Novel Short-term Hypothermic Oxygenated Perfusion (HOPE) System Prevents Injury in Rat Liver Graft From Non-Heart Beating Donor

Abstract

Objective:

To assess a machine perfusion system in rescuing liver grafts from non-heart-beating donors (NHBD).

Summary Background Data:

The introduction of extracorporeal liver perfusion systems in the clinical routine depends on feasibility. Conceivably, perfusion could be performed during recipient preparation. We investigated whether a novel rat liver machine perfusion applied after in situ ischemia and cold storage can rescue NHBD liver grafts.

Methods:

We induced cardiac arrest in male Brown Norway rats by phrenotomy and ligation of the subcardial aorta. We studied 2 experimental groups: 45 minutes of warm in situ ischemia + 5 hours cold storage versus 45 minutes of warm in situ ischemia + 5 hours cold storage followed by 1 hour hypothermic oxygenated extracorporeal perfusion (HOPE). In both groups, livers were reperfused in a closed sanguineous isolated liver perfusion device for 3 hours at 37°C. To test the benefit of HOPE on survival, we performed orthotopic liver transplantation in both experimental groups.

Results:

After cold storage and reperfusion, NHBD livers showed necrosis of hepatocytes, increased release of AST, and decreased bile flow. HOPE improved NHBD livers significantly with a reduction of necrosis, less AST release, and increased bile flow. ATP was severely depleted in cold-stored NHBD livers but restored in livers treated by HOPE. After orthotopic liver transplantation, grafts treated by HOPE demonstrated a significant extension on animal survival.

Conclusions:

We demonstrate a beneficial effect of HOPE by preventing reperfusion injury in a clinically relevant NHBD model.

We assessed a new approach of machine perfusion (HOPE) in rescuing liver grafts from non-heart-beating donors. The results demonstrate a beneficial effect of HOPE by improving reperfusion injury in cold-stored non-heart-beating donor livers.

After establishment of brain death criteria in 1968, donor organs for orthotopic liver transplantation (OLT) have been almost exclusively taken from heart-beating donors. But in face of the increasing shortage of donor organs for clinical transplantation, there is now a resurgence of interest in non-heart-beating donor (NHBD) transplantation. Several centers have initiated programs for the procurements of livers and kidneys from NHBDs.^1–4 However, preservation of organ viability during cold storage of these marginal grafts is particularly endangered by the apparently reduced ischemic tolerance.⁵

To improve graft function, several strategies have been used. With respect to renal transplantation, it could be shown that viability of NHBD kidneys was improved when conventional cold storage was replaced by continuous hypothermic perfusion.^6–8 Correspondingly, Butler et al,⁹ Schön et al,¹⁰ and St. Peter et al¹¹ showed advantages of normothermic (37°C) extracorporeal machine liver perfusion after cold or warm ischemia. Other groups demonstrated the superiority of cold (4°C) machine liver perfusion over cold storage upon ex vivo reperfusion^12–14 or even in vivo reperfusion.¹⁵ The benefit of this machine perfusion during the preservation is the removal of toxic catabolites, oxygenation of the tissue, and maintenance of vascular flow. All these models are consecutively based on machine perfusion during the entire cold or warm preservation. However, the clinical situation does not allow such a strategy, as perfusion systems may not be available at the place of organ harvest and machine perfusion during transport of an NHBD organ may not be feasible. Considering these practical aspects, a certain period of cold storage appears unavoidable to transfer the graft.

We showed recently, that metabolic depletion due to cold storage of fresh resected rat livers can be effectively reversed by applying an oxygenated machine perfusion for 3 hours following cold storage.¹⁶ This approach, termed hypothermic oxygenated perfusion (HOPE), improved several critical parameters. Treatment by HOPE resulted in less oxidative stress, improved liver function, and decreased cell death as tested during ex vivo reperfusion on the isolated perfused rat liver (IPRL).¹⁶

In a next step, the presented study focused on the clinically relevant situation of a NHBD liver and asked whether HOPE would be beneficial. We therefore chose a period of warm in situ ischemia combined with a period of cold storage and applied HOPE afterward. We investigate in this study ex vivo (by IPRL) and in vivo (by OLT), whether this new approach may rescue NHBD liver grafts.

METHODS

Animals

We used male Brown Norway rats (180–200 g). The animals were maintained on standard laboratory diet and water ad libitum according to the Swiss Animal Health Care law. All animal experiments were approved by the animal ethics committee. Anesthesia during liver harvest and liver transplantation was performed with isoflurane.

Experimental Design

To achieve a clinically relevant NHBD model, we induced warm ischemia by incision of the diaphragm. Five minutes after apnea, we ligated the subcardial aorta. The period of non-heart-beating (warm ischemia) started from this point of cardiac arrest. During warm ischemia, we closed the abdomen and recorded the temperature of the liver surface (29.2°C ± 1.5°C). No manipulation at the portal vessels was performed before and during warm ischemia. We gave no heparin to the donor animals before or during the warm ischemia period. After a designated period of warm ischemia, we flushed the liver grafts with 5 mL of cold heparinized (1 U/mL) UW solution at 4°C via the portal vein. After organ harvest, we stored the livers in original UW solution for 5 hours at 4°C. We tested liver viability by reperfusion on the isolated perfused liver device (I) and in additional experiments by orthotopic liver transplantation (II).

Study I

We compared the following groups: 45 minutes of warm in situ ischemia + 5 hours cold storage + 3 hours normothermic (37°C) reperfusion (n = 9) versus 45 minutes of warm in situ ischemia + 5 hours cold storage + 1 hour hypothermic oxygenated extracorporeal perfusion (HOPE) + 3 hours normothermic (37°C) reperfusion (n = 9). Before reperfusion we introduced a 15-minutes period of ischemic rewarming to simulate organ transplantation with a rise of liver temperature from 4°C to 22°C. Fresh unpreserved livers (cold storage <30 minutes) reperfused under the same conditions served as controls (n = 5).

Hypothermic Oxygenated Machine Perfusion (HOPE)

Hypothermic machine perfusion was performed according to earlier studies^16,17 with some modifications. The same apparatus was used for cold machine perfusion as well as for normothermic reperfusion. Livers to be perfused hypothermically were connected to the perfusion device while the temperature of the perfusion box was maintained at 4°C by a Kryothermostat (Huber, Germany). Perfusion started after cold storage with a flow of 0.2 mL/min/g and was pressure controlled (3 mm Hg). We oxygenated the perfusate up to 350 mm Hg (pO₂). The first 20 mL of perfusate was collected for analysis (Table 1) and were not recirculated. The recirculating perfusion volume of the cold perfusate was 50 mL. We used a modified starch omitted UW solution containing 3 mmol/L reduced glutathione, 5000 IU SOD/L, 10 mmol/L deferoxamine, and 2 mmol/L CaCl₂ for cold machine perfusion.

TABLE 1. Biochemical Parameter After Warm Ischemia and Cold Preservation

Ex Vivo Reperfusion (IPRL)

In both experimental groups, we reperfused livers after ischemic rewarming in a closed isolated perfusion device for 3 hours at 37°C using Krebs Henseleit bicarbonate buffer + rat blood (hematocrit, 6.4 ± 1.7). Before reperfusion, we flushed all livers with 20 mL Ringer solution, which was analyzed for AST release and TBARS. The perfusion medium also contained taurocholate (0.5 g/L), indocyanine green (10 mg/L), hyaluronic acid (150 μg/L), gentamycin (80 mg/L), and heparin (2000 U/L). We performed isolated liver perfusion according to Gores et al.¹⁸ The perfusate was oxygenated by a pressurized membrane oxygenator (pO₂, 400–500 mm Hg). The total volume of the warm perfusate was 50 mL. We used 2 reservoirs (reservoirs I and II) and perfused the liver via the portal vein by gravity from reservoir I at a height of 12 cm (pressure controlled system) resulting in a constant perfusion pressure of 6.6 ± 0.5 mm Hg and a perfusion flow of 1.65 ± 0.4 mL/g liver wet weight. The liver effluent was allowed to escape freely from the inferior vena cava into reservoir II and recycled back to reservoir I by a computerized pump. Automatic niveau control of reservoir I was achieved by a computerized fluid sensor detecting changes in the perfusate level. We measured on line perfusion flow, perfusate temperature, liver temperature, perfusate pH, perfusate pressure, and perfusate oxygen inflow and outflow (AOIP software, France). The pH was adjusted to 7.4 in both experimental groups by adding bicarbonate during reperfusion.¹⁸

Assessment of Hepatocyte Function and Cell Injury

Bile flow was assessed every 30 minutes by collecting the bile in preweighed tubes.

The O₂ consumption during HOPE and reperfusion was calculated by the difference of pO₂ outflow and inflow considering the perfusion flow and liver weight:

HOU (μL/min × g liver wet weight) = portal flow × S×(pO_2inflow − pO_2outflow)/g liver wet weight¹⁹ (S, solubility constant of oxygen in water at 37°C = 0.031 μL/mL/mm Hg).

Glycolytic Metabolites and ATP

Total liver glycogen was determined after homogenization of liver samples in phosphate buffered saline (1:5), hydrolysis at 80°C, and measurement of the difference of glucose in samples treated or not treated with amyloglucosidase. For analysis of ATP and lactate, we homogenated livers (1:10) in cold 4% HClO₄. After centrifugation and pH adjustment (pH 8.5), we measured ATP by UV spectroscopy (340 nm) with hexokinase and glucose-6-phosphate dehydrogenase. Lactate was determined according to earlier studies²⁰ by UV spectroscopy (340 nm) with lactate dehydrogenase and glutamate pyruvate transaminase.

Lipid Peroxidation and Total Glutathione

Lipid peroxidation was analyzed by the amount of thiobarbituric acid reactive substances (TBARS). Homogenized liver tissue (1:5, PBS) or effluate and perfusate samples were incubated for 1 hour at 95°C after addition of the following components: 8.1% SDS, 20% acetic acid, and 0.8% thiobarbituric acid. The resulting TBARS were extracted with butanol/pyridine and measured at 532 nm. Total glutathione was determined by using a commercially available GSH-Kit (OxisResearch).

AST Release

AST release in effluate and perfusate samples was determined by using an automatic analyser.

Histology

At the end of ex vivo reperfusion (3 hours), each liver was divided into multiple random samples, which were fixed in formalin. At least 5 sections were examined by blinded observers.

Statistical analysis were performed using the nonparametric Mann-Whitney-Wilcoxon U test or 2-way analysis of variance (ANOVA) (Graph Pad Prism, version 4.0). Results are given as mean ± SD (SD). A P value below 0.05 was regarded as significant.

Study II

To test whether the results of study I would have an impact on graft survival, we performed OLT after warm ischemia + cold storage in both experimental groups. Male Brown Norway rats (180 g) were used as transplant recipients. Liver transplantation was performed by an experienced microsurgeon (Y.T.) with nonarterialized technique according to Kamada and Calne.²¹ We implanted grafts by connecting the suprahepatic vena cava with running suture, inserting cuffs into the portal vein and the subhepatic vena cava and splints into the bile duct. The rewarming time of the graft, ie, clamping of the portal vein, was adjusted to 15 minutes for comparability with the ex vivo experiments of study I. Immune rejections did not occur because of isograft transplant model.

RESULTS

Does HOPE Improve Graft Viability During Reperfusion of Cold Stored NHBD Livers?

Survival of a graft after ischemia is dependent on the extent of injury occurring after reperfusion. Therefore, AST release, bile flow, and histology of liver samples were determined during and after normothermic reperfusion. In fresh resected livers AST release during 3 hours remained very low (1.24 ± 0.9 U/g liver wet weight) (Fig. 1) and bile flow was high (1.10 ± 0.2 mL/3 hours) (Fig. 2) (Table 2). Histologic samples confirmed normal endothelial cells and hepatocytes of fresh resected livers after 3 hours reperfusion on the IPRL system (Fig. 3B). In contrast, the non-heart-beating situation (45 minutes warm in situ ischemia) combined with 5 hours cold storage induced liver injury evident by macroscopic inspection (Fig. 3C). Histologically, the livers appeared severely injured (Fig. 3D). Biochemical analysis demonstrated a significant increase in AST release during 3 hours reperfusion (7.28 ± 1.9 U/g liver wet weight) (P < 0.001) (Fig. 1), and decreased bile flow to 0.17 ± 0.1 mL/3 hours (P < 0.001, Fig. 2).

FIGURE 1. Fresh resected reperfused livers () showed no significant AST release during 3 hours normothermic ex vivo reperfusion. Livers exposed to 45 minutes of warm ischemia + 5 hours cold storage (▴) showed significantly increased AST release, which could be prevented by HOPE before reperfusion (▪).

FIGURE 2. Bile flow was high in fresh resected livers () and severely decreased in livers after 45 minutes of warm ischemia + 5 hours cold storage (▴). HOPE resulted in significantly improvement of bile flow (▪).

TABLE 2. Biochemical Parameters After Ex Vivo Reperfusion (IPRL)

FIGURE 3. A, Fresh resected livers showed normal gross appearance after 3 hours reperfusion on the IPRL. B, Histologic samples confirmed no necrosis of fresh resected reperfused livers. C, Livers after 45 minutes of warm ischemia + 5 hours cold storage showed macroscopically inhomogeneous perfusion pattern. D, Histologic samples showed multiple necrosis after reperfusion (45 minutes of warm ischemia + 5 hours cold storage). E, Livers treated by HOPE after 45 minutes of warm ischemia + 5 hours cold storage appeared macroscopically similar like fresh resected livers. F, Histologic samples demonstrated no necrosis after reperfusion (45 minutes of warm ischemia + 5 hours cold storage + 1 hour HOPE).

In a next step, we tested whether HOPE would reverse preservation injury of stored NHBD livers. Indeed, all parameters tested were markedly improved. Macroscopically, the liver appeared comparable to fresh resected livers (Fig. 3E vs. Fig. 3A). Histologic samples confirmed no necrosis (Fig. 3F); biochemical analysis showed significantly less AST release (3.94 ± 1.3 U/g wet weight) (Fig. 1) and increased bile flow (0.50 ± 0.2 mL/3 hours) (P < 0.001, respectively) (Fig. 2; Table 2) during reperfusion.

We conclude that liver injury caused by warm ischemia and cold storage can be significantly decreased by treatment with HOPE.

Does HOPE Influence Oxidative Stress During Normothermic Reperfusion of Cold Stored NHBD Livers?

Oxidative stress plays a decisive role during reperfusion. We postulated that either HOPE may have a negative impact on reactive oxygen species (ROS) due to oxygenation during hypothermia²² or may even decrease oxygen free radicals due to improved mitochondrial integrity.²³ To test both hypotheses, we determined TBARS in liver tissue and effluate samples to indicate extracellular lipid peroxidation. In addition, we analyzed the loss of total cellular glutathione (GSH) in liver tissue as a marker of intracellular oxygen radical scavenger. HOPE did not increase TBARS levels in effluate after cold machine perfusion (Table 1). Also, TBARS of liver extracts were not increased in either experimental groups (Table 2). In contrast, total liver glutathione was depleted to 40.6% of the baseline (reperfused fresh resected liver, Table 2) after 45 minutes warm ischemia + 5 hours cold storage. Treatment with HOPE, however, increased total liver glutathione significantly to 72.6% of the baseline (Table 2). The decrease of cellular GSH can be interpreted as oxidation of GSH to GSSG and subsequently loss of GSSG across cellular membrane.

We conclude that oxygen mediated stress during reperfusion of cold stored NHBD livers can be prevented by HOPE.

Does HOPE Improve Mitochondrial Function During Normothermic Reperfusion of Cold Stored NHBD Livers?

A number of studies have suggested that liver cell viability is dependent on ATP during ischemia-reperfusion.²⁴ To analyze, whether ATP recovery was depressed after warm ischemia and cold storage, we determined tissue levels of ATP after 3 hours of reperfusion. Compared with fresh resected livers, ATP was significantly decreased (Table 2). Applying HOPE significantly improved ATP recovery during reperfusion, resulting in ATP levels that were 10 times higher compared with cold-stored NHBD livers without treatment of HOPE (P < 0.001, Table 2).

To test glycolysis end products, we determined lactate in the perfusate in both experimental groups. Lactate accumulated in the perfusate during 3 hours of reperfusion after warm ischemia and cold storage to 16.5 mmol/L despite oxygenation during reperfusion. In contrast, NHBD livers treated by HOPE were able to degrade perfusate lactate down to physiologic values (2.65 mmol/L), which implies intact mitochondrial function after treatment with HOPE (Fig. 4). To support these results, we determined O₂ consumption during reperfusion. Livers treated with HOPE showed significantly lower O₂ consumption during reperfusion (Fig. 5). The reduced hepatic oxygen uptake and the high ATP levels after HOPE suggest remarkably improved mitochondrial efficiency.

FIGURE 4. Perfusate lactate degradation during reperfusion was not different between fresh resected livers () and livers after 45 minutes of warm ischemia + 5 hours cold storage if HOPE was performed before reperfusion (▪). In contrast, ischemic livers not treated by HOPE exhibited a high amount of lactate, which could not be degraded to physiologic levels during 3 hours of reperfusion (▴).

FIGURE 5. Hepatic oxygen uptake (HOU) was low and constant during reperfusion of fresh resected livers (). After 45 minutes of warm ischemia + 5 hours cold storage, oxygen consumption was significantly increased (▴). HOPE before reperfusion decreased hepatic oxygen uptake during reperfusion (▪).

Finally, we tested glycogen levels of liver tissue after reperfusion. They were significantly decreased in both experimental groups (Table 2). Despite increased energy resynthesis by HOPE, glycogen levels were not different in the 2 experimental groups (Table 2), implicating only minor loss of glycogen during HOPE. This can be explained by glucose excess in the perfusate of the liver during cold machine perfusion (Table 1).

In conclusion, treatment by HOPE after 45 minutes of warm in situ ischemia + 5 hours cold storage was highly effective in terms to ATP stores and glucose metabolism during reperfusion. Mitochondrial injury was prevented by HOPE.

Does HOPE Improve Survival After Transplantation of Cold Stored NHBD Livers?

In further experiments, we attempted to show that HOPE was beneficial for survival after transplantation of a NHBD liver. Under experimental conditions (45 minutes of warm in situ ischemia + 5 hours cold storage), all animals survived liver transplantation irrespective of the use of HOPE (n = 6). However, animals pretreated with HOPE disclosed normal mobility and adequate behavior rapidly after surgery; for example, these animals could resume enteral feeding within 2 hours of surgery. On the other hand, animals without HOPE treatment disclosed a sleepy behavior for the first 48 hours after surgery, but eventually recovered fully. Therefore, we extended the period of warm ischemia to 90 minutes, and again important differences in behavior were observed in favor of the use of HOPE, although all animals recovered (n = 6). With 120 minutes of warm ischemia, all recipient rats died between 2 and 4 hours of surgery (n = 5) if HOPE was not performed. Of note, these grafts exhibited severe tissue injury, which was macroscopically visible prior to implantation. The use of HOPE under similar conditions resulted in significantly longer survival, although all animals died within 24 hours of surgery (range, 15–24 hours, n = 5) (P = 0.008). These grafts appeared macroscopically much improved at the time of implantation.

DISCUSSION

In this study, we demonstrate that short-term hypothermic oxygenated perfusion (HOPE) improves rat livers injured by a combination of warm and cold ischemia. Livers, taken from NHBDs after 45 minutes of in situ warm ischemia and cold storage (5 hours), exhibited significant macroscopic and microscopic damage. HOPE reversed this damage resulting in viable livers.

Up to now, machine perfusion is not used in routine clinical practice because it requires sophisticated machines that might not be available at the site of harvest, and transport may not be feasible. Hence, we set out to establish a rat model that meets several criteria: clinical relevance, feasibility, and simplicity of use. From a recent study, we knew that metabolic reloading after a short time of cold storage (10 hours) is possible by HOPE. Those livers showed less reperfusion injury on the IPRL system.¹⁶

The goal of this study was to provide tools that can extend the pool of livers for organ transplantation. For the first part of the study, we chose 45 minutes of warm ischemia, a period reflecting clinical practice of NHBDs. We furthermore assumed that the harvested liver has to be transported to a larger center. Therefore, we added 5 hours of cold ischemia. Consequently, HOPE, ie, machine perfusion, was performed afterward. In former studies, we decided to perform HOPE over 3 hours after cold storage to achieve an optimal ATP loading prior to reperfusion.¹⁶ In this study, we reduced the machine perfusion time to 1 hour in order not to delay liver transplantation. We demonstrated that ATP recovery during reperfusion can also satisfactorily be improved by 1-hour machine perfusion after 45 minutes of warm ischemia + 5 hours cold storage. In addition, this simple intervention by HOPE changed the antioxidative status of the depleted liver and reduced cell death during reperfusion. These results are new and exciting, pointing to the potential to improve NHBD organs.

Numerous strategies have already been reported to rescue NHBD livers. So far 3 different methods have been presented: improvement of liver microcirculation by thrombolysis during initial flush, oxygenation of the graft during preservation after organ harvest, and cytoprotective strategies in the donor and/or recipients.

The first 2 strategies describe effective improvement of graft quality if either cold or warm ischemia precedes reperfusion. Administration of streptokinase during the cold flush of the ischemic liver has been shown to improve microvascular perfusion and to reduce hepatocellular enzyme release in rat ex vivo reperfusion after warm ischemia.^25–27 Also, the use of low viscosity cold flush solutions (polysol, celsior)^28–30 has led to lower vascular resistance and improved survival of NHBD livers if applied directly after harvest with minimal cold storage time. Oxygenation of the graft either by cold or warm machine perfusion or oxygen persufflation did successfully rescue warm or cold ischemically damaged livers if applied during the entire time of preservation.^8–15 No success was shown if warm ischemia was combined with cold storage,^31,32 a scenario that, however, reflects the real situation of multiorgan retrieval followed by organ transport. The authors concluded that even a brief period of cold ischemia after warm ischemia is deleterious and abrogated the previously shown benefit of normothermic preservation of ischemically damaged livers.³¹

Cytoprotective strategies were reported to be effective in various animal models.^33–36 But most clinical trials would require donor pretreatment and could be tested only if pretreatment is allowed. Few studies yet describe improvement of cold stored NHBD rat livers without pretreatment of the donor. Gu et al³⁷ reported porcine liver graft protection after 45 minutes of warm in situ ischemia + 8 hours cold storage by applying an endothelin antagonist (TAK-044) and a platelet activating factor antagonist (E5880) to the preservation solution and the recipient. Warm in situ ischemia in this model was induced by disconnection of the respirator, which led to cardiac arrest after additional 10.5 minutes, resulting in actual non-heart-beating period of 35 minutes. Other techniques to induce NHBD include ex situ ischemia,¹⁵ phrenotomy,^14,38,39 injection of KCl,^40,41 ligation of the heart base,³⁶ exsanguination,^10,31 and overdose of ether.⁴² Most of the published results concerning NHBD livers are therefore difficult to compare.

We used an apnea induced NHBD model where the graft is receiving portal flow through hypoxic or ischemic intestine. This reflects the clinical situation of an uncontrolled NHBD.⁴³ In pilot experiments, however, we found a large variability of ongoing cardiac action in spite of anoxia, ranging from 10 to 60 minutes after apnea. To get a reliable starting point of, ie, cardiac arrest we decided to ligate the subcardial aorta 5 minutes after apnea. Forty-five minutes after cardiovascular arrest followed by 5 hours of cold storage, livers were treated with HOPE. They exhibited an effective recovery of ATP, protection of cellular glutathione, and reduced cell death, which was not observed in untreated livers.

Survival experiments are one of the important endpoints to demonstrate whether HOPE would be beneficial in NHBDs. Therefore, we attempted to investigate whether grafts treated by HOPE would improve survival time after transplantation. To establish a lethal injury, we had to extend the period of warm ischemia up to 120 minutes. In pilot experiments, there was a macroscopic improvement by HOPE as well as a benefit on the length of rat survival. At this point, animals of both experimental groups died. However, HOPE was only used 1 hour in this study instead of 3 hours in previous reports.¹⁶ Therefore, future studies will establish the time with maximal benefit for the graft. Whether this approach will result in better graft survival in comparison with other modalities (machine perfusion using low viscosity solutions, eg, polysol;³⁸ venous oxygen persufflation during cold preservation⁴¹) remains to be seen.

The underlying protective mechanism of HOPE remains unclear. In ischemia-reperfusion injury, a decrease in ATP synthesis can be explained by a reduced function of mitochondrial NADH dehydrogenase but also by an increase of ROS. The first mechanism prevents optimal generation of an electrochemical gradient across the inner mitochondrial membrane and results in impaired ATP synthase.⁴⁴ The second mechanism is an increase in mitochondrial ROS by electron leakage that is injurious to mitochondrial function.^45,46 Former studies have shown that mitochondrial injury can be prevented by HOPE after cold storage.²³ We suggest therefore that HOPE may be helpful in maintaining or stabilizing an optimal mitochondrial electrochemical gradient during reperfusion. This implies a central role of oxygenation during preservation.

CONCLUSION

We could demonstrate in an animal model the feasibility of applying HOPE. Rescue of NHBD livers was performed in a manner mimicking the clinical situation. Thus, it may be possible to harvest livers from NHBD by the conventional approach and add machine perfusion during the period of recipient preparation. This clinically attractive strategy will be further tested in a large animal model with the aim to expand the donor liver pool.

Discussions

Dr. Thomas van Gulik: Thank you very much. It is a great pleasure to open the discussion on this paper. Obviously, the authors address a very important issue in organ transplantation, that is, the use of organs retrieved from non-heart-beating donors. That is a concept that is widely accepted in kidneys but is slow to follow in liver transplantation obviously because the liver is much more susceptible to ischemic injury. It has been shown in the setting of kidney transplantation that machine perfusion has a benefit in improving preservation and improving function of the graft after transplantation.

Now in this paper, the authors have very nicely shown, in a clinically relevant model, that there is a benefit of a one-time perfusion after a period of cold storage because the premise of their study is that machine perfusion is difficult, complex, and probably not as feasible in the setting of a non-heart-beating donor. So the sequence of the model is very relevant in that a period of 45 minutes ischemia is introduced, after which there is 5 hours of cold storage, then a reperfusion period of 3 hours and then the comparison is whether 1 hour machine perfusion improves the organ. It has been shown on the energy parameter level that there is a benefit of this one-time perfusion and also on the level of parenchymal damage as you showed, in that the AST levels were lower.

I have a comment and a few questions. The first comment is that I argue that machine perfusion should be complex. It is complex because it involves machinery that is too difficult to handle. What we need are simple machines that are ready to use, disposable, available in every hospital, and would be as easy to use as putting an organ into a bag with a preservation solution as we do with cold storage. That is a comment, but that means that we have to develop those machines and we have to educate the transplantation community about that.

Now the questions. You emphasize that you use HOPE, which is hypothermic oxygenated extracorporeal perfusion. Now there is a contradiction between hypothermia and oxygenation because what you do in hypothermia is that you decrease metabolism to a level of 4°C where there is probably no need for oxygen. So my first question is: how do you look at oxygenation at 4°C when we all know that oxygen supply is probably more important when you go to subnormothermic levels in which you have a certain level of metabolism that is still preserved? That is the first question.

The next question addresses the composition of your perfusion solution. You use an UW solution in which the starch has been omitted. Now that is quite an unorthodox view because, in many studies, it has been shown that what you need is a colloid and, if you omit the hydroxy-ethyl starch, you omit one of the most essential components of a perfusion solution. So my question is: what was the rationale to do that and wouldn't you have a much better perfusion solution if you used the gold standard solution for machine perfusion? This would be the modified UW solution in which lactobionic acid is replaced by gluconate. And as a machine perfusion solution, it of course contains hydroxy-ethyl starch.

And the third question addresses one of the most crucial factors in organ preservation injury, and that is the state of the microvasculature. You showed that the sinusoidal endothelial cells, morphologically, were improved in your transmission electron microsocopy pictures. On the other hand, you showed in your hyaluronic acid uptake experiments that there was no difference. So my question is: how can you reconcile this result with the improved morphology?

But the most important thing is, of course, what is the intravascular pressure during perfusion. Now, your experimental setup is very nice and you showed in your methods slide that it is a pressure-controlled perfusion system. You showed that you measure flow, but you didn't show us the outcome of the flow parameter. This parameter shows what the intravascular resistance of the organ is, and this is one of the paramount parameters for success of preservation.

Now, finally, you showed in the rat liver transplantation experiments, which are the single most important, to prove that using the setup of the previous experiments there was no survival benefit from your extracorporeal perfusion. The question is: whether when you look instead of survival at, let's say, soft parameters, such as AST release, you would find a difference? But you went on to 120 minutes of warm ischemia and now you could show a benefit. I doubt if that is realistic because when you subject a rat liver to 120 minutes of warm ischemia, in my experience, it is almost all necrotic.

All together, I think that you made some very important points. I want to congratulate you and your group with these results, and I would like to thank the organization for allowing me to put forward these comments.

Dr. Philipp Dutkowski: Thank you very much for these interesting and good questions. We investigated the level of oxygenation in several prestudies, which showed that oxygen was necessary to achieve a very fast ATP increase or energy resynthesis. Because our interest was not to delay the transplantation procedure, we tried to perfuse as short as possible with a maximum increase in ATP. Our previous data suggested that ATP resynthesis clearly increased during machine perfusion with an increase of oxygen concentration. That was the reason for oxygenation during hypothermic perfusion.

Regarding the second question, we did not choose the gluconate-based UW solution for liver perfusion because we started our first machine liver perfusion experiments with the original, lactobionate based UW solution, which was, at that time, the most efficient liver preservation solution. However, when we machine-perfused rat livers with lactobionate UW, we found a severe increase in vascular resistance after half an hour of perfusion. In contrast, when we omitted starch in original UW, we achieved no increase of vascular resistance. In spite of no presence of starch, we never observed with our technique of short machine perfusion an increase of tissue edema.

The third question was about hyaluronic acid uptake, which was not different in both experimental groups, although electron microscopy showed a clear difference regarding endothelial cells. In our opinion, hyaluronic acid uptake is one test for endothelial function, but this does not necessarily completely exclude endothelial cell injury. It is just monitoring one aspect of endothelial function in the liver.

The fourth question focuses on parameters of perfusion flow. We observed no decrease in perfusion flow during the whole 3 hours of extracorporeal reperfusion in all experimental groups. According to this, vascular resistance also did not increase during the reperfusion period. To achieve these results, we first tested our perfusion apparatus in fresh resected livers before we started experiments with non-heart-beating donor livers.

The last question was about parameters after orthotopic transplantation. When we started our experiments, we assumed that 45 minutes and 5 hours of cold storage should be a lethal injury. There are several studies that clearly support this hypothesis. We were very surprised that we had no death after this remarkable period of warm and cold ischemia, and this is probably based on our experienced microsurgeon. Then we went to 60 minutes of warm ischemia and to 90 minutes warm ischemia plus 5-hour cold storage and the rats did not die. But we wanted to achieve a lethal condition, and so we jumped to 120 minutes. We agree that this is probably too much, but, in the next step, we will go back and check how this will work. We did not take blood samples after the transplantation because we did not want to confuse the experiment. So we cannot show AST release after liver transplantation.

Dr. Peter Friend: Thank you very much for really fascinating data, which are spectacular—I am most impressed. My question is whether in fact it is the delivery of oxygen or the perfusion, which has given you this extraordinary benefit. I don't think you described any studies where you just perfused the organs without delivering oxygen. Of course, quite a lot of the data now emerging on machine perfusion suggest that it's the business of perfusing and removing metabolites which is perhaps the most important. The role of oxygen delivery at low temperature is perhaps a little more controversial. Have you done any experiments in which you have separated those 2 variables?

Dr. Philipp Dutkowski: No. We will do it, of course. This is an important point if we want to understand the underlying mechanism of these results. One possible effect is in regard to mitochondrial function. To prove this, we will perform experiments without oxygen. For this purpose, we included in our machine the possibility to exchange oxygen with nitrogen.