Hepcidin: what every gastroenterologist should know

In health, the body content of iron is maintained within fairly narrow limits to provide sufficient iron for synthesis of ferroproteins essential for oxygen transport and catalysis yet avoid the toxic effects of excess. Early experiments provided tantalising evidence for a humoral regulatory factor. Serum obtained from iron repleted rats inhibited iron absorption in normal rat duodenum.¹ For many years this putative factor remained elusive. However, recent studies indicate that a peptide hormone, hepcidin, may play a crucial regulatory role in normal iron homeostasis, haemochromatosis, and the anaemia of chronic disease.²

Hepcidin was identified not from studies of iron homeostasis but from investigation of novel antimicrobial peptides in body fluids. Krause and colleagues³ screened human blood ultrafiltrate (a source of the antimicrobial peptides defensins) for small cysteine rich peptides. A 25 amino acid peptide with a mass of 2.7 kDa was found, containing a remarkable eight disulphide bonded cysteine residues. It showed antimicrobial activity against some bacteria and the yeast Saccharomyces cerevisiae. Quantitative reverse transcription-polymerase chain reaction showed predominant expression in liver. The peptide was called LEAP-1 (liver-expressed antimicrobial peptide).³ The cDNA sequence predicted an 84 amino acid pre-propeptide with two cleavage sites. The first predicts cleavage of the N terminal endoplasmic reticulum targeting signal sequence (amino acids 1–24). The second consensus cleavage site for prohormone convertases would yield the central prodomain (amino acids 25−59) plus the 25 amino acid C terminal peptide (also called hepcidin-25; amino acids 60–84). A slightly truncated 20 amino acid peptide was also found (hepcidin 20; amino acids 65–84). Independently, Park and colleagues⁴ isolated the same cysteine rich peptide from human urine and named it hepcidin, reflecting its hepatic expression and bactericidal properties. In addition to hepcidin-20 and hepcidin-25, a minor 22 amino acid form was found (hepcidin-22; amino acids 63–84). The antimicrobial activity of urinary hepcidin was inhibited by 100–150 mM NaCl,⁴ similar to the normal range for plasma sodium (~135–145 mM). Whether circulating plasma hepcidin plays a significant antimicrobial role in vivo therefore remains unclear.

Experiments in mice have shown that hepcidin plays a role in iron regulation. Pigeon et al identified mouse liver hepcidin mRNA because it was induced by carbonyl iron overload.⁵ The mouse has two hepcidin genes in tandem (probably arising from a duplication event), located adjacent to the upstream stimulatory factor 2 (Usf2) transcription factor gene.⁵ Mice that lack hepatic hepcidin expression, due to targeted disruption of the adjacent Usf2 gene, developed predominant iron loading of parenchymal cells of the liver and pancreas with relative sparing of the macrophage rich spleen.⁶ Similar patterns of iron overload in Hfe knockout and Usf2 (hepcidin deficient) knockout mice suggested that hepcidin and Hfe may function in the same regulatory pathway. Hepcidin was proposed to be a negative regulator (that is, “inhibitor”) of iron absorption in the duodenum and of iron release from macrophages.⁶ This was supported by demonstration that transgenic mice constitutively expressing liver hepcidin died within a few hours of birth with decreased body iron levels and severe microcytic hypochromic anaemia.⁷

Further observations implicate hepcidin in the anaemia of chronic disease. This is the most common anaemia of hospitalised patients, often associated with infection, cancer, and autoimmune diseases. There is a diminished response of erythroid precursors to erythropoietin, decreased survival of erythrocytes, impaired iron absorption, and retention of iron in macrophages. The net effect would be to decrease the iron available to erythroid precursor cells.^8,9 Two patients with severe iron refractory anaemia associated with hepatic adenoma showed spontaneous resolution of anaemia after removal of the tumour.⁹ The adenomatous tissue expressed inappropriately high levels of hepcidin mRNA. On the basis of these observations, it was proposed that hepcidin may be important in the pathogenesis of anaemia of chronic disease.⁹ Analysis of urinary hepcidin excretion in patients with anaemia due to chronic infection or severe inflammatory disease provided some confirmatory evidence of increased hepcidin levels in response to inflammation in humans.¹⁰ In these patients, hepcidin excretion correlated with serum ferritin concentration which, like hepcidin, increases in response to both elevated body iron stores and inflammation. Interestingly, injection of erythropoietin into mice was shown to decrease dramatically liver hepcidin expression. Therefore, the efficacy of erythropoietin in non-renal anaemias such as cancer and autoimmune diseases may result from both stimulation of erythropoiesis and inhibition of hepcidin expression, leading to increased plasma iron levels.¹¹

Regulation of hepcidin expression has been investigated in cultured cells. Surprisingly, iron loading of primary human hepatocytes with either ferric ammonium citrate or diferric transferrin decreased hepcidin mRNA levels.¹⁰ This is in contrast with experiments in mice in which dietary or parenteral iron loading increased hepcidin mRNA levels.⁵ This suggested that during iron overload in the whole animal, a signal from non-parenchymal cells may induce hepatocytes to express hepcidin.¹⁰ Mouse liver hepcidin mRNA is induced not only by iron overload but also by treatment with bacterial lipopolysaccharide, hinting that iron and inflammatory cytokines may have a common signalling pathway.⁵ Hepcidin mRNA in primary human hepatocytes was induced either by medium conditioned by monocytes exposed to lipopolysaccharide or by interleukin 6 (IL-6).¹⁰ At the level of regulation by transcription factors, the promoters of human and mouse hepcidin genes contain at least one functional binding site for CCAAT/enhancer binding protein α (C/EBPα), a transcription factor highly expressed in adult liver which appears to stimulate hepcidin expression. Hepatocyte nuclear factor 4α (HNF4α) appears to reduce hepcidin expression.¹²

The medical importance of hepcidin is emphasised by recent studies in human iron overload. Homozygous mutations of hepcidin have been demonstrated in two families with juvenile haemochromatosis.¹³ This disorder is characterised by a rapid rate of iron loading, usually causing presentation before the age of 30 years. There is a high frequency of endocrine and cardiac iron overload, and failure of early diagnosis and treatment may be fatal.¹⁴ Juvenile haemochromatosis is genetically heterogeneous with a second form mapping to chromosome 1q21. Whether this second gene will encode another protein in the hepcidin-HFE pathway remains to be resolved.¹⁵ The potential interaction between hepcidin and HFE has also been suggested by the description of haemochromatosis associated with one mutated allele of both hepcidin and HFE, termed “digenic inheritance”.¹⁶

The classical HFE related form of the disease accounts for more than 90% of well characterised haemochromatosis patients in the UK.¹⁷ In HFE related haemochromatosis and in Hfe knockout mice, there is failure to upregulate hepcidin levels despite hepatic iron overload. This indicates that the HFE protein is involved in regulation of hepcidin levels in response to iron. Disruption of the normal hepcidin response to body iron stores may contribute to the pathogenesis of iron overload seen in HFE related haemochromatosis.^18–20

Recognition of liver produced hepcidin as an iron regulatory factor has challenged the hypothesis that the control of iron absorption in relation to body stores is mediated in duodenal crypts. Hepcidin has revived the concept that some regulation may be mediated by the liver. Rats switched from a control to an iron deficient diet showed changes in hepcidin expression in close temporal relationship with changes in duodenal iron transporter expression. This suggested that hepcidin may act directly on villus enterocytes, pointing to the liver as a regulatory site.²¹ Certainly, siderosis in Kupffer cells, often on a background of mild iron deposition in hepatocytes, is a familiar feature of common liver diseases such as hepatitis C infection and alcoholic liver disease.²² Whether increased concentration of hepcidin, perhaps in response to inflammation, has a role in the hepatic siderosis observed in viral and alcoholic liver disease remains to be demonstrated.

To date, investigation of hepcidin in various clinical conditions has been impeded by the lack of anti-hepcidin antibodies and suitable assays. Recently, a western blot densitometric assay was reported using an antibody raised against hepcidin-25, allowing quantitation of human hepcidin in serum samples.²³ In this issue of Gut, Kulaksiz and colleagues²⁴ describe antisera raised against two unique synthetic peptides corresponding to the predicted human hepcidin prodomain (two antisera to the same peptide) and the mature hepcidin C terminal peptide [see page 735]. All three antisera detected immunoreactive bands estimated at 10 and 20 kDa on western blots of guinea pig and human liver and HepG2 cells. The 10 kDa band only was detected faintly in human serum. Immunohistochemistry of guinea pig and human liver with all three antisera showed staining in hepatocytes with a basolateral expression pattern. This confirms and extends earlier mRNA results, and is consistent with basolateral release of hepcidin into the liver sinusoids.²⁴

A new enzyme linked immunosorbent assay (ELISA) was developed using the prodomain antiserum. The ELISA was used to investigate serum levels of pro-hepcidin immunoreactivity in controls and patients with haemochromatosis, renal anaemia, and chronic renal insufficiency. The mean level of pro-hepcidin immunoreactivity detected in the serum of healthy volunteers was an order of magnitude greater than levels of hepcidin detected in human urine.⁴ Pro-hepcidin immunoreactivity was increased in chronic renal insufficiency patients who had normal haemoglobin levels compared with the healthy control group, despite erythropoietin treatment, which has been observed to decrease hepatic hepcidin expression.¹¹ These observations are therefore consistent with renal metabolism and/or elimination of peptide. Pro-hepcidin was decreased in HFE related haemochromatosis patients (C282Y homozygous) compared with healthy controls,²⁴ consistent with earlier mRNA studies,¹⁸ and indicating defective hepcidin regulation in HFE related haemochromatosis.

Questions and inconsistencies remain regarding the size and identity of hepcidin immunoreactive bands on western blots. The sequence of hepcidin-25 predicts a mass of 2.7 kDa, corresponding to the result determined by mass spectrometry.³ Furthermore, the oxidised synthetic hepcidin-25 co-migrates with native hepcidin-25 in analytical high pressure liquid chromatography and capillary electrophoresis.²⁵ On western blots however, serum hepcidin-25 immunoreactivity has been detected with an apparent mass of 10 kDa, co-migrating with the synthetic peptide,²³ possibly reflecting aggregation of hepcidin monomers.²⁶ Kulaksiz et al detected bands of 10 kDa and 20 kDa with their three antisera on western blots.²⁴ They attributed the 10 kDa band to pro-hepcidin on the grounds that it was detected by antisera to both the mature hepcidin peptide and to the prodomain. However, the predicted mass of pro-hepcidin is 6.9 kDa; that of pre-pro-hepcidin is 9.4 kDa (ExPASy programme: http://ca.expasy.org/tools/peptide-mass.html).²⁷ It might be informative to compare the mobility of synthetic peptides versus the bands observed in human serum in this system.²⁴ Perhaps attention to denaturation conditions and accurate verification of experimentally determined molecular masses may help to clarify the discrepancies.

Is pro-hepcidin an active player in iron regulation or merely a precursor and an indirect measure of hepcidin potential? Interestingly, a homozygous point mutation of the human hepcidin prodomain has been detected, predicting a threonine to methionine substitution at amino acid 31 (T31M). This individual had normal serum iron indices, normal haemoglobin, and normal mean corpuscular volume.²⁸ However, this residue is not conserved in evolution: both wild-type mouse hepcidin genes have a methionine at the position equivalent to human threonine 31. The normal phenotype associated with homozygous T31M mutation therefore does not resolve the role of pro-hepcidin.

In summary, hepcidin is a peptide hormone with important biological effects on iron metabolism. It has sequence homology to antimicrobial peptides, although whether this function is conserved in vivo remains to be demonstrated. Hepcidin is derived by cleavage of a pre-propeptide; questions remain about the processing and activity of the peptide products. It has been implicated as a regulator of body iron stores and mutations have been found in some families with juvenile haemochromatosis. Iron balance is probably sensed by Kupffer cells and signalled to hepatocytes by cytokines such as IL-6. Hepcidin expression is also induced by inflammatory agents giving it a potential role in the anaemia of chronic disease. The reagents and assays described²⁴ will permit investigations of the role of hepcidin in clinical samples in relation to the siderosis observed in viral and alcoholic liver diseases. This peptide is a fascinating addition to the family of molecules involved in iron metabolism. Future studies of its action and regulation in vivo and in vitro should fill in some of the tantalising gaps in our understanding of this increasingly complex field.