This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Original Contribution

Gemcitabine response in pancreatic adenocarcinoma cells is synergistically enhancedby dithiocarbamate derivatives

Elisa Dalla Pozza a,1, Massimo Donadelli a,⁎,1, Chiara Costanzo a, Tatyana Zaniboni a, Ilaria Dando a,Marta Franchini b, Silvia Arpicco c, Aldo Scarpa b, Marta Palmieri a

a Department of Life and Reproduction Sciences, Biochemistry Section, University of Verona, 37134 Verona, Italyb Department of Pathology and Diagnostics, University of Verona, 37134 Verona, Italyc Department of Science and Drug Technology, University of Torino, Torino, Italy

a b s t r a c ta r t i c l e i n f o

Article history:Received 27 July 2010Revised 17 December 2010Accepted 3 January 2011Available online 12 January 2011

Keywords:Pancreatic adenocarcinomaGemcitabineZincDithiocarbamateOxidative stressApoptosisFree radicals

Pancreatic adenocarcinoma is a common malignancy that remains refractory to all available therapies,including the gold standard drug gemcitabine (GEM). We investigated the effect of the combination of GEMand each of the ionophore compounds pyrrolidine dithiocarbamate (PDTC) and disulfiram [DSF;1-(diethylthiocarbamoyldisulfanyl)-N,N-diethylmethanethioamide] on p53−/− pancreatic adenocarcinomacell growth. PDTC or DSF synergistically inhibited cell proliferation when used in combination with GEM byinducing apoptotic cell death. This effect was associated with an increased mitochondrial O2

•− production andwas further enhanced by zinc ions. Basal levels of mitochondrial O2

•− or manganese superoxide dismutase(MnSOD) strictly correlated with the IC50 for GEM or the percentage of synergism. Thus, the most relevantvalues of the antiproliferative synergism were obtained in GEM-resistant pancreatic adenocarcinoma celllines. Interestingly, the GEM-sensitive T3M4 cells transfected with MnSOD expression vector showedmitochondrial O2

•− and IC50 for GEM similar to those of resistant cell lines. In vivo experiments performed onnude mice xenotransplanted with the GEM-resistant PaCa44 cell line showed that only the combinedtreatment with GEM and DSF/Zn completely inhibited the growth of the tumoral masses. These results andthe consideration that DSF is already used in clinics strongly support the GEM and DSF/Zn combination as anew approach to overcoming pancreatic cancer resistance to standard chemotherapy.

© 2011 Elsevier Inc. All rights reserved.

Pancreatic adenocarcinoma is one of the most aggressive humancancers, with a long-term overall survival (OS) lower than 5% [1,2].Standard treatments for advanced disease include monotherapy withgemcitabine (GEM), which has, however, a response rate of less than20% [3]. Many clinical trials have failed to demonstrate an improve-ment in OS with the addition of various drugs to GEM. Nevertheless,some modest but interesting advances have been provided by drugcombination therapies such as GEM/erlotinib, GEM/capecitabine, andGEM plus a platinum salt [4]. Nowadays, research is focused on theidentification of other potential targets of response, the regulation ofwhich may improve GEM antitumoral activity.

Reactive oxygen species (ROS) have recently emerged as prom-ising targets for anticancer drug discovery. Indeed, constitutivelyelevated levels of ROS represent a specific vulnerability of malignantcells that can be selectively targeted by pro-oxidant chemotherapeu-tics [5]. We have recently demonstrated that the induction of ROS isone of the mechanisms of GEM antitumoral action and that pancreaticadenocarcinoma cell lines with lower basal levels of ROS are more

resistant to GEM compared to cells with higher ROS levels [6]. Wehave also shown that an increase of the level of zinc transported intothe cell by the ionophore compound pyrrolidine dithiocarbamate(PDTC) strongly enhances ROS production and pancreatic adenocar-cinoma cell growth inhibition by apoptosis-inducing factor-mediatedapoptosis [7,8].

Disulfiram [DSF; 1-(diethylthiocarbamoyldisulfanyl)-N,N-diethyl-methanethioamide] is a drug with molecular structure and zincionophore activity [9] similar to those of PDTC. It has been used forover half a century for alcohol aversion therapy and has an excellentsafety record [10,11].

The main purpose of this study was to investigate whether DSF isable to increase GEM sensitivity of pancreatic adenocarcinoma celllines in vitro and in vivo.

In vitro results show that both GEM/PDTC and GEM/DSFtreatments give rise to strong synergistic cell growth inhibition,which is associated with apoptotic cell death and mitochondrial O2

•−

production. Interestingly, synergistic growth inhibition is significantlyhigher in GEM-resistant compared to GEM-sensitive cell lines. Thus, infour cell lines examined, GEM IC50 or the percentage of synergismcorrelates directly with manganese superoxide dismutase (MnSOD)expression and inversely with mitochondrial O2

•− production. In vivo

Free Radical Biology & Medicine 50 (2011) 926–933

⁎ Corresponding author. Fax: +39 045 8027170.E-mail address: massimo.donadelli@univr.it (M. Donadelli).

1 These authors contributed equally to this work.

0891-5849/$ – see front matter © 2011 Elsevier Inc. All rights reserved.doi:10.1016/j.freeradbiomed.2011.01.001

Contents lists available at ScienceDirect

Free Radical Biology & Medicine

j ourna l homepage: www.e lsev ie r.com/ locate / f reeradb iomed

Author's personal copy

experiments on nudemice xenotransplanted with the PaCa44 cell lineshow that the combined treatment with GEM and DSF/Zn has astronger inhibitory effect on the tumoral growth compared to singletreatments.

Materials and methods

Chemicals

PDTC and DSF (Antabuse) were obtained from Sigma–Aldrich Co.,solubilized in sterile water and ethanol, respectively, and stored at−80 °C until use. GEM (Gemzar; Lilly) and zinc sulfate (Sigma) wereprepared in sterile water.

Cell culture

Four human pancreatic adenocarcinoma cell lines were studied:PaCa44, T3M4, Panc1, and MiaPaCa2 (see [12] for genetic character-ization and primary tissue source). All cell lines were grown in RPMI1640 supplemented with 2 mM glutamine, 10% fetal bovine serum,and 50 μg/ml gentamicin sulfate (BioWhittaker) and were incubatedat 37 °C with 5% CO2.

Cell proliferation assay

Cells were plated in 96-well plates (4×103 cells/well) and 24 hlater treated with the various compounds and further incubated forthe indicated times (see figure legends). At the end of the treatmentsthe cells were stained with a crystal violet solution (Sigma). The dyewas solubilized in PBS containing 1% SDS and measured photomet-rically (A595 nm) to determine cell growth.

Drug combination studies

The combination index (CI) was calculated using the Chou–Talalayequation, which takes into account both the potency (IC50) and theshape of the dose–effect curve [13,14], taking advantage of the CalcuSynsoftware (Biosoft, Cambridge, UK). The general equation for the classicisobologram is given by CI=(D)1/(Dx)1+(D)2/(Dx)2+[(D)1×(D)2]/[(Dx)1×(Dx)2], where (Dx)1 and (Dx)2 in the denominators are thedoses (or concentrations) of drug1 anddrug2 alone that give x% growthinhibition, whereas (D)1 and (D)2 in the numerators are the doses ofdrug 1 and drug 2 in combination that also inhibited x% cell growthinhibition (i.e., isoeffective). CIb1 and CIb0.3 indicated synergism andstrong synergism, and CI=1 and CIN1 indicated additive andantagonism, respectively. CI/effect curves represent the CI versus thefraction (0→1) of cells killed by drug combination. The synergismpercentagewasobtainedbyanalyzing theCI/effect curve andmeasuringthe CI values at each 0.05 fraction, i.e., 5% growth inhibition, of theantiproliferative effect.

Drug combination studieswere performed using the concentrationratios [GEM]:[DSF]=3.6:1 and [GEM]:[PDTC]=1.8:1, which took intoaccount the fact that a single zinc ion binds one molecule of DSF ortwo molecules of PDTC. The ratio of 3.6:1 for [GEM]:[DSF] was chosenfollowing U.S. Food and Drug Administration (FDA) directives thatestablish, for humans, doses of 1800 mg (25 mg/kg) and 500 mg forGEM and DSF, respectively. Taking into account the drug molar ratios,the in vitro ranges of concentration used for each compound were 10nM→100 μM for GEM, 5.56 nM→55.6 μM for PDTC, and 2.78nM→27.8 μM for DSF.

Apoptosis

Cells were plated in 96-well plates (4×103 cells/well) and, the dayafter, were treated with the various compounds at the indicatedconcentrations for 16 h. At the end of the treatment, the cells were

fixed with 2% paraformaldehyde in PBS at room temperature for30 min. The cells were washed twice with PBS and stained withannexin V/FITC (Bender MedSystem) in binding buffer (10 mMHepes/NaOH, pH 7.4, 140 mM NaOH, and 2.5 mM CaCl2) for 10 minat room temperature in the dark. Finally, the cells were washed withbinding buffer solution and fluorescence was measured by using amultimode plate reader (ex 485 nm and em 535 nm). The values werenormalized on cell proliferation by crystal violet assay.

Measurement of mitochondrial superoxide ion production

The nonfluorescent MitoSOX red probe (Molecular Probes) wasused to evaluate mitochondrial O2

•− production. The probe is live-cellpermeative and is rapidly and selectively targeted to themitochondriawhere it becomes fluorescent after oxidation by O2

•−, but not by otherROS or reactive nitrogen species. Briefly, cells were plated in 96-wellplates (4×103 cells/well) and, the day after, were treated with thevarious compounds at the indicated concentrations for 16 h. At theend of the various treatments, the cells were incubated in culturemedium with 0.5 μM MitoSOX at 37 °C for 15 min. The cells werewashed with Hanks’ buffer (20 mM Hepes, pH 7.2, 10 mM glucose,118 mM NaCl, 4.6 mM KCl, and 1 mM CaCl2) and fluorescence wasmeasured by using a multimode plate reader (ex 535 nm and em590 nm). The values were normalized on cell proliferation by crystalviolet assay.

Immunoblot analysis

Cells were collected, washed in phosphate-buffered saline, andresuspended in 50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 1× Triton X-100, 1 mM phenylmethylsulfonyl fluoride, and 1× protease inhibitorcocktail (Roche). After three freeze/thaw cycles and incubation on icefor 15 min, the lysate was centrifuged at 14,000g for 10 min at 4 °Cand the supernatant was used for Western blotting. Proteinconcentration was measured with the Bradford protein assay reagent(Pierce) using bovine serum albumin as a standard. Ten microgramsof protein extracts was electrophoresed through a 12% SDS–polyacrylamide gel and electroblotted onto polyvinylidene difluoridemembranes (Millipore). Membranes were then incubated overnightat 4 °C with blocking solution (5% low-fat milk in 100 mM Tris, pH 7.5,0.9% NaCl, 0.1% Tween 20) and probed for 1 h at room temperaturewith the mouse monoclonal anti-MnSOD antibody (1:2000 inblocking solution; Abcam). Horseradish peroxidase-conjugated anti-mouse IgG (1:1000 in blocking solution; Upstate Biotechnology) wasused to detect specific proteins. Immunodetection was carried outusing chemiluminescent substrates (Amersham Pharmacia Biotech)and recorded using HyperfilmECL (Amersham Pharmacia Biotech).The bands were scanned as digital peaks and the areas of the peakswere calculated in arbitrary units using the public domain NIH Imagesoftware (http://rsb.info.nih.gov/nih-image/). The value of Ponceau Sdye was used as a normalizing factor.

Transfection experiments

Exponentially growing T3M4 cells were seeded at a density of4×103 cells/well in 96-well plates for proliferation assays and at2.5×105 cells/plate in 60-mm cell culture plates for proteinextraction. Twenty-four hours later, transfections were carried outwith a pCR3.1 expression vector containing the human MnSOD cDNAusing TransIT-LT1 transfection reagent (Mirus) according to themanufacturer's directions. Cells were incubated for 24 h and thentreated with GEM at 10 μM for 16 h, to evaluate the effect of GEMtreatment on mitochondrial superoxide production, and with in-creasing doses of the drug for 48 h to identify the IC50 value. After48 h, transfected cells were analyzed for basal levels of mitochondrialsuperoxide ions and of MnSOD production. Transfection efficiency

927E. Dalla Pozza et al. / Free Radical Biology & Medicine 50 (2011) 926–933

Author's personal copy

was assessed by cytofluorimetric analysis and ranged between 85 and90%. T3M4 cells transfected with the empty pCR3.1 vector were usedas a control and behaved as the untransfected cells (data not shown).The expression vector for the human MnSOD was kindly provided byDr. Akashi (National Institute of Radiological Sciences, Chiba, Japan).

In vivo studies

PaCa44 cells (2×106 cells/mouse) were subcutaneously (sc)injected into female nude mice (4 weeks of age; Harlan Laboratories).One week after cell inoculation, five randomized animals for eachexperimental group received vehicle solution (PBS) or 25 mg/kg GEMand/or 6.9 mg/kg DSF in the presence of 3.75 mg/kg ZnSO4 byintraperitoneal injection biweekly for 4 weeks. Drug doses werechosen on the basis of U.S. FDA directives. As for the in vitro studies,the drug concentration ratio was [GEM]:[DSF]=3.6:1. DSF wasadministered in the presence of ZnSO4 using a 1:1 molar ratio.Tumor volume and body mass were recorded biweekly for eachanimal. Animals were sacrificed at the end of the 4-week study periodand the tumors were resected and weighed. The animal studies wereapproved by the Verona University Review Board.

Statistical analysis

ANOVA (post hoc Bonferroni) and graphical presentations wereperformed by GraphPad Prism 5. P values of b0.05, b0.01, or b0.001are indicated on the figures. Throughout all the experiments shown inFig. 5D and Supplementary Figs. 2 and 3 we obtained a mean of thelinear correlation coefficient corresponding to r=0.91±0.05.

Results

The combined treatments GEM/PDTC and GEM/DSF synergisticallyinhibit pancreatic adenocarcinoma cell proliferation in vitro

The antiproliferative effects of PDTC or DSF in combination withGEM, both in the absence or in the presence of 30 μM ZnSO4, wereexamined on four pancreatic adenocarcinoma cell lines characterizedby different sensitivities to GEM (T3M4NMiaPaCa2NPaCa44NPanc1),as reported in our previous papers [6,15] (Figs. 1A and B and data notshown). In all cell lines, growth inhibition by GEM was significantlyincreased by PDTC or DSF and further enhanced by ZnSO4. GEMsensitivity was not altered by the presence of zinc ions, according to

our previous observation that only the presence of ionophorecompounds can efficiently enhance intracellular zinc concentration[8]. It is worth noting that the combined treatment GEM/dithiocar-bamate/ZnSO4 was able to almost totally inhibit cell growth.

Dose-dependent analyses of cell growth inhibition were per-formed at 48 h with various concentrations of GEM and/or PDTC orDSF, in the absence or presence of a fixed dose of zinc ions. Figs. 2Aand B report the isobolograms of the CI values versus the fraction(0→1) of PaCa44 cells killed by drug combinations (results on T3M4,MiaPaCa2, and Panc1 cell lines were similar and are reported inSupplementary Fig. 1). Most of the CI values were lower than 1,indicating the presence of synergism with most of the concentrationsused. Zinc significantly increased the number of CI values lower than0.3, as calculated in Figs. 2C and D, indicating its ability to stronglyenhance the synergistic effect by GEM/PDTC or DSF. When thepercentages of CIb1 values were plotted versus the GEM IC50 of eachcell line, GEM-resistant cell lines (Panc1 and PaCa44) showed anantiproliferative synergism higher than those displayed by GEM-sensitive cell lines (MiaPaCa2 and T3M4), both in the absence and inthe presence of zinc ions (see Supplementary Fig. 2).

GEM/PDTC and GEM/DSF induce apoptosis and mitochondrialsuperoxide production

Fluorescence analyses with annexin V/FITC show that GEM/PDTCor GEM/DSF treatment significantly increased the apoptotic cell deathinduced by the single drugs. The addition of zinc enhanced theapoptotic effect of dithiocarbamates and dithiocarbamates/GEM, butnot that induced by GEM (Figs. 3A and B). These results are inagreement with those obtained from cell growth inhibition analysis(see Fig. 1).

Our previously published data reported that pancreatic adenocar-cinoma cell resistance to GEM is associated with low levels ofconstitutive ROS [6] and that oxidative stress generation plays a keyrole in pancreatic adenocarcinoma cell growth inhibition by GEM [6],PDTC [8], or PDTC/Zn [7]. To verify whether the association of GEMwith dithiocarbamates enhances ROS levels, we measured mitochon-drial O2

•− production in PaCa44 cells. Figs. 4A and B show that GEM/PDTC or GEM/DSF determined an induction of O2

•− higher than thatobtainedwith the single drugs. This effect was further increased in thepresence of zinc. Altogether, the above results indicate that DSFinduced a much stronger O2

•− production and apoptosis than PDTCwhen associated with GEM and ZnSO4.

50

60

50

60

A B***

***

**

****

*

10

20

30

40

Cel

l gro

wth

(%

)

10

20

30

40

Cel

l gro

wth

(%

)

** *

00

- + + - + + DSF+ - + + - + GEM GEM

- + + - + + PDTC

+ - + + - +

+ ZnSO4 + ZnSO4

Fig. 1. In vitro effects of (A) GEM+PDTC or (B) GEM+DSF treatment on PaCa44 cell growth. Cells were seeded in 96-well plates, incubated overnight, and treated with 10 μMGEMand/or 5.56 μM PDTC or 2.78 μM DSF for 16 h, in the absence or presence of 30 μM ZnSO4. Cell proliferation was determined using the crystal violet colorimetric assay as describedunder Materials and methods. Values are the means (±SD) of three independent experiments each performed in triplicate. ***pb0.001, **pb0.01, *pb0.05.

928 E. Dalla Pozza et al. / Free Radical Biology & Medicine 50 (2011) 926–933

Author's personal copy

Constitutive levels of MnSOD and mitochondrial superoxide correlatewith resistance to GEM and synergism between GEM and PDTC or DSF

As mentioned above, pancreatic adenocarcinoma cell lines containvarious basal levels of ROS that correlate with GEM resistance [6]. Toexamine whether MnSOD is responsible for this feature, wemeasuredboth the constitutive expression of the enzyme and the basalmitochondrial O2

•−. Fig. 5 shows that there was a direct correlationbetween GEM IC50 and MnSOD expression (Figs. 5A and D) and aninverse correlation between GEM IC50 and mitochondrial O2

•− basallevels (Figs. 5B and D). The same correlation was also found betweenthe percentage of synergism (CIb1) and the MnSOD or mitochondrialO2•− basal levels, respectively, both in the absence and in the presence

of zinc ions (Supplementary Figs. 3A and B). T3M4 cells transfectedwith the MnSOD expression vector (T3M4-T) showed decreased

mitochondrial superoxide production, both basal (Figs. 5B and D) andGEM-induced (Fig. 5C), and increased GEM resistance (Fig. 5D)compared to the untransfected cells.

GEM and DSF synergistically inhibit growth of human pancreaticadenocarcinoma cells in vivo

The effects of GEM and/or DSF/Zn on growth inhibition of PaCa44cells sc xenografted in nude mice were investigated. Examination ofthe volume–time curve (Fig. 6A) reveals that the volume of thetumors in mice treated with the combination GEM+DSF/Znremained practically unchanged, whereas it increased considerablyin the controls or after DSF/Zn treatment and to a lower extent afterGEM treatment. Fig. 6B shows that mouse body masses did notchange during the experiment, suggesting that the treatments did

A3

+ ZnSO4

0 0.2 0.4 0.6 1.00

1

2C

ombi

nati

on I

ndex

0.3

3

0

1

2

Com

bina

tion

Ind

ex

0.3

3

0

1

2

Com

bina

tion

Ind

ex0.3

B+ ZnSO4

0.8

0 0.2 0.4 0.6 1.00.8

Effect (GEM+PDTC)

3

0 0.2 0.4 0.6 1.00

1

2

Com

bina

tion

Ind

ex

0.3

0.8

Effect (GEM+PDTC)

Effect (GEM+DSF)0 0.2 0.4 0.6 1.00.8

Effect (GEM+DSF)

C

75

100D

75

100

GE

M+P

DT

C%

str

ong

syne

rgis

m (

CI<

0.3)

25

50

25

50

GE

M+D

FS

% s

tron

g sy

nerg

ism

(C

I<0.

3)

0ZnSO4

0- + - + - + - + - + - + - + - + ZnSO4

Fig. 2. Analysis of PaCa44 synergistic cell growth inhibitionby the combined treatments (A)GEM+PDTC or (B)GEM+DSF, in the absence or presenceof 30 μMZnSO4. The values on the xaxis correspond to the fraction of growth inhibition (0→1) given by increasing concentrations of drug combinations after 48 h treatment. The values on the y axis correspond to themeasurement of CI (see Materials and methods). Values are the means of three independent experiments each performed in triplicate. Effects of ZnSO4 addition to (C) GEM+PDTC or(D) GEM+DSF on strong antiproliferative synergism (CIb0.3) in T3M4,MiaPaCa2, PaCa44, and Panc1 cell lines. Themean values (±SD) of the percentages of strong synergismare shownat 48 h for each pancreatic adenocarcinoma cell line. ***pb0.001, **pb0.01, *pb0.05.

929E. Dalla Pozza et al. / Free Radical Biology & Medicine 50 (2011) 926–933

Author's personal copy

not produce any apparent toxicity. At the end of the treatmentperiod, the combination GEM+DSF/Zn determined a reduction inthe mean tumor mass of about 40- or 14-fold compared to control orGEM single treatment, respectively (Figs. 6C and D). It is noteworthy

that although DSF/Zn alone did not cause any significant suppres-sion of tumor growth, its combination with GEM determined a quitetotal inhibition of tumor growth as representatively shown inFig. 6D.

BA

250

***

250

175

200

225

***

175

200

225

**

100

125

150

100

125

150 ***

***

25

50

75 **

25

50

75

Ann

exin

V/F

ITC

bin

ding

(fo

ld in

duct

ion)

Ann

exin

V/F

ITC

bin

ding

(fo

ld in

duct

ion)

00

- + + - + + DSF

+ - + + - + GEM

- + + - + + PDTC + - + + - + GEM

+ ZnSO4 + ZnSO4

Fig. 3. Effects of (A) GEM+PDTC or (B) GEM+DSF treatment on apoptotic cell death in the PaCa44 cell line. Cells were seeded in 96-well plates, incubated overnight, and treatedwith 10 μM GEM and/or 5.56 μM PDTC or 2.78 μM DSF for 16 h, in the absence or presence of 30 μM ZnSO4. Apoptotic cell death was analyzed by using a multimode plate reader todetermine the fold induction of annexin V/FITC binding, as described under Materials and methods. Values are the means (±SD) of three independent experiments each performedin triplicate. ***pb0.001, **pb0.01, *pb0.05.

BA

300

325

…

******

20

****

100

125

…

**

**10

15

**

2-.pr

oduc

tion

0

5

10

0

5

Mit

ocho

ndri

al O

(fol

d in

duct

ion)

2-.pr

oduc

tion

Mit

ocho

ndri

al O

(fol

d in

duct

ion)

+ - + + - + GEM- + + - + + DSF

+ - + + - + GEM- + + - + + PDTC

+ ZnSO4 + ZnSO4

Fig. 4. Effects of (A) GEM+PDTC or (B) GEM+DSF treatment on mitochondrial superoxide production in the PaCa44 cell line. Cells were seeded in 96-well plates, incubatedovernight, and treated with 10 μM GEM and/or 5.56 μM PDTC or 2.78 μM DSF for 16 h, in the absence or presence of 30 μM ZnSO4. The fluorescence intensity of the MitoSOX redprobe, corresponding to the level of mitochondrial superoxide production, was measured by using a multimode plate reader, as described under Materials and methods. Values arethe means (±SD) of three independent experiments each performed in triplicate. ***pb0.001, **pb0.01, *pb0.05.

930 E. Dalla Pozza et al. / Free Radical Biology & Medicine 50 (2011) 926–933

Author's personal copy

Discussion

In this study, we have demonstrated that the combination of thestandard chemotherapeutic agent GEM and the ionophore compoundsPDTC or DSF synergistically inhibited pancreatic adenocarcinoma cellgrowthby inducingapoptotic cell death. Theantiproliferative synergismoccurs either in the absence or in the presence of a fixed dose of zincions, although it is stronger in the presence of zinc, and correlates withthe basal levels ofmitochondrial O2

•− and ofMnSOD expression. Growthof PaCa44 cells xenotransplanted into nude mice is strongly inhibitedafter treatment with GEM+DSF/Zn.

Recently, our research group has demonstrated that p53−/− pancre-atic adenocarcinoma cell lines, but not normal cells, are sensitive to PDTCtreatment via a ROS-mediated cytostatic mechanism dependent on theuptakeof zincpresent in the culturemedium[7].Wehavealso shownthatthe addition of nontoxic doses of zinc ions to PDTC induces apoptotic cell

death by further enhancing both intracellular zinc and ROS levelscompared to PDTC alone [8].Moreover,wehave previously demonstratedthat pancreatic adenocarcinoma cell growth inhibition by GEM is due toROS induction, at least to some extent, and that cell lines with a lowerbasal level of ROSaremore resistant toGEMcompared to cellswithhigherROS levels [6].

Our previous data and the consideration that cancer cells generallyexhibit an intrinsic oxidative stress higher than that of normal cells[5,16] encouraged us to investigate whether the exposure to furtherROS insults could efficiently inhibit pancreatic adenocarcinoma cellproliferation by exceeding the antioxidant cellular system activities.The present study reports that a combination of ROS-generating drugs,i.e., PDTC/Zn or DSF/Zn and GEM, synergistically inhibits pancreaticadenocarcinoma cell growth by activating apoptotic cell death andthat this apoptosis correlates with an increased mitochondrial O2

•−

production. Zinc has been shown to induce mitochondrial ROS mainly

Fig. 5. (A)Western blot analysis of constitutive MnSOD in the cell lines Panc1, PaCa44, MiaPaCa2, T3M4, and T3M4 transfected with theMnSOD expression vector (T3M4-T). The resultsshownare representative of three independent experiments. Thedensitometric analysis ofMnSODexpression levels, normalizedusing Ponceau S dye and reportedas arbitraryunits (a.u.),shows the mean values (±SD) of three independent experiments. (B) Basal levels of mitochondrial superoxide. The MitoSOX red fluorescence intensity, corresponding to the level ofmitochondrial superoxide production, is reported as relative mean fluorescence values (RMF). RMF is the ratio between fluorescence intensity of cells treated and untreated(autofluorescence) with theMitoSOX probe. Values are the means (±SD) of three independent experiments each performed in triplicate. (C) Effects of GEM treatment onmitochondrialsuperoxide production in T3M4 cells untransfected or transfected with the MnSOD expression vector (T3M4-T) or with the pCR3.1 empty vector (T3M4-T CTRL). Values are the means(±SD) of three independent experiments each performed in triplicate. ***pb0.001 for T3M4-T vs T3M4-T CTRL. (D) Correlation between the mean values of MnSOD basal levels (lightsymbols) or mitochondrial superoxide (dark symbols) and the GEM IC50 values at 48 h for each pancreatic adenocarcinoma cell line (T3M4, square; MiaPaCa2, triangle; PaCa44, circle;Panc1, rhombus). MnSOD basal levels of T3M4-T are indicated by a striped square and mitochondrial superoxide by a dotted square.

931E. Dalla Pozza et al. / Free Radical Biology & Medicine 50 (2011) 926–933

Author's personal copy

by inhibiting mitochondrial complex III, and GEM generates ROS viaacid sphingomyelinase activation or neutral ceramidase inhibition andsubsequent ceramide production [17,18], which directly affects themitochondrial electron transport chain [19]. Interestingly, interrup-tion of electron flow at complex III by antimycin A potentiates ROSinduction by ceramide, suggesting that the combination GEM+DSF/Zn operates with a similar mechanism.

Our findings show that MnSOD basal levels directly correlate withGEM resistance, further supporting the observation that low basallevels of ROS identify the GEM-resistant phenotype. Consistently, theGEM-sensitive T3M4 cells transfected with MnSOD expression vectoracquired properties similar to those of resistant cell lines.

Although MnSOD has been described as a tumor suppressor [20]capable of reducing proliferation of many cell lines, its high expressionreduces the efficacy of chemotherapy [21]. However, cells resistant toGEM and expressing higher levels of MnSOD present a strongersynergistic response to the combination GEM+DSF/Zn. Thus, ourresults strongly support the idea that the increase in ROS productionmay be a good strategy to overcome GEM resistance in the therapeuticmanagement of pancreatic cancer.

Our in vivo experiments show that intraperitoneal injections of GEM+DSF/Zn into nude mice bearing a subcutaneous mass of humanpancreatic adenocarcinoma cells completely inhibit tumor growth. Incontrastwith the in vitro results, GEMprovided a stronger inhibition thanDSF/Zn, which had no effect compared to control, probably because of

differences inpharmacokinetics. No apparent formof toxicity in vivo, suchas mouse death, body mass variations, or other apparent toxicity-relatedfeatures, was observed in mice treated with GEM+DSF/Zn.

Some recent studies have demonstrated that DSF displays a verystrong anticancer effect on cellular or xenograft models of severaltumor types, such as myeloma, leukemia, lymphoma, neuroblastoma,and colorectal cancer [22,23]. Moreover, it has been shown that DSFinhibits invasion of tumor cells at nontoxic concentrations anddisplays antiangiogenic effects [24]. A case report for the first use ofdisulfiram and Zn2+ (zinc gluconate) to treat advanced stage IVmetastatic melanoma has shown a N50% reduction in tumor size after3 months of therapy [25]. On the basis of all these observations, manyclinical trials have been approved and are currently in progress forvarious tumors to evaluate the anticancer efficacy of DSF alone or inaddition to standard chemotherapy. Recently, Guo et al. havepublished that DSF/copper complex is able to potentiate the cytotoxiceffect of GEM on colon and breast cancer cell lines [26], confirmingtheir previous study showing that DSF sensitizes colon cancer celllines to 5-fluorouracil (5-FU) and efficiently reverts 5-FU resistance[22]. Iljin et al. have recently identified DSF as the best drug in a panelof 4910 known compounds, including the most currently marketeddrugs, for its preferential inhibitory effect on prostate cancer cellsrelative to nonmalignant epithelial cells and for its relative safety,being used for decades as a Food and Drug Administration-approvedalcohol abuse deterrent [10,27].

Fig. 6. Effects of GEM+DSF/Zn on xenografts of PaCa44 cells in nudemice. PaCa44 cells were subcutaneously injected into female nude mice. After 1 week, intraperitoneal injectionswith PBS (vehicle solution), GEM, and/or DSF/Zn were carried out twice a week for 4 weeks, as described under Materials and methods. (A) Values are the means of mouse tumorvolume measured 3 days after each injection. (B) Values are the means of mouse body mass measured 3 days after each injection. (C) Values are the means of mouse tumor mass(±SD) measured after eight injections. (D) Representative photographs of tumor masses derived from mice treated with the indicated drugs after eight injections. ***pb0.001,**pb0.01, *pb0.05.

932 E. Dalla Pozza et al. / Free Radical Biology & Medicine 50 (2011) 926–933

Author's personal copy

In conclusion, our study provides the first evidence that DSF/Znand GEM efficiently inhibit pancreatic tumor cell growth in nudemice, strongly supporting a GEM/DSF-based clinical trial for pancre-atic adenocarcinoma patients.

Supplementarymaterials related to this article can be found onlineat doi:10.1016/j.freeradbiomed.2011.01.001.

Acknowledgments

We thank Dr. Akashi (National Institute of Radiological Sciences,Chiba, Japan), who kindly provided us with the expression vector forthe human MnSOD. This work was supported by the AssociazioneItaliana Ricerca Cancro, Milan, Italy; Fondazione CariPaRo, Padova,Italy; Joint Project of the University of Verona, Verona, Italy; andMinistero della Salute, Rome, Italy.

References

[1] Li, D.; Xie, K.; Wolff, R.; Abbruzzese, J. L. Pancreatic cancer. Lancet 363:1049–1057;2004.

[2] Neoptolemos, J. P.; Cunningham, D.; Friess, H.; Bassi, C.; Stocken, D. D.; Tait, D. M.;Dunn, J. A.; Dervenis, C.; Lacaine, F.; Hickey, H.; Raraty, M. G.; Ghaneh, P.; Büchler,M. W. Adjuvant therapy in pancreatic cancer: historical and current perspectives.Ann. Oncol. 14:675–692; 2003.

[3] Burris, H. A.; Moore, M. J.; Andersen, J.; Green, M. R.; Rothenberg, M. L.; Modiano,M. R.; Cripps, M. C.; Portenoy, R. K.; Storniolo, A. M.; Tarassoff, P.; Nelson, R.; Dorr,F. A.; Stephens, C. D.; Von Hoff, D. D. Improvements in survival and clinical benefitwith gemcitabine as first-line therapy for patients with advanced pancreascancer: a randomized trial. J. Clin. Oncol. 15:2403–2413; 1997.

[4] Ducreux, M.; Boige, V.; Malka, D. Treatment of advanced pancreatic cancer. Semin.Oncol. 34:S25–S30; 2007.

[5] Laurent, A.; Nicco, C.; Chereau, C.; Goulvestre, C.; Alexandre, J.; Alves, A.; Levy, E.;Goldwasser, F.; Panis, Y.; Soubrane, O.; Weill, B.; Batteux, F. Controlling tumorgrowth by modulating endogenous production of reactive oxygen species. CancerRes. 65:948–956; 2005.

[6] Donadelli, M.; Costanzo, C.; Beghelli, S.; Scupoli, M. T.; Dandrea, M.; Bonora, A.;Piacentini, P.; Budillon, A.; Caraglia, M.; Scarpa, A.; Palmieri, M. Synergisticinhibition of pancreatic adenocarcinoma cell growth by trichostatin A andgemcitabine. Biochim. Biophys. Acta 1773:1095–1106; 2007.

[7] Donadelli, M.; Dalla Pozza, E.; Costanzo, C.; Scupoli, M. T.; Piacentini, P.; Scarpa, A.;Palmieri, M. Increased stability of P21(WAF1/CIP1) mRNA is required for ROS/ERK-dependent pancreatic adenocarcinoma cell growth inhibition by pyrrolidinedithiocarbamate. Biochim. Biophys. Acta 1763:917–926; 2006.

[8] Donadelli, M.; Dalla Pozza, E.; Scupoli, M. T.; Costanzo, C.; Scarpa, A.; Palmieri, M.Intracellular zinc increase inhibits p53(−/−) pancreatic adenocarcinoma cellgrowth by ROS/AIF-mediated apoptosis. Biochim. Biophys. Acta 1793:273–280; 2009.

[9] Sauna, Z. E.; Shukla, S.; Ambudkar, S. V. Disulfiram, an old drug with new potentialtherapeuticuses forhumancancersand fungal infections.Mol.Biosyst.1:127–134;2005.

[10] Eneanya, D. I.; Bianchine, J. R.; Duran, D. O.; Andresen, B. D. The actions ofmetabolic fate of disulfiram. Annu. Rev. Pharmacol. Toxicol. 21:575–596; 1981.

[11] Fuller, R. K.; Gordis, E. Does disulfiram have a role in alcoholism treatment today?Addiction 99:21–24; 2004.

[12] Moore, P. S.; Sipos, B.; Orlandini, S.; Sorio, C.; Real, F. X.; Lemoine, N. R.; Gress, T.;Bassi, C.; Kloppel, G.; Kalthoff, H.; Ungefroren, H.; Lohr, M.; Scarpa, A. Geneticprofile of 22 pancreatic carcinoma cell lines: analysis of K-ras, p53, p16 and DPC4/Smad4. Virchows Arch. 439:798–802; 2001.

[13] Bruzzese, F.; Di Gennaro, E.; Avallone, A.; Pepe, S.; Arra, C.; Caraglia, M.; Tagliaferri,P.; Budillon, A. Synergistic antitumor activity of epidermal growth factor receptortyrosine kinase inhibitor gefitinib and IFN-alpha in head and neck cancer cells invitro and in vivo. Clin. Cancer Res. 12:617–625; 2006.

[14] Chou, T. C.; Talalay, P. Quantitative analysis of dose–effect relationships: thecombined effects of multiple drugs or enzyme inhibitors. Adv. Enzyme Regul. 22:27–55; 1984.

[15] Piacentini, P.; Donadelli, M.; Costanzo, C.; Moore, P. S.; Palmieri, M.; Scarpa, A.Trichostatin A enhances the response of chemotherapeutic agents in inhibitingpancreatic cancer cell proliferation. Virchows Arch. 448:797–804; 2006.

[16] Pelicano, H.; Carney, D.; Huang, P. ROS stress in cancer cells and therapeuticimplications. Drug Resist. Update 7:97–110; 2004.

[17] Modrak, D. E.; Leon, E.; Goldenberg, D. M.; Gold, D. V. Ceramide regulatesgemcitabine-induced senescence and apoptosis in human pancreatic cancer celllines. Mol. Cancer Res. 7:890–896; 2009.

[18] Dumitru, C. A.; Sandalcioglu, I. E.; Wagner, M.; Weller, M.; Gulbins, E. Lysosomalceramide mediates gemcitabine-induced death of glioma cells. J. Mol. Med. 87:1123–1132; 2009.

[19] Garcia-Ruiz, C.; Colell, A.; Mari, M.; Morales, A.; Fernandez-Checa, J. C. Direct effectof ceramide on the mitochondrial electron transport chain leads to generation ofreactive oxygen species: role of mitochondrial glutathione. J. Biol. Chem. 272:11369–11377; 1997.

[20] Hurt, E. M.; Thomas, S. B.; Peng, B.; Farrar, W. L. Molecular consequences of SOD2expression in epigenetically silenced pancreatic carcinoma cell lines. Br. J. Cancer97:1116–1123; 2007.

[21] Kinnula, V. L.; Crapo, J. D. Superoxide dismutases in malignant cells and humantumors. Free Radic. Biol. Med. 36:718–744; 2004.

[22] Wang, W.; McLeod, H. L.; Cassidy, J. Disulfiram-mediated inhibition of NF-kappaBactivity enhances cytotoxicity of 5-fluorouracil in human colorectal cancer celllines. Int. J. Cancer 104:504–511; 2003.

[23] Wickstrom, M.; Danielsson, K.; Rickardson, L.; Gullbo, J.; Nygren, P.; Isaksson, A.;Larsson, R.; Lovborg, H. Pharmacological profiling of disulfiram using humantumor cell lines and human tumor cells from patients. Biochem. Pharmacol. 73:25–33; 2007.

[24] Shian, S. G.; Kao, Y. R.; Wu, F. Y.; Wu, C. W. Inhibition of invasion and angiogenesisby zinc-chelating agent disulfiram. Mol. Pharmacol. 64:1076–1084; 2003.

[25] Brar, S. S.; Grigg, C.; Wilson, K. S.; Holder Jr., W. D.; Dreau, D.; Austin, C.; Foster, M.;Ghio, A. J.; Whorton, A. R.; Stowell, G. W.; Whittall, L. B.; Whittle, R. R.; White, D.P.; Kennedy, T. P. Disulfiram inhibits activating transcription factor/cyclic AMP-responsive element binding protein and human melanoma growth in a metal-dependent manner in vitro, in mice and in a patient with metastatic disease. Mol.Cancer Ther. 3:1049–1060; 2004.

[26] Guo, X.; Xu, B.; Pandey, S.; Goessl, E.; Brown, J.; Armesilla, A. L.; Darling, J. L.;Wang, W. Disulfiram/copper complex inhibiting NFkappaB activity and potenti-ating cytotoxic effect of gemcitabine on colon and breast cancer cell lines. CancerLett. 290:104–113; 2010.

[27] Iljin, K.; Ketola, K.; Vainio, P.; Halonen, P.; Kohonen, P.; Fey, V.; Grafstrom, R. C.;Perala, M.; Kallioniemi, O. High-throughput cell-based screening of 4910 knowndrugs and drug-like small molecules identifies disulfiram as an inhibitor ofprostate cancer cell growth. Clin. Cancer Res. 15:6070–6078; 2009.

933E. Dalla Pozza et al. / Free Radical Biology & Medicine 50 (2011) 926–933