Are prolyl-hydroxylase inhibitors potential alternative treatments for anaemia in patients with chronic kidney disease? Free

Abstract

Prolyl-hydroxylase (PHD) inhibitors (PHD-I) are the most appealing drugs undergoing clinical development for the treatment of anaemia in patients with chronic kidney disease. PHD inhibition mimics the exposure of the body to hypoxia and activates the hypoxia-inducible factor system. Among many other pathways, this activation promotes the production of endogenous erythropoietin (EPO) and the absorption and mobilization of iron. PHD-I are given orally and, differing from erythropoiesis-stimulating agents (ESAs), they correct and maintain haemoglobin levels by stimulating endogenous EPO production. Their efficacy and safety are supported by several Phases I and II studies with relatively short follow-up. This class of drugs has the potential to have a better safety profile than ESAs and there may be additional advantages for cardiovascular disease (CVD), osteoporosis and metabolism. However, possible adverse outcomes are feared. These span from the worsening or occurrence of new cancer, to eye complications or pulmonary hypertension. The data from the ongoing Phase III studies are awaited to better clarify the long-term safety and possible advantages of PHD-I.

INTRODUCTION

Erythropoiesis-stimulating agents (ESAs) have been used for nearly three decades to treat anaemia in patients with chronic kidney disease (CKD). In more recent years, safety concerns surrounded their use, especially when given at high doses in hyporesponsive patients or to correct anaemia until normal haemoglobin (Hb) levels. These resulted in changes to the labels of ESAs and treatment guidelines [1, 2], and to worldwide decreases of Hb values and ESA doses.

The relative limitations of current treatment options were incentives for developing new drugs.

Among others, prolyl-hydroxylase (PHD) inhibitors (PHD-I) are now in advanced clinical development. These drugs not only correct anaemia, but may also have possible advantages compared with ESAs. Possible concerns also surround this class of drugs.

PHD-INHIBITORS WHAT ARE THEY?

The hypoxia-inducible factor (HIF) system regulates the cellular transcriptional response to hypoxia. This includes the stimulation of erythropoiesis, the absorption and mobilization of iron, and the coordination of other cell-specific responses (angiogenesis, lipid and glucose metabolism, glycolysis, mitochondrial function, cell growth and survival, vasodilation, cell migration and inflammation).

HIF is a heterodimer made of one oxygen-sensitive α subunit and a β subunit [3]; three different HIFα isoforms exist: HIF1α, HIF2α and HIF3α [4]. HIFs are controlled by a family of PHD: PHD1, PHD2 and PHD3; PHD2 is considered as the main regulator of the pathway [5].

In normal oxygen conditions, HIFα is quickly hydroxylated by PHD and degraded via interaction with the von Hippel–Lindau tumour suppressor protein [6, 7]. Under hypoxia, PHD activity decreases. HIFα can then bind to the HIFβ subunit, translocate into the nucleus and control the physiological response to hypoxia.

All PHD-I are pan-inhibitors with different levels of selectivity, possibly leading to different final effects on the HIF pathway [8].

ERYTHROPOIETIN STIMULATION

Differing from ESA, PHD-I correct anaemia by stimulating the production of endogenous erythropoietin (EPO) in the kidney cortex. HIF activation can also promote the hepatic production of EPO (normally occurring only during the embryonal life) [9]. Accordingly, PHD-I correct anaemia in anephric patients [10].

The normal concentration of endogenous EPO in human plasma amounts to about 15 IU/L (∼5 pmol/L). Following PHD-I therapy, peak EPO levels are within the physiological range or well below those observed with ESA; their magnitude depends on the molecule and dose (Table 1), with high inter-individual variability.

In healthy subjects, the administration of daprodustat determined high EPO peaks exceeding 500 IU/L only in those receiving doses ≥50 mg/day; with lower doses, the maximum observed peak was around 100 IU/L [16].

In haemodialysis (HD) patients, the administration of roxadustat caused peak EPO levels of ∼130 mIU/mL; they were ∼700 mIU/mL with epoetin alfa (median dose, 90 IU/kg/week) [12]. Similarly, HD patients treated with daprodustat at 5 mg daily had much lower median EPO peak levels than those receiving epoetin (24.7 versus 424.9 IU/L, respectively), with high peaks in selected patients (max 1786.5 IU/L) [13]. In non-dialysis (ND) patients, the same dose produced a median EPO peak of only 34.4 IU/L, with maximal values not exceeding 100 IU/L [13].

The administration of vadadustat [17] and roxadustat [11] also caused intermittent and low-level EPO stimulation in ND-CKD patients.

There is no explanation for why some patients obtained much higher EPO peak levels than the average. This is not explained by kidney function (in theory, the healthier the kidney, the higher its capability to produce EPO). Indeed, high EPO peak levels were observed in HD patients, either hyporesponsive to ESA or not [13, 15].

The fact that PHD-I do not expose patients to high peak EPO concentrations represents a theoretical advantage, given the association between high ESA doses and increased all-cause and cardiovascular (CV) mortality [18–20]. Indeed, the EPO receptor is not only expressed in the bone marrow, but also on endothelial cells and in several other tissues. High EPO levels can more easily stimulate these vascular receptors, possibly leading to negative effects such as hypertension or thrombosis after the release of vasoconstrictors (endothelin-1 in the first place) or vasodilators (mainly NO) and altering their balance.

PHD-INHIBITORS AND IRON METABOLISM

HIF is an important regulator of iron metabolism [21]. Because of increased erythropoiesis, hepcidin expression is reduced as an indirect effect [22, 23]. HIF1α can also downregulate hepcidin through Type II transmembrane serine protease matriptase-2 [24]. The lower the hepcidin level, the higher the quantity of iron available for erythropoiesis. In addition, the HIF system has a direct effect on other regulators of iron transport, since it regulates the expression of the molecules involved in intestinal iron uptake (duodenal cytochrome b, apical divalent metal transporter1, ferroportin), and stimulates the synthesis of transferrin and its receptor and that of ferrochelatase (the enzyme that catalyses the insertion of iron into protoporphyrin to form haem) [25].

Iron is also a co-substrate of PHD enzymes, influencing their activity.

Overall, PHD-I improve iron availability. This could be an advantage compared with ESA, especially in patients having functional iron deficiency (little iron available for erythropoiesis despite high ferritin levels).

Phase II clinical studies of PHD-I showed significant and dose-dependent decreases of serum hepcidin levels compared with placebo [13, 26, 27]. However, this could be simply explained by increased erythropoiesis. In few studies comparing PHD-I to ESA, data are less homogeneous (Table 2). Roxadustat was associated with a mean reduction of hepcidin (−60.4 ± 187.8 ng/mL), which slightly increased with epoetin alfa (+35.6 ± 123.4 ng/mL) [12]; a numerical lower percentage of roxadustat-treated patients received intravenous (IV) iron for rescue (2% versus 13%; P = not significant) [12].

Conversely, in a trial comparing daprodustat to epoetin alfa in HD patients, non-significant dose-related trends towards a decrease in hepcidin were noted with increasing daprodustat doses; however, the follow-up was only for 4 weeks [12].

One trial tested the efficacy of roxadustat >12 weeks in three different groups of incident HD patients according to iron treatment (no iron, oral or IV iron) [31]. The rates of increase in mean Hb were similar through 7 weeks for all treatment groups and thereafter lower for the no-iron versus oral or IV iron groups. However, the percentage of Hb responders was similar in the two groups.

Altogether, preliminary data may indicate less iron needs compared with ESA. However, iron therapy is still likely to be required in the majority of patients during PHD-I therapy.

PHD-INHIBITORS AND INFLAMMATION

The HIF pathway plays an important role in adaptation to inflammation [32]. HIF-1 is upregulated by bacterial and viral compounds to prepare cells of the innate and adaptive immune system to migrate to hypoxic and inflamed tissues. HIF activity on immunity and inflammation is probably dependent upon the type of hypoxic stimulus and immune cells [33]. HIF1α is suppressed in patients with sepsis [34], possibly to counter regulating an overwhelming inflammatory response and increase hepcidin levels (and thus limit iron availability for bacteria replication). Conversely, impairment of PHD3 activity aggravates the clinical course of abdominal sepsis in mice [35]. Interestingly, some mediators of inflammation [36] or uraemic toxins [37], which decrease EPO production, mediate their effect through a suppression of the HIF system.

The effect of PHD-I on inflammation is still an open question. Preliminary data suggest that they stimulate erythropoiesis also in inflamed patients. Indeed, dose needs with roxadustat seem to be unrelated to CRP values [12]. Recently, a pilot study tested the efficacy of daprodustat (mean dose of 12.1 mg ± 2.9 day) in 15 hyporesponsive patients [15]. At Week 16, half of the patients had suboptimal efficacy, partially because of insufficient compliance in some patients and by the choice of a too low starting dose. Interestingly, high-sensitive CRP decreased during the study.

More robust findings are needed to confirm or deny the hypothesis that PHD-I are effective in inflamed patients.

PHD-INHIBITORS AND CV RISK

There are discrepancies on the protective role of the HIF system during tissue ischaemia, depending on the pathological context, cell types and differences in activation of HIF isoforms [38, 39]. Controversial data have also been published on cardiac remodelling, atherosclerosis plaque formation and brain ischaemia [40, 41].

Interestingly, PHD-I improve or have a neutral effect on some traditional CV risk factors.

Hypertension is a possible side effects of ESAs [42]. Apparently, PHD-I have a neutral effect on blood pressure in clinical trials. Interestingly, the administration of molidustat led to a sustained reduction in mean systolic blood pressure in a model of subtotal nephrectomy [43].

High serum cholesterol is a well-known CV risk factor. Deficiency of PHD-I attenuates hypercholesterolaemia, atherosclerosis and hyperglycaemia in mice [44]. In these animals, trans-intestinal cholesterol excretion is increased without changes in cholesterol synthesis. Similarly, murine PHD2 hypomorphism lowers serum cholesterol but also high-density lipoproteins (HDLs) [45]. Coherently, PHD-I reduces serum cholesterol in humans. Provenzano et al. [28] showed a significant decrease of serum cholesterol by 26 (±30) mg/dL overall after 8 weeks of therapy with roxadustat; the reduction persisted over time and returned to baseline levels at treatment end. The effect was independent of statin use. As in animal studies, PHD-I decrease not only low-density lipoprotein (LDL) cholesterol, but also HDL cholesterol and the LDL to HDL ratio [14]. Cholesterol decrease is not observed with all the molecules, probably because of different levels of inhibition of the three PHD isoforms [17]. The final effect of these opposing trends on CVD is unknown, also in view of the fact that HDLs are not necessarily protective in CKD patients [46].

Another possible CV protective effect of PHD-I is the lowering of platelet count [47].

Until now, it remains difficult to predict the overall clinical impact of HIF activation on CVD in CKD patients; the results of Phase III trials on hard CV endpoints are awaited.

POSSIBLE NEGATIVE EFFECTS OF PHD-INHIBITORS

HIFs are involved in the regulation of a number of biological processes other than erythropoiesis, some of them still unknown; their activation over prolonged time periods may have adverse effects.

The primary fear is the promotion of tumour growth. The HIF system is upregulated in several cancers and it is one target of anti-cancer therapy [47]. In addition, the HIF system upregulates vascular endothelial growth factor (VEGF), which is a potent angiogenic factor involved in tumour growth and metastasis formation. Data obtained in a VEGF-sensitive model of spontaneous breast cancer are reassuring [48]. No significant increases of VEGF have been shown following short-term treatment with PHD-I, excepting small increases in single patients [14, 17]. Of note, no clear cut-off for normal values exists for VEGF levels, so it is difficult to define the magnitude of VEGF increase to be defined as significant. It should be considered that differing from what happens in the neoplastic tissue, the pharmaceutical activation of the HIF system is intermittent and of much lower magnitude. Careful monitoring on the oncologic potential of this new class of drugs is required.

VEGF has been implicated in the pathogenesis of several eye diseases, including diabetic retinopathy and macular degeneration; these conditions will also require regular assessment.

Chronic hypoxia is a model of pulmonary hypertension; studies in mice demonstrated a role of the HIF system in hypoxic pulmonary hypertension [49]. At present, no data are available about possible development or worsening of pulmonary hypertension following PHD-I.

Experimental studies showed that daprodustat inhibited human ether-a-go-go-related gene current and human sodium channel, possibly affecting cardiac repolarization. A trial on healthy volunteers did not confirm this safety issue; however, 500 mg daprodustat dose was associated with an increase in heart rate with an unknown mechanism [50]. At present, it is difficult to find a cause–effect relationship in sporadic cases of arrhythmia in patients treated with PHD-I.

The HIF system is a modulator of the anaerobic metabolism occurring during hypoxia. Chronic intermittent hypoxia and sustained hypoxia induce body weight reductions and insulin resistance of different magnitudes, suggesting different HIF1α-related activity.

HIF1α is also known to promote obesity and age-dependent expansion of visceral white adipose tissue. This effect seems to be mediated by mitochondrial dysfunction [51]. The clinical implications of the activation of these metabolic pathways are unknown. Of note, a trend for increased glucose levels from baseline was observed following treatment with daprodustat at high doses [26]. However, patients were not stratified by pre-existing diabetes.

Controversial data exist regarding the HIF system and osteoporosis. In osteoclasts, HIF1α is stabilized under oestrogen deficiency and promotes bone loss [52]. Conversely, hypoxia could improve angiogenesis and the differentiation and activity of osteoblasts via up-regulating VEGF [53]. The final net effect in the human bone is still unknown.

Obstructive sleep apnoea is associated with the progression of non-alcoholic fatty liver disease; liver tissue hypoxia possibly activates HIF-1 and mediates the progression of liver fibrosis [54]. The development of the first PHD-I was halted years ago following increases in transaminase levels and a case of fatal hepatitis (then not found drug-related). Data from Phase II clinical studies of the newer molecules did not show any signs of hepatic toxicity.

Following the occurrence of cases of pure red cell aplasia with some ESAs and of severe allergic reactions with peginesatide, the risk of immunogenicity and allergic reactions is to be considered with new ESAs. In the case of PHD-I, the risk appears quite low for both since these drugs are administered orally and are the product of chemical synthesis.

CONCLUSIONS

PHD-I are a novel class of drugs under investigation for the treatment of anaemia in CKD patients. They are orally active and expose patients to lower EPO levels than ESA. Given the broad downstream effects of the HIF system, its stabilization may have beneficial effects other than stimulating erythropoiesis. Even if available data are conflicting, these agents could have a favourable CV profile and possibly improve osteoporosis.

However, they also have open issues regarding safety, considering that they act on not fully elucidated pathways.

Preliminary data do not suggest either significant increases of VEGF following their short-term use or relevant safety concerns. The data of the Phase III studies will give important information on their long-term safety (Table 3).

At present, it is difficult to foresee the future level of penetrance of PHD-I in the ESA market following their approval. This will depend on several factors other than simply anaemia correction, including selling price, oral versus parenteral administration, patient preference and adherence, and possible protective effects from the cardiovascular point of view.

CONFLICT OF INTEREST STATEMENT

F.L. is or was member of an advisory board of Akebia, Amgen, Astellas and Vifor Fresenius medical care Pharma, and was speaker at meetings supported by Amgen, Bayer, GSK, Roche, Takeda and Vifor Fresenius Medical Care Pharma.

L.D.V. was member of an advisory board of Roche, Astellas and DOC, and received speaker fees at meetings supported by Roche and Vifor Fresenius Medical Care Pharma.

We declare that the results presented in this article have not been published previously in whole or part, except in abstract format.

REFERENCES