RED CELLS

Expression of hepcidin in hereditary hemochromatosis: evidence for a regulation in response to the serum transferrin saturation and to non–transferrin-bound iron

Sven G. Gehrke, Hasan Kulaksiz, Thomas Herrmann, Hans-Dieter Riedel, Karin Bents, Claudia Veltkamp, and Wolfgang Stremmel

Experimental data suggest the antimicrobial peptide hepcidin as a central regulator in iron homeostasis. In this study, we characterized the expression of human hepcidin in experimental and clinical iron overload conditions, including hereditary hemochromatosis. Using quantitative reverse transcriptase–polymerase chain reaction (RT-PCR), we determined expression of hepcidin and the most relevant iron-related genes in liver biopsies from patients with hemochromatosis and ironstain–negative control subjects. Regulation of hepcidin mRNA expression in response to transferrin-bound iron, non–

Introduction

transferrin-bound iron, and deferoxamine was analyzed in HepG2 cells. Hepcidin expression correlated significantly with serum ferritin levels in controls, whereas no significant up-regulation was observed in patients with hemochromatosis despite iron-overload conditions and high serum ferritin levels. However, patients with hemochromatosis showed an inverse correlation between hepcidin transcript levels and the serum transferrin saturation. Moreover, we found a significant correlation between hepatic transcript levels of hepcidin and transferrin receptor-2 irrespective of the iron status.

In vitro data indicated that hepcidin expression is down-regulated in response to non–transferrin-bound iron. In conclusion, the presented data suggest a close relationship between the transferrin saturation and hepatic hepcidin expression in hereditary hemochromatosis. Although the causality is not yet clear, this interaction might result from a down-regulation of hepcidin expression in response to significant levels of non–transferrinbound iron. (Blood. 2003;102:371-376)

© 2003 by The American Society of Hematology

Abnormal iron homeostasis is found in many common disorders. These disorders include the iron storage disease, hereditary hemochromatosis, chronic viral hepatitis, alcoholic liver disease, chronic inflammation, or anemias with ineffective erythropoiesis such as thalassemia.1-5

Although several new elements of iron metabolism have been characterized over the past years, for most disorders the exact pathophysiology of iron overload is still unclear. It has been shown that the limiting step in iron homeostasis, the intestinal absorption of dietary ferrous Fe[II] iron, seems to be mediated by 2 iron transport proteins. Dietary Fe[II] is transported into the enterocytes by the apical transporter divalent metal ion transporter 1 (DMT1; formerly called Nramp2, DCT1).6,7 The basolateral transporter iron-regulated transporter 1 (IREG1; also known as ferroportin, MTP1) stimulates iron efflux and, therefore, might export the absorbed Fe[II] from the enterocyte into the plasma.8-10 Most absorbed plasma iron then binds to transferrin and circulates as diferric transferrin (Fe[III]2-Tf).11,12In addition, a small proportion of iron exported into the plasma is found as non–transferrin-bound iron (NTBI).13

Most of the absorbed iron is used in the bone marrow, where transferrin-bound iron is needed for erythropoiesis and taken up by the classical transferrin receptor 1 (TfR1) pathway. The excess iron is stored in the liver.2,11,12 This cellular iron uptake mechanism might also include the identified transferrin-receptor 2 (TfR2) that shows a high hepatic expression.14-16

From the Department of Internal Medicine IV, University Hospital Heidelberg, Germany.

Submitted December 2, 2002; accepted February 27, 2003. Prepublished online as Blood First Edition Paper, March 13, 2003; DOI 10.1182/blood-200211-3610.

Supported by grant STR 216/10-1 from the Deutsche Forschungsgemeinschaft and by the Dietmar Hopp Foundation.

BLOOD, 1 JULY 2003 · VOLUME 102, NUMBER 1

The transferrin receptor pathway seems to play a central role in the pathogenesis of the most common iron storage disease, hereditary hemochromatosis. This disorder is associated with a homozygous Cys282Tyr mutation in the hemochromatosis gene HFE.17 The HFE protein is homologous to class I major histocompatibility complex (MHC) molecules and requires [2-microglobulin ([2m) for surface presentation.17-19 Experimental studies have shown that isolated overexpression of wild-type HFE leads to a decreased cellular uptake of transferrin-bound iron by binding to homodimeric TfR1 and lowering the affinity for iron-saturated transferrin.20,21 However, coexpression of both wild-type HFE and [2m has the reverse effect and results in an increase in TfR1dependent cellular iron uptake.22

The Cys282Tyr substitution in HFE disrupts the association with [2m and, therefore, prevents surface presentation of HFE.18,19 Although the exact mechanism is still incompletely understood, the homozygous Cys282Tyr mutation is associated with an increased intestinal iron absorption, resulting in parenchymal iron overload and the clinical syndrome of hemochromatosis.23

A phenotype similar to classical hereditary hemochromatosis is also observed in individuals with mutations in TfR2 (hemochromatosis type 3)24,25 or IREG1 (autosomal dominant hemochromatosis; type 4).26,27 An identified antimicrobial peptide, named hepcidin, represents another strong candidate putatively involved in the etiology of iron overload syndromes.28-30 Such a hypothesis is supported by the observation that a hepcidin knockout leads to

Reprints: Wolfgang Stremmel, University Hospital Heidelberg, Department of Internal Medicine IV, Bergheimer Strasse 58-69115, Heidelberg, Germany; e-mail:[email protected]

The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked ‘‘advertisement’’ in accordance with 18 U.S.C. section 1734.

© 2003 by The American Society of Hematology

372 GEHRKE et al BLOOD, 1 JULY 2003 · VOLUME 102, NUMBER 1

severe iron overload and a hepcidin overexpression to severe iron deficiency.31-33 In addition, hepcidin mutations were found in 2 families with a new type of juvenile hemochromatosis not linked to chromosome 1q.34

Because hepcidin is induced by iron stores30,35 and inflammation,36,37 it might act as a central iron-regulatory hormone important in the pathogenesis of iron overload and the anemia of chronic disease. Therefore, the aim of the present study was to evaluate the regulation of hepcidin in response to iron loading and its role in hereditary hemochromatosis.

Materials and methods

Cells

HepG2 cells were obtained from DSMZ (Braunschweig, Germany) and grown in RPMI 1640 medium containing 10% fetal calf serum, penicillin (100 U/mL), and streptomycin (100 fg/mL). Incubation was performed with 65 fM Fe-NTA (1:1), 50 fM deferoxamine (DFO), and 2.5 g/L iron-saturated human diferric transferrin (Fe[III]2-Tf).

Liver biopsies

Liver biopsy samples, obtained for diagnostic purposes, were from 10 patients with iron-overloaded hereditary hemochromatosis (HH) homozygous for the Cys282Tyr mutation in HFE prior to phlebotomy. Control samples derived from 20 patients (7 with isolated ‘-glutamyl transferase elevation and healthy liver histology, 2 with mild steatosis, and 11 with chronic hepatitis C and mild histological activity) (Table 1). All control samples showed no significant fibrosis and were negative for stainable iron in liver biopsy. Samples were stored at -20°C in RNAlater solution (Ambion, Austin, TX) prior to RNA isolation. The study was approved by the local ethics committee of the University of Heidelberg. Informed consent was obtained from all patients.

Chemicals

RPMI 1640 medium, penicillin, and streptomycin were from Life Technologies (Paisley, United Kingdom). Ferric nitrate nonahydrate, nitrilotriacetic acid disodium salt (NTA), and iron-saturated diferric transferrin were obtained from Sigma-Aldrich (Steinheim, Germany). Deferoxamine was from Novartis (Nuernberg, Germany).

Determination of serum iron parameters

Serum ferritin levels were measured by electrochemiluminescence immunoassay (ECLIA) technology on an Elecsys analyzer (Roche Diagnostics, Mannheim, Germany). Transferrin saturation was calculated from serum iron, determined photometrically on an LX-Analyzer (Beckman-Coulter, Krefeld, Germany), and serum transferrin, determined by nephelometry on a BNAII analyzer (Dade-Behring, Schwalbach, Germany), as iron (fg/ dL) X 100/transferrin (mg/dL) X 1.4. Serum samples were obtained in the fasting state at the same time of day.

Quantitative RT-PCR

Total RNA was isolated from liver biopsies and from cell culture using the RNAeasy Mini Kit (Qiagen, Hilden, Germany) including DNAse digestion

Table 1. Demographic characteristics and iron parameters of patients with hemochromatosis and control individuals

Patients with HH, Control Individuals, n = 10 n = 20

Age, y* 27-63 (44 ± 11) 24-60 (41 ± 9) Sex, male/female 8/2 15/5 Serum ferritin, fg/L* 313-3750 (2061 ± 1030) 11-357 (145 ± 102) Transferrin saturation, %* 80-100 (92 ± 7) 4-51 (29 ± 11)

*Range (mean ± standard deviation).

according to manufacturer’s instructions. Real-time quantification of mRNA transcripts was performed with a 2-step reverse transcriptase–polymerase chain reaction (RT-PCR) using the LightCycler system and Relative Quantification Software Version 1.0 (Roche Molecular Biochemicals, Mannheim, Germany). In a first step, cDNA synthesis was performed with the First Strand cDNA Synthesis Kit for RT-PCR (Roche Molecular Biochemicals) according to manufacturer’s instructions. In a second step, transcripts of hepcidin (Hepc), transferrin receptor (TfR1), transferrin receptor-2 (TfR2), iron-regulated transporter (IREG1), the IRE and non-IRE splice variant of the divalent metal-ion transporter (DMT1-IRE, DMT1nonIRE), and ceruloplasmin (Cp) were amplified in duplicates with specific sense and antisense primers (Table 2). All transcripts were detected using SYBR Green I according to manufacturer’s instructions and were normalized to actin ([-actin) as internal control. Therefore, actin transcripts were amplified in duplicates with sense primer ACTB-502 (5′-AGG ATG CAG AAG GAG ATC ACT G) and antisense primer ACTB-302 (5’-GGG TGT AAC GCA ACT AAG TCA TAG) and detected using SYBR Green I. Hepc/Actin, TfR1/Actin, TfR2/Actin, IREG1/Actin, DMT1-IRE/Actin, DMT1nonIRE/Actin, and Cp/Actin ratios were calculated using LightCycler Relative Quantification Software Version 1.0 (Roche Molecular Biochemicals), which provides a fully automated, efficiency-corrected, relative quantification normalized to calibrators. According to manufacturer’s instructions, calibrators for Hepc, TfR1, TfR2, IREG1, DMT1-IRE, DMT1nonIRE, Cp, and [-actin were generated from expressed sequence tag (EST) clones (obtained from RZPD, Berlin, Germany, followed by sequence analyses to verify the proposed insert). In addition, standard curves were prepared according to accurately determined dilutions of the plasmids containing cDNA sequences of Hepc, TfR1, TfR2, IREG1, DMT1-IRE, DMT1-nonIRE, Cp, and [-actin as templates. Plasmid dilutions covered a dynamic range of 5 logarithmic orders.

Statistical analysis

Statistical analysis of quantitative variables was performed using the nonparametric Mann-Whitney test. To study the linear relationship between continuous variables, Pearson correlation coefficients were calculated. P < .05 was considered significant. All statistical analyses were performed using StatView Version 5.0 (SAS Institute, Cary, NC).

Results

Transcript levels of iron-related genes in liver biopsies from patients with HH and iron-stain–negative control subjects

Hepatic expression of the iron-related genes Hepc, TfR1, TfR2, IREG1, DMT1-IRE, DMT1-nonIRE, and Cp normalized to actin transcript levels was analyzed in liver biopsy samples from patients with untreated hereditary hemochromatosis and in liver biopsy samples from control individuals negative for iron staining (Table 1). Differences between patients with HH and control subjects were found for the mean TfR1/Actin ratio. As expected, the TfR1/Actin ratio was significantly decreased in patients with untreated HH (P < .001) (Figure 1). In contrast, mean ratios (± standard deviation) for Hepc/Actin (0.53 ± 0.26 versus 0.47 ± 0.52) (Figure 1), TfR2/Actin (0.90 ± 0.37 versus 0.73 ± 0.18), IREG1/ Actin X 10 (1.35 ± 0.30 versus 1.27 ± 0.26), DMT1-IRE/Actin X 10(0.99 ± 0.56 versus 0.87 ± 0.49), DMT1-nonIRE/ Actin X 10(3.45 ± 0.72 versus 4.62 ± 1.75), and Cp/Actin

(0.50 ± 0.21 versus 0.62 ± 0.24) did not differ significantly between patients with HH and control subjects.

Hepatic expression of hepcidin in relation to serum ferritin levels and the transferrin saturation

In control individuals, the hepatic Hepc/Actin ratio correlated significantly with serum ferritin levels (r = 0.713, P < .001)

BLOOD, 1 JULY 2003 · VOLUME 102, NUMBER 1

Table 2. Primers for quantification of Hepc, TfR1, TfR2, IREG1, DMT1-IRE, DMT1-nonIRE, and Cp transcripts using the LightCycler RT-PCR assay

Primer Sequence Orientation
Hepc
Hepc-501 5′-CTG CAA CCC CAG GAC AGA G-3′ Sense
Hepc-301 5′-GGA ATA AAT AAG GAA GGG AGG GG-3′ Antisense
TfR1
TfR1-502 5′-TAT AGA AGG TTT GGG GGC TGT G-3′ Sense
TfR1-302 5′-GAG ACC CTA TGA ACT TTT CCC TAG-3′ Antisense
TfR2
TfR2-501 5′-GAT TCA GGG TCA GGG AGG TG-3′ Sense
TfR2-301 5′-GAA GGG GCT GTG ATT GAA GG-3′ Antisense
IREG1
IREG1-501 5′-CTT CAG CCT GGC AAG TTA CAT G-3′ Sense
IREG1-301 5′-TTC TCA AAG GCA TTT GAA AGG G-3′ Antisense
DMT1-IRE
DMT-IRE-502 5′-CTT CAT ATC TGC CTC TTC CCC-3′ Sense
DMT-IRE-301 5′-AAA TCT GAG ACT GAC TGG ACC C-3′ Antisense
DMT-nonIRE
DMT-nIRE-501 5′-TGG TGT GAT CTC AGC TCA CTG-3′ Sense
DMT-nIRE-301 5′-GGC CAG CAG ATT ACT TGA GC-3′ Antisense
Cp
Cp-501 5′-ATG GGA ATG GGC AAT GAA ATA G-3′ Sense
Cp-301 5′-GCA TGA ATG TGG TCG GTC AC-3′ Antisense

(Figure 2A). In contrast, an inverse correlation between the hepatic Hepc/Actin ratio and serum ferritin levels was observed in patients with iron overloaded HH (r =-0.715, P <.05) (Figure 2B). We also analyzed the association between the hepatic Hepc/Actin ratio and the serum transferrin saturation. Patients with HH with a transferrin saturation above 80% showed a strong inverse correlation between the Hepc/Actin ratio and the serum transferrin saturation (r =-0.861, P <.01) (Figure 3B); no significant correlation was found in control patients with a transferrin saturation between 4% and 51% (Figure 3A).

Because the hepatic Hepc/Actin ratio was found to correlate with serum ferritin levels and the transferrin saturation, multiple regression analyses were performed. These analyses confirmed the significant correlation between the Hepc/Actin ratio and serum ferritin levels in control individuals (P <.001) (Figure 2A) and the significant inverse correlation between the Hepc/Actin ratio and the serum transferrin saturation in patients with untreated HH (P <.05) (Figure 3B). The inverse correlation between the Hepc/Actin ratio and serum ferritin levels in patients with HH (Figure 2B) did not remain statistically significant using multiple regression analysis. As the serum ferritin levels correlate with the serum transferrin saturation in our

Figure 1. Hepatic TfR1/Actin � 103 and Hepc/Actin ratios in patients with HH and control individuals. The mean values are shown (±95% confidence intervals). The TfR1/Actin X103 ratio differed significantly between patients with HH and control individuals (P <.001).

REGULATION OF HUMAN HEPCIDIN EXPRESSION 373

Figure 2. Linear regression analysis of the correlation between the hepatic Hepc/Actin ratio and serum ferritin levels in control individuals and untreated HH patients. (A) In control individuals, the hepatic Hepc/Actin ratio correlated significantly with serum ferritin levels. (B) In iron overloaded HH patients, an inverse correlation between the hepatic Hepc/Actin ratio and serum ferritin levels was observed. * indicates statistically significant using multiple regression analysis (serum ferritin levels and serum transferrin saturation).

patients with HH (r =0.662, P <.05), the impaired Hepc/Actin ratio in patients with HH with high serum ferritin levels (Figure 2B) might be indeed related to a high transferrin saturation.

Hepatic Hepc/Actin ratio correlates significantly with the hepatic TfR2/Actin ratio

To evaluate whether the expression of hepcidin in liver correlates with the expression of iron-related genes, Hepc/Actin ratios of all patients (control subjects and patients with HH) were plotted against TfR1/Actin, TfR2/Actin, IREG1/Actin, DMT1-IRE/Actin, DMT1-nonIRE/Actin, and Cp/Actin ratios. These analyses revealed a strong correlation between the Hepc/Actin ratio and the TfR2/ Actin ratio (r =0.777, P <.0001) (Figure 4). In addition, data from control subjects and patients with HH were analyzed separately. These analyses also demonstrated a significant correlation between the Hepc/Actin ratio and the TfR2/Actin ratio in control subjects (r =0.636, P =.014) and patients with HH (r =0.823, P <.01). However, no significant correlation between the Hepc/ Actin ratio and TfR1/Actin, IREG1/Actin, DMT1-IRE/Actin, DMT1nonIRE/Actin, and Cp/Actin ratios was found in patients with HH and control subjects.

Hepcidin is down-regulated in HepG2 cells in response to non–transferrin-bound iron but not in response to diferric transferrin

For in vitro analysis of hepcidin regulation in response to iron, HepG2 cells were incubated for 72 hours with different concentrations of non–transferrin-bound ferric iron (Fe-NTA). As demonstrated in Figure 5, the Hepc/Actin ratio decreased after incubation