-fected by HIF-1α-siRNA indicating HIF-1α was responsible only forinsulin-induced TfR1 synthesis (Fig. 6C). When transferrin-boundiron uptake capacity of HIF-1α siRNA transfected cells was tested byincubating with 55Fe-Tf, more than 80% of insulin-induced iron up-take was blocked in compared to control siRNA transfected cells(Fig. 6D). A similar result was also obtained with another set ofsiRNA targeted to HIF-1α coding region (target sequence: 5′ AACACA CAG CGG AGC TTT TTT 3′) (S1). All these results strongly suggestinsulin regulates TfR1 expression in HepG2 cells by a HIF-1 depen-dent mechanism to stimulate iron uptake.

4. Discussion

Glucose metabolism is critically dependent on cellular iron avail-ability [1,13] and insulin [8–12]. Hepatic iron store also influences in-sulin sensitivity and signaling [13–18]; however, any direct role ofinsulin on hepatic iron homeostasis has not been explored. In thisstudy we demonstrated a novel role of insulin in promoting iron

Fig. 5. Identification of HIF-1 as transcription factor mediating insulin-induced TfR1promoter activity. A. EMSA was performed with nuclear extract isolated from insulin(30 nM, 12 h) treated and untreated cells. Induction of insulin-induced DNA-proteincomplex was shown with arrow (closed). NS denotes non-specific binding. Competi-tion experiment was carried out with 10- or 30- fold molar excess unlabeled probescontaining either wild HRE (W) or mutant-HRE (M). Unlabeled probes were added1-minute prior the addition of radiolabled probe. B. Insulin treated nuclear extractwas incubated with either HIF-1α and/or HIF-1β antibody for 1 h prior to EMSA. ‘SS’denotes super-shifted HIF-1 complex. Results are representative of one of the threeindependent experiments.

Fig. 6. HIF-1α siRNA blocks insulin-mediated TfR1 expression and iron uptake. A. Cellswere transfected with either control (si cont) or HIF-1α specific siRNA (si HIF-1α) andthen treated with insulin (30 nM) for 16 h. HIF-1α and actin expressions were deter-mined in nuclear extracts by Western blot analyses. B. After 24 h of transfection withHIF-1α specific siRNA, cells were retransfected with either pcDNA3 (vector control)or pcDNA3-HIF-1α (P/A) for 24 h as described details in Materials and methods.Then cells were treated with insulin (30 nM) or remain untreated for 16 h and West-ern blot analyses were performed in nuclear extracts with HIF-1α and actin antibodies.C. In control (si cont) or HIF-1α specific siRNA (si HIF-1α) transfected cell lysates (asdescribed in ‘A’) TfR1 and actin expressions were determined. D. Similarly, controland HIF-1α siRNA transfected cells were treated with insulin for 16 h and 55Fe-Tfuptake was determined as described earlier. Results were from three independentexperiments performed in triplicate, pb0.04.

299S. Biswas et al. / Biochimica et Biophysica Acta 1832 (2013) 293–301

uptake in human hepatic HepG2 cell by regulating TfR1 synthesis. Asimilar regulation of TfR1 was also detected in insulin treatedhuman primary hepatocytes. We further revealed that insulin-induced TfR1 synthesis was mediated by transcription factor HIF-1by binding to the single HRE present in the TfR1 promoter. Our resultestablished a direct and novel link between iron and energy metabo-lism in liver cell in response to insulin.

We detected about 1.8-fold increase in radiolabeled transferrin-bound iron uptake after 16 h of insulin treatment (Fig. 1) that wasconsistent with increase in TfR1 expression by insulin (Fig. 2B) indi-cating TfR1 synthesis was the key cellular event for insulin-inducediron uptake. This was further confirmed by blocking insulin-inducedTfR1 expression using HIF-1α siRNA, and 55Fe-transferrin uptake(Fig. 6C–D; S1). All these results strongly suggest that insulin pro-motes iron uptake in hepatic cell by HIF-1 dependent synthesis ofTfR1. Earlier, we reported the ability of insulin in ceruloplasmin

(Cp) synthesis in HepG2 cells by HIF-1 dependent mechanism [33].Given the role of Cp in loading iron into transferrin; the binding com-ponent of TfR1, a global role of insulin-mediated HIF-1 activation inpromoting hepatic iron uptake could be assumed.

The regulation of TfR1 due to stability of its mRNA has been studiedoverwhelmly [3,21,41]. Our result indicated that insulin apparently hadno effect on TfR1 mRNA stability as measured by semi-quantitativeRT-PCR (Fig. 3B–C). Even though the method we used wassemi-quantitative in nature, the half-life of TfR1 was detected atabout 2 h that closely matched with the earlier results [35,36]. Butgiven the earlier report of increased IRP1 expression at mRNA level indifferentiated adipocytes resulting into increased IRP–IRE interaction[42] we are not completely ruling out the possibility of post-transcriptional regulation of TfR1 mRNA in insulin-induced hepaticcells. Of note, that insulin was used as one of the components alongwith isobutylmethylxanthine and dexmathosone for differentiation of

300 S. Biswas et al. / Biochimica et Biophysica Acta 1832 (2013) 293–301

the adipocyte [42]; however, our experiment was conducted only withinsulin in HepG2 cells.

Interestingly, regulation of TfR1 at the transcription level is lessreported. The presence of an active HRE in TfR1 promoter and its in-volvement in TfR1 regulation by iron depletion or hypoxic exposurewas reported in hepatic cells [25,26]. Basal synthesis of TfR1 is depen-dent on GC-rich Sp1 binding site [32]. A rapid insulin-induced nuclearlocalization of Sp1 was detected in skeletal muscle cells [31]; where-as, role of Ets-1 in TfR1 regulation was reported in erythroid cells[43,44]. Among these transcription factors reported to control TfR1expression in various cell types we could only detect the involvementof HIF-1 in insulin-induced TfR1 synthesis in HepG2 cells. Although,the ability of insulin in HIF-1 activation in muscle and hepatic cellwas reported earlier [30,33,34], but the actual mechanism of activa-tion had not been elucidated so far. Post-translational protein stabilitymechanism of regulatory subunit HIF-1α during hypoxia, exposuresto other hypoxia mimetic like cobalt or iron depletion is the paradigmof HIF-1 activation. In normal condition, hydroxylation of two prolineresidues of HIF-1α (Pro402 and Pro564) promotes ubiquitination andsubsequent proteasomal degradation [27–29]. Three different HIFprolyl-hydroxylases termed PHD1, PHD2, and PHD3 are able to hy-droxylate these HIF-1α proline residues. PHDs hydroxylate HIF-1αusing oxygen and 2-oxoglutarate as substrates and iron and ascorbateas cofactors [45]. Since cellular iron level reciprocally regulatesPHD activity affecting HIF-1α stabilization; thus, insulin-inducediron-uptake should inhibit HIF-1 activation by increasing PHD activi-ty. As a result there should be a decrease in TfR1 expression in case ofHIF-1 activation by insulin which depends on PHD activity exclusive-ly. We detected TfR1 expression even after 24 h of insulin treatment(Fig. 2B) indicating involvement of a different or additional mecha-nism other than PHD activity for insulin-induced HIF-1 activation. Arecent report demonstrated that insulin can increase HIF-1α mRNAand protein level in 3T3-L1 adipocyte [46]. When we examinedHIF-1α mRNA level in insulin treated HepG2 cell, about 2-fold in-crease in HIF-1α mRNA was detected by Northern blot analysis (S2)suggesting that insulin-induced increase in HIF-1α mRNA levelcould be one of the mechanisms of HIF-1 activation by insulin. Thedegree of involvement of cellular PHD activity and increased HIF-1αtranscript in insulin-induced HIF-1 activation is not clear so far. Itneeds further study to unravel the detailed mechanism of HIF-1 acti-vation by insulin.

Earlier studies reported the ability of insulin in increased TfR1recycling [47,48] in adipocyte. But the role of insulin in TfR1 synthesishas not been reported so far in any other cells including adipocyte.We detected insulin-induced TfR1 synthesis and 55Fe-transferrin up-take only after 8 h of treatment indicating TfR1 recycling may nothave any significant role in insulin-induced 55Fe-transferrin uptakein HepG2 cell. The previous study in adipocyte and the currentstudy in HepG2 cell though suggest a general role of insulin in cellulariron uptake in sensitive cells. But the exact physiological importanceof insulin-mediated iron uptake is not clear so far. Insulin as a growthfactor may need the appropriate amount of micronutrient like iron forits critical anabolic activity and energy generation from glucose.Interestingly, redistribution and localization of TfR1 and glucosetransporters along with IGF-II receptor in to the surface from intracel-lular compartment in insulin treated 3T3-L1 adipocytes [49] furthersupport that insulin-induced iron uptake in sensitive cells may beimportant for energy generation.

The role of HIF-1 in iron homeostasis is well established duringhypoxia and iron deficiency [50]. But its role in cellular iron uptakein normoxic condition has not been envisaged and shown before.Our result also indicates an important role of HIF-1 in iron physiologyeven in other conditions than hypoxia or iron deficiency. Taken to-gether, we hypothesize that activation of HIF-1 and subsequent regu-lation of iron homeostasis by insulin may help in the redistribution ofiron within and outside the tissues for cellular growth.

The insulin-induced HIF-1 activation and resulting TfR1-mediatediron uptake may be important for cellular growth but in a conditionlike IR or hyperinsulinemia, this may contribute to the pathophysiol-ogy of insulin resistance related HIO. When peripheral tissues like themuscle and adipocyte become resistant to insulin [9,11], then thebody tries to compensate the ineffectiveness by increasing synthesisand secretion of more insulin [9,11]. Thus, physiological concentra-tion of insulin becomes higher than normal (hyperinsulinemia). Inthe early stage of hyperinsulinemia hepatocyte remains sensitive toinsulin [9,11,51]. In this condition, higher insulin concentration mayincrease hepatic iron uptake by increasing TfR1 expression resultinginto higher hepatic iron level as reported in IR-HIO patients [19,52].Intracellular increase in iron level regulates iron-storage protein ferri-tin synthesis [3,21,41]. Thus, insulin-induced TfR1mediated increasein cellular iron level may result into increased ferritin level asreported in IR-HIO patient [19,52]. However, cellular iron load alsodepends on the expression of ubiquitous iron release componentferroportin, which contains an active IRE in its 5′UTR [3,21,41].Thus, increase in iron pool in hepatocyte should increase ferroportinexpression by decreasing IRP interaction with IRE present in its 5′UTR. However, hepcidin degrades ferroportin by a posttranslationalmechanism [3,21,41]. About 50% increase in serum prohepcidinlevel was reported in type 2 diabetes patients than normal glucosetolerant subjects [53]. Higher cellular iron load or hyperferritinemiawas reported with higher hepcidin level in serum [19,53]. Increasedhepcidin level should decrease ferroportin expression to restrictiron within cell. All these events may be initiated with chronic in-crease in hepatic iron load from increased TfR1 expression by insulin.Although, our study in HepG2 cell culture may be extrapolated to thismodel of hepatic iron overload; but it is not possible to examine allthese chronic events leading to pathological condition like IR-HIO incell culture study. To overcome this limitation it needs a combinationof in vivo and in vitro studies for better understanding of the molec-ular mechanism of IR-HIO in the future.

In summary, we identified a novel function of insulin in TfR1 me-diated iron uptake in hepatic cell mediated by HIF-1. The importanceof which in physiology is not clear so far but this may be one of thecontributors of hepatic iron overload condition in hyperinsulinemiaor IR.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbadis.2012.11.003.

AbbreviationsDFO DeferrioxamineFAC Ferric ammonium citrateFt-H Ferritin-HHIO Hepatic iron overloadHRE Hypoxia responsive elementHIF-1 Hypoxia inducible factor-1IDDM Insulin-dependent diabetes mellitusIR Insulin resistanceIRP Iron regulatory proteinIRE Iron responsive elementLIP Labile iron poolNAFLD Non-alcoholic fatty liver diseaseNIDDM Non-insulin-dependent diabetes mellitusNTA NitrilotriacetateNTBI Non transferrin-bound ironROS Reactive oxygen speciesRT-PCR Reverse transcriptase polymerase chain reactionSDS-PAGE Sodium dodecyl sulfate-polyacrylamide gel electrophoresisTBI Transferrin-bound ironTf TransferrinTfR1 Transferrin receptor-1UTR Untranslated regionVEGF Vascular endothelial growth factor

301S. Biswas et al. / Biochimica et Biophysica Acta 1832 (2013) 293–301

Contributors

S.B., N.T., R.M., and R.K. performed the experiments, organized thedata and statistical analysis. C.K.M. conceived, organized, supervisedthe study and wrote the paper.

Acknowledgements

This work is supported by grants from the Council of Scientific andIndustrial Research and Indian council of Medical Research of Indiaand Department of Science and Technology-PURSE program to CKM.SB acknowledges CSIR for Senior Research Fellowship. MutantHIF-1α cDNA (P/A) was a kind gift from Dr. Ritu Kulshreshtha, IndianInstitute of Technology, New Delhi, India.

References

[1] H. Oexle, E. Gnaiger, G. Weiss, Iron-dependent changes in cellular energy metab-olism: influence on citric acid cycle and oxidative phosphorylation, Biochim.Biophys. Acta 1413 (1999) 99–107.

[2] B. Halliwell, J.M.C. Gutteridge, Role of free radicals and catalytic metal ions inhuman disease, in: L. Packer (Ed.), Oxygen Radicals, Part B, Methods Enzymol.,186, Academic Press, San Diego, 1990, pp. 1–85.

[3] M.W. Hentze, M.U. Muckenthaler, N.C. Andrews, Balancing acts: molecularcontrol of mammalian iron metabolism, Cell 117 (2004) 285–297.

[4] Y. Kohgo, K. Ikuta, T. Ohtake, Y. Torimoto, J. Kato, Iron overload and cofactors withspecial reference to alcohol, hepatitis C virus infection and steatosis/insulinresistance, World J. Gastroenterol. 13 (2007) 4699–4706.

[5] R. Moirand, A.M. Mortaji, O. Loreal, F. Palliard, P. Brissot, Y. Deugnier, A newsyndrome of liver iron overload with normal transferrin saturation, Lancet 349(1997) 95–97.

[6] M.H. Mendler, B. Turlin, R. Moirand, A.M. Jouanolle, T. Sapey, D. Guyader, J.Y. LeGall, P. Brissot, V. David, Y. Deugnier, Insulin resistance-associated hepatic ironoverload, Gastroenterology 117 (1999) 1155–1163.

[7] D.F. Wallace, V.N. Subramaniam, Co-factors in liver disease: the role ofHFE-related hereditary hemochromatosis and iron, Biochim. Biophys. Acta 1790(2009) 663–670.

[8] S.E. Leney, J.M. Tavaré, The molecular basis of insulin-stimulated glucose uptake:signalling, trafficking and potential drug targets, J. Endocrinol. 203 (2009) 1–18.

[9] M.D. Michael, R.N. Kulkarni, C. Postic, S.F. Previs, G.I. Shulman, M.A. Magnuson,C.R. Kahn, Loss of insulin signaling in hepatocytes leads to severe insulin resis-tance and progressive hepatic dysfunction, Mol. Cell 6 (2000) 87–97.

[10] M.K. Cavaghan, D.A. Ehrmann, K.S. Polonsky, Interactions between insulinresistance and insulin secretion in the development of glucose intolerance,J. Clin. Invest. 106 (2000) 329–333.

[11] C.A. Baumann, A.R. Saltiel, Spatial compartmentalization of signal transduction ininsulin action, Bioessays 23 (2001) 215–222.

[12] K.F. Petersen, G.I. Shulman, Etiology of insulin resistance, Am. J. Med. 119 (2006)S10–S16.

[13] J.M. Fernández-Real, A. López-Bermejo, W. Ricart, Cross-talk between iron metab-olism and diabetes, Diabetes 51 (2002) 2348–2354.

[14] L. Tilbrook, Cross talk between iron metabolism and diabetes, Ann. Clin. Biochem.41 (2004) 255.

[15] P. Dongiovanni, L. Valenti, A.L. Fracanzani, S. Gatti, G. Cairo, S. Fargion, Irondepletion by deferoxamine up-regulates glucose uptake and insulin signaling inhepatoma cells and in rat liver, Am. J. Pathol. 172 (2008) 738–747.

[16] L. Valenti, A.L. Fracanzani, P. Dongiovanni, E. Bugianesi, G. Marchesini, P. Manzini,E. Vanni, S. Fargion, Iron depletion by phlebotomy improves insulin resistance inpatients with nonalcoholic fatty liver disease and hyperferritinemia: evidencefrom a case–control study, Am. J. Gastroenterol. 102 (2007) 1251–1258.

[17] J.M. Fernández-Real, A. López-Bermejo, W. Ricart, Iron stores, blood donation, andinsulin sensitivity and secretion, Clin. Chem. 51 (2005) 1201–1205.

[18] A. Piperno, A. Vergani, A. Salvioni, P. Trombini, M. Viganò, A. Riva, A. Zoppo, G.Boari, G. Mancia, Effects of venesections and restricted diet in patients with theinsulin-resistance hepatic iron overload syndrome, Liver Int. 24 (2004) 471–476.

[19] P. Dongiovanni, A.L. Fracanzani, S. Fargion, L. Valenti, Iron in fatty liver and in themetabolic syndrome: a promising therapeutic target, J. Hepatol. 55 (2011) 920–932.

[20] J. Gao, N. Zhao, M.D. Knutson, C.A. Enns, The hereditary hemochromatosis protein,HFE, inhibits iron uptake via down-regulation of Zip14 in HepG2 cells, J. Biol.Chem. 283 (2008) 21462–21468.

[21] N.C. Andrews, P.J. Schmidt, Iron homeostasis, Annu. Rev. Physiol. 69 (2007)69–85.

[22] T.A. Rouault, The role of iron regulatory proteins in mammalian iron homeostasisand disease, Nat. Chem. Biol. 2 (2006) 406–414.

[23] J.P. Liuzzi, F. Aydemir, H. Nam, M.D. Knutson, R.J. Cousins, Zip14 (Slc39a14) medi-ates non-transferrin-bound iron uptake into cells, Proc. Natl. Acad. Sci. U. S. A. 103(2006) 13612–13617.

[24] Y. Deugnier, P. Brissot, O. Loréal, Iron and the liver: update 2008, J. Hepatol. 48(2008) S113–S123.

[25] L. Bianchi, L. Tacchini, G. Cairo, HIF-1-mediated activation of transferrin receptorgene transcription by iron chelation, Prog. Nucleic Acid Res. 27 (1999) 4223–4227.

[26] C.N. Lok, P. Ponka, Identification of a hypoxia response element in the transferrinreceptor gene, J. Biol. Chem. 274 (1999) 24147–24152.

[27] L.E. Huang, J. Gu, M. Schau, H.F. Bunn, Regulation of hypoxia-inducible factor 1alphais mediated by an O2-dependent degradation domain via the ubiquitin-proteasomepathway, Proc. Natl. Acad. Sci. U. S. A. 95 (1998) 7987–7992.

[28] M. Ivan, K. Kondo, H. Yang, W. Kim, J. Valiando, M. Ohh, A. Salic, J.M. Asara, W.S.Lane, W.G. Kaelin Jr., HIF alpha targeted for VHL-mediated destruction by prolinehydroxylation: implications for O2 sensing, Science 292 (2001) 464–468.

[29] P. Jaakkola, D.R. Mole, Y.M. Tian, M.I. Wilson, J. Gielbert, S.J. Gaskell, Av.Kriegsheim, H.F. Hebestreit, M. Mukherji, C.J. Schofield, P.H. Maxwell, C.W.Pugh, P.J. Ratcliffe, Targeting of HIF-alpha to the von Hippel–Lindauubiquitylation complex by O2-regulated prolyl hydroxylation, Science 292(2001) 468–472.

[30] E. Zelzer, Y. Levy, C. Kahana, B.Z. Shilo, M. Rubinstein, B. Cohen, Insulin inducestranscription of target genes through the hypoxia-inducible factorHIF-1alpha/ARNT, EMBO J. 17 (1998) 5085–5094.

[31] M. Horovitz-Fried, A.I. Jacob, D.R. Cooper, S.R. Sampson, Activation of the nucleartranscription factor SP-1 by insulin rapidly increases the expression of proteinkinase C delta in skeletal muscle, Cell. Signal. 19 (2007) 556–562.

[32] W.K. Miskimins, A. McClelland, M.P. Roberts, F.H. Ruddle, Cell proliferation andexpression of the transferrin receptor gene: promoter sequence homologies andprotein interactions, J. Cell Biol. 103 (1986) 1781–1788.

[33] V. Seshadri, P.L. Fox, C.K. Mukhopadhyay, Dual role of insulin in transcriptionalregulation of the acute phase reactant ceruloplasmin, J. Biol. Chem. 277 (2002)27903–27911.

[34] S. Biswas, M.K. Gupta, D. Chattopadhyay, C.K. Mukhopadhyay, Insulin-inducedactivation of hypoxia-inducible factor-1 requires generation of reactive oxygenspecies by NADPH oxidase, Am. J. Physiol. Heart Circ. Physiol. 292 (2007)H758–H766.

[35] N.K. Das, S. Biswas, S. Solanki, C.K. Mukhopadhyay, Leishmania donovani depleteslabile iron pool to exploit iron uptake capacity of macrophage for its intracellulargrowth, Cell. Microbiol. 11 (2009) 83–94.

[36] N. Tapryal, C. Mukhopadhyay, D. Das, P.L. Fox, C.K. Mukhopadhyay, Reactive oxy-gen species regulate ceruloplasmin by a novel mRNA decay mechanism involvingits 3′-untranslated region: implications in neurodegenerative diseases, J. Biol.Chem. 284 (2009) 1873–1883.

[37] C.K. Mukhopadhyay, B. Mazumder, P.L. Fox, Role of hypoxia-inducible factor-1 intranscriptional activation of ceruloplasmin by iron deficiency, J. Biol. Chem. 275(2000) 21048–21054.

[38] R.D. Klausner, J.V. Renswoude, G. Ashwell, C. Kempf, A.N. Schechter, A. Dean, K.R.Bridges, Receptor-mediated endocytosis of transferrin in K562 cells, J. Biol. Chem.258 (1983) 4715–4724.

[39] A. Sturrock, J. Alexander, J. Lamb, C.M. Craven, J. Kaplan, Characterization of atransferrin-independent uptake system for iron in HeLa cells, J. Biol. Chem. 265(1990) 3139–3145.

[40] R.G. Bennett, F.G. Hamel, W.C. Duckworth, Insulin inhibits the ubiquitin-dependentdegrading activity of the 26S proteasome, Endocrinology 141 (2000) 2508–2517.

[41] I. De Domenico, D. McVeyWard, J. Kaplan, Regulation of iron acquisition and storage:consequences for iron-linked disorders, Nat. Rev. Mol. Cell Biol. 9 (2008) 72–81.

[42] M. Festa, G. Ricciardelli, G. Mele, C. Pietropaolo, A. Ruffo, A. Colonna, Overexpressionof H ferritin and up-regulation of iron regulatory protein genes during differentia-tion of 3T3-L1 pre-adipocytes, J. Biol. Chem. 275 (2000) 36708–36712.

[43] C.N. Lok, P. Ponka, Identification of an erythroid active element in the transferrinreceptor gene, J. Biol. Chem. 275 (2000) 24185–24190.

[44] G. Marziali, E. Perrotti, R. Ilari, V. Lulli, E.M. Coccia, R. Moret, L.C. Kühn, U. Testa, A.Battistini, Role of Ets-1 in transcriptional regulation of transferrin receptor anderythroid differentiation, Oncogene 21 (2002) 7933–7944.

[45] E.L. Pagé, D.A. Chan, A.J. Giaccia, M. Levine, D.E. Richard, Hypoxia-induciblefactor-1{alpha} stabilization in nonhypoxic conditions: role of oxidation and in-tracellular ascorbate depletion, Mol. Biol. Cell 19 (2008) 86–94.

[46] Q. He, Z. Gao, J. Yin, J. Zhang, Z. Yun, J. Ye, Regulation of HIF-1{alpha} activity inadipose tissue by obesity-associated factors: adipogenesis, insulin, and hypoxia,Am. J. Physiol. Endocrinol. Metab. 300 (2011) E877–E885.

[47] R.J. Davis, S. Corvera, M.P. Czech, Insulin stimulates cellular iron uptake andcauses the redistribution of intracellular transferrin receptors to the plasmamembrane, J. Biol. Chem. 261 (1986) 8708–8711.

[48] L.I. Tanner, G.E. Lienhard, Insulin elicits a redistribution of transferrin receptors in3T3-L1 adipocytes through an increase in the rate constant for receptor external-ization, J. Biol. Chem. 262 (1987) 8975–8980.

[49] L.I. Tanner, G.E. Lienhard, Localization of transferrin receptors and insulin-likegrowth factor II receptors in vesicles from 3T3-L1 adipocytes that contain intra-cellular glucose transporters, J. Cell Biol. 108 (1989) 1537–1545.

[50] C. Peyssonnaux, V. Nizet, R.S. Johnson, Role of the hypoxia inducible factors HIF iniron metabolism, Cell Cycle 7 (2008) 28–32.

[51] M.J. Birnbaum, Turning down insulin signaling, J. Clin. Invest. 108 (2001)655–659.

[52] B. Turlin, M.H. Mendler, R. Moirand, D. Guyader, A. Guillygomarc'h, Y. Deugnier,Histologic features of the liver in insulin resistance-associated iron overload. Astudy of 139 patients, Am. J. Clin. Pathol. 116 (2001) 263–270.

[53] J.M. Fernández-Real, F. Equitani, J.M. Moreno, M. Manco, F. Ortega, W. Ricart,Study of circulating prohepcidin in association with insulin sensitivity and chang-ing iron stores, J. Clin. Endocrinol. Metab. 94 (2009) 982–988.