The Nephroprotective Properties of Recombinant Human Erythropoietin in Kidney Transplantation: Experimental Facts and Clinical Proofs


  • N. Pallet,

    Corresponding author
    1. INSERM U775, Centre Universitaire des Saints Pères, et Université Paris Descartes, Paris, France
    Search for more papers by this author
  • M. Rabant,

    1. Service d’anatomopathologie, Hôpital Necker, Assistance Publique Hôpitaux de Paris et Université Paris Descartes, Paris, France
    Search for more papers by this author
  • C. Legendre,

    1. Service de Transplantation Rénale et Soins Intensifs, Hôpital Necker, Assistance Publique Hôpitaux de Paris et Université Paris Descartes, Paris, France
    Search for more papers by this author
  • F. Martinez,

    1. Service de Transplantation Rénale et Soins Intensifs, Hôpital Necker, Assistance Publique Hôpitaux de Paris et Université Paris Descartes, Paris, France
    Search for more papers by this author
  • G. Choukroun

    1. Service de Néphrologie, Dialyse, Transplantation et Réanimation, CHU d’Amiens, Université Jules Verne et INSERM ERI-12, Amiens, France
    Search for more papers by this author

Nicolas Pallet,


Adaptive responses to hypoxia, including hypoxia-inducible factor signaling, allow the cell to satisfy its basal metabolic demand and avoid death, but these responses can also be deleterious by promoting inflammation, cell dedifferentiation and fibrogenesis. Therefore, targeting hypoxia constitutes a promising therapeutic avenue. Recombinant human erythropoietin (rhEPO) appeared as a good candidate therapy because its hematopoietic properties could reverse anemia, and its tissue-protective properties could reduce cell death and limit maladaptive cellular responses to hypoxia. Despite experimental evidence on the nephroprotecive properties of rhEPO, recent clinical trials provided evidence that rhEPO was ineffective in preventing delayed graft function after ischemic acute injury but that the normalization of hemoglobin values preserved kidney function deterioration and reduced graft loss. Our aim here is to provide a survey of the rationale for evaluating the administration of rhEPO in the setting of kidney transplantation. We will discuss the intriguing findings that emerged from the clinical trials and the discrepancies between promising experimental results and negative clinical studies, as well as the differences in terms of the benefits and safety profiles of the normalization of hemoglobin values in chronic kidney disease patients and kidney transplant patients.


and activator protein 1


heat shock protein 70


B cell lymphoma


carbamylated EPO


chronic kidney disease


connective tissue growth factor


donation after cardiac death


delayed graft function


eNOsynthase (eNOS)


EPO receptor


erythropoiesis stimulating agent


end-stage renal disease




hypoxia inducible factor


interstitial fibrosis-tubular atrophy




Janus kinase


mitogen-activated protein kinase


nuclear factor-κB


platelet-derived growth factor


phosphatidylinositol-3 kinase


posttransplant anemia


renin–angiotensin–aldosterone system


recombinant human erythropoietin


Src homology-2


X-linked inhibitor of apoptosis protein


Tissue hypoxia is a common denominator in the progression of several, if not all, chronic kidney diseases (CKD) that result in adaptive tissue responses, such as angiogenesis, metabolic reprogramming and survival but also in maladaptive responses, including inflammation, the epithelial-to-mesenchymal transition and fibrosis (1). Hypoxia of the transplanted kidney is a consequence of systemic (anemia) and local (tissue ischemia) factors. These factors interact in a vicious cycle, with local hypoxia fueling kidney function deterioration, which results from the reduced secretion of erythropoietin (EPO), leading to anemia that, in turn, enhances tissue hypoxia. Therefore, the therapeutic means that aim at reducing kidney hypoxia could potentially limit maladaptive tissue remodeling and preserve graft function and survival. The hematopoietic and nonhematopoietic properties of recombinant human erythropoietin (rhEPO) confer a prominent place in the nephroprotective armamentarium in kidney transplantation to this class of drugs, as confirmed by strong experimental evidence. However, the hope for a clinical translation of the protection afforded by rhEPO use has been tempered by the puzzling results of recent clinical trials. This review aims at outlining our current knowledge on the causes and consequences of kidney allograft hypoxia, which led to the development of trials testing the role of rhEPO in the preservation of kidney allograft function. In light of the available experimental and clinical data, we will discuss important questions including the discrepancies between experimental facts and clinical evidence. Furthermore, we will emphasize the fact that the efficacy and safety profiles of trials using high-dose rhEPO reinforce the idea that kidney transplant patients constitute a very particular group of patients with chronic kidney disease.

Origins and Consequences of Hypoxic Stress

The origins of hypoxic stress

Hypoxia occurs when oxygen supply to the tissue cannot meet cellular metabolic demand. Oxygen deprivation is the consequence of anemia and kidney allograft ischemia, this last entity being defined as the shortage of blood supply, either acute or chronic, to the organ. Anemia and kidney ischemia are frequently encountered in transplant medicine (2,3) and interact to generate a vicious circle that maintains a permanent level of hypoxic tissue injury, leading to maladaptive tissue remodeling and structural deterioration, the loss of kidney function and EPO secretion (1). In turn, the decline of EPO secretion contributes, among many other factors, to posttransplantation anemia (PTA), with its contingent of deleterious systemic effects, including kidney hypoxia and interstitial fibrosis and tubular atrophy (IF-TA) (4,5). After the first year posttransplantation, the prevalence of PTA varies from 25% to 40% (6,7). PTA contributes to the progression of IF-TA (4,5) with a positive correlation between hemoglobin (Hb) levels and renal function in this population (5–7).

Aside from the contribution of anemia to hypoxic stress, ischemia, either acute or chronic, is another hypoxic challenge to the success of the kidney allograft. Acute ischemia occurs during the perioperative period and cause delayed graft function (DGF), which is the clinical consequence of ischemic acute tubular necrosis (2). The incidence of DGF varies depending on numerous existing definitions, but a recent comprehensive survey showed that DGF occurred in 21.3% of US transplant patients in 2008 ( DGF is a strong risk factor for IF-TA and acute rejection, and negatively influence graft survival (2).

Although not specific to the kidney allograft, the chronically diseased kidney is ischemic, and hypoxia drives maladaptive tissue healing and inflammation, which in turn accelerates fibrogenesis and cell dedifferentiation (1). Chronic structural changes embedded with chronic allograft nephropathy contribute to tissue ischemia: chronic vasoconstriction, arteriosclerosis, capillary rarefaction and interstitial fibrosis (which reduce oxygen diffusion) are the main factors that contribute to the maintenance of a hypoxic state. Chronic hypoxia in the human-transplanted kidney is not well appreciated due to the lack of direct read-outs and sensitive methods of detection, and its role in the progression of chronic allograft nephropathy is mostly based on experimental evidence.

The (mal)adaptive responses to hypoxia

Cells subjected to hypoxic stress engage in adaptive responses, which aim to maintain basal metabolic functions (e.g. anaerobic glycolysis), to increase oxygen and nutrient supplies (e.g. neoangiogenesis, EPO synthesis) and to activate antiapoptotic molecular programs. Notably, HIFs (hypoxia inducible factors) are central regulators of cellular adaptation to hypoxia (8). Under hypoxic conditions, HIFα induce the transcription of hundreds of genes that encode proteins responsible for regulating glucose uptake and metabolism; angiogenesis, inflammatory cell chemotaxis; cell proliferation or survival, extracellular matrix formation and turnover (8).

If adaptation fails, the cell will die, by apoptosis or necrosis. The permeabilization of the plasma membrane that occurs during necrosis and late apoptosis allows for the release of intracellular compounds that will prime antigen-presenting cells through the activation of pattern-recognition receptors. Moreover, HIFα activates nuclear factor-κB (NF-κB) signaling and enhances the immune functions of myeloid cells (9). Aside from any inflammation, adaptive responses to hypoxia fuel tissue remodeling, leading to interstitial fibrosis and tubular atrophy. HIFα regulates the expression of profibrotic cytokines including TGFα, platelet-derived growth factor (PDGF), connective tissue growth factor (CTGF) and VEGF (8), as well as activation of the epithelial-to-mesenchymal transition. Therefore, nephroprotective strategies must be designed to increasing oxygen supply to the tissue, improving cell survival and dampening detrimental cellular responses to hypoxia.

Strategies for Avoiding Structural Deterioration of the Kidney Transplant

The concept of nephroprotection

Strategies that aim at limiting kidney allograft structural deterioration can be classified into two categories: treatment of the cause and nonspecific treatments. Nephroprotective strategies target nonspecific pathological processes, independently of their causes, which are often latent and insidious. These pathological processes result from maladaptive responses, triggered by various causes and fueled by their incomplete resolution and maintained in a self-autonomous manner. The most advanced success of nephroprotective strategies concerns blockade of the renin–angiotensin–aldosterone system (RAAS), which can slow the evolution of numerous chronic kidney diseases (10,11). Intriguingly, there are few, if any, examples of established nephroprotective strategies, except RAAS blockade (12,13), in kidney transplantation.

Therapeutic approaches that aim at limiting the consequences of ischemic insults during the perioperative period are one field of intensive research that has generated a great amount of experimental and clinical data. These include organ preservation techniques (preservation solutions, organ storage techniques) (14), delayed introduction of calcineurin inhibitors (15), ischemic preconditioning (16), vasodilatory agents (17,18), induction with antilymphocyte antibodies (19) or the modulation of HIF-1α stabilization (20). Although some of these approaches have led to important progress in the management of organ preservation against hypoxic insults, one of their limitations is that they focus on the immediate perioperative period and only target kidney ischemia-reperfusion (IR) injury. Consequently, the hypoxic stresses that may occur afterward will not be supported, and thus these strategies may fail in preventing chronic structural deterioration.

Rationale for the use of rhEPO in kidney transplant nephroprotection

The hematopoietic and nonhematopoietic properties of rhEPO make this molecule an exciting potential nephroprotective agent, which targets multiple processes that mediate maladaptive hypoxic tissue remodeling. rhEPO have become standard for the treatment of renal and in some instances nonrenal anemia including anemia associated with chronic diseases or myelodysplatic syndrome. EPO binds to and activates its homodimerized receptor (EPOR) on erythroid precursors. The EPO/EPOR interaction triggers conformational changes in the extracellular domain of the receptor and consequently the activation of the EPOR-associated Janus Kinase (JAK)-2 (21). JAK2 activation results in the phosphorylation of eight tyrosine residues on the cytoplasmic region of EPOR (22). These phosphotyrosine residues recruit a variety of Src homology-2 (SH2) domain containing proteins that initiate various signaling pathways. One of the main activated signaling pathway is the phosphatidylinositol-3 kinase (PI3K)/AKT pathway, which is implicated in cell proliferation and survival (23). This signaling subsequently leads to the activation of a cell survival factor, Bcl-XL and the down regulation of the cell cycle inhibitor protein p27Kip1, resulting in cell proliferation and protection against programmed cell death. Another EPO-mediated signaling pathway is the RAS/RAF/mitogen-activated protein kinase (MAPK) pathway (24), which inhibits caspase-3 activity. Finally, JAK2/STAT5 is activated by EPO, which leads to the upregulation of the antiapoptotic Bcl-XL gene (25).

Aside from the hematopoietic properties of rhEPO, EPOR-mediated signaling activates antiapoptotic and proliferative pathways which confers clinically relevant tissue-protective effects to rhEPO in cases of nonhematological experimental disorders such as stroke or retinal degeneration (26,27). Numerous encouraging clinical studies have tested the neuro-protective effects afforded by high-dose rhEPO (28,29). The tissue protective activity of EPO may be due to the presence of the heterodimeric receptor consisting of EPOR and the β common receptor (βcR, CD131) (18). This alternative receptor could be responsible for the majority of the EPO protective activity (30), which constitutes an interesting therapeutic approach (see below). However, recent data questions the exclusivity of this receptor (31). The affinity of EPO for the EPOR/βcR receptor is low, so the tissue-protective properties of EPO are reached with higher dosage than needed for its circulating hormonal effects. Early experiments on cardiac ischemia showed that JAK2, STAT5, PI3K, RAF, MEK1/2, p42/44-MAPK and p38-MAPK were involved in tissue protection (32). Most of the tissue protective properties of EPO are mediated by Akt, which increases the expression of antiapoptotic Bcl2 and reduces caspase 3, resulting in an antiapoptotic effect (32). EPO reduces inflammation by signaling pathways that involve Akt-mediated induction of eNOsynthase (eNOS) and activator protein 1 (AP1), as well as the inhibition of the expression of IL6 and tumor necrosis factor α (TNFα) by JAK2-STAT5 (26,32–34). The immunoregulatory properties of rhEPO seem to be complex because subpopulations of dendritic cells express the EPOR and are activated by EPO with increased CD86, CD40 and IL-12 expression levels (35). Finally, EPO induces angiogenesis, through ill-defined mechanisms, that involve the production of VEGF and angiopoietin-1 (36,37).

Therefore, the rationale for the use of rhEPO in the prevention of the structural deterioration of the allograft mediated by hypoxia involves the correction of anemia, the tissue protective properties of EPO-mediated signaling, and the reduction of biological, deleterious adaptive responses activated by hypoxia.

The Nephroprotective Properties of rhEPO

Experimental facts

EPOR is expressed by numerous kidney cells, including podocytes, endothelial and epithelial cells, and constitutes a potential target of rhEPO. As a nephroprotective strategy in animal models, rhEPO protects against acute kidney injuries, including cisplatin and cyclosporine nephrotoxicity, IR injury, contrast media-induced renal failure and ureteral ligature (38–44). The mechanisms by which rhEPO protects cells after acute kidney injuries is due to the tissue-protective of EPO and not to the correction of anemia. rhEPO protects against kidney injury by inhibiting apoptosis, which as a consequence, reduces the inflammatory consequences of programmed cell death in kidney tissue. These effects are associated with a reduction of polymorphonuclear cells infiltration, a reduction of lipid peroxydation, the activation Akt and JAK2, which induce X-linked inhibitor of apoptosis protein (XIAP) and BCL2, and reduce the activity of caspases, the activation of c-Jun N-terminal kinases (JNK) signaling, and the induction of expression heat shock protein 70 (HSP70) (38–44).

The administration of rhEPO in models of chronic nephropathies clarified important issues on the nephroprotective properties of rhEPO. First, rhEPO is effective in preventing maladaptive remodeling and fibrosis in chronically injured kidneys, as in rats that have undergone five-sixths nephrectomy and diabetic nephropathy (45,46). However, it is not clear if the antifibrotic effects of rhEPO are related to the inhibition of cell death, a reduction of VEGF secretion, or more specific antifibrotic effects on TGFβ signaling (45). Next, the tissue-protective effects of rhEPO are also effective in animal models of nonischemic chronic kidney injury, including diabetic nephropathy (45). Finally, the finding that the administration of hematologically noneffective doses of rhEPO provides long-term tissue protection suggests that the chronic stimulation of EPOR in tissues provides protective effects (45,46). The nonhematopoietic, tissue-protective properties of rhEPO have been demonstrated to prevent chronic allograft injury in fully mismatched rat allograft kidneys. This effect was associated with the upregulation of tubular phospho-AKT. The correction of anemia per se had no impact on chronic graft injury or dysfunction (47). These experimental studies demonstrate that rhEPO prevents chronic kidney structural deterioration, including kidney allograft, and that this effect is mediated by the tissue protective rather that hematopoietic properties of EPO.

Clinical evidence

The clinical therapeutic application of rhEPO in the prevention of kidney allograft structural deterioration is supported by numerous experimental data, which, in turn, define two main avenues that have been tested in clinical trials: (1) acute kidney injury and the prevention of delayed graft function, and (2) chronic hypoxic stress and slowed progression of chronic allograft nephropathy.

Acute ischemic kidney injury and delayed graft function

Clinical trials that tested whether the administration of high doses of rhEPO during the perioperative period could provide beneficial short- or long-term effects on allograft function yielded negative results. An open-label multicenter randomized study evaluated the effect of high doses of epoetin β during the first 2 weeks of renal transplantation on renal function and anemia correction in 104 patients at risk for delayed graft function. Patients randomized in the treatment group received 4 injections of rhEPO (30 000 UI each): one before surgery, one 12 h after surgery, and 7 and 14 days posttransplant (48). At 1 and 3 months posttransplant, the estimated glomerular filtration rates were similar in both groups, as was the incidence of DGF. Another double blind, placebo-controlled trial evaluated short-term (6 weeks) and long-term (1 year) effects of high-dose epoetin α in 90 deceased donors (49). The first dose (40 000 IU) was administered intraarterially just before the transplantation; the second and third doses were administered intravenously on days 3 and 7 posttransplantation. This study did not show a significant effect of rhEPO on kidney transplant function or on the frequency of DGF, and no significant differences between treatment groups with respect to acute and chronic histological changes were recorded. A single-center prospective, randomized, double blind, placebo-controlled study evaluated the efficacy and safety of high-dose EPO in 92 recipients form cardiac death kidney donors (50). EPO was administered as an intravenous bolus of 33 000 IU, 3–4 h before the transplantation, and 24 and 48 h after reperfusion. EPO did not influence the incidence or duration of DGF. A significant improvement in the EPO group 1 year after transplantation (68 ± 23 vs. 57 ± 25 mL/min) (p < 0.05) was observed, whereas 1-year patient and graft survivals remained unchanged. These results are corroborated by the negative findings of another double blind, placebo-controlled trial investigating the effects of a treatment with rhEPO on the development of acute kidney injury in 162 nontransplanted patients in general intensive-care units (51). Of note, rhEPO was well tolerated, and no difference in the incidence of serious adverse events including cardiovascular events, and cancer progression, with a small increase in vascular thrombosis, was observed in these studies. Together, these adequately powered studies demonstrate that the administration of high-dose rhEPO during the perioperative period does not provide any structural or functional protection against acute kidney injury.

Chronic allograft nephropathy

Retrospective and prospective studies have shown that anemia is robustly associated with an increased greater risk of allograft loss over the long term (4,5,52), supporting the potential beneficial effects of the administration of rhEPO and anemia correction in IF-TA progression and kidney allograft survival. We undertook an open-label, multicenter, randomized study designed to investigate the effect of epoetin-β in the normalization of Hb values (target Hb: 13.0–15.0 g/dL, group A) compared with the partial correction of anemia (target Hb: 10.5–11.5 g/dL, group B) with respect to the deterioration of renal function in 125 transplant patients (53). After a 2-year follow-up period, the mean estimated creatinine clearance had decreased less in the group A than in group B. At year 2, cumulative death-censored graft survival was 94.6% in group A and 80.0% in group B (p < 0.01). The number of cardiovascular events was low and similar between groups. Therefore, this study demonstrates that the correction of anemia with rhEPO slows the progression of chronic allograft nephropathy and prolongs graft survival. The reasons why rhEPO could protect against chronic allograft nephropathy remain speculative. The respective contributions of the correction of anemia, and the direct effects of EPO on apoptosis, angiogenesis, inflammation and fibrogenesis, in the progression of chronic allograft nephropathy remain to be explored.

These results are in apparent discrepancy with an observational study recently published which measured survival time and Hb concentration after treatment with EPO in 1794 renal transplant recipients recorded in the Austrian Dialysis and Transplant Registry who received a transplant between 1992 and 2004 (54). This study showed that increasing Hb concentrations to above 14 g/dL with EPO in renal transplant recipients was significantly associated with an increase in mortality. Although these two studies cannot be compared because of their different designs (prospective and interventional vs. retrospective and observational), it should be noticed that the mean Hb level in the rhEPO-treated group in the CAPRIT study was 13.1 ± 1.7 g/dL in group A (vs. 11.4 ± 1.0 g/dL), which is less that the cutoff associated with higher mortality in the observational study. A recently published study raised concern regarding the safety profile of the use of rhEPO in kidney transplant patients, which contradicts the safety results of the CAPRIT study. EPO treatment significantly increased the risk of thrombotic events in a population of 92 donation after cardiac death (DCD) kidney transplant recipients (50). The reasons for these discrepancies are not known, but methodological issues, including differences of population characteristics or differences of power to detect adverse events can be considered. Therefore, the safety profiles of high Hb levels under rhEPO treatment in kidney transplant patients remains unclear.

Lessons from Recent Clinical Trials

More than a decade of experimental and clinical research allowed rhEPO to become the first nephroprotective strategy ever validated in kidney transplantation. However, the clinical trials detailed above yielded unexpected results and provide important lessons and questions that need to be addressed.

Whereas numerous models of acute kidney injury demonstrated a protective effect of high doses of rhEPO, clinical trials failed to confirm these findings in the clinical setting. This discrepancy illustrates the fact that the relevance of in vivo findings should always be questioned. From a conceptual point of view, the IR injury model usually used in animals cannot be directly translated in the context of the IR injury that occurs during kidney transplantation. Indeed, in animals, a short (usually 30-min) period of warm ischemia precedes reperfusion, whereas a prolonged (usually >12 h) cold ischemia period precedes reperfusion in the clinical setting. One cannot exclude the possibility that kidney injury is less severe in such animal models of IR injury. Moreover, animal models of IR are ‘pure’ models of kidney injury, without ageing lesions, preexisting arteriosclerosis or IF-TA, allograft rejection or calcineurin inhibitor nephrotoxicity, which are often encountered in renal transplantation with suboptimal allografts. Rodent kidneys are known to be more resistant to injury with a greater regenerating capacity than human kidneys (55). Methodological issues could also be considered: the interindividual variability in preclinical studies is very small because they are almost always monocentric, and the characteristics of the animals are similar, except the treatment of interest. This could explain the fact that in preclinical studies, the number of animal is often less than 10, whereas hundreds of patients are required in clinical trials to reach adequate power.

Pharmacological reasons may explain the lack of efficacy of rhEPO in human kidney injury. The protective activity of EPO may be due to the presence of the EPOR-βcR heterodimeric receptor with a low affinity for EPO (18,56). Therefore, the tissue-protective properties of rhEPO are reached with higher dosages of rhEPO that may promote an increase in cardiovascular and thromboembolic events (57,58). To circumvent the side effects while preventing the cytoprotective activities of EPO, different nonerythropoietic EPO derivatives have been developed either by chemically modifying or mutating EPO. Carbamylated EPO (CEPO) lacks erythropoietic activity but maintains the tissue-protective effect of EPO (59). It has been demonstrated that CEPO protects kidneys from ischemic injury (60). Based on the tertiary structure of EPO, small peptides have been developed and are more stable. Moreover, they display less immunogenicity and have tissue-protective effects (61). AsialoEPO lacks the sialic acid from rhREPO, and its binding affinity to EPOR is the same that rhREPO. However, its shorter half-life reduces its erythropoietic effects, so the Hb concentration does not increase. AsialoEPO is neuroprotective (62).

Another intriguing finding are the differences of high Hb targets under erythropoiesis stimulating agents (ESA) therapy in patients with CKD patients and kidney transplant patients with respect to the benefits and safety profiles. Anemia correction with a target Hb level of 13 g/dL using ESA does not prevent progression to ESRD (57,63), whereas it has been proven to be beneficial for transplant patients (53). The study populations and design should be considered to explain the observed differences in risk between the CAPRIT study and the CHOIR and TREAT studies. Patients in the CAPRIT study had a far less severe cardiovascular profile with an annual cardiovascular event rate that was extremely low compared to those of patients in the CHOIR and TREAT studies (57,58). Taking into account the size of our population and the lack of power to detect low-rate events, this safety profile of the CAPRIT study should be interpreted with caution. However, the rhEPO doses were more than twice those used in the CAPRIT study to achieve similar Hb levels, indicating a more severe cardiovascular risk profile, a more pronounced inflammatory state and a higher state of resistance to ESA therapy in the two aforementioned studies (57,58,64,65).

Concluding Remarks

The clinical development of nephroprotective strategies involving the administration of rhEPO in preventing kidney allograft deterioration has progressed very recently, and safety and benefit profiles are being delineated. Clearly, high-dose rhEPO failed to show any benefit in the prevention of DGF. However, some questions remains unanswered in the prevention of chronic allograft nephropathy with EPO derivates, and it is of importance to separate the beneficial effects of the correction of anemia to the tissue-protective properties of EPO. Given the potential side effects of the use of high-doses EPO in kidney transplant (50,54) and the encouraging results of the CAPRIT study on the prevention of allograft function loss (48), the clinical efficacy and safety of nonhematopoietic EPO derivates including CEPO, EPO-derivative peptides, asialoEPO in the prevention of chronic allograft damage should be evaluated in the future. The safety profile of high-dose rhEPO in kidney transplant patients remains to be fully characterized.


The authors of this manuscript have no conflict of interest to disclose as described by the American Journal of Transplantation.