Rescue of the Cold Preserved Rat Liver by Hypothermic Oxygenated Machine Perfusion



The aim of the study was to investigate whether hypothermic oxygenated liver perfusion after cold liver preservation resuscitated metabolic parameters and whether this treatment had a benefit for liver viability upon reperfusion.

We preserved rat livers either by cold storage (UW) for 10 h, or by perfusion for 3 h (oxygenated modified UW) after 10 h cold storage. We assessed viability of livers after preservation and after ischemic rewarming + normothermic reperfusion ex vivo. Ten hour cold storage reduced mitochondrial cytochrome oxidase and metabolically depleted the livers. Oxygenated perfusion after cold storage resulted in uploaded cellular energy charge and oxidized mitochondrial cytochrome oxidase. Reperfusion after 10 h cold storage increased formation of superoxid anions, release of cytosolic LDH, lipid peroxidation, caspase activities and led to disruption of sinusoidal endothelial cells. In contrast, reperfusion after 10 h cold storage + 3 h hypothermic oxygenated perfusion resulted in no changes of lipid peroxidation, bile flow, energy charge, total glutathione, LDH release and of caspase activation, as compared to fresh resected livers.

This study demonstrates, that a metabolically depleted liver due to cold storage can be energy recharged by short-termed cold machine perfusion. The machine perfused graft exhibited improved viability and functional integrity.


During cold storage of livers, a number of important cellular components gradually deteriorate with a loss of energy stores, a drastic decrease in cellular glycogen and ATP contents and with an alteration in biochemical functions and cellular architecture (1). Concurrently, the cellular and mitochondrial redox status is affected resulting in a change of the electrochemical potential of cytochrome c and cytochrome oxidase during cold storage of livers (2). Studies on the cellular architecture have clearly demonstrated injury to the sinusoidal endothelial cells (SEC), which become detached and rounded during cold ischemia (3,4). The degree of endothelial cell detachment correlates with the duration of cold ischemia (3). After cold storage, rewarming and reperfusion are regarded as critical points for the viability of these cells as well as for the whole graft. Reoxygenation leads to oxygen free radical production and to disruption of the sinusoidal endothelial cell wall, which can be mediated synergistically by platelets and leukocytes (3,5). The relationship between metabolic depletion of hepatocytes and SEC damage is still not fully understood.

Experiments on the isolated perfused rat liver (IPRL) suggested that a corrected ATP loss before rewarming could help the liver to overcome reperfusion injury (4,6). Several groups could restore the energy reserve and components of the cellular redox system through the use of machine liver perfusion systems under either hypothermic (7,8) or normothermic (9,10) conditions, but these models have been challenged and found unattractive by many due to the complexity of the proposed system and the need to run machine perfusion during cold or warm preservation. However, if a machine perfusion system could be applied to the donor liver graft only for a limited time during the preparation of the recipient, the procedure could rapidly gain wide acceptance in the clinical setting. Therefore, we designed a study to evaluate such a novel strategy in a rat model. We tested whether an energy-depleted liver can be “recharged” by short-termed cold machine perfusion.

Material and Methods


Male Brown Norway rats (250–300 g of body weight) were maintained on standard laboratory diet and water ad libitum according to the law of Animal Health Care. The animals were anesthetized with ether and the liver was exposed through a midline incision. The liver was cannulated, flushed and resected as previously described (6,11).

Study design

The harvested livers were divided into four groups:

  • • cold (4°C) storage over a period of 10 h (n = 8).
  • • cold storage over a period of 10 h followed by machine perfusion (hypothermic oxygenated perfusion: HOPE) over a period of 3 h (n = 8).
  • • 10 h cold storage followed by 20 min of poikilothermic ischemic rewarming and 40 min of acellular normothermic (37°C) perfusion (n = 12).
  • • 10 h cold storage + 3 h hypothermic oxygenated perfusion (HOPE) followed by 20 min of poikilothermic ischemic rewarming and 40 min of acellular normothermic (37°C) perfusion (n = 12).

Fresh resected livers served as controls (n = 6).

Hypothermic oxygenated liver perfusion was performed according to previous experiments with an oscillating perfusion technique (6) via the portal vein. The perfusion chamber and ancillary devices were kept in an industrial type refrigerator with the temperature maintained between 3 and 5°C. The computer controlled machine perfusion was activated intermittently, resulting in an oscillating flow (2.75 mL/min–3.25 mL/min) with a low pressure of 4.4 ± 0.5 mm Hg. The perfusate was recirculated (perfusate volumen 450 mL) and oxygenated by passing through a thin silicon tube under exposure of 95% oxygen and 5% CO2. The pO2 of the cold perfusate during hypothermic perfusion was at 313.4 ± 24.6 mm Hg.

A modified UW solution for machine perfusion was used as described previously, omitting hydroxyethyl starch to reduce viscosity (6). Cold storage was performed with standard UW solution. The mean liver weight of all livers was 10.8 ± 1.4 g after harvest.

Rewarming and reperfusion

Between preservation and reperfusion a 20-min interval of poikilothermic ischemic rewarming was chosen to simulate graft transplantation. During this ischemic rewarming the liver temperature rose in all experiments from 4 ± 0.3°C to 25 ± 0.4°C prior to reoxygenation.

After the ischemic rewarming period, livers in both groups were perfused with non-recirculating oxygenated (Carbogen) Krebs-Henseleit bicarbonate buffer (flow rate 15 mL/min) for 40 min at 37°C (8). The pO2 during reperfusion increased to 534 ± 46.3 mm Hg. Separate experiments were used for measuring superoxide anion formation upon reperfusion by adding 50 mM ferricytochrome c to the perfusate during reperfusion (2).

Assessment of hepatocyte cell injury

Glykolytic metabolites:  Liver glycogen, ATP, ADP, AMP, energy charge and lactate were determined according to earlier studies (6,11): for analysis of nucleotides livers were homogenated (1:10) in cold 4% HClO4 with a Potter teflon homogenizer. After centrifugation and pH adjusting (pH 8.5) ATP was measured by UV spectroscopy (340 nm) with hexokinase and glucose-6-phosphate dehydrogenase. For ADP and AMP measurement we adjusted the pH of liver homogenates at 7.2 and determined ADP, AMP by UV spectroscopy (340 nm) with pyruvat kinase and myokinase. Glycogen was determined after homogenization of liver samples in phosphated buffered saline (1:5) and measuring the difference of glucose in samples treated or not treated with amyloglucosidase.

Apoptosis, lipid peroxidation:  DNA fragmentation, caspase activities, total glutathione and lipidperoxidation were determined according to earlier studies (2): for detection of DNA fragmentation we used a cell Death ELISA (Roche molecular biochemicals). Caspase activities were measured by colorimetric assays (R&D systems). The amount of thiobarbituric acid reactive substances (TBARS) were determined after homogenization of liver samples (1:5) with cold phosphated buffered saline + 0.01 mM butylated hydroxy toluene. After centrifugation samples were incubated with 8.1% SDS, 20% acetic acid and 0.8% thiobarbituric acid for 1 h at 95°C. The resulting TBARS were extracted with butanol/pyridine and measured at 532 nm.

Superoxide anion/LDH release:  Perfusate outflow during reperfusion was collected every minute and superoxide anion formation was determined by the reduction of ferricytochrome c included in the perfusate (2). LDH release were determined according to previous studies (8).

Mitochondrial cytochrome redox state:  The change in absorbance of liver tissue was measured by a tissue spectrophotometer (O2C, LEA Medizintechnik, Germany) during hypothermic oxygenated perfusion (HOPE) as well as during normothermic reperfusion. A flat signal probe (LEA Medizintechnik, Germany,) was mountained under the central rat liver lobe and introduced monochromatic xenon light in the liver tissue. Light absorbance was detected by another transducer and coupled to a photometric device. A reproducible monitoring of absorbance values of the liver tissue within 500 and 620 nm (depth of invasion: 2 mm) was achieved. The signal was transformed using arbitrary units and recorded on a PC (LEA Medizintechnik, Germany). The resulting spectra allowed visualization of the redox state of mitochondrial cytochrome oxidase (2) (Figure 1).

Figure 1.

Transhepatic spectroscopy. Transhepatic spectroscopy during cold and warm perfusion enabled monitoring of mitochondrial cytochrome redox state: changes in light absorbance at 605 nm (cytochrome oxidase peak) correlate with the redox state of cytochrome aa3. During ischemic rewarming after CS + HOPE cytochrome oxidase redox state switched from oxidized to reduced.

Reversed transcriptional PCR–analysis (RT PCR):  Total RNA was isolated according to previous studies (2). Primers specific for rat genes were designed using the Primer 3 Software (Whitehead Institute, Boston, USA) and synthesized by MWG Biochemical, Germany. Glyceraldehyd–3-phosphate dehydrogenase (GAPDH) was used as house keeping gene.

The following primers were used:

  • - TNFα forward: 5′-ATGTGGAACTGGCAGAGGAG-3′ and TNFα reverse: 5′-GGCCATGGAACTGATGAGGA-3′ (expected product 200 bp)
  • - NFκB forward: 5′-CGATCTGTTTCCCCTCATCT-3′ and Nf κB reverse: 5′-ATTGGGTGCGTCTTAGTGGT–3′ (expected product 174 bp)
  • - procaspase 9 forward: 5′-TTTGAGGTGGCCTTCACTTC-3′ and procaspase 9 reverse: 5′-CAGGAACCGCTCTTCTTGTC-3′ (expected product 200 bp)
  • - c-JNK forward: 5′–ACAAGCGGATCTCTGTGGAC-3′ and JNK reverse: 5′-TTTCACCCCATTCTTGCTTC-3′ (expected product 197 bp)
  • - MIP-2 forward: 5′-AGGGTACAGGGGTTGTTGTG-3′ and MIP-2 reverse: 5′-TTTGGACGATCCTCTGAACC-3′ (expected product 204 bp)
  • - BAK-forward: 5′-CCTGCTAACCCTGAGATGGA-3′ and bak reverse: 5′-AATAGGCTGGAGGCGATCTT-3′ (expected product 194 bp)
  • - HGF-forward: 5′-CTTCCTGTCACCATCCCCTA-3′ and hgf reverse: 5′-AAAGGCCTTGCAAGTGAATG-3′ (expected product 198 bp)

All sequences were compared to the complete rat Genbank library to ensure that each primer was unique to its intended target and that areas of sequence polymorphism were avoided. Equivalent levels of starting cDNA were used for amplification of PCR products. Master reagent mix consisted of 4 μL dNTP-Mix (1.25 mM), 1 μL specific primer (50 pmol/μL), 1 μL Taq-polymerase (Boehringer, 5 U/μL) and 10 μL PCR buffer (1.5 mM) were heated to 93°C over 4 min and cycled (n = 34) for 1 min at 93°C, for 1 min at 59°C and 1 min at 72°C. After cycling the samples were heated to 72°C for 7 min and then cooled to 4°C. The resulting PCR products were separated by agarose gel electrophoresis and by staining with ethidium bromide without quantification

Statistical analysis were performed using the non-parametric Mann-Whitney-Wilcoxon U-test or two way Analysis of variance (ANOVA) (Graph Pad Prism, version 3.0). Results are given as mean ± standard deviation (SD)(Tables 1, 2, Figures 2, 4–6). A p-value below 0.05 was regarded as significant.

Table 1. Biochemical parameters after preservation
 Fresh resected liver10h CS10h CS + 3h HOPEp*
  1. Values are given as means ± SD.

  2. *CS vs CS + HOPE.

Glycogen[μmol/g ww]261.4 ± 47.3176.4 ± 28.596.4 ± 17.90.001
[% of fresh resected liver]  67.5 36.9  
Lactate[μmol/g ww]1.21 ± 0.49.38 ± 1.50.61 ± 0.30.001
[% of fresh resected liver]  775.2 50.4  
Energy charge[% of fresh resected liver]0.76 ± 0.040.17 ± 0.030.79 ± 0.050.002
22.4 103.9  
TBARS (MDA)[nmol/ g ww]194.4 ± 38.5266.2 ± 35.4227.5 ± 67.60.272
[% of fresh resected liver]  136.9 117.0  
Total glutathione[μmol/g ww]3.29 ± 0.22.64 ± 0.52.53 ± 0.30.485
[% of fresh resected liver]  80.2 76.9  
DNA fragments[enrichment factor]10.55 ± 0.150.56 ± 0.20.635
Caspase 3 activity[fold increase]11.1 ± 0.50.9 ± 0.40.525
Table 2. Biochemical parameters after rewarming and reperfusion
 Fresh resected liver10h CS10h CS + 3h HOPEp**
  1. Values are given as means ± SD.

  2. *pilot experiments, freshly resected livers subjected to rewarming and reperfusion, see result section.

  3. **CS vs CS + HOPE.

Glycogen[μmol/g ww]261.4 ± 47.315.7 ± 8.831.9 ± 18.20.189
[% of fresh resected liver]  6.0 12.2  
Lactate[μmol/g ww]1.21 ± 0.41.81 ± 0.941.27 ± 0.260.485
[% of fresh resected liver]  149.6 104.9  
Energy charge[% of fresh resected liver]0.76 ± 0.040.36 ± 0.080.58 ± 0.080.01
47.4 76.3  
Bile flow[μL/h]*65.3 ± 14.895.4 ± 20.20.003
[% of fresh resected liver]  54.0 78.9  
TBARS (MDA)[nmol/ g ww]194.4 ± 38.5458.8 ± 58.9254.4 ± 37.40.003
[% of fresh resected liver]  236.0 130.9  
Total glutathione[μmol/g ww]3.29 ± 0.21.38 ± 0.52.64 ± 0.40.001
[% of fresh resected liver]  41.9 80.2  
Total LDH release[U/L]*143.3 ± 31.667.6 ± 13.20.008
[% of fresh resected liver]  192.8 90.9  
DNA fragments[enrichment factor]12.71 ± 0.80.63 ± 0.20.0007
Caspase 3 activity[fold increase]14.7 ± 1.31.2 ± 0.50.004
Caspase 8 activity[fold increase]10.8 ± 0.41.1 ± 0.40.419
Caspase 9 activity[fold increase]13.5 ± 1.20.92 ± 0.50.002
Figure 2.

Redox state of rat liver cytochrome aa3 (ΔA 605 nm) after 10 h cold storage (CS). Cytochrome oxidase (cytochrome aa3) was oxygenated during 3 h of hypothermic oxygenated perfusion after cold storage (CS + HOPE). During reperfusion cytochrome oxidase was fully oxidized after CS + HOPE and only partially oxidized after cold storage alone.

Figure 4.

O2 -consumption during reperfusion after preservation and ischemic rewarming. Oxygen consumption was significantly higher during acellular ex vivo reperfusion in cold stored livers (CS) as compared to cold stored and machine perfused livers (CS + HOPE)

Figure 5.

Superoxide anion formation during reperfusion after preservation and ischemic rewarming. Superoxide anion formation, detected by the reduction of 50 μM ferricytochrome (ΔA 555 nm) included in the perfusate, was increased during reperfusion after cold storage (CS) as compared to cold storage followed by oxygenated machine perfusion (CS + HOPE).

Figure 6.

LDH release during reperfusion after preservation and ischemic rewarming. LDH release was increased during reperfusion after cold storage (CS) as compared to cold storage followed by oxygenated machine perfusion (CS + HOPE).


Are glycolytic metabolism and mitochondrial electron transfer affected during hypothermic oxygenation after cold storage?

We (8,11) and others (12) have previously shown that 10 h cold storage causes significant depletion of glycolytic metabolites. To test whether energy stores could be rapidly restored in cold preserved rat livers, we performed an experiment in which livers were stored in the cold for 10 h followed by 3 h of hypothermic machine perfusion. Glycogen, one of the primary energy stores, decreased to 68% of the baseline value during cold preservation, and further decreased following oxygenated perfusion (p < 0.01, Table 1). Lactate accumulated during cold preservation, but was metabolized to physiological levels by the machine perfusion system (p < 0.01, Table 1). Similarly, cellular energy charge was depleted during preservation but was fully restored by oxygenated machine perfusion (p < 0.01, Table 1). Cytochrome oxidase redox status, as evaluated by transhepatic measurement of light absorbencies at 605 nm (2), was reduced at the end of cold preservation (2281 ± 31 A.U., Figure 2), but was reoxygenated by machine perfusion to 2519 ± 45 A.U. (Figure 2).

We conclude from this set of experiments that cellular energy is mainly synthesized by glycolysis during cold liver storage resulting in reduced mitochondrial cytochrome oxidase and lactate accumulation. Perfusion of the liver after cold storage in the HOPE system enabled to switch the mitochondrial redox state with a return to physiologic cellular energy charge and oxidized mitochondrial cytochrome oxidase.

Does hypothermic oxygenated machine perfusion induce oxidative stress prior to normothermic reperfusion?

Although oxygenated machine perfusion has a beneficial effect on the liver by reloading the cellular energy, oxygenation may concurrently cause oxidative stress with deleterious effects on liver cells. Therefore, we determined TBARS in liver tissue, an established indicator of lipid peroxidation. We failed to detect any significant changes of TBARS either after cold storage or after hypothermic oxygenated machine perfusion indicating that oxidative stress was not a relevant source of tissue injury under hypothermic conditions. This observation was further supported by glutathione levels, which similarly were not affected by this procedure (Table 1).

Is there evidence for induction of apoptosis during hypothermic oxygenated machine perfusion?

Previous studies in cultured hepatocytes and endothelial cells demonstrated that cold ischemia followed by cold oxygenation caused a significant induction of apoptosis (13). We tested DNA fragmentation and caspase activity, two parameters indicative of apoptosis, and found no significant changes during cold storage and HOPE of the whole graft (Table 1). Electron microscopy showed well- preserved hepatocytes and sinusoidal endothelial cells (Figure 3). LDH release during cold machine perfusion was 9.8 ± 2.4 U/l.

Figure 3.

Figure 3.

Transmission electron microscopy of liver tissue after preservation. A: 10 h cold storage with UW solution resulted in well-preserved hepatocytes and sinusoidal endothelial cells. B: hypothermic oxygenated machine perfusion after cold storage did not change hepatocyte or sinusoidal structure.

Figure 3.

Figure 3.

Transmission electron microscopy of liver tissue after preservation. A: 10 h cold storage with UW solution resulted in well-preserved hepatocytes and sinusoidal endothelial cells. B: hypothermic oxygenated machine perfusion after cold storage did not change hepatocyte or sinusoidal structure.

These results support the concept that major tissue injury is not induced during cold oxygenated perfusion with the described technique.

Does hypothermic oxygenated machine perfusion improve metabolic parameters during normothermic reperfusion ?

As demonstrated above, oxygenated machine perfusion is highly effective in restoring the energy charge during cold preservation. However, as tissue damage predominantly occurs during rewarming and reperfusion, we tested the effects of hypothermic oxygenated machine perfusion (HOPE) on livers subjected to normothermic oygenated reperfusion in the IPRL model: the livers were stored in the cold for 10 h followed by 20 min of ischemic rewarming and 40 min of reperfusion. This scenario was chosen to best mimic the clinical situation of transplantation.

The same parameters as studied above were evaluated. Reperfusion had a dramatic effect on glycogen contents, which deteriorated to 6% of the baseline liver values (Table 2). Oxygenated machine perfusion did not prevent glycogen depletion (Table 2). Lactate levels dropped to low physiological values in both experimental groups, but the energy charge improved from 47.4% to 76.3% of the baseline with the use of HOPE (p = 0.01) (Table 2).

In pilot experiments we tested bile flow in freshly resected livers subjected to rewarming and reperfusion: bile flow was 120.9 ± 17.4 μL/h (n = 5). When compared to unpreserved livers (cold ischemia time < 30 min), bile flow decreased by about half in livers preserved for 10 h, but decreased only by 21% when the same livers were perfused in the HOPE system. (p = 0.003) (Table 2).

Transhepatic measurement of cytochrome oxidase redox state during reperfusion showed that oxygenated machine perfusion after cold storage led to fully oxidized cytochrome aa3. After cold storage alone without hypothermic oxygenated perfusion only partial recovery of the cytochrome oxidase system occurred (Figure 2). Simultaneously we found lower O2 consumption with the use of the HOPE apparatus (Figure 4).

We concluded from this set of experiments that dysfunction of mitochondrial electron transfer occurs during reperfusion after cold storage that leads to increased oxygen uptake in spite of low electron transfer at stage 3 of the respiratory chain.

Does HOPE decrease oxidative stress during normothermic reperfusion?

After rewarming and normothermic reperfusion of cold stored livers we found an increase of TBARS and depletion of total glutathione as compared to fresh resected livers (Table 2). In contrast, cold stored livers treated by oxygenated machine perfusion showed normal levels of lipid peroxidation and glutathione (Table 2). As an additional and more independent parameter for oxidative stress we determined the formation of superoxide anions: hypothermic machine perfusion after cold storage decreased superoxid anion release during reperfusion (Figure 5). These results suggest that if cold stored livers are hypothermically machine perfused prior to normothermic reperfusion they exhibit a significantly reduced oxidative stress.

Does HOPE improve graft viability after reperfusion?

Survival of a graft after ischemia is dependent on the extent of injury occurring after reperfusion. Therefore LDH release, caspase activities, DNA fragmentation and electron microscopy were determined after normothermic reperfusion in the IPRL: in pilot experiments LDH release in freshly resected livers accumulated to 74.3 ± 15.6 U/L (n = 5). Cold storage significantly increased LDH release (Table 2). In contrast, cold storage + HOPE resulted in LDH levels similar to reperfused fresh resected livers (Table 2) (Figure 6).

Cold storage also increased caspase 3 and caspase 9 activities, caspase 8 activity was unaffected (Table 2). HOPE after cold storage prevented the activation of caspase 3 and caspase 9 (Table 2) and inhibited increase of histone associated DNA fragments in contrast to cold storage alone (Table 2).

Electron microscopy after rewarming and reperfusion of 10 h cold stored livers showed significant cell injury such as cell blebbing, loss of microvilli, the absence of endothelial cell process and signs of apoptosis (Figure 7A). In contrast, sinusoidal cells and hepatocytes showed less injury when machine perfused (CS + HOPE) prior to rewarming and reperfusion (Figure 7B).

Figure 7.

Figure 7.

Transmission electron microscopy of liver tissue after preservation, ischemic rewarming and reperfusion. A: After 20 min ischemic rewarming and 40 min reperfusion the 10 h cold stored livers (CS) demonstrated severe cell blebbing and sinusoidal detachment (arrows). B: 10 h cold stored + 3 h machine perfused livers (CS + HOPE) showed enlarged sinusoids but intact endothelial microvilli (arrows) after 20 min ischemic rewarming and 40 min reperfusion. K: Kupffer cell, H: hepatocyte.

Figure 7.

Figure 7.

Transmission electron microscopy of liver tissue after preservation, ischemic rewarming and reperfusion. A: After 20 min ischemic rewarming and 40 min reperfusion the 10 h cold stored livers (CS) demonstrated severe cell blebbing and sinusoidal detachment (arrows). B: 10 h cold stored + 3 h machine perfused livers (CS + HOPE) showed enlarged sinusoids but intact endothelial microvilli (arrows) after 20 min ischemic rewarming and 40 min reperfusion. K: Kupffer cell, H: hepatocyte.

Finally, cellular components known to respond to tissue injury were determined by RT PCR. Selection of PCR products was based on recent studies by Bradham et al. and Shinoda et al., who demonstrated increased transcription of tumor necrosis factor α (TNFα), nuclear factor κB (NFκB) and c-Jun N-terminal kinase (c-JNK) during early reperfusion after orthotopic rat liver transplantation (14,15). Upregulation of chemokines including macrophage inflammatory protein-2 (MIP-2) and hepatocyte growth factor (HGF) was also described after cold liver ischemia reperfusion, most probably depending on TNFα release (16,17). Transcription of procaspase 9 is known to be related to increased apoptotic cell death via an intrinsic apoptotic pathway in the cold ischemic liver (18). Finally, it has been shown that decreased levels of the Bcl-2 antagonist killer (BAK) in the outer mitochondrial membrane is associated with the inability of truncated BID (tBID) to induce cytochrome c release from mitochondria (19).

Ten-hour cold storage led to detectable transcripts of the above mentioned apoptotic proteins (procaspase 9, BAK), inflammatory mediators (TNFα, NFκB, cJNK), chemokine (MIP-2) and growth factor (HGF) during rewarming and reperfusion (Figure 8). In contrast, oxygenated machine perfusion after 10 h cold storage (CS + HOPE) prevented induction of transcription of all these proteins during the cold machine perfusion period as well as during the following rewarming and reperfusion (Figure 8).

Figure 8.

RT PCR analysis after preservation, ischemic rewarming and reperfusion. Specific transcripts of procaspase 9, NFκB, TNFα, BAK, c-JNK, HGF and MIP-2 were detectable after 10 h cold storage, 20 min ischemic rewarming and 40 min of reperfusion (CS). In contrast, no transcripts were detectable after rewarming and reperfusion of hypothermic oxygenated livers following cold storage (CS + HOPE).

We conclude that both types of cell injury necrosis and apoptosis were significantly decreased by a treatment of HOPE after cold storage.


There are three main findings in this study. First, we demonstrated that cellular energy recharge of cold stored (10 h) livers is feasible within 3 h of cold oxygenated machine perfusion and without induction of apoptosis or increase in lipid peroxidation. Second, we found that oxidative stress during normothermic reperfusion was decreased when cold stored livers were hypothermically machine perfused. Finally, we demonstrated that cell death decreased during normothermic reperfusion by the use of hypothermic oxygenated machine perfusion.

The benefit of oxygenated machine perfusion applied during the entire duration of storage is well established (20–22). However, this approach has not gained wide popularity due to its complexity and the need for transportable perfusion devices. Considering practical aspects, the stored organ usually is implanted within three additional h after arrival, while the recipient is being prepared. This would give the opportunity to intervene during this period to minimize graft injury after reperfusion.

Our strategy of 3 h cold perfusion would fit with such a design, and may be feasible in a clinical setting. Furthermore, this procedure might help against vascular endothelial damage previously reported as a decisive factor in early tissue damage (23,24).

The selection of 3 h of hypothermic oxygenated perfusion is based on previously unpublished results: 1 h HOPE after 10 h cold storage resulted in minor ATP and energy charge reloading after 10 h cold storage: 0.82 ± 0.2 μmol ATP/g liver ww and 0.54 ± 0.1 energy charge (n = 3). After 2 h HOPE we achieved an increase to 1.49 ± 0.4 μmol ATP and 0.61 ± 0.2 energy charge (n = 3). After 3 h of cold oxygenated perfusion ATP increased to 1.81 ± 0.3 μmol ATP and the energy charge to 0.79 ± 0.05 (Table 1), which was not significantly different from fresh resected livers (Table 1). Therefore we decided to choose a 3 h period of cold oxygenated perfusion after 10 h cold storage in order to get physiologic levels of cellular energy charge prior to rewarming and reperfusion.

Although only minor morphological alterations are known during the first 10 h of cold rat liver storage with UW solution (25), a number of studies clearly demonstrated severe metabolite depletion already during short periods of cold static preservation, for example, significant ATP depletion and energy charge depletion of hepatic tissue were reported after only 1 h of cold storage (6). The amount of cellular free chelatable iron in isolated rat hepatocytes increased after 30 min of cold preservation and stayed at this level for 6 h (13). Cold preservation of rat livers in UW solution for as few as 4 h led to a significant loss of hepatocyte viability on re-oxygenation (26). In line with these findings, Vreugdenhil et al. (27) reported a significant impairment of protein synthesis after 4 h of cold storage. Isolated mitochondria from rat liver hepatocytes also appeared to be compromised by short periods of liver storage in UW solution as shown by Ohkohchi et al. (28), who described impaired mitochondrial function within 6 h and 12 h of liver preservation. In addition, a recent study indicated that a 10 h preservation period impaired the regenerative ability of the recipient rat liver (29). Our own results demonstrated mitochondrial dysfunction after 10 h cold storage (2).

The current study showed that restoration of cellular energy charge and lactate in cold preserved liver with only 3 h of hypothermic oxygenated perfusion prior to implantation was highly successful. (Table 1). Glycogen was depleted during HOPE because of aerobic glycolysis and ongoing mitochondrial respiration despite hypothermia (Table 1). This explains the effect of ATP and energy charge loading and why lactate did not increase during glycogen degradation and glucose metabolism to pyruvate (Table 1). On the other side when we measured glucose in livers treated by HOPE we found high levels of glucose, which did not prevent glycogen depletion via feedback under cold oxygenated conditions. We cannot satisfactorily explain why glycogen depletion was not inhibited in spite of high glucose levels in the cold.

Analysis of the livers treated by HOPE showed after 20 min of ischemic rewarming and 40 min of reperfusion significantly improved cytosolic enzyme release (Figure 6), oxidative stress (Table 2, Figure 5), bile flow (Table 2), sinusoidal integrity (Figure 7B) and responses to pro-inflammatory mediators (Figure 8).

In order to achieve clear results concerning release of reactive oxygen species during the early reperfusion after cold oxygenated perfusion we decided to choose the isolated reperfused rat liver with ex vivo asanguineous short reperfusion. Thus an extrapolation to overall statements concerning reperfusion injury after in vivo liver transplantation is not justified. But our results underline the contribution of an ischemic rewarming period in enhancing injury prior to reperfusion. Also, Vadojva et al. (4,30) showed that reperfusion injury after 9 h and even 18 h CS of rat livers was significantly decreased if no ischemic rewarming period was interposed between cold storage and normothermic reperfusion. According to this data, other investigators reported that an ischemic temperature rise (4 to 37°C) over 15 min after preservation is critical for both, mitochondrial and organ energy contents (31). Organ protection, demonstrated by the recovery of a high ATP content, has been shown in the presence of low concentrations of cyclosporin A (CsA), a highly specific inhibitor of mitochondrial permeability transition (MPT). Mitochondrial damage after transition of the liver to normothermia was therefore thought to be caused by MPT opening during rewarming after cold storage (31,32). In this study, oxygen consumption during reperfusion was significantly increased after cold storage (Figure 4), whereas cytochrome oxidase was significantly reduced after cold storage (Figure 2). Taken together these observations strongly support the hypothesis that mitochondrial dysfunction occurs after cold storage + ischemic rewarming. Similar findings that cold storage with UW solution of rat liver affects mitochondrial electron transfer during reperfusion has been reported by other investigators (33).

Measurements of the initiator caspases 8 and 9 showed no increase of caspase 8 activity after rewarming and reperfusion in both experimental groups, i.e cold stored and machine perfused livers. On the other hand, caspase 9 activity was increased only in reperfused cold stored livers in contrast to those treated with the machine perfusion (Table 2). These results indicate induction of apoptosis after cold storage by an intrinsic apoptotic pathway under the described conditions of acellular early reperfusion.

Of note, no evidence of oxidative or apoptotic damage was observed during hypothermic oxygenation after cold storage. This is in contrast to previous studies from Rauen et al. who suggested that oxygen supply under cold conditions should be very carefully managed because of possible ROS formation via classical Fenton chemistry (13) of increasing intracellular amounts of chelatable free iron. This phenomenon was shown in cultured hepatocytes and endothelial cells. Our results demonstrate that whole organ low pressure perfusion with oxygenated modified cold UW solution after cold storage did not increase apoptosis or lipid peroxidation (Table 1). While we were not able to fully understand this contradictory results, we would postulate that the iron chelating properties of lactobionic acid (34) during machine perfusion may contribute to different amounts of chelatable cellular iron as compared to cold stored cultured hepatocytes.

Several recent studies clearly showed the absence of apoptotic or necrotic cell death during oxygenated machine liver perfusion. Butler et al. (10), Schön et al. (9) and Imber et al. (35) reported increased liver viability after normothermic (not hypothermic) extracorporeal machine perfusion of warm ischemic or fresh harvested porcine livers. Other groups have shown advantages of hypothermic oxygenated machine preservation in the rat liver (22,36). Minor et al. found decreased cell death by cold venous retrograde oxygen insufflation prior to reperfusion after warm ischemia (37) and also after very long (47 h) cold storage (38). Bessems et al recently reported about a new machine perfusion solution, polysol, which was superior to cold storage of heart beating (39) or non-heart-beating (40) donors. In addition, Lee et al. demonstrated in a rat non-heart-beating-donor transplantation model that a 5 h machine liver perfusion improved survival and reduced cellular damage in the graft (41). The same group, however, noted that cold machine liver perfusion unfortunately increased endothelial damage (23,24).

In summary, conclusions from the reports on machine liver perfusion are difficult to interpret because of varying study designs e.g. different perfusion solutions, normothermic versus hypothermic perfusion, heart-beating versus non-heart-beating donors, reperfusion with versus without blood. In addition, most investigators did not perform an ischemic rewarming period between preservation and normothermic reperfusion, and apply their perfusion system during the entire duration of preservation; both are important shortcomings, which have prevented application in the clinical setting. To the best of our knowledge, the oxygenated machine liver perfusion approach proposed in this study appears clinically attractive because of a short perfusion period after transport of the cold stored organ. In contrast to the normothermic perfusion approach it avoids repeated temperature variations because of cold perfusion after cold storage.

Hypothermic oxygenated machine perfusion may thus serve as an important tool in the clinic for optimizing marginal organs.