Testosterone perturbs systemic iron balance through activation of epidermal growth factor receptor signaling in the liver and repression of hepcidin

Authors

  • Chloé Latour,

    1. Inserm, U1043, Toulouse, France
    2. CNRS, U5282, Toulouse, France
    3. Université de Toulouse, UPS, Centre de Physiopathologie de Toulouse Purpan (CPTP), Toulouse, France
    Search for more papers by this author
  • Léon Kautz,

    1. Department of Pathology, David Geffen School of Medicine, University of California, Los Angeles, CA
    2. Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA
    Search for more papers by this author
  • Céline Besson-Fournier,

    1. Inserm, U1043, Toulouse, France
    2. CNRS, U5282, Toulouse, France
    3. Université de Toulouse, UPS, Centre de Physiopathologie de Toulouse Purpan (CPTP), Toulouse, France
    Search for more papers by this author
  • Marie-Laure Island,

    1. Inserm, UMR991, Rennes, France
    2. Université de Rennes 1, Rennes, France
    Search for more papers by this author
  • François Canonne-Hergaux,

    1. Inserm, U1043, Toulouse, France
    2. CNRS, U5282, Toulouse, France
    3. Université de Toulouse, UPS, Centre de Physiopathologie de Toulouse Purpan (CPTP), Toulouse, France
    Search for more papers by this author
  • Olivier Loréal,

    1. Inserm, UMR991, Rennes, France
    2. Université de Rennes 1, Rennes, France
    Search for more papers by this author
  • Tomas Ganz,

    1. Department of Pathology, David Geffen School of Medicine, University of California, Los Angeles, CA
    2. Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, CA
    Search for more papers by this author
  • Hélène Coppin,

    Corresponding author
    1. Inserm, U1043, Toulouse, France
    2. CNRS, U5282, Toulouse, France
    3. Université de Toulouse, UPS, Centre de Physiopathologie de Toulouse Purpan (CPTP), Toulouse, France
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Marie-Paule Roth

    Corresponding author
    1. Inserm, U1043, Toulouse, France
    2. CNRS, U5282, Toulouse, France
    3. Université de Toulouse, UPS, Centre de Physiopathologie de Toulouse Purpan (CPTP), Toulouse, France
    Search for more papers by this author
    • These authors contributed equally to this work.


  • Supported in part by a grant from the French National Research Agency (ANR, programme Genopat, project ANR-09-GENO-016).

  • Potential conflict of interest: Nothing to report.

Abstract

Gender-related disparities in the regulation of iron metabolism may contribute to the differences exhibited by men and women in the progression of chronic liver diseases associated with reduced hepcidin expression, e.g., chronic hepatitis C, alcoholic liver disease, or hereditary hemochromatosis. However, their mechanisms remain poorly understood. In this study we took advantage of the major differences in hepcidin expression and tissue iron loading observed between Bmp6-deficient male and female mice to investigate the mechanisms underlying this sexual dimorphism. We found that testosterone robustly represses hepcidin transcription by enhancing Egfr signaling in the liver and that selective epidermal growth factor receptor (Egfr) inhibition by gefitinib (Iressa) in males markedly increases hepcidin expression. In males, where the suppressive effects of testosterone and Bmp6-deficiency on hepcidin expression are combined, hepcidin is more strongly repressed than in females and iron accumulates massively not only in the liver but also in the pancreas, heart, and kidneys. Conclusion: Testosterone-induced repression of hepcidin expression becomes functionally important during homeostatic stress from disorders that result in iron loading and/or reduced capacity for hepcidin synthesis. These findings suggest that novel therapeutic strategies targeting the testosterone/EGF/EGFR axis may be useful for inducing hepcidin expression in patients with iron overload and/or chronic liver diseases. (Hepatology 2014;59:683–694)

Abbreviations
BMP

bone morphogenetic protein

EGF

epidermal growth factor

EGFR

epidermal growth factor receptor

NTBI

nontransferrin-bound iron

Hepcidin, a circulating hormone produced primarily by the liver, plays a central role in the regulation of systemic iron homeostasis[1] necessary to ensure sufficient availability of iron for hemoglobin synthesis and other metabolic processes while avoiding the oxidative damage to cells that can result from excess free iron.[2] Hepcidin triggers internalization and degradation of ferroportin, the only known iron export channel from cells into the plasma, which leads to the decrease of dietary iron absorption from duodenal enterocytes and to the sequestration of iron recycled from senescent blood cells within the macrophages.[3] The essential role of hepcidin in the maintenance of systemic iron balance has been demonstrated in mouse models. Mice lacking hepcidin expression develop systemic iron overload, whereas transgenic mice overexpressing hepcidin exhibit severe iron deficiency anemia.[4] In humans, loss-of-function mutations in the hepcidin gene HAMP cause juvenile hemochromatosis, an autosomal recessive disorder characterized by severe iron deposition in multiple organs, including the liver, heart, and endocrine tissues.[5]

Recent advances have been made in our understanding of the molecular mechanisms through which hepcidin expression is modulated to influence systemic iron balance. Iron overload induces the expression of bone morphogenetic protein 6 (BMP6), a member of the transforming growth factor beta (TGF-β) superfamily of ligands,[6] which activates a signaling cascade leading to hepcidin gene transcription. Hemojuvelin (HJV) functions as an essential coreceptor for BMP6. Mice with disruption of either Bmp6[7, 8] or the hemojuvelin gene[9, 10] exhibit hepcidin deficiency and severe iron overload, confirming the central role of these two molecules in the hepatic BMP signaling pathway that promotes hepcidin expression. Interestingly, however, there are considerable, and still unexplained, gender differences in residual hepcidin expression and in the severity of tissue iron loading in both hemojuvelin-[11] and Bmp6-deficient mice (our unpublished data).

Clinical data have shown that men and women exhibit important disparities in the progression of chronic liver diseases such as alcoholic liver disease, chronic hepatitis C, or genetic hemochromatosis, all of which have been reported associated with reduced hepcidin expression.[12] Gender-related variations in the regulation of iron metabolism may contribute to these differences and further understanding of their underlying mechanisms lead to the development of novel treatment strategies for chronic liver diseases. In the present study, we took advantage of the major differences in hepcidin expression and tissue iron stores observed between Bmp6-deficient males and females to explore the role of sex hormones in the regulation of iron metabolism by the liver.

Materials and Methods

Animals and Treatments

Bmp6 null mice (Bmp6m1Rob), obtained from E. Robertson (Dunn School of Pathology, University of Oxford, UK), were maintained on a CD1 outbred background. Wild-type CD1 (outbred) mice were purchased from the Centre d'Elevage Robert Janvier (Le Genest St-Isle, France). Hamp-deficient mice, originally provided by S. Vaulont (Institut Cochin, Paris), were backcrossed onto the C57BL/6 background (N4, 99% gene marker identity) using marker-assisted accelerated backcrossing. Protocols for gonadectomies, irradiation, and testosterone or gefitinib administration are detailed in the Supporting Methods. They were approved by the Midi-Pyrénées Animal Ethics Committee.

Blood Analysis

Serum iron and unsaturated iron-binding capacity (UIBC) were determined using BIOLABO (Maizy, France) kits. Transferrin saturations were deduced from serum iron and UIBC values. Serum hepcidin levels were measured by sandwich enzyme-linked immunosorbent assay (ELISA) as described.[13]

Tissue Iron Staining and Quantitative Iron Measurement

Liver, spleen, heart, pancreas, and kidney samples were fixed in 10% buffered formalin and embedded in paraffin. Deparaffinized tissue sections were stained with the Perls' Prussian blue stain for nonheme iron and counterstained with nuclear fast red. Quantitative measurement of nonheme iron in the liver was performed according to the method recommended by Torrance and Bothwell.[14]

Quantitation of Messenger RNA (mRNA) Levels

Total RNA from mouse liver was extracted using Trizol (Invitrogen, Carlsbad, CA). Complementary DNA (cDNA) was synthesized using MMLV-RT (Promega, Madison, WI). The sequences of the primers for target genes and the reference gene Hprt are listed in Supporting Table 1. Quantitative polymerase chain reaction (PCR) reactions were prepared with LightCycler 480 DNA SYBR Green I Master reaction mix (Roche, Mannheim, Germany) and run in duplicate on a LightCycler 480 Instrument (Roche).

Protein Extraction

Livers and spleens were homogenized in a FastPrep-24 Instrument (MP Biomedicals, Solon, OH) for 15 seconds at 4 m/s. The lysis buffer (50 mM Tris-HCl, pH 8, 150 mM NaCl, 5 mM EDTA, pH 8, 0.1% NP-40) included inhibitors of proteases (complete protease inhibitor cocktail, Roche) and of phosphatases (phosphatase inhibitor cocktail 2, Sigma-Aldrich). Proteins were quantified using a protein assay kit (DC Protein Assay, Bio-Rad, Hercules, CA).

Western Blot Analysis

Fresh protein extracts were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes (GE Healthcare, Little Chalfont, UK). Membranes were incubated with phospho-EGFR (Tyr1068), EGFR, ferroportin, vinculin, phospho-Smad5 (Ser463/465), or Smad5 antibodies, as detailed in the Supporting Methods.

Statistical Analyses

Data were normalized to the invariant control Hprt and means of ΔCt values compared by Student t tests as detailed in the Supporting Methods.

Results

Bmp6-Deficiency Leads to a Much Greater Hepcidin Decrease in Males Than in Females

Bmp6 plays a critical role in the maintenance of iron homeostasis. Indeed, 7-week-old Bmp6−/− mice present with marked iron accumulation in liver parenchymal cells, reduced hepcidin expression compared with wild-type mice, and stabilization of ferroportin at the membrane of enterocytes, tissue macrophages, and hepatocytes.[8] However, although 7-week-old Bmp6−/− males and females have about the same amount of liver iron (Fig. 1C), males have much lower hepcidin mRNA compared with females (Fig. 1A). This prompted us to examine the gender differences in hepcidin regulation further. We quantified hepcidin expression and assessed splenic iron content and liver iron accumulation in older (12- and 30-week-old) mice of both genders. Bmp6−/− males have consistently lower hepcidin mRNA expression than Bmp6−/− females (Fig. 1A) and, as could be expected, less iron in the spleen than females (Fig. 1B). The gender-related differences in hepcidin levels previously reported in wild-type mice[15, 16] and confirmed in this study (Fig. 1A) are thus magnified in Bmp6−/− mice. We did not observe significant gender differences in serum iron or transferrin saturation, which are particularly high in Bmp6−/− mice (Supporting Table 2).

Figure 1.

Bmp6−/− males have consistently lower hepcidin mRNA expression than females. Groups of 6 Bmp6−/− and 6 wild-type (Bmp6+/+) mice of each gender were compared at 7, 12, and 30 weeks of age. (A) Hepcidin (Hamp) mRNA levels were measured by quantitative PCR. Values shown are means of −ΔCt (i.e., Ct Hprt − Ct Hamp) ± SEM. The higher the −ΔCt, the greater is the amount of Hamp amplicon. Means of −ΔCt values in males and females of each age and genotype combination were compared by Student t tests (***P < 0.001; **P < 0.01). (B) Spleen nonheme iron content (mean ± SEM) is reported as micrograms of iron per gram of tissue. At 12 and 30 weeks of age, males have a lower splenic iron content than females (**P < 0.01; *P < 0.05). (C) Liver nonheme iron content (mean ± SEM) is reported as micrograms of iron per gram dry weight of tissue. None of the male/female comparisons were statistically significant.

Male But Not Female Bmp6−/− Mice Accumulate Iron in Pancreas, Heart, and Kidneys

We next assessed the sites of iron accumulation in males and females by staining histological sections for iron. Interestingly, whereas iron deposition was restricted to the liver in 12-week-old females, males of the same age had major iron loading in other tissues, most notably the exocrine pancreas, the heart, and the proximal and distal convoluted tubules of the kidney (Fig. 2). These gender differences in tissue iron deposition were exacerbated with age and particularly striking in 30-week-old mice (Supporting Fig. S1).

Figure 2.

Histological examination of iron loading in 12-week-old Bmp6−/− males and females (intact or gonadectomized). Tissue iron was detected by staining with Perls' Prussian blue (blue stain). Original magnification ×50 for liver and pancreas, ×100 for heart and kidney.

Castration of Bmp6−/− Males Increases Hepcidin Expression and Strongly Reduces Tissue Iron Deposition

To investigate the reasons for these important gender differences, 4-week-old Bmp6−/− animals were ovariectomized or castrated and examined at 7, 12, or 30 weeks of age. Hepcidin expression is similar in ovariectomized and nonovariectomized Bmp6−/− females (Fig. 3A). Ovariectomized Bmp6−/− females still exclusively accumulate iron in their liver (Fig. 2). In contrast, castrated Bmp6−/− males have much higher hepcidin expression than noncastrated animals (Fig. 3A). Their hepcidin levels are similar to those of Bmp6−/− females of the same age, indicating that male gonadal hormones are responsible for the inhibition of hepcidin expression. Not surprisingly, castration led to a significant increase in the splenic iron content of 12- and 30-week-old males (Fig. 3B) and a slight but nonsignificant decrease in liver iron accumulation of 30-week-old males (Fig. 3C). Most remarkably, 12-week-old castrated Bmp6−/− males have virtually no iron in organs other than the liver (Fig. 2) and 30-week-old castrated males have considerably lower iron accumulation in their pancreas and heart than noncastrated males (Fig. S1). The same experiments were carried out on wild-type controls. Similar to Bmp6−/− mice, hepcidin expression was comparable in ovariectomized and nonovariectomized wild-type females, but significantly induced in castrated versus noncastrated wild-type males (Fig. S2).

Figure 3.

Hepcidin mRNA levels in Bmp6−/− males increase after castration. Twenty-four Bmp6−/− mice were either gonadectomized (N = 12) or sham operated (N = 12). A liver biopsy was performed on 12 mice (6 per group) at 7 weeks of age. These mice were then sacrificed at 30 weeks. The other 12 mice (6 per group) were sacrificed at 12 weeks. (A) Hepcidin (Hamp) mRNA levels were measured by quantitative PCR. Values shown are means of −ΔCt (i.e., Ct Hprt − Ct Hamp) ± SEM in females and males. Means of −ΔCt values in gonadectomized and intact mice of each sex and age combination were compared by Student t tests (***P < 0.001;**P < 0.01). (B) Spleen nonheme iron content (mean ± SEM) is reported as micrograms of iron per gram of tissue. At 12 and 30 weeks of age, castrated males have a higher splenic iron content than intact males (***P < 0.001; *P < 0.05). (C) Liver nonheme iron content (mean ± SEM) is reported as micrograms of iron per gram dry weight of tissue. In none of the sex/age group did gonadectomy induce a significant change in liver iron content.

Testosterone Is the Major Hormone Responsible for the Observed Gender Differences in the Regulation of Iron Metabolism

To examine the role of testosterone on hepcidin production further, 7-week-old ovariectomized Bmp6−/− females received daily injections of testosterone propionate (10 mg/kg subcutaneously) or vehicle for a week. As shown in Fig. 4A, hepcidin mRNA expression was repressed on average 15.4-fold after testosterone treatment. Hepcidin expression was reduced in the same proportions (on average 15.8-fold) in 7-week-old Bmp6−/− males compared with females (Fig. 1A), suggesting that testosterone is the major hormone responsible for the inhibition of hepcidin in males. As reported recently,[17] treatment of wild-type females with daily injections of testosterone for a week not only significantly reduced their hepcidin expression, compared with injection of vehicle only (Fig. S3), but also significantly increased their serum iron and transferrin saturation (Supporting Table 2).

Figure 4.

Testosterone administration to ovariectomized Bmp6−/− females represses hepcidin expression. (A) Nonirradiated 7-week-old ovariectomized mice received daily injections of testosterone propionate (10 mg/kg sc; N = 6) or vehicle (N = 4) for 7 days. Hepcidin (Hamp) mRNA levels were measured by quantitative PCR. Values shown are means of −ΔCt (i.e., Ct Hprt − Ct Hamp) ± SEM. (B) Whole-body irradiated mice received daily injections of testosterone (N = 8) or vehicle (N = 5) from days 2 to 8. Means of −ΔCt values in testosterone or vehicle-treated mice were compared by Student t tests (***P < 0.001).

Testosterone-Induced Repression of Hepcidin Has a Functional Impact on Cell Surface Ferroportin Expression

As shown in Fig. 5A, there is a good correlation between hepcidin protein abundance measured by sandwich ELISA and liver hepcidin mRNA levels assessed by quantitative PCR. Because hepcidin triggers ferroportin for degradation, we tested whether there was an inverse relationship between serum hepcidin and expression of ferroportin. As expected, the mice with the lowest amounts of circulating hepcidin, i.e., Bmp6−/− males and Bmp6−/− gonadectomized females treated with testosterone, were the mice with the highest expression of ferroportin in the liver (Fig. 5B) and in the spleen (Fig. 5C). Thus, the differences in hepcidin expression between males and females are functionally important, as they impact the quantity of ferroportin present at the cell surface.

Figure 5.

Serum hepcidin concentration correlates with Hamp mRNA expression in the liver and with ferroportin expression at the cell membrane. Scatterplot showing the relationship between serum hepcidin levels measured by sandwich ELISA and hepcidin mRNA expression assessed by quantitative PCR for the different groups of mice (A). Membrane protein extracts were prepared from the livers and the spleens of Bmp6−/− males (4 intact and 4 gonadectomized) and Bmp6−/− gonadectomized females (4 treated with vehicle and 4 treated with testosterone). Immunoblot techniques were used to detect ferroportin and vinculin in the liver (B) and the spleen (C). Results for two representative mice per group are shown on the blots.

Residual Hepcidin Levels in Bmp6−/− Females Are Sufficient to Prevent Massive Tissue Iron Loading

Differences in tissue iron deposition between males and females could be the consequence of reduced production of hepcidin, increased iron absorption, and higher circulating amounts of nontransferrin-bound iron (NTBI) in males compared with females. Alternatively, these differences could be independent of the levels of hepcidin but due to the influence of male gonadal hormones on the expression of iron transporters into storage tissues. To discriminate between these two possibilities, we compared tissue iron accumulation of 12-week-old hepcidin (Hamp)-deficient males and females. In contrast to Bmp6−/− females, Hamp−/− females accumulate iron not only in the liver, but also in the pancreas, heart, and kidneys (Fig. S4), suggesting that the residual hepcidin levels found in Bmp6−/− females protect them against massive iron loading of organs other than the liver. Importantly, the differences in tissue iron accumulation observed between Bmp6−/− and Hamp−/− females are not due to differences in their genetic background, as there is no extrahepatic iron loading in 12-week-old Bmp6−/− females on a C57BL/6 background recently obtained in our laboratory (data not shown).

Testosterone-Induced Suppression of Hepcidin Expression Is Not Due to Its Ability to Stimulate Erythropoiesis

Transcription of the hepcidin gene is controlled negatively by the rate of erythropoiesis.[18] Men and women exhibit differences in hemoglobin concentration and during puberty hemoglobin levels increase only in males. Moreover, hemoglobin levels decline after castration or antitestosterone therapy.[19] These observations suggest that androgens play a role in erythropoiesis. We first tested whether testosterone had an influence on Epo transcription in the liver and/or the kidney but did not find a significant difference in Epo mRNA levels between Bmp6−/− males and females (data not shown). In humans, levels of erythropoietin are also similar in men and women and it is assumed that testosterone increases the sensitivity of erythroid progenitors to erythropoietin.[20] To test whether the down-regulation of hepcidin expression by testosterone in Bmp6−/− mice was due to the stimulation of erythropoiesis, we irradiated ovariectomized females (60Co, 6Gy) to inhibit erythropoiesis. Testosterone propionate (10 mg/kg) or vehicle was then administered on days 2 to 8. Giemsa stain and flow cytometry analysis of bone marrow at day 8 showed massive depletion of nucleated cells in irradiated mice. The spleens of these mice were atrophic and erythropoiesis was absent (Fig. S5). In the absence of testosterone, hepcidin expression was not reduced in irradiated mice compared with nonirradiated mice, confirming that erythopoiesis was inhibited (Fig. 4). Interestingly, irradiation did not prevent testosterone-induced down-regulation of hepcidin expression to levels similar to those observed in nonirradiated control mice (Fig. 4B). These results demonstrate that the observed effects of testosterone on hepcidin expression are not caused by the negative control of erythropoietic regulators.

Activation of Egfr Signaling in the Liver Is Testosterone-Dependent

The growth factors EGF and HGF were recently shown to suppress hepatic hepcidin synthesis[21] and, interestingly, hepatic EGF receptor expression in the mouse was reported to be induced by testosterone.[22, 23] We therefore hypothesized that the suppressive role of EGF and HGF on hepcidin transcription might depend on the expression of their receptors in the liver and thus be modulated by endocrine influences. To confirm this hypothesis, we first compared Egfr and Met mRNA expression between males and females. There was no influence of gender on liver expression of Met (data not shown). In contrast, mRNA expression of Egfr was sexually dimorphic, and higher in males than in females, both in Bmp6−/− (Fig. 6A) and in wild-type mice (Fig. S6A). Moreover, expression of Egfr mRNA was reduced in the liver of castrated males, and induced in ovariectomized females treated with testosterone for a week, both in Bmp6−/− (Fig. 6B) and in wild-type mice (Fig. S6B). There was a good correspondence between Egfr mRNA expression levels, protein abundance, and the amount of phosphorylated Egf receptors, suggesting a role for testosterone in the activation of the EGFR signaling pathway in the liver (Fig. 6C; Fig. S6C). In addition, there was a strong negative correlation between Egfr and Hamp mRNA levels in mice from the different groups (R2 = −0.87; P < 0.001; Fig. 6D), suggesting that the suppressive effect of testosterone on hepcidin expression might be due to its capacity to induce Egfr expression and activate Egfr signaling in the liver.

Figure 6.

Hepatic expression of Egfr is higher in Bmp6−/− males than in females and is induced by testosterone. (A) Egfr mRNA levels were measured by quantitative PCR in 6 males and 6 females Bmp6−/−. Values shown are means of −ΔCt (i.e., Ct Hprt − Ct Egfr) ± SEM. Means of −ΔCt values in males and females were compared by Student t tests (***P < 0.001). (B) Egfr mRNA levels were measured in 6 gonadectomized and 6 sham-operated Bmp6−/− males, as well as in gonadectomized females treated with vehicle (N = 6) or testosterone (N = 11) for a week. Egfr expression was decreased by castration and increased by testosterone administration. Means of −ΔCt values in gonadectomized and intact mice or in testosterone and vehicle-treated mice were compared by Student t tests (***P < 0.001). (C) Membrane protein extracts were prepared from the mouse livers of Bmp6−/− intact males, castrated males, and gonadectomized females treated with either vehicle or testosterone (4 mice/group). Phospho-Egfr, total Egfr, and vinculin were detected by immunoblot techniques. Results for two representative mice per group are shown on the blots. (D) Scatterplot showing the negative correlation (R2 = −0.87; P < 0.001) between Egfr and Hamp mRNA levels assessed by quantitative PCR in mice from the different groups.

Selective Inhibition of Egfr Signaling in Males Markedly Increases Hepcidin Expression

To definitely confirm that the effect of testosterone on hepcidin down-regulation was mediated by an increase in Egfr signaling, we treated 7-week-old males with the selective EGFR-tyrosine kinase inhibitor gefitinib or vehicle daily for 5 days. As expected, phosphorylation of the Egf receptors was virtually abolished in the liver of Bmp6−/− (Fig. 7A) and wild-type (Fig. S7A) mice treated with gefitinib. Most remarkably, repression of Egfr signaling by gefitinib led to a significant induction of Hamp mRNA levels both in Bmp6−/− (Fig. 7B) and in wild-type mice (Fig. S7B). These data demonstrate that the observed gender differences in hepcidin expression are mediated by testosterone by way of Egfr signaling in the liver.

Figure 7.

Selective inhibition of Egfr in mice prevents hepcidin down-regulation by testosterone. Seven-week-old Bmp6−/− males were treated with the selective EGFR-tyrosine kinase inhibitor, gefitinib (N = 5), or vehicle (N = 3) daily for 5 days. (A) Fresh membrane protein extracts were prepared from mouse livers. Phospho-Egfr and vinculin were detected by immunoblot techniques. (B) Hamp mRNA levels were measured by quantitative PCR in Bmp6−/− males treated with gefitinib or vehicle. Means of −ΔCt values in mice treated with or without gefitinib were compared by Student t tests (**P < 0.01).

Phosphorylation of Smad1/5 Is Lower in Bmp6−/− Males Than in Females and Is Influenced by Testosterone Levels

We finally tested whether testosterone-induced hepcidin repression was due to EGF-mediated perturbation of Smad1/5/8 signaling. Interestingly, Bmp6−/− males had lower amounts of Smad5 phosphorylation than females (Fig. 8A). Moreover, Smad5 phosphorylation was increased by castration in males, and reduced by administration of testosterone to females, which parallels changes in hepcidin expression. Smad5 phosphorylation was also increased by treating males with gefitinib (Fig. 8B). The gender differences in Smad5 phosphorylation observed in Bmp6−/− mice are not explained by differences in gene expression of any of the Bmp ligands (Bmp2, Bmp4, Bmp5, Bmp7, or Bmp9) between males and females (Fig. S8) and the exact molecular mechanisms linking testosterone-induced Egfr signaling to reduced Smad5 phosphorylation remain to be elucidated.

Figure 8.

Activation of Egfr signaling by testosterone coincides with low levels of Smad5 phosphorylation in Bmp6−/− mice. (A) Total protein extracts were prepared from the livers of intact males, gonadectomized males, and gonadectomized females treated with either vehicle or testosterone (4 mice/group). Immunoblot techniques were used to detect phospho-Smad5 and total Smad5. Results for two representative mice per group are shown on the blots. (B) Total protein extracts were prepared from the livers of 2 males treated with vehicle or 3 with gefitinib. Immunoblot techniques were used to detect phospho-Smad5 and total Smad5.

Discussion

In this study we investigated the molecular mechanisms of the sexual dimorphism in hepcidin expression and tissue iron loading observed in Bmp6-deficient mice. We report that testosterone robustly suppresses hepcidin expression in vivo by enhancing EGFR signaling in the liver. EGF, which is produced by the liver and regulates hepatic regeneration, is also secreted in relatively large quantities by salivary glands and kidneys. It was recently shown to block iron-induced hepcidin mRNA in mice by suppressing BMP signaling upstream of the hepcidin promoter.[21] We therefore hypothesized that the previously described changes in hepatic EGF receptor expression induced by testosterone[22, 23] could modulate the autocrine, paracrine, and possibly endocrine effects of endogenous EGF on the liver, particularly on hepcidin expression. We explored this possibility further and indeed observed that hepcidin and Egfr mRNA levels in the liver vary in opposite directions. In contrast to hepcidin, hepatic Egfr expression and the amount of phosphorylated Egf receptors are higher in both wild-type and Bmp6−/− males than in females. They are also reduced by castration but stimulated by testosterone. Most interestingly, the selective inhibition of Egfr signaling by gefitinib in males markedly induced hepcidin mRNA levels, confirming that, by increasing hepatic Egfr expression, testosterone amplifies the inhibitory effect of EGF on hepcidin transcription.

The molecular mechanisms linking testosterone-induced Egfr signaling to hepcidin down-regulation in vivo are not yet understood and remain to be elucidated. We observed that, in Bmp6−/− mice, activation of Egfr signaling (in noncastrated males or in females treated with testosterone) coincided with a decrease in the levels of Smad5 phosphorylation. Therefore, repression of hepcidin expression in these mice is most likely the consequence of reduced Smad1/5/8 protein binding to the BMP/Smad response elements in the hepcidin promoter, as suggested by recent chromatin immunoprecipitation assays performed with liver tissue of testosterone-treated mice.[17]

In contrast to males, female Bmp6−/− mice do not massively accumulate iron in their pancreas, their heart, their kidneys, or, as reported in a separate study, in their retina.[24] Although the same sexual dimorphism in tissue iron accumulation was observed in hemojuvelin (Hjv)-deficient mice (J. Krijt, pers. commun.), it is not seen, as shown here, in Hamp−/− mice. Similar to Bmp6−/− mice, Hjv−/− males have much lower Hamp mRNA levels than females.[11] It can therefore be deduced from these observations that protection against iron loading of extrahepatic tissues requires a certain level of hepcidin expression, a condition met in females lacking either Bmp6 or Hjv, but not in males, nor in Hamp-deficient mice. It is possible that, when hepcidin levels are negligible, NTBI accumulates in the blood circulation and is taken up by extrahepatic tissues through transmembrane transporters such as ZIP14, up-regulated in pancreas and heart during iron overload.[25, 26] Notably, liver and extrahepatic tissues of chronically transfused thalassemia major patients also have different kinetics of iron uptake. Whereas the liver is the storage depot for both transferrin- and nontransferrin-bound iron, extrahepatic tissues predominantly load circulating NTBI, and this difference is believed to be responsible for the threshold behavior exhibited by pancreatic and cardiac iron accumulation with respect to liver iron concentration.[27]

Interestingly, the modulation of Egfr signaling and hepcidin levels by testosterone is not restricted to Bmp6−/− mice but is also seen in wild-type mice. However, in the presence of functional Bmp6, basal hepcidin levels are high even in males and, as could be expected, the observed gender differences in hepcidin expression have no impact on ferroportin expression or tissue iron loading (data not shown). These findings suggest that testosterone-induced repression of hepcidin expression becomes functionally important in disorders characterized by reduced capacity for hepcidin synthesis.

Several examples in the literature suggest that our observations in mice are relevant to humans. Testosterone stimulates erythropoiesis in both sexes[28] and excessive erythrocytosis is the most common serious adverse event associated with testosterone therapy in older men.[29] However, the mechanisms by which testosterone stimulates erythropoiesis do not involve, as could be expected, stimulation of erythropoietin secretion[30] or erythroid progenitor cells[31] but, interestingly, suppression of hepcidin.[32] Another example of the suppressive role of testosterone on hepcidin in humans comes from investigations on women with polycystic ovary syndrome, a frequent androgen excess disorder, who present with mild iron overload.[33] Notably, there is a negative correlation between serum hepcidin levels and serum free testosterone concentrations in these patients, suggesting that increased body iron stores are due to the down-regulation of hepcidin by testosterone.[34] The mechanisms by which testosterone suppresses hepcidin levels in humans are still unknown but, as suggested by the present study, may involve testosterone induction of EGFR expression in the liver and amplification of the inhibitory effect of EGF on hepcidin transcription.

In contrast to rodents, which do not have menstruation cycles, regular menstrual blood losses in premenopausal women have a suppressive effect on hepcidin transcription that may be equivalent to that of testosterone in men. This may explain why, in the general population of Val Borbera, serum hepcidin levels are about the same in men and women with ferritin ≤250 ng/mL.[36] However, as shown in Bmp6−/− animals, testosterone-induced repression of hepcidin expression in humans appears to have a functional impact during homeostatic stress from various disorders that result in iron loading and/or reduced capacity for hepcidin synthesis. Indeed, men of Val Borbera with serum ferritin >250 ng/mL had much lower hepcidin levels than women with similarly elevated ferritin levels.[35] Furthermore, among patients with HFE-related hereditary hemochromatosis, a prevalent iron overload disorder where hepcidin levels are low relative to the iron load, many more males than females present with symptoms and signs of hemochromatosis even though the pathogenic genotypes are equally distributed between genders. Men accumulate more iron and have a higher prevalence of liver injury.[12] Although blood loss experienced with menstruation and childbirth may account for at least part of these gender differences, a direct contribution of testosterone, as shown in the present study, to the more severe phenotypic expression of hemochromatosis in males cannot be ruled out. Juvenile hemochromatosis is a rare but severe form of the disease caused by mutations in the HJV gene.[36] Interestingly, cardiac failure appears more frequent in males than in females,[37] and further investigations are needed to determine whether, as in Hjv−/− mice, males have higher cardiac iron content than females. Patients with alcoholic liver disease or chronic hepatitis C frequently display, similar to animal models of alcohol and hepatitis C, reduced hepcidin expression.[12] As in hereditary hemochromatosis, the relative lack of hepcidin induction by iron in chronic hepatitis results in chronic hyperabsorption of dietary iron. Parenchymal iron deposition added to preexisting liver injury from hepatitis worsens disease prognosis and increases the risk of developing cirrhosis and hepatocellular carcinoma. Notably, men and women exhibit clinical differences in the severity of these liver diseases[12] and, in mice, males have been reported to display significantly lower hepcidin expression compared to females following acute alcohol exposure.[38] Elevated liver tissue concentrations of EGF in chronic viral and alcoholic hepatitis could have a higher impact on hepcidin transcription in males who express more EGF receptors than females, and this could contribute to the gender differences reported in the progression of chronic liver diseases.

In summary, this study provides evidence for negative regulation of hepcidin transcription by testosterone, which is mediated through the testosterone-dependent up-regulation of EGF receptors in the liver. Testosterone-induced repression of hepcidin is responsible for major tissue iron accumulation, particularly in the heart and the pancreas of Bmp6−/− mice, and may contribute to the sexual dimorphism observed in diseases with altered hepcidin expression such as hereditary hemochromatosis, chronic hepatitis, or alcoholic liver disease. In the future, novel therapeutic strategies targeting the testosterone/EGF/EGFR axis may thus be useful for inducing hepcidin expression in patients with iron overload and/or chronic liver diseases.

Acknowledgment

The authors thank Pr. Elizabeth Robertson (Dunn School of Pathology, University of Oxford, UK) and Dr. Sophie Vaulont (Institut Cochin, Paris) for kindly providing the Bmp6−/− and Hamp−/− mice, respectively. They also thank Dr. Tara Arvedson and Dr. Barbra Sasu for making available their mouse hepcidin-1 ELISA, Dr. Martin Vokurka and Dr. Jan Krijt for sharing their irradiation protocols and data on Hjv-deficient mice, Dr. Jerome Boué for help with flow cytometry analysis of bone marrow, Florence Capilla and Dr. Talal Al Saati (Experimental Histopathology Platform, Toulouse Purpan) for skilled advice, and Maryline Calise (Inserm, US006, Toulouse) for technical assistance and help in the mouse breeding.

Ancillary