Colin J Forman, Department of Renal Transplantation, University of Queensland at Princess Alexandra Hospital, Brisbane, Australia. e-mail: email@example.com
To determine whether the administration of erythropoietin at the time of ischaemic renal injury (IRI) inhibits apoptosis, enhances tubular epithelial regeneration and promotes renal functional recovery, as it does in rodent models, in a higher mammalian model.
MATERIALS AND METHODS
The model of IRI involved unilateral nephrectomy in pigs, followed a week later by renal artery occlusion for 1 h, followed by reperfusion for 5 days. Pigs were randomized to receive erythropoietin 5000 units/kg intravenously at the time of ischaemia, followed by 1000 units/kg subcutaneously daily, or no treatment (six pigs each). Renal function and structure were analysed; blood and urine were collected daily to determine serum creatinine level, blood urea nitrogen, and creatinine clearance. Animals were killed after 5 days to obtain the injured kidneys. The kidneys were examined histologically for evidence of cellular mitosis, apoptosis and necrosis.
Erythropoietin significantly abrogated renal dysfunction after IRI compared with controls at 12 h after injury; the mean (sem) creatinine clearance (as a percentage of baseline) for IRI was 68.2 (6)% vs erythropoietin-IRI 94.9 (8.9)% (P = 0.027), although by 36 h this was no longer significant, with values of 73.8 (12.7)% vs 95.9 (12)%, respectively (P = 0.23). Erythropoietin also significantly reduced the amount of cell death on histological analysis after 5 days of reperfusion, with a median (range) for IRI of 5.5 (1–45) vs erythropoietin-IRI of 1.5 (0–4) (P = 0.043).
This study confirms the potential clinical applications of erythropoietin as a novel therapeutic agent in patients at risk of IRI.
Acute renal failure due to ischaemic renal injury (IRI) is a common disease, with an overall mortality of nearly 50%[1–3]. Although some progress has been made in reducing this proportion since the introduction of dialysis , acute renal failure remains a significant clinical problem, and thus far an effective specific treatment for the disease has remained elusive. Recently attention focused on the potential of erythropoietin, a 34-kDa glycoprotein hormone, traditionally thought only to control the proliferation, differentiation and survival of erythroid progenitor cells through an anti-apoptotic mechanism . In the past decade erythropoietin was shown to have effects beyond the maintenance of red blood cell mass. Erythropoietin receptors have been found in numerous adult tissues, where their stimulation appears to confer protection against apoptotic cell death . Human, porcine and rodent renal tubular cells are amongst those known to express the erythropoietin receptor [7,8], and there is strong in vivo evidence from rodent models, and in vitro evidence in human cell lines, that erythropoietin has a protective effect against ischaemic renal damage [8–11]. However, for reasons of differences in renal embryology, architecture and lymphatic pattern, relative medullary thickness, urinary concentrating ability and tolerance to ischaemia, results obtained from rodent studies cannot always be reliably extrapolated to humans . The pig kidney is thought to be a better match to the human kidney, especially in renal haemodynamic responses to ischaemia . This is thought to be the reason why several agents found to protect the rat kidney from ischaemia fail to protect either pig or human kidneys . Clinical trials involve considerable expenditure of resources, and therefore a case can be made for trialling novel therapeutic agents in pigs that show promise in rodent IRI, before they are transferred to a clinical setting. Thus the aim of the present study was to determine whether the beneficial effect of erythropoietin found in rats can be replicated in pigs.
MATERIALS AND METHODS
All experiments were carried out with the approval of the University of Queensland animal ethics committee; 12 Large White pigs (15–26 kg) had a left nephrectomy under general anaesthetic, and were allowed to recover for 1 week. Four of the kidneys removed were retained to act as uninjured controls. The nephrectomized pigs then had the remaining renal artery occluded for 1 h under general anaesthetic, using a suitable vascular clamp. Half the pigs (six) were randomized to receive 5000 units/kg i.v. human recombinant erythropoietin (Eprex®, Janssen-Cilag Pty. Ltd, Australia), administered 60 min before occlusion of the renal artery, followed by daily injections with 1000 units/kg s.c. erythropoietin. The other six pigs received no treatment. The pigs were killed after 5 days of reperfusion to allow harvesting of the injured kidney. Blood samples were taken before injury, and at 24-h intervals after reperfusion; these were analysed for serum creatinine levels, blood urea nitrogen and haemoglobin concentration. Complete 24-h urinary collections, with determination of urinary creatinine concentration, were used to calculate creatinine excretion.
Harvested kidneys were bisected and fixed in buffered formalin at 4 °C and prepared routinely for histology. Formalin-fixed tissue was embedded in paraffin wax using routine methods and 4 µm sections cut onto Superfrost Plus histology slides and stained with haematoxylin and eosin (H&E). All H&E-stained tissue sections were viewed using light microscopy and counted for apoptosis and necrosis in the tubular epithelium of the outer medulla, in 10 random microscope fields at × 400 (high-power fields). Apoptosis was recognized by its morphological characteristics : (i) shrunken eosinophilic cells with condensed, marginated nuclear chromatin and intact cell membranes; (ii) discrete apoptotic bodies comprising large, dense, pyknotic, nuclear fragments usually surrounded by a narrow eosinophilic cytoplasmic rim; and (iii) clusters of small apoptotic bodies (assessed as one apoptotic occurrence). Features of necrosis were also sought, i.e. swollen cytoplasm with disrupted cell and organelle membranes and lytic nuclear changes . Similar methods for assessing these variables were published previously .
For the functional analysis, arterial blood was taken to measure plasma creatinine, blood urea nitrogen and haemoglobin concentrations. Blood was analysed using a Modular Blood Analyser (Hitachi Australia, Milton, Australia) within 12 h of collection; 24-h urine collections were analysed for creatinine concentration by the Queensland Health Pathology and Scientific Services Unit, Royal Brisbane Hospital, Brisbane, Australia. The mean creatinine clearance for the collection period was calculated as (urinary flow rate × urinary creatinine concentration)/(mean plasma creatinine concentration) . The clearances for each animal at each time point were expressed as a percentage of the value before injury.
Each study group at each time point consisted of six pigs; serum creatinine, blood urea nitrogen, haemoglobin concentration and creatinine clearance are expressed as mean (sem). Groups were compared at each time point using an independent-sample two-tailed t-test. The within-group haemoglobin concentration changes were compared with a paired-sample two-tailed t-test. Histological cell death counts were compared using a two-tailed Mann–Whitney U-test, and in all tests P < 0.05 was considered to indicate significant differences.
Erythropoietin at 5000 units/kg at the time of IRI, followed by daily doses of 1000 units/kg, abrogated the subsequent decrease in creatinine clearance when compared to untreated controls (Fig. 1a). The difference was statistically significant at 12 h (P = 0.027), but no longer significant by 36 h (P = 0.236). Values for creatinine clearance were not determined beyond this time due to technical difficulty with urine collection. The serum creatinine concentration peaked 1 day after IRI in both groups, and had recovered to baseline level by 5 days (Fig. 1b). Although the mean creatinine concentration was greater in the untreated controls than in the erythropoietin group at all sample times, the difference was not statistically significant. The serum urea concentration after IRI followed a similar pattern, peaking at 1 day in both groups, being consistently higher in the untreated controls but not statistically significantly, and recovering to baseline by 5 days (Fig. 1c).
Examples of the histopathology are shown in Fig. 2, showing a normal high-power field taken from an undamaged control kidney (A), apoptosis (arrowed) in the tubular epithelium of a control kidney after IRI (B) and necrosis (arrowed), in a control kidney after IRI (C), and (D) a mitotic figure in the tubular epithelium of a kidney after IRI and treated with erythropoietin.
Cell counts for the undamaged control (four pigs), IRI with no treatment (six), and erythropoietin-IRI (six) kidneys are shown in Fig. 3A,B. For the statistical analysis the sum of 10 random high-power fields per kidney was used. Apoptosis and necrosis were considered together as ‘cell-death’. Inter-group comparisons were made among all three groups. Cell death was significantly less in the erythropoietin-IRI group than in the untreated IRI group (P = 0.042). Neither of the other comparisons, i.e. undamaged control vs IRI, and undamaged control vs erythropoietin-IRI, were statistically significant, possibly because there were too few control kidneys. Mitosis was significantly more prevalent in control than in IRI kidneys (P = 0.02). There was no significant difference between control and erythropoietin-IRI kidneys (P = 0.45) or between IRI and erythropoietin-IRI kidneys (P = 0.12)
The haemoglobin concentration in the erythropoietin group increased over the 5 days of the study, from 9.6 (0.33) g/L before to 10.6 (0.36) g/L at 5 days after IRI (P = 0.129). The haemoglobin concentration in the control group decreased over the 5 days of the study, from 10.2 (0.23) g/L before to 8.8 (0.79) g/L (P = 0.193); the haemoglobin concentration was not significantly different between the groups at 5 days (P = 0.058).
In rodent models of renal injury, erythropoietin and its derivatives confer protection against renal damage whether due to ischaemia/reperfusion injury or due to toxins. For example, Vesey et al. reported decreased tubular apoptosis, increased tubular regeneration and decreased serum creatinine concentration in a rat model of IRI, while Bagnis et al. showed increased GFR, increased tubular regeneration and increased renal blood flow after cisplatinum renal toxicity in a rat model. A search showed a further 14 studies reporting similar results, using a variety of rodent models of IRI and nephrotoxicity . The study by Vesey et al. showed an anti-apoptotic effect in vitro for human renal tubular cells exposed to hypoxia, implying that the results might be generalizable to humans. The present study confirmed these findings in the pig, which is thought to model in vivo human IRI more accurately, for reasons discussed earlier . We showed that erythropoietin not only reduced the functional impairment, as measured by changes in creatinine clearance, but also reduced cell death in the outer renal medulla, as assessed by a pathologist unaware of the treatment groups. The correlation between functional and anatomical benefit increases the robustness of the results. The study was limited by the choice of arterial clamp time, 1 h in this case. In the pig this resulted in only modest rises in urea and creatinine levels, and only modest evidence of anatomical damage. This degree of derangement in a clinical setting would be unlikely to justify the administration of an expensive treatment. Thus the model does not cover the group of patients who stand to benefit most from this potential treatment. Nevertheless useful as a proof of concept, the study lends weight to the case for instituting clinical trials.
Erythropoietin is thought to exert its renoprotective effect independent of its effects on renal haemodynamics and haematocrit, by directly inhibiting cellular apoptosis . In rodent models of IRI the effect is most marked in the outer medulla, which is the region with the lowest blood supply . This pattern was also apparent in pigs, the damage being largely limited to tubules, with the appearance of the glomeruli seemingly unaffected.
The mechanism most commonly advanced for renoprotection is that erythropoietin inhibits apoptosis in tubular cells in a way analogous to its effects on erythroid precursor cells. Briefly, mitochondria appear to be the final arbiters of whether a cell damaged by IRI undergoes apoptosis, necrosis, or survives. Research shows the key event in triggering apoptosis to be an increase in mitochondrial permeability, which is multifactorial, and leads to the leakage of mitochondrial cytochrome C into the cytosol . Cytochrome C activates a common pathway in apoptosis; cytosolic caspases are activated, leading to a proteolytic cascade, resulting in the degradation of DNA and the autodigestion of cellular proteins . Administration of erythropoietin is known to increase the expression of the anti-apoptotic proteins Bcl-XL and XIAP , which inhibit the triggering of this proteolytic cascade. Erythropoietin is though to accomplish this through activation of the Jak-2/STAT-5 and inositol triphosphate kinase pathways .
The present study is the first to show an in vivo functional or structural benefit for erythropoietin in the prophylaxis and treatment of IRI in a higher mammal. It thus opens the way for trials to investigate this effect in a clinical setting. Potential clinical applications include its use in circumstances where blood flow to the kidney is temporarily interrupted, such as partial nephrectomy (especially laparoscopic), renal transplantation, and aortic surgery, as well as in acute tubular necrosis secondary to renal ischaemia.
The study was funded by a research grant from Janssen-Cilag Pty. Ltd.
CONFLICT OF INTEREST
Prof Johnson has received consultancy fees and research grants from Janssen-Cilag (which markets Eprex®, erythropoietin, in Australia). Dr Forman’s salary was funded through a research grant from Janssen-Cilag Pty. Ltd.