Abstract
Iron overload results in cellular toxicity, tissue injury, organ fibrosis and increased risk of neoplastic transformation. SerpinB3 is a serine protease inhibitor overexpressed in the liver in oxidative stress conditions, able to induce fibrosis and increased risk of malignant transformation. Aim of the present study was to assess the effect of iron overload on SerpinB3 expression in the liver using in vivo and in vitro models.
The expression of Serpinb3 was assessed in the liver of hemojuvelin knockout mice (Hjv−/−), an established model of hereditary hemochromatosis, and of wild type control mice, following dietary or pharmacological iron manipulation. To assess the direct effect of iron in vitro, cell lines were treated with different concentration of hemin or with an iron chelator.
Hepatic Serpinb3 mRNA and protein were highly expressed in Hjv−/− mice, but not in wild type controls fed with a standard diet. Serpinb3 became detectable in wild type mice fed with a high iron diet or injected with iron dextran; these treatments further induced Serpinb3 expression in Hjv−/− mice. Livers expressing Serpinb3 showed a positive staining also for HIF-2α in the same areas. Hemin promoted induction of SerpinB3 mRNA in HeLa and HA22T/VGH cells, but a mild stimulation of SerpinB3 promoter activity in HeLa and Huh7 cells. In conclusion, Serpinb3 is strongly induced by iron in the mouse liver. The molecular link between iron, ROS and SerpinB3 seems to be HIF-2α, which is induced by iron overload and was previously found capable of up-regulating SerpinB3 at the transcriptional level.
Significance of this study
What is already known about this subject?
Iron overload may cause cellular toxicity, organ fibrosis and increased risk of neoplastic transformation.
SerpinB3 is overexpressed in the liver under oxidative stress conditions and has the capacity to induce fibrosis and promote malignant transformation.
HIF-2α is induced by reactive oxygen species (ROS) and can directly activate the SerpinB3 promoter.
What are the new findings?
SerpinB3 is upregulated in the liver of mice with severe iron overload (hemojuvelin knockout mice, a model of hereditary iron overload).
Dietary or parenteral iron loading induced Serpinb3 overexpression also in wild-type control mice.
SerpinB3 expression is induced by iron treatment in vitro.
HIF-2α was also induced in the liver of iron-overloaded mice.
How might these results change the focus of research or clinical practice?
The molecular link between iron, ROS and SerpinB3 seems to be HIF-2α, which is induced by iron overload and was previously found capable of upregulating SerpinB3 at the transcriptional level.
Introduction
Iron overload may cause severe disease and death in patients with hereditary hemochromatosis. The key molecular event that brings to excessive iron accumulation is inappropriate decrease of hepcidin, a liver peptide that acts as the central regulator of iron homeostasis.1 2 Iron-dependent hepcidin expression is mediated by the bone morphogenetic proteins/SMAD (BMP/SMAD) signaling cascade, which involves bone morphogenetic proteins (BMP2 and BMP6), the BMP coreceptor hemojuvelin (HJV) and other cofactors. From the functional point of view, the intact BMP signalling pathway is essential for hepcidin expression,2 and hepatic HJV is indispensable for BMP activation.3
Deregulation of this pathway in hereditary hemochromatosis leads to hepatic iron overload and iron-mediated liver injury. This is primarily caused by overproduction of reactive oxygen species (ROS), catalyzed by excess redox-active iron,4 and can lead to liver fibrosis, cirrhosis and hepatocellular cancer.5
SerpinB3 is a member of the ov-serpin family of serine protease inhibitors. It is expressed in normal epithelial cells and overexpressed in cancerous cells and damaged hepatocytes.6–8 SerpinB3 was recently shown to prevent oxidative stress-induced cell death.9 10 This antioxidant effect was interpreted as the result of SerpinB3 binding to intramitochondrial respiratory complex I, leading to a decrease of complex I activity and mitochondrial ROS generation, resulting in an antiapoptotic outcome. Another study provided a link between SerpinB3, hypoxia and ROS by showing that hypoxia inducible factor 2α (HIF-2α) acts as a direct transcriptional inducer of the SerpinB3 gene in liver cancer cells.10 HIF-2α exerts a central role in coordinating oxygen sensing and is stabilized by hypoxia and ROS.11 Interestingly, HIF-2α regulates transcription of several iron metabolism genes12 13 and is subjected to regulation by iron regulatory protein 1 (IRP1). Thus, IRP1 binds with high affinity to the 5’UTR site of HIF-2α mRNA, inhibiting its translation. The inhibition is relieved in hypoxic and iron-repleted cells, allowing HIF-2α synthesis.14 15
In light of these findings, the aim of the present study was to assess whether iron overload can modulate the expression of SerpinB3 in vitro and in vivo.
Materials and methods
Animals model
Hemojuvelin knockout mice (Hjv−/−), a model of hereditary iron overload (hemochromatosis) were originally provided by Dr N. Andrews16 and backcrossed onto the C57BL/6 background.17 C57BL/6 mice were purchased from Jackson Laboratories. Animals were housed in macrolone cages (up to five mice/cage, 12∶12 hours light–dark cycle: 07:00–19:00; 22°C±1°C, 60%±5% humidity) with free access to food and water, according to standard institutional guidelines. At the age of 6 months, male Hjv−/− and wild-type mice (n=10 for each genotype) were sacrificed; the livers were snap frozen in liquid nitrogen and stored at −80˚C for molecular studies. To assess the possible modulatory effect of iron on SerpinbB3 expression, 10-week-old male C57BL/6 Hjv−/− mice (n=22) and isogenic wild-type controls (n=24) were placed on diets with variable iron content (Harlan Laboratories, Indianapolis, Indiana, USA), or intraperitoneally injected with iron dextran (1 g/kg body weight) once per week for 4 weeks. The standard diet contained 225 ppm iron (2018 Teklad) and the low-iron diet 75–100 ppm iron (TD.05616). The high-iron diet was the standard, enriched with 2% carbonyl iron (TD.09521).17 At the endpoint, the animals were sacrificed by cervical dislocation. All experimental procedures were approved by the Animal Care Committee of McGill University (protocol 4966).
Cell culture
Huh7 (human hepatoma cell line) and HA22T/VGH (liver endothelial cell line) were maintained in Dulbecco’s modified Eagle’s medium 1× with 4.5 g/L glucose, L-glutamine, and sodium pyruvate supplemented with 10% fetal bovine serum (FBS) 1× penicillin/streptomycin, and 1× non-essential amino acids in a humidified incubator at 37°C with 5% CO2. HeLa cells (epitheloid cervix carcinoma cell line) were cultured in the same manner, excluding 1× non-essential amino acids. To determine iron-dependent regulation, cells were treated overnight with 50 µM hemin (Sigma), or with 100 µM of the iron chelator deferoxamine (DFO, Pharma Science).
Quantitative real-time (qRT)-PCR
Total RNA was extracted using RNase Trizol (Invitrogen, Carlsbad, California, USA), according to the manufacturer’s instructions. After determination of the purity and the integrity total RNA, complementary DNA synthesis, qRT-PCR reactions were carried out as previously described.18 To characterize the mouse-homologous serpin isoforms, the following set of conserved primers was used: sense: 5′-TTTTACACAAGTCCTTTGTGGAGG-3′, antisense: 5′-CTGGACACATGGAAGAGACACCAC-3′, which allows the amplification of the mouse Serpinb3a, -b3b and -b3c, homologous to human SerpinB3 and SerpinB4.19 To characterize the expression of human SerpinB3 in human cell lines the following set of specific primer was used: sense: 5'-aactcctgggtggaaagtcaa-3'; antisense: 5'- accaatgtggtattgctgccaa-3'. β-Actin (for mouse models) and GAPDH (for human cell lines) were used as internal reference and coamplified with target samples using identical qRT-PCR conditions.
Immunohistochemistry (IHC)
IHC was performed on paraffin liver sections specimens. Briefly, paraffin sections (2 µm thick), mounted on poly-L-lysine coated slides, were incubated with polyclonal antibody specific anti-SerpinB3 protein (Hepa-Ab, kind gift of Dr Fassina, Xeptagen, Venice, Italy) or against HIF-2α (Novus Biologicals, Cambridge, UK), as previously described.10 After blocking endogenous peroxidase activity with 3% hydrogen peroxide and performing microwave antigen retrieval, primary antibodies were labelled by using EnVision, HRP-labeled System (DAKO) and visualized by 3′-diaminobenzidine substrate. For negative controls, the primary antibodies were replaced by isotype-matched and concentrations-matched irrelevant antibody.
Promoter activity
To investigate the mechanism of induction, an in vitro reporter assay to test the iron responsiveness of the SerpinB3 promoter was generated by luciferase assay. Huh7 or HeLa cells were seeded at a density of 7.5×104 cells/mL or 1.5×105 cells/mL, respectively, in a 12-well dish in media lacking antibiotics and incubated overnight. Transient transfections were performed the next day using Lipofectamine2000 (Invitrogen) as per the manufacturer’s protocol for a 24-well dish. To generate the promoter constructs of SerpinB3, fragments immediately upstream of the transcription start site of SerpinB3 (online supplementary figure 1) were cloned into the promoterless pGL3 basic luciferase (luc) reporter vector. Luminescence from luciferase was measured and normalized to renilla luminescence using the Dual Luciferase Reporter Assay (Promega). Values were reported as percentages of luc activity. The 2.7 Kb Hepc-pGL3 plasmid, containing 2.7 Kb of the hepcidin promoter cloned into the pGL3 basic vector, used as positive control, was kindly provided by Dr M. Muckenthaler.20
Supplementary file 1
Western blot
Cells proteins were extracted on ice using RIPA (50 mM Tris-HCL pH 7.5, 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, and 1% Triton X-100) and boiled at 95°C for 5 min in sample buffer (62.5 mM Tris-HCL pH 6.8, 50% glycerol, 2% SDS, and 50 mM DTT) as previously described.21 Sodium dodecyl sulphate - polyAcrylamide gel electrophoresis (SDS-PAGE) gels of 10%–15% were used to separate proteins, which were transferred onto nitrocellulose membranes (BioRad) and blocked in 5% milk TBS-T (Tris Buffered Saline – tween 20) for 1 hour at room temperature. Membranes were probed overnight at 4°C with primary antibodies: anti-β actin (1:5000, Sigma-Aldrich, Milan, Italy), heme oxygenase 1 (HO-1, 1:1000, Thermo Fisher Scientific, Milan, Italy), ferritin (1:1000, Novus Biologicals, Cambridge, UK), SerpinB3 (Hepa-Ab, 8 µg/mL, Xeptagen, Venice, Italy), HiF2 (1:500, Novus Biologicals, Cambridge, UK) in 5% milk TBS-T. Membranes were washed for 3×10 mins with fresh tris-buffered saline-Tween 20 (TBS-T) and incubated with appropriate HRP-conjugated secondary antibodies, washed and the detection of peroxidase-coupled antibodies was performed with the enhanced chemiluminescence method (Perkin-Elmer). Sample loading was evaluated by reblotting the same membrane with β-actin antibody.
Detection of intracellular generation of ROS
Huh7 cells transfected with empty vector (Huh7/pCMV6, control) or Huh7 cells overexpressing HIF-2α were seeded in 35×10 mm culture dishes (5×105 cells per well), for 24 hours. For comparative purposes, control cells were treated with 50 µM H2O2 under normoxic conditions for 3 hours (as positive control).
Intracellular generation of ROS was detected evaluating the conversion of 2′,7′-dichlorodihydrofluorescein diacetate (used at 5 µM concentration), once taken up by cells and deacetylated by esterase, into the corresponding fluorescent derivative21 in flow cytometric analysis. The samples were analyzed by flow cytometer (BD Accurri C6); 10 000 events were collected per sample. Results were provided as mean±SD of two experiments conducted in triplicate.
Statistical analysis
Data are expressed as mean±SE mean (SEM). Analysis of multiple groups was performed with one-way analysis of variance. A probability value of p<0.05 was considered as statistically significant. The calculations were carried out with GraphPad InStat Software (San Diego, California, USA).
Results
SerpinB3 is upregulated in livers of Hjv−/− mice in an iron-dependent manner
To explore the effects of genetic iron overload in the in vivo expression of Serpinb3, we evaluated Serpinb3 mRNA levels in livers of Hjv−/− mice and isogenic wild-type control mice (wt) on a standard diet. The data in figure 1A show a 15-fold increase in expression of Serpinb3 in Hjv−/− livers compared with controls (p<0.01) (figure 1A).
To assess possible effects of dietary or pharmacological iron on Serpinb3 regulation, wt and Hjv−/− mice were fed with standard, low or high iron diets or injected intraperitoneally with iron dextran (to bypass the hepcidin-controlled enterocyte gate). In wt mice, liver Serpinb3 mRNA became detectable only in animals fed with high iron diet or intraperitoneally injected with iron dextran. In Hjv−/− mice, Serpinb3 mRNA expression was already detectable in the liver of mice fed with the low diet and progressively increased proportionally to dietary iron content. Iron dextran injection induced Serpinb3 mRNA similar to high iron diet (figure 1B).
Immunohistochemical analysis in serial paraffin sections was consistent with Serpinb3 mRNA expression, showing that hepatic SerpinB3 immunopositivity was virtually absent in sections from control WT mice, increased in WT mice following treatment with iron dextran and was strongly exacerbated in the liver of Hjv−/− mice fed with iron supplemented diet or treated with iron dextran (figure 2).
HIF-2α and Serpinb3 in human liver cells and in livers of iron-overloaded mice
In order to confirm the recent study describing SerpinB3 as a novel hypoxia-responsive and ROS-responsive protein in the liver,10 the tumor liver cell line Huh7, stably transfected to overexpress HIF-2α, was used. At protein level, we observed an increase of SerpinB3 (figure 3A), associated with intracellular generation of ROS (figure 3B) in cells overexpressing HIF-2α, compared with parental control cells.
Considering that HIF-2α mRNA translation is derepressed by iron15 and that HIF-2α, in turn, can induce the expression of SerpinB3,10 immunohistochemistry analysis was carried out to investigate possible coexpression of these two molecules in the liver, relative to iron accumulation. These analyses revealed the presence of Serpinb3 and HIF-2α in the same areas of the livers with high iron deposition (figure 4).
SerpinB3 mRNA is upregulated by iron
To explore if SerpinB3 transcription was iron-dependent in vitro, different cell lines were treated overnight with 50 µM hemin or with 100 µM DFO. Hemin treatment of HeLa cells induced a sixfold induction of SerpinB3 mRNA, while no significant change was observed with DFO treatment (figure 5A). A similar trend was observed in the HA22T/VGH cells, although SerpinB3 increase was not statistically significant (figure 5B). Also in this case DFO did not modify SerpinB3 mRNA expression and this is likely to reflect low cellular iron levels.
As expected, hemin administration to HeLa cells promoted induction of HO-1, the enzyme-catalyzing heme degradation, and ferritin, the intracellular iron storage protein (figure 5C).
Iron stimulates the Serpinb3 promoter
To determine whether iron directly modulates Serpinb3 mRNA transcription, we used Huh7 and HeLa cells transfected with a luciferase indicator under the control of a 5 kb Serpinb3 promoter fragment (onlinesupplementary figure 2). Overnight treatment with 50 µM hemin triggered 1.5-fold induction of Serpinb3 promoter activity in both Huh7 and HeLa cells (figure 6). DFO did not elicit any significant effect of Serpinb3 transcription, in agreement with the Serpinb3 mRNA data shown in figure 5.
Supplementary file 2
It is worth noting that the 2.7 Kb Hepc promoter (containing 2.7 Kb of the hepcidin promoter cloned into the pGL3 basic vector), used as positive control, showed 15-fold increased activity in Huh7 compared with HeLa cells, consistent with the fact that hepcidin is predominantly expressed in hepatocytes.
Taken together, these findings uncover the role of iron as an inducer of Serpinb3 transcription.
Discussion
Iron is vital for normal cellular physiology, but iron overload results in cellular toxicity, tissue injury, organ fibrosis and increased risk of neoplastic transformation.22 Since SerpinB3 is profoundly involved in liver fibrogenesis18 23 and cancer development,24 we investigated whether iron affects SerpinB3 expression. In the present study, we demonstrate that SerpinB3 is induced by iron accumulation in the liver. Accordingly, we observed high SerpinB3 mRNA and protein expression in the livers of Hjv−/− mice, a model of hereditary hemochromatosis. Furthermore, liver expression of SerpinB3 was positively regulated by dietary or parenteral iron in both Hjv−/− and wild-type mice. In vitro experiments further showed that iron induces SerpinB3 transcription. Nevertheless, the degree of transcriptional induction of SerpinB3 in iron-treated cultured cells was substantially lower to that observed in iron-loaded mouse livers. Thus, our data suggest the involvement of additional mechanisms underlying iron-dependent SerpinB3 induction in vivo.
The iron-induced expression of SerpinB3 in the liver was associated with a parallel increase of HIF-2α expression in the same histological areas. A recent study reported that HIF-2α directly binds to SerpinB3 gene promoter, highlighting SerpinB3 as a novel hypoxia-responsive and ROS-responsive protein in the liver, especially in primary liver cancer.10 Based on these findings, we speculate that iron could indirectly promote the robust SerpinB3 induction observed in vivo via the generation of ROS and subsequent HIF-2α stabilization, leading to SerpinB3 transcriptional activation. This hypothesis was supported by the increase of ROS and SerpinB3 found in Huh7 cell line genetically modified to overexpress HIF2-α. The increased expression of SerpinB3 in response to iron overload might contribute to hepatocyte survival9 and fibrosis deposition.18 23 Along this line, hepatic iron overload was found to increase chemically induced liver fibrogenesis by promoting an oxidative injury and, similarly to SerpinB3, premature induction of profibrogenic cytokines, especially transforming growth factor beta 1 (TGF-β1), endothelin-1 and platelet-derived growth factor (PDGF).25
The molecular link between iron, ROS and SerpinB3 seems to be HIF-2α, which is induced by iron overload, and it is capable of upregulating SerpinB3 at the transcriptional level.10
This suggested pathway could represent a novel biological mechanism, allowing cell survival under hypoxic and oxidative stress conditions, like iron overload, and supports the pathophysiological involvement of SerpinB3 in different chronic diseases leading to progressive fibrosis and, eventually, to neoplastic transformation.
Footnotes
Contributors CT, PK, PP, GS, and KP conceived and designed the experiments; CT, PK, SC and EM performed the experiments; CT, LT, AB, and DS analyzed the data; CT wrote the paper; PP, MP, and KP revised the paper. All Authors had final approval of the submitted manuscript.
Funding This work obtained financial support by the Associazione Italiana per la Ricerca sul Cancro and the Canadian Institutes for Health Research.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Correction notice This article has been corrected since it was published Online First. The affiliation for the first author has been updated to Veneto Institute of Oncology, IOV - IRCCS, Padua, Italy.