PED mediates AKT-dependent chemoresistance in human breast cancer cells

PED Mediates AKT-Dependent Chemoresistance in

Human Breast Cancer Cells

Giorgio Stassi,3Michela Garofalo,

1Monica Zerilli,

3Lucia Ricci-Vitiani,

5

Ciro Zanca,1Matilde Todaro,

3Federico Aragona,

4Gennaro Limite,

2

Giuseppe Petrella,2and Gerolama Condorelli

1

Departments of 1Cellular and Molecular Biology and Pathology and 2Endocrinology and Molecular and Clinical Oncology, ‘‘Federico II’’University of Naples, Naples; 3Department of Surgical and Oncological Sciences, 4Pathology Institute, University of Palermo,Palermo; 5Department of Hematology, Oncology and Molecular Medicine, Istituto Superiore di Sanita, Rome, andInstituto Oncologico del Mediterraneo, Catania, Italy

Abstract

Killing of tumor cells by cytotoxic therapies, such as chemo-therapy or gamma-irradiation, is predominantly mediated bythe activation of apoptotic pathways. Refractoriness toanticancer therapy is often due to a failure in the apoptoticpathway. The mechanisms that control the balance betweensurvival and cell death in cancer cells are still largely unknown.Tumor cells have been shown to evade death signals throughan increase in the expression of antiapoptotic molecules orloss of proapoptotic factors. We aimed to study the involve-ment of PED, a molecule with a broad antiapoptotic action, inhuman breast cancer cell resistance to chemotherapeuticdrugs–induced cell death. We show that human breast cancercells express high levels of PED and that AKT activity regulatesPED protein levels. Interestingly, exogenous expression of adominant-negative AKT cDNA or of PED antisense in humanbreast cancer cells induced a significant down-regulation ofPED and sensitized cells to chemotherapy-induced cell death.Thus, AKT-dependent increase of PED expression levelsrepresents a key molecular mechanism for chemoresistancein breast cancer. (Cancer Res 2005; 65(15): 6668-75)

Introduction

PED (known also as PED/PEA-15) is a death effector domain(DED)-family member of 15 kDa with different effects on cellgrowth and metabolism (1–3). The protein consists of an NH2-terminal DED of 80 amino acids and a COOH-terminal region inwhich are located the phosphorylation sites for protein kinase C(Ser104), calcium calmodulin kinase II and AKT (Ser116; refs. 4–6).PED inhibits the formation of a functional death-inducingsignaling complex and caspase 3 activation following treatmentwith different apoptotic cytokines including CD95/FasL, tumornecrosis factor-a (TNF-a) and TNF-related apoptosis-inducingligand (TRAIL; refs. 2, 7, 8). At least in part, the antiapoptoticaction of PED is accomplished through its DED domain, whichlikely acts as a competitive inhibitor for proapoptotic moleculesduring the assembly of the death-inducing signaling complex(2, 9). Protein kinase C and AKT play an important role inenabling the antiapoptotic function of PED (5, 6, 9) by regulatingits specific interactions and localization within the cell. In order

to be recruited into the TRAIL receptor death-inducing signalingcomplex in TRAIL-resistant gliomas, PED has to be doublyphosphorylated (9). Furthermore, PED inhibits the induction ofdifferent stress-activated protein kinases triggered by growthfactor deprivation, hydrogen peroxide, and anisomycin (10). PEDhas been found overexpressed in a number of different tumors,including human gliomas (7) and squamous carcinoma (11).However, the meaning and the molecular mechanisms ofderegulation of PED expression in cancer cells are presentlyunknown. We have recently shown in vitro as well as in intactcells that PED is a substrate of AKT that phosphorylates it onSer116, and that, in turn, phosphorylation of S116 increases PEDprotein stability (6). AKT plays a key role on cell survival (12–14).Furthermore, up-regulation of AKT activity has been implicated ina number of human cancers (15, 16), including breast cancer (17).Converging evidences suggest that deregulation of PED expressionlevels may play an important role in breast cancer (18–20). The 3Vuntranslated region of the 2.5-kb spliced isoform of PED containsthe mammary transforming (MAT1) proto-oncogene. This se-quence, isolated from an experimental mouse mammary tumor,was reported to induce the oncogenic transformation of NIH-3T3cells and the mammary epithelial cell line TM3 (18). The MAT1-containing PED mRNA is expressed weakly during certainphysiologic circumstances, such as pregnancy and lactation, butbecomes strongly expressed in murine mammary tumors (19).Furthermore, ped/pea-15 gene maps at 1q21-q22 between themarkers D1S2635 and D1S4841 (1) a region frequently gained (upto 50%) in breast cancer (20) and rearranged in leukemia (21). Wehave previously shown that in the breast cancer cell line MCF-7,PED overexpression induces resistance toward TNF-a and FasL/CD95-induced apoptosis (2).We therefore investigated if PED could play a role in the

resistance to cell death in breast tumors. We additionally studiedthe contribution of AKT in PED overexpression and in breastcancer resistance to chemotherapy.

Materials and Methods

Cell culture. Human MCF7, MDA-MB-231, MDA-MB-131, ZR-78 breastcancer cell lines were grown in RPMI 1640 (MDA) or DMEM (MCF7 and

ZR-78; Life Technologies, Grand Island, NY) containing 10% heat-

inactivated fetal bovine serum supplemented with 2 mmol/L L-glutamine

and 100 units/mL penicillin-streptomycin.Human SK Br-3 and MDA-MB-468 were kindly provided by Dr.

Normanno (Istituto Pascale, University of Naples, Naples, Italy) and grown

in RPMI. MCF-10A cell line was kindly provided by Dr. Normanno andgrown in 50% MF12-50% DMEM (Life Technologies) containing 5% horse

serum supplemented with 2 mmol/L L-glutamine and 100 units/mL

penicillin-streptomycin, 10 Ag/mL insulin, 0.5 Ag/mL hydrocortisone,

Note: G. Stassi and M. Garofalo contributed equally to this work.Requests for reprints: Gerolama Condorelli, Dipartimento di Biologia e Patologia

Cellulare e Molecolare, Facolta di Scienze Biotecnologiche, Universita degli Studi diNapoli ‘‘Federico II’’, Via Pansini 5, 80131 Naples, Italy. Phone: 39-81-746-2768; Fax: 39-81-770-1016; E-mail: [email protected].

I2005 American Association for Cancer Research.

Cancer Res 2005; 65: (15). August 1, 2005 6668 www.aacrjournals.org

Research Article

10 Ag/mL epidermal growth factor (Sigma, St. Louis, MO). Cells were kept ina 5% CO2 atmosphere and routinely passaged when 80% to 85% confluent.

Breast cells purification from human tissue. Specimens were

manually cut in small pieces and digested for 4 hours at 37jC in a shaking

incubator with 0.5 mg/mL collagenase (Type II, Life Technologies) inDMEM/F12 (Life Technologies) supplemented with 10% fetal bovine serum.

Following enzymatic digestion, cells were centrifuged at 1,500 � g and

plated in 75 cm2 culture flasks coated with 5 Ag/mL type I collagen

(Calbiochem-Novabiochem, Darmstadt, Germany). Purified cells wereallowed to grow in monolayer and detached with trypsin plus EDTA for

functional and protein expression analyses.

Materials. Media, sera, and antibiotics for cell culture were from Life

Technologies. Protein electrophoresis reagents were from Bio-Rad Labora-tories (Richmond, VA) and Western blotting and enhanced chemilumines-

cence reagents from Amersham (Arlington Heights, IL). All other chemicals

were from Sigma.Flow cytometry. Cytometric staining was analyzed on a FACSCalibur

Instrument (Becton Dickinson, Franklin Lakes, NJ) and data were analyzed

with Cell Quest software (Becton Dickinson).

Protein isolation and Western blotting. Breast tissue specimens werecollected from surgical mastectomy of 20 patients affected by adenocar-

cinoma, in accordance with the ethical standards of the institution

responsible (Committee on Human Experimentation). From the same

patient, we collected both neoplastic and adjacent normal tissue. Thepatients did not receive any medical treatment at the time of the operation.

Clinical and pathologic data, including tumor-node-metastasis stage and

estrogen receptor status are shown in Table 1. Tissues were homogenized in1 mL of Harvest buffer [30 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl,

10% glycerol, and 0.1% Triton X-100] containing proteinase inhibitor

cocktail (Sigma), with a tissue homogenator. Cell pellets were washed twice

with cold PBS and resuspended in Harvest buffer. Solubilized proteins wereincubated for 30 minutes on ice. After centrifugation at 10,000 � g for

30 minutes at 4jC supernatants were collected. To remove excess fat,

samples derived from human tissue were precipitated by adding 4 volumesof cold acetone. Briefly, the samples were kept for 30 minutes at �20jC in

the presence of cold acetone and centrifuged at 15,000 � g . The pellets were

dried and resuspended in Harvest buffer. Fifty micrograms of sample

extract were resolved on 12% SDS-polyacrylamide gels using a mini-gelapparatus (Bio-Rad Laboratories) and transferred to Hybond-C extra

nitrocellulose (Amersham Pharmacia Biotech, Piscataway, NJ). Membranes

were blocked for 1 hour with 5% nonfat dry milk in TBS containing 0.05%

Tween 20 and incubated for 2 hours with specific antibodies. The followingantibodies were used for immunoblotting: anti-AKT (Cell Signaling

Technology, Inc., Charlottesville, VA); antiphospho AKT (Ser473; rabbit

polyclonal IgG, Cell Signaling Technology); anti-PED serum previously

described (1) and anti-h actin (Ab-1, mouse IgM; Oncogene, Darmstadt,Germany), anti-p53 from (Santa Cruz Biotechnology, Santa Cruz, CA; sc-

6243). The washed membranes were then incubated for 45 minutes with

horseradish peroxidase–conjugated anti-rabbit or anti-mouse secondaryantibodies (Amersham Pharmacia Biotech) and visualized by using a

chemiluminescence detection system (Amersham Pharmacia Biotech).

Immunohistochemistry. Immunohistochemical staining was done on

5-Am-thick human breast cancer and normal paraffin sections. Deparaffi-nized sections were heated in a microwave oven for antigen unmasking in

10 mmol/L sodium citrate (pH 6.0) for 1 minute at 450 W followed by 5

minutes at 100 W. The slides were then cooled for 20 minutes washed in

distilled water and PBS, and permeabilized with 0.1% Triton X-100/PBS for3 minutes. After rinsing with PBS, sections were incubated with 3%

hydrogen peroxide for 3 minutes and then with blocking solution for 10

minutes. The sections were exposed overnight at 4jC to rabbit anti-PEDantibody (1:100) in 3% bovine serum albumin/PBS and then for 30 minutes

at room temperature with a biotinylated secondary antibody. Finally,

sections were rinsed and incubated with streptavidin for 30 minutes. Nuclei

were counterstained using hematoxylin.Cell death and cell proliferation quantification. Cell viability was

evaluated with the CellTiter 96 AQueous One Solution Cell Proliferation

Table 1. Clinical features of breast cancer patients and PED and p-AKT expression levels

Patient

no.

Age Histologic

type

TNM stage Estrogen

receptor (%)

Progesterone

receptor (%)

erbB2

receptor (%)

PED expression

levels

p-AKT

1 50 CDI, G3 pT2aN1B 20 — — � �2 58 CDI, G3 pT2aN1B 40 — 100 ++++ ++++3 43 CDI, G3 pT1pN1 70 80 — ++ ++4 51 CLI pT2aN0 55 25 — ++ ++5 79 CDI, G2 pT1aN0 50 35 — ++++ ++++6 49 CDI, G1 pT1aN0 50 50 — ++++ +++7 60 CDI, G1 pT1aN0 20 40 — � �8 59 CDI, G2 pT1aN1 60 10 — ++ ++9 56 CDI, G2 pT1aN1B 65 — 30 +++ +++10 67 CDI, G3 pT1bN1B 80 80 — ++ ++11 66 CDI, G2 pT1bN1B 70 65 — +++ +++12 65 CLI pT2aN0 30 45 — � �13 47 CDI, G1 pT1aN0 60 50 — + +14 70 CDI, G3 pT2aN1B 70 40 — +++ +++15 65 CDI, G1 pT1aN0 20 — — � �16 55 CDI, G2 pT1bN1B 60 — 60 ++++ +++17 49 CDI, G1 pT1aN0 70 40 — +++ +++18 52 CDI, G2 pT1aN0 70 50 — ++ ++19 60 CDI, G1 pT1aN0 80 45 — +++ ++20 55 CDI, G2 pT1bN1B 60 — 50 ++++ ++++

NOTE: PED and phospho-AKT expression levels were estimated by densitometric scanning and normalized for the corresponding actin expressionlevels. Data represent fold increase of PED and p-AKT in cancer samples compared with normal.

Abbreviations: CDI, ductal infiltrating carcinoma; CLI, lobular infiltrating carcinoma (49); (�) no increase; (+) 1- to 3-fold increase; (++) 4- to 6-fold

increase; (+++) 6- to 10-fold increase; (++++) >10-fold increase.

PED and Breast Cancer

www.aacrjournals.org 6669 Cancer Res 2005; 65: (15). August 1, 2005

Assay (Promega, Madison, WI), according to the manufacturer’s protocol.

Cells were plated in 96-well plates in triplicate, stimulated and incubated at37jC in a 5% CO2 incubator. Etoposide (7 Amol/L), vincristine (1 Amol/L),

paclitaxel (5 Amol/L), cisplatin (300 ng/mL), and doxorubicin (5 Amol/L)

were used in vitro at doses compatible with the levels reached in vivo during

cancer treatment for 24 hours. Metabolically active cells were detected byadding 20 AL of MTS to each well. After 2 hours of incubation, the plates

were analyzed on a Multilabel Counter (Bio-Rad).

Transduction of breast cells with lentiviral vectors. Gene transfer wasdone using a variant of the third-generation lentiviral vector (the advanced

generation) recently described (22). In order to simultaneously transduce

both reporter and target gene, a new lentiviral vector, Tween, was generated

by engineering pRRLsin.cPPT.hCMV.hPGKGFP.Wpre. In this vector, theoriginal hCMV.GFP cassette was substituted with hCMV.hPGK.GFP. A

multiple cloning site was inserted downstream of hCMV.

PED antisense cDNA was obtained by PCR amplification of the human

PED cDNA using the following primers: 5V-CCCGCTAGCGCTCAATGTAG-GAGAGGTTG-3V and 5V-CCCCTCGAGGCCAGAGCGCGCGGGGCAGTGTG-3Vcontaining the NheI and XhoI cloning sites, respectively (7). The amplified

fragment was subcloned in the XbaI site of the Tween vector. Myc-taggedPED cDNA (1) was subcloned in the XhoI site of the Tween lentivector.

AKT E40 K cDNA (constitutively active AKT cDNA, HA-AKTDP) and AKT

K179M cDNA (dominant-negative AKT cDNA, HA-AKTDN) with an HA-tag

were a kind gift from Prof. G.L. Condorelli (University of Rome ‘‘LaSapienza’’) and were subcloned in the Tween vector.

Lentiviral supernatants were produced by calcium phosphate transient

cotransfection of a three-plasmid expression system in the packaging

human embryonic kidney cell line, 293T. The calcium-phosphate DNAprecipitate was removed after 14 to 16 hours by replacing the medium.

Viral supernatant was collected 48 hours after transfection, filtered through

0.45-Am pore nitrocellulose filters and frozen in liquid nitrogen. On the

same day of transfection, MDA MB-231 and MCF-10A cells were plated insix-well plates in the presence of viral supernatant. Polybrene (4 Ag/mL) was

added to the viral supernatant to improve the infection efficiency (22). Cellswere then centrifuged for 45 minutes at 1,800 rpm and incubated for

75 minutes in a 5% CO2 atmosphere. Following the infection cycles, cells

were washed twice and fresh medium was added. Infection efficiency was

evaluated after 48 hours by flow cytometry.c-Jun-NH2-kinase phosphorylation. The cells were kept in medium

supplemented with 1% serum for 18 hours and then incubated with 0.3 Ag/mL paclitaxel as indicated in the description of the experiment. The cells

were harvested and the lysates immunoblotted as described above. Anti-phospho (Thr183/Tyr185) c-jun-NH2-kinase (JNK; Cell Signaling Technology)

or anti-JNK (Santa Cruz Biotechnology) antibodies were used.

Results

PED expression in breast cancer cell lines. To evaluate PEDexpression levels, Western blot analysis was first done on a panel ofhuman breast cancer–derived cell lines with anti-PED antibodies(Fig. 1A and B). We used the MCF-10A cell line, an immortalizedhuman mammary epithelial-derived cell line that retains manycharacteristics of the normal mammary gland in culture, such asthe ability to form acinar structures. PED was clearly detected in allthe transformed cell lines analyzed, but not in MCF-10A lysates,thus indicating that PED protein expression is restricted to breasttumor–derived cell lines.PED expression in human breast cancer. We next determined

the expression levels of PED in human breast cancer specimens

Figure 1. PED expression in breast cancer cell lines. PED expression levelswere analyzed by Western blotting in different breast cancer cell lines: A, ZR-78,MCF-7, MDA MB-131; B, SK Br-3, MDA MB-231 and MDA MB-468 and in theimmortalized human mammary cell line MCF10-A (A ). Loading control wasassessed on the same blot with anti-h-actin antibody.

Figure 2. PED expression levels in breast cancer samples. Western blotanalysis of breast tissue samples for PED expression and AKT activity. A, breastcarcinoma (T) or adjacent normal (N) tissue samples obtained from patientswere homogenized as described in Materials and Methods. AKT activity wasassessed with specific anti-phospho-Akt (Ser473) antibody. B, in sampleswhere AKT was not found increased, we could not detect any increase inPED expression. Loading control was assessed on the same blot with humanh-actin antibody.

Cancer Research


from 20 patients. All the patients did not receive any medicaltreatment at the time of the operation. Immunoblot analysisshowed that PED was strongly up-regulated in breast cancer cellscompared with cells from the adjacent normal tissue (Fig. 2A ;Table 1). However, in four tumor specimens—out of the 20analyzed—(patients no. 1, 7, 12, and 15), we did not detect anydifference in PED protein expression compared with thecorresponding normal tissues (Table 1; Fig. 2B). Interestingly, thesefour patients did not exhibit erbB2 expression or strong expressionof estrogen receptors (Table 1) at histologic analysis. These dataindicate that PED is preferentially overexpressed in breast tumorscompared with normal epithelium and that this overexpression iscorrelated with either erB2 or estrogen receptor levels. Immuno-histochemistry analysis confirmed the presence of PED in breastcancer tissue (Fig. 3A).

Up-regulation of AKT in breast cancer correlates withoverexpression of PED. We have recently shown that PED is asubstrate of AKT, which may regulate its levels by increasing theprotein’s stability upon phosphorylation of Ser116 (6). AKT has beenfrequently found activated in breast cancer (17, 23, 24). Therefore,it is plausible that AKT is responsible for the accumulation of PEDprotein. In order to address this issue, we analyzed AKT activationin breast tumor samples in which PED was found to beoverexpressed. As expected, Western blot and immunohistochem-ical analysis with anti-phospho-AKT (Ser473) antibodies revealedthat AKT was phosphorylated to a greater extent in tumor-derivedsamples which expressed a higher amount of PED (Table 1; Figs. 2Aand 3B). In samples from patients no. 1, 7, 12, and 15, we could notdetect any difference in AKT activity and PED expression levelsbetween cancerous areas of the specimen (Table 1; Fig. 2B). Thus,AKT activation in human breast cancer, as revealed by bothWestern blot and immunohistochemistry, correlates to the increaseof PED protein levels.PED mediates breast cancer cell death resistance. If AKT

activity is a major determinant for the increased stability ofPED in breast cancer, interfering with AKT signaling shouldmodulate PED levels. We therefore tested this hypothesisevaluating PED expression and cell viability following exposureto different chemotherapeutic agents in primary breast tumorcultures and in the MDA-MB231 cell line infected with alentiviral vector that expresses the HA-tagged dominant-negativeor the dominant-active mutant of AKT (HA-AKTDN, HA-AKTDP).Western blot analysis revealed that viral infection with HA-AKTDN resulted in a significant expression of the kinase and ina reduction of PED expression levels (Fig. 4A and B).Furthermore, infection of with HA-AKTDN significantly increasedbreast cancer cell sensitivity to drug-induced cell death (Fig. 4Cand D). Interestingly, infection of MDA cells with HA-AKTDPincreased PED expression levels (Fig. 4B) and significantlydecreased the sensitivity to drug-induced cell death (Fig. 4D).Finally, to prove that overexpression of PED is responsible forresistance to apoptosis in breast cancer, a PED antisense cDNAwas cloned into the Tween lentiviral vector carrying a greenfluorescent protein (GFP)-fusion protein and then transducedinto the human breast cancer cell line, MDA MB-231.Fluorescence-activated cell sorting analysis of MDA MB-231cells transduced with PED antisense-containing or a controllentiviral vector, showed that the infection produced a virtuallypure population of GFP-positive cells as indicated by thehomogenous right-shifted peak (Fig. 5A). The left-shifted peakrepresents noninfected cells. Transduction of MDA MB-231 withPED antisense considerably reduced PED expression as revealedby immunoblot analysis (Fig. 5B). Accordingly, MDA MB-231transduced with PED antisense showed an increase in drugsensitivity (Fig. 5C ). These data further suggest thatPED protects breast cancer cells from chemotherapy-inducedapoptosis.Inhibition of PED expression in MDA cells reestablishes

c-Jun-NH2-kinase activation. One of the mechanisms ofchemotherapy-induced cell death is activation of the JNK pathway(25, 26). We have recently showed that in 293 cells overexpressingPED, JNK activation following different stress stimuli is impaired(10). To test whether this may also occur in breast cancer, weincubated MDA-MB231 cells expressing PED antisense cDNA(MDA PEDa) in the presence of paclitaxel for different lengths oftime. JNK activation was evaluated by Western blotting with an

Figure 3. Immunohistochemical analysis of PED expression and AKT activityin breast cancer. A, immunohistochemical analysis of paraffin-embeddednormal or breast cancer sections labeled with control preimmune IgG (NC),cytokeratin (CK), or anti-PED antibody (1:100) and revealed by AEC(red staining). One representative of two independent experiments isshown. B, immunohistochemical analysis of phospho Akt (Ser473) done onparaffin-embedded sections of normal or cancer breast tissues (red staining).Hearts from AKT transgenic mice (48) were used as positive controls. Onerepresentative of two independent analyses using specimens from differentpatients is shown.



anti-phospho-JNK antibody that recognizes two JNK isoforms. Asshown in Fig. 6, MDA PEDa, different from MDA control cells, hada consistent and sustained increase in JNK phosphorylation. Asrevealed with Western blot analysis, in both cell types, treatmentwith paclitaxel at longer time points induced a slight decrease inPED expression levels.Thus, in breast cancer, PED overexpression may determine

increased resistance to chemotherapeutic drug-induced cell deaththrough an inhibition of the stress kinase pathway.Effect of PED overexpression in normal mammary cells. In

order to study the role of PED in normal breast tissue, Myc-tagged PED cDNA, cloned into the Tween lentiviral vector, wastransduced into the human breast primary cells obtained fromnormal tissue and into MCF-10A cells (Figs. 7 and 8,respectively). We also transduced these cells with the HA-AKTDP lentiviral vector, previously described. Fluorescence-activated cell sorting analysis of transduced cells showed thatthe infection produced a virtually pure population of GFP-positive cells (data not shown). Western blot analysis with anti-PED antibody or with anti-HA antibody revealed that viralinfection resulted in a significant expression of the proteins(Figs. 7A and 8A).First, we studied the effect of PED on expression and

activation of AKT, JNK, and p53 in normal conditions or understress (paclitaxel treatment). AKT, p-AKT (Figs. 7B and 8B), p53,and JNK levels (Figs. 7C and 8C), were not affected by PEDoverexpression either in the presence or in the absence ofpaclitaxel. This strongly suggests that in breast cancer, activationof AKT is responsible for the increase in the expression of PEDand not vice versa. The overexpression of PED in epithelial cellsfrom normal breast induced a slight but significant increase inthe rate of cell proliferation, as revealed in Figs. 7D and 8D ,indicating that PED may also participate in the regulation of

proliferation beside that of cell death. Interestingly, in MCF-10Acells, PED and AKT overexpression induced an increase in celldeath resistance upon 24 hours incubation in serum deprivationconditions (Fig. 8E).

Discussion

The selective induction of cell death by drugs or cytokines is thefinal goal of any new therapeutic strategy for cancer. Apoptosis isbelieved to be the main mechanism of cell death induced bychemotherapeutics in cancer cells (27–29). However, in a numberof patients, tumor cells evade death signals generated by drugsthrough the activation of effective antiapoptotic mechanisms, suchas increased levels of caspase inhibitors or proteins of the Bcl-2family (30, 31).Increased expression of DED family members with antiapoptotic

functions such as c-FLIP and PED could also represent amechanism of apoptosis resistance in cancer cells (2, 3, 32–34).Both these molecules act primarily by preventing the interactionbetween the adaptor molecule FADD and procaspase-8. PED is arecently identified protein featuring a broad antiapoptotic function(6–9). We have previously shown that PED inhibits the anti-apoptotic signal of CD95/FasL, TNF-a, and TRAIL in manydifferent cell types, including breast carcinoma (2, 7, 8). PEDinteracts in vitro and in intact cells with the serine-threoninekinase AKT (6). AKT is also able to phosphorylate PED at Ser116 (6).AKT-mediated phosphorylation of PED increases its stability, thusdetermining an increase in the content of this protein within thecell (6).In this work, we show that the differential activation of AKT in

normal and transformed breast tissues represents the mainmechanism responsible for PEDoverexpression and thus contributesto the relative increase in the resistance to chemotherapy-induced

Figure 4. Inhibition of AKT activity results in reductionof PED expression. Immunoblot analysis of freshlypurified primary breast epithelial cancer cells (A) andMDA-MB231 cells (B) transduced with empty vector,with hemagglutinin (HA)-tagged dominant-negativeAKT mutant (AKTDN) or with a dominant-active AKTmutant (AKTDP). Twenty micrograms of proteins wereloaded for each sample. Inhibition of AKT activityresults in reduction of PED expression, whereasactivation of AKT results in an increase. AKTexpression was assessed by blotting for HA. Loadingcontrol was assessed by h-actin staining. C,percentage of viable primary breast epithelial cancer or(D ) MDA-MB231 cells transduced as in (A and B ) after24 hours of exposure to chemotherapeutic drugs.

Cancer Research


apoptosis observed in breast cancer. Inmost of the patients analyzed,we identified a tight correlation between AKT activation and PEDlevels. Furthermore, in thiswork, we show that exogenous expressionof a dominant-negative AKT cDNA or of PED antisense cDNA inhuman breast cancer cells induced a significant down-regulation ofPED and sensitized cells to chemotherapy-induced cell death. Thus,the levelofPEDseemstobedirectly correlated toapoptosis resistancein breast cancer.Genetic and biochemical evidence suggests that activation of the

phosphoinositide-3-kinase/AKT pathway contributes to breastcancer tumorigenesis. Up-regulation of AKT activity in breastcancer could be the result of mutations in the AKT-phosphatase

Figure 6. Down-regulation of PED in MDA MB-231 cells reestablishes the abilityof paclitaxel to induce phosphorylation of JNK. Western blot analysis of JNK,phospho-JNK, and PED expression in MDA-MB231 cells transduced with emptyvector (MDA) or with Tween/PED antisense (MDA PEDa) and treated for 2 or4 hours with 0.3 Ag/mL paclitaxel.

Figure 5. Down-regulation of PED in MDA MB 231 cells increases sensitivityto chemotherapy-induced cell death. A, flow cytometry analysis. An empty‘‘Tween’’ lentiviral vector carrying a GFP-fusion protein (top ) or the PEDantisense cDNA (bottom ) were transduced into MDA MB-231 cells. FACSanalysis of transduced cells shows that the infection produced a virtually purepopulation of GFP-positive cells as indicated by the right-shifted peak (blackpeak). B, immunoblot analysis of PED expression in MDA-MB231 transduced asabove. C, percentage of viable cells in MDA vector or MDA PEDa transducedcells exposed to vincristine or paclitaxel for 24 hours. One of four typicalexperiments is shown.

Figure 7. Effect of PED overexpression in primary cultures from normal breasttissue. A, immunoblot analysis of freshly purified primary breast epithelial cellstransduced with empty vector (Tween), with Myc-Tagged PED cDNA (Myc-PED)or with (HA)-tagged dominant-active AKT mutant (AKTDP). Twenty microgramsof proteins were loaded. AKT expression was assessed by blotting for HA.B, Western blot analysis of AKT, phospho-AKT and (C ) JNK and p53 expressionin cells infected with PED cDNA (PED) or control empty vector (vector), in basalconditions (NT) or treated with 0.3 Ag/mL paclitaxel for 12 hours (T). Loadingcontrol was assessed by h-actin staining. D, proliferation rate of primary breastepithelial cells transduced with empty vector or PED as indicated. Cells wereplated at a density of 5,000 cells per well and growth rate was evaluated by thecell titer up to 12 days. One of three representative experiments is shown.



PTEN (35–39). Furthermore, overexpression of the epidermalgrowth factor receptor family member erbB2 (HER-2/neu) receptor(40–42), and overexpression of estrogen receptors have been shownto increase AKT activity (43, 44). 17h-Estradiol reduces apoptosisboth in vitro and in animal models through estrogen receptors andphosphoinositide-3-kinase/AKT-dependent pathways. Treatmentof cells with the estrogen receptor antagonist, ICI, blocked theincrease in phospho-AKT after stimulation with 17h-estradiol,supporting the hypothesis that AKT activation may occur throughan estrogen receptor–dependent mechanism (43). In addition,several compelling evidences support a key role played by AKT indetermining tumor cell resistance to chemotherapy (45–47). Thus,one of the mechanisms of apoptosis resistance observed in humancancer characterized by constitutive AKT activation might involvean increase of PED levels.PED overexpression was also observed in non–small cell lung

cancer, in thyroid cancer and in B-lymphocytes from patientsaffected by B cell chronic leukemia thus suggesting that PED mayplay a general role in cancer development.6 Whether the increase inPED protein level is also mediated by AKT phosphorylation inthese other tumors remains to be determined.Chemotherapeutics induce cell death by activating JNK and p38

kinases. Activation occurs in a dose- and time-dependent manner(25, 26). Resistance to cell death may implicate an impairment inthe activation of these stress kinase pathways. Indeed, theinhibition of JNK activity greatly reduces paclitaxel-induced celldeath in cancer cells (25). We have previously shown that PEDoverexpression inhibits the induction of different stress-activatedprotein kinases triggered by growth factor deprivation, hydrogenperoxide and anisomycin (10). Interestingly, in this work, we showthat in breast cancer, inhibition of PED expression results insustained JNK activation.Because expression levels of PED are a focal point for apoptosis

regulation, PED represents an attractive target for therapeuticintervention in cancer treatment.

Acknowledgments

Received 11/18/2004; revised 4/29/2005; accepted 5/16/2005.Grant support: MIUR legge 449/97, Ministero della Salute, MIUR-PRINN 2004 pret.

2004060785-002, EU grant European Molecular Imaging Laboratories Network contractno. 503569, and Associazione Italiana per la Ricerca sul Cancro to G. Stassi.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Vittorio de Franciscis and Prof. Gianluigi Condorelli for suggestionsand constructive criticism, Prof. Giampaolo Tortora for critical reading of the manu-script, and Profs. Roberto Di Lauro and Silvestro Formisano for continuous support.

Figure 8. Effect of PED overexpression in MCF-10A cells. A, immunoblotanalysis of freshly purified primary breast epithelial cells transduced with emptyvector (Tween), with Myc-Tagged PED cDNA (Myc-PED) or with (HA)-taggeddominant-active AKT mutant (AKTDP). Twenty micrograms of proteins wereloaded. AKT expression was assessed by blotting for HA. B, Western blotanalysis of AKT, phospho-AKT and (C ) JNK and p53 expression in cells infectedwith PED cDNA in basal conditions (NT) or treated with 0.3 Ag/mL paclitaxel for12 hours (T). Loading control was assessed by h-actin staining. D, proliferationrate of MCF-10A cells transduced with empty vector or PED as indicated.Cells were plated at a density of 10,000 cells per well and growth rate wasevaluated by the cell titer up to 4 days. E, experiments were done in triplicate andwere repeated thrice (*, P < 0.005). Percentage of apoptotic cells in MCF-10Avector, MCF-10A PED, or AKTDP transduced cells in the absence of serum for24 hours. One of three representative experiments is shown. 6 G. Condorelli, M. Garofalo, and C. Zanca, unpublished data.

References1. Condorelli G, Vigliotta G, Iavarone C, et al. PED/PEA-15 gene controls glucose transport and is overexpressedin type 2 diabetes mellitus. EMBO J 1998;17:3858–65.

2. Condorelli G, Vigliotta G, Cafieri A, et al. PED/PEA-15:an anti-apoptotic molecule that regulates FAS/TNFR1induced apoptosis. Oncogene 1999;18:4409–15.

3. Renault F, Formstecher E, Callebaut I, Junier MP,Chneiweiss H. The multifunctional protein PEA-15 is

involved in the control of apoptosis and cell cycle inastrocytes. Biochem Pharmacol 2003;66:1581–8.

4. Araujo H, Danziger N, Cordier J, Glowinski J,Chneiweiss H. Characterization of PEA-15, a majorsubstrate for protein C in astrocytes. J Biol Chem 1993;268:5911–20.

5. Kubes M, Cordier J, Glowinski J, Girault JA,Chneiweiss HJ. Endothelin induces a calcium-dependentphosphorylation of PEA-15 in intact astrocytes: identi-fication of Ser104 and Ser116 phosphorylated, respec-

tively, by protein kinase C and calcium/calmodulinkinase II in vitro . J Neurochem 1998;71:1307–14.

6. Trencia A, Perfetti A, Cassese A, et al. Protein kinaseB/AKT binds and phosphorylates PED/PEA-15, stabi-lizing its antiapoptotic action. Mol Cell Biol 2003;23:4511–21.

7. Hao C, Beguinot F, Condorelli G, et al. Inductionand intracellular regulation of TRAIL-induced apo-ptosis in human malignant glioma cells. Cancer Res2001;61:1–9.

Cancer Research


8. Kitsberg D, Formstecher E, Fauquet M, et al.Knock-out of the neural death effector domain proteinPEA-15 demonstrates that its expression protectsastrocytes from TNFa-induced apoptosis. J Neurosci1999;19:8244–51.

9. Xiao C, Yang BF, Asadi N, Beguinot F, Hao C. Tumornecrosis factor–related apoptosis-inducing ligand-induced death-inducing signaling complex and itsmodulation by c-FLIP and PED/PEA-15 in glioma cells.J Biol Chem 2002;277:25020–5.

10. Condorelli G, Trencia A, Vigliotta G, et al. Multiplemembers of the mitogen-activated protein kinase familyare necessary for PED/PEA-15 anti-apoptotic function.J Biol Chem 2002;277:11013–8.

11. Dong G, Loukinova E, Chen Z, et al. Molecularprofiling of transformed and metastatic murine squa-mous carcinoma cells by differential display and cDNAmicroarray reveals altered expression of multiple genesrelated to growth, apoptosis, angiogenesis, and theNF-nB signal pathway. Cancer Res 2001;61:4797–808.

12. Dudek H, Datta SR, Franke TF. Regulation ofneuronal survival by the serine-threonine protein kinaseAkt. Science 1997;275:661–5.

13. Li B, Desai SA, MacCorkle-Chosnek RA, Fan L,Spencer DM. A novel conditional Akt ‘survival switch’reversibly protects cells from apoptosis. Gene Ther 2002;9:233–44.

14. Franke TF, Kaplan DR, Cantley LC. PI3K: down-stream AKTion blocks apoptosis. Cell 1997;88:435–7.

15. Nicholson KM, Anderson NG. The protein kinaseB/Akt signalling pathway in human malignancy. CellSignal 2002;14:381–95.

16. Testa JR, Bellacosa A. AKT plays a central role intumorigenesis. Proc Natl Acad Sci U S A 2001;98:10983–5.

17. Bellacosa A, de Feo D, Godwin AK. Molecularalterations of the AKT2 oncogene in ovarian and breastcarcinomas. Int J Cancer 1995;64:280–5.

18. Bera TK, Guzman RC, Miyamoto S, et al. Identifica-tion of a mammary transforming gene (MAT1) associ-ated with mouse mammary carcinogenesis. Proc NatlAcad Sci U S A 1994;91:9789–93.

19. Tsukamoto T, Yoo J, Hwang SI, et al. Expression ofMAT1/PEA-15 mRNA isoforms during physiological andneoplastic changes in the mouse mammary gland.Cancer Lett 2000;149:105–13.

20. Hwang S, Kuo WL, Cochran JF. Assignment ofHMAT1, the human homolog of the murine mammarytransforming gene (MAT1) associated with tumorigen-esis, to 1q21.1, a region frequently gained in humanbreast cancers. Genomics 1997;42:540–2.

21. Craig RW, Jabs EW, Zhou P, et al. Human and mousechromosomal mapping of the myeloid cell leukemia-1gene: MCL1 maps to human chromosome 1q21, a regionthat is frequently altered in preneoplastic and neoplasticdisease. Genomics 1994;23:457–63.

22. Follenzi A, Ailles LE, Bakovic S, Geuna M, Naldini L.Gene transfer by lentiviral vectors is limited by nucleartranslocation and rescued by HIV-1 pol sequences. NatGenet 2000;217–22.

23. Nakatani K, Thompson DA, Barthel A. Up-regulationof Akt3 in estrogen receptor-deficient breast cancersand androgen-independent prostate cancer lines. J BiolChem 1999;274:21528–32.

24. Bacus SS, Altomare DA, Lyass L. AKT2 is frequentlyupregulated in HER-2/neu-positive breast cancers andmay contribute to tumor aggressiveness by enhancingcell survival. Oncogene 2002;21:3532–40.

25. Lee LF, Li G, Templeton DJ, Ting JP. Paclitaxel(Taxol)-induced gene expression and cell death are bothmediated by the activation of c-Jun NH2-terminal kinase(JNK/SAPK). J Biol Chem 1998;273:28253–60.

26. MacKeigan JP, Collins TS, Ting JP. MEK inhibitionenhances paclitaxel-induced tumor apoptosis. J BiolChem 2000;275:38953–6.

27. Debatin KM, Krammer PH. Death receptors inchemotherapy and cancer. Oncogene 2004;23:2950–66.

28. Kaufmann SH, Earnshaw WC. Induction of apoptosisby cancer chemotherapy. Exp Cell Res 2000;256:42–9.

29. Kaufmann SH, Vaux DL. Alterations in the apoptoticmachinery and their potential role in anticancer drugresistance. Oncogene 2003;22:7414–30.

30. Johnstone RW, Ruefli AA, Lowe SW. Apoptosis: a linkbetween cancer genetics and chemotherapy. Cell 2002;108:153–64.

31. Okada H, Mak TW. Pathways of apoptotic and non-apoptotic death in tumor cells. Nat Rev Cancer 2004;4:592–603.

32. Tschopp J, Irmler M, Thome M. Inhibition of Fasdeath signals by FLIPs. Curr Opin Immunol 1998;10:552–8.

33. French LE, Tschopp J. Inhibition of death receptorsignaling by FLICE-inhibitory protein as a mechanismfor immune escape of tumors. J Exp Med 1999;190:891–4.

34. Irmler M, Thome M, Hahne M, et al. Inhibition ofdeath receptor signals by cellular FLIP. Nature 1997;388:190–5.

35. Marsh DJ, Dahia PL, Zheng Z. Germline mutations inPTEN are present in Bannayan-Zonana syndrome. NatGenet 1997;16:333–4.

36. Singh B, Ittmann MM, Krolewski JJ. Sporadic breastcancers exhibit loss of heterozygosity on chromosome

segment 10q23 close to the Cowden disease locus.Genes Chromosomes Cancer 1998;21:166–71.

37. Rhei E, Kang L, Bogomolniy F, Federici MG, BorgenPI, Boyd J. Mutation analysis of the putative tumorsuppressor gene PTEN/MMAC1 in primary breastcarcinomas. Cancer Res 1997;57:3657–9.

38. Feilotter HE, Coulon V, McVeigh JL. Analysis of the10q23 chromosomal region and the PTEN gene inhuman sporadic breast carcinoma. Br J Cancer 1999;79:718–23.

39. Depowski PL, Rosenthal SI, Ross JS. Loss ofexpression of the PTEN gene protein product isassociated with poor outcome in breast cancer. ModPathol 2001;14:672–6.

40. Bacus SS, Stancovski I, Huberman E, et al. Tumor-inhibitory monoclonal antibodies to the HER-2/Neureceptor induce differentiation of human breast cancercells. Cancer Res 1992;52:2580–9.

41. Kim R, Tanabe K, Uchida Y, Osaki A, Toge T. The roleof HER-2 oncoprotein in drug-sensitivity in breastcancer. Oncol Rep 2002;9:3–9.

42. Menard S, Pupa SM, Campiglio M, Tagliabue E.Biologic and therapeutic role of HER2 in cancer.Oncogene 2003;22:6570–8.

43. Patten R, Pourati I, Aronovitz MJ, et al. 17B-Estradiolreduces cardiomyocyte apoptosis in vivo and in vitro viaactivation of phospho-Inositide-3 kinase/Akt signaling.Circ Res 2004;95:692–9.

44. Alexaki V, Charalampopoulos I, Kampa M, et al.Estrogen exerts neuroprotective effects via membraneestrogen receptors and rapid Akt/Nos activation. FASEBJ 2005;18:1594–6.

45. Clark AS, West K, Streicher S, Dennis PA.Constitutive and inducible Akt activity promotesresistance to chemotherapy, trastuzumab, or tamox-ifen in breast cancer cells. Mol Cancer Ther 2002;1:707–17.

46. Knuefermann C, Lu Y, Liu B. HER2/PI-3K/Aktactivation leads to a multidrug resistance in humanbreast adenocarcinoma cells. Oncogene 2003;22:3205–12.

47. Perez-Tenorio G, Stal O. Southeast Sweden breastcancer group. Activation of AKT/PKB in breast cancerpredicts a worse outcome among endocrine treatedpatients. Br J Cancer 2002;86:540–5.

48. Condorelli G, Drusco A, Stassi G, et al. Akt inducesenhanced myocardial contractility and cell size in vivoin transgenic mice. Proc Natl Acad Sci U S A 2002;99:12333–8.

49. Benson JR, Weaver DL, Mittra I, Hayashi M. The TNMstaging system and breast cancer. Lancet Oncol 2003;4:56–60.



Documents

PED mediates AKT-dependent chemoresistance in human breast cancer cells