Detection of HDM2 and VEGF co-expression in cancer cell lines: novel effect of HDM2 antisense treatment on VEGF expression

Available online at www.sciencedirect.com

Life Sciences 81 (2007) 1362 – 1372 www.elsevier.com/locate/lifescie

Detection of HDM2 and VEGF co-expression in cancer cell lines: novel effect of HDM2 antisense treatment on VEGF expression Madhusudhanan Narasimhan a , Rajiv Rose b , Muthusamy Karthikeyan a , Appu Rathinavelu a,b,⁎ a

Department of Pharmaceutical Sciences, College of Pharmacy, Health Professions Division, Nova Southeastern University, Ft. Lauderdale, FL 33328, USA b Rumbaugh-Goodwin Institute for Cancer Research, Nova Southeastern University, Plantation, FL 33313, USA Received 10 August 2007; accepted 29 August 2007

Abstract The human homologue of murine double minute 2 (HDM2) oncogene is amplified in approximately 7% of all human cancers. Overexpression of HDM2 protein impairs cell cycle control and confers growth advantage to cancer cells. In several cancers the progression of tumor growth and formation of distant metastases are found to be dependent on tumor angiogenesis, a process that is regulated by vascular endothelial growth factor (VEGF). In this study, we have investigated the co-expression of HDM2 and VEGF in various types of human cancer cell lines and have shown that the co-expression is not cell-type-specific. Furthermore, when different types of cell lines were treated with a HDM2 gene specific antisense phosphorothioate oligodeoxynucleotide (HDMAS5), the expression of VEGF mRNA as well as the levels of VEGF protein was found to be decreased. Interestingly, the higher basal levels of VEGF mRNA and the protein observed in HDM2 transfected LNCaP-MST cells were effectively suppressed by HDMAS5 treatment. On the contrary, the mutant oligodeoxynucleotide containing 4 mismatched bases (M4) did not alter the expression of either HDM2 or VEGF in any of the cell lines tested. In conclusion, our findings are the first time evidence showing that HDM2 and VEGF are co-expressed in various cancer cell lines that have aggressive growth and high metastatic abilities. Furthermore, the decrease in VEGF expression observed at the transcriptional as well as translational levels, subsequent to HDMAS5 treatment of p53 null cells, strongly suggests that HDM2 has a regulatory role on VEGF expression in a p53 independent manner. © 2007 Elsevier Inc. All rights reserved. Keywords: HDM2; VEGF; Co-expression; Antisense oligonucleotide; Cancer

Introduction The HDM2 gene was originally detected as an amplified DNA sequence on double minute chromosomes in the spontaneously transformed 3T3DM murine cell line (Cahilly-Snyder et al., 1987; Fakharzadeh et al., 1991). Subsequently, it was shown that amplification or overexpression of the HDM2 gene could increase the tumorigenic potential of the NIH3T3 and Rat2 cells (Fakharzadeh et al., 1991). To date, HDM2 is found to be

⁎ Corresponding author. Rumbaugh-Goodwin Institute for Cancer Research, 1850, NW 69th Avenue, Suite 5, Plantation, FL 33313, USA. Tel.: +1 954 587 9020; fax: +1 954 321 5311. E-mail address: [email protected] (A. Rathinavelu). 0024-3205/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.lfs.2007.08.029

overexpressed in more than forty different types of malignancies, including solid tumors, sarcomas and leukemias, thus qualifying itself as a classical oncogene. Although there are some conflicting data concerning the effects of HDM2 expression in individual cancers exist, the overall evidence is convincingly indicative of the fact that increased HDM2 expression is frequently related to a worse clinical prognosis (Onel and Cordon-Cardo, 2004; Rayburn et al., 2005). Additionally, growing number of evidences have strongly supported a pathologically important role for the physical interaction between HDM2 and the tumor suppressor p53 protein. During a direct interaction with p53 through its Nterminal domain, HDM2 forms an autoregulatory loop that results in the inhibition of transcriptional activity and degradation of p53 by the ubiquitin proteasome pathway (Jimenez et al., 1999; Lakin and Jackson, 1999; Sionov and Haupt, 1999). Furthermore, HDM2 shuttles p53 from the nucleus to cytoplasm (Kubbutat

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372

et al., 1999; O'Keefe et al., 2003) thereby shortening the half-life of p53 (Momand et al., 1992; Haupt et al., 1997; Vargas et al., 2003) through a mechanism involving functional nuclear export signal (NES) within HDM2 protein (Geyer et al., 2000). Alternatively, the RNA-binding activity of the RING-finger domain of HDM2 protein might play an important role in the nuclear exclusion of p53 (Boyd et al., 2000). While there is a strong interest existing to fine tune and target the HDM2–p53 interplay, for the purpose of antiproliferative effect, it has been discovered that HDM2 also has effects beyond disabling p53 (Ganguli and Wasylyk, 2003). Recent evidences indicate that HDM2 can interact with CBP/p300, pRB, p73, E2F1, DP1, the L5 ribosomal ribonucleoprotein particle, Numb, G3BP, etc. and influence various intrinsic pathways such as cell cycle, apoptosis, and tumorigenesis. Several in vivo studies have addressed a p53-independent role for HDM2 in processes other than tumorigenesis (Jones et al., 1998; Alkhalaf et al., 1999; Paliwal et al., 2007; Kim et al., 2007). HDM2 amplification has been detected more frequently in metastatic or recurrent tumors than in primary tumors. Therefore, it is widely speculated that HDM2 protein is not only responsible for tumorigenesis via p53 inactivation, but also for increasing the metastatic ability of originally non-metastatic tumor cells (Rathinavelu et al., 1999). Thus, HDM2 qualifies itself as a classical oncogene by virtue of its unique abilities to integrate multiple pathways involved in regulating cell cycle control, cell differentiation, gene transcription, signal transduction, receptors regulation etc. (Leng et al., 1995; Strous and Schantl, 2001; Gu et al., 2002; Yogosawa et al., 2003; Sdek et al., 2004). Typically, activation of oncogenes such as ras, raf, and src, or inactivation of tumor suppressor genes such as p53, VHL, PTEN are associated with VEGF overexpression in cancer cells that subsequently leads to loss of cell cycle control (Van Meir et al., 1994; Linderholm et al., 2001; Narendran et al., 2003; Na et al., 2003; Pore et al., 2003; Fukuda et al., 2003; Flaxenburg et al., 2004; Wang et al., 2004; Liu et al., 2004). Data derived from various experimental models clearly demonstrates that VEGF expression and its receptor function play a crucial role in tumor growth progression, invasion and metastasis (Cheng et al., 1996; Brekken et al., 2000; Lu et al., 2002; Baritaki et al., 2007; Adair, 2005; Belinsky et al., 2005). Furthermore, some of the recent experiments have shown that overexpression of VEGF could significantly enhance cell survival even after growth factor withdrawal and provide resistance to apoptosis induced by strong chemotherapeutic agents such as cisplatin (Zhang et al., 2002). Our own laboratory has reported a high level expression of VEGF consistent with the rapidly growing ability of GI-101A and HL-60 cells (Ramakrishnan et al., 2000). To a greater surprise, both GI-101A and HL-60 cells used in our experiments were also found to over express HDM2. On the basis of the known roles of HDM2 in both the growth and metastasis of tumors, we strongly speculated that a possible association might exist between HDM2 expression and VEGF levels in cancer cells. It is also clear that elucidating the pathway that is linking HDM2 to VEGF production could open up new avenues in the treatment of cancers by preventing tumor angiogenesis. Towards achieving this goal, as a first step, we have made an attempt to demonstrate the co-expression of the

1363

suspected cohorts — HDM2 and VEGF, in eight different cancer cell lines. To further verify the possibility of HDM2 being involved in regulating VEGF expression, we investigated the status of VEGF gene transcription as well as VEGF protein levels after treating the cells with HDMAS5, a HDM2 specific antisense oligonucleotide. Though there are many other options available for inhibiting HDM2 function, including the use of siRNAs, antibodies, peptides, and small molecules, we chose to use HDMAS5 due to its proven efficacy. The results of our experiments are presented here. Materials and methods Materials All reagents used in this study were of analytical grade and prepared in HPLC grade water whenever necessary. The antisense oligodeoxynucleotide HDMAS5 and its 4 base mismatch containing mutant oligonucleotide (M4) were synthesized from Biosynthesis (Lewisville, TX). The sequence of HDMAS5 is 5′GATCACTCCCACCTTCAAGG-3′, and the sequence of M4 containing 4 base mismatches (underlined) is 5′-GATGACTCACACCATCATGG-3′, that was used as a control in our treatments (Chen et al., 1998). Fetal bovine serum (FBS) was purchased from Hyclone. Lipofectin, penicillin, streptomycin, G418 were from GIBCO BRL, Life Technologies (Gaithersburg, MD). Complete RNeasy kit for total RNA preparation was purchased from Qiagen Inc. (Valencia, CA). The reverse transcriptase-polymerase chain reaction (RT-PCR) kit including 5× AMV/Tfl reaction buffer, dNTPs, MgSO4, Tfl DNA polymerase and AMV-reverse transcriptase enzyme was purchased from Promega (Madison, WI). Sense and Antisense primers for HDM2, VEGF were also obtained from Biosynthesis Inc. (Lewisville, TX). Cell culture The GI-101A cells were derived from a recurrent infiltrating ductal adenocarcinoma tumor (stage III a, T3N2MX) of a 57year-old woman (Rathinavelu et al., 1999). The HL-60, OVCAR3 and H358 cells were obtained from ATCC. MCF-7 was a kind gift from Dr. Subbarayan Pochi of the University of Miami School of Medicine, (Miami, FL). A2780/CP70 and the prostate cancer LNCaP cells were generous gifts from Dr. Bing Hua Jiang (West Virginia University, Morgan Town, WV) and Dr. Thomas Powell (Cleveland Clinic Foundation, Cleveland, OH) respectively. The HDM2-transfected LNCaP (LNCaP-MST) cells were kindly provided by Dr. Alan Pollack (Fox Chase Cancer Center, Philadelphia, PA). All the aforementioned cells were maintained in RPMI medium containing 10% (v/v) fetal bovine serum, 50 U of penicillin/ml and 50 mg of streptomycin/ml. The HDM2 gene transfected LNCaP-MST and LNCaP were maintained in Dulbecco's modified Eagle's medium (DMEM)-F12 containing 10% fetal bovine serum, 1% L-glutamine, 1% penicillin– streptomycin and 500 μg Geneticin (G418)/ml medium. All the cancer cells were grown in a humidified air/CO2 (19:1) atmosphere at 37 °C and replenished with the respective medium before any treatment.

1364

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372

Transfection of cells with antisense oligonucleotide The cells were grown to 50–60% confluency and changed to serum free medium on the day before transfection and allowed to stabilize overnight. Before transfecting the cells, they were replenished with the respective medium containing 1% FBS. Lipofectin suspension was prepared by incubating with serum free medium for 45 min. Oligonucleotides were then added to the lipofectin suspension in medium and left at room temperature for 10 min. At the end of incubations the lipofectin containing oligonucleotides (HDMAS5 or M4) was added to the respective cell culture flasks. The final concentration of lipofectin and FBS was adjusted to 7 μg/ml and 0.75% respectively. All the cell lines used in our study were subjected to transfection with HDMAS5 (500 nM) or M4 (500 nM) for 24 h at 37 °C. Detection of HDM2 and VEGF mRNA by RT-PCR After HDMAS5 and M4 treatment, the total RNAwas extracted from cells using the procedure provided with the RNeasy kit. The total number of cells used in all our treatments was approximately 106 cells; they were lysed with 350 μl of RLT buffer containing β-mercaptoethanol, after each treatment the RNA was extracted using RNeasy Kit according to the manufacturer's instructions (Qiagen Inc.). The RT-PCR was performed with 500 ng of total RNA using Promega PCR buffers and enzymes in 50 μl total volume. The reaction mixture consisted of 1× buffer, 0.2 mM dNTPs, 1 mM MgSO4, 50 pmol of each sense and antisense primers, and 5 U/μl each of AMV-RT and Tfl DNA polymerase. The PCR mix was reverse transcribed at 48 °C for 45 min and then the amplification was initiated at 95 °C for 10 min as AMV-RT activation step. The PCR thermocycles consisted of denaturation at 94 °C for 30 s, annealing at 59 °C for 1 min and synthesis at 68 °C for 2 min. After completing the cycles, an additional extension step was carried out for 7 min at 68 °C. The following forward: 5′-TCGGGCCTCCGAAACCATGA-3′, and reverse: 5′-CTGGTGAGAGATCTGGTTC-3′ primers (Yoshiji et al., 1996) were used in the VEGF RT-PCR reactions. The HDM2 cDNA was amplified using the following forward: 5′-CTGGGGAGTCTTGAGGGACC-3′ and the reverse: 5′-CAGGTTGTCTAAATTCCTAG-3′ primers (Matsumoto et al., 1998). β-actin gene expression was performed in all the RNA samples using conditions as described for VEGF, with the change of annealing temperature to 67 °C. For β-actin amplification the forward primer: 5′-GTGGGGCGCCCCAGGCACCA-3′ and the reverse primer: 5′-CTCCTTAATGTCACGCACGATTTC-3′ (Yamamura et al., 1991) were used. RT-PCR amplified products were separated on 1% agarose gel containing ethidium bromide. The agarose gel pictures were scanned using a HP flat bed scanner and the intensity of each band was determined using Scion Image (Frederick, MD) gel analysis program. The relative levels of RNA were calculated by utilizing the band intensity values. Western blot analysis of HDM2 and VEGF The cells treated with HDMAS5 or M4 for 24 h were lysed by sonication on ice and then the cell lysates were used for

western blotting experiments. The protein concentrations in the homogenates were measured and then exactly 750 μg of protein from each sample was incubated with anti-HDM2 monoclonal antibody (Ab-1) at 4 °C for 12 h, according to the method of Rathinavelu et al. (1999). At the end of incubation the immunoprecipitates were collected by centrifugation and resuspended in the homogenizing buffer. About 30 μg of the immunoprecipitated proteins were resolved in 7.5% SDS– polyacrylamide gel. The proteins were then transferred onto the nitrocellulose membrane and probed with 1:200 dilution of antiHDM2 monoclonal antibody (Ab-1) from Santacruz Biotechnologies (Santa Cruz, CA) for western blot analysis of HDM2 expression. The immunoreactive HDM2 protein signals were detected using ECL blot-developing system obtained from Amersham Corporation (Piscataway, NJ). For the detection of VEGF, aliquots (50 μg) of the protein were subjected to electrophoresis on 10% polyacrylamide gel and then they were transferred onto the nitrocellulose membrane. Following blocking with 6% milk, the membrane was probed with 1:250 dilution of the primary antibody (rabbit anti-VEGF antibody, Santa Cruz Biotechnologies, CA). The membranes were washed and treated with the goat anti-rabbit secondary antibody labeled with horseradish peroxidase (Sigma, St. Louis, MO). The VEGF protein bands were visualized using a commercially available chemiluminescence kit (Amersham, Piscataway, NJ). As a control, β-actin western blots were performed using a 1:2000 dilution of anti-β-actin monoclonal antibody (Sigma, St. Louis, MO). Statistical analysis The results are expressed as mean ± SD. Statistical significance between groups was analyzed by one-way analysis of variance (ANOVA) followed by Student–Newman–Keuls Multiple Comparisons tests. The P value of b 0.05 was considered significant and presented in our results. Results Effect of HDMAS5 on HDM2 mRNA levels RT-PCR analysis was performed with total RNA from either lipofectin control or HDMAS5 treated cells to assess the levels of HDM2 and VEGF mRNA expression. We first examined the effect of HDMAS5 on HDM2 mRNA expression in all the eight different cancer cell lines. The primer set selected for the RT-PCR of HDM2 (as described in Materials and methods) amplified two distinct cDNA fragments of 1.6 kb and 350 bp length in almost all the cell lines tested. Each of these cDNA fragments from one of the cell lines — A2780/CP70, was individually cloned into plasmids and confirmed as HDM2 splice variant (data not presented). As shown in Fig. 1a and b the GI-101A and HL-60 control cells expressed noticeable levels of 1.6 kb full-length HDM2 mRNA. Subsequent analysis of the other cells namely H358, A2780/CP70, MCF-7, OVCAR-3, LNCaP and the HDM2 gene transfected LNCaP-MST cells also showed high level expression of full-length HDM2 mRNA as well as one of its splice

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372

1365

Fig. 1. Effect of antisense oligonucleotides on HDM2 mRNA expression in various cancer cells. The cells were pretreated with or without HDMAS5 or M4 (500 nM) included in Lipofectin for 24 h. After treatment the total RNA was extracted from cells and subjected to RT-PCR analysis using HDM2 and β-actin specific primers. Graphs in lower panels show the relative intensity of bands corresponding to HDM2 mRNA expression. The band intensity was analyzed using Image J software, where an arbitrary value of one was assigned to the average intensity of HDM2 mRNA expression in control cells. These experiments were performed in triplicate, and the results shown are representative. The ⁎ indicates P b 0.05 when compared with control. (a) HDM2 expression in GI 101A. Molecular weight marker (lane 1), Control (lane 2), HDMAS5 (lane 3), M4 (lane 4). (b) HDM2 expression in HL-60 cells. Molecular weight marker (lane 1), Control (lane 2), HDMAS5 (lane 3), M4 (lane 4). (c) HDM2 expression in different cell lines. Molecular weight markers (lane 1), Positive Control (lane 2); H358 cells (lanes 3,4, and 5), A2780/CP70 cells (lanes 6,7, and 8), MCF-7 (lanes 9,10, and 11). The first three lanes under each cell line represent Control, HDMAS5 and M4 treatment respectively. (d) HDM2 expression in different cell lines. Molecular weight markers (lane 1), Positive Control (lane 2); OVCAR-3 cells (lanes 3,4, and 5), LNCaP-MST cells (lanes 6,7, and 8), LNCaP (lanes 9,10, and 11). The first three lanes under each cell line represent Control, HDMAS5 and M4 treatment respectively.

variants (Fig. 1c and d). It is important to note that the A2780/ CP70 ovarian cancer cells expressed the highest levels of HDM2 mRNA when compared to the other cell lines even in an unstimulated condition. As anticipated, HDMAS5 pretreatment

significantly reduced the full-length HDM2 mRNA expression by about 50% and 80% in unstimulated GI-101A cells and HL-60 cells respectively (Fig. 1a and b). Similar inhibitory patterns of HDM2 mRNA levels were observed in other cell lines such as

1366

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372

H358, A2780/CP70, MCF-7, OVCAR-3, LNCaP-MST and LNCaP and the percentage reduction was found to be 49%, 40%, 55%, 58%, 50% and 60% respectively. Furthermore, the level of 350 bp HDM2 mRNA splice variant was reduced remarkably (P b 0.05) by HDMAS5 pretreatment in the tested cancer cells. These results confirmed that HDMAS5 could effectively downregulate the expression of HDM2 mRNA, which might be due to the RNA-targeted degradation by RNase-H. The mutant M4 oligonucleotide that was used as a control displayed no inhibitory effects on the basal level expression of HDM2 mRNA. Our results clearly suggested that HDMAS5 treatment could effectively inhibit HDM2 mRNA expression in all the cancer cell lines tested in our experiments. Inhibition of HDM2 protein levels by HDMAS5 We then conducted western blot experiments to determine whether the expression of HDM2 mRNA, as detected by RT-

PCR, was reflected by expression of HDM2 protein in all the cancer cells. The basal level expression of HDM2 mRNA concurred with the expression of HDM2 protein after the antisense treatment, as detected by western blot in all the cell lines analyzed (Fig. 2a,b,c and d). As expected, when the cells were transiently transfected with HDMAS5 for 24 h, the HDM2 protein levels were also clearly reduced by 95% in GI-101A cells and 70% in HL-60 cells compared to HDMAS5 nontransfected cells (Fig. 2a and b). In HDMAS5 treated H358 and LNCaP cells, HDM2 protein was noticeably reduced by an average of 55% compared to lipofectin control (Fig. 2c and d). In MCF-7 and A2780/CP70 cells, the HDMAS5 treatment caused only an average reduction of around 45% in HDM2 protein levels (Fig. 2c). Parallel to the inhibition observed in other cell lines, the HDM2 protein expression at the baseline was dramatically decreased by around 70% in HDMAS5 transfected OVCAR-3 and LNCaP-MST cells (Fig. 2d). However, the treatment of cells with M4 did not change the

Fig. 2. Western blot analysis of HDM2 protein before and after antisense HDMAS5 treatment in various cancer cells. After 24 h transfection of the cells with HDMAS5 or M4, the whole immunoprecipitates containing HDM2 proteins were separated on 7.5% SDS-PAGE gel, transferred to nitrocellulose membrane and detected by using HDM2 specific monoclonal antibody. The first three lanes under each cell line represent the HDM2 protein expression in Control, HDMAS5 and M4 treatment respectively. The expression of β-actin was examined as a quantity control. The graphs below each plate show the relative intensity of HDM2 protein expression that was determined using Image J software. The ⁎ indicates P b 0.05 when compared with control. (a) HDM2 expression in GI-101A cells. (b) HDM2 expression in HL-60 cells. (c) HDM2 expression in H358, A2780/CP70, MCF-7 cells. (d) HDM2 expression in OVCAR-3, LNCaP-MST, LNCaP cells.

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372

1367

Fig. 3. Agarose gel electrophoresis pictures showing the effect of HDMAS5 or M4 on VEGF mRNA expression in various cancer cells. The cells were pretreated with or without HDMAS5 or M4 (500 nM) for 24 h in the presence of 7 μg/ml Lipofectin. After treatment the total RNA was extracted from the cells and reverse transcribed to cDNA, which was used as template to detect VEGF and β-actin. The first three lanes under each cell line represent Control, HDMAS5 and M4 treatment respectively. In all plates, the graphs on lower panel show the relative intensity of bands corresponding to VEGF mRNA expression and the band intensity was analyzed using Image J software. RT-PCR was repeated three times on different experiments and this result is representative. ⁎The difference from cells treated with HDMAS5 alone was statistically significant at P b 0.05 when compared with Control. In all the experiments equal amounts of RNA were analyzed, as confirmed by the intensity of β-actin. (a) VEGF expression in GI-101A cells. Molecular weight marker (lane 1), Positive Control (lane 2), Control (lane 3), HDMAS5 (lane 4), M4 (lane 5). (b) VEGF expression in HL-60 cells. Molecular weight marker (lane 1), Positive Control (lane 2), Control (lane 3), HDMAS5 (lane 4), M4 (lane 5). (c) VEGF expression in H358, A2780/CP70, MCF-7 cells. Molecular weight markers (lane 1), Positive Control (lane 2); H358 cells (lanes 3,4, and 5), A2780/CP70 cells (lanes 6,7, and 8), MCF-7 (lanes 9,10, and 11). The first three lanes under each cell line represent Control, HDMAS5 and M4 treatment respectively. (d) VEGF expression in OVCAR-3, LNCaP-MST, LNCaP cells. Molecular weight markers (lane 1), Positive Control (lane 2); OVCAR-3 cells (lanes 3,4, and 5), LNCaP-MST cells (lanes 6,7, and 8), LNCaP (lanes 9,10, and 11). The first three lanes under each cell line represent Control, HDMAS5 and M4 treatment respectively.

basal levels of HDM2 protein. Thus, our immunoblotting analyses clearly suggest that interference and degradation of HDM2 mRNA by HDMAS5 might have led to the decrease in HDM2 protein levels in tumor cells.

Effect of HDMAS5 on VEGF mRNA levels The mammalian VEGF gene is known to consist of at least 8 exons that can be assembled, through alternative splicing into

1368

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372

a number of variant mRNA molecules. In our present study, all the human cancer cell lines that were shown to express HDM2 mRNA also expressed high levels of VEGF mRNA. In each case, the amplification resulted in visualization of three bands in the range of 400, 520 and 650 bp corresponding to the mRNA encoding VEGF121, VEGF165 and VEGF189 respectively. In breast cancer (GI-101A and MCF-7), lung cancer (H358) and the leukemic (HL-60) cells, the splice variants VEGF121 and VEGF165 were abundantly expressed than the splice variant VEGF189. In ovarian cancer (OVCAR-3) and in prostate cancer (LNCaP and LNCaP-MST) cell lines, the expression levels of VEGF isoforms were found to be in the order of 189 N 165 N 121 (Fig. 3d). In A2780/CP70 cells a higher level expression was observed for the isoforms 189 and 121 when compared to the isoform 165. Though the average levels of individual splice variants of VEGF appeared to be relatively comparable among

all the cancer cells tested in our experiment, at present the physiological significance of the presence of varied levels of isoforms are unclear. Notably, the cells that possess p53 mutation/lost function namely GI-101 A, OVCAR-3 and A2780/CP70 cells and the p53 null(−/−) cells HL-60, H358, expressed very high basal levels of VEGF mRNA. In fact, in A2780/CP70 cells that have lost the function of p53 protein, the basal VEGF mRNA levels were found to be highest when compared to the rest of the cell lines (Fig. 3c, lane 6). This shows that the transcriptional activation may be very high in p53 mutant cell lines, even in an unstimulated basal state, that correlates very well with the high level expression of HDM2 mRNA also. The HDM2 gene transfected prostate cells (LNCaP-MST) showed higher level expression of VEGF mRNA than the non-transfected LNCaP cells as a result of HMM2 gene transfection (Fig. 3d, compare lane 6 vs lane 9).

Fig. 4. VEGF protein expression in various cancer cells before and after treating with HDM2 mRNA specific antisense HDMAS5. After 24 h transfection of the cells with HDMAS5 or M4, the protein extracts from the cells were resolved by 10% SDS-PAGE gel, probed with VEGF polyclonal antibody and revealed by chemiluminescence. The first three lanes under each cell line represent the VEGF protein expression in Control, HDMAS5 and M4 treatment respectively. Quantification of the relative intensity of VEGF protein expression was achieved using Image J program. Values expressed in the graph are mean ± SD and the ⁎ indicates P b 0.05 when compared with control. Equal protein loading was confirmed by actin blot. (a) VEGF expression in GI-101A cells. (b) VEGF expression in HL-60 cells. (c) VEGF expression in H358, A2780/CP70, MCF-7 cells. (d) VEGF expression in OVCAR-3, LNCaP-MST, LNCaP cells.

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372

The HDMAS5 pretreatment almost completely attenuated the expression of 650 bp splice variant of VEGF in GI-101A cells, while the level of 400 bp (VEGF121) and 520 bp (VEGF165) splice variants was effectively blocked by about 70% and 90% below the untreated control (Fig. 3a). Similar inhibitory pattern was observed in the p53(−/−) null HL-60 cells. In other cells such as H358 and MCF-7, OVCAR-3, LNCaP-MST and LNCaP a similar downregulation of VEGF mRNA isoforms expression in the range of 20–70% was observed when treated with HDMAS5. Interestingly, the levels of VEGF mRNA isoforms 189, 165 and 120 observed in A2780/CP70 cells were attenuated by 55%, 45% and 35% (compare lane 7 vs lane 6; Fig. 3c) suggesting that HDM2 might play a major role in the transcriptional regulation of VEGF. At the same time our present study with HDMAS5 indicates that it has the ability to negatively influence the expression of all the three splice variants of VEGF, with no inhibitory effects shown by the mutant M4 oligonucleotide. Thus, the results ensued from our current study confirms the ability of HDM2 gene specific antisense HDMAS5's effect over VEGF expression. In all our RT-PCR experiments βactin mRNA level was used as an internal control and its level was not altered in both cell lines during any of the treatments. Furthermore, comparison of data shown in Fig. 3a,b,c and d, derived from HDMAS5-pretreated cancer cells with those from the untreated cancer cells provides a relative estimate of the abundance of the basal level expression of VEGF mRNAs in all the cancer cells that express HDM2. Inhibition of VEGF protein expression by HDM2 antisense — HDMAS5 In order to confirm the impact of observed changes in VEGF mRNA level on its translation and to determine whether the reduction of HDM2 protein expression has any association with decreased VEGF transcription, we analyzed the levels of VEGF protein in various cancer cells before and after HDMAS5 pretreatment. The western blot analysis confirmed a high basal expression of VEGF protein under normal culture conditions in all the cancer cells lines analyzed (Fig. 4a,b,c and d). Our data also indicates that HDM2 overexpression could result in more pronounced VEGF translation as can be noticed from higher expression of VEGF protein in HDM2 gene transfected LNCaPMST cell line (compare lane 4 vs lane 7; Fig. 4d). Furthermore, pretreatment with HDM2 gene specific antisense HDMAS5 abrogated the VEGF protein signal in the p53 positive cell lines such as MCF-7, LNCaP, and LNCaP-MST by 45%, 55% and 70% respectively. We subsequently made an attempt to determine whether the ability of HDMAS5 in decreasing the expression of VEGF through interference with HDM2 mRNA is p53 independent. To our surprise, in p53 mutant (GI-101A, OVCAR3, A2780/CP70) and p53(−/−) null (HL-60, H358) cells the HDMAS5 treatment remarkably decreased the VEGF protein levels (P b 0.05) (Fig. 4a,b,c and d). Thus, our findings confirmed that the VEGF suppression induced by HDM2 mRNA specific antisense oligodeoxynucleotides (AS-ODNs) was due to the attenuation of HDM2 protein expression and thereby deranging the transcriptional activation of VEGF gene. Furthermore, our

1369

study strongly suggests that this regulatory effect could be achieved by both p53 independent as well as dependent mechanisms. Discussion A potentially relevant role in cancer progression and prognosis has been attributed to two important proteins namely HDM2 and VEGF that are involved in multiple functions. In our present study, for the first time we have showed the co-expression of HDM2 and VEGF in eight different cancer cell lines irrespective of p53 status. As a negative regulator of p53, HDM2 is known to play an important role in tumor formation and growth. In addition to its p53-associated functions, HDM2 has been reported to transform cells, independent of p53 status (Dubs-Poterszman et al., 1995; Nelson et al., 2006). Furthermore, in support of HDM2's independent tumorigenic transforming potential it was shown that the transgenic mice that overexpress HDM2 developed sarcomas even in the p53 null mice (Jones et al., 1998). With at least 40 alternatively and aberrantly spliced transcripts of HDM2 mRNA have been identified in tumors so far, little is known about their functions. In the present study, we have reported the detection of a 350 bp splice variant of HDM2 apart from the full-length HDM2 (1.6 kb) in majority of the cancer cells tested. The proangiogenic cytokine, VEGF, also has been frequently found to be overexpressed in different cancers and, the consequent elevated levels of VEGF in serum have proven its utility as a marker of tumor progression (Poon et al., 2001). Analogous to HDM2 overexpression, we have detected high basal levels of three major isoforms of VEGF 121, 165 and 189 in almost all the cancer cell lines used in our study. Since all three splice variants were detected in majority of the cancer cells used in our experiment, it appears that the splice variant expression may not have any association with organ types such as breast, ovary, prostate, and lung. The functional significance of the splice variants of VEGF is also an undefined aspect similar to the role for HDM2 splice variants in cancer; however, they are all reported to be present in various tumor tissues. Usually the VEGF121 and VEGF165 isoforms appear to predominate quantitatively and functionally in most angiogenic states, while VEGF121 was estimated to be the most angiogenic and tumorigenic than the other isoforms (Zhang et al., 2000; Catena et al., 2007). It has been shown that the VEGF189 isoform more strongly binds to the cell membrane than other isoforms and therefore, has a high potential in local angiogenesis and tumor growth in the cancer inductive microenvironment (Tomii et al., 2002). Thus, our data could reveal that the co-expression of HDM2 and VEGF might represent a useful combination of prognostic markers in different cancer types and could also serve as potential markers for targeted antiangiogenic therapies. Consistent with the oncogenic potential of HDM2, its ability to interact with other critical growth suppressors make HDM2 a valuable candidate to be targeted for cancer treatment. On the basis of previous studies and our present results, which show coexpression of HDM2 and VEGF in various cancer cells, we speculated that there might exist an intracellular link between the multifunctional HDM2 and VEGF at the transcriptional level

1370

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372

favoring towards proangiogenic abilities. In the present study, the antisense HDMAS5 that was directed against a sequence within exon 7 of the HDM2 mRNA effectively decreased the basal levels of both HDM2 and VEGF mRNA in all the cancer cells tested. Interestingly, HDMAS5 treatment not only inhibited the expression of the target protein HDM2 but also VEGF. Among the various approaches for targeting HDM2 for cancer therapy, small molecule antagonists have recently featured as effective agents in experimental models. Various small molecule inhibitors have been shown to disrupt the p53–HDM2 binding, thereby enhancing p53 activity to elicit anticancer effect. Though the small molecule strategy is promising, yet it suffers a serious setback in that they don't seem to influence the human cancers that have high mutation frequency in p53 gene. Therefore, use of AS-ODN directed against HDM2 could still be a safer cancer therapy strategy, because it is possible to achieve sequence specific inhibition of the HDM2 mRNA expression and the resultant protein synthesis. As a consequence of this advantageous applicability of HDM2 gene specific oligodeoxynucleotide (ODN) to curb the p53 independent functions of HDM2, unlike the small molecules that are designed to disrupt only the HDM2 binding to p53, multiple effects can be achieved. A second generation ODN specific for HDM2 has also been shown to exhibit a potent antitumor activity against a large variety of human cancer types in vitro and in mice, regardless of the p53 status (Zhang and Wang, 2000; Tortora et al., 2000). Particularly relevant in this discussion is the fact, that the cells used in our present study namely GI-101A, OVCAR-3 possess mutant p53 status while HL-60 and H358 are p53 null cells. The overexpression of HDM2 in these cells certainly should have p53 independent functions. Therefore, obviously ODNs are going to be more effective than HDM2 specific small molecule inhibitors. Furthermore, in A2780/CP70 cells that have lost function of p53 protein, the robust expression of VEGF mRNA was effectively attenuated by HDMAS5. Thus, it is apparent that the observed decrease in VEGF mRNA and protein levels in these p53 mutant and null cells could be directly attributed to the suppression of HDM2 expression achieved by HDMAS5 transfection independent of p53 status. It is interesting to note that a protype antiangiogenic compound SU5416 has been shown to inhibit VEGF at the transcriptional level by reducing the hypoxia inducible factor-1α (HIF-1α) expression (Zhong et al., 2004a). Interestingly the same compound was able to inhibit the expression of G1 cell cycle checkpoint regulators such as, HDM2, p21, and p27 apart from p53 in ovarian carcinoma cells (Zhong et al., 2004b). While correlating the findings of these studies, it is reasonable to presume that blockade of HDM2 could result in decreased HIF-1α mediated VEGF transcription independent of p53 status. Overall our results prompt us to suggest that there might exist a unique intracellular link between HDM2 and VEGF independent of p53. However, further efforts along this line are required to confirm the potential candidates involved in this regulatory pathway. Herein we have also presented data showing that the transfection of HDM2 gene to the p53 positive LNCaP prostate cancer cells could produce HDM2 and VEGF overexpressing cell line (LNCaP-MST). As a conclusion to our hypothesis, the

detection of an overall increase in VEGF mRNA as well as VEGF protein levels found in HDM2 gene transfected cell line (LNCaP-MST) clearly substantiates our hypothesis that there might exist an intracellular link between HDM2 and VEGF. Additionally, the results presented here by us clearly demonstrate that HDMAS5 pretreatment could effectively inhibit both the transcription as well as the translation of VEGF in HDM2 overexpressing LNCaP-MST cells via inhibiting HDM2 expression. It has been previously shown by other groups also that the HDMAS5-ODN treatment could significantly reduce the HDM2 protein level by 5-fold in JAR, SJSA, and MCF-7 cells that resulted in activation of p53 (Chen et al., 1998). Though it may appear that the regulation of VEGF expression could be influenced by p53 dependent pathway in some cell lines, including LNCaP, LNCaP-MST cell lines, it cannot be ruled out at the same time that the p53 independent regulation of VEGF transcription by HDM2 could also be possible. At present, although we have not determined the precise mechanism, our data could argue that HDM2 might act as a positive regulator of VEGF both in p53 dependent and p53 independent fashion. Nevertheless, additional research efforts are in progress in our laboratory to elucidate the actual pathway that is linking HDM2 to VEGF production as it could lead to the development of new prognostic markers to assess tumor angiogenesis and cancer metastasis. The present study thus suggests that in most of the cancer types, the co-expression of HDM2 and VEGF might occur. In HDM2 and VEGF co-expressing cells blocking HDM2 could negatively influence VEGF transcription, which in turn can inhibit tumor angiogenesis. In light of the current findings, it is safer to propose that the future cancer therapy could involve agents that are capable of inhibiting HDM2 expression, in addition to blocking the binding of HDM2-p53 interaction. In this regard, our future studies would be to assess the effectiveness of combinatorial therapy using HDM2-based antisense with other conventional drugs to achieve synergistic therapeutic efficacy. Acknowledgement Financial supports from the Center of Excellence for Marine Biology and Biotechnology by the State of Florida through the Florida Atlantic University, Boca Raton, FL., the Presidents Faculty Research and Development Grant of Nova Southeastern University, Fort Lauderdale, FL., and the Royal Dames of Cancer Research Inc., Fort Lauderdale, FL., are gratefully acknowledged. The authors also thank Dr. Subbarayan Pochi (University of Miami school of Medicine, USA), Dr. Bing Hua Jiang (West Virginia University, USA), Dr. Thomas Powell (Cleveland Clinic Foundation, USA) and Dr. Alan Pollack (Fox Chase Cancer Center, USA) for their kind gifts of MCF-7, A2780/CP70, LNCaP and LNCaP-MST cells respectively. References Adair, T.H., 2005. Growth regulation of the vascular system: an emerging role for adenosine. American Journal of Physiology. Regulatory, Integrative and Comparative Physiology 289, R283–R296.

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372 Alkhalaf, M., Ganguli, G., Messaddeq, N., Le Meur, M., Wasylyk, B., 1999. MDM2 overexpression generates a skin phenotype in both wild type and p53 null mice. Oncogene 18, 1419–1434. Baritaki, S., Sifakis, S., Huerta-Yepez, S., Neonakis, I.K., Soufla, G., Bonavida, B., Spandidos, D.A., 2007. Over expression of VEGF and TGF-beta 1 mRNA in Pap smears correlates with progression of cervical intraepithelial neoplasia to cancer: implication of YY1 in cervical tumorigenesis and HPV infection. International Journal of Oncology 31 (1), 69–79. Belinsky, G.S., Claffey, K.P., Nambiar, P.R., Guda, K., Rosenberg, D.W., 2005. Vascular endothelial growth factor and enhanced angiogenesis do not promote metastatic conversion of a newly established azoxymethaneinduced colon cancer cell line. Molecular Carcinogenesis 43, 65–74. Boyd, S.D., Tsai, K.Y., Jacks, T., 2000. An intact HDM2 RING-finger domain is required for nuclear exclusion of p53. Nature Cell Biology 2, 563–568. Brekken, R.A., Overholser, J.P., Stastny, V.A., Waltenberger, J., Minna, J.D., Thorpe, P.E., 2000. Selective inhibition of vascular endothelial growth factor (VEGF) receptor 2 (KDR/Flk-1) activity by a monoclonal anti-VEGF antibody blocks tumor growth in mice. Cancer Research 60, 5117–5124. Cahilly-Snyder, L., Yang-Feng, T., Francke, U., George, D.L., 1987. Molecular analysis and chromosomal mapping of amplified genes isolated from a transformed mouse 3T3 cell line. Somatic Cell and Molecular Genetics 13, 235–244. Catena, R., Muniz-Medina, V., Moralejo, B., Javierre, B., Best, C.J.M., EmmertBuck, M.R., Green, J.E., Baker, C.C., Calvo, A., 2007. Increased expression of VEGF121/VEGF 165–189 ratio results in a significant enhancement of human prostate tumor angiogenesis. International Journal of Cancer 120, 2096–2109. Chen, L., Agrawal, S., Zhou, W., Zhang, R., Chen, J., 1998. Synergistic activation of p53 by inhibition of MDM2 expression and DNA damage. Proceedings of the National Academy of Sciences of the United States of America 95, 195–200. Cheng, S.Y., Huang, H.J., Nagane, M., Ji, X.D., Wang, D., Shih, C.C., Arap, W., Huang, C.M., Cavenee, W.K., 1996. Suppression of glioblastoma angiogenicity and tumorigenicity by inhibition of endogenous expression of vascular endothelial growth factor. Proceedings of the National Academy of Sciences of the United States of America 93, 8502–8507. Dubs-Poterszman, M.C., Tocque, B., Wasylyk, B., 1995. MDM2 transformation in the absence of p53 and abrogation of the p107 G1 cell-cycle arrest. Oncogene 11, 2445–2449. Fakharzadeh, S.S., Trusko, S.P., George, D.L., 1991. Tumorigenic potential associated with enhanced expression of a gene that is amplified in a mouse tumor cell line. EMBO Journal 10, 1565–1569. Flaxenburg, J.A., Melter, M., Lapchak, P.H., Briscoe, D.M., Pal, S., 2004. The CD40-induced signaling pathway in endothelial cells resulting in the overexpression of vascular endothelial growth factor involves Ras and phosphatidylinositol 3-kinase. Journal of Immunology 172, 7503–7509. Fukuda, R., Kelly, B., Semenza, G.L., 2003. Vascular endothelial growth factor gene expression in colon cancer cells exposed to prostaglandin E2 is mediated by hypoxia-inducible factor 1. Cancer Research 63, 2330–2334. Ganguli, G., Wasylyk, B., 2003. p53-independent functions of MDM2. Molecular Cancer Research 1, 1027–1035. Geyer, R.K., Yu, Z.K., Maki, C.G., 2000. The MDM2 RING-finger domain is required to promote p53 nuclear export. Nature Cell Biology 2, 569–573. Gu, L., Findley, H.W., Zhou, M., 2002. MDM2 induces NF-kappaB/p65 expression transcriptionally through Sp1-binding sites: a novel, p53independent role of MDM2 in doxorubicin resistance in acute lymphoblastic leukemia. Blood 99, 3367–3375. Haupt, Y., Maya, R., Kazaz, A., Oren, M., 1997. Mdm2 promotes the rapid degradation of p53. Nature 387, 296–299. Jimenez, G.S., Khan, S.H., Stommel, J.M., Wahl, G.M., 1999. p53 regulation by post-translational modification and nuclear retention in response to diverse stresses. Oncogene 18, 7656–7665. Jones, S.N., Hancock, A.R., Vogel, H., Donehower, L.A., Bradley, A., 1998. Overexpression of Mdm2 in mice reveals a p53-independent role for Mdm2 in tumorigenesis. Proceedings of the National Academy of Sciences of the United States of America 95, 15608–15612. Kim, M.M., Wiederschain, D., Kennedy, D., Hansen, E., Yuan, Z.M., 2007. Modulation of p53 and MDM2 activity by novel interaction with Ras–GAP binding proteins (G3BP). Oncogene 29, 4209–4215.

1371

Kubbutat, M.H., Ludwig, R.L., Levine, A.J., Vousden, K.H., 1999. Analysis of the degradation function of Mdm2. Cell Growth & Differentiation: The Molecular Biology Journal of The American Association for Cancer Research 10, 87–92. Lakin, N.D., Jackson, S.P., 1999. Regulation of p53 in response to DNA damage. Oncogene 18, 7644–7655. Leng, P., Brown, D.R., Shivakumar, C.V., Deb, S., Deb, S.P., 1995. N-terminal 130 amino acids of MDM2 are sufficient to inhibit p53-mediated transcriptional activation. Oncogene 10, 1275–1282. Linderholm, B.K., Lindahl, T., Holmberg, L., Klaar, S., Lennerstrand, J., Henriksson, R., Bergh, J., 2001. The expression of vascular endothelial growth factor correlates with mutant p53 and poor prognosis in human breast cancer. Cancer Research 61, 2256–2260. Liu, J., Shibata, T., Qu, R., Ogura, M., Hiraoka, M., 2004. Influences of the p53 status on hypoxia-induced gene expression. Journal of Radiation Research 45, 333–339. Lu, D., Jimenez, X., Zhang, H., Bohlen, P., Witte, L., Zhu, Z., 2002. Selection of high affinity human neutralizing antibodies to VEGFR2 from a large antibody phage display library for antiangiogenesis therapy. International Journal of Cancer 97, 393–399. Matsumoto, R., Tada, M., Nozaki, M., Zhang, C.L., Sawamura, Y., Abe, H., 1998. Short alternative splice transcripts of the mdm2 oncogene correlate to malignancy in human astrocytic neoplasms. Cancer Research 58, 609–613. Momand, J., Zambetti, G.P., Olson, D.C., George, D., Levine, A.J., 1992. The mdm-2 oncogene product forms a complex with the p53 protein and inhibits p53-mediated transactivation. Cell 69, 237–245. Na, X., Wu, G., Ryan, C.K., Schoen, S.R., di'Santagnese, P.A., Messing, E.M., 2003. Overproduction of vascular endothelial growth factor related to von Hippel–Lindau tumor suppressor gene mutations and hypoxia-inducible factor-1 alpha expression in renal cell carcinomas. Journal of Urology 170, 588–592. Narendran, A., Ganjavi, H., Morson, N., Connor, A., Barlow, J.W., Keystone, E., Malkin, D., Freedman, M.H., 2003. Mutant p53 in bone marrow stromal cells increases VEGF expression and supports leukemia cell growth. Experimental Hematology 31, 693–701. Nelson, D.M., Bhaskaran, V., Foster, W.R., Lehman-Mc Keeman, L.D., 2006. p53 Independent induction of rat hepatic Mdm2 following administration of phenobarbital and pregnenolone 16 alpha-carbonitrile. Toxicological Sciences: An Official Journal of the Society of Toxicology 94 (2), 272–280. Poon, R.T., Fan, S.T., Wong, J., 2001. Clinical implications of circulating angiogenic factors in cancer patients. Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology 19, 1207–1225. O'Keefe, K., Li, H., Zhang, Y., 2003. Nucleocytoplasmic shuttling of p53 is essential for MDM2-mediated cytoplasmic degradation but not ubiquitination. Molecular and Cellular Biology 23, 6396–6405. Onel, K., Cordon-Cardo, C., 2004. MDM2 and prognosis. Molecular Cancer Research 2, 1–8. Paliwal, P., Radha, V., Swarup, G., 2007. Regulation of p73 by Hck through kinasedependent and independent mechanisms. BMC Molecular Biology 8, 45. Pore, N., Liu, S., Haas-Kogan, D.A., O'Rourke, D.M., Maity, A., 2003. PTEN mutation and epidermal growth factor receptor activation regulate vascular endothelial growth factor VEGF. mRNA expression in human glioblastoma cells by transactivating the proximal VEGF promoter. Cancer Research 63, 236–241. Ramakrishnan, R., Zell, J.A., Malave, A., Rathinavelu, A., 2000. Expression of vascular endothelial growth factor mRNA in GI-101A and HL-60 cell lines. Biochemical and Biophysical Research Communications 270, 709–713. Rathinavelu, P., Malave, A., Raney, S.R., Hurst, J., Roberson, C.T., Rathinavelu, A., 1999. Expression of mdm-2 oncoprotein in the primary and metastatic sites of mammary tumor (GI-101) implanted athymic nude mice. Cancer Biochemistry Biophysics 17, 133–146. Rayburn, E., Zhang, R., He, J., Wang, H., 2005. MDM2 and human malignancies: expression, clinical pathology, prognostic markers, and implications for chemotherapy. Current Cancer Drug Targets 5, 27–41. Sdek, P., Ying, H., Zheng, H., Margulis, A., Tang, X., Tian, K., Xiao, Z.X., 2004. The central acidic domain of MDM2 is critical in inhibition of retinoblastoma-mediated suppression of E2F and cell growth. Journal of Biological Chemistry 279, 53317–53322.

1372

M. Narasimhan et al. / Life Sciences 81 (2007) 1362–1372

Sionov, R.V., Haupt, Y., 1999. The cellular response to p53: the decision between life and death. Oncogene 18, 6145–6157. Strous, G.J., Schantl, J.A., 2001. Beta-arrestin and Mdm2, unsuspected partners in signaling from the cell surface. Science's STKE: Signal Transduction Knowledge Environment 2001, PE41. Tomii, Y., Yamazaki, H., Sawa, N., Ohnishi, Y., Kamochi, J., Tokunaga, T., Osamura, Y., Sadahiro, S., Kijima, H., Abe, Y., Ueyama, Y., Tamaoki, N., Nakamura, M., 2002. Unique properties of 189 amino acid isoform of vascular endothelial growth factor in tumorigenesis. International Journal of Oncology 21, 1251–1257. Tortora, G., Caputo, R., Damiano, V., Bianco, R., Chen, J., Agrawal, S., Bianco, A.R., Ciardiello, F., 2000. A novel MDM2 anti-sense oligonucleotide has anti-tumor activity and potentiates cytotoxic drugs acting by different mechanisms in human colon cancer. International Journal of Cancer 88, 804–809. Van Meir, E.G., Polverini, P.J., Chazin, V.R., Su Huang, H.J., de Tribolet, N., Cavenee, W.K., 1994. Release of an inhibitor of angiogenesis upon induction of wild type p53 expression in glioblastoma cells. Nature Genetics 8, 171–176. Vargas, D.A., Takahashi, S., Ronai, Z., 2003. Mdm2: a regulator of cell growth and death. Advances in Cancer Research 89, 1–34. Wang, F.S., Wang, C.J., Chen, Y.J., Chang, P.R., Huang, Y.T., Sun, Y.C., Huang, H.C., Yang, Y.J., Yang, K.D., 2004. Ras induction of superoxide activates ERK-dependent angiogenic transcription factor HIF-1alpha and VEGF-A expression in shock wave-stimulated osteoblasts. Journal of Biological Chemistry 79, 10331–10337.

Yamamura, M., Uyemura, K., Deans, R.J., Weinberg, K., Rea, T.H., Bloom, B.R., Modlin, R.L., 1991. Defining protective responses to pathogens: cytokine profiles in leprosy lesions. Science 254, 277–279. Yogosawa, S., Miyauchi, Y., Honda, R., Tanaka, H., Yasuda, H., 2003. Mammalian Numb is a target protein of Mdm2, ubiquitin ligase. Biochemical and Biophysical Research Communications 302, 869–872. Yoshiji, H., Gomez, D.E., Shibuya, M., Thorgeirsson, U.P., 1996. Expression of vascular endothelial growth factor, its receptor, and other angiogenic factors in human breast cancer. Cancer Research 56, 2013–2016. Zhang, Wang, H., 2000. MDM2 oncogene as a novel target for human cancer therapy. Current Pharmaceutical Design 6, 393–416. Zhang, H.T., Scott, P.A., Morbidelli, L., Peak, S., Moore, J., Turley, H., Harris, A.L., Ziche, M., Bicknell, R., 2000. The 121 amino acid isoform of vascular endothelial growth factor is more strongly tumorigenic than other splice variants in vivo. British Journal of Cancer 83, 63–68. Zhang, L., Yang, N., Garcia, J.R., Mohamed, A., Benencia, F., Rubin, S.C., Allman, D., Coukos, G., 2002. Generation of a syngeneic mouse model to study the effects of vascular endothelial growth factor in ovarian carcinoma. American Journal of Pathology 161, 2295–2309. Zhong, X.S., Zheng, J.Z., Reed, E., Jiang, B.H., 2004a. SU5416 inhibited VEGF and HIF-1alpha expression through the PI3K/AKT/p70S6K1 signaling pathway. Biochemical and Biophysical Research Communications 324, 471–480. Zhong, X., Li, X., Wang, G., Zhu, Y., Hu, G., Zhao, J., Neace, C., Ding, H., Reed, E., Li, Q.Q., 2004b. Mechanisms underlying the synergistic effect of SU5416 and cisplatin on cytotoxicity in human ovarian tumor cells. International Journal of Oncology 25, 445–451.