Targeting the Active B-Catenin Pathway to Treat Cancer Cells

2861

Targeting the active B-catenin pathway to treat cancer cells

Hadas Dvory-Sobol,1,2 Eyal Sagiv,1,2 PUMA, whereas that of cells with low levels of B-catenin Diana Kazanov,1 Avri Ben-Ze’ev,3 and signaling was not. Growth inhibition was associated with Nadir Arber1,2 induction of apoptosis. Chemotherapy synergistically enhanced the effect of AdTOP-PUMA. A combination of 1Integrated Cancer Prevention Center, Tel Aviv Sourasky the adenovirus system with standard therapy may 2 Medical Center; Sackler School of Medicine, Tel-Aviv University, improve the efficacy and reduce the toxicity of therapy Tel-Aviv, Israel; 3Department of Molecular Cell Biology, Weizmann Institute of Science, Rehovot, Israel in humans. [Mol Cancer Ther 2006;5(11):2861–71]

Abstract Introduction The adenomatous polyposis coli or b-catenin genes are h-Catenin is a multifunctional protein serving as a major frequently mutated in colorectal cancer cells, resulting in structural component of cell-to-cell adherens junctions. In oncogenic activation of B-catenin signaling. We tried to addition, it also acts as an important signaling molecule in establish in vitro and in vivo models for selectively the Wnt pathway that plays a key role in embryogenesis killing human cancer cells with an activated B-catenin/ and tumorigenesis (1–3). T-cell factor (Tcf) pathway. We used a recombinant In the absence of Wnt signaling, the cytoplasmic level of adenovirus that carries a lethal gene [p53-up-regulated h-catenin is kept low through interaction with a protein modulator of apoptosis (PUMA)] under the control of a complex [containing GSK3h-glycogen synthase kinase 3h, B-catenin/Tcf–responsive promoter (AdTOP-PUMA) to axin and adenomatous polyposis coli (APC)] that can selectively target human colorectal cancer cells phosphorylate h-catenin and target it to ubiquitin-mediated (SW480, HCT116, DLD-1, and LS174T), hepatocellular proteasomal degradation (4). Activation of Wnt signaling carcinoma (HepG2), and gastric cancer cells (AGS) in leads to inactivation of GSK3h, resulting in cytoplasmic which the B-catenin/Tcf pathway is activated, and accumulation of h-catenin (5). The increase in h-catenin compared its efficiency in killing cancer cells in which level is followed by its translocation into the nucleus, this pathway is inactive or only weakly active. AdFOP- where in complex with members of the T-cell factor (Tcf)/ PUMA, carrying a mutant Tcf-binding site, was used as lymphocyte enhancer–binding factor family of transcrip- control virus. The combined effect of AdTOP-PUMA with tion factors it activates the expression of target genes (6). several chemotherapeutic agents (5-florouracil, doxorubi- The APC tumor suppressor is mutated in f80% of the cin, and paclitaxel) was also evaluated. The effect of familial adenomatous polyposis syndrome and sporadic AdTOP-PUMA on colorectal cancer cells was also colorectal cancer patients (7). Loss of APC is believed to be examined in nude mice: SW480 cells were infected with one of the earliest initiating events in multistage colorectal the AdTOP-PUMA and AdFOP-PUMA, and then inoculat- carcinogenesis (8). Mutant APC loses its ability to direct h- ed s.c. into nude mice. The TOP-PUMA adenovirus catenin to degradation, resulting in nuclear accumulation inhibited cell growth in a dose-dependent fashion, and inappropriate activation of h-catenin–mediated trans- depending on the signaling activity of B-catenin. The activation. Mutations in h-catenin in the GSK-3h phosphor- growth of cells displaying high levels of active B-catenin/ ylation sites have been identified in 50% of colorectal Tcf signaling was inhibited after infection with AdTOP- cancer cases that retain wild-type APC (9–11). c-MYC (12) and cyclin D1 (13, 14), which positively regulate cell proliferation, are target genes of h-catenin/Tcf with direct implications in tumorigenesis (15–19). Activating h-catenin Received 3/6/06; revised 8/19/06; accepted 9/11/06. mutations have also been identified in a variety of other Grant support: Israel Cancer Association (N. Arber), Israel Science Foundation (A. Ben-Ze’ev), and the German-Israel Foundation for Scientific tumors, including melanomas (20), hepatocellular carcino- Research and Development (A. Ben-Ze’ev). mas (21), skin (22), breast (23), and prostate cancers (24), The costs of publication of this article were defrayed in part by the whereas the h-catenin–Tcf pathway is not activated in payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to most normal tissues. Therefore, a therapeutic strategy that indicate this fact. targets this pathway could be applied to patients with Note: This work was part of the requirements of Hadas Dvory-Sobol for primary or metastatic colorectal cancer. her Ph.D. degree at the Sackler School of Medicine at Tel Aviv University. p53-up-regulated modulator of apoptosis (PUMA) is a Requests for reprints: Nadir Arber, Director-Integrated Cancer Prevention potent mediator of the p53 apoptotic response (25, 26). Center, Tel-Aviv Medical Center, 6 Weizmann Street, Tel-Aviv 64239, Israel. Phone: 972-3-6974968; Fax: 972-3-6950339. It belongs to the group of BH3-only proteins that have E-mail: [email protected] or [email protected] been shown to function by dimerization with other BH3 Copyright C 2006 American Association for Cancer Research. domain–containing proteins, including Bcl-2 and Bcl-XL, doi:10.1158/1535-7163.MCT-06-0122 that results in the release of cytochrome c from

Mol Cancer Ther 2006;5(11). November 2006

mitochondria and induction of apoptosis by activation of the human PUMA cDNA fused to a double hemagglutinin- caspase-3 and caspase-9 (27). epitope tag) from pCEP4-PUMA (a generous gift from Bert A substantial limitation of conventional cancer chemo- Vogelstein, Johns Hopkins Oncology Center, Baltimore, therapy and radiotherapy is the toxicity of these agents to MD) was cloned downstream to the TOP/FOP elements normal tissue. The toxicity of currently available gene in the pAd-Track vector. The resultant plasmids were delivery systems to the normal cell population results from designated pAdTrack-TOP/FOP-PUMA. These shuttle their toxicity to the normal cell population. Here, we propose vectors were linearized with PmeI and cotransformed with a novel gene therapy approach that selectively expresses a E1-deleted adenoviral backbone AdEasy-1 into the compe- lethal gene by targeting the active h-catenin–Tcf pathway in tent bacterial strain BJ5183, which enables efficient recom- human colorectal, gastric, and hepatic cancer cells. More- bination. A panel of Ad-TOP-PUMA and Ad-FOP-PUMA over, we show that combining this strategy with standard recombinant adenoviruses were generated. chemotherapy results in a synergistic growth inhibition of Adenovirus Production and Titering colorectal cancer cells that overcomes cancer cell resistance To produce viruses, 4 Ag PacI-linearized adenoviral to therapy. This approach may pave the way to a novel DNA was transfected into 50% to 70% confluent 293 cells treatment of primary and metastatic colorectal cancer. in 10-cm dishes using LipofectAMINE and Plus Reagents (Invitrogen Life Technologies, Carlsbad, CA). Between 5 and 7 days posttransfection, colonies expressing GFP were Materials and Methods observed under a fluorescent microscope, the cells were Cell Culture harvested and lysed in PBS by four cycles of freeze/thaw/ Human colorectal cancer (SW480, DLD-1, HT-29, vortex (Fig. 1C). The supernatant was collected and half of HCT116, LS174T), gastric (AGS), hepatic (HepG2, it was used to reinfect 50% to 70% confluent 293 cells. SK-Hep-1), pancreatic (Colo357, Panc-1), and embryonic Viruses were collected 2 to 3 days postinfection when a kidney (293) cell lines were obtained from the American cytopathic effect became evident. Further amplification Type Culture Collection (Manassas, VA). They were and concentration of the virus stocks was achieved cultured in DMEM (Sigma, Rehovot, Israel) containing through several rounds of infection. To titer the viruses, 5% to 10% fetal bovine serum (Biological Industries, Beit 50% to 70% confluent 293 cells in 96-well dishes were Haemek, Israel), 1% penicillin, and 1% streptomycin, at infected with serial dilutions of the virus stocks. GFP- j 37 C, in an atmosphere of 95% oxygen and 5% CO2 positive colonies were counted 5 days postinfections. The (complete medium). control Ad-CMV-GFP adenovirus containing the GFP gene Construction of Plasmids and Adenoviral Vectors under the control of a full-length CMV promoter was a Two sets of h-catenin/Tcf–responsive promoters were generated: one contains wild-type Tcf/lymphocyte enhancer–binding factor binding sites fused with cFos (TOP-cFos- Luc-TOPFLASH) and the other, the SV40 (TOP-SV40-Luc) minimal promoter upstream to a luciferase (Luc) reporter gene. The corresponding control plasmids were constructed for each promoter by replacing the TOP oligomers with mutant Tcf-binding oligomers (FOP), e.g., FOP-cFos-Luc (FOPFLASH) and FOP-SV40-Luc (see the TOP and FOP sequences in Fig. 1A). To construct the TOP/FOP-cFos-Luc plasmids, an XbaI fragment containing the TOP-cFos, and the FOP-cFos from TOPFLASH and FOPFLASH plasmids (generous gifts from Hans Clevers, Utrecht University, Utrecht, the Netherlands), was cloned into the NheI site upstream to the Luc gene in the pGL3-basic plasmid (Promega, Rehovot, Israel). To construct the TOP/FOP- SV40-Luc plasmids, the BglII-NheI TOP and FOP fragments were cloned into the NheI and BglII sites upstream of the SV40 minimal promoter in the pGL3-promoter plasmid (Promega). The AdEasy system (28) was used to generate the AdTOP-PUMA and AdFOP-PUMA (AdTOP/FOP-PUMA) adenoviruses. The TOP and FOP sequences were obtained Figure 1. Construction of adenoviruses. A, h-catenin/Tcf activatable from the TOP-cFos-Luc and FOP-cFos-Luc plasmids and promoters containing TOP or FOP sequences. B, schematic representa- cloned into the shuttle vector-pAd-Track. The pAdTrack tion of different adenovirus constructs. C, the PacI-digested recombi- also contains a green fluorescent protein (GFP) gene under nant adenoviral vector pAdTOP-PUMA was transfected into 293 cells and GFP expression was visualized by fluorescence microscopy at the the control of the cytomegalovirus (CMV) promoter indicated times. Comet-like adenovirus – producing foci became apparent (Fig. 1B). The blunted HA-PUMA fragment (containing after 5 to 7 d.

Mol Cancer Ther 2006;5(11). November 2006

kind gift of Hila Giladi (Hadassah School of Medicine, Western Blot Analysis Jerusalem, Israel) and was amplified in 293 cells. Infected cells were harvested and protein concentrations Human cDNA of caspase-8, Bak, and Bax were a kind gift were determined using the Bio-Rad protein assay kit (Bio- from Atan Gross (Weizmann Institute of Science, Rehovot, Rad, Hercules, CA). An equal amount of protein from Israel). The expression constructs of PUMA and PUMA- each lysate was analyzed by SDS-PAGE and the proteins DBH3, a mutant PUMA without activity, were generous were transferred to hybond-C extra nitrocellulose mem- gifts from Bert Vogelstein. PKGIh that encodes a mutant branes (Amersham Life Science, Buckinghamshire, United PKG sequence with an NH2-terminal truncation (29), was a Kingdom). Membranes were blocked with buffer contain- gift from I. Bernard Weinstein (Columbia University, New ing 5% low-fat milk and 0.05% Tween 20 in PBS for 1 hour, York, NY). This deletion renders PKG independent of incubated with primary antibodies for 1 hour with cyclic guanosine 3¶,5¶-monophosphate, and it is constitu- peroxidase-conjugated secondary antibodies, and devel- tively active. oped with a Supersignal West Pico chemiluminescent Luciferase Assays substrate (Pierce, Rockford, IL). Antibodies against hemag- Transfections were done using LipofectAMINE and Plus glutinin, actin, h-catenin, and caspase-3 were purchased Reagents (Invitrogen) according to the instructions from from Santa Cruz Biotechnology (Santa Cruz, CA). the manufacturer. A total of 5 Â 105 cells were seeded in Apoptosis Analysis six-well plates. The next day, 50% confluent dishes were Flow Cytometry. Cells were plated at 5 Â 106/10-cm cotransfected with 1 Ag vectors plus 0.1 Ag pRL-TK dish 24 hours before infection, and were infected with (Promega). Luc assay was done 24 hours posttransfection. recombinant adenoviral vectors at 5 MOI. Twenty-four Briefly, cells were washed once with PBS and lysed in hours later, both adherent and floating cells were 400 AL lysis buffer for 15 minutes at room temperature. harvested, washed with PBS, and fixed in 80% ethanol The lysates were centrifuged at 14,000 rpm for 5 minutes, for 1 hour and stained with propidium iodide for anal- and 20 AL of each lysate were used to measure Luc ysis of DNA content. The number of subdiploid cells, reporter gene expression. Luc activity was normalized to representing apoptotic cells, was quantified by FACScan Renilla Luc activity from a parallel cotransfection of pRL- using the CellQuest software (Becton Dickinson Immuno- TK (Dual Luc system, Promega). All experiments were cytometry Systems, San Jose, CA). Necrotic cells were done in triplicate at least thrice and gave similar results. excluded by staining with trypan blue. The average of at The Potency of PUMA in Cell Killing least three independent experiments with two replicates SW480 cells were transfected with constructs encoding was recorded. for PUMA or a mutant PUMA-DBH3, without activity. ssDNA. For the ssDNA assay, 104 cells were seeded in Cells were harvested 24 hours after transfection, and an 96-well microplates, and after 24 hours infected with equal number of cells was diluted in duplicates (1:10 and recombinant adenoviral vectors at 5 MOI. The following 1:25) into 10-cm dishes and grown under hygromycin B day, the ssDNA Apoptosis ELISA kit was used (Chemicon selection for 3 weeks, after which the cells were fixed and International, Inc., Temecula, CA). Based on the selective, stained with 0.2% Coomassie blue, 50% methanol, 10% formamide-induced denaturation of DNA, this method acetic acid, and 40% H2O. identifies apoptotic cells (30), by staining ssDNA using a Cell Viability Assays mixture of anti-ssDNA monoclonal antibody and peroxi- Between 2 Â 104 and 5 Â 104 cells in 100 AL complete dase-conjugated anti-mouse IgM. The average of at least medium were plated in 96-well dishes. The next day, six two independent experiments with two replicates was wells were infected with each adenovirus at a different recorded. multiplicity of infection (0.1–50 MOI). Cell viability was Fluorogenic assay for caspase-3 activity. assessed by methylene blue staining after 48 hours. The cells Cells were plated at 1 Â 106 per well onto a six-well were washed once with PBS and fixed in 150 AL plate 24 hours before infection, and were infected with formaldehyde (4%) for 2 hours at room temperature, recombinant adenoviral vectors at 5 MOI. Cells were washed with 0.1 mol/L sodium borate (pH 8.5), and stained harvested with a rubber policeman; washed; resuspended with 0.5% methylene blue for 10 minutes, then washed with in 50 mmol/L Tris-HCl buffer (pH 7.4), 1 mmol/L EDTA, tap water and 150 AL of 0.1 mol/L HCl to dilute the cell- and 10 mmol/L EGTA; and lysed by three successive bound dye. Absorbance was measured at 590 nm. Cell freeze-thaw cycles on dry ice. Cell lysates were centrifuged viability is expressed as percentage absorbance relative to at 20,000 Â g for 5 minutes, and the supernatants were mock-infected cells. The average of at least two independent stored at À70jC. The protein concentration of each sample experiments with six replicates was recorded. was determined using the Bradford Bio-Rad protein assay. Chemicals For caspase-3 activity, a total of 50 Agproteinwas Paclitaxel, doxorubicin, and 5-florouracil were obtained incubated with 50 mmol/L ac-DEVD-AMC (from BIOMOL from Sigma. Cells were infected with AdTOP-PUMA, Research Laboratories, Plymouth Meeting, PA) at 37jC, for AdFOP-PUMA, or Ad-CMV-GFP (5 MOI); after 5 hours 30 minutes in the dark. The release of 7-amino-4-methyl- were treated with 0.05 Amol/L paclitaxel, 1 Amol/L coumarine was monitored by a spectrofluorometer using doxorubicin, or 0.05 Amol/L 5-florouracil; and were then an excitation wavelength of 360 nm and an emission cultured for 48 hours. wavelength of 460 nm.

Mol Cancer Ther 2006;5(11). November 2006

Figure 2. PUMA suppresses the growth of colon cancer cells. A, comparison of h-catenin levels in different cell types. Cell extracts were prepared from colorectal cancer lines (SW480, HCT116, DLD-1, and HT-29). Twenty micrograms of protein from each sample were subjected to Western blot analysis using anti-h-catenin and anti-h-actin antibodies. B, 293 cells were transiently transfected with human cDNA of caspase-8, Bid, Bak, Bax, PUMA, PKGIh, and pcDNA3. Forty-eight hours after transfection, the percentage of apoptotic (sub-G1) cells was determined by FACSanalysis. *, P < 0.05, significantly different from control. **, P < 0.01, significantly different from control. C, SW480 cell lines were transfected with constructs encoding PUMA and PUMA-DBH3. Cells were harvested 24 h after transfection, and equal cell numbers were diluted in duplicates in 10-cm dishes and grown under selection in hygromycin B for 3 wks, then fixed and stained with Coomassie blue. Colony formation in representative dishes (1:10 and 1:25 dilution) of transfected SW480 cells is shown. D, the number of colonies formed with cells infected with PUMA and PUMA-DBH3 is shown. *, P < 0.05, significantly different from PUMA-DBH3.

Tumorigenic Assays Results CD1 nude mice housed in sterile cages were handled B-Catenin/Tcf ^ Mediated Luc Activity in Different under aseptic conditions. The animals were maintained in Colorectal Cancer Cells facilities approved by the Israeli Association for Accredi- h-Catenin/Tcf–dependent activity was determined in tation of Laboratory Animal Care and in accordance with human cell lines displaying different levels of h-catenin current regulations and the standard of care of the Israeli (Fig. 2A). These cell lines harbored mutant APC proteins, Ministry of Health. SW480 colorectal cancer cells were except for HCT116 that have a deletion at residue S45 of infected with 5 MOI of adenoviral vectors in serum-free the h-catenin protein (11). To evaluate which h-catenin/Tcf medium. Five hours after infection, the medium was reporter construct is more readily detectable in these cells, changed to 10% fetal bovine serum medium and the the h-catenin/Tcf–responsive promoters fused to SV40 infected cells were incubated at 37jC overnight. Twenty- and cFos minimal promoters were used using the luci- four hours after adding the virus, the cells were trypsinized ferase (Luc) assay. h-Catenin–activated promoters contain- and inoculated s.c. on the backs of nude mice (3 Â 106 ing four copies of the Tcf-binding site (TOP) fused to either per animal). The number of mice with tumors (incidence) the cFos or the minimal SV40 promoter were analyzed. and their volume (tumor burden) were determined after The TOP-cFos (TOPFLASH) construct exhibited higher different times. activity than TOP-SV40 and was therefore used in the Statistical Analysis following experiments. The relative activity of TOPFLASH Statistical analysis was done by using InStat software is shown in Table 1. Luc activity, determined as fold version 3.01 (GraphPad Software, Inc., San Diego, CA). In induction of TOPFLASH, was 4.1- to 25.3-fold higher than the tissue culture experiments, the comparison between that of the control FOPFLASH in colorectal cancer cells two samples was done using Student’s t test and between (SW480, DLD-1, HCT116, and LS174T), whereas HT-29 more than two samples using one-way ordinary parametric cells did not show significant transcriptional activation of ANOVA followed by Tukey-Kramer multiple comparison this reporter construct (Table 1). test. For all statistical tests, preliminary evaluation of the PUMA Induces Cell Death in SW480 Cells homoscadacity and normality of the compared samples To identify the most potent proapoptotic gene in was done using Bartlett and Kolmogorov-Smirnov tests, colorectal cancer cells, we tested several full-length cDNAs, respectively. including those encoding for Bak, caspase-8, PUMA,

Mol Cancer Ther 2006;5(11). November 2006

Table 1. B-catenin/Tcf–regulated transcription activity

Cell line Fold Luc activity (TOP/FOP) Cell line Fold Luc activity (TOP/FOP)

SW480 25.3 HepG2 63.5 HCT116 4.1 SK-Hep-1 1.6 DLD-1 4.8 AGS 5.5 HT-29 1.5 Colo357 1.1 LS174T 12.9 Panc-1 0.9

NOTE: Luc activities were assayed after 24 hours and plotted as fold activation of TOP-cFos-Luc relative to that of FOP-cFos-Luc.

PKGIh, Bax, and Bid that we transfected into 293 cells. B –resistant gene (pCEP4-PUMA) was transfected into Forty-eight hours after transfection, the number of sub- SW480 cells. PUMA-DBH3 encoding for a nonfunctional diploid DNA-containing cells, representing apoptotic cells, PUMA (without its BH3 domain) was used as control. The was quantified by FACScan (Fig. 2B). PUMA and Bax results shown in Fig. 2C and D show a drastic reduction in induced the highest apoptotic activity. We have chosen colony formation by cells after transfection with PUMA PUMA because it was more effective in SW480 cells, which compared with the mutant PUMA vector. we used in the following studies (Fig. 2C). Next, we used an adenoviral vector selected for gene To determine the effect of PUMA expression on colon delivery with PUMA placed downstream to the cFos cancer cell growth, an expression vector containing PUMA minimal promoter. The promoter contained either the under the control of the CMV promoter and a hygromycin wild-type (AdTOP-PUMA) or the mutant (AdFOP-PUMA)

Figure 3. AdTOP-PUMA suppresses the survival of colon cancer cells. A, SW480, DLD-1, HCT116, and HT-29 cells were infected with AdTOP-PUMA, AdFOP-PUMA, and Ad-CMV-GFP adenoviruses in 96-well culture plates. Cell viability expressed as percentage absorbance relative to mock-infected cells was measured by methylene blue staining 48 h after adenoviral infection. Average of at least two independent experiments with six replicates. Statistical difference was observed between AdTOP-PUMA and the control groups (AdFOP-PUMA and Ad-CMV-GFP; ***P < 0.001; **P < 0.01). B, the number of viable cells is proportional to the intensity of methylene blue staining shown here for SW480 cells 48 h after adenoviral infection. C, SW480 cells 48 h following infection with either AdFOP-PUMA (left) or AdTOP-PUMA (right). AdTOP-PUMA and AdFOP-PUMA also contain a GFP gene under the control of a CMV promoter. GFP expression was visualized by fluorescence microscopy. D, HT-29 and DLD-1 cells were infected (at 5 MOI) with either AdFOP-PUMA (left) or AdTOP-PUMA (right) and GFP expression was visualized by fluorescence microscopy after 48 h.

Mol Cancer Ther 2006;5(11). November 2006

Tcf/lymphocyte enhancer–binding factor binding sites. apoptotic cells) 24 hours after treatment with AdFOP- The Ad-CMV-GFP vector was used as control for viral PUMA or Ad-CMV-GFP (Fig. 4A). Apoptosis induced by toxicity. The ability of AdTOP-PUMA and AdFOP-PUMA AdTOP-PUMA was detected in SW480, DLD-1, and adenoviral vectors to kill cells with different levels of HCT116 cells 48 hours after infection, by the ssDNA assay, h-catenin signaling was evaluated by cell viability assays but not in HT-29 cells (Fig. 4B). The number of apoptotic 48 hours after infection with adenoviruses at varying cells was proportional to the increase in methylene blue doses (Fig. 3A). Cells displaying elevated h-catenin tran- color intensity as shown for SW480 cells infected with sactivation (Table 1), such as SW480, HCT116, and DLD-1, adenoviral constructs and stained with this dye after were killed efficiently by infection with AdTOP-PUMA, 48 hours (Fig. 4C). Taken together, the results of these in a dose-dependent manner. Although the infection different approaches suggest that AdTOP-PUMA is capa- efficiency of the adenoviral vectors in the human colorectal ble of inducing the death of cells that have elevated h- cancer cell line HT-29 was high (shown by the number of catenin/Tcf transcriptional activity. GFP positive cells in Fig. 3D), neither AdTOP-PUMA nor Activated B-Catenin/Tcf Signaling Induces Apoptosis AdFOP-PUMA caused cell death in this cell line (Fig. 3A), by Up-regulating PUMA Expression most probably because the level of h-catenin/Tcf trans- In SW480, HCT116, and DLD-1 cells, high levels of activation in these cells is low (Table 1). The number PUMA protein were detected after infection with AdTOP- of viable cells after infection with AdTOP-PUMA was PUMA, but not after AdFOP-PUMA or Ad-CMV-GFP proportional to the methylene blue color intensity as shown infection (Fig. 5A). HT29 colorectal cancer cells that express for SW480 cells infected with adenoviral constructs lower levels of h-catenin and low levels of h-catenin–Tcf (Fig. 3B). Representative pictures of SW480, DLD-1, and transactivation did not express the PUMA protein after HT-29 cells, 48 hours after infection with either AdTOP- infection with AdTOP-PUMA (Fig. 5A). The infection of PUMA or AdFOP-PUMA, are shown in Fig. 3C. SW480 cells with AdTOP-PUMA induced PUMA expres- Next, the number of cells in the sub-G1 population was sion as early as 12 hours after infection, as shown in determined by fluorescence-activated cell sorting (FACS) Fig. 5B. Induction of apoptosis by AdTOP-PUMA was analysis in the different colorectal cancer cell lines infected apparently mediated by the activation of caspases. Western with the various adenovirus constructs. By this method, blot analysis of SW480 cells after infection with AdTOP- SW480 and HCT116 cells displayed only a very low PUMA, but not with AdFOP-PUMA or Ad-CMV-PUMA, number of cells in the sub-G1 population (representing detected a cleaved form of caspases-3 (p17; Fig. 5C). Also,

Figure 4. Cell killing by AdTOP-PUMA adenovirus. A, SW480 and HCT116 cells were infected with AdTOP-PUMA, AdFOP-PUMA, or Ad-CMV-GFP adenovirus constructs at 5 MOI. Percentage apoptotic (sub-G1) cells was determined by FACSanalysis 24 h after treatment. Significantlydifferent from the control groups (no virus, AdFOP-PUMA, Ad-CMV-GFP; *P < 0.05; **P < 0.01). B, induction of apoptosis determined by the apoptosis ELISA kit described in Materials and Methods in cultures grown in 96-well plates and infected with adenoviruses for 48 h. *, significantly different from the control groups (no virus, P < 0.01; AdFOP-PUMA, P < 0.05; Ad-CMV-GFP, P < 0.01). #, significantly different from the control groups (no virus, P < 0.01; AdFOP-PUMA, P < 0.05; Ad-CMV-GFP, P < 0.05). &, significantly different from the control groups (no virus, P < 0.001; AdFOP-PUMA, P < 0.001; Ad-CMV-GFP, P < 0.001). C, color intensity is proportional the number of apoptotic SW480 cells 48 h after adenoviral infection.

Mol Cancer Ther 2006;5(11). November 2006

Figure 5. Up-regulation of PUMA after infection of cells with AdTOP-PUMA adenovirus. A, SW480, HCT116, DLD-1, and HT-29 cells were infected with AdTOP-PUMA, AdFOP-PUMA, or Ad-CMV-GFP (Ad-GFP) adenoviruses. Cells were harvested after 48 h and the lysates, normalized for protein concentration, were analyzed by Western blotting using antihemagglutinin (for PUMA detection) and anti-h-actin antibodies. Cells were treated with PBS (left lane in each blot). B, immunoblot analysis of PUMA protein in SW480 cells 12, 24, 36, and 48 h postinfection with AdTOP-PUMA (5 MOI). C, Western blot analysis of caspase-3 activation in SW480 cells 24 h after infection. The antibody detected both the procaspase-3 (32 kDa) and the cleaved fragment of caspase-3 (17 kDa). caspase-3 activation occurred 48 hours following infection basal levels of h-catenin/Tcf activity. Next, we examined with AdTOP-PUMA, but not after infection with control the effect of Ad-PUMA constructs on these cells. As shown viruses (AdFOP-PUMA or Ad-CMV-PUMA; Table 2). in Fig. 6A, AGS, HepG2, and LS174T were efficiently killed AdTOP-PUMA Inhibits Tumor Growth In vivo by AdTOP-PUMA infection, whereas Colo357 and Panc-1 Next, we evaluated whether AdTOP-PUMA infection can cells were resistant, despite the high adenovirus infection induce a cell killing effect in s.c. tumors established in nude efficiency in these cells (Fig. 6A and B). mice with colorectal cancer cells. SW480 cells were infected AdTOP-PUMA and Chemotherapy Synergistically with adenoviruses in vitro, harvested after 24 hours, and Induce Tumor Cell Death then inoculated s.c. into nude mice. As shown in Table 3, We investigated whether the combination of AdTOP- mice that were injected with SW480 cells that were infected PUMA adenovirus and chemotherapeutic agents more with AdTOP-PUMA virus failed to develop tumors. In contrast, AdFOP-PUMA and Ad-CMV-GFP did not suppress the growth of the implanted SW480 cells. Table 2. Caspase-3is activated by the AdTOP-PUMA construct B-Catenin Activates Expression of PUMA from Wild-Type Tcf/Lymphocyte Enhancer ^ Binding Factor ^ Construct Caspase activity Responsive Constructs in Other Tumor Cells F The effect of AdTOP-PUMA was determined in other AdTOP-PUMA 38,070 193 F tumor cell lines with hyperactive h-catenin/Tcf signaling. AdFOP-PUMA 9,342 1744 AdCMV-GFP 14,515 F 409 Luc activity was analyzed in AGS (gastric cancer), HepG2 No treatment 11,363 F 317 (hepatocellular carcinoma), and LS174T (a colorectal cancer) cell lines. The results (Table 1) show that these NOTE: SW480 cells (1 Â 106 per well) were seeded onto a six-well plate. cell lines displayed significant levels of h-catenin/Tcf Twenty-four hours later, 5 MOI of the various viruses were added to the medium for 48 hours. Caspase-3 activity was measured using the activity. In contrast, pancreatic cells (Colo357 and Panc-1) fluorometric assay described in Materials and Methods. The numbers and SK-Hep-1 (hepatocellular carcinoma) showed only represent the fluorometric signal F SD (P < 0.01).

Mol Cancer Ther 2006;5(11). November 2006

Table 3. AdTOP-PUMA suppresses the growth of tumors formed Adenoviral systems are easy to produce and they have B by SW480 cells displaying hyperactive -catenin a highly effective nuclear entry mechanism and display very low pathogenicity in humans (36). The adenoviral No. mice Adenovirus Mice with tumor vectors can transduce cells in vivo and they do not integrate 8 AdTOP-PUMA 0* (0%) into the host cell genome (36). Mulvihill et al. (37) 8 AdFOP-PUMA 3 (37.5%) conducted a phase I clinical trial of adenoviral administra- 7 Ad-CMV-PUMA 4 (57.1%) tion using ONYX-015 that consists of an adenovirus expressing a deletion mutant of E1B-55 that selectively *Number of mice with s.c. tumors after 8 weeks. replicates in and lyses tumor cells displaying a mutant p53. The virus was injected through hepatic arterial catheters efficiently induce apoptosis in colorectal cancer cells than to patients with colorectal cancer liver metastases, and was each treatment alone. HCT116, SW480, and HT29 cells well tolerated at doses up to 1011 plaque-forming units. A were infected with either AdTOP-PUMA, AdFOP-PUMA, phase I clinical trial using the Escherichia coli cytosine or Ad-CMV-GFP adenoviruses (at 5 MOI) and cultured for deaminase for the treatment of metastatic colon cancer is 48 hours in the presence or absence of paclitaxel (0.05 currently being conducted (38). Amol/L), doxorubicin (1 Amol/L), or 5-florouracil (0.05 In the present study, we show that targeting h-catenin/ A mol/L). AdTOP-PUMA adenovirus dramatically en- Tcf responsive transcription can selectively and efficiently hanced the killing effect by all three chemotherapeutic kill tumor cells displaying high levels of h-catenin/Tcf agents in both SW480 and HCT116 cells, but not in HT29 transactivation, with minimal toxicity to cells displaying cells (Fig. 7). The efficacy of the chemotherapeutic agents low levels of h-catenin–Tcf signaling. Similar strategies was not augmented by exposure of the different cells to were used and recently reported by others (39, 40). In Chen either Ad-CMV-GFP or AdFOP-PUMA adenovirus con- and McCormick’s work (39), an adenoviral vector, AdWt- structs (Fig. 7). Fd, containing the thymidine kinase promoter carrying the proapoptotic gene Fadd, selectively killed colorectal cancer Discussion cells in vitro. Kwong et al. (40) used an in vitro-in vivo Although gene therapy–based clinical trials showed animal model similar to the one used in the present study. significant success (31–33) in a variety of human tumors, They showed selective killing of DLD-1 colorectal cancer thus far they failed to show a significant therapeutic effect cells in an ex vivo animal model, by the adenoviral vector in colorectal cancer. By contrast, in cell lines and in AdTOP-CMV-TK that contains a herpes simplex virus experimental animals, remarkable results, including com- thymidine kinase gene under the control of a h-catenin/Tcf– plete regression of the tumor, have been shown by this responsive promoter linked to a minimal CMV promoter. approach (34, 35). The lack of efficacy of this approach in Lipinski et al. (41) optimized the activity and specificity humans might result from the poor targeting selectivity of profile of a synthetic catenin-dependent promoter by the vectors that could lead to low levels of expression varying its basal promoter, the number of Tcf-binding sites, of the transferred gene in tumor cells and to high toxicity and the distance between them and the basal promoter. The of the gene product in normal cells. optimal promoter showed virtually undetectable expression

Figure 6. The effect of AdTOP-PUMA in different tumor cell lines. A, AGS(gastric cancer), HepG2 (hepatocellular carcinoma), LS174T(colon cancer), and Colo357 and Panc-1 (pancreatic cancer) cells were infected with AdTOP-PUMA, AdFOP-PUMA, and Ad-CMV-GFP viruses. Cell viability was measured 48 h after adenoviral infection by methylene blue staining. Cell viability is expressed as percentage absorbance relative to mock-infected cells. Average of at least two independent experiments with six replicates. Statistical difference was observed between AdTOP-PUMA and the control groups (AdFOP- PUMA and Ad-CMV-GFP) at 5, 10, and 25 MOI (***P < 0.001, **P < 0.01, *P < 0.05). B, Colo357 and Panc-1 cells were infected (at 5 MOI) with either AdFOP-PUMA (left) or AdTOP-PUMA (right), and GFP expression was visualized by fluorescence microscopy after 48 h.

Mol Cancer Ther 2006;5(11). November 2006

Figure 7. Combination effect of AdTOP-PUMA and chemotherapeutic agents in colon cancer cells. Cells were infected with the different viruses (5 MOI) and treated with various anticancer agents [paclitaxel, doxorubicin, and 5-florouracil (5-FU)]. One hundred percent indicates the viability of cells treated with PBS, without any agent or viral vector. Cell viability was determined 48 h after treatment with the various agents. Columns, mean of two quadruplicate assays; bars, SD. **, P < 0.01, significantly different from the AdTOP-PUMA (virus alone). in cells with normal h-catenin regulation, but displayed and gastric cancer cells (AGS that harbor activating high activity in cells expressing deregulated h-catenin. h-catenin mutations; refs. 21, 47). We found that both cell Malerba et al. (42) inserted Tcf-binding sites into the viral P4 lines were very sensitive to treatment with the AdTOP- promoter and showed that reduction of the number of Tcf PUMA adenovirus. In particular, HepG2 cells were sites from four to two leads to an increase in the efficiency of 10 times more sensitive than the other cell lines, including replication and toxicity of the virus in Co115 colon cancer SW480 cells that contain a very potent transactivating h- cells. catenin/Tcf complex (Fig. 6). In contrast, pancreatic cancer The current study took this strategy further in several cell lines (Colo357 and Panc-1) that do not possess an aspects: (1) Several genes known to be associated with active h-catenin signaling pathway (48–51) were not apoptosis induction (Bax, Bak, Bid, and caspase-8) were affected by this therapy. Of special note is that although evaluated for their apoptotic effect. Among these, we SW480 cells have the highest h-catenin/Tcf activity found that PUMA, a critical mediator of p53-dependent (Table 1), the number of apoptotic SW480 cells is lower apoptosis following DNA damage (25, 26), was the most after 24 hours (Fig. 4A) than in HCT116 cells that have effective one. There are only two more recent reports weaker h-catenin/Tcf transactivation. This suggests that demonstrating the therapeutic value of PUMA: one in the effect of AdTOP-PUMA virus most probably depends malignant glioma (43) and another in esophageal carcino- on additional regulatory mechanism(s) in the cell besides ma cells (44). (2) Many cancers harbor mutations in the transactivation by h-catenin/Tcf, and different cancer cells p53 gene. Adenovirus-mediated p53 gene transfer has display varying sensitivities to different killing agents been extensively studied as a plausible novel gene therapy as was shown in Fig. 7. strategy in various cancers. However, a variety of cancer A proof of concept of this gene therapy strategy was cells are resistant to this novel therapy (45, 46). We shown in the in vitro-in vivo experiments. Indeed, s.c. hypothesized that the PUMA gene located downstream to injection of SW480 cells infected with AdTOP-PUMA did p53 pathways may serve as a good alternative strategy to not produce tumors in any of the animals, whereas p53 gene therapy as it can kill cancer cells directly. (3) In injection of SW480 cells infected with AdFOP-PUMA or this study, PUMA expression was induced in four human Ad-CMV-GFP induced tumors in the majority of nude colorectal cancer cell lines (SW480, HCT116, DLD-1, and mice. This confirms that this strategy could be useful in LS174T) that display hyperactive h-catenin/Tcf signaling. targeting human cancer cells. The induction of PUMA expression in these cells using the We have also shown that a combination of the gene TOP, but not the FOP, construct resulted in significant cell therapy approach with chemotherapeutic agents that have killing. PUMA expression was not induced in cells with a distinct mechanism of action might be an effective route low or undetectable h-catenin signaling (HT-29). (4) in achieving a better antitumor response. Thus, the Hyperactivity of h-catenin signaling is not unique to combination of AdTOP-PUMA adenovirus and paclitaxel, colorectal cancer cells. Hence, the effect of these viruses doxorubicin, and 5-florouracil dramatically enhanced the was also evaluated in hepatocellular carcinoma (HepG2) cell killing effect. This strategy could be particularly useful

Mol Cancer Ther 2006;5(11). November 2006

in the treatment of chemotherapy-resistant colorectal 14. Shtutman M, Zhurinsky J, Simcha I, et al. The cyclin D1 gene is a target of the h-catenin/LEF-1 pathway. Proc Natl Acad Sci U S A 1999;96: cancer (52) because it might minimize the toxicity of this 5522 – 7. regimen. Adenoviruses practically infect all types of cells 15. Sikora K, Chan S, Evan G, et al. c-Myc oncogene expression in in the body and, therefore, PUMA will probably be colorectal cancer. Cancer 1987;59:1289 – 95. activated only in cells with hyperactive h-catenin signaling. 16. Erisman MD, Scott JK, Watt RA, Astrin SM. The c-myc protein is Hence, in the clinical setting, combination therapy using constitutively expressed at elevated levels in colorectal carcinoma cell lines. Oncogene 1998;2:367 – 78. adenovirus and chemotherapeutic agents would be suit- 17. Finley GG, Schulz NT, Hill SA, Geiser JR, Pipas JM, Meisler AI. able not only for the metastatic disease, but also at stages II Expression of the myc gene family in different stages of human colorectal and III of the disease, aiming to eradicate microscopic cancer. Oncogene 1989;4:963 – 71. residual tumor cells. 18. Imaseki H, Hayashi H, Taira M, et al. Expression of c-myc oncogene in Recently, we showed that a Ras-responsive element colorectal polyps as a biological marker for monitoring malignant potential. Cancer 1989;64:704 – 9. selectively kills tumor cells without affecting the growth 19. Arber N, Hibshoosh H, Moss SF, et al. Increased expression of cyclin and survival of normal cells (53). Introduction of cell death D1 is an early event in multistage colorectal carcinogenesis. Gastroenter- genes (bax, caspase-8, and mutant PKG) under the control of ology 1996;110:669 – 74. a promoter containing wild-type Ets and activator protein-1 20. Rubinfeld B, Robbins P, El-Gamil M, Albert I, Porfiri E, Polakis P. binding sites resulted in preferential killing of cells dis- Stabilization of h-catenin by genetic defects in melanoma cell lines. Science 1997;275:1790 – 2. playing a mutant, but not a normal, Ras pathway. h 21. de La Coste A, Romagnolo B, Billuart P, et al. Somatic mutations of In conclusion, this strategy involving an active -catenin/ the h-catenin gene are frequent in mouse and human hepatocellular Tcf pathway is not a cell- or organ-specific approach, but carcinomas. Proc Natl Acad Sci U S A 1998;95:8847 – 51. could be applied to a much wider range of cases because 22. Chan EF, Gat U, McNiff JM, Fuchs E. A common human skin tumour hyperactive h-catenin/Tcf signaling is found in a significant is caused by activating mutations in h-catenin. Nat Genet 1999;21: 410 – 3. percentage of almost all types of cancer (2, 54). The results 23. Lin SY, Xia W, Wang JC, et al. h-Catenin, a novel prognostic marker of the current study may pave the way to new combined for breast cancer: its roles in cyclin D1 expression and cancer progression. approaches with chemotherapy in the setting of all stages Proc Natl Acad Sci U S A 2000;97:4262 – 6. of colorectal cancer with activated h-catenin signaling. 24. Voeller HJ, Truica CI, Gelmann EP. h-Catenin mutations in human prostate cancer. Cancer Res 1998;58:2520 – 3. Acknowledgments 25. Yu J, Zhang L, Hwang PM, Kinzler KW, Vogelstein B. PUMA We thank Moshe Oren (Weizmann Institute of Science, Israel) for helpful induces the rapid apoptosis of colorectal cancer cells. Mol Cell 2001;7: advice and thorough discussions, Hila Giladi (Hebrew University, Jerusa- 673 – 82. lem, Israel) for assistance with adenovirus production, and Naham Kariv 26. Nakano K, Vousden KH. PUMA, a novel proapoptotic gene, is induced (Tel Aviv University, Israel) for assistance with the animal studies. by p53. Mol Cell 2001;7:683 – 94. 27. Green DR, Reed JC. Mitochondria and apoptosis. Science 1998;281: References 1309 – 12. 1. Aberle H, Schwartz H, Kemler R. Cadherin-catenin complex: protein 28. He TC, Zhou S, da Costa LT, Yu J, Kinzler KW, Vogelstein B. interactions and their implications for cadherin function. J Cell Biochem A simplified system for generating recombinant adenoviruses. Proc Natl 1996;61:514 – 23. Acad Sci U S A 1998;95:2509 – 14. 2. Conacci-Sorrell M, Zhurinsky J, Ben-Ze’ev A. The cadherin-catenin 29. Soh JW, Mao Y, Liu L, Thompson WJ, Pamukcu R, Weinstein IB. adhesion system in signaling and cancer. J Clin Invest 2002;109: Protein kinase G activates the JNK1 pathway via phosphorylation of 987 – 91. MEKK1. J Biol Chem 2001;276:16406 – 10. 3. Wodarz A, Nusse R. Mechanisms of Wnt signaling in development. 30. Frankfurt OS, Krishan A. Identification of apoptotic cells by Annu Rev Cell Dev Biol 1998;14:59 – 88. formamide-induced DNA denaturation in condensed chromatin. J Histo- 4. Polakis P. Wnt signaling and cancer. Genes Dev 2002;14:1837 – 51. chem Cytochem 2001;49:369 – 78. 5. Eastman Q, Grosschedl R. Regulation of LEF-1/Tcf transcription factors 31. Nemunaitis J, Swisher SG, Timmons T, et al. Adenovirus-mediated by Wnt and other signals. Curr Opin Cell Biol 1999;11:233 – 40. p53 gene transfer in sequence with cisplatin to tumors of patients with non-small-cell lung cancer. J Clin Oncol 2000;18:609 – 22. 6. Behrens J, von Kries JP, Kuhl M, et al. Functional interaction of h-catenin with the transcription factor LEF-1. Nature 1996;382: 32. Buller RE, Runnebaum IB, Karlan BY, et al. A phase I/II trial of rAd/p53 638 – 42. (SCH 58500) gene replacement in recurrent ovarian cancer. Cancer Gene Ther 2002;9:553 – 66. 7. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990;61:759 – 67. 33. Swisher SG, Roth JA, Komaki R, et al. Induction of p53-regulated genes and tumor regression in lung cancer patients after intratumoral 8. Kinzler KW, Vogelstein B. Lessons from hereditary colorectal cancer. delivery of adenoviral p53 (INGN 201) and radiation therapy. Clin Cancer Cell 1996;87:159 – 70. Res 2003;9:93 – 101. 9. Morin PJ, Sparks AB, Korinek V, et al. Activation of h-catenin-Tcf h 34. Mullen JT, Donahue JM, Chandrasekhar S, et al. Oncolysis by viral signaling in colon cancer by mutations in -catenin or APC. Science 1997; replication and inhibition of angiogenesis by a replication-conditional 275:1787 – 90. herpes simplex virus that expresses mouse endostatin. Cancer 2004;101: 10. Sparks AB, Morin PJ, Vogelstein B, Kinzler KW. Mutational analysis of 869 – 77. h the APC/ -catenin/Tcf pathway in colorectal cancer. Cancer Res 1998;58: 35. Qin XQ, Tao N, Dergay A, et al. Interferon-h gene therapy inhibits 1130 – 4. tumor formation and causes regression of established tumors in immune- 11. Ilyas M, Tomlinson IP, Rowan A, Pignatelli M, Bodmer WF. h-Catenin deficient mice. Proc Natl Acad Sci U S A 1998;95:14411 – 6. mutations in cell lines established from human colorectal cancers. Proc 36. Volpers C, Kochanek S. Adenoviral vectors for gene transfer and Natl Acad Sci U S A 1997;94:10330 – 4. therapy. J Gene Med 2004;6 Suppl 1:S164 – 71. 12. He TC, Sparks AB, Rago C, et al. Identification of c-MYC as a target of 37. Mulvihill S, Warren R, Venook A, et al. Safety and feasibility of the APC pathway. Science 1998;281:1509 – 12. injection with an E1B-55 kDa gene-deleted, replication-selective adenovi- 13. Tetsu O, McCormick F. h-Catenin regulates expression of cyclin D1 in rus (ONYX-015) into primary carcinomas of the pancreas: a phase I trial. colon carcinoma cells. Nature 1999;398:422 – 6. Gene Ther 2001;8:308 – 15.

Mol Cancer Ther 2006;5(11). November 2006

38. Crystal RG, Hirschowitz E, Lieberman M, et al. Phase I study of 46. Kagawa S, Gu J, Swisher SG. Antitumor effect of adenovirus- direct administration of a replication deficient adenovirus vector contain- mediated Bax gene transfer on p53-sensitive and p53-resistant cancer ing the E. coli cytosine deaminase gene to metastatic colon carcinoma of lines. Cancer Res 2000;60:1157 – 61. the liver in association with the oral administration of the pro-drug 47. Ebert MP, Yu J, Hoffmann J, et al. Loss of h-catenin expression in 5-fluorocytosine. Hum Gene Ther 1997;8:985 – 1001. metastatic gastric cancer. J Clin Oncol 2003;21:1708 – 14. 39. Chen RH, McCormick F. Selective targeting to the hyperactive 48. Lowy AM, Fenoglio-Preiser, Kim OJ, et al. Dysregulation of h-catenin h-catenin/T-cell factor pathway in colon cancer cells. Cancer Res 2001; expression correlates with tumor differentiation in pancreatic duct 61:4445 – 9. adenocarcinoma. Ann Surg Oncol 2003;10:284 – 90. 40. Kwong KY, Zou Y, Day CP, Hung MC. The suppression of colon h cancer cell growth in nude mice by targeting h-catenin/Tcf pathway. 49. Gerdes B, Ramaswamy A, Simon B, et al. Analysis of -catenin gene Oncogene 2002;21:8340 – 6. mutations in pancreatic tumors. Digestion 1999;60:544 – 8. h 41. Lipinski KS, Djeha HA, Gawn J, et al. Optimization of a synthetic 50. Ikenoue T, Ijichi H, Kato N, et al. Analysis of the -catenin/T cell factor h-catenin-dependent promoter for tumor-specific cancer gene therapy. signaling pathway in 36 gastrointestinal and liver cancer cells. Jpn Mol Ther 2004;10:150 – 61. J Cancer Res 2002;93:1213 – 20. 42. Malerba M, Nikolova D, Cornelis J, Iggo R. Targeting of autonomous 51. Caca K, Kolligs FT, Ji X, et al. Beta- and g-catenin mutations, but not parvoviruses to colon cancer by insertion of Tcf sites in the P4 promoter. E-cadherin inactivation, underlie T-cell factor/lymphoid enhancer factor Cancer Gene Ther 2005;13:273 – 80. transcriptional deregulation in gastric and pancreatic cancer. Cell Growth Differ 1999;10:369 – 76. 43. Ito H, Kanzawa T, Miyoshi T, et al. Therapeutic efficacy of PUMA for malignant glioma cells regardless of p53 status. Hum Gene Ther 2005;16: 52. de Angelis PM, Fjell B, Kravik KL, Haug T, Tunheim SH, Reichelt W. 685 – 98. Molecular characterizations of derivatives of HCT116 colorectal cancer 44. Wang H, Qian H, Yu J, et al. Administration of PUMA adenovirus cells that are resistant to the chemotherapeutic agent 5-fluorouracil. Int increases the sensitivity of esophageal cancer cells to anticancer drugs. J Oncol 2004;24:1279 – 88. Cancer Biol Ther 2006;5:380 – 5. 53. Dvory-Sobol H, Kazanov D, Arber N. Gene targeting approach to 45. Harris MP, Sutjipto S, Wills KN, et al. Adenovirus-mediated p53 gene selectively kill colon cancer cells, with hyperactive K-Ras pathway. transfer inhibits growth of human tumor cells expressing mutant p53 Biomed Pharmacother 2005;59 Suppl 2:S370 – 4. protein. Cancer Gene Ther 1996;3:121 – 30. 54. Polakis P. Wnt signaling and cancer. Genes Dev 2000;14:1837 – 51.

Mol Cancer Ther 2006;5(11). November 2006

Hadas Dvory-Sobol, Eyal Sagiv, Diana Kazanov, et al.

Mol Cancer Ther 2006;5:2861-2871.

Updated version Access the most recent version of this article at: http://mct.aacrjournals.org/content/5/11/2861

Cited articles This article cites 54 articles, 20 of which you can access for free at: http://mct.aacrjournals.org/content/5/11/2861.full#ref-list-1

Citing articles This article has been cited by 7 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/5/11/2861.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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/5/11/2861. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.