Use of a hanging-weight system for liver ischemic preconditioning in mice

Melanie L. Hart, Chressen Much, David Köhler, Jens Schittenhelm, Iris C. Gorzolla, Gregory L. Stahl, Holger K. Eltzschig

Abstract

Ischemic preconditioning (IP) represents a powerful experimental strategy to identify novel molecular targets to attenuate hepatic injury during ischemia. As a result, murine studies of hepatic IP have become an important field of research. However, murine IP is technically challenging, and experimental details can alter the results. Therefore, we systematically tested a novel model of hepatic IP by using a hanging-weight system for portal triad occlusion. This system has the benefit of applying intermittent hepatic ischemia and reperfusion without manipulation of a surgical clamp or suture, thus minimizing surgical trauma. Systematic evaluation of this model revealed a close correlation of hepatic ischemia time with liver damage as measured by alanine (ALT) and aspartate (AST) aminotransferase serum levels. Using different numbers of IP cycles and times intervals, we found optimal liver protection with four cycles of 3 min ischemia/3 min reperfusion as measured by ALT, AST, lactate dehydrogenase, and interleukin-6. Similarly, ischemia-associated increases in hepatic infarct size, neutrophil infiltration, and histological injury were maximally attenuated with the above regimen. To demonstrate transcriptional consequences of liver IP, we isolated RNA from preconditioned liver and confirmed transcriptional modulation of known target genes (equilibrative nucleoside transporters, acute-phase complement genes). Taken together, these studies confirm highly reproducible liver injury and protection by IP when using the hanging-weight system for hepatic ischemia and intermittent reperfusion. Further studies of murine IP may consider this technique.

  • preconditioning
  • liver protection
  • targeted gene deletion
  • murine
  • ischemia reperfusion

ischemia-reperfusion (I/R) injury is a phenomenon whereby cellular damage in an ischemic organ is elevated after the reestablishment of oxygen flow. The discovery of the cellular protective mechanism known as ischemic preconditioning (IP) has risen hopes that natural pathways could be activated to help prevent cells from death. IP is a technique whereby an organ is rendered resistant to the damaging effects of I/R by prior exposure to brief periods of vascular occlusion. This phenomenon was first described by Murry et al. (31) in the myocardium but has been subsequently observed in other tissues (42).

I/R of the liver is a clinically significant manifestation of several surgical procedures, such as liver transplantation, partial hepatic resection, hepatic tumor, or trauma repair (17). I/R injury is a major cause of morbidity and mortality following liver surgery and transplantation, and IP is a promising strategy to identify novel molecular targets for improving outcomes of liver surgery (17, 23). Although the benefits of IP in the liver have been suggested clinically (5, 6, 34), knowledge of the molecular mechanisms remains vague (3, 23, 35). Furthermore, despite the recent improvements in liver preservation and surgical techniques, hepatic I/R remains an important clinical problem (24, 33). However, recent advances in transgenic mice may help to unravel the molecular mechanisms of hepatic protection afforded by IP. Moreover, the use of tissue-specific knockout mice using the Cre/loxP system (15, 27) or chimeric (8, 41) mice may yield additional insight into the contribution of individual tissues or cell lines. For example, a recent study using chimeric mice suggests that I/R injury of the liver can be diminished via activation of adenosine receptors located on bone marrow-derived cells (7). Such experimental studies are important for the design of therapeutics, since drug design clearly depends on the tissue of interest.

Previous studies have demonstrated that liver IP can be performed in mice (23). However, these studies differ in their experimental approach, particularly with regard to details of intermittent occlusion of blood flow to the liver. Furthermore, occlusion is achieved by use of a surgical clamp that can be technically challenging. In particular, placement of the clamp without injury to the hepatic lobes is technically difficult, and this technique may be associated with further tissue trauma during IP by removing and replacing the clamp. Repeated removal and replacement of the surgical clamp also makes it difficult to produce exact reproducible durations of IP intervals since the lobes of the liver frequently shift and thus disrupt the view of the portal triad. Moreover, if the clamp is not strong enough or is misplaced due to slight movements, inadvertent reperfusion due to imperfect occlusion may affect the results. Based on these potential problems associated with portal triad occlusion by a clamp, we developed a model of IP using a hanging-weight system for intermittent occlusion of the portal triad. We hypothesized that this model would yield results with highly reproducible injury markers and hepatic protection by IP, since it combines the advantage of immediate and reliable occlusion/reperfusion with avoidance of tissue trauma due to manipulation of the hepatic lobes by reapplication of a clamp.

MATERIALS AND METHODS

Anesthesia and surgery.

All animal experiments were in accordance with the German guidelines for use of live animals approved by the Institutional Animal Care and Use Committee of the Tübingen University Hospital and the Regierungspräsidium Tübingen. C57BL/6 mice (8–12 wk old) were purchased from Charles River Laboratory and maintained on a 12:12-h light-dark cycle at an ambient temperature of 24°C and 60% humidity. Food and water were provided ad libitum.

Mice were anesthetized with pentobarbital sodium (70 mg/kg body wt ip) and placed on a temperature-controlled heated table with a rectal thermometer probe attached to a thermal feedback controller to maintain body temperature at 37°C. Mice were then secured in a supine position, with the upper and lower extremities attached to the table with removable tape.

Technique of portal triad occlusion.

Operations were performed under an upright dissecting microscope (model SZX7; Olympus, Hamburg, Germany). After a midline laparotomy and incision of the linea alba, the peritoneal cavity was exposed. The stomach and duodenum were caudally displaced using a wet cotton tip swab to expose the portal triad and caudate lobe. The caudate lobe was gently separated from the left lobe, and the right lobe was then slightly shifted to clearly view the portal triad above the bifurcation of right, median, and left lobes. We used a partial hepatic ischemia model that avoids mesenteric congestion by allowing blood flow through the right lobe. Once visually identified, the needle, followed by suture (7/0 nylon suture; Ethicon, Norderstedt, Germany), was placed under the portal triad, including the hepatic artery, hepatic vein, and common bile duct (Fig. 1, A and B). The left end of the suture was then placed over the right pole, whereas the right end of the suture was placed over the left pole, and a weight of 2 grams was attached to each end (Fig. 1, A and C). Although the weights were placed over the poles, the triad was immediately occluded, causing blood supply to the left, median, and caudate lobes of the liver to be interrupted (Fig. 1D). Successful occlusion was confirmed by visual inspection of pale blanching in the ischemic lobes (i.e., a change in color from red to a pale color). In contrast, the change of color immediately disappeared when the hanging weights were removed from the poles and the liver was reperfused. In an additional experiment, successful occlusion to the left, median, and caudate lobes was confirmed by injection of Evan's blue in the portal vein caudal to the liver (Fig. 1E). During surgery, the liver was kept wet and warm with a wet swab soaked with saline at 37°C. The surgical wound was closed using continuous sutures of the muscle wall and skin. After surgery, mice were allowed to recover for 3 h of reperfusion under a heating lamp. Sham-operated mice served as the control and underwent anesthesia, laparotomy, and exposure of the portal triad without I/R or IP. All animals survived the surgical procedure, and no complications were observed with portal triad occlusion using the hanging-weight system or in control mice.

Fig. 1.

Model of liver ischemic preconditioning (IP) using a hanging-weight system for portal triad occlusion. After a midline laparotomy, the stomach and duodenum were caudally displaced to expose the portal triad. A: the caudate lobe was gently separated from the left lobe, and the right lobe was then slightly shifted to clearly view the portal triad above the bifurcation of right, median, and left lobes. B–C: once visually identified, a 7/0 needle and nylon suture were placed under the portal triad, including the hepatic artery, hepatic vein, and common bile duct. The left end of the suture was then placed over the right pole while the right end of the suture was placed over the left pole, and a weight of 2 grams was attached to each end. D: while the weights were suspended by the poles, the triad was immediately occluded, causing blood supply to the left, median, and caudate lobes of the liver to be interrupted. Successful occlusion was confirmed by a change of color from red to a pale color. In contrast, the change of color immediately disappeared when the hanging weights were removed from the poles and the liver was reperfused. E: in an additional experiment, successful occlusion to the left, median, and caudate lobes was confirmed by injection of Evan's blue in the portal vein caudal to the liver. The ischemic (pale colored) area represents the area at risk (AAR).

Changes in liver perfusion.

Changes in liver perfusion were determined by injecting a solution of fluorescein isothiocyanate (FITC)-labeled dextran (0.06 g/ml; FITC-dextran; mol wt 4,000; Sigma, St. Louis, MO) in the carotid artery (200 μl dissolved in saline), followed by 200 μl of saline. FITC-dextran was allowed to perfuse through the entire mouse for ∼2 min to ensure maximal perfusion of the liver. The right followed by median lobe was quickly excised from the mouse before ischemia, at 30 min ischemia, or following 30 min ischemia and 15 min reperfusion and placed in formamide (100 mg tissue/ml; Sigma) for 2 h at 55°C. The fluorescence of FITC-dextran was determined using a fluorescence spectrophotometer with 485 nm excitation and 535 nm emission.

Preconditioning protocols.

First, the influence of different ischemia times (10, 20, 30, 40, and 50 min) on hepatic function was investigated. Then hepatic protection by IP was assessed in this model comparing different IP cycles consisting of 3 or 5 min ischemia and 3 or 5 min reperfusion, followed by 30 min ischemia and 3 h reperfusion.

Serum enzymatic and cytokine measurements.

Serum aspartate (AST) and alanine (ALT) aminotransferase levels were measured using a microtiter plate adaptation of a commercially available kit (Teco Diagnostics, Anaheim, CA). Lactate dehydrogenase (LDH) activity was measured using a kit purchased from Randox (Crumlin, United Kingdom). A commercially available enzyme-linked immunosorbent assay kit for measuring mouse interleukin (IL)-6 was obtained from R&D Systems (Minneapolis, MN), and tests were performed according to the manufacturer's instructions.

Triphenyltetrazolium chloride staining for determining percent infarct.

After mice underwent 30 min ischemia, followed by 3 h reperfusion (with or without prior IP), the ischemic area (area at risk, AAR, Fig. 1E) and the size of the infarct itself were determined using an adaptation of the previously described triphenyltetrazolium chloride (TTC) staining technique (14). In brief, the left and median lobes of the liver were carefully removed following reperfusion, washed in ice-cold 0.9% saline, placed on parafilm, frozen at −20°C for 30 min, and cut into 1-mm slices. The slices were then incubated with 1% TTC at 37°C for 30 min and fixed in 10% formaldehyde to allow differences between viable and necrotic tissue to become apparent. Thus TTC stains all cells red except those that are depleted in NADPH and therefore allows one to visualize the viable tissue that appears red vs. the infarcted tissue, which appears pale. Using computer-assisted planimetry, the boundaries of the whole lobe vs. the infarcted areas were defined and calculated using National Institutes of Health software Image 1.0. The infarct size and therefore degree of damage were calculated as the percentage of infarcted AAR compared with the whole lobe (=100%).

Myeloperoxidase activity.

Myeloperoxidase (MPO) analysis was performed using the Fluorescent Myeloperoxidase Detection Kit (Cell Technology, Mountain View, CA). In brief, 0.5 g of the median and left liver lobes was homogenized in 500 μl 1× Assay Buffer that was provided with the kit. Diamide (10 mM; Sigma), a glutathione inhibitor, was added to each sample and incubated for 30 min at room temperature (RT). Homogenates were centrifuged, and 0.5 ml solubilization buffer (provided in the kit) was added to the pellet. Samples were homogenized for an additional 30 s followed by 30 s of sonication and two cycles of freezing/thawing. After centrifugation, 20 mM 3-amino-1,2,4-triazole (Sigma), a catalase inhibitor, was added to the supernatant and incubated at RT for 1 h. A standard curve was prepared by using the standard provided in the kit, and the value for each sample was read from this curve. Fluorescence was measured at excitation of 550 nm and emission at 590 nm using a fluorescent plate reader.

Histological assessment of damage.

The median and left liver lobes were harvested and placed in cryomolds containing optimum cutting temperature Tissue-Tek embedding medium (Sakura Finetek Europe, Zoeterwoude, NL). Samples were immediately frozen in 2-methylbutane (Sigma), precooled in liquid nitrogen, and stored at −80°C until further processing. Frozen tissues were subsequently sectioned on a cryostat, 10 μm thick, collected on (+) charge slides, and stained with hematoxylin and eosin. Examination and scoring of each lobe was carried out by a pathologist who was blinded to the experimental group. A semiquantitative grading scale of 0–4, as outlined by Suzuki et al. (38), was used for the histopathological assessment of liver necrosis where 0 = no liver necrosis, 1 = single cell necrosis, 2 = up to 30% lobular necrosis, 3 = up to 60% lobular necrosis, and 4 = >60% lobular necrosis.

Gene regulation by hepatic IP.

To test the usefulness of this model and to assess transcriptional consequences of IP, we used real-time RT-PCR to demonstrate regulation of two different groups of genes: equilibrative nucleoside transporters (ENT) 1, 2, 3, and 4 and complement system genes including mannose-binding lectin (MBL)-A, MBL-C, C3, C5, and C9. We performed four cycles of IP (3 min ischemia, 3 min reperfusion) and excised the median and left lobes after different reperfusion times, followed by isolation of RNA, reverse transcription, and quantification with ENT- or complement-specific primers by real-time RT-PCR (iCycler; Bio-Rad Laboratories, Munich, Germany). In short, total RNA was isolated from liver tissue using the total RNA isolation NucleoSpin RNA II Kit according to the manufacturer's instructions (Macherey and Nagel, Dueren, Germany). For this purpose, tissue from the liver was homogenized in the presence of RA1 lysis buffer (Micra D8 homogenizer; ARTLabortechnik, Muellheim, Germany), and after filtration lysates were loaded on NucleoSpin RNA II columns, followed by desalting and DNase I digestion (Macherey and Nagel). RNA was washed, and the concentration was quantified. cDNA synthesis was performed by using reverse transcription according to the manufacturer's instructions (i-script Kit; Bio-Rad Laboratories). The primer sets for the PCR reaction contained 10 pM sense and 10 pM antisense with SYBR Green I (Molecular Probes, Leiden, Netherlands). The following murine primer sequences were used (sense and antisense, respectively): ENT1 (5′-CTTGGGATTCAGGGTCAGAA-3′, 5′-TCAGGTCACACGACACCAA-3′); ENT2 (5′-CATGGAAACTGAGGGGAAGA-3′, 5′-GTTCCAAAGGCCTCACAGAG-3′); ENT3 (5′AACCTGGGCTACAGGAGACA-3′, 5′-TAGAACAGGGAGCCCTGAGA-3′), ENT4 (5′-AGGGGGCGTTTATTCA-GTCT-3′, 5′-AGAACGGAGTTGGGGAC-TTT-3′), MBL-A (5′-CCAAAGGGG-AGAAGGGAGAAC-3′, 5′-GCCTCGTCCG-TGATGCCTAG-3′); MBL-C (5′-GACGTGACGGTGCCAAGGG-3′, 5′-CTTTCTGGATGGCCGAGTTTTC-3′); C3 (5′-CACCGCCAAGAATCGCTAC-3′, 5′-GATCAGGTGTTTCAGCCGC-3′); C5 (5′-CAAAGGATCCAGAAAAGAAGCCTGTAAACC-3′, 5′-CCTTAAGCTTCGTGCA-GCAGAACTTTTCATTC-3′); and C9 (5′-CCACCGAAGTACCTGAAAAG-3′, 5′-AGGAAAGTTGACCTCAGCAC-3′). The primer set was amplified by using increasing numbers of cycles of 94°C for 1 min, 58°C for 0.5 min, and 72°C for 1 min. Murine β-actin (sense primer, 5′-ACATTGGCATGGCTTTGTTT-3′ and antisense primer, 5′-GTTTGCTCC-AACCAACTGCT-3′) in identical reactions was used to control for the starting template.

Data analysis.

Hepatic injury score data are given as median (range), and all other data are presented as means ± SE. We performed statistical analysis using one-way analysis of variance to determine group differences. Hepatic injury was analyzed with a Kruskal-Wallis rank test.

RESULTS

Technique of portal triad occlusion.

Based on potential problems associated with portal triad occlusion by a clamp, we developed a model of hepatic IP using a hanging-weight system for intermittent occlusion of the portal triad (Fig. 1, AD). Once the portal triad was visually identified, the needle with suture was placed under the portal triad as indicated (Fig. 1, A and B). The sutures, with weights attached to each end, were then placed over opposite poles (Fig. 1, A and C). Once the weights were placed over the poles, the triad was immediately occluded, causing blood supply to the left, median, and caudate lobes of the liver to be interrupted, with blood flow continuing through the right lobe (Fig. 1D). Successful occlusion was confirmed by visual inspection of pale blanching in the ischemic lobes (i.e., a change in color from red to a pale color). In contrast, the change of color immediately disappeared when the hanging weights were removed from the poles and the liver was reperfused.

In an additional experiment, the portal triad was occluded, and ischemia was confirmed via injection of Evan's blue in the portal vein caudal to the liver (Fig. 1E). Evan's blue was unable to enter the left, median, and caudate ischemic lobes; therefore, these lobes appeared pale colored, while the nonischemic right lobe was intensely stained with Evan's blue. To quantify the observed changes in hepatic perfusion, mice were given FITC-dextran via a carotid artery catheter before ischemia (sham), at 30 min ischemia, or following 30 min ischemia and 15 min reperfusion, and the perfused-to-nonperfused FITC-dextran ratio of the liver lobes was calculated. As shown in Fig. 2A, during ischemia (30 min I) perfusion to the ischemic lobe was significantly decreased (P < 0.001) compared with the sham control or following 30 min ischemia and 15 min reperfusion (30 min I/15 min R).

Fig. 2.

Changes in hepatic perfusion and intestinal congestion. A: mice were given fluorescein isothiocyanate (FITC)-dextran via a carotid artery catheter before ischemia (sham), at 30 min ischemia (30 min I), or following 30 min ischemia and 15 min reperfusion (30 min I/15 min R) and the perfused-to-nonperfused FITC-dextran ratio of the liver lobes was calculated. Results are expressed as means ± SE of 3–4 mice/group. *P < 0.05 compared with sham or 30 min I/15 min R groups. B: for partial I, the portal triad was occluded above the bifurcation of the right, median, and left lobes, causing blood supply to the left, median, and caudate lobes of the liver to be interrupted. For total I, the portal triad was occluded below the bifurcation of the right, median, and left lobes, causing blood supply to all lobes of the liver to be interrupted.

Total hepatic ischemia has been shown to produce several pathological events, including splanchnic congestion, severe intestinal ischemia, mesenteric congestion, and systemic shock, thereby resulting in a high mortality rate (21, 22, 32, 37). Furthermore, occlusion to all lobes of the liver causes venous congestion in the mesenteric bed and compromises the intestinal mucosa, resulting in bacterial translocation and onset of systemic inflammatory response syndrome (4, 25). To confirm these results, we compared intestinal congestion following partial vs. total hepatic ischemia using the hanging-weight system. As shown in Fig. 2, in sharp contrast to mice that underwent total hepatic ischemia, mice that underwent partial hepatic ischemia revealed little to no intestinal/mesenteric congestion. Therefore, we used a partial hepatic ischemia model that avoids congestion to the intestine by allowing blood flow through the right lobe.

Influence of ischemia time on hepatic injury.

Because of the fact that previous studies in murine liver IP suggest dissimilar ischemia times (23, 35), we first tested the effect of different ischemia times on hepatic injury using the hanging-weight system model. Identification of an ischemia time resulting in a medium range of hepatic damage is important for the study of liver protection by IP, since it allows detection of changes in both directions, e.g., smaller degree of injury with hepatic IP or larger degree of injury with experimental therapeutics or a specific gene deletion. Elevated serum AST and ALT concentrations are commonly detected after hepatic injuries and therefore are reliable markers for assessing recent liver parenchymal cell membrane integrity and liver injury (20, 26). As shown in Fig. 3, ischemia times from 0 to 50 min followed by 3 h of reperfusion were associated with increased ALT and AST serum enzymatic levels. In fact, over the examined time range (0–50 min), hepatic ischemia time closely correlated with ALT and AST levels (R2 = 0.941, P < 0.001, R2 = 0.957, P < 0.001, respectively). Furthermore, our AST and ALT results compared with the results of others using relatively similar ischemia times in rodents (19, 36). Taken together, these results provide feasibility of using portal triad occlusion via a hanging-weight system for inducing highly reproducible and time-dose-dependent hepatic injury. Because we observed a “medium” degree of liver injury with 30 min of ischemia time, all further studies were performed using 30 min of portal triad occlusion followed by 3 h of reperfusion.

Fig. 3.

Effect of different I times on alanine (ALT) and aspartate (AST) aminotransferase in mice. Portal triad I was induced as indicated (10–50 min). After 3 h of R, ALT (A) and AST (B) were measured. Results are expressed as means ± SE of 6–12 mice/group. *P < 0.05 compared with 0 min I.

Influence of different IP cycles.

After having demonstrated reproducible hepatic injury with ischemia and reperfusion, we next used the hanging-weight occlusion system in experiments of liver protection by IP. To test the influence of different cycle numbers of IP, mice underwent two to five cycles of IP in which each cycle consisted of 3 or 5 min ischemia, followed by 3 or 5 min reperfusion. IP was followed by 30 min of ischemia and 3 h reperfusion.

There was a significant increase in ALT and AST following I/R compared with sham controls (Fig. 4, A and B, respectively). However, a decreasing trend in ALT and AST was apparent in mice that underwent increasing cycles of IP consisting of 3 min ischemia, followed by 3 min reperfusion before I/R, with a significant and maximal level of protection occurring with (3 min I/3 min R) × 4 cycles of IP. Interestingly, ALT and AST levels appeared to increase with 5 cycles of IP [(3 min I/3 min R) × 5 cycles]. As shown in Fig. 4B, mice that underwent three to five cycles of (5 min I/5 min R) IP were also afforded a significant protection for AST. However, this protection did not occur with ALT (Fig. 4A). In fact, all of the mouse groups that underwent IP consisting of 5 min ischemia followed by 5 min reperfusion showed no significant difference in ALT compared with the I/R group. Because of the fact that both ALT and AST showed similar protective effects for IP cycles of 3 min I/3 min R, but not for IP cycles of 5 min I/5 min R, we further investigated the influence of different IP cycle numbers using 3 min ischemia followed by 3 min reperfusion. In agreement with our ALT and AST results, mice undergoing 3 min I/3 min R showed a similar decrease in proinflammatory cytokine IL-6 and liver enzyme LDH serum levels with increasing cycle numbers (Fig. 5, A and B, respectively). In these studies, one or two cycles of IP (3 min I/3 min R) were not associated with a significant attenuation of LDH or IL-6 compared with unpreconditioned animals. In contrast, 3 cycles of IP resulted in a significant decrease in LDH (Fig. 5B), whereas three or four cycles of IP afforded significant liver protection as measured by IL-6 (Fig. 5A). As with ALT and AST levels (Fig. 4, A and B, respectively), LDH and IL-6 appeared to increase with five cycles of IP. Collectively, these results demonstrate that a maximum level of protection using the hanging-weight system occurs when mice are subjected to four cycles of IP consisting of 3 min ischemia and 3 min reperfusion.

Fig. 4.

Effect of different IP cycles on ALT and AST in mice. Mice underwent I/R alone or 2–5 cycles of IP in which each cycle consisted of 3 or 5 min I, followed by 3 or 5 min R. IP was followed by 30 min portal triad I and 3 h R. Sham mice underwent the same surgical procedure but without IP or I/R. Injury was assessed by measuring ALT (A) or AST (B). Results are expressed as means ± SE of 4 mice/group. P < 0.05 compared with sham (*) and compared with I/R (†).

Fig. 5.

Effect of 1–5 cycles of IP consisting of 3 min I and 3 min R on injury markers in mice. Mice underwent I/R alone or 1–5 cycles of IP consisting of 3 min I and 3 min R, followed by 30 min portal triad I and 3 h R. Sham mice underwent the same surgical procedure but without IP or I/R. Injury was assessed by measuring interleukin-6 (IL-6, A) or lactate dehydrogenase (LDH, B). Results are expressed as means ± SE of 6–12 mice/group. P < 0.05 compared with sham (*) and compared with I/R (†).

Liver protection by IP.

Because maximal protection was observed with four cycles of IP (3 min I/3 min R), further analysis of liver injury was assessed using this method, followed by 30 min ischemia and 3 h reperfusion (Fig. 6A). Results demonstrated that this robust improvement in hepatic injury by IP was also observed in additional tests. Thus, as shown in Fig. 6B, mice subjected to IP before 30 min of ischemia also showed a significant decrease in the percent infarction. Because leukocytes have been shown to be responsible for the acute inflammatory response during I/R (11, 13), we measured MPO as an indicator of neutrophil infiltration. Mice subjected to IP demonstrated a significant decrease in infiltration of neutrophils into the ischemic lobes compared with I/R mice (Fig. 6C). These studies demonstrate that the degree of damage as measured by percent infarct and MPO are reliable readouts for liver injury and protection by IP in this model.

Fig. 6.

IP model and IP protective effects on infarct size and myeloperoxidase (MPO). A: schematic illustration of the experimental protocol for hepatic IP. One IP cycle consisted of 3 min (′) I followed by 3 min of R. B: to document hepatic protective effects of IP, infarct size was measured using TTC staining. C: MPO was measured as an indicator of neutrophil infiltration. Results are expressed as means ± SE of 6–12 mice/group. P < 0.05 compared with sham (*) and compared with I/R (†).

As demonstrated in Fig. 7A, histological signs of ischemic injury were also attenuated by IP. Thus 30 min of ischemia resulted in hepatocyte liver necrosis (I/R). In contrast, mice with IP before ischemia showed only mild to moderate histological signs of injury similar to sham-operated control mice. In fact, semiquantitative histological analysis demonstrated a reduction in the Suzuki index (38) from three (range 3–4) without IP to two (range 2–3, Fig. 7B, P < 0.01) with IP.

Fig. 7.

Histological signs of liver injury are attenuated following IP. Representative hematoxylin and eosin (H&E)-stained sections (×200, A) and quantification of ischemic injury (B). Results are expressed as medians ± range; n = 6 mice/group. P < 0.05 compared with sham (*) and compared with I/R (†).

Comparison of clamping vs. hanging-weight system methods.

As the next step, we compared hepatic protection by IP via portal triad occlusion using the hanging-weight system vs. conventional clamping. As shown in Fig. 8, IP (4 cycles of IP consisting of 3 min I and 3 min R) using the hanging-weight system resulted in robust reduction of hepatic injury. In contrast, clamping of the portal triad (using a similar IP and ischemia protocol) was not associated with a statistically significant improvement of liver function. In fact, clamping the portal triad resulted in significantly less injury compared with occlusion using the hanging-weight system with 30 min ischemia alone (I/R). Furthermore, repeated clamping during IP appeared to cause more injury than I/R alone. These results suggest that occlusion of the portal triad using the hanging-weight system is more reliable than occlusion with a clamp.

Fig. 8.

Comparison of hepatic protection from I by IP using the hanging-weight system vs. clamping methods. IP was performed using the hanging-weight system for occlusion of the portal triad with 4 IP cycles (3 min I, 3 min R) before 30 min of I and 3 h R. Alternatively, the same IP protocol was used with clamping the portal triad. Injury was assessed by measuring ALT (A), AST (B), and LDH (C). Results are expressed as means ± SE of 4–8 mice/group. *P < 0.05 compared with I/R using the hanging-weight system.

Effect of genetic background.

Based on previous reports suggesting murine strain-specific differences in sensitivity to I/R injury (2, 9, 10, 29, 40), we compared hepatic injury using the hanging-weight system in two different mouse strains (C57BL/6 and SV129). As shown in Fig. 9, results demonstrated that four cycles of hepatic IP consisting of 3 min ischemia and 3 min reperfusion before 30 min ischemia [(3 min/3 min) × 4 IP] did not protect SV129 mice, as it did for C57BL/6 mice, from injury. However, three cycles of hepatic IP consisting of 5 min ischemia and 5 min reperfusion before 30 min ischemia [(5 min/5 min) × 3 IP] protected SV129 mice. Taken together, these data further demonstrate marked differences between different murine genetic backgrounds. In fact, these results underline the critical importance of performing control experiments in closely matched littermate controls of a similar genetic background.

Fig. 9.

Effect of genetic background on hepatic injury. To assess the influence of different murine genetic backgrounds (C57BL/6 or SV129) on hepatic injury, we compared mice that underwent I/R alone vs. mice that underwent either 4 IP cycles consisting of 3 min I/3 min R or 3 IP cycles consisting of 5 min I/5 min R before 30 min of I and 3 h R. Sham mice underwent the same surgical procedure but without IP or I/R. Injury was assessed by measuring ALT (A) or AST (B). Results are expressed as means ± SE of 4–12 mice/group. P < 0.05 compared with sham (*) and compared with respective I/R group (†).

Modulation of gene expression by liver IP.

As the last step, we measured gene regulation by IP in this model. To test the usefulness of this model and to assess transcriptional consequences of IP, we used real-time RT-PCR to demonstrate regulation of a group of genes known as ENTs, which have previously been shown to be hypoxia regulated (12). Although ENTs are highly expressed in the liver (28), very little is known about their physiological role in the liver. Under the hypothesis that IP would also repress these genes, we performed four cycles of IP and excised the liver after 180 min. Similar to what is known about ENT regulation by hypoxia (12), liver ENT1–4 are transcriptionally repressed 3 h after IP (P < 0.01 compared with control; Fig. 10A).

Fig. 10.

Repression of the equilibrative nucleoside transporters (ENTs) and complement genes by IP. Mice were subjected to 4 cycles of IP (3 min I, 3 min R) and 30 or 180 min R. Total RNA was isolated, and ENT1–4 (A) or mannose-binding lectin (MBL)-A, MBL-C, C3, C5, and C9 (B) mRNA levels were determined by real-time PCR. Data were calculated relative to β-actin and expressed as the degree of change in transcript relative to control (C) samples. Results are expressed as means ± SE of 3 mice/group. *P < 0.05 compared with control.

Complement activation following oxidative stress is an early event, and inhibition of complement activation or its components may offer tissue protection (1, 16). Tanhehco et al. (39) demonstrated that preconditioning of the heart reduces myocardial complement gene expression. To determine if preconditioning of the liver reduces hepatic complement gene expression, we transcriptionally assessed expression of MBL-A and MBL-C, as well as C3, C5, and C9 after four cycles of IP. All complement genes were significantly repressed 30 min after IP (Fig. 10B). Taken together, these results highlight the usefulness of this model to measure transcriptional effects of liver IP.

DISCUSSION

Liver protection from ischemia by IP is an area of intense investigation. Genetically engineered mice may provide additional insight into molecular mechanisms of hepatic protection by IP. Because of the technical difficulty associated with manually clamping the portal triad, we performed a systematic evaluation using a novel model for portal triad occlusion in mice, which specifically avoids the use of a clamp. By using a hanging-weight system, the portal triad is only distressed once throughout the entire surgical procedure, causing significantly less damage to the hepatic lobes. In addition, no hepatic or intestinal congestion occurs with this technique. In the present study, we demonstrate time-dependent and highly reproducible liver injury with ischemia and protection by IP using a hanging-weight system. We also demonstrate that this model can be used for the investigation of gene regulation by IP, since ENT1–4 and complement (MBL-A, MBL-C, C3, C5, and C9) message levels were repressed by IP. Taken together, the present study provides feasibility of the hanging-weight system for portal triad occlusion during IP, minimizing the variability and limitations associated with clamping. Thus this technique may be useful for future investigations involving the protective effects of IP in murine models.

Despite the growing number of reports investigating the mechanisms leading to hepatic protection by IP, the present understanding of the events that promote tolerance to I/R damage is still quite preliminary (3). Thus fundamental questions such as how IP works and which of the multiple hepatic cell types the trigger mechanism resides still remain unanswered. Similar to the present study, previous investigations have demonstrated hepatic protective effects of IP in mice (23, 35). These studies were performed by using a clamp to occlude the portal triad. Even with these successful studies of murine liver ischemia and reperfusion, we show that using a clamp-free system of portal triad occlusion may be superior and yield more reliable and reproducible results. In fact, it has been our experience with clamp systems in mice that it is hard to guarantee reliable portal triad occlusion during ischemia. Furthermore, more tissue trauma occurs by removing and replacing the clamp, especially with reapplication of the clamp during multiple IP cycles. The use of hanging weights that are in a remote location from the liver tissue, according to our observations, provides the advantage of reliable occlusion while preventing tissue trauma due to manipulation of the liver lobes by reapplication of a clamp. Perhaps this novel hanging-weight system model can be explored in murine and other small animal models to answer fundamental questions regarding the mechanism of IP and cells responsible for protection. The recent development of cell- or tissue-specific knockout and transgenic mice will also help to answer these central questions.

Despite many advantages associated with using targeted gene deletion in mice for studying liver protection by IP, some limitations of this approach have to be pointed out. Although it is likely that the use of genetically modified mice may yield important information about IP, biological compensation for gene deletion is known to occur (30). Furthermore, our results show strain-specific differences in response to hepatic IP when comparing C57BL/6 with SV129 mice. Moreover, it is appreciated that different responses to liver ischemia have been observed, not only with regard to different genetic backgrounds but also between different species (9, 18). For example, one study showed that intravenous administration of a gas-carrier contrast reagent used in ultrasound imaging caused intravascular expansion of the gas-carrier reagent in the portal vein and therefore ischemia to the liver and subsequent midzonal patterns of hepatic necrosis (9). Interestingly, the differences in the incidence of these lesions were highly dependent on the strain of mice and rats used. Furthermore, no lesions were found in guinea pigs and rabbits even upon repeated administrations, and dogs appeared to be a markedly less sensitive species than the rat and mouse. Of note, the tolerance of the murine liver against ischemia appears to be comparable to that of the human liver (6). However, such studies emphasize that, despite multiple similarities and molecular mechanisms, potential therapeutic targets identified in murine models cannot be directly transferred to the clinical setting but first require further testing in other models or species.

An additional limitation of the present study is its focus on the early phase of preconditioning. In addition to the early preconditioning phase, there is a late preconditioning phase that begins 12–24 h from the transient ischemia and lasts for 3–4 days (3). However, both phases of preconditioning can be initiated by the same stimuli and may partially share the same intracellular signaling pathways (3). To further unravel mechanisms of hepatic protection by IP, experimental approaches may need to examine both phases of preconditioning to fully understand the molecular mechanisms of protection afforded by IP.

In summary, the present study describes a novel technique of performing IP in an intact murine model by using a hanging-weight system for occlusion of the hepatic portal triad. This study demonstrates highly reproducible injury and liver protection by IP, minimizing the variability and potential damage associated with clamping of the portal triad. Investigators who consider studying hepatic protection by IP may benefit from this model.

GRANTS

This work was funded in part by National Institutes of Health Grants DK-067782 (M. L. Hart), HL-52886 (G. L. Stahl), HL-56086 (G. L. Stahl), DE-17821 (G. L. Stahl), and DE-16191 (G. L. Stahl), Fortune Grant F1211269 (M. L. Hart and H. K. Eltzschig), a European Society for Anesthesiology Research Grant D3008762 (M. L. Hart and H. K. Eltzschig), and a German Research Foundation Grant EL274/2-2 (H. K. Eltzschig).

Acknowledgments

We acknowledge Angelina Falk and Bernd Rolauffs for artwork and photography, respectively.

Current address for M. L. Hart: Dept. of Anesthesiology and Intensive Care Medicine, Center for Inflammation and Hypoxia, Tübingen University Hospital, Wilhelmstr. 56, Lothar Meyer Bau, 72074 Tübingen, Germany (e-mail: melaniehar@googlemail.com).

Footnotes

  • * M. L. Hart and C. Much contributed equally to this work.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract