Liver specific overexpression of platelet-derived growth factor-B accelerates liver cancer development in chemically induced liver carcinogenesis

Liver specific overexpression of platelet-derived growth factor-Baccelerates liver cancer development in chemically inducedliver carcinogenesis

Thorsten Maass1, Florian R. Thieringer1, Amrit Mann2, Thomas Longerich3, Peter Schirmacher3, Dennis Strand1,

Torsten Hansen4, Peter R. Galle1, Andreas Teufel1 and Stephan Kanzler1,5

1 Department of Medicine I, Johannes Gutenberg-University, Mainz, Germany2 Center for Biotechnology and Biomedicine, University of Leipzig, Leipzig, Germany3 Institute of Pathology, Ruprecht-Karls University, Heidelberg, Germany4 Institute of Pathology, Johannes Gutenberg-University, Mainz, Germany5 Department of Medicine II, Leopoldina Hospital, Schweinfurt, Germany

A genetic basis of hepatocellular carcinoma (HCC) has been well-established and major signaling pathways, such as p53,

Wnt-signaling, transforming growth factor-b (TGF-b) and Ras pathways, have been identified to be essential to HCC

development. Lately, the family of platelet-derived growth factors (PDGFs) has shifted to the center of interest. We have

reported on spontaneously developing liver fibrosis in PDGF-B transgenic mice. Since HCC rarely occurs in healthy liver, but

dramatically increases at the cirrhosis stage of which liver fibrosis is a preliminary stage, we investigated liver cancer

development in chemically induced liver carcinogenesis in these mice. HCC induction was performed by treatment of the mice

with diethylnitrosamine and phenobarbital. At an age of 6 months, the tumor development of these animals was analyzed.

Not only the development of dysplastic lesions in PDGF-B transgenic mice was significantly increased but also their malignant

transformation to HCC. Furthermore, we were able to establish a key role of PDGF-B signaling at diverse stages of liver cancer

development. Here, we show that development of liver fibrosis is likely through upregulation of TGF-b receptors by PDGF-B.

Additionally, overexpression of PDGF-B also leads to an increased expression of b-catenin as well as vascular endothelial

growth factor and platelet endothelial cell adhesion molecule-1 (PECAM-1/CD31), all factors with established roles in

carcinogenesis. We were able to extend the understanding of key genetic regulators in HCC development by PDGF-B and

decode essential downstream signals.

Hepatocellular carcinoma (HCC) is among the most commonmalignancies worldwide with approximately 550,000 newpatients each year. However, the incidence of HCC has sig-nificantly increased over the past decades, due to an increas-ing prevalence of chronic viral hepatitis, especially chronichepatitis C. Additional causes leading to HCC are alcohol,toxins, hemochromatosis, a1-antitrypsin deficiency and non-alcoholic fatty liver disease.1–4 Despite major efforts to improvediagnosis and treatment of HCC, therapeutic options remainlimited. The main therapeutic strategies are surgical resectionof the tumor or liver transplantation and to a lesser extentnonsurgical treatments including radiofrequency ablation and

percutaneous ethanol injection.5 Most patients are diagnosedat late stages of the disease or with underlying liver cirrhosisand consequently surgical options may no longer be indicated.Although palliative treatments are needed, they remain verylimited. It was only 2 years ago, that an effort to establish effi-cient systemic therapy regimens has succeeded.6 Nevertheless,besides sorafenib, best supportive care remains standard oftreatment. The need for novel therapeutic strategies is obviousand therefore, a better understanding of the underlying patho-mechanisms is imperative.

A genetic basis of the formation of hepatocellular carci-noma has been well established and major signaling pathways,such as p53, Wnt-signaling, transforming growth factor-b(TGF-b), Ras and retinoblastoma protein pathways, have beenidentified to be essential to HCC development.7

Lately, the family of platelet-derived growth factors (PDGF)has shifted to the center of interest. At present, four membersof the PDGF family have been identified, PDGF-A, PDGF-B,PDGF-C and PDGF-D.8 PDGF plays an important role duringembryonic development. Targeted disruption of PDGF-B or thePDGF-b receptor caused ablation of pericytes, normally formingpart of the capillary wall,9 PDGF-overexpression has been linkedto different types of fibrotic disorders and malignancies.10 It has

Key words: HCC, PDGF-B, DEN, TGF-b, b-catenin

Thorsten Maass and Florian R. Thieringer contributed equally to

this work.

DOI: 10.1002/ijc.25469

History: Received 6 Oct 2009; Accepted 29 Mar 2010; Online 20

May 2010

Correspondence to: Stephan Kanzler, MD, Department of

Medicine II, Leopoldina Hospital, Gustav-Adolf-Straße 8, 97422

Schweinfurt, Germany, Tel.: þ49-(0)9721-7202482, Fax:

þ49-(0)9721-7202484, E-mail: [email protected]

Carcinog

enesis

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC

International Journal of Cancer

IJC

been demonstrated to be a key pathogenic factor in multiplesolid tumors. Biological relevance of this signaling pathway hasbeen demonstrated by therapeutic strategies targeting PDGF sig-naling and thereby inhibiting tumor growth.6,11,12 The functionof PDGF family members is mediated through PDGF tyrosinekinase receptors a and b (PDGF-Ra/b), which are activated byPDGFs by forming homo- or heteroreceptor dimers.12 Bothreceptors can activate major signal transduction pathways,including phosphatidylinositol 3-kinase, Ras mitogen-activatedkinase and phospholipase Cc pathway.13 These pathways canaffect various cellular functions including cell growth, prolifera-tion and differentiation. Accordingly, the biologic role of PDGFsignaling may vary from autocrine stimulation of cancer cellgrowth to subtler paracrine interactions involving adjacentstroma and vasculature.12,13

With respect to chronic liver disease and liver cancer,others and we have previously demonstrated an essential roleof all PDGF family members in liver fibrosis, a prerequisiteof HCC.14–17 PDGF-B transgenic mice were demonstrated tospontaneously develop liver fibrosis within a period of 6months.15 PDGF-B is a potent mitogen for hepatic stellatecells.18 In normal liver, there is only weak expression ofPDGF-B. However, during fibrogenesis increased immunore-activity for PDGF-B was shown in human liver.19 Inflamma-tory liver disease like hepatitis B can cause increased expres-sion of PDGF-B.20

As HCC is often observed following liver fibrosis, and livercirrhosis, it can be speculated that PDGF-B overexpressionmay also lead to an increased development of HCC. Liver spe-cific overexpression of PDGF-C in a transgenic mouse modelinduces liver fibrosis, steatosis and HCC.14 Growth factors likePDGF-B are potential targets for treatment of HCC.21 In ourstudy, we report on an essential role of PDGF-B in HCC de-velopment in vivo, investigated by using transgenic mice,which overexpress PDGF-B specifically in liver. Undergoingchemically induced liver carcinogenesis PDGF-B transgenicmice showed an accelerated development of HCC. This findingwas accompanied by changes in b-catenin distribution and up-regulation of vascular endothelial growth factor (VEGF),PECAM/CD31 and fibroblast growth factor (FGF).

Material and MethodsAnimals

PDGF-B transgenic animals had previously been generatedin our group and genotyping was performed as describedbefore.15 All animals were maintained as hemizygous on a FVB/N background. Animal care and animal procedures were in ac-cordance with the governmental and institutional guidelines.

Induction of carcinogenesis

HCC development in mice was achieved by diethylnitros-amine (DEN) and phenobarbital as described.16 To induceliver carcinogenesis, a single DEN (0.05 mg per mouse) i.p.injection was performed at day 7 post partum. Four experi-mental groups were used: heterozygous PDGF-B transgenic

mice, that received DEN and phenobarbital, age-matchedwild-type (WT) mice, that received DEN and phenobarbital,heterozygous PDGF-B transgenic mice without treatment,and age-matched wild-type (WT) mice without treatment. Topromote carcinogenesis in these animals, phenobarbital(0.05%) was continuously added to drinking water.

Initial analysis of livers

Livers were assessed visually and the numbers of tumor nod-ules appearing at the surface of these livers were counted.For histological analysis and measuring the size of lesions,formalin-fixed and paraffin-embedded sections (5 lm) werestained with hematoxylin and eosin.

Immunohistochemistry

Routine immunohistochemistry was performed on frozen sec-tions (7 lm), fixed in 4% paraformaldehyde. Following pri-mary antibodies were used: mouse-anti-b-catenin (1:250,Sigma-Aldrich, Taufkirchen, Germany), rat-anti-CD31 (1:100,BD Pharmingen, Heidelberg, Germany) and rabbit-anti-PDGFR-b (1:100 Santa Cruz, Heidelberg, Germany). Cy3-conjugated rat-anti-mouse (1:200, Jackson Immunoresearch,Suffolk, UK) and AP-conjugated goat-anti-rat (1:400, Sigma-Aldrich) were used as secondary antibodies. Signal detectionwas performed by fluorescence detection or by using FastRed (Roche, Mannheim, Germany) as substrate.

BrdU labeling

Bromodeoxyuridine (5-bromo-2-deoxyuridine, BrdU) labelingexperiments were performed using ‘‘In Situ Cell ProliferationKit’’ (Roche) on 7-lm frozen sections. Mice were injected i.p.30 lg of BrdU per gram body weight and sacrificed after alabeling period of 1 hr. Fixation and processing of the sam-ples were performed according to the manufacturer’s instruc-tions. Signal detection was performed using Fast Red (Roche)as substrate. As a measurement of cell proliferation, the totalnumbers of BrdU-labeled and nonlabeled nuclei in liver tis-sue were counted.

Real-time PCR

Total RNA was isolated using Tri Reagent (Sigma-Aldrich).cDNA was synthesized from total RNA with oligo-dT-pri-mers by using a cDNA Kit (Roche) according to the manu-facture’s manual. Specific mRNA transcripts were quantifiedusing ‘‘Lightcycler FastStart DNA Master SYBR Green I’’(Roche) and with a LightCycler (Roche). Primers used seeTable 1. Determination of gene expression was performedwith the help of LightCycler software package (Roche). Rela-tive gene expression was given as x-fold expression of theused housekeeping gene GAPDH.

Statistical analysis

Mean 6 standard error of the mean are given. For compari-son of groups, the nonparametric Mann–Whitney U-test was

Carcinog

enesis

1260 PDGF-B in liver cancer development

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC

applied. A p < 0.05 was considered to be significant and ap < 0.01 was considered to be highly significant.

ResultsTo validate our PDGF-B overexpressing mouse model, wemeasured the extent of PDGF-B overexpression in PDGF-Btransgenic mice. The expression of PDGF-B was found to bebetween 463- and 1150-fold increased as compared to wild-type animals depending on whether these mice underwentDEN/phenobarbital treatment or not (Fig. 1a).

As the DEN/phenobarbital-treated mice displayed a lowerPDGF-B expression, we investigated the albumin expression inthese mice, since the transgenic PDGF-B construct is driven byan albumin promoter. This analysis revealed that the expressionof albumin in PDGF-B treated animals was 1.9-fold downregu-lated in comparison to the untreated PDGF-B animals (Fig. 1b).However, whether this is the only reason for a less extensiveoverexpression of PDGF-B remains elusive. Besides PDGF-Bbeing upregulated, the expression of the two PDGF-receptorvariants a and b (PDGF-Ra and b) was also found to be up-regulated significantly in PDGF-B transgenic animals. Theexpression of PDGF-Ra was upregulated 11.0-fold in PDGF-B

transgenic nontreated animals and 8.1-fold in DEN/phenobarbi-tal-treated animals. Also, for the PDGF-Rb variant a 4.5-foldand 3.0-fold upregulation was observed in PDGF-B transgenicnontreated and DEN/phenobarbital-treated mice (Figs. 1c and1d). Elevated PDGF-Rb expression was further confirmed byimmune fluorescence. The increased PDGF Rb expression wasfound to be not hepatocyte specific but rather correlated tofibrotic areas and nonhepatocyte cells (Fig. 1e).

Six months after DEN/phenobarbital treatment, macro-scopic analysis of livers of PDGF-B and WT mice revealed thatuntreated animals, neither WT nor PDGF transgenic animals,exhibited any signs of liver cancer and no macroscopic mor-phological changes could be observed. In contrast, wild-typeanimals treated with DEN and phenobarbital developed in av-erage one neoplastic lesion visible on the liver surface. How-ever, development of neoplastic lesions in liver was signifi-cantly higher in PDGF-B transgenic mice. These animalsdeveloped on an average seven lesions on the liver surface after6 months of treatment with DEN and phenobarbital (Fig. 2a).

Representative liver sections were examined with respectto dysplastic nodules and HCC (Figs. 2e, 2f and 2g). Again,untreated WT and PDGF-B transgenic animals did not show

Table 1. Primer used for real-time PCR analyses

Gene Primer sequence Genebank Acc. No.

GAPDH Fw: GGCATTGCTCTCAATGACAA Pos.: 942–961 NM_008084

Product size: 200 bp Rev: TGTGAGGGAGATGCTCAGTG Pos.:1141–1122

rS6 Fw: GTCCGCCAGTATGTTGTCAG Pos.:498–517 NM_009096

Product size: 103 bp Rev: GTTGCAGGACACGAGGAGTA Pos.:600–581

PDGF-B Fw: TCCAGATCTCTCGGAACCTC Pos.:1121–1140 NM_011057

Product size: 178 bp Rev: GGCTTCTTTCGCACAATCTC Pos.:1298–1279

PDGF receptor a Fw: TGGCATGATGGTCGATTCTA Pos.:2870–2889 NM_011058

Product size: 152 bp Rev: CGCTGAGGTGGTAGAAGGAG Pos.:3021–3002

PDGF receptor b Fw: TCAACGACTCACCAGTGCTC Pos.:2801–2820 NM_008809

Product size: 229 bp Rev: TTCACAGGCAGGTAGGTGCT Pos.:3029–3010

TGF-b Fw: TTGCTTGAGCTCCACAGAGA Pos.:1719–1738 NM_011577

Product size: 183 bp Rev: TGGTTGTAGAGGGCAAGGAC Pos.:1901–1882

TGF-b receptor type I Fw: GGTCTTGCCCATCTTCACAT Pos.:941–960 NM_009370

Product size: 211 bp Rev: CAGGGGCCATGTACCTTTTA Pos.:1151–1132

TGF-b receptor type II Fw: GCAAGTTTTGCGATGTGAGA Pos.:553–572 NM_009371

Product size: 197 bp Rev: GGCATCTTCCAGAGTGAAGC Pos.:749–730

Albumin Fw: TCCAGAGAAGGAGAAGCA Pos.:1666–1683 NM_009654

Product size: 157 bp Rev: GAAGCAGGTGTCCTTGTCAG Pos.:1822–1803

VEGF Fw: CAAGATCCGCAGACGTGTAA Pos.:1484–1503 NM_009505

Product size: 339 bp Rev: TTAATCGGTCTTTCCGGTGA Pos.:1822–1803

b-catenin Fw: CTCTTCAGGACAGAGCCAATG Pos.:2331–2351 NM_007614

Product size: 168 bp Rev: ATGCTCCATCATAGGGTCCA Pos.:2498–2479

FGF-2 Fw: CCTTGCTATGAAGGAAGATGG Pos.:440–460 NM_008006

Product size: 110 bp Rev: TCCGTGACCGGTAAGTATTG Pos.:549–530

For real-time PCR, a touchdown procedure with an annealing temperature of 64�C was used.

Carcinog

enesis

Maass et al. 1261

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC

any histological evidence of tumor formation, whereas bothDEN/phenobarbital-treated WT and PDGF-B animals devel-oped dysplastic nodules. Altered/dysplastic nodules werediagnosed significantly more frequent in PDGF-B transgenicanimals. They were diagnosed if nodular hepatocellular pro-liferations with dysplastic hepatocytes were seen. Although intreated PDGF-B transgenic animals 87% of the nodulesshowed such premalignant transformation, only 9% of thenodules in treated WT animals exhibited premalignant path-ologic changes (Fig. 2b). In addition, the size of these lesionswas significantly greater in PDGF-B transgenic animals ascompared to WT animals (Fig. 2c). Also, development ofHCC was observed only in PDGF-B transgenic mice treatedwith DEN/phenobarbital. HCC development was found in25% of the treated PDGF-B transgenic mice, whereas the WTmice treated with DEN and phenobarbital as well as non-

treated WT and PDGF-B transgenic mice did not exhibit anyHCC development (Fig. 2d). HCC was diagnosed if the tu-mor size exceeded one liver acinus, and if significant cellularatypia as well as disruption of liver architecture was seen.

To evaluate the mechanisms leading to a significantlyincreased incidence of dysplastic nodules and HCC develop-ment in the livers of PDGF-B transgenic animals, we analyzedthe rate of proliferation in livers of these animals. Comparingthe proliferation rate of WT animals treated with DEN/pheno-barbital to those not treated with the carcinogens no significantchange in proliferation rate was observed. Thus, DEN/pheno-barbital treatment did not lead to an increased proliferation.Comparing PDGF-B transgenic mice to WT animals (treatedas well as nontreated), the rate of proliferation in the liver ofboth treated and untreated PDGF-B transgenic animals wassignificantly increased. PDGF-B transgenic animals displayed a

Figure 1. Expression of PDGF-B, albumin and PDGF-receptors. PDGF-B transgenic mice express significantly more PDGF-B (PDGF-B�/þ) than

WT animals (WT�/þ). Untreated PDGF-B transgenic animals express more PDGF-B than DEN/phenobarbital-treated transgenic animals (a).

Additionally, the expression of albumin is elevated only in untreated PDGF-B animals (b). PDGF-B transgenic mice showed upregulated

expression of both PDGF-receptors a (c) and b (d). Immunofluorescent-staining of PDGF-receptor b. Bars ¼ 50 lm (e). [Color figure can be

viewed in the online issue, which is available at wileyonlinelibrary.com.]

Carcinog

enesis

1262 PDGF-B in liver cancer development

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC

3.8-fold increase in proliferation index. Still higher prolifera-tion rates were detected in PDGF-B transgenic animals aftercontinuous treatment with DEN/phenobarbital. These animalsexhibited a 5.0-fold increased proliferation index when com-pared to WT animals (Fig. 3). Proliferation was seen predomi-nantly in hepatocytes but not exclusively restricted to these.Taken together, PDGF-B transgenic overexpression leads to anincreased proliferation index but PDGF-B overexpression incombination with DEN/phenobarbital treatment leads to evenhigher rate of proliferation in liver tissue.

Although searching for changes in extra cellular matrix(ECM) related to the observed increased incidence in dysplasticliver lesions and malignant transformation, a significant changein localization and expression of b-catenin, an ECM-relatedprotein, was observed. A massive dislocation of b-catenin fromthe plasma membrane to cytoplasm and nucleus was demon-strated in PDGF-B transgenic mice by means of immunohisto-chemistry (Figs. 4b and 4c). Also, PDGF-B transgenic animalsshowed 2-fold higher b-catenin expression (Fig. 4a). Since inuntreated and treated WT animals no significant changes in

expression or localization of b-catenin were observed, thesechanges could not be attributed to DEN/phenobarbital treat-ment. PDGF-B transgenic animals without treatment showedupregulation of b-catenin expression in comparison to wild-type mice. In contrast, PDGF-B transgenic animals that hadbeen treated with DEN/phenobarbital showed no upregulationof b-catenin expression in comparison to wild-type control,and a decrease in b-catenin expression in comparison tountreated transgenic mice (Fig. 4a). However, in PDGF-Btransgenic animals that had been treated with DEN/phenobar-bital delocalization of b-catenin from the plasma membrane tothe nucleus was found (Figs. 4b and 4c). Platelet endothelialcell adhesion molecule-1 (PECAM-1/CD31) has been shown topromote angiogenesis. As PDGF-B has been proven to play arole in angiogenetic mechanisms, livers of all animals wereimmunohistochemically stained for the blood vessel-related en-dothelial marker CD31. Both PDGF-B transgenic animals ei-ther untreated or treated with DEN/phenobarbital exhibited anincrease in the appearance of CD31-positive regions. CD31staining was specific to endothelial cells like blood vessels and

Figure 2. Macroscopical and pathological analysis of liver and appearance of HCC. Treated wild-type (WTþ) and transgenic (PDGF-Bþ) animals

develop visible tumors on the liver surface whereas untreated wild-type (WT�) and transgenic (PDGF-B�) animals lack these lesions (a). WT

treated animals have less altered lesions than PDGF-B treated animals, whereas untreated WT and PDGF-B animals develop no lesions at all

(b). Comparison of the size of the lesions shows that lesions of treated PDGF-B animals are significantly larger than those of treated WT

animals (c). Pathological analysis and evaluation of the livers shows that only DEN/phenobarbital-treated PDGF-B transgenic animals develop

HCCs (d). H&E staining of liver sections. An early HCC in the right portion and a small clear cell altered/dysplastic nodule in the lower left (e).

Overview of an early HCC and a small rim of surrounding dysplastic liver parenchyma on the left (f). Large altered/dysplastic nodule with

pronounced fatty change (g). Bars ¼ 200 lm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Carcinog

enesis

Maass et al. 1263

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC

sinusoids. We did not find a correlation of CD31 staining withdysplastic lesions or fibrosis. In contrast, the livers of WT ani-mals did not exhibit such a CD31 staining pattern, irrespectiveof DEN/phenobarbital treatment (Fig. 5a).

An angiogenetic background of the increased malignanttransformation potential in PDGF-B transgenic mice was fur-ther supported by expression analysis of VEGF. PDGF-Btransgenic animals exhibited a 2.5-fold higher level of VEGFexpression. Comparing PDGF-B transgenic animals treated

with DEN and phenobarbital to WT animals no significantchanges could be demonstrated (Fig. 5b).

Expression of TGF-b, a known key regulator in HCCdevelopment was also investigated. Contrary to our previousfindings,13 PDGF-B transgenic animals showed a clearly up-regulated TGF-b expression. TGF-b was upregulated 2.3-foldas compared to WT animals. However, TGF-b expression inPDGF-B transgenic animals treated with DEN/phenobarbitalwas comparable to wild-type animals (Fig. 6a). At the sametime, measurement of TGF-b protein by ELISA exhibited nosignificant changes, which is in accordance with our previousfindings (Fig. 6b). Expression profiling of the TGF-b recep-tors (TGF-RI/II) type I and II showed no significant changesof TGF-RI expression between the experimental groups (Fig.6c), TGF-RII expression was observed to be increased 24.5-fold and 15.7-fold in nontreated and DEN/phenobarbital-treated PDGF-B transgenic mice, respectively (Fig. 6d).

Analysis of the expression of FGF-2 showed a 6.27- and5.20-fold upregulation in gene expression for untreated andDEN/phenobarbital-treated PDGF-B transgenic mice. How-ever, there was no significant difference between the PDGF-Btransgenic mice that had received DEN/phenobarbital anduntreated transgenic mice (Fig. 7).

DiscussionBetter understanding of HCC pathogenesis and the mediatorsinvolved is essential for development of new preventive andtherapeutic strategies. There is strong evidence that the fam-ily of PDGFs plays a central role in the course of liver cirrho-sis, and HCC development. PDGF-C has been demonstratedto promote not only liver fibrosis but also development ofhepatocellular carcinoma. At 9 months of age, PDGF-Ctransgenic mice developed HCC.14 In accordance, approxi-mately 70% of HCC tissues had elevated PDGF-Ra levels due

Figure 3. Proliferation rate in the liver. Transgenic PDGF-B animals

both treated and untreated (PDGF-Bþ and PDGF-B�) exhibit a

significant increase in the proliferation in the liver as compared

untreated and treated to WT (WT� and WTþ) animals. In addition,

proliferation in PDGF-B treated animals is also significantly higher

in comparison to PDGF-B untreated animals.

Figure 4. Analysis of b-catenin expression and localization. Relative expression of b-catenin is increased in PDGF-B untreated animals (a).

Whereas WT animals exhibit a regular localization of b-catenin, both treated and untreated PDGF-B transgenic animals demonstrate a

massive dislocation of b-catenin from the plasma membrane of hepatocytes into cytoplasm and nucleus. Nuclei were counterstained with

DAPI (b, c). Bars ¼ 100 lm (b); ¼ 50 lm (c). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Carcinog

enesis

1264 PDGF-B in liver cancer development

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC

to diverse mechanisms associated with a higher metastaticpotential pointing toward an increased PDGF downstreamsignaling in HCC.22,23

Our own group showed development of liver fibrosis inPDGF-A transgenic mice. However, PDGF-A transgenic micedid not show spontaneous development of HCC.24 In a modelof CCl4-induced liver fibrogenesis, induction of all four knownPDGF-isoforms, and the PDGF-Ra and b was shown.25 In ain vivo HCC model, a central role of PDGF signaling in thecourse of hepatocarcinogenesis has been shown.26

We have previously reported on spontaneously developingliver fibrosis in PDGF-B transgenic mice.15 Since HCC onlyrarely occurs in healthy liver, and the cancer risk increasesdramatically at the cirrhosis stage of which liver fibrosis is apreliminary stage, it seemed reasonable to postulate thatPDGF-B may also be involved in the development of livercancer. Since the development of spontaneous fibrosis inPDGF-B transgenic mice already took 6 months, it can bespeculated that spontaneous development of HCC could be along lasting process. Indeed, initial results of PDGF-B trans-genic animals at an age of 8 month showed no spontaneoustumor formation (data not shown).

Six months after HCC induction, not only the developmentof tumors in PDGF-B transgenic mice was significantlyincreased but also malignant transformation of these tumorswas significantly more frequent. As these malignant changes inPDGF-B transgenic livers occurred by a large margin of signifi-cance, this growth factor must be considered a major regulatorin HCC development. These findings are in accordance withthe earlier mentioned data, where overexpression of PDGF-Clead to enhanced HCC development and the role of PDGF-Bin other malignancies such as colon cancer, where the growthfactor has been shown to have tumor promoting function.27,28

Both, PDGF-B and PDGF-C transgenic mice showed spontane-

ous development of liver fibrosis. However, in contrast toPDGF-C transgenic mice PDGF-B transgenic mice did notshow spontaneous development of liver steatosis and HCC.14

Establishing an important role for PDGF-B in HCC devel-opment in vivo, the downstream mechanisms of PDGF-B sig-naling were of high interest and even more a linkage to char-acterized genetic pathways that are known to be involved inHCC development. Several lines of evidence support anessential role of the Wnt-/b-catenin signaling pathway inHCC.29 These include an increased expression and nuclearaccumulation of b-catenin as a feature of an activated Wnt-signaling pathway.30,31 Up to 62% of all HCC were shown todisplay such a dysregulation of b-catenin. In addition, a mul-tivariate analysis has demonstrated poorer prognosis andhigher rate of tumor recurrence in patients with nuclearaccumulation of b-catenin.31,32

Our findings of a massive dislocation of b-catenin fromthe plasma membrane to the cytoplasm and nucleus inPDGF-B transgenic mice point toward a novel pathomechan-ism in HCC development, connecting PDGF-B expression toWnt-/b-catenin signaling and ultimately disembogue into acommon end stage-signaling cascade. Dislocation of b-cateninfrom the plasma membrane to the nucleus has been describedin various studies.33,34 In addition, mutations of Axin-1, anegative regulator of the Wnt-signaling pathway, have alsobeen reported to be highly prevalent in human HCC. Trans-fection of wild-type Axin-1 leads to reconstitution of Wnt-sig-naling and apoptosis in cancer cells.33,34 At a lower frequency,Axin-2 mutations may contribute to HCC as well.34

Significant influence of TGF-b and TGF-b receptor signal-ing in HCC development has been demonstrated previ-ously.35 Current concepts on liver carcinogenesis throughTGF-b signaling postulate a decreased expression of TGF-bRII to be key to the evasion of tumor cells from a generally

Figure 5. Analysis of CD31 positive regions and VEGF-expression. PDGF-B transgenic animals demonstrate increased CD31-positive areas as

compared to WT groups (a). Relative expression of VEGF is upregulated in PDGF-B untreated animals (b). Bars ¼ 100 lm. [Color figure can

be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Carcinog

enesis

Maass et al. 1265

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC

growth limiting function of TGF-b.36 This hypothesis is sup-ported by in vitro and clinical data. In most expression stud-ies on HCC tissue, expression of TGF-b RII in liver tissueswas significantly decreased in patients with HCC in compari-son to patients with chronic hepatitis or liver cirrhosis. How-ever, there is also data where enhanced TGF-b RII expressionin HCC has been described.37 In light of these current con-cepts on TGF-b signaling and liver cancer, we are of theopinion that significantly higher expression of TGF-b andTGF-b RII is associated with the development of liver fibro-sis/cirrhosis, which is a prerequisite of liver cancer. It has tobe pointed out that, in contrast to expression studies onhuman HCC, we did not measure expression in clearly iden-tified HCC tissue due to the small size of HCC nodules.Rather mixed liver tissue was analyzed.

Another fundamental step in tumor growth is angiogene-sis, the growth of new blood vessels from pre-existing vessels.

Without angiogenesis tumors are not able to grow beyond asize of 1–2 mm3, due to a lack of nutrients and oxygen.Angiogenesis is induced by tumors in multiple ways. One ofthe best-characterized angiogenetic factors is VEGF, stimulat-ing endothelial cell proliferation and induces the formation ofnew blood vessels.38 Earlier studies have demonstrated thatVEGF expression and angiogenesis occur already in the pre-neoplastic lesions of dysplastic nodules and that the degree ofVEGF expression and angiogenesis increases according to theprogression of multistep hepatocarcinogenesis.39,40

PECAM-1/CD31 is a major constituent of the endothelialcell intercellular junction, where it is highly concentrated.41

Given its abundant expression in endothelial cells, PECAM-1/CD31 has been demonstrated to be involved in a variety ofendothelial functions among them the formation of newblood vessels in angiogenesis.42 Both VEGF and PECAM-1/CD31 were regulated in PDGF-B transgenic animals. A lackof significant VEGF upregulation in WT DEN/phenobarbital-treated mice may be due to the early tumor stages and onlyfew tumors in WT animals. These findings support an essen-tial role of PDGF-B in tumor angiogenesis. Thus, PDGF-Bmay not only be involved in the induction of HCC but alsopromote further growth of initial neoplastic lesions by pro-moting angiogenesis. This data is in accordance with previousreports from PDGF-B knock out mice. Lindahl et al. foundthat mouse embryos deficient in PDGF-B lack microvascularpericytes, which normally form part of the capillary wall, anddevelop numerous capillary micro-aneurysms that rupture atlate gestation.9 Endothelial cells of the sprouting capillaries inmutant mice appeared to be unable to attract PDGF-Rb-posi-tive pericyte progenitor cells. In both rat and rabbit ischemichind limb models, PDGF-B and FGF-2 together markedlystimulated collateral arteriogenesis after ligation of the

Figure 6. Analysis of TGF-b and its receptors. Untreated PDGF-B

transgenic mice show increased relative expression of TGF-b (a)

but no difference between the different experimental groups in the

amount of total TGF-b protein could be detected (b). No changes in

expression of TGF-b receptor type I could be observed (c) whereas

in PDGF-B transgenic animals the expression of TGF-b receptor type

II is significantly upregulated (d).

Figure 7. Analysis of FGF-2 mRNA expression. In comparison to

treated and untreated wild-type mice (WT�/WTþ), both treated

and untreated PDGF-B transgenic mice (PDGF-B�/PDGF-Bþ) show

an increased expression of FGF-2 in liver tissue.

Carcinog

enesis

1266 PDGF-B in liver cancer development

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC

femoral artery, with a significant increase in vascularizationand improvement in paw blood flow. A possible mechanismof angiogenic synergism between PDGF-B and FGF-2involves upregulation of the expression of PDGF-Ra andPDGF-Rb by FGF-2 in newly formed blood vessels.43 Indeed,PDGF-B transgenic mice showed an increased expression ofFGF-2 mRNA in the liver.

As demonstrated above, PDGF-B is involved in differentstages of liver cancer development. PDGF-B is an essentialregulator in the development of liver fibrosis, preparing theground for liver carcinogenesis. A possible mechanism could

be TGF-b receptor upregulation. In addition, PDGF-B leadsto an increased expression of b-catenin as well as VEGF,PECAM/CD31 and FGF, all factors with established rolesin carcinogenesis. Together, these factors may be assumedto induce HCC on the basis of an existing liver fibrosis/cir-rhosis. In this novel study, we have been able to establish akey regulatory function of PDGF-B in multiple aspects ofHCC development. The new aspects about the role ofPDGF-B in liver carcinogenesis could be the basis for de-velopment of new preventive as well as treatment strategiesin hepatocarcinogenesis.

References

1. Motola-Kuba D, Zamora-Valdes D, UribeM, Mendez-Sanchez N. Hepatocellularcarcinoma: an overview. Ann Hepatol 2006;5:16–24.

2. McGlynn KA, London WT. Epidemiologyand natural history of hepatocellularcarcinoma. Best Pract Res ClinGastroenterol 2005;19:3–23.

3. Teo EK, Fock KM. Hepatocellularcarcinoma: an Asian perspective. Dig Dis2001;19:263–8.

4. Clark JM. The epidemiology ofnonalcoholic fatty liver disease in adults. JClin Gastroenterol 2006;40 (Suppl 1):S5–10.

5. Schwartz M, Roayaie S, Konstadoulakis M.Strategies for the management ofhepatocellular carcinoma. Nat Clin PractOncol 2007;4:424–32.

6. Llovet JM, Ricci S, Mazzaferro V, HilgardP, Gane E, Blanc JF, de Oliveira AC,Santoro A, Raoul JL, Forner A, SchwartzM, Porta C, et al. Sorafenib in advancedhepatocellular carcinoma. N Engl J Med2008;359:378–90.

7. Teufel A, Staib F, Kanzler S, Weinmann A,Schulze-Bergkamen H, Galle PR. Geneticsof hepatocellular carcinoma. World JGastroenterol 2007;13:2271–82.

8. Li X, Eriksson U. Novel PDGF familymembers: PDGF-C and PDGF-D. CytokineGrowth Factor Rev 2003;14:91–8.

9. Lindahl P, Johansson BR, Leveen P,Betsholtz C. Pericyte loss andmicroaneurysm formation in PDGF-B-deficient mice. Science 1997;277:242–5.

10. Dirks RP, Bloemers HP. Signals controllingthe expression of PDGF. Mol Biol Rep1995;22:1–24.

11. Ding J, Feng Y, Wang HY. From cellsignaling to cancer therapy. ActaPharmacol Sin 2007;28:1494–8.

12. Heldin CH, Westermark B. Mechanism ofaction and in vivo role of platelet-derivedgrowth factor. Physiol Rev 1999;79:1283–316.

13. Alvarez RH, Kantarjian HM, Cortes JE.Biology of platelet-derived growth factorand its involvement in disease. Mayo ClinProc 2006;81:1241–57.

14. Campbell JS, Hughes SD, Gilbertson DG,Palmer TE, Holdren MS, Haran AC, OdellMM, Bauer RL, Ren HP, Haugen HS, YehMM, Fausto N. Platelet-derived growthfactor C induces liver fibrosis, steatosis,and hepatocellular carcinoma. Proc NatlAcad Sci USA 2005;102:3389–94.

15. Czochra P, Klopcic B, Meyer E, Herkel J,Garcia-Lazaro JF, Thieringer F,Schirmacher P, Biesterfeld S, Galle PR,Lohse AW, Kanzler S. Liver fibrosisinduced by hepatic overexpression ofPDGF-B in transgenic mice. J Hepatol2006;45:419–28.

16. Borkham-Kamphorst E, van Roeyen CR,Ostendorf T, Floege J, Gressner AM,Weiskirchen R. Pro-fibrogenic potential ofPDGF-D in liver fibrosis. J Hepatol 2007;46:1064–74.

17. Friedman SL. Mechanisms of hepaticfibrogenesis. Gastroenterology 2008;134:1655–69.

18. Pinzani M, Gesualdo L, Sabbah GM,Abboud HE. Effects of platelet-derivedgrowth factor and other polypeptidemitogens on DNA synthesis and growth ofcultured rat liver fat-storing cells. J ClinInvest 1989;84:1786–93.

19. Pinzani M, Milani S, Herbst H, DeFrancoR, Grappone C, Gentilini A, Caligiuri A,Pellegrini G, Ngo DV, Romanelli RG,Gentilini P. Expression of platelet-derivedgrowth factor and its receptors in normalhuman liver and during active hepaticfibrogenesis. Am J Pathol 1996;148:785–800.

20. Lou SM, Li YM, Wang KM, Cai WM,Weng HL. Expression of platelet-derivedgrowth factor-BB in liver tissues of patientswith chronic hepatitis B. World JGastroenterol 2004;10:385–8.

21. Hopfner M, Schuppan D, Scherubl H.Growth factor receptors and relatedsignalling pathways as targets for noveltreatment strategies of hepatocellularcancer. World J Gastroenterol 2008;14:1–14.

22. Stock P, Monga D, Tan X, Micsenyi A,Loizos N, Monga SP. Platelet-derived

growth factor receptor-alpha: a noveltherapeutic target in human hepatocellularcancer. Mol Cancer Ther 2007;6:1932–41.

23. Zhang T, Sun HC, Xu Y, Zhang KZ, WangL, Qin LX, Wu WZ, Liu YK, Ye SL, TangZY. Overexpression of platelet-derivedgrowth factor receptor alpha in endothelialcells of hepatocellular carcinoma associatedwith high metastatic potential. Clin CancerRes 2005;11:8557–63.

24. Thieringer F, Maass T, Czochra P, KlopcicB, Conrad I, Friebe D, Schirmacher P,Lohse AW, Blessing M, Galle PR, Teufel A,Kanzler S. Spontaneous hepatic fibrosis intransgenic mice overexpressing PDGF-A.Gene 2008;423:23–8.

25. Borkham-Kamphorst E, Kovalenko E, vanRoeyen CR, Gassler N, Bomble M,Ostendorf T, Floege J, Gressner AM,Weiskirchen R. Platelet-derived growthfactor isoform expression in carbontetrachloride-induced chronic liver injury.Lab Invest 2008;88:1090–100.

26. van Zijl F, Mair M, Csiszar A, Schneller D,Zulehner G, Huber H, Eferl R, Beug H,Dolznig H, Mikulits W. Hepatic tumor-stroma crosstalk guides epithelial tomesenchymal transition at the tumor edge.Oncogene 2009;28:4022–33.

27. Hsu S, Huang F, Friedman E. Platelet-derived growth factor-B increases coloncancer cell growth in vivo by a paracrineeffect. J Cell Physiol 1995;165:239–45.

28. Pietras K, Ostman A, Sjoquist M,Buchdunger E, Reed RK, Heldin CH,Rubin K. Inhibition of platelet-derivedgrowth factor receptors reduces interstitialhypertension and increases transcapillarytransport in tumors. Cancer Res 2001;61:2929–34.

29. Luu HH, Zhang R, Haydon RC, RayburnE, Kang Q, Si W, Park JK, Wang H, PengY, Jiang W, He TC. Wnt/beta-cateninsignaling pathway as a novel cancer drugtarget. Curr Cancer Drug Targets 2004;4:653–71.

30. Willert K, Jones KA. Wnt signaling: is theparty in the nucleus? Genes Dev 2006;20:1394–404.

Carcinog

enesis

Maass et al. 1267

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC

31. Wong CM, Fan ST, Ng IO. beta-Cateninmutation and overexpression inhepatocellular carcinoma: clinicopathologicand prognostic significance. Cancer 2001;92:136–45.

32. Inagawa S, Itabashi M, Adachi S,Kawamoto T, Hori M, Shimazaki J,Yoshimi F, Fukao K. Expression andprognostic roles of beta-catenin inhepatocellular carcinoma: correlation withtumor progression and postoperativesurvival. Clin Cancer Res 2002;8:450–6.

33. Satoh S, Daigo Y, Furukawa Y, Kato T,Miwa N, Nishiwaki T, Kawasoe T, IshiguroH, Fujita M, Tokino T, Sasaki Y, Imaoka S,et al. AXIN1 mutations in hepatocellularcarcinomas, and growth suppression incancer cells by virus-mediated transfer ofAXIN1. Nat Genet 2000;24:245–50.

34. Taniguchi K, Roberts LR, Aderca IN, DongX, Qian C, Murphy LM, Nagorney DM,Burgart LJ, Roche PC, Smith DI, Ross JA,Liu W. Mutational spectrum of beta-catenin, AXIN1, and AXIN2 in

hepatocellular carcinomas andhepatoblastomas. Oncogene 2002;21:4863–71.

35. Tannapfel A, Wittekind C. Genes involvedin hepatocellular carcinoma: deregulationin cell cycling and apoptosis. VirchowsArch 2002;440:345–52.

36. Paik SY, Park YN, Kim H, Park C.Expression of transforming growth factor-beta1 and transforming growth factor-betareceptors in hepatocellular carcinoma anddysplastic nodules. Mod Pathol 2003;16:86–96.

37. Torbenson M, Marinopoulos S, Dang DT,Choti M, Ashfaq R, Maitra A, Boitnott J,Wilentz RE. Smad4 overexpression inhepatocellular carcinoma is stronglyassociated with transforming growth factorbeta II receptor immunolabeling. HumPathol 2002;33:871–6.

38. Senger DR, van de Water L, Brown LF,Nagy JA, Yeo KT, Yeo TK, Berse B,Jackman RW, Dvorak AM, Dvorak HF.Vascular permeability factor (VPF, VEGF)

in tumor biology. Cancer Metastasis Rev1993;12:303–24.

39. Park YN, Kim YB, Yang KM, Park C.Increased expression of vascularendothelial growth factor and angiogenesisin the early stage of multistephepatocarcinogenesis. Arch Pathol Lab Med2000;124:1061–5.

40. Yamaguchi R, Yano H, Iemura A,Ogasawara S, Haramaki M, Kojiro M.Expression of vascular endothelial growthfactor in human hepatocellular carcinoma.Hepatology 1998;28:68–77.

41. Newman PJ. The role of PECAM-1 invascular cell biology. Ann NY Acad Sci1994;714:165–74.

42. Jackson DE. The unfolding tale ofPECAM-1. FEBS Lett 2003;540:7–14.

43. Cao R, Brakenhielm E, Pawliuk R, WariaroD, Post MJ, Wahlberg E, Leboulch P, CaoY. Angiogenic synergism, vascular stabilityand improvement of hind-limb ischemiaby a combination of PDGF-BB and FGF-2.Nat Med 2003;9:604–13.

Carcinog

enesis

1268 PDGF-B in liver cancer development

Int. J. Cancer: 128, 1259–1268 (2011) VC 2010 UICC