Expression of SFT (stimulator of Fe transport) is enhanced by iron chelation in HeLa cells and by hemochromatosis in liver.

SFT (stimulator of Fe transport) is a novel transport protein that has been found to facilitate uptake of iron presented to cells as either Fe(II) or Fe(III). When HeLa cells are exposed to the iron chelator desferrioxamine, levels of SFT mRNA increase in an actinomycin D-sensitive manner. In contrast, cells exposed to high levels of iron down-regulate SFT expression in a time-dependent and reversible fashion. Thus, homeostatic regulation of SFT expression not only ensures that sufficient levels of iron are maintained but also limits excessive assimilation to prevent potentially harmful effects of this toxic metal. The unexpected observation that SFT transcript levels are up-regulated in hemochromatosis patients therefore suggests that enhanced SFT expression contributes to the etiology of this iron overload disorder.


From the Department of Nutrition, Harvard School of Public Health, Boston, Massachusetts 02115
SFT (stimulator of Fe transport) is a novel transport protein that has been found to facilitate uptake of iron presented to cells as either Fe(II) or Fe(III). When HeLa cells are exposed to the iron chelator desferrioxamine, levels of SFT mRNA increase in an actinomycin D-sensitive manner. In contrast, cells exposed to high levels of iron down-regulate SFT expression in a time-dependent and reversible fashion. Thus, homeostatic regulation of SFT expression not only ensures that sufficient levels of iron are maintained but also limits excessive assimilation to prevent potentially harmful effects of this toxic metal. The unexpected observation that SFT transcript levels are up-regulated in hemochromatosis patients therefore suggests that enhanced SFT expression contributes to the etiology of this iron overload disorder.
Transport of iron by microbes, plants, and animals is a tightly regulated process that is limited to prevent harmful effects due to overload of this toxic metal (for a comprehensive review of this field, please see Ref. 1). Iron is taken up by bacteria through direct ferrous import systems (e.g. Feo of Escherichia coli) as well as siderophore-based mechanisms. Iron-responsive gene transcription through the regulator Fur is largely responsible for the control of bacterial iron uptake (2). As reviewed by Eide (3) and Askwith and Kaplan (4), studies in yeast have further characterized the role of transport proteins (e.g. Fet4 and Ftr1), ferrireductases (e.g. Fre1 and Fre2), and oxidases (e.g. Fet3 and Fet5) in the membrane transport of iron and also have revealed iron-responsive transcriptional control involving the element Aft1. Like microbes and yeast, plants too display iron-regulated uptake (e.g. IRT1) (5). Finally, advances made through functional expression cloning have most recently identified two mammalian iron transport systems. DCT1/ Nramp2 mediates Fe(II) uptake (6) and is thought to be important for iron assimilation by intestinal and erythroid cells, since defects in its gene promote microcytic anemia in the mk mouse (7) and Belgrade (b) rat (8). In contrast, SFT 1 facilitates the import of nontransferrin-bound iron presented as Fe(II) or Fe(III) (9,10). The latter appears to be first reduced in a copper-requiring process prior to SFT-mediated translocation of Fe(II) across the membrane bilayer (10). Gunshin et al. (6) have reported the presence of DCT1/Nramp2 message in rat intestine, kidney, thymus, and brain, with low levels detected in testis, liver, colon, heart, spleen, skeletal muscle, lung, bone marrow, and stomach. This profile appears to overlap with the appearance of SFT transcripts, which have been found in human intestine, spleen, thymus, prostate, testis, ovaries, and peripheral blood lymphocytes, with low levels observed for colon (9). However, while DCT1/Nramp2 transcripts are known to be increased in the intestine of iron-deficient animals (6), regulation of SFT expression in response to iron levels has yet to be characterized.
The critical need for iron transport regulation can be most readily recognized by the pathological consequences observed when control is lost in the human disease hemochromatosis. This genetic disorder promotes increased intestinal absorption and progressive tissue deposition of iron resulting in cirrhosis of the liver, hepatic carcinoma, congestive heart failure, endocrinopathies, and premature death. The recent elucidation of the gene for this disease led to the discovery that defects in a nonclassical major histocompatibility complex (MHC) class I protein, HFE, are responsible for the iron overload exhibited by patients (11). Although the mechanistic basis for the disease remains obscure, the fact that ␤ 2 -microglobulin knockout mice display characteristics of iron overload associated with hemochromatosis patients confirms the relationship of an MHC class I-like molecule with iron metabolism (12,13). It is estimated that 1 in 10 -20 individuals carry the defective HFE allele and that 1 in 200 -400 Caucasians are homozygous for the disease; thus, it is the most common defective genetic trait known in humans, more prevalent than cystic fibrosis, phenylketonuria, and muscular dystrophy combined. The recent confirmation of its prevalence has prompted some to refer to hemochromatosis as "the genetic disorder of the 21st century" (14).
We have investigated control of SFT expression by iron levels and as shown by the results presented here, we observe that iron chelation enhances accumulation of SFT transcripts and that SFT protein and message levels both decline when cells are exposed to high iron. Remarkably, however, we find that SFT mRNA is significantly higher in liver from hemochromatosis patients despite massive iron deposition within this tissue. Because mucosal uptake of Fe(III), but not Fe(II), is stimulated in the ␤ 2 -microglobulin knockout mouse (15), it is likely that a ferrireductase, an Fe(II)/Fe(III) transporter, or a complex contributing both functions is up-regulated when HFE function is compromised. Thus, our findings suggest that upregulated expression of the Fe(II)/Fe(III) transporter SFT may contribute to the etiology of hemochromatosis.

EXPERIMENTAL PROCEDURES
HeLa cells were grown in Dulbecco's modified Eagle's medium supplemented with 300 g/liter L-glutamine, 10% fetal bovine serum, 50 units/ml penicillin, and 50 g/ml streptomycin as detailed previously (10). HepG2 cells, obtained from American Type Tissue Collection (Manassas, MD), were grown in ␣-minimal essential medium containing 10% fetal bovine serum, 50 units/ml penicillin, and 50 g/ml streptomycin. Cells were iron-loaded with addition of 65 M Fe-NTA (1:1) to the culture medium for the times shown. These conditions are not necessarily selective for nontransferrin-bound iron uptake, but transferrin-independent iron transport has been demonstrated in studies employing Fe-NTA (16 -18), as well as Fe-citrate (18,19), Fe-ascorbate (19 -23), and Fe-diethylenetriaminepentaacetic acid (24). Fe-NTA was chosen for this investigation, because NTA is membrane-impermeable and provides a stably chelated form of Fe(III); others have shown a 4-fold increase in cellular iron content under these experimental conditions (25). This level of Fe-NTA (65 M) had no apparent toxicity based on cell viability and growth. Although confluent cells downregulate SFT expression in response to iron, the extent of down-regulation is greater in dividing cells and therefore all results shown are from experiments performed on subconfluent cultures except where noted. Tissue from hemochromatosis patients and control individuals was obtained from the NIH-funded Liver Tissue Procurement and Distribution System directed by Dr. Harvey Sharp (Fairview University Hospitals and Clinics, Minneapolis, MN).
Total cellular RNA was isolated using RNAzol B according to the manufacturer's direction (Tel-test, Friendswood, TX). For Northern analysis, samples were electrophoresed on denaturing formaldehydeagarose gels, transferred to Nytran (Schleicher & Schuell), UV-crosslinked, then probed with randomly primed 32 P-labeled SFT cDNA under high stringency conditions. Blots were washed twice with 6 ϫ SSPE (1 ϫ SSPE: 0.18 mM sodium chloride, 10 mM sodium phosphate, 1 mM EDTA) containing 0.5% SDS at room temperature, twice with 1 ϫ SSPE containing 0.25% SDS at 37°C, followed by a final wash in 0.1 ϫ SSPE containing 0.1% SDS for 30 min at 65°C. Blots were exposed to film 5-14 days for autoradiography. Ribonuclease protection assays utilized a cDNA fragment of SFT containing the entire 3Ј-noncoding region (9) subcloned into pBSK(Ϫ) (Stratagene, La Jolla, CA) such that EcoRIlinearized plasmid could be used as a template for in vitro transcription of 32 P-labeled antisense probe; in addition to ϳ400 nucleotides of noncoding SFT sequence, probes contained ϳ30 bases of vector sequence such that complete digestion could be verified by the difference in electrophoretic mobility of the protected fragment. Twenty g of total RNA was hybridized to this probe (5 ϫ 10 5 cpm) for 16 h at 50°C. Nonhybridized probe was digested with a combination of ribonuclease A (40 g/ml) and ribonuclease T1 (2 g/ml) for 60 min at 30°C; nucleases were subsequently digested with proteinase K (125 g/ml) at 37°C for 15 min. After phenol/chloroform extraction, the hybridized probes were precipitated with ethanol and heat-denatured at 85°C for 5 min prior to electrophoresis on a 6% polyacrylamide-8 M urea gel. The gel was then dried and exposed to film for autoradiography. Molecular biology reagents were obtained from Boehringer Mannheim and U. S. Biochemical Corp. The data shown are from individual experiments and represent results obtained on multiple occasions under various experimental conditions for the different cell lines and tissue specimens.

RESULTS AND DISCUSSION
To investigate regulation of SFT expression, Northern analysis was performed with total RNA from HeLa cells grown under low, normal, and high iron conditions (Fig. 1). Treatment with the iron chelater desferrioxamine to deplete cellular iron increases SFT transcript levels, while iron-loading induces down-regulation. As shown in Fig. 2A (top panel), SFT transcripts decline to barely detectable levels within 2 days of culture in medium supplemented with 65 M Fe-NTA. Western analysis confirms the loss of SFT synthesis over precisely the same time frame (Fig. 2B), indicating that levels of the transport protein closely correlate with the amount of its transcript. Down-regulation of SFT message in response to high iron is completely reversible, and basal levels are restored when HeLa cells are returned to normal culture conditions (data not shown; see below).
Transferrin receptor synthesis also is down-regulated in response to iron-loading. When cellular iron content is low, the binding of iron-responsive regulatory protein (IRP) to ironresponsive elements (IREs) present in the 3Ј-untranslated region protects the receptor message against degradation. The loss of IRP mRNA binding activity under high iron conditions leads to destabilization and transcript decay (26). Under our experimental conditions, loss of transferrin receptor mRNA upon iron-loading is significantly slower than the time course observed for SFT mRNA reduction ( Fig. 2A), suggesting that the latter is regulated via different mechanism(s).
HeLa cells were exposed to desferrioxamine in the presence or absence of actinomycin D to study the effects of the transcription inhibitor on the accumulation of SFT mRNA. As shown in Fig. 3, the SFT message is significantly reduced in both control and desferrioxamine-treated cells upon treatment with this drug, indicating that the transcript half-life is Ͻ16 h. Time course experiments revealed a t1 ⁄2 ϳ 6 h for SFT mRNA isolated from cells treated with or without desferrioxamine, confirming no significant change in decay rate upon iron chelation (not shown). Although desferrioxamine-induced up-regulation of SFT mRNA levels appears to be sensitive to actinomycin D, it is important to note that our data do not directly discriminate between transcriptional and post-transcriptional events mediating SFT down-regulation under high iron conditions. However, because the IRE hexanucleotide consensus sequence, which contributes to a stem-loop structure recognized by IRPs, does not appear within the 3Ј-untranslated region of SFT (9), it is unlikely that modulation of SFT synthesis involves stabilization of its message through IRP-controlled mechanisms.
Hemochromatosis is the most common genetic disorder known and results in massive iron deposition in the parenchyma of liver, kidney, pancreas, and other tissues (11,14). Because of the functional role in iron transport implicated for SFT and the observed iron-dependent regulation of its expression, it was of interest to study levels of SFT in liver from hemochromatosis patients. Based on the results discussed above, a prediction would be that SFT transcripts are downregulated in the disease state. Unexpectedly, ribonuclease pro- Briefly, RNA was isolated, electrophoresed on denaturing formaldehyde-agarose gels, and transferred as described under "Experimental Procedures." Randomly primed probes were hybridized overnight at 42°C in 50% formamide containing 5 ϫ Denhardt's solution; the blot was then washed twice with 6 ϫ SSPE-0.5% SDS for 15 min at room temperature, twice in 6 ϫ SSPE-0.25% SDS at 37°C, then once in 0.1 ϫ SSPE-0.1% SDS for 30 min at 65°C. Shown are the autoradiographic results obtained for 32 P-labeled probes using SFT or 36B4 cDNA as templates for random priming; the latter ribosomal protein is expressed constitutively and confirms equivalent loading in all lanes. In addition to the native 1.5-kilobase message observed for SFT, a 2.4-kilobase band is detected as reported previously (9). Levels of the latter are not modulated by iron depletion or iron loading, although the cross-hybridization of this species with the SFT probe suggests homology between these two transcripts.

Expression of SFT 34676
tection experiments reveal that liver samples from hemochromatosis patients (Fig. 4, lanes 1-6) contain 5-fold higher levels of SFT mRNA compared with control individuals (Fig. 4, lanes  7-10). Western blot analysis confirmed the up-regulation of SFT protein in patient liver as well (not shown). One explanation for these observations is that liver cells differentially regulate SFT expression, and to examine this possibility, the properties of SFT expression in HepG2 cells were investigated. This well differentiated human hepatoma cell line synthesizes transferrin, expresses the transferrin receptor, and distributes iron between intracellular pools in a manner similar to liver. Nonetheless, ribonuclease protection experiments confirm the loss of SFT transcripts when HepG2 cells are grown in medium supplemented with 65 M Fe-NTA for 8 days (Fig. 4, lane 11  versus lane 12). This effect is entirely reversible: basal levels of SFT are observed in cells cultured in high iron medium for 4 days, then returned to normal medium for 4 days (lane 13), while under the same regimen, cells first grown in normal medium, then in iron-supplemented medium, have reduced SFT transcript levels. Thus, the pattern of SFT expression in HepG2 cells markedly contrasts with the increased levels of SFT mRNA observed for hemochromatosis patients.
Many studies of hemochromatosis patients have evaluated differences in other molecules involved in iron metabolism, including transferrin, transferrin receptor, ferritin, and IRPs, but strong evidence to support abnormal biosynthetic control of these factors in the disease is lacking. Pietrangelo et al. (27) specifically examined expression levels of factors involved in hepatic iron metabolism and concluded that their regulation was normal in hemochromatosis patients. Nevertheless, these individuals display high levels of transferrin saturation, serum ferritin, and non-transferrin-bound iron; accumulation of iron from the latter fraction may play a major role in tissue damage (28). Clearly, the deposition of non-transferrin bound iron into the liver is consistent with the now well characterized function of SFT to stimulate transferrin-independent uptake of Fe(III) (9,10).
The discovery that SFT expression is up-regulated in the liver of hemochromatosis patients suggests that it contributes to the etiology of this disease. While the observed iron-induced down-regulation of SFT expression in HepG2 hepatocytes argues that the patients may have lost some form of iron-regulated biosynthetic control, our studies do not exclude the pos-  Fig. 1 is presented to demonstrate the time course of down-regulation of SFT transcripts compared with transferrin receptor message (TfR) in HeLa cells. At time 0, culture medium was supplemented with 65 M Fe-NTA, and cells were grown for the times shown prior to RNA isolation. Detection of transferrin receptor transcripts was as described previously (36). B, Western analysis of membrane protein from HeLa cells exposed to 65 M Fe-NTA for the times shown (0 -72 h). Briefly, confluent cells from a 100-mm dish were harvested in phosphate-buffered saline containing 1 mM EDTA, then washed three times with ice-cold phosphate-buffered saline and suspended in 25 mM HEPES, pH 7.4, 85 mM sucrose, 100 mM KCl, and 20 M EGTA. After disruption of the cells by repeated freeze/thaw, postnuclear supernatants were collected and membrane was then isolated by ultracentrifugation with a Sorval RP100AT3 rotor at 95,000 rpm for 15 min at 4°C. One-hundred g of membrane protein was loaded in each lane and electrophoresed on a 6% polyacrylamide gel under nonreducing conditions. After transfer to nitrocellulose, the blot was blocked, washed, and cut to probe with anti-SFT L4 antisera (1:200 dilution) and anti-Na,K-ATPase antisera (1:5000) as indicated in the figure. Following incubation with secondary antibody conjugated with horseradish peroxidase (1:2000), the staining pattern shown was developed by enhanced chemiluminescence (ECL) according to the manufacturer's directions (Amersham Pharmacia Biotech). The anti-L4 antisera has been characterized previously (9). The anti-Na,K-ATPase antisera, which is specific for the ␣-subunit, was a generous gift of Dr. Lorraine Santy (Harvard). The latter can be detected as a broad band on nonreducing gels, but aggregates under reducing conditions. For this experiment, detection of the ␣-subunit serves as a sample-loading control for each gel lane. Expression of SFT 34677 sibility that up-regulation of SFT message levels is indicative of a specific liver detoxification response that is necessary for the clearance of excess iron incorporated due to enhanced dietary absorption. Unfortunately, the recent elucidation of the disease gene, HFE, yields little insight about how the reciprocal relationship between storage and transport of iron is lost in hemochromatosis. HFE is an MHC class I-like molecule that appears to be expressed in nearly all tissues (11); the defective allele contains a mutation that alters Cys282 such that newly synthesized protein can not stably associate with ␤ 2 -microglobulin for appropriate assembly and function (29,30). Consistent with the loss of HFE function, the ␤ 2 -microglobulin knockout mouse displays characteristics of iron overload resembling the human disease (12, 13) as well as those described for the HFE knockout mouse (31). It is known that HFE associates with the transferrin receptor (32,33), and a role for the receptor in HFE trafficking has been suggested (34). Gross et al. (34) further found that cells overexpressing HFE have increased levels of transferrin receptors but reduced levels of ferritin, supporting the idea that HFE modulates the level of iron within the cell. However, whether the absence of this factor alters expression of iron transport and storage factors or whether SFT expression is directly modulated by HFE remains unknown. Duodenal samples from hemochromatosis patients do display greater Fe(III) reducing and uptake activity compared with controls (35); ␤ 2microglobulin knockout mice also show higher intestinal uptake of Fe(III) but not Fe(II) (15). These findings are consistent with the likelihood that a ferrireductase, an Fe(II)/Fe(III) transporter, or a complex contributing both functions is upregulated when HFE function is compromised. Because previous studies have revealed that SFT is ubiquitously expressed with abundant transcript levels in intestinal tissue (9) and that SFT-mediated iron transport involves a ferrireductase activity (10), future investigations should examine whether hemochromatosis patients display abnormal expression of this Fe(II)/ Fe(III) transporter in the duodenum as well as liver.