Disrupted signaling and inhibited regeneration in obese mice with fatty livers: implications for nonalcoholic fatty liver disease pathophysiology

Disrupted Signaling and Inhibited Regeneration in Obese Mice With Fatty Livers: Implications for Nonalcoholic Fatty Liver Disease Pathophysiology SHI QI YANG,1 HUI ZHI LIN,1 ALOKE K. MANDAL,2 JIAWEN HUANG,1 AND ANNA MAE DIEHL1

The impaired regenerative capacity of fatty livers might promote the progression of nonalcoholic fatty liver disease (NAFLD). To identify mechanisms involved, regenerative responses were compared in normal mice and ob/ob mice (a model for NAFLD) after partial hepatectomy (PH). We hypothesized that the usual PH activation of oxidant-sensitive, growth-regulatory kinase cascades would be abnormal in fatty hepatocytes, which have adapted to chronic oxidant stress, and expected that this might interfere with the induction of proliferative- and stress-related genes. The normal coordinated induction of Jun N-terminal kinases (Jnks) and extracellular regulated kinases (Erks) does not occur after PH in ob/ob mice, which cannot activate Jnks but can superinduce Erks. Jnk inhibition is associated with enhanced activation of Akt, which inhibits phosphoenolpyruvate carboxykinase (PEPCK) induction, causing severe hypoglycemia and increased lethality in the ob/ob group. Activation of nuclear factor ␬B (NF-␬B) is also inhibited, but liver damage is increased only modestly, perhaps because Akt-regulated survival factors are protective. Despite enhanced Erk activity, induction of cyclin D-1, an NF-␬B target gene, is abolished and this, together with hyperphosphorylated signal transducer and activator of transcription-3 (Stat-3) and reduced adenosine triphosphate (ATP) levels, arrests fatty hepatocytes in G1. Thus, in mice with NAFLD that have adapted hepatocyte signaling mechanisms to survive chronic oxidative stress, the cellular response to an acute regenerative stimulus is altered. This contributes to NAFLD pathophysiology by in-

Abbreviations: NAFLD, nonalcoholic fatty liver disease; PH, partial hepatectomy; BrdU, bromodeoxyuridine; PCNA, proliferating cell nuclear antigen; ATP, adenosine triphosphate; NF-␬B, nuclear factor ␬B; Stat-3, signal transducer and activator of transcription-3; Jnk, Jun N-terminal kinase; Erk, extracellular regulated kinase; EMSA, electrophoretic mobility shift assay; TNF-␣, tumor necrosis factor ␣; IL-6, interleukin 6; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; mRNA, messenger RNA; SAPK, stress-activated protein kinases; MAPK, mitogen-activated protein kinases; cdk, cyclindependent kinases; INK, inhibitor of cyclin dependent kinases; PEPCK, phosphoenolpyruvate carboxykinase. From the Departments of 1Medicine and 2Surgery, The Johns Hopkins University, Baltimore, MD. Received April 23, 2001; accepted July 10, 2001. Supported by NIH ROI AA10154 and DK3457 (A.M.D.). A.K.M.’s current address is the Division of Transplant Surgery, Department of Surgery, Hennepin County Medical Center, 701 Park Avenue, Minneapolis, MN 55415-2810. Address reprint requests to: Anna Mae Diehl, M.D., The Johns Hopkins University GI Division, 912 Ross Bldg., 720 Rutland Street, Baltimore, MD 21205. E-mail: [email protected]; fax: 410-955-9677. Copyright © 2001 by the American Association for the Study of Liver Diseases. 0270-9139/01/3404-0012$35.00/0 doi:10.1053/jhep.2001.28054

hibiting proliferation, increasing injury, and limiting function in fatty livers. (HEPATOLOGY 2001;34:694-706.) Obesity and type 2 diabetes are risk factors for nonalcoholic fatty liver disease (NAFLD),1-3 a spectrum of liver pathology that resembles alcohol-induced liver damage.4 Fatty liver is the earliest and most prevalent stage of NAFLD.5 Hepatic steatosis generally has a benign outcome,6 but some individuals with this lesion develop progressive liver injury and eventually experience significant liver-related morbidity and increased mortality.1 The mechanisms that promote disease progression in NAFLD are poorly understood. It has been suggested that more advanced stages of NAFLD might be triggered when an acute inflammatory insult, i.e., “second hit,” is superimposed on hepatic steatosis.7 Because healthy livers typically regenerate and recover completely from acute inflammation,8 this suggests that the normal regenerative response to injury might be impaired in fatty livers. Hepatic regenerative capacity is often assessed experimentally by monitoring the response to PH.8 Liver regeneration after two thirds (partial) hepatectomy (PH) is inhibited in obese, diabetic Zucker fa/fa rats with fatty livers,9 supporting the concept that the liver’s ability to regenerate is decreased in NAFLD. However, the mechanism(s) that impair liver regeneration in fatty livers have not been identified. The present study uses PH as a tool to compare the regenerative activation of oxidant-sensitive, growth-regulatory protein kinase cascades in genetically obese, ob/ob mice, an animal model for NAFLD,10 and lean control mice. After PH, these enzymes are normally activated by cytokines and growth factors,11,12 and their interactions help to orchestrate the changes in hepatocyte gene expression13-19 that are required to maintain hepatocyte viability,20,21 induce proliferation,19,22-24 and preserve tissue-specific functions18,19 during the regenerative response. We hypothesize that the normal, acute oxidant-mediated, activation of the mitogen- and stress-activated protein kinases will be disturbed in fatty livers, which over-produce oxidants chronically.25 If this concept is confirmed, then secondary objectives are to identify possible mechanisms for this dysregulated signaling and to evaluate its impact on hepatocyte viability, liver-specific gene expression, and animal survival. MATERIALS AND METHODS Partial Hepatectomy Experiments. Fifty young adult (aged 8-10

weeks) male lean and 50 age- and gender-matched, ob/ob C57BL-6 mice were purchased from Jackson Laboratories (Bar Harbor, ME), housed in a temperature-controlled animal facility with 12-hour light-dark cycles, and given ad libitum access to water and pellet type

694

HEPATOLOGY Vol. 34, No. 4, 2001

chow. At the onset of the PH experiments, 2 lean mice and 3 ob/ob mice were sacrificed to obtain baseline (i.e., pre-PH) liver tissue. Seventy percent PH was performed in the remaining 48 lean mice and 47 ob/ob mice according to the method of Higgins and Andersen26 after metaphane anesthesia. One lean mouse died intraoperatively. There was no operative mortality in the ob/ob group. Approximately half (25 lean mice and 25 ob/ob mice) of the mice were permitted to survive until 24 to 36 hours after PH so that blood glucose concentrations and animal survival could be monitored during liver regeneration. Blood was obtained from the tail veins before PH and at 24 and 36 hours post-PH, and glucose concentrations were assessed immediately with a glucometer. During the postoperative monitoring period 7 of 25 (28%) ob/ob mice died. All deaths occurred from 8 to 36 hours after PH, with the majority occurring between 12 and 24 hours post-PH. Two hours before sacrifice, all surviving mice were injected intraperitoneally with bromodeoxyuridine (BrdU) to evaluate hepatocyte DNA synthesis. At the time of sacrifice, liver tissues were fixed in buffered formalin. Subsequently, these formalin-fixed samples were embedded in paraffin and then sectioned and evaluated for hepatic nuclear BrdU incorporation by light microscopy using a Zeiss microscope. Coded sections from each mouse were inspected by an observer who counted the total number of hepatocyte nuclei and the number of BrdU (⫹) nuclei in 10 microcope fields (⫻400)on each section to determine the average number BrdU-labeling index (BrdU ⫹ hepatocytes/total hepatocytes) in each mouse. The mean (⫾ SEM) labeling index for each group was derived from data obtained from the 25 surviving lean mice and 18 surviving ob/ob mice. Hepatocyte proliferating cell nuclear antigen (PCNA) expression was evaluated on other sections by using a similar approach. Liver damage was assessed on hematoxylin-eosin–stained sections by counting foci of inflammation, hemorrhage, and necrotic- (i.e., ballooned) or apoptotic- (i.e., shrunken with condensed nuclear chromatin) appearing hepatocytes at ⫻200 magnification. Livers were considered “injured” if any of these lesions were detected. Injury was scored as 0 (none), 1⫹ (rare focal hepatocyte death, inflammation or hemorrhage), 2⫹ (focal liver cell death, inflammation, and/or hemorrhage in 25%-50% of the fields), 3⫹ (focal liver cell death, inflammation, and/or hemorrhage in ⬎50% of the fields). The other 22 to 23 mice in each group were sacrificed at various intervals (i.e., 0.5, 1, 6, 24, and 36 hours) after PH to obtain liver tissue for subsequent evaluation of hepatic kinases, transcription factors, gene expression, and adenosine triphosphate (ATP) content. Four to 5 mice per group were sacrificed at each of these time points in the lean group. A similar number of ob/ob mice were sacrificed at the first 3 time points, but delayed post-PH mortality in the ob/ob group reduced the numbers of surviving ob/ob mice to 4 and 2 at 24 and 36 hours, respectively. All animal studies fulfilled NIH and Johns Hopkins University guidelines for humane animal experimentation. Liver Protein Isolation, Immunoblot, and Immune Complex Assays. Liver tissue from each mouse was used to obtain whole liver protein and nuclear protein. Whole liver homogenates were prepared by homogenizing frozen tissue in homogenization buffer (50 mmol/L HEPES, pH 7.5, 150 mmol/L NaCl, 1.5 mmol/L MgCl2, 10% glycerol, 5 mmol/L EGTA, 1% Triton X-100) on ice for 10 minutes, followed by centrifugation in a microfuge for 15 minutes. Nuclear proteins were prepared from the same tissues according to the method of Lavery and Schibler27 as we have described.25 Proteins were isolated from lean and ob/ob mouse liver tissues concurrently, and the protein concentration in each extract was determined with dye-binding assays using reagents from Pierce Chemical Co. (Rockford, IL). To evaluate variations in the expression of specific proteins, immunoblot analysis was performed. Liver proteins in Lamelli sample buffer were separated by electrophoresis on polyacrylamide gels and transferred to nylon membranes. After a brief incubation with 5% low fat milk to block nonspecific binding, membranes were exposed overnight to specific antisera at 4°C. Next the membranes were washed and exposed to secondary, peroxidase-conjugated antisera, and antigens were visualized by enhanced chemiluminescence using

YANG ET AL.

695

reagents from Amersham (Arlington Heights, IL).25 Signal intensity was quantified by a densitometer (Molecular Dynamics, Sunnyvale, CA), and variations in protein expression were normalized to the values in the lean time 0 (i.e., pre-PH) sample on each blot. For immunoblot analysis, primary antisera were used at the following concentrations: nuclear factor ␬B (NF-␬B) p50 and p65 (Santa Cruz Diagnostics, San Diego, CA) 1:1,000, total and phosphorylated signal transducer and activator of transcription-3 (Stat-3) (Santa Cruz Diagnostics), total and phosphorylated p42 and p44 MAPK (Cell Signaling Technology, Lexington, KY) 1:500, total and phosphorylatedAkt (Cell Signaling) 1:500, cyclin D-1 (Santa Cruz Biotechnology) 1:400, and caspase 3 (Pharmingen, San Diego, CA) 1:1,000. Jun N-terminal kinase (Jnk) activity was evaluated with minor modification of assays we described previously.11 In the present experiments, the Jnk substrate, GST-c-Jun fusion peptide (Santa Cruz Diagnostics) and ␥-32P ATP (Dupont, New England Nuclear, Boston, MA) were added directly to hepatic nuclear extracts; after separating the reaction products by electrophoresis on 10% acrylamide gels, gels were dried, and total Jnk activity was quantified by phosphoimager analysis of the phosphorylated substrate using ImageQuant software (Molecular Dynamics). Immune complex assays were performed with nuclear extracts from 3 different mice/group at each time point. Electrophoretic Mobility Shift Assays. Nuclear extracts (10 ␮g/assay) from individual mice in each group were used for electrophoretic mobility shift assay (EMSA). EMSAs were done with doublestranded oligonucleotide fragments that contain the core ␬B-binding motif from the ␬-light chain enhancer element. Probe preparation and EMSA were performed according to the methods we described.28 One microliter of antisera to NF-␬B p65 or p50 or nonimmune sera were used in supershift EMSAs to verify that the DNA binding complexes contained NF-␬B. Isolation of Hepatic RNA, Northern Blot, and RNAase Protection Analyses.

By using minor modifications29 of the method of Chomzynski and Sacchi,30 total RNA was isolated from the liver of each mouse. RNA was quantified by spectroscopy and its quality was evaluated by agarose gel electrophoresis and subsequent ethidium bromide staining. UCP-2 and PEPCK gene expression were evaluated by Northern blot analysis using 20 ␮g total RNA/lane as we have described.31 To compare lane-lane equivalency in total RNA content, the content of 18S and 28S RNA was evaluated on ethidium bromide-stained membranes. C-jun, cyclin D-1, tumor necrosis factor ␣ (TNF-␣), tumor necrosis factor receptor type 1, and interleukin 6 (IL-6) expression were evaluated by RNAase protection assays by using reagents from Pharmingen (San Diego, CA) according to the manufacturer’s specifications, as we have described.25 This procedure permits the concurrent evaluation of transcripts for several different genes, and a housekeeping gene (glyceraldehyde 3-phosphate dehydrogenase; GAPDH) in each 10-␮g RNA sample. RNA samples from 3 to 4 mice per group were assayed at each time point. After electrophoretic separation of the products on 5% acrylamide gels, gels were dried, evaluated by phosphoimager by using ImageQuant software (Molecular Dynamics), and then exposed to Kodak X-ray film in cassettes with intensifying screens. Evaluation of Hepatic ATP Content. Hepatic ATP content was evaluated by luciferase assay by using reagents from Sigma Chemical, Co (St. Louis, MO) according to the manufacturer’s instructions. Immediately before performing the assay, freeze clamped liver tissues (50 mg/mouse) that had been stored at ⫺80°C were pulverized in a mortar and pestle under liquid nitrogen, suspended in chilled lysis buffer, and centrifuged for 5 minutes at 4°C and supernatant luciferase activity was evaluated in a Berthold luminometer at 37°C. ATP content in the samples were determined by comparison to a concurrent standard curve. Results are expressed as ␮g ATP/g liver. RESULTS

Increased Post-PH Mortality and Decreased Liver Regeneration in ob/ob Mice. To verify that liver regeneration is inhibited in the ob/ob mouse model of NAFLD, as has been shown

696 YANG ET AL.

in the fa/fa rat model for this disease,9 post-PH mortality and hepatocyte proliferation were compared in 26 lean and 25 ob/ob mice. During the initial 36 hours after PH, no lean mice died. In contrast, 7 of 25 (28%) ob/ob mice died during this same time period. Hepatocyte proliferation was assessed by comparing the induction of PCNA and the nuclear incorporation of BrdU in surviving ob/ob and control mice. Nuclear PCNA expression normally increases as hepatocytes exit the pre-replicative (G1) phase of the cell cycle and remains increased throughout the S phase and into the early postreplicative (G2) period.8 The numbers of hepatocytes expressing nuclear PCNA increased about 4-fold during the initial 36 hours after PH in lean mice. Induction of hepatocyte PCNA expression was delayed and inhibited in ob/ob mice, such that by 36 hours after PH, the numbers of PCNA-expressing hepatocytes had only doubled in this group (Fig. 1A). The incorporation of BrdU into nuclear DNA is a good marker for S phase.8 In lean mice, the numbers of BrdU (⫹) hepatocytes increased 4- to 5-fold after PH, peaking at 24 hours and declining slightly by 36 hours (Fig. 1B). In contrast, the number of BrdU (⫹) hepatocytes actually declined after PH in ob/ob mice and there were significantly fewer BrdU (⫹) hepatocytes in ob/ob mice than in lean mice at both 24 and 36 hours post-PH. Although both markers of cellular replication (i.e., PCNA and BrdU) were decreased in ob/ob mice compared with lean controls, the discordance between PCNA expression and BrdU incorporation within the ob/ob group suggests

HEPATOLOGY October 2001

that ob/ob hepatocytes might have become trapped in G1 (where PCNA expression occurs but before BrdU incorporation begins). G1 Arrest in ob/ob Hepatocytes. To evaluate this possibility, hepatic expression of c-jun, an immediate early gene that is induced in early G1, and cyclin D-1, a delayed-early gene that is induced at the G1-S boundary,8 were compared in an additional 24 lean and 24 ob/ob mice at various time points after PH (4 mice/group/time point). Steady state messenger RNA (mRNA) levels of c-jun increased transiently in the livers of lean mice from 0.5 to 1 hour after PH. PH also induced c-jun mRNA levels in ob/ob mice (Fig. 2A). In the ob/ob group, c-jun induction was prolonged after PH, and c-jun mRNA levels became greater in ob/ob mice than in lean controls at 6 and 24 hours, suggesting that a greater proportion of ob/ob hepatocytes were in an early, prereplicative phase of the cell cycle. However, in contrast to the 8-fold induction of cyclin D-1 which occurred in lean mice at 24 hours post-PH, cyclin D-1 induction did not occur after PH in ob/ob mice (Fig. 2B). Additional analyses showed that in lean mice, the accumulation of cyclin D-1 transcripts increased even more (to almost 16-fold greater than pre-PH values) at 36 hours after PH. Cyclin D-1 mRNA expression was about 80% lower in ob/ob mice than in lean mice at the 36-hour time point (P ⱕ .01). The inhibited accumulation of cyclin D-1 transcripts post-PH in ob/ob mice was associated with significant reductions in the hepatic content of cyclin D-1 protein at these time points (Fig. 2C). The impaired induction of cyclin D-1 is consistent with the observation that BrdU incorporation failed to increase post-PH in ob/ob hepatocytes and, together with the PCNA and c-jun data, indicate that ob/ob hepatocytes are arrested in the prereplicative (G1) phase of the cell cycle after PH. Differential Induction of Oxidant-Regulated Stress and MitogenActivated Kinases in ob/ob Mice. In normal mice, hepatic mito-

FIG. 1. Hepatocyte replication after PH. (A). Nuclear expression of PCNA. (B). Nuclear incorporation of BrdU. Mice were injected intraperitoneally with BrdU 2 hours before sacrifice to pulse-label S-phase hepatocyte nuclei. Formalin-fixed, paraffin-embedded liver sections from lean and ob/ob mice were examined for evidence of PCNA expression or BrdU incorporation in hepatocyte nuclei as described in the Methods. Results (mean ⫾ SEM) from 2 to 3 mice per group at time 0 (pre-PH) and 9 to 13 mice per group at 24 and 36 hours post-PH.

chondrial oxidant generation increases almost immediately after PH,28 and this is followed by the activation of oxidantsensitive, stress- and mitogen-activated kinases during G1.11,12,32 To determine if dysregulated kinase induction might contribute to the mechanisms that mediate ob/ob cell cycle arrest, the activities of growth-regulatory kinase cascades were compared in the same 24 lean and 24 ob/ob mice that had been used to evaluate post-PH gene expression. In lean mice, both the stress-activated protein kinases (SAPK), Jnk (Fig. 3A), and the mitogen-activated protein kinases (MAPK), Erk-1 and Erk-2 (Fig. 3B) occurred within an hour after PH in lean mice. A second peak in Erk activation also occurred later, i.e., from 24 to 36 hours, in this group. In contrast, Jnk was barely induced in ob/ob mice after PH. Thus, Jnk activity in ob/ob mice was significantly lower than in lean mice at all post-PH time points evaluated (Fig. 3A). However, induction of Erk-1/Erk-2 was greater in ob/ob mice than in lean controls at both early and late time points after PH (Fig. 3B). Thus, PH led to the normal, coordinated induction of SAPK and MAPK in lean mice, but not in ob/ob mice. Rather, the ob/ob group experienced super-induction of the MAPK cascade but inhibition of the SAPK cascade. These findings support our hypothesis that the normal, acute oxidant-mediated activation of certain redox-sensitive kinases is disturbed in fatty livers, which overproduce oxidants chronically. Unimpaired Post-PH Induction of TNF-␣ and IL-6 in ob/ob Mice.

Subsequent experiments were directed to identifying mechanisms for the abnormal kinase response to PH in ob/ob mice. Because activation of Jnk after PH requires TNF-␣,11 but many

HEPATOLOGY Vol. 34, No. 4, 2001

YANG ET AL.

697

FIG. 2. Induction of proliferation-associated genes after PH. Total hepatic RNA was isolated from 2 lean and 2 ob/ob mice before (0) PH and from 4 lean and 4 ob/ob mice at various time points after (0.5, 1, 6, and 24 hours) PH. Changes in the mRNA levels of c-jun (A) and cyclin D-1 (B) were evaluated by RNAase protection assays. Each lane contains 20 ␮g total hepatic RNA from a different mouse. GAPDH expression was evaluated concurrently in the same samples to show lane-to-lane equivalency of RNA loading/transfer. Representative autoradiographs are shown. The graphs summarize the data from 18 ob/ob mice and 18 lean mice. (C) Changes in cyclin D-1 protein content were evaluated by immunoblot analysis (40 ␮g protein/lane). A representative blot is shown. Each lane contains a sample from a single mouse. Analysis of cyclin D-1 protein expresssion in 3 mice/group/time point showed significantly reduced cyclin D-1 content in ob/ob livers at 24 hours (P ⱕ .01 ob/ob vs. lean).

mitogens and cytokines can induce MAPK,33 it is conceivable that PH-induction of TNF-␣ is impaired selectively in ob/ob mice, which fail to activate Jnk post-PH. To evaluate this possibility, hepatic expression of TNF-␣ was compared in lean and ob/ob mice at various time points after PH. No evidence for inhibited TNF-␣ expression (Fig. 4) or tumor necrosis factor receptor type 1 (data not shown) were noted by RNAase protection assays. Consistent with these findings, mRNA levels of IL-6, a TNF-inducible gene, increased in the liver after PH in both lean and ob/ob mice (Fig. 4). Thus, the impaired post-PH induction Jnk in ob/ob mice cannot be attributed to TNF␣ deficiency per se. Selective Inhibition of Postreceptor Cytokine Signaling in ob/ob Mice. It remains possible, however, that TNF-␣ signaling

might be interrupted selectively at postreceptor levels. The early induction of NF-␬B binding activity that follows PH15 is a TNF-dependent process.24 Thus, if TNF-␣ signaling was interrupted in regenerating ob/ob hepatocytes, then the post-PH activation of NF-␬B might be impaired in ob/ob livers. To evaluate this possibility directly, hepatic nuclear extracts from lean and ob/ob mice were analyzed for NF-␬B DNA binding activity by EMSA. NF-␬B binding activity increased transiently from 0.5 to 1 hour post-PH in lean mice, but this response was much less robust in ob/ob mice (Fig. 5A). The EMSA results suggested that the nuclear accumula-

tion of NF-␬B p65 and p50 subunits might be inhibited selectively in the latter group. Consistent with this concept, immunoblot analysis of hepatic nuclear extracts showed a significantly reduced content of these NF-␬B subunits in the ob/ob mice compared with the lean controls during the early prereplicative period after PH (Fig. 5B and C). Thus, this evidence of inhibited NF-␬B induction in ob/ob mice complements the earlier finding of inhibited Jnk activation in this group and supports the concept that certain TNF-␣–initiated signals are interrupted in ob/ob mice after PH. To determine if the inhibition of cytokine signaling is generalized or specific for TNF-regulated responses, our next experiments evaluated post-PH activation of Stat-3, an IL-6 – regulated kinase that is believed to play a key role in cell cycle progression after PH.16 In contrast to the impaired induction of 2 different TNF-␣ targets in ob/ob livers, the activation of Stat-3 was enhanced in ob/ob mice after PH (Fig. 6). Therefore, the postreceptor impairments in cytokine signaling appear to be restricted to TNF-regulated events. Moreover, because IL-6 signaling is actually enhanced in ob/ob mice after PH, inhibition of the IL-6 –Stat-3 pathway cannot explain why ob/ob hepatocytes fail to induce cyclin D-1 and progress into S phase after PH. In this regard, ob/ob mice resemble fa/fa rats, which also exhibit IL-6 –independent inhibition of post-PH

698 YANG ET AL.

HEPATOLOGY October 2001

FIG. 3. Post-PH activation of oxidant-sensitive kinases. (A) Jnk activity was evaluated in nuclear extracts from 14 ob/ob and 14 lean mice using immune complex assays as described in the Methods. A representative autoradiograph is shown. Each lane contains 10 ␮g nuclear protein from a different mouse. The graph summarizes the data from 4 immune complex assays. (B) MAPK activation was evaluated by immunoblot analysis of whole liver homogenates from 14 ob/ob and 14 lean mice. A representative immunoblot is shown. Each lane contains 40 ␮g protein from a different mouse. The membrane was exposed to antisera specific for phosphorylated p44 and p42 MAPK (i.e., Erk-1 and Erk-2) and then stripped and exposed to antisera to total (i.e., phosphorylated and nonphosphorylated) MAPK. The graphs summarizes the data from 4 immunoblot analyses.

HEPATOLOGY Vol. 34, No. 4, 2001

YANG ET AL.

699

FIG. 4. Induction of cytokine genes after PH. Total hepatic RNA was isolated from 18 lean and 18 ob/ob mice and cytokine mRNA levels were evaluated by RNAase protection assays. Expression of TNF-␣, IL-6, and GAPDH were assessed concurrently in each sample. A representative autoradiograph is shown. Each lane contains 20 ␮g RNA from a different mouse. The graphs summarize the data from duplicate RNAase protection analyses.

liver regeneration.9 On the other hand, the inhibited post-PH activation of the redox-sensitive factor, NF-␬B, in ob/ob livers might contribute to hepatocyte cell cycle arrest, because NF-␬B is known to activate cyclin D-1 transcription.34-36 Cyclin D-1 is required to activate cyclin-dependent kinases that are necessary for hepatocyte DNA replication. As such, cyclin D-1 is a component of the complex cellular machinery that carefully regulates the transition of hepatocytes from prereplicative to replicative phases of the cell cycle.37-39 Taken together, the results of the previous experiments identify a potential mechanism for the inhibited regeneration in fatty livers, i.e., selective inhibition of TNF-␣ signaling inhibits activation of certain redox-sensitive factors, including NF-␬B. Decreased NF-␬B activity, in turn, inhibits the induction of NF-␬B–regulated genes, such as cyclin D-1, which are necessary for hepatocytes to progress from the prereplicative (G1) to replicative (S) phase of the cell cycle. Additional Mechanisms That Potentiate Cell Cycle Arrest in ob/ob Mice. In addition to the regulation of its promoter by several re-

dox-sensitive DNA binding proteins, including NF-␬B,34,36,40 cyclin D-1 is regulated by post-transcriptional mechanisms that modulate cyclin D-1 interactions with cyclin-dependent kinases (cdk). Cellular ATP stores play a key role in regulating the posttranscriptional activation of cyclin D-1/cdk protein complexes because ATP binding is required to activate the complex.41,42 Other factors, such as inhibitors of cyclin-dependent kinases (INK) block ATP binding and inhibit cyclin D-1/cdk function.43 INK expression is regulated by Stat-3.44,45 Our results show that the normal post-PH induction of Stat-316 is sustained and significantly enhanced in ob/ob mice (Fig. 6). Thus, excessive activation of Stat-3 might also contribute to mechanisms that arrest fatty hepatocytes in G1 after PH. This concept is supported by evidence that experimental over-expression of constitutively phosphorylated forms of Stat-3 inhibits cellular proliferation.46 Moreover, INK-4, a Stat-3 target gene, is known to cause growth arrest in hepatocytes.47 Given that Stat-3–inducible INKs inhibit the cyclin D-1/cdk complex by preventing ATP binding,43 their effect is expected to be amplified when hepatic ATP stores are reduced. Thus, if ATP stores are decreased in ob/ob hepatocytes, this might

potentiate G1-arrest in fatty livers. To evaluate this possibility, ATP concentrations were measured in freeze-clamped liver samples from both groups at various time points. As previously reported,28 PH led to a rapid, dramatic reduction in hepatic ATP content in normal, lean mice, but ATP stores recovered to about 70% of preoperative levels by 24 hours post-PH (Fig. 7A). In ob/ob mice, hepatic ATP content is reduced relative to controls both before and after PH (Fig. 7A). Therefore, in ob/ob mice, PH-activation of cyclin D-1/cdk complexes is prevented by mechanisms that limit the transcription of cyclin D-1 (e.g., decreased NF-␬B), as well as by mechanisms that promote the synthesis of cyclin/cdk inhibitors (e.g., excessive Stat-3 induction) and mechanisms that potentiate the actions of these inhibitors (e.g., ATP depletion). Increased Mitochondrial Uncoupling Proteins and ATP Depletion in ob/ob Mice. When UCP-2, a mitochondrial uncoupling pro-

tein that decreases the efficiency of mitochondrial oxidative phosphorylation, is transferred into rapidly proliferating yeast colonies, colony growth arrests because ATP stores become depleted.48 We reported previously that UCP-2 mRNA and protein expression are increased in ob/ob hepatocytes.31 Therefore, increased UCP-2 might contribute to hepatocyte ATP depletion in ob/ob livers if increases in UCP-2 gene expression persist after PH. In ob/ob mice, the hepatic expression of UCP-2 is significantly greater than controls both before and after PH (Fig. 7B). Also, unlike control mice, which recover from early post-PH ATP depletion despite the accumulation of UCP-2 transcripts, increases in UCP-2 expression are accompanied by sustained ATP depletion in ob/ob mice (Fig. 7A). Consistent with these differences in hepatic ATP content, foci of apparent hepatocyte necrosis increased significantly in the livers of ob/ob mice by 24 hours post-PH but were rarely noted in the livers of control mice at this time point (Fig. 7C). Despite focal hepatic necrosis in fatty livers after PH, massive hepatocyte death was not identified by light microscopy of hematoxylin-eosin–stained sections (Fig. 7C). To evaluate the possibility that more subtle differences in apoptotic activity might have occurred, the hepatic accumulation of caspase 3 cleavage products was assessed by immu-

700 YANG ET AL.

HEPATOLOGY October 2001

FIG. 5. Activation of NF-␬B after PH. (A) NF-␬B DNA binding activity was evaluated by EMSA using nuclear extracts from 8 lean and 8 ob/ob mice. A representative autoradiograph is shown. Each lane contains 10 ␮g nuclear protein from a different mouse. The panel on the right shows that antisera to either NF-␬B p50 or p65 disrupt the complexes that are induced at 1 hour after PH, confirming that these bands contain NF-␬B isoforms. Nuclear content of NF-␬B p50 (B) and NF-␬B p65 (C) was evaluated in the same extracts by immunoblot analysis. Representative blots are shown. Each lane contains 40 ␮g nuclear protein from a different mouse. The graphs summarize the data from triplicate experiments.

noblot analysis using commercial antisera that identify both 32-kd procaspase 3 and its 11-kd cleavage product, which is the biologically active executioner caspase during hepatocyte apoptosis. The ratio of activated caspase 3 to procaspase 3 increased by almost 2-fold after PH in lean mice (P ⱕ .05 post-PH vs. pre-PH values), whereas no increase in this ratio was noted in ob/ob mice. Thus, at every time point evaluated after PH, caspase 3 activation was less in ob/ob mice than in lean control mice. Therefore, mechanisms other than massive necrosis or caspase-3–dependent apoptosis must account for the increased post-PH mortality rates observed in the ob/ob group. Hypoglycemia and Prolonged Glycogen Depletion in ob/ob Mice.

After PH, stress-related changes in hepatic gene expression are necessary to induce tissue-specific functions, such as glycogenolysis and gluconeogenesis, that prevent lethal hypoglycemia after 70% hepatectomy.49 In contrast to lean mice, which exhibited a 30% reduction in blood glucose levels and

ambulated freely within an hour after PH, the diabetic ob/ob mice became much more hypoglycemic after PH (Fig. 8A). They assumed a hunched posture and were virtually immobile throughout the post-PH recovery period; their coats also became ruffled. Mortality began to occur around 6 hours after PH, coincident with the onset of hypoglycemia in the ob/ob group. Moreover, in the surviving ob/ob mice, serum glucose concentrations continued to decline throughout the remainder of the follow-up period. To determine if defects in hepatic glycogen mobilization might have contributed to the inability of ob/ob mice to remain euglycemic post-PH, liver glycogen content was assessed by PAS-staining. No obvious differences were noted in the preoperative hepatic glycogen content of lean (Fig. 8B, top panels) and ob/ob (Fig. 8B, bottom panels) mice. Moreover, both groups appeared to mobilize hepatic glycogen efficiently

HEPATOLOGY Vol. 34, No. 4, 2001

FIG. 6. Activation of Stat-3 after PH. Phosphorylated-Stat-3 and total (phospho-Stat-3 ⫹ nonphosphorylated Stat-3) were evaluated in nuclear extracts from 11 lean and 11 ob/ob mice using immunoblot analysis. A representative immunoblot is shown. Each lane contains 40 ␮g nuclear protein from a different mouse. The membrane was exposed to antisera specific for phospho-Stat-3 and then stripped and exposed to antisera for total Stat-3. The graphs summarize the data from triplicate experiments.

during the early prereplicative period. However, the livers of ob/ob mice remained profoundly depleted of glycogen at 24 hours after PH, whereas glycogen stores had almost normalized in the normal lean mice by this time point. The persistent glycogen depletion in fatty livers suggested that ob/ob mice experienced a sustained stimulus for glycogenolysis after PH. Post-PH Activation of Protein Kinase B/Akt and Impaired PEPCK Induction in ob/ob Mice. It is well known that insulin levels fall

transiently after PH and this permits glucagon, norepineph-

YANG ET AL.

701

rine, glucocorticoids and other stress hormones to induce phosphoenolpyruvate carboxykinase (PEPCK),49 a rate-limiting enzyme in hepatic gluconeogenesis. More recently, it has been shown that insulin blocks the stress-related induction of PEPCK by activating protein kinase B/Akt.50 Circulating levels of insulin and glucocorticoids are increased basally in ob/ob mice.51-53 Thus, it is difficult to predict how these mice might respond to the acute stress imposed by PH. To determine if inhibited hepatic induction of PEPCK might contribute to post-PH hypoglycemia in ob/ob mice, hepatic PEPCK mRNA levels and Akt activity were compared in lean and ob/ob mice before and after PH. PEPCK expression was 2- to 3-fold greater in ob/ob mice than in lean controls before PH but post-PH induction of this gene was inhibited in the ob/ob group (Fig. 9A). The post-PH changes in PEPCK mRNA levels were reciprocally related to changes in Akt activity in both groups. In lean mice, the ratio of activated, phospho-Akt/total Akt was highest before PH and fell quickly during the early prereplicative period before PEPCK gene expression increased (Fig. 9B). In contrast, ob/ob mice had a lower ratio of activated, phospho-Akt/total Akt and higher PEPCK mRNA levels before PH but exhibited rapid activation of Akt followed by inhibition of PEPCK gene expression after PH. Thus, after PH, ob/ob mice appeared to become more insulin-sensitive, activating Akt and inhibiting PEPCK gene expression. Given that PEPCK induction is important for maintaining euglycemia after PH,49 it is likely that inhibition of PEPCK contributed to post-PH hypoglycemia and mortality in the ob/ob group. DISCUSSION

These studies show that the regenerative response to liver injury is impaired in ob/ob obese mice with fatty livers, pro-

FIG. 7. Changes in ATP content, UCP-2 expression and hepatocyte viability after PH. (A) Hepatic ATP content was evaluated in 18 lean and 18 ob/ob mice as described in the Methods. Each data point is the mean (⫾ SEM) of results from 2 mice per group before PH or 4 mice per group after PH. (B) UCP-2 mRNA levels were assessed by Northern blot. A representative autoradiograph is shown. Each lane contains 20 ␮g total RNA from a different mouse. The content of 18S RNA in the same samples is also shown to demonstrate lane-to-lane equivalency of RNA loading and transfer. The graph summarizes the data from all 36 mice. (C) Representative photomicrographs showing liver histology in a lean mouse (left) and an ob/ob mouse (right) 24 hours after PH. Hepatic steatosis and focal liver injury are more prominent in the ob/ob liver. Final magnification ⫻200.

702 YANG ET AL.

HEPATOLOGY October 2001

FIG. 8. Changes in blood glucose concentrations and hepatic glycogen content after PH. (A) Glucose concentrations were measured in tailvein blood from 55 mice (27 lean and 28 ob/ob) at various time points after PH. (B) Formalin-fixed, paraffin-embedded liver sections from lean (top) and ob/ob (bottom) mice were stained with periodic acid Schiff (PAS) reagent to evaluate hepatic glycogen stores before (0) and after (1 and 24 hours) PH. Results from 1 representative mouse per group are shown at each time point.

viding further support for the concept that inhibited regeneration contributes to disease progression in NAFLD. Our results indicate that fatty hepatocytes in ob/ob mice become trapped in G1, the prereplicative phase of the cell cycle, and fail to enter S phase after PH. Although G1 arrest has already been reported to occur in fa/fa rats after PH,9 that work did not suggest mechanisms for the inhibited proliferation. Our results identify several mechanisms that are likely to contribute to this cell cycle arrest. First, consistent with our hypothesis, the coordinated activation of the oxidant-sensitive, stress- and mitogen-activated protein kinase cascades that normally occurs early in G128 does not occur in fatty hepatocytes, which have adapted to a state of chronic oxidant stress.25 Rather, in fatty livers, activation of the mitogen-activated kinases, Erk 1 and Erk 2, is enhanced whereas induction of the stress-activated kinase, Jnk, is abolished. Erk and Jnk have multiple cellular targets and interact to orchestrate the complex hepatic response to injury, inducing genes that regulate cell survival, proliferation, and certain tissue-specific functions.33,54-59 The importance of coordinated activation of these kinases has been illus-

trated by studies in cultured fibroblasts, breast cancer cells, endothelial cells, and embryonic mice, all of which show that activation of the Erk cascade cannot induce a proliferative response when Jnk induction is inhibited.40,56-58,60 The present results show that mature adult hepatocytes resemble the cells in these other systems because fatty ob/ob hepatocytes fail to proliferate despite robust Erk activation when Jnk induction is blocked. Therefore, these experiments identify a mechanism (i.e., Jnk inhibition) that is likely to inhibit proliferation in fatty hepatocytes. Additional studies with Jnk inhibitors are planned to test this theory more directly. However, such work may be difficult because our results also suggest that Jnk plays an important, albeit indirect, role in other aspects of the post-PH stress response, such as the induction of PEPCK, that are required for survival after PH (see below). Our study also identifies several other mechanisms, including inhibited induction of the cyclin D-1 gene, enhanced Stat-3 activation, and hepatic ATP depletion, that are likely to inhibit fatty hepatocytes from progressing into replicative stages of the cell cycle after PH. Moreover, the results suggest that these processes, which were initiated by diverse signals,

HEPATOLOGY Vol. 34, No. 4, 2001

FIG. 9. Changes in PEPCK gene expression and protein kinase B/Akt activity after PH. (A) PEPCK mRNA levels were evaluated in 18 lean and 18 ob/ob mice by Northern blot analysis. A representative autoradiograph is shown. Each lane contains 20 ␮g RNA from a different mouse. The content of 18S RNA in each sample is also shown. The graph summarizes data from all 36 mice. (B) Phosphorylated and total Akt were evaluated in whole liver homogenates from the same mice. A representative immunoblot is shown. Each lane contains 40 ␮g protein from a different mouse. The blot was exposed to antisera specific for phospho-Akt and then stripped and exposed to antisera for total (phospho-Akt ⫹ nonphosphorylated Akt). The graph summarizes the data from all 36 mice.

converge to block cyclin D-1/cdk complex activation at multiple levels during late G1. The decreased accumulation of cyclin D-1 transcripts in ob/ob livers at 24 and 36 hours post-PH is an expected consequence of the inhibited induction of NF-␬B in this group, because NF-␬B is necessary to trans-activate the cyclin D-1 promoter.34-36 Consequent decreases in cyclin D-1 protein content occur and this will impair G1-S transition because hepatocytes require cyclin D-1– dependent kinases to enter S phase.37,38 However, the failure to up-regulate cyclin D-1 gene expression after PH is probably not sufficient to arrest cell cycle progression in fatty hepatocytes because these cells express increased amounts of cyclin

YANG ET AL.

703

D-1 at baseline and during the early prereplicative period after PH. Therefore, other post-transcriptional mechanisms must also have inhibited the activity of the existing cyclin D-1 molecules in ob/ob hepatocytes. Activated Stat-3 could do this by inducing INK-4,44 a cyclin/cdk inhibitor.43 Although we did not assess INK-4 expression directly, our results show that hyper-phosphorylation of Stat-3 occurs in fatty livers after PH. Hyperphosphorylation of Stat-3 is known to inhibit proliferation in certain cell lines,44 so the possibility that increased Stat-3 activation contributes to cell cycle arrest in fatty hepatocytes is plausible. Moreover, because ob/ob hepatocytes have reduced ATP stores and the mechanism for Stat3–mediated growth arrest involves inhibition of ATP binding to regulatory sites in the cyclin/cdk enzyme complex,43 ATP depletion in ob/ob hepatocytes is likely to potentiate Stat-3– mediated cell-cycle arrest in these cells. In addition to cyclin D-1/cdk activation, several other events that are required for G1-S transition, including the activation of certain ion channels, ornithine decarboxylase, thymidine kinase, and chromatin remodeling enzymes, also require ATP.61 Pharmacologic agents that inhibit ATP binding sites in these enzymes have potent antiproliferative effects in most cells.42 Therefore, hepatic ATP depletion probably played a general role in inhibiting post-PH hepatocyte proliferation in ob/ob mice. In addition, our studies suggest a mechanism, i.e., over-expression of UCP-2, that helps to explain why hepatic ATP stores are lower in ob/ob than control livers. Consistent with this concept, growth arrest due to ATP depletion in proliferating yeast colonies is the routine bioassay for UCP-2, a mitochondrial protein that inhibits oxidative phosphorylation.62 Assuming that UCP-2 exerts a similar effect in proliferating yeast and hepatocytes, then UCP-2–related limitations in ATP availability may be a general mechanism that retards cell cycle progression in fatty livers. ATP depletion is not uniformly deleterious, however, because it might actually protect fatty hepatocytes from apoptotic death by inhibiting caspase-3 activation.63 Jnk inhibition also prevents caspase 3 activation.64-66 Consistent with those reports, Jnk activation is required for apoptosis in many types of cells.64-66 Thus, our studies identify 2 mechanisms, i.e., the inhibition of Jnk and the impaired activation of caspase 3, that are expected to protect ob/ob hepatocytes from apoptosis. In addition, inhibition of Jnk permits activation of the protein kinase B/Akt survival factor.67 The latter response provides a third mechanism that helps fatty hepatocytes to escape apoptosis when cell cycle progression is blocked. This helps to explain why ob/ob livers did not undergo massive apoptosis despite the inhibited post-PH induction of NF-␬B, a transcription factor that normally provides survival signals for regenerating hepatocytes.20 Apparently, however, in some hepatocytes, ATP depletion becomes so severe after PH that membrane integrity is compromised, generating the focal areas of ballooned hepatocytes that we observed in ob/ob livers. On the other hand, most fatty hepatocytes, although growth-arrested, remain viable during the initial 36 hours after PH. Therefore, the increased post-PH mortality that the ob/ob group experienced during this time period cannot be attributed to massive liver cell death. Rather, mortality in the ob/ob group appears to result from hepatic dysfunction because it correlated with severe post-PH hypoglycemia and inhibited regenerative induction of PEPCK, a rate-limiting enzyme for hepatic gluconeogenesis during liver regenera-

704 YANG ET AL.

tion. Our results suggest a mechanism for PEPCK inhibition by showing that differential regulation of Akt activity occurs in normal and fatty livers after PH. Another group has already reported that induction of Akt is sufficient to abolish stress hormone-related increases in PEPCK transcription.50 Therefore, the premature and sustained activation of Akt that occurs after PH in ob/ob mice is likely to repress the normal stress-related induction of PEPCK in ob/ob mice, exacerbating hypoglycemia after PH. Ironically, the super-induction of Akt (which promotes lethal hypoglycemia) can be a consequence of Jnk inhibition67 and, thus, a component of the attenuated stress response that protects fatty hepatocytes from apoptosis during prolonged cell cycle arrest.68-71 This observation shows that mechanisms that improve hepatocyte survival can also limit liver-specific functions and thereby increase net morbidity and mortality after liver injury. The present work and other systematic studies of the regenerative response in animal models of human liver diseases are informative because they provide an opportunity to assess the relative importance of putative growth-regulatory and survival factors in biologically relevant contexts. This is important because the findings in animals with “naturally occurring” liver diseases do not always match the outcomes predicted by more contrived gene transfer or gene disruption experiments in healthy animals. For example, although adenovirally mediated delivery of I␬B to normal mice inhibits hepatic nuclear accumulation of NF-␬B p50 and p65 and results in massive hepatocyte apoptosis after PH,20 extensive liver cell death did not follow the inhibited nuclear accumulation of NF-␬B p50/p65 or decreased NF-␬B– binding activity in ob/ob mice. Similarly, Erk activity, which generally regulates proliferation in normal liver cells,8 did not promote hepatocyte proliferation in the ob/ob mice with NAFLD. These discrepancies underscore the complexity of the mechanisms that regulate liver regeneration in diseased livers and emphasize the importance of work in animal models of human liver disease, if an ultimate objective is to develop treatments for patients with established NAFLD. In this regard, it is disappointing that our studies failed to identify one particular extracellular factor to blame for the adverse outcomes in mice with NAFLD. Liver regeneration is inhibited in ob/ob mice despite increased levels of cytokines and hormones, such as TNF-␣, IL-6, and insulin, that normally promote hepatocyte proliferation after PH.8 Although ob/ob mice lack leptin,72 and leptin generally enhances cellular proliferation73-75 and wound healing,76 leptin deficiency is not likely to explain the decreased regenerative capacity of ob/ob fatty livers because regeneration is also inhibited in obese diabetic fa/fa rats9 and normal rats that have been fed ethanol,77 both of which have fatty livers and high leptin levels.78,79 Thus, our data suggest that treatment of NAFLD will not be as simple as supplementing or reducing any of these extracellular factors to normalize regeneration in injured livers. Rather, effective treatments will probably have to adjust the “cross-talk” among multiple, overlapping signaling pathways that mediate hepatocellular proliferation, differentiated functions, and viability.33,40,54,57,58,60,80 In naturally occurring animal models for fatty liver disease, such as ob/ob mice, and probably in people that develop similar disorders, signaling has been modified to permit hepatocyte survival during chronic oxidative stress.25,81 Additional studies are necessary to determine if such adaption reduces the acute burst of mi-

HEPATOLOGY October 2001

tochondrial oxidant release that normally occurs as hepatocytes enter G1 or increases the threshold of oxidants that are required to activate redox-sensitize target molecules. At this point, however, it is clear that the cellular response to acutely superimposed stressors is altered considerably in fatty hepatocytes and this is likely to contribute to the pathophysiology of NAFLD. REFERENCES 1. Matteoni C, Younossi ZM, McCullough A. Nonalcoholic fatty liver disease: a spectrum of clinical pathological severity. Gastroenterology 1999; 116:1413-1419. 2. Angulo P, Deach JC, Batts KP, Lindor KD. Independent predictors of liver fibrosis in patients with nonalcoholic steatohepatitis. HEPATOLOGY 1999; 30:1356-1362. 3. Ratziu V, Giral P, Charlotte F, Bruckert E, Thibault V, Theodorou I, Khalil L, et al. Liver fibrosis in overweight patients. Gastroenterology 2000;118:1117-1123. 4. Ludwig J, Viggiano RT, McGill DB. Nonalcoholic steatohepatitis. Mayo Clinic experiences with a hitherto unnamed disease. Mayo Clin Proc 1980;55:342-348. 5. Sheth SG, Gordon FD, Chopar S. Nonalcoholic steatohepatitis. Ann Intern Med 1997;126:137-145. 6. Teli M, James OF, Burt AD, Bennett MK, Day CP. A natural history of nonalcoholic fatty liver: a follow-up study. HEPATOLOGY 1995;22:17141717. 7. Day CP, James O. Steatohepatitis: a tale of two “hits”? Gastroenterology 1998;114:842-845. 8. Michalopoulos GK, DeFrances MC. Liver regeneration. Science 1997; 276:60-66. 9. Selzner M, Clavien PA. Failure of regeneration of the steatotic rat liver: disruption at two different levels in the regeneration pathway. HEPATOLOGY 2000;31:35-42. 10. Diehl AM. Animal models of hepatic steatosis. Semin Liv Dis 2001;21: 89-104. 11. Diehl AM, Yin M, Fleckenstein J, Yang SQ, Lin HZ, Brenner DA, Westwick J, Bagby G, Nelson S. Tumor necrosis factor alpha induces c-jun during the regenerative response to liver injury. Am J Physiol 1994; 267:G552-G561. 12. Westwick J, Weitzel C, Leffert HL, Brenner DA. Activation of Jun kinase is an early event in hepatic regeneration. J Clin Invest 1995;95:803-810. 13. Mohn KL, Laz TM, Melby AE, Taub R. Immediate-early gene expression differs between regenerating liver, insulin-stimulated H-35 cells, and mitogen-stimulated Balb/c 3T3 cells. Liver-specific induction patterns of gene 33, phosphoenolpyruvate carboxykinase, and the jun, fos, and egr families. J Biol Chem 1990;265:20914-29121. 14. Diamond RH, Du HK, Lee VM, Mohn KL, Haber BA, Tewari DS, Taub R. Novel delayed-early and highly insulin-induced growth response genes. J Biol Chem 1993;268:15185-15192. 15. Cressman DE, Greenbaum LE, Haber BA, Taub R. Rapid activation of PHF/NF kappa B in hepatocytes, a primary response in the regenerating liver. J Biol Chem 1994;269:30429-30435. 16. Cressman DE, Diamond RH, Taub R. Rapid activation of Stat3DNA binding in liver regeneration. HEPATOLOGY 1995;21:1443-1449. 17. Greenbaum LE, Cressman DE, Haber BA, Taub R. Coexistence of C/EBP alpha, beta, growth-induced proteins and DNA synthesis in hepatocytes during liver regeneration. Implications for maintenance of the differentiated state during liver growth. J Clin Invest 1995;96:1351-1365. 18. Diehl AM, Yang SQ, Yin M, Lin HZ, Nelson S, Bagby G. Tumor necrosis factor alpha modulates CCAAT/Enhancer binding proteins activities and promotes hepatocyte-specific gene expression during liver regeneration. HEPATOLOGY 1995;22:252-261. 19. Greenbaum LE, Li W, Cressman DE, Peng Y, Ciliberto G, Poli B, Taub R. CCAAT enhancer-binding protein beta is required for normal hepatocyte proliferation in mice after partial hepatectomy. J Clin Invest 1998;102: 996-1007. 20. Iimuro Y, Nishiura T, Hellerbrand C, Behrns KE, Schoonhoven R, Grisham JW, Brenner DA. NF kappa B prevents apoptosis and liver dysfunction during liver regeneration. J Clin Invest 1998;101:802-811. 21. Rai RM, Lee FYJ, Rosen A, Yang SQ, Lin HZ, Koteish Q, Liew FY, et al. Impaired liver regeneration in inducible nitric oxide synthase-deficient mice. Proc Natl Acad Sci U S A 1998;95:13829-13834. 22. Akerman P, Cote P, Yang SQ, McClain C, Nelson S, Bagby GJ, Diehl AM. Antibodies to tumor necrosis factor alpha inhibit liver regeneration after partial hepatectomy. Am J Physiol 1992;263:G579-G585.

HEPATOLOGY Vol. 34, No. 4, 2001 23. Cressman DE, Greenbaum LE, DeAngelis RA, Ciliberto G, Furth EE, Poli V, Taub R. Liver failure and defective hepatocyte regeneration in interleukin-6-deficient mice. Science 1996;274:1379-1383. 24. Yamada Y, Kirillova I, Peschon JJ, Fausto N. Initiation of liver growth by tumor necrosis factor: deficient liver regeneration in mice lacking type I tumor necrosis factor receptor. Proc Natl Acad Sci U S A 1997;94:14411446. 25. Yang SQ, Zhu H, Li Y, Lin HZ, Gabrielson K, Trush MA, Diehl AM. Mitochondrial adaptations to obesity-related oxidant stress. Arch Biochem Biophys 2000;378:259-268. 26. Higgins GM, Andersen RM. Experimental pathology of liver: restoration of liver of the white rat follwoing partial surgical removal. Arch Pathol 1931;12:186-202. 27. Lavery DJ, Schibler U. Circadian transcription of the cholesterol 7 alpha hydroxylase gene may involve the liver-enriched bZIP protein DBP. Gen Dev 1993;7:1871-1884. 28. Lee FYJ, Li Y, Zhu H, Yang SQ, Lin HZ, Trush M, Diehl AM. Tumor necrosis factor alpha increases mitochondrial oxidant production and induces expression of uncoupling protein-2 expression in the regenerating mouse liver. HEPATOLOGY 1999;29:677-687. 29. Lin HZ, Yang SQ, Kujhada F, Ronnet G, Diehl AM. Metformin reverses nonalcoholic fatty liver disease in obese leptin-deficient mice. Nat Med 2000;6:998-1003. 30. Chomczynski P, Sacchi N. Single step method of RNA isolation by acid guanidine thiocyanate-phenol-chloroform extraction. Anal Biochem 1987;162:156-161. 31. Chavin K, Yang SQ, Lin HZ, Chatham J, Chacko VP, Hoek JB, WalajtyzsRode E, et al. Obesity induces expression of uncoupling protein-2 in hepatocytes and promotes liver ATP depletion. J Biol Chem 1999;274: 5692-5700. 32. Zeldin G, Rai R, Yang SQ, Lin HZ, Yin M, Diehl AM. Effects of alcohol on cytokine-inducible transcription factors. Alcohol Clin Exp Res 1996;20: 1639-1645. 33. Talarmin H, Rescan C, Cariou S, Glaise D, Zanninelli G, Bilodeau M, Loyer P, et al. The mitogen-activated protein kinase kinase/extracellular signal-regulated kinase cascade activation is a key signalling pathway involved in the regulation of G1 phase progression in proliferating hepatocytes. Mol Cell Biol 1999;19:6003-6011. 34. Guttridge DC, Albanese C, Reuther JY, Pestell RG, Baldwin AS. NFkappaB controls cell growth and differentition through transcriptional regulation of cyclin D1. Mol Cell Biol 1999;19:5785-5799. 35. Hinz M, Krappmann D, Eichten A, Heder A, Scheidereit C, Strauss M. NF-kappa B function in growth control: regulation of cyclin D1 expression and G0/G1-to-S-phase transition. Mol Cell Biol 1999;19:2690-2698. 36. Joyce D, Bouzahzah B, Fu M, Albanese C, D’Amico M, Steer J, Klein A, et al. Integration of Rac-dependent regulation of cyclin D1 transcription through a nuclear factor-kappaB-dependent pathway. J Biol Chem 1999; 274:25245-25249. 37. Koch KS, Lu XP, Leffert HL. Primary rat hepatocytes express cyclin D1 messenger RNA during their growth cycle and during mitogenic transitions induced by transforming growth factor-alpha. Biochem Biophys Res Commun 1994;204:91-97. 38. Nishida N, Fukuda Y, Komeda T, Kita R, Sando T, Furukawa M, Amenomori M, et al. Amplification and overexpression of the cyclin D1 gene in aggressive human hepatocellular carcinoma. Cancer Res 1994;54:31073110. 39. Wu H, Wade M, Krall L, Grisham J, Xiong Y, Van Dyke T. Targeted in vivo expression of the cyclin-dependent kinase inhibitor p21 halts hepatocyte cell-cycle progression, postnatal liver development, and regeneration. Gen Devel 1996;10:245-260. 40. Lee RJ, Albanese C, Stenger RJ, Watanae G, Inghirami G, Haines GKr, Webster M, et al. pp60 (v-src) induction of cyclin D1 requires collaborative interactions between the extracellular signal-regulated kinases, p38, and Jun kinase pathways. A role for cAMP response element-binding protein and activating transcription factor-2 in pp60 (v-src) signaling in breast cancer cells. J Biol Chem 1999;274:7341-7350. 41. Laird AD, Vajkoczy P, Shawver LK, Thurnher A, Liang C, Mohamm X, Schlessinger J, et al. SU6668 is a potent antiangiogenic and antitumor agent that induces regression of established tumors. Cancer Res 2000;60: 4152-4160. 42. Fischer PM, Lane DP. Inhibitors of cyclin-dependent kinases as anticancer therapeutics. Curr Med Chem 2000;7:1213-1245. 43. Jeffrey PD, Tong L, Pavletich NP. Structural basis of inhibition of CDKcyclin complexes by INK4 inhibitors. Gen Dev 2000;14:3115-3125.

YANG ET AL.

705

44. O’Farrell AM, Parry DA, Zindy F, Roussel MF, Lees E, Moore KW. Stat3dependent induction of p19INK4D by IL-10 contributes to inhibition of macrophage proliferation. J Immunol 2000;164:4607-4615. 45. de Koning JP, Soede-Bobok AA, Ward AC, Schelen AM, Antonissen K, van Leeuwen D, Lowenberg B, et al. STAT3-medicated differentiation and survival and of myeloid cell response to granulocyte colony-stimulating factor: role for the cyclin-dependent kinase inhibitor p27 (Kip1). Oncogene 2000;19:3290-3298. 46. Spitto MT, Chung TD. STAT3 mediates IL-6-induced neuroendocrine differentiation in parostate cancer cells. Prostate 2000;42:186-195. 47. Awad MM, Sanders JA, Gruppuso PA. A potential role for p15 (Ink4b) and p57 (Kip2) in liver development. FEBS Lett 2000;483:160-164. 48. Skulachev VP. Uncoupling: new approaches to an old problem of bioenergetics. Biochem Biophys Acta 1998;1363:100-124. 49. Bucher NLR. Liver regeneration: an overview. J Gastroenterol Hepatol 1991;6:615-624. 50. Liao J, Barthel A, Nakatani K, Roth RA. Activation of protein kinase B/Akt is sufficient to repress the glucocorticoid and cAMP induction of phosphoenolpyruvate carboxykinase gene. J Biol Chem 1998;273:2732027324. 51. Rossier J, Rogers J, Shibasaki T, Guillemin R, Bloom FE. Opiod peptides and alpha-melanocyte-stimulating hormone in genetically obese (ob/ob) mice during development. Proc Natl Acad Sci U S A 1979;76:2077-2080. 52. Pelleymounter MA, Cullen MF, Baker MB, Hecht R, Winters D, Boone T, Collins F. Effects of the obese gene product on body weight regulation in ob/ob mice. Science 1995;269:540-543. 53. Seufert J, Kieffer TJ, Habener JF. Leptin inhibits insulin gene transcription and reverses hyperinsulinemia in leptin-deficient ob/ob mice. Proc Natl Acad Sci U S A 1999;96:674-679. 54. Yan M, Dal T, Deak JC, Kyriakis JM, Zon LI, Woodgett JR, Templeton DJ. Activation of stress-activated protein kinase by MEKK1 phosphorylation of its activator SEK1. Nature Lond 1994;372:798-1994. 55. Davis RJ. Transcriptional regulation of MAP kinases. Mol Reprod Dev 1995;42:459-467. 56. Bost F, McKay R, Dean N, Mercola D. The JUN kinase/stress-activated protein kinase pathway is required for epidermal growth factor stimulation of growth of human A549 lung carcinoma cells. J Biol Chem 1997; 272:33422-33429. 57. Ganiatsas S, Kwee L, Fjuiwara Y, Perkins A, Ikeda T, Labor MA, Zon LI. SEK1 deficiency reveals mitogen-activated protein kinase cascade crossregulation and leads to abnormal hepatogenesis. Proc Natl Acad Sci U S A 1998;95:6881-6886. 58. Potapova O, Gorospe M, Dougherty RH, Dean NM, Gaarde WA, Holbrook NJ. Inhibition of c-Jun N-terminal kinase 2 expression suppresses growth and apoptosis of human tumor cells in a p53-dependent manner. Mol Cell Bio 2000;5:1713-1722. 59. Fan H, Derynck R. Ectodomain shedding of TGF-alpha and other transmembrane proteins is induced by receptor tyrosine kinase activation and MAP kinase signaling cascades. EMBO J 1999;18:6962-6972. 60. Pedram A, Razandi M, Levin ER. Extracellular signal-regulated protein kinase/Jun kinase cross-talk underlies vascular endothelial cell growth factor-induced endothelial cell proliferation. J Biol Chem 1998;273: 26722-26728. 61. Lee DG, Bell SP. ATPase switches controlling DNA replication initiation. Curr Opin Cell Biol 2000;12:280-285. 62. Boss O, Muzzin P, Giacobino JP. The uncoupling proteins, a review. Eur J Endocrinol 1998;139:1-9. 63. Eguchi Y, Shimizu S, Tsujimoto Y. Intracellular ATP levels determine cell death fate by apoptosis or necrosis. Cancer Res 1997;57:1835-1840. 64. Aggarwal BB. Tumour necrosis factors receptor associated signaling molecules and their role in activation of apoptosis, JNK and NK-kappa B. Ann Rheum Dis 2000;59(Suppl 1):I6-I16. 65. Takada E, Toyota H, Suzuki J, Mizuguchi J. Prevention of anti-IgMinduced apoptosis accompanying G(1) arrest in B lymphoma cells overexpressing dominant-negative mutant form of c-Jun N-terminal kinase 1. J Immunol 2001;166:1641-1649. 66. Tournier C, Hess P, Yang DD, Xy J, Turner TK, Nimnual A, Bar-Sagi R, Flavell RA, David RJ. Requirement of JNK for stress-induced activation of the cytochrome c-mediated death pathway. Science 2000;288:870-874. 67. Levress V, Butterfield L, Zentrich E, Heasley LE. Akt negatively regulates the cJun N-terminal pathway in PC12 cells. J Neurosci Res 2000;62:799808. 68. Romashkova JA, Makarov SS. NF-kappaB is a target of AKT in antiapoptotic PDGF signalling. Nature 1999;401:86-90.

706 YANG ET AL. 69. Beraud C, Henzel WJ, Baeuerle PA. Involvement of regulatory and catalytic subunits of phosphoinositide 3-kinase in NF-␬B activation. Proc Natl Acad Sci U S A 1999;96:429-434. 70. Madrid LV, Wang C-Y, Guttridge DC, Schottelius AJG, Baldwin AS, Mayo MW. Akt suppresses apoptosis by stimulating the transactivation potential of the relA/p65 subunit of NF-kB. Mol Cel Biol 2000;20:16261638. 71. Zhou BP, Hu MC-T, Miller SA, Yu Z, Xia W, Lin S-Y, Hung M-C. Her-2/ neu blocks tumor necrosis factor-induced apoptosis via the Akt/NF-kB pathway. J Biol Chem 2000;275:8027-8031. 72. Friedman JM, Leibel R, Siegel DS, Walsh J, Bahary N. Molecular mapping of the mouse ob mutation. Genetics 1991;11:1054-1062. 73. Gainsford T, Willson TA, Metcalf D, Handman E, McFarlane C, Ng A, Nicola NA, et al. Leptin can induce proliferation, differentiation, and functional activation of hemopoietic cells. Proc Natl Acad Sci U S A 1996;93:14564-14568. 74. Tanabe K, Okuya S, Tanizawa Y, Matsutani A, Oka Y. Leptin induces proliferation of pancreatic beta cell line MIN6 through activation of mitogen-activated protein kinase. Biochem Biophys Res Commun 1997; 241:765-768. 75. Howard JK, Lord GM, Matarese G, Vendetti S, Ghatei MA, Ritter MA, Lechler RI, et al. Leptin protects mice from starvation-induced lymphoid

HEPATOLOGY October 2001

76. 77. 78. 79.

80.

81.

atrophy and increases thymic cellularity in ob/ob mice. J Clin Invest 1999;104:1051-1059. Frank S, Stallmeyer B, Kampfer H, Kolb N, Pfeilschifter J. Leptin enhances wound re-epithelialization and constitutes a direct function of leptin in skin repair. J Clin Invest 2000;106:501-509. Wands JR, Carter EA, Bucher NLR, Isselbacher KJ. Inhibition of hepatic regeneration in rats by acute and chronic ethanol intoxication. Gastroenterology 1979;77:528-531. Phillips MS, Liu Q, Hammond HA, Dugan V, Hey PJ, Caskey CJ, Hess JF. Leptin receptor missense mutation in the fatty Zucker rat. Nat Genet 1996;13:18-19. Lin HZ, Yang SQ, Zeldin G, Diehl AM. Chronic ethanol consumption induces the production of tumor necrosis factor-alpha and related cytokines in liver and adipose tissue. Alc Clin Exp Res 1998;22(Suppl):231S247S. Kaliman P, Canicio J, Testar X, Palacin M, Zorzano A. Insulin-like growth factor-II, phosphatidylinositol 3-kinase, nuclear factor-␬B and inducible nitric-oxide synthase define a common myogenic signaling pathway. J Biol Chem 1999;274:17437-17444. Yang SQ, Lin HZ, Huang X, Diehl AM. Fatty liver vulnerability to lipopolysaccaride despite NF-␬B induction and caspase 3 inhibition. Am J Physiol 2001 281:G382-392.