Transactivation of cyclin E gene by EWS-Fli1 and antitumor effects of cyclin dependent kinase inhibitor on Ewing's family tumor cells

Authors


Abstract

Chromosomal translocation t(11; 22)(q24; q12) is detected in approximately 90% of Ewing's family tumors (EFTs) including Ewing's sarcoma and primitive neuroectodermal tumor. This results in the formation of the EWS-Fli1 fusion gene, which produces EWS-Fli1 fusion protein. This chimerical gene product acts as an aberrant transcriptional activator, which may be responsible for the tumorigenesis of EFTs. We have previously reported that cyclin E expression was upregulated in EFT cells and in EWS-Fli1 transformed fibroblastic cells. However, the mechanism of the overexpression of cyclin E by EWS-Fli1 is still unknown. In our study, we investigated the mechanism of transactivation of the cyclin E gene in EFT cells. We found that EWS-Fli1 enhanced the activity of the cyclin E gene promoter partially through E2F binding sites in the promoter. In addition, the basic transcriptional factor, Sp1, might also be involved in the transactivation of the cyclin E gene by EWS-Fli1. To study the biological significance of cyclin E overexpression in EFT cells, we used flavopiridol, a pan-cyclin-dependent kinase (CDK) inhibitor and found that flavopiridol efficiently suppressed the growth of EFT cells in vitro and in vivo by the inhibition of cyclinE/CDK2 kinase activity and the induction of apoptosis. These results suggest that targeting of the cyclin/CDK complex may provide new insight into treatment of EFTs. © 2005 Wiley-Liss, Inc.

A specific chromosomal translocation t(11; 22)(q24; q12) is detected in approximately 90–95% of Ewing's family tumors (EFTs) including Ewing's sarcoma and primitive neuroectodermal tumor.1 This results in the formation of EWS-Fli1 gene fusion, which produces EWS-Fli1 fusion protein.2 EWS-Fli1 functions as an aberrant transcriptional activator, which may be responsible for the tumorigenesis of EFTs.3, 4 However, the mechanisms leading to oncogenesis of EFTs by EWS-Fli1 are not well understood. Since the 5-year survival rate for patients with EFTs is approximately 50%,5 there is an immediate need to further study the disease mechanisms and develop new treatment strategies for EFTs.

We have previously reported that targeting of the EWS-Fli1 fusion gene by antisense oligonucleotides inhibited the growth of various EFT cell lines and arrested the cell cycle progression at G1 phase.6 In addition, the expression levels of G1 cyclins, including cyclin D1 and cyclin E, were markedly decreased by the reduction of EWS-Fli1 fusion protein.7 On the other hand, the expression of 2 important cyclin-dependent kinase inhibitors for G1-S transition, p21WAF1/CIP1 and p27KIP1, was dramatically increased after treatment with the antisense oligonucleotides for EWS-Fli1.7 We have also demonstrated that EWS-Fli1 fusion protein directly bound to the p21WAF1/CIP1 promoter through ETS-binding sites and negatively regulated p21WAF1/CIP1 gene expression by the suppression of p300-mediated HAT activity.8 Furthermore, we have shown that EWS-Fli1 might attenuate p27KIP1 protein levels via activation of the proteasome-mediated degradation pathway.9

Recently, it has been reported that complexes of cyclins and cyclin dependent kinases (CDKs) are involved in the development of human tumors.10 The overexpression of positive cell cycle regulators results in the elevation of CDK activities and subsequent functional inactivation of Rb by its hyperphosphorylation. Among these cell cycle regulators, cyclin E, which makes a complex with CDK2, plays a critical role in the progression of the cell cycle at the G1/S transition. The overexpression of cyclin E is implicated in the development of several cancers.11, 12 Although we have detected the strong induction of cyclin E gene expression in EWS-Fli1 transformed cells, the mechanism of transactivation of the cyclin E gene by EWS-Fli1 is still unknown.

In our study, we showed that the cyclin E gene might be 1 of the targets of EWS-Fli1 fusion protein in EFTs. The basic transcription factor Sp1 and E2F1 might be involved in the transactivation of the cyclin E gene by EWS-Fli1 fusion protein. In addition, flavopiridol, a pan-CDK inhibitory compound, efficiently inhibited cyclin E/CDK2 activity and the growth of EFT cells in vitro and in vivo.These results suggest that cyclin E might play an important role in unlimited growth of EFTs and that CDK inhibitors might be important for molecular targeting therapy against EFTs.

Material and methods

Reagents

EWS-Fli1 antisense and sense phosphorothioate oligodeoxynucleotides purified by high performance liquid chromatography were purchased from Kurabo Industries, Ltd. (Osaka, Japan). The sequence of the antisense oligonucleotides was ATCCGTGGACGCCATTTTCTCTCCT, and the corresponding sense sequence was used as a control.

Flavopiridol compound was kindly provided by Dr. Jose-Ramon, Aventis Pharma, Inc. (Bridgewater, NJ). A 10 mM stock solution of flavopiridol was prepared in DMSO and diluted in the culture media to working concentrations immediately before use. The final concentration of DMSO in the medium was less than 0.1%.

Cell culture

Human EFT cell lines SK-N-MC and RD-ES, and a murine fibroblast cell line MEF were obtained from American Type Culture Collection (Manassas, VA). A human EFT cell line WE-68 was kindly provided by Dr. F. van Valen. PNKT-1, a human primitive neuroectodermal tumor cell line, was established and characterized in our laboratory.13 BJ-1, a human normal fibroblast cell line was obtained from Clontech (Palo Alto, CA). WE-68 was grown in RPMI medium supplemented with 10% FBS, and the other cell lines were cultured in DMEM supplemented with 10% FBS at 37°C and 5% CO2.

Plasmids

Human cyclin E promoter/luciferase-reporter gene constructs, p10-4-Luc (the full length human cyclin E promoter) and pMUT(I + II + III)-Luc (−363/+1007, with point mutations in the 3 upstream E2F binding sites in the promoter) were kindly provided by Dr. E.A. Thompson.14 The other reporter gene constructs, DE1 (−159/+1007), DE2 (−7/+1007) and DE3 (+219/1007), were produced by the subcloning of PCR products into the HindIII/KpnI sites of pGl2 basic vector (Promega, Madison WI). An expression vector for human Sp1 protein, pCMV-Sp1, was kindly provided by Dr. R. Tjian.

Luciferase assay

EFT cell lines and MEF cells were seeded into 6-well plates (1 × 105 cells/well). Twenty-four hours later, transfection (1.0 μg plasmid/well) was carried out using FuGENE 6 (Roche Molecular Biochemicals, Indianapolis, IN) according to the manufacturer's protocol. For the forced expression experiments, MEF cells were cotransfected with 1 μg of each reporter construct and various doses of expression vector for EWS-Fli1 (kindly provided by Dr. C.T. Denny) and pCMV-Sp1. We used pGl2 basic vector and pRL-SV40 (Promega) as a positive control and an internal control for normalization of the transfection efficiency, respectively. The luciferase activity was assayed using Dual-Luciferase Reporter Assay Systems (Promega) and a Microlumat Plus LB 96V (EG&G Berthold, Germany).

Chromatin immunoprecipitation assay

Chromatin immunoprecipitation (ChIP) assays were carried out using a Chromatin Immunoprecipitation Assay Kit (Upstate Biotechnology, Waltham, MA), according to the manufacturer's protocol. Briefly, genomic DNA and protein were cross-linked by addition of formaldehyde to 2 × 106 WE-68 cells. The cells were lysed in SDS-lysis buffer with protease inhibitors and sonicated to generate DNA fragments of approximately 300–1,000 bp in size. After centrifugation, the cleared supernatant was incubated with anti-Fli1, anti-Sp1 (Santa Cruz Biotechnology, Santa Cruz, CA) and anti-E2F1 (Upstate Biotechnology) antibodies. A ChIP assay without antibody was performed as a negative control. The immune complexes were precipitated, washed and eluted according to the manufacturer's protocol. The DNA-protein cross-linking was reversed by heating and DNA was recovered by phenol/chloroform extraction and ethanol precipitation, followed by resuspension in 50 μl TE buffer. Five microliters of each sample was used as a template for PCR amplification. PCR amplification of the cyclin E promoter sequence containing 3 upstream E2F binding sites and 7 Sp1 sites was performed on the immunoprecipitated chromatin using the specific primers; 5′-GCCAGCCACGCGGCTTTTTGCCGC-3′ (forward) and 5′-CCAGCGAGGCGC AGGGACGGGGAA-3′; (reverse).

Protein extraction and Western blot analysis

Cell lysates from transfected cells or flavopiridol-treated cells were prepared and subjected to Western blot analysis as described previously.15 Briefly, 20 μg of each protein sample was electrophoresed on a 4–12% gradient polyacrylamide gel (Invitrogen, Carlsbad, CA) and transferred onto a nitrocellulose membrane (Amersham, Arlington Heights, IL). The membranes were immunoblotted with primary antibodies for Fli1, Sp1, cyclin E, cyclin A, CDK4, CDK2 (Santa Cruz), PARP-85 (Promega) and actin (Pharmingen, San Diego, CA) for 1 hr at room temperature. After several washes in TBST, the filters were treated with mouse or rabbit-radish peroxidase-conjugated secondary antibodies (Santa Cruz) at room temperature for 1 hr. After the final washes with TBST, the immunoreactivity of the blots was detected using an enhanced chemiluminescence (ECL) detection system (Amersham).

RNA extraction and quantitative real-time RT-PCR

Real-time RT-PCR (TaqMan PCR) was performed using an ABI PRISM 7700 Sequence Detection System (Perkin-Elmer Applied Biosystems, Foster City, CA). SK-N-MC cells (2 × 104) were cultured for 24 hr in 35 mm dishes (BD Falcon, Bedford, MA) and treated with 10 mM antisense or sense oligonucleotides against EWS-Fli1 for the required time. Total RNAs from the treated cells were isolated using RNeasy Mini kit (Qiagen, Valencia, CA), according to the manufacturer's instructions. The extracted total RNA (1 μg) was reverse-transcribed using Superscript II reverse transcriptase (Invitrogen). The following primers were used for PCR amplification of EWS-Fli1; forward primer, 5′-GGCAGCAGCCTCCCACTAG-3,′ and reverse primer, 5′-CCATGCTCCTCTTCTGACTGAGT-3′. The sequence of the TaqMan probe used to quantify EWS-Fli1 mRNA was 5′-(Fam) CCACCCCAAACTGGATAATACAGCC (Tamra)-3′. The probe and primer mixture for the human cyclin E gene amplification were purchased from ABI assay-on-Demand system. The standard cDNA templates, containing EWS-Fli1, cyclin E and glyceraldehyde-3-phosphate dehydrogenase (GAPD) sequences, were generated by PCR and subcloned into pCR2.1 TOPO vector (Invitrogen). The mixture of the primers and probe for GAPD amplification, which was used as an internal control, was purchased from PE Applied Biosystems. Triplicate 50 μl reactions containing 1 μl of cDNA template, 25 μl of 52 Taqman Universal Master mix, probe (0.1 μM final concentration) and primers (0.1 μM final concentration) were prepared and were cycled in 96-well plates according to the manufacturer's instructions. The PCR data were analyzed using SDS Software (Applied Biosystems). The relative amounts of EWS-Fli1 and cyclin E mRNAs that were standardized by the amount of GAPD mRNA were then calculated.

Cell viability assay

The cell viability assay was carried out using a CellTiter-Glo Luminescent Cell Viability Assay Kit (Promega) as previously described.16 Briefly, the cells (1.5 × 103/well) were grown in 96-well plates and treated with flavopiridol (50–5,000 nM) or DMSO (0.1% final concentration). Twenty-four hours later, the plates were equilibrated at room temperature for approximately 30 min, and an equal volume of Celltiter-Glo reagent was added to the volume of the cell culture medium present in each well. The plates were further shaken for 2 min to induce cell lysis. After a final incubation for 10 min at room temperature, the luminescence was measured using the Microlumat Plus LB 96V (EG&G Berthold).

Cell cycle analysis

Cell cycle analysis was performed as described previously.17 In brief, control and treated cells were harvested, washed in cold PBS and fixed in 70% ethanol. DNA was stained by incubating the cells in PBS containing propidium iodide (50 μg/ml) and RNase A (1 mg/ml) for 30 min at 37°C. Alterations in cell distribution were analyzed using an Epics-XL Flow Cytometer (Beckman Coulter, Fullerton, CA), and the cell proportions were calculated using EXPO32 software (Beckman Coulter). Ten thousand events were scored for each sample.

Immunoprecipitation and histone-H1 kinase assay

The in vitro kinase activity assay was carried out as described previously.18 Briefly, after completion of drug treatment, the cells were lysed in lysis buffer and disrupted by passing through a 26-gauge syringe 10 times. The lysates were then clarified by centrifugation (10 min at 10,000g). Two hundred micrograms of soluble protein were incubated with anti-cyclin E, anti-cyclin A, anti-cyclin D1 or anti-CDK2 antibody (Santa Cruz) at 4°C for 2 hr, respectively. The immune complexes were then precipitated with 40 μl of protein A/G PLUS-agarose beads overnight at 4°C and washed 3 times with lysis buffer and twice with kinase assay buffer. The kinase assay was carried out by combining the washed protein beads with 20 μl of kinase buffer with 10 μCi of [γ-32p] ATP, 15 μ ATP, and 50 μg/ml Histone-H1 (Roche). The reaction was allowed to proceed for 20 min at 30°C and was terminated by the addition of 10 μl of 3× Laemmli sample buffer and boiling for 5 min. The products were resolved by 12% SDS-PAGE. The gels were dried and the radioactivity was visualized using a Bio-Image analyzer (Fuji Photo Film, Tokyo, Japan) and then analyzed by NIH image software.

DNA fragmentation assay

The cells treated with various doses of flavopiridol were scraped and collected for the DNA fragmentation assay. The cells were dissolved in 100 μl of cell lysis buffer [10 mM Tris (pH 7.5), 10 mM EDTA (pH 8.0) and 0.5% Triton X-100] and were then incubated on ice for 10 min. The lysates were centrifuged at 16,000g for 5 min at 4°C. The supernatant was recovered and treated with RNase A (Roche) for 1 hr at 37°C and then with protease K (Roche) for 1 hr at 50°C. DNA was precipitated by the addition of isopropanol and 5 M NaCl to the mixture, and the precipitants were resuspended in TE buffer [10 mM Tris (pH 8.0)/1 mM EDTA]. The samples were electrophoresed on 1.2% agarose gels and stained with 0.5 mg/ml ethidium bromide (Sigma Chemical Co., St. Louis, MO) for 10 min to visualize the DNA ladders.

Animal experiments

WE-68 cells (1.5 × 107) were mixed with a collagen gel as 1:1 volume, and then injected s.c. in the flank of 5-week-old female athymic nu/nu mice. The treatment with flavopiridol was initiated 10 days after the tumor cell inoculation. Mice were then randomized, placed into 3 groups (N=4) and were treated by intraperitoneal injections of flavopiridol (1 mg/kg/day or 3 mg/kg/day) in a 100 μl volume for 5 days. The sizes of tumors were measured with calipers twice weekly, and tumor volume was calculated as length × width2 × 0.52. Statistical analysis was performed using a 1-factor ANOVA test. For demonstration of significance, Fisher's protected least significant difference test was used. p < 0.05 was considered as statistically significant.

Results

Downregulation of the cyclin E gene by inhibition of EWS-Fli1 expression in EFT cells

We have previously reported that the suppression of EWS-Fli1 expression by treatment with antisense oligonucleotides led to growth inhibition of EFT cells. In our study, the quantitative PCR analysis showed that in parallel with the down-regulation of EWS-Fli1 mRNA expression, the treatment of EFT cells by antisense oligonucleotides significantly downregulated the expression of cyclin E mRNA (Fig. 1b). Consistent with the mRNA data, the treatment with 20 μM of antisense oligonucleotides also downregulated cyclin E protein expression in SK-N-MC cells (Fig. 1a), as we have previously reported.7

Figure 1.

The effect of antisense oligonucleotides against EWS-Fli1 on the expression of cyclin E in SK-N-MC cells. (a) SK-N-MC cells were exposed to 20 μM sense or antisense oligonucleotides against EWS-Fli1 for 48 hr. The cell lysate was extracted and subjected to Western blot analysis using anti-Fli1, cyclin E and actin antibodies. The protein expression of EWS-Fli1 and cyclin E was inhibited by treatment with the antisense oligonucleotides. Ctr, untreated control; S, treated with sense oligonucleotides; AS, treated with antisense oligonucleotides. (b) Total RNAs were extracted from SK-N-MC cells treated with oligonucleotides and subjected to Quantitative real-time RT-PCR (TaqMan PCR). The mRNA expression levels relative to that in the untreated control cells are indicated. Columns represent means of 3 independent experiments, and bars represent S.E. Ctr, untreated control; S, treated with sense oligonucleotides; AS, treated with antisense oligonucleotides.

Activation of cyclin E promoter activity by EWS-Fli1 fusion protein

EWS-Fli1 is thought to be a member of the Ets family of DNA-binding proteins, which can bind to DNA in a sequence specific manner since the fusion protein has an intact Ets DNA binding domain derived from Fli1. However, no Ets binding sites are reported in the cyclin E gene promoter. It has previously been reported that there are 6 E2F and 12 Sp1 binding sites in the promoter, and that the 3 upstream E2F binding sites are important for the regulation of cyclin E gene expression.14, 19 Thus, we used the wild-type promoter and the promoter with mutations in E2F binding sites to investigate the effect of EWS-Fli1 on cyclin E promoter activity. The human cyclin E promoter-reporter construct, designated p10-4-Luc, included the nucleotides from −363 to +1007 of the promoter. In addition, pMUT(I + II + III)-Luc contained nucleotides −363 to +1007 of the human cyclin E promoter with point mutations in the 3 upstream E2F binding sites, which is known to inhibit the binding of E2F1 to the promoter (Fig. 2a). Each reporter construct was transiently transfected into mouse embryo fibroblast MEF cells with mock vector or EWS-Fli1 expression vector. The cells were harvested 36 hr after the transfection and subjected to luciferase assay. As shown in Figure 2b, when the cells were transfected with p10-4Luc and EWS-Fli1 vector, the promoter activity in the cells was increased to 4-fold that of those transfected with p10-4Luc and mock vector. When pMUT(I + II + III)-Luc construct was transfected to MEF cells instead of p10-4Luc, the activation of cyclin E promoter activity was also observed. However, the level of promoter activation was less than that in the p10-4Luc-transfected cells (p<0.05). We then transfected these 2 promoter constructs into various EFT cell lines and found that pMUT(I + II + III)-Luc also exhibited less promoter activity than the wild-type promoter p10-4Luc (p<0.05) (Fig. 2c). These results indicate that EWS-Fli1 can activate the cyclin E promoter. However, the 3 upstream E2F binding sites in the promoter might play a limited role in the EWS-Fli1-mediated cyclin E transactivation.

Figure 2.

The effect of EWS-Fli1 expression on cyclin E promoter activity. (a) Schema of cyclin E promoter/luciferase constructs. There are 6 wild-type E2F binding sites in the cyclin E promoter in p10-4-Luc, whereas pMut(I+II+III)-Luc has point mutations in the 3 upstream E2F binding sites. Open circle, the wild-type E2F binding site; filled circle, the mutated E2F binding site. (b) The Effect of EWS-Fli1 on the wild-type and the mutant cyclin E promoter constructs. Each promoter-reporter construct was cotransfected with the expression vector for full-length EWS-Fli1 or mock vector in MEF cells. The cells were harvested 36 hr after the transfection and assayed for luciferase activity. The dose-dependent activation of both the wild-type and mutated cyclin E promoter by EWS-Fli1 was demonstrated. Columns represent means of 3 independent experiments, and bars represent S.E. A Renilla luciferase expression vector, RL-SV40, was used as an internal control of the transfection efficiency. The relative luciferase activity between wild-type and mutant cyclin E promoter in MEF cells expressing EWS-Fli1 was statistically analyzed using an 1-factor ANOVA test. *p<0.05. (c) Cyclin E promoter activities in EFT cell lines SK-N-MC, WE-68, RD-ES and PNKT-1. The cells were transfected with each promoter-reporter construct and harvested 36 hr after the transfection for luciferase assay. Columns represent means of the triplicate experiments, and bars represent S.E. The relative luciferase activity between wild-type and mutant cyclin E promoter in the cells expressing EWS-Fli1 was statistically analyzed using a 1-factor ANOVA test. *p<0.05.

The effect of Sp1 binding sites on the activation of the cyclin E promoter by EWS-Fli1

We generated a series of promoter truncation constructs in order to define the elements in the cyclin E promoter that are responsible for its inducibility by EWS-Fli1 (Fig. 3a). The Sp1-binding sites and down stream E2F binding sites were serially deleted from the cyclin E promoter; i.e., DE1 (the 3 upstream E2F binding sites and 7 Sp1 sites adjacent to the 3 E2F sites were deleted), DE2 (the 3 upstream E2F binding sites and all of the Sp1 sites were deleted) and DE3 (5 E2F binding sites and all Sp1 sites were deleted). The constructs were then assayed for Sp1- and E2F-dependent activation of the promoter. When we transfected DE-1 into EFT cells, the cyclin E promoter activity was reduced to approximately 20% of the wild-type construct (Fig. 3a). Further deletions of all Sp1 binding sites from the promoter (DE-2 and -3) showed little change in luciferase activity from the promoter.

Figure 3.

The effect of EWS-Fli1 and Sp1 on cyclin E promoter activities in EFT and MEF cells. (a) Activities of the mutated or deleted cyclin E promoters in EFT cells. WE-68 cells were transfected with the constructs containing the wild-type promoter (p10-4-Luc), mutation in E2F binding sites (pMut(I+II+III)-Luc) or deletion of Sp1 binding sites (DE-1, -2 and -3). Open circle, the wild-type E2F binding site; filled circle, the mutated E2F binding site; thick line, Sp1 binding site. Columns represent means of the triplicate experiments, and bars represent S.E. The relative luciferase activity between p10-4-LUC and pMUT(I+II+III)-LUC in the cells expressing EWS-Fli1 was statistically analyzed using a 1-factor ANOVA test. *p<0.05. (b) The effect of EWS-Fli1 on the activities of the mutated or deleted cyclin E promoters in MEF cells. MEF cells were transfected with 1 μg of each reporter construct along with the expression vector of EWS-Fli1 or mock vector. Open circle, the wild-type E2F binding site; thick line, Sp1 binding site. Columns represent means of the triplicate experiments, and bars represent S.E. (c) The effect of EWS-Fli1 and Sp1 on the activity of the cyclin E promoter. MEF cells were transiently transfected with 1 μg of the p10-4-LUC and pMUT(I+II+III)-Luc construct with EWS-Fli1 and Sp1 expression vectors. The cells were harvested 36 hr after transfection and assayed for luciferase activity. EWS-Fli1 and Sp1 showed synergistic activation of the promoter of the cyclin E gene. These experiments were repeated at least 3 times with different plasmid preparations. Bars represent S.E. (d) Chromatin immunoprecipitation (ChIP) assay of the cyclin E promoter. A ChIP assay using a genomic fragment (nucleotide position from −362 to −133 of the promoter) containing 3 upstream E2F binding sites and 7 Sp1 binding sites was carried out in WE-68 cells. Immunoprecipitation was performed using anti-Fli1 (Fli1), anti-Sp1 (Sp1) and anti-E2F1 (E2F1) antibodies, followed by PCR amplification. Input represents the sonicated chromatin prior to immunoprecipitation. A ChIP assay without antibody (noAb) was performed as a negative control. (e) The expression status of EWS-Fli1 and Fli1 in WE-68 cells. EWS-Fli1 and Fli1 expressions were examined by Western blot analysis using anti-Fli1 antibody. There was no expression of endogenous Fli1 in WE-68 cells, which exhibited strong expression of EWS-Fli1 fusion protein. We used 293-T cells transfected with the expression vector for Fli1 (293-Flag-Fli1) as a positive control of Fli1 expression.

We also carried out forced expression experiments to investigate the effects of EWS-Fli1 on the deleted cyclin E promoter in MEFs. The deletion of 7 Sp1 binding sites adjacent to the 3 upstream E2F sites (DE-1) resulted in the loss of the effect of EWS-Fli1 on the activation of the cyclin E promoter (Fig. 3b). The further truncation of all of the Sp1 sites in the promoter also resulted in the loss of induction of promoter activity by EWS-Fli1 fusion protein. These results suggest that the sequences containing 7 Sp1 binding sites can mediate the response of the cyclin E promoter to EWS-Fli1.

The effect of EWS-Fli1 and Sp1 on the cyclin E promoter

We analyzed the effects of overexpression of Sp1 on the EWS-Fli1-mediated activation of the cyclin E promoter. EWS-Fli1 and Sp1 synergized to induce the promoter activity of pMUT(I + II +III) (Fig. 3c, right panel). EWS-Fli1 and Sp1 alone enhanced cyclin E promoter activity in MEF cells up to 3-fold and 4-fold that of control cells, respectively. However, when EWS-Fli1 expression vector was cotransfected with Sp1 expression vector into MEF cells, cyclin E promoter activity was increased to approximately 12-fold that of cells transfected with the mock vector. To exclude the possibility that E2F1 would affect the synergism, we repeated the experiments with the wild-type cyclin E reporter construct. As shown in Fig.3c (left panel), the result was almost same as that with the mutant construct. Thus we concluded that the transactivation of cyclin E promoter is mainly mediated by the synergy between EWS-Fli1 and Sp1.

We performed ChIP assays in order to determine the binding of EWS-Fli1 and Sp1 to the cyclin E promoter sequence in vivo. Immunoprecipitation of DNA-protein complexes were performed on cross-linked extracts from WE-68 cells using antibodies against Fli1, Sp1 and E2F1. We then measured the abundance of genomic DNA sequences of the cyclin E promoter within the immune complexes by PCR amplification. The Sp1-immunoprecipitated complex contained a similar amount of the amplifiable DNA sequence of the promoter as the E2F1-immunoprecipitated complex. This result suggests that E2F1 and Sp1 can bind to the cyclin E promoter through their binding sites in the promoter. Interestingly, the Fli1-immunoprecipitated complex also contained a slight amount of cyclin E promoter sequence. Thus, EWS-Fli1 might be able to interact with the sequence in the cyclin E promoter (Fig. 3d). We confirmed the expression of EWS-Fli1 and Fli1 in WE-68 cells in order to examine whether endogenous Fli1 participates in this interaction. Western blot analysis revealed that EWS-Fli1 fusion protein was expressed in WE-68 cells, whereas the expression of Fli1 was not detected. This suggests that EWS-Fli1, but not Fli1, was involved in the complex with the cyclin E promoter immunoprecipitated with the Fli1 antibody (Fig. 3e). Taken together, our results indicate that EWS-Fli1 and Sp1 might synergistically activate cyclin E promoter activity and that Sp1 binding sites in the promoter sequence might be important for the synergy.

Antiproliferative effect of the CDK inhibitor flavopiridol on EFT cells

We and others previously reported that both cyclin E and cyclin D1 were strongly induced in EWS-Fli1 transformed cells.7, 20 Therefore we hypothesized that CDK could be used as a target for novel molecular based therapy of EFTs, since the induced cyclins activate their counterpart CDKs and subsequently inhibit Rb function in cell cycle regulation. We used flavopiridol (HMR1275), a pan-CDK inhibitor, to investigate whether inhibition of the function of cyclin E could suppress the growth of EFT cells. Flavopiridol inhibited the proliferation of EFT cells in a dose-dependent manner (Fig. 4a). All EFT cell lines tested were sensitive to flavopiridol in the nanomolar range with IC50 values of 75–150 nM. On the other hand, human fibroblast cell line BJ-1 was not efficiently affected at these doses. These results indicate that flavopiridol might exert selective growth inhibitory effects on EFT cells. The time course effects of flavopiridol on BJ-1 and 2 EFT cell lines, SK-N-MC and WE-68, were then examined. The growth inhibition profiles over a 3-day exposure period are depicted in Figure 4b. The challenge of 150 nM flavopiridol remarkably inhibited the proliferation of EFT cells in a time-dependent manner. The growth inhibition of EFT cells by treatment with flavopiridol ranged from 93 to 99.0% of controls at 72 hr after treatment, whereas BJ-1 exhibited little inhibition of cell growth by flavopiridol.

Figure 4.

Anti-proliferative effect of flavopiridol on EFT cells. (a) Dose-response effects of flavopiridol on EFT cells. BJ-1, SK-N-MC, VH64, RDES, PNKT-1 and WE-68 cells (1.5–3×103/well) were seeded into 96-well plates and grown for 24 hr. The cells were further incubated for 24 hr in the presence of flavopiridol (50–5000 nM) or DMSO (0.1%). The cell viability assay was then performed as described in Material and Methods. The growth inhibitory effects of flavopiridol are shown as percent inhibitions of ATP-luminescence relative to the control. The results are the means of 3 independent experiments and the value of IC50 for each cell line was calculated. Bars represent S.E. (b) Time-course effect of flavopiridol on EFT cells. The human fibroblast cell line BJ-1 and EFT cell lines, SK-N-MC and WE-68, were cultured with either 0.1% DMSO (filled boxes) or 150 nM flavopiridol (filled triangles). The growth of BJ-1 cells was not inhibited by flavopiridol, whereas EFT cells showed significant growth inhibition with the treatment. The data represent the means of triplicate experiments and bars represent S.E. (c) The effect of flavopiridol on the cell cycle distribution of BJ-1 and EFT cells. The cells were treated with various concentrations of flavopiridol and subjected to flow cytometry analysis. The cell population in the sub-G1 DNA content in EFT cells dose-dependently increased with treatment of flavopiridol, whereas BJ-1 cells exhibited partial growth inhibition at the G2/M phase.

We also analyzed the cell cycle profiles of EFT cells exposed to flavopiridol. The treatment of BJ-1 cells by flavopiridol induced a partial inhibition of the cell cycle at the G2/M phase (Fig. 4c). In contrast, flavopiridol-treated EFT cells (SK-N-MC and WE-68) exhibited a significant increase in cell population in the sub-G1 fraction (37.7% and 58.4% at 300 nM, respectively), indicating the possible induction of apoptosis by flavopiridol in EFT cells.

Inhibition of cyclin E/CDK2 activity in EFT cells by flavopiridol

It was previously reported that cyclin E/CDK2 complex plays an essential role in the transition from G1 to S phase in the cell cycle progression.21 Since the effect of flavopiridol on the cell cycle might be, at least in part, important for the antiproliferative response of EFT cells, we examined whether the activity of cyclin E/CDK2 would be altered by flavopiridol. The immunoprecipitants of cyclin E and CDK2 were prepared from WE-68 cells treated with various concentrations of flavopiridol for 15 hr, and the in vitro immunocomplex kinase reactions were performed using Histone H1 as a substrate (Fig. 5a and b). The data demonstrated that the kinase activity of cyclin E/CDK2 in the flavopiridol-treated cells was decreased to approximately 15% of that in control cells (Fig. 5b, left panel). Total-CDK2 kinase activity was also calculated as the background of the kinase activity of cyclin E/CDK2 complex. The total kinase activity of CDK2 immunoprecipitated with CDK2 antibody was also decreased with high concentrations of flavopiridol (150 and 300 nM) (Fig. 5b, left panel). The results suggest that other cyclins/CDK2 complexes in EFT cells might also be inhibited by high doses of flavopiridol. However, since 75 nM flavopiridol could significantly inhibit the growth of WE-68 cells and the kinase activity of CDK2 when associated with cyclin E, cyclin E/CDK2 complex might be the main target of flavopiridol among cyclins/CDK2 for G1/S transition in EFT cells.

Figure 5.

CDK activity in EFT cells exposed to flavopiridol. (a) Two hundred mircrograms of total cell lysate was prepared from the cells treated with flavopiridol (0–300 nM) or DMSO and immunoprecipitated for CDK complexes. The histone H1 kinase reaction assay was carried out as described in Material and Methods. The reactions were resolved by SDS-PAGE and autoradiographed. The total CDK2 (CDK2) and cyclin E/CDK2 complex (Cyclin E) and cyclin D1/CDKs activities in WE-68 cells are shown (left panel). Western blots were also performed on the same immunoprecipitants to show the amounts of proteins (right panel). (b) The activities of total CDK2 and cyclinE/CDK2 and cyclin D1/CDKs complexes shown in (a) were quantified using the image analyzer. The CDK activities were significantly inhibited by flavopiridol in a dose-dependent manner. Data represent 3 independent experiments.

It is previously reported that cyclin D1 is 1 of the target genes of EWS-Fli1 and can be inhibited by flavopiridol.7, 22 Therefore, we also investigated the effect of flavopiridol on cyclin D1 in EWS-Fli1 expressing cells. As shown in Figure 5, although the expression of cyclin D1 protein was not affected by flavopiridol, the kinase activity of cyclin D1/CDKs was remarkably decreased in the flavopiridol-treated EFT cells.

Induction of apoptosis in EFT cells by flavopiridol

Recent studies have shown that flavopiridol has the ability to induce apoptosis in several human cancer cells.23, 24 Since we observed an increase in sub-G1 populations of EFT cells treated with flavopiridol in the cell cycle analysis, we investigated whether these cells really underwent apoptosis using a DNA fragmentation assay. When EFT cell lines were treated by flavopiridol, the formation of DNA ladders was observed in the cells (Fig. 6a). Western blot analysis also revealed that cleavage of nuclear protein poly (ADP-ribose) polymerase (PARP), an indicator of apoptosis,25 occurred in flavopiridol-treated EFT cells in a dose-dependent manner. On the other hand, the cleaved PARP was not detected in BJ-1 cells, although the treatment with flavopiridol exhibited partial growth inhibition of BJ-1 cells (Fig. 6b). These data demonstrate that there might be selectivity in the induction of apoptosis by flavopiridol between EFT and normal cells.

Figure 6.

Induction of apoptosis in EFT cells by flavopiridol. (a) Flavopiridol-induced DNA fragmentation in SK-N-MC and WE-68 cells. The cells were treated with flavopiridol (0–300 nM) for 24 hr. The dose-dependent formation of DNA ladders in the treated cells was observed. (b) PARP cleavage in flavopiridol-treated SK-N-MC and WE-68 cells. The cells were treated with flavopiridol (0–300 nM) for 24 hr. The expression of cleaved-PARP, an indicator of apoptosis, was dose-dependently induced in EFT cells by flavopiridol, whereas no PARP cleavage was observed in BJ-1 cells.

The effect of flavopiridol on EFT xenografts

Since flavopiridol displayed antiproliferative and apoptotic effects on EFT cell lines in vitro, we set out to investigate the antitumor activity of flavopiridol in EFT tumor xenografts. The subcutaneous tumors of WE-68 cells were established in BalbC nu/nu mice within 10 days after inoculation (7–9 mm in diameter) and were subsequently treated with an intraperitoneal injection of flavopiridol. After the administration of flavopiridol, the animals were monitored for changes in body weight and tumor growth for a further 3 weeks. An immediate effect of flavopiridol was detectable on Day 6 after the treatment, at which time the tumor growth in the treated group (3 mg/kg/day) showed a reduction to 34% of that in the control group. This reduction was sustained for 3 weeks after the treatment, and the tumor size in the mice treated with 3 mg/kg/day flavopiridol was approximately 60% smaller than that in the control animals (Fig. 7a). Dose-dependent growth inhibition of the xenografts was observed and there was a significant reduction of tumor growth in mice treated with 3 mg/kg/day of flavopiridol (p<0.05) (Fig. 7b). The drug toxicity was modest since the animals in the treated group did not present significant weight loss (data not shown).

Figure 7.

Antitumor activity of flavopiridol in vivo. (a) WE-68 cells were xenografted in Balb/C nu/nu mice. The animals were treated (intraperitoneally for 5 days) with either 100 μl of flavopiridol (1 mg/kg/day, N=4; 3 mg/kg/day, N=4) or an equivalent volume of saline (N=4). The treatment period (from day 1 to day 5) is indicated. The tumor sizes in both groups were assessed twice weekly and the tumor weights were calculated as described in Methods. Bars represent S.E. The statistic significance between control group and 3 mg/kg/day group after Day 6 from the treatment was demonstrated using an 1-factor ANOVA test. *p<0.05. (b) Tumor sizes measured on day 21 from treatment. The tumor growth was dose-dependently inhibited by treatment with flavopiridol. There was a significant difference in tumor sizes between controls and the 3 mg/kg/day groups (p<0.05). Data represent the mean of 4 mice in each group. Bars represent S.E.

Discussion

In the present study, we demonstrated that EWS-Fli1 upregulates the expression of cyclin E via modulation of the promoter activity of the cyclin E gene. Together with our previous report, our data strongly suggest that the unbalanced expression of G1/S regulatory factors in the cell cycle caused by EWS-Fli1 may lead to the tumorigenesis of EFTs.

CDKs play a pivotal role in controlling progression of the cell cycle.26 CDKs form complexes with their regulatory subunits, cyclins, to function as kinases of specific proteins at different phases of the cell cycle. Current evidence suggests that the S-phase promoting function of cyclin D and cyclin E is mediated by the kinase activities of cyclin-CDK complexes to phosphorylate Rb protein (pRb) to release E2F from an inactive Rb-E2F complex.27, 28 It has been reported that, at least in some instances, the expression of cyclin E is able to substitute for cyclin D29 and that cyclin E itself is 1 of the target genes of E2F.14, 19 When the inactivation of Rb occurs via its phosphorylation and E2F is subsequently activated, E2F upregulates the expression of target genes including cyclin E gene required for the G1/S transition of the cell cycle. The increased cyclin E expression results in an enhancement of the activity of cyclin E/CDK2 complex, leading to further phosphorylation of Rb and the reinforcement of the cell cycle.30 We have previously found that EWS-Fli1 induces the expression of the E2F1 gene via upregulation of promoter activity, suggesting that E2F1 might be a target gene of EWS-Fli1 fusion protein (Tanaka et al., unpublished data). In our study, we demonstrated that the introduction of mutations in the 3 E2F binding sites in the cyclin E promoter partially inhibited the promoter activity. Thus, there is a possibility that EWS-Fli1 might indirectly transactivate the cyclin E promoter through overexpression of E2F1 interacting with the 3 upstream E2F binding sites.

We further demonstrated the 7 Sp1 biding sites, which were adjacent to the 3 E2F sites, and they appeared to be more important for the activation by EWS-Fli1 of the cyclin E promoter. There could be several possible roles of Sp1 in the transactivation of the cyclin E gene. It has been shown that Sp1 and E2F1 act synergistically to induce transcription of murine cyclin E and DHFR genes and that the synergistic activation of the target genes might be mediated by a physical interaction between Sp1 and E2F1.31, 32, 33 Sp1 can also transactivate or interact with several G1/S regulators and promote G1/S transition and S phase progression in the cell cycle.34 Therefore, Sp1 might associate with E2F1 and/or other G1/S regulators in the EWS-Fli1-mediated transactivation of the cyclin E gene. On the other hand, Ets transcription factor cooperates with Sp1 to activate promoters of several cell cycle regulator genes.35 Notably, a recent report demonstrated that Sp1 could recruit EWS-Ets fusion protein, including EWS-Fli1 and EWS-ER81, to gene promoters by protein–protein interaction, thereby providing one possible explanation why EWS-Ets fusion protein does not require direct DNA binding for the promoter activation of several genes.36 In our study, there was a strong interaction of Sp1 and E2F1 with the cyclin E promoter as demonstrated by ChIP assay. Interestingly, the complex immunoprecipitated with the Fli1 antibody from WE-68 cells, which did not express Fli1, and also exhibited a small amount of the cyclin E promoter sequence, suggesting the possible interaction of EWS-Fli1 with the cyclin E promoter. These data suggest that EWS-Fli1 might associate with the cyclin E promoter via an interaction with Sp1 and/or direct binding to the promoter. Further study is required to elucidate the precise mechanisms through which EWS-Fli1 fusion protein synergistically activates the cyclin E promoter with Sp1 and E2F1.

Despite the advances in multimodal therapy or the application of high-dose chemotherapy followed by peripheral blood stem cell transfusion for EFTs,37 the survival rate of the metastatic or recurrent cases is still poor. To improve the outcome of the disease, several gene-targeting therapies have been considered for EFTs, including the introduction of viral vectors encoding antisense-EWS-Fli1 or treatment with antisense oligonucleotides against EWS-Fli1.7, 38 However, it is still difficult to deliver high amounts of these molecules specifically to the tumor cells to achieve a safe and effective tumor response.

We have previously demonstrated that EWS-Fli1 inhibited the expression of p21WAF1/CIP1 and p27KIP1, whereas the fusion protein enhanced the expression of cyclin E and cyclin D1.7, 38 In addition, loss of the p16 gene or suppression of the expression of p57KIP2 might lead to EWS-Fli1-mediated transformation of EFTs.39, 40 Since all of the alterations correlate with the activation of CDKs, we hypothesized that inactivation of CDKs may be effective as a molecular-targeting therapy against EFTs. A pan-CDK inhibitor flavopiridol is the first small-molecule CDK modulator tested in a clinical trial.41 Recent reports have indicated that flavopiridol suppresses the activity of CDKs by interacting with the ATP-binding site of CDKs,42 and that flavopiridol promotes cell cycle arrest, apoptosis, differentiation and antiangiogenic properties against several types of cancers. In our study, flavopiridol effectively inhibited the growth of EFTs via the inhibition of cyclin E/CDK2 kinase activity. We also tested whether the kinase activity of other major CDK complexes was also inhibited by flavopiridol. It is well known that CDK4/CDK6 bind to cyclin D1 and promote progression through mid G1 of the cell cycle, and that cyclin A/CDK2 and cyclinE/CDK2 complexes promote the progression of G1/S transition and the S-phase.43. In our study, cyclin D1/CDKs activity was remarkably decreased in the flavopiridol-treated EFT cells. However, cyclin A/CDK2 activity was not affected by flavopiridol (data not shown). Therefore, the stimulation of cell cycle progression in EFT cells might be mainly mediated via cyclin E and cyclin D1 induced by EWS-Fli1 fusion protein.

It is still unclear whether the inhibition of CDK activity is required for the induction of apoptosis by flavopiridol. We also investigated the effects of flavopiridol on apoptosis induced in EFT cells. The apoptosis mediated by flavopiridol in our study appeared to be independent of p53, since apoptosis occurred in all the cell lines tested in which p53 was deleted or was inactivated by mutation.44 The same phenomena have been reported in lung cancer45 and oral squamous cell carcinoma.24 The Bcl-2 family of proteins consisting of both inhibitors and promoters of apoptosis, are postulated to play a central role in the cytotoxicity of flavopiridol and the downregulation of some anti-apoptotic proteins, including Bcl-2, XIAP and Bcl-xL.24, 46, 47. In our study, the expressions of Bcl-2, XIAP and Bcl-xL proteins were not affected by flavopiridol, whereas the protein expression of an anti-apoptotic factor Mcl-1 decreased in flavopiridol-treated SK-N-MC and WE-68 cells (data not shown). Although the mechanisms of the induction of apoptosis by flavopiridol require further investigation, this is the first report demonstrating significant antitumor effects of a CDK inhibitor on EFTs.

In summary, we demonstrated that EWS-Fli1 fusion protein could transactivate the cyclin E gene in EFTs mainly via Sp1 binding sites and partially via E2F sites in the promoter. This may partly explain the tumorgenetic activity of EWS-Fli1 fusion protein. In addition, we also found that flavopiridol, a pan-CDK inhibitor, can inhibit the growth of EFT cells in vitro and in vivo by the inhibition of cyclin E/CDK2 kinase activity and the induction of apoptosis. Thus, CDK inhibitors including flavopiridol might have potential for testing in clinical trials for treatment of cases with advanced EFTs.

Acknowledgements

We thank Dr. F. van Valen for the WE-68 cells, Dr. E. Aubrey Thompson for the human cyclin E promoter constructs, p10-4-Luc and pMUT(I + II + III)-Luc, Dr. R. Tjian for pCMV-Sp1, and Dr. C.T. Denny for the expression plasmid for EWS-Fli1. We also thank Dr. Y. Maehara for the analyses with the ABI PRISM 7700 Sequence Detection System.

Ancillary