Antiproliferative_cancro Da Mama

7/27/2019 Antiproliferative_cancro Da Mama

1/12

Anti-Cancer Drugs 1997, 8, pp. 470-481

Antiproliferative effect of curcumin(diferuloylmethane) against human breast tumorcell lines

Kapil Mehta, Panayotis Pantazis,2 Theresa McQueen and Bharat B Aggarwal1Department of Bioimmunotherapy, and 1Cytokine Research Section, Department of Molecular Oncology,The University of Texas MD Anderson Cancer Center, and 2The Stehlin Foundation for CancerResearch, Houston, TX 77030, USA. Tel: (+1) 713792-2649; Fax: (+1) 713796-1731.

Pharmacologically safe compounds that can inhibit theproliferation of tumor cells have potential as anticanceragents. Curcumin, a diferuloylmethane, is a major activecomponent of the food flavor turmeric (Curcuma longa) thatexhibits anticarcinogenic properties in vivo. In vitro, itsuppressed c-junIAp-1 and NF-KB activation and type 1human immunodeficiency virus long-terminal repeat-direc-ted gene expression. We examined the anti proliferativeeffects of curcumin against several breast tumor cell lines,including hormone-dependent and -independent and multi-drug-resistant (MDR) lines. Cell growth inhibition wasmonitored by [3H]thymidine incorporation, Trypan blueexclusion, crystal violet dye uptake and flow cytometry. Allthe cell lines tested, including the MDR-positive ones, werehighly sensitive to curcumin. The growth inhibitory effectof curcumin was time- and dose-dependent, and correlatedwith its inhibition of ornithine decarboxylase activity.Curcumin preferentially arrested cells in the G2/S phase ofthe cell cycle. Curcumin-induced cell death was neither dueto apoptosis nor to any significant change in the expres-sion of apoptosis-relatedgenes, including Bcl-2, p53, cyclinBand trans glutaminase. Overall our results suggest thatcurcumin is a potent antiproliferative agent for breasttumor cells and may have potential as an anticancer agent.Key words: Cell cycle, breast carcinoma, drugresistance, flavonoids, ornithine deoxycarboxylase.

IntroductionCurcumin (diferuloylmethane), a non-nutritive foodchemical present in turmeric (Curcuma longa),inhibits lipid-peroxide-induced DNA damage andtumor initiation induced by benzo[a]pyrene and7, 12-dimethylbenz[a]anthracene. -4 Phorbol ester-induced tumor promotion is also inhibited bycurcumin.5.6 Besides its anticarcinogenic effects,curcumin exhibits anti-inflammatory properties in

ViVO!-9 Pharmacological safety of curcumin is wellestablished due to the fact that people in certaincountries have for centuries consumed curcumin asa dietary spice in amounts up to 100 mg/day with-7out any harm.

In vitro curcumin inhibits neutrophil activation,suppresses mitogen-induced proliferation of bloodmononuclear cells, inhibits the mixed lymphocytereaction and inhibits proliferation of smooth musclecells.10,11 It is also a potent scavenger of reactiveoxygen species, protects hemoglobin from nitrite-induced. oxidation to methemoglobin and inhibitslipid peroxidation.12-14 Some of these activities arealso responsible for its ability to protect DNA fromfree radical-induced dama~e and hepatocytes fromd b ..14.15mage y VarIOUS tOXInS. ,

How curcumin induces such a wide range ofeffects is not fully understood. Among the possibili-ties are its profound ability to inhibit the phorbolester-induced expression of the nuclear transcriptionfactor c-jun/AP-l,16 inhibition the nuclear transcrip-tion factor NF-KB,17 and inhibition of protein kinaseC (PKC) and xanthine dehydrogenase/oxidase.18,19Ornithine decarboxylase (ODC; EC 4.1.1.17), a rate-limiting enzyme in polyamine synthesis, and tyrosineprotein kinase are inhibited by curcumin:o It alsoinhibits the activity of several different serine/threonine and tyrosine kinases in vitrO.21 Morerecently, curcumin was shown to inhibit both theepidermal growth factor (EGF) receptor intrinsickinase activity and EGF-mediated activation of EGFreceptor phosphorylation.22,23 These studies suggestthat curcumin may serve as a growth inhibitoryagent by interfering with certain signal transductionpathways that are critical for cell growth.

The potential of clircumin as an anticancer agentled us to investigate the effect of this compound onproliferation and cell growth of several breast tumorcells. The results described in this paper suggest thatcurcumin could serve as an effective antitumor agent

Supported, n part, by the Clayton Foundation for Research BBA),the Stehlin Foundation for Cancer Research PP) and a grant fromthe TexasHigher Education Coordinating Board (KM).Correspondence to K Mehta

470 Anti-Cancer Drugs. Vol8 .1997 1997 Rapid Science Publishers


2/12

Antiproliferative effectsof curcuminmedium (RpMI 1640 with 10% FBS) in 96-wellComing plates in the presence or absence ofindicated concentrations of the drug. At appropriatetimes, the medium was removed and cells wereeither counted by a hemocytometer or monitored bycrystal violet staining. In preliminary experimentswe determined that the crystal violet stainingmethod to determine cell viability correlates wellwith cell number determined by detachment with atrypsin solution and counting with a hemocyto-meter. Relative cell viability was calculated by divid-ing optical density in the presence of the testsample by optical density in the absence of the testsample (medium) and multiplying the results by 100.For [3H]thymidine incorporation, cells were cul-tured and treated with curcumin as ndicated above.During the last 6h, [3H]thymidine (6.7 Ci/mmol;New England Nuclear, Boston, MA) was added toeach well (0.5 f.lCi/well). Thereafter, the culturemedium was removed, and the cells washed twicewith phosphate-buffered saline (PBS) and detachedin a 0.5% trypsin solution containing 5.3 mM EDTA.The cell suspensionwas then harvested with the aidof a Filtermate 196 cell harvester (packard Instru-ments, Canberra, Australia) and lysed by washingwith distilled water. Radioactivity bound to the filterwas measured directly by a Direct Beta CounterMatrix 9600 (Model 1600 TR; Packard,Meriden, CT).Thymidine incorporation determined by this methodhas been shown to correlate with cell growth.26 Alldeterminations were made n triplicate.

against both hormone-dependent and -independenthuman breast cancer cells even when they becomerefractory to conventional chemotherapy.

Materials and methodsMaterialsHighly purified curcumin was purchased from Sigma(St Louis, MO). Dulbecco's modified Eagle's medium(DMEM/F12) was obtained from ICN Biomedical(Irvine, CA); RPMI 1640, DMEM and EMEM fromWhit taker MA Bioproducts (Walkersville, MD); fetalbovine serum (FBS) and gentamicin from Gibco(Grand Island, NY). Tetrachloro-diphenylglycouril,glycine and 3-(4-5-dimethylthiozol-2-yl)2-5-diphenyl-tetrazolium bromide (MTT) were obtained fromSigma. The antibodies to human Bcl-2, p53 andguinea-pig liver tissue type transglutanaU1ase TGase)were purchased from Neo-Markers (Fremont, CA)and anti-cyclin B from Santa Cruz Biotechnology(Santa Cruz, CA).

CellsThe human breast tumor cell lines BT-20,T-47D, SK-BR3 and MCF- were obtained from the ATCC(Rockville, MD). The MCF- cells were selected forresistance o adriamycin (MCF- ADR) nd BT-20 cellswere selected or resistance o tumor necrosis factor(BT-201NF)as previously described.24 Cells weretested for Mycoplasma contamination using eitherthe DNA-basedassaykit purchased from Gen-Probe(SanDiego, CA) or the Hoechst stain.

MTT assayThe number of viable cells remaining after appro-priate treatment was determined by using themodified tetrazolium salt (MTI) assayas described.24Briefly, 5 X 103 cells/well were incubated in thepresence or absence of the indicated test sample ina final volume of 0.2 mi for 72 hat 37C. Thereafter,0.1 mi of cell medium was removed and 0.025 mi ofMTT solution (5 mg/ mi in PBS) was added to eachwell. After 2 h incubation at 37C, 0.1 mi of theextraction buffer (20% sodium dodecyl sulfate, 50%dimethyl formamide) was added. After an overnightincubation at 37C, the optical densities at 570 nmwere measured using a 96-well multiscanner auto-reader (Dynatech MR 5000), with the extractionbuffer serving as a blank. The cell viability wasexpressed as a percentage using the following equa-tion:

Cell cultureAll breast tumor cell lines were routinely grown inRPMI 1640 medium supplemented with 10 mMHEPES buffer, 2 mM glutamine, 50 ,ug/ml gentamicinand 10% FCS. The cells were cultured in a humidi-fled incubator in 5% CO2 in air and were maintainedin continuous exponential growth by twice a weekpassage.

Cell proliferation assaysCell growth assays were carried out essentiallyaccording to the procedure described:5 Briefly, thecells (5 X 103/well) were plated in 0.2 ml of the

471nti-Cancer Drugs. VolB .1997


3/12

K Mehta et alODC assay (20 mM, pH 7.2) containing 1 mM EDTA, 1 mMEGTA, 100 mM NaCI, 1 mM phenylmethylsulfonylfluoride, 0.1% aprotinin, 0.1% leupeptin, 0.1% pep-statin and 0.1% Triton X-100. The Iystes werecentrifuged at 16000 9 for 45 mill, and the clearsupernatants were assayed for protein using theamido black dye staining procedure and bovineserum albumin as standard. Fifty micrograms ofprotein from each sample was electrophoresed on10% SDS-polyacrylamide gel and subsequently elec-trotransferred onto nitrocellulose membranes. Themembranes were blocked with 5% non-fat milk andprobed with anti-Bcl-2, anti-p53, anti-TGaseor anti-cyclin Bl antibody. The immunoreactive bands weredetected by using horseradish peroxidase-conjugatedanti-mouse gG as secondary antibody and enhancedchemiluminescence detection system according tothe manufacturer's instructions (Amersham, Arling-ton Heights, It). The membranes were stripped,using a procedure recommended by the manufac-

turer (Amersham), and reprobed with antibody toactin (SantaCruz Biotechnology).

ODC activity in curcumin-treated breast tumor cellswas determined according to a method described byGrewal et al.27 Briefly, after appropriate treatment,the cells (8 XI 06) were lysed by sonication in25 mM Tris-HCI buffer containing 2.5 mM DTT and0.1 mM EDTA.The reaction was carried out in 15 mIFalcon tubes containing 150 mM Tris-HCI (pH 7.5),1 mM EDTA, 22.5 mM DTT, 0.4 mM pyridoxal phos-phate and 4 mM L-r4C]ornithine (New EnglandNuclear, Boston, MA) in a 0.2 mI volume. Thereaction was initiated by adding 0.05 mI of the celllysate (1 mg protein). The tubes were immediatelycapped and incubated at 37C in a shaking waterbath for 1 h. The cap contained a Whatman 3 mmfilter disc saturated with barium hydroxide to trap14CO2produced during the reaction. The reactionwas stopped by adding 0.5 mI of 5 N HCI andincubtaion continued for an additional 1 h to ensurecomplete entrapment of the CO2. The filter discswere recovered, dried and counted on a {3 scintilla-tion counter. A parallel reaction with heat-inacti-vated (100C for 10 min) cell lysate served as acontrol for background counts. A unit of enzymeactivity represented the amount of enzyme thatcatalyzes 1 nmol of 14CO2 released from orni-thine/h/mg protein.

ResultsCurcumin inhibits the proliferation ofdifferent breast tumor cell linesWe first investigated the effect of curcumin ongrowth of different breast tumor cell lines by thethymidine incoporation assay.The results of theseexperiments are shown Table I. Treatment of cellswith 1 ,ug/mI (2.7 ,uM) curcumin for 72 h inhibitedthe growth of all the seven breast tumor cell linestested with a maximum effect on BT-20 cells.Interestingly, the cell line (MCF- ADR) hat exhibits

Cell cycle analysisCurcumin-induced changes n the cell cycle of tumorcells were determined by using an Epics-Elite Lasterflow cytometer (Coulter, Hialeah, FL). Briefly, tryp-sin-detachedcells were washed in PBSand fixed inabsolute methanol. About 1 X 106 alcohol-fixed cellswere then washed in PBS and resuspended n 1 miof lysis-staining solution that contained propidiumiodide and a detergent (Coulter). The stained cellswere analyzed for relative DNA content. Chickenerythrocytes (Coutler) were used as a standard. Thenumber of cells in various phases of the cell cyclewas determined with the aid of the Multicycleprogram (phoenix Flow Systems,SanDiego, CA). Allexperiments were repeated at least twice. Most ofthe data shown are averagesof duplicate or triplicatedeterminations with below 10%variations.

Table 1. Antiproliferative effect of curcumin againstbreast tumor cell linesCell lines Relative cell viability(% of control)BT -20BT-2~NFSK-BR3MCF-7MCF-7ADRT -47 DZR-75-1

1869151326

Western blots Cells (5 X 103) were plated overnight and then incubated withcurcumin (1 Itg/ml). After 66 h at 3rc. the cells were pulsedwith [3H]thymidine for 6 hand then harvested. Thymidineincorporation was determined and normalized against untreatedcells (100%). All determinations were made in triplicate.

Following treatment with curcumin under appropri-ate conditions, cells were lysed in Tris-HCI buffer

472 AntI-Cancer Drugs. Vol8 .1997

:i:O:i:O:!::1:i: 1:i: 6:i:O:i: 1


4/12

an 80- to lOO-fold ncrease in resistance to adriamy-cin was also senstive o curcumin-induced inhibitionof cell growth. Similarly, a TNF-resistantsubclone ofBT-20 cells (BT-200rNF) emonstrated significantgrowth inhibition when cultured in the presence ofcurcumin.

Curcumin inhibits growth of hormone-independent and -dependent breasttumor cellsThe MCF-7 and T-47D cell lines are estrogen-depen-dent, whereas the SK-BR3 ine is estrogen-indepen-dent. The growth of the estrogen-dependent celllines was inhibited in a dose-dependentmanner withmaxinlum inhibition observed at less than a 1 I"g/ miconcentration of curcumin (Figure lA and B). Asimilar dose-dependent inhibition was observedwhen the estrogen-independent SK-BR3cells wereincubated in the presence of curcumin (Figure 1C).The effect of curcumin against hormone-depen-dent MCF- and hormone-independent BT-20 andMDA231 cell lines was time-dependent (Figure 2). A

Antiproliferative effectsof curcuminsignificant growth inhibitory effect of curcumin onthymidine incorporation by MCF- and DT-20 cellscould be observed as early as 20 h after treatment(Figure 2A and 2D). When examined for the growthrate, 1 fig/ mI of curcumin was sufficient to com-pletely inhibit the growth of MCF- cells for up to 7days (Figure 2C), whereas MDA231 cells required ahigh concentration of curcumin (3 fig/ mI) to showcomplete inhibition of cell growth (Figure 2D). Thedrug-induced inhibition in thymidine incorporationwas due to a net decrese in the number of viablecells remaining after curcumin treatment.When compared to untreated cells, MCF- cellstreated with curcumin for 24 h did not show anyobvious changes n their morphology. However, 48 htreatment with curcumin in general, and 72 h treat-ment in particular, induced some characteristicsmorphological changes in MCF- cells (Figure 3).The treated cells typically appeared enlarged, occa-sionally with two to four nuclei, and showed signifi-cant changes n nuclear morphology (Figure 3D).Next we determined the effect of curcumin onthe ability of MCF- cells to form anchorage-depen-dent colonies. Ten thousand cells were seeded per

120. 120 120B

1001

~ 81

~:0ta.>= 610)()

80 8

60 6

41 40

21 20

0 , , 0 00 0..., 00 0000000 10-1 10 101 0 10-1 10 101 0 10-1 10 101

Curcumin (!lg/ml) Curcumin (!lg/ml) Curcumin (!lg/ml)Figure 1. Oose-dependent growth inhibition of hormone-dependent MCF-7 (A) and T -470 (8), and of hormone-independent SK-8R3 (C) human breast adenocarcinoma cells. Cells (5 x 103 cells/0.1 mi/well) were plated overnightat 37C, washed and then incubated with different concentrations of curcumin for 72 h. Viable cells were counted asdescribed in Materials and methods. All determinations were made in triplicate. The results are expressed aspercentages of the control (untreated cells).

473nti-Cancer Drugs. Val8 .1997


5/12

K Mehta et al.

12000

10000

2000

00 20 40

Time (h)60 80

45

35

'.0E 25~~c"'_0

""Qj.-UXQ)~~ 15>

5

-50 2 4

Days in culture6 8

Figure 2. Time-dependent growth inhibition of human breast adenocarcinoma cells, MCF-7 (A and C), 8T -20 (8) andMDA-231 (D) by curcumin. Cells (5 X 103 cells/0.2 mi/well) were plated overnight at 37C, washed and then incubatedwith different concentrations of curcumin for indicated times. Viability of cells was examined either by thymidineincorporation (A and 8) or by counting viable cells (C and D). For thymidine incorporation, 0.5 flCi [3H]thymidine wasadded to the culture during the last 6 h of the incubation. The cells were washed, harvested and monitored forincorporation as described in Materials and methods. All determinations were made in triplicate. The results areexpressed as percentage of the control (untreated cells).

well in 6-well plates in medium alone or mediumcontaining increasing amounts of curcumin. Theplates were incubated for 7 days to allow individualcells to form colony forming units (c.f.u.). At theend of incubation period, the number of c.f.u. werescored after staining the plates with amino black.Results of this experiment are shown in Figure 4.The majority of cells grew into c.f.u. after 7 days of

culture in medium alone. However, the simultaneouspresence of curcumin during the incubation periodcaused a dose-dependentdecrease n the number ofc.f.u.. The minimum effective concentration thatresulted in a significant decrease in c.f.u. was0.13.ug/mI and 1 .ug/ml concentration of curcuminresulted in more than 90% nhibition in c.f.u. (Figure4). These results suggested hat the growth inhibi-474 AntI-Cancer Drugs. VolB .1997


6/12

Antiproliferative effectsof curcuminCurcumin

(~g/ ml)

0

!1 0.03

0.13

Figure 3. Effect of curcumin treatment on the morphol-ogy of MCF-7 cells. Cells cultured on glass cover slipswere left untreated (A) or incubated with 2.5 f1g/mlconcentration of curcumin (8) for 3 days. At the end ofthe incubation period, the cell monolayers were washedwith P8S, fixed in 100% methanol and stained with Diff-Quick stain (Dade Diagnostic, Aguada, Puerto Rico). o!t,//6

tory effect of curcumin against breast tumor cellswas cytotoxic rather than cytostatic. Figure 4. Dose-dependent inhibition of anchorage-de-pendent c.f.u.s of MCF-7 cells by curcumin. Cells wereplated overnight at 37C in 6-well plates (5 x 103/well/3 ml), washed and. then incubated with differentconcentrations of curcumin for 72 h. At the end ofincubation period, cell cultures were washed and stainedwith amido black.Optimum antiproliferative effects requirecontinuous presence of curcuminTo determine the minimum time required for curcu-mill to induce a growth inhibitory signal, we incu-bated MCF- cells with the drug for various lengthsof time. At the end of each incubation period, thecells were washed and reincubated in drug-freemedium for a total period of 72 h and cell growthwas determined by (3H]thymidine incorporation.The results shown in Figure S reveal that a minimumof 3 h preincubation with the drug was essential toproduce a small but significant growth inhibition ofMCF- cells. Twenty-four hour treatment with curcu-mill induced maximum inhibition, suggesting thatthe presence of curcumin for at least 24 h is neededfor optimum antitumor effect againstMCF- cells.

Curcumin blocks cell growth at the G2/Sphase of the cell cycleWe next determined if growth inhibitory effects ofcurcumin could be accounted for by effects onregulatory cell cycle check points. Asynchronouslygrowing MDA-231 cells were incubated in thepresence or absenceof increasing concentrations ofcurcumin. On different days of treatment, cellfractions were analyzed or the relative DNA contentby flow cytometry. The data showed that untreatedcells were evenly distributed in the G1 and S phasesof the cell cycle (Figure 6A, histograms A, G, M and



7/12

K Mehta et a1.

120

100

l~15(13.>""05(.).~~""05[I:

20

0 0 0.5 1 3 6 24Preincubation time (h)

48 72Figure 6. Growth arrest of curcumin-treated MDA-231cells in G2/S phase of cell cycle. Culture of MDA-231cells were treated with curcumin at 1 (histograms B, H, Nand T), 2 (histograms C, I, O and U), 3 (histograms D, J,P and V), 4 (histograms E, K, Q and W) or 5 f.tg/ml(histograms F, L, R and X) concentration for 1 (histo-grams A-F), 2 (histograms G-L), 3 (histograms M-S) or4 days (histograms S-X). Cells from curcumin-treated orcontrol cultures were analyzed by flow cytometry forrelative DNA content, as described in Materials andMethods. G1 = Go + G1 cells; G2 = G2 = M cells;AP = apoptotic cells.

Figure 5. Continuous presence of curcumin is requiredto inhibit the growth of MCF-7 cells. Cells (5 X 103 cells/0.1 mi/well) were plated overnight at 37C, washed andthen incubated with curcumin (1 f.lg/ml) for indicated timeperiods. At the end of each incubation, curcumin wasremoved by washing, the cells resuspended in curcumin-free medium and incubation continued for a total of 72 h.During the last 6 h of incubation, 0.5 f.lCi [3H]thymidinewas added to the culture, and the cells were washed,harvested and monitored for incorporation as indicated inMaterials and methods. All determinations were made intriplicate. The results are expressed as percentage of thecontrol (untreated cells). Table 2. Stage-specific growth arrest of MDA-231 cellsafter treatment with curcuminS). However, treatment with curcumin caused asignificant increase in G2 checkpoint arrest (Figure6, histograms EL, R and X), with more than 75%cells arrested n G2/S by 24 h after treatment (Figure6, histograms E and F; and Table 2). The extent ofG2 checkpoint arrest was concentration dependent;at 3 /-lg/mI or higher concentration, the curcumininduced rapid and pronounced growth arrest inMDA-231 cells (Figure 6, vertical histograms of anyday; and Table 2). Curcumin-induced cell growtharrest in G2 was reversible with time in culture(Figure 6, horizontal histograms for any concentra-tion). For example, the difference between curcu-min-treated and untreated cells became obvious at24 h after treatment (Figure 6, histograms A versushistograms C-E), but at later time points, cell cycleprofiles of the curcumin-treated cells were virtuallyindistinguishable from that of the untreated asyn-chronously growing cells (Figure 6, histogram Sversus histogram U-W), suggesting that the cells

Curcumin(ug/ml)

Treatmenttime(days)

Cell cycle stage (%)8G1 s G2 Apoptosis

None 1 (A)4 (8)1 (8)4 (T)1 (0)4 (V)1 (F)4 (X)

43493441253329

4838514348564525

861482535159

7,0

82826

3.05.0a Percentage of cells in various phases of cell cycle werecalculated from the histograms (shown in parentheses) of Figure5.

had successfully re-entered the cell cycle. At higherconcentrations (5 /lg/ml), curcumin appeared tocause the cells to arrest in late S phase of the cellcycle (Figure 6, histograms 1':L, R and X).

476 Anti-Cancer Drugs. VolB .1997

80

60

40


8/12

Curcumin-induced growth arrest isassociated with ODC inhibition

Antiproliferative effectsof curcuminsignificant decrease in ODC activity was observed2 h after treatment, but by 24 h after treatment thecells had fully recovered from this inhibitory effectand the level of the enzyme activity was comparableto that observed in untreated cells (Figure 7A). At ahigher concentration of 10,ug/ml, curcumin treat-ment caused a rapid and nearly complete inhibitionof ODC activity that persisted even 24 h after thetreatment (Figure 7B).

Treatment of MCF-7 cells with curcumin caused adose- and time-dependent inhibition in ODC activity(Figure 7). At a low concentration of 2.5 Jlg/m1, a

1000

750 Curcumin can overcome resistance toadriamycin in MCF-7 cells

500

250~.>.'8 cm .Q)0>-(/) eroc.-s:,0! 0E:c'NO 1000()OE.3:

242 5

We extended our initial observation regarding thegrowth inhibitory effect of curcumin to the adriamy-cin-resistant breast tumor cell line MCF- ADR Table1). This drug-resistant subclone of MCF- cells wasselected by continuous culture in the presence ofincreasing concentrations of adriamycin over aperiod of time. The establishment and characteristicsof this cell line have been described elsewhere.24MCF- ADR ells exhibit a 100- to 120-fold ncrease nresistance to adriamycin and express high levels ofp-glycoprotein and TGase.24Unlike their differentialsensitivity towards adriamycin, both the resistantand sensitive MCF- cells showed almost equalsusceptibility to curcumin-induced growth inhibition(Figure 8).To examine further the mechanism by whichcurcumin inhibits growth and to determine ifdrug-resistance in MCF- cells reflects theirgeneral resistance to apoptosis, we evaluated thelevels of several proteins whose expression ismodulated during apoptosis. Some striking differ-ences in the levels of Bcl-2, p53 and TGaseexpression became apparent in drug-sensitiveand -resistantcells (Figure 9). For instance, Bcl-2 wasover produced in sensitive cells as compared toresistant cells, whereas p53, and TGase wereproduced only in the adriamycin-resistant cells.Treatment with curcumin of either adriamycin-sensitive or -resistant cells failed to alter the produc-tion of any of these proteins, suggesting thatcurcumin-mediated growth inhibition of resistantcells may operate by a mechanism other thanapoptosis. Consistent with these results are thoseobtained by flow cytometry data (Table 2 and Figure6), morphological alterations (Figure 3) and DNAladdering data (not shown) which showed no signifi-cant increase in the proportion of cells undergoingapoptosis even after prolonged treatment withcurcumin.

~.0CauQ)"0Q)c:.2a 750

500

250

2Time(h)

5 24

Figure 7. Effect of curcumin treatment on OOC activityin MCF-7 cells. Cells (8 x 106) were incubated inmedium alone (hollow bars) or medium containing either2.5 (A) or 10 f1g/ml (8) curcumin (filled bars). Atindicated time intervals, the cells were washed andprocessed for assaying the OOC activity as described inMaterials and methods. The results shown are theaverages from triplicate values :J::SO from the mean.

477nti-Cancer Drugs. Vol8 .1997


9/12

MW (kDa),cl226

P5353

TGase85

62

13-Actin-42,

Curcumin concentration (Jlg/ml)- -

11 2 3 1MCF-7WTigure 8. Effect of curcumin treatment on cell growth ofdrug-sensitive and adriamycin-resistant MCF-7 cells.Cells (5 X 103/well/0.2 ml) were grown overnight in 96-well plates and then treated with increasing concentra-tions of curcumin for 3 days. The number of survivingcells at the end of incubation was determined by MTT

assay as described in Materials and methods.

6 7

Figure 9. Effect of curcumin treatment on apoptosis-related proteins in MCF-cells. Drug-sensitive (lanes 1-4)and adriamycin-resistant (lanes 5-8) MCF- 7 cells wereincubated in medium alone (lanes 1 and 5) or mediumcontaining 2.5 ,ug/ml curcumin (lanes 2, 3, 4, 6, 7 and 8)for 24 (lanes 2 and 6), 48 (lanes 3 and 7) and 72 h(lanes 4 and 8). At the end of incubation period, cellswere washed, lysed and subjected (50 ,ug protein) toWestern blotting by using specific antibodies as primaryantibody and horseradish peroxidase-conjugated anti-mouse IgG as secondary antibody. The antigen-anti-body reaction was detected by ECL as described inMaterials and methods.

DiscussionThe search for new chemopreventive and antitumoragents that are more effective and less toxic haskindled great interest in phytochemicals. Curcuminwhich is derived from the root of a plant, c. tonga,is one such compound. It has been used as a dietaryfactor and as a herbal medicine for centuries inseveral Asian countries. In the present report, wedescribe the potent antiproliferative effects of cur-cumin against a wide variety of breast tumor celllines. The antiproliferative effects were observedagainst hormone-independent and -dependent andadriamycin-sensitive and -resistant breast tumorcells. The antiproliferative effects of curcuminwere observed at concentrations of 2-15 .uM whichhave been shown to inhibit PKC activity in vitro,EBV-DRpromoter, SV 40 promoter enhancer, AP-1activation, NF-KB activation and HIV-LTR-mediated..16,17,28,29 N 1 II 1 . 1ranscnptlon. orma ce s were re atlve yresistant to the toxic effects of curcumin (data notshown), suggesting that the growth inhibitory ef-

fects of this drug against tumor cell lines werespecific.How different antiproliferative agents suppresscellular growth is not well understood. Certaincytokines such as interferon-a, interleukin-6 andtransforming growth factor-{3 nhibit cellular prolif-eration by inducing tumor suppressor genes such asretinoblastoma (Rb) or by suppressing the expres-sion of genes nvolved in cellular proliferation or bymodulating the phosphorylation of the geneproduct.30.31 Since phorbol ester-mediated tumorpromotion is inhibited by curcumin and this wasshown to be due to the inhibitory effects ofcurcumin on protein kinase C,18 t is possible thatthe growth-inhibitory effects of curcumin are also

478 Anti-Cancer Drugs. VolB .199;


10/12

Antiproliferative effectsof curcuminthan those treated at higher concentrations (Figure6, horizontal groups of histograms). ODC plays apivotal role in cell growth and proliferation sincedisruption of its functions by agents such as DFMOresults in growth arrest.35,36 verexpression of ODChas also been linked to cell transformation andcarcinogenesis. Most transformed cell lines andtumors exhibit high levels of basal ODC expression,and its induction has been suggested o be a criticalevent for tumor promotion in a variety of experi-mental models including skin, breast and coloncarcinogenesis.37 rom these studies, it is temptingto speculate that anticarcinogenic properties ofcurcumin may be attributed to its ability to inhibitODC activity.Another interesting feature of this study was theobservation that curcumin could circumvent thedrug-resistant phenotype in breast tumor cells.Curcumin was as effective in arresting the growth ofmultidrug-resistant and TNF-resistant cells as it waswild-type MCF- cells (Table 1 and Figure 8). Thesestudies suggested hat curcumin does not serve as asubstrate for P-glycoprotein and that it uses uniquetargets in resistant cells. Previous studies havedemonstrated elevated levels of PKC activity inmulti-drug resistant MCF- cells.38 The agents thatalter PKC activity have been shown to modulate thedrug-resistantphenotype.39 t is likely that curcuminmediates its effects against drug-resistant cells byvirtue of its ability to inhibit PKC. Furthermore, thegrowth inhibitory effects of certain cytokines andchemotherapeutic agents correlate with their abilityto modulate intracellular glutathione levels.40,41heinhibitory effect of curcumin against PKC activity issuppressed by thiol compounds such as mercap-toethanol, cysteine and dithiothreitol.18 Whether thesensitivity of tumor cells and resistance of normalcells to curcumin is dependent on cellular gluta-thione levels s not clear.Previously we reported that curcumin is a potentinhibitor of NF-KB activation induced by a widevariety of agents including TNF, phorbol ester andhydrogen peroxide.17 Therefore, it is equally possi-ble that the target for curcumin action in trans-formed cells could lie upstream in the signaltransduction pathway involving NF-KB,which servesas a central coordinating regulator in signal-respon-sive transduction of several genes, ncluding c-mycand p53, which in turn play important roles in cellgrowth, differentiation, activation and chemotaxis.42The absence or inhibition of NF-KB n normal ortransformed cells has recently been shown to renderthem more sensitive to chemotherapy, radiation andTNF-induced apoptosis.43-47 hus, it is conceivable

mediated through its ability to inhibit specific pro-tein kinases. Our studies clearly suggest hat curcu-mill treatment results in growth arrest of breasttumor cells in the G2 phase of the cell cycle (Figure6 and Table 2). In higher eukaryotes, several cyclin-dependent kinases (Cdks) playa crucial role duringcell proliferation. For example, at G2/M phasetransition, mitosis is initiated by a Cdk-cyclin com-plex that is composed of a Cdc2 protein kinase anda B-type cyclin. Cdc2 kinase phosphorylates a varietyof cellular proteins, including the tumor suppressorgene products Rb and p53 that regulate criticalevents during cell growth.32.33 It is likely thatcurcumin affects the growth of tumor cells byinhibiting the activity of Cdc2 or some other kinasesthat have a role in initiating mitosis. We havepreviously reported that curcumin can inhibit bothprotein tyrosine kinases and serine/threonine pro-tein kinases in vitro.21 Other investigators demon-strated that curcumin could also inhibit in vivo theintrinsic kinase activity of the epidermal growthfactor (EGF) receptor,22 which may result in anti-growth action of curcumin. In addition, curcuminwas also shown to inhibit EGF-inducedactivation ofEGF eceptor phosphorylation.23In general, G2 arrest of proliferating cells is causedby DNA-damagingagents and represents a mechan-ism of negative eedback control for the induction ofgene products that facilitate the repair of DNAlesions. Therefore, G2/M phase arrest is thought toensure that DNA replication will proceed withfidelity and to avoid segregation of defectivechromosomes.34 f DNA repair is successful duringG2 phase, the cells re-enter the cell cycle; otherwise,they are eliminated via apoptosis. Our studiessuggested that curcumin treatment causes G2/Marrest of the human breast tumor cells without anyevidence of apoptosis (Table 2). Indeed, the tumorcells rescued the growth inhibitory effect with timein culture (Figure 6). The recovery from growtharrest depended on the concentration of curcuminused. Moreover, curcumin treatment did not alterthe expression of the apoptosis-relatedproteins Bcl-2, p53 and TGase Figure 9).ODC, a key regulatory enzyme in polyaminebiosynthesis, was strongly inhibited by curcumin(Figure 7). At lower concentrations, curcumin tran-siently inhibited ODC activity in MCF- cells. How-ever, at higher concentrations the inhibition wasacute and persisted for at least 24 h after treatment.The ability of curcumill to inhibit ODC activitycorrelated well with its growth inhibitory activity.Cells treated at lower concentrations of curcuminwere able to escape more rapidly from G2 arrest



11/12

K Mebta et al.that the anti-proliferative effect of this compoundmay be linked to this pathway. In any case, the lowtoxicity at pharmacological doses and potent anti-proliferative activity against several breast tumor celllines exhibiting various phenotypes lead us toconclude that the therapeutic potential of curcuminin breast cancer warrants further investigation.

AcknowledgmentsWe would like to thank Dr Sadhana Gupta forassistance at the early phase of these studies, DrMadan Chaturvedi and Raj Pandita with the ODCassays,Klara Totpal with the clonogenic assays,andMr Walter Pagel or editorial assistance.

References

damage: protection by tUrnIeric (Curcuma longa).Mol Cell Bio11987; 77: 3-10.13. Unnikrishnan MK, Rao MNA. Curcumin inhibits nitrite-induced methemoglobin formation. FEBS Lett 1992;301: 195-6.14. Donatus lA, Vermeulen S, Vermeulen NPE. Cytotoxicand cytoprotective activities of curcumin. BiocbemPbarmacol1990; 39: 1869- 75.15. Shalini VK, Srinival L. Fuel smoke condensate nducedDNA damage n human lymphocytes and protectionby tumeric (Curcuma longa). Mol Cell Bio11990; 95:21-30.16. Huang T-S,Lee s-c, LiD ]-K. Suppressionof c-]un/AP-1activation by an inhibitor of tumor promotion inmouse fibroblast cells. Proc Natl Acad Sci USA 1991;88: 5292-6.17. Singh S, Aggarwal BB. Activation and transcriptionfactor NF-KB s suppressedby curcumin. J Bioi Cbem1995; 270: 24995-5000.18. Liu ]-Y; ]u S, LiD-]-K. nhibitory effects of curcumin onprotein kinase C activity induced by 12-0-tetradeca-noyl-phorbol-13 acetate n NIH 3T3 cells. Carcinogen-esis 1993; 14: 857-61.19. LiD ]K, Shih CA. Inhibitory effect of curcumin onxanthine dehydrogenase/oxidase nduced by 12-0-tetradecanoyl-phorbol-13-acetaten NIH3T3 cells. Car-cinogenesis 1994; 15: 1717-21.20. Rao CV; Simi B, Reddy BS. Inhibition by dietarycurcumin of azoxymethane-induced ornithine decar-boxylase, yrosine protein kinase, and arachidonic acidmetabolism and aberrant crypt foci formation in ratcolon. Carcinogenesis 1993; 14: 2219-25.21. Reddy S, Aggarwal BB. Curcumin is a non-competitiveand selective inhibitor of phosphorylase kinase. FEBSLett 1994; 341: 19-22.22. Korutla L, Kumar R. Inhibitory effect of curcumin onepidermal growth factor receptor kinase activity inA431 cells. Biocbem Biopbys Acta 1994; 1224: 597-600.23. Korutla L, Cheung JY; Mendelsohn , Kumar R. Inhibi-tion of ligand-induced activation of epidermal growthfactor receptor tyrosine phosphorylation by curcumin.Carcinogenesis 1995; 16: 1741-5.24. Mehta K. High levels of transglutaminase expressionin doxorubicin-resistant human breast carcinoma cells.IntJCancer 1994; 58: 400-6.25. SugarmanB], Aggarwal BB, HassPE, Figari IS, PalladinoMA, ShepardHM. Recombinant human tumor necrosisfactor: Effects on proliferation of normal and trans-formed cells in vitro. Science 1985; 230: 943-5.26. Vilcek ], Palombella VJ, Henriksen-DeStefanoD, et al.Fibroblasts growth enhancing activity of tumor necro-sis factor and its relationship to other polypeptidegrowth factors.J Exp Med 1986; 163: 632-43.27. Grewal HS, Sloan D, Sampliner RE, Fennerty B.Ornithine decarboxylase assay in human colorectalmucosa. Methodologic issues of importance to qualitycontrol. Int J Cancer 1992; 52: 355-8.28. Bouvier G, Hergenhahn M, Polack A, Bornkamm GW;Bartsch H. Validation of two test systems or detectingtumor promoting EBV inducers: chemopreventiveresponses of several agents in DR-aT Raji cells andin human granulocytes. Carcinogenesis 1993; 14:1573-8.

1. Huang M-T, Wang ZY; Georgiadis CA, Laskin ]D,Conney AH. Inhibitory effects of curcumin on tumorinitiation by benzo[a]pyrene and 7,12-dimethylbenz-[a]anthracene. Carcinogenesis 1992; 13: 2183z-6.2. Azuine MA, Bhide SV: Chemopreventive effect oftumeric against stomach and skin tumors induced bychemical carcinogens in swiss mice. Nutr Cancer1992; 17: 77-83.3. Azuine MA, Bhide SV: Protective single/combinedtreatment with betel leaf and turmeric against methyl(acetoxymethyl) nitrosamine-induced hamster oralcarcinogenesis. nt] Cancer 1992; 51: 412-5.4. NagabhushanM, Bhide SV:Curcumin as an inhibitor ofcancer. Am CoIl Nutr 1992; 11: 192-8.5. Huang M-T, Smart RC, Wong C-Q, Conney AH. Inhibi-tory effect of curcumin, chlorogenic acid, caffeic acid,and ferulic acid on tumor promotion in mouse skin by12-O-tetradecanoylphorbol-13-acetate. Cancer Res1988; 48: 5941-6.6. Conney AH, Lysz T, Ferraro T, et at. Inhibitory effectof curcumin and some related dietary compounds ontumor promotion and arachidonic acid metabolism inmouse skin. Adv Enzyme Regull991; 31: 385-96.7. Ammon HP, Wahl MA. Pharmacology of Curcumatonga. PlantMed 1991; 57: 1-7.8. Satoskar RR, Shah S], Shenoy SG. Evaluation of anti-inflammatory property of curcumin (diferuloylmethane) n patients with postoperative inflammation.Int] Clin Pbarmacoll986; 24: 651-4.9. Shankar TNB, Shantha Nv; Ramsh Hp, Murthy IAS,Murthy VS. Toxicity studies on turmeric (Curcumatonga): acute toxicity studies in rats, guinea pigs andmonkeys. Indian] Exp BioI 1980; 18: 73-5.10. Srivastava R. Inhibition of neutrophil response bycurcumin. Agents Action 1989; 38: 298-303.11. Huang H-C, ]an T-R, Yeh S-E Inhibitory effect ofcurcumin, an anti-inflammatory agent, on vascularsmooth muscle cell proliferation. Eur ] Pbarmacol1992; 221: 381-4.12. Shalini VK, Srinivas L. Lipid peroxide induced DNA

480 Anti-Cancer Drugs. VolB .1997


12/12

Antiproliferative effects of curcuminlevel of nuclear protein kinase C in multidrug-resistantMCF- human breast carcinoma cells. Cancer Res1996; 52: 3750-3759.39. Fine RL, Patel J, Chabner BA. Phorbol ester inducemultidrug resistance in human breast cancer cells.Proc Natl Acad Sci USA 1988; 85: 582-6.40. A1i-Osman Stein DE, Renwick A. Glutathione contentand glutathione-S-transferase xpression in 1,3-bis(2-chloroethyl)-l-nitrosourea-resistant human malignantastrocytoma cell lines. Cancer Res 1990; 50: 6976-80.41. 1shii Y; Partridge CA, Del Vecchio PJ,Malik AB. Thmornecrosis factor-a-mediated decrease n glutathione in-creases he sensitivity of pulmonary vascular endothe-lial cells to H2O2. Clin Invest 1992; 89: 794-802.42. Siebenlist U, Franzoso G, Brown K. Structure, regula-tion and function of NF-KB.Annu Rev Cell Bio11994;10: 405-55.43. Wu M, Lee H, Bellas RE, et al. Inhibition of NF-KB/Relinduces apoptosis of murine B cells. EMBO J 1996;15: 4682-90.44. Begg AA, Baltimore D. An essential role for NF-KB npreventing TNF-a-induced cell death. Science 1996;247: 782-4.

45. Wang C-Y;Mayo MW; Baldwin AS, Jr. TNF and ~ancertherapy-induced apoptosis: potentiation by inhibitionofNF-KB. Science1996; 247: 784-7.46. Van Antwerp DJ, Martin SJ,Kafri T, Green DR, Verma1M. Suppressionof TNF-a-inducedapoptosis by NF-KB.Science1996; 247: 787-9.47. Liu Z-G, Hsu H, Goeddel D": Karin M. Dissection ofTNF receptor 1 effector functions: JNK activation isnot linked to apoptosis while NF-KB activation pre-vents cell death. Cell 1996; 87: 565- 76.

(Received 6 March 1997; accepted 20 March 1997)

29. ti CJ, Zh31lg LJ, Dezube BJ, Crumpacker CS, PardeeAB. Three inhibitors of type 1 human immunodefi-ciency virus long tenninal repeat-directedgene expres-sion and virus replication. Proc Natl AcadSci USA1993; 90: 1839-42.30. Kumar R, Atlas I. Interferon-a the expression ofretinoblastoma gene product in human Burkitt'slymphoma Daudi cells: role in growth regulation. ProcNatl Acad Sci USA 1992; 89: 6599-603.31. Massague , Cheifetz S, Lahio M, Ralph DA, Weis FMB,Zentalla A. Transfonning growth factor-fJ. In: LevineAJ, ed. Tumor suppressor genes, the cell cycle andcancer. Cold Spring Harbor, NY: Cold Spring HarborLaboratory Press 1992; 81-103.32. Sturzbecher ~ Mainets T, Chumakov P, et al. P53interacts with P34cdc2n mammalian cells: implicationfor cell cycle control and oncogenesis. Oncogene1990; 5: 795-801.33. tin BT, Gruenwald SD, Marla A, Lee W-H, Wang JYRetinoblastoma cancer suppressor gene product is asubstrate of the cell cycle regulator cdc kinase. EMBOJ 1991; 10: 857-64.34. Kahn Kw; JackmanJ, O'Conner PM. Cell cycle controland chemotherapy.J Cell Biochem 1994; 54: 440-52.35. Heby 0, Persson L. Molecular genetics of polyaminesynthesis in eukaryotic cells. Trends Biochem Sci1990; 172: 153-8.36. HayashiS-I. Ornithine decarboxylase; biology, enzym-ology, and molecular genetics. New York: PergamonPress 1989.37. Verma AK, Boutwell RK. Inhibition of carcinogenesisby inhibitors of putrescine biosynthesis. In: McCannPp, Pegg AE, Sjoerdsma A, eds. Inhibition of poly-amine metabolism; biological significance and basisfor new therapies. New York: Academic Press 1987249-275.38. Lee SA, Karaszkiewicz ]w; Anderson WB. Elevated


Documents

Antiproliferative_cancro Da Mama