Liver donor shortages stimulate the development of strategies that incorporate damaged organs into the donor pool. Herein we present a simplified machine perfusion system without the need for oxygen carriers or temperature control, which we validated in a model of orthotopic liver transplantation.
Rat livers were procured and subnormothermically perfused with supplemented Williams E medium for 3 hours, then transplanted into healthy recipients (Fresh-SNMP group). Outcome was compared with static cold stored organs (UW-Control group). In addition, a rat liver model of donation after cardiac death was adapted using a 60-minute warm ischemic period, after which the grafts were either transplanted directly (WI group) or subnormothermically perfused and transplanted (WI-SNMP group).
One-month survival was 100% in the Fresh-SNMP and UW-Control groups, 83.3% in the WI-SNMP group and 0% in the WI group. Clinical parameters, postoperative blood work and histology did not differ significantly between survivors.
This work demonstrates for the first time in an orthotopic transplantation model that ischemically damaged livers can be regenerated effectively using practical subnormothermic machine perfusion without oxygen carriers.
KeywordsLiver Transplantation Rat Machine Perfusion Preservation Ischemia Donor Shortage
- H & E
Hematoxylin and eosin
The discrepancy between the amount of livers transplanted annually and patients on the waiting list has been growing over the past decade. In 2009, the waiting list mortality rate was 15% in the United States (approximately 1,500 patients), whereas the transplantation rate was about 38% (5,000 transplantations) . This shortage has driven scientific efforts to increase organ availability targeted at upgrading conventional storage methods (static cold storage (SCS) in University of Wisconsin (UW) solution) or employing graft optimization to allow for extension of donor criteria. First reported in 1967 , machine perfusion is a modality whereby organs can be assessed, preserved and treated. In 2009, the first clinical trial with hypothermic machine perfusion (HMP) of human livers was conducted . Besides preserving organs , HMP can regenerate ischemic damage and shows potential for reclaiming livers derived from donation after cardiac death (DCD) [5, 6]. Although an environment of hypothermia facilitates preservation by slowing cellular metabolism and thereby reducing the need for nutrients and oxygen, the low temperature can cause damage to the microvasculature. Intravital microscopy of hypothermically perfused rat livers has demonstrated that temperature-dependent cellular deformation of sinusoidal endothelial cells obstructs flow, which congests the sinusoid and decreases parenchymal perfusion [4, 7]. In addition, the lower temperature decreases the fluidity of the plasma membrane, which can lead to cell lysis and increased enzyme leakage . It follows that because HMP is superior to SCS at temperatures where metabolism is slowed, a separate, perhaps mechanistic phenomenon contributes to this advantage [9, 10]. One possibility is that during machine perfusion, harmful waste products are flushed out of the organ and the continuous flow facilitates better availability of the beneficial solution to the peripheral parenchyma of the organ. In addition to HMP, researchers in many studies have explored normothermic machine perfusion (NMP) [11–17]. Because higher temperature involves increased metabolism, it becomes more essential to promote oxygen delivery, cellular function and bile homeostasis while restricting adverse processes such as nutrient depletion, apoptosis and cellular swelling. Raising the temperature extends both the potential for resuscitation and the risk of organ damage [11, 13, 14]. Although few studies have compared HMP with NMP in the same setting, NMP has been successful in regenerating warm ischemic grafts and preserving livers in rat and porcine models, proving its superiority over SCS [17–24].
Studies of machine perfusion have led to the need for additional research, but new trials have been slow to emerge. A potential reason lies in the relative complexity of machine perfusion, which also has hampered its progress when introduced in kidney transplantation [25, 26]. A simplified machine perfusion protocol might expedite its implementation, benefiting from the practicality of cold storage while not relinquishing the advantages of machine perfusion. In this context, subnormothermic machine perfusion (SNMP) may offer multiple advantages: no temperature control and a moderate rate of metabolism in which beneficial processes still occur while adverse cellular processes may be controlled.
Coupled with metabolism at the temperatures used in SNMP and NMP is the need for aerobic respiration. Although erythrocytes or artificial oxygen carriers added to the perfusate are effective in delivering oxygen to the liver [11, 18, 27, 28], Vairetti et al. successfully used oxygenated SNMP without an oxygen carrier to investigate the temperature-dependence and functional integrity of the graft postperfusion [29, 30]. Recently, Gringeri et al. applied SNMP without an oxygen carrier to liver grafts for 120 minutes in a porcine model of autotransplantation . Moreover, Tolboom et al. calculated the oxygen use of livers during machine perfusion at 20°C and 30°C with an oxygen carrier and concluded that though livers consume more oxygen at 37°C, the addition of the oxygen carrier was not required for adequate liver metabolism at room temperature . Currently, no studies are underway to investigate the validity of SNMP without an oxygen carrier in a long-term transplantation survival model. Moreover, it has not been shown whether, similarly to HMP [3, 8, 33] and NMP [13, 15, 16], ischemically damaged organs can be reclaimed and successfully transplanted using this technique.
In this study, we used a SNMP protocol that did not involve dialysis, oxygen carriers or donor pretreatments such as anticoagulants or preconditioning, enabling high-fidelity simulation of DCD. We report herein for the first time that both fresh and warm ischemic livers can be successfully treated by using such a system and transplanted into healthy recipients with good 30-day survival.
Male Lewis rats weighing 250 to 300 g (Charles River Laboratories, Boston, MA, USA) were used for transplantation. The animals were maintained in accordance with National Research Council guidelines, and the experimental protocols were approved by the Subcommittee on Research Animal Care, Committee on Research, Massachusetts General Hospital (Boston, MA, USA). The animals were divided into four groups: the UW-Control group (n = 4, 3 hours at 4°C, SCS preservation and subsequent transplantation), the WI group (n = 4, 60 minutes at 34°C, warm ischemia and transplantation), the Fresh-SNMP group (n = 6, 3 hours of SNMP and transplantation) and the WI-SNMP group (n = 6, 60 minutes at 34°C, warm ischemia followed by 3 hours of SNMP and then transplantation).
Liver procurement (all groups)
All surgical procedures were performed under aseptic conditions. Each animal was weighed and anesthetized with isoflurane (Baxter, Deerfield, IL, USA). The right phrenic vein was ligated, and the liver was freed of its surrounding ligaments. The infrahepatic vena cava (IHVC) was mobilized and elongated by ligation and dissection of the adrenal vein, lumbar plexus and right renal vein. The bile duct was cannulated (SURFLO 28-gauge polyethylene stent; Terumo Medical Corp, Somerset, NJ, USA) and dissected. The portal vein (PV) was mobilized, and its most proximal side vessels (gastroduodenal and splenic veins) were ligated and cut. The hepatic artery was ligated and dissected. The PV and IHVC were cross-clamped, marking the start of ischemic time, and the liver was excised from its recess and weighed. Cuffs fashioned from 16- and 14-gauge intravenous catheters (Becton Dickinson, Franklin Lakes, NJ, USA) were applied to the PV and IHVC as described by Delrivière et al. . The suprahepatic vena cava (SHVC) was tailored for a sutured anastomosis.
Static cold storage group
SCS was performed using UW solution (CoStorSol; Preservation Solutions, Inc, Elkhorn, WI, USA). No other preservatives, pharmaceuticals or anticoagulants were used during SCS. Rat livers were procured, prepared for transplantation and cold-stored for 3 hours (matching the established machine perfusion time), then transplanted as described below.
DCD model (WI group and WI-SNMP group)
After procurement, the liver was placed without flushing into a chamber filled with saline and maintained at 34.0 ± 0.1°C for 60 minutes. Ex vivo warm ischemia has the advantage of temperature control and has been shown to be a relevant and severe model of warm ischemia .
Subnormothermic machine perfusion (Fresh-SNMP group and WI-SNMP group)
Perfusate samples collected every 5 minutes for the first 20 minutes and on the half hour thereafter were tested for alanine aminotransferase (ALT) and aspartate aminotransferase (AST) using InfinityTM ALT(GPT) and AST(GOT) liquid stable reagent kits (Cellomics/Thermo Electron, Pittsburgh, PA, USA). Every 30 minutes PV inflow and inferior vena cava (IHVC) outflow were sampled and analyzed using a Bayer Rapidlab 845 blood gas analyzer (Siemens Medical Solutions, Malvern, PA, USA) to record the partial pressure of oxygen (pO2). Those values were used to calculate the oxygen uptake rate (OUR) as detailed previously . Bile was collected, quantified and reentered into the system.
Tissue ATP content
Livers (n = 3 per time point) were flash-frozen in liquid nitrogen directly after procurement (Fresh-SNMP group), after 60 minutes of warm ischemia (WI group/WI-SNMP group) and after 1, 2 and 3 hours of SNMP (Fresh-SNMP group/WI-SNMP group). Various sections (n = 7 ± 1) were sampled per liver. Cellular ATP levels were measured using the ApoSENSOR ATP Luminescence Assay Kit (BioVision Inc, Milpitas, CA, USA). The results were normalized for protein content using a Bradford assay (Fisher Scientific, Pittsburgh, PA, USA).
Orthotopic liver transplantation
Surgery was performed by TB in all cases according to the technique invented by Kamada  and described in detail by Delrivière et al. . During the final 20 minutes of perfusion, the recipient hepatectomy was carried out up to the start of the anhepatic phase. At the end of machine perfusion, the liver was weighed and flushed with sterile PBS containing 10 U/ml heparin. The recipient hepatectomy was completed, and the donor graft was transplanted, starting with the SHVC anastomosis using a Prolene 7–0 polypropylene suture (Ethicon, Inc, Somerville, NJ, USA). Portal reperfusion occurred 13 to 17 minutes after the start of anhepatic time and approximately 20 to 30 minutes after the end of machine perfusion. Anastomoses of the IVC and bile duct were completed, the animal was administered 1 to 2 ml of sterile PBS intravenously and the abdomen was irrigated with warm PBS and closed. The skin was closed, and the animal was placed under a heat lamp to recover.
Blood samples (about 80 μl) were taken from transplantation recipients by way of tail-snip hourly after the procedure for 3 hours, then daily for 7 days after the surgery and finally after 30 days. The blood samples were analyzed using a Piccolo xpress blood chemistry analyzer and a metabolic panel (Abaxis North America, Union City, CA, USA). The animals were inspected regularly for signs of jaundice or infection and were killed 30 days posttransplantation. The liver was mobilized, resected and prepared for histology.
Livers from each group were sacrificed for histological analysis. Samples were incubated overnight in formalin and embedded in paraffin, then processed and stained with H & E and terminal deoxynucleotidyltransferase 2′-deoxyuridine 5′-triphosphate nick-end labeling (TUNEL). As a negative control for cellular damage, stains derived from fresh liver tissue were used, in comparison to samples from the preserved, treated and transplanted groups. As a positive control for apoptosis, the TUNEL stains from the WI group were used because it has been reported that tissue from warm ischemic liver can exhibit extensive cellular degeneration and apoptosis [16, 38, 39].
A paired Student’s t-test was applied to compare values for each parameter independently. The results were deemed statistically significant if time point comparisons had a P-value <0.05. Statistical significance is indicated in each figure where applicable.
Subnormothermic machine perfusion
The pO2 in the perfusate was used to calculate the OUR (Figure 5C). The OUR was stable during SNMP, without differences between the groups. The pO2 in the outflow remained above 200 mmHg, indicating that the pO2 in the inflow perfusate was sufficient for hepatic oxygenation.
As a determinant of synthetic function, bile production is a prerequisite for successful transplantation . In vivo bile flow ranges from 15 to 20 μl/g liver in healthy rats. During SNMP, bile flow averaged about 10 μl/g liver. It seems likely that the lower bile flow is due to the lower temperature during SNMP. Bile production was not significantly higher in the Fresh-SNMP group than in the WI-SNMP group (Figure 5D).
Posttransplantation blood analysis
TB increased mildly over the first 7 days, which was more apparent in the Fresh-SNMP group (Figure 6C). At the end of the 30-day period, TB was increased to approximately 2 mg/dl in all surviving animals. BUN levels were stable during the postoperative phase and throughout the 30-day period (Figure 6D).
SCS is practical and cost-effective but does not allow the resuscitation of damaged grafts, which is vital to the extended incorporation of marginal organs . In addition, evaluation of organ quality during SCS is limited to subjective assessments. To combine the simplicity of SCS with the therapeutic impact of machine perfusion, we established a practical SNMP system that requires no temperature control, donor blood or artificial oxygen carriers. In an analogous study by our group, this system was used to optimize livers for hepatocyte isolation (ML Izamis, personal communication). Herein we have shown that fresh livers could be preserved for 3 hours and transplanted successfully, with survival rates and secondary outcomes equivalent to conventional cold storage parameters. Furthermore, we used this system to treat livers with sustained warm ischemia, which were then successfully transplanted.
In the animal from the WI-SNMP group that died within 72 hours after transplantation, the highly elevated levels of AST and ALT point to primary nonfunction. During treatment, the portal pressure of this liver was significantly increased, suggesting a correlation between high vascular resistance and graft failure. Although this is a singular finding in our study, the principle of vascular resistance as a predictor of graft viability has been suggested elsewhere . We conclude that high vascular resistance during machine perfusion could serve as a warning sign indicative of a mechanical or physiological problem. However, specific research into this phenomenon should be continued on a broad scale.
Cumulative ALT levels during SNMP were highest in the WI-SNMP group. After 1 hour of SNMP, these levels reached a plateau, signifying that the hepatocyte damage caused by the WI period had been contained. Warm ischemic and fresh livers consumed similar amounts of oxygen during the SNMP period. We observed that oxygen consumption rates were not affected by the organ’s condition. We can conclude that resuscitation of a warm ischemic liver to a transplantable state does not require more oxygen than the SNMP preservation of a fresh liver. In addition, the high residual oxygen content in the outflow perfusate indicates that none of the livers exhausted the supply of oxygen. Moreover, the ATP recovery of warm ischemic livers beyond fresh levels also points to adequate aerobic respiration in the ischemic grafts. Thus we conclude that our system did not require an oxygen carrier.
The WI-SNMP group showed higher levels of transaminase leakage posttransplantation, indicating that although the livers were reclaimed adequately for recipient survival, there was still regenerative potential that remained unfulfilled. These observations are concurrent with the body weight trends: animals in all groups initially lost weight, but the UW-Control and Fresh-SNMP groups regained that weight sooner than the WI-SNMP group. Possibly, SNMP was responsible for only part of the restoration process and the remainder was carried out at the expense of the recipient. Another explanation may be that certain specific effects of warm ischemia were not mitigated by the SNMP and could be only restored by the recipients themselves, which in this case were young and healthy animals. As rats have great regenerative capacity, this phenomenon may be of greater consequence in a higher-order species. Although patients with end-stage liver disease are often in a fragile physical state, it has been shown that human livers from patients with elevated transaminases  or sustained warm ischemia  can be successfully transplanted. Whether our findings call for a longer or enhanced SNMP period will be clarified by continued investigation.
In this study, we have established a protocol for SNMP, without oxygen carriers and without temperature control, that uses oxygenated and supplemented cell culture medium as perfusate. We applied this system to demonstrate feasibility against the current gold standard for the preservation of fresh organs, as well as the potential for regeneration of DCD grafts.
Funding from the National Institutes of Health (R01DK096075, R01EB008678, R00DK080942) and the Shriners Hospitals for Children are gratefully acknowledged. This work was funded in part by Health Resources and Services Administration contract 234-2005-37011C. The content is the responsibility of the authors alone and does not necessarily reflect the views or policies of the US Department of Health and Human Services, nor does mention of trade names, commercial products or organizations imply endorsement by the US government. TB was supported in part by the Michaël-van Vloten Foundation. The authors gratefully acknowledge Dr Charles Lee for instruction on orthotopic liver transplantation in rats.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.