Normothermic machine perfusion (NMP) technologies are being trialled in clinical practice to improve early graft function and assess the quality of kidneys for transplantation (1–3). An oxygenated red cell-based solution is circulated through the kidney under near physiological conditions. Cellular function is restored to replenish ATP and minimize the effects of cold ischemia.

NMP can be carried out using a number of different strategies. Experimentally, prolonged durations of NMP have shown benefit in improving early graft function (4). However, in clinical practice shorter durations performed after hypothermic preservation at the recipient center are logistically more practical. Preliminary studies have demonstrated the safety and feasibility of this technique in clinical practice (5) and the results of a RCT in DCD kidneys (1) are expected later this year.

Compatible packed red blood cells (pRBCs) from a local blood bank are used in most NMP circuits as an oxygen carrier. RBCs are also important regulators of vascular tone and nitric oxide bioavailability (6). Although pRBCs can be stored for up to 42 days, over time their condition deteriorates (7). During storage the destruction of the RBC membrane causes the release of haemoglobin and free heme (8, 9). This is further exacerbated during NMP when RBCs come into contact with artificial surfaces and by the perfusion pump (10, 11). Excess free heme is highly toxic and has the ability to induce oxidative stress, inflammation, cellular injury and apoptosis (10, 12). Free heme also binds nitric oxide derived from the endothelium reducing nitric oxide availability to cause vasoconstriction, thrombin formation, fibrin deposition, platelet activation and aggregation that leads to organ dysfunction (10, 12).

The aim of this study was to measure the levels of free heme during NMP and assess the effect on kidney function and to determine any association with inflammatory and stress related gene expression.

Kidneys undergoing 1h NMP from a RCT (ISRCTN15821205) were used to determine the effect of free heme on perfusion parameters (n = 42). The study protocol and trial documents were approved by the NHS Health Research Authority East of England, Cambridge Central Research Committee (15/EE/0356).

Human kidneys rejected for transplantation and offered for research were used to investigate the transcriptional effects of free heme on gene expression and cellular damage (n = 15). Consent for research was obtained by specialist nurses in organ donation (SNODs). The study was approved by the National Research Ethics committee and Research and Development office at the University of Cambridge (NRES: 15/NE/0408).

The NMP system uses adapted clinical grade pediatric cardiopulmonary bypass technology designed to minimize hemolysis. However, factors such as the heat generated by the centrifugal pump, shear stress, negative pressures, turbulent flow and gravity assisted venous drainage contribute to hemolysis and release of free heme. In cardiac bypass procedures excess levels of heme are associated with organ damage and increased mortality (11, 14). The use of banked RBCs for NMP is problematic in that during storage the deformability and surface charge of RBCs decreases and fragility and aggregability increases (15, 16). This makes them more susceptible to damage during NMP. During storage there is also an accumulation of lipids and RBC derived microparticles or macrovesicles that act as damage-associated molecular patterns (DAMPs), which can stimulate the immune system and production of reactive oxygen species (17, 18). In the clinical series, we found that older units of pRBCs had significantly higher levels of free heme at baseline and levels increased further during NMP. Furthermore, older units of pRBCs also had higher levels of potassium and lactate before and after NMP. This is presumably due to leakage from the fragile RBCs during storage. Perfusate levels of lactate increased significantly during perfusion but it is difficult to determine whether this was due to further damage of the red cells or increased lactate production due to inadequate tissue oxygenation (19). Interestingly, in the discard kidney series the age of pRBCs was not associated with increased levels of free heme but levels increased significantly during NMP. This may be due to the small sample size and variation within this group. Nonetheless, in both the transplanted and discarded kidneys no association was found between the levels of free heme and renal perfusion parameters including oxygen consumption which suggests no detrimental effect and adequate tissue oxygenation. There was also no evidence of significant oxidative damage at the level of protein expression or in the number of apoptotic cells within this short NMP timeframe.

To determine the molecular effects of heme, changes in transcriptional gene expression in a number of the discarded kidneys were assessed. There was significant upregulation of cellular stress related genes that are consistent with ischemic injury. Genes upregulated after NMP included heat shock proteins, oxygen free radicals (KLF4), hypoxia and inflammatory genes (IL-6, SOCS3, ATF3) and chemokines (CXCL8, CXCL2, CC3/L1). FOS and JUN gene transcription were amongst the top five significantly upregulated differentially expressed genes. They regulate activator protein 1, a prominent transcription factor involved with regulating cellular fate involving the mitogen-activated protein kinases (MAPKs) pathways (ERK, p38 and JNK). Together with the NLR pathway, the MAPK pathway is involved in the downstream signalling of the innate immune response and pro-inflammatory cytokine production with activation of NF-κB/NLRP3 inflammasome (20).

To quantify the findings from the transcriptional data, a number of genes were investigated using qRT-PCR. In agreement with the NanoString transcriptomic data, there was a significant upregulation of IL-6, FOS and JUN. We also measured HMGB1, a nuclear protein that in situations of stress and ischemia, translocates from the nucleus to the cytoplasm and is excreted into the extracellular space to act as a DAMP (21, 22). HMGB1 has proinflammatory interactions with downstream receptors such as toll like receptor-4 (TLR4). The HMGB1/TLR4 pathway is well recognized in the pathophysiological process of renal ischemic injury and acute kidney injury activating NF-κB mainly through the MyD88 pathway (23). In this study we found a significant upregulation of TLR4 after NMP but not HMGB1. Dexamethasone is a glucocorticoid receptor that suppress NF-κB and MAPK-ERK activation and the subsequent translocation of HMGB1 (24). It is added to the perfusate during NMP as an anti-inflammatory agent and may therefore account for our findings.

There was no association in levels of free heme and gene expression in the series of discarded kidneys. However, based on the known effects of free heme we cannot dismiss the possibility of its contribution to the stimulation of these inflammatory and oxidative stress pathways. With such a short perfusion time, we were unable to quantify the protein expression of these pathways. As a DAMP, free heme influences the expression of many genes and, in the NMP environment in the absence of plasma proteins, free heme can interact with TLR4 to upregulate inflammatory mediators (23, 25, 26). Pathway analysis showed upregulation of NLR, oxidative stress, apoptosis, Th17 and TNF family signalling, all of which can be stimulated by free heme (27). Surprisingly, hemeoxygenase-1 (HO-1), an important enzyme rapidly stimulated to counteract the detrimental effects of free heme by catalyzing heme to carbon monoxide, ferrous iron and biliverdin (28), was not upregulated during NMP. Again, this may be due to the addition of dexamethasone during NMP which can suppress its expression (29).

The upregulation of inflammatory mediators during NMP has been associated with prolonged early graft dysfunction (30) and therefore efforts should be made to improve the NMP environment, including reducing levels of free heme, this may be particularly important for longer periods of NMP. Younger units of pRBCs would be advantageous but free heme, potassium, lactate and microparticles can be removed by washing the RBCs before use. However, this may affect the quality of the RBCs increasing their fragility and susceptibility to mechanical stress during NMP (8). Dilution of the RBCs with plasma or albumin can reduce inflammation and oxidative damage by binding to free heme (29).

The addition of an oxygen carrier during NMP is considered essential for adequate oxygen delivery to the tissues. Venema L et al, found that the addition of RBCs during more prolonged periods of NMP of porcine kidneys resulted in higher levels of oxygen consumption and function compared to conditions without RBCs (19). Artificial oxygen carriers such as hemopure are an alternative to RBCs and are being trialled in kidney NMP (31, 32). Subnormothermic temperatures or the combination of subnormothermic temperature and artificial oxygen carriers have recently been shown to reduce inflammation and pulmonary vascular resistance during ex vivo lung perfusion (EVLP) (33) and should be explored in kidney NMP.

Another strategy is to use an oxygenated acellular solution to rewarm kidneys in a controlled manner. Zlatev H et al. recently published the results of six extended criteria donor kidneys demonstrating the safety and feasible of the technique (34). Nonetheless, this has not be trialled for longer periods of perfusion.

This is the first study to examine the effects of free heme during kidney NMP. It involved two populations of kidneys: to determine the effects of free heme on kidney function during NMP a series of DCD kidneys included in a RCT were used; to determine the effect of free heme at a molecular level a series of DBD and DCD kidneys rejected for transplantation and offered for research were used. Although the research kidneys had longer cold ischemic times and were from DBD and DCD donors, levels of free heme and functional parameters during NMP were similar to the RCT kidneys. The population of DCD research kidneys were further disadvantaged with more of the DCD kidneys being perfused with older units of pRBCs. The small sample size limits any conclusion on the effect of free heme on DCD compared to DBD kidneys in this study but previous analysis of transcriptional gene expression after NMP has found results to be similar between DBD and DCD kidneys (30).

In conclusion, high levels of free heme are found in units of stored RBCs and increase significantly during 1h of NMP. There was no direct association between levels of free heme, kidney function or the upregulation of inflammatory and stress related genes during perfusion. However, this may have consequences for longer periods of perfusion and levels of hemolysis should be minimized during NMP.

The data presented in the study are deposited in the GEO repository, accession number (GSE197304) https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE197304)

The studies involving human participants were reviewed and approved by East of England, Cambridge Central Research Committee (15/EE/0356), National Research Ethics committee and Research and Development office at the University of Cambridge (NRES: 15/NE/0408). The patients/participants provided their written informed consent to participate in this study.

SH conceived the research idea, planned and conducted the experiments, analysed the results and co-wrote the manuscript; TE conducted the experiments, analysed the results and co-wrote the manuscript; NJ conducted the experiments, analysed the results and revised the manuscript; MN conceived the research idea and revised the manuscript. All authors contributed to the article and approved the submitted version.

The research was funded Kidney Research UK and by the National Institute for Health Research Blood and Transplant Research Unit (NIHR BTRU) in Organ Donation and Transplantation at the University of Cambridge in collaboration with Newcastle University and in partnership with NHS Blood and Transplant (NHSBT) Grant RG75628.

The views expressed are those of the authors and not necessarily those of the NIHR, the Department of Health and Social Care or NHSBT.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.