SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Although it is becoming increasingly common to accept livers from older donors for transplantation, old livers are more damaged by hepatic ischemia and reperfusion injury (HIRI) than young livers. We hypothesized that this age-related susceptibility to HIRI is due to increased hepatocellular apoptosis driven by tumor necrosis factor α (TNFα). Young (6-week-old) and old (60-week-old) mice underwent 60 minutes of hepatic ischemia and increasing periods of reperfusion. TNFα was determined by enzyme-linked immunosorbent assay. Liver injury (enzyme release), apoptosis (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate–digoxigenin nick-end labeling staining, cytochrome C release, and caspase activation), and necrosis (hematoxylin and eosin staining) were assessed. We assessed the impact of apoptosis by blocking TNFα production or effect (pentoxifylline and TNFα receptor knockout), inhibiting apoptotic pathways (caspase inhibition), or imposing a hepatic protective strategy [glucose infusion with ischemic preconditioning (Glc/PC)]. In comparison with young livers, old livers subjected to HIRI had more pronounced liver aspartate aminotransferase release (6200 versus 3900 U/L, P = 0.02), necrosis (45% versus 25%, P = 0.03), and apoptosis with increased 30-minute TNFα release (19.02 versus 10.62 pg/mg, P = 0.03). Eliminating TNFα production reversed the effect of age, as did inhibition of apoptotic pathways with caspase inhibition. Glc/PC of old mice attenuated TNFα release (9.56 versus 19.02 pg/mg, P = 0.001) and age-related exaggerated HIRI and improved survival (60% versus 0%). In conclusion, the age-related susceptibility to HIRI is driven by an exaggerated induction of TNFα-dependent hepatocellular apoptosis. Targeting the apoptotic cascade has implications for the older donor liver population. Liver Transpl 15:1594–1604, 2009. © 2009 AASLD.

Until recently, it was thought that the liver is relatively unaffected by aging,1 but it is now clear that the older liver does not tolerate pathological insults as well as the younger liver.2–9 Clinical and animal studies provide strong evidence for an age-related susceptibility to hepatic ischemia and reperfusion injury (HIRI).10 A prospective randomized study in humans investigated ischemic preconditioning as a strategy of protection against reperfusion injury.11 Preconditioning had a strong protective effect in young patients, but it worsened reperfusion injury in elderly patients. Taken together, these results suggest that age contributes to clinical outcome following ischemic injury and that there are different mechanisms of reperfusion injury in old livers versus young livers.

A number of mechanisms may underlie the age-related susceptibility to HIRI. A previous study by members of our group demonstrated in a mouse model that HIRI is increased in old livers versus young livers.10 Aging was associated with decreased adenosine triphosphate (ATP) and glycogen levels in old livers versus young livers at baseline and after ischemia and reperfusion. Mitochondria from old livers had a decreased ability to produce ATP and consume oxygen. Pretreatment of old mice with intravenous glucose/preconditioning improved hepatic energy levels and protected the old mice against HIRI. Others have shown in several organ systems that aging is associated with altered mitochondrial function. For example, mitochondria in old hearts have altered cytochrome C binding sites, and this results in decreased respiratory chain function, less ATP generation, and increased injury following ischemia and reperfusion.12 Another group used a model of isolated rat hepatocytes and found that aging reduces the mitochondrial membrane potential and increases mitochondrial size; these changes were associated with disturbed gluconeogenesis.13 The heightened susceptibility of old livers to reperfusion injury was confirmed in a separate study5: the authors found higher neutrophil activation in old livers versus young livers with a suppression of nuclear factor kappa B in old mice, an effect that was interpreted as neutrophil dysfunction related to aging. Old mouse livers had decreased expression of the cytoprotective mediator heat shock protein 70, which has been strongly linked to antiapoptotic properties. Thus, mitochondrial function, inflammatory pathways, and apoptosis might play roles in the age-related susceptibility to HIRI.

There is evidence indicating that neutrophil production of important regulators of apoptosis such as nuclear factor kappa B, cytochrome c, and hydrogen peroxide (H2O2) is altered in the aging liver.5 In addition, the regulation of apoptosis is energy-dependent, and a low intrahepatic ATP content in aging hepatocytes might change the balance of proapoptotic and antiapoptotic mediators and contribute to increased HIRI in older livers. Whether apoptosis contributes to the age-related susceptibility to HIRI is unknown.

Tumor necrosis factor α (TNFα) is a key mediator of apoptosis after reperfusion injury,14, 15 and it is released by activated Kupffer cells after neutrophil infiltration and release of oxygen free radicals.16 TNFα binds to tumor necrosis factor receptor 1 (TNFR1) and activates the intracellular component of the apoptotic cascade, resulting in caspase 8 activation. Inhibition of TNFα signaling by TNFR1 knockout (KO) mice,14 deletion of Kupffer cells,17 inhibition of TNFα production by pentoxifylline,18, 19 and blockage of TNFα-induced caspase activation20 have all been demonstrated to decrease reperfusion injury in animal models. Although TNFα clearly can drive HIRI, how or whether it contributes to the additional injury seen in older livers has not been examined.

The aim of this study was to clarify the role of apoptosis in the exaggerated injury experienced by older livers following reperfusion injury and to determine the contribution of TNFα to the heightened effect of HIRI with aging. We further aimed to evaluate in old livers whether enhanced susceptibility to apoptosis could be modulated in a clinically relevant manner to reverse the age-related susceptibility to HIRI.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Animals

All experiments were performed in male C57/BL6 mice that were 6 weeks old (young mice) or 60 weeks old (old mice). The animals received a standard laboratory diet that was 12% fat, 28% protein, and 60% carbohydrates (5001 rodent diet, PMI, Inc., Brentwood, MO) with water and food ad libitum. Mice were kept under constant environmental conditions with a 12-hour light-dark cycle (light from 7:00 A.M. to 7:00 P.M.). All surgical procedures were performed under aseptic conditions between 7:00 A.M. and 11 A.M. to avoid circadian variations. Survival studies were performed with 10 mice per group to adjust for variations that might occur in long-term experiments. All other experiments used 5 mice per group. Euthanasia was performed with exsanguination. Animals received humane care in compliance with the Toronto General Research Institute Animal Care Committee.

Hepatic Ischemia

A model of segmental (70%) hepatic ischemia was used. Mice were anesthetized with inhalation of isoflurane (Pittman-Moore, Chicago, IL). After a midline laparotomy, the structures to the left of the porta hepatis (hepatic artery, portal vein, and bile duct), serving the left and median liver lobes, were occluded for 60 minutes. This method of partial hepatic ischemia prevents mesenteric venous congestion by permitting portal decompression through the right and caudate lobes.21 Reperfusion was initiated by removal of the clamp. The animals were again anesthetized at different periods after reperfusion, and liver biopsy specimens were taken for further evaluation. Mice receiving ischemic preconditioning received 10 minutes of ischemia and 10 minutes of reperfusion prior to 60 minutes of ischemic insult. Five mice were used in each group.

Glucose/Preconditioning Protocol

The glucose/preconditioning treatment was performed in old mice as previously described.10 Briefly, 0.1 cc of 20% D-glucose (D-Glc) was injected into the vena cava of the mice 1 hour prior to ischemia and just prior to reperfusion. L-Glucose is not metabolized by mice and was used in the control group. Preconditioning was performed by 10 minutes of clamping of the portal vein and hepatic artery with a microclamp followed by 10 minutes of reperfusion.

Parameters of Hepatocyte Injury

The degree of hepatic injury was assessed from measurements of the serum levels of aspartate aminotransferase (AST), an established marker of hepatocellular injury in the rodent liver.22, 23 AST serum levels were analyzed with a serum multiple analyzer (Ektachem DTSC II multianalyzer, Johnson & Johnson).

Terminal Deoxynucleotidyl Transferase–Mediated Deoxyuridine Triphosphate–Digoxigenin Nick-End Labeling (TUNEL) Staining

At various times after ischemia and reperfusion, mice were euthanized, and their livers were perfused with freshly prepared 4% paraformaldehyde in phosphate-buffered saline (pH 7.2) under a constant pressure of 10 cm H2O for 5 minutes through the portal vein. Each liver was cut into 3- to 5-mm sections that were stored in 70% alcohol after additional overnight fixation in 4% paraformaldehyde. Tissues were then incubated in 30% sucrose/phosphate-buffered saline, embedded in 7.5% gelatin, and finally frozen in isopentane submerged in dry ice and 95% alcohol slush. Sections of 5 μm were placed on silanized slides and were treated with terminal deoxynucleotidyl transferase from calf thymus in the presence of fluorescein–deoxyuridine triphosphate and deoxyribonucleotide triphosphate according to the supplier's recommended protocol (catalog 1767305, Boehringer Mannheim, Indianapolis, IN). This was followed by poststaining with horseradish peroxidase–conjugated anti-fluorescein antibody and then developed with diaminobenzidine/H2O2. Positive and negative controls were done with test sections pretreated with DNase I and staining without deoxynucleotide substrate, respectively. Thirty randomly selected fields were investigated per slide by a blinded investigator (M.S.) to determine the percentage of TUNEL-positive cells.

Caspase Assays

Caspase 3–like activity was determined by measurement of the proteolytic cleavage of the specific substrates N-acetyl-Asp-Glu-Val-Asp-7-amino-4-trifluoromethyl coumarin (Biomol, Plymouth Meeting, PA) and N-acetyl-Ile-Glu-Thr-Asp-7-amino-4-trifluoromethyl coumarin in the presence or absence of the specific caspase 3 aldehyde inhibitor based on the same amino acid sequence (Ac-DEVD-CHO and Ac-IETD-CHO, respectively; Biomol, Plymouth Meeting, PA). For these studies, liver tissue was quickly excised and sonicated in an assay buffer {1 mM ethylene diamine tetraacetic acid, 145 mM sodium chloride, 100 mM trishydroxymethylaminomethane, 0.1 mM dithiothreitol, 0.1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, and 10% glycerol}. The protein content was determined with the Bradford protein assay. The samples were diluted and incubated at room temperature with N-acetyl-Asp-Glu-Val-Asp-7-amino-4-trifluoromethyl coumarin substrate in the presence or absence of the inhibitor Ac-DEVD-CHO. 7-Amino-4-trifluoromethyl coumarin release was assayed over 2 hours with a fluorometer using 400-nm excitation and measuring 505-nm emissions. 7-Amino-4-trifluoromethyl coumarin release was expressed as arbitrary fluorescence units per milligram of protein after the reading in the inhibited sample was subtracted from that in the uninhibited sample.

Western Blot for Cytochrome c

Livers from transgenic and nontransgenic mice were collagenase-perfused via the portal vein, and the hepatocytes were separated from the nonparenchymal cells by density centrifugation. The tissue was homogenized in 10 vol/wt buffer A [20 mM 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid–potassium hydroxide (pH 7.5), 10mM potassium chloride, 1.5 mM magnesium dichloride, 1 mM ethylene diamine tetraacetic acid, 1 mM ethylene glycol tetraacetic acid, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, and 250 mM sucrose] with a Dounce grinder. Unbroken cells and nuclei were pelleted at 400g for 5 minutes at 4°C. The supernatants were further centrifuged at 12,000g to pellet the mitochondria. The supernatant was boiled with an equal volume of 2× sodium dodecyl sulfate sample buffer for 90 seconds and subjected to 4% to 20% sodium dodecyl sulfate–polyacrylamide gel electrophoresis. The proteins were transferred onto a nitrocellulose filter. Blots were probed with a rabbit polyclonal anti–cytochrome c antibody (Santa Cruz Biotechnology) followed by a secondary antibody conjugated to horseradish peroxidase and detected with enhanced chemiluminescence western blotting detection reagents (Amersham, Little Chalfont, United Kingdom).

TNFα Levels in Plasma

Whole blood (0.5 mL), obtained by vena cava puncture, was centrifuged at 4000g for 15 minutes at 4°C to obtain plasma, and 50-μL aliquots were snap-frozen for TNFα assay. Plasma TNFα levels were determined by enzyme-linked immunosorbent assay with a commercial kit (Quantikine M murine TNFα assay, R&D, Minneapolis, MN).

Determination of Necrosis

Histology was evaluated by a blinded investigator (M.S.). Hepatocellular necrosis was determined in hematoxylin and eosin–stained tissue sections by a point counting method with a semiquantitative scale as previously described.24 Thirty random fields were investigated per slide to determine the area of necrosis. In this study, only grade 3 injury with disintegration of hepatic cords was counted as necrosis.

Isolation of Peritoneal Macrophages

Peritoneal macrophages were harvested as described by Liu et al.25 Briefly, 2 cc of a 3% thioglycolate solution was injected inside the peritoneum of old and young C57BL/6 mice. Four days later, the peritoneal cells were harvested by lavage with 10 cc of Hank's buffered saline solution (HBSS). The extracted solution was centrifuged for 5 minutes at 1000 rpm, and red blood cells were lysed with 5 cc of endotoxin-free water for 3 minutes. The remaining cells were washed with Hank's buffered saline solution 3 times. Finally, the purity of the obtained macrophages was confirmed with CD11b staining by flow cytometry.

Statistical Analysis

The data were analyzed with the SPSS 11.5 statistic package. For continuous variables, a Mann-Whitney U test was used for the comparison of old and normal livers as indicated. A log rank analysis was performed for survival experiments. The results are presented as means and standard deviations and were considered significant at the level of P < 0.05.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Confirmation of Increased Reperfusion Injury in Old Mice Versus Young Mice

In the first set of experiments, we confirmed that reperfusion injury is worsened in old mice versus young mice. AST release and necrosis (hematoxylin and eosin staining) were determined as markers of hepatocyte death after 1 hour of ischemia and 1, 3, and 24 hours of reperfusion. At 1 hour of reperfusion, old and young mice had comparable AST levels (1100 ± 380 versus 1000 ± 320 mol/L, P = 0.9). Similarly, only minimal necrosis was present in both old and young livers at this time point (10% ± 0.05% versus 15% ± 0.03%, P = 0.9). In contrast, after 3 hours of reperfusion, old mice had significantly increased AST release in comparison with young mice (3900 ± 550 versus 6200 ± 650, P = 0.025; Fig. 1), which indicated more severe hepatocyte injury in the older animals. After 24 hours of reperfusion, necrosis had increased to 45% in the old animals, whereas only 25% of hepatocytes were necrotic in the young group (P = 0.03).

thumbnail image

Figure 1. AST release in young and old livers after 60 minutes of ischemia and 1 and 3 hours of reperfusion. After 1 hour of reperfusion, a minimal elevation of AST was present in young and old mice. In contrast, after 3 hours of reperfusion, old mice demonstrated significantly more AST release than the young control group (*P = 0.025, n = 5 in each group, Mann-Whitney U test). Abbreviation: AST, aspartate aminotransferase.

Download figure to PowerPoint

TNFα Release Is Increased in Old Livers Versus Young Livers After HIRI

TNFα is a critical regulator of hepatic apoptosis, has been shown to be an important mediator of the apoptotic pathway following reperfusion injury, and is an early and upstream component in the cascade of cell death.14 Several groups have shown that after hepatic ischemia, the serum TNFα level peaks within the first hour of reperfusion and returns to baseline within 3 hours.26, 27 TNFα release in the serum was determined as a possible early event in the induction of apoptosis following ischemia and reperfusion. After 1 hour of ischemia and 30 minutes of reperfusion, old mice had increased levels of TNFα in the serum in comparison with young mice (19.02 ± 5.03 versus 10.62 ± 5.4, P = 0.02; Fig. 2).

thumbnail image

Figure 2. TNFα in serum after 60 minutes of ischemia and 30 minutes of reperfusion in young and old mice. Old mice had significantly increased TNFα release in comparison with the young control group (P = 0.006, n = 5 in each group, Mann-Whitney U test). Abbreviation: TNF, tumor necrosis factor.

Download figure to PowerPoint

Because macrophages are an important source of TNFα after ischemia and reperfusion, we evaluated TNFα release of peritoneal macrophages derived from old and young mice. After 24 hours, H2O2 (40μM) exposure in cell culture macrophages from older mice versus younger mice produced an increased TNFα level in the supernatant (0.09 ± 0.002 versus 0.07 ± 0.003 pg/mL, P = 0.01).

Activation of the Apoptotic Pathway in Old Mice Versus Young Mice

TNFα, a key upstream regulator of hepatocellular apoptosis, is increased more in old livers than in young livers after HIRI. We next sought to determine if hepatocellular apoptosis was in fact increased more in the old livers versus the young livers following HIRI. Three lines of evidence suggest that this is the case. First, cytochrome c release into the cytosol is an important mechanism for the activation of caspases. After 1 hour of ischemia and 3 hours of reperfusion, old mice had increased cytochrome c in the cytosol in comparison with young mice (Fig. 3A). Second, we investigated activation of caspase 3 as an important mediator of programmed cell death. Caspase 3 activation after 1 hour of ischemia and 3 hours of reperfusion was significantly higher in old mice versus young mice (110 ± 15 versus 55 ± 11 AUF/mg, P = 0.04; Fig. 3B). Third, TUNEL staining of hepatocytes is a late event during apoptosis and represents DNA fragmentation. After 1 hour of ischemia and 3 hours of reperfusion, 82% of the hepatocytes in old mice were TUNEL-positive versus 50% in the young group (P = 0.04; Fig. 3C). Thus, there is strong evidence that HIRI-induced apoptosis is increased more in old mouse livers versus young mouse livers.

thumbnail image

Figure 3. (A) Cytochrome c in the cytoplasm and mitochondria of young and old mice after 1 hour of ischemia and 3 hours of reperfusion. (B) Caspase 3 activation after 60 minutes of ischemia and 3 hours of reperfusion in young and old mice. The caspase 3 activity was twice as high in old mice as in the young group (P = 0.03, n = 5 in each group, Mann-Whitney U test). (C) TUNEL staining as a late event of apoptosis. The percentage of TUNEL-stained hepatocytes was increased in old livers versus young livers after 60 minute of ischemia and 3 hours of reperfusion (P = 0.04, n = 5 in each group, Mann-Whitney U test). Abbreviation: TUNEL, terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate–digoxigenin nick-end labeling.

Download figure to PowerPoint

Inhibition of Apoptosis Protects the Old Liver Against Reperfusion Injury

If the increased activation of apoptosis is an important mechanism for the higher susceptibility of old mice to reperfusion injury, then the inhibition of apoptosis should provide a strong protective effect against the age-related susceptibility to HIRI in old mice. Three independent antiapoptotic strategies were investigated to determine if inhibition of apoptosis has a stronger protective effect against reperfusion injury in old livers versus young livers. Two of these were directed against the early (30-minute) increase in TNFα noted previously: TNFR1 KO mice and the administration of pentoxifylline. The third strategy was to pharmacologically inhibit the caspase pathway, a common downstream mediator of apoptosis.

First, we investigated hepatocyte injury and activation of the apoptotic pathway in young and old TNFR1 KO mice. After 1 hour of ischemia and 3 hours of reperfusion, AST release was dramatically decreased in older KO mice (6200 ± 810 versus 2800 ± 665 U/L, P = 0.02), whereas the effect was less prominent in younger KO mice (3900 ± 599 versus 3200 ± 550 U/L, P = 0.08; Fig. 4). Caspase 3 activity was largely decreased in old TNFR1 KO mice versus wild-type mice (115 ± 14.9 versus 31 ±3 AUF/mg, P = 0.02), whereas a smaller effect was observed in young TNFR1 KO mice versus wild-type mice (55 ± 4 versus 35 ± 4 AUF/mg, P = 0.05; Fig. 5). After 3 hours of reperfusion, TUNEL staining in old mice decreased from 82% ± 12% in the wild-type mice to 45% ± 9% in the TNFα KO mice, whereas TUNEL staining was reduced proportionally less in young mice from 45% ± 9% (wild-type) to 32% ± 3% (TNFR1 KO).

thumbnail image

Figure 4. AST release as a marker of hepatocyte injury after 60 minutes of ischemia and 3 hours of reperfusion in young and old mice with and without antiapoptotic strategies. AST release was significantly reduced in young mice treated with the pan-caspase inhibitor or the preconditioning/glucose protocol and in young TNFR1 knockout mice. No protection was observed in young mice with the application of PTx. Strong protective effects against reperfusion injury were found in old TNFR1 knockout mice and in old mice treated with PTx administration, caspase inhibition, or PC/Gluc (*P <0.05 in comparison with the young or old control group, n = 5 in each group, analysis of variance). Abbreviations: AST, aspartate aminotransferase; Casp Inh, pancapase inhibitor N-benzyloxycarbonyl-valyl-alanyl-aspartyl-fluoromethyl ketone; PC/Gluc, preconditioning and glucose treatment; PTx, pentoxifylline; TNFR1, tumor necrosis factor receptor 1.

Download figure to PowerPoint

thumbnail image

Figure 5. Reduction of caspase 3 activity in young and old mice treated with 3 different antiapoptotic strategies after 1 hour of ischemia and 3 hours of reperfusion injury in comparison with mice experiencing ischemia and reperfusion alone. Deletion of the TNFR1 receptor, administration of PTx, or Gluc-PC resulted in a dramatic decrease in caspase 3 activity in old mice, whereas the effect was marginal in young mice (*P < 0.05 and **P < 0.01 for young mice versus old mice, Mann-Whitney U test). Abbreviations: Gluc-PC, glucose/preconditioning pretreatment; KO, knockout; PTx, pentoxifylline; TNFR1, tumor necrosis factor receptor 1.

Download figure to PowerPoint

As a second antiapoptotic strategy, we administered pentoxifylline prior to ischemic injury. Pentoxifylline is an inhibitor of TNFα synthesis, and protective effects have been suggested in several models of reperfusion injury.14 In this series, administration of pentoxifylline decreased the TNFα levels in young and old mice, resulting in comparable levels following reperfusion (6.05 ± 1.1 versus 6.01 ± 1.2 pg/mL, P = 0.9). The level of TNFα reduction by pentoxifylline was stronger in the old group (19.02 ± 5.03 versus 6.01 ± 1.2 pg/mL, P = 0.0001), whereas only a moderate reduction was achieved in the young group (10.62 ± 5.4 versus 6.05 ± 1.1 pg/mL, P = 0.05). Pentoxifylline treatment resulted in strong protection against reperfusion injury in old mice, resulting in the postoperative AST level decreasing by more than half (6200 ± 810 versus 2500 ± 310 U/L, P = 0.025; Fig. 4). In contrast, no protective effect was observed in young mice with or without pentoxifylline treatment (3900 ± 599 versus 3780 ±488 U/L, P = 0.3). Caspase 3 activation was largely prevented by pentoxifylline treatment in old mice in comparison with untreated old animals (115 ± 14.9 versus 15 ± 2 AUF/mg, P = 0.01), whereas young mice demonstrated only a minimal reduction of the caspase 3 activity by pentoxifylline (55 ± 4 versus 45 ± 4 AUF/mg, P = 0.1; Fig. 5). Similarly, although pentoxifylline treatment decreased TUNEL staining from 82% ± 12% to 30% ± 5% in old mice (P = 0.02), no effect was observed in the young group (45% ± 9% versus 40% ± 6%, P = 0.4).

As a third antiapoptotic strategy, the pan-caspase inhibitor N-benzyloxycarbonyl-valyl-alanyl-aspartyl-fluoromethyl ketone was injected 1 hour prior to ischemia and reperfusion. Caspase inhibition in old mice resulted in a decrease in AST 3 hours after reperfusion from 6200 ± 810 to 2650 ± 450 U/L (P = 0.02), whereas in young mice, only a smaller reduction of AST release by caspase inhibition could be achieved (3900 ± 599 versus 2800 ± 250 U/L, P = 0.045; Fig. 4). TUNEL staining was decreased by caspase inhibition in old mice from 82% ± 12% to 15% ± 3% (P = 0.015), whereas in young mice, caspase inhibition decreased TUNEL staining from 45% ± 9% to 20% ± 2% (P = 0.02).

Inhibition of Apoptosis Improves Hepatocellular Necrosis Following HIRI

Because hepatocellular apoptosis can induce necrotic cell death at later stages, we also investigated how inhibition of apoptosis affected HIRI-induced hepatic necrosis. All 3 antiapoptotic strategies had marked effects on the induction of hepatic necrosis in old mice but not in young mice. After 24 hours of reperfusion, old TNFR1 KO mice tended to demonstrate less necrosis in comparison with the old wild-type (45% ± 4% versus 20% ± 3%, P = 0.35), whereas in young mice, no difference between wild-type and KO mice was observed (25% ± 2% versus 20% ± 2%, P = 0.5). Necrosis was significantly reduced by pentoxifylline administration in old mice (45% ± 9% versus 10% ± 2%, P = 0.01), but only a small effect was determined in young mice with or without pentoxifylline administration, and this difference did not reach statistical significance (25% ± 3% versus 15% ± 2%, P = 0.06). Necrosis was decreased by caspase inhibition from 45% ± 9% to 25% ± 5% in old mice (P = 0.03); there was no effect of caspase inhibition on necrosis in young animals (25% ± 3% versus 20% ± 2%, P = 0.7). These results highlight the importance of the age-related exaggeration in HIRI-induced apoptosis and hepatic necrosis.

The Protective Effect of Ischemic Preconditioning and D-Glc Is Mediated at the Level of Apoptosis

Members of our group have previously demonstrated that glucose infusion with ischemic preconditioning (Glc/PC) reverses the exaggerated HIRI in older murine livers.10 We therefore examined whether ischemic preconditioning (10 minutes of ischemia and 10 minutes of reperfusion) plus an intravenous D-Glc injection (0.1 mL of 20% D-Glc) in old mice prior to ischemia and prior to reperfusion resulted in altered levels of hepatocellular apoptosis. A similar regimen with an injection of L-glucose (0.1 mL L-glucose 20%) was used in control animals because rodents can metabolize only D-Glc. As we have previously reported, combined preconditioning/D-Glc treatment resulted in a dramatic AST decrease in old mice (6200 ± 810 versus 1800 ± 220 U/L, P = 0.02) and provided moderate protection in the young group (3900 ± 599 versus 2600 ± 230 U/L, P = 0.04; Fig. 4). Caspase 3 activity was reduced in old mice by ischemic preconditioning plus glucose treatment from 115 ±14.9 to 28 ± 3 AUF/mg (Fig. 5). In young mice, preconditioning plus glucose treatment diminished caspase 3 activation from 55 ± 4 to 25 ± 3 AUF/mg (Fig. 5). TUNEL staining decreased with glucose/preconditioning treatment from 82% ± 12% to 10% ± 2% in old mice (P = 0.03) and from 45% ± 9% to 15% ± 2% in young mice (P = 0.04; Fig. 6A,B). In old mice, glucose preconditioning treatment decreased necrosis from 45% ± 9% to 10% ± 1% in old mice (P = 0.02) and from 25% ± 2% to 15% ± 2% in the young group (P = 0.05; Fig. 6C,D). These results suggest that glucose/preconditioning reverses both the exaggerated HIRI response in older livers and the underlying hepatocellular apoptosis.

thumbnail image

Figure 6. TUNEL staining and hematoxylin and eosin staining in old mice after 60 minutes of ischemia and 3 hours of reperfusion. (A) In old mice without a protective strategy, most hepatocytes became TUNEL-positive. (B) In contrast, ischemic preconditioning plus glucose treatment prevented TUNEL staining in old mice. (C) The liver tissue of old mice without a protective treatment developed large areas of necrosis. (D) The pretreatment of old mice with ischemic preconditioning plus glucose injection prevented hepatocyte necrosis. Abbreviation: TUNEL, terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate–digoxigenin nick-end labeling.

Download figure to PowerPoint

We next investigated whether the protective effect of Glc/PC is mediated through alterations in the early mediators of apoptosis after reperfusion in old and young mice. First, TNFα was measured in the serum of young and old mice after 1 hour of ischemia and 30 minutes of reperfusion. Old mice had significantly increased TNFα levels 30 minutes after reperfusion in comparison with young mice (19.8 ± 2.2 versus 10.6 ± 2.4 pg/mg, P = 0.007). Pretreatment of old mice with glucose preconditioning resulted in a decrease of the TNFα release in old mice to levels comparable to those of young mice (9.56 ± 1.2 pg/mg; Fig. 7).

thumbnail image

Figure 7. TNF in serum after a sham operation, after 1 hour of ischemia and 30 minutes of reperfusion, and after 1 hour of ischemia and 30 minutes of reperfusion in mice pretreated with glucose preconditioning. Old mice had 2-fold higher TNFα levels after ischemia and reperfusion in comparison with young mice (P = 0.006). Although Gluc-PC reduced TNFα levels in old mice (P = 0.019), it had no effect in young animals (P = 0.2, Mann-Whitney U test). Abbreviations: Gluc-PC, glucose/preconditioning pretreatment; I/Rp, ischemia and reperfusion; TNF, tumor necrosis factor.

Download figure to PowerPoint

Does Blockage of Apoptosis in Old Mice Improve Survival Following Ischemia and Reperfusion?

Animal survival was compared in young wild-type, old wild-type, and old TNFα KO mice after 1 hour of total hepatic ischemia. In this model, ischemia was induced in 70% of the liver (left and median lobes) with resection of the 30% nonischemic liver lobes (caudate and right lobes) prior to reperfusion. One hour of total hepatic ischemia resulted in 90% animal survival in the young group, whereas all old wild-type mice died within 3 days (P = 0.001). Pretreatment of old wild-type mice with Glc/PC and deletion of the TNFR1 receptor in KO mice increased animal survival to 60% and 70%, respectively (P = 0.016 for old wild-type mice versus old wild-type mice ± Glc/PC; P = 0.001 for old wild-type mice versus old TNFR1 KO mice; Fig. 8).

thumbnail image

Figure 8. Survival of young wild-type mice, old wild-type mice, old wild-type mice pretreated with glucose preconditioning, and old TNFR1 knockout mice after 60 minutes of hepatic ischemia and reperfusion. All old mice died within 3 days, whereas 9 of 10 young mice survived permanently (P = 0.001, log rank). Glc/PC for old mice improved the survival of the old animals to 60% (P = 0.016 for old mice versus old mice ± Glc/PC, log rank). Old TNFR1 knockout mice had 70% long-term survival (P = 0.001 for old mice versus old TNFR1 knockout mice). Abbreviations: Glc/PC, glucose infusion with ischemic preconditioning; TNFR1, tumor necrosis factor receptor 1.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

In this study, we have demonstrated that TNFα-dependent alterations in hepatocellular apoptosis underlie the age-related susceptibility to HIRI. There is considerably more apoptosis following HIRI in older livers than in younger livers, and this is associated with increased TNFα release. Inhibition of the apoptotic pathway had a strong effect on reperfusion injury and survival in old mice but only a minor effect in young mice. Moreover, Glc/PC, which has been previously demonstrated by members of our group to reverse the age-related susceptibility to HIRI, appears to mediate its protective effects via decreases in TNFα release and downstream apoptotic pathways. These results suggest that inhibiting mediators of apoptosis might have beneficial effects in modulating the age-related susceptibility to HIRI.

There is emerging evidence that older livers are less tolerant of warm and cold ischemia and subsequent reperfusion. We recently demonstrated that aging is associated with decreased hepatic mitochondrial function and decreased intrahepatic ATP levels, resulting in increased reperfusion injury.4 In the present study, we asked whether apoptosis or necrosis is the predominant form of cell death in old and young livers following reperfusion injury, and we conclude that apoptosis is the predominant mechanism in the increased susceptibility to permanent liver damage in old livers subjected to HIRI.

Apoptosis and necrosis were initially considered to be 2 distinct pathways, with apoptosis representing an active, ATP-requiring form of cell suicide and necrosis reflecting passive cell death with rupture of the cellular membranes and inflammation. Apoptosis is a highly regulated ATP-dependent mechanism, and we have previously demonstrated that aging is associated with decreased intrahepatic ATP levels.10 In this study, we determined that despite the decreased ATP levels in older livers, apoptosis is a predominant feature of cell death after ischemic injury. We have to assume that the lower ATP content in the older liver does not directly affect the apoptotic cascade but probably affects the regulation of cell death mechanisms. It is possible that an ATP level below a critical threshold facilitates the activation of apoptotic mediators and inhibits survival mechanisms. Glc/PC increases the hepatic ATP content10 and might affect cell death regulation. In the last 5 years, most groups have agreed that necrotic cell death and apoptotic cell death can overlap in later stages, resulting in a mixed picture of necro-apoptosis. However, although the final stage of cell death often contains features of both apoptosis and necrosis, the specific and early induction of apoptotic mediators distinguishes the active process of apoptosis from the passive process of necrosis. Our results indicate that 1 fundamental difference between young and old livers is that the initiation of apoptosis by nonparenchymal cells is increased in old livers and is driven by the enhanced release of TNFα.

There are several mechanisms that might underlie the heightened release of TNFα after reperfusion injury in old mice versus young mice. TNFα is mainly released by Kupffer cells after ischemia and reperfusion, and it is possible that aging directly affects Kupffer cells, resulting in increased TNFα release. Hilmer et al.28 investigated Kupffer cell function in young and old rats. In their series, aging was associated with increased Kupffer cell density in the liver parenchyma and increased phagocytic activity of Kupffer cells in older rats. Similarly, we determined that peritoneal macrophages obtained from old mice versus young mice had increased TNFα release following H2O2 stimulation. This finding was confirmed by de la Fuente et al.,29 who found that after stimulation with lipopolysaccharide, macrophages from older mice versus younger mice had significantly increased TNFα production. Furthermore, lipopolysaccharide stimulation resulted in a stronger release of oxidants and increased macrophage adherence in the old group versus the young group. These studies support our finding that TNFα-dependent apoptosis is an important mechanism for increased reperfusion injury in the old liver and suggest that the effects of aging on macrophages could be an important underlying mechanism of the decreased tolerance of old livers against ischemic injury.

In contrast to our findings, Okaya et al.5 did not detect increased TNFα release in old mice versus young mice after hepatic ischemia and reperfusion. This discrepancy may be due to differences between the experimental models used in the 2 studies. One fundamental difference is the length of ischemia: Okaya et al. used a longer ischemia time than that in our study (90 versus 60 minutes). The longer ischemia time induced a considerably more severe degree of hepatic injury following reperfusion, which might have resulted from different patterns of Kupffer cell activation and different mechanisms of endothelial cell and hepatocyte death. Indeed, during the ischemic phase, anaerobic glycolysis and ATP hydrolysis lead to decreased tissue pH and ATP reserves.30–32 The availability of ATP and the degree of acidosis are important factors for the activation of the apoptotic death cascade versus necrotic death cascade: differences in the length of ischemic time will affect cell and tissue outcomes.33 In addition, Okaya et al. determined TNFα levels 1, 4, and 8 hours after reperfusion, whereas our study focuses on early TNFα release 30 minutes after reperfusion. There is no consensus about the best time point for TNFα measurement following ischemia and reperfusion, and different time points have been used by various groups.34–39 In our model, caspase activation and DNA fragmentation are present 90 minutes after reperfusion in sinusoidal endothelial cells and 3 hours post-reperfusion in hepatocytes. We believe that TNFα levels present soon after reperfusion best reflect its function as a mediator of reperfusion injury, whereas at later time points, TNFα levels might reflect the degree of ongoing inflammatory response to the injury.

Kupffer cell activation after ischemia and reperfusion is a downstream effect of a complex nonparenchymal cell cascade, and additional effects of aging on other cell types cannot be excluded. Ischemia and reperfusion are associated with adhesion of neutrophils, lymphocytes, and platelets to the hepatic sinusoidal endothelial cells. Subsequently, neutrophils infiltrate the sinusoidal tissue, resulting in the production of reactive oxygen species, activation of Kupffer cells, and release of TNFα. It is possible that aging affects bone marrow–derived cells, such as lymphocytes, neutrophils, and platelets, as important components of the nonparenchymal cell cascade. Okaya et al.5 reported that neutrophils derived from old mice had an increased ability to generate H2O2 in comparison with neutrophils from young animals. The differences in H2O2 formation between neutrophils from old and young mice were even more pronounced after neutrophil stimulation with N-formyl-methionyl-leucyl-phenylalanine. Similarly, platelets are also affected by the aging process, and like neutrophils, they are important mediators of the nonparenchymal response to HIRI. Goubareva et al.40 found that aging is associated with decreased nitric oxide synthesis and lower nitric oxide levels in platelets from old humans versus young humans. This difference was associated with increased monocyte-platelet aggregation in cells derived from older subjects versus younger subjects. These effects of age on neutrophils and platelets might also contribute to the increased TNFα release in older livers subjected to HIRI.

Our results demonstrate that blocking TNFα signaling or downstream mediators of the apoptotic cascade is equally effective in reversing the age-related effects of HIRI. Of the 4 strategies affecting hepatocellular apoptosis that we applied to reperfusion injury in old mice, the most effective was Glc/PC. Glc/PC treatment strongly decreased the TNFα levels in old mice to levels identical to those of the young group, and this resulted in strong protection against hepatocyte death. We have previously demonstrated that glucose/preconditioning has additional effects, such as increasing the hepatic ATP content, which could improve the resistance of old hepatocytes to prolonged ischemia. In addition, members of our group41 and others42, 43 have described improved hepatic microcirculation and decreased expression of endothelial adhesion molecules by ischemic preconditioning alone. The second strategy used in this series, pentoxifylline administration, is well known to facilitate microcirculation and inhibit TNFα synthesis. Although these and other mechanisms are likely to contribute to the effect of each strategy, the fact that all 4 strategies intersect at the level of inhibiting apoptosis argues that apoptosis is of fundamental importance in driving the age-related susceptibility to HIRI.

In conclusion, aging is associated with increased activation of the apoptotic pathway after ischemia and reperfusion, and this activation is driven by increased TNFα release. Strategies blocking TNFα signaling were strongly protective against reperfusion injury in old mice. These results suggest that the performance of older livers exposed to ischemia and reperfusion particularly benefits from protective strategies against TNFα signaling and apoptosis, and they argue for a clinical trial of the same.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES