Current Basic Research in Normothermic Machine Perfusion

The advancing age of the deceased donor population poses significant challenges in the field of transplantation. Suboptimal function and early graft loss are more common in organs from older donors [1, 2]. These organs are also more likely to be declined for transplantation due to concerns about their quality. To address these challenges, the application of machine perfusion technologies is a promising new approach in solid organ transplantation.

Hypothermic machine perfusion (HMP) allows the circulation of cold preservation solution through an organ and is clinically established as a better method of preservation in kidney transplantation compared to static cold storage (SCS) [3]. The addition of oxygen during HMP in kidney and liver transplantation has proved to be beneficial in supporting a low level of metabolism to improve early graft function [4‒6].

Normothermic machine perfusion (NMP) is a technique that involves the circulation of an oxygenated solution through an isolated organ to re-establish metabolic function. Its utility has clear advantages compared to SCS and HMP techniques. NMP supports a level of aerobic metabolism to minimise ischaemic tissue damage, allows the monitoring of organ function, and provides an opportunity for therapeutic intervention. The reduced level of metabolism during HMP limits the ability of many therapeutics to be actively taken up and synthesised by the cells. Furthermore, parameters assessed during HMP have proved to be poor predictors of outcome.

To increase organ utilisation, NMP has become an important adjunct in clinical transplantation in the heart, lung, and liver, particularly for organs from donation after circulatory death (DCD) and older marginal donors. In liver transplantation, a large randomised controlled trial found that NMP reduced liver injury compared to SCS [7]. NMP in heart transplantation has been proved to be non-inferior to SCS [8] and is mainly reserved for DCD and marginal donors. Ex vivo lung perfusion (EVLP) is well established clinically and is beneficial for organ assessment [9]. In clinical kidney transplantation, the safety and feasibility of NMP have been established but, at present, they are not used routinely [10]. Hypothermic techniques of preservation are standard practice in most centres, but the benefits of NMP are gradually being recognised. Standardisation of protocols and the identification of better markers of viability are needed to further refine NMP techniques so that they can be adopted more readily.

This review outlines the latest understanding on the mechanistic effects of NMP in the heart, lung, liver, and kidney. The review also details recent advancements in the ability to preserve organs for days rather than hours and the use of NMP to introduce therapeutic interventions, with a focus on organ modulation (Fig. 1).

The clinical application of machine perfusion technology has led to the manufacture of a number of different clinical grade perfusion systems. A recent review by López-Martínez et al. [11] details the different clinical devices and technical specifications for NMP in all organs.

In lung, liver, and kidney transplantation, NMP is normally commenced upon arrival of the organ at the recipient centre. For simplicity, SCS or HMP facilitates transportation of the organ. The heart is extremely susceptible to cold ischaemic injury and can only tolerate 4–6 h of SCS. Therefore, if NMP is utilised, the heart is placed on the NMP device immediately after retrieval and transported to the recipient centre.

Once an organ is placed on an NMP system, the continuous circulation of an oxygenated solution re-establishes cellular metabolism and restores oxidative phosphorylation and adenosine triphosphate synthesis. In the heart, whole blood from the donor is used as an oxygen carrier during perfusion [12] but red cells obtained from local blood banks are normally used for the liver and kidney. This is supplemented with a crystalloid or a colloid solution. In the lung, there are 2 main protocols: one uses a red cell-based perfusate and the other an acellular solution [13, 14].

Nutrients, glucose, vasodilators, anti-inflammatory agents, and antibiotics are common additives to support perfusion and metabolism. Although the conditions are designed with preservation in mind, a consequence of NMP is the initiation of ischaemia-reperfusion injury (IRI), with the upregulation of inflammatory and immune mediators once circulation is restored [15, 16]. Donor comorbidities and warm and cold ischaemic injury contribute to the level of IRI and can have a profound effect on the function of the organ during NMP.

Further understanding of the mechanistic response of an organ during NMP has been gained with the application of transcriptomics and single-cell RNA sequencing. In a porcine DCD heart model, Saemann et al. [17] performed transcriptomics on samples taken from the heart before and after NMP. NMP resulted in the upregulation of genes involved in mitochondrial calcium accumulation, reactive oxygen species production and inflammation, and the downregulation of genes essential for cardiac function. In the lung, transcriptomic analysis has shown the activation of innate inflammation, cell death, and heat stress proteins and the downregulation of metabolism and protein synthesis during NMP [15]. An inflammatory signature during NMP was associated with poor graft survival. A similar finding has been reported in the kidney. Ferdinand et al. [18] reported that in DCD kidneys with prolonged delayed graft function there was significant upregulation of tumour necrosis factor α (TNFα) signalling via the nuclear factor kappa B (NF-κB) pathway and downregulation of genes relating to oxidative phosphorylation.

With single-cell RNA sequencing, more insight has been gained into the response of different populations of cells during NMP. In the lung, pathways related to inflammation were found to be activated in epithelial and endothelial cells, monocyte-derived macrophages, and alveolar macrophages after 4 h of NMP [15]. In the liver, Hautz et al. [19] characterised and tracked immune cell populations during NMP. Interestingly, they found that the most abundant immune cells were neutrophils and that they shifted from a proinflammatory state towards an aged/chronically activated/exhausted phenotype. They also identified interleukin 6 (IL-6) as a key inflammatory cytokine. IL-6 is upregulated rapidly in response to infection and tissue injury. It is a master regulator of immune and physiological processes such as innate and adaptive immunity, apoptosis, and cellular metabolism [20]. High levels of IL-6 in circulating perfusate during NMP have been identified as important predictors of outcome. Mathis et al. [21] found that levels of IL-6 during NMP were associated with catecholamine levels and mean arterial pressure in the reperfusion phase after liver transplantation. Nonetheless, with longer periods of NMP, Gilbo et al. [22] found that although levels of cytokines increased with time, by 24 h of NMP there was a shift towards an anti-inflammatory signalling profile. This suggests that longer rather than shorter durations of NMP may be beneficial for repair.

Urbanellis et al. [23] provided some insight into the transcriptional actions of NMP after transplantation using a porcine kidney autotransplant model. Kidneys were subjected to 30 min of warm ischaemia followed by 8 h of NMP and compared to an SCS control group. Transcriptomic analysis of cortical kidney samples at day 3 post-transplant revealed the enrichment of fatty acid metabolism, oxidative phosphorylation, TCA cycle, and pyruvate metabolism pathways in NMP kidneys, whereas in the cold-stored kidneys there was enrichment of genes related to mitosis, cell cycle checkpoint, and reactive oxygen species. Overall, the profile of NMP kidneys was similar to that of recovered kidneys, whereas cold-stored kidneys had similar gene expression to IRI. The study also found increased transcripts of key mitochondrial metabolic pathways following NMP [23]. This suggests that NMP kidneys were able to recover better than SCS kidneys.

Further research is needed to establish the role of the tissue-resident lymphocytes that are released into the circulating perfusate during NMP [24]. In the kidney, populations of cells include CD4+ and CD8+ tissue-resident memory T cells, natural killer cells, and helper-like innate lymphoid cells. Activated neutrophils appear to be the main source of immune activation, but more research is needed to determine their role.

Non-invasive techniques such as magnetic resonance imaging can also provide insight into the impact of ischaemic injury during NMP. The rate of perfusate flow into the kidney via the renal artery and calculation of intra-renal resistance based on the mean arterial pressure are used as global measures of function. Using magnetic resonance imaging, Hamelink et al. [25] found significant differences in the distribution of flow between in vivo and ischaemically damaged ex vivo perfused kidneys. Alterations in flow were not evident using the conventional flow measurement. Applying this type of technology enhances our understanding of the mechanistic responses of NMP and may be developed to assess viability.

The administration of therapeutics to an organ during NMP has several advantages: the therapy can be targeted precisely to the isolated organ; the effects of the therapy can be monitored; and the organ can be treated successfully without any risk to the patient. With some therapies, such as cell-based treatments, it may be necessary to carry out NMP for an extended period to provide the opportunity for repair to enhance an organ’s suitability for transplantation.

Clinically, in the heart, NMP has extended the preservation interval up to 8 h using the TransMedics Organ Care System (OCS) [26]. The heart is perfused in a resting mode with aortic perfusion (Langendorff method). There is limited application or research on attempts to extend the duration further, but there are 2 case reports of successful transplantation after 10 and 16 h of NMP [27]. There is 1 experimental report of the stable NMP of porcine hearts for 24 h using a Langendorff system with hemofiltration. The filter removed waste products and provided a more stable environment for the heart [28].

Lung NMP, or EVLP as it is more commonly known, is well established in several centres across the world using the TransMedics OCS and XVIVO perfusion devices. There are several different protocols based on an acellular or red cell-based perfusate [14]. A new NMP device that uses negative pressure ventilation to reduce ventilator lung injury has shown promising results [29]. The longest duration of clinical NMP for a transplanted lung is 17 h, although 24 h of NMP has been achieved in porcine models [30, 31]. To extend the preservation interval further, a combination of intermittent static storage at 10°C and NMP enabled porcine lungs to be preserved for 3 days [32]. This is an interesting approach that may be more practical than continual perfusion.

Clinically, livers have been transplanted after more than 24 h of NMP [33, 34]. In the experimental setting, there has been success in extending the duration of NMP further. Eshmuminov et al. [35] engineered a system designed to provide core body functions critical to liver health during NMP. The device included the automated control of glucose levels, a dialysis membrane to remove metabolic waste products, and regulation of oxygen/carbon dioxide levels. The liver was placed on a silicone mat and a balloon positioned beneath it. This was connected to an air oscillator to simulate the movement of the diaphragm and rise and fall of the liver to prevent pressure necrosis. Livers produced bile, maintained cell metabolism, and regulated blood proteins. Immediate liver function was gained in 5 pigs after transplantation following 7 days of NMP. There was no difference between levels of liver enzymes or histology compared to livers with only 2–3 h of SCS. The authors then went on to perfuse 10 human livers that were declined for transplantation following a 12–18 h SCS period. The level of injury and poor quality of these livers posed a significant challenge for perfusion. As such, the protocol was modified and livers were perfused at 34°C to provide some additional protection against IRI. Six out of the 10 livers were viable after 7 days of NMP, with decreasing injury and inflammatory markers and a reduction in the number of non-viable cells during perfusion. The authors advocated the system for the repair of declined livers with the potential to treat steatosis or modify the immunogenicity of the organ. Successful 7-day perfusion opens many avenues of research and may even facilitate the growth of new liver segments for transplantation. Although the ability of 7-day NMP to sustain liver function after transplantation has not been translated clinically, Clavien et al. [36] reported the successful transplantation of a liver perfused for 3 days using the same system. The recipient had an excellent recovery with no complications and only a minimal requirement for immunosuppressive therapy post-transplant.

The safety and feasibility of durations of NMP from 1 to 17 h have been reported in the kidney by a number of groups [10, 37‒39]. In the experimental setting, NMP durations have been extended to 48 h using a red cell-based perfusate and urine recirculation to provide metabolic stability at 37°C [40]. Each of the five declined human kidneys demonstrated different levels of function and inflammation reflecting the level of injury and condition of the organ. Montagud-Marrahi et al. [41] recently used the new ARK kidney device made by EBERS Medical^® to perfuse a kidney from an uncontrolled DCD donor for 73 h. Again, a red cell-based perfusate was used and the kidney perfused at 37°C but without urine recirculation. There was reduction in the initial acute kidney injury after 60 h but damage markers increased thereafter. The study demonstrated the feasibility of perfusion for 3 days. The use of red cells particularly with kidney NMP can be problematic [39, 42, 43]. Haemolysis occurs during perfusion due to contact with the artificial surfaces of the circuit and mechanical trauma inflicted by the perfusion pump. The release of free haem from damaged red cells can cause oxidative damage and inflammation for prolonged periods [42, 43]. To counteract this, recent research has explored the use of acellular perfusion solutions. These are either highly oxygenated tissue culture medias or simple perfusates based on human serum albumin preparations. Proof of principle has been established for up to 24 h of perfusion at a subnormothermic temperature of 22–32°C in human-declined kidneys and following transplantation of porcine kidneys in a DCD model [44, 45]. Longchamp et al. [46] also demonstrated the feasibility of acellular perfusion in the kidney with William’s E medium for 6 h at 37°C compared to NMP with red cells.

The longest duration of kidney NMP has recently been reported by de Haan et al. [47]. Declined human kidneys were perfused with a cell-free nutrient-supplemented perfusate and were metabolically active at 25°C for up to 4 days. Beyond this, kidneys demonstrated progressive injury. Unfortunately, the technique was not translated effectively in a porcine model, with kidneys showing progressive oedema during perfusion and early graft failure post-transplant. This may have been a consequence of the tissue culture-like media used for perfusion and the different requirements for porcine and human kidneys [47].

Mesenchymal stem cells (MSCs) are multipotent adult cells that can differentiate into cells of their origin. Despite their inherent progenitor properties, MSCs mediate tissue repair through paracrine factors and direct cell interactions [48]. The delivery of MSCs during NMP offers an exciting opportunity to improve donor organ quality and reduce immune activation [48]. The cells are administered directly to an organ during NMP. However, the fate of the cells during perfusion appears to differ with the composition of the perfusate and with each organ [48].

The Toronto group has recently used genetic engineering to enhance the therapeutic potential of MSCs [49]. MSCs derived from human umbilical cords with augmented production of human anti-inflammatory interleukin-10 were genetically engineered by adenoviral transduction and cryopreservation [49]. Adenoviruses are DNA viruses with double-stranded genomes that efficiently deliver genetic material to both non-dividing and dividing cells. They are commonly used to deliver genes to targeted cells. Adenoviruses have a large carrying capacity and result in high and transient transgene expression. Nonetheless, due to their capsid proteins and remaining viral DNA, they can elicit an immune response [50, 51]. In human lungs that were declined for transplantation, elevated levels of IL-10 were evident within minutes of intravascular administration of human anti-inflammatory interleukin-10 indicating their successful application [52]. Nonetheless, several lungs were of poor quality and the reduced metabolism and acidic environment during NMP led to decreased IL-10 production. The group’s recent research used a porcine lung model to explore dose-related effects of the modified MSCs. A low (20 × 10⁶) or high dose (40 × 10⁶) of cells was delivered during NMP for 6 h, after which a single-lung transplant was performed [53]. There was a dose-dependent immunomodulatory effect, with the low dose resulting in reduced apoptosis and reduced macrophage activity. Although lung function was not affected, the high dose induced inflammatory and cytotoxic CD8+ T-cell activation. MSCs were detected in the transplanted lungs after 1 h of reperfusion but were not detectible at 3 days post-transplant [53]. Genetic modification of MSCs has been tested in numerous disease models, including heart IRI, myocardial infarction, acute lung injury, liver cirrhosis, and acute kidney injury [54], but has not yet been applied during NMP in organs other than the lung.

One of the hurdles in transplantation is the development of acute or chronic rejection mediated by complement and T-cell responses. Human leucocyte antigen molecules play a central role in regulating T-cell-mediated antigen recognition, and therefore, human leucocyte antigen mismatching poses a significant problem in transplantation. Strategies to induce immunologic tolerance would reduce the immunological risk of rejection and open new opportunities for highly sensitised recipients.

The Hanover group in Germany carried out a series of experiments in heart, lung, and kidney NMP using genetic engineering to reduce organ immunogenicity. They used a lentiviral vector encoding for short hairpin RNAs (shRNAs) targeting beta-2 microglobulin and class II transactivator administered to porcine hearts during 2 h of NMP [55]. shRNA is a single-stranded molecule that folds in on itself to form a stem-loop structure. The construct for shRNA expression can be made under promoters that are transcribed by RNA polymerase II. The constructs can be packaged into viral systems and transfected to mammalian cells [56]. Lentiviral vectors are retroviruses with single-stranded RNA that infect non-dividing and dividing cells. They have the advantage over adenoviruses in that they enable long-term transgene expression and have relatively low immunogenicity [57]. Nonetheless, they are incorporated into the host genome, leading to a higher susceptibility to oncogenesis [57].

The group described stable reporter gene expression in endothelial cells and cardiomyocytes indicating successful transduction [55]. Swine leucocyte antigen (SLA) class I and class II expressions were reduced by 66% and 67%, respectively, in the vascular endothelium. Furthermore, there was no indication of increased immune activation [55].

In porcine lungs using the same shRNAs with lentiviral vector, silencing was achieved for 67% of class I and 52% of class II SLA [58]. In the most recent publication by the same group, permanent downregulation of SLA during EVLP by lentiviral transduction of shRNAs targeting mRNAs encoding beta-2 microglobulin and class II transactivator reduced the incidence of rejection in a porcine allogeneic lung transplant model. Treated lungs also had lower donor-specific antibodies and lower levels of proinflammatory cytokines post-transplant [59].

In a rodent model, the lentiviral vectors encoding shRNAs targeting the same beta-2 microglobulin and class II transactivator were introduced into the kidney [60]. The vector also contained the sequence for a secreted nanoluciferase, a reporter gene to quantify expression. Kidneys were treated during 2 h of subnormothermic machine perfusion. Following perfusion, kidneys were transplanted into a recipient and monitored for 2 weeks. Levels of MHC class I and class II transcripts were stably downregulated at 2 weeks post-transplant. Furthermore, the lentivirus DNA was exclusively restricted to the genetically engineered organ, which supports the safe delivery during ex vivo perfusion [60]. Collectively, these studies show that the combination of NMP and RNA interference technology offers a promising strategy to permanently reduce organ immunogenicity by targeting MHC class I and class II expressions.

Organ transplantation requires matching of donor and recipient ABO blood group antigens to avoid damage and hyperacute antibody-mediated rejection. There is a lower percentage of ABO compatible donors for patients who are blood group O or B. These patients can wait up to three times longer for a suitable organ compared with patients who are blood group A or AB. In deceased donor transplantation, one potential solution involves the enzymatic blood group conversion of an organ to universal blood group O. Bacteria-derived glycoside hydrolase enzymes that alter the blood group antigens on cell surfaces can be administered to an organ using ex vivo perfusion.

Based on research on the enzymatic conversion of red blood cells, Wang et al. [61] administered two enzymes, FpGalNAc deacetylase and FpGalactosaminidase, to human blood group A lungs during EVLP. Over 97% of endothelial A antigen was removed within 4 h. When challenged in an ABO incompatible ex vivo perfusion model, the treated lungs demonstrated reduced antibody binding, complement deposition, and anti-body-mediated injury compared to untreated lungs [61]. MacMillan et al. [62] recently investigated the same 2 enzymes administered to human blood group A kidneys during NMP and HMP. Blood group A antigen loss was 80% within 2 h of perfusion in each of the perfusion modalities. They also demonstrated the efficiency of the enzymes in an ABO incompatible model with reduced antibody binding and reduced activation of the classical complement pathway compared to untreated control kidneys. MacMillan et al. [63] have also demonstrated proof of principle in blood group B human kidneys using an α-galactosidase enzyme from Bacteroides fragilis. This strategy has important implications for the future of transplantation and could also be applied to the heart and liver. Modifying organs to create a universal compatible transplant using ex vivo technology within a short time will improve access for patients.

The use of NMP technology in solid organ transplantation offers many exciting opportunities for the future. New devices are being developed to facilitate the routine use of the technology in clinical practice. This is particularly important for kidney NMP which has been underdeveloped and lacks widespread clinical application. There is an abundance of research to refine protocols and determine how organs respond during NMP. This will be important in the development of targeted therapies and to ensure that organ preservation is optimal. The ability to preserve organs for days may change how transplantation is delivered in the future. Without the pressure of time, donor and recipient matching may be improved and marginal organs could be repaired or altered to improve function and increase utilisation. Techniques such as cell-based therapies, gene therapies, and specialist enzymes are continuing to be developed, and their application to an organ during NMP holds great potential.

We would like to thank Rachel Brown for reviewing and editing the manuscript.

Sarah A. Hosgood received honoraria from CytoSorbents and Aferetica. Professor Michael L. Nicholson has no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

S.A.H. conceived and wrote the manuscript. Professor M.L.N. co-wrote and reviewed the manuscript.