Pioglitazone promotes survival and prevents hepatic regeneration failure after partial hepatectomy in obese and diabetic KK-Ay mice

Authors

  • Tomonori Aoyama,

    1. Department of Gastroenterology, Juntendo University School of Medicine, Tokyo, Japan
    2. Sportology Center, Juntendo University Graduate School of Medicine, Tokyo, Japan
    Search for more papers by this author
  • Kenichi Ikejima,

    Corresponding author
    1. Department of Gastroenterology, Juntendo University School of Medicine, Tokyo, Japan
    2. Sportology Center, Juntendo University Graduate School of Medicine, Tokyo, Japan
    • Department of Gastroenterology, Juntendo University School of Medicine, 2-1-1 Hongo, Bunkyo-ku, Tokyo, 113-8421 Japan
    Search for more papers by this author
    • fax: (81)-3-3813-8862

  • Kazuyoshi Kon,

    1. Department of Gastroenterology, Juntendo University School of Medicine, Tokyo, Japan
    Search for more papers by this author
  • Kyoko Okumura,

    1. Department of Gastroenterology, Juntendo University School of Medicine, Tokyo, Japan
    Search for more papers by this author
  • Kumiko Arai,

    1. Department of Gastroenterology, Juntendo University School of Medicine, Tokyo, Japan
    Search for more papers by this author
  • Sumio Watanabe

    1. Department of Gastroenterology, Juntendo University School of Medicine, Tokyo, Japan
    2. Sportology Center, Juntendo University Graduate School of Medicine, Tokyo, Japan
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

Abstract

Pathogenesis of metabolic syndrome–related nonalcoholic steatohepatitis (NASH) involves abnormal tissue-repairing responses in the liver. We investigated the effect of pioglitazone, a thiazolidinedione derivative (TZD), on hepatic regenerative responses in obese, diabetic KK-Ay mice. Male KK-Ay mice 9 weeks after birth underwent two-thirds partial hepatectomy (PH) after repeated intragastric injections of pioglitazone (25 mg/kg) for 5 days. Almost half of the KK-Ay mice died within 48 hours of PH;however, mortality was completely prevented in mice pretreated with pioglitazone. In KK-Ay mice, bromodeoxyuridine (BrdU) incorporation to hepatocyte nuclei 48 hours after PH reached only 1%; however, pioglitazone pretreatment significantly increased BrdU-positive cells to 8%. Cyclin D1 was barely detectable in KK-Ay mice within 48 hours after PH. In contrast, overt expression of cyclin D1 was observed 24 hours after PH in KK-Ay mice pretreated with pioglitazone. Hepatic tumor necrosis factor alpha (TNF-α) messenger RNA (mRNA) was tremendously increased 1 hour after PH in KK-Ay mice, the levels reaching ninefold over C57Bl/6 given PH, whereas pioglitazone blunted this increase by almost three-fourths. Pioglitazone normalized hypoadiponectinemia in KK-Ay mice almost completely. Serum interleukin (IL)-6 and leptin levels were elevated extensively 24 hours after PH in KK-Ay mice, whereas the levels were largely decreased in KK-Ay mice given pioglitazone. Indeed, pioglitazone prevented aberrant increases in signal transducers and activators of transcription (STAT)3 phosphorylation and suppressor of cytokine signaling (SOCS)-3 mRNA in the liver in KK-Ay mice. Conclusion: These findings indicated that pioglitazone improved hepatic regeneration failure in KK-Ay mice. The mechanism underlying the effect of pioglitazone on regeneration failure most likely involves normalization of expression pattern of adipokines and subsequent cytokine responses during the early stage of PH. (HEPATOLOGY 2009.)

Nonalcoholic fatty liver disease (NAFLD) is one of the important manifestations of metabolic syndrome, a clinical entity comprising obesity, hypertension, glucose intolerance, and hyperlipidemia.1–3 NAFLD includes a spectrum of liver pathology ranging from simple steatosis to nonalcoholic steatohepatitis (NASH), which demonstrates progressive diseases, including advanced hepatic fibrosis and hepatocellular carcinoma.1–3 Pathogenesis of metabolic syndrome–related NAFLD/NASH involves a variety of environmental and nutritional factors as well as genetic susceptibilities. Among these, insulin resistance is believed to be the most profound metabolic abnormality.4 It is postulated that adipokines, a group of cytokines produced exclusively from adipose tissue (in other words, leptin, adiponectin, resistin, plasminogen activator inhibitor-1), regulate both metabolic balance and progression of hepatic disorder.4 The mechanisms underlying the disease progression of NASH involve several key biological responses such as excess oxidative stress, alteration in innate immunity, and abnormal tissue repairing responses. Especially, impairment of hepatic regeneration seems to be critical because regeneration failure often results in progression of hepatic fibrogenesis and subsequent carcinogenesis. Indeed, it has been reported that progression of fibrosis is associated with altered regeneration in NAFLD patients.5

It has been well documented that various types of hepatic steatosis and steatohepatitis demonstrate poor hepatic regeneration. For instance, poor hepatic regeneration has been demonstrated in experimental models of hepatic steatosis/steatohepatitis such as ob/ob mice and Zucker rats, which carry genetic defects of leptin and its receptor (ObR), respectively.6, 7 These observations suggest that leptin is an important regulator of liver regeneration; however, the role of adipokines in hepatic regenerative responses without defects in leptin/ObR genes has not been well elucidated. Given the evidence that genetic defects in leptin/ObR genes are seldom found in humans, more clinically relevant animal models are required to investigate the pathophysiology and experimental therapeutics of NAFLD/NASH.

KK-Ay mice are a cross-strain of diabetic KK mice8 and lethal yellow (Ay) mice, which carry mutation of the agouti(a) gene in mouse chromosome 2.9 KK-Ay mice develop maturity-onset obesity, dyslipidemia, and insulin resistance, in part because of the antagonism of melanocortin receptor-4 by ectopic expression of the agouti protein.9 Importantly, these mice present hyperleptinemia and leptin resistance without defects in the ObR gene, and the expression of adiponectin is conversely down-regulated.10, 11 The phenotype of KK-Ay mice, including altered adipokine expression, quite resembles metabolic syndrome in humans, indicating the potential usefulness of this strain as a model of metabolic syndrome–related NASH. Indeed, we have previously demonstrated that KK-Ay mice are more susceptible to experimental steatohepatitis induced by a methionine-deficient, choline-deficient diet.11

The current study aimed to clarify the causal relationship between liver regeneration and metabolic background related to obesity, insulin resistance, and expression of adipokines. Here we investigated the changes in hepatic regenerative response in KK-Ay mice after two-thirds partial hepatectomy (PH). Furthermore, we evaluated the effect of a thiazolidinedione derivative (TZD) pioglitazone, which improves insulin resistance through actions as a peroxisome proliferator–activated receptor (PPAR)γ agonist, on regeneration failure in these animals. Through this study, we tested the hypothesis that pioglitazone improves regenerative responses by modulating inflammation and aberrant adipokine expression in a model of metabolic syndrome–related steatohepatitis.

Abbreviations

ANOVA, analysis of variance; BrdU, bromodeoxyuridine; ELISA, enzyme-linked immunosorbent assay; IL, interleukin; JAK, Janus kinase; NASH, nonalcoholic steatohepatitis; mRNA, messenger RNA; NAFLD, nonalcoholic fatty liver disease; ObR, leptin receptor; PCNA, proliferating cell nuclear antigen; PH, partial hepatectomy; PPAR, peroxisome proliferator-activated receptor; RT-PCR, reverse transcription polymerase chain reaction; SEM, standard error of the mean; SOCS, suppressor of cytokine signaling; STAT, signal tranducers and activators of transcription; TNF-α, tumor necrosis factor alpha; TZD, thiazosidinedione derivative.

Materials and Methods

Animal Experiments and Operative Procedure.

Male KK-Ay and C57Bl/6 mice 8 weeks after birth were obtained from CLEA Japan Inc. (Tokyo, Japan). Mice were housed in air-conditioned specific pathogen–free animal quarters with lighting from 8:00 AM to 9:00 PM and given unrestricted access to a standard laboratory chow and water for 1 week before and during experiments. All animals received humane care in compliance with the experimental protocol approved by the Committee of Laboratory Animals according to institutional guidelines. C57Bl/6 mice, which are the grandparental strain of KK-Ay mice, were selected as nonobese and nondiabetic controls. Some KK-Ay mice were treated with 25 mg/kg pioglitazone (a gift from Takeda Pharmaceutical Co., Ltd., Tokyo, Japan) or vehicle by intragastric injection once daily for 5 days before operation. After overnight fasting, 70% PH was performed in the mice according to the Higgins and Anderson method.12 Mortality was observed up to 48 hours after PH. Mice were sacrificed by exsanguinations from inferior vena cava, and wet weight of the whole remaining liver was measured. Some mice were pulse-labeled with a single intraperitoneal injection of bromodeoxyuridine (BrdU; Sigma Chemical Co., St. Louis, MO; 50 mg/kg in phosphate-buffered saline) 2 hours before sacrifice, and liver specimens were fixed in buffered formalin for immunohistochemistry. Serum and liver samples were kept frozen at −80°C until assayed.

Immunohistochemistry.

For immunohistochemistry, formalin-fixed and paraffin-embedded tissue sections were deparaffinized and incubated with 3% H2O2for 10 minutes. To examine BrdU incorporation to hepatocyte nuclei, tissue sections were incubated with 2N hydrochloric acid for 30 minutes. After blocking with normal horse serum for 60 minutes, tissue sections were incubated with a mouse monoclonal anti-BrdU antibody (DakoCytomation Norden A/S, Glostrup, Denmark). After rinsing the primary antibody, the sections were incubated with secondary biotinylated anti-mouse immunoglobulin G antibody, and the specific binding was visualized with the avidin–biotin complex solution followed by incubation with a 3,3-diaminobenzidine tetrahydrochloride solution using Vectastain Elite ABC kit (Vector Laboratories, Burlingame, CA). BrdU-positive hepatocytes were counted in five 100× fields on each slide to determine the average number BrdU-labeling index (BrdU-positive hepatocytes/total hepatocytes). Expression of proliferating cell nuclear antigen (PCNA) in hepatocytes was evaluated similarly by immunohistochemistry as previously described elsewhere.13 Specimens were observed and photographed using a microscope (BH-2, Olympus Corp., Tokyo, Japan) equipped with a digital imaging system (VB-6010, Keyence Corp., Osaka, Japan).

Western Blot Analysis.

Whole-liver protein extracts were prepared by homogenizing frozen tissue in a buffer containing 50 mM Tris-hydrochloric acid (pH 8.0), 150 mM sodium chloride, 1 mM ethylenediaminetetraacetic acid, 1% Triton X-100, protease inhibitors (Complete Mini, Roche Diagnostics Co., Mannheim, Germany), and a phosphatase inhibitor Na3VO4 (50 μM, Sigma Chemical Co.), followed by a centrifugation at 10,400 g for 10 minutes. Protein concentration was determined by Bradford assay using Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA). Fifty micrograms protein was separated in 10% to 15% sodium dodecyl sulfate polyacrylamide gel electrophoresis and electrophoretically transferred onto polyvinylamide fluoride membranes. After blocking with 5% nonfat dry milk in Tris-buffered saline, membranes were incubated overnight at 4°C with rabbit polyclonal anti-cyclin D1, anti-phospho-signal tranducers and activators of transcription 3 (STAT3) (Tyr705) or anti-STAT3 (Cell Signaling Technology Inc., Beverly, MA), followed by a secondary horseradish peroxidase–conjugated anti-rabbit immunoglobulin G antibody (DakoCytomation Norden A/S). Subsequently, specific bands were visualized using the enhanced chemiluminescence detection kit (Amersham Pharmacia Biotech, Piscataway, NJ).

Measurement of Blood Glucose and Enzyme-Linked Immunosorbent Assay.

Blood glucose levels were measured enzymatically using standard glucose test strips and a glucometer (Glutest Sensor and Glutest Ace; Sanwa Kagaku Kenkyusho Co. Ltd., Nagoya, Japan). Serum insulin, adiponectin, interleukin (IL)-6, and leptin were determined using enzyme-linked immunosorbent assay (ELISA) kits (insulin and leptin: Seikagaku Corp., Tokyo, Japan; adiponectin: Otsuka Pharmaceutical Co.Ltd., Tokyo, Japan; IL-6: eBiocience, San Diego, CA) according to the manufacturer's instruction.

RNA Preparation and Real-Time Reverse Transcription Polymerase Chain Reaction.

Total liver RNA was prepared from frozen tissue samples by guanidium/cesium trifluoroacetete extraction method using Quick Prep total RNA extraction kit (Amersham Pharmacia Biotech). The concentration and purity of isolated RNA were determined by measuring optical density at 260 and 280 nm. Furthermore, the integrity of RNA was verified by electrophoresis on formaldehyde denaturing agarose gels.

For real-time reverse transcription polymerase chain reaction (RT-PCR), total RNA (1 μg) was reverse-transcribed using Moloney murine leukemia virus transcriptase (Super-Script II, Invitrogen Corp., Carlsbad, CA) and a deoxythymidine oligonucleotides [oligo(dT)12-18] primer (Invitrogen Corp.) at 42°C for 1 hour. Obtained complementary DNA (1 μg) was amplified using SYBR Premix Ex TaqTM (Takara Bio, Tokyo, Japan) and specific primers as appropriate. Primer sets for tumor necrosis factor alpha (TNF-α), suppressor of cytokine signaling 3 (SOCS3), and glyceraldehyde-3-phosphate dehydrogenase are shown in Table 1. After a 10-second activation period at 95°C, 40 cycles of 95°C for 5 seconds, and 60°C for 31 seconds, followed by the final cycle of 95°C for 15 seconds, 60°C for 1 minute, and 95°C for 15 seconds, were performed using ABI Prism 7700 Sequence Detection System (PE Applied Biosystems, Foster City, CA), and the threshold cycle (CT) values were obtained.

Table 1. Primer Sets for Real-Time RT-PCR
Gene (GeneBank Accession)Primer SequencesProduct Size
TNF-αForward: 5′-AAGCCTGTAGCCCACGTCGTA-3′122 bp
(NM_013693)Reverse: 5′-GGCACCACTAGTTGGTTGTCTTTG-3′ 
SOCS-3Forward: 5′-CAATACCTTTGACAAGCGGACTCTC-3′146 bp
(NM_007707)Reverse: 5′-TCAAAGCGCAAACAAGTTCCAG-3′ 
GAPDHForward: 5′-AAATGGTGAAGGTCGGTGTG-3′108 bp
(NM_001001303)Reverse: 5′-TGAAGGGGTCGTTGATGG-3′ 

Image and Statistical Analysis.

Densitometrical analysis was performed using Scion Image (version Beta 4.0.2, Scion Corp., Frederick, MD). Data were expressed as means ± standard error of the mean (SEM). Statistical differences between means were determined using Mann-Whitney rank sum test or analysis of variance (ANOVA) on ranks followed by a post hoc test (Student-Newman-Keuls all pairwise comparison procedures) as appropriate. Statistical significance for percent survival was determined by log-rank test. A value of P < 0.05 was selected before the study to reflect significance.

Results

Liver Regeneration After PH Is Impaired in KK-Ay Mice.

In this study, we first investigated the strain differences in liver regeneration after 70% PH between KK-Ay mice and C57Bl/6 mice as controls. Because impaired liver regeneration often results in death, we observed the mortality after PH in both strains of mice (Fig. 1A). All of the C57Bl/6 mice were alive after PH as expected; however, 50% of KK-Ay mice died within 48 hours after PH. In C57Bl/6 mice, the wet weight of the remnant liver was increased 2.1-fold within 48 hours; in contrast, the values were only 1.4-fold in KK-Ay mice (Fig. 1B). To evaluate DNA synthesis during liver regeneration, we next examined BrdU uptake into hepatocyte nuclei by immunohistochemistry (Fig. 1C-E). BrdU incorporation was observed in 22% of hepatocytes 48 hours after PH in C57Bl/6 mice; however, the levels only reached 1% in KK-Ay mice given the same procedure. To further confirm the impaired regeneration in KK-Ay mice, hepatic expression of cyclin D1 was detected by western blotting (Fig. 2). In C57Bl/6 mice, cyclin D1 expression was increased almost ninefold over basal levels 48 hours after PH; however, KK-Ay mice almost completely lacked increases in cyclin D1 after PH. Taken together, these findings clearly indicated that KK-Ay mice demonstrate extremely poor regenerating response in the liver after PH.

Figure 1.

Impaired liver regeneration after PH in KK-Ay mice. Both C57Bl/6 and KK-Ay mice underwent 70% partial hepatectomy (PH). Percentages of survival in C57Bl/6 controls (closed circle) and KK-Ay mice (closed square) up to 48 hours after PH are plotted (A; n = 14; *P < 0.05 versus C57Bl/6 by log-rank test). Wet weight of the liver was measured before and 48 hours after PH (B; n = 5; *P < 0.05, by Mann-Whitney rank-sum test). Representative photomicrographs of BrdU staining 48 hours after PH from C57Bl/6 (C) and KK-Ay mice (D) are shown (original magnification 100×). Mean BrdU indices are plotted (E; n = 5; mean ± SEM, *P < 0.05, by ANOVA on ranks and Student-Newman-Keuls post-hoc test).

Figure 2.

Hepatic cyclin D1 expression after PH. Experimental designs as in Fig. 1. Expression of cyclin D1 in the liver was detected by western blotting. Representative photographs of specific bands from five individual experiments are shown (A). The densitometrical results for cyclin D1 (expressed as fold over C57Bl/6 controls) are plotted (B; n = 5; mean ± SEM, ***P < 0.001, by one-way ANOVA and Student-Newman-Keuls post-hoc test).

Effect of Pioglitazone on Liver Regeneration After PH in KK-Ay Mice.

We next investigated the effect of pioglitazone on regeneration failure and mortality after PH in KK-Ay mice. Interestingly, the mortality after PH was completely prevented in KK-Ay mice given pioglitazone for 5 consecutive days before operation (Fig. 3A). Pioglitazone pretreatment significantly increased BrdU-positive hepatocytes in KK-Ay mice 48 hours after PH to 7.2% (Fig. 3B, C, F). Similarly, pioglitazone increased the percentages of PCNA-positive hepatocytes almost 10-fold in KK-Ay mice 48 hours after PH (Fig. 3D-E, G). Furthermore, overt expression of cyclin D1 was observed 24 hours after PH in KK-Ay mice when they were pretreated with pioglitazone (Fig. 4A, B), confirming that pioglitazone improves liver regeneration failure after PH in these mice.

Figure 3.

Effect of pioglitazone on liver regeneration after PH in KK-Ay mice. KK-Ay mice pretreated with pioglitazone or vehicle alone underwent 70% PH, and mortality was observed up to 48 hours later. Percentages of survival are plotted (A; n = 10; *P < 0.05 by log-rank test). Representative photomicrographs of BrdU (B, C) and PCNA (D, E) staining 48 hours after PH from KK-Ay mice pretreated with vehicle alone (left panels; B, D) and pioglitazone (right panels; C, E) are shown (original magnification 100×). Mean indices of BrdU (F) and PCNA (G) in each group are plotted (n = 5; mean ± SEM, ***P < 0.001, by Mann-Whitney rank-sum test).

Figure 4.

Effect of pioglitazone on cyclin D1 expression in the liver after PH in KK-Ay mice. Experimental designs as in Fig. 3. Expression of cyclin D1 in the liver was detected by western blotting. Representative photographs of specific bands from five individual experiments are shown (A). The densitometrical results for cyclin D1 (expressed as fold over C57Bl/6 controls) are plotted (B; n = 5; mean ± SEM, *P < 0.05, by one-way ANOVA and Student-Newman-Keuls post-hoc test).

Effect of Pioglitazone on Blood Glucose/Insulin Regulation After PH in KK-Ay Mice.

Because one of the characteristic phenotypes of KK-Ay mice is glucose intolerance, we monitored blood glucose levels during the time course of PH (Fig. 5A). Fasting blood glucose levels in KK-Ay mice are significantly higher than those in C57Bl/6 mice as expected, the values being 177.8 ± 3.9 and 80.8 ± 6.2 mg/dL, respectively. Pretreatment with pioglitazone decreased fasting blood glucose levels in KK-Ay mice to values reaching 122.2 ± 11.4 mg/dL. Blood glucose levels in C57Bl/6 mice were slightly increased after PH and maintained the levels around 118.8 ± 8.9 mg/dL up to 48 hours after PH. In sharp contrast, blood glucose levels in KK-Ay mice were drastically degreased after PH, the values being as low as 58.8 ± 8.1mg/dL at 48 hours. Interestingly, pioglitazone pretreatment paradoxically prevented this hypoglycemia with values of 113.6 ± 2.0 mg/dL at 48 hours after PH.

Figure 5.

Effect of pioglitazone on blood glucose and serum insulin levels after PH in KK-Ay mice. Blood glucose levels were detected enzymatically, and serum insulin levels were measured by ELISA. Mean blood glucose levels (A) and serum insulin levels (B) in C57Bl/6 mice (closed circle), KK-Ay mice (open square), and KK-Ay mice pretreated with pioglitazone (gray diamond) are plotted (n = 5; mean ± SEM, *P < 0.05 versus C57Bl/6 mice, #P < 0.05 versus KK-Ay mice in each time point by ANOVA on ranks and Student-Newman-Keuls post-hoc test).

Next, we measured serum insulin levels in these animals (Fig. 5B). Fasting serum insulin levels in KK-Ay mice were significantly higher than those in C57Bl/6 mice, the levels being 1.64 ± 0.25 and 0.13 ± 0.03 ng/mL, respectively (Fig. 5B). Pioglitazone pretreatment decreased fasting serum insulin levels in KK-Ay mice to the values of 0.43 ± 0.06 ng/mL. After PH, serum insulin levels in C57Bl/6 mice were slightly elevated in 3 hours to the values of 0.7 ± 0.05 ng/mL, and kept constant levels thereafter. In sharp contrast, serum insulin levels in KK-Ay mice were tremendously increased after PH, with maximum values of 16.6 ± 6.1 ng/mL 12 hours after PH. In KK-Ay mice pretreated with pioglitazone, increases in serum insulin levels after PH were significantly blunted as compared with those in mice without pioglitazone, with the peak values reaching only 4.13 ± 1.29 ng/mL at 24 hours.

Effects of Pioglitazone on Hepatic TNF-α Messenger RNA and Serum Adiponectin, IL-6, and Leptin Levels After PH in KK-Ay Mice.

Because TNF-α is believed to promote liver regeneration in the early phase,14, 15 hepatic TNF-α messenger RNA (mRNA) were measured by real-time RT-PCR (Fig. 6A). TNF-α mRNA levels in the liver before PH were quite low, and no difference was observed between two strains. After PH, hepatic TNF-α mRNA was swiftly increased with a single, small peak within 1 hour in C57Bl/6 mice as expected. In contrast, increases in TNF-α mRNA 1 hour after PH were greatly enhanced, with the maximal levels reaching almost ninefold over C57Bl/6 peaks, followed by the second peak 24 hours later. Pretreatment with pioglitazone to KK-Ay mice blunted the initial peak of TNF-α mRNA by approximately three fourths, and the second peak almost completely.

Figure 6.

Effects of pioglitazone on hepatic TNF-α mRNA and serum adiponectin, IL-6, and leptin levels after PH in KK-Ay mice. The levels of TNF-α mRNA in the liver (A) were determined by real-time RT-PCR, and serum adiponectin (B), IL-6 (C), and leptin (D) levels were measured by ELISA. These parameters in C57Bl/6 mice (closed circle), KK-Ay mice (open square), and KK-Ay mice pretreated with pioglitazone (gray diamond) are plotted (n = 5; mean ± SEM, *P < 0.05 versus C57Bl/6 mice, #P < 0.05 versus KK-Ay mice in each time point by ANOVA on ranks and Student-Newman-Keuls post-hoc test).

Next, we measured serum adiponectin levels by ELISA (Fig. 6B). Serum adiponectin levels were significantly lower in KK-Ay mice than in C57Bl6 mice before PH, as expected. In both strains of mice, serum adiponectin levels tended to decrease gradually after PH; however, the levels in KK-Ay mice were lower than those in C57Bl/6 mice throughout the time course after PH. Interestingly, hypoadiponectinemia observed in KK-Ay mice was reversed by pioglitazone pretreatment almost completely. Serum adiponectin levels in KK-Ay mice pretreated with pioglitazone were almost similar to C57Bl/6 controls in all times after PH.

IL-6 and the Janus kinase (JAK)-STAT pathway also play a pivotal role in liver regeneration16, 17; therefore, we measured serum IL-6 levels after PH by ELISA (Fig. 6 C). In C57Bl/6 mice, serum IL-6 levels peaked at 3 hours after PH with values reaching 2.3 ± 0.7 ng/mL, and then decreased thereafter. In KK-Ay mice, serum IL-6 levels were drastically elevated, with maximal values of 178.3 ± 38.9 ng/mL at 24 hours after PH. In contrast, pioglitazone pretreatment significantly blunted this increase in serum IL-6 after PH in KK-Ay mice almost completely.

Because leptin shares the same JAK-STAT pathway through Ob-Rb, which is predominantly expressed in sinusoidal cells in the liver,18 we also measured serum leptin levels after PH in KK-Ay mice (Fig. 6D). Serum leptin levels in untreated KK-Ay mice were remarkably high compared with the levels in C57Bl/6 controls, as expected. In KK-Ay mice pretreated with pioglitazone, serum leptin levels before PH were further increased to 45.4 ± 3.1 ng/mL. In C57Bl/6 mice, serum leptin levels were slightly elevated 3 hours after PH to values of 1.0 ± 0.2 ng/mL, and sustained until 48 hours. In sharp contrast, serum leptin levels were tremendously increased in KK-Ay mice, with maximal values of 176.9 ± 13.7 ng/mL 24 hours after PH. Pioglitazone pretreatment blunted the peak serum leptin levels in hepatectomized KK-Ay mice by almost half, with a shifted peak time to 12 hours after PH.

Effects of Pioglitazone on Phosphorylation of STAT3 and SOCS-3 mRNA in the Liver After PH in KK-Ay Mice.

Because remarkable differences were observed in serum IL-6 and leptin levels after PH, we further investigated the activation of the downstream JAK-STAT pathway by detecting phosphorylation of STAT3 (Tyr705) in the liver, using western blot analysis (Fig. 7A, B). In C57Bl/6 mice, the levels of STAT3 phosphorylation in the liver were increased 3 hours after PH, followed by a gradual decrease to near basal levels within 24 hours. Hepatic phospho-STAT3 levels in KK-Ay mice were also elevated 3 hours after PH but sustained persistently until 24 hours later. Pioglitazone pretreatment, however, inhibited this sustained increase in phospho-STAT3 in KK-Ay mice almost completely.

Figure 7.

Effects of pioglitazone on phosphorylation of STAT3 and SOCS-3 mRNA in the liver after PH in KK-Ay mice. Phosphorylation of STAT3 and SOCS-3 mRNA in the liver were detected by western blotting and real-time RT-PCR, respectively. Representative photographs of specific bands for phospho (p)-STAT3, and total STAT3 from five individual experiments are shown (A). The densitometrical results for pSTAT3/STAT3 ratio (B) and hepatic SOCS-3 mRNA levels (C) in C57Bl/6 mice (closed circle), KK-Ay mice (open square), and KK-Ay mice pretreated with pioglitazone (gray diamond) are plotted (n = 5; mean ± SEM, *P < 0.05 versus C57Bl/6 mice, #P < 0.05 versus KK-Ay mice in each time point by ANOVA on ranks and Student-Newman-Keuls post-hoc test).

Finally, we evaluated mRNA levels of SOCS-3, a negative regulator of the JAK-STAT pathway, in the liver by real-time RT-PCR (Fig. 7C). In C57Bl/6 mice, hepatic SOCS-3 mRNA levels were increased 3 hours after PH, followed by gradual decrease to near basal levels at 24 hours. In KK-Ay mice that received PH, the peak levels at 3 hours were almost 2.8-fold higher than those in C57Bl/6 mice, and the higher levels of SOCS-3 mRNA were thereafter sustained until 24 hours after PH. Pioglitazone pretreatment, however, prevented increases in hepatic SOCS-3 mRNA in KK-Ay mice almost completely throughout the time course after PH. Taken together, these findings indicated that pretreatment with pioglitazone normalizes the pattern of the JAK-STAT signal activation in KK-Ay mice after PH.

Discussion

Hepatic regeneration is one of the most typical and well-characterized tissue-repairing responses in the mammalian body.16, 17 Under normal conditions, two-thirds PH in rodents is not lethal, and the remnant liver spontaneously regenerates to regain the original organ size with proper functions within several days. However, hepatic regeneration is impaired in the fatty liver in various genetically obese animals such as ob/ob mice and Zucker (fa/fa) rats.6, 7 In the current study, we demonstrated that KK-Ay mice, which present obese and diabetic phenotypes, also lack normal hepatic regenerative response, with increased mortality after two-thirds PH (Figs. 1, 2). It is striking that pioglitazone, a TZD, not only prevented hepatic regeneration failure but also promoted survival after PH in KK-Ay mice almost completely (Figs. 3, 4).

Regarding the mechanisms of regeneration failure, KK-Ay mice demonstrated some outstanding abnormal responses after PH. First, KK-Ay mice showed excessive induction of TNF-α mRNA in the liver (Fig. 6A). Although TNF-α produced by Kupffer cells plays an important role in the initiation of normal hepatic regeneration,14, 19 excess induction of TNF-α in Kupffer cells might interfere with the regenerative responses. Indeed, we confirmed that BrdU incorporation in the liver 48 hours after PH in C57Bl/6 mice was blunted significantly by simultaneous treatment with recombinant TNF-α (data not shown). In KK-Ay mice, augmented TNF-α production seems to inhibit regeneration without inducing apparent hepatocellular injury, because no obvious increases in apoptotic hepatocytes were detected after PH by using M30 CytoDEATH immunohistochemistry (Roche Diagnostics Co.), which visualizes caspase cleavage product of cytokeratin 18 (data not shown). Conversely, adiponectin has been shown to inhibit lipopolysaccharide induction of macrophage activation both in vitro and in vivo.10, 20 Because KK-Aymice present significant hypoadiponectinemia (Fig. 6B), it is reasonable that Kupffer cells in these animals are more susceptible to certain stimuli such as gut-derived endotoxin (lipopolysaccharide). In contrast, mice pretreated with pioglitazone showed nearly normal serum adiponectin levels (Fig. 6B), which most likely prevented excess production of TNF-α from Kupffer cells during regeneration in KK-Ay mice. It is also possible that pioglitazone inhibits activation of Kupffer cells independent of adiponectin, because it has also been reported that TZDs inhibit activation of Kupffer cells in a direct manner.21, 22 Taken together, these findings support the hypothesis that the abnormal innate immune responses are critical for the regeneration failure in KK-Ay mice.

Second, KK-Ay mice showed remarkable increases in serum IL-6 and leptin levels after PH (Fig. 6C, D), which share the same intracellular signaling pathway involving JAK-STAT3. Given the evidence that IL-6, leptin, and the JAK-STAT signals play a pivotal role in liver regeneration,23–26 it is conceivable that dysregulation in JAK-STAT signaling is profoundly involved in regeneration failure observed in KK-Ay mice. The role of IL-6 and the JAK-STAT pathway in hepatic regeneration, however, appears to be complex. Knockout mice deficient in IL-6 have been shown to lack hepatocyte proliferation after PH,23 whereas mice lacking its receptor glycoprotein (gp)130 demonstrate only minor changes in cell cycle and DNA synthesis in hepatocytes after PH.27 In this study, we demonstrated that increases in phospho-STAT3, as well as the induction of SOCS-3 mRNA, in the hepatectomized liver were sustained in KK-Ay mice (Fig. 7). Furthermore, pretreatment with pioglitazone normalized phosphorylation of STAT3 (Fig. 7A, B) and increased cyclin D1 expression after PH in KK-Ay mice. Because the JAK-STAT pathway is involved in tissue protection in various kinds of hepatic injuries,28, 29 up-regulation of the JAK-STAT signaling in KK-Ay mice is considered as a stress-related protective response in the liver. Importantly, it has been reported that excess phosphorylation of STAT3 results in poor hepatic regeneration because of direct down-regulation of cyclin D1 expression26, 30, 31 and that higher SOCS-3 expression in hepatocytes lacking gp130-dependent Ras correlates with delayed hepatocyte proliferation.32 Collectively, these findings are consistent with the hypothesis that sustained activation of the JAK-STAT pathway leads to inhibition of cyclin D1, thereby causing regeneration failure in KK-Ay mice.

It is not clear whether the increase in mortality in KK-Ay mice is caused by hepatic failure after PH; however, it is obvious that KK-Ay mice developed progressive hypoglycemia after PH, even though they showed significant hyperglycemia before operation (Fig. 5A). In fact, KK-Ay mice showed extremely high serum insulin levels after PH (Fig. 5B), and it is considerable that the liver in KK-Ay mice failed to provide adequate amount of glucose by inhibiting glycogenesis and facilitating gluconeogenesis. In this aspect, the mortality after PH in KK-Ay mice, especially in the late phase, might be attributable to impaired glucose metabolism. Interestingly, pretreatment with pioglitazone, an insulin sensitizer, paradoxically prevents hypoglycemia after PH in KK-Ay mice (Fig. 5A). This is also consistent with the fact that pioglitazone blunted increases in serum insulin levels after PH (Fig. 5B). It is postulated that pioglitazone normalizes serum adiponectin levels, thereby avoiding irregular increases in serum insulin levels and subsequent hypoglycemia, thus in part promoting survival, after PH in KK-Ay mice.

PPAR-γ is a key nuclear receptor/transcription factor for transcriptional regulation of various genes related to glucose/lipid metabolism, immune responses, and tissue repair.33 Because TZDs, synthetic PPAR-γ ligands, not only improve insulin resistance but also inhibit activation of Kupffer cells21, 22 and transactivation of hepatic stellate cells,34–36 these chemicals are believed to be suitable for prevention/treatment of hepatic inflammation and fibrogenesis in NASH. Indeed, a placebo-controlled randomized study has demonstrated the therapeutic efficacy of pioglitazone on NASH in terms of both metabolic and histological improvement.37 It has been reported that TZDs rather inhibit liver regeneration after PH in normal rodents.38, 39 It is important to note, however, that TZDs are regularly applied for pathophysiological conditions involving insulin resistance. Our data added new experimental evidence that pioglitazone prevents hepatic regeneration failure in steatohepatitis. In metabolic syndrome–related steatohepatitis, such as in KK-Ay mice, pioglitazone most likely exerts beneficial effects, rather than inhibitory actions, on hepatic regeneration failure. It is postulated that therapeutic effects of pioglitazone on NASH in part involve normalization of tissue repairing responses in the liver.

In conclusion, KK-Ay mice, which present phenotypes resembling metabolic syndrome in humans, demonstrated severe hepatic regeneration failure with high mortality after PH. This regeneration failure was prevented partially, and the mortality was completely prevented by pretreatment with pioglitazone. The mechanisms underlying the regeneration failure in KK-Ay mice most likely involve alteration in innate immune responses and abnormal JAK-STAT signaling based on imbalance of adipokine expression. Pioglitazone improves the expression pattern of adipokines and normalizes innate immune responses and the aberrant JAK-STAT signaling, thereby facilitating regenerative responses in KK-Ay mice. These findings add a new aspect of therapeutic advantages of pioglitazone for treatment of NASH.

Ancillary

Advertisement