To avoid iron deficiency and overload, iron availability is tightly regulated at both the cellular and systemic levels. The liver peptide hepcidin controls iron flux to plasma from enterocytes and macrophages through degradation of the cellular iron exporter ferroportin. The hepcidin-ferroportin axis is essential to maintaining iron homeostasis. Genetic inactivation of proteins of the hepcidin-activating pathway causes iron overload of varying severity in human and mice. Hepcidin insufficiency and increased iron absorption are also characteristic of anemia due to ineffective erythropoiesis in which, despite high total body iron, hepcidin is suppressed by the high erythropoietic activity, worsening both iron overload and anemia in a vicious cycle. Hepcidin excess resulting from genetic inactivation of a hepcidin inhibitor, the transmembrane protease serine 6 (TMPRSS6) leads to a form of iron deficiency refractory to oral iron. Increased hepcidin explains the iron sequestration and iron-restricted erythropoiesis of anemia associated with chronic inflammatory diseases. In mice, deletion of TMPRSS6 in vivo has profound effects on the iron phenotype of hemochromatosis and beta-thalassemia. Hepcidin manipulation to restrict iron is a successful strategy to improve erythropoiesis in thalassemia, as shown clearly in preclinical studies targeting TMPRSS6; attempts to control anemia of chronic diseases by antagonizing the hepcidin effect are ongoing. Finally, the metabolic pathways identified from iron disorders are now being explored in other human pathologic conditions, including cancer.
Iron is essential for multiple cell functions, but is also potentially deleterious because of its ability to generate free oxygen radicals. Due to the absence of an active excretory mechanism, iron balance in mammals is maintained by limiting its intestinal uptake and by continuously recycling and reusing cellular iron. Multiple safety mechanisms, such as binding to chaperone proteins, storage in ferritin, and export through ferroportin (FPN), protect cells from free iron toxicity. The mechanisms of cellular iron handling are summarized in Figure 1. Iron is used in mitochondria for heme synthesis and iron sulfur cluster biogenesis. There is increasing interest in the latter pathway because iron sulfur clusters are prosthetic groups for key enzymes of DNA duplication, repair, and epigenetics. Iron-regulatory proteins (IRPs) and hepcidin exert iron homeostatic control at the cell and systemic levels, respectively.1 Disruption of iron control mechanisms leads to genetic iron disorders and may also contribute to the pathophysiology of common pathologic conditions including inflammation, neurodegeneration, metabolic disorders, and cancer.
At the cellular level, IRP1 and IRP2 orchestrate the coordinated expression of iron importers (transferrin receptor 1 [TFR1] and divalent metal transporter 1 [DMT1]) and of storage (ferritin light and heavy chains) and export (FPN) proteins. IRPs regulate their targets posttranscriptionally by binding to special stem loop elements in the untranslated regions of mRNA-encoding proteins involved in iron metabolism; binding activity is high in iron deficiency and hypoxia and is suppressed by iron and oxygen (for review, see Hentze et al1 ). Recently, differential target specificity of the 2 IRPs has been identified, with IRP1 specifically controlling the hypoxia mediator HIF2-alpha2 and IRP2 controlling ferritin.3 Control of HIF2-alpha by IRP is one of the multiple links between iron and hypoxia. Undoubtedly, conditional deletion of either IRP in animal models will clarify other tissue- and IRP-specific roles.
At the systemic level, the liver peptide hepcidin regulates iron homeostasis by binding and degrading the sole cellular iron exporter FPN, which is highly expressed at the basolateral surface of duodenal enterocytes and on the cell membrane of macrophages. In this way, hepcidin restricts the amount of iron delivered to its plasma carrier transferrin.4 The concentration of both circulating and tissue iron provides distinct signals that modulate hepcidin. The result is low hepcidin and active iron delivery to plasma in iron deficiency and high hepcidin with reduced iron flux to plasma in iron overload (Figure 2).
How IRP-based and systemic regulatory pathways interconnect and work together in general iron homeostasis is only partially understood and is a subject of intensive investigation.
Hepcidin transcription in hepatocytes is dependent on the bone morphogenic protein (BMP)-SMAD signaling cascade (Figure 3A).1 BMP6 is the iron-related BMP receptor (BMPR) ligand in vivo, as shown by Bmp6−/− mice, which have severe iron overload and very low hepcidin.5 In the liver, BMP6 is mainly expressed in nonparenchymal cells such as sinusoidal endothelial and Kupffer cells.6 Binding of the ligand to BMPR complex on the hepatocyte surface triggers phosphorylation of SMAD proteins, which translocate to the nucleus to activate target genes including hepcidin (Figure 3A). In mice, liver-specific disruption of the BMPR ALK2 and ALK-3 or of SMAD4 molecule results in iron overload with low hepcidin.7 Hemojuvelin (HJV), a protein mutated in juvenile hemochromatosis type A (Table 1), is the essential BMP coreceptor in this pathway. In humans, its inactivation causes severe, early onset iron overload indistinguishable from hemochromatosis caused by inactivation of the hepcidin gene itself.8 Hemochromatosis type 1, 2 and 3 (Table 1) and their corresponding murine models show defective BPM signaling that results in hepcidin insufficiency. Whereas the function of membrane-HJV in hepcidin activation is clearly defined, that of other hemochromatosis proteins (TFR2 and HFE) remains uncertain. To add further complexity, in mice, hepcidin, Hfe, and Hjv are modulated by microRNA miR-122.9
In inflammation, hepcidin is activated by IL-6, IL-1-beta, and other cytokines as well as by lipopolysaccharide through the JAK2/STAT3 signaling pathway (Figure 3B). The integrity of the BMP pathway is essential for a full hepcidin response in inflammation1 and the cross-talk between the 2 pathways is the subject of intensive investigation.
Erythropoiesis consumes most of the recycled and absorbed iron (∼25 mg/daily). Therefore, it is not surprising that hepcidin expression is suppressed in all conditions characterized by increased erythroid demand for iron and elevated circulating erythropoietin levels, such as iron deficiency, hypoxia, and erythropoietic expansion. A key hepcidin inhibitor was discovered by positional cloning of the gene responsible for the phenotype of the Mask mice, which have microcytic anemia, high hepcidin, and are unable to absorb oral iron.10 Matriptase-2, encoded by TMPRSS6, is a liver-expressed type II transmembrane serine protease that is inactivated in Mask mice due to truncation of the protein and loss of the catalytic domain. TMPRSS6 interacts with and cleaves the BMP coreceptor HJV, switching off BMP signaling and thereby decreasing hepcidin transcription (Figure 3A).11 It is unknown whether, in vivo, TMPRSS6 cleaves other substrates. In iron deficiency, the function of TMPRSS6 is essential to suppress hepcidin and to allow iron absorption. In vitro, the expression of TMPRSS6 is up-regulated by hypoxia and iron deficiency and its proteolytic activity is inhibited by hepatocyte growth factor activator inhibitor type-2 (HAI-2), an inhibitor of the homologous protease matriptase-1.12 The regulation of TMPRSS6 in vivo is largely unknown.
Several other candidate hepcidin inhibitors have been proposed. Any direct effect of the hypoxia mediator HIF-1-alpha on the hepcidin promoter remains uncertain. Growth differentiation factor 15 (GDF15) and twisted gastrulation protein homolog 1 (TWGS1) are released by erythroblasts and have been proposed as hepcidin suppressors, especially in thalassemia, but their physiologic or pathologic role remains uncertain.1 The longstanding known soluble (s)TFR1 and the recently identified sTFR2 (C.C., unpublished data, April 2013) could in theory contribute to hepcidin suppression, but experimental proof is lacking. In vitro, soluble HJV produced by furin cleavage13 competes as a decoy molecule with membrane-HJV for BMP6 binding, but its in vivo function is unclear. HJV is highly expressed in muscle, where it might have an iron-unrelated function, because muscle-specific Hjv-knockout mice have normal iron homeostasis.14 It is conceivable that multiple inhibitors triggered by different signals from iron-deficient, expanded, or abnormal erythropoiesis all converge on the same final pathway.
Lessons from human iron disorders
Genetic disruption of the finely tuned regulation of the hepcidin-FPN axis causes either iron overload or iron deficiency (Table 1). Hemochromatosis, a heterogeneous genetic disorder leading to iron overload and ultimately to organ failure, provided important clues to understanding hepcidin up-regulation, and autosomal recessive iron-refractory iron-deficiency anemia (IRIDA) provided clues to the mechanisms of hepcidin suppression.
Hemochromatosis patients are natural human mutants of iron homeostasis. In the recessive disease, mutations affect genes involved in hepcidin activation (HFE, HJV, HAMP, TFR2); in the dominant type, they affect the hepcidin receptor FPN (Table 1; for review, see Pietrangelo15 ). In hemochromatosis, the severity of iron overload correlates with the degree of hepcidin deficiency, suggesting a hierarchy of the corresponding proteins in the regulatory pathway. HAMP (hepcidin) and HJV, the genes of the most severe juvenile form of hemochromatosis, have a central role in hepcidin regulation, whereas HFE and TFR2 have ancillary roles. In the current model, HFE and TFR2 function as a complex to activate hepcidin1 (Figure 3A); however, distinct disorders result from their individual inactivation. HFE disease has adult onset, low penetrance, and male predominant expression, suggesting a modest biological effect of the protein. Although the number of cases reported is limited, TFR2 hemochromatosis affects both genders and has early onset, but its clinical course is not as severe as the juvenile form. These clinical observations suggest 2 distinct, perhaps age-dependent, mechanisms of hepcidin regulation. In addition, the hepcidin response after an oral iron challenge that increases only plasma and not tissue iron is blunted in HFE but absent in TFR2 patients,16 and the same response is observed in the corresponding animal models.17 The evidence that normal HFE and TFR2 do not compensate for deficient HJV suggests that all of these hepcidin-regulatory proteins converge on the same pathway. Indeed, HFE, TFR2, and HJV in vitro interact to form a cell surface complex.18 The simplest explanation is that TFR2, which binds iron-loaded transferrin, controls hepcidin according to plasma iron levels to avoid iron overload when iron demands are high, as in young individuals. HFE might control hepcidin according to tissue iron concentration to avoid iron accumulation in adults. An additive effect of HFE and TFR2 is evident in a patient who presented with mutations in both genes and had severe disease.15
On the opposite side, the only known genetic disorder with high hepcidin is the recessive IRIDA19 due to TMPRSS6 inactivation, pointing to the unique role of this protease in hepcidin suppression. Inhibition of hepcidin expression in iron deficiency serves to increase the iron supply to plasma. IRIDA patients have moderate anemia, severe microcytosis and hypochromia, very low transferrin saturation, and inappropriately normal/high hepcidin levels.19 Anemia is more severe in children than in adults,20 supporting the concept that TMPRSS6 mediates the physiologic response to increased iron demand.
All of these clinical observations point to a single iron-responsive hepcidin regulatory pathway with hepcidin production reflecting the balance between positive (BMP6) and negative (TMPRSS6) hepcidin regulators (Figure 4). Interestingly, hepcidin up-regulation by increased plasma iron does not require BMP6 to increase. One possible mechanism causing a rapid increase in hepcidin when BMP6 is low is blocking TMPRSS6 activity. Hemochromatosis resulting from mutations of FPN is dominant (Table 1) and the homozygous state has not been reported, likely because homozygous inactivation of FPN would be incompatible with life; this again points to the critical importance of cellular iron export and macrophage iron recycling. Disorders of FPN are heterogeneous: mutations that decrease its surface expression or the ability to export iron result in relatively benign iron accumulation in macrophages.15 In contrast, mutations at the hepcidin-binding site of FPN cause true hemochromatosis with parenchymal iron overload because FPN is not degraded by hepcidin (hepcidin resistance). The similarity of the phenotype caused by hepcidin deficiency and by FPN mutations that cause resistance to hepcidin attests to the essential role of the hepcidin-FPN interaction in iron homeostasis.15
Hepcidin insufficiency causes the development of secondary iron overload also in “iron-loading anemias,” which include the non-transfusion-dependent thalassemia syndromes such as thalassemia intermedia21 and congenital dyserythropoietic and inherited nonsyndromic sideroblastic anemias (Table 1). Beta-thalassemia intermedia, which has a clinical course of severity intermediate between transfusion-dependent patients and asymptomatic carriers, is the prototype of conditions characterized by ineffective erythropoiesis and high iron stores. Despite often severe iron overload, hepcidin is suppressed by the expanded erythropoiesis. The observation of increased iron absorption irrespective of high iron stores in iron-loading anemias antedated the discovery of hepcidin.
We learned from patients with inflammatory disorders that hepcidin production is up-regulated by cytokines. Hepcidin is an acute phase protein and an essential mediator of the complex anemia of inflammation or anemia of chronic diseases, a multifactorial form of anemia present in numerous disorders22 in which the blockage of iron absorption and recycling plays a major role. In addition to the systemic effects of increased hepcidin production by the liver, inflammatory macrophages also express hepcidin and may induce iron retention by an autocrine mechanism. The blockage of macrophage iron recycling and the resulting hypoferremia is considered a protective mechanism against extracellular pathogens, likely reflecting the true “antimicrobial” function of hepcidin. Recent studies indicate that serum ferritin is predominantly secreted by macrophages; if so, hepcidin-induced iron sequestration in macrophages and the resulting stimulation of ferritin synthesis would explain the high serum ferritin observed in inflammation and also the high correlation between serum hepcidin and ferritin levels reported not only in healthy subjects23 but also in inflammation.
Lessons from animal models of TMPRSS6 inactivation
Tmprss6−/− mice show the same phenotype of iron deficiency with high hepcidin described in Mask mice. Crossing Tmprss6−/− mice with iron-loaded mice provided important insights into the hierarchy of the hepcidin pathway proteins (Figure 5). Tmprss6-Bmp6 double knockout mice are as severely iron loaded as Bmp6−/− mice, implying that the function of Tmprss6 requires an active Bmp-Smad pathway.24 Tmprss6-Hjv double mutant mice are as iron loaded as Hjv−/− mice, a finding consistent with Hjv being downstream of Tmprss6 and likely its substrate.25 On the contrary, the genetic loss of Tmprss6 in Hfe−/− mice reverts the body iron status to an IRIDA-like phenotype,26 indicating that Hfe acts genetically upstream of Tmprss6. Hfe was hypothesized to reduce the activity of Tmprss6, promoting Bmp/Smad signaling and increasing hepcidin.26 In addition, Tmprss6-TFR2–null mice have iron deficiency, with increased RBC and reticulocyte count and even more severe microcytosis than Tmprss6−/− mice27 (C.C., unpublished data, April 2013). Hepcidin is increased, although it does not attain the levels found in Tmprss6−/− mice, likely because the expanded erythropoiesis in the double mutants drives partial hepcidin suppression. Interestingly, TFR2 is a component of the erythropoietin receptor complex and is necessary for its efficient transport to the erythroblast surface.28 The development of a chimeric mouse with TFr2-deficient BM is in progress in our laboratory and hopefully will help to clarify the function of TFR2 in erythropoiesis.
Unexpected results were obtained when crossing Tmprss6−/− with Hbbth3/+ mice, which recapitulate features of thalassemia intermedia patients, such as microcytic anemia, ineffective erythropoiesis, splenomegaly, low hepcidin levels, and liver iron overload. As expected, the ablation of Tmprss6 increased hepcidin expression and reduced liver iron concentration, but, surprisingly, it also decreased ineffective erythropoiesis and spleen size, improving anemia and erythrocyte survival.29 These results provided the proof of principle for novel thalassemia therapeutic strategies based upon Tmprss6 manipulation. In addition, the characterization of this double mutant mouse showed that full hepcidin suppression by ineffective erythropoiesis requires a functional Tmprss6, in agreement with the observation that Tmprss6−/− mice are resistant to the effect of exogenous erythropoietin.27,29 Only hypoxia seems to fully suppress hepcidin in Tmprss6−/− animals,27 suggesting an effect downstream of Tmprss6 and consistent with the concept of multiple inhibitors of the pathway.
Lessons from genome-wide association studies for iron and erythrocyte parameters
Several genome-wide association studies have shown that common TMPRSS6 genetic variants are associated with erythrocyte traits and iron parameters, highlighting a role for TMPRSS6 in the control of erythropoiesis in healthy individuals.30 A common associated single nucleotide polymorphism (rs855791) causes a nonsynonymous alanine to valine change in the catalytic domain of the protease (A736V), leading to the hypothesis that the hematological effects are mediated by variations of hepcidin levels secondary to the different inhibitory activity of TMPRSS6 polymorphic alleles. The rs855791 allele (736V), associated with low MCV and MCH, induced the hepcidin promoter more efficiently in an in vitro assay and was associated with higher serum hepcidin and lower iron levels than the 736A allele in a large series of healthy individuals.31 Variants of iron genes, such as TMPRSS6 and HFE, are associated with serum hepcidin, iron parameters and erythrocyte traits, underlining the influence of hepcidin and iron levels on normal erythropoiesis.31 These results further reinforce the concept that TMPRSS6 modulates the BMP-SMAD pathway in physiological conditions and that hepcidin production results from the balance between opposing forces (Figure 4). This also implies that the role of TMPRSS6 is especially relevant when BMP6 is low. As a corollary, in the latter condition, the only possibility for rapidly increasing hepcidin expression is to suppress TMPRSS6 activity. From these data, TMPRSS6 appears to be an important target for the manipulation of hepcidin levels in different pathological conditions.
TMPRSS6 as a therapeutic target in disorders of low hepcidin
Because the hepcidin pathway is deranged in several disorders, its manipulation is an attractive novel therapeutic strategy. Hepcidin agonists might replace hepcidin when insufficient, restoring the correct balance and controlling iron overload in disorders with low hepcidin (Table 1). Small synthetic peptides (minihepcidins) may decrease serum iron in healthy mice and prevent iron overload and promote at least partial iron redistribution in hepcidin-deficient mice (Table 2).32 The pharmacologic inhibition of TMPRSS6 is emerging as an alternative strategy to increase hepcidin levels. This strategy has been exploited in preclinical studies using Tmprss6 small interfering RNA formulated in lipid nanoparticles with high liver affinity33 or by Tmprss6 allele-specific oligonucleotides34 (Table 2). Tmprss6 silencing successfully increased hepcidin levels and reduced serum and liver iron concentration in both hemochromatosis Hfe−/− and thalassemia Hbbth3/+ murine models. In thalassemic mice, the effect overlaps that observed in the Tmprss6−/−Hbbth3/+ double mutant,29 with improvement of RBC maturation and survival and partial correction of anemia. The positive effect was shown to be thalassemia specific because Hfe−/− mice showed mild anemia as a side effect of Tmpss6 inhibition.
On the opposite side, excessive hepcidin production can be antagonized at different levels. Antagonists are under development with the aim of reducing excessive hepcidin in anemia of chronic diseases and are being tested in animal models of inflammation (Table 2). However, it remains to be proven that reduction of hepcidin alone can improve anemia in inflammatory disorders because of the multifactorial nature of anemia of inflammation.
Iron as a modifier of erythropoiesis
Why does enhancing hepcidin production improve anemia in beta-thalassemia? The analysis of the cross-talk between iron homeostasis and erythropoiesis in Hbbth3/+ mice led to the conclusion that limiting iron may be beneficial in transfusion-independent beta-thalassemia. Beta-thalassemia erythroblasts are unable to produce adequate amounts of beta globin chains but synthesize normal amounts of alpha chains and heme. Excess unpaired alpha chains causes severe oxidative damage through hemichrome formation that induces apoptosis and death of the erythroid precursors, leading to “a high proliferation low differentiation state.”21 Limiting iron might reduce the proliferation of immature cells and increase RBC maturation. Positive effects were reported first in Hbbth1/th1 mice, a model of mild thalassemia intermedia, treated by infusions of exogenous transferrin35 ; then in Hbbth3/+ mice engineered to overexpress moderate amounts of hepcidin36 ; and finally by Tmprss6 pharmacological ablation.33,34 In all cases, the RBC count was increased, although MCV and MCH were further reduced. This suggests that in all of these situations, less iron was delivered to a higher number of erythroid precursors than in control thalassemia animals. Because heme is a strong regulator of protein translation in erythroblasts, it is highly likely (but not yet proven) that decreased heme synthesis due to low iron impairs alpha globin translation, reducing globin chain imbalance, oxidative damage, and ineffective erythropoiesis.37
Anemia may also improve after iron removal by chelators in sideroblastic anemia, which is characterized by mitochondrial iron accumulation in ringed sideroblasts and excessive reactive oxygen species formation, and in some myelodysplastic syndromes.38 Limiting iron excess improves erythropoiesis in a proportion of these patients, reinforcing the concept that iron overload has a negative effect on erythropoiesis and that iron is a modifier of erythroid maturation.
Potential role of hepcidin in common nonhematologic disorders
Derangement of iron metabolism at the systemic and/or cellular level may accompany common pathological processes such as infections, inflammatory and metabolic disorders, neurodegeneration, and cancer. Hepcidin production by inflammatory macrophages to control local iron availability is a phenomenon that might be relevant in disorders characterized by low-grade inflammation, such as obesity, diabetes, and metabolic syndrome.
The hepcidin-FPN axis has been proposed to mediate the acute and chronic changes in iron distribution that contribute to host defense in major infections.39 The relevance of iron for cell proliferation and its propensity to generate reactive oxygen species explain the increasing interest for the study of iron metabolism in cancer, considering that hypoxia and inflammation create a microenvironment that favors iron supply to neoplastic cells. The hepcidin-FPN axis plays an important role in breast cancer, in which cell retention of iron due to low FPN expression or the high hepcidin/high FPN combination is a prognostic signature of cancer itself,40 an issue worth investigating in hematologic malignancies as well.
The lessons learned from rare genetic iron disorders are beginning to shed light on important pathogenic or protective mechanisms in major human diseases.
This work was partially supported by Telethon Fondazione Onlus Rome (Grants GGP12025 and MIUR PRIN 2010-2011) and the Italian Ministry of Health (Grant RF-2010-2312048).
Conflict-of-interest disclosure: The author declares no competing financial interests. Off-label drug use: None disclosed.
Clara Camaschella, MD, Vita-Salute University and San Raffaele Scientific Institute, Via Olgettina, 60, 20132 Milano, Italy; Phone: 39-02-2643-7782; Fax: 39-02-2641-2640; e-mail: firstname.lastname@example.org.