Static cold storage (CS) using the University of Wisconsin solution (Viaspan^®) (UW) is the gold standard of preservation of kidneys obtained from deceased donors 1 . Its introduction in the late nineteen eighties has reduced the incidence of delayed graft function (DGF) and improved graft survival of kidneys obtained from donations after brain death 2 . Nevertheless, the growing use of expanded criteria donors (ECD), donors with acute renal failure 3 4 and deceased after cardiac death donors (DCD) has increased the DGF incidence of graft preserved by UW 5 or by CS in general 6 7 .

Use of DCD grafts in the clinic is limited by a high rate of primary non function and DGF 7 8 9 , in correlation with the length of the warm ischemia period 6 . However, as they represent a significant increase in the pool of donors (30%), which is of particular importance in the current shortage (only one out of three patients on the waiting list receives a kidney), finding the optimal way to preserve these organs and improve their quality as become a first order issue.

Hypothermic machine perfusion (MP) preservation is increasingly being used as an alternative preservation method to CS. Studies have reported a reduction of DGF after MP compared to CS 10 11 12 13 14 15 16 17 18 , however the solutions used were different, and some studies lacked proper randomization. These early clinical data were supported by experimental studies, conducted in large animal models of DCD using different preservation solutions, reporting improvements of kidney function after MP 19 20 21 22 . Nevertheless, not all animal studies support the superiority of MP over CS in DCD models. Indeed, MP of pig DCD kidneys using a combination of Belzer machine perfusion solution (MPS) and Viaspan^® did not reveal any superior effect to ViaspanUW-CS 20 and when the same preservation solution was used in both the CS- and MP- groups, no significant difference between MP and CS preservation could be observed in dogs 21 or pigs 23 for WI times of up to 60 min. A better performance of ViaspanUW-MP was, however, reported for longer WI times in dogs 21 . These experimental data question the necessity of MP for DCD kidneys. Clinical evidence on the use of MP and its benefits can be conflicting 24 25 26 , however recent clinical trials show small but significant benefits of MP over CS 27 in terms of DGF rate and one year survival of grafts from all categories of donors and further studies demonstrated some benefits from MP in terms of DGF and function in a DCD subset 28 .

Hence, clinical evidence for the superiority of MP over CS in DCD kidney transplantation is accumulating and interest in MP is still growing 29 30 31 32 as new machines 33 34 and preservation concepts 35 are being developed. Nevertheless there is also a need for preclinical studies in a standardized transplantation model to investigate the benefits of MP on both acute and chronic kidney injury.

The present study uses a recently developed porcine model mimicking conditions of DCD class I and II 36 37 , by 60 minutes of WI before organ collection and storage. We propose a four-way comparison using preservation with Viaspan^® (ViaspanUW), the gold standard in CS, either by CS or MP, and preservation with Kidney preservation solution-1^® (KPS), recommended for MP, either by CS or MP.

We will measure function recovery, quality of function, chronic immune response development, chronic fibrosis development and animal survival. This will allow us to determine a 'machine effect' independently of the solution used, as well as measure benefits of clinical MP (KPS-MP) versus clinical CS (ViaspanUW-CS).

Herein, we demonstrate in a preclinical study using a highly reproducible swine model of transplantation the benefits of MP over CS, particularly in regards to chronic outcome.

We performed static preservation with both ViaspanUW and KPS, demonstrating the superiority of KPS in terms of function recovery, histology at D7 and survival. Comparisons of these two groups offers a perspective on studies generally performed on machine perfusion: when two different solutions are used for static and machine preservation, the observed effect is not solely due on perfusion but also depends significantly on the solution used. Our 4 groups/2 variables approach circumvents this bias, highlighting the importance of animal studies in large animals to assess the benefits of novel therapies, as indeed such setting is impossible in the clinic.

Weight variations of kidney grafts are classically observed during preservation. Our observation of weight loss for CS and weight gain for MP are consistent with a similar experimental design in pigs 23 . In addition, increases in kidney weight after MP have been previously reported to have no significant impact on the graft outcome 40 .

Comparing ViaspanUW-CS to ViaspanUW-MP allows us to determine the benefits of machine perfusion with the current high-K⁺ gold standard in static preservation. Although ViaspanUW is not used for MP in clinical settings, using identical preservation solution focuses the analysis solely on the effect of perfusion. Early follow up with classical tools such as serum creatinine do not allow to determine differences between the two methods. In our setting, pigs were not dialyzed thus analysis of diuresis was pertinent, but this would not be the case in the clinic. Interestingly, measure of peripheral blood gluthathion red/ox status provided discriminating information between the groups, which was enhanced by analysis of histology at day 7. Use of UW demonstrates in the clearest fashion the benefits of MP: while high concentration of potassium induce vasoconstriction, as seen in the resistance index at beginning of perfusion, the machine is able to rescue this negative effect and regulate flow, allowing the organ to better face the stress of reperfusion, with dramatic benefits on outcome particularly survival, such as found in the clinic 27 28 . This model thus offers a unique opportunity for further clarification of the exact mechanisms through which MP provides this protection.

Benefits of machine perfusion were also not immediately obvious between KPS-CS and KPS-MP groups: diuresis and creatinine levels were close, as were other functional parameters usually available in the clinic. Here also, discrimination was possible with measure of Glutathion red/ox. Moreover, since both groups produced urine, proximal tubule enzymes activity assay in the urine was invaluable. Alanine aminopeptidase and β-N-acetylglucosaminidase are found in kidney tubular cells brush border and their presence in urine is a commonly accepted sign of tubular damage 41 , their activity level in the urine revealed a superiority of MP in maintaining tissue integrity at all time point, which was confirmed by histological analysis of the grafts parenchyma.

Early follow up of ViaspanUW-MP and KPS-CS showed similar values on the tests we performed, highlighting the existence of a solution bias when comparing preservation strategies. Altogether, results from the early follow up do not permit a clear discrimination between CS and MP, unless we consider less orthodox tests such as glutathione red/ox or urinary tubular enzyme activity assays. Excretion of Na⁺ and glycosuria, in a context of normoglycemia, also offered a degree of discrimination between experimental groups for tubular necrosis and tubular dysfunction.

In the case of glutathione red/ox, a clear correlation was drawn between the use of MP and lower oxidative stress, and both solution and perfusion technique demonstrated an effect on this parameter. However, addition of effects was not found until day 7. We thus identify an independent machine effect, however the relatively small differences observed herein would likely not be present in the clinic due to disparities in patients population, while in identical pigs statistical significance is obtainable. Grafts histology analysis confirmed the superiority of MP over CS, however these tests may not be standard in clinical practice. Thus, measurement of the benefits of MP is difficult in short follow up studies, particularly if the preservation solution bias is not circumvented.

We followed animals for 3 month post reperfusion. In this large animal model, this length permits us to follow the development of chronic lesions such as immune response and interstitial fibrosis and tubular atrophy (IFTA). The summated effects of damage sustained by organ preservation and reperfusion 42 lead to loss of graft function, and ultimately loss of the grafts itself, often due to the development of IFTA 43 . This pathology is also strongly correlated with immune response 42 44 45 46 . Herein, KPS-MP showed less innate and adaptative invasion compared to KPS-CS, which showed lower levels that ViaspanUW-MP. Use of KPS correlated with lower invasion, and within the KPS groups we showed that the use of MP correlated with better outcome. Unfortunately, absence of data from the UW-CS group did not allow us to perform further statistical analysis. This confirms the benefits of the machine on chronic immune response development. The ViaspanUW-MP fared poorly compared to KPS groups, however its superiority to ViaspanUW-CS is demonstrated in terms of animal survival. These results are in contradiction to a study conducted on dogs 21 , however the setting of the study and the anatomy of the dog kidney render the comparison of data difficult.

Epithelial to mesenchymal transition (EMT), a process through which polarized tubular cells are driven to de-differentiate and alter their phenotype towards that of a mobile and fast proliferating mesenchymal cell 47 , is shown to be a repair mechanism that can be deregulated during injury and promote interstitial fibrosis 48 49 50 . Our results show that Vimentin staining, a marker of EMT, is high in ViaspanUW-MP, lower in KPS-CS and close to control levels in KPS-MP. Thus, the machine effect is also found in a major pathway leading to fibrosis and graft loss. We measured the extend of fibrosis using Sirius red and showed a similar order in the grade of lesion: ViaspanUW-MP was highest and KPS-CS showed half the degree of fibrosis of ViaspanUW-MP. KPS-MP group did not show a degree of fibrosis higher than control.

Considering no ViaspanUW-CS animal survived to the end of the follow up, no comparison is possible in regards to chronic lesions such as immune response or fibrosis, however previous studies using the same protocol as ViaspanUW-CS showed a 27% survival rate with important immune response and IFTA (47%) 51 52 . ViaspanUW-MP showed better survival, strengthening the results of a similar study investigating the short-term effects (7 days) of ViaspanUW-MP in a pig model 19 , also reporting trends towards a better early kidney function after MP 19 23 . Our results demonstrate superiority of KPS over ViaspanUW solution in our animal model, independently of the preservation strategy. UW is a high K⁺ and low Na⁺ solution 53 , proposed to maintain intracellular ionic balance. However high potassium has been shown to induce cellular depolarization, decrease cellular ATP content and activates voltage-dependent channels, such as calcium channels 54 55 . Influx of calcium can result in vasoconstriction impairing organ perfusion during washout and reperfusion, participating in the 'no reflow' phenomenon 56 57 58 . Recently, studies have shown equal or improved results of low potassium/high sodium ratio such as KPS 1 , consistent with our findings. Use of Mannitol instead of lactobionate in KPS may also account for the better performance, as this compound has reactive oxygen species scavenging properties 1 .

The present study uses large white pigs, an animal well suited for preclinical studies as it is close to humans, particularly in regards to the multipapillar multilobular organization of its kidney, only found in higher mammals, implying a complex vascular bed making these organs particularly sensitive to IRI 59 . In this setting, we determined that the benefits of machine perfusion, with a machine currently used in the clinic, are most evident on chronic graft outcome. Indeed, discrimination between the groups in the early time points was only possible through assays rarely performed in transplant centers and thus could explain the relatively small benefits found in clinical studies investigating the machine effect 27 . However, our results suggest that chronic follow up of these patients will uncover a wider rift between MP and CS, as chronic lesions start to develop.

The exact mechanisms by which MP minimizes the activation of lesional pathways in our study remain to be elucidated. MP actions may include a complete perfusion of the organ promoting a thorough washout of blood and subsequent tissue equilibration with the preservation solution. This more efficient washout has been previously reported to limit the aggregation of erythrocytes 60 . Finally, the maintenance of a flow may protect against depolarization of the endothelial cell membrane which is linked to generation of ROS, increased intracellular Ca²⁺ concentration, and activation of NO synthases 61 . Hence, more mechanistic studies are necessary to unravel the exact mechanism of action in MP, in order to focus on improvement and optimal application of this technique.

The present study appears limited by the use of an isograft model, devoid of the influence of immunosuppressants. However, machine perfusion has been developed to optimize graft preservation, hence address ischemia reperfusion injury. Thus, we felt that an allograft model, with the addition of immunosuppressors and their own set of deleterious side effects, would dilute the impact of our results. We thus sacrificed relevance to the clinic by the use of isograft in order to obtain clarity of our results in regards to the benefits of machine perfusion. Another limitation is the fact that our model does not follow exactly the setting of classes I and II of the Maastricht criteria. Indeed, it normally includes no more than 30 min arrest before starting the CPR procedure, which is then continued during the transport to the hospital (generally with a machine); then as failure to resuscitate is pronounced there is a 5 min no touch period. All these steps should not exceed 150 min. The patient is then either cold perfused or a extracorporeal circuit is put in place, giving enough time to secure consent from the family and collect the organs, which are then machine perfused. It is obvious that a correct modelling of this situation should include all these steps, and we are actually in the process of adapting such procedures on the pig. However in the meantime we are using 60 min WI as it reproduces as closely as possible the conditions of DCD.

In a study using a preclinical model of DCD kidney transplantation, we demonstrate the superiority of MP over CS independently of the solution used for perfusion. Our results suggest significant benefits on graft outcome, particularly evident on the chronic effects of IRI with a protection against chronic immune response, EMT and IFTA.

The authors declare that they have no competing interests.

NV carried out the animal experiment design, surgery, data gathering. RT carried out the data management and processing, writing the paper. FF carried out experimental design, running experiments, data gathering. NC carried out the experiments, data gathering. SM carried out the experimental design, running experiments, data gathering. ME carried out the study design. TH carried out the study design, writing the paper. BB carried out the animal experimental design, study design. All authors read and approved the final manuscript.