The Role of Hepcidin and Iron Homeostasis in Atherosclerosis

Graphical abstract

Atherosclerotic cardiovascular disease is a major burden on global health and a leading cause of death worldwide. The pathophysiology of this chronic disease is complex, involving inflammation, lipoprotein oxidation and accumulation, plaque formation, and calcification. In 1981, Dr. Jerome Sullivan formulated the ‘Iron Hypothesis’, suggesting that higher levels of stored iron promote cardiovascular diseases, whereas iron deficiency may have an atheroprotective effect. This hypothesis has stimulated research focused on clarifying the role of iron in the development of atherosclerosis. However, preclinical and clinical studies have produced contradictory results and the observation that patients with hemochromatosis do not appear to have an increased risk of atherosclerosis seemed incongruous with Sullivan’s initial hypothesis. The ‘paradox’ of systemic iron overload not being accompanied by an increased risk for atherosclerosis led to a refinement of the iron hypothesis focusing on intracellular macropage iron. More recent in vitro and animal studies have elucidated the complex signaling pathways regulating iron, with a particular focus on hepcidin, the master regulator of body iron homeostasis. Bone morphogenetic protein (BMP) signaling is the major pathway that is required for induction of hepcidin expression in response to increasing levels of iron. Strong links between iron homeostasis, BMP signaling, inflammation and atherosclerosis have been established in both mechanistic and human studies. This review summarizes the current understanding of the role of iron homeostasis and hepcidin in the development of atherosclerosis and discusses the BMP-hepcidin-ferroportin axis as a novel therapeutic target for the treatment of cardiovascular disease.

Introduction

Iron is an essential regulator of many cellular and biological processes including oxygen transport and energy metabolism. Intracellular and circulating iron levels are maintained by the hormone hepcidin, a master regulator of iron homeostasis. The transmembrane protein ferroportin is the only known “iron exporter”, which serves to transport ferrous iron from within cells to the plasma.¹ Hepcidin regulates the amount of ferroportin at the surface of enterocytes, hepatocytes, macrophages and enterocytes, and thereby controls the concentration of iron in the circulation.²

Because free iron participates in the Fenton reaction, which leads to the generation of reactive oxygen species (ROS), iron at high levels may contribute to DNA, protein or lipid damage and thereby cause cell death.³ ROS-mediated changes to lipoproteins have a central role in the development of atherosclerosis. Oxidized low-density lipoproteins (oxLDL) trigger endothelial activation and induce inflammatory responses, which promote macrophage recruitment. Macrophages internalize oxLDL and become foam cells within vascular plaques. Vascular plaques can increase in size and thicken the vessel wall causing lumen narrowing. Plaques may also rupture and trigger thrombosis, which leads to acute ischemia and infarction.^4–6 Because of the involvement of iron in the formation of ROS, iron may contribute to the development and progression of atherosclerosis. Approximately 40 years ago, the ‘Iron Hypothesis’ was formulated, suggesting that higher levels of stored iron promote cardiovascular diseases, whereas iron deficiency may have an atheroprotective effect.

Atherosclerosis

Atherosclerosis is a major burden on global health and cardiovascular disease (CVD) is the leading cause of death worldwide.⁷ Modifiable risk factors for the development of atherosclerosis include smoking, hypertension, diabetes, and hyperlipidemia.⁸ Over the last two decades, there has been increased appreciation of the role of oxidative stress and inflammation in the development of atherosclerosis. The opportunity to inhibit inflammation represents a potentially novel approach to prevent and treat atherosclerotic heart disease.⁹

The pathophysiology of atherosclerosis is characterized by a chronic inflammatory response to endothelial injury with associated lipid deposition and infiltration by monocyte-derived macrophages, followed by calcification of arterial vessels^.5,10 In the early stages of atherogenesis, intimal accumulation of apolipoprotein B (apoB)-containing lipoproteins, particularly low-density lipoprotein cholesterol ( LDL-C), activates the overlying endothelium.¹¹ Activated endothelial cells express leukocyte adhesion molecules that bind circulating monocytes, which then diapedese into the intimal wall where they differentiate into macrophages.^12,13 Atherosclerotic plaque progression is characterized by the intracellular accumulation of oxidized lipoproteins (oxLDL) and cholesterol esters within macrophages, leading to foam cell formation, with eventual apoptosis and growth of a necrotic core within the plaque.^9,14 Atheromatous plaques consist of lipids, fibrosis, and inflammatory cells.⁵ Intra-plaque necrosis perpetuates the recruitment of more monocytes, which differentiate into macrophages and release cytokines and growth factors, further fueling the inflammatory process. The density of foam cells within. atherosclerotic plaques and the size of the necrotic core are thought to be key determinants of the vulnerability of plaques to rupture with subsequent vessel occlusion.^4,15

In addition to treating hypertension and diabetes, intensive control of plasma lipid concentration is considered part of the first-line medical treatment for primary and secondary prevention of CVD.^16–18 Numerous randomized clinical trials support the benefits of decreasing the level of harmful lipids (LDL-C) using-3-hydroxy-3-methyglutaryl-CoA (HMG-CoA) reductase inhibitors (statins), to prevent CVD.^19,20 Nevertheless, in clinical practice, some patients do not reach the desired LDL-C target level using statins alone. In addition, some patients still have cardiovascular (CV) events, despite adequate control of LDL-C, diabetes, and hypertension. Efforts have been made to find and test alternative lipid-lowering agents for the primary and secondary prevention of CV-events. Novel lipid-lowering agents, such as cholesterol absorption inhibitors and proprotein convertase subtilisin/kexin type 9 (PCSK-9) inhibitors, have been developed and used in clinical settings.²¹

Along with the increasing appreciation of the role of inflammation in atherosclerosis, the inflammatory response has become a promising target for novel therapies. However, because inflammation is often required for an adequate defense against infections, there is a narrow therapeutic window for the use of anti-inflammatory medications for the treatment of atherosclerotic disease.^{22, 23} The use of aggressive immunosuppressive therapy, as an approach to inhibiting inflammation in atherosclerosis, has only gradually been pursued as a therapeutic option.²⁴ Novel approaches are needed to specifically target aspects of inflammatory processes that may be unique to CVD.

Iron modulates Toll-like receptor (TLR) 4-activated inflammatory responses and increases oxidative stress through the generation of reactive oxygen and nitrogen species.^25,26 ROS activate the transcription factor NF-kB (nuclear factor kappa-B), which regulates expression of genes important for inflammation.²⁷ Because iron is a regulator of the body’s inflammatory response, one promising approach to the prevention of athrosclerosis is to modify iron homeostasis.²⁸

Regulation of iron homeostasis

In a healthy individual, plasma iron levels are maintained at a constant level to provide iron for essential biological and cellular processes, while preventing excess iron accumulation, which can lead to the production of ROS. The central regulator of systemic iron homeostasis is the hepatic peptide hormone hepcidin (Figure 1).²⁹ Initially identified as a defensin-like antimicrobial peptide,^{2, 30,31} hepcidin received its name because of high expression levels in the liver (hep-) and microbicidal activity (-cidin)³¹. By targeting and down-regulating ferroportin (FPN), hepcidin reduces the export of ion from within duodenal enterocytes, macrophages, and hepatocytes.^32,33 Disturbances in the tight regulation of hepcidin expression are associated with various diseases including iron overload disorders, such as hereditary hemochromatosis^2,34, as well as anemia of inflammation.³⁵ The expression of the gene encoding hepcidin (HAMP) is regulated at the level of transcription. ^29,36 Iron deficiency, hypoxia, and/or erythropoiesis inhibit the transcription of , HAMP, whereas increased iron levels and inflammation stimulate hepcidin transcription ³⁷ Inflammatory mediators, such as IL-6, increase hepcidin expression through the Janus kinase/signal Transducer and Activator of Transcription 3 (JAK/STAT3) pathway, resulting in sequestration of iron in macrophages, reduced duodenal iron absorption, and subsequent hypoferremia, which can lead to the development of anemia.^38,39 Increasing the level of hepcidin in states of inflammation (and thereby decreasing serum iron) may have evolved as a defense mechanism against invading pathogens, because iron is an obligate cofactor for the proliferation and growth of some microorganisms.³³

Figure 1. Overview of the regulation of iron homeostasis.

The hepatic hormone hepcidin is the central regulator of systemic iron homeostasis. The expression of the gene encoding hepcidin (hepcidin antimicrobial protein, “HAMP”) is regulated at the transcriptional level by iron via the BMP/Smad signaling pathway, by inflammatory cytokines such as IL-6 and by IL-1β via the JAK/STAT3 pathway, and by erythropoietic demand. While iron and inflammation increase the expression of hepcidin, erythropoietic demand and iron deficiency cause suppression of hepcidin. Hepcidin regulate, the level of the sole known iron exporter, ferroportin, at the cell surface of enterocytes, hepatocytes, erythrocytes and macrophages, and thereby controls the amount of iron in the circulation. IL-6, Interleukin-6; BMP, Bone morphogenetic protein; Smad, mothers against decapentaplegic; JAK/STAT3, Janus kinase/Signal Transducer and Activator of Transcription 3.

The bone morphogenetic protein (BMP) signal transduction system is the major pathway that is required for induction of hepcidin expression in response to increasing levels of iron, with BMP2 and BMP6 being the predominant ligands.^33–40–43 Upon binding of BMP2 or BMP6 to the BMP receptor complex, type II receptors (ActRIIA, ActRIIB, and BMPR2) phosphorylate type I receptors ALK2 (ACVR1) and ALK3 (BMPR1B). Activated type I BMP receptors phosphorylate SMAD1/5/8 proteins, which translocate to the nucleus together with SMAD4, bind to BMP response elements in the hepcidin promoter and induce the expression of hepcidin.⁴⁴ In mice, liver-specific disruption of BMP type I receptors ALK2 or ALK3 markedly decreases hepcidin expression, resulting in iron overload.⁴⁵ This effect is more severe in hepatocyte-specific Alk3 knockout mice, indicating a predominant role for ALK3 in basal hepcidin expression. ALK3 is also required for induction of hepcidin expression by IL-6, showing an essential crosstalk between inflammatory-mediated (JAK/STAT) and iron-mediated (BMP/SMAD) hepcidin regulatory pathways.⁴⁶ Pharmacological inhibition of BMP type I receptors, with small molecules such as dorsomorphin and its derivative LDN-193189, inhibits hepcidin expression and decreases anemia of chronic disease in rodents treated with inflammatory agents.^47,48 Therefore, inhibitors of the BMP pathway may represent new therapeutic options for the treatment of (chronic) inflammatory processes with associated anemia of chronic disease.

The Iron Hypothesis

In 1981, Dr. Jerome Sullivan formulated the ‘Iron Hypothesis’, suggesting that higher levels of stored iron promote cardiovascular diseases, whereas iron deficiency may have an atheroprotective effect.^49,50 The hypothesis was primarily based on the observation that women have a lower risk of cardiovascular disease compared to men. The increased incidence of CVD that occurs in women after menopause, and the associated increase in iron stores, further suggested a link between iron and CVD.⁵¹ Increased iron, in the form of free non-transferrin bound iron (NTBI), hemoglobin or heme, may catalyze the formation of ROS and promote lipid oxidation, thus enhancing intimal LDL-C retention and macrophage progression to foam cells.^52–54

The iron hypothesis was supported by clinical studies that reported an association between atherosclerosis and either increased serum iron⁵⁵ or increased levels of serum ferritin^56–59 (the latter is a marker of tissue iron stores). Studies demonstrated an association between iron depletion, (either by blood donation^60–62 or iron chelation⁶³) and a reduced risk of cardiovascular disease. however, several other studies found no evidence for, or demonstrated results that were even contrary to, the iron hypothesis:^64–67 Mendelian randomization analysis showed that higher levels of serum iron, as indicated by an increased percentage of transferrin saturation or increased serum ferritin, were associated with a reduced risk of coronary artery disease; ⁶⁴ In a cohort study, transferrin saturation was inversely associated with cardiovascular mortality in men and post-menopausal women;⁶⁷ and in a randomized controlled study, reduction of iron stores by phlebotomy in patients with peripheral arterial disease did not decrease all-cause mortality (Table 1).⁶⁵

Table 1.

Studies on the association or effect of iron measures on atherosclerotic cardiovascular disease

Study	Year	Sample Size	Design	Iron measure	Association with clinical disease
Salonen et al.56	1992	1,931 (men)	Prospective cohort study Endpoint: myocardial infarction	Serum ferritin	Men with high serum ferritin levels had a 2.2-fold (95% CI, 1.2–4.0) higher risk of acute myocardial infarction
Kiechl et al.57	1997	826	Prospective cohort study Endpoint: progression of carotid atherosclerosis	Serum ferritin	Serum ferritin is a strong risk factor for carotid atherosclerosis progression (odds ratio, 1.50 per SD increase in ferritin; P=0.0002)
Salonen et al. 60	1998	2,862	Prospective cohort study Endpoint: risk of myocardial infarction	Iron depletion by blood donation	Frequent blood donation was associated with an 88% reduced risk of acute myocardial infarction (95% CI for hazard ratio, 0.02–0.86)
Duffy et al. 63	2001	54	Acute intervention clinical study with CAD patients and healthy controls. Endpoint: vasomotor function in forearm resistance vessels	Iron depletion by iron chelation	Iron chelation with one dose of deferoxamine improved endothelium-dependent vasodilation in patients with coronary artery disease (p<0.01 by 2-way repeated-measures ANOVA)
Meyers et al.61	2002	2,104	Retrospective cohort study Endpoint: cardiovascular events	Iron depletion by blood donation	Frequent blood donation was associated with reduced risk of cardiovascular events (odds ratio, 0.60; CI 0.43–0.83)
Zacharski et al.65	2007	,277	Multicenter, randomized, controlled, single-blinded clinical trial Endpoint: all-cause mortality	Iron depletion by phlebotomy	Phlebotomy in patients with symptomatic peripheral arterial disease did not decrease all-cause mortality (hazard ratio, 0.85; 95% CI, 0.67–1.08)
Menke et al.58	2009	2,662	Cross-sectional analysis Endpoint: association of ferritin and transferrin saturation with PAD	Serum ferritin and transferrin saturation	Borderline association in men of serum ferritin with peripheral arterial disease (PAD) (men: OR, 1.18; 95% CI, 1.00–1.41; women: OR, 1.04; 95% CI, 0.87–1.25) but not transferrin saturation (men: OR, 1.45; 95% CI, 0.83–2.51; women: OR, 1.55; 95% CI, 0.98–2.45)
Sung et al.59	2012	12,033	Cross-sectional analysis Endpoint: association of ferritin and coronary artery calcium score (marker of atherosclerosis)	Serum ferritin	Increased ferritin levels were associated with the presence of coronary artery calcium (OR, 1.66; 95% CI, 1.3–1.98)
Houschyar et al.62	2012	64	Randomized, controlled, single-blind clinical trial Endpoint: clinical markers of metabolic syndrome	Iron depletion by phlebotomy in metabolic syndrome patients	In patients with metabolic syndrome, phlebotomy reduced systolic blood pressure by 16.6 mmHg compared to the control group (95% CI, −20.7 to −12.5 mmHg)
Kim et al.67	2012	5695	Prospective cohort study Endpoint: all cause-, cancer-, and cardiovascular mortality	Transferrin saturation, serum ferritin	Increasing quintiles of transferrin saturation were associated with reduced all-cause, cancer, and CV mortality, but there was no association with serum ferritin
Bagheri et al.55	2013	337	Cross sectional analysis Endpoint: association of iron levels with severity of coronary artery disease (CAD)	Serum iron	Increased serum iron levels in patients with severe atherosclerosis (p=0.01) compared to those with mild and moderate amounts of coronary disease
Gill et al.64	2017	294,223	Mendelian randomization analysis Endpoint: Coronary artery disease risk	Serum iron, transferrin saturation, and ferritin	Protective effect of higher iron status on risk of coronary artery disease: OR 0.94 per SD increase in iron; 95% CI, 0.88–1.00; OR, 0.95 per SD increase in transferrin saturation; 95% CI, 0.910.99; OR, 0.85 per SD increase in log- transformed ferritin; 95% CI, 0.73–0.98

Murine models examining the role of iron in the development of atherosclerosis also produced conflicting results. In mice with iron overload disease, administration of iron chelators protected against the development of atherosclerosis.⁶⁸ Reduced macrophage iron, achieved through either an iron-deficient diet⁶⁹ or treatment with an iron chelator⁷⁰, was associated with decreased atherosclerosis suggesting a protective effect of decreased intracellular iron levels. Dietary iron loading worsened atherosclerosis by inducing inflammation in macrophages.⁷¹ However, another study found that a high-iron diet decreased atherogenesis.⁷² Kautz and colleagues reported that iron loading of macrophages had no effect on atherosclerosis. The authors suggested that excess iron was effectively stored in combination with ferritin within macrophages, thereby inhibiting ROS formation.⁷³

Patients with hereditary hemochromatosis (HH), a genetic disorder associated with hepcidin deficiency and progressive iron overload, do not appear to have an increased risk of atherosclerosis and ischemic heart disease.^74–76 The low hepcidin levels in hemochromatosis patients and the resulting increase in the amount of cell surface ferroportin leads to iron depletion of monocytes. Iron-depleted monocytes isolated from hemochromatosis patients were found to have a decreased ability to upregulate the cytokines monocyte chemoattractant protein (MCP)-1 and interleukin (IL)-6.⁷⁷ The ‘paradox’ that chronic systemic iron overload in hemochromatosis was not accompanied by an increased risk for atherosclerosis, and even perhaps conferred protection from atherosclerotic heart disease, was taken as evidence against the original iron hypothesis. The hypothesis has subsequently been refined to focus on the level of intracellular macrophage iron and its impact on the development of atherosclerosis.⁷⁸

Macrophages in atherosclerosis

Macrophages have a vital role in systemic immunity and tissue remodeling as well as iron homeostasis. Macrophages have pro-and anti-inflammatory effects and these cells both protect from infection and resolve inflammation Multiple macrophage phenotypes have been detected in atherosclerotic lesions, including macrophages with M1, M2, or M(Hb) phenotype (Table 2).⁷⁹

Table 2. Characteristics of macrophages with a M1, M2, and M(Hb) phenotype.

Multiple macrophage phenotypes have been detected in atherosclerotic lesions, including macrophages with a M1, M2, and M(Hb) phenotype. The different types of macrophages have different effects on plaque progression. IFNγ, Interferon gamma; IL, Interleukin; TNFα, Tumor necrosis factor alpha; iNOS, inducible nitric oxide synthase; PDGF, Platelet-derived growth factor; MMP, Matrix metallopeptidase; VEGF-A, vascular endothelial growth factor; HIF, Hypoxia inducible factor; ABC transporter; ATP-binding cassette transporter; oxLDL, Oxidized low-density lipoprotein; FPN, Ferroportin; HO-1, Heme oxygenase 1.

	Classically activated M1 macrophages	Alternatively activated M2 Macrophages	Hemorrhage-associated M(Hb) macrophages
Activation by	INFγ, fatty acids, oxidized lipids79	IL-4, IL-10, IL-1387	Hemoglobin97
Characteristics	Pro-inflammatory80,85,86 Promote inflammation Promote collagen hydrolysis ↑TNFα, IL-1, IL-6, IL-12, iNOS ↑TGF-β, PDGF production ↑secretion of MMP1, MMP3 and MMP9	Anti-inflammatory88,89,93,94 Promote tissue repair and vascular stalility Promote collagen production ↑IL-10, arginase-1, CD206 ↑ornithine, VEGF-A production	Anti- as well as pro-inflammatory properties97,105,106 Promote tissue repair and vascular stability Promote angiogenesis, vessel permeability, & inflammatory cell recruitment ↑IL-10, CD206 ↑HIF-α and VEGF-A
Lipid handling	Increased intracellular lipid levels84 ↑oxLDL scavenger receptors CD36 ↑foam ceil formation	Decreased intracellular lipid levels93,96 ↑ABC transporters ↓CD36 ↑efferocytosis	Decreased intracellular lipid levels97 ↑ABC transporters ↓CD36
Iron phenotype	Iron sequestration phenotype82,83 ↑Ferritin ↓FPN, HO-1, CD163 ↑ intracellular iron	Iron releasing phenotype83,90 ↑Ferritin ↓FPN, HO-1, CD163 ↑intracellular iron	Iron releasing phenotype103,104 ↓Ferritin ↑FPN, HO-1, CD163 ↓intracellular iron
Atherosclerosis	Pro-atherogenic	Atheroprotective	Pro-atherogenic and Atheroprotective

Classically activated (‘M1’) macrophages promote inflammation and are characterized by high expression of tumor necrosis factor alpha (TNF-α), IL-6, IL-1, and inducible nitric oxide synthase (iNOS).⁸⁰ M1 macrophages are the predominant phenotype found in advanced, unstable, atherosclerotic plaques, which have a large lipid-rich necrotic core.⁸¹ Endogenous TLR ligands, such as free fatty acids and oxidized lipids, as well as cytokines such as interferon Y (IFN-γ), contribute to activation of macrophages to a M1-phenotype.⁷⁹ M1 polarized macrophages express high levels of the iron storage protein ferritin and low levels of FPN.^82,83 M1 macrophages express scavenger receptors on their surface, which internalize oxLDL, and M1 macrophages can transform into foam cells by lipid accumulation, which is a critical step in atherogenesis.⁸⁴ Factors such as platelet-derived growth factor (PDGF) or TGF-β produced by M1 macrophages prompt proliferation of vascular smooth muscle cells (VSMC) and their migration from the arterial media into the intima.⁸⁵ Matrix metalloproteases (MMP-1, MMP-3, MMP-9), secreted by M1 macrophages, may destabilize the atherosclerotic plaque by hydrolysis of collagen within the fibrous cap.⁸⁶ In general, classically activated M1 macrophages have a pro-atherogenic phenotype.

Alternatively activated (M2) macrophages are induced by different cytokines, including IL-4 and IL-13, and produce anti-inflammatory cytokines, such as IL-10.⁸⁷ M2 macrophages contribute to the resolution of vascular inflammation and to the stabilization of atheromatous plaques by clearing apoptotic cells and debris, by increasing the production of type I collagen within the plaque, and possibly by inducing the proliferation of VSMC within the fibrous cap (before activation of the MMP cascade).^80,88,89 Compared to M1 macrophages, M2 macrophages have high constitutive expression of arginase I, mannose receptor (CD206), hemoglobin-haptoglobin complex scavenging receptor (CD163) and IL-10. Relatively high levels of FPN and low levels of ferritin in M2 macrophages favor cellular iron efflux resulting in lower intracellular iron levels.^83,90 The activation of heme oxygenase 1 (HO-1), which catalyzes the degradation of he me into ferrous iron and carbon monoxide, further favors iron release.^91,92

In contrast to M1 macrophages, M2 macrophages have lower levels of intracellular lipids and do not transform into foam cells. In a process referred to as efferocytosis M2 marophages engulf apoptotic foam cells within atherosclerotic plaques, thereby promoting tissue repair and inhibiting further lesion progression.^93,94 However, in advanced atherosclerotic plaques, increased accumulation of cellular debris and lipids derived from apoptotic foam cells may exceed the efferocytosis capacity of M2 macrophages and promote necrotic core formation, contributing to further inflammation, necrosis, and thrombosis.^95,96

Hemorrhage-associated macrophages (M(Hb), also referred to as M(hem)), are induced by hemoglobin and are found in areas of neoangiogenesis aM hemorrhage.⁹⁷ In humans, but not mice, intraplaque hemorrhage (IPH), as a consequence of neovascularization due to local hypoxia from intimal thickening, is observe’ in the late stages of atherosclerosis. IPH is associated with increased vessel permeability, red blood cell lysis, and free hemoglobin (Hb) release.^{98,99, 100} The CD163 heme-scavenger receptors on the surface of M(Hb) bind and internalize free hemoglobin-haptoglobin complexes. M(Hb) macrophages produce anti-inflammatory factors such as IL-10 ^101,102 In addition, M(Hb) macrophages are resistant to foam cell formation due to LXRα-induced up-regulation of cholesterol-exporting ATP-binding cassette (ABC) transporters ABCA-1 and ABCG-1.⁹⁷ M(Hb) macrophages were therefore originally classified as atheroprotective and anti-inflammatory. However, LXRα also upregulates FPN mRNA expression and promotes iron export.^{103, 104} The resulting iron decrease in M(Hb) macrophages increases argiogenesis, vessel permeability, and inflammatory cell recruitment, leading to plaque progression.¹⁰⁵ M(Hb) macrophages are therefore classified as both atheroprotective and pro-atherogenic, as they have anti- as well as pro-inflammatory propertes.¹⁰⁶

Differentiated macrophages have a “plastic” cell type in that they maintain the ability to change from one phenotype to another. For example, M1 macrophages can develop features of M2 macrophages, and may then revert to the M1 phenotype, depending on environmental signals.^107,108 Macrophage plasticity is reported to occur during atherosclerotic plaque progression, in both humans and mice.¹⁰⁹ Iron is one important factor that affects macrophage plasticity.^110–112

Effect of intracellular iron on macrophage polarization and function

Intracellular iron levels in macrophages are regulated by the hepcidin-ferroportin axis.³⁷ Iron handling and immune functions of macrophages are linked, and intracellular iron levels directly regulate macrophage polarization.^110–112 Although some studies have yielded conflicting results on the directionality of the effect of iron on the inflammatory response,^113,114 most studies indicate that increased intracellular iron induces a pro-inflammatory M1 phenotype, ^115–119 potentially by increasing ROS production.¹²⁰ Intracellular iron per se has direct effects on macrophage polarization, as increased iron loading of M2 macrophages induces a rapid switch from an M2 to an M1 phenotype.¹¹⁷ Furthermore, intracellular iron levels contribute alterations in cytokine production. Fpn-deficient macrophages, which are characterized by increased intracellular iron levels due to lack of iron export, have an increased expression of inflammatory cytokines. Moreover, intracellular macrophage iron may catalyze the oxidation of LDL thereby promoting the development of foam cells.¹²¹ In contrast, treatment of macrophage with the iron chelator deferoxamine (DFO) has an anti-inflammatory effect and decreases the release of TNFα, IL-6, and the potent monocyte-attracting chemokine MCP-1.¹²²

The pro-inflammatory effects of increased intracellular iron concentration were observed in a mouse model of wound healing. In mice with a macrophage-specific inactivation of FPN, wound closure was delayed, while the inflammatory response was prolonged. In these mice, high intracellular iron content inhibited the switch from pro-inflammatory M1 macrophages to anti-inflammatory M2 phenotype, and prolonged the wound healing phase.¹²³ In humans, an “unrestrained” population of pro-inflammatory M1 macrophages was identified adjacent to chronic venous leg ulcers. Localized iron overloaded induced by release of iron from degenerating, extracellular erythrocytes may contribute to the information of M1 macrophages and impair wound healing.¹¹⁵

Mice that lack the Hfe gene are a well-studied murine model of hemochromatosis. Although the plasma iron concentrations in these mice are high, the intracellular iron levels in macrophages from Hfe^−/− mice are low due to systemic hepcidin deficiency. Associated with the low intracellular iron levels ii Hfe^−/− macrophages, the cells have an impaired response to activation by TLR4 and impaired expression of inflammatory cytokines, including TNFα and IL-6. The reduced inflammatory response is likely the result of low intracellular iron levels within macrophages, as pro-inflammatory responses of macrophages can be increased by administration of exogenous iron.^26,90,124 The results of these studies suggest that macrophage iron content influences cytokine production in macrophages. The effect is, at least in part, attributed to the RNA-binding protein tristetraprolin (TTP). TTP is induced by iron deficiency ^125,126 and destabilizes mRNA of pro-inflammatory cytokines including TNFα and IL-6.¹²⁷

Because high intracellular iron levels induce the pro-inflammatory, pro-atherogenic M1 macrophage phenotype, factors that decrease intracellular iron may promote the switch from the M1 phenotype to the M2 phenotype and thereby protect against the development of atherosclerosis.

The role of hepcidin in atherosclerosis

Hepcidin gene expression is induced by increased levels of iron and by inflammation (Figure 2). High levels of iron increase hepcidin by stimulating the bone morphogenetic protein (BMP) signal transduction pathway. Inhibition of the BMP signaling pathway has been demonstrated to effectively hinder atherogenesis in different mouse models of atherosclerosis.^128–131

Figure 2. Schematic representation of the role of hepcidin in foam cell formation and the pathogenesis of atherosclerosis.

Hepcidin gene expression in the liver is induced by inflammatory cytokines IL-6 and IL1b. Hepcidin binds to ferroportin, the only-known exporter of iron, and induces the degradation of ferroportin. Iron export from macrophages is therefore reduced, and high intracellular iron levels induce a pro-inflammatory, pro-atherogenic M1 macrophage phenotype. As a result of increased intracellular iron, the production of inflammatory cytokines, including IL-6 and TNFα, is increased. High levels of reactive oxygen species (ROS) are generated, which inhibit LXRα-induced expression of the gene encoding the ABCA-1 cholesterol transporter. As a consequence, cholesterol efflux is decreased, and lipids accumulate in the cell, leading to foam cell formation within the vascular plaque. Vascular plaques increase in size, causing lumen narrowing. The dotted grey arrow indicates reduced export. Fe²⁺, ferrous iron; TNFα, Tumor necrosis factor α; IL-6, Interleukin-6; ROS, Reactive oxygen species; LXRα, Liver X receptor α; ABCA-1, ATP-binding cassette transporter 1; oxLDL, Oxidized low-density lipoprotein.

LDN-193189 is a small molecule inhibitor of the BMP type I receptor. Treatment of Apolipoprotein E deficient (Apoe^−/−) mice (a model for the development of atherosclerosis) with LDN-193189 significantly reduced vascular macrophage accumulation and atherosclerotic lesion formation.¹³⁰ Inhibition of BMP signaling reduced hepcidin levels and increased the expression of FPN, which resulted in macrophage iron depletion and reduced intracellular ROS production. These changes were associated with an increase in transcription of genes encoding ABC transporters, which raised cholesterol efflux from macrophages and decreased the formation of foam cells and atherosclerosis.¹³⁰ Finn and colleagues further. supported the interplay between cholesterol and iron metabolism in macrophages. In human monocytes derived from atherosclerotic plaques, hepcidin-induced down-regulation of FPN increased intracellular iron and increased lipid accumulation in macrophages.⁹⁷ By promoting the production of ROS, high intracellular iron levels inhibited LXRα-mediated transcription of genes encoding ABC transporters and hence hindered cholesterol efflux.

In LDL receptor-deficient (Ldlr^−/−) mice, a second model of atherosclerosis, inhibition of BMP signaling, either by treatment with LDN 193189 or with a soluble receptor-antibody fusion protein that sequesters BMP ligands (ALK3-Fc) reduced vascular calcification, vascular inflammation and atheroma formation.¹²⁹ Besides the direct effect of BMP inhibition on the vasculature, Derwall and colleagues found a potent effect of LDN-193189 on lipoprotein biosynthesis. LDN-193189 treatment reduced serum levels of total cholesterol and LDL, suggesting an additional and previously unknown role for BMP signaling in lipoprotein homeostasis.¹²⁹

These studies performed in Apoe^−/− and Ldlr^−/− mice described the essential role of BMP signaling in the development of atherosclerosis with intimal calcification. In addition, we and others have identified he BMP pathway as an important mediator of calcification of the vascular media both in humans and in mouse models with matrix Gla protein (MGP) deficiency or overexpression ^128,132,133 Pharmacological inhibition of BMP signaling with LDN-193189 or ALK3-Fc reduced medial vessel calcification in Mgp^−/− mice and was associated with a decreased VSMC osteogenic phenotype.¹²⁸ Furthermore, BMP inhibition by overexpression of MGP, an antagonist o BMP signaling, reduced BMP activity, atherosclerotic lesion size and vascular calcification, without having an effect on cholesterol levels.¹³¹ Intimal and medial calcification results from different underlying pathogenic mechanisms: Intimal calcification is associated with inflammation, lipid deposition and macrophage accumulation; medial calcification results from metabolite-induced (e.g., phosphate-induced) upregulation of osteogenic gene programs in the vasculature¹³⁴. Despite the differing mechanisms of calcification, both processes are linked to BMP signaling.

BMP signaling also has a critical role in maintaining homeostasis in many tissues, including bone and cartilage formation.¹³⁵ Therefore, chronic inhibition of BMP signal transduction might have a broad range of adverse effects. To specifically determine the role of hepcidin-ferroportin signaling in the development of atherosclerosis, we investigated the effects of hepcidin deficiency on the progression of atherosclerosis in hepcidin- and LDL receptor-deficient (Hamp^−/−/Ldlr^−/−) mice.¹¹⁹ The absence of hepcidin in these mice, and the resulting increase in ferroportin on the surface of macrophages, was associated with decreased atherosclerosis. A regression analysis showed that the benefits of hepcidin deficiency were not caused by differences in serum iron levels, body weight or serum LDL levels. The results strongly suggest that the protective effect of hepcidin deficiency was caused by macrophage iron depletion. Decreased intracellular macrophage iron was associated with decreased levels of M1 phenotypic markers in aortic macrophages as well as reduced oxidized LD uptake, ¹¹⁹

Although these studies show that mice treated with inhibitors of BMP signaling (thereby decreasing hepcidin levels) or mice with genetic deletion of hepcidin have decreased atherosclerosis, the precise role of hepcidin in the pathogenesis of atherosclerosis remains controversial. In a study published by Vinchi et al., ApoE-deficient/FPN ^wT/C326S knock-in mice, which carry a mutation in ferroportin that disrupts hepcidin binding, had significantly increased atherosclerosis compared to age-matched ApoE-deficient mice, despite reduced intracellular iron levels in macrophages.⁶⁸ Resolution of these seemingly contradictory results requires further investigation. One possible explanation for these disparate results^68,119 may be the intrinsic difference between the ApoE- and Ldlr-deficient mouse models of atherosclerosis.¹³⁶ ApoE, unlike LDLR, exerts additional atheroprotective effects (beyond its capacity to lower lipid levels), which include direct anti-inflammatory properties on macrophages that favor M2 polarization.¹³⁷ A second possible explanation or the differences between the two models of atherosclerosis is the complex interplay between serum and intracellular iron levels and the progression of atherosclerosis. The protective benefits of low macrophage iron, along with reduced hepcidin levels, may be partially offset by the detrimental effects of high serum iron. The degree of intracellular macrophage iron content might affect the anti-inflammatory properties of macrophages and explain the controversial results of the two models: while macrophages of homozygous FPN^C326S knock-in mice are iron depleted,¹³⁸ heterozygous FPN^wT/C326S knock-in mic may still have residual intracellular iron in macrophages as indicated by the presence of detectable iron within atherosclerotic plaques in ApoE-deficient/FPN^WT/C326S knock-in mice.⁶⁸ Residual macrophage iron content may counteract the protective effect on atherosclerosis. It is also possible that the effect of hepcidin may be dependent on the stage of atherosclerotic disease. ¹³⁹ In early- to mid-stage disease, low hepcidin levels might have beneficial effects by constraining the effects of pro-inflammatory macrophages.⁹⁷ In late-stage lesions however, depletion of macrophage iron might lead to HIF (hypoxia-inducible factor)-α and VEGGF (vascular endothelial growth factor)-A mediated increased angiogenesis, vessel permeability, and inflammatory cell recruitment resulting in plaque progression.^105,106

Hepcidin and atherosclerosis in clinical studies

Inflammation induces hepcidin expression, and a variety of chronic inflammatory diseases are associated with increased levels of hepcidin, including rheumatoid arthritis¹⁴⁰, heart failure¹⁴¹, chronic kidney disease,¹⁴² and non-alcoholic fatty liver disease (Table 3).^143,144 In a study of patients with rheumatoid arthritis, increased serum hepcidin levels were associated with higher coronary artery calcium scores.¹⁴⁵ In patients with non-alcoholic fatty liver disease, higher serum ferritin and hepcidin levels were associated with increased vascular damage and the presence of carotid plaques.¹⁴⁶ In individuals with metabolic syndrome (diabetes mellitus, hypertension, and hyperlipidemia), who have a high risk of atherosclerosis, serum hepcidin levels and intracellular macrophage iron levels correlated with release of the pro-inflammatory cytokine MCP-1 and vascular damage.^77,146 Studies in patients with chronic kidney disease undergoing hemodialysis showed a relationship between elevated hepcidin levels and arterial stiffness¹⁴⁷ as well as between hepcidin levels and cardiovascular events.¹⁴⁸ In a recent population-based cohort study, increased hepcidin and an increased ratio of heicidin to ferritin (an index of the amount of hepcidin relative to iron stores) were significantly associated with the presence of plaques in the carotid artery of post-menopausal women.¹⁴⁹ Together, these findings strongly suggest an important role for hepcidin in the pathogenesis of atherosclerosis and CVD events.

Table 3.

Epidemiological studies showing a correlation between circulating hepcidin levels and sub-clinical or clinical atherosclerosis

Study	Year	Underlying disease	Sample size	Atherosclerosis Marker or Clinical Event
Abdel-Khalek et al.145	2011	Rheumatoid Arthritis	80	Coronary calcium score (CCS)
Valenti et al.146	2011	Nonalcoholic fatty liver disease (NAFLD)	143	Carotid plaque
Kuragano et al.147	2011	Chronic Kidney Disease patients on maintenance hemodialysis	168	Arterial stiffness
Van der Weerd et al.148	2013	Chronic Kidney Disease patients on chronic hemodialysis	405	Fatal and non-fatal CV events
Galesloot et al.149	2014	General population	766 (346 women)	Presence of plaque

To date, there have been no clinical studies addressing the specific role of hepcidin in atherosclerosis. However, modulating the inflammatory response has become a promising target to reduce mortality and morbidity in CVD (Table 4 ).^22,150,151 Increased expression of hepcidin is induced by inflammatory cytokines, inclining IL-6 and IL-1β.¹⁵² Canakinumab is an inhibitor of IL-1β, a cytokine that is highly expressed in atherosclerotic lesions¹⁵³ and induces the expression of IL-6 and downstream CRP and hepcidin.¹⁵⁴ The CANTOS trail was designed to test whether canakinumab prevents recurrent cardiovascular events in men and women with prior myocardial infarction who have elevated levels of inflammation. Treatment with canakinumab resulted in reduction of IL-6 and CRP levels, with no effects on lipid levels. The incidence rate for myocardial infarction stroke, and cardiovascular death was reduced in the canakinumab-treated group compared to the placebo group.²² Reducing the inflammatory mediators IL-1β and IL-6 seems to be one of the key factors that were beneficial in reducing CVD events. The CANTOS trial is the first clinical trial to report that inhibition of inflammatory mediators improved CVD outcome.

Table 4.

Clinical trials on the effect of targeting inflammation in cardiovascular disease

Study	Year	Sample Size	Type of study	Drug	Result
Klevelnad O et al.158 and Holte E et al.159	2016, 2017	117	Randomized, double-blind placebo-Controlled phase II trial	Tocilizumab	Tocilizumab attenuated inflammation and Troponin-T release (primarily PCI-related, p<0.001), but did not improve coronary flow reserve (a measure of endothelial function) in NSTEMI patients
Ridker P et al.22 CANTOS-trial	2017	10,061	Randomized, double-blind placebo-controlled phase III trial	Canakinumab	Treatment with canakinumab (at either a 150 mg or 300 mg dose) led to a statistically significant 15% or 14% reduction, respectively, in the composite of nonfatal MI, nonfatal stroke, or CV death
Ridker P et al.162 C’RT-trial	2019	4786	Randomized, double-blind, placebo-controlled trial	Methotrexate	Treatment with low-dose methotrexate did not reduce the occurrence of cardiovascular events including nonfatal myocardial infarction, nonfatal stroke or cardiovascular death (HR, 1.01; 95% CI, 0.82 to 1.25).

The significance of the inflammatory cytokine IL-6 in CVD became evident with the finding that the ASP358Ala hypofunctional variant of the IL-6 receptor in humans is associated with reduced risk coronary artery disease.^155–157 Expression of this variant in hepatocytes, monocytes and macrophages reduces IL-6 signaling and leads to an attenuated expression of IL-6-induced proteins, including CRP. In a double-blind, randomized, placebo-controlled phase II trial, patients with non-ST-elevation MI received the IL-6 receptor antagonist tocilizumab to analyze the effect of IL-6 inhibition on the acute inflammatory response (primary endpoint), and to reduce troponin T release.

Treatment with tocilizumab attenuated inflammation and reduced percutaneous coronary intervention (PCI)-related troponin T release compared to the placebo group, which may indicate that IL-6 receptor antagonism is cardioprotective in ischemia-reperfusion injury. However, tocilizumab did not improve coronary flow reserve (a measure of endothelial function).^158,159 The clinical trial was not powered to determine the effect of IL-6 inhibition on clinical outcome.¹⁵⁸ A definitive clinical trial to test the effects of specific IL-6 antagonism on CVD outcomes has not yet been performed. However, previous studies^160,161 found an association between high IL-6 levels and increased rates of coronary heart disease and CV-related mortality, as well as all-cause mortality. Increased hepcidin levels and iron deficiency were associated with higher IL-6 levels¹⁶¹, suggesting a link between hepcidin, inflammation and CV outcomes. Further investigation is needed to determine if the beneficial clinical effects of decreased IL-6 or IL-1β signaling, as observed in humans carrying the Asp358Ala variant or in patients treated with tocilizumab or canakinumab, are in part mediated by decreased hepcidin expression, increased ferroportin levels, and low intracellular macrophage iron.

Clinical implications/Conclusions

The “Iron Hypothesis” has led to research focused on clarifying the role of iron in the development of atherosclerotic heart disease. Early preclinical and clinical studies produced contradictory results and the observation that patients with hemochromatosis do not appear to have an increased risk of atherosclerosis seemed especially difficult to explain. This ‘paradox’ of systemic iron overload not being accompanied by an increased risk to atherosclerosis led to a refinement of the iron hypothesis focusing particularly on intracellular macrophage iron. More recent in vitro and animal studies have identified intracellular iron as an effector of macrophage polarization, cytokine production as well as lipid metabolism. High intracellular iron promotes a pro-inflammatory M1 macrophage phenotype with an increased expression of cytokines and reduced efflux of cholesterol. The effect on cholesterol transport is a key determinant of foam cell formation. Intracellular iron levels in macrophages are regulated by the hormone hepcidin, which induces the degradation of ferroportin and decreases iron export. BMP signaling is essential for the induction of hepcidin expression, and inhibition of BMP signaling in in vivo models has an atheroprotective effect. The strong links between iron homeostasis, BMP signaling, inflammation and atherosclerosis have been established in both mechanistic and human studies. However, while recent clinical trial results have underscored the benefits of targeting inflammation in cardiovascular disease, the precise clinical impact of modifying iron homeostasis warrants further investigation. The BMP-hepcidin-ferroportin axis may represent a novel therapeutic target for the treatment of cardiovascular disease.