Introduction

Hepcidin is a hepatic iron regulatory hormone critical for maintaining systemic iron homeostasis. It is secreted into the blood stream by hepatocytes and binds to the iron exporter ferroportin located on the surface of macrophages, hepatocytes, and intestinal enterocytes to cause its internalization and proteolysis [1, 2]. As a consequence, hepcidin reduces intestinal iron absorption and macrophage-mediated iron recycling. Hepcidin levels are inappropriately low in juvenile and adult forms of hereditary hemochromatosis (HH), a disease caused by mutations in HFE, transferrin receptor 2 (TfR2), hemojuvelin (HJV, HFE2), or hepcidin itself [3–5]. These mutations thus lead to increased duodenal iron absorption and macrophage iron release resulting in elevated serum iron levels and liver iron overload [6]. Recently, the pathogenesis of HH was linked to the bone morphogenetic proteins (BMPs) [7], a group of cytokines that belongs to the TGF-ß superfamily. BMPs initiate intracellular signaling by binding to complexes of type I and type II serine/threonine kinase receptors, which once activated phosphorylate intracellular SMAD proteins [8]. These translocate to the nucleus to control gene transcription. The HH protein HJV is a BMP co-receptor that enhances BMP signaling to regulate hepcidin expression and iron metabolism [7]. Consistently, livers of HJV knockout mice present significantly lower levels of basal phosphorylated SMAD1/5/8 than wild-type mice [7]. In addition, a liver-specific conditional knockout of the SMAD4 transcription factor that mediates BMP signaling is associated with low hepcidin expression and iron overload [9].

The discovery of a BMP-responsive element (BMP-RE1) at position −84/−79 of the hepcidin promoter that controls HJV and BMP-2-dependent hepcidin transcription, directly links the HJV/BMP signaling pathway to hepcidin promoter activation [10]. The mutation of BMP-RE1 also severely impairs the activation of hepcidin in response to IL-6 [10]. Additionally, induction of hepcidin mRNA expression by IL-6 is controlled via the JAK/STAT pathway and a STAT-binding site (STAT-BS) located at position −72/−64 of the hepcidin promoter [11–13].

Recently, Truksa et al. identified an additional BMP-responsive region within the distal part of the murine hepcidin promoter, located 1.6 to 1.8 kb upstream of the translation start of the murine hepcidin gene [14]. Interestingly, the same region also controls the hepcidin response to iron overload in mice hydrodynamically transfected with murine hepcidin promoter reporter constructs [15].

In this study, we identify a highly conserved BMP-responsive element (BMP-RE2) which is located in the distal part of the human hepcidin promoter at position −2,255/−2,250. We show that BMP-RE1 and BMP-RE2 function together to control basal hepcidin mRNA expression under control conditions and the hepcidin response to BMP-2 and BMP-6. In contrast to the proximal BMP-RE1, the distal BMP-RE2 does not influence the inflammatory response to IL-6.

Materials and methods

Cell culture

The human hepatocarcinoma HuH7 cell line was cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, high glucose; Invitrogen) supplemented with 10% heat-inactivated low-endotoxin fetal bovine serum (FBS, Invitrogen), 100 U/ml penicillin, and 100 μg/ml streptomycin. Cell cultures were maintained at 37°C under 5% CO2.

RNA isolation, reverse transcription, and quantitative real-time PCR analysis

Total RNA was isolated using the Qiagen RNAeasy kit according to manufacturer’s instruction (Qiagen). The concentration and purity of the RNA was determined by OD260/280 reading. Two micrograms of total RNA were reverse transcribed using 10 μM each dCTP, dGTP, dATP, dTTP, 100 ng random primers, 5× reaction buffer (Fermentas) and 100 U of RevertAid H Minus M-MuLV reverse transcriptase (Fermentas) in a 25-μl reaction for 90 min at 42°C. Real-time polymerase chain reaction (qPCR) was performed as previously described [10]. Amplification reactions were carried out on the ABI Prism 7500 Applied Biosystems (Applera Deutschland) system in a 20-μl volume using SYBR Green I dye and the following amplification conditions: 50°C 2 min, 95°C 10 min, (95°C 15 s, 60°C 15 s) ×40 cycles. Primers were designed to specifically amplify 123 bp of hepcidin cDNA (forward 5′-CTCTGTTTTCCCACAACAGAC-3′, reverse 5′-TAGGGGAAGTGGGTGTCTC-3′), 113 bp of GAPDH cDNA (forward 5′-CATGAGAAGTATGACAACAGCCT-3′, reverse 5′-AGTCCTTCCACGATACCAAAGT-3′), 126 bp of the HJV cDNA (forward 5′-CACCCGGAAGCTCACCAT-3′, reverse 5′-GGTCGGTCACCTCCATTGAT-3′), and 71 bp of the SMAD4 cDNA (forward 5′-TGGCCCAGGATCAGTAGGT-3′, reverse 5′-CATCAACACCAATTCCAGCA-3′). The mRNA/cDNA abundance of each gene was calculated relative to the expression of a house-keeping gene, GAPDH (glyceraldehyde-3-phosphate-dehydrogenase).

Promoter analysis

A restriction fragment containing 2,762 bp (2.7 kb) of the 5′-flanking genomic region of the human hepcidin gene plus its 5′ untranslated region were inserted into the promoterless luciferase reporter vector pGL3_Basic (Promega) to generate WT_2.7kb. The generation of the WT_1kb construct, as well as the deletion of the STAT-binding site (position −72/−64; STAT_BS_1kb) and the mutation of the BMP-responsive element (position −84/−79; BMP_RE1_1kb) were reported previously [10, 11]. Additional luciferase reporter constructs were generated by site-directed mutagenesis (Invitrogen) of the 2.7 kb hepcidin promoter construct using the primer pairs listed in Supplementary Table 1. Deletion of the STAT-binding site (STAT_BS_2.7kb) and mutation of the BMP-RE1 (BMP_RE1_2.7kb) within the context of the 2.7kb promoter fragment were carried out as described for the WT_1kb promoter construct [10, 11]. To generate BMP_RE3_2.7kb the nucleotide sequence GGCGCC at position −2,301/−2,296 bp was exchanged for TACGAA. To generate BMP_RE2_2.7kb the sequence GGCGCC at position −2,255/−2,250 was substituted by TCAGGC. To generate BMP_RE1_BMP_RE2_2.7kb, base substitutions were introduced into the sequence of BMP-RE2 within the construct BMP_RE1_2.7kb. The identity of all constructs was confirmed by DNA sequencing.

Luciferase assays

Hepatic HuH7 cells (1.5 × 105 per well) were seeded onto six-well plates. The next day, 500 ng of pGL3-hepcidin promoter containing reporter vectors were transfected, together with 10 ng of a control plasmid containing the Renilla gene under the control of the CMV promoter (details of the control plasmid are available upon request). Plasmid transfections were performed using Lipofectamine 2000 (Invitrogen) according to manufacturer’s instructions. After 24 h, the cells were harvested for measurement of luciferase activity. Cells were lysed in Passive Lysis Buffer (Promega) and cellular extracts were analyzed using the Dual-Luciferase-Reporter assay system (Promega) and a Centro LB 960 luminometer (Berthold Technologies).

Treatment with BMP-2, BMP-6, and IL-6

Twenty-four hours after transfection with Lipofectamine, cells were treated with human BMP-2 (50 ng/ml, 14 h), BMP-6 (50 ng/ml, 24 h), or IL-6 (2 ng/ml, 24 h). Cells were harvested in PBS. Half of the cells were used for extraction of total RNA and the other half for the measurement of luciferase activity.

siRNA-mediated knockdown, reporter plasmid transfection, and luciferase assay

HuH7 cells were seeded at 30% confluency in 12-well plates, in DMEM supplemented with 10% low endotoxin FBS and antibiotics. After 24 h, cells were transfected using Oligofectamine Reagent (Invitrogen) and 100 nM of siRNA directed against HJV (Dharmacon), SMAD4 (Dharmacon), or scrambled siRNA (Dharmacon) as a control. Forty-eight hours after siRNA transfection, 200 ng of pGL3-hepcidin promoter containing reporter vectors were transfected into the same cells together with a control plasmid containing the Renilla gene under the control of the CMV promoter. Plasmid transfections were performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. After 24 h, the cells were harvested using one half for RNA extraction and the other half for luciferase assays. The efficiency of the knockdown was analyzed at the mRNA level by quantitative real-time PCR.

Statistics

Results are expressed as mean ± SD. Student’s t test was used for estimation of statistical significance.

Results

Two BMP-responsive elements at positions −84/−79 (BMP-RE1) and −2,255/−2,250 (BMP-RE2) of the human hepcidin promoter regulate basal hepcidin mRNA expression under control conditions

A 200 bp sequence within the murine hepcidin promoter located 1.6 to 1.8 kb upstream of the translation start site is critical for BMP-responsiveness and the in vivo response to iron overload [14, 15]. A sequence located between 2.2 and 2.4 kb upstream of the transcription start site of the human hepcidin promoter shares a high degree of homology with this murine region (Fig. 1). Careful inspection of the conserved human promoter fragment revealed two BMP-REs with sequence identity to the BMP-RE1 identified previously at position −84/−79 bp of the human hepcidin promoter (GGCGCC, [10]; Fig. 1a). Whereas the most distal motif, BMP-RE3 (at position −2,301/−2,296) is not fully conserved between species, BMP-RE2 (at position −2,255/−2,250) is highly conserved between human, mouse, rat, and cow (Fig. 1). Importantly, the 6 bp motif is not present elsewhere within the 2.7 kb hepcidin promoter.

Fig. 1
figure 1

Scheme of the luciferase reporter vectors containing the human hepcidin promoter. a Top Sequence alignment of the distal hepcidin promoter (bp 2,200–2,400) revealed the phylogenetically conserved BMP-RE2 at position −2,255/−2,250 bp and a non-conserved BMP-RE at position −2,301/−2,296 bp of the human hepcidin promoter. Bottom The proximal region of the hepcidin promoter contains the previously described BMP-RE1 and STAT-BS motifs [10]. Boxed sequences indicate the BMP-REs and STAT-BS. Base substitutions analyzed experimentally are shown in bold. b Luciferase (Firefly) reporter vectors. The BMP_RE1_1kb and the BMP_RE1_2.7kb constructs contain base substitutions (GGCGCC was exchanged for AGAACC) at position −84/−79 bp within the human hepcidin promoter containing 942 bp (WT_1kb) or 2,762 bp (WT_2.7kb) upstream of the transcription start site, respectively. The BMP_RE2_2.7kb construct contains base substitutions at position −2,255/−2,250 (TCAGGC instead of GGCGCC) of WT_2.7kb. The BMP_RE1_BMP_RE2_2.7kb construct contains base substitutions in both BMP-REs of the WT_2.7kb construct. The BMP_RE3_2.7kb construct contains base substitutions (TACGAA instead of GGCGCC) at position −2,301/−2,296 of WT_2.7kb promoter. The STAT_BS_1kb and the STAT_BS_2.7kb constructs contain a deletion of the STAT-BS (TTCTTGGAA) located at position −72/−64 bp within WT_1kb or WT_2.7kb, respectively

Full size image

To assess the role of these motifs for hepcidin promoter activity, we generated a luciferase reporter (WT_2.7kb) that contains the firefly luciferase coding sequence fused to the entire 2,762 bp intergenic region between the 3’end of the USF2 gene and the transcription start site of the hepcidin gene. The WT_2.7kb and the WT_1kb (containing 942 bp of the hepcidin promoter; [11]) reporters were transfected into the human hepatocyte cell line HuH7 and luciferase activity was analyzed 24 h later. Promoter activity of the WT_2.7kb construct is approximately three times higher than that of the shorter, WT_1kb hepcidin promoter, showing that the distal region of the hepcidin promoter confers high basal promoter activity under control conditions. We next mutated the BMP-RE1 sequence at position −84/−79 bp or deleted the STAT-binding site at position −72/−64 within the context of the WT_2.7kb construct to generate constructs BMP_RE1_2.7kb and STAT_BS_2.7kb, respectively. While the deletion of the STAT-binding site in the context of the 1 kb promoter (STAT_BS_1kb) decreases hepcidin promoter activity (Fig. 2; [10]), deletion of the STAT-BS in the context of the 2.7 kb human hepcidin promoter does not affect luciferase expression. By contrast, mutation of the BMP-RE1 (BMP_RE1_2.7kb) greatly reduces luciferase activity (Fig. 2). Interestingly, the ‘fold change’ between the WT_1kb and the WT_2.7kb constructs and their respective mutant BMP-RE versions is very similar (12.3-fold versus 13.8-fold, respectively), suggesting that BMP-RE1 plays the same critical role in the context of the 2.7 kb promoter as in the context of the 1 kb human hepcidin promoter.

Fig. 2
figure 2

BMP-RE1 and BMP-RE2 mediate high baseline hepcidin promoter activity. HuH7 cells were transfected with wild-type or mutant luciferase reporter vectors and luciferase activity was measured 24 h later. Experiments were performed at least in triplicate and results of the luciferase reporter assay are presented as fold change ± SD of the Firefly/Renilla (F/R) luciferase activities. Statistically significant expression changes between the WT promoter and the mutated promoter of the same length, are marked by an asterisk. Asterisk represents p values <0.05 and double asterisk represents p values <0.005

Full size image

We next investigated whether the newly identified BMP-RE2 (BMP_RE2_2.7kb) or BMP-RE3 (BMP_RE3_2.7kb) motifs affect promoter activity of the human hepcidin promoter. While mutations within the non-conserved BMP-RE3 do not significantly alter luciferase activity, mutation of the conserved BMP-RE2 reduces human hepcidin promoter activity to levels similar to the WT_1kb construct. The BMP-RE2 thus plays a key role in determining the high level promoter activity of the WT_2.7kb promoter. Mutation of both the BMP-RE1 and BMP-RE2 in construct BMP_RE1_BMP_RE2_2.7kb reduces the activity of the human hepcidin promoter even further, to levels similar to those of the BMP_RE1_1kb construct. These data suggest that the two highly conserved BMP-REs 1 and 2 jointly control hepcidin transcription.

BMP-RE1 and BMP-RE2, but not BMP-RE3, mediate BMP signaling

BMP-2, BMP-4, and BMP-6 are all expressed in human liver and are endogenous ligands for HJV in hepatoma-derived cells [16]. BMP-6 expression is also modulated in response to dietary iron levels [17], suggesting a role for this cytokine in the iron-mediated regulation of hepcidin transcription.

We have previously shown that the BMP-2 response of the human hepcidin promoter in HuH7 cells requires an intact BMP-RE1 sequence [10]. To determine whether BMP-RE2 and/or BMP-RE3 are critical for hepcidin promoter activation by BMPs, we analyzed the respective mutant luciferase reporter constructs (Fig. 1) in HuH7 cells. Following transfection, HuH7 cells were treated with 50 ng/ml of BMP-2 and luciferase activity was assessed (Fig. 3). In agreement with our earlier data [10], deletion of the STAT-binding site leads to an increase of the responsiveness to BMP-2, which might be explained by an enhanced accessibility of the BMP-RE to interacting transcription factors. BMP-2-induced hepcidin promoter activation of the WT_2.7kb promoter is elevated more than ninefold compared to the WT_1kb promoter. This BMP-inducibility is partially ablated by the BMP-RE1 or BMP-RE2 mutations, whereas mutation of the BMP-RE3 does not significantly impair the BMP-2 response. Similar to the observations with basal promoter activity, mutation of BMP-RE2 reduces the BMP responsiveness to the levels obtained with the WT_1kb vector, suggesting that BMP-RE2 is the major element regulating the BMP response within the distal region of the human hepcidin promoter. Interestingly, mutations of BMP-RE1 and BMP-RE2 within the hepcidin 2.7 kb promoter almost completely ablate the BMP response (Fig. 3).

Fig. 3
figure 3

BMP-dependent activation of the hepcidin promoter is mediated by BMP-RE1 and BMP-RE2. HuH7 cells were transfected with wild-type and mutant luciferase reporter vectors and treated a with BMP-2 (50 ng/ml for 14 h) or b with BMP-6 for 24 h. Results of the luciferase reporter assays are presented as fold change ± SD of the BMP-treated versus non-treated samples. Statistically significant changes between the WT promoter and the mutated promoter of the same length are marked by an asterisk. A single asterisk represents p values <0.05 and a double asterisk represents p values <0.005

Full size image

Essentially, identical results were obtained when transfected HuH7 cells were treated with BMP-6 instead of BMP-2 (Fig. 3), further corroborating the essential function of BMP-RE1 and BMP-RE2 in mediating the BMP response. These data suggest a common signaling pathway for the BMPs that regulate hepcidin expression at least in cultured cells.

BMP-RE1, but not BMP-RE2 or BMP-RE3, is required for the inflammatory response to IL-6

Activation of the hepcidin promoter by BMPs does not require IL-6 or the STAT-BS that controls the IL-6 response [10, 18]. By contrast, the IL-6 response of the hepcidin promoter requires the BMP-RE1 located close to the STAT-BS in the proximal region of the hepcidin promoter [10]. A cross-talk of the BMP and IL-6 signaling pathways is further supported by the finding that hepatocyte-specific SMAD4-deficiency abolishes hepcidin activation by IL-6 in mice [9].

To investigate whether the BMP-RE2 or BMP-RE3 motifs are involved in the IL-6 response of the hepcidin promoter, we transfected HuH7 cells with WT or mutated luciferase reporter constructs (Fig. 1), and treated the cells with IL-6 (2 ng/ml). As shown in Fig. 4, the WT_1kb and WT_2.7kb constructs are similarly inducible by IL-6. Deletion of the STAT-BS or mutation of the BMP-RE1 effectively ablates the IL-6 response both in the context of the 1 kb and the 2.7 kb hepcidin promoter. By contrast, mutations in BMP-RE2 or BMP-RE3 display no effect on IL-6 inducibility. These findings demonstrate that the inflammatory response involves the BMP-RE1 and STAT-BS in the proximal promoter region, suggesting a specific cross-talk between HJV/BMP/SMAD and IL-6/STAT pathways involving these promoter elements.

Fig. 4
figure 4

BMP-RE1 and the STAT-BS in the proximal region of the hepcidin promoter are required for its IL-6 response. HuH7 cells were transfected with the luciferase reporter vectors and treated with IL-6 (2 ng/ml, 24 h). Results of the luciferase reporter assays are presented as fold change ± SD of the IL-6-treated versus non-treated samples. Statistically significant changes between the WT promoter and the mutated promoter of the same length are marked by an asterisk. A single asterisk represents p values <0.05 and a double asterisk represents p values <0.005

Full size image

BMP-RE1 and BMP-RE2 are critical for the HJV/SMAD4-mediated response of the hepcidin promoter

To determine whether BMP-RE1 and BMP-RE2 represent the sites through which HJV and SMAD4 control hepcidin promoter activity, we performed siRNA-mediated knockdowns of HJV and SMAD4 followed by transfection with luciferase reporter vectors. Consistent with our previous data [10], the knockdown of HJV (to 46% of the control siRNA) decreases basal hepcidin promoter activity of the WT_1kb and WT_2.7kb constructs to 54% and 42%, respectively. For comparison, endogenous hepcidin mRNA expression is reduced to 50%. The knockdown of SMAD4 (to 35% of the control siRNA) decreases luciferase activity to 42% (for WT_1kb), to 15% for WT_2.7kb, and to 32% for endogenous hepcidin mRNA expression, consistent with observations in mice deficient in hepatocytic SMAD4 expression. As observed previously, mutation of the BMP-RE1 in the context of the 1 kb hepcidin promoter (WT_1kb) fully abrogates the decrease in hepcidin promoter activity upon HJV and SMAD4 knockdown. By contrast, mutations within the sequence of the BMP-RE1 or BMP-RE2 in the context of the 2.7 kb hepcidin promoter only partially relieve the consequences of the SMAD4 or HJV knockdown on hepcidin promoter activity. Only when both BMP-RE1 and BMP-RE2 are mutated, hepcidin promoter activity is fully unaffected by HJV knockdown and partially unaffected following SMAD4 knockdown. These findings show a critical role for BMP-RE1 and BMP-RE2 in controlling hepcidin promoter activation by HJV and SMAD4, and suggest that additional elements within the distal region of the hepcidin promoter may control the SMAD4 response.

Discussion

The precise transcriptional control of the iron-regulatory hormone hepcidin coordinates body iron needs in response to iron availability, hypoxia, erythropoietic drive, and inflammatory stimuli. Genetic impairment of hepcidin activators such as the HH proteins HFE, TfR2, or HJV causes inappropriately low hepcidin mRNA expression. As a consequence, membrane expression of the iron exporter ferroportin is high and duodenal iron absorption is increased. Ultimately, tissue iron overload will develop. On the contrary, inflammatory cytokines activate hepcidin mRNA expression which in turn triggers ferroportin protein degradation and reduces both duodenal and macrophage iron export. If the inflammatory condition persists the anemia of inflammation will develop, a common clinical disorder that affects patients with acute and chronic infections, trauma, inflammatory disorders, and malignancies [1]. Because of hepcidin’s critical role in maintaining body iron homeostasis, it is important to understand how hepcidin expression is regulated.

We and others showed that the transcription factor STAT-3 activates hepcidin transcription in response to IL-6 treatment and possibly other inflammatory stimuli via a promoter element located at −72/−64 bp [11–13]. In addition, a bone morphogenetic protein (BMP)-responsive element (BMP-RE) at position −84/−79 bp of the hepcidin promoter directly links the HJV/BMP/SMAD pathway to the hepcidin promoter and also affects its response to IL-6. The HH proteins HFE, TfR2, and HJV together with the BMPs are implicated in the iron-dependent regulation of hepcidin expression, activating hepcidin transcription via the SMAD signaling pathway [3–7].

In this study, we have investigated whether the distal part of the hepcidin promoter contains sequence elements that confer BMP-responsiveness or that are involved in the inflammatory response of the hepcidin promoter. We show that the 2.7 kb hepcidin promoter confers three times higher luciferase activity in non-stimulated cells and mediates nine times higher responsiveness to stimulation of the cells by BMP-2 and BMP-6 compared to the 942 bp promoter region characterized previously [10]. This finding is consistent with previous observations that identify a region within the murine hepcidin promoter located between positions 1.6 kb and 1.8 kb as critical for its BMP response in cultured cells [14] and the hepcidin response to iron availability in mice [15].

Alignment of the mouse and human hepcidin promoters revealed a highly conserved region in the distal part of the human HAMP promoter (−2.2 to 2.4 kb), which precisely corresponds to the BMP/iron-responsive region of the murine hepcidin1 promoter (Fig. 1). Interestingly, the conserved region of the human hepcidin promoter contains two BMP-RE motifs with sequence identity to the proximal BMP-RE [10] that we then renamed BMP-RE1. Only one of the two novel motifs, BMP-RE2, is conserved between the hepcidin promoters of human, mouse, rat, and cow, while BMP-RE3 is not. In addition to BMP-RE1, only the phylogenetically conserved BMP-RE2 is functional: base substitutions within BMP-RE2 confer 3.7 times lower luciferase activity under control conditions (Fig. 2; BMP_RE2_2.7kb) and seven times reduced inducibility of the hepcidin promoter by BMP-2 and BMP-6 (Fig. 3; BMP_RE2_2.7kb). The finding that the BMP-RE3 does not affect hepcidin promoter activity (Fig. 2) is surprising, since the substitution of the STAT-BS in the proximal region of the promoter by a second BMP-RE1 that is sequence-identical to BMP-RE3 clearly increases the basal hepcidin promoter activity, as well as its response to BMP-2 [10]. This finding suggests that the palindromic hexanucleotide of the BMP-RE is essential but insufficient to control hepcidin promoter activity by itself and that surrounding sequences may play a role.

While a hepcidin promoter deficient in BMP-RE1 or BMP-RE2 shows decreased basal expression under control conditions (Fig. 2) and decreased inducibility by BMPs (Fig. 3), the combined mutagenesis of the proximal BMP-RE1 and the distal BMP-RE2 leads to a more drastic, additive reduction (47-fold) of the basal expression level and a complete loss of response to BMP stimuli (Figs. 2 and 3).

Additionally, BMP-RE1 and BMP-RE2 equally contribute to link the signaling pathway triggered by HJV to hepcidin promoter activation, as the knockdown of HJV remains without consequence for a hepcidin promoter lacking functional BMP-RE1 and BMP-RE2. By contrast, the knockdown of SMAD4 still reduces luciferase activity even if the BMP-RE1 and BMP-RE2 are mutated (Fig. 5). This may be explained by roles of SMAD4 in transcriptional hepcidin activation that are independent of HJV. Whether or not this involves additional responsive elements located between 1 kb and 2.5 kb of the hepcidin promoter is currently unclear. It is of note that bioinformatic analysis that applies the motif search program MatInspector [19] detects an additional, putative SMAD4 binding site at position −1,414/−1,406 of the hepcidin promoter.

Fig. 5
figure 5

HJV and SMAD4-mediated hepcidin expression requires both BMP-RE1 and BMP-RE2. a Effect of siRNA-mediated knock down of SMAD4 and hemojuvelin (HJV) on endogenous hepcidin mRNA expression. Hepcidin mRNA expression was assayed in HuH7 cells after transfection with SMAD4-specific, HJV-specific or control (scrambled) siRNAs, followed by transfection with the luciferase constructs 48 h later. Experiments were performed at least in triplicate. Hepcidin mRNA expression was analyzed by quantitative real-time PCR and data were normalized to mRNA expression of a house-keeping gene, GAPDH. Data are presented as fold change, whereby HuH7 cells transfected with the scrambled control siRNA are set to 100%. b Luciferase promoter constructs activity under siRNA treatment. HuH7 cells were transfected with siRNAs specific for SMAD4, HJV, or a control siRNA. Forty-eight hours later, cells were transfected with the luciferase reporter vectors containing the wild-type or mutant hepcidin promoter. Luciferase activity was measured 24 h later. Knockdown experiments were performed at least in triplicate. Results are presented as ratios between the luciferase activities (±SD of Firefly/Renilla) obtained from samples transfected with specific-siRNAs and samples transfected with control siRNA. Significant changes are marked by an asterisk. A single asterisk represents p values <0.05 and a double asterisk represents p values <0.005

Full size image

It is of great interest that the BMP-RE2 is located within the region of the hepcidin promoter that was recently linked to iron-dependent hepcidin regulation in mice [15]. Furthermore, BMP-6, an endogenous ligand for HJV was recently implicated in iron sensing in mice and in communication of this signal to the hepcidin promoter [17]. Our data show that BMP-RE2 and BMP-RE1 are essential for both BMP-6 signaling (Fig. 3) and the HJV/SMAD response (Fig. 5). This supports the hypothesis that BMP-6 mediates the physiological response to changes in iron levels and acts via the HJV/SMAD signaling pathway to regulate hepcidin transcription via BMP-RE1 and BMP-RE2 in the proximal and distal regions of the hepcidin promoter, respectively.

During the preparation of this manuscript, Truksa et al. reported the identification of BMP-RE2 in the distal part of the murine and human hepcidin promoters [20]. Although the experimental approaches chosen by Truksa et al. and us differ, most of the main conclusions obtained by our independent studies mutually confirm each other. For example, Truksa et al. investigated the link between the HJV/SMAD signaling pathway and the hepcidin promoter by protein overexpression of HJV, SMAD4, and SMAD1 in cells transfected with wild-type and mutant hepcidin promoter vectors. While they were unable to see an effect of SMAD4 over expression on hepcidin promoter activity, the response of the hepcidin promoter triggered by HJV and SMAD1 overexpression fully depended on functional BMP-RE1 and BMP-RE2. By contrast, in the study reported here the knockdown of SMAD4 strongly decreased both endogenous hepcidin expression and luciferase counts obtained from the WT_1kb and the WT_2.7kb promoters (Fig. 5), consistent with findings in hepatocyte-specific SMAD4-deficient mice [9]. The reduction of hepcidin promoter activity following SMAD4 knockdown was largely dependent on BMP-RE1 and BMP-RE2. Differences between the two studies regarding SMAD4 may be related to the relative abundance of SMAD4 compared to the receptor-regulated SMADs (R-SMADS), whose heterodimerization with SMAD4 is critical to activate transcription, and which may act as limiting factors in the regulation of hepcidin expression.

In addition to the BMP-RE2, Truksa et al. further analyzed its surrounding promoter region and identified bZIP, COUP, and HNF4α binding motifs that modulate the response of the hepcidin promoter to BMP-2 and BMP-9 and HJV/SMAD1 overexpression. Whether or not these motives can explain the small decrease of hepcidin promoter activity upon SMAD4 knockdown when BMP-RE1 and BMP-RE2 are mutated requires further investigation.

Our study further shows that the inflammatory IL-6 response is exclusively mediated by the proximal region of the hepcidin promoter. Consistent with our previous analysis, mutation of the BMP-RE1 or the STAT-BS leads to a nearly complete ablation of IL-6 responsiveness, both in the context of the 942 bp and the 2.7 kb promoters (Fig. 4), which suggests that BMP and IL-6-mediated control of hepcidin transcription overlap at the point of SMAD4. This finding supports previous data that report an inadequate response to inflammatory stimuli in mice deficient for SMAD4, HJV, or HFE [9, 21, 22]. However, while in the context of the 942 bp promoter deletion of the STAT-BS affects basal transcription of the hepcidin promoter, this effect is lost in the context of the 2.7 kb promoter. We speculate that the formation of a protein complex involving transcription factors bound to BMP-RE1 and BMP-RE2 may alter transcription factor composition within the hepcidin promoter compared to a promoter context that only contains BMP-RE1. Transcription factors that bind to the STAT-BS may act to stabilize a transcription factor complex in the context of the 942 promoter, a function that may be redundant in the context of the 2.7 kb promoter. Analysis in animal models may be useful to further elucidate the function of this element. Compared to previous reports, in this study, the difference in activity between the WT_1kb and the STAT_BS_1kb promoter construct was less pronounced. The diminished difference correlated with the use of a new batch of serum that is required to maintain the cultured cells, suggesting that one of the serum components may be responsible for this effect. All together, these results differ from those obtained by Truksa et al. In their assay system based on HepG2 cells, deletion and mutation of the STAT-BS reduced basal hepcidin levels under control conditions also in the context of the murine 2.5 kb hepcidin promoter. More importantly, the BMP-RE1 did not show a role in mediating the response to IL-6 in the murine hepcidin1 promoter. These discrepant observations might be explained by the use of the human and mouse promoters or the cell lines used.

Taken together, our data uncover the link between the BMP/HJV/SMAD pathway and hepcidin promoter activation by two sequence-identical and highly conserved BMP-REs located in the proximal and distal part of the hepcidin promoter and suggest a possible pathway by which iron levels may activate hepcidin transcription. Our data further show that only the proximal region of the hepcidin promoter controls the inflammatory response and further characterize the link between BMP and IL-6 signaling. As such, these data provide novel insights into the regulation of the iron hormone hepcidin by iron and/or inflammatory stimuli and shed new light on diseases of systemic iron metabolism as well as the anemias of inflammation.