Tissue iron stores and circulating transferrin-bound iron exert distinct signals that regulate hepcidin expression in hepatocytes. Hepcidin gene transcription is stimulated by the dual effect of liver iron stores and the concentration of plasma holotransferrin (iron-saturated transferrin), conveyed through iron-regulated production of bone morphogenetic proteins (BMP) acting on BMP receptors and the associated mothers against decapentaplegic-related protein (SMAD) pathway (13). Intracellular iron stores interact with hepcidin via the BMP6 pathway activating SMAD and increasing hepcidin levels. Circulating transferrin-bound iron exerts its effects via the haemochromatosis protein (HFE)/transferrin receptor 2 (TFR2) pathway (14, 15). Mutations of these receptors are associated with hereditary haemochromatosis resulting in iron overload via dysregulated hepcidin expression (16, 17).

Hepcidin levels are increased by states of inflammation, and this is thought to have evolved as a host defence mechanism. Interleukin-6, acting through the JAK2/STAT3 pathway and, to a lesser extent, interferon γ, and interleukin-1 are the primary inflammatory inducers of hepcidin expression (18, 19). Recently, a new inflammatory signalling pathway was identified, stimulating hepcidin production via activin B, BMP receptors and SMAD (20).

This is a potent inhibitor of hepcidin production, even in the absence of anaemia, and thus increases iron availability (18). Hypoxia-inducible factors (HIFs) are transcription factors that regulate expression of genes in response to hypoxia including genes required for iron metabolism and erythropoiesis. EPO synthesis is regulated in the liver and kidney via HIF-2α. HIF activity is controlled by prolyl-4-hydroxylase domains (PHD), which act as oxygen sensors. At normal oxygen concentrations, PHD enzymes hydroxylate the HIF-α subunit resulting in its rapid degradation. At lower concentrations of oxygen, HIF–PH activity is reduced, and there is accumulation of HIF-α, leading to increased levels of EPO and its receptor, and decreased hepcidin levels, ultimately increasing iron availability and erythropoiesis (21–24). Similar effects were seen during hypoxia at high altitude. In a study of healthy volunteers who were exposed to high altitude levels (3400–5400 m above sea level), hypoxia induced a marked suppression of hepcidin, which appeared to result from the combined action of hypoxia-induced increased erythropoiesis and iron depletion (25).

Increased erythropoiesis appears to suppress hepcidin levels, allowing for higher iron bioavailability to meet increased demands for red blood cell production in such states. Erythroferrone (ERFE), a relatively new hormone identified in 2014, was found to regulate iron metabolism by decreasing hepcidin levels during periods of stress erythropoiesis (26). This protein is thought to be the long-sought erythroid factor that inhibits hepcidin during increased erythropoietic activity and may contribute to the pathogenesis of iron-loading anaemias. Kautz et al. showed in animal models that bleeding or administration of EPO leads to release of ERFE from erythroblasts, which acts directly on hepatocytes to suppress hepcidin (26). In ineffective erythropoiesis, ERFE secreted by the massively increased numbers of erythroid precursors may overwhelm the iron storage signal and shut off hepcidin production (26, 27). Erythropoietin-stimulating agents (ESAs) are widely used to treat anaemia of CKD. ESAs significantly suppress levels of serum hepcidin and ferritin, resulting in effective erythropoiesis and release of stored iron (28, 29). Honda et al. examined the association between ERFE and biomarkers of iron metabolism in haemodialysis patients and found that levels of ERFE were inversely correlated with levels of hepcidin and ferritin and positively correlated with soluble transferrin receptor. They also showed that the use of ESAs increased the levels of ERFE that regulated hepcidin and led to iron mobilisation from body stores during erythropoiesis (30). Additional studies of this pathway and its potential effects in CKD are warranted.

With regard to iron administration during episodes of acute infection or inflammation which results in elevated serum ferritin levels, opinions suggest that iron therapy should be withheld under such circumstances, citing a concern that iron may further help in the proliferation of microorganisms. Iron loading has been shown to be associated with worse outcomes in infectious diseases such as malaria, tuberculosis and HIV (48–50). However, surprisingly, the recent intravenous iron or placebo for anaemia in the intensive care unit (IRONMAN) clinical trial demonstrated no adverse effects of iron administration in acutely unwell intensive care patients but there were significant improvements in haemoglobin levels (51).

Moreover, there are concerns related to iron therapy-induced iron overload, with a study by Barany et al. suggesting that haemodialysis patients with very high ferritin levels have a mean liver iron concentration similar to that of patients with untreated idiopathic haemochromatosis (52). Iron deposition has been associated with the pathogenesis of several other disorders such as diabetes mellitus, neurodegenerative diseases and atherosclerosis (53–55).

Iron could be a potential link between oxidative stress and cardiovascular disease (56, 57). Iron has been found in advanced human atherosclerotic plaques (58), and free iron may play a role in plaque destabilisation post intra-plaque haemorrhage (59), although, despite pathogenic hypotheses, hard evidence linking iron, oxidative stress and cardiovascular disease is limited (60). Also, the long-term effects of high-dose iron therapy remain unclear.

Large prospective randomised controlled trials in the CKD population are long overdue to assess the efficacy of recurrent iron infusions with regard to long-term safety, mortality and morbidity. In the absence of clear target values for serum ferritin and transferrin, clinicians continue to make a case-by-case decision on the best treatment option for their patients.

The treatment of anaemia of CKD was revolutionised in the 1980s with the development of recombinant ESAs, which has reduced the need for blood transfusions (which in turn reduces the chances of acquiring blood-borne infection and avoids sensitisation in potential renal transplant candidates) (61, 62) and improves exercise tolerance, quality of life symptoms and left ventricular hypertrophy (63, 64).

The target haemoglobin level in the treatment of CKD has been debated for some time, and a number of clinical trials have sought to assess whether full correction of anaemia to normal levels confers benefits; however, results have been disappointing. In 2006, two large randomised control trials in CKD showed that complete correction of anaemia had no effect (65) or conferred a greater risk for attaining the primary composite cardiovascular endpoint (66). The Trial to Reduce Cardiovascular Events with Aranesp Therapy (TREAT)-diabetic patients showed no survival benefit in patients when a haemoglobin target of 13 g/l was set and a secondary analysis suggested higher risk for stroke, death resulting from cancer in patients with a history of malignancies, and venous and arterial thromboembolic events (67, 68).

Another limitation of ESAs is that they also necessitate regular injections, either subcutaneous for patients with CKD and on peritoneal dialysis or via the intravenous route for patients on haemodialysis. Supra-physiologic effects of ESAs, especially at high doses, have off-target effects on other cell types expressing EPO receptors including endothelial cells. This results in adverse effects such as hypertension (69), intimal hyperplasia especially in the setting of inflammation (70) and promotion of tumour growth (71). Other drawbacks include the development of ESA hyporesponsiveness (which can occur in 10%–20% of patients with end-stage kidney disease). Patients needing greater doses are those with concomitant infectious, inflammatory or malignant conditions resulting in relative ESA resistance, which may contribute to increased mortality (72).

Given the high costs and potential disadvantages of ESAs, further elucidation of the molecular mechanisms of anaemia in CKD and the development of better targeted therapies have become a priority.

Given that inflammation contributes to elevated hepcidin levels and may contribute to ESA hyporesponsiveness, it was thought that pentoxifylline might partially correct pro-inflammatory cytokines levels in CKD, resulting in improved iron utilisation and erythropoiesis. The drug has been shown to have anti-inflammatory properties (anti-apoptosis, anti-oxidant, anti-TNF-α and anti-IFN-γ) (73–75). In the handling erythopoietin resistance with oxpentifylline (HERO) trial, which was a double-blind, randomised, placebo-controlled trial, Johnson et al. studied the effects of pentoxifylline on ESA hyporesponsive anaemia in 53 patients with CKD stage 4 or 5 (including dialysis) (76). Although pentoxifylline did not significantly modify ESA hyporesponsiveness as measured by the erythropoiesis resistance index, it did safely increase mean haemoglobin concentration significantly, relative to the control group. A smaller sub-study of the HERO trial examined the effect of pentoxifylline on serum hepcidin level but found no significant difference in patients who received the drug as compared to those on placebo (77). A small uncontrolled pilot study that looked at the effect of this drug on inflammation showed that pentoxifylline reduced levels of interleukin-6 and improved haemoglobin levels in non-inflammatory moderate to severe CKD (78). A systematic review and meta-analysis of 11 studies did not demonstrate conclusive effects of pentoxifylline on haematocrit and ESA dosing (79). Whether this drug may provide any benefits remains to be seen in larger randomised trials.

The role of hypoxia in hepcidin regulation has been briefly explained above. Small-molecule inhibitors of the PHD enzymes mimic the response to a cellular reduction in oxygen levels, increase HIF levels and thereby increase EPO production, thus promoting erythropoiesis. These drugs are in various phases of clinical development for the treatment of renal anaemia.

One such agent roxadustat, an oral HIF–PH inhibitor, was trialled in 60 incident dialysis patients in a phase 2 clinical trial and was shown to increase haemoglobin levels by ≥2.0 g/l within 7 weeks, regardless of baseline iron repletion status, C-reactive protein levels, iron regimen or dialysis modality. It was also found to reduce serum hepcidin levels. Roxadustat by inhibiting HIF–PHs results in increased levels of HIF and stimulates erythropoiesis (80).

More recently, Pergola et al. studied the effect of vadadustat as compared to placebo in 210 non-dialysis-dependent CKD patients (stages 3–5) in a 20-week multi-centre phase 2b study (81). They showed that 55% of the candidates who received vadadustat achieved the primary end point (percentage of participants who during the last 2 weeks of the treatment achieved or maintained a mean haemoglobin level of >11 g/dl or an increase in haemoglobin level of ≥1.2 g/dl over the pre-dose average) as compared to 10% of the placebo-treated candidates, and the drug raised and maintained haemoglobin in a predictable manner with no significant side effects as compared to placebo. They noted significant increases in both reticulocyte and total iron-binding capacity and significant decreases in both serum hepcidin and ferritin levels.

Similar small preclinical and clinical studies have demonstrated some pleiotropic effects of this class of drug (molidustat corrected anaemia in rat studies and also helped normalise blood pressure; daprodustat was shown to improve cholesterol levels) (82, 83). Advantages of this new class of drug include oral administration, low immunogenicity, product stability and possibly lower costs as well as potential cardiovascular benefits (84).

Safety concerns of targeting the HIF pathway include the potential for raised Vascular endothelial growth factor (VEGF) production, which is a HIF-related angiogenic growth factor known to be associated with vasculopathies and progression of tumour growth (85), and the potential for development of pulmonary and systemic hypertension, given the role of HIF in regulation of vascular tone (86, 87). Results of phase 3 trials that continue to monitor for these side effects are awaited.

Several agents that can antagonise hepcidin are under development, including neutralising hepcidin peptide by anti-hepcidin antibodies or by engineered hepcidin binders such as anticalins (engineered human proteins that can bind specific target molecules). Cooke et al. demonstrated an increase in haemoglobin by 1.5 g/dl within 1 week of injection of an anti-hepcidin antibody in humanised murine models where endogenous mouse hepcidin was replaced by human hepcidin and inflammation was induced using heat-killed Brucella abortus (88). The most effective method was the combination of ESA and anti-hepcidin antibody, which increased haemoglobin by >3 g/dl after 1 week compared with inflamed mice injected with a control antibody. The improvement in haemoglobin resulted from increased serum iron levels and better haemoglobinisation of erythroid precursors, without affecting inflammatory responses in these mice. Pieris pharmaceuticals are conducting a phase 1b placebo-controlled study using a hepcidin-antagonist in patients on dialysis, after demonstrating that the drug was shown to reduce hepcidin levels and increase serum iron and transferrin saturation in 48 healthy male subjects in a single ascending dose study with no significant adverse effects (89).

Recent studies have shown that vitamin D concentrations are inversely associated with hepcidin levels and positively associated with haemoglobin and iron concentrations (36, 90, 91). Zughaier et al. demonstrated in vitro that vitamin D is associated with reduced production of pro-hepcidin cytokines such as IL-6 and interleukin-1β. In their in vivo pilot study of 38 patients with early stage (2/3) CKD who received high doses of oral vitamin D3 as compared to placebo, the percent change from baseline to 3 months in serum 25-hydroxy cholecalciferol concentrations was inversely associated with the percent change in serum hepcidin levels (92). These findings are relevant as a large majority of patients with CKD are vitamin D deficient, and correction of vitamin D levels, as an adjunct therapy, is attractive, given the inexpensive cost, easy availability, favourable safety profile and potential to reduce dependence on other more expensive therapies.