Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Primary nonfunction and dysfunction occur in 5 to 30% of liver transplantation cases, resulting either in the need for retransplantation or death of the recipient (Ploeg et al., 1993). Because liver transplantation is the therapy of choice in an increasing number of liver diseases (Pichlmayr et al., 1987) and the organ pool is limited, there is an urgent need to understand underlying mechanisms responsible for graft failure.

It has been demonstrated that Kupffer cell-dependent reperfusion injury to endothelial cells is detrimental for graft survival. Once activated, Kupffer cells release numerous inflammatory mediators, including tumor necrosis factor-, interleukins, and prostaglandin E₂. Furthermore, free radicals are produced upon reperfusion. These substances mediate injury to the graft after transplantation (Lemasters and Thurman, 1997). Recently, in situ manipulation by touching, retracting, and moving liver lobes gently during harvest, which is difficult to prevent with standard harvesting techniques, reduced survival dramatically after transplantation (Schemmer et al., 1998). Gadolinium chloride (GdCl₃), a rare earth metal and Kupffer cell toxicant, and glycine, a nontoxic amino acid given to donors before organ harvest, totally prevented the development of pathology upon reperfusion, suggesting a role for Kupffer cells.

Several lines of evidence suggest that microcirculatory disturbances are a key factor in enhancing donor organ susceptibility to both cold and warm ischemia in livers (Teramoto et al., 1993; Hui et al., 1994; Husberg et al., 1994). A number of studies show that microcirculatory disturbances cause hypoxia, which leads to activation of Kupffer cells, free radical production, and reperfusion injury after cold storage (Lemasters et al., 1995; Lemasters and Thurman, 1997). Thus, this study was designed to test the hypothesis that organ manipulation causes hypoxia and free radical production during harvest by disturbing the hepatic microcirculation. To avoid difficulties with interpretation, which could occur in vivo, a blood-free liver perfusion model was used here.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Experimental Animals and Treatment. Female Sprague-Dawley rats (200-230 g) were allowed free access to standard laboratory chow (Agway PROLAB RMH 3000, Syracuse, NY) and tap water. Some donor animals were given a single injection of GdCl₃ (10 mg/kg in acidic saline) through the tail vein 24 h before harvesting. This treatment destroys all large Kupffer cells (Hardonk et al., 1992; Koop et al., 1997). Other donor rats were fed a powdered chow diet containing 5% glycine for 3 days, which blunts the response of Kupffer cells to endotoxin (Ikejima et al., 1996). Animals were anesthetized with methoxyflurane before surgery.

Harvest Procedure. To determine the influence of gentle manipulation on reperfusion injury, donor livers were harvested within 25 min before perfusion with cold University of Wisconsin (UW) solution. Minimal dissection was performed in a standardized fashion during the first 12 min, including freeing the organ from ligaments and cannulation of the bile duct. During the next 13 min, livers were either left alone or manipulated gently. To maintain standard conditions, gentle manipulation was carried out by the same surgeon touching, retracting, and moving the liver lobes in situ for a specified time interval. Care was taken to use the same number of manipulations in each experiment with similar pressures. Serum transaminases at the end of manipulation were identical regardless of pretreatment, validating the manipulation technique (Schemmer et al., 1998). At 25 min reperfusion with 8 ml of cold Ringer's solution followed by 3 ml of cold UW solution was performed in situ via the portal vein. Livers were subsequently stored for 24 h in cold UW solution.

Liver Perfusion. Livers were perfused via the portal vein at 3 to 4 ml/min/g liver with Krebs-Henseleit bicarbonate buffer (118 mM NaCl, 25 mM NaHCO₃, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 4.7 mM KCl, and 1.3 mM CaCl₂) at pH 7.6, saturated with 95% N₂ and 5% CO₂ or 95% O₂ and 5% CO₂ at 37°C using a peristaltic pump after 24 h of cold storage of the liver in UW solution (Scholz, 1968; Brouwer and Thurman, 1996).

Trypan Blue Infusion and Histology. Following each experiment, trypan blue (500 µM; Aldrich Chemical Co., Milwaukee, WI) was infused into the liver to assess viability of cells. Livers were then flushed with additional perfusate to remove excess dye and fixed by perfusion with 4% paraformaldehyde in Krebs-Henseleit bicarbonate buffer at pH 7.6, embedded in paraffin, and processed for light microscopy using an eosin counterstain. The presence of trypan blue in the nuclei is indicative of irreversible loss of cell viability (Belinsky et al., 1984; Bradford et al., 1986). Five pericentral and five periportal fields (100× magnification) were selected at random from at least four different sections per sample, and mean values of stained nuclei from nonparenchymal and parenchymal cells were calculated.

Surface Fluorescence Measurement. Fluorescein isothiocyanate-dextran (mol wt 70,000, catalog no. FD-70S, Sigma Chemical Co., St. Louis, MO) was dissolved in perfusate at a concentration of 12 µM and perfused for 3 min at the end of the donor operation to assess microcirculation. Because fluorescein isothiocyanate-dextran is confined to the vascular space in liver, the rate of fluorescence wash-in and -out as well as the percentage of increase of surface fluorescence over basal fluorescence is indicative of microcirculation and vascular space, respectively. A mercury arc lamp equipped with a glass filter was used to produce excitation wavelengths of 430 nm. Fluorescence of fluorescein-dextran (560 nm) was measured via a light guide (tip diameter, 2 mm) placed on the surface of the perfused liver with a micromanipulator. The signal was amplified and recorded as described elsewhere (Conway et al., 1985). Fluorescein-dextran was infused and changes of fluorescence were detected from the surface of the liver. The rate of fluorescence wash-in and -out as well as the percentage of increase of surface fluorescence was calculated from changes of fluorescence. Basal fluorescence is the value detected from the liver surface due to endogenous fluorophores (e.g., flavoproteins) before the dye was administered. Because fluorescence values are presented as percentage of basal, this normalizes for day-to-day and organ-to-organ variation.

Portal Pressure. Livers were perfused in situ with oxygenated Krebs-Henseleit bicarbonate buffer (3-4 ml/min/g liver at 37°C) and the donor operation was performed as described above. Portal pressure was monitored continuously using a Digi-Med Low-Pressure-Analyzer model 200 (Micro-Med, Louisville, KY).

Determination of Reduced, Protein-Bound Pimonidazole by Enzyme-Linked Immunoassay (ELISA) and Immunohistochemistry. Pimonidazole, a 2-nitroimidazole marker for viable hypoxic cells (Durand and Raleigh, 1998), was given to donors i.v. 5 min before the donor operation. Pimonidazole adduct accumulation in vivo was measured in tissue homogenates with a competitive ELISA procedure described previously (Raleigh et al., 1994) as modified for liver tissue (Arteel et al., 1995). Protein levels in tissue homogenates were determined with the bicinchoninic acid assay using a commercially available kit (Pierce Chemical Company, Rockford, IL). Paraffin blocks of formalin-fixed liver tissue were sectioned at 6 µm and pimonidazole was detected with a biotin-streptavidin-peroxidase indirect immunostaining method using diaminobenzidine as a chromogen as described previously (Arteel et al., 1995). After the immunostaining procedure, a counterstain of hematoxylin was applied. A Universal Imaging Corp. (Chester, PA) Image-1/AT image acquisition and analysis system incorporating an Axioskop 50 microscope (Carl Zeiss, Inc., Thornwood, NY) was used to capture and analyze the immunostained tissue sections at 100× magnification (Arteel et al., 1997). Although results of ELISA give only the quantity of bound pimonidazole, immunohistochemical analysis shows the lobular pattern of binding.

Free Radical Detection. To detect free radical adducts, C-phenyl-N-tert butylnitrone (PBN; 0.1 g/kg, i.p.) dissolved in dimethyl sulfoxide was injected 5 min before surgery, the bile duct was cannulated with PE10 tubing, and bile was collected for 1 h into 30 µl of deferoxamine mesylate solution (5 mM) to prevent ex vivo radical formation. Samples were frozen immediately on dry ice. Subsequently, bile samples were thawed, placed in a quartz flat cell, and bubbled with oxygen for 10 min and nitrogen for 5 min to remove interfering ascorbate. ESR spectra were obtained using a Bruker ECS-106 spectrometer operating at 9.77 GHz with a 50-kHz modulation frequency. Instrument conditions were as follows: 20-mW microwave power, 1.0-G modulation amplitude, 80-G scan width, 8-min scan, and 0.5-s time constant. Spectra were analyzed for hyperfine coupling constants by computer simulation (Duling, 1994).

Statistical Analyses. Mean values ± S.E.M. for various groups were compared using two-way ANOVA with Student-Newman-Keuls post hoc and t test as appropriate. p < .05 was selected before the study as the criterion for significance.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Gentle Organ Manipulation on Hepatic Microcirculation. To determine the influence of manipulation during organ harvest on microcirculation, livers were perfused with fluorescein-dextran, a dye that is confined to the vascular space because of its large size (mol wt 70,000) and highly charged nature (Conway et al., 1985). In unmanipulated control livers, fluorescein-dextran fluorescence from the surface was increased about 360 ± 27% over basal fluorescence; however, it increased after manipulation only 190 ± 10%, reflecting a reduced intrahepatic vascular volume. Treatment with GdCl₃ and dietary glycine blunted the effect of manipulation, reflected by an increase of fluorescence to 327 ± 70% and 362 ± 55%, respectively (Figs. 1 and 2). Furthermore, manipulation markedly affected the microcirculation, as shown by a 50 to 70% decrease in the rate of wash-in and -out of fluorescein-dextran (Figs. 1 and 2). These effects were also prevented by pretreatment with GdCl₃ or glycine (Figs. 1 and 2).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Fluorescence increases due to fluorescein-dextran. Donor livers were harvested within 25 min. Briefly, after minimal dissection during the first 12 min, livers were left alone or manipulated for the next 13 min. Some donor rats were pretreated with GdCl3 or dietary glycine as described in Materials and Methods. At the end of harvest, livers were perfused with fluorescein-dextran via the portal vein for 3 min as indicated by horizontal bars and vertical arrows. Data are representative of typical experiments.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Effect of gentle organ manipulation on fluorescein-dextran distribution. Conditions were as described in Fig. 1. Fluorescein-dextran (12 µM) was infused after manipulation for 3 min and percentage of increase over basal fluorescence (A) as well as the rate of fluorescence wash-in and -out (B) was recorded for each liver as described in Materials and Methods. Rates of fluorescein-dextran entry and exit were determined by measuring the slope of the linear portion of the change in fluorescence upon infusion (wash-in) and after termination (wash-out). Changes of fluorescence are presented as percentage of basal fluorescence/time. Values are mean ± S.E.M. (p < .05 by two-way ANOVA with Student-Newman-Keuls post hoc test, n = 5-8). *p < .05 for comparison to nonmanipulated group; p < .05 compared with manipulated group without pretreatment.

Gentle Organ Manipulation Causes Hypoxia in the Liver. In this study, pimonidazole, a 2-nitroimidazole marker that binds to hypoxic liver cells, was used to detect hypoxia in vivo (Raleigh et al., 1994; Arteel et al., 1995; Durand and Raleigh, 1998). Pimonidazole (120 mg/kg, i.p.) was injected 5 min before the donor operation. Figure 3 shows the pattern of binding of pimonidazole in livers from representative unmanipulated controls and in livers manipulated during harvest. Binding of pimonidazole in naive livers was significantly lower than in any other group; values were 85 ± 5 pmol/mg protein and less than about 1% area of pimonidazole-labeled cells detected by ELISA or immunohistochemistry, respectively. In contrast to naïve unmanipulated controls in which only marginal staining (i.e., hypoxia) was observed, pimonidazole staining was localized extensively in pericentral regions of the liver lobule after manipulation (Fig. 3). Manipulation of livers from rats treated with GdCl₃ or glycine produced a pattern of pimonidazole binding comparable with unmanipulated controls (Fig. 3).

View larger version (184K): [in this window] [in a new window]   Fig. 3.   Effect of gentle organ manipulation on lobular pattern of pimonidazole binding. Conditions were as described in Fig. 1. To detect hypoxia in liver tissue, pimonidazole (120 mg/kg, i.p.), a 2-nitroimidazole hypoxic marker, was injected 5 min before the donor operation. At the end of experiments, livers were fixed with 10% formalin by perfusion. Immunohistochemistry was performed using antibodies to bound pimonidazole as described in Materials and Methods. Photomicrographs depict patterns of pimonidazole binding in livers after harvest. Data are typical of representative experiments.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Effect of gentle organ manipulation on lobular pattern of hypoxia reflected by pimonidazole binding. Conditions were as described in Fig. 1. A, quantitation of immunohistochemistry: sections were taken after surgery, immunohistochemistry was carried out as described in Fig. 3, and image analysis for pimonidazole labeling was conducted as described in Materials and Methods. Furthermore, competitive ELISA was used as described in Materials and Methods to detect bound pimonidazole (B). Values are mean ± S.E.M. (p < .05 by two-way ANOVA with Student-Newman-Keuls post hoc test, n = 5). *p < .05 for comparison to nonmanipulated group; p < .05 compared with manipulated group without pretreatment.

Effect of Gentle Organ Manipulation on Free Radical Formation. Because free radicals are produced during hypoxia (Lemasters et al., 1981; Gao et al., 1991), experiments using electron spin resonance spectroscopy were designed to determine whether free radicals were formed early during harvest. When livers were manipulated gently during harvest in the presence of the spin trap, radical adducts were detected in bile. Figure 5 depicts typical six-line electron spin resonance spectroscopy signals after surgery. Quantitative analysis indicates that graft manipulation significantly increased electron spin resonance spectroscopy signal magnitude from 3.5 ± 1.6 in unmanipulated controls to 11.9 ± 2.3 arbitrary units following manipulation (p < .05). The same radicals were found in bile from some animals treated with GdCl₃ (7.5 ± 4.1) or glycine (8.3 ± 3.8) after manipulation of the liver; however, because of variability, these values were not significantly different from the manipulated group. Computer simulation of the data indicated that manipulation produced two radical adducts. The first possessing N-hyperfine coupling of 16.1 G and -hydrogen hyperfine coupling of 3.6 G, most likely attributable to a carbon-centered radical adduct with a -hydroxyl group. The second adduct had N-hyperfine coupling of 15.4 G and -hydrogen hyperfine coupling of 2.03 G, probably from an oxygen-centered radical adduct.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Effect of gentle organ manipulation on ESR spectrum of radical adducts. Conditions were as described in Fig. 1. Spin trap (PBN; 0.1 g/kg i.p.) was given before harvest as described in Materials and Methods. Bile was sampled during harvest and free radical adducts were measured. Upper panel, spectrum of samples from controls; middle panel, radical adducts from manipulated livers; lower panel, computer simulation of spectrum from manipulated liver.

Effect of Gentle Organ Manipulation on Reperfusion Injury after Cold Storage. After 24 h of cold storage in cold UW solution, livers were perfused with oxygenated buffer for 10 min. To determine whether organ manipulation during harvest causes reperfusion injury to sinusoidal lining cells after cold storage, trypan blue was added to the perfusate (Belinsky et al., 1984; Bradford et al., 1986). Gentle organ manipulation during harvest increased the number of trypan blue-positive sinusoidal lining cells more than 5-fold (Fig. 6). Parenchymal cells were not affected by manipulation (data not shown). Cell death upon reperfusion in livers manipulated during harvest was totally prevented by perfusion with O₂-free (i.e., N₂-saturated) buffer or if donors were pretreated with GdCl₃ or glycine before harvest (Fig. 6).

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Effect of gentle organ manipulation on trypan blue uptake by hemoglobin-free perfused livers. Conditions were as described in Fig. 1. Liver grafts were stored for 24 h in UW solution at 0 to 4°C and subsequently perfused with Krebs-Henseleit buffer containing trypan blue as described in Materials and Methods. *p < .05 compared with no manipulation; p < .05 compared with manipulation without pretreatment by two-way ANOVA with Student-Newman-Keuls post hoc test, n = 6-7.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Gentle Manipulation Causes Hypoxia and Free Radical Formation During Organ Harvest. Preparation of the portal vein alone before perfusion with cold UW solution has been demonstrated to impair the intrahepatic circulation (D'Alessandro et al., 1989; Klar et al., 1995). Even laparatomy and mild abdominal exploration resulted in decreased flow in both the hepatic artery and portal vein (Lautt, 1983; Ji et al., 1984). Recently, in situ organ manipulation during harvest was shown to dramatically reduce survival after transplantation via mechanisms including disturbed microcirculation and Kupffer cells (Schemmer et al., 1998). Kupffer cells play a key role in reperfusion injury; however, pathophysiological changes in the liver due to manipulation before cold storage remain unknown. Because several lines of evidence suggest that microcirculatory disturbances cause hypoxia, free radical production and reperfusion injury after cold storage (Lemasters et al., 1995; Lemasters and Thurman, 1997), this study was designed to test the following working hypothesis: Manipulation of the liver causes disturbances in hepatic microcirculation in situ during harvest, which contributes to a steeper oxygen gradient along the hepatic sinusoid leading to hypoxia in pericentral areas. At reperfusion after cold storage, activated Kupffer cells cause oxygen-dependent injury involving free radicals causing injury to endothelial cells. In support of this hypothesis, organ manipulation during harvest disturbed the hepatic microcirculation (Figs. 1 and 2) and caused hypoxia in this study (Figs. 3 and 4). Hypoxia was detected with pimonidazole, which binds to viable hypoxic cells in vivo. These data are consistent with effects of low flow in isolated perfused rat liver, where areas of hypoxia develop around central veins, whereas tissue surrounding periportal zones remain normoxic (Lemasters et al., 1981) (Fig. 3). Furthermore, because low-flow hypoxia produces anoxia in pericentral zones of the liver lobule, ATP decreases and hypoxanthine accumulates. At the midzonal border between anoxic and normoxic areas, enough oxygen is present to support superoxide formation via xanthine oxidase (Marotto et al., 1988). Oxygen may also be increased transiently by fluctuations in the microcirculation, further promoting oxygen radical formation, which recently has been demonstrated directly by fluorescence microscopy (Suematsu et al., 1992). Interestingly, in this study both oxygen- and carbon-centered free radicals (see Results and Fig. 5) increased significantly in bile collected during organ manipulation.

Gentle Organ Manipulation During Harvest Dramatically Increases Reperfusion Injury. Important factors in graft failure include the donor condition, cold and warm ischemic times, operative complications in the recipient, surgical experience, and the immune status of the recipient (Starzl et al., 1987; D'Alessandro et al., 1989; Strasberg et al., 1994; Imagawa et al., 1996). Some of these factors could potentially contribute to the development of Kupffer cell-dependent reperfusion injury after cold storage, a key event for primary nonfunction, which is characterized by microcirculatory disturbances and oxygen-dependent death of endothelial lining cells (Caldwell-Kenkel et al., 1988; Lemasters and Thurman, 1997). Indeed, in this study gentle organ manipulation dramatically increased the number of dead sinusoidal lining cells after 24 h of cold storage, assessed by nuclear staining with trypan blue (Fig. 6). Because endothelial cells comprise about 40% of sinusoidal lining cells, and Kupffer cells and stellate cells initially retain viability (Lemasters et al., 1995; Lemasters and Thurman, 1997), the exacerbated damage in manipulated livers shown here corresponds to virtually complete denudation of the endothelium (Fig. 6). Furthermore, removal of oxygen totally prevented trypan blue uptake in manipulated livers (see Results), confirming the oxygen-dependence of pathology.

Role of Kupffer Cells in Mechanisms of HarvestInduced Injury upon Reperfusion. To test the hypothesis that Kupffer cells are involved in mechanisms of harvest-related reperfusion injury, donors were pretreated with GdCl₃, a rare earth metal and Kupffer cell toxicant (Hardonk et al., 1992), or dietary glycine, a nonessential amino acid that prevents activation of Kupffer cells (Ikejima et al., 1997). These treatments prevented effects of manipulation on reperfusion injury (Fig. 6), suggesting a role for Kupffer cells (Bremer et al., 1994). This study demonstrates that manipulation causes alterations in microcirculation and hypoxia, which stimulates Kupffer cells that actually cause injury. Using pimonidazole as a marker for hypoxia, it is clear that hypoxia occurs to some degree in all manipulated livers, whether or not Kupffer cells are present; however, hypoxia was blunted when Kupffer cells were inactivated by GdCl₃ or glycine.

Conclusion and Clinical Implications. Collectively, data from this study clearly demonstrate that gentle manipulation of livers during organ retrieval increases oxygen-dependent injury after cold storage and reperfusion. Disturbed hepatic microcirculation in situ during harvest causes hypoxia and free radical formation via mechanisms involving Kupffer cells. This causes oxygen-dependent injury to endothelial cells upon reperfusion after cold storage. Effects of gentle organ manipulation can be prevented by depletion of Kupffer cells with GdCl₃, a rare earth metal, and glycine given to donors before organ harvest. If effects of organ manipulation are confirmed in humans, these donor pretreatments could improve the overall outcome of liver transplantation, because it would reduce the effect of physical organ manipulation, which is inevitable with standard harvesting techniques.