Hepatocyte-specific deletion of the antiapoptotic protein myeloid cell leukemia-1 triggers proliferation and hepatocarcinogenesis in mice

Hepatocyte-Specific Deletion of the AntiapoptoticProtein Myeloid Cell Leukemia-1 Triggers Proliferation

and Hepatocarcinogenesis in MiceAchim Weber,1* Regina Boger,2* Binje Vick,2 Toni Urbanik,2 Johannes Haybaeck,3 Stefan Zoller,4 Andreas Teufel,2

Peter H. Krammer,5 Joseph T. Opferman,6 Peter R. Galle,2 Marcus Schuchmann,2 Mathias Heikenwalder,3 andHenning Schulze-Bergkamen2,7

Regulation of hepatocellular apoptosis is crucial for liver homeostasis. Increased sensitivity ofhepatocytes toward apoptosis results in chronic liver injury, whereas apoptosis resistance islinked to hepatocarcinogenesis and nonresponsiveness to therapy-induced cell death. Recently,we have demonstrated an essential role of the antiapoptotic Bcl-2 family member Myeloid cellleukemia-1 (Mcl-1) in hepatocyte survival. In mice lacking Mcl-1 specifically in hepatocytes(Mcl-1�hep), spontaneous apoptosis caused severe liver damage. Here, we demonstrate thatchronically increased apoptosis of hepatocytes coincides with strong hepatocyte proliferationresulting in hepatocellular carcinoma (HCC). Liver cell tumor formation was observed in >50%of Mcl-1�hep mice already by the age of 8 months, whereas 12-month-old wild-type (wt) andheterozygous Mcl-1flox/wt mice lacked tumors. Tumors revealed a heterogenous spectrum rangingfrom small dysplastic nodules to HCC. The neoplastic nature of the tumors was confirmed byhistology, expression of the HCC marker glutamine synthetase and chromosomal aberrations.Liver carcinogenesis in Mcl-1�hep mice was paralleled by markedly increased levels of Survivin, animportant regulator of mitosis which is selectively overexpressed in common human cancers.Conclusion: This study provides in vivo evidence that increased apoptosis of hepatocytes not onlyimpairs liver homeostasis but is also accompanied by hepatocyte proliferation and hepatocarci-nogenesis. Our findings might have implications for understanding apoptosis-related humanliver diseases. (HEPATOLOGY 2010;51:1226-1236.)

See Editorial on Page 1110.

The survival of multicellular organisms depends onthe maintenance of tissue homeostasis. Underphysiological conditions, apoptosis contributes

to liver homeostasis by removing damaged hepatocytes.

Proliferation, growth, and programmed hepatocyte celldeath are highly coordinated and tightly controlled eventsin the normal liver.1

On the one hand, increased apoptosis sensitivity con-tributes to liver injury. On the other hand, defectiveapoptosis was demonstrated to lead to excessive hepato-

Abbreviations: aCGH, array-based comparative genomic hybridization; ALT, alanine aminotransferase; AST, aspartate aminotransferase; Bcl-2, B cell lymphoma-2;BrdU, bromodeoxyuridine; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; HCC, hepatocellular carcinoma; IAP, inhibitors of apoptosis protein; IFN, interferon; IL,interleukin; mRNA, messenger RNA; Mcl-1, myeloid cell leukemia-1; RT-PCR, real-time polymerase chain reaction; TGF, transforming growth factor; TUNEL, terminaldeoxynucleotidyl transferase deoxyuridine triphosphate nick-end labeling; WT, wild-type.

From the 1Department of Pathology, Institute of Surgical Pathology, University Hospital Zurich, Switzerland; 2First Department of Medicine, Johannes Gutenberg-University Mainz, Mainz, Germany; 3Department of Pathology, Institute of Neuropathology, University Hospital Zurich, Switzerland; 4Functional Genomics CenterZurich, University of Zurich, Switzerland; 5German Cancer Research Center, Tumor Immunology Program, Heidelberg, Germany; 6Department of Biochemistry, St. JudeChildren’s Research Hospital, Memphis, TN; and 7National Center of Tumor Diseases, Department of Medical Oncology, University Clinic of Heidelberg, Germany.

Received September 15, 2009; accepted October 29, 2009.*These authors contributed equally to this study.Financial Support: This work was supported by grants of the Deutsche Forschungsgemeinschaft, DFG SCHU 1443/3-1, to H.S.B.; Oncosuisse foundation, OCS

02113-08-2007, to M.H. and A.W.; the research foundation at the Medical Faculty Zurich (M.H.), the Prof. Dr. Max-Cloetta foundation (M.H.) and the “Julius-Muller-Stiftung” (M.H.) and the “Kurt-Senta-Hermann”-Stiftung to M.H. and J.H.

Address reprint requests to: Henning Schulze-Bergkamen, National Center of Tumor Diseases, Department of Medical Oncology, University Clinic of Heidelberg, ImNeuenheimer Feld 350, 69120 Heidelberg, Germany. E-mail: [email protected]; telephone: �49 6221 560.

Copyright © 2010 by the American Association for the Study of Liver Diseases.Published online in Wiley InterScience (www.interscience.wiley.com).DOI 10.1002/hep.23479Potential conflict of interest: Nothing to report.Additional Supporting Information may be found in the online version of this article.

1226

cellular survival and has emerged as a major mechanismby which premalignant hepatocytes obtain a competitiveadvantage over normal liver cells.2 Various molecular al-terations have been characterized that cause an imbalancein the regulation of apoptosis. Among these are alterationsin p53 signaling, expression of death receptors, growthfactors, and mitochondrial integrity.3 Decreased activityof proapoptotic signaling as well as increased activity ofantiapoptotic events are associated with hepatocellularcarcinoma (HCC) development and progression.4

Among the main cellular changes that trigger apoptosisof hepatocytes is the permeabilization of the outer mito-chondrial membrane followed by the release of proapop-totic factors.5 The B cell lymphoma-2 (Bcl-2) proteinfamily plays a pivotal role for mitochondrial integrity andthe selective interactions between proapoptotic and anti-apoptotic family members regulate mitochondrial activa-tion.6 Bcl-2 family members are similar within the Bcl-2homology (BH) regions (BH1 through BH4) and can bedivided into proapoptotic and antiapoptotic Bcl-2 pro-teins.

Proapoptotic Bcl-2 proteins comprise (1) multido-main members, which lack the BH4 domain (e.g., Bcl-2–associated X protein [Bax], Bcl-2 homologous antagonist/killer [Bak]) and (2) BH3-only proteins, which lack BH1,BH2, and BH4 domains (e.g., Bid, Noxa, Puma [p53up-regulated modulator of apoptosis]). BH3-only pro-teins initiate the mitochondrial signaling cascade by sens-ing cellular damage.7 After activation, BH3-only proteinsare released to neutralize antiapoptotic Bcl-2 proteins.Subsequently, Bax and Bak trigger mitochondrial mem-brane leakage and the release of mitochondrial proteins,including cytochrome c, Smac/DIABLO (second mito-chondria-derived activator of caspases/direct IAP-bindingprotein with low pI), and apoptosis-inducing factor(AIF). Smac/DIABLO proteins inactivate the inhibitorsof apoptosis protein (IAP) family, which consists ofIAP1/2, BRUCE, NAIP, ILP2, ML-IAP, Survivin, andX-linked IAP (XIAP). XIAP is a direct caspase inhibitor.Other IAPs including Survivin have several functionsapart from caspase inhibition, e.g., triggering of ubiquiti-nation processes.8 Antiapoptotic Bcl-2 family members(e.g., Bcl-2, Bcl-xL, and myeloid cell leukemia-1 [Mcl-1]),interact with Bax and Bak to inhibit the activation ofmitochondria.7

Both Bcl-xL and Mcl-1 have been identified as majorantiapoptotic Bcl-2 proteins in the liver.9-11 Liver ho-meostasis is severely disturbed in Mcl-1�hep mice.10,11

Spontaneous hepatocyte apoptosis was observed in liversof Mcl-1�hep mice resulting in profound liver cell damageand increased susceptibility of hepatocytes toward pro-apoptotic stimuli.10 In addition, Mcl-1 has been shown to

be highly expressed in a subset of human HCC, contrib-uting to apoptosis resistance of cancer cells.12,13 Thus,abrogation of the prosurvival function of Mcl-1, either by(1) diminishing its levels or (2) by inactivating its func-tion, have shown promising results with regards to treat-ment of HCC.12,13

In this study, we show that liver-specific depletion ofMcl-1 increases hepatocyte apoptosis, induces hepatocel-lular proliferation, and causes HCC in the absence ofovert inflammation.

Materials and Methods

Mcl-1 Knockout Mice. Conditional liver-specificMcl-1 knockout mice (homozygous: Mcl-1flox/flox-Alb-Cre, referred to as Mcl-1�hep; heterozygous: Mcl-1flox/wt

-AlbCre referred to as Mcl-1flox/wt; and control littermates:Mcl-1wt/wt) were generated and genotyped as described.10

Animal experiments were performed as described else-where.10 All animals received humane care according tothe criteria published by the National Institutes ofHealth.14

Aminotransferase Levels. Levels of alanine amino-transferase (ALT) and aspartate aminotransferase (AST)were determined as described.10

TUNEL Assay. Terminal deoxynucleotidyl trans-ferase deoxyuridine triphosphate nick-end labeling(TUNEL) assays were performed as described.10

Caspase Assay. Caspase 3, caspase 8, and caspase 9activity was measured as previously described.15

Histology and Immunohistochemistry. Histologyand immunohistochemistry on a Ventana stainer fromRoche were performed as described.10,16 Transforminggrowth factor beta 1 (TGF�1) was stained with a poly-clonal antibody (immunoglobulin G, species rabbit) fromSanta Cruz Biotechnology Inc. (Code: sc-146, 1:200),p53 was stained with a polyclonal antibody (immuno-globulin G, species rabbit) from Imgenex (IMG-80442,1:200). A6 antibody was kindly provided by ValentinaFactor.

DNA and RNA Isolation, Quantitative ReverseTranscription Polymerase Chain Reaction. DNA andRNA isolation as well as quantitative reverse transcriptionpolymerase chain reaction (RT-PCR) were performedas described.10 For quantitative RT-PCR, QuantiTectprimers (Qiagen, Hilden, Germany) for murine Mcl-1,10 Bcl-xL (QT00149254), XIAP (QT01755649),Noxa (QT00142940), Puma (QT01657432), Mcl-1(QT01038730), CD95/APO-1/Fas (QT00095333),CD95L (QT00104125), Ciapin1 (QT00118986), Bid(QT00145061), cFLIP (QT01776033), cIAP1 (forward:5�-cgaggaggaggagtcagatg-3�; reverse: 5�-gtgatggcccttg-

HEPATOLOGY, Vol. 51, No. 4, 2010 WEBER, BOGER, ET AL. 1227

cacttag-3�), Il1� (forward: 5�-gcaactgttcctgaactcaact-3�;reverse: 5�-atcttttggggtccgtcaact-3�), and glyceraldehyde3-phosphate dehydrogenase (GAPDH, NM_008084,XM_001003314, XM_990238, QT01658692) wereused.

Tissue Lysis and Immunoblotting. Immunoblotanalyses were performed as described,10 using the follow-ing primary antibodies: Mcl-1 (Rockland, Gilbertsville,PA), Bcl-xL (H-62; Santa Cruz Biotechnology, Heidel-berg, Germany), Survivin (NB500-201; Novus Biologi-cals, Littleton, CO), lamin (2032; Cell SignalingTechnology, Danvers, MA) and �-tubulin (Sigma).

Detection of Cell Proliferation. Cell proliferationwas assessed by bromodeoxyuridine (BrdU) and Ki67staining as reported previously.10

Array-Based Comparative Genomic Hybridiza-tion. Agilent oligonucleotide array-based comparativegenomic hybridization (aCGH) for genomic DNA anal-ysis for formalin-fixed paraffin-embedded samples(Mouse Genome CGH Microarray 4�44K) was per-formed on paraffin-embedded liver tissues according tothe protocol provided by Agilent Technologies. Chromo-somal copy number aberration in HCC samples of Mcl-1�hep mice in relation to wild-type (WT) samples wereinvestigated by aCGH.17 Log2-ratios of signal intensityvalues of WT (Cy5) versus signal intensity values of HCC(Cy3) samples were computed with Agilent Feature Ex-traction software, version 9.5.3.1. Log2 ratios were im-ported into the DNA Analytics Software 4.0.76 (AgilentTechnologies, Santa Clara, CA). Saturated and nonuni-form data points were filtered out. Values of probes thatoccurred several times within one chip were combinedand averaged. The aCGH data were then normalized in alinear way using the DNA Analytics centralizationmethod. Aberrations were detected using the AberrationDetection Method Nr.1 (ADM-1) as implemented in theDNA Analytics software17 with standard settings. Thoseaberrations that were not covered by more than twoprobes were filtered out. Single log2 ratio intensities, mov-ing average of these ratios, and aberration detection re-sults were graphically displayed in the genome browser ofthe DNA Analytics software. Statistical significance ofamplification and deletion patterns in aCGH for mono-clonal tumors was calculated by applying a permutationtest. The samples were compared pairwise as follows, us-ing a program written in-house. First, the sequence over-lap (o) of amplifications/deletions was calculated for thetwo samples. Then, the amplifications/deletions of onesample were kept but randomly distributed on the othersample and the new overlap (ri) was calculated. This stepwas repeated n � 1 � 107 times, and r � sum (ri � o) was

computed. Finally, the P value for the pairwise compari-son was estimated as p � r/n.

Original data from aCGH on HCC tissues of Mcl-1�hep mice as well as liver tissues of wild-type control miceare available in the Gene Expression Omnibus (GEO)database under accession number GSE16580.

Data Analysis. All graphs represent at least three in-dependent experiments. Histological images show repre-sentative results. Data were analyzed by Mann-WhitneyU test using SPSS software, with P � 0.05 consideredsignificant.

Results

Sustained Chronic Liver Injury and Increased He-patocyte Proliferation in Aged Mcl-1�hep Mice. Wehave previously shown that deletion of Mcl-1 in hepato-cytes results in liver injury of mice �6 months of age,caused by spontaneous induction of hepatocellular apo-ptosis.10 In the present study, we examined the long-termconsequences of liver-specific deletion of Mcl-1 by inves-tigating the liver phenotype of adult Mcl-1�hep mice �6months of age. First, we tested the ablation efficiency ofMcl-1 in livers of 12-month-old Mcl-1�hep mice. Mcl-1protein and messenger RNA (mRNA) expression werestrongly reduced or virtually absent in whole-liver extractsof Mcl-1�hep mice compared to age matched wild-typeanimals (Fig. 1A,B). The residual low expression levels ofMcl-1 were most likely attributed to nonparenchymalliver cells.

Although no significant differences in body weightwere detected between Mcl-1�hep and control mice from0-12 months of age, liver weight was significantly reduced(P � 0.05) in 2-month-old and 4-month-old Mcl-1�hep

animals.10 These differences decreased with age:8-month-old and 12-month-old Mcl-1�hep animals didno longer show a significant reduction of liver/bodyweight ratio compared to age-matched controls (Fig. 1C).Furthermore, aminotransferase levels were determined asa surrogate marker for liver cell damage. No significantelevation of AST and ALT levels was found in 8-month-old Mcl-1�hep mice. This was in contrast to the significantdifferences (P � 0.05) in aminotransferase levels observedin sera of Mcl-1�hep mice at 2 and 4 months of life shownpreviously (Fig. 1D).10 Interestingly, a significant rise inserum aminotransferase levels was again reproducibly de-tected in 12-month-old Mcl-1�hep mice (P � 0.05; Fig.1D).

As already observed at the age of 4 months,10 8-month-old to 12-month-old Mcl-1�hep mice still showed an ab-normal liver morphology. Mcl-1�hep livers displayednumerous pleomorphic and atypical hepatocytes and an

1228 WEBER, BOGER, ET AL. HEPATOLOGY, April 2010

altered, remarkably nodular liver structure (Fig. 1E, rightpanel), which was in contrast to livers of age-matchedwild-type and heterozygous Mcl-1flox/wt mice. This wasaccompanied by a subtle, mostly pericellular fibrosis, notfound in control littermates (Fig. 1E).

Similar to 4-month-old and younger mice,10 8-month-old and 12-month-old Mcl-1�hep mice histologically alsostill showed an increased rate of liver cell apoptosis, high-lighted by Caspase 3 staining and TUNEL assay (Fig.2A), which was not observed in wild-type and heterozy-gous Mcl-1flox/wt mice (data not shown). This was paral-leled by an increased activity of Caspase 3 and Caspase 9in liver homogenates of Mcl-1�hep mice (Fig. 2B). Caspase8 activity in livers of Mcl-1�hep mice, however, was notsignificantly different compared to wild-type and Mcl-1flox/wt mice (Fig. 2B).

To test for potential compensatory antiapoptoticmechanisms in livers of Mcl-1�hep mice, the expression of

several apoptosis-related factors was analyzed. Remark-ably, strongly elevated transcript levels of Survivin, a pro-tein of the IAP family associated with hepatocyteproliferation and carcinogenesis,18,19 were detected inMcl-1flox/wt and Mcl-1�hep livers (Fig. 2D). On the proteinlevel, Survivin expression was significantly higher in nu-clear fractions of Mcl-1�hep compared to WT livers (Fig.2C). No differences in Survivin protein expression wereobserved in the cytosolic fraction (data not shown). Incontrast to Survivin, XIAP and cIAP-1—other membersof the IAP family—were not up-regulated (Fig. 2D). Pre-vious reports suggested that proteins of the death-induc-ing signaling complex (DISC), such as CD95/APO-1/Fasand cellular FLICE-inhibitory protein, long isoform (c-FLIP), can couple cell death and proliferation in hepato-cytes.20 However, neither transcript levels of CD95 norcFLIP, nor transcript levels of the ligand of CD95,CD95L, were significantly different comparing livers of

Fig. 1. Chronic liver injury in Mcl-1�hep

mice. (A) Whole-liver extracts derived from12-month-old wild-type (WT) and Mcl-1�hep

mice were analyzed by immunoblot analysisfor the protein expression of Mcl-1 andTubulin (loading control). kDa, kilodalton.(B) Whole-liver extracts from 12-month-oldWT, heterozygous Mcl-1flox/wt, and Mcl-1�hep

mice were analyzed for the mRNA expres-sion levels of Mcl-1 by quantitative RT-PCR.(C) Liver/body weight ratio of control (graybars) and Mcl-1�hep (black bars) mice wasdetermined. Boxes represent the average.Standard deviation is indicated by errorbars. Asterisk (*) indicates a statistical sig-nificance of P � 0.05; n.s., statistically notsignificant. (D) Serum AST and ALT levels ofcontrol (gray bars) and Mcl-1�hep (blackbars) mice. Mean levels of control and Mcl-1�hep mice are shown. Boxes represent theaverage. Standard deviation is indicated byerror bars. Asterisk (*) indicates a statisti-cal significance of P � 0.05; n.s., statisti-cally not significant. Aminotransferaselevels of 1-month-old to 4-month-old ani-mals have been published previously10 andare shown for completeness. (E) Livers ofWT, heterozygous Mcl-1flox/wt, and Mcl-1�hep

mice, all 12 months old, were analyzedhistologically. Consecutive paraffin sectionswere stained: hematoxylin/eosin (HE), col-lagen IV (col IV), and Sirius red. Scale bar:100 �m.


Mcl-1�hep mice to wild-type and Mcl-1flox/wt mice (Fig.2D, middle panel). Furthermore, hepatic mRNA expres-sion of the antiapoptotic Bcl-2 proteins Bcl-xL, and Bcl-2(Fig. 2D, middle panel; Supporting Fig. 1), the BH3-onlyproteins Bid, Puma, and Noxa, as well as Bax and Bak(Fig. 2D, lower panel), respectively, was analyzed. Nosignificantly different expression levels were found com-paring livers of 12-month-old Mcl-1�hep mice to age-matched WT and Mcl-1flox/wt mice.

Chronic liver injury potentially causes an inflamma-tory reaction. Although no overt inflammatory infiltrateswere detectable histologically (Fig. 1E; data not shown),the expression of several inflammatory mediators wasstudied. Increased levels of IL6, a trigger of Survivin ex-pression and an important regulator of proliferation in theliver,21,22 were found in liver lysates of 12-month-old

Mcl-1�hep mice compared to WT and Mcl-1flox/wt mice(Fig. 3A). Albeit remarkable, these differences were notstatistically different. Similarly, a slight up-regulation oftumor necrosis factor � mRNA (two-fold to six-fold) wasdetected in Mcl-1�hep mice compared to WT and Mcl-1flox/wt mice (data not shown). In contrast, mRNA levelsof interleukin 1� (IL1� and interferon gamma (IFN�)were not different (Fig. 3A). Remarkably, livers of Mcl-1�hep mice revealed scattered cells immunoreactive for thecytokine TGF�, an important inducer of carcinogenesis,which were not detectable in control mice (Fig. 3B).

Next, we addressed whether the relative increase ofliver weight in Mcl-1�hep animals and the strong up-reg-ulation of Survivin might be linked to a higher hepatocyteproliferation rate, which we had observed previously in4-month-old mice.10 Indeed, Mcl-1�hep mice revealed a

Fig. 2. Increased apoptosis in Mcl-1�hep

mice. (A) Liver histology of 8-month-old and12-month-old Mcl-1�hep mice revealed anincreased basal rate of apoptosis that wasnot observed in livers of WT and heterozy-gous Mcl-1flox/wt (data not shown). Hema-toxylin/eosin (HE), activated caspase 3,and TUNEL staining. Arrowheads indicateapoptotic hepatocytes. Scale bar: 20 �m.(B) Analysis for Caspase 3, Caspase 8, andCaspase 9 activity in whole-liver lysates ofWT, heterozygous Mcl-1flox/wt, and Mcl-1�hep

mice. Boxes represent the average. Stan-dard deviation is indicated by error bars.Asterisk (*) indicates a statistical signifi-cance of P � 0.05. (C) Nuclear extracts oflivers from 12-month-old WT and Mcl-1�hep

mice were analyzed by immunoblot analysisfor the expression of Survivin and lamin.kDa, kilodalton. (D) mRNA expression anal-ysis of several apoptosis-related factors(CD95, CD95L, cFLIP, BclxL, Bcl2, Bid,Noxa, Puma, Bax, Bak, survivin, XIAP, andcIAP1) by quantitative RT-PCR. Boxes rep-resent the average expression level of con-trol and Mcl-1flox/wt, Mcl-1�hep mice. Allvalues are normalized to GAPDH mRNA ex-pression. Standard deviation is indicated byerror bars. Asterisk (*) indicates a statisti-cal significance of P � 0.05.


highly significant increase of Ki67� hepatocytes com-pared to WT and heterozygous Mcl-1flox/wt mice at the ageof 8 and 12 months (P � 0.0001, and P � 0.001, respec-tively; Fig. 3C). Remarkably, heterozygous Mcl-1flox/wt

livers also displayed increased proliferation indices com-pared to WT livers. Quantification by BrdU incorpora-tion corroborated this finding: Livers of 8-month-oldMcl-1�hep mice still showed a significantly higher prolif-eration rate when compared to age-matched heterozygousMcl-1flox/wt mice (P � 0.05; Fig. 3D).

Development of Liver Cell Tumors in Mcl-1�hep

Mice. WT and heterozygous Mcl-1flox/wt mice revealedmacroscopically normal livers at the age of 8 and 12months. This was in contrast to age-matched Mcl-1�hep

livers which contained tumors in �50% of Mcl-1�hep

livers (Table 1). Liver tumors ranged from �2 mm to 3

cm in diameter (Fig. 4A). In addition, nontumorous partsof Mcl-1�hep livers revealed a spectrum of findings rangingfrom a macroscopically unremarkable (some animals) to astrongly nodular (most animals) structure (Fig. 3A). His-tologic analysis confirmed that �50% of all Mcl-1�hep

mice (11/21) displayed liver tumors (Fig. 4B). Larger tu-

Fig. 3. Increased hepatocyte prolifera-tion in Mcl-1�hep mice. (A) mRNA expres-sion analysis of whole liver homogenatesderived from WT, Mcl-1flox/wt, and Mcl-1�hep

mice for the inflammatory cytokines IL6,IL1�, and IFN� by quantitative RT-PCR.Boxes represent the average expressionlevel. Standard deviation is indicated byerror bars. All values are normalized toGAPDH mRNA expression. (B) Livers of WT,heterozygous Mcl-1flox/wt, and Mcl-1�hep

mice, all 12 months old, were analyzedhistologically for the expression of TGF�.Arrowheads indicate positive cells, whichmorphologically appeared to be nonparen-chymal cells. Scale bar: 50 �m. (C, upperrow) To assess hepatocyte proliferation inWT, heterozygous Mcl-1flox/wt, and Mcl-1�hep

mice, Ki67 staining was performed. Arrow-heads indicate positive cells, which mor-phologically appeared to be hepatocytes.Scale bar: 100 �m. (C, lower row) fiveindependent high-power fields (400�) offour to seven mice per genotype were quan-tified for Ki67-positive hepatocytes and sta-tistically analyzed. Asterisks (*** and **)indicate a statistical significance of P �0.0001 and P � 0.001, respectively. (D)Livers of 8-month-old to 12-month-old Mcl-1flox/wt and Mcl-1�hep mice were stained forBrdU incorporation (left panels), and theratio of BrdU-positive nuclei to total livercells was calculated (right panel). Boxesrepresent the average. Standard deviationis indicated by error bars. Asterisk (*) indi-cates a statistical significance of P � 0.05.

Table 1. Development of Liver Tumors in Mcl-1�hep Mice

Genotype Mice (n)Incidence of Liver

TumorsNumber of TumorNodules (Range)

Wild-type 9 0/9 0Mcl-1flox/wt 15 0/15 0Mcl-1�hep 21 11/21 1–14

At the age of 8 and 12 months, no tumors were detected in wild-type orheterozygous Mcl-1flox/wt mice. More than 50% of the Mcl-1�hep mice revealed upto 14 macroscopically detectable tumor nodules per liver.


mors showed cellular atypia, altered liver-architecturewith broadening of liver cell cords (highlighted by colla-gen IV staining), and loss of reticulin fibers (shown byGomori staining). In addition, the proliferation rate againincreased compared to nontumorous areas, and a focal

pattern of strong immunoreactivity for glutamine syn-thetase was observed (Fig. 4C). These findings supportthat the tumors histologically qualified as HCC.

Mcl-1–deficient livers displayed different staining pat-terns for the oval cell marker A6. Mostly, tumors andnontumorous tissues were A6-negative (A6). However,in several instances A6� tumors, surrounded by A6 livertissue, were detected. Besides, we could also detect oneA6 tumor surrounded by A6� liver tissue (Fig. 4D).

Because tumors might have originated from a minutesubset of Mcl-1–positive hepatocytes due to a lack of Cre-recombination, quantitative real-time RT-PCR and im-munoblot analyses of tumor nodules in Mcl-1�hep micewas performed. These analyses revealed only marginalMcl-1 mRNA and protein expression compared to WTlivers (Fig. 4E; data not shown). This strongly suggeststhat tumors did not originate from a conceivable subset ofhepatocytes with a growth advantage due to leaky knock-out of Mcl-1, but rather from Mcl-1–deficient hepato-cytes.

Finally, we set out to investigate whether HCC nod-ules of Mcl-1�hep mice contain chromosomal aberrations.Five HCCs (ranging from 5-30 mm in diameter) derivedfrom independent Mcl-1�hep livers were analyzed byaCGH analysis. This revealed numerous, chromosomalaberrations with amplifications and deletions on severalchromosomal regions that were statistically significant(P � 0.05; Fig. 5). No clearly mutual pattern of chromo-somal aberrations was detected in Mcl-1�hep HCCs.These observations not only confirmed the neoplastic na-ture of the tumor nodules, but also indicated that HCCscontained different chromosomal aberrations.

To further explore possible signaling mechanisms, whichmay contribute to hepatocarcinogenesis in the presentedmodel, p53 expression was analyzed. No significantly differ-ent mRNA expression levels were detected when livers ofMcl-1�hep mice were compared to WT and Mcl-1flox/wt mice.In addition, no p53 accumulation was detectable by immu-nostaining, in neither tumor nor nontumor tissues of Mcl-1�hep mice (Supporting Fig. 1B). Based on these findings,there was no evidence for p53 being a key factor for HCCformation in the presented model.

Enhanced expression of the vascular endothelialgrowth factor-A (VEGF-A) has been discussed as beinginvolved in hepatocarcinogenesis.23 Although a few tu-mors revealed a slightly enhanced VEGF-A expression byimmunohistochemistry, this was not a constant finding(Supporting Fig. 1C).

DiscussionHCC is one of the most common cancers worldwide

and frequently develops in the context of chronic liver

Fig. 4. Development of liver cell tumors (HCC) in Mcl-1�hep mice. Inthe picture 55 for Tubulin (A) Macroscopic view of livers from WT,Mcl-1flox/wt, and Mcl-1�hep mice are shown. The macroscopic spectrum ofMcl-1�hep livers with tumors is shown. Scale bar: 1 cm. (B) Represen-tative HE-stained liver sections confirm hepatocellular tumors rangingfrom small nodules to tumors of several centimeters in diameter. Scalebars: 500 �m (left), 1 mm (middle), 500 �m (right). Tumor borders aremarked by asterisks. (C) Tumor sections were stained for Collagen IV (colIV), reticulin fibers (Gomori staining [gom]), Ki67, and glutamine syn-thetase (GS). Scale bars: 500 �m (upper panel), 100 �m (lower panel).(D) Presence of A6-positive cells was investigated in livers of Mcl-1�hep

mice containing HCC. T, tumor. Tumor borders are marked by asterisks.Scale bar: 50 �m. (E) Whole liver homogenates of 12-month-old WTmice and tumor nodules from Mcl-1�hep mice were analyzed for Mcl-1and Tubulin expression. kDa, kilodalton.


disease and cirrhosis.24 However, the molecular mecha-nisms causing this sequence of events are still poorly un-derstood. In this study, we describe HCC development inmice with hepatocyte-specific depletion of the antiapop-totic Bcl-2 family member Mcl-1.

Apoptosis is generally considered a tumor-preventingmechanism, because it removes unwanted or dangerouscells, e.g., those with oncogenic alterations. Conversely,evasion of apoptotic cell death is considered a basic cellu-lar feature contributing to cancer.25 We have recentlyshown that Mcl-1 is a crucial antiapoptotic factor in hepa-tocytes.10,26 It is well known that liver cell death throughapoptosis is a key pathogenic feature of acute and chronicliver diseases, including cholestasis, hepatitis C virus in-fection, as well as alcoholic and nonalcoholic steatohepa-titis.27 Because mitochondrial activation is a central eventin the induction of hepatocellular apoptosis, Bcl-2 familymembers play a pivotal role for the apoptosis regulation ofhepatocytes. To date, five antiapoptotic Bcl-2 familymembers are known: Bcl-2, Bcl-w, Bfl-1, Bcl-xL, andMcl-1. However, only deletion of Bcl-xL or Mcl-1 re-sulted in severe liver phenotypes due to apoptosis induc-tion.9-11

Controlled hepatocyte apoptosis is essential for liverhomeostasis. However, apoptosis can also induce com-pensatory proliferation of hepatocytes. Here, we showthat increased apoptosis of hepatocytes, due to the lack of

Mcl-1, finally results in hepatocarcinogenesis. At firstglance, our data indicating that increased hepatocyte ap-optosis can cause liver cancer are seemingly incongruouswith the known phenomenon of apoptosis resistance ofpremalignant and malignant hepatocytes. However, ourresults suggest a link between uncontrolled hepatocyteapoptosis, hepatocyte proliferation, and hepatocarcino-genesis. In line with these findings, our study demon-strates chronic liver damage and aberrant liverarchitecture in 8-month-old and 12-month-old Mcl-1�hep mice. In addition, pericellular fibrosis triggered bychronic liver injury could be observed. Similar data wereobtained in Bcl-xL–deficient livers, demonstrating a linkbetween apoptosis induction and fibrogenesis.9,11

In a previous study, we reported that liver injury causedby apoptosis induction was accompanied by a decreasedrelative liver weight in 1-month-old to 4-month-old Mcl-1�hep mice.10 Here, we demonstrate that relative liverweight was back to normal in 12-month-old mice Mcl-1�hep mice. The transient increase in relative liver weightcompared to 1-month-old to 4-month-old Mcl-1�hep

mice is presumably caused by a compensatory hepatocel-lular proliferation: Indeed, we observed significantly in-creased proliferation rates of hepatocytes in Mcl-1�hep

mice. The observation we made in heterozygous Mcl-1flox/wt mice further supports this interpretation: Thesemice revealed an increased hepatocyte apoptosis rate com-

Fig. 5. Chromosomal aberrations in HCC of Mcl-1�hep mice shown by aCGH analysis. The q-arm of each chromosome is shown and chromosomenumbers are indicated. Gray and black bars within the symbolized chromosomes represent G bands. Deletions according to the genomic segmentationworkflow are indicated in blue, amplifications in red (see Materials and Methods for details). HCC of five individual Mcl-1hep mice were hybridizedand normalized to liver tissue of age-matched wild-type mice and analyzed by aCGH analysis. Columns next to each chromosome represent individualMcl-1�hep HCC (1, 2, 3, 4, and 5).


pared to WT mice, but significantly lower than Mcl-1�hep

mice. This was paralleled by an increased hepatocyte pro-liferation rate in heterozygous Mcl-1flox/wt mice comparedto WT mice which again was significantly lower com-pared to Mcl-1�hep mice.

In this study, we found sustained caspase 3 and caspase9 activity in 8-month-old and 12-month-old Mcl-1�hep

mice. In contrast, caspase 8 activity was not different inMcl-1�hep hepatocytes, most likely due to the fact thatcaspase 8 activation mainly occurs upstream of mitochon-drial activation. Caspases may not only trigger controlledcell death but also proliferation to preserve homeostasisafter tissue damage.28 Distinct mechanisms of compensa-tory proliferation are well described in Drosophila melano-gaster.29,30 Our data do not prove a direct causalitybetween apoptosis induction in the liver and hepatocyteproliferation. However, it is very likely that the chronicinduction of apoptosis and chronically elevated caspaseactivities, which are observed in livers of Mcl-1�hep mice,may not only cause hepatocyte apoptosis, but also induc-tion of compensatory proliferation.

Further putative mechanisms contributing to the pro-liferative state of Mcl-1�hep hepatocytes and HCC devel-opment might be linked to the strong up-regulation of theIAP family member Survivin we observed. Recent studieshave shown that Survivin is involved in carcinogenesis,not only by its antiapoptotic effects, but also by regulationof mitosis and angiogenesis.31

In contrast to Survivin, mRNA expression of other IAPfamily members (e.g., XIAP, cIAP-1) was not up-regu-lated in Mcl-1�hep livers. Further, potentially interestingplayers of the apoptotic network were studied, but foundto be of minor or no relevance for hepatocarcinogenesis inMcl-1�hep mice: (1) Absence of Mcl-1 was not compen-sated by enhanced expression of other antiapoptotic Bcl-2proteins (Bcl-xL, Bcl-2, or A1). (2) Moreover, the mul-tidomain proapoptotic Bcl-2 members Bax and Bak werenot significantly changed, as described in hepatocytes de-ficient in NEMO/IKK�.32 (3) Besides, no down-regula-tion of the BH3-only protein Bid, essential for deathreceptor-induced activation of mitochondria in hepato-cytes,5 or other BH3-only proteins (Noxa, Puma) wasobserved in of Mcl-1�hep livers.

Persistent apoptosis of hepatocytes could lead to com-pensatory hepatocyte proliferation probably due to an in-creased activation of hepatic progenitor (oval) cells. Ovalcells are located in the periphery of the biliary tract andrepresent a constant source to restore the pool of hepato-cytes. The finding that most of the liver tumors fromMcl-1�hep mice lacked A6� cells argues against a promi-nent involvement of oval cells and liver tumor formationin Mcl-1�hep mice. However, it is also possible that A6

positivity of HCC, which demarcates oval cell origin, islost in the environment of Mcl-1�hep HCC, and thereforea link between oval cell proliferation and HCC develop-ment cannot be absolutely excluded.

Remarkably, HCC development in Mcl-1�hep mice oc-curred independently of overt hepatitis. This is in contrastto recently published mouse models that link HCC for-mation to inflammation, e.g., deletion of nuclear factor�B essential modifier/I�B kinase-gamma (NEMO/IKK�).16,32 In line with the observed absence of morpho-logically overt inflammation, we could also not detect anup-regulation of IFN� or IL1� in livers of Mcl-1�hep micewhen compared to WT controls. Only a slightly increasedexpression of the proinflammatory cytokine IL6 wasfound in livers of Mcl-1�hep mice. Increased levels of IL6in Mcl-1�hep livers are very likely to be produced by acti-vated Kupffer cells, the main source of cytokines in theliver.33 IL6 may also co-contribute to hepatocarcinogen-esis in Mcl-1�hep mice, as described in other HCC models.For example, in dimethylnitrosamine-induced HCC inmice, triggering of IL6 production is considered a keymechanism for chemically induced hepatocarcinogen-esis.34

Because chronic (viral) hepatitis is not only character-ized by inflammation, but also by an increased rate ofapoptosis, we believe that useful insight can also be gainedfrom our model for understanding the impact of in-creased apoptosis on liver cancer development, includingthose diseases with concomitant severe chronic inflamma-tion as well as those diseases without overt chronic in-flammation (e.g., alcoholic liver disease withoutsteatohepatitis). In our model, proinflammatory cyto-kines such as IL6, likely derived from Kupffer cells, as wellas regulators of mitosis such as Survivin may play a majorrole for hepatocarcinogenesis. Indeed, increased levels ofSurvivin and IL6 have also been described in humanHCC tissues and sera of patients at increased risk forHCC development, respectively.35,36

Hepatocellular proliferation in livers of Mcl-1�hep micepreceded HCC development detected in 8-month-oldand 12-month-old Mcl-1�hep animals. These regenerativeresponses may represent a risk factor for HCC formationby triggering genomic aberrations. In line with this hy-pothesis, we found genetic aberrations in tumor nodulesof Mcl-1�hep mice by aCGH. Various genomic alter-ations, including amplifications and deletions, were ob-served.

We detected liver tumors of varying size from the ageof 8 months in Mcl-1�hep mice. Some liver tumors weresmall in size and did not (yet) meet the histological criteriaof HCC. However, morphology on a macroscopic andhistological level of larger tumors, combined with the ex-


pression of glutamine synthetase and overt chromosomalaberrations, confirmed them as HCC.

In summary, HCC were heterogenous as corroboratedby different patterns of morphology, immunohistochem-istry (glutamine synthetase, A6) as well as chromosomalaberrations.

This argues against one particular molecular pathwayinvolved in HCC formation in Mcl-1�hep mice. In con-trast it favors the notion that compensatory mechanismsunderlie HCC formation in Mcl-1�hep mice.

We show that HCC of Mcl-1�hep mice revealed nosignificant expression of Mcl-1, emphasizing that hepato-carcinogenesis in Mcl-1�hep mice occurs in the absence ofthe prosurvival protein Mcl-1. Remarkably, Mcl-1 wasalso found to be highly expressed in human malignancies,including HCC, and is discussed in terms of contributingto apoptosis resistance of HCC cells.12,13 Thus, depend-ing on the context, Mcl-1 plays seemingly contrary rolesin hepatocarcinogenesis. On the one hand, as observed inlivers of Mcl-1�hep mice, Mcl-1 deficiency can result in ahyperapoptotic environment, which provokes compensa-tory up-regulation of antiapoptotic pathways (e.g., Sur-vivin) and compensatory hyperproliferation, finallyresulting in the outgrowth of a malignant cell population.On the other hand, in tumors with Mcl-1–independentinitiation, Mcl-1 overexpression can be acquired duringtumor progression as an antiapoptotic factor of cancercells. Whether Mcl-1 up-regulation in human malignan-cies is causally linked to carcinogenesis or a correlativefinding still has to be examined.

Taken together, we provide the first in vivo evidencethat deletion of an antiapoptotic protein in the liver notonly induces hepatocyte apoptosis but is also accompa-nied by compensatory proliferation in the liver, resultingin HCC formation.

Acknowledgment: We thank the Dana Farber CancerInstitute, Boston, MA, for providing Mcl-1flox/flox mice. Inaddition, we thank Sandra Heine, Silvia Behnke, BirgitRiepl, and Fian K. Mirea for excellent technical assis-tance.

References1. Galle PR. Apoptosis in liver disease. J Hepatol 1997;27:405-412.2. Schulte-Hermann R, Bursch W, Low-Baselli A, Wagner A, Grasl-Kraupp

B. Apoptosis in the liver and its role in hepatocarcinogenesis. Cell BiolToxicol 1997;13:339-348.

3. Schattenberg JM, Galle PR, Schuchmann M. Apoptosis in liver disease.Liver Int 2006;26:904-911.

4. Fabregat I, Roncero C, Fernandez M. Survival and apoptosis: a dysregu-lated balance in liver cancer. Liver Int 2007;27:155-162.

5. Li S, Zhao Y, He X, Kim TH, Kuharsky DK, Rabinowich H, et al. Reliefof extrinsic pathway inhibition by the Bid-dependent mitochondrial re-lease of Smac in Fas-mediated hepatocyte apoptosis. J Biol Chem 2002;277:26912-26920.

6. Youle RJ, Strasser A. The BCL-2 protein family: opposing activities thatmediate cell death. Nat Rev Mol Cell Biol 2008;9:47-59.

7. Huang DC, Strasser A. BH3-only proteins—essential initiators of apopto-tic cell death. Cell 2000;103:839-842.

8. Eckelman BP, Salvesen GS, Scott FL. Human inhibitor of apoptosis pro-teins: why XIAP is the black sheep of the family. EMBO Rep 2006;7:988-994.

9. Takehara T, Tatsumi T, Suzuki T, Rucker EB 3rd, Hennighausen L,Jinushi M, et al. Hepatocyte-specific disruption of Bcl-xL leads to contin-uous hepatocyte apoptosis and liver fibrotic responses. Gastroenterology2004;127:1189-1197.

10. Vick B, Weber A, Urbanik T, Maass T, Teufel A, Krammer PH, et al.Knockout of myeloid cell leukemia-1 induces liver damage and in-creases apoptosis susceptibility of murine hepatocytes. HEPATOLOGY

2009;49:627-636.11. Hikita H, Takehara T, Shimizu S, Kodama T, Li W, Miyagi T, et al. Mcl-1

and Bcl-xL cooperatively maintain integrity of hepatocytes in developingand adult murine liver. HEPATOLOGY 2009;50:1217-1226.

12. Fleischer B, Schulze-Bergkamen H, Schuchmann M, Weber A, BiesterfeldS, Muller M, et al. Mcl-1 is an anti-apoptotic factor for human hepatocel-lular carcinoma. Int J Oncol 2006;28:25-32.

13. Sieghart W, Losert D, Strommer S, Cejka D, Schmid K, Rasoul-Rocken-schaub S, et al. Mcl-1 overexpression in hepatocellular carcinoma: a po-tential target for antisense therapy. J Hepatol 2006;44:151-157.

14. National Institutes of Health. Guide for the Care and Use of LaboratoryAnimals. NIH Publication No. 86-23. United States Department ofHealth and Human Services, National Institutes of Health, Revised 1985.

15. Schulze-Bergkamen H, Fleischer B, Schuchmann M, Weber A, Wein-mann A, Krammer PH, et al. Suppression of Mcl-1 via RNA interferencesensitizes human hepatocellular carcinoma cells towards apoptosis induc-tion. BMC Cancer 2006;6:232.

16. Haybaeck J, Zeller N, Wolf M, Weber A, Wagner U, Kurrer M, et al. Alymphotoxin-driven pathway to hepatocellular carcinoma. Cancer Cell2009;16:295-308.

17. Agilent DNA Analytics 4.0 CGH Module User Guide. Santa Clara, CA:Agilent Technologies Inc.; 2008.

18. Baba HA, Wohlschlaeger J, Schmitz KJ, Nadalin S, Lang H, Benesch A, etal. Survivin is upregulated during liver regeneration in rats and humans andis associated with hepatocyte proliferation. Liver Int 2009;29:585-592.

19. Altieri DC. New wirings in the survivin networks. Oncogene 2008;27:6276-6284.

20. Gilot D, Serandour AL, Ilyin GP, Lagadic-Gossmann D, Loyer P, Corlu A,et al. A role for caspase-8 and c-FLIPL in proliferation and cell-cycleprogression of primary hepatocytes. Carcinogenesis 2005;26:2086-2094.

21. Taub R. Liver regeneration: from myth to mechanism. Nat Rev Mol CellBiol 2004;5:836-847.

22. Gao B. Cytokines, STATs and liver disease. Cell Mol Immunol 2005;2:92-100.

23. Yang JC, Teng CF, Wu HC, Tsai HW, Chuang HC, Tsai TF, et al.Enhanced expression of vascular endothelial growth factor-A in groundglass hepatocytes and its implication in hepatitis B virus hepatocarcinogen-esis. HEPATOLOGY 2009;49:1962-1971.

24. Motola-Kuba D, Zamora-Valdes D, Uribe M, Mendez-Sanchez N. Hep-atocellular carcinoma. An overview. Ann Hepatol 2006;5:16-24.

25. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100:57-70.

26. Schulze-Bergkamen H, Brenner D, Krueger A, Suess D, Fas SC, Frey CR,et al. Hepatocyte growth factor induces Mcl-1 in primary human hepato-cytes and inhibits CD95-mediated apoptosis via Akt. HEPATOLOGY 2004;39:645-654.

27. Guicciardi ME, Gores GJ. Apoptosis: a mechanism of acute and chronicliver injury. Gut 2005;54:1024-1033.

28. Lamkanfi M, Festjens N, Declercq W, Vanden Berghe T, Vandenabeele P.Caspases in cell survival, proliferation and differentiation. Cell Death Dif-fer 2007;14:44-55.


29. Fan Y, Bergmann A. Apoptosis-induced compensatory proliferation. Thecell is dead. Long live the cell! Trends Cell Biol 2008;18:467-473.

30. Huh JR, Guo M, Hay BA. Compensatory proliferation induced by celldeath in the Drosophila wing disc requires activity of the apical celldeath caspase Dronc in a nonapoptotic role. Curr Biol 2004;14:1262-1266.

31. Mita AC, Mita MM, Nawrocki ST, Giles FJ. Survivin: key regulator ofmitosis and apoptosis and novel target for cancer therapeutics. Clin CancerRes 2008;14:5000-5005.

32. Luedde T, Beraza N, Kotsikoris V, van Loo G, Nenci A, De Vos R, et al.Deletion of NEMO/IKKgamma in liver parenchymal cells causes steato-hepatitis and hepatocellular carcinoma. Cancer Cell 2007;11:119-132.

33. Ramadori G, Armbrust T. Cytokines in the liver. Eur J GastroenterolHepatol 2001;13:777-784.

34. Sander LE, Trautwein C, Liedtke C. Is interleukin-6 a gender-specific riskfactor for liver cancer? HEPATOLOGY 2007;46:1304-1305.

35. Augello C, Caruso L, Maggioni M, Donadon M, Montorsi M, Santam-brogio R, et al. Inhibitors of apoptosis proteins (IAPs) expression and theirprognostic significance in hepatocellular carcinoma. BMC Cancer 2009;9:125.

36. Nakagawa H, Maeda S, Yoshida H, Tateishi R, Masuzaki R, Ohki T, et al.Serum IL6 levels and the risk for hepatocarcinogenesis in chronic hepatitisC patients: an analysis based on gender differences. Int J Cancer 2009;125:2264-2269.


Documents

Hepatocyte-specific deletion of the antiapoptotic protein myeloid cell leukemia-1 triggers proliferation and hepatocarcinogenesis in mice