Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
Hepcidin, the iron hormone, is produced by the liver in response to iron and inflammation. Its synthesis during inflammation is triggered by cytokines, but the details of iron activation are obscure. We tested the role of Kupffer cells and macrophages by studying iron-loaded or inflamed mice with selective inactivation of Kupffer cells or the in vitro effect of conditioned human macrophages on hepcidin expression. Hepcidin messenger RNA (mRNA) expression was studied by Northern blot and reverse transcriptase polymerase chain reaction analysis in mice that were treated with 40 mg/kg gadolinium (III) chloride (GdCl3) as a Kupffer cell inactivating agent and subjected to inflammatory challenges with either lipopolysaccharide (LPS) and turpentine or iron overload by iron-dextran administration. Similar analyses were performed in human hepatoma cells (HepG2) cultured with medium from LPS- or iron-conditioned macrophages from blood donors or patients with HFE-linked hereditary hemochromatosis (HH). In vivo, LPS and particularly turpentine stimulated hepcidin mRNA expression, and this effect was prevented by the inactivation of Kupffer cells. Also, iron overload markedly upregulated hepatic hepcidin mRNA, but this activity persisted in spite of Kupffer cell blockade. In vitro, the medium of LPS-treated normal or hemocromatotic macrophages turned on hepcidin expression. On the contrary, medium of iron-manipulated macrophages, regardless of their HFE status, did not affect hepcidin mRNA steady-state levels. In conclusion, Kupffer cells are required for the activation of hepcidin synthesis during inflammation, and HH inflamed macrophages are capable of mounting a normal response, eventually leading to hepcidin stimulation. However, both Kupffer cells and human macrophages are dispensable for the regulatory activity exerted by iron on hepatic hepcidin. (HEPATOLOGY 2005;41:545–552.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Hepcidin, the product of the HAMP gene, was independently discovered by two groups searching for novel antimicrobial peptides.1, 2 It is produced almost exclusively by the hepatocytes in response to iron3 and inflammatory stimuli.3, 4 A positive correlation between hepcidin expression and iron burden5–7 or inflammatory states7 has also been reported in humans.
Studies in transgenic mice indicate that hepcidin is the principal downregulator of the transport of iron across the small intestine and its release from macrophages. Mice with genetic disruption of the Upstream Stimulatory Factor 2 (USF2) that fail to express either copy of the duplicated hepcidin genes immediately downstream of USF28 and mice with genetic disruption of the CCAAT/enhancer-binding protein α (C/EBPα)9 required for hepcidin transcriptional control develop iron overload resembling hemochromatosis. Transgenic animals that overexpress hepcidin die perinatally from severe iron deficiency anemia occurring in the context of reticuloendothelial cell iron overload.10 Hepcidin is currently seen as a key pathogenic factor in both anemia of inflammation and hemochromatosis.11
In both mice and humans the stimulation of hepcidin synthesis by inflammatory stimuli seems to be indirect and mediated mainly by the inflammatory cytokine interleukin 6 (IL-6) produced by macrophages.7, 12, 13 It is controversial whether HFE, the hemochromatosis gene product, is required for hepcidin activation in response to inflammatory stimuli in mice.13–15
On the other hand, the details of the stimulatory effect of iron on hepcidin synthesis are entirely obscure. Hepatocytes appear to respond to iron in vivo,6 but exposure of cultured hepatocytes to ferric iron3 or iron-saturated transferrin7 does not increase hepcidin messenger RNA (mRNA), and it may even reduce it. Thus, the role of iron sensing may fall to other cells. The key role of macrophage-derived IL-6 in the induction of hepcidin during infection has led to the suggestion that reticuloendothelial macrophages (e.g., Kupffer cells) might be good candidates.16 We have tested this hypothesis by studying hepcidin expression in mice with selective inactivation of Kupffer cells and in cultured hepatoma cells exposed to conditioned medium of human macrophages. We used gadolinium (III) chloride (GdCl3) as a specific Kupffer cell–blocking agent in vivo. This agent inhibits phagocytosis and selectively eliminates liver macrophages while reducing cytokine and free radical production by Kupffer cells in response to injury.17–19 The results indicate that Kupffer cells in vivo and macrophages ex vivo are essential for hepcidin expression in hepatocytes in response to inflammatory stimuli but are not required for iron activation. This finding has important implications for understanding the role of hepcidin in physiology but also in pathophysiological states such as hereditary hemochromatosis (HH).
BALB/c male mice (ages 6-10 weeks) weighing 22 ± 3 g were fed a standard chow diet in pellets and were allowed free access to water. All animals received humane care according to the criteria outlined in the “Guide for the Care and Use of Laboratory Animals” published by the National Institutes of Health (NIH publication 86-23, revised 1985).
Ten mice received one intramuscular injection of 1 g/kg dextran, and after 24 hours received one intraperitoneal injection of either pyrogen-free phosphate-buffered saline (n = 5; group A1) or 40 mg/kg GdCl3 (n = 5; group A2). Twelve mice received one intramuscular injection of 1 g/kg iron-dextran, and after 24 hours received one intraperitoneal injection of either saline (n = 6; group B1) or 40 mg/kg GdCl3 (n = 6; group B2). All animals were sacrificed three days after the GdCl3 treatment for analysis.
We used two different strategies to induce inflammation: lipopolysaccharide (LPS) or turpentine injections. Ten mice received one intraperitoneal injection of saline and after three days received an additional intraperitoneal injection of either saline (n = 5; group C1) or 0.1 mg/kg body weight LPS (Escherichia coli serotype 0111:B4, Sigma, St. Louis, MO) (n = 5; group D1) and were sacrificed for analysis after 90 minutes. Twelve mice received one intraperitoneal injection of 40 mg/kg GdCl3 and after three days received an additional intraperitoneal injection of either saline (n = 6; group C2) or 0.1 mg/kg body weight LPS (n= 6; group D2) and were sacrificed for analysis after 90 minutes.
Tor the turpentine treatment groups, 10 mice received one intraperitoneal injection of saline and after 2 days received an additional injection of either saline (n = 5; group E1) or 0.1 mL/20 g body weight injection of turpentine oil (Sigma) in the interscapular fat pad (n = 5; group E2) and were sacrificed after 24 hours for analysis. Twelve mice received one intraperitoneal injection of 40 mg/kg GdCl3 and after 2 days received an additional interscapular fat pad injection of either saline (n = 5; group F2) or 0.1 mL/20 g body weight injection of turpentine oil (n = 5; group F1) and were sacrificed after 16 hours for analysis.
In blood samples from all animals serum iron and transaminases were measured using a spectrophotometric assay. Serum ferritin was measured by a turbidimetric assay in an automated laboratory device (Modular Analytics System, Roche, Basel, Switzerland) after calibration with recombinant mouse ferritin (kindly provided by P. Arosio, University of Brescia, Brescia, Italy).
Statistical differences of results between various treatment groups and relevant control groups were analyzed using the Student t test.
Cell Culture Studies
Informed consent in writing was obtained from each subject participating in the study, and the study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the review committee at the University of Modena and Reggio Emilia.
Isolation and Culture of Human Monocytes.
Monocytes from eight healthy blood donors (2 women and 6 men; age 35 ± 10 years; serum ferritin < 250 ng/mL; transferrin saturation < 35%) and from 6 individuals with untreated HFE-related HH (homozygotes for the C282Y HFE genotype) and clinically expressing the disease (1 women and 5 men; age 39 ± 10 years; serum ferritin > 1500 ng/mL; transferrin saturation > 90%) were obtained by gradient centrifugation of peripheral white blood cells as previously described.20 None of the subjects suffered from anemia or inflammatory disorder at the time of monocyte isolation. Lymphocytes were depleted by complement mediated lysis using a monoclonal antibody to CD3 (clone HIT3a, PharMingen, San Diego, CA) and rabbit complement (Cederlane, Hornby, ON, Canada). Monocytes (∼2.5 × 106 cells/mL) were maintained in culture in RPMI 1640 medium containing 2 mmol/L glutamine, antibiotics, and 20% heat-inactivated human serum, 5% CO2 at 37°C. To induce differentiation to macrophages, 100 μg/mL granulocyte-macrophage colony-stimulating factor (GM-CSF) (PeproTech EC Ltd., London, UK) was added on day 2 of culture. Medium and all reagents were endotoxin free. The macrophage-conditioned medium (CM) was prepared by incubating monocytes on day 4 with either 0.1 μg/mL LPS or 10 μmol/L ferric ammonium citrate (FAC; Sigma) or 20 μmol/L deferoxamine (DFO; Sigma) for 24 hours and collecting the cell-free supernatant.
Isolation and Culture of Mouse Hepatocytes.
Hepatocytes were isolated by enzymatic dissociation from the liver of normal BALB/cJ mice. Briefly, after cannulation of the portal vein the liver was first perfused with calcium-free buffer (154 mmol/L NaCl, 5.6 mmol/L KCl, 5.5 mmol/L Glucose, 25 mmol/L NaHCO3, 20 mmol/L Hepes), then with the same buffer (0.33 mmol/L, pH 7.6) containing calcium chloride (0.5 mmol/L) and collagenase (0.04%) for 5 to 10 minutes, at 8 mL/min. After enzymatic digestion, livers were removed and transferred to a Petri dish containing 25 mL of Hepatozime-S medium (GIBCO-BRL Life Technologies, Gathersburg, MD). Dissociated cells were isolated by gentle rubbing and hepatocytes were obtained by centrifugation (500 rpm; 2 minutes). Mouse hepatocytes were seeded in a Hepatozime-S medium supplemented with 100 IU/mL penicillin, 30 mg/mL streptomycin sulfate, and 10% fetal calf serum. Four hours later, medium was renewed by an identical medium. Thereafter, hepatocytes were maintained in culture for 2 days.
Human Hepatoma Cell Culture.
Human hepatoma cells (HepG2) were cultured in Minimum Essential Medium containing 10% fetal calf serum and antibiotics at 37°C, 5% CO2. Cells were then incubated for 24 hours in the presence of either 10 μmol/L FAC, 20 μmol/L DFO, 0.1 μg/mL LPS alone, or medium from macrophages obtained from healthy blood donors (CM-C) or patients with HH (CM-HH) conditioned with LPS, FAC, or DFO at 12.5% final concentration.
Both HepG2 and primary mouse hepatocytes were cultured for 24 hours in the absence or presence of various concentrations of GdCl3 (0.01-1 mmol/L) and processed for hepcidin mRNA expression (see below) or cell viability using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. The MTT assay is based on the reduction of the tetrazolium salt MTT into a crystalline blue formazan product by the cellular oxidoreductases.21 Therefore, the amount of formazan produced is proportional to the number of viable cells. Cells were incubated for the last 4 hours of the gadolinium treatment with a solution of MTT (0.5 mg/mL). Then this solution was removed, the resulting blue formazan was solubilized in DMSO, and the optical density was read at 540 nm using a microplate reader (Fluostar-Galaxy, BMG Labtech GmbH, Offenburg, Germany).
RNA Isolation and Analysis.
RNA was prepared using TRIzol reagent according to the manufacturer's instructions (Invitrogen, Carlsbad, CA).
Northern Blot Analysis.
Total RNA (15 μg per lane) was separated on a 1.2% agarose formaldehyde gel, transferred to nylon membrane (Amersham Biosciences, Piscataway, NJ), and ultraviolet light crosslinked. Human and mouse hepcidin cDNA were generated by polymerase chain reaction (PCR) from human and mouse hepatic RNA with the following primers: human, forward 5′GCACGATGGCACTGAGCTCC3′ and reverse 5′TAGGTTCTACGTCTTGCAGC3′; mouse, forward 5′ACCATGGCACTCAGCACTCG3′ and reverse 5′GCGGCTCTAGGCTATGTTTTG3′ and cloned in pcDNA3.1/ V5/His-TOPO. Hybridization was performed at 42°C with random-primed 32P-labeled cDNA probes for hepcidin and β-actin.22
The cDNA was generated by reverse transcription of 5 μg of total mouse liver RNA (for in vivo mouse studies), primary mouse hepatocytes (for cell culture studies), or HepG2 RNA (for cell culture studies), with 100 ng random examer (Roche, Mannheim, Germany), 250 μmol/L dNTPs (Promega Corp., Madison, WI) and 200 units M-MLV Reverse Transcriptase (Promega Corp.) in 1× reverse transcriptase (RT) buffer for 1 hour at 42°C. Expression of human or mouse hepcidin, GAPDH, IL-6 and tumor necrosis factor α (TNFα) were analyzed using Platinum SYBR Green SuperMix (Invitrogen, Carlsbad, CA). The primers were as follows: human hepcidin, forward 5′TGTTTTCCCACAACAGACGGG3′ and reverse 5′CGCAGCAGAAAATGCAGATGG3′; mouse hepcidin, forward 5′GCCTGTCTCCTGCTTCTCCT3′ and reverse 5′GCTCTGTAGTCTGTCTCATCTGTT3′; human GAPDH, forward 5′GGACCTGACCTGCCGTCTAG3′ and reverse 5′TAGCCCAGGATGCCCTTGAG3′; mouse GAPDH, forward 5′CAATGTGTCCGTCGTGGATCT3′ and reverse 5′GTCCTCAGTGTAGCCCAAGATG3′; mouse IL-6, forward 5′CAATTCCAGAAACCGCTATGAAG3′ and reverse 5′GACTTGTGAAGTAGGGAAGGC3′; mouse TNFα, forward 5′ATGGCCTCCCTCTCATCAGTT3′ and reverse 5′GGCTACAGGCTTGTCACTCG3′. Amplification was performed at 60°C to 64°C for 45 cycles in iCycler Thermal Cycler (Bio-Rad, Hercules, CA), and data were analyzed using iCycler iQ Optical System Software. The relative expression in each sample was calculated by a mathematical method based on the real-time PCR efficiencies using as references GAPDH mRNA.23 All samples were assayed in triplicate. After 45 amplification cycles, threshold cycle values were automatically calculated, and femtograms of starting cDNA were calculated from a standard curve covering a range of four orders of magnitude. Both hepcidin and GAPDH standard curves ranged from 1 to 1000 femtograms per 25-μL reaction. Ratios of hepcidin to GAPDH starting quantity were calculated. Statistical differences of results between various treatment groups and relevant control groups were analyzed using the Student t test.
To test the role of Kupffer cells in hepcidin activation we used the GdCl3 method in inflamed and iron-loaded mice. In order to rule out the possibility that GdCl3 depressed hepcidin mRNA expression per se or through a direct toxic effect on hepatocytes, we tested the response of human and mouse hepatocytes to increasing concentrations of GdCl3. As no reliable method to measure serum GdCl3 exists, it was calculated that a steady-state extracellular concentration in the range of approximately 50 to 100 μmol/L would result from the in vivo dose of 40 mg/kg.24 We found that expression of hepcidin mRNA in both mouse and human hepatocytes was preserved up to 250 μmol/L GdCl3 concentration, and declined thereafter (Fig. 1). Hepatocyte viability as assessed by MTT assay was significantly affected only at 1 mmol/L GdCl3 (Fig. 1).
Kupffer Cells Are Not Required for Hepcidin Activation During Iron Overload In Vivo.
We had preliminarily evaluated the time course of Kupffer cell inactivation by analyzing hepatic TNFα and IL-6 mRNA steady-state levels at days 1, 2, 3, 4, 5, and 10 after GdCl3 dosing. In agreement with published data,24 we found a marked drop of TNFα (50%-70%) and IL-6 (60%-75%) mRNA at day 1 through day 3 and complete recovery after day 5. We therefore performed all mRNA analyses at day 3 after GdCl3 injection. Northern blot and quantitative RT-PCR data of LPS-, turpentine-, and iron-treated mice are reported in Figs. 2 and 3, respectively. Iron overload in vivo greatly stimulated hepcidin mRNA expression (compare lanes 3 to 1 in Fig. 2 and column B1 to A1 in Fig. 3A). However, under these circumstances inactivation of Kupffer cells did not abrogate hepcidin upregulation by iron (compare lane 4 to 2 in Fig. 2 and column B2 to A2 in Fig. 3A). Similar results were obtained when iron-dextran was administered at the time of GdCl3 injection or one day after (data not shown).
LPS injection led to a significant rise in hepcidin mRNA expression (compare lane 7 to 5 in Fig. 2 and column D1 to C1 in Fig. 3B), and pretreatment of animals with GdCl3 completely prevented activation (compare lanes 8 to 6 in Fig. 2 and column D2 to C2 in Fig. 3B). The effect of turpentine on upregulation of hepcidin expression was even more pronounced than that of LPS because a 3- to 5-fold hepcidin mRNA increase was consistently found in inflamed animals (compare lane 11 to 9 in Fig. 2 and column F1 to E1 in Fig. 3C). Nevertheless, pretreatment with GdCl3 appeared to significantly prevent the increase of hepcidin mRNA induced by turpentine (compare lane 12 to 10 in Fig. 2 and column F2 to E2 in Fig. 3C).
Figure 3 also reports serum chemistry data collected in various experimental groups. Serum iron and ferritin levels were significantly increased by iron regardless of GdCl3 treatment (Fig. 3A). As expected, in inflamed animals, particularly in those treated with turpentine, hypoferremia developed (Fig. 3B-C). This effect is likely mediated by hepcidin upregulation and consequent iron trapping in macrophages and enterocytes. GdCl3 treatment reverses, at least in part, the hypoferremia (Fig. 3B-C). Interestingly, transaminase levels are significantly increased in inflamed mice, and this effect is conteracted by GdCl3 pre-treatment (Fig. 3B-C). This suggests that inactivation of Kupffer cells with GdCl3 greatly limits the cytotoxic effect exerted on hepatocytes by macrophage-released inflammatory and necrogenic mediators (particularly evident in turpentine-treated animals).
Interestingly, GdCl3 treatment per se seemed to decrease the steady-state level of hepcidin mRNA expression also in control animals from various treatment groups (see both Northern blot and RT-PCR data in Figs. 2 and 3), suggesting that Kupffer cell/macrophage factors sustain a basal activity of hepatic hepcidin gene expression. However, iron overload overrides this effect, leading to a comparable activation of hepcidin mRNA in mice treated with iron alone or iron and GdCl3 (compare lanes 3 and 4 to 1 and 2 in Fig. 2 and columns B1 and B2 to A1 and A2 in Fig. 3A).
Iron-Conditioned Human Macrophage Medium is Unable to Turn on Hepcidin mRNA Expression.
To test the potential modulatory effect of factors released by human macrophages on hepcidin expression, we used in vitro–differentiated primary human monocyte-macrophages as a surrogate experimental model for human tissue macrophages. To this end, human hepatoma (HepG2) cells were incubated in the presence of culture medium from iron- or deferoxamine- or LPS-conditioned human macrophages obtained from healthy subjects or subjects with circulatory and tissue iron overload (HH patients). Northern blot and quantitative RT-PCR data are reported in Figs. 4 and 5, respectively. CM-LPS from healthy blood donors significantly increased hepcidin mRNA expression (compare lane 2 to 1 in Fig. 4 and column 2 to 1 in Fig. 5). On the other hand, CM-FAC or CM-DFO did not modify appreciably the baseline level of hepcidin mRNA expression (compare lanes 3 and 4 to 1 in Fig. 4 and columns 3 and 4 to 1 in Fig. 5). Interestingly, we obtained similar results with CM of HH macrophages: CM of inflamed macrophages from patients with HH readily upregulated hepcidin mRNA expression (compare lane 6 to 5 in Fig. 4 and column 6 to 5 in Fig. 5), whereas CM from iron-loaded or iron-starved HH macrophages had no effect (compare lanes 7 and 8 to 5 in Fig. 4 and columns 7 and 8 to 5 in Fig. 5). Direct exposure of HepG2 cells to iron (10 μM) or LPS did not significantly modify hepcidin mRNA level (see Fig. 4 lanes 9 through 12 and Figure 5 columns 9 through 12).
The defensin-like circulatory peptide hepcidin has been hailed as the iron-regulatory hormone that links innate immunity and iron metabolism. In response to inflammatory stimuli the liver produces hepcidin a peptide that likely interacts with iron-export protein (such as ferroportin)25 and decreases iron release/transfer from enterocytes and macrophages and causes secondary hypoferremia.11 This limits vital iron needed by invading pathogens but may lead to reduced iron availability for erythropoiesis and anemia (anemia of chronic disease).16 Previous studies have clearly indicated IL-6 as the key mediator of hepcidin activation in response to LPS and inflammatory stimuli7, 12 and suggested that the cytokine released from Kupffer cells/macrophages triggers hepcidin transcription in nearby hepatocytes. Here, we prove that this hypothesis was in fact true because in vivo inactivation of Kupffer cells abrogates the stimulatory effect of both LPS and turpentine on hepcidin expression. Interestingly, hypoferremia caused by turpentine was reversed only to a limited extent by GdCl3 pretreatment. This might be due to the fact that in inflamed GdCl3–treated animals, in spite of normalized hepcidin levels, inactivation depletion of Kupffer cells might deprive the circulatory iron pool of an important source of iron. In addition, during inflammation, the downregulation of ferroportin expression in macrophages,26 the main iron exporter macrophages, likely persists even in the presence of GdCl3 treatment.
In agreement with our ex vivo experiments, LPS-treated macrophages appear to release soluble factors able to turn on hepcidin mRNA in human hepatoma cells. Interestingly, also LPS-conditioned medium from HFE-defective macrophages (i.e., from patients with HH) is able to activate hepcidin mRNA. There exists in the literature a controversy on the actual requirement of HFE for hepcidin activation during inflammation, as studies in inflamed Hfe knockout mice have reported that Hfe is essential14 or redundant.13, 15 Our experiments in human cells add an important piece of information by showing that HFE in human macrophages is not limiting for the early regulatory steps leading to hepcidin activation in nearby hepatocytes, that is, the response of macrophages to the inflammatory stimulus and the release of soluble mediators. Our approach cannot address the potential regulatory role of HFE in the late regulatory steps of hepcidin activation, that is, response of the hepatocytes to the mediators and upregulation of hepcidin gene transcription.
When excess iron enters the circulation, hepcidin transcription is turned on and iron release from intestine and macrophages abrogated.11 Circumstantial evidence indicates that the effect of circulatory iron on hepcidin requires functional HFE, TfR2 and hemojuvelin (HJV). In fact, lack of functional HFE, TfR2, or HJV as found in patients with HH associates with inappropriately low serum level of hepcidin.5–7, 27–29 Therefore, hepcidin now stands out as a common denominator to all forms of HH.11 As to the source of signal for hepcidin upregulation during iron overload, Kupffer cells/macrophages might well be involved, as occurs during inflammatory settings (Ganz16 and this paper). However, we show here that iron loading in mice upregulates hepatic hepcidin mRNA in vivo in spite of Kupffer cell inactivation. Seemingly, human macrophages, regardless of the presence of functional HFE, do not appear to release soluble factors able to activate hepcidin mRNA when iron-loaded or iron-starved. As published in mouse primary hepatocytes,3 neither iron nor LPS stimulated hepcidin mRNA in HepG2 cells. Yet, we found that higher concentration of iron (50-100 μmol/L) led to a consistent decrease of hepcidin mRNA expression (data not shown), in agreement with a previous study.7 This inhibitory effect may be due to a toxic activity of excess free iron in vitro and may not be physiologically relevant since in vivo in both humans and rodents iron excess has always been associated with high hepcidin expression.3, 5–7
In conclusion, although the source of the factor(s) responsible for hepcidin upregulation in response to iron remain elusive, our data indicate that Kupffer cells/macrophages are not involved in this regulatory activity. Gehrke et al.6 found that hemochromatosis patients show an inverse correlation between hepatic hepcidin transcript levels and the serum transferrin saturation. Moreover, in vitro, hepcidin expression was downregulated in response to non-transferrin-bound iron.6 Therefore, it might be not that factors released by Kupffer cells but circulatory forms of iron appearing in transferrin-saturated serum (or the extent of transferrin saturation itself) signal the hepatocytes to upregulate hepcidin transcription. HFE in hepatocytes (and not in Kupffer cells) would be required for this activation and lack of functional HFE might be the reason for inappropriate hepcidin upregulation in HFE-related HH. Further studies are necessary to prove this hypothesis and dissect the complex molecular pathways for abnormal hepcidin regulation in HH.