* Abbreviations: ATP, adenosine triphosphate; DPG, 2,3 diphosphoglycerate; GPC, glycerophosphorylcholine; GPE, glycerophosphorylethanolamine; HtR, hypothermic reperfusion; NDP, nucleotide disphosphate; NTP, nucleotide triphosphate; PDE, phosphodiester; 3PG, 3 phosphoglycerate; Pi, inorganic phosphate; PME, phosphomonoester.
Liver transplantation is the treatment of choice for patients with end-stage acute or chronic liver disease. Various studies have been carried out to optimize organ viability and to extend the preservation time of the liver, once harvested (1-3). Refinements in organ preservation fluids have resulted in the reduction of cell injury by the inclusion of osmotically active substances, which prevent hypothermically induced cell swelling and free radical scavengers, which prevent cell membrane oxidation and protein degradation (4). Despite these developments primary nonfunction remains a clinical problem that is largely unexplained. Further research is needed to assess the metabolic requirements of the liver during cold storage and immediately following transplantation in order to improve ultimate clinical outcome. An important consideration is the ability of the liver graft to quickly regenerate sufficient high energy phosphate compounds to sustain metabolic integrity. Adenosine triphosphate(ATP)* is the major energy-rich phosphate source required for most hepatic synthetic functions including protein, bile, and urea production as well as important intracellular transport mechanisms. Marubayashi et al.(5) have shown in experimental transplantation rat models that ATP regeneration is an important prerequisite for hepatic allograft function. Further studies using high pressure liquid chromatography of human donor liver biopsies following cold ischemia have shown similar correlations between ATP levels and graft integrity (6, 7).
Our previous studies and those of others using 31P-NMR have shown that it is possible to regenerate hepatic ATP following 24 hr cold storage (4°C) using a method of brief hypothermic reperfusion(3, 8, 9) in the rat. More recently, Busza et al. (10) used 31P-NMR techniques for evaluating different preservation fluids and found that a modified lactobionate/raffinose(University of Wisconsin [UW]) solution was beneficial for hepatic ATP regeneration in the rat following a period of 48 h cold storage, compared with a composite gelatine/citrate buffer. Most experimental liver studies have been conducted on rats, with very few clinically related animal models being employed. A study by Morita et al. (13) used dog pancreas to determine viability using a two-layer cold storage method(14) that allowed continuous ATP production following a prolonged period of warm ischemia. Results indicated that tissue ATP levels, at the end of the preservation period using the two layer technique, correlated well with pancreatic functional success rates posttransplantation. If progress in human liver preservation is to be achieved, more studies on larger animals with anatomy similar to humans are essential. It is imperative that the energetic status of the larger organ be assessed during the period of retrieval, storage, and subsequent reperfusion to provide more concise knowledge for the development of better clinical preservation fluids and surgical techniques.
The pig has a liver of a size comparable to humans with a similar structural and functional anatomy. The pig model allows human donor liver harvesting techniques to be employed and therefore is of greater clinical relevance than rat studies. This study, therefore, has used this animal model to delineate simultaneous changes seen in three important moieties of hepatic metabolism by using in vitro 31P-NMR. First, the high energy rich phosphate compounds of the nucleotide tri- and di- phosphates (NTP and NDP, which include ATP and ADP) were studied together with their intracellular exchange with inorganic phosphate (Pi). This provides information on the energetic status of the liver during storage and subsequent reperfusion. Second, the glycolytic/gluconeogenic intermediates 3-phosphoglycerate (3PG) and 2,3 diphosphoglycerate (DPG) were investigated. Finally, glycerophosphorylcholine (GPC) and glycerophosphorylethanolamine (GPE)-both of which are membrane derived, reflecting cell integrity following ischemic insult-were followed.
MATERIALS AND METHODS
The animals were housed and maintained in accordance with procedures outlined in the Home Office (Animals Protection) Act, 1986.
Liver isolation, storage, and reperfusion. Landrace/large white cross-pigs (n=5) (30 kg) were fasted for 12 hr. Following sedation with an intramuscular injection of azaperone (0.1 ml/kg), an intravenous injection of ketamine (10 mg/kg) was administered. Intubation of the spontaneously ventilated pigs was via an endotracheal tube supplying 0.5% halothane and a mixture of oxygen and nitrous oxide. Midline and lateral abdominal incisions were made to fully expose the 4 anatomical lobes of the liver. A 2-3 g wedge of liver was then removed by sharp dissection and immediately freeze-clamped in liquid nitrogen. The biopsy site was then sutured. A further liver biopsy was taken following the exposure of the hilar structures and ligation of all fascial tissue surrounding the bile duct, hepatic artery, and portal vein. These two initial biopsy samples were obtained at a time designated as Period A. Subsequent sampling intervals B-D are described below. Following hepatic arterial occlusion the portal vein was cannulated. The liver was immediately perfused with 1 L ice-chilled citrate buffer and then 1 L ice-chilled modified UW preservation fluid (11). Liver biopsies were taken immediately postperfusion (0 min) and then sequentially at 15, 30, 45, 60, 90, 120, and 240 min following harvest (Period B). All sampling points had randomized selections of biopsy sites to ensure that no zonal attributes were incorporated into the results and that the results were a true representation of the whole liver. Typically, the perfusion time was 3-5 min, after which the liver was excised and placed in a bowl that was surrounded with ice. The liver was stored on ice for a total of 4 hr, after which the organ was reperfused with ice-cold oxygenated UW solution. The reperfusion circuit was recirculatory through the portal vein at a constant flow rate of 500 ml/min. Following 1.5 and 2 hr of hypothermic reperfusion (HtR), liver biopsies were taken (Period C), after which the pump was switched off to induce a second period of cold hypoxia and further biopsies removed every 15 min for 1 hr(Period D).
NMR spectroscopy. All liver biopsies were deproteinized with perchloric acid (12%) and lyophilized for 48 hr following neutralization with KOH. The lyophilized liver extracts were resuspended in deuterium oxide (1.5 ml) and EDTA was added to a final concentration of 100 mmol/L to chelate any ions (which cause spectral line broadening) such as Mn2+, Fe3+ and Mg2+, which are bound to 93-97% of total ATP. The samples were neutralized with KOH to pH 7.45 and were run on a Varian 500 MHz 11.7T NMR machine, following the addition of 50 μl of phosphocreatine (PCr) to a final concentration of 67 mmol/L which acted as a reference standard for the chemical shift phosphorus spectrum. Fully relaxed proton decoupled (WALTZ 16 modulated) spectra were accumulated using an acquisition time of 0.8s, a relaxation time of 20s, and a pulse width of 13 μs corresponding to a pulse angle of 45 °. Following identification of metabolite peaks(15), absolute quantification was accomplished by measuring the integrals of the resonances associated with individual metabolites with respect to the known PCr standard concentration. Individual metabolite integrals, including those that overlap with one another, were measured using a dedicated curve fitting program (NMR1, New Methods Research., Inc., Syracuse, NY).
Statistical analysis. Metabolite concentrations were calculated as μmol/g wet weight of tissue extracted and represented as means ± SEM. Significant differences between the means of preharvest hepatic metabolite concentrations and subsequent means of hepatic metabolite concentrations from the cold storage and HtR periods were calculated using the Student's paired t test with weighting for multiple comparisons.
Figure 1 shows typical high resolution phosphorus spectra obtained at 500 MHz from the pig liver biopsies. The spectra contain four distinct classifications: (A) phosphomonosters-consisting of glycolytic intermediates such as 3PG, 2,3 DPG, and various sugar phosphates (6-8 ppm);(B) inorganic phosphate; (C) phosphodiesters-GPC and GPE (1.5-4 ppm); and (D) NTP and NDP. The NTP is represented by three distinct reasonances consisting of various contributions from both tri and diphosphate nucleotides. Theγ-NTP resonance contains contributions from both β-NDP andγ-NTP, due to the similarity of the terminal phosphate group. The same is true for α-NTP, which shares a similar resonance with α-NDP and other nucleotides such as NAD+/NADH. As a result both γ-NTP andα-NTP are not true indicators of total nucleotide triphosphate concentration. However, β-NTP is solely due to the resonance of theβ-phosphate group of NTP and provides an excellent means of assessing nucleotide triphosphate levels.
Prior to harvesting, the 31P-NMR spectrum (Fig. 1a) shows 3 distinct resonances from the three NTP components-however following 4 hr of cold storage NTP levels were seen to reduce dramatically(Fig. 1b). After a subsequent 2 hr period of oxygenated HtR, NTP regenerated to almost preharvest levels.
Absolute changes in NTP and Pi. Preharvest β-NTP concentrations were calculated as 1.10±0.13 μmol/g(Fig. 2). Immediate perfusion of the liver with citrate and UW solutions (between 3-5 min) resulted in dropping β-NTP levels by 28% to 0.79±0.14 μmol/g. After 15 min storage, β-NTP levels had fallen by 61% to 0.43±0.22 μmol/g (P<0.05), and at 45 min by 88% to 0.13±0.08 μmol/g (P<0.001). The observed decline in β-NTP was exponential throughout period B which extended to 4 hr postharvest when there were no detectable β-NTP levels. Upon oxygenated HtR for 1.5 and 2 hr, β-NTP was regenerated to levels that were 43% and 30% lower than preharvest values, respectively. These levels(0.63±0.28 μmol/g and 0.77±0.14 μmol/g) were not significantly different. It is of interest to note that preharvest NTP levels were not achieved during the period of oxygenated hypothermic reperfusion. This was probably because of ischemic cell and mitochondrial damage sustained during harvesting and cold storage of the liver. The initial rate at whichβ-NTP levels fell following 15 min cessation of HtR (Period D) appeared to be similar to that during storage at the same time point (Period B)-however, this was not sustained and the decline slowed thereafter.
The response of γ-NTP to cold storage and subsequent HtR mirrored that exhibited by β-NTP-however, distinct differences in the rate and magnitude of change were evident (Fig. 3). The level of preharvest γ-NTP was approximately 1 μmol/g and equivalent toβ-NTP. It was not until 45 min postharvest that a significant decline of 47% was observed (0.75±0.12 μmol/g;P<0.05) compared with 88% of the β-NTP at the same time point (0.13±0.08 μmol/g;P<0.001). Even after 4 hr of storage, the levels still remained at 0.75±0.21 μmol/g (P<0.1), approximately 45% lower than preharvest values. Upon HtR for 1.5 hr (Period C) γ-NTP levels increased to 1.14±0.25 μmol/g-not significantly different from initial levels (NS)-representing an increase of 52% from 4 hr cold storage. After 2 hr of HtR γ-NTP levels had reached preharvest levels(1.33±0.21 μmol/g; NS), an increase of 77% over 4 hr storage values. Following 30 min cessation of HtR (Period D), γ-NTP levels had fallen to 0.77±0.25 μmol/g. It is interesting to note that 60 min after cessation of HtR, values of γ-NTP were not significantly different from preharvest values. However, at this same time point following liver harvest,γ-NTP was 46% lower (P<0.05) than the preharvest value.
There appeared to be no significant change associated with the α-NTP moiety (Fig. 4), due to its overlap with NAD+ and NADH, which provide a constant and substantial contribution. Levels during Period A and Period B ranged between 2.50 μmol/g and 1.5 μmol/g. During HtR (Period C) α-NTP increased slightly to preharvest levels of approximately 2.50 μmol/g and remained at this level even after cessation of HtR.
Changes seen by intracellular inorganic phosphate are strongly associated with the phosphorylation state of the nucleotides. However, the large phosphate component of the UW preservation buffer will affect intracellular and vascular phosphate levels, especially following reperfusion. Preoperative Pi levels were approximately 1.8 μmol/g (Period A); this rapidly increased by 31% to 2.4±0.3 μmol/g immediately following harvest and by 96%(P<0.02) to 3.6±0.55 μmol/g after 15 min storage.Figure 5 shows a steady increase in Pi levels during the course of storage (Period B), and after 2 and 4 hr, levels had increased by 179% (P<0.001) and 223% (P<0.001) to 5.1±0.5μmol/g and 5.92±0.52 μmol/g, respectively. Following 1.5 hr of HtR (Period C), Pi continued to increase to 6.93±1.28 μmol/g (+278% cf. preharvest levels). After a further 30 min HtR, Pi levels began to fall to a level that was only 10% lower than 4 hr storage values. Interestingly, the fall in Pi continued for a further 15 min after HtR (Period D), but returned to preHtR values after 30 min.
Absolute changes in 3 phosphoglycerate and 2,3 diphosphoglycerate. Prior to harvesting the liver (Period A), 3-phosphoglycerate was not detectable, but immediately after flushing the liver in situ (Period B) levels increased to 0.31±0.08 μmol/g, and at 45 min postharvest the levels had increased to 0.88±0.07 μmol/g(Fig. 6). After 45 min the increase plateaued and remained relatively constant for the period leading up to 4 hr storage, apart from 60 min postharvest when values reached 1.16±0.17 μmol/g. Following hypothermic reperfusion (Period C) 3PG levels reduced by 44% and 61% of the 4 hr storage value. During Period D, after stopping HtR, 3PG did not increase at the same rate seen during the initial stages of storage, and it remained fairly constant throughout the second period of hypoxia at approximately 0.4 μmol/g. This is more than half the concentration seen in the later stages of cold storage (Period B).
The absolute changes associated with the 2,3 DPG moiety reflect the changes seen by the 3PG due to their close relationship in the glycolysis pathway(Fig. 7). Prior to liver harvest (Period A) DPG levels were very low (ca. 0.10 μmol/g)-however, at 30 min postharvest (Period B) they had increased to 0.58±0.24 μmol/g. Following storage for 4 hr, levels increased further to 0.90±0.18 μmol/g. During the period of HtR (Period C) DPG levels fell to 0.47±0.18 μmol/g and 0.40±0.25 μmol/g after 1.5 and 2 hr, respectively. Following cessation of HtR DPG levels increased.
Absolute changes in GPC and GPE. The changes associated with GPC and GPE could be partially a consequence of phospholipid breakdown of plasma membranes (17), or they could be a consequence of cell swelling. Figure 8 shows that preharvest, GPC and GPE values were 1.69±0.14 μmol/g and 1.07±0.16 μmol/g, respectively. At 15 min storage (Period B) there was a transient increase in both levels to 2.09±0.18 μmol/g and 1.40±0.16, respectively, but in neither case was this deemed to be significant. Thereafter, preharvest values were maintained throughout the 4 hr storage period of both metabolites. Following HtR the GPC moiety decreased significantly to 1.41±0.10μmol/g- i.e., a decline of 23% (P<0.05) but GPE was reduced only slightly.
Previous studies have shown that it is possible to resuscitate high energy phosphorylated nucleotide (8) in the rat liver after cold preservation by hypothermic reperfusion. Studies using the two layer technique in dog pancrease have shown improved graft viability corresponding to high levels of ATP (13). However, there have been no studies, to our knowledge, that have used a larger animal model that more closely resembles human metabolism and anatomy to study the liver. Studies on rat liver have shed important light on the process of preservation and subsequent hypothermic reperfusion to regenerate NTP-however, a clinical approach cannot be simulated. The pig has allowed techniques of anesthesia, surgery, and perfusion employed for human liver harvesting to be used.
This study has shown for the first time that it is possible to regenerate NTP levels following a period of hypothermic reperfusion in this clinically related model. Our results indicate that many metabolic functions are affected by organ retrieval, storage, and HtR. However, one important prerequisite for the long-term survival of the organ is the potential to produce high energy phosphorylated nucleotides to sustain metabolic integrity following storage. An important observation was the speed with which NTP levels were depleted immediately following in situ perfusion during the harvesting procedure. This rapid change was also reflected in changes in the Pi, PME, and PDE resonances, all of which increased during this period. Following 2 hr of HtR it was apparent that all these metabolites, including NTP, returned toward their preharvest values but none actually reached these original levels. This would indicate that, during in situ perfusion and storage, damage has occurred to cellular integrity. Various factors, such as nucleotide depletion, cell swelling, mitochondrial damage, and plasma membrane damage may be contributory to this phenomenon.
Preharvest β-NTP levels seen in this study appear low, compared with studies in the rat, where hepatic ATP levels of 2.5 μmol/g are reported(16). However, the method we employed involved careful cutting of a wedge biopsy followed by freeze-clamping, unlike most rat studies where freeze-clamping can be achieved either in situ or within a few seconds of tissue retrieval. The time taken to freeze-clamp the liver biopsy was the minimum possible-however, a longer than desired period of tissue ischemia, and hence hypoxia, was inevitable. From previous studies it has been noted that short periods of ischemia (<60s) reduce hepatic ATP levels considerably(16), hence lower preharvest NTP levels were obtained in the present study.
The elevation of 3PG and DPG following liver harvest are important indicators of the changes associated with the glycolytic pathway. The conversion of phosphoenolpyruvate to pyruvate requires substrate level phosphorylation-i.e., the transfer of Pi to ADP-which may be blocked due to the absence of ADP, and therefore causing a build-up in 3PG and DPG. The increases in 3PG and 2,3 DPG may also be a consequence of pH sensitive glycolytic enzymes such as 6-phosphofructokinase that become inactive due to low pH and can also be allosterically inactivated due to ATP and AMP changes(12). It is interesting to note that the increases seen in the 3PG moiety occur at a faster rate than that of the DPG. The increase in DPG is not clear since this intermediate is transient during glycolysis and remains at very low concentrations under normal circumstances.
Another important observation from this study is the behavior of GPC and GPE, which mimic each other. Throughout the period of storage, both GPC and GPE remain fairly constant, resembling preharvest levels. However, following hypothermic reperfusion both moieties decreased in concentration, indicating a possible flushing away of these accumulated metabolites during the period of cold storage. Accumulation of GPC and GPE during storage may have a number of cause. First, if membrane damage is occuring, these two metabolites would accumulate, but following reperfusion they may be flushed away, or, due to the tight control of phospholipid degradation and synthesis, may be converted to a variety of phospholipid metabolites (17). Second, if these two metabolites are acting as osmolytes, an increase in GPC and GPC would be expected with any cell swelling during the period of cold storage. Then upon reperfusion cellular swelling would reduce and GPC and GPE levels would drop.
This study has shown that NTP regeneration is possible in the larger animal model following HtR-hence its clinical importance for liver transplantation. Although further studies need to be carried out, we propose that hypothermic reperfusion of human liver prior to transplantation may be beneficial for postoperative liver function. It is recognized that many advances in the procedure of liver harvesting are required to produce livers with a consistently high level of initial graft function. The technique described here may facilitate this possibility. The method will allow assessment of methods for prevention of cellular damage from hypoxia/ischemia, such as the use of free radical scavengers. These agents may provide beneficial effects on the outcome of hypothermic reperfusion such as faster NTP regeneration at levels more closely resembling preharvest values. These studies show that it may be possible to conduct ex vivo NMR studies on intact livers to investigate the metabolic integrity of the liver during the window of opportunity prior to implantation, provided suitably adapted NMR hardware is available. Furthermore, our model may be used to assess NTP changes with variations in preservation fluids and organ storage that may have an impact on the current clinical methods of organ harvesting, retrieval, storage, and implantation.
Acknowledgment. We would like thank the University of London Intercollegiate Research Service for machine time at the Institute of Child Health and Birkbeck College (London).
1. Sumimoto R, Kamada N, Jamieson NV, Fukuda Y, Dohi K. A comparison of a new solution combining histidine and lactobionate with UW solution and Eurocollins for rat liver preservation. Transplantation 1991; 51: 589.
2. Palombo JD, Pomposelli JJ, Hirschberg Y, Blackburn GL, Bistrian BR. Glycolytic support of adenine nucleotides in rat liver flush-preserved with UW or Collins' II. Transplantation 1989; 48: 901.
3. Palombo JD, Pomposelli JJ, Fechner KD, Blackburn GL, Bistrian, BR. Enhanced restoration of adenine nucleotides in rat liver following extended preservation in UW solution by provision of adenosine during reperfusion. Transplantation 1991; 51: 867.
4. Southard JH, van Gulik TM, Ametani MS, et al. Important components of the UW solution. Transplantation 1990; 49: 251.
5. Marubayashi S, Takenaka M, Dohi K, Ezaki H, Kawasaki T. Adenine nucleotide metabolism during hepatic ischemia and subsequent blood reflow periods and its regulation to organ viability. Transplantation 1980; 30: 294.
6. Kamiike W, Burdelski M, Steinhoff G. Adenine nucleotide metabolism and its relation to organ viability in human liver transplantation. Transplantation 1980; 45: 138.
7. Lanir A, Jenkins RL, Caldwell C, Lee RGL, Khettry U, Clouse ME. Hepatic transplantation survival: correlation with adenine nucleotide level in donor liver. Hepatology 1988; 8: 471.
8. Fuller BJ, Busza AL, Proctor E. Possible resuscitation of liver function by hypothermic reperfusion in vitro after prolonged (24-hour) preservation-a 31
P NMR study. Transplantation 1990; 50: 511.
9. Delmas-Beauvieux M-C, Gallis J-L, Clerc M, Cannioni P. A nuclear magnetic resonance study of thermal transition during reperfusion of rat liver preserved in Bretschneider's solution. Transplantation 1992; 53: 709.
10. Busza AL, Fuller BJ, Proctor E. Evaluation of cold reperfusion as an indicator of viability in stored organs: a 31
P NMR study in rat liver. Cryobiology 1994; 31: 26.
11. Busza AL, Fuller BJ, Proctor E. The response of the liver to lactobionate/raffinose (University of Wisconsin-UW) solution during hypothermic preservation: a study using 31
phosphorus nuclear magnetic resonance. Cryobiology 1989; 26: 273.
12. Sakakibara R, Uyeda K. Differences in the allosteric properties of pure low and high phosphate forms of phosphofructokinase from rat liver. J Biol Chem 1983; 258: 8656.
13. Morita A, Kuroda Y, Fujino, Y, Tanioka Y, Ku Y, Saitoh Y. Assessment of pancreas graft viability preserved by a two-layer (University of Wisconsin solution/perfluorochemical) method after significant warm ischemia. Transplantation 1993; 55: 667.
14. Kuroda Y, Kawamura T, Suzuki Y, Fujiwara H, Yamamoto K, Saitoh Y. A new simple method for cold storage of the pancreas using perfluorochemical. Transplantation 1988; 46: 457.
15. Bell JD, Cox IJ, Sargentoni J, et al. A 31
H investigation of normal and abnormal human liver. Biochim Biophys Acta 1993; 1225: 71.
16. Hems DA, Brosnan JT. Effects of ischemia on content of metabolites in rat liver and kidney in vivo. Biochem. J 1970; 120: 105.
© Williams & Wilkins 1996. All Rights Reserved.
17. Ruiz-Cabello J, Cohen JS. Phospholipid metabolites as indicators of cancer cell function. NMR Biomed 1992; 5: 226.