Male Sprague-Dawley rats (weighing 200~250 g) purchased from Hyo-Chang Science (Daegu, Korea) were used in this study. Prior to the experiment, all rats were fasted overnight with free access to water. The care, maintenance, and treatment of all animals used in this study complied with protocol 2008-11 issued by the Institutional Animal Care and Use Committee of Yeungnam University School of Medicine.

A model of 70% partial hepatic ischemia was used in this study.(7) After being anesthetized with intraperitoneal ketamine (100 mg/kg) and xylazine® (10 mg/kg), each rat underwent a midline laparotomy, and the hepatoduodenal ligament was dissected. Ischemia was induced in median and left liver lobes by clamping the portal triad (portal vein, hepatic artery, and bile ducts) with a microvascular clamp. Livers were subjected to 120 minutes of warm ischemia. BEIs were measured at 0.12 KHz using a LCR meter (GS-4311B, Ando, Japan) every 10 minutes (n=6), as previously described.(3) The distance between the 2 platinum electrodes (length 25 mm, diameter 0.45 mm) inserted into the median lobe was 10 mm. The electrodes were coated with Teflon except for a 2 mm region at their tips. A heat mattress and transparent acryl case were used to maintain rat body temperature and ambient humidity.

Approximately 100 mg of fresh liver tissue from the median lobe (n=6) was excised every 30 minutes during the 120 minute ischemic period and immediately placed into ice-cold 10% HClO₄ buffer and homogenized. The ATP contents of liver tissue were measured as described by Khan.(8) Briefly, homogenates were centrifuged at 4,500 rpm for 10 minutes at 4℃. A 500 µl aliquot of supernatant was then neutralized with 200 µl of 2.5 mol/L KOH, and the precipitate was removed by centrifugation at 4,500 rpm for 5 minutes at 4℃. 10 µl of supernatant was then mixed (1:9) with 10 mmol/L of Tris-acetate buffer, and ATP relative luminescence units were measured using ATP Assay Kits (ENLITEN® ATP Assay System Bioluminescence Detection Kit for ATP, Promega, Madison, WI, USA) by using a luminometer (TD-20/20, Turner Designs, Sunnyvale, CA, USA).

A microdialysis technique was used to determine recovery rate and cation concentrations in ECF. The microdialysis probe (CMA 30 Linear Microdialysis Probe, CMA/Microdialysis, Solna, Sweden) with a cuprophane membrane (length 10 mm, molecular cut off 6,000 Daltons, outer diameter 0.24 mm) was used. The inlet of the probe was connected to a perfusion pump (CMA 102 Microdialysis Pump, CMA/Microdialysis AB, Solna, Sweden). To remove the glycerol in which the probe was originally packaged, it was placed in 70% aqueous ethanol and perfused with Hartmann's solution at a rate of 8 µl/minute for 10 minutes.

The construction of Zero net flux (ZNF) curve: The probe used had a dialysis membrane, which enables ion movement. After perfusing the microdialysis probe with 4 different perfusates, depending upon the ion concentration gradient, we can have 4 different values which enable to construct the ZNF curve. For example, when the ion concentration in the dialysate is higher than that in the perfusate, we can have a positive value on the ordinate, conversely if the ion concentration in the dialysate is lower than that in the perfusate, we have a negative value on the ordinate. When no ions diffuses from ECF or probe (y=0), the value is equal to the ion concentration in the ECF (x value) Ion recovery was calculated from the slope of the line in the ZNF curve.(9)

ZNF curves for [Na⁺] and [K⁺] in Hartmann's solution: 4 microdialysis probes were placed in Hartmann's solution, and perfused with 4 known different concentrations of perfusates (0, 75, 150, 225 mM for Na⁺ and 0, 5, 10, 15 mM for K⁺) (Table 1). The flow rate through the dialysis probe was increased to 8µl/minute for 10 minutes to ensure that liquid was flowing, and then reduced to a flow rate of 1µl/minute. After a 60-minute equilibrium period, dialysates were collected in 1.5 ml e-tubes every 30 minutes (n=6). Recoveries of Na⁺ and K⁺ from Hart-mann's solution were calculated from the slope of the line in the ZNF curve.

ZNF curves for [Na⁺] and [K⁺] in liver tissue ECF in vivo prior to ischemia: Prior to inserting the microdialysis probe into liver tissue, a rat was anesthetized and the abdomen was opened. Four microdialysis probes were inserted as described by the manufacturer (CMA/Microdialysis AB, Solna, Sweden). Briefly, the introducing needle was past through the median lobe of liver and the probe was slid through the needle end of the introducer. While ensuring that the membrane was located in the middle of the introducer and the introducer needle was carefully withdrawn. The 4 microdialysis probes were perfused with 4 known concentrations of perfusates (0, 75, 150, 225 mM for Na⁺ and 0, 5, 10, 15 mM for K⁺), respectively (Table 1). The flow rates were same to the above explanation in Hartmann's solution. After a 60-minutes equilibrium period, dialysates were collected in 1.5 ml e-tubes for 30 minutes (n=6). The dead space between the tip of the dialysis probe and the end of the outlet tube was 15µl. Sample collection was allowed for 15 minutes to compensate for the dead space volume in the microdialysis probe. Recoveries of Na⁺ and K⁺ from ECF in liver tissue in vivo were calculated from the slope of the line in the ZNF curve.

Measurement of [Na⁺] and [K⁺] in liver tissue ECF in vivo: After perfusing 4 known perfusates, 4 dialysates were collected during the following ischemic periods (0, 0~30, 30~60, 60~90, and 90~120 minutes). At each ischemic period, we can draw different ZNF curve. Depending on the ZNF curve at each ischemic time period, we obtained the concentrations of Na⁺ and K⁺ in liver tissue ECF in vivo (x value when y=0). Perfusates and dialysates were analyzed for Na⁺ and K⁺ using a Radiometer (ABL800 FLEX, Radiometer America Inc, Westlake, OH, USA) to construct the ZNF curves.

Median lobe liver tissue (n=6) was excised every 30 minutes during the 120 minute ischemic period, fixed in 10% formalin, dehydrated, and embedded in paraffin. Four micron-thick sections were stained with hematoxylin and eosin (H&E) and subjected to terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL) assays (ApopTag Peroxidase in situ Apoptosis Detection Kit, Chemicon, NY, USA). Hepatocyte diameters were measured using the Image J program provided by National Institutes of Health.

All data are expressed as means±SEs. Data were analyzed using SPSS v. 14.0. Mean values were compared using ANOVA with repeated measures to determine the significances differences at different time points during ischemia. P-values of less than 0.05 were considered statistically significant.

In our previous study,(3,4) we found that the BEI of hepatic tissue increased significantly during ischemia at low frequency versus high frequency. In the present study, BEI (kΩ) at 0.12 KHz were 3.1±0.2, 5.3±0.6, 6.8±0.7, 8.1±0.3, 9.1±0.3, 9.7±0.4, 10.1±0.5, 10.4±0.6, 10.8±0.7, 10.8±0.8, 11.1±0.9, 11.2±0.9, and 11.3±0.9 every 10 minutes in 120 minutes of ischemia, respectively. BEI increased gradually during 60 minutes of ischemia and then tended to plateau (Fig. 1).

The ATP contents decreased to below 20% of baseline after 60 minutes of ischemia, but did not change significantly thereafter. ATP (µmol/g wet liver tissue) contents were 2.65±0.34, 0.61±0.07, 0.45±0.07, 0.25±0.09, and 0.22±0.09, at 0, 30, 60, 90, and 120 minutes of ischemia, respectively (Fig. 2).

Fig. 3 shows the ZNF curves for Na⁺ and K⁺ in Hartmann's solution. After perfusing the 4 different perfusates, 4 dialysates were collected and analyzed, and then depending upon the ion concentration gradients 4 points were plotted to construct ZNF curve. The linear equation of the Na⁺ curve was y=-0.6693x+87.8 R²=0.9849 and that of the K⁺ curve was y=-0.748x+3.86 R²=0.9905. The concentrations of Na⁺ and K⁺ in Hartmann's solution calculated using the ZNF curve were 131.18 mM and 5.16 mM, respectively. These values were close to real concentrations in Hartmann's solution (Table 1). The recoveries of Na⁺ and K⁺ were 66.93% and 74.8% as calculated using the slope of the line in the ZNF curves.

Fig. 4 shows of ZNF curves for Na⁺ and K⁺ in liver tissue ECF in vivo prior to ischemia. According to the method described above, 4 points were plotted to construct ZNF curve. For Na⁺, the linear equation was y=-0.224x+33.7 R²=0.9931, and or K⁺ the linear equation was y=-0.244x+1.83 R²=0.9818. Hepatic extracellular [Na⁺] and [K⁺] values calculated from the ZNF curve were 150.4 mM and 7.5 mM. Recoveries of Na⁺ and K⁺ were 22.4% and 24.4% as calculated from the slope of the line in the ZNF curve. Recoveries of Na⁺ and K⁺ in liver tissue ECF in vivo were significantly less than those in Hartmann's solution.

Concentrations of Na⁺ and K⁺ in ECF in liver tissue in vivo were calculated using the ZNF curves made at each time point during ischemia (Fig. 5). [Na⁺] (mM) were 150.4±3.8, 128.9±5.1, 97.8±10.6, 90.7±4.9, and 88.7±5.0, and [K⁺] (mM) were 7.5±0.3, 22.1±0.7, 34.3±5.5, 40.4±7.0, and 48.3±1.7, at 0, 0~30, 30~60, 60~90, and 90~120 minutes of ischemia, respectively (Fig. 5).

BEI (kΩ) at 0.12 KHz were 3.1±0.2, 8.1±0.3, 10.1±0.5, 10.8±0.8, and 11.3±0.9 at 0, 30, 60, 90, and 120 minutes of ischemia, respectively (Fig. 6). [Na⁺]+[K⁺] (mM) were 157.8±3.8, 151.4±5.9, 135.7±7.6, 131.0±6.2, and 137.0±6.3 at 0, 30, 60, 90, and 120 minutes of ischemia, respectively (Fig. 6). There was inverse symmetry between BEI changes and [Na⁺]+[K⁺].

Hepatocyte diameter increased by about 20% of normal after 60 minutes of ischemia, but did not change significantly thereafter (Fig. 7). No significant differences histological differences were detected by H&E or TUNEL staining during the ischemic period (image data not shown). No apoptotic or necrotic changes were observed during 120 minutes of ischemia.