The anti-apoptotic protein Bcl-2 is overexpressed in a majority of breast cancers, and is associated with a diminished apoptotic response and resistance to various anti-tumor agents. Bcl-2 inhibition

The antiapoptotic protein Bcl-2 is overexpressed in amajority of breast cancers, and is associated with adiminished apoptotic response and resistance to variousantitumor agents. Bcl-2 inhibition is currently beingexplored as a possible strategy for sensitizing breastcancer cells to standard chemotherapeutic agents. Anti-sense Bcl-2 oligonucleotides represent one method forblocking the antiapoptotic effects of Bcl-2. In this study,we show that antisense Bcl-2 efficiently blocks Bcl-2expression, resulting in the apoptosis of breast cancercells. Antisense Bcl-2-mediated cytotoxicity was associ-ated with the induction of the B cell translocation gene 1(BTG1 ). Importantly, knockdown of BTG1 reducedantisense Bcl-2-mediated cytotoxicity in breast cancercells. Furthermore, BTG1 expression seems to be nega-tively regulated by Bcl-2, and exogenous expression ofBTG1 induced apoptosis. These results suggest thatBTG1 is a Bcl-2-regulated mediator of apoptosis in breastcancer cells, and that its induction contributes toantisense Bcl-2-mediated cytotoxic effects. [Mol CancerTher 2006;5(6):1593–601]Introduction Chemoresistance is a major clinical problem in thetreatment of breast cancer, and is associated withdecreased apoptosis in response to chemotherapeuticagents (1). Apoptosis occurs through two major path-ways—the extrinsic or cytoplasmic pathway, which isregulated by the Fas death receptor, and the intrinsicpathway, which is controlled in part by the Bcl-2 family ofproteins (2, 3). This family is composed of variousproapoptotic and antiapoptotic proteins that heterodimer-ize and modulate each other’s function. Thus, the relativeconcentration of each Bcl-2 family member is thought todetermine whether cell suicide will occur. The ratio ofantiapoptotic Bcl-2 to proapoptotic Bax is a criticaldeterminant of apoptosis, as Bcl-2 heterodimerizes withBax, blocking apoptosis (4).Overexpression of Bcl-2 occurs in 40% to 80% of primaryinvasive breast carcinomas, depending on the methods ofmeasurement and quantification (5–9). Although Bcl-2expression is associated with estrogen receptor–positivedisease, which has a highly favorable prognosis, itsoverexpression is associated with a diminished apoptoticresponse and resistance to various antitumor agents(10–12). Thus, the therapeutic suppression of Bcl-2 levelsis a feasible approach toward improving the outcome ofcurrent standard chemotherapies in breast cancer.Therapeutic approaches targeting Bcl-2 include antisensestrategies to inhibit translation of the Bcl-2 protein andsmall-molecule inhibitors that recognize the surface pocketof Bcl-2, blocking the interaction of Bcl-2 with the BH3domain of related death agonists such as Bax. AntisenseBcl-2 has been shown to effectively suppress Bcl-2expression in vitro and in vivo (13–16), and to promotechemosensitization and tumor regression of human breastcancer xenografts (13, 14). Interestingly, the quantity ofendogenous Bcl-2 was not a critical determinant ofantisense efficacy in breast cancer models, as cell survivaland Bcl-2 levels were reduced by >80% in breast cancerlines expressing high (MCF-7) or low (MDA435/LCC6)Bcl-2 levels (15, 16).In the present study, we sought to determine thebiological effects of antisense Bcl-2 in breast cancer cells,and to investigate the molecular mechanisms associatedwith these effects. We found that antisense Bcl-2-mediatedcytotoxicity of breast cancer cells was associated withinduction of the B cell translocation gene 1 (BTG1).Knockdown of BTG1 reduced antisense Bcl-2-mediatedcytotoxicity, suggesting that the induction of BTG1 is animportant molecular mechanism contributing to the cyto-toxic effects of antisense Bcl-2 therapy. In addition, BTG1expression was suppressed by Bcl-2, and BTG1 inducedapoptosis, suggesting that BTG1 is a Bcl-2-regulatedmediator of apoptosis in breast cancer cells.Received 3/13/06; accepted 4/19/06. Grant support: U.S. Department of Defense Concept awardW81XWH0510419 (R. Nahta); NIH grant K23 CA82119 (F.J. Esteva);funds from the University Cancer Foundation at the University of TexasM.D. Anderson Cancer Center (F.J. Esteva and R. Nahta); the Nellie B.Connally Breast Cancer Research Fund, which supports the Breast CancerTranslational Research Laboratory at M.D. Anderson Cancer Center; andNIH Cancer Center Support grant CA 16672-27. The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely toindicate this fact. Requests for reprints: Rita Nahta or Francisco J. Esteva, Department ofBreast Medical Oncology, Unit 1354, The University of Texas M.D.Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX77030-4009. Phone: 713-792-2817; Fax: 713-563-0739.E-mail: [email protected] or [email protected] Copyright C 2006 American Association for Cancer Research. doi:10.1158/1535-7163.MCT-06-01331593 Mol Cancer Ther 2006;5(6). June 2006 Materials andMethodsMaterialsThe antisense Bcl-2 molecule used here is an 18-merphosphorothioate DNA oligonucleotide that is complemen-tary to the first six codons of the human Bcl-2 open readingframe (G3139, Genasense, oblimersen sodium), and thecontrol oligonucleotide G4126 contains a 2 bp mismatchversus G3139. Antisense Bcl-2 and control oligonucleotidewere provided by Genta, Inc. (Berkeley Heights, NJ) at100 Amol/L stock concentrations. The sequences of theoligonucleotides are as follows: Bcl-2 antisense G3139(5¶-TCTCCCAGCGTGCGCCAT-3¶); mismatch controlG4126 (5¶-TCTCCCAGCATGTGCCAT-3¶). Small interferingRNA (siRNA) against BTG1 was purchased from Dharma-con (Lafayette, CO) with the following sequence: sensestrand, 5¶-UUGUUGGGUCUCACACUCAA-3¶; antisensestrand, 5¶-CAACCCAGAGUGUGAGUUCUU-3¶. Negativecontrol siRNA was purchased from Ambion (Austin, TX).OligofectAMINE transfection reagent (Invitrogen, Carlsbad,CA) was used according to the manufacturer’s guidelines.Cell CultureEstrogen receptor–positive MCF-7 and MDA-231 (MDA-MB-231) breast cancer cells were purchased from theAmerican Type Culture Collection (Manassas, VA). Bcl-2-overexpressing MDA-231 clones (referred to as clone 4 andclone 5 cells) and the Neo control stable transfectant werecreated as previously described (17). Briefly, the pCl-neo/Bcl-2 (CMV-Bcl-2) plasmid encoding the full-length Bcl-2coding sequence or the backbone vector pCl-neo wastransfected into MDA-231 cells with subsequent neomycin(G418) antibiotic selection and screening of clones byimmunoblot. All cells were maintained in DMEM supple-mented with 1% penicillin-streptomycin and 10% FCS.ImmunofluorescenceMCF-7, MDA-231, and MDA-231 Bcl-2-overexpressingclone 5 cells were seeded at 10,000 cells per well in chamberslides. Cells were transfected the next day using Oligofect-AMINE with antisense Bcl-2 oligonucleotide labeled with6-fluorescein on the 5¶-T residue (FAM-G3139) at dosesranging from 150 to 600 nmol/L. After 48 hours, cells werefixed in 4% formaldehyde, washed with PBS, and viewedby fluorescence microscopy.ImmunoblottingTotal protein lysates were obtained using 1% NP40 lysisbuffer [150 mmol/L NaCl, 50 mmol/L Tris (pH 8), 1%NP40], and immunoblotted (50 Ag) with an anti-Bcl-2monoclonal antibody (Oncogene Research Products, SanDiego, CA) diluted 1:1,000, anti-BTG1 polyclonal (N-20;Santa Cruz Biotechnology, Santa Cruz, CA) used at 1:100, oran anti-h-actin monoclonal antibody (Santa Cruz Biotech-nology) at 1:5,000. Secondary antibodies were chosenaccording to species of origin and detected using theOdyssey Imaging system (Li-Cor Biosciences, Lincoln, NE).Analysis of Cell DeathCell Survival Assays. MCF-7, MDA-231 parental, andMDA-231 clone 5 cells were seeded at 5 10 cells/well in12-well dishes. After 24 hours, cells were transfected with2-fold serial dilutions of mismatch control oligonucleotideor antisense Bcl-2 using OligofectAMINE reagent. Trans-fection control cultures were treated with OligofectAMINEalone. Cells were trypsinized after 72 hours and counted bytrypan blue exclusion. For each dose, the ratio of viableantisense Bcl-2-transfected cells to viable mismatch controloligonucleotide–transfected cells was determined, andthen graphed as a percentage of viable cells in theOligofectAMINE control group. All experiments were donein triplicate (at least twice). Error bars represent SDbetween replicates.To establish the importance of BTG1 to antisense Bcl-2-mediated cytotoxicity, 400 nmol/L of control siRNA or 400nmol/L of BTG1 siRNA were transfected into MCF-7 cells.After 24 hours, 400 nmol/L of mismatch control oligonu-cleotide or 400 nmol/L of antisense Bcl-2 was transfectedfor an additional 48 hours, at which point cells weretrypsinized and counted by trypan blue exclusion. Cellviability was determined as a percentage of the controlsiRNA/mismatch control oligonucleotide transfectiongroup. All experiments were done in triplicate (at leasttwice). Error bars represent SD between replicates.DNA Fragmentation Cell Death Detection ELISA. Toestablish the importance of BTG1 to antisense Bcl-2-mediated cytotoxicity, 400 nmol/L of control siRNA or400 nmol/L of BTG1 siRNA were transfected into MCF-7cells. After 24 hours, 400 nmol/L of mismatch controloligonucleotide or 400 nmol/L of antisense Bcl-2 wastransfected for an additional 48 hours, at which pointprotein lysates were obtained and analyzed for cytoplasmichistone-associated DNA fragments (mononucleosomes andoligonucleosomes) representative of apoptosis using theCell Death Detection ELISA (Roche, Indianapolis, IN)according to manufacturer guidelines. Absorbance wasmeasured at 490 nm. The DNA fragmentation enrichmentfactor was determined as a ratio of absorbance per group toabsorbance of the control siRNA/mismatch control oligo-nucleotide group. All experiments were done in triplicate,with error bars representing the SD between replicates.To establish the effect of BTG1 on apoptosis, MCF-7 cellswere transfected with an empty control vector or pXT-BTG1 expression plasmid (a gift from Dr. Cabello,Montpellier, France; ref. 18). After 72 hours, protein lysateswere obtained and analyzed for cytoplasmic histone-associated DNA fragments (mononucleosomes and oligo-nucleosomes) representative of apoptosis using the CellDeath Detection ELISA (Roche) according to manufacturerguidelines. Absorbance was measured at 490 nm. TheDNA fragmentation enrichment factor was determined asa ratio of absorbance of the BTG1-transfected group to theabsorbance of the control transfectant group. All experi-ments were done in triplicate, with error bars representingthe SD between replicates.MicroarrayAnalysisMCF-7 cells were transfected with 200 nmol/L ofmismatch control oligonucleotide for 24 hours or 200, 400,and 600 nmol/L of antisense Bcl-2 oligonucleotide for 24and 48 hours. Total RNA was extracted using the RNeasyBTG1, a Bcl-2-Regulated Mediator of Apoptosis1594 Mol Cancer Ther 2006;5(6). June 2006 Mini Kit (Qiagen, Valencia, CA), converted to cRNA andhybridized (10 Ag per sample) onto human genome arrayU133A (Affymetrix, Santa Clara, CA) containing f16,900well-characterized gene sequences. Hierarchical gene clus-ter analysis was done, in which genes were filteredaccording to a mean log expression level >5.5.ReverseTranscriptase-PCRMCF-7, MDA-231 parental, and clone 5 cells weretransfected for 48 hours with 400 of nmol/L mismatchcontrol oligonucleotide or 400 of nmol/L antisense Bcl-2oligonucleotide. Total RNA was extracted using RNeasyMini Kit (Qiagen). Expression levels of h-actin, RTP801,and BTG1 were measured by reverse transcription-PCR(RT-PCR). For the reverse transcription procedure, 200 ngof thawed RNA from the extracted RNA was combinedwith 0.5 AL of RNase inhibitor (Sigma, St. Louis, MO),1.0 AL of anchored oligo (dT)23 (Sigma), and 6.5 AL ofRNase-free dH2O. This mixture was incubated for10 minutes at 70jC. This solution was then added to2.0 AL of reverse transcriptase buffer, 1.0 AL of 5 mmol/Ldeoxynucleotide mix (combination of dATP, dCTP, dGTP,dTTP—5 mmol/L of each), 1.0 AL of Sensiscript reversetranscriptase (all from Qiagen Sciences, Baltimore, MD),and 6.0 AL of RNase-free dH2O. This was followed byincubation at 25jC for 15 minutes, then 50 minutes ofincubation at 42jC.The following primers (Sigma Genosys, The Woodlands,TX) were used for PCR: RTP801 forward, 5¶-GGGGTAC-CATGCCTAGCCTTGG-3¶; RTP801 reverse, 3¶-TAAAGCGG-CCGCTCACAACATGTCAATGAGCAGCTG-5¶; BTG1forward, 5¶-ACTAGTAAGCATGACCTGGGGA-3¶; BTG1reverse, 5¶-ACAAAATAGATGGTGGTTTGTGG-3¶; h-actinforward, 5¶-GCGGGAAATCGTGCGTTGACATT-3¶; h-actinreverse, 3¶-GTGCTTTGATGGAAGTTGAGGTAG-5¶. ThePCR mixture consisted of 3.0 AL of cDNA from the abovereverse transcriptase procedure, 5.0 AL of PCR buffer (100mmol/L Tris-HCl, 500 mmol/L KCl, 15 mmol/L MgCl2,and 0.01% gelatin; Sigma), 1.0 AL of 10 mmol/L deoxy-nucleotide mix (combination of dATP, dCTP, dGTP,dTTP—10 mmol/L of each; Sigma), 1.0 AL of REDTaqDNA Polymerase (Sigma), 38.0 AL of RNase-free dH2O,1.0 AL each of 15 mmol/L forward and reverse primers.For PCR conditions, the mixture was incubated in a PCRMasterCycler Personal (Eppendorf Scientific, Inc., West-bury, NY) with the first denaturation at 94jC for 2 minutes,the following denaturation at 94jC for 15 seconds, anneal-ing at 60jC for 1 minute, extension at 72jC for 1 minute andfinal extension at 68jC for 5 minutes for a total of 35amplification cycles. The final PCR product was visualizedin a 2% agarose gel containing ethidium bromide. Quanti-tation of bands was calculated as the band net intensity, andwas done using the Kodak DC-290 digital camera withKodak 1D Image Analysis software (Kodak, Rochester, NY).The band net intensity for the PCR product band of RTP801or BTG1 was divided by the band net intensity of the h-actinPCR product band for that RNA sample run in a parallelPCR reaction at the same time to yield a relative band netintensity ratio as previously described (19).Luciferase ReporterAssayMDA-231 Neo, Bcl-2 clone 4, and Bcl-2 clone 5 cells weretransfected with 1 Ag of pGL3-BTG1 (a gift from Dr. vonLindern, Erasmus Medical Center Rotterdam, Rotterdam,the Netherlands), which encodes the 1033/+82 BTG1promoter (and part of the first exon) upstream of theluciferase reporter gene, as previously described (20), or1 Ag of pGL3 control luciferase reporter plasmid (Promega,Madison, WI). All cells were also transfected with 1 ngof hRL4 (Promega), a Renilla luciferase plasmid. After48 hours, cells were lysed and luciferase values weremeasured using the Dual Luciferase Reporter Kit (Prom-ega) and luminometer. All pGL3 or pGL3-BTG1 promoterluciferase values were normalized to the Renilla luciferasevalue per group. The ratio of normalized BTG1 promoter-luciferase to normalized pGL3 control luciferase was thendetermined, and graphed as the BTG1 promoter-luciferaseactivity. ResultsAntisenseBcl-2Down-Regulates Bcl-2 andDecreasesBreastCancerCellViabilityThe MDA-231 breast cancer cell line, stable control, andBcl-2 transfectants derived from MDA-231 (17), and MCF-7breast cancer cells were used for our study. Parental MDA-231 and control Neo MDA-231-derived stable transfectantcells express relatively low levels of endogenous Bcl-2,whereas MCF-7 cells overexpress Bcl-2 (Fig. 1A). The stableMDA-231-derived clone 5 cells overexpress Bcl-2 at a levelsimilar to MCF-7 cells, whereas clone 4 cells showsignificantly higher levels. Using these cell lines as modelsof Bcl-2 overexpression in breast cancer, we studied thecellular and molecular effects of antisense Bcl-2 treatment.Although the antisense Bcl-2 used in this study has beenshown to down-regulate Bcl-2 in various breast cancer celllines (15, 16), we wanted to ensure that the cell lines used inthis study were able to internalize the antisense Bcl-2oligonucleotide with subsequent down-regulation of Bcl-2.The breast cancer cell lines MCF-7, MDA-231 parental, andMDA-231-derived Bcl-2-overexpressing stable clone 5 weretransfected with antisense Bcl-2 oligonucleotide labeledwith 6-fluorescein on the 5¶-T residue (FAM-G3139). After48 hours, cells were fixed and viewed by fluorescencemicroscopy (Fig. 1B). All cell lines efficiently internalizedantisense Bcl-2 at low concentrations (200–250 nmol/L),confirming that this antisense Bcl-2 could be delivered tobreast cancer cells in vitro using liposome-mediatedtransfection.Next, Bcl-2 down-regulation by antisense Bcl-2 wasassessed by immunoblotting (Fig. 1C). MCF-7 cells showeddecreased Bcl-2 levels upon transfection with 600 nmol/Lof antisense Bcl-2. Similar to MCF-7 cells, concentrations of400 to 800 nmol/L of antisense Bcl-2 reduced Bcl-2 proteinlevels in MDA-231 parental cells. In contrast, clone 5 cellsrequired much higher doses (1,500 nmol/L) to down-regulate Bcl-2. Complete suppression of Bcl-2 expressionwas not achieved in clone 5 cells even with such high dosesMolecular Cancer Therapeutics 1595 Mol Cancer Ther 2006;5(6). June 2006 of antisense Bcl-2. However, MCF-7 cells, which expresssimilar levels of Bcl-2 as clone 5 cells, showed moresignificant down-regulation of Bcl-2 upon treatment with alower dose (600 nmol/L) of antisense Bcl-2. Thus, factorsother than baseline Bcl-2 levels seem to affect antisense-mediated down-regulation of Bcl-2.The dose-response profiles of MCF-7, MDA-231 parental,and clone 5 cells to antisense Bcl-2 were obtained by trypanblue exclusion assays. Cells were transfected with 2-foldserial dilutions of mismatch control oligonucleotide orantisense for 72 hours. For each dose, the ratio of viableantisense Bcl-2-transfected cells to viable mismatch controloligonucleotide–transfected cells was determined, and thisratio was then graphed as a percentage of the viable cells inthe OligofectAMINE control group. MCF-7 cells showed anIC50 (50% inhibition of viability) at f400 to 600 nmol/L(Fig. 1D), which is consistent with immunoblots thatshowed reduced Bcl-2 expression in MCF-7 cells treatedwith 600 nmol/L antisense Bcl-2 (Fig. 1C). Also consistentwith immunoblots of Bcl-2 down-regulation, MDA-231parental cells showed significantly higher sensitivity toantisense Bcl-2, with an IC50 of f200 nmol/L andessentially all cells nonviable by 1,600 nmol/L, comparedwith clone 5 cells, which maintained f50% cell viability at1,600 nmol/L (Fig. 1E).Antisense Bcl-2 Induces the Expression of BTG1To better understand the molecular mechanisms bywhich antisense Bcl-2 induces a cytotoxic response inbreast cancer cells, we did microarray analysis of MCF-7cells transfected with antisense Bcl-2 in comparison toMCF-7 cells transfected with mismatch control oligonu-cleotides. Hierarchical gene cluster analysis was done toexamine relative changes in gene expression per treatmentgroup (Fig. 2A). Only 10 genes were predicted by micro-array to be induced by 48 hours of antisense Bcl-2transfection (Table 1). Gene induction did not seem to be Figure 1. Antisense Bcl-2 down-regulates Bcl-2 protein levels and inhibits cell survival. A, to assess baseline expression of Bcl-2 protein, MDA-231parental, MDA-231/Neo control stable transfectant, MDA-231/Bcl-2 clone 4 stable transfectant, MDA-231/Bcl-2 clone 5 stable transfectant, and MCF-7cells were lysed for protein and immunoblotted (50 Ag) for Bcl-2 and actin.B, to show internalization of antisense Bcl-2, cells were transfected withantisense Bcl-2 labeled with 6-fluorescein (FAM-G3139) at doses ranging from 150 to 600 nmol/L. After 48 h, cells were fixed in 4% formaldehyde,washed with PBS, and viewed by fluorescence microscopy. Representative fields are shown for MCF-7 cells transfected with 250 nmol/L of FAM-G3139,MDA-231 parental cells with 200 nmol/L of FAM-G3139, and clone 5 cells with 200 nmol/L of FAM-G3139.C, MCF-7 and MDA-231 parental and clone5 cells were transfected with mismatch control oligonucleotide or antisense Bcl-2 at doses ranging from 200 to 1,500 nmol/L for 72 h or treated withOligofectAMINE alone (O) for 72 h. Total protein lysates were immunoblotted for Bcl-2 and actin. MCF-7(D) and MDA-231 (E) parental and clone 5 cellswere transfected with mismatch control oligonucleotide or antisense Bcl-2 at doses ranging from 200 to 1,500 nmol/L for 72 h. Cell viability was assessedby trypan blue exclusion. For any given concentration of antisense Bcl-2 oligonucleotide, cell viability of antisense Bcl-2-transfected cells is shown as apercentage of cells transfected with the same dose of mismatch control oligonucleotide. All assays were done in triplicate and repeated at least thrice.Bars, SD between replicates.BTG1, a Bcl-2-Regulated Mediator of Apoptosis1596 Mol Cancer Ther 2006;5(6). June 2006 dose-dependent. Two of the genes have a putative role ingrowth arrest and apoptosis. RTP801 is a hypoxia-induciblefactor-I–responsive gene whose overexpression is associ-ated with apoptosis, particularly in response to cellularstress (21). According to microarray studies, the RTP801transcript was induced f5-fold in antisense Bcl-2-treatedMCF-7 cells (Table 1). RT-PCR analysis of MCF-7, MDA-231 parental, and clone 5 cells transfected for 48 hours withFigure 2. Antisense Bcl-2 induces expression of BTG1. A, MCF-7 cells were treated with OligofectAMINE alone (lane 6 ), transfected with 200, 400, or600 nmol/L of antisense Bcl-2 for 24 h (lanes 1–3) or 48 h (lanes 4, 5, and 8 ), or transfected with 200 nmol/L of mismatch control oligonucleotide for24 h (lane 7 ). Microarray analysis was done using human genome array U133A (Affymetrix), which containsf16,900 well-characterized gene sequences.Hierarchical gene cluster analysis was done, in which genes were filtered according to a mean log expression level >5.5. Red, gene up-regulation; green,gene down-regulation.B, MDA-231 parental, clone 5, and MCF-7 cells were transfected with 400 nmol/L of mismatch control oligonucleotide or 400 nmol/Lof antisense Bcl-2 for 48 h, at which time total RNA was extracted. RT-PCR analysis of transcripts for RTP801 and BTG1 were done to validate themicroarray findings. The band net intensity (BNI ) was determined as the ratio of RTP801 or BTG1 to actin per sample. RTP801 induction was not observedby RT-PCR. BTG1 induction was validated in all cell lines on transfection with 400 nmol/L of antisense Bcl-2 on the order of 2to 3-fold versus cellstransfected with 400 nmol/L of mismatch control oligonucleotide.C, immunoblot analysis of MCF-7 cells (100 Ag) treated with OligofectAMINE alone (O),or transfected with 400 nmol/L of mismatch control oligonucleotide or 400 nmol/L of antisense Bcl-2 for 48 h confirmed that antisense Bcl-2 induces BTG1protein expression. Table 1. Genes predicted by microarray to be induced by antisense Bcl-2 GeneFold increase (dose) 200 nmol/L400 nmol/L600 nmol/L Hypoxia-inducible factor-1–responsive RTP801 gene4.934.524.95Solute carrier family 7, member 113.023.003.77Phospholipid scramblase 12.694.063.74Asparagine synthetase3.792.173.61BTG12.742.983.10Stanniocalcin 22.582.912.74Amino acid transporter 22.502.292.78Ornithine decarboxylase inhibitor2.202.172.66CD24 antigen2.771.892.58v-myc2.741.942.53 NOTE: Genes that were predicted by microarray analysis to be induced with a fold increase >2.5 relative to mismatch control oligonucleotide – transfected cellsare shown for MCF-7 cells transfected with 200, 400, or 600 nmol/L antisense Bcl-2 for 48 hours.Molecular Cancer Therapeutics 1597 Mol Cancer Ther 2006;5(6). June 2006 400 nmol/L of mismatch control oligonucleotide or 400nmol/L of antisense Bcl-2 was done to validate microarrayfindings. RT-PCR failed to show induction of RTP801 byantisense Bcl-2 oligonucleotide (Fig. 2B). The second genewith a putative role in cell death predicted to be induced3-fold by antisense Bcl-2 on microarray was BTG1 , which isa potential mediator of chemotherapy-induced apoptosis(22). RT-PCR analysis of MCF-7, MDA-231 parental, andclone 5 cells transfected for 48 hours with 400 nmol/L ofmismatch control oligonucleotide or 400 nmol/L ofantisense Bcl-2 validated induction of BTG1 on the orderof 2to 4-fold (Fig. 2B). Immunoblotting further confirmedup-regulation of BTG1 in response to antisense Bcl-2treatment (Fig. 2C). Thus, down-regulation of Bcl-2 resultedin increased expression of BTG1.BTG1 siRNA Reduces Antisense Bcl-2-MediatedApoptosis of Breast Cancer CellsTo determine if the induction of BTG1 contributes toantisense Bcl-2-mediated cytotoxic effects, we suppressedBTG1 expression using BTG1 siRNA and then examinedthe effects of antisense Bcl-2 on cell viability. MCF-7 cellswere transfected with 200 or 400 nmol/L of BTG1 siRNA or400 nmol/L of control siRNA for 48 hours each. BTG1siRNA at 400 nmol/L reduced BTG1 protein levels,whereas the equivalent concentration of control siRNAdid not alter BTG1 levels (Fig. 3A).We next transfected MCF-7 cells with 400 nmol/L ofcontrol siRNA or 400 nmol/L of BTG1 siRNA. After 24hours, mismatch control oligonucleotide or antisense Bcl-2was transfected at 400 nmol/L each. Cell viability wasmeasured after an additional 48 hours by trypan blueexclusion (Fig. 3B). Knockdown of BTG1 partially inhibitedantisense Bcl-2-mediated cell death. To specifically assesseffects on apoptosis, ELISA-based analysis of DNAfragmentation was done. Similar to trypan blue assays,MCF-7 cells were transfected with 400 nmol/L of controlsiRNA or BTG1 siRNA for 24 hours, then with mismatchcontrol oligonucleotide or antisense Bcl-2 for an additional48 hours. Protein lysates were then evaluated for levels ofDNA fragmentation (Fig. 3C). Knockdown of BTG1blocked antisense Bcl-2-mediated DNA fragmentation,indicating that BTG1 contributes to antisense Bcl-2-mediated apoptosis.BTG1Is a Bcl-2-RegulatedMediator of ApoptosisOur results demonstrating that antisense Bcl-2 inducesBTG1 suggest that BTG1 is negatively regulated by Bcl-2.To test this hypothesis, MDA-231-derived Neo control, Bcl-2-overexpressing clone 4, and Bcl-2-overexpressing clone5 stable transfectants were transiently transfected with aBTG1 promoter-luciferase reporter construct. Increasedexpression of Bcl-2 in the stable transfectants resulted inan f2-fold reduction in BTG1 promoter activity versusNeo control cells (Fig. 4A). These results show that Bcl-2suppresses BTG1 promoter activity, supporting a role forBcl-2 as a negative transcriptional regulator of BTG1.Our results suggest that BTG1 contributes to antisenseBcl-2-mediated apoptosis. To determine if BTG1 itselfinduces apoptosis, MCF-7 cells were transiently transfectedwith a control vector or a vector containing the BTG1coding sequence. Immunoblotting confirmed increasedexpression of BTG1 after 48 hours of transfection(Fig. 4B). To assess BTG1-mediated apoptosis, cells werelysed and analyzed for DNA fragmentation after 48 hoursof transfection with BTG1 or control vector. BTG1 induceda 2.5-fold increase in fragmented DNA (Fig. 4C), indicatingthe induction of apoptosis. These results support BTG1 as anovel Bcl-2-regulated mediator of apoptosis in breastcancer.Figure 3. BTG1 knockdown blocks antisense Bcl-2-mediated cell death.A, MCF-7 cells were transfected with control siRNA (si-C ) at 400 nmol/L,or BTG1 siRNA (si-BTG ) at 200 or 400 nmol/L, each for 48 h. Proteinlysates were immunoblotted (100 Ag) for BTG1 and actin, and showedknockdown of BTG1 using 400 nmol/L BTG1 siRNA.B, MCF-7 cells weretreated in the following groups: transfection of control siRNA or BTG1siRNA for 24 h, followed by transfection of mismatch control oligonucle-otide or antisense Bcl-2 for an additional 48 h. Cells were counted bytrypan blue exclusion. All experiments were done in triplicate (at leasttwice). Cell viability is expressed as a percentage of the control siRNA/mismatch control oligonucleotide transfection control group. Bars, SDbetween triplicates. Knockdown of BTG1 reduced antisense Bcl-2-mediated cell death.C, the same transfection groups as in B were lysedfor protein and analyzed by ELISA for DNA fragmentation. DNAfragmentation is shown relative to the control siRNA/mismatch controloligonucleotide transfection control group. Error bars represent SDbetween triplicates. The control siRNA/antisense Bcl-2 group showed2-fold induction of DNA fragmentation. Knockdown of BTG1 inhibitedantisense Bcl-2-mediated DNA fragmentation.BTG1, a Bcl-2-Regulated Mediator of Apoptosis1598 Mol Cancer Ther 2006;5(6). June 2006