significance of insulin signaling in liver regeneration triggered by portal vein ligation

10
Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation Jeng-Hwei Tseng, M.D.,* ,2 Chun-Hsiang Ouyang, M.D.,,2 Kun-Ju Lin, M.D., Ph.D.,and Ta-Sen Yeh, M.D., Ph.D.,1 *Department of Radiology; Department of Surgery; and Department of Nuclear Medicine, Chang Gung Memorial Hospital, Chang Gung University, Taipei Submitted for publication February 18, 2009 Background. To investigate the role of insulin signaling in liver regeneration following portal vein ligation (PVL). Materials and Methods. Streptozotocin-induced insulin-deficient rats underwent PVL, and were sacri- ficed at indicated time points. Liver regeneration indi- ces, including volumetric shifting, BrdU, proliferative cell nuclear antigen (PCNA), and Ki-67 labeling index, were determined. Cell cycle markers, telomerase reverse transcriptase (TERT), and apoptosis-related genes were detected using quantitative real time poly- merase chain reaction (PCR). Cell cycle analysis was determined using flow cytometry. Expression of insu- lin-like growth factor receptor (IGFR)-2 and TGFb1 were determined using Western blot. Results. Restituted liver mass and redistributed volume ratio of insulin-deficient rats were decreased compared with those of normal rats. Labeling index of BrdU and PCNA of insulin-deficient rats were increased compared with normal rats, evidenced by an increased S-phase fraction detected by flow cytome- try. Expression of cyclin A2, cyclin B1, TERT mRNA, and telomerase activity were decreased in insulin- deficient rats. Increased Bax, Daxx, and JNK mRNA expression and decreased Bcl X L expression in insu- lin-deficient rats, led to increased hepatocyte apopto- sis than normal rats. Finally, expression of IGFR-2 was increased in insulin-deficient rats. Conclusions. Insulin signaling plays an important role in liver regeneration triggered by portal vein ligation. Ó 2011 Elsevier Inc. All rights reserved. Key Words: insulin; liver regeneration; portal vein ligation. INTRODUCTION Portal vein embolization has been applied prior to major hepatectomy for hepatocellular carcinoma, bile duct malignancy, and metastatic liver tumors [1–3]. This procedure acts to embolize the portal branch of lobes targeted for resection, and induces compensatory hypertrophy in future remnant lobes, which largely involves process of liver regeneration. In a two-thirds partial hepatectomized animal model, the majority of liver parenchymal cells that are normally quiescent rapidly reenter the cell cycle. Several growth factors have been implicated in this process, including TGFa, EGF, HGF, neural stimuli, glucagon and insulin [4]. Of them, insulin supplied by the pancreatic islets perfuses hepatocytes continually through the portal vein. If the amount of portal circulation to the liver is decreased, the liver atrophies; in contrast, injection of insulin prevents or reverses this process [5, 6]. Thus, insulin by itself is not a primary mitogen for hepato- cytes, however, insulin signaling must be present for the mitogenic signal to proceed normally. In the clinical setting, diabetes has been shown as an independent fac- tor to reduce hypertrophy ratio of unembolized lobe after portal vein embolization [7, 8]. Furthermore, given that up to 30% of patients with liver malignancy are associated with diabetes and their livers are poten- tially less able to regenerate [9, 10], evaluating the physiological significance of insulin signaling on liver regeneration following portal vein embolization and relevant molecular events is thus intriguing. 1 To whom correspondence and reprint requests should be addressed at Surgical Laboratory, Department of Surgery, Chang Gung Memorial Hospital, 5 Fu-Hsing Street, Kwei-Shan Shiang, Taoyuan, Taiwan. E-mail: [email protected], tsy471027@ adm.cgmh.org.tw. 2 These two authors contributed equally to this work. 0022-4804/$36.00 Ó 2011 Elsevier Inc. All rights reserved. 77 Journal of Surgical Research 166, 77–86 (2011) doi:10.1016/j.jss.2009.06.043

Upload: jeng-hwei-tseng

Post on 27-Oct-2016

214 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

Journal of Surgical Research 166, 77–86 (2011)doi:10.1016/j.jss.2009.06.043

Significance of Insulin Signaling in Liver Regeneration Triggered

by Portal Vein Ligation

Jeng-Hwei Tseng, M.D.,*,2 Chun-Hsiang Ouyang, M.D.,†,2 Kun-Ju Lin, M.D., Ph.D.,‡and Ta-Sen Yeh, M.D., Ph.D.†,1

*Department of Radiology; †Department of Surgery; and ‡Department of Nuclear Medicine, Chang Gung Memorial Hospital,Chang Gung University, Taipei

Submitted for publication February 18, 2009

Background. To investigate the role of insulinsignaling in liver regeneration following portal veinligation (PVL).

Materials and Methods. Streptozotocin-inducedinsulin-deficient rats underwent PVL, and were sacri-ficed at indicated time points. Liver regeneration indi-ces, including volumetric shifting, BrdU, proliferativecell nuclear antigen (PCNA), and Ki-67 labeling index,were determined. Cell cycle markers, telomerasereverse transcriptase (TERT), and apoptosis-relatedgenes were detected using quantitative real time poly-merase chain reaction (PCR). Cell cycle analysis wasdetermined using flow cytometry. Expression of insu-lin-like growth factor receptor (IGFR)-2 and TGFb1were determined using Western blot.

Results. Restituted liver mass and redistributedvolume ratio of insulin-deficient rats were decreasedcompared with those of normal rats. Labeling indexof BrdU and PCNA of insulin-deficient rats wereincreased compared with normal rats, evidenced byan increased S-phase fraction detected by flow cytome-try. Expression of cyclin A2, cyclin B1, TERT mRNA,and telomerase activity were decreased in insulin-deficient rats. Increased Bax, Daxx, and JNK mRNAexpression and decreased Bcl XL expression in insu-lin-deficient rats, led to increased hepatocyte apopto-sis than normal rats. Finally, expression of IGFR-2was increased in insulin-deficient rats.

Conclusions. Insulin signaling plays an importantrole in liver regeneration triggered by portal veinligation. � 2011 Elsevier Inc. All rights reserved.

1 To whom correspondence and reprint requests should beaddressed at Surgical Laboratory, Department of Surgery, ChangGung Memorial Hospital, 5 Fu-Hsing Street, Kwei-Shan Shiang,Taoyuan, Taiwan. E-mail: [email protected], [email protected].

2 These two authors contributed equally to this work.

77

Key Words: insulin; liver regeneration; portal veinligation.

INTRODUCTION

Portal vein embolization has been applied prior tomajor hepatectomy for hepatocellular carcinoma, bileduct malignancy, and metastatic liver tumors [1–3].This procedure acts to embolize the portal branch oflobes targeted for resection, and induces compensatoryhypertrophy in future remnant lobes, which largelyinvolves process of liver regeneration. In a two-thirdspartial hepatectomized animal model, the majority ofliver parenchymal cells that are normally quiescentrapidly reenter the cell cycle. Several growth factorshave been implicated in this process, including TGFa,EGF, HGF, neural stimuli, glucagon and insulin [4].Of them, insulin supplied by the pancreatic isletsperfuses hepatocytes continually through the portalvein. If the amount of portal circulation to the liver isdecreased, the liver atrophies; in contrast, injection ofinsulin prevents or reverses this process [5, 6]. Thus,insulin by itself is not a primary mitogen for hepato-cytes, however, insulin signaling must be present forthe mitogenic signal to proceed normally. In the clinicalsetting, diabetes has been shown as an independent fac-tor to reduce hypertrophy ratio of unembolized lobeafter portal vein embolization [7, 8]. Furthermore,given that up to 30% of patients with liver malignancyare associated with diabetes and their livers are poten-tially less able to regenerate [9, 10], evaluating thephysiological significance of insulin signaling on liverregeneration following portal vein embolization andrelevant molecular events is thus intriguing.

0022-4804/$36.00� 2011 Elsevier Inc. All rights reserved.

Page 2: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

JOURNAL OF SURGICAL RESEARCH: VOL. 166, NO. 1, MARCH 201178

MATERIALS AND METHODS

Induction of Insulin-Deficient Rats

Male Sprague-Dawley rats weighing 250–300 g were used. Rats werehoused under a 12-h light/dark cycle with free access to a standard ratchow and water. All experiments were conducted in accordance withthe Ethics Review Committee for Animal Experimentation of ChangGung University. Diabetes was induced by intraperitoneal injection ofstreptozotocin (STZ) (66 mh/kg, Sigma, St. Louis, MO) dissolved insodium citrate buffer, pH 4.5. STZ-injected rats were considered insu-lin-deficient when exhibiting blood glucose levels> 250 mg/mL within3 d after the induction [11].

Portal Vein Ligation (PVL)

Under general anesthesia with an intraperitoneal injection of ket-amine, the portal branches that perfuse the median and left lobes ofthe liver were ligated with fine silk [12, 13]. Animals of the sham-operated group were subjected to laparotomy, and their livers weremanipulated. At indicated times (d 0, 1, 3, 5, and 7) after PVL, theanimals with and without insulin-deficiency were sacrificed, and bloodand livers were collected for further analysis (n¼ 5–6, for each groupand each time point). The ligated lobes (left and median lobes) and un-ligated lobes (right and caudate lobe) were retrieved and weighed, re-spectively. A portion of the unligated lobes of the liver wasimmediately frozen in liquid nitrogen and stored at –80 �C for subse-quent investigation. For histologic and immunohistochemical exami-nation, tissue samples were fixed in 10% neutral-buffered formalin,processed with graded ethanol solutions, and embedded in paraffin.

Small-Animal Single Photon Emission Computed tomography/

Computed Tomography (SPECT/CT)

On purpose of longitudinal observation of liver regeneration ofunligated lobes, another independent experiment was conducted.Insulin-deficient and normal rats before and d 1 and 7 after PVL,respectively, underwent NanoSPECT/CT (Bioscan Inc., Washington,DC) examination (n¼ 6 in each group) [14, 15]. Liver SPECT scanwith 74 MBq of 99mTc sulfur colloid (i.v.) was performed using 5 cmbed position. The 99mTc sulfur colloid (concentrated in a total volumeof 0.2 mL) was injected through the tail vein of the animal. LiverSPECT was performed 10 min after tracer injection. The total scan-ning time period lasted for 5 min. The final images were reconstructedusing iterative method. The images were displayed in PMOD (PMODTechnologies Ltd., Zurich, Switzerland) and volumes of interest(VOIs) were drawn manually. The volume of four liver segments (me-dian, right, left, and caudate lobe) at each indicated time point wasmeasured for analysis.

Evaluation of Liver Regeneration

Restituted liver mass (%) following PVL was calculated using theformula: volume of right lobe þ caudate lobe at d 7/volume of rightlobe þ caudate lobe at d 0 3 100%, based on the results of SPECT/CT analysis. Redistributed volume ratio of unligated lobes to thewhole liver following PVL at each time point was calculated usingthe formula: volume of right lobe þ caudate lobe/volume of the wholeliver 3 100%, based on the results of SPECT/CT analysis. The resti-tuted liver mass and redistributed volume ratio calculated fromSPECT/CT examination were compared with the data derived fromautopsy examination.

Surface Area of Hepatocytes of Un-Ligated Lobes

Paraffin sections (4–5 mm thickness) of the liver were cut andstained with HE. Light microscopic images of the hepatic lobules, at

a magnification of 3400, were analyzed. Ten randomly chosen peri-portal and pericentral visual fields per section were examined. Gener-ally, 200 hepatocytes per section were examined. Surface area of eachhepatocyte was estimated using the following equation: length (themaximal dimension) 3 width (perpendicular dimension) mm2. Then,the mean surface area of hepatocytes of each experimental animalwas calculated.

BrdU (Bromodeoxyuridine) Assay for Evaluation

of DNA Synthesis

The experimental animals received BrdU (Sigma) at a dose of30 mg/kg body weight, intraperitoneal injection 2 h before sacrifice.The livers were removed, fixed in 95% ethanol containing 1% aceticacid at 4 �C for 2 h, and then overnight in ethanol 4 �C, and processedto paraffin sections. Incorporated BrdU was immunohistochemicallystained using a monoclonal anti-BrdU antibody (Becton Dickinson,Mountain View, CA). Ten randomly chosen visual fields per sectionwere examined at a magnification of 3200. Labeling index wasdefined as the mean number of BrdU-positive cells per HPF (3200).

Immunohistochemical Stainings for PCNA (Proliferative Cell

Nuclear Antigen) and Ki-67

Formalin-fixed, paraffin-embedded tissues were cut into 4-mmsections and mounted on glu-coated slides. A modification of theavidin-biotin-peroxidase complex immunohistochemical method wasperformed. Slides were heated at 60 �C for 60 min, then deparaffi-nized in xylene, and rehydrated in graded alcohols. Endogenousperoxidase was blocked by incubation in 0.3% hydrogen peroxidasein methanol, and slides were rehydrated and washed in PBS, pH7.4 for 15 min. Sections were then blocked with 10% normal goatserum in PBS with 1% bovine serum albumin for 15 min. The blockingserum was decanted, and various primary antibodies (PCNA andKi-67) were applied, respectively, in PBS with 1% bovine serum albu-min for 16 h at 4 �C. Slides were washed in PBS with 1% Tween 20 for10 min. After three PBS rinses, biotinylated goat anti-rabbit immuno-globulin (Vector Labs, Burlingame, CA) at a dilution of 1/500 wasapplied for 30 min at room temperature. Following another PBSrinse, avidin DH:biotinylated horseradish peroxidase complex (VectorLabs) was applied for 30 min. After a final PBS rinse, the tissuesections were reacted with 0.06% diaminobenzidine (Sigma) for5 min, rinsed, counter-stained with hematoxylin, dehydrated withgraded alcohols, cleared xylene, and coverslipped with permount.Ten randomly chosen visual fields per section were examined ata magnification of 3200. Labeling index was defined as the meannumber of PCNA-positive cells or Ki-67-positive cells per HPF(3200), respectively.

Quantitative Real-Time PCR for Telomerase Reverse

Transcriptase (TERT), Cell Cycle-Related Genes and Apoptosis-

Related Genes

Two mg of total RNA was treated with DNAse 1 (Invitrogen, Carls-bad, CA) to remove DNA contamination. Then, the RNA was reverse-transcribed using MMLV reverse transcriptase (Invitrogen) IN A 50-mL reaction mixture. TaqMan primers and probes for quantitative de-tection of cyclin D1, cyclin A2, cyclin B1, PCNA, TERT, bax, Bcl 2, BclXL, Daxx, Fas L, JNK, and 18S rRNA were designed with Primer Ex-press (ABI-Perkin-Elmer, Applied Biosystems, Branchburg, NJ) us-ing the GenBank accession number. cDNA samples were mixedwith 2 3 Universal TaqMan buffer containing the Taq enzyme,primers, and probes in a total volume of 25 mL. The thermal cycle con-ditions were 50 �C 2 min, 95 �C 10 min, and 42 cycles of 95 �C 15 s and60 �C 1 min. All PCRs and analysis were performed by using an ABIPRISM 7700 sequence Detection System (Applied Biosystems). Allsamples were run in triplicate.

Page 3: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

TSENG ET AL.: INSULIN SIGNALING AND PORTAL VEIN LIGATION 79

Telomeric Repeat Amplification Protocol (TRAP) for

Telomerase Activity

Frozen liver tissues were used. The modified Telo TAGGG telome-rase PCR ELISA kit (cat. no. 11854666910; Roche, Mannheim, Ger-many) was performed according to manufacturer’s instruction.

Isolation of Rat Hepatocytes and Cell Cycle Analysis Using Flow

Cytometry

Hepatocyte was isolated from adult male Sprague-Dawley rats(300–400 g) by two-step method as previously described. The cellsuspension was filtered through 70 mm nylon mesh (BD Falcon, [Bec-ton, Dickinson, Bedford, MA]), and washed twice with serum-freeDMEM at 200 3 g for 10 min. For ethanol fixation, we resuspendedhepatocyte in 70% ethanol for 1 d and filtered through 60 mm nylonmesh, and centrifuged at 1000 rpm for 5 min twice. The hepatocyteswere treated with RNase (Sigma) and propidiumiodide (Sigma Chem-ical Co.) at a final concentration of 2 3 106 cell per mL. The sampleswere run on BD FACSCalibur (Becton, Dickinson). Ten thousandstained hepatocyte were analyzed from each sample, and a DNA his-togram was generated using a computer with BD CellQuest Pro (Bec-ton, Dickinson).

Detection of Apoptosis

The DNA fragmentation of the specimen was monitored by terminaldeoxynucleotidyltransferase-mediated UTP end labeling (TUNEL)staining using an Apoptag peroxidase in-situ apoptosis-detection kit(Intergen, Purchase, NY) according to the manufacturer’s instruc-tions.

Western Blotting for IGFR (Insulin-Like Growth Factor

Receptor)-2 and TGFb1

Frozen tissue was cut and homogenized in 200 mL of ice-cold lysisbuffer (10 mM Tris-HCl, pH 7.5, 1 mM MgCl2, 1 mM EGTA, 0.1 mMphenylmethylsulfonyl fluoride, 5 mM b-mercaptoethanol, 0.5%Chaps, 10% glycerol). Following incubation for 30 min on ice, the sam-ples were centrifuged at 14,000 rpm for 20 min at 4�C, and the super-natant then was transferred to a new tube. Total protein wasmeasured using a Bio-Rad Bradford kit (Bio-Rad Laboratories, Hercu-les, CA), and 30 mg of total protein then was run on a 12% SDS-poly-acrylamide gel and transferred to a nitrocellulose membrane, whichwas blocked using non-fat dry milk in TBS-T (20 mM Tris-HCl, pH7.6, 150 mM NaCl, 0.05% Tween 20) overnight at 4�C. The membranewas then probed using a primary antibody, washed several times withTBS-T, and incubated using a horseradish peroxidase (HRPO)-conju-gated secondary antibody (goat anti-mouse HRPO, transduction). Fi-nally, the membrane was washed and developed using an enhancedchemiluminescence system (Pierce, Rockford, IL). The primary anti-bodies used were against IGFR -2 and TGFb1. Tubulin-a served as in-ternal control. Each targeted protein from the same animal wasnormalized using the tubulin-a, and then the ratio of normalized tar-get protein to the sham was determined. Quantification of the autora-diographs was conducted using a Bio-Rad/GS 700 imagingdensitometer.

Statistical Analysis

All continuous variables were expressed as the mean 6 SD. Statis-tical analysis for continuous variables was performed using a Stu-dent’s t-test. A P value< 0.05 was considered to representstatistically significant difference between tested data sets.

RESULTS

Evaluation of Liver Regeneration Following PVL

Following PVL, the ligated lobes (median and leftlobe) became atrophic; while unligated lobes (rightand caudate lobes) became hypertrophic (Fig. 1A).The effect of compensatory mass shifting from ligatedlobes to unligated lobes of the liver reached the plateauaround d 7 following PVL. The represented features of99mTc sulfur colloid NanoSPECT/CT performed beforeand d 1 and 7 after PVL were shown (Fig. 1B). Therestituted liver mass of unligated lobes in normal ratswas up to 270% at d 7 following PVL compared with198% in insulin-deficient rats, based on SPECT/CTexamination (P< 0.01) (Fig. 1C). Redistributed volumeratio of unligated lobe to the whole liver in normal ratswas up to 80% 6 3% at d 7 following PVL compared to69% 6 5% in insulin-deficient rats, based on SPECT/CT examination (P< 0.01) (Fig. 1D). The data of resti-tuted liver mass and redistributed volume ratio calcu-lated from SPECT/CT examination were wellcorrelated with those derived from the autopsy exami-nation. Microscopically, the hepatocytes of normalrats swiftly became hypertrophic and arrived at theirplateau around d 1 and 3 following PVL, then shrankgradually and restored to the original size at the d 7,comparable to that before PVL. In contrast, the hepato-cytes of insulin-deficient rats following PVL did nothave a compensatory hypertrophic effect (Fig. 1E).Following PVL, serum ALT surged to 2700 U/L at d 1following PVL in insulin-deficient rats compared withthat of 780 U/L in normal rats (P< 0.01) (Fig. 1F).

BrdU Assay, PCNA, and Ki-67 Staining

Labeling index of BrdU was remarkably increased ininsulin-deficient rats at d 1 and d 3 following PVLcompared with that of normal rats (103 6 61/HPFversus 61 6 6/HPF, P< 0.01; 70 6 22/HPF versus45 6 13/HPF, P< 0.01) (Fig. 2A). Labeling index ofPCNA was remarkably increased in insulin-deficientrats at d 1 and y 3 compared with normal rats(78 6 13/HPF versus 27 6 0.5/HPF, P< 0.01; 59 6 16/HPF versus 6 6 0.5/HPF, P< 0.01) (Fig. 2B). In con-trast, Labeling index of Ki-67 was modest increased ininsulin-deficient rats at d 1 and 3 compared with nor-mal rats (59 6 29/HPF versus 30 6 7/HPF, P< 0.05;109 6 39/HPF versus 93 6 24/HPF, P> 0.05) (Fig. 2C).

Expression of Cell Cycle Markers Detected by Quantitative PCR

The expression of cell-cycle related genes, includingcyclin D1, cyclin A2, PCNA, and cyclin B1, followingPVL in insulin-deficient and normal rats was shown(Fig. 3A–D). Of them, expression of cyclin D1 mRNA

Page 4: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

FIG. 1. (A) Gross appearance of the liver at d 7 following portal vein ligation in Sprague-Dawley rat. (B) The represented feature of 99mTcsulfur colloid Nano SPECT/CT performed before and d 1 and 7 after portal vein ligation in S-D rats. (C) The comparison of restituted liver massafter portal vein ligation among rats with and without insulin-deficiency. (D) The comparison of redistributed volume ratio of the liver afterportal vein ligation among rats with and without insulin-deficiency. (E) The hepatocytes of normal rats became hypertrophic and arrived attheir plateau around d 1–3, then restored to the original size at d 7. In contrast, the hepatocytes of insulin-deficient rats following PVL failedto have a compensatory hypertrophic effect. (F) Comparison of serum alanine transferase among rats with and without insulin-deficiencyfollowing portal vein ligation. Non-DM ¼ non-diabetes; DM ¼ diabetes. * represented p<0.05.

JOURNAL OF SURGICAL RESEARCH: VOL. 166, NO. 1, MARCH 201180

(G1 phase) of normal rats was quickly induced andpeaked at d 1 following PVL, and this was sustainedfrom d 1 to 7, while cyclin D1 expression of insulin-deficient rats peaked around d 1and 3, but decreasedswiftly thereafter. Expression of cyclin A2 (S phase)and cyclin B1 (G2/M phase) of normal rats were quicklyinduced and peaked around d 1 and 3 following PVL,

and decreased swiftly thereafter. The expression ofcyclin A2 and cyclin B1 of insulin-deficient rats gener-ally followed this trend, however, their magnitudewas much decreased compared with those of normalrats. Finally, expression of PCNA in both normal andinsulin-deficient rats was rapidly induced and peakedat d 1, and declined thereafter.

Page 5: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

TSENG ET AL.: INSULIN SIGNALING AND PORTAL VEIN LIGATION 81

Hepatocyte Isolation and Cell Cycle Analysis Using Flow

Cytometry

Fractions of G0/G1-, S-, and G2/M-hepatocytes in in-sulin-deficient and normal rats at d 3 following PVL areshown (Table 1). Of them, the fraction of S-hepatocytesin insulin-deficient rats was statistically increasedcompared with that of normal rats.

Telomerase Activity and Expression of Telomerase Reverse

Transcriptase (TERT) mRNA

Telomerase activity of normal rats was steadily in-creased and peaked at d 5 following PVL, while insu-lin-deficient rats did not have this surge phenomenon(Fig. 4A). Expression of TERT mRNA of normal ratswas in parallel with telomerase activity and peaked atd 5 following PVL, while expression of TERT mRNAof insulin-deficient rats declined dramatically fromd 5 (Fig. 4B).

Hepatocyte Apoptosis of Unligated Lobes Following PVL

Representative features of hepatocyte apoptosis inunligated lobes of insulin-deficient and normal rats fol-lowing PVL detected by TUNEL assay were shown(Fig. 5A and B). The hepatocyte apoptosis of insulin-deficient rats following PVL was generally increasedcompared with that of normal rats (at d 5, 99 6 21/0HPF versus 72 6 29/10 HPF, P< 0.01) (Fig. 5C).

FIG. 2. Comparison of BrdU (A), PCNA (B), and Ki-67 (C) labeling ivein ligation. **, * represented p<0.01 and 0.05, respectively.

Expression of Apoptosis-Related Genes Detected by Quantitative

PCR

The expression of Bax, Bcl 2, Bcl-XL, Daxx, Fas L, andJNK mRNA detected by quantitative PCR in insulin-deficient and normal rats following PVL were shown(Fig. 6A–F). Of them, the pro-apoptotic genes, includingBax, Daxx, and JNK mRNA were increased in insulin-deficient rats following PVL compared with those ofnormal rats; while the anti-apoptotic gene, Bcl XL

mRNA, was decreased in insulin-deficient rats follow-ing PVL compared with that of normal rats.

Expression of IGFR-2 and TGFb1

Representative expression of IGFR-2 and TGFb1 ofinsulin-deficient and normal rats following PVL de-tected by Western blotting is shown (Fig. 7A). Expres-sion of TGFb1 was rapidly induced up to 1.3-foldcompared with that of sham subjects from d 1 followingPVL in both insulin-deficient and normal rats (Fig. 7B).Expression of IGFR-2 of normal rats peaked at d 3 fol-lowing PVL and increased 2.1-fold compared withthat of shamed subjects; in contrast, expression ofIGFR-2 of insulin-deficient rats was increased dramat-ically from d 1and peaked 6.7-fold on d 5 following PVLcompared with that of sham subjects (Fig. 7C).

ndex among rats with and without insulin-deficiency following portal

Page 6: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

TABLE 1

Fraction of Replicating Hepatocytes Following PortalVein Ligation Among Rats With and Without

Insulin-Deficiency

G1/G0-phase (%) S-phase (%) G2/M-phase (%)

Insulin-deficientrats

76.45 6 1.54 13.2 6 0.28* 10.48 6 1.63

Normal rats 82.54 6 2.05 7.49 6 0.028 9.97 6 2.08Sham 91.41 6 2.38 4.71 6 0.03 3.88 6 0.05

*P< 0.05 (insulin-deficient rats versus normal rats).

FIG. 3. Comparisons of cyclin D1 (A), cyclin A2 (B), PCNA (C), and cyclin B1 (D) mRNA expression among rats with and without insulin-deficiency following portal vein ligation.

JOURNAL OF SURGICAL RESEARCH: VOL. 166, NO. 1, MARCH 201182

DISCUSSION

Several important findings were highlighted asfollows. First, hepatocellular damage following PVLwas more severe in insulin-deficient rats comparedwith normal rats, as reflected by the higher value ofserum ALT in the former group. Second, it has beendebated whether liver regeneration following PVL ismainly attributed to hepatocyte hypertrophy (cellularenlargement) or hyperplasia (cellular proliferation)[12, 15–17]. Most studies have examined hepatocyteproliferation as the major mechanism for liver regener-ation, and have focused mainly on the cell-cycle regula-tory system. In the present study, we examined the cellgrowth of unligated lobe in response to PVL by directlymeasuring the cellular size of the hepatocytes. Wefound that transient hepatocyte enlargement emergesswiftly in normal rats, which might provide a support-ive system to maintain appropriate hepatic physiologicfunction before cellular replication is completed. Incontrast, the hepatocytes of insulin-deficient ratsfollowing PVL failed to acquire this compensatoryhypertrophic effect, which otherwise put the insulin-deficient rats in a hazardous situation during earlytime period. Of utmost significance, the restituted livermass of nonligated lobes in normal rats was strikinglyup to 270% on d 7 following PVL compared with 198%in insulin-deficient rats. Further, redistributed volumeratio of nonligated lobes to the whole liver in normal

rats was up to 80% on d 7 following PVL comparedwith 69% in insulin-deficient rats. This discrepancy ofliver regeneration among insulin-deficient and normalsubjects certainly influences the safety of the on-goingoperation, if an extensive hepatectomy is attempted.The data presented in the current study is in accor-dance with Starzl’s studies [5, 6]. Using a specialprocedure called splanchnic division, Starzl et al.proposed that insulin is the most important specifichepatotrophic factor in portal venous blood. Therefore,it is as expected that insulin-deficient rats exhibiteda less regenerated liver mass after PVL.

The principal task of the cell division cycle is to repli-cate DNA (without error during S phase) and to segre-gate the duplicated chromosomal DNA equally to twodaughter cells (during M phase).

Page 7: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

FIG. 4. (A) Telomerase activity among rats with and without insulin-deficiency following portal vein ligation. (B) TERT mRNA expressionamong rats with and without insulin-deficiency following portal vein ligation. **, * represented p<0.01 and 0.05, respectively.

TSENG ET AL.: INSULIN SIGNALING AND PORTAL VEIN LIGATION 83

Unexpectedly, both of BrdU and PCNA labelingindex were remarkably increased in insulin-deficientrats following PVL compared with normal ones, whileKi-67 labeling index was modestly increased in insu-lin-deficient rats compared with normal rats. PCNA isan auxiliary factor for DNA polymerase-d, whose maxi-mal expression correlates with the late G1/S phase, andPCNA expression is linked closely to the incorporationof bromodeoxyuridine, a well accepted marker of DNAsynthesis. As another proliferating nuclear antigen,Ki-67 is present in cells that are replicating in the G1,S, G2, and M stages of the cell cycle [18]. Accordingly,it seemed reasonable to speculate that a considerableportion of G1/S-hepatocytes failed to enter into G2/M

FIG. 5. Representative features of hepatocyte apoptosis at d 5 followdeficiency detected by TUNEL assay. Original magnification 3200. Apcompared to that of normal rats (C). **, * represented p<0.01 and 0.05,

phase. In addition to the molecular regulators thatdrive these processes, a monitoring circuitry ensuresthat S phase is completed before mitosis begins. Theseholoenzymes contain both regulatory (cyclin) and cata-lytic (cdk) subunits. Our data showed remarkablydecreased expression of cyclin A2 (S-phase) and cyclinB1 (G2/M phase) in insulin-deficient rats during liverregeneration, suggesting the presence of stagnatedcell cycle progression in the S-phase.

This speculation was confirmed by the cell cycle anal-ysis using flow cytometry, which indicated that thefraction of S-hepatocytes in insulin-deficient rats was2-fold greater than normal rats, whereas the fractionof G2/M-hepatocytes among insulin-deficient and

ing portal vein ligation among rats with (A), and without (B), insulin-optosis index of hepatocytes was increased in insulin-deficient ratsrespectively.

Page 8: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

FIG. 6. Comparisons of Bax (A), Bcl 2 (B), Bcl XL (C), Daxx (D), Fas L (E), and JNK (F) mRNA among rats with and without insulin-defi-ciency following portal vein ligation.

JOURNAL OF SURGICAL RESEARCH: VOL. 166, NO. 1, MARCH 201184

normal rats were similar. Telomeres are specializednucleoprotein structures at the end of eukaryotic chro-mosomes [19]. Loss of telomeric DNA via the end repli-cation problem limits the proliferative capacity ofprimary cells in vitro at the stage of cellular senescence[20]. Liver regeneration was impaired in mTERT–/–

mice in different model systems of liver regeneration[21]. Specifically, in response to partial hepatectomy,liver mass restoration was delayed, and hepatocytesshowed signs of telomere dysfunction (anaphasebridge) and a suppressed progression through G2/Mstage of the cell cycle. In the present study, expressionof telomerase activity and its catalytic subunit,mTERT, of insulin-deficient rats during liver regenera-tion were generally lower compared with those ofnormal rats, which might lead the blockade of G2/Mstage of the cell cycle in insulin-deficient rats. However,stagnation in the S-phase of replicating hepatocytes ininsulin-deficient subjects alone could not explain theirphenotype of deranged liver regeneration.

After a spectacular phase of hepatic growth and re-structuring, liver regeneration eventually stops [4]. Ithas been demonstrated that FAS system is up-regulatedin the late stage of liver regeneration, with an increasedhepatocyte apoptosis and a down-regulation of Bcl-2[22–24]. In the present study, the pro-apoptotic genes,including Bax, Daxx, and JNK, were increased in insu-lin-deficient rats following PVL compared with those ofnormal rats, while the anti-apoptotic gene, Bcl XL, wasdecreased in insulin-deficient rats compared with thatof normal rats. In net, increased hepatocyte apoptosisoccurred in insulin-deficient group at late stage of liverregeneration compared with nondiabetic group, partlyevidenced by the data of TUNEL assay. The deathdomain-associated protein (Daxx) was originally clonedas a CD95 (FAS)-interacting protein and modulator ofFAS-induced cell death [25]. Daxx function as a pro-apoptotic protein downstream of FAS through activationof c-Jun-N-terminal kinase (JNK) pathway in a FADD-independent manner. Accordingly, we proposed that

Page 9: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

FIG. 7. (A) Representative feature of IGFR-2 and TGFb1 of insulin-deficient and normal rats following portal vein ligation detected byWestern blot. Tubulin-a served as internal controls. (B) Comparison of IGFR-2 expression among insulin-deficient and normal rats. (C)Comparison of TGFb expression among insulin-deficient and normal rats. **, * represented p<0.01 and 0.05, respectively.

TSENG ET AL.: INSULIN SIGNALING AND PORTAL VEIN LIGATION 85

impaired liver regeneration in insulin-deficient ratsmight be partly due to increased hepatocyte apoptosisexecuted by Daxx-JNK pathway, although other formsof cellular death, such as hepatocyte necrosis due touncertain mechanism, cannot be completely excluded.

Several lines of evidence indicate that TGFb1 isinvolved in preventing uncontrolled hepatocyte prolifer-ation following a partial hepatectomy [25, 26]. Further-more, a marked increase in the IGF-II/Man-6-Preceptor co-localizes in those hepatocytes with elevatedintracellular TGFb1. Since the growth factor, IGF-II, isa ligand for the IGF-II/Man-6-P receptor, an increase inreceptor number may serve to enhance the probabilityof hepatocyte replication. IGF-II has been proposed toact as mitogen during fetal development, since its levelis high in fetal and neonatal rats [27]. Nevertheless, therole of IGF-II in liver regeneration is less clear sincethe expression of IGF-II in the liver does not changesignificantly during liver regeneration [28]. On the con-trary, it has been demonstrated that IGF-II/Man-6-Preceptor bi-functionally plays a role in binding and pro-cessing free form of TGFb1 into active form [26, 29].Taken together, TGFb1-IGF-II/Man-6-P receptor alto-gether might act as a brake system of liver regeneration.In the present study, the expression of TGFb1 amonginsulin-deficient and normal group appeared equivalent,whereas the expression of IGF-II/Man-6-P receptor ininsulin-deficient rats was much increased comparedwith that of normal rats. Accordingly, it might implythat increased IGF-II/Man-6-P receptor, which activatedlatent TGFb1 complex, substantially limited liver regen-eration of insulin-deficient rats.

In summary, we demonstrated the significance ofinsulin signaling upon liver regeneration triggered by

portal vein ligation. The net effect of decreased hepato-cyte proliferation due to suppressed telomerase activityand S-phase stagnation, and increased hepatocyteapoptosis due to increased Daxx-JNK activation andIGF-II/Man-6-P receptor expression, contributes tothe deranged liver restitution of impaired insulinsignaling subjects.

REFERENCES

1. Wakabayashi H, Ishimura K, Okano K, et al. Application ofpreoperative portal vein embolization before major hepaticresection in patients with normal or abnormal liver parenchyma.Surgery 2002;131:26.

2. De Baere T, Roche A, Elias D, et al. Preoperative portal veinembolization for extension of hepatectomy indications. Hepatol-ogy 1996;24:1386.

3. Hemming AW, Reed AI, Howard RJ, et al. Preoperative portalvein embolization for extended hepatectomy. Ann Surg 2003;237:686.

4. Michalopoulus GK, DeFrances MC. Liver regeneration. Science1997;276:60.

5. Starzl TE, Porter KA, Kashiwagi N. Surg Gynecol Obstet 1975;141:843.

6. Starzl TE, Watanabe K, Porter KA, et al. Lancet 1976;1:821.

7. Imamura H, Kawasaki S, Brandi Y, et al. Morphometry ofsinusoids and portal hypertension in non-alcoholic cirrhosis.J Hepatol 1994;21:167.

8. Nagino M, Nimura Y, Kamiya J, et al. Changes in hepatic lobevolume in biliary tract cancer patients after right portal veinembolization. Hepatology 1995;21:434.

9. Ikeda Y, Shimada M, Hasegawa H, et al. Prognosis of hepatocel-lular carcinoma with diabetes mellitus after hepatic resection.Hepatology 1998;27:1567.

10. Yanaga K, Matsumata T, Akazawa K, et al. Effect of diabetesmellitus on hepatic resection. Arch Surg 1993;128:445.

11. Matsumoto T, Yamaguchi M, Kuzume M, et al. Insulin genetransfer with adenovirus vector via the spleen safely and

Page 10: Significance of Insulin Signaling in Liver Regeneration Triggered by Portal Vein Ligation

JOURNAL OF SURGICAL RESEARCH: VOL. 166, NO. 1, MARCH 201186

effectively improves post-hepatectomized conditions in diabeticrats. J Surg Res 2003;110:228.

12. Nagano Y, Nagahori K, Kamiyama M, et al. Improved functionalreserve of hypertrophied contralateral liver after portal veinligations in rats. J Hepatol 2002;37:72.

13. Haga S, Ogawa W, Inoue H, et al. Compensatory recovery of livermass by Akt-mediated hepatocellular hypertrophy in liver-specific STAT3-deficient mice. J Hepatol 2005;43:799.

14. Graaf W, Vetelainen RL, Bruin K, et al. 99mTc-GSA scintigra-phy with SPECT for assessment of hepatic function andfunctional volume during liver regeneration in a rat model ofpartial hepatectomy. J Nucl Med 2008;49:122.

15. Satoh K, Yamamoto Y, Nishiyama Y, et al. 99mTc-GSA liverdynamic SPECT for the preoperative assessment of hepatec-tomy. Ann Nucl Med 2003;17:61.

16. Nagy P, Teramoto T, Factor VM, et al. Reconstitution of livermass via cellular hypertrophy in the rat. Hepatology 2001;33:339.

17. Harada H, Imamura H, Miyagawa S, et al. Fate of human liverafter hemihepatic portal vein embolization: Cell kinetic and mor-phometric study. Hepatology 1997;26:1162.

18. Kaita KDE, Pettigrew N, Minuk GY. Hepatic regeneration inhumans with various liver diseases as assessed by Ki-67 stainingof formalin-fixed paraffin-embedded liver tissue. Liver 1997;17:13.

19. Kim NW, Piatyszek MA, Prowse KR, et al. Specific association ofhuman telomerase activity with immortal cells and cancer.Science 1994;266:2011.

20. Satyanarayana A, Wiemann SU, Buer J, et al. Telomere shorten-ing impairs organ regeneration by inhibiting cell cycle re-entryof a subpopulation of cells. EMBO J 2003;22:4003.

21. Rudolph KL, Chang S, Millard M, et al. Inhibition of experimen-tal liver cirrhosis in mice by telomerase gene delivery. Science2000;287:1253.

22. Vail ME, Chaisson ML, Thompson J, et al. Bcl-2 expressiondelays hepatocyte cell cycle progression during liver regenera-tion. Oncogene 2002;21:1548.

23. Takushi Y, Shiraishi M, Nozato E, et al. Expression of anti-apoptotic protein, Bcl-2, in liver regeneration after a partialhepatectomy. J Surg Res 2006;134:93.

24. Tzung SP, Fausto N, Hockenbery DN. Expression of Bcl-2 familyduring liver regeneration and identification of Bcl-x as a delayedearly response gene. Am J Pathol 1997;150:1985.

25. Yang X, Khosravi-Far R, Chang HY, et al. Daxx, a novelFas-binding protein that activates JNK and apoptosis. Cell1997;89:1067.

26. Jirtle RL, Carr BI, Scott CD. Modulation of insulin-like growthfactor-II/Mannose 6-phosphate receptors and transforminggrowth factor-b1 during liver regeneration. J Biol Chem 1991;266:22444.

27. Bondy CA, Werner H, Roberts CT, et al. Cellular pattern ofinsulin-like growth factor-I (IGF-I) and type I IGF receptorgene expression in early organogenesis: Comparison with IGF-IIgene expression. Mol Endocrinol 1990;4:1386.

28. Braun L, Mead J, Panzica M, et al. Transforming growth factor-bmRNA increases during liver regeneration: A possible paracrinemechanism of growth regulation. Proc Natl Acad Sci USA 1988;85:1539.

29. Villevalois-Cam L, Rescan C, Gilot D, et al. The hepatocyte isa direct target for transforming- growth factor-b activation viathe insulin-like growth factor II/mannose 6-phosphate receptor.J Hepatology 2003;38:156.