Defective bone morphogenic protein signaling underlies hepcidin deficiency in HFE hereditary hemochromatosis


  • Potential conflict of interest: Nothing to report.


Hereditary hemochromatosis (HH) is a common inherited iron overload disorder. The vast majority of patients carry the missense Cys282Tyr mutation of the HFE gene. Hepcidin, the central regulator of iron homeostasis, is deficient in HH, leading to unchecked iron absorption and subsequent iron overload. The bone morphogenic protein (BMP)/small mothers against decapentaplegic (Smad) signaling cascade is central to the regulation of hepcidin. Recent data from HH mice models indicate that this pathway may be defective in the absence of the HFE protein. Hepatic BMP/Smad signaling has not been characterized in a human HFE-HH cohort to date. Hepatic expression of BMP/Smad-related genes was examined in 20 HFE-HH males with significant iron overload, and compared to seven male HFE wild-type controls using quantitative real-time reverse transcription polymerase chain reaction. Hepatic expression of BMP6 was appropriately elevated in HFE-HH compared to controls (P = 0.02), likely related to iron overload. Despite this, no increased expression of the BMP target genes hepcidin and Id1 was observed, and diminished phosphorylation of Smad1/Smad5/Smad8 protein relative to iron burden was found upon immunohistochemical analysis, suggesting that impaired BMP signaling occurs in HFE-HH. Furthermore, Smad6 and Smad7, inhibitors of BMP signaling, were up-regulated in HFE-HH compared to controls (P = 0.001 and P = 0.018, respectively). Conclusion: New data arising from this study suggest that impaired BMP signaling underlies the hepcidin deficiency of HFE-HH. Moreover, the inhibitory Smads, Smad6, and Smad7 are identified as potential disruptors of this signal and, hence, contributors to the pathogenesis of this disease. (HEPATOLOGY 2010;)

Hereditary hemochromatosis (HH) is an autosomal recessive disorder characterized by iron overload. Unregulated iron absorption from the intestine and release from macrophages primarily affects the liver, the main storage site of this essential mineral. Left untreated, iron excess may progress to hepatic fibrosis, cirrhosis, and hepatocellular carcinoma.1, 2

The most common form of HH (type 1) results from the missense Cys282Tyr (C282Y) mutation of the HFE (hemochromatosis) gene. Although it is a disease of variable penetrance and considerable heterogeneity, the vast majority of patients with HH are homozygous for the C282Y mutation.3, 4 The mutant C282Y HFE protein is unable to bind beta-2-microglobulin and fails to reach the cell membrane, resulting in a misfolded, nonfunctional protein.5 HFE represents a nonclassical major histocompatability complex type 1 molecule expressed in several different tissues. Liver-specific HFE knockout in animal models resulted in a phenotype similar to HFE-HH, suggesting the liver is where HFE exerts its main effect on iron metabolism.6, 7 Upon interacting with diferric transferrin and transferrin receptor 1 (TfR1) at the hepatocyte cell surface, HFE is thought to shift to form part of an iron-sensing complex through its interaction with TfR2.8, 9

The key defect underlying the pathogenesis of all forms of HH is a deficiency of hepcidin, the small peptide hormone produced by the liver, and central regulator of iron homeostasis.10 How the mutant HFE protein can result in deficient hepcidin production remains uncertain, and undoubtedly involves a multifactorial process. Hepcidin controls iron metabolism by targeting ferroportin, the iron exporter located on duodenal enterocytes and macrophages, inducing its internalization and degradation, thus preventing iron absorption.11 Despite significant systemic and tissue iron overload, patients with HFE-HH have inappropriately low levels of hepcidin and continue to absorb excessive amounts of iron.12 HFE knockout mice mirror the human HH phenotype, exhibiting hepcidin deficiency and hepatic iron overload,13, 14 yet curiously do not develop hepatic fibrosis.15-17

Hepcidin is regulated by several factors, including systemic iron and oxygen levels, inflammation, and oxidative stress.18-21 The bone morphogenic protein (BMP)/small mothers against decapentaplegic (Smad) pathway has emerged as the signaling cascade central to the regulation of hepcidin.22-24 Studies from knockout mice have revealed BMP6 and Smad4 as central players in this signaling pathway, as evidenced by severe hepcidin deficiency and massive iron overload in their absence.25-27 Briefly, the BMP6 ligand, induced by iron, engages hepatocyte cell surface receptors BMPR-I and BMPR-II together with the BMP coreceptor hemojuvelin, initiating a signal conveyed intracellularly by phosphorylation of the Smad proteins Smad1, Smad5, and Smad8, which form a complex with the common mediator Smad4, before translocating to the nucleus and activating hepcidin expression.28 Genome-wide liver transcription profiling of mice with varying iron diets recently led to the identification of specific BMP target genes regulated by iron in a similar manner to hepcidin, namely BMP6, the inhibitory Smad molecule Smad7, and inhibitor of differentiation 1 (Id1).29

The association of single-nucleotide polymorphisms (SNPs) in genes of the BMP pathway with HFE-HH disease phenotype has been described previously, although this finding was not substantiated in a follow-up study.30, 31 Recently, impaired BMP/Smad signaling was described in HFE knockout mice models of hemochromatosis of varying genetic backgrounds. By demonstrating inappropriately low levels of the BMP target genes hepcidin (HAMP) and Id1, along with reduced phosphorylation of the Smad1/Smad5/Smad8 complex in HFE knockout mice, these studies revealed a novel role for the HFE molecule in the regulation of iron homeostasis.32, 33 To date, the BMP/Smad signaling pathway has not been characterized in liver tissue from HFE-HH patients.

In this study, we sought to examine the hepatic expression of key molecules of the BMP/Smad pathway in a homogeneous group of untreated C282Y homozygote males with significant iron overload. The data demonstrate an appropriate induction of BMP6, yet failure of up-regulation of BMP target genes hepcidin(HAMP) and Id1 in patients with the mutated HFE protein, in addition to reduced Smad1/Smad5/Smad8 phosphorylation relative to hepatic iron burden. Moreover, up-regulation of Smad6 and Smad7, inhibitors of BMP signaling, occurs in HFE-HH, identifying these molecules as potential aggravators of disease pathogenesis which may act by preventing appropriate induction of hepcidin in the setting of hepatic iron overload.


ALT, alanine aminotransferase; BMP, bone morphogenic protein; HAMP, hepcidin antimicrobial peptide; HH, hereditary hemochromatosis; Id1, inhibitor of differentiation 1; pSmad, phosphorylated Smad; RT-PCR, reverse transcription polymerase chain reaction; Smad, small mothers of decapentaplegic; SNP, single-nucleotide polymorphism; TfR, transferrin receptor; TGFβ, transforming growth factor β.

Patients and Methods


Ethical approval for this study was obtained from the Research Ethics Committee of the Mater Misericordiae University Hospital, Dublin, Ireland. Informed written consent was obtained from all patients involved. Liver tissue was collected from 20 male C282Y homozygotes with HH prior to venesection therapy. Control liver tissue was obtained from four donor livers at time of transplant and three biopsies were from patients undergoing liver biopsy for investigation of abnormal liver function tests that demonstrated no inflammation or fibrosis, and were negative for iron staining. The immunohistochemical component of this study was extended with the addition of a further 10 untreated C282Y homozygotes and three individuals with non-HFE hepatic hemosiderosis associated with chronic viral hepatitis, all of whom had hepatic iron concentrations measured. All study subjects were male and had no cirrhosis. Controls were negative for the HFE mutations C282Y and His63Asp (H63D).

Laboratory Measurements

Following an overnight fast, blood samples were obtained from all patients with HH for serum ferritin, transferrin saturation, iron, total iron binding capacity, full blood count, and liver function tests (including alanine aminotransferase [ALT]). HFE genetic analysis for C282Y and H63D mutations was performed using LightCycler technology (Roche Diagnostics) with Genes-4U ToolSets. Hepatic iron concentration was measured as described.34

Liver Histology

Liver biopsies were independently evaluated by a single histopathologist (A.F.) for grading of hepatocellular iron staining (Perl's Prussian blue stain) and fibrosis (METAVIR score).35

Quantitative Real-Time Reverse Transcription Polymerase Chain Reaction

Liver samples were snap-frozen or placed in RNAlater and stored at −80°C prior to use. Total RNA was extracted using the RNAeasy kit (Qiagen, UK). Reverse transcription was performed using the high-capacity complementary DNA reverse transcription kit (Applied Biosystems [AB], Carlsbad, CA). Gene expression analysis for hamp(hepcidin antimicrobial peptide), bmp6, smad4, smad6, smad7, and Id1 was performed using AB gene expression assay systems, using AB 7000 sequence detector. Samples were analyzed in triplicate. As a validated endogenous control, we used 18S ribosomal RNA.


Immunohistochemistry was performed on formalin-fixed, paraffin-embedded liver tissue (5 μm sections) from the HFE-HH patient cohort, compared to normal liver tissue from hepatectomy specimens remote from colorectal cancer metastases. Tissue was deparaffinized in xylene, antigen-retrieval was performed in citrate buffer by microwave, and tissue was blocked with Powerblock solution (BioGenex Laboratories, Inc., San Ramon, CA). Slides were incubated with rabbit polyclonal anti-BMP6 antibody (1:50 dilution; ProSci, Inc., Poway, CA) at room temperature for 6 hours. In addition, Smad1/Smad5/Smad8 phosphorylation was assessed in formalin-fixed, paraffin-embedded liver tissue from 10 patients with HFE-HH compared to three non-HFE control individuals with hepatic iron excess in whom hepatic iron concentrations were also available. Immunostaining for Smad1/Smad5/Smad8 phosphorylation was performed using a rabbit polyclonal anti-phosphorylated Smad1/Smad5/Smad8 antibody (1:50 dilution; Millipore, Billerica, MA), incubated overnight at 4°C. Immunohistochemistry was performed using the alkaline phosphatase Super Sensitive Link-Label IHC Detection System (BioGenex, Inc.) according to the manufacturer's instructions. Slides were counterstained with hematoxylin. BMP6 and pSmad1/pSmad5/pSmad8 staining was assessed by a single pathologist (A.F.), who was blinded to clinical data.

Statistical Analysis

Differences between HFE-HH and control groups were examined using the Student t test or Mann-Whitney U test where appropriate, and correlations performed using the Spearman Rank method. Gene expression levels were calculated using the delta-delta cycle threshold (Ct) method as previously described,36and normalized to 18S ribosomal RNA. Data analysis was performed using SPSS 13.0 for Windows. A P value of <0.05 was deemed significant.


Clinical and biochemical characteristics of all 20 HFE-HH males are illustrated in Table 1(Mean ± standard deviation unless specified). All HFE-HH patients had significant systemic and hepatic iron overload, as evidenced by elevated serum ferritin (median = 1518 μg/L), transferrin saturation (mean ± standard deviation = 85% ± 15%), and a mean hepatocellular iron-staining grade of 2+ (out of 4). Two patients were found to have precirrhotic livers (grade 3 METAVIR fibrosis) at biopsy. Of note, it was not possible to obtain corresponding data from control liver transplant donors because of confidentiality reasons. Of the three control patients undergoing liver biopsy for abnormal liver function tests, mean age was 51 (±7) years, and serum ferritin was 172 (±51)(serum ALT = 81 ± 31 IU/L). Two patients had minimal fatty change without inflammation or fibrosis and one patient had an entirely normal liver biopsy. All control patients had no hepatocellular iron staining and were negative for the HFE mutations C282Y and H63D.

Table 1. Clinical and Biochemical Characteristics of 20 HFE-HH Males
CharacteristicHH (n = 20)Normal Values
  • *

    Iron deposition graded using Perl's Prussian blue staining (0−4+).

  • Hepatic fibrosis (METAVIR) graded 0−4 (0 = no fibrosis, 1 = mild fibrosis, 2 = moderate with rare septa formation, 3 = bridging fibrosis, 4 = cirrhosis).

Age at diagnosis49 (±9) yearsNA
Ferritin (median)1518 μg/L20-330 μg/L
Transferrin saturation (%)85%(±15)25%-56%
Serum iron42 (±9)μmol/L11-33 μmol/L
ALT90 (±52) IU/L5-40 IU/L
Hepatocellular iron staining*2+ (Range 2-4)0
Fibrosis grade1 (Range 0-3)0

Figure 1 shows the hepatic expression of BMP6 and Smad4 in patients with HFE-HH compared to controls. As expected, expression of BMP6 was significantly elevated in the setting of iron overload (P = 0.019), whereas Smad4 was not up-regulated in HFE-HH compared to controls (P = 0.11). Surprisingly, BMP6 expression did not correlate significantly with serum iron parameters or degree of hepatic iron staining. Diffuse hepatocytic staining for BMP6 was evident at immunohistochemical analysis, without specific cellular or zonal patterns, in contrast to that of normal liver tissue, where BMP6 staining appeared less prominent and was localized to periportal zones (Fig. 2).

Figure 1.

Hepatic expression of key regulators of hepcidin expression, BMP6 and Smad4 in patients with HFE-HH compared to controls. Liver tissue from untreated male patients with C282Y HFE-HH (n = 20) and seven male controls were assessed for expression of BMP6 and Smad4 by quantitative RT-PCR and normalized to 18S endogenous control. Results are mean (± standard deviation). *P = 0.019; NS, nonsignificant.

Figure 2.

BMP6 staining in normal and HFE-HH liver tissue. (A)(Magnification = ×4) Positive normal control liver tissue. BMP6 staining was cytoplasmic in hepatocytes, and appeared restricted to periportal areas (thick black arrows) with a reduction in staining around centrilobular veins (thin black arrows). (B)(Magnification = ×4) Low-power image of Perl's stain on HFE-HH liver biopsy with increased iron deposition (in blue) evident throughout each zone from the portal tract to centrilobular vein, most marked in periportal areas. (C)(Magnification = ×4) BMP staining in HFE-HH liver tissue appears diffusely hepatocellular (thick black arrows) and the zonal pattern seen in (A) is lost. (Inset) Negative control. (D)(Magnification = ×10): High-power image of BMP6 staining in HFE-HH liver demonstrating positive staining around the centrilobular vein (thick black arrows). No cellular polarization was observed. CV, centrilobular vein; PT, portal tract.

Figure 3 illustrates immunostaining for pSmad1/pSmad5/pSmad8 protein in HFE-HH compared with non-HFE iron overload. Although the pattern of positive nuclear staining differed between groups, with patchy immunostaining observed in HFE-HH, contrasted with a diffuse pattern in non-HFE iron overload, no significant difference in the total number of positive-staining cells was found between groups (Fig. 4A). However, allowing for the degree of hepatic iron burden, which was significantly higher in the HFE-HH cohort (Fig. 4B), the amount of pSmad1/pSmad5/pSmad8 staining relative to hepatic iron burden was significantly lower in HFE-HH compared to controls (P = 0.007, Fig. 4C).

Figure 3.

Phosphorylated Smad1/Smad5/Smad8 immunohistochemistry in untreated HFE-HH compared to non-HFE iron overload. Positive nuclear staining in hepatocytes (thick black arrows) contrasted with negative nuclear staining (thin black arrows). Hepatic iron deposition (white arrow). (A) Low-power (×10) and (B) high-power (×20) images of pSmad1/pSmad5/pSmad8 staining in untreated HFE-HH, with contained areas of positive and negative staining. (C) Low-power (×10) and (D) high-power (×20) images of pSmad1/pSmad5/pSmad8 staining in non-HFE iron overload demonstrate a diffuse immunostaining pattern.

Figure 4.

(A) Analysis of pSmad1/pSmad5/pSmad8 immunostaining in HFE-HH compared to non-HFE iron overload. The number of positive pSmad1/pSmad5/pSmad8 cells was not significantly different between HH individuals and secondary iron overload controls. Positive cells were counted by a blinded pathologist (A.F.) using a manual counter, with an average of 300 cells (from three separate fields, at either end and middle of the specimen) taken for each biopsy. (B) Hepatic iron concentration (HIC) was significantly higher in the HFE-HH compared to non-HFE iron overload (*P = 0.01). (C) The ratio of pSmad1/pSmad5/pSmad8 staining to hepatic iron concentration was significantly lower in the HFE-HH cohort compared to controls.

Despite appropriate up-regulation of BMP6 in untreated HFE-HH, Fig. 5 shows hepatic expression of BMP target genes hepcidin (HAMP) and Id1 were not elevated. Hepcidin expression was inappropriately low given the amount of iron-loading in the HFE-HH cohort, although this did not achieve statistical significance (P = 0.097).

Figure 5.

Hepatic expression of BMP target genes HAMP (hepcidin) and Id1 were not elevated in patients with HFE-HH compared to controls. Liver tissue from untreated male patients with C282Y HFE-HH (n = 20) and seven male controls were assessed for expression of HAMP and Id1 by quantitative RT-PCR and normalized to 18S endogenous control. No significant difference for HAMP and Id1 expression was seen in HFE-HH compared to controls (P = 0.097, and P = 0.54, respectively). Results are mean (±standard deviation); NS, nonsignificant.

Expression of Smad7, another BMP target gene and inhibitory Smad (I-Smad), was assessed by quantitative reverse transcription polymerase chain reaction (RT-PCR) in patients with HFE-HH compared to controls. Smad7 was found to be significantly up-regulated in the patient cohort (P = 0.018). Expression of the other principal I-Smad, Smad6, was also significantly elevated in the same group (P < 0.001, Fig. 6).

Figure 6.

Hepatic expression of the inhibitory Smad molecules, Smad6 and Smad7, are significantly up-regulated in patients with untreated HFE-HH compared to controls. Liver tissue from untreated male patients with C282Y HFE-HH (n = 20) and seven male controls were assessed for expression of Smad6 and Smad7 by quantitative RT-PCR and normalized to 18S endogenous control. Results are mean (± standard deviation). *P = 0.018, ***P < 0.001.


Hepcidin deficiency has been demonstrated to be the chief mechanism underlying tissue iron overload seen in patients with HFE-HH. Although hepcidin continues to be synthesized by the liver, its levels are inappropriately low for the systemic iron burden, fueling a cycle of excessive iron absorption and hepatic iron accumulation. Data from mouse models of HFE-HH have suggested that HFE plays a role in the main regulatory pathway of hepcidin production, the BMP/Smad pathway. In this human study, examination of specific genes central to the BMP/Smad pathway and BMP target genes in liver tissue from a homogeneous cohort of untreated male patients with overt HFE-HH indicates that impaired BMP/Smad signaling underlies the hepcidin deficiency seen in this disorder, and corroborates recent findings from HFE knockout mice.

The patient cohort in this study was composed solely of individuals who were male and of Irish origin, a population previously reported to have a 40%-80% disease penetrance.37, 38 All HFE-HH patients had chronic and significant iron overload, and liver biopsies were performed prior to initiation of therapeutic venesection. Since discovery of the HFE gene, the role of liver biopsy in the diagnosis of HH has diminished considerably, and thus cohorts of patients with complete data including histology and hepatic iron concentrations are less available than in the past.

The results outlined in this study confirm several findings from animal models of hemochromatosis. First, BMP6 was up-regulated in iron-loaded patients with HFE-HH compared to controls. As outlined by both Kautz et al.32 and Corradini et al.33, BMP6 expression was induced by iron in both HFE-deficient mice and HFE-wild type mice maintained on an iron-enriched diet, and correlated with increased hepatic iron concentration. Hepatic BMP6 staining displayed a diffuse intracellular pattern and was present in all zones in HFE-HH liver tissue, whereas it was mostly centrilobular and localized to the hepatocyte basolateral membrane in mice with hepatic iron overload. This may reflect the chronicity of iron loading along with the greater extent of iron deposition seen in these patients.39 Iron excess further induced phosphorylation of Smad1/Smad5/Smad8 and expression of the BMP target genes hepcidin (HAMP) and Id1 in HFE-wild type mice, but importantly, this was not seen in HFE-deficient mice.32, 33 These latter findings were mirrored in the HFE-HH patient cohort, because levels of both hepcidin and Id1 remained similar to controls despite iron-loading and elevated BMP6 levels. A nonsignificant trend toward reduced hepcidin expression that was observed in the HFE-HH group was similar to other reports of reduced serum hepcidin levels in HFE-HH, which could be expected to fall further following venesection therapy.40, 41 Moreover, this study (as previously shown in HFE-deficient mice) suggests that induction of BMP6 by iron is not dependent on a functional HFE protein. Expression of Smad4, the central mediator of the BMP signal, was not significantly elevated in the HFE-HH cohort compared to controls. This finding may relate to the abrogated BMP signal, as levels of Smad1/Smad5/Smad8 phosphorylation were inappropriately low relative to iron burden in the HFE-HH cohort. Furthermore, the pattern of pSmad1/pSmad5/pSmad8 immunostaining evident in HFE-HH liver tissue may be relevant to the impairment of the BMP signal, possibly reflecting local regulatory mechanisms at play.

Up-regulation of other BMP target genes, the inhibitory Smad proteins Smad6 and Smad7, was demonstrated in untreated HFE-HH. Indeed, Smad7 expression was seen to follow BMP6 gene expression in mice fed an iron-enriched or iron-deficient diet.29 Smad6 and Smad7 are inhibitors of the transforming growth factor β (TGFβ) family signaling pathway (which includes BMP), and act by preventing phosphorylation of receptor-regulated Smads such as Smad1, Smad5, and Smad8.42 Although Smad6 and Smad7 share close homology, their functions differ.43 Smad6 primarily inhibits BMP signaling (by preventing Smad1 and Smad2 phosphorylation), whereas Smad7 inhibits all TGFβ family members (through effect on Smad2 and Smad3 phosphorylation).44-47 Importantly, Smad7 has been recently identified as a potent suppressor of BMP-mediated hepcidin activation in primary murine hepatocytes, forming part of a negative feedback regulatory loop of hepcidin regulation.48

Smad7 has also been implicated in hepatic fibrosis through alteration of the TGFβ signaling pathway, and its up-regulation in hepatic stellate cells and hepatocytes was associated with a protective effect in animal models of liver fibrosis.49, 50 The degree of fibrosis in this HFE-HH patient cohort was generally mild despite significant iron-loading, and increased Smad7 may have a beneficial role in this disease. Interestingly, overexpression of hepatic TGFβ1, which is associated with hepatic fibrosis51 and known to activate I-Smads,44, 52 was previously reported in iron-loaded patients with HH, and normalized following therapeutic venesection.53 Overexpression of the inhibitory Smads in HFE-HH suggests a specific role for these molecules in interfering with the BMP6 signal induced by iron, preventing an appropriate induction of hepcidin despite iron excess, and leading to self-perpetuation of disease.

In summary, this study demonstrates that failure of iron to induce hepcidin synthesis in the setting of HFE hemochromatosis may result from impaired BMP/Smad signaling, and corroborates recent findings of defective BMP signaling in hemochromatosis mouse models. Furthermore, the inhibitory Smad molecules Smad6 and Smad7 are revealed as potentially important players in the suppression of hepcidin which underlies this disorder.


The authors thank Dr. Jennifer Russell for excellent technical assistance and advice. We also are indebted to Professor Martina Muckenthaler and Dr. Maja Vujic-Spasic for their invaluable correspondence and advice.