Human and Mouse VEGFA -Amplified Hepatocellular Carcinomas Are Highly Sensitive to Sorafenib Treatment

Published OnlineFirst March 31, 2014; DOI: 10.1158/2159-8290.CD-13-0782

RESEARCH ARTICLE

Human and Mouse VEGFA -Ampliﬁ ed Hepatocellular Carcinomas Are Highly Sensitive to Sorafenib Treatment

Elad Horwitz 1 , Ilan Stein 1 , 3, Mariacarla Andreozzi 5 , Julia Nemeth 6, Avivit Shoham 1 , Orit Pappo 3, Nora Schweitzer 12 , Luigi Tornillo 5 , Naama Kanarek 1 , Luca Quagliata 5 , Farid Zreik 3 , Rinnat M. Porat 3, Rutie Finkelstein3 , Hendrik Reuter 7 , Ronald Koschny 11 , Tom Ganten 11 , Carolin Mogler 9 , Oren Shibolet 4 , Jochen Hess 6 , 8 , 10, Kai Breuhahn 9 , Myriam Grunewald 2 , Peter Schirmacher 9 , Arndt Vogel12 , Luigi Terracciano 5 , Peter Angel 6, Yinon Ben-Neriah 1 , and Eli Pikarsky 1 , 3

ABSTRACT Death rates from hepatocellular carcinoma (HCC) are steadily increasing, yet therapeutic options for advanced HCC are limited. We identify a subset of mouse and human HCCs harboring VEGFA genomic amplifi cation, displaying distinct biologic characteristics. Unlike common tumor amplifi cations, this one seems to work via heterotypic paracrine interactions; stromal VEGF receptors (VEGFR), responding to tumor VEGF-A, produce hepatocyte growth factor (HGF) that reciprocally affects tumor cells. VEGF-A inhibition results in HGF downregulation and reduced proliferation, specifi cally in amplicon-positive mouse HCCs. Sorafenib—the fi rst-line drug in advanced HCC—targets multiple kinases, including VEGFRs, but has only an overall mild benefi cial effect. We found that VEGFA amplifi cation specifi es mouse and human HCCs that are distinctly sensitive to sorafenib. FISH analysis of a retrospective patient cohort showed markedly improved survival of sorafenib-treated patients with VEGFA -amplifi ed HCCs, suggesting that VEGFA amplifi cation is a potential biomarker for HCC response to VEGF-A–blocking drugs.

SIGNIFICANCE: Using a mouse model of infl ammation-driven cancer, we identifi ed a subclass of HCC carrying VEGFA amplifi cation, which is particularly sensitive to VEGF-A inhibition. We found that a similar amplifi cation in human HCC identifi es patients who favorably responded to sorafenib—the fi rst-line treatment of advanced HCC—which has an overall moderate therapeutic effi cacy. Cancer Discov; 4(6); 730–43. ©2014 AACR.

See related commentary by Luo and Feng, p. 640.

Authors’ Afﬁ liations: 1 The Lautenberg Center for Immunology; 2 Depart- Note: Supplementary data for this article are available at Cancer Discovery ment of Developmental Biology and Cancer Research, IMRIC, Hadassah Online (http://cancerdiscovery.aacrjournals.org/). 3 Medical School, Hebrew University; Department of Pathology, Hadassah- Corresponding Authors: Eli Pikarsky, The Lautenberg Center for Immunol- 4 Hebrew University Medical Center, Jerusalem; Liver Unit, Tel Aviv Sour- ogy, IMRIC, Hadassah Medical School, Hebrew University, Kiryat Hadas- 5 asky Medical Center, Tel Aviv, Israel; Institute of Pathology, University sah, P.O.B. 12272, Jerusalem 91120, Israel. Phone: 972-2-675-8202; Fax: 6 Hospital Basel, Basel, Switzerland; Division of Signal Transduction and 972-2-642-6268; E-mail: [email protected] ; and Yinon Ben-Neriah, 7 8 Growth Control (A100), Division of Molecular Genetics (B060), and Jun- The Lautenberg Center for Immunology, IMRIC, Hadassah Medical School, ior Group Molecular Mechanisms of Head and Neck Tumors (A102), Ger- Hebrew University, Kiryat Hadassah, P.O.B. 12272, Jerusalem 91120, 9 man Cancer Research Center (DKFZ), DKFZ-ZMBH Alliance; Institute of Israel. Phone: 972-2-675-8718; E-mail: [email protected] Pathology, University Hospital Heidelberg; Departments of 10Otolaryn gol- doi: 10.1158/2159-8290.CD-13-0782 ogy, Head and Neck Surgery and 11Internal Medicine, University Hospital Heidelberg, Heidelberg; and 12 Department of Gastroenterology, Hepatology ©2014 American Association for Cancer Research. and Endocrinology, Hannover Medical School, Hannover, Germany

730 | CANCER DISCOVERYJUNE 2014 www.aacrjournals.org

INTRODUCTION HCC development, we applied array-based comparative genomic hybridization (aCGH) to 10 HCCs obtained from Hepatocellular carcinoma (HCC) is the third leading cause of − − 16-month-old Mdr2 / mice (Supplementary Fig. S1A). We cancer-related mortality worldwide, with the highest increase detected several amplifi cations and deletions, including an rate in North America (1, 2). Sorafenib is the mainstay of amplifi cation in the qB3 band of murine chromosome 17 therapy for advanced HCC and the only systemic drug that has (Chr17qB3; Supplementary Table S1), encoding among oth- shown any survival advantage in HCC so far (3, 4). However, ers the gene for VEGF-A. Genomic amplifi cation of Vegfa was patients’ response is modest, and sorafenib treatment is asso- of interest as it is a cytokine gene that can modulate several ciated with side effects (3–8 ). Thus, several studies looked for components of the tumor microenvironment and induce liver predictive markers for sorafenib response (9–11 ), yet no such cell growth (24, 25). To determine the frequency of this ampli- biomarkers have entered the clinical setting. Predictive biomar- − − fi cation, we tested a larger cohort of Mdr2 / tumors by quan- kers, identifying patient subsets to guide treatment choices, are titative PCR (qPCR) of tumor DNA (Supplementary Fig. S1B) usually based on distinct pathogenetic mechanisms, and are and chromogenic in situ hybridization (CISH; Fig. 1A ); 13 of the cornerstone of personalized medicine (12, 13). Prominent 93 (∼14%) HCCs harbored this amplicon. To map the minimal examples include ERBB2 amplifi cation and KRAS mutations amplifi ed region of this amplicon, we used DNA qPCR directed that serve as key determinants of treatment with trastuzumab at several loci along murine chromosome 17 in a cohort of or cetuximab, respectively (14, 15). tumors bearing this amplifi cation ( Fig. 1B ). We found that Sorafenib is a multikinase inhibitor, the targets of which the common proximal border lies between 43.3 and 48.5 mega include BRAF, CRAF, PDGFR2, c-KIT, and the VEGFA recep- base pairs (Mbp) from the chromosome 17 start. This minimal tors (VEGFR; refs. 10 , 16 ). Although VEGFRs were postulated amplifi ed region spanned 53 genes (Supplementary Table S2). as mediators of sorafenib response in HCC, testing for VEGF-A Amplifi cation of human Chr6p21 (the region syntenic for serum levels was not found to be predictive of sorafenib murine Chr17qB3) and whole chromosome gains were previ- treatment success ( 10 ). Moreover, bevacizumab, an antibody ously reported in several whole-genome analyses of human against VEGF-A, only shows minimal responses in HCC HCC with a frequency ranging between 7% and 40% (29–33 ). (17–20 ). A possible explanation for this is that the mecha- Accordingly, through FISH, we found VEGFA amplifi cations nism of action of sorafenib is predominantly independent of and chromosome 6 polysomies in 11% of human HCCs we VEGF-A inhibition. Alternatively , a small subset of patients tested (21 of 187; Fig. 1C ; Supplementary Table S3). who can respond to VEGF-A blockers may exist that was To elucidate the tumor relevance of these chromosomal underrepresented in bevacizumab studies. gains, we analyzed the expression of mouse Chr17qB3 ampli- VEGF-A is a master regulator of angiogenesis whose role con genes in amplicon-positive (herein Amppos ) and amplicon- in tumor vessel recruitment is well established ( 21 ). However, negative (Ampneg ) tumors. Matched increases in mRNA and additional roles for VEGF-A in tumorigenesis are emerging. DNA levels were found for Vegfa, Tjap1 , and Xpo5 (Fig. 1D VEGF-A was shown to promote tumor cell growth in an and Supplementary Fig. S2A). We further found a correlation autocrine manner, in skin and lung cancer cells expressing between Chr17qB3 amplifi cation and VEGF-A protein levels VEGFRs (22, 23). Moreover, VEGF-A also has nonangiogenic in both tumor extracts and serum ( Fig. 1E and Supplemen- functions in normal physiology. In liver tissue, VEGF-A elicits tary Fig. S2B). Furthermore, double immunostaining for hepatocyte proliferation by elevating expression of hepato- VEGF-A and E-cadherin in Amppos tumors showed that the cyte mitogens in liver sinusoidal endothelial cells ( 24, 25 ). tumor cells are indeed the main origin of VEGF-A in these HCC is most commonly the outcome of chronic injury and tumors (Fig. 1F). Thus, amplifi cation of murine Chr17qB3 is infl ammation, resulting in hepatocyte regeneration and dys- a recurrent event in HCC associated with an elevated expres- regulated growth factor signaling (1 , 26 ). It has become clear sion of several resident genes, including Vegfa . that infl ammatory signaling pathways can support survival, growth, and progression of cancer. Accordingly, secreted Amp pos Tumors Display Distinct cytokines and effector molecules, which are abundant in the Histologic Features tumor microenvironment, could constitute suitable targets for treatment and primary prevention of HCC (26 ). Here, we Tumor cell proliferation is a strong and consistent marker used Mdr2 -defi cient mice (Mdr2 −/− ), which develop chronic of poor prognosis in human HCC ( 34 ). Bromodeoxyuridine liver infl ammation, eventually leading to infl ammation- (BrdUrd ) immunostaining revealed that Amppos mouse HCCs induced liver tumors similar to human HCC (27, 28), to look displayed a 2-fold higher proliferation index compared with for candidate treatment targets modulating the tumor micro- Amp neg tumors ( Fig. 2A and B , top). No differences were environment that are relevant to human HCC. noted in apoptosis (Supplementary Fig. S2C) or in the presence of neutrophils or fi broblasts (data not shown). Other RESULTS histologic features, differing between Amppos and Amp neg mouse HCCs, include a 6-fold higher vessel density (Fig. Array CGH Reveals Recurrent Gains in the VEGFA 2A and B , middle) and 4-fold higher macrophage content Locus Identifying a Molecularly Distinct Tumor (Fig. 2A and B, bottom). Moreover, several signature genes Subpopulation of tumor-associated macrophages (TAM), including those In search of microenvironment-affecting factors whose encoding Arginase 1, TGFβ, and YM1 ( 35, 36 ), were elevated amplifi cation or deletion plays a role in infl ammation-induced in the Amppos group (Fig. 2C), whereas markers of classically

JUNE 2014CANCER DISCOVERY | 731

RESEARCH ARTICLE Horwitz et al.

A B pos Murine HCCs Murine Amp HCCs Nonamplified 43.3 Tumor #1 Tumor #2 Amplified 43.8 Minimal amplified region neg 44.4 Cdc5L 44.36–44.41 Amp Aars2 45.64–45.66 Tumor #3 Tumor #4 45.0 Nfkbie 45.69–45.70 Hsp90ab1 45.71–45.71 Capn11 45.77–45.79 Tmem63b 45.78–45.82

pos 45.7 45.9 Mrpl14 45.82–45.84 Vegfa 46.15–46.17

Amp 46.16 Mrps18a 46.24–46.26 46.4 Xpo5 46.33–46.37 Tjap1 46.39–46.41 46.9 Egfl9 46.43–46.44 Human HCCs Parc 46.63–46.68 Gnmt 46.86–46.87

C Mbp from chromosome 17 start 47.4

47.9

48.5

Normal Polysomic Amplified

D E 7 Cdc5L Hsp90ab1 Vegfa Vegfa mRNA (qPCR) 6 5 10 VEGF-A protein (ELISA) ** n.s. *** 6 4 8 4 3 6 5 2 4 2 1 2 µ g protein (ELISA) 4 Relative expression Relative expression Relative expression 0 0 0 pos pos pos WT Ampneg Amp WT Ampneg Amp WT Ampneg Amp 3 liver tumor tumor liver tumor tumor liver tumor tumor

Cul9 2 6 Tjap1 8 Xpo5 40

*** 6 *** 30 *** 1 4

4 20 Fold induction (qPCR) pg/ 0 2 2 10 1212341234 Relative expression Relative expression Relative expression 0 0 0 WT Ampneg Amppos pos pos pos WT Ampneg Amp WT Ampneg Amp WT Ampneg Amp liver –/– liver tumor tumor liver tumor tumor liver tumor tumor Mdr2 tumors

F Hoechst E-cadherin VEGF-A Merge tumor pos Amp

Figure 1. A recurrent gain in the VEGFA locus identifying a molecularly distinct tumor subpopulation. A, representative photomicrographs of CISH using probes specifi c for the murine Chr17qB3. B, DNA qPCR analysis using primers specifi c for different loci on the qB3 arm of chromosome 17. Each vertical line represents a single Amppos tumor. The thin line represents nonamplifi ed regions, and the thick line represents amplifi ed regions. The list includes several of the residing genes (a full list is available as Supplementary Data). C, representative photomicrographs of FISH of human HCC using Chr6p12 probe (red) and chromosome 6 centromere probe (green). D, qPCR analysis of the mRNA levels of murine genes encoded on the amplifi ed region. Each dot represents a different tumor. The cross line signifi es geometric mean (n.s., not signifi cant; **, P < 0.01; ***, P < 0.0001). E, qPCR and ELISA analyses of VEGF-A performed on extracts of wild-type (WT) livers, Ampneg , and Amp pos Mdr2 −/− tumors in matching pairs show a correlation between the increase in mRNA and protein levels. F, confocal microscopy images of an Mdr2 − /− Amp pos tumor immunostained for E-cadherin and VEGF-A. Hoechst 33342 marks nuclei. Scale bar, 20 μm.

732 | CANCER DISCOVERYJUNE 2014 www.aacrjournals.org

VEGFA Ampliﬁ cation in HCC RESEARCH ARTICLE

A Ampneg tumor Amppos tumor B BrdUrd 20 WT liver ** neg 15 Amp tumor Amppos tumor 10 BrdUrd 5

Percentage positive nuclei 0 vWF vWF 14 * 12 10 8

vWF 6 4 2

Percentage positive area 0

F4/80 14 12 * 10 8

F4/80 6 4 2

Percentage positive area 0

C Arg1 Tgfb Ym1 1.5 840 ** ** 6 *** 30 1.0 4 20 0.5 markers 2 10 Protumorigenic Relative expression Relative Relative expression Relative 0.0 0 expression Relative 0 neg pos pos pos WT Amp Amp WT Ampneg Amp WT Ampneg Amp liver tumor tumor liver tumor tumor liver tumor tumor

Tnfa Nos2 Cxcl10 600 n.s. 60 50 400 n.s. n.s. 200 40 15 40 30 10 20 20 markers 5 10 Antitumorigenic Relative expression Relative expression Relative Relative expression Relative 0 0 0 pos pos pos WT Ampneg Amp WT Ampneg Amp WT Ampneg Amp liver tumor tumor liver tumor tumor liver tumor tumor

D Ampneg tumor Amppos tumor E MRC1 15 **

MRC1 5

Percentage positive area positive Percentage 0 Ampneg Amppos

Figure 2. Amp pos tumors are a distinct tumor subpopulation. A, representative IHC photomicrographs for BrdUrd, vWF, and F4/80. Scale bar, 50 μm (BrdUrd) and 100 μm (vWF and F4/80). B, IHCs were quantifi ed using automated image analysis (n ≥ 7; *, P < 0.01; **, P < 0.05). C, qPCR analysis of tumor-associated (protumorigenic) and classically activated (antitumorigenic) macrophage markers. Each dot represents a different tumor. The cross line signifi es geometric mean (n.s., not signifi cant; **, P < 0.01; ***, P < 0.001). D, immunofl uorescent stain for MRC1 on Ampneg and Amppos tumors. Scale bar, 40 μm. E, quantifi cation of MRC1 immunofl uorescence. Bars represent geometric mean; **, P < 0.01.

JUNE 2014CANCER DISCOVERY | 733

RESEARCH ARTICLE Horwitz et al. activated macrophages, including TNFα, inducible nitric tion in vitro, we treated peritoneal macrophages with recom- oxide synthase (iNOS), CXCL10, and IL12p40, were similar binant VEGF-A and detected an increase in Hgf mRNA levels in both tumor groups (Fig. 2C and Supplementary Fig. S3A). ( Fig. 3H ). This raises the possibility that VEGF-A in Amppos Along with these, immunofl uorescent staining for the protu- tumors does not provide autocrine signals to hepatocytes, morigenic macrophage marker MRC1 revealed a higher pres- but rather acts through manipulation of the microenviron- ence of MRC1-expressing cells in Amppos tumors (Fig. 2D and ment to induce HGF secretion. E ). Of note, VEGF-A was shown to act as a chemoattractant for naïve myeloid cells, which facilitate the generation of new VEGF-A Increases Cellular Proliferation in HCC blood vessels ( 37 ). Together, these data signify Amppos tumors To prove the functional role of VEGFA in this genomic ampli- as a distinct subgroup of HCCs characterized by enhanced fi cation, we set out to inhibit VEGF-A in Mdr2 −/− tumors. To presence of specifi c microenvironmental components and this end, we injected intravenously adenoviral vectors encoding increased proliferation rate. GFP alone or GFP and the soluble VEGFR1 (sFLT), a potent Histologic analysis of human HCCs revealed several char- inhibitor of VEGF-A (38 ), into 56 tumor-harboring Mdr2 − /− acteristic traits of HCCs harboring VEGFA gains. A higher mice of ages 14 to 18 months and sacrifi ced them 10 days fol- incidence of vascular invasion was found in the Amppos lowing injection (Supplementary Fig. S5A). Vegfa amplifi cation group (9 of 20 Amp pos tumors vs. 1 of 20 Ampneg tumors; status was determined after sacrifi ce by both DNA qPCR and P < 0.01; Supplementary Table S4). A lower incidence of fi brosis Vegfa mRNA expression ( Fig. 4C and data not shown). Liver within tumor tissue was found in the Amppos group as well. Few damage, measured through plasma aspartate aminotransferase additional traits nearly reaching statistical signifi cance were (AST) activity, was similar in all groups (Supplementary Fig. identifi ed (Supplementary Table S4). We found no differences S5B). Immunostaining for BrdUrd, Ki67, and phosphorylated in several clinical characteristics of HCC including underlying histone H3 (pHH3) revealed that blocking VEGF-A markedly disease, gender, or tumor size (Supplementary Fig. S4A–S4C). inhibited tumor cell proliferation in Amppos tumors, but not Altogether, these results indicate that in both murine Mdr2− /− in Ampneg ones (Fig. 4A and B and Supplementary Fig. S5C and human HCCs, tumors that harbor genomic gains in the and S5D). This decrease in proliferation was accompanied by VEGFA locus are distinct from those that do not. reduced Hgf mRNA levels ( Fig. 4C ). Treatment with adenovirus encoding GFP alone did not induce any change in either Macrophage–Tumor Cell Cross-talk group. Macrophage infi ltration and protumorigenic macro- pos in Amp Tumors phage expression profi le did not decrease following the sFLT Previous studies have delineated hepatocyte–endothelial treatment (Supplementary Fig. S5C and S5D and data not cross-talk taking place in non-neoplastic liver, wherein VEGF-A shown). Macroscopic and histologic analyses revealed multiple stimulates endothelial cells to secrete several mitogens includ- foci of coagulative necrosis only in sFLT-treated Amppos tumors ing hepatocyte growth factor ( HGF; refs. 24, 25 ). We hypothe- (3 of 6 vs. 0 of 10 in Ampneg ; P < 0.05; Fig. 4D ). In these specifi c sized that Amppos HCCs exploit this interaction for promoting tumors, we also found an elevation of the hypoxia-inducible tumor cell proliferation. Following this notion, we detected a factor-1α (HIF1α) target genes Glut1 and Pgk1 (Fig. 4C), indica- 3-fold elevation of Hgf mRNA levels in Amppos versus Ampneg tive of tissue hypoxia. Immunostaining for vWF revealed a mouse tumors ( Fig. 3A ). We did not fi nd signifi cant changes trend of decrease in vasculature only in VEGF-A–blocked in other angiocrine-produced molecules ( 24, 25 )—Wnt2, IL6, Amppos tumors, particularly in the hypoxic tumors (Supple- and heparin binding EGF-like growth factor (HB-EGF; data mentary Fig. S5C and S5D). These data denote Amppos tumors not shown). Immunostaining detected HGF expression only as hypersensitive to direct inhibition of VEGF-A. in Amppos tumors, exclusively in the non-neoplastic stromal cells ( Fig. 3B ). Immunofl uorescent staining for von Willebrand Overexpression of VEGF-A in HCC Cells Increases Factor (vWF), F4/80, and HGF suggested that macrophages are Proliferation Only In Vivo the major cell type expressing HGF ( Fig. 3C ). To further substantiate the tumor–stroma relationship with To understand the VEGF-A–HGF relationship, we isolated respect to VEGF amplifi cation, in particular hepatocyte VEGF- hepatocytes and macrophages from Mdr2 −/− livers at the age A eliciting a macrophage HGF–heterotypic circuit in tumor of 8 months ( Fig. 3D and Supplementary Fig. S3B), a time growth, we injected human Hep3B HCC cells transduced with point signifi ed by marked dysplasia, yet no overt HCC for- a lentivector overexpressing human VEGF-A into immuno- mation (27 ). mRNA profi ling of these fractions showed that defi cient mice (Supplementary Fig. S6A). The in vivo growth the genes encoding VEGFRs (FLT1 and KDR) and corecep- rate of VEGF-A–overexpressing cells was higher than control tors [Neuropilin (Nrp)1 and 2] were higher in macrophages vector–transduced cells (Fig. 5A and Supplementary Fig. S6B). whereas the HGF receptor (c-MET) was more abundant in This correlated with higher proliferation rate and increased hepatocytes (Fig. 3E). This aligns with previous work showing HGF expression ( Fig. 5B–D and Supplementary Fig. S6C and that hepatocytes are inert to direct activation with VEGF-A S6D), phenocopying the Amppos Mdr2 −/− HCCs. Supporting (24 ). Immunostaining for KDR in Amp pos tumors demon- the non–cell-autonomous role, VEGF-A overexpression did strated that its expression in these tumors was restricted to not affect the in vitro growth rate of Hep3B cells ( Fig. 5E ). We non-neoplastic cells (Fig. 3F). This correlated with a modest also found no difference in expression of the hypoxia markers increase in mRNA levels of both Kdr and Flt1 in total tumor prolyl hydroxylase domain 3 (PHD3) and lactate dehydroge- lysates, which was comparable with the increase in mRNA nase A (LDHA) in these xenografts, lending further support to levels of recruited macrophage and endothelial markers Msr1 a nonangiogenic role for VEGF-A in HCC (Supplementary Fig. and Cd105 , respectively ( Fig. 3G ). Recapitulating this interac- S6E). Immunofl uorescence confi rmed VEGF-A expression in

734 | CANCER DISCOVERYJUNE 2014 www.aacrjournals.org

VEGFA Ampliﬁ cation in HCC RESEARCH ARTICLE

A B Ampneg tumor Amppos tumor Hgf Figure 3. A macrophage- 6 tumor cell cross-talk within ** Amppos tumors. A, qPCR 4 analysis of Hgf mRNA in WT livers, Ampneg , and Amppos 2

HGF tumors. The cross line signi-

Relative expression 0 ﬁ es geometric mean (**, pos WT Ampneg Amp P < 0.001). B, representative liver tumor tumor photomicrographs of IHC for HGF on Amp neg and Amp pos tumors. Scale bar, 100 μm. C vWF F4/80 HGF Merge C, representative confocal microscopy images of Amp pos tumor immunostained for vWF, F4/80, and HGF. Hoechst

tumor 33342 marks nuclei. Scale bar, 40 μm. D, representative pos cytospin preparations of

Amp macrophage and hepatocyte fractions from Mdr2− /− livers. E, qPCR analysis of hepatocyte and macrophage D fractions ( n = 11 different mice), isolated from livers of Mdr2 −/− mice. Bars represent geometric mean (**, P < 0.001;

H&E ***, P < 0.0001). F, representative photomicrograph of IHC Macrophages Hepatocytes for the VEGFR KDR in Ampneg and Amppos tumors. Note that the expression of KDR is Ampneg Amppos E 18 F conﬁ ned to endothelial and ** Hepatocytes stromal cells. Scale bar, 100 μm. 16 ** Macrophages G, mRNA qPCR analysis 14 for the VEGFRs Kdr and Flt1 12 and the macrophage and KDR 10 endothelium markers Msr1 and Cd105 on tissue lysates from 8 ** the indicated groups. Each dot 6 ** represents a different tumor.

Relative expression 4 *** The cross line represents geometric mean (***, P < 0.0001). 2 H, peritoneal macrophages 0 were cultured under serum-free Flt1 Kdr Nrp1 Nrp2 Met conditions followed by 8 hours G H of exposure to recombinant Kdr Flt1 Hgf murine VEGF-A (100 ng/mL). 35 2.5 Hgf expression was measured *** 4 *** * by qPCR. Bars represent 2 3 2.0 geometric mean; n ≥ 3; < 2 *, P 0.05. 1 1.5 1 1.0

Relative expression 0 Relative expression 0 neg pos neg pos WT Amp Amp WT Amp Amp 0.5

liver tumor tumor liver tumor tumor Relative expression 0.0 Msr1 Cd105 NT 8h rVEGF-A 10 3 Peritoneal 8 *** *** 2 macrophages 6 4 1 2

Relative expression 0 Relative expression 0 pos pos WT Ampneg Amp WT Ampneg Amp liver tumor tumor liver tumor tumor

transduced tumor cells and revealed HGF expression in mac- successful separation of cell populations (Supplementary rophages (Supplementary Fig. S7A). Fig. S7B). Although VEGF-A and control tumor groups Next, we generated single-cell suspensions from VEGF-A– did not show signifi cant differences in the TAM genes Tgf b , overexpressing or control Hep3B ZsGreen-labeled xenografts Ym1, Tnf a , and Nos2 (Supplementary Fig. S7C), we found and isolated macrophages (ZsGreen − CD45 + F4/80 + ) and endo- a 2-fold increase in Hgf expression in macrophages, but thelial cells (ZsGreen− CD45 −Meca32 +) using FACS sorting. not endothelial cells, isolated from VEGF-A–overexpress- qPCR analysis of the macrophage-specifi c and endothelium- ing tumors ( Fig. 5F ). Thus, VEGF-A overexpression by HCC specifi c genes, Msr1 and Cd105 , respectively, confi rmed cells is suffi cient to induce upregulation of HGF, mostly in

JUNE 2014CANCER DISCOVERY | 735

RESEARCH ARTICLE Horwitz et al.

A B neg neg pos pos Amp GFP Amp sFLT Amp GFP Amp sFLT BrdUrd 5 Ampneg GFP neg * 4 Amp sFLT Amppos GFP 3 Amppos sFLT 2 BrdUrd 1

Percentage positive nuclei 0

pos Vegfa Hgf Amp sFLT C 8 5 D

4 *** 6 3 4 2 2 1 Relative expression Relative expression

0 0 H&E neg neg pos Amp Amp Amppos Amppos Ampneg Ampneg Amppos Amp GFP sFLT GFP sFLT GFP sFLT GFP sFLT

Glut1 Pgk1 30 8

6 20 4 10 2 Relative expression Relative expression 0 0 pos pos pos pos Ampneg Ampneg Amp Amp Ampneg Ampneg Amp Amp GFP sFLT GFP sFLT GFP sFLT GFP sFLT

Figure 4. VEGF-A inhibition impedes proliferation in Amppos tumors. Mdr2 −/− mice were treated with adenovectors expressing either GFP alone or GFP and sFLT for 10 days. A, representative photomicrographs of IHC for BrdUrd. Tumor-infi ltrating cells remain proliferative. Scale bar, 100 μm. B, BrdUrd immunostaining was quantifi ed using automated image analysis (n ≥ 6; *, P < 0.05). C, mRNA qPCR analysis of the indicated genes in Ampneg and Amp pos tumors treated with the indicated adenovectors. Each dot represents a different tumor. The cross line signifi es geometric mean (***, P < 0.0001). D, left, histologic section stained with H&E depicting necrosis, representing three out of the six sFLT-treated Amppos tumors. Scale bar, 500 μm. Right, a macroscopic picture of a tumor with hemorrhagic necrosis. Scale bar, 0.5 cm.

A B C Control 2,000 VEGF-A 1,800 Control Control VEGF-A

) 35 3 1,600 * VEGF-A 30 1,400 25 1,200 20 1,000 15 800 ** 600 * BrdUrd 10 400 ** 5 Tumor volume (mm volume Tumor 200 * 0 Percentage positive nuclei positive Percentage 0 BrdUrd 01020 30 Days D E F Hgf 6 1.9 2.5 ** Control * 5 VEGF-A 1.7 2.0 1.5 4 * 1.3 1.5 3 1.1 Control 1.0 0.9 2 VEGF-A 0.7 0.5 Relative expression Relative 1 Absorbance (450 nm) ND ND Relative expression Relative 0.5 0.0 01234 56 Mac. End. Mac. End. 0 Days in culture hVEGFA mHgf Control VEGF-A tumor tumor

Figure 5. VEGF-A overexpression enhances tumor cell proliferation. Immunodefi cient mice were subcutaneously injected with Hep3B cells transduced with control vector or with a human VEGF-A vector. A, growth curve of xenografts transduced with control vector (black line) or with VEGF-A lentivector (gray line). Tumor volumes were measured topically with a caliper (*, P < 0.05; **, P < 0.01). B, representative photomicrographs of IHC for BrdUrd in control or VEGF-A lentivector-transduced xenografts. Scale bar, 100 μm. C, BrdUrd immunostaining was quantifi ed using automated image analysis (n ≥ 3; *, P < 0.05; a representative experiment of two performed is shown). D, mRNA qPCR analysis in control or VEGF-A lentivector-transduced xenografts. Bars represent geometric mean (n ≥ 4; *, P < 0.05; **, P < 0.01). E, XTT in vitro proliferation assay of the indicated lentivector-transduced cultured cells. F, expression of Hgf in macrophage and endothelial fractions from tumors overexpressing VEGF-A and controls determined by qPCR. Bars represent geometric mean; n ≥ 3; *, P < 0.05; ND, not detected—amplifi cation did not occur in any of the sample’s wells.

736 | CANCER DISCOVERYJUNE 2014 www.aacrjournals.org

VEGFA Ampliﬁ cation in HCC RESEARCH ARTICLE

50 * n.s. 40 * n.s. 30 n.s. 20 n.s. n.s. n.s. n.s. 10 n.s. 5 n.s. n.s. 4 3 n.s.

Relative to normal liver 2 *** ** 1 0 Ampneg Amppos Ampneg Amppos Ampneg Amppos Ampneg Amppos Ampneg Amppos Angpt1 Angpt2 Angptl2 Fgf1 Fgf2

n.s. n.s. 80 * n.s. 70 n.s. 60 * * 50 n.s. 40 n.s. 30 * 20 ** ** 10 5 4 n.s. 3 n.s. n.s. 2 Relative to normal liver 1 0 Ampneg Amppos Ampneg Amppos Ampneg Amppos Ampneg Amppos Ampneg Amppos Pdgfa Pdgfb Pdgfc Pgf Vegfb

Figure 6. Angiogenic factors elevated in murine Mdr2 − /− HCC. mRNA qPCR analysis of Mdr2 −/− tumors for the indicated angiogenesis regulators. Values shown are fold over normal livers’ average. Each dot represents a different tumor. The cross line signifi es geometric mean, signifi cance marks immediately above each group refer to comparison with normal livers. Topmost signifi cance marks represent the comparison between Ampneg and Amp pos tumors (n = 4 normal liver; n ≥ 6 in tumor groups; n.s., not signifi cant; *, P < 0.05; **, P < 0.01; ***, P < 0.001). PDGF, platelet-derived growth factor.

macrophages, and leads to increased proliferation of tumor ble for local therapies and is a fi rst-line treatment for these cells. patients (3 ). As sorafenib inhibits VEGFRs and BRAF—a downstream effector of both VEGFRs and the HGF receptor Profi ling Proangiogenic-Factor c-MET (21 , 40 , 41 )—we tested whether sorafenib may have a −/− Expression in Mdr2 HCC selective advantage in Amp pos tumors. We treated 58 Mdr2 −/− As Amp neg tumors did not respond to sFLT treatment, we mice ages 14 to 18 months with sorafenib or vehicle alone profi led other proangiogenic factors that can support tumor for 3 days, after which they were sacrifi ced and tumor tissue vascularization. mRNA qPCR analysis of the genes encod- was analyzed (experimental design depicted in Supplemen- ing angiopoietin 1 and 2, angiopoietin like 2, FGF 1 and 2, tary Fig. S5A). Amplifi cation status was assessed after sac- platelet-derived growth factor (PDGF)-A, -B, and -C, placental rifi ce by DNA qPCR and verifi ed by Vegfa mRNA expression growth factor (PLGF) , and VEGF-B revealed that several of ( Fig. 7C and data not shown). Immunostaining for BrdUrd these factors were overexpressed in Mdr2 − /− HCCs compared and pHH3 demonstrated decreased proliferation in soraf- with normal livers, irrespective of the amplicon status. Nota- enib-treated mice in Amp pos but not Amp neg HCCs (Fig. 7A bly, Pdgfc levels were signifi cantly higher (2.4-fold) in Ampneg and B and Supplementary Fig. S8A and S8B). Similar to compared with Amppos tumors ( Fig. 6 ). This is in line with VEGF-A inhibition, Hgf levels were decreased only in soraf- a previous report showing that PDGF-C can promote ang- enib-treated Amp pos tumors ( Fig. 7C ). Although no effects iogenesis in a VEGF-A–independent manner ( 39 ) and could were observed on tumor macrophage density (Supplemen- provide a plausible explanation for the lack of effect on vessel tary Fig. S8A and S8B), expression of the TAM marker TGFβ density in Ampneg tumors in response to sFLT. was decreased in sorafenib-treated Amp pos tumors (Sup- plementary Fig. S8C). Blood vessel density changes or signs pos Murine Amp Tumors Are Uniquely of hypoxia were absent (Supplementary Fig. S8A–S8C), Sensitive to Sorafenib possibly due to the short treatment duration, implying Sorafenib is the only systemic drug showing a clinical that the early inhibitory effect of sorafenib in VEGF-A– advantage in patients with advanced HCC who are not eligi- amplifi ed tumors may be independent of angiogenesis.

JUNE 2014CANCER DISCOVERY | 737

RESEARCH ARTICLE Horwitz et al.

A B neg neg pos pos Amp Amp Amp Amp BrdUrd vehicle sorafenib vehicle sorafenib 2 * Ampneg Veh 1.8 neg 1.6 Amp Sor 1.4 Amppos Veh 1.2 Amppos Sor 1 0.8

BrdUrd 0.6 0.4 0.2

Percentage positive nuclei 0

Vegfa Hgf C D 1,500 10 4 n.s. Control - NT

* ) 2 Control - Sorafenib 8 3 VEGFA - NT 6 1,000 2 VEGFA - Sorafenib * 4 * 1 2 500 Relative expression Relative expression

0 0 Tumor volume (mm Ampneg Ampneg Amppos Amppos Ampneg Ampneg Amppos Amppos vehicle sorafenib vehicle sorafenib vehicle sorafenib vehicle sorafenib 0 0 51015 20 Days

Daily treatment period

E Resected HCC patients F Resected patients treated with sorafenib

VEGFA gains 100 VEGFA gains 100 Plogrank = 0.029 Normal VEGFA Normal VEGFA

50 50 Percentage survival Percentage survival

0 0 0 1,000 2,000 3,000 4,000 5,000 0 500 1,000 1,500 2,000 Days Days G H

VEGFA–amplified VEGF-A VEGFA–amplified malignant hepatocyte malignant hepatocyte Soluble VEGF receptor VEGF receptors Sorafenib VEGF genomic region HGF C-Met

VEGF-A VEGF-A Angiocrine Angiocrine Protumorigenic Decrease in mitogens factors factors HGF macrophages HGF Decreased Activated vascularity endothelial cells

Figure 7. Amp pos tumors are uniquely sensitive to sorafenib. A, representative BrdUrd immunostains. Scale bar, 100 μm. B, BrdUrd immunostaining was quantifi ed using automated image analysis (n ≥ 6; *, P < 0.05). C, qPCR analysis of Ampneg and Amp pos tumors treated as indicated. Each dot represents a different tumor. The cross line signifi es geometric mean (n.s., not signifi cant; *, P < 0.01). D, growth curves of xenografts transduced with control (dashed lines) or VEGF-A (solid lines) lentivectors, treated daily with sorafenib (gray lines) or nontreated (NT, black lines). Tumor volumes were measured with a caliper (n ≥ 6; *, P < 0.05). E, Kaplan–Meier curves showing survival of resected HCC patients negative (n = 96) or positive ( n = 14) for VEGFA gain (p log-rank > 0.05). F, Kaplan–Meier curves showing survival of patients with HCC treated with sorafenib, negative (n = 47, 10 months median) or positive for VEGFA gain ( n = 7; median undefi ned; plog-rank = 0.029). G and H, genomic gains in VEGFA promote tumorigenesis through the microenvironment. G, an increase in VEGFA gene copy number in liver tumor cells leads to elevated VEGF-A secretion. VEGF-A modulates the tumor microenvironment in favor of tumor cell growth through several modes: (i) recruitment of TAMs expressing the mitogen HGF and (ii) activating the liver endothelium to secrete angiocrine factors and enhance the tumor blood supply. H, inhibition of VEGF-A through soluble receptor or sorafenib results in decreased HGF signaling and blood supply, impeding tumor growth.

738 | CANCER DISCOVERYJUNE 2014 www.aacrjournals.org

VEGFA Ampliﬁ cation in HCC RESEARCH ARTICLE

Amp neg tumors did not show any measurable response to detected in a subset of HCCs, can promote tumor growth sorafenib. through a similar cell–cell interaction module. In addition, we treated mice bearing Hep3B xenografts, TAMs are key players in tumor progression and are known with or without VEGF overexpression, with sorafenib for 10 to modulate invasion, angiogenesis, immune response, and days. Similar to the Mdr2−/ − HCCs, this treatment markedly metastasis (36 ). TAMs are characterized by a specifi c phe- reduced growth, proliferation, and HGF expression in VEGF- notype, distinguished by a unique expression signature. A–overexpressing HCCs but did not affect control HCCs ( Fig. Previous studies have shown that VEGF-A recruits myeloid 7D and Supplementary Fig. S9A–S9D). Despite the longer cells that play an active part in vessel growth processes ( 37 , duration of treatment, we still could not detect changes in 42 ). In agreement, we observed that Amp pos tumors harbor macrophage or blood vessel densities (Supplementary Fig. higher numbers of TAMs compared with Ampneg tumors and S10A–S10C). Of note, a differential response to sorafenib was detected features of protumorigenic macrophages. Our fi nd- not evident in vitro, emphasizing the importance of microen- ings therefore raise the interesting possibility that the VEGFA vironmental factors (Supplementary Fig. S10D). amplicon contributes to the infl ammatory microenvironment, which supports developing tumors. Benefi cial Effect of Sorafenib Treatment in Human Our data suggest that VEGFA genomic gains facilitate Patients with HCCs Bearing VEGFA Gains tumor development by several different modes: (i) providing Noting the predictive potential of Vegfa gains in the mouse a microenvironment rich in TAMs; (ii) promoting prolif- model, we analyzed samples from patients with HCC who eration via stroma-derived HGF secretion (and possibly other underwent tumor resection. This retrospective cohort was col- cytokines as well); and (iii) enhancing tumor angiogenesis. lected from three different centers. To assess the correlation Interestingly, all these modes entail heterotypic cellular inter- between VEGFA gain and survival, we analyzed human tumor actions, distinguishing this particular genomic gain from samples by FISH (Fig. 1C). Survival of patients with HCC who other studied amplicons ( 43 ). We maintain that the unique did not receive sorafenib was independent of VEGFA status sensitivity of tumors with VEGFA gains to VEGF-A block- ( Fig. 7E ). However, markedly improved survival was seen in ade stems from these multiple protumorigenic functions of the VEGFA-gain group compared with the non-gain group VEGF-A ( Fig. 7G and H ). in sorafenib-treated patients (indefi nable median survival An amplicon spanning VEGFA was noted in several differ- from sorafenib treatment start and 11 months, respectively; ent human cancers ( 29–31 , 33 , 43–51 ). A linear correlation p log-rank = 0.029; Fig. 7F). Taken together, our mouse data was found between mRNA levels of VEGFA and extent of and retrospective analysis of a human cohort imply that amplifi cation in human HCC (29 ). Amplifi cations in the VEGFA gains correlate with a particularly benefi cial response VEGFA locus and juxtaposed regions were associated with to sorafenib, and possibly other VEGF-A inhibitors, in HCC. advanced-stage HCC ( 30 ). In colorectal carcinoma and breast cancer, this amplifi cation was found in correlation with vascular invasion and shorter survival ( 46 , 51 ). Cumulative anal- DISCUSSION ysis of these reported human studies shows that gains and Using a mouse model of infl ammation-induced HCC, amplifi cations of VEGFA are found in 7% to 30% of human we identifi ed and characterized a unique group of HCC in HCCs ( 29–33 ). Our interventional studies in the Mdr2 − /− humans and mice. These tumors are defi ned by genomic gains model indicate that Vegfa is a major driver of this amplicon, of a region encompassing VEGFA, and are distinct in histologic which minimally harbors 53 genes in Mdr2 − /− murine tumors appearance, rate of proliferation, and microenvironmental and 11 genes in humans ( 32 ). Xenograft experiments reveal content. We delineated cytokine-based heterotypic cross-talk that overexpression of VEGF-A in a human HCC cell line is between malignant Amppos hepatocytes and tumor stromal suffi cient to upregulate HGF expression and increase tumor cells. Importantly, we showed that mouse Amppos tumors are cell proliferation in vivo and that the tumor growth advantage uniquely sensitive to VEGF-A inhibition and to sorafenib. A gained through this overexpression can be negated by sora- retrospective analysis of human HCCs indicates that genomic fenib. Nevertheless, we cannot completely exclude the contri- gains of VEGFA can predict response to sorafenib. bution of other amplicon genes to tumorigenesis. The Mdr2 −/− model of infl ammation-induced HCC yields The fi rst line of treatment for advanced HCC is the multi- primary tumors each holding specifi c genetic changes. There- kinase inhibitor sorafenib, which prolongs median survival fore, we denote it as a sound platform to test the pro- by 10 to 12 weeks (3, 4). The response to sorafenib seems tumorigenic effects of recurring changes in an unbiased to be variable, and treatment is associated with signifi cant manner. Moreover, the infl ammatory background in this side effects (3–8 ). Importantly, there are no clinically applied model is particularly relevant for studying tumor–microenvi- biomarkers for predicting sorafenib response in HCC (10 ). ronment interactions in clinically relevant settings. Previous Among the multiple targets of sorafenib (16 ) are VEGFRs work showed that systemic elevation of VEGF-A induces and BRAF—a downstream effector of both VEGFRs and the proliferation of normal hepatocytes through factors secreted HGF receptor c-MET ( 21 , 40 ). Indeed, in line with our mouse from endothelial cells ( 24 ). Although hepatocytes are inert results, sorafenib treatment in patients led to a decrease in to VEGF-A, liver sinusoidal cells respond to VEGF-A with serum HGF (10 ). Unlike most tumors, advanced HCC is counter-secretion of HGF ( 24 ). Furthermore, in regenerat- usually diagnosed and treated without obtaining tumor tis- ing livers, hepatocyte proliferation was shown to depend on sue, making it diffi cult to establish tissue-based predictive VEGF-A–induced expression of HGF from endothelial cells biomarkers. On the basis of our small-scale retrospective (25 ). Here, we show that recurring genomic gains in VEGFA , study, we show that VEGFA gains may predict response to

JUNE 2014CANCER DISCOVERY | 739

RESEARCH ARTICLE Horwitz et al. sorafenib in HCC, thus enabling the tailoring of treatment 60% adenovector transduction effi ciency (by tissue staining) were to only those patients who may benefi t from this side effect– excluded. Lentiviral-based vectors were prepared by subcloning the prone therapy. Notably, the same amplifi cation was also PCR products of the human VEGFA 165 gene [from cDNA of decidual found in lung, colorectal, bone, and breast cancers (44–48 ); natural killer (NK ) cells; a kind gift from Ofer Mandelboim, Hebrew thus, it is plausible that similar considerations could be University, Jerusalem, Israel] into the pSC-B plasmid using the Strata- Clone Kit (Stratagene), subsequently digested with Bam HI and Not I applied to other tumors harboring VEGFA gains. and subcloned into the self-inactivating lentiviral vector pHAGE (gift of Gustavo Mostoslavsky, Boston University School of Medicine, Bos- METHODS ton, MA) digested with Bam HI and Not I. Lentivectors were produced by cotransfection of the backbone vector plasmid with the gag-pol Human Tissue Samples and Tissue Microarray and pMD.G plasmids and using standard procedures. Hep3B cells Human HCC tissues were obtained from resected patients from (obtained from the ATCC) were grown in Dulbecco’s Modifi ed Eagle the institutes of Basel University Hospital (Basel, Switzerland), Han- Medium (DMEM; 10% FBS). Cells were tested free of Mycoplasma nover Medical School Hospital (Hannover, Germany), and Heidel- before transduction and injection. Lentivector transduction effi ciency berg University Hospital (Heidelberg, Germany). Clinical information was assessed by fl uorescent microscopy and was estimated as 80%. included age at diagnosis, tumor diameter, gender, and survival time In vitro proliferation was determined through XTT assay (Biological information. Examination of tumor hematoxylin and eosin (H&E) Industries) using the manufacturer’s protocol. Murine recombinant sections was performed by an expert liver pathologist (O. Pappo). VEGF-A (R&D Systems) was used at the concentration of 100 ng/mL. Samples from Heidelberg were also reviewed by Heidelberg pathologists (C. Mogler and P. Schirmacher). The study was approved by IHC, Immunofl uorescence, and ELISA each of the institutions’ ethics committees—numbered 206/05 Antibodies used for tissue immunostaining throughout the (Heidelberg), 660-2010 (Hannover), and EKBB20 (Basel). Required work were—vWF (dilution at 1:300; Dako), pHH3 (1:800; Upstate), cohort size was computed by power analysis to yield a power of at cleaved caspase-3 (1:200; Cell Signaling Technology), F4/80 (1:300; least 90% with an α value of 0.05. Construction of the tissue micro- Seroteq), HGF (1:100; R&D Systems), BrdUrd (1:200; NeoMarkers), array (TMA) was performed as follows: tissue samples were fi xed KDR (1:400; Cell Signaling Technology), Ki67 (1:100; NeoMarkers), in buffered 10% formaldehyde and embedded in paraffi n. H&E- E-cadherin (1:100; Cell Signaling Technology), and VEGF (1× as stained sections were made from each selected primary block (named supplied; Spring). IHC was performed on 5-μm paraffi n sections. donor blocks) to defi ne representative tissue regions. Tissue cylinders Antigen retrieval was performed in a decloaking chamber (Biocare (0.6 mm in diameter) were then punched from the region of the Medical) in citrate buffer for all antibodies except vWF and F4/80, donor block with the use of a custom-made precision instrument for which retrieval was performed with pronase (Sigma). Horserad- (Beecher Instruments). Afterward, tissue cylinders were transferred to ish peroxidase (HRP)–conjugated secondary antibodies for all IHC × a 25 mm 35 mm paraffi n block to produce the TMAs. The resulting antibodies used were Histofi ne (Nichirei Biosciences), except for anti- TMA block was cut into 3-μm sections that were transferred to glass mouse–derived antibodies that were detected with Envison (Dako). slides by use of the Paraffi n Sectioning Aid System (Instrumedics). 3,3′-Diaminobenzidine (DAB; Lab Vision) was used as a chromogen. Sections from the TMA blocks were used for FISH analysis. Immunohistochemical stainings were quantitated when indicated using an Ariol SL-50 system (Applied Imaging). For quantifi cation Mice of nuclear immunostaining, the ki-sight module of the Ariol-SL50 Male and female Mdr2 − /− mice on FVB background were held in robotic image analysis system was applied. This system designates specifi c pathogen-free conditions. Experiments on Mdr2 − /− mice were classifi ers for positive (red-brown) and negative (azure) nuclei defi ned performed in cohorts of 10 to 20 mice each time; results show the by color intensity, size, and shape. Each tumor cell nucleus (distin- combined data from at least four different experiments. Wild-type guished by morphology) was designated as either positive or negative (WT) controls were age-matched FVB mice. Mouse body weights dur- by these parameters. The fraction of positive cells was calculated ing the experiments were 30 to 45 g. Sorafenib (Xingcheng Chemp- from counting at least fi ve randomly selected fi elds in each tumor. harm Co., Ltd.) was administered daily (50 mg/kg) by oral gavage. Immunofl uoresence was performed on snap-frozen tissue embed- Cremophor EL (Sigma)/ethanol/water (1:1:6) was used as a vehicle. ded in OCT gel (Sakura Finetek) and sectioned to 8-μm slices. Two hours before sacrifi ce, mice were injected with 10 μL BrdUrd Slides were incubated at 37°C and fi xed with both acetone and 4% (Cell Proliferation labeling reagent; Amersham) per 1 g body weight. paraformaldehyde sequentially. Fluorophore-conjugated secondary Mice were anesthetized with ketamine and xylazine and the liver was antibodies used were donkey anti-goat Alexa Fluor 647 (Invitro- perfused via the heart with PBS–Heparin solution followed by 4% gen), donkey anti-rabbit Cy2/Cy5, donkey anti-mouse Cy3, and goat formaldehyde. Following perfusion, livers were removed and sub- anti-rat Cy3 (The Jackson Laboratory). Hoechst 33342 was used as jected to standard histologic procedures. Xenograft experiments were a nuclear marker (Invitrogen). Antibodies used for fl ow FACS sort- performed by subcutaneous injection of transduced Hep3B cells (2.5 ing were CD45-pacifi c blue, F4/80-PE (both 1:50; BioLegend), and × 106 ) suspended in 100 μL PBS and 100 μL Matrigel (Becton Dickin- Meca32-biotin (1:50; BioLegend) used with streptavidin APC-Cy7 son) into NOD/SCID mice. Tumor volumes were assessed by external (BD Biosciences). Flow cytometry–based cell sorting was performed measurement with calipers. All animal experiments were performed in a FACSAria III cell sorter (BD Biosciences). VEGF-A ELISA was in accordance with the guidelines of the institutional committee for performed using Quantikine mouse ELISA kit (R&D Systems). the use of animals for research. In all mouse experiments, the different groups were housed together in the same cages. DNA In Situ Hybridizations Probes for CISH analysis of mouse tumors were prepared from the Viral Vectors and Cultured Cells BAC clones RP24-215A3 for the murine Chromosome 17 (BACPAC Adenoviral vectors encoding GFP or GFP and sFLT—a kind gift resources center). BAC clones were labeled with DIG using Nick- from David Curiel (Washington University, St. Louis, MO) and Yosef Translation mix (Roche). Mouse Cot-1 DNA (Invitrogen) and soni- Haviv (Hadassah Hospital, Jerusalem, Israel)—were prepared in GH354 cated murine genomic DNA were added to the probe for background cells using standard procedures. A titer of 109 transducing units was block. Tissues were prepared by boiling in pretreatment buffer and injected into mice tail veins. Mice whose livers did not yield minimal digestion with pepsin (Zymed). Hybridization was performed at 37°C

740 | CANCER DISCOVERYJUNE 2014 www.aacrjournals.org

VEGFA Ampliﬁ cation in HCC RESEARCH ARTICLE

overnight after denaturation at 95°C for 5 minutes. The SPoT-Light applied. Primers detecting the murine chromosome 17 pericentro- Detection Kit (Invitrogen) was used for anti-DIG antibody and HRP- meric region were used as references in DNA qPCR analyses. conjugated secondary antibody. FISH analysis for human HCCs was performed as follows. The Cell Separation genomic BAC clone RPCIB753M0921Q (imaGENES), which covers the Hepatocytes and macrophages were isolated from Mdr2 −/ − mice human VEGFA gene region, was used for preparation of the FISH probe. livers essentially as described by Kamimura and Tsukamoto ( 54 ). BAC-DNA was isolated using the Large-Construct Kit (Qiagen) accord- Briefl y, livers were digested enzymatically with pronase and colla- μ ing to the instructions of the manufacturer. Isolated BAC-DNA (1 g) genase (Sigma) by in situ perfusion. Hepatocytes were isolated by cen- was digested with Alu I restriction enzyme (Invitrogen) and labeled with trifugation at 50 × g for 2 minutes, and after three washes were frozen Cy3-dUTP (GE Healthcare) using the BioPrime Array CGH Kit (Invit- immediately in liquid nitrogen for RNA preparation. Nonparenchy- rogen). The labeling reaction was assessed by NanoDrop. Labeled DNA mal cells were pelleted by centrifugation at 150 × g for 8 minutes, laid was purifi ed with the FISH Tag DNA Kit (Invitrogen). TMAs and whole- on top of a four-density gradient of Larcoll (Sigma) and centrifuged tissue sections were deparaffi nized in xylene for 20 minutes and subse- at 20,000 rpm at 25°C for 30 minutes using a SW41Ti rotor (Beck- quently washed with 100%, 96%, and 70% ethanol followed by a wash man). Liver macrophages were recovered from the interface between ° with tap water (2 minutes each step). Slides were air dried at 75 C for 3 8% and 12% Larcoll, washed three times and immediately frozen in minutes. Slides were then boiled in pretreatment buffer [70% formamide, liquid nitrogen. Purity of hepatocytes and macrophage fractions was × ° 2 saline-sodium citrate buffer (SSC)] at 100 C for 15 minutes followed determined by H&E staining of cytospin preparations and always by a wash with tap water. Tissue was then subjected to Proteinase K exceeded 90%. Dissociation of cells from tumor xenografts was per- ° (Sigma) treatment at 37 C for 70 minutes followed by a wash in tap formed using the gentleMACS dissociator (Miltenyi Biotec) accord- water (2 minutes). Dehydration of slides was performed by serial immer- ing to the manufacturer’s protocol. sion of slides in 70%, 96%, and 100% ethanol (2 minutes each step). Slides ° were then air dried at 75 C for 3 minutes. The FISH probe was applied Statistical Analysis and slides were sealed with rubber cement. Following a denaturation < step (10 minutes at 75°C), slides were incubated overnight at 37°C. Data were analyzed using a paired two-tailed Student t test at P χ2 Slides were washed in Wash Buffer (2× SSC, 0.3% NP40, pH 7–7.5) and 0.05. Histologic differences were analyzed using the Pearson test < counterstained with 4′,6-diamidino-2-phenylindole (DAPI) I solution at P 0.05. Data were processed using Microsoft Excel 2007. Graphs (1,000 ng/mL; Vysis Abbott Molecular). As reference, a Spectrum Green- were generated using either GraphPad Prism 5.0 or Excel software. labeled chromosome 6 centromeric probe (Vysis Abbott Molecular) was Kaplan–Meier calculations and graphs were performed in GraphPad used. Images were obtained with a Zeiss fl uorescence microscope using Prism 5.0. Log-rank (Mantel–Cox) was used to determine survival P a 63× objective (Zeiss) and the Axiovision software (Zeiss). value. Throughout the work, error bars represent 1 SEM. FISH results were evaluated according to (i) absolute VEGFA gene copy number and chromosome 6 copy number and (ii) VEGFA gene/ Disclosure of Potential Confl icts of Interest chromosome 6 copy number ratio. The following classifi cation was R. Koschny has received travel grants from Bayer Company. used: not amplifi ed—VEGFA /Chr6 ratio of less than 1.8; equivocal/ E. Pikarsky has fi led for a provisional patent application based on borderline—VEGFA /Chr6 ratio between 1.8 and 2.2; and amplifi ed— this article. No potential confl icts of interest were disclosed by the VEGFA/Chr6 ratio higher than 2.2, as proposed by the ASCO/CAP other authors. (American Society of Clinical Oncology/College of American Pathol- ogists) guidelines for HER2 amplifi cation in breast cancer. High poly- Authors’ Contributions somy was defi ned as >3.75 copies of the CEP6 probe, low polysomy Conception and design: E. Horwitz, I. Stein, J. Hess, P. Angel, was defi ned as cases displaying between 2.26 and 3.75 copies of the Y. Ben-Neriah, E. Pikarsky CEP6 probe (52, 53). All cases displaying either amplifi cation or poly- Development of methodology: E. Horwitz, I. Stein, M. Andreozzi, somy were collectively defi ned as VEGFA gain. FISH quantifi cation R.M. Porat, M. Grunewald and classifi cation were done by an expert molecular pathologist who Acquisition of data (provided animals, acquired and managed had no access to the clinical data (L. Tornillo). patients, provided facilities, etc.): E. Horwitz, I. Stein, J. Nemeth, aCGH and qPCR A. Shoham, O. Pappo, N. Schweitzer, L. Tornillo, N. Kanarek, L. Quagliata, F. Zreik, R.M. Porat, R. Finkelstein, R. Koschny, T. Ganten, C. Mogler, Genomic DNA was isolated using the QIAGEN DNAeasy Tissue O. Shibolet, K. Breuhahn, M. Grunewald, P. Schirmacher, A. Vogel, Kit. Samples were hybridized to mouse CGH 60-mer oligonucle- L. Terracciano otides microarrays (Agilent Technologies), washed, and scanned Analysis and interpretation of data (e.g., statistical analysis, according to the Agilent Technologies’ instructions. Data were ana- biostatistics, computational analysis): E. Horwitz, A. Shoham, lyzed using Feature Extraction software V8.1 (Agilent), GeneSpring O. Pappo, N. Schweitzer, L. Quagliata, H. Reuter, C. Mogler, O. Shibolet, GX V7.3.1, and CGH Analytics V3.4.27 (Agilent) software. RNA J. Hess, A. Vogel, L. Terracciano, P. Angel, Y. Ben-Neriah, E. Pikarsky was extracted from tissues by mechanical grinding in TriReagent Writing, review, and/or revision of the manuscript: E. Horwitz, (Sigma) with a Polytron tissue homogenizer (Kinematica) at maxi- I. Stein, N. Schweitzer, R. Koschny, T. Ganten, C. Mogler, O. Shibolet, mum speed. cDNA was prepared with Moloney murine leukemia J. Hess, P. Schirmacher, P. Angel, Y. Ben-Neriah, E. Pikarsky virus (MMLV) reverse transcriptase (Invitrogen). qPCR analyses were Administrative, technical, or material support (i.e., reporting or carried out with SYBR green (Invitrogen) in 7900HT Fast Real-Time organizing data, constructing databases): E. Horwitz, A. Shoham, PCR System (Applied Biosystems). Results were analyzed using the N. Kanarek, L. Quagliata, T. Ganten, J. Hess, P. Schirmacher, A. Vogel, qBase v1.3.5 software. Primer sequences are available in Supplemen- L. Terracciano tary Table S5. In the xenografts experiment, murine Hgf mRNA levels Study supervision: Y. Ben-Neriah, E. Pikarsky were assessed with TaqMan probe (Life Sciences). Human hypox- anthine phosphoribosyltransferase (HPRT ) and ubiquitin C ( UBC ) were used as reference genes in the xenograft experiment. Hprt and Acknowledgments peptidylprolyl isomerase A (Ppia ) combined were used as reference The authors thank Rivka Ben-Sasson, Reba Condioti, Shaffi ka El- genes in all murine analyses except for the hepatocyte versus mac- Kawasmi, Etti Avraham, Mohamad Juma’, and Drs. Hila Giladi, Gus- rophage comparison in which Ubc, B2m, and Tbp were additionally tavo Mostoslavsky, David Curiel, Yosef Haviv, Shemuel Ben-Sasson,

JUNE 2014CANCER DISCOVERY | 741

RESEARCH ARTICLE Horwitz et al. and Sharon Elizur for providing expertise and reagents. The authors 13. Hoshida Y , Toffanin S , Lachenmayer A , Villanueva A , Minguez B , also thank Drs. Jacob Hannah, Hidekazu Tsukamoto, Ofer Man- Llovet JM . Molecular classifi cation and novel targets in hepatocellular delboim, Noam Stern-Ginosar, Noa Stanietsky, Rachel Yamin, Seth carcinoma: recent advancements. Semin Liver Dis 2010 ; 30 : 35 – 51 . Salpeter, Temima Schnitzer-Perlman, Dan Lehmann, Rachel Horwitz, 14. Piccart-Gebhart MJ , Procter M , Leyland-Jones B , Goldhirsch A , Untch and Chamutal Gur for expert advice and kind assistance. The authors M , Smith I , et al. Trastuzumab after adjuvant chemotherapy in are indebted to Dr. Daniel Goldenberg for supplying aged Mdr2 − /− HER2-positive breast cancer. N Engl J Med 2005 ; 353 : 1659 – 72 . mice and are grateful to Drs. Christoffer Gebhardt, Robert Goldstein, 15. Karapetis CS , Khambata-Ford S , Jonker DJ , O’Callaghan CJ , Tu D , Moshe Biton, Zvika Granot, and Tzachi Hagai for fruitful discussions. Tebbutt NC , et al. K-ras mutations and benefi t from cetuximab in advanced colorectal cancer. N Engl J Med 2008 ; 359 : 1757 – 65 . Grant Support 16. Wilhelm SM , Carter C , Tang L , Wilkie D , McNabola A , Rong H, et al. BAY 43-9006 exhibits broad spectrum oral antitumor activity and tar- The research leading to these results has received funding from gets the RAF/MEK/ERK pathway and receptor tyrosine kinases involved the European Research Council under the European Union’s Seventh in tumor progression and angiogenesis. Cancer Res 2004 ; 64 : 7099 – 109 . Framework Programme (FP/2007-2013)/ERC Grant Agreements 17. Govindarajan R , Siegel E , Makhoul I , Williamson S . Bevacizumab and 281738 to E. Pikarsky and 294390 to Y. Ben-Neriah; the Dr. Miriam erlotinib in previously untreated inoperable and metastatic hepato- and Sheldon G. Adelson Medical Research Foundation (AMRF), the cellular carcinoma. Am J Clin Oncol 2013 ; 36 : 254 – 7 . German Research Foundation (DFG, SFB-TR77), the Cooperation 18. Siegel AB , Cohen EI , Ocean A , Lehrer D , Goldenberg A , Knox JJ , Program in Cancer Research of the Deutsches Krebsforschungszen- et al. Phase II trial evaluating the clinical and biologic effects of trum (DKFZ), and Israel’s Ministry of Science, Culture and Sport bevacizumab in unresectable hepatocellular carcinoma . J Clin Oncol (MOST) to E. Pikarsky, Y. Ben-Neriah, and P. Angel; the Israel Sci- 2008 ; 26 : 2992 – 8 . ence Foundation (ISF) Centers of Excellence to E. Pikarsky and 19. Thomas MB , Morris JS , Chadha R , Iwasaki M , Kaur H , Lin E , et al. Y. Ben-Neriah; and an ISF Project Grant to I. Stein and O. Pappo. Phase II trial of the combination of bevacizumab and erlotinib in patients who have advanced hepatocellular carcinoma . J Clin Oncol 2009 ; 27 : 843 – 50 . Received October 21, 2013; revised March 26, 2014; accepted 20. Yau T , Wong H , Chan P , Yao TJ , Pang R , Cheung TT , et al. Phase II March 26, 2014; published OnlineFirst March 31, 2014. study of bevacizumab and erlotinib in the treatment of advanced hepatocellular carcinoma patients with sorafenib-refractory disease . Invest New Drugs 2012 ; 30 : 2384 – 90 . REFERENCES 21. Ferrara N . Vascular endothelial growth factor. Arterioscler Thromb 1. Farazi PA , DePinho RA . Hepatocellular carcinoma pathogenesis: Vasc Biol 2009 ; 29 : 789 – 91 . from genes to environment. Nat Rev Cancer 2006 ; 6 : 674 – 87 . 22. Lichtenberger BM , Tan PK , Niederleithner H , Ferrara N , Petzelbauer P, 2. El-Serag HB . Hepatocellular carcinoma. N Engl J Med 2011 ; 365 : Sibilia M . Autocrine VEGF signaling synergizes with EGFR in tumor 1118– 27 . cells to promote epithelial cancer development. Cell 2010 ; 140 : 268 – 79 . 3. Llovet JM , Ricci S , Mazzaferro V , Hilgard P , Gane E , Blanc JF, et al. 23. Chatterjee S , Heukamp LC , Siobal M , Schöttle J , Wieczorek C , Peifer Sorafenib in advanced hepatocellular carcinoma . N Engl J Med M , et al. Tumor VEGF: VEGFR2 autocrine feed-forward loop triggers 2008 ; 359 : 378 – 90 . angiogenesis in lung cancer. J Clin Invest 2013 ; 123 : 1732 – 40 . 4. Cheng AL , Kang YK , Chen Z , Tsao CJ , Qin S , Kim JS , et al. Effi cacy 24. LeCouter J , Moritz DR , Li B , Phillips GL , Liang XH , Gerber H P, and safety of sorafenib in patients in the Asia-Pacifi c region with et al. Angiogenesis-independent endothelial protection of liver: role advanced hepatocellular carcinoma: a phase III randomised, double- of VEGFR-1. Science 2003 ; 299 : 890 – 3 . blind, placebo-controlled trial. Lancet Oncol 2009 ; 10 : 25 – 34 . 25. Ding BS , Nolan DJ , Butler JM , James D , Babazadeh AO , Rosenwaks 5. Iavarone M , Cabibbo G , Piscaglia F , Zavaglia C , Grieco A , Villa Z , et al. Inductive angiocrine signals from sinusoidal endothelium are E , et al. Field-practice study of sorafenib therapy for hepatocellu- required for liver regeneration. Nature 2010 ; 468 : 310 – 5 . lar carcinoma: a prospective multicenter study in Italy. Hepatology 26. Hernandez-Gea V , Toffanin S , Friedman SL , Llovet JM . Role of the 2011 ; 54 : 2055 – 63 . microenvironment in the pathogenesis and treatment of hepatocel- 6. Jia N , Liou I , Halldorson J , Carithers R , Perkins J , Reyes J , et al. Phase I lular carcinoma. Gastroenterology 2013 ; 144 : 512 – 27 . adjuvant trial of sorafenib in patients with hepatocellular carcinoma 27. Mauad TH , van Nieuwkerk CM , Dingemans KP , Smit JJ , Schinkel AH , after orthotopic liver transplantation. Anticancer Res 2013 ; 33 : 2797 – 800 . Notenboom RG , et al. Mice with homozygous disruption of the mdr2 7. Balsom SM , Li X , Trolli E , Rose J , Bloomston M , Patel T , et al. A P-glycoprotein gene. A novel animal model for studies of nonsup- single-institute experience with sorafenib in untreated and previously purative infl ammatory cholangitis and hepatocarcinogenesis . Am J treated patients with advanced hepatocellular carcinoma. Oncology Pathol 1994 ; 145 : 1237 – 45 . 2010 ; 78 : 210 – 2 . 28. Shukla R , Upton KR , Munoz-Lopez M , Gerhardt DJ , Fisher ME , 8. Brunocilla PR , Brunello F , Carucci P , Gaia S , Rolle E , Cantamessa Nguyen T , et al. Endogenous retrotransposition activates oncogenic A , et al. Sorafenib in hepatocellular carcinoma: prospective study pathways in hepatocellular carcinoma. Cell 2013 ; 153 : 101 – 11 . on adverse events, quality of life, and related feasibility under daily 29. Chiang DY , Villanueva A , Hoshida Y , Peix J , Newell P , Minguez B , et al. conditions. Med Oncol 2013 ; 30 : 345 . Focal gains of VEGFA and molecular classifi cation of hepatocellular 9. Dai X , Schlemmer HP , Schmidt B , Hoh K , Xu K , Ganten TM , et al. carcinoma. Cancer Res 2008 ; 68 : 6779 – 88 . Quantitative therapy response assessment by volumetric iodine- 30. Chochi Y , Kawauchi S , Nakao M , Furuya T , Hashimoto K , Oga A, e t a l . uptake measurement: initial experience in patients with advanced A copy number gain of the 6p arm is linked with advanced hepatocel- hepatocellular carcinoma treated with sorafenib. Eur J Radiol lular carcinoma: an array-based comparative genomic hybridization 2013 ; 82 : 327 – 34 . study. J Pathol 2009 ; 217 : 677 – 84 . 10. Llovet JM , Pena CE , Lathia CD , Shan M , Meinhardt G , Bruix J , et al. 31. Patil MA , Gutgemann I , Zhang J , Ho C , Cheung ST , Ginzinger D , e t a l . Plasma biomarkers as predictors of outcome in patients with advanced Array-based comparative genomic hybridization reveals recurrent hepatocellular carcinoma. Clin Cancer Res 2012 ; 18 : 2290 – 300 . chromosomal aberrations and Jab1 as a potential target for 8q gain in 11. Tsukui Y , Mochizuki H , Hoshino Y , Kawakami S , Kuno T , Fukasawa hepatocellular carcinoma. Carcinogenesis 2005 ; 26 : 2050 – 7 . Y , et al. Factors contributing to the overall survival in patients with 32. Beroukhim R , Mermel CH , Porter D , Wei G , Raychaudhuri S , Dono - hepatocellular carcinoma treated by sorafenib. Hepatogastroenterol- van J , et al. The landscape of somatic copy-number alteration across ogy 2012 ; 59 : 2536 – 9 . human cancers. Nature 2010 ; 463 : 899 – 905 . 12. Strebhardt K , Ullrich A . Paul Ehrlich’s magic bullet concept: 100 33. Moinzadeh P , Breuhahn K , Stutzer H , Schirmacher P . Chromosome years of progress. Nat Rev Cancer 2008 ; 8 : 473 – 80 . alterations in human hepatocellular carcinomas correlate with aetiology

742 | CANCER DISCOVERYJUNE 2014 www.aacrjournals.org

VEGFA Ampliﬁ cation in HCC RESEARCH ARTICLE

and histological grade—results of an explorative CGH meta-analysis . 46. Vlajnic T , Andreozzi MC , Schneider S , Tornillo L , Karamitopoulou Br J Cancer 2005 ; 92 : 935 – 41 . E , Lugli A , et al. VEGFA gene locus (6p12) amplifi cation identifi es 34. Ouchi K , Sugawara T , Ono H , Fujiya T , Kamiyama Y , Kakugawa Y , e t a l . a small but highly aggressive subgroup of colorectal patients . Mod Mitotic index is the best predictive factor for survival of patients with Pathol 2011 ; 24 : 1404 – 12 . resected hepatocellular carcinoma. Dig Surg 2000 ; 17 : 42 – 8 . 47. Horlings HM , Lai C , Nuyten DS , Halfwerk H , Kristel P , van Beers E , 35. Qian BZ , Pollard JW . Macrophage diversity enhances tumor progres- et al. Integration of DNA copy number alterations and prognostic sion and metastasis. Cell 2010 ; 141 : 39 – 51 . gene expression signatures in breast cancer patients. Clin Cancer Res 36. Sica A , Larghi P , Mancino A , Rubino L , Porta C , Totaro MG , et al. 2010 ; 16 : 651 – 63 . Macrophage polarization in tumour progression. Semin Cancer Biol 48. Lau CC , Harris CP , Lu XY , Perlaky L , Gogineni S , Chintagumpala 2008 ; 18 : 349 – 55 . M , et al. Frequent amplifi cation and rearrangement of chromosomal 37. Grunewald M , Avraham I , Dor Y , Bachar-Lustig E , Itin A , Jung S , et al. bands 6p12-p21 and 17p11.2 in osteosarcoma . Genes Chromosomes VEGF-induced adult neovascularization: recruitment, retention, and Cancer 2004 ; 39 : 11 – 21 . role of accessory cells. Cell 2006 ; 124 : 175 – 89 . 49. Andreozzi M , Quagliata L , Gsponer JR , Ruiz C , Vuaroqueaux V , 38. Mahasreshti PJ , Navarro JG , Kataram M , Wang MH , Carey D , Siegal Eppenberger-Castori S , et al. VEGFA gene locus analysis across 80 GP , et al. Adenovirus-mediated soluble FLT-1 gene therapy for ovar- human tumour types reveals gene amplifi cation in several neoplastic ian carcinoma. Clin Cancer Res 2001 ; 7 : 2057 – 66 . entities. Angiogenesis. 2013 Oct 11. [Epub ahead of print]. 39. Crawford Y , Kasman I , Yu L , Zhong C , Wu X , Modrusan Z , et al. 50. Yang J , Yang D , Sun Y , Sun B , Wang G , Trent JC , et al. Genetic ampli- PDGF-C mediates the angiogenic and tumorigenic properties of fi cation of the vascular endothelial growth factor (VEGF) pathway fi broblasts associated with tumors refractory to anti-VEGF treat- genes, including VEGFA, in human osteosarcoma . Cancer 2011 ; 117 : ment. Cancer Cell 2009 ; 15 : 21 – 34 . 4925 – 38 . 40. Trusolino L , Bertotti A , Comoglio PM . MET signalling: principles 51. Schneider BP , Gray RJ , Radovich M , Shen F , Vance G , Li L , et al. and functions in development, organ regeneration and cancer. Nat Prognostic and predictive value of tumor vascular endothelial growth Rev Mol Cell Biol 2010 ; 11 : 834 – 48 . factor gene amplifi cation in metastatic breast cancer treated with 41. Whittaker S , Marais R , Zhu AX . The role of signaling pathways in the paclitaxel with and without bevacizumab; results from ECOG 2100 development and treatment of hepatocellular carcinoma. Oncogene trial. Clin Cancer Res 2013 ; 19 : 1281 – 9 . 2010 ; 29 : 4989 – 5005 . 52. Salido M , Tusquets I , Corominas JM , Suarez M , Espinet B , Corzo C , 42. Wyckoff JB , Wang Y , Lin EY , Li JF , Goswami S , Stanley ER, et al. et al. Polysomy of chromosome 17 in breast cancer tumors showing Direct visualization of macrophage-assisted tumor cell intravasation an overexpression of ERBB2: a study of 175 cases using fl uorescence in mammary tumors. Cancer Res 2007 ; 67 : 2649 – 56 . in situ hybridization and immunohistochemistry . Breast Cancer Res 43. Albertson DG . Gene amplifi cation in cancer. Trends Genet 2006 ; 22 : 2005 ; 7 : R267 – 73 . 447– 55 . 53. Wolff AC , Hammond ME , Schwartz JN , Hagerty KL , Allred DC , 44. Weir BA , Woo MS , Getz G , Perner S , Ding L , Beroukhim R , et al. Cote RJ , et al. American Society of Clinical Oncology/College of Characterizing the cancer genome in lung adenocarcinoma . Nature American Pathologists guideline recommendations for human epi- 2007 ; 450 : 893 – 8 . dermal growth factor receptor 2 testing in breast cancer . J Clin Oncol 45. Tsafrir D , Bacolod M , Selvanayagam Z , Tsafrir I , Shia J , Zeng Z , et al. 2007 ; 25 : 118 – 45 . Relationship of gene expression and chromosomal abnormalities in 54. Kamimura S , Tsukamoto H . Cytokine gene expression by Kupffer cells colorectal cancer. Cancer Res 2006 ; 66 : 2129 – 37 . in experimental alcoholic liver disease. Hepatology 1995 ; 22 : 1304 – 9 .

JUNE 2014CANCER DISCOVERY | 743

Human and Mouse VEGFA-Amplified Hepatocellular Carcinomas Are Highly Sensitive to Sorafenib Treatment

Elad Horwitz, Ilan Stein, Mariacarla Andreozzi, et al.

Cancer Discovery 2014;4:730-743. Published OnlineFirst March 31, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-13-0782

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2014/04/01/2159-8290.CD-13-0782.DC1 http://cancerdiscovery.aacrjournals.org/content/suppl/2021/03/04/2159-8290.CD-13-0782.DC2

Cited articles This article cites 53 articles, 14 of which you can access for free at: http://cancerdiscovery.aacrjournals.org/content/4/6/730.full#ref-list-1

Citing articles This article has been cited by 18 HighWire-hosted articles. Access the articles at: http://cancerdiscovery.aacrjournals.org/content/4/6/730.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/4/6/730. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.