Overexpression of caspase-3 restores sensitivity for drug-induced apoptosisin breast cancer cell lines with acquired drug resistance

Katrin Friedrich1,3, Thomas Wieder1,3, Clarissa Von Haefen1, Silke Radetzki1, Reiner JaÈ nicke2,Klaus Schulze-Ostho�2, Bernd DoÈ rken1 and Peter T Daniel*,1

1Department of Hematology, Oncology and Tumor Immunology, University Medical Center ChariteÂ, Humboldt University Berlin,13125 Berlin, Germany; 2Department of Immunology and Cell Biology, University of MuÈnster, 48149 MuÈnster, Germany

In this study, we asked whether overexpression ofcaspase-3, a central downstream executioner of apoptoticpathways, might sensitize breast cancer cells withacquired drug resistance (MT1/ADR) to drug-inducedapoptosis. As control, we employed caspase-3 negativeand caspase-3-transfected MCF-7 cells. Whereas mock-transfected MCF-7 cells were resistent to epirubicin,etoposide and paclitaxel (taxol), the same drugs led tobreakdown of nuclear DNA in caspase-3-transfectedMCF-7 cells. MT1/ADR cells express low levels of wildtype caspase-3 but show defective caspase activation andapoptosis upon drug exposure. These cells also display aless e�cient activation of the mitochondrial permeabilitytransition. Caspase-3-transfected MT1/ADR clonesshowed a 2.8-fold increase in the protein level and a3.7-fold higher speci®c enzyme activity. Procaspase-3overexpression was not toxic and did not a�ect back-ground apoptosis. Interestingly, procaspase-3-transfectedMT1/ADR cells were more sensitive to cytotoxic drugsas compared with vector-transfected controls and DNAfragmentation nearly reached the levels of the originaldrug sensitive MT1 cells. Thus, overexpression ofcaspase-3 enhances chemosensitivity especially in situa-tions where activation of the mitochondrial apoptosomeis disturbed. Oncogene (2001) 20, 2749 ± 2760.

Keywords: caspase-3; chemotherapeutics; apoptosis;drug resistance; MCF-7 cells

Introduction

Apoptosis, a morphologically and biochemicallyde®ned form of cell death (Kerr et al., 1972), hasbeen demonstrated to play a role in a wide variety ofbiological systems, including physiological cell turn-over, the immune system, embryonic development,and hormone-dependent atrophy (Cohen et al., 1992;

Ellis et al., 1991; Krammer et al., 1994). Although theprimary targets of chemotherapeutic antitumor agentscomprise distinct intracellular structures, such as thenucleus, the cell membrane and the cytoskeleton,considerable evidence exists that drug-induced cyto-toxicity leads to the activation of common apoptoticpathways (Fisher, 1994; Hannun, 1997). However,acquired drug resistance represents a major drawbackin chemotherapy and is the main reason for the failureof chemotherapy to cure cancer. Dysfunction ofapoptotic pathways has been implicated in themechanisms leading to the selection of drug resistantcell clones. For example, the deregulated expression ofanti-apoptotic Bcl-2 family members can confer drugresistance in transfected tumor cells (Lotem andSachs, 1993; Miyashita and Reed, 1993). On the otherhand, overexpression of Bax-a, a proapoptoticmember of the Bcl-2 family, in breast cancer cellsreversed the malignant phenotype with regard totumorigenicity in a SCID mouse xenotransplant modeland increased the sensitivity of these cells to apoptosis(Bargou et al., 1996). In this line, we and othersestablished that the loss of Bax expression inmetastatic breast (Krajewski et al., 1995; Sturm etal., 2000) or colorectal cancer (Sturm et al., 1999) is anegative prognostic factor for response to cancertherapy. Inactivation of the apoptosis machinery byloss of Bax may occur at the time of tumor relapseafter chemotherapy as recently demonstrated in child-hood acute lymphoblastic leukemia (Prokop et al.,2000).

The central component of the apoptotic machineryis a proteolytic system consisting of a family ofcysteinyl proteases with an absolute requirement forcleavage after aspartic acid, therefore called caspases(for reviews see Cohen, 1997; Thornberry andLazebnik, 1998). In mammalian cells, at least 14di�erent caspases exist. The members of the caspasefamily can be divided into two groups: (i) upstreaminitiator caspases, such as caspase-8 and caspase-9,which cleave and activate other caspases; and (ii)downstream e�ector caspases, such as caspase-3 andcaspase-6, which cleave cellular substrates therebydisassembling cellular structures or inactivating en-zymes. Caspase-3 is the most intensively studiede�ector caspase. It has been shown that depletion of

Oncogene (2001) 20, 2749 ± 2760ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00

www.nature.com/onc

*Correspondence: PT Daniel, Department of Hematology, Oncologyand Tumor Immunology, University Medical Center Charite ,Campus Berlin-Buch, Humboldt University Berlin, LindenbergerWeg 80, 13125 Berlin, Germany3These authors contributed equally to the present studyReceived 7 November 2000; revised 1 December 2000; accepted 6February 2001

caspase-3 by homologous recombination leads toaccumulation of neuronal cells whereas other tissueswere not a�ected (Kuida et al., 1996). Furthermore,we provided evidence that development of therapy-refractory relapse of childhood acute lymphoblasticleukemia is accompanied by a complete loss ofspontaneous processing of caspase-3 (Prokop et al.,2000). Mutations of caspase-3 have been described incell line models. Recently, loss of caspase-3 due to a47-base pair deletion within exon 3 of the CASP-3gene in the breast carcinoma cell line MCF-7 wasreported leading to a blockade of DNA fragmentationin response to tumor necrosis factor (TNF) a andstaurosporine (JaÈ nicke et al., 1998a). In this line, thehierarchical activation of caspases-2, -3, -6, -7, -8 and-10 in a caspase-9 dependent manner has been shownin a cell-free system and a feedback ampli®cation loopinvolving caspase-3 and caspase-9 has been proposed(Slee et al., 1999). Indeed, results from di�erentlaboratories con®rmed these in vitro data and evidencewas provided for a caspase-3-driven ampli®cation loopduring drug-induced apoptosis (Slee et al., 2000;Wieder et al., 2001). Thus, the intracellular level ofprocaspase-3 might be of importance for the impact ofapoptotic stimuli. We assumed that the forcedexpression of this downstream apoptosis e�ector couldnot only be used to restore apoptosis sensitivity incaspase-3 de®cient MCF-7 cells but could also beemployed to overcome drug resistance in breast cancercell lines.

To investigate whether caspase-3 might revert drugresistance and in¯uence apoptotic signaling, theenzyme was stably overexpressed by retroviral genetransfer in the drug resistant breast cancer cell line,MT1/ADR, which was selected from the drug-sensitive MT1 cell line by culture in the presence ofdoxorubicin (Stein et al., 1997). MT1/ADR cells carrythe p53 wild type gene and do not show increaseddrug e�ux, but instead display a defect in theactivation of the mitochondrial apoptosis machinerywhich can be reverted by overexpression of Bak orBik/Nbk (PT Daniel and S Radetzki, unpublishedobservation). For example, speci®c apoptosis afterchallenge with 2.5 mg/ml epirubicin was increasedfrom 13.2% in pcDNA3-transfected MT1/ADR cellsto 52.5% in bak-transfected cells and to 21.3% inbik/nbk-transfected cells. Additionally, these cellsexpress low levels of functional, i.e. wild typeprocaspase-3.

In the present study, we demonstrate that restorationof caspase-3 expression is necessary for apoptotic DNAfragmentation when MCF-7 cells were treated withepirubicin, etoposide and taxol. In the case of MT1/ADR cells, we show that the sensitivity against allthree drugs was signi®cantly enhanced by overexpres-sion of procaspase-3 and nearly reached the same levelof sensitivity as compared with the original drugsensitive MT1 cell line. Thus, procaspase-3 could serveas a future target gene to enhance chemosensitivityeven in cells with defective activation of upstreame�ectors of the mitochondrial apoptosome.

Results

Chemotherapeutic drugs induce apoptotic cell death inprocaspase-3-transfected MCF-7 cells

MCF-7 breast carcinoma cells are widely used as amodel system for breast cancer. Recently, it has beendemonstrated that MCF-7 cells do not expresscaspase-3, and MCF-7 clones stably transfected withfull length procaspase-3 cDNA were generated(JaÈ nicke et al., 1998a). To determine the role ofcaspase-3 in such a knock-out model, we testedcaspase-3-de®cient and caspase-3-reconstituted MCF-7cells for drug-induced apoptosis. In a ®rst set ofexperiments, we checked the expression of caspase-3and caspase-3-like activity in pcDNA3-transfected andcaspase-3-transfected MCF-7 cells. As shown in Figure1a, vector-transfected MCF-7 cells did not express anyimmuno-detectable procaspase-3 whereas caspase-3-transfected MCF-7 cells contained signi®cant amountsof the proenzyme. Furthermore, we found onlymarginal caspase-3-like enzyme activity in vector-transfected MCF-7 cells below 0.1 mM/min, even whenthe apoptotic cascade was triggered in vitro by additionof cytochrome c/dATP as previously described inHeLa cell extracts (Liu et al., 1996). On the otherhand, basal caspase-3-like enzyme activity was0.29 mM/min in caspase-3-transfected MCF-7 cells.This basal activity was stimulated by about 13-foldin the presence of cytochrome c/dATP (Figure 1b)thereby indicating that the components required foractivation of caspase-3 by mitochondrial release ofcytochrome c are present and fully functional in MCF-7 cells.

We employed caspase-3-transfected MCF-7 cells as apositive control and compared the results withessentially caspase-3-free MCF-7 vector controls toshow the e�ect of defective caspase-3 activation ondrug-induced cell death. The nuclear alterationsassociated with apoptosis are characterized by inter-nucleosomal cleavage of DNA (Wyllie, 1980) leadingto the appearance of a sub G1-peak in apoptotic cellswhen cellular DNA is analysed by ¯ow cytometry.Figure 2 shows the nuclear DNA fragmentationpro®les of mock-transfected MCF-7 cells as comparedwith caspase-3-transfected MCF-7 cells after treatmentwith di�erent drugs. As expected, vector-transfectedMCF-7 cells were very insensitive in response toepirubicin, etoposide and taxol and DNA fragmenta-tion did not exceed 30%, even at high concentrationsof the respective drugs (Figure 3). In contrast, caspase-3-transfected MCF-7 cells displayed high sensitivitytowards epirubicin and etoposide and incubation with0.3 mg/ml epirubicin or 40 mg/ml etoposide led to anaccumulation of 80% apoptotic cells. Although thee�ect was not as dramatic as compared with the otherdrugs, taxol-induced apoptosis was also enhanced incaspase-3-transfected MCF-7 cells. Furthermore, wedetermined the cytotoxicity of all three drugs asassessed by the release of cytosolic LDH in thisexperimental system. However, caspase-3-transfection

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2750

Oncogene

Figure 1 Expression of caspase-3 in mock- and caspase-3-transfected MCF-7 cells. (a) Extracts from 26106 mock-transfected cells(A) and 26106 caspase-3-transfected cells (B) were prepared and 30 mg of cellular protein were loaded per lane. Proteins wereseparated by SDS±PAGE and immunoblotting using an anti-caspase-3 antibody was performed as described in Materials andmethods. Positions of molecular weight standards are shown. (b) Caspase-3 activity was measured in extracts from vector-(pcDNA3) or caspase-3-transfected MCF-7 (Casp.3) cells after in vitro activation with 1 mM dATP and 10 mM cytochrome c (+).Controls were incubated in the absence of dATP and cytochrome c (7). Data are given in mM DEVD-pNA cleaved per min andrepresent the mean of two determinations with an error of less than 10%. The experiment was repeated and similar results wereobtained

Figure 2 FACS pro®les of mock- and caspase-3-transfected MCF-7 cells after epirubicin, etoposide and taxol treatment. Mock-transfected (MCF-7/pcDNA3) and procaspase-3-transfected MCF-7 cells (MCF-7/Casp.3) were treated with 1 mg/ml epirubicin,40 mg/ml etoposide or 10 mg/ml taxol as indicated. After 72 h of incubation, DNA fragmentation was determined as described inMaterials and methods. A representative FACS pro®le of each experiment is shown and cells displaying hypodiploid DNA are givenin per cent of the total population

Oncogene

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2751

in MCF-7 cells only in¯uenced the general cytotoxice�ects of epirubicin, etoposide or taxol at concentra-tions 41 mg/ml, 410 mg/ml or 41 mg/ml, respectively,whereas at very high drug concentrations 100% of thecells died irrespective of caspase-3 transfection vianecrotic cell death (Figure 3d ± f). In the case of taxoltreatment, induction of necrotic cell death mightexplain the bell shaped dose/response curve whichwas observed in pcDNA-3-mock-transfected MCF-7cells (Figure 3c). In the mock-transfectants, taxol

preferentially induces apoptosis at concentrations40.1 mg/ml. However, the relative amount of apopto-tic cells after challenge of the mock transfectants with1 mg/ml or 10 mg/ml taxol even declines whereas theproportion of dead cells increases from 20% aftertreatment with 1 mg/ml taxol to 100% after treatmentwith 10 mg/ml taxol. The `dip' observed in the dose/response curve for apoptosis induction in the procas-pase-3-transfected MCF-7 and in the LDH measure-ments for both cell lines may therefore be the

Figure 3 Induction of apoptosis and cell death by epirubicin, etoposide and taxol in mock- and caspase-3-transfected MCF-7 cells.Mock-transfected (*, &) and procaspase-3-transfected MCF-7 cells (*, &) were treated with di�erent concentrations of epirubicin(a,d), etoposide (b,e) or taxol (c,f) as indicated in the ®gure. After 72 h of incubation, DNA fragmentation (a ± c) and cell death (d ±f) were determined as described in Materials and methods. Values of apoptosis are given as per cent hypoploidy and values of celldeath as per cent dead cells. Data points represent the mean of two determinations with an error of less than 10%. The experimentswere repeated and similar results were obtained

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2752

Oncogene

consequence of di�erent modes of cell death inductionat high or at low taxol concentrations.

Characterization of MT1/ADR clones

Drug resistance is often correlated with dysfunction ordysregulation of apoptotic pathways and overexpres-sion of downstream e�ectors, such as caspase-3, mightrestore or enhance chemosensitivity. To investigate thishypothesis, we used the breast carcinoma cell line MT1(Naundorf et al., 1992) and its drug resistant counter-part MT1/ADR which was selected by treatment withincreasing concentrations of doxorubicin for severalweeks (Stein et al., 1997). Interestingly, the drugresistance of MT1/ADR cells is due to a defect at thelevel of the mitochondrial apoptosome and transfectionof proapoptotic genes, e.g. bak and nbk, restoresmitochondrial activation and drug sensitivity of thesecells (PT Daniel and S Radetzki, unpublished observa-tion). Three clones of MT1/ADR cells stably expres-sing exogenous procaspase-3 were generated asdemonstrated by vector speci®c RT±PCR analysis(Figure 4). In contrast, exogenous procaspase-3 wasabsent in MT1, vector-transfected MT1/ADR anduntransfected MT1/ADR cells (Figure 4). To con®rmthat expression of exogenous procaspase-3 leads tohigher protein levels in positive clones, Western blotanalyses and measurements of caspase-3-like enzymeactivities were performed. As shown in Figure 5a, MT1and MT1/ADR contained relatively low endogenouslevels of procaspase-3. This level was not signi®cantlyaltered by transfection of MT1/ADR with the controlvector LXSN. In contrast, the caspase-3-transfectedMT1/ADR clones 11, 12 and 14 contained signi®cantlyhigher procaspase-3 levels. Quantitation of the procas-pase-3 band by videodensitometry revealed a 2.8+0.3-fold increase in caspase-3-transfected MT1/ADR cellsas compared with the parental MT1/ADR cell line

(Figure 5b). Overexpression of procaspase-3 proteinwas not accompanied by spontaneous caspase-3activation and we could not detect mature p17 subunitsin the procaspase-3-transfected clones (Figure 5a). As aconsequence, the stable overexpression of procaspase-3alone induced neither apoptosis-like cell death normorphological changes (data not shown) therebycon®rming similar observations in NIH3T3 cells usingprocaspase-1 (Hiwasa et al., 1998). Nevertheless, wetested whether the higher levels of procaspase-3 inpositive clones could be in vitro activated by additionof 1 mM dATP and 10 mM cytochrome c thus resultingin higher speci®c enzyme activities. In accordance withthe absence of mature p17 subunits of caspase-3 undercontrol conditions, neither caspase-3- nor LXSN-mock-transfected MT1/ADR cells contained signi®cantamounts of caspase-3-like activity when assayed in

Figure 4 Detection of the procaspase-3 transgene in MT1/ADRclones. Stable transfectants were tested for procaspase-3 transgeneexpression by vector-speci®c reverse transcription and PCRampli®cation as described in Materials and methods (right laneof each clone). Three clones of caspase-3-transfected (MT1/ADR/C3), and two clones of mock-transfected cells (MT1/ADR/LXSN) are shown. In addition, untransfected MT1 and MT1/ADR cells were included. Actin was ampli®ed in parallel and usedas a loading control (left lane of each clone)

Figure 5 Overexpression of procaspase-3 in MT1/ADR. (a)Extracts from 26106 caspase-3-transfected (MT1/ADR/C3,clones 11, 12 and 14) and 26106 mock-transfected cells (MT1/ADR/LXSN, clones 9 and 10) were prepared. In addition,extracts from 26106 untransfected MT1 and 26106 MT1/ADRcells were included. Thirty mg of cellular protein were loaded perlane, proteins were separated by SDS±PAGE and immunoblot-ting using anti-caspase-3 antibody was performed as described inMaterials and methods. Positions of molecular weight standardsare shown. (b) The expression of procaspase-3 was quanti®ed byvideodensitometry using a Gel Doc 2000 apparatus (Biorad;MuÈ nchen, Germany) and optical density of the procaspase-3 bandis given in arbitrary units. The experiment was repeated andsimilar results were obtained

Oncogene

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2753

extracts which had not been in vitro activated (data notshown). On the other hand, caspase-3-like enzymeactivity after in vitro activation was signi®cantlyincreased in caspase-3-transfected MT1/ADR clones(Figure 6). As compared with untransfected MT1/ADR cells this increase reached 3.7+0.7-fold. InLXSN-mock-transfected MT1/ADR cells, enzyme ac-tivity was slightly higher as compared with untrans-fected MT1/ADR cells (Figure 6). This increase,however, was not signi®cant. Thus, expression levelsof procaspase-3 nicely correspond to the induciblecaspase-3-like activity in a cell-free system.

Enhancement of apoptotic DNA fragmentation bydifferent cytotoxic drugs in procaspase-3-transfectedMT1/ADR cells

As demonstrated in MCF-7 cells, caspase-3 is necessaryfor drug-induced DNA fragmentation and apoptoticcell death. We now asked whether overexpression ofprocaspase-3 might sensitize the drug-resistant breastcancer cell line MT1/ADR for drug-induced apoptosisand thereby overcomes drug resistance in this cell linemodel. Figure 7 shows the apoptosis induction ofmock-transfected MT1/ADR cells or caspase-3-trans-

fected MT1/ADR cells after treatment with di�erentdrugs as determined by analysis of nuclear DNAfragmentation by ¯ow cytometry on the single celllevel. As expected, MT1/ADR cells were very insensi-tive to treatment with epirubicin, etoposide and taxol(Figure 8). Apoptotic cells were below 25% even whenthe cells were treated with 10 mg/ml epirubicin (Figure8a), 80 mg/ml etoposide (Figure 8b) or 0.8 mg/ml taxol(Figure 8c). In contrast, the original cell line MT1already showed induction of apoptosis in the presenceof lower drug concentrations and in all cases thenumber of cells with hypoploid DNA reached 75%(Figure 8d ± f). When caspase-3-transfected MT1/ADRcells were treated with di�erent concentrations ofepirubicin (Figure 8a), etoposide (Figure 8b) or taxol(Figure 8c), DNA fragmentation was enhanced by atleast twofold as compared with the respective vector-transfected control cells. Apoptosis nearly reached thesame levels as in the original MT1 cell line indicatingthat overexpression of procaspase-3 in MT1/ADR cellsleads to ampli®cation of drug-induced apoptotic path-ways culminating in internucleosomal breakdown ofDNA. Interestingly, we could demonstrate that thestable transfection of procaspase-3 did not in¯uencecells with intact apoptotic signaling. For this, wegenerated stable clones of drug-sensitive MT1 cellsoverexpressing procaspase-3. These cells showed a4.5+0.6-fold increase of procaspase-3 as comparedwith the parental MT1 cell line (Figure 9). Thus, theoverexpression of procaspase-3 was in the same rangeas described for the drug-resistant MT1/ADR cells. Incontrast to drug-resistant MT1/ADR cells, overexpres-sion of procaspase-3 in drug-sensitive MT1 cells didnot further sensitize the cells for epirubicin-inducedapoptosis (Figure 9c). Similar results were obtainedusing etoposide or taxol and a sensitizing e�ect ofprocaspase-3 was not observed in MT1 cells (data notshown).

Discussion

Drug resistance is one of the major draw backs ofchemotherapy and new therapeutical strategies toovercome this problem are necessary. Several anti-cancer drugs have been demonstrated to induceapoptotic cell death and to activate caspases in avariety of cellular systems (Essmann et al., 2000;Ibrado et al., 1996a,b; Suzuki et al., 1998; Wesselborget al., 1999). Although these cytotoxic agents head fordi�erent primary targets it has become obvious that inmost cases drug-induced apoptosis converges on acommon pathway, causing mitochondrial membranepermeabilization (Kroemer and Reed, 2000) andsubsequent activation of downstream caspases, suchas caspase-3. Dysfunction of apoptotic pathways hasbeen implicated in the development of drug resistance.In this line, we recently demonstrated that therapy-refractory relapse of childhood acute lymphoblasticleukemia is accompanied by a drastic decline of theBax/Bcl-2 ratio and loss of spontaneous caspase-3

Figure 6 Determination of caspase-3-activity in MT1/ADR.Caspase-3 activity was measured in extracts from procaspase-3-transfected (MT1/ADR/C3, clones 11, 12, and 14) or mock-transfected cells (MT1/ADR/LXSN, clones 9 and 10) and inextracts from untransfected MT1 or MT1/ADR cells after in vitroactivation with 1 mM dATP and 10 mM cytochrome c as describedin Materials and methods. Data are given in mM DEVD-pNAcleaved per (min6mg cellular protein)+s.d. (n=3). **Signi®-cantly di�erent from MT1/ADR cells at P50.01, ***Signi®cantlydi�erent from MT1/ADR cells at P50.001

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2754

Oncogene

processing in vivo (Prokop et al., 2000). In the samecontext, we have previously shown that overexpressionof proapoptotic regulators, such as Bax (Bargou et al.,1996; Wagener et al., 1996) or Bik/Nbk (Daniel et al.,1999), increases the sensitivity of cancer cell lines fordrug-induced apoptosis. However, Bcl-2 type regula-tors of apoptosis pathways are located upstream of themitochondrion which represents a central checkpointof these pathways.

In the present study, we asked whether genetransfer of the downstream e�ector procaspase-3would also lead to sensitization of drug-resistantcancer cells. As model systems we used the caspase-3-de®cient breast cancer cell line MCF-7 (JaÈ nicke etal., 1998a) and drug-resistant MT1/ADR cells (Naun-dorf et al., 1992; Stein et al., 1997) which show adefect in the formation of the mitochondrial death-inducing signaling complex and impaired activation ofcaspase-3 (PT Daniel and S Radetzki, unpublishedobservation). In a ®rst set of experiments, we couldclearly demonstrate that only MCF-7 cells which hadbeen transfected with procaspase-3 underwent apop-tosis after challenge with di�erent chemotherapeutics.Caspase-3 is therefore necessary for epirubicin-,etoposide- and taxol-induced apoptosis, at least inthis particular caspase-3 knock-out model system. Ourresults are consistent with previous reports showingthat caspase-3 is indispensable for DNA fragmenta-

tion and a-fodrin cleavage after TNF a- andstaurosporine-treatment but not for cell death ingeneral (JaÈ nicke et al., 1998a,b).

In a second experimental model, we investigated thebreast cancer cell line MT1 and its drug resistantcounterpart MT1/ADR. We generated di�erent MT1/ADR clones carrying exogenous procaspase-3 byretroviral transfer of the procaspase-3 gene intoMT1/ADR cells. Stable transfection of the procas-pase-3 gene led to moderate but signi®cantly higherexpression of the proenzyme. The very similar andrather low levels of procaspase-3 in the MT1/ADR/C3clones analysed in our study can be explained by theassumption that cells with higher levels of procaspase-3have been eliminated during the selection of the clones.However, stable transfection of procaspase-3 did nota�ect background apoptosis levels of the cells (seeFigure 8). This di�ers from the results of previousstudies which used the proapoptotic proteins Bax orapoptotic protease activating factor-1 (Apaf-1). It wasshown that controlled expression of Bax, withoutanother death stimulus, is su�cient to induce apoptosisin Jurkat cells (Xiang et al., 1996). In the case of Apaf-1, overexpression of this adaptor molecule not onlysensitized HL-60 cells to taxol- and etoposide-inducedapoptosis but also signi®cantly increased backgroundapoptosis in transfected cells (Perkins et al., 1998). Allthe more, our procaspase-3-overexpressing MT1/ADR

Figure 7 FACS pro®les of mock- and caspase-3-transfected MT1/ADR cells after epirubicin, etoposide and taxol treatment.Mock-transfected (MT1/ADR/LXSN) and procaspase-3-transfected MT1/ADR cells (MT1/ADR/Casp.3) were treated with 1 mg/mlepirubicin, 40 mg/ml etoposide or 0.5 mg/ml taxol as indicated. After 72 h of incubation, DNA fragmentation was determined asdescribed in Materials and methods. A representative FACS pro®le of each experiment is shown and cells displaying hypodiploidDNA are given in per cent of the total population

Oncogene

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2755

clones represent an interesting cellular system toinvestigate whether enforced expression of this down-stream e�ector might overcome drug resistance. Wecould demonstrate that MT1/ADR cells stably trans-fected with procaspase-3 were signi®cantly moresusceptible to treatment with epirubicin, etoposideand taxol as compared with the mock-transfectedMT1/ADR cells. The enhanced susceptibility ofprocaspase-3 overexpressing cells can be explained bythe ®ndings that caspase-3 is involved in di�erentampli®cation loops thereby multiplying the impact of agiven death stimulus. For example, it has been shownthat drug treatment triggers caspase-3-mediated clea-vage of Bid. Truncated Bid then enhances theapoptotic signal upstream of mitochondria (Slee etal., 2000). Furthermore, as evidenced recently, themolecular order of drug-induced apoptosis in Blymphoid cells is represented by the Fas/CD95-

independent, hierarchical activation of mitochondrialcytochrome c release, caspase-3 and caspase-8 (Wiederet al., 2001). In this system, caspase-8 functions as anampli®er and not as an initiator caspase.

As already mentioned above, caspase-3-transfectedMT1/ADR cells were more sensitive to cytotoxic drugsas compared with vector-transfected controls and DNAfragmentation nearly reached the levels of the originaldrug sensitive MT1 cells. Interestingly, the drugresistance of MT1/ADR cells is due to a defect at thelevel of the mitochondrial apoptosome. Theoretically,transfection of caspase-3 should therefore have noe�ect since activation of the mitochondrial apoptosomeis the dominant signaling pathway for drug-inducedapoptosis (Kroemer and Reed, 2000) and caspase-3activation is presumed to occur downstream ofcaspase-9 activation. In our system, however, caspase-3 overexpression appears to bypass this e�ect since the

Figure 8 Induction of apoptosis by epirubicin, etoposide and taxol in mock- and procaspase-3-transfected MT1/ADR cells. Clones9, 10 and 11 of mock-transfected (*) or clones 11, 12 and 14 of procaspase-3-transfected MT1/ADR cells (*) were treated withdi�erent concentrations of epirubicin (a), etoposide (b) or taxol (c). Additionally, MT1/ADR (*) or MT1 cells (*) were treatedwith di�erent concentrations of epirubicin (d), etoposide (e) or taxol (f). After 72 h of incubation, DNA fragmentation wasdetermined as described in Materials and methods. Data are given as per cent hypoploidy and represent the mean of threeindependent experiments+s.d.

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2756

Oncogene

MT1/ADR cells, which are defective at the level ofcytochrome c release and mitochondrial permeabilityshift transition, were sensitized for apoptotic cell deathupon transduction of the caspase-3 construct. This, inturn, suggests that activation of caspase-3 in MT1/ADR cells may even occur in the presence ofmitochondrial signaling defects.

From our results, we conclude that enhancedexpression of the central executioner of apoptoticpathways, caspase-3, in the form of its inactiveproenzyme can overcome acquired drug resistance.Thus, caspase-3 could serve as a target gene to enhancechemosensitivity not only by gene transfer but also bynew drugs even in situations where the activation ofthe mitochondrial apoptosome is defective, e.g. asconsequence of a loss of Bax or overexpression of Bcl-2which are both frequently observed in breast cancer

and other tumors (Bargou et al., 1995; Sturm et al.,1999, 2000; Prokop et al., 2000).

Materials and methods

Materials

Polyclonal rabbit anti-human caspase-3 antibody (developedagainst human recombinant protein and recognizing unpro-cessed procaspase-3 and the 17 kDa subunit of active caspase-3) was from PharMingen (Hamburg, Germany) and used at adilution of 1 : 1000. Secondary anti-rabbit, horse radishperoxidase conjugated antibody was from Promega (Mann-heim, Germany). Caspase-3 substrate (Ac-DEVD-pNA) wasfrom Calbiochem-Novabiochem GmbH (Bad Soden, Ger-many). RNase A was from Roth (Karlsruhe, Germany).Epirubicin was purchased from Pharmacia Upjohn (Erlangen,

Figure 9 Overexpression of procaspase-3 in MT1 and induction of apoptosis by epirubicin. (a) Extracts from each 26106

procaspase-3-transfected (MT1/C3), mock-transfected (MT1/LXSN) and untransfected MT1 cells (MT1) were prepared. Thirty mgprotein were loaded per lane, proteins were separated by SDS±PAGE and immunoblotting using anti-caspase-3 antibody wasperformed as described in Materials and methods. (b) The expression of procaspase-3 was quanti®ed by videodensitometry andoptical density of the procaspase-3 band is given in arbitrary units. (c) MT1/LXSN (*) or MT1/C3 cells (*) were treated withdi�erent concentrations of epirubicin as indicated in the ®gure. After 72 h, DNA fragmentation was determined. The data of arepresentative experiment are given in per cent hypoploidy. Data points represent the mean of two determinations with an error ofless than 10%. The experiment was repeated and similar results were obtained

Oncogene

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2757

Germany), etoposide and taxol (paclitaxel) were from BristolArzneimittel GmbH (MuÈ nchen, Germany).

Cell culture

Control vector (pcDNA3-transfected) and pcDNA3-caspase-3-transfected MCF-7 cells (JaÈ nicke et al., 1998a) were grownin Roswell Park Memorial Institute (RPMI) mediumsupplemented with 10% fetal calf serum (FCS), 0.56 g/l L-glutamine, 100 000 international units (IU)/penicillin and0.1 g/l streptomycin. Media and culture reagents were fromLife Technologies GmbH (Karlsruhe, Germany). Con¯uentcells were subcultured every 5 days after detaching the cellswith a Trypsin/EDTA solution (Sigma; MuÈ nchen, Germany).

Vector construction and transfection of packaging cells

The caspase-3 cDNA was ampli®ed by polymerase chainreaction (PCR) from human cDNA using the followingprimers: sense: 5'-AAC GAA TTC CCA CCA TGG AGAACA CTG AAA ACT CAG-3' antisense: 5'-AAC GAA TTCTTA GTG ATA AAA ATA GAG TTC-3'. The resultingproduct was digested with EcoRI and the 847 bp caspase-3fragment was cloned nondirectedly into the correspondingsites of retroviral vector pLXSN (Miller and Rosman, 1989)which was kindly provided by T Blankenstein (Max DelbruÈ ckCentrum; Berlin, Germany). The orientation of the caspase-3fragment was tested by HindIII digestion and sequenceanalysis of the caspase-3/pLXSN construct was performedon a ABI PRISM 310 Genetic Analyzer (Applied Biosystems;Weiterstadt, Germany) using a standard protocol.

The amphotropic packaging cell line FlyA (Cosset et al.,1995) was transfected with the caspase3/pLXSN-expressionconstruct (7.5 mg and 15 mg) by Ca3(PO4)2 precipitation usingthe Mammalian Transfection Kit (Stratagene; Heidelberg,Germany) according to the manufacturer's instructions.Stable virus-producing transfectants were selected using1 mg/ml geneticin 418 (G418) (Gibco ±BRL; Karlsruhe,Germany) and maintained in Dulbecco s minimal essentialmedium (DMEM) supplemented with 10% heat-inactivatedFCS, 0.56 g/l L-glutamine, 100 000 IU/penicillin and 0.1 g/lstreptomycin. All products were purchased from Gibco ±BRL. The colonies were expanded for analysis and virus-production. Control transfectants carrying the pLXSN-vectoralone were generated and selected in parallel.

Reverse transcription and PCR ampli®cation for testingvirus-producing transfectants were performed as described(Sturm et al., 1999). The following primers were used forampli®cation: sense: 5'-CCG GAA TTC CCA CCA TG-3'antisense: 5'-AAC GAA TTC TTA GTG ATA AAA ATAGAG TTC- 3'. PCR products were separated on 1% agarosegel and DNA was visualized by ethidiumbromide staining.

Virus titration

The titers of the virus-containing supernatants of positiveclones were determined on D17 cells (Uckert et al., 1998).Cells were grown in a-MEM (Gibco ±BRL) and infectedusing di�erent virus concentrations in the presence of 8 mg/mlpolybrene. After selection using 0.8 mg/ml G418, colonieswere counted.

Infection of target cell lines

MT1 and MT1/ADR mammary carcinoma cells (Naundorfet al., 1992; Stein et al., 1997) were cultured in RPMI 1640supplemented with 10% heat-inactivated FCS, 0.56 g/l L-

glutamine, 100 000 IU/penicillin and 0.1 g/l streptomycin(Gibco ±BRL). Cells were infected three times with 2 mlcaspase-3 virus-containing supernatant (at 106 plaque formingunits (p.f.u.)/ml) or pLXSN control virus-containing super-natant (mock) in the presence of 8 mg/ml polybrene and thenselected in medium containing 1 mg/ml G418. The colonieswere expanded for analysis and the supernatants were foundto be free of replication-competent viruses.

Transgene detection by PCR

Stable transfectants were tested for expression of thetransgene by reverse transcription and PCR ampli®cation asdescribed (Sturm et al., 1999) using the following primerscontaining sequences of caspase-3 as well as sequences of theLXSN-vector: sense: 5'-CCG GAA TTC CCA CCA TG-3'antisense: 5'-TTG TGA GCA TGG AAA CAA TAC-3'. ThePCR products were separated on 1% agarose gel and theDNA was visualized by ethidiumbromide staining.

Measurement of cell death

Cytotoxicity of the di�erent drugs was measured by therelease of lactate dehydrogenase (LDH) as describedpreviously (Wieder et al., 1998). After incubation withdi�erent concentrations of epirubicin, etoposide or taxol,LDH activity released by the cells was measured in the cellculture supernatants using the Cytotoxicity Detection Kitfrom Boehringer-Mannheim (Mannheim, Germany). Thesupernatants were centrifuged at 300 g for 5 min. Twenty mlof cell-free supernatants were diluted with 80 ml phosphatebu�ered saline (PBS) and 100 ml reaction mixture containing2-[4-Iodophenyl]-3-[4-nitrophenyl]-5-phenyltetrazolium chlor-ide (INT), sodium lactate, NAD+ and diaphorase wereadded. Then, time-dependent formation of the reactionproduct was quanti®ed photometrically at 490 nm. Themaximum amount of LDH activity released by the cellswas determined by lysis of the cells using 0.1% Triton X-100in culture medium and set as 100% cell death.

Measurement of apoptotic cell death by flow cytometry

Apoptosis was determined on a single cell level by measuringthe DNA content of individual cells on a Fluorescenceactivated cell (FAC)Scan (Becton Dickinson; Heidelberg,Germany) as described with some modi®cations (Nicoletti etal., 1991). After incubation of the cells with di�erentconcentrations of epirubicin, etoposide or taxol, cells were®xed with 2% formaldehyde in PBS for 30 min on ice.Thereafter, cells were spun down at 300 g for 3 min.Supernatants were removed and cell pellets were resuspendedby vortexing. Fifty ml PBS and 100 ml ice-cold ethanol wereadded and incubated for 15 min on ice. After centrifugationat 300 g for 3 min, supernatants were removed and 50 mlDNase-free RNase (40 mg/ml) from Boehringer Mannheimwere added and incubated for 30 min at 378C. After washingwith PBS, 200 ml propidium iodide solution (50 mg/ml inPBS) were added to the resuspended cell pellets. CellularDNA content was measured with logarithmic ampli®cation inthe FL-3 channel of the FACScan. Single cells were identi®edand gated by pulse code processing of the area and width ofthe signal. Cell debris was excluded by appropriately raisingthe forward scatter threshold. Apoptotic nuclei display adecreased DNA content below the G1 peak paralleled by anincrease of the side scatter. Data are given in per centhypoploidy, i.e. the percentage of nuclei sub G1, whichre¯ects the number of apoptotic cells.

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2758

Oncogene

Immunodetection of caspase-3 protein in stable clones

Cells in log phase were harvested, washed twice with PBS andlysed in bu�er L (10 mM Tris/HCl, pH 7.5, 300 mM NaCl,1% Triton X-100, 2 mM MgCl2, 5 mM ethylenediaminetetraacetic acid (EDTA), 1 mM pepstatin, 1 mM leupeptin,0.1 mM phenylmethylsulfonyl¯uoride (PMSF)). Protein con-centration was determined using the bicinchoninic acid(BCA) assay (Smith et al., 1985) from Pierce (Rockford,IL, USA) and equal amounts of protein (20 mg per lane) wereseparated by sodium dodecylsulphate (SDS)-polyacrylamidegel electrophoresis (Laemmli, 1970). Then, immunoblottingwas performed essentially as described (Wieder et al., 1994).Membranes (Schleicher & Schuell; Dassel, Germany) wereswollen in 3-[Cyclohexylamino]-1-propanesulfonic acid(CAPS)-bu�er (10 mM CAPS, pH 11, 10% MeOH) forseveral minutes and blotting was performed at 1 mA/cm2

for 1 h in a transblot SD cell (BioRad; MuÈ nchen, Germany).The membrane was blocked for 1 h in blocking bu�er(16PBS, 0.05% Tween-20, 3% non fat dry milk) andincubated with polyclonal rabbit anti-human caspase-3antibody in blocking bu�er for 1 h. After washing themembrane three times in PBST (16 PBS, 0.05% Tween-20)secondary antibody in PBST was applied for 1 h. Finally, themembrane was washed in PBST again and bands weredetected using the enhanced chemiluminescence system(Amersham Buchler; Braunschweig, Germany) according tothe manufacturer's protocol.

Measurement of caspase-3-like activity

Caspase-3 activity was measured as described (Zhou et al.,1997) with some modi®cations. Cells were washed twice withPBS and carefully collected in PBS by use of a cell scraper(Costar; Cambridge, MA, USA). Cells were spun down at200 g, 48C for 5 min and cell pellets were resuspended inbu�er A containing 20 mM N-[2-hydroxyethyl]piperazine-N'-

[2-ethanesulfonic acid] (HEPES), pH 7.4, 10 mM KCl, 2 mM

MgCl2, 1 mM EDTA, 1 mM dithiothreitol (DTT) andincubated on ice for 15 min. Cell extracts were prepared byhomogenizing cells through a 21 G61 1/2 needle (Braun;Melsungen, Germany). The homogenates were cleared bycentrifugation at 16 000 g, 48C for 15 min. The centrifugationstep was repeated and the clear supernatants were used forthe caspase assay. After determination of the proteinconcentration using BCA, caspase-3 activity was measuredby mixing 10 ml of extract, 90 ml of bu�er B containing50 mM HEPES, pH 7.4, 100 mM NaCl, 1 mM EDTA,0.1% 3-[(3-Cholamidopropyl)dimethylammonio]-1-propane-sulfonate (CHAPS), 10% saccharose, 5 mM DTT, and 2 mlcolorimetric substrate (10 mM Ac-DEVD-pNA in dimethyl-sulfoxide). Samples were incubated at 378C and theabsorbance at 405 nm was measured in a microplate readerevery 10 min. From the data, speci®c caspase-3 activities werecalculated in mmol Ac-DEVD-pNA cleaved per (min6mgprotein).Additionally, cellular extracts were preincubated with

1 mM dATP and 10 mM cytochrome c for 1 h at 308C tofully in vitro activate procaspase-3 prior to measurement ofthe enzyme activity.

Other procedures

Statistical comparisons were made in these studies by the useof Student's t-test.

AcknowledgmentsThis work was supported by research grants from theDeutsche Forschungsgemeinschaft (SFB 273 and SFB 506),the European Community TMR Program, the UniversityMedical Center Charite and the Berliner Krebsgesellschaft.

References

Bargou RC, Daniel PT, Mapara MY, Bommert K, WagenerC, Kallinich B, Royer HD and DoÈ rken, B. (1995). Int. J.Cancer, 60, 854 ± 859.

Bargou RC, Wagener C, Bommert K, Mapara MY, DanielPT, Arnold W, Dietel M, Guski H, Feller A, Royer HDand DoÈ rken B. (1996). J. Clin. Invest., 97, 2651 ± 2659.

Cohen GM. (1997). Biochem. J., 326, 1 ± 16.Cohen JJ, Duke RC, Fadok VA and Sellins KS. (1992).

Annu. Rev. Immunol., 10, 267 ± 293.Cosset FL, Takeuchi Y, Battini JL, Weiss RA and Collins

MK. (1995). J. Virol., 69, 7430 ± 7436.Daniel PT, Pun KT, Ritschel S, Sturm I, Holler J, DoÈ rken B

and Brown R. (1999). Blood, 94, 1100 ± 1107.Ellis RE, Yuan JY and Horvitz HR. (1991). Annu. Rev. Cell

Biol., 7, 663 ± 698.Essmann F, Wieder T, Otto A, MuÈ ller EC, DoÈ rken B and

Daniel PT. (2000). Biochem. J., 346, 777 ± 783.Fisher DE. (1994). Cell, 78, 539 ± 542.Hannun YA. (1997). Blood, 89, 1845 ± 1853.Hiwasa T, Tokita H, Sakiyama S and Nakagaware A. (1998).

Anticancer Drugs, 9, 82 ± 87.Ibrado AM, Huang Y, Fang G and Bhalla K. (1996a). Cell.

Growth. Di�er., 7, 1087 ± 1094.Ibrado AM, Huang Y, Fang G, Liu L and Bhalla K. (1996b).

Cancer Res., 56, 4743 ± 4748.

JaÈ nicke RU, Sprengart ML, Wati MR and Porter AG.(1998a). J. Biol. Chem., 273, 9357 ± 9360.

JaÈ nicke RU, Ng P, Sprengart ML and Porter AG. (1998b). J.Biol. Chem., 273, 15540 ± 15545.

Kerr J, Wyllie A and Currie A. (1972). Br. J. Cancer, 26,239 ± 257.

Krajewski S, Blomqvist C, Franssila K, Krajewska M,Wasenius VM, Niskanen E, Nordling S and Reed JC.(1995). Cancer Res., 55, 4471 ± 4478.

Krammer PH, Behrmann I, Daniel P, Dhein J and DebatinKM. (1994). Curr. Opin. Immunol., 6, 279 ± 289.

Kroemer G and Reed JC. (2000). Nat. Med., 6, 513 ± 519.Kuida K, Zheng TS, Na S, Kuan C, Yang D, Karasuyama H,

Rakic P and Flavell RA. (1996). Nature, 384, 368 ± 372.Laemmli UK. (1970). Nature, 227, 680 ± 685.Liu X, Kim CN, Yang J, Jemmerson R and Wang X. (1996).

Cell, 86, 147 ± 157.Lotem J and Sachs L. (1993). Cell Growth Di�er., 4, 41 ± 47.Miller AD and Rosman GJ. (1989). Biotechniques, 7, 980 ±

982, 984 ± 986, 989 ± 990.Miyashita T and Reed JC. (1993). Blood, 81, 151 ± 157.Naundorf H, Rewasowa EC, Fichtner I, Buttner B, Becker

M and Gorlich M. (1992). Breast Cancer Res. Treat., 23,87 ± 95.

Oncogene

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2759

Nicoletti I, Migliorati G, Pagliacci MC, Grignani F andRiccardi C. (1991). J. Immunol. Methods, 139, 271 ± 279.

Perkins C, Kim CN, Fang G and Bhalla KN. (1998). CancerRes., 58, 4561 ± 4566.

Prokop A, Wieder T, Sturm I, Eûmann F, Seeger K, WuchterC, Ludwig W-D, Henze G, DoÈ rken B and Daniel PT.(2000). Leukemia, 14, 1606 ± 1613.

Slee EA, Harte MT, Kluck RM, Wolf BB, Casiano CA,Newmeyer DD, Wang HG, Reed JC, Nicholson DW,Alnemri ES, Green DR and Martin SJ. (1999). J. CellBiol., 144, 281 ± 292.

Slee EA, Keogh SA andMartin SJ. (2000). Cell Death Di�er.,7, 556 ± 565.

Smith PK, Krohn RI, Hermanson GT, Mallia AK, GartnerFH, Provenzano MD, Fujimoto EK, Goeke NM, OlsonBJ and Klenk DC. (1985). Anal. Biochem., 150, 76 ± 85.

Stein U, Walther W, Lemm M, Naundorf H and Fichtner I.(1997). Int. J. Cancer, 72, 885 ± 891.

Sturm I, KoÈ hne C-H, Wol� G, Petrowsky H, Hillebrand T,Hauptmann S, Lorenz M, DoÈ rken B and Daniel PT.(1999). J. Clin. Oncol., 17, 1364 ± 1374.

Sturm I, Papadopoulos S, Hillebrand T, Benter T, LuÈ ck H-J,Wol� G, DoÈ rken B and Daniel PT. (2000). Int. J. Cancer,87, 517 ± 521.

Suzuki A, Kawabata T and Kato M. (1998). Eur. J.Pharmacol., 343, 87 ± 92.

Thornberry NA and Lazebnik Y. (1998). Science, 281,1312 ± 1316.

Uckert W, Willimsky G, Pedersen FS, Blankenstein T andPedersen L. (1998). Hum. Gene Ther., 9, 2619 ± 2627.

Wagener C, Bargou RC, Daniel PT, Bommert K, MaparaMY, Royer HD and DoÈ rken B. (1996). Int. J. Cancer, 67,138 ± 141.

Wesselborg S, Engels IH, Rossmann E, Los M and Schulze-Ostho� K. (1999). Blood, 93, 3053 ± 3063.

Wieder T, Geilen CC, Wieprecht M, Becker A and OrfanosCE. (1994). FEBS Lett., 345, 207 ± 210.

Wieder T, Orfanos CE and Geilen CC. (1998). J. Biol. Chem.,273, 11025 ± 11031.

Wieder T, Eûmann F, Prokop A, Schmelz K, Schulze-Ostho� K, Beyaert R, DoÈ rken B and Daniel PT. (2001).Blood, 97. 1378 ± 1387.

Wyllie AH. (1980). Nature, 284, 555 ± 556.Xiang J, Chao DT and Korsmeyer SJ. (1996). Proc. Natl.

Acad. Sci. USA, 93, 14559 ± 14563.Zhou Q, Snipas S, Orth K, Muzio M, Dixit VM and Salvesen

GS. (1997). J. Biol. Chem., 272, 7797 ± 7800.

Caspase-3 gene transfer and chemoresistanceK Friedrich et al

2760

Oncogene