Cancer Letters 332 (2013) 35–45

Contents lists available at SciVerse ScienceDirect

Cancer Letters

journal homepage: www.elsevier .com/locate /canlet

Original Articles

Cucurbitacin B potently suppresses non-small-cell lung cancer growth:Identification of intracellular thiols as critical targets

0304-3835/$ - see front matter � 2013 Elsevier Ireland Ltd. All rights reserved.http://dx.doi.org/10.1016/j.canlet.2013.01.008

⇑ Corresponding author. Address: James Graham Brown Cancer Center, Universityof Louisville, 304E Delia Baxter II, 580 S. Preston St., Louisville, KY 40202, UnitedStates.

E-mail address: rcgupta@louisville.edu (R.C. Gupta).

Hina Kausar a, Radha Munagala a, Shyam S. Bansal b, Farrukh Aqil a, Manicka V. Vadhanam a,Ramesh C. Gupta a,b,⇑a James Graham Brown Cancer Center, University of Louisville, Louisville, KY 40202, United Statesb Department of Pharmacology and Toxicology, University of Louisville, Louisville, KY 40202, United States

a r t i c l e i n f o

Article history:Received 31 October 2012Received in revised form 24 December 2012Accepted 4 January 2013

Keywords:Cucurbitacin BLung cancerCellular thiolsApoptosis

a b s t r a c t

Cucurbitacin B (CuB), has recently emerged as a potent anticancer agent; however, its efficacy in non-small-cell lung cancer (NSCLC) and the mechanism(s) initiating its biological effects remain largely unclear. In thisstudy, CuB potently suppressed the growth of four NSCLC cells (H1299, A549, HCC-827 and H661) in vitroand the highly aggressive H1299 xenograft in vivo. CuB significantly altered the actin cytoskeletal assembly,induced G2/M cell-cycle arrest and mitochondrial apoptosis through the modulation of several key molec-ular targets mediating the aforementioned processes. Interestingly, all cellular effects of CuB were com-pletely attenuated only by the thiol antioxidant N-acetylcysteine (NAC). Furthermore, pretreatment withglutathione synthesis inhibitor butithione-sulfoxime (BSO), significantly exacerbated CuB’s cytotoxiceffects. To this end, cells treated with CuB revealed a rapid and significant decrease in the levels of proteinthiols and GSH/GSSG ratio, suggesting disruption of cellular redox balance as the primary event in CuB’scytotoxic arsenal. Using UV and FTICR mass spectrometry we also demonstrate for the first time a physicalinteraction of CuB with NAC and GSH in a cell-free system suggesting that CuB interacts with and modulatescellular thiols to mediate its anti-cancer effects. Collectively, our data sheds new light on the working mech-anisms of CuB and demonstrate its therapeutic potential against NSCLC.

� 2013 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Despite recent advances in modern medicine, non-small celllung cancer (NSCLC) still continues to be a radical health concern[4]. Dysregulation of several key signal transduction pathways, de-fects in cell cycle progression and resistance to apoptotic signalinghas largely been attributed to the uncontrolled proliferation ob-served in lung tumor cells [16,34], thus warranting the need tocharacterize new agents that can effectively re-enforce the defensemachinery of tumor cells propelling them to undergo apoptosis.

Over the recent years, apoptosis induction due to alterations inintracellular redox status by certain oxidative or non-oxidativestimuli has become a focus of extensive research [11]. Besidesreactive oxygen species (ROS), intracellular thiols – GSH and theprotein thiols are the two major determinants of cellular redoxequilibrium, functioning as regulators of cell survival through theireffects on gene activation, cell cycle arrest and apoptosis [14,26].

Hence, manipulation of cellular redox status represents a viablestrategy for apoptosis induction by anti-cancer agents.

Lately, the natural triterpenoids, cucurbitacins have become asubject of intense investigation because of their high efficacy to in-hibit varied cancers [30,38]. Based on molecular structural differ-ences, the cucurbitacins are classified into twelve groups(denoted A through T) [7]. Amongst them, the terpenoid cucurbita-cin B (CuB) has recently gained increasing attention because of itspotent in vitro and in vivo antitumor effects [8,42,45]. The biologicaleffects of CuB has been shown to be mediated mainly through its ef-fects on the actin cytoskeleton, intracellular ROS levels, cell-cyclearrest and apoptosis through the modulation of COX2, NF-kappaB,JAK2/STAT3, ERK and MAPK pathways, [18,21,22,42,45,48,50].While many of these studies have mostly delineated events occur-ring relatively late after the exposure of cells to CuB, the mecha-nisms initiating CuB’s biological effects are still not clear yet. Alsowith the exception of a few published studies exploring the cyto-toxic effects of CuB and CuI (another member of the cucurbitacinfamily) on lung cancer growth [20,21,27], detailed mechanisticevaluation of CuB’s antiproliferative effects in lung cancer is stilllacking. The present study was therefore designed to investigatethe anticancer potential of CuB against NSCLC and identify theunderlying mechanisms initiating its cytotoxic effects.

36 H. Kausar et al. / Cancer Letters 332 (2013) 35–45

2. Materials and methods

2.1. Reagents and culture materials

CuB (95% pure) was purchased from Phytomyco Research (Greenville, NC). NACand BSO (Sigma–Aldrich, St. Louis, MO), MTT (Calbiochem, Gibbstown, NJ), floures-cent probe H2DCFDA, cell culture medium and other supplements (Invitrogen,Grand Island, NY), were obtained from the sources cited.

2.2. Cells, culture conditions and treatments

H1299 (p53null/EGFRWT), H661 (p53mutant/EGFRWT), and A549 (p53WT/EGFRWT) cells were a kind gift from Drs. Paula Bates and Keith Davis (Universityof Louisville) and Dr. C.G. Gairola (University of Lexington KY) respectively. HCC-827 (p53WT/EGFR mutant) cells were obtained from ATCC. The cell lines wereauthenticated by short tandem repeat sequencing and mycoplasma detection anal-ysis performed at the John Hopkins genetic resources core facility (Baltimore, MD).Cells were maintained in recommended medium at 37 �C in 5% CO2 and humidifiedair. For all studies, cells at 70–75% confluency were treated with vehicle (0.1%DMSO) or indicated concentrations of CuB and 5 mM each of, NAC and/or BSO.

2.3. Cell-viability assay

Cells (4.5 � 103/well) were cultured at 37 �C in 96-well plates with vehicle,varying concentrations of CuB (0.025–1 lM) for 24, 48 or 72 h, 5 mM NAC + CuB,or 5 mM BSO + CuB + 5 mM NAC for 72 h and cell viabilities were measured usingthe MTT reduction assay [24].

2.4. In vivo xenograft assay

Female athymic nude (nu/nu) mice (5–6-week old), from Harlan laboratories(Indianapolis, IN) were maintained in accordance with the Institutional Animal Careand Use Committee guidelines. A suspension of 1 � 106 H1299 cells in serum-freemedia were mixed with Matrigel matrix (Becton Dickinson, Bedford, MA) and sub-cutaneously injected into the left flank of each mouse. The following day, mice wererandomly divided into 2 groups of 9 mice each and treated intraperitoneally witheither vehicle (10% DMSO in PBS) or CuB (1 mg/kg body weight) on alternate days.Tumor size and animal weights were monitored weekly. After 5 weeks of treatment,the experiment was terminated due to severe scabbing of few tumors in the controlgroup. Post-euthanasia, tumors were harvested and snap frozen in liquid nitrogenfor further analysis.

2.5. Confocal laser scanning microscopy for F-actin staining

H1299 cells grown in 8-well chambered slides were treated with vehicle, CuB orthe p38 inhibitor SB203580 (SelleckChem, Boston MA) for 2 h and then labeled withrhodamine–phalloidin as described previously [45]. Post-staining, the cells wereexamined with a Zeiss LSM410 confocal system.

2.6. Intracellular ROS measurement

Early levels of intracellular ROS generation were assessed spectrofluorimetrical-ly by oxidation of H2DCFDA [1]. Briefly, cells seeded on black 96-well tissue cultureplates were washed once with PBS, then incubated concomitantly with differentconcentrations of CuB and 100 lM H2DCFDA in PBS. Fluorescence intensity wasmeasured immediately after addition (0 min), then at 15, 45, 90 and 180 min atexcitation and emission wavelengths of 490 nm and 535 nm, respectively.

2.7. Cell-cycle analysis

Flow cytometric evaluation of cell cycle distribution in CuB- or vehicle-treatedH1299 cells was performed as described previously [36]. The population of cells ineach cell-cycle phase was determined using Flow Jo software (Becton Dickinson).

2.8. Assays for apoptosis

Flow cytometric analysis: Apoptotic populations of vehicle- or CuB-treated cellswere quantified using the dual staining annexin V-FITC/PI apoptosis detection kit(Invitrogen, Grand Island, NY) as per manufacturer’s instructions. Morphologicalexamination: Vehicle- and CuB-treated cells were stained with 1:1 mixture of ethi-dium bromide (EB) and acridine orange (AO) (5 lg/ml each) for 5 min, and thenanalyzed for morphological changes under a fluorescence microscope (NikonEclipse TS100).

2.9. Western-blot analysis

Western-blot analysis was done as previously described [36]. Blots were probedfor Bcl2, Bax, CyclinB1, cdc2p34, p21 (WAF1), pERK, cytochrome C, Hsp27 antibod-ies (Santa Cruz Biotechnology, Santa Cruz, CA). Caspase-9, cleaved caspase-3, PARP,PCNA, pSTAT3, STAT3, pHSP27, pP38, p38, pJNK, Rho, pRac, Cdc42, ERK1/2 antibod-ies (Cell Signaling Technology, Beverly, MA) and gelsolin antibody (R&D systems,Minneapolis, MN). b-Actin (Sigma, St. Louis, MO) was used as loading control.

2.10. Determination of GSH and protein thiol content

GSH levels in H1299 cells treated with vehicle or 0.35 lM CuB in the presenceand absence of 5 mM NAC for 1 and 4 h was determined with GSH/GSSG-GloTM as-say kit (Promega, Madison, WI) as per manufacturer’s instructions. Cellular proteinthiol content was measured as described [13]. Briefly, post-treatment, H1299 cellpellets (obtained after deproteination with sulfosalicylic acid) were finally resus-pended in 0.5 M Tris–HCl, (pH 7.6). DTNB (100 lM) was then added and 30 min la-ter the absorbance was measured at 410 nm. The protein concentrations of thesolutions were determined using BCA protein assay kit (Thermo Fisher Scientific,Rockford, IL) and the thiol content of each sample was normalized to their respec-tive protein content.

2.11. Determination of CuB-NAC binding in vitro

To evaluate CuB-NAC interaction, increasing concentrations of NAC (0.98, 1.9and 3.9 mM) were added to 100 ll of CuB (9 lM) in a 96-well microplate. Theintrinsic absorbance of CuB was monitored by an absorbance scan at 200–450 nmusing M2 spectramax spectrophotometer. A red-shift in the spectrum indicatescomplex formation.

2.12. Conjugation of CuB with NAC and GSH

CuB (5 mM stock in DMSO diluted 1:10 in PBS) was incubated with NAC(500 mM stock in DMSO diluted 1:100 in PBS) or GSH (5 mM stock in DMSO diluted1:10 in PBS) for 10 min and 30 min at room temp, respectively. Post-incubation, thesamples were further diluted 1:10 in methanol and analyzed by direct injectionusing nano-spray (TriVersa NanoMate, Advion BioSciences, Ithaca, NY) on a hybridlinear ion trap FT-ICR-MS (Fourier-transform Ion cyclotron resonance mass spec-trometer, Finnigan LTQ FT, Thermo Electron, Bremen, Germany). The electron-sprayionization was performed in positive ion mode by applying 1.85 kV with no headpressure. MS runs were recorded over a mass range from 100 to 1000 Daltons.MS/MS analysis was done on select ions to confirm adduct formation.

2.13. Statistical analysis

Statistical differences between groups were assessed by Students t-test. Differ-ences between mean were considered significant if p < 0.05.

3. Results

3.1. CuB suppresses the growth of lung cancer cells in vitro and in vivo

CuB dose-dependently inhibited the proliferation of all NSCLCcells (H1299, A549, H661 and HCC-827) irrespective of their p53or EGFR status; time-dependent inhibition was also evident asinvestigated with H1299 cells (Fig. 1a and b). CuB was effectivein inhibiting the proliferation of all the cell lines with IC50 valuesranging between 0.05 and 0.130 lM. Since the H1299 cells arethe most aggressive in terms of tumor xenograft growth, the effectof CuB on H1299 xenograft growth in nude mice was further ex-plored. As shown in Fig. 1c and 1d, CuB significantly bluntedH1299 tumor xenograft growth throughout the duration ofthe study. The mean tumor volume at the end of the study inCuB-treated mice was 200 ± 111 mm3 as compared to 684 ± 321 mm3

in the control group which corresponds to an average reduction of70% in tumor volume (p < 0.005). No visible sign of toxicity was ob-served in CuB-treated mice based on gross visual pathology ofinternal organs and body weight gains between both the groups.

3.2. CuB induces morphological changes and actin cytoskeletalremodeling via p38MAPK/Hsp27 activation

Mechanistically, rapid alterations in H1299 cellular morphologywere observed within minutes of CuB treatment as evidenced by

Fig. 1. Effect of CuB on NSCLC cell viability and tumor xenograft growth. (a) Cell viability of vehicle or CuB-treated H1299 cells following 24, 48 and 72 h of CuB treatment.Data represent mean (±) of 3 independent experiments. a, p < 0.001, and b, p < 0.0001 compared to vehicle treatment. (b) Effect of 72 h of CuB treatment on cell viabilities ofA549, HCC-827 and H661 cells. a, p < 0.001, and b, p < 0.0001 compared to vehicle treatment. (c) Mean tumor volumes of mice receiving intraperitoneal injections of eitherCuB (1 mg/kg) or vehicle (10% DMSO in PBS) on alternate days for 5 weeks. Bars represents SD, ��(p < 0.005). (d) Representative image of tumor-bearing mice and excisedtumors from vehicle- or CuB-treatment groups.

H. Kausar et al. / Cancer Letters 332 (2013) 35–45 37

the appearance of moderate to severe surface–plasma membraneprotrusions, called blebs, following 30 min of CuB treatment(Fig. 2a). A correlation between surface bleb formation and pertur-bation of normal actin cytoskeletal organization is reported [3]hence, the effect of CuB on F-actin cytoskeletal microfilaments inH1299 cells was analyzed. A 2 h CuB treatment dramatically chan-ged the distribution of F-actin from the homogeneous network incontrol cells to large intensely-fluorescent cytoplasmic aggregates,pointing to an increase in F-actin polymerization/stabilization(Fig. 2b(i)). Evaluation of signaling molecules (Fig. 2c) demon-strated early activation of p38 MAPK which led to an increase inthe phosphorylation of the F-actin polymerization modulatorHsp27 that could be in part responsible for the observed polymer-ization and stability of the F-actin microfilaments [29]. Further ver-ification for involvement of p38 MAPK signaling was obtained withthe use of p38 MAPK inhibitor SB203580, which mostly preventedCuB-induced F-actin polymerization (Fig. 2b(ii)).

Actin-microfilament remodeling has also been shown to be reg-ulated either by the Rho-family of GTPases or the ROS [12,17],

hence the effect of CuB on Rho-family signaling molecules (Rac,Rho and cdc42) and intracellular ROS levels was further evaluated.As shown in Fig. 2d, while CuB had no obvious effect on the levelsof RhoA, Rac, and cdc42 in the early time points (15 min, 2 h),downregulation of both pRac and cdc2 was observed following 8and 24 h of CuB treatment. Additionally, up to 3 h post-CuB treat-ment except at very high (10 lM) CuB concentration, no increasein intracellular ROS levels was observed at 0.1 lM concentrationat which CuB elicited the cytoskeletal changes (Fig. 2e). These datasuggest the possible unlikeliness of the Rho family GT-Pases andintracellular ROS levels in mediating CuB-induced actin-cytoskele-tal changes at the earlier time points.

3.3. CuB induces G2/M cell cycle arrest and mitochondrial apoptosis

CuB has been shown to inhibit G2/M cell-cycle transition andinduce mitochondrial apoptosis in other cell types [8,41]. Flow-cytometric analysis of CuB-treated H1299 cells showed an earlyand sustained G2/M cell arrest (40–60%) throughout the treatment

Fig. 2. Effect of CuB on cellular morphology actin cytoskeleton and intracellular ROS generation. (a) Morphologic changes in H1299 cells treated with CuB compared withvehicle treatment (magnification 20�). (b) (i and ii) Representative confocal images of F-actin assembly in H1299 cells exposed to (i) vehicle or CuB (ii) vehicle or CuB after30 min pre-incubation with 10 lM of p38 MAPK inhibitor SB203580 at 100� magnification. (c) Immuno-blot analysis of MAPKs after 15 min and 2 h vehicle and CuBtreatment, respectively. The phosphorylated proteins were normalized with their corresponding total proteins. b-actin has been shown for equal loading. Numbers aboveband depict changes in phosphorylated protein levels relative to the corresponding DMSO-treated controls. (d) Immuno-blot analysis of Rho-family GTPases as mediators ofcytoskeletal alterations after vehicle and CuB treatment, respectively. (e) Early H2O2 accumulation in H1299 cells exposed to 0.1, 0.5 and 10 lM concentrations of CuB. Therelative H2O2 production is expressed as the % increase in fluorescence in CuB-treated cells compared to cells exposed to PBS alone. Bars represent SD of 3 independentexperiments performed in triplicates. a, p < 0.05, and b, p < 0.005 compared to vehicle treatment. ⁄Fold change values shown for phosphorylated proteins only.

38 H. Kausar et al. / Cancer Letters 332 (2013) 35–45

H. Kausar et al. / Cancer Letters 332 (2013) 35–45 39

duration (Fig. 3a). Additionally, apoptotic populations of 10% and14% at 24 h and 22% and 33% at 48 h were also observed following0.1 lM and 0.35 lM of CuB-treatment, respectively (Fig. 3b). Vali-dation for apoptosis was further obtained by AO/EB staining to de-tect apoptotic-morphological alterations such as membraneblebbing, chromatin aggregation and nuclear condensation(Fig. 3c). Similar effects of CuB on cell cycle arrest and apoptosiswere also observed in A549 cells (data not shown).

Evaluation of signaling molecules revealed a dose- and time-dependent decrease in the activation of ERK and STAT3 proteins,induction in the levels of cell cycle regulatory cyclin kinase inhib-itor, p21 (WAF1) and down regulation of cyclin B1, cdc2p34(Cdk1), cyclin D1 and PCNA proteins (Fig. 3d(i and ii). Furthermore,a decrease in the expression of anti-apoptotic protein Bcl2 accom-panied by an increase in pro-apoptotic Bax, an increase in the cyto-solic levels of cytochrome c, a significant and progressiveactivation of caspase-9 and caspase-3 and cleavage of their down-stream targets PARP and gelsolin were also observed (Fig. 3e).These data collectively suggest that CuB alters the levels of cellularproteins to induce G2/M arrest and mitochondrial apoptosis in thecells.

Additionally, the expression levels of these proteins in the tu-mor lysates of the CuB-treated mice were also analyzed. In corrob-oration with the in vitro studies, a similar modulation in expressionof the biomarkers was observed (Fig. 3f(i and ii) and g(i and ii)).

3.4. CuB’s cytotoxicity and other cellular effects are blocked by the thiolantioxidant NAC and not vitamin C and ascorbic acid

The dose response for NAC chemoprotection was evaluated inH1299 cells treated concurrently with varying concentrations ofNAC (0.1, 1, 5 and 10 mM) and CuB (0.1–1 lM) for 72 h. NAC signif-icantly protected cells from CuB-induced cytotoxicity dose-depen-dently (Fig. 4a). Notably, other antioxidants ascorbic acid and‘‘vitamin E’’ at 0.2 mM concentration failed to afford much protec-tion (data not shown). To further comprehend the protective roleof NAC, the effect of NAC adding time on CuB’s cytotoxicity was as-sessed (Fig. 4b). Incubating cells with a mixture of 0.35 lM CuBand 5 mM NAC or a 2 h pretreatment with 5 mM NAC followedby addition of 0.35 lM CuB in the continued presence of NAC affor-ded complete protection. However, removal of NAC prior to theaddition of CuB or addition of NAC, 2 h post-treatment with CuB of-fered either null or minimal protection, thus suggesting (i) theneed for concurrent presence of both NAC and CuB to prevent CuB’scytotoxicity and (ii) irreversibility of CuB’s effects.

Because CuB’s cytotoxic effects were mediated by the onset ofcellular morphological changes, actin-cytoskeletal alterations, G2/M arrest and apoptosis, the effect of NAC (5 mM) on these eventswas explored. As shown in Fig. 4c–e, NAC completely blocked allthe cellular processes modulated by CuB.

3.5. CuB-induces depletion of intracellular thiol content and alterscellular redox balance

In the absence of significant ROS generation and preferentialinhibition of CuB’s cytotoxicity by the thiol antioxidant NAC, wesought to determine CuB’s effects on the levels of intracellular thi-ols in H1299 cells. Following treatment with 0.35 lM CuB, an in-crease in GSSG (108% at 1 h and 131% at 4 h) paralleled the lossof GSH (92% at 1 h and 79% at 4 h) (Fig. 5a, data shown only for4 h), causing a decrease in the GSH/GSSG ratio of 6.95 at 1 h and4.51 at 4 h as compared to 8.74 in vehicle-treated sample(Fig. 5b). Additionally, a rapid and significant depletion of proteinthiols was also observed following 1 h CuB treatment (Fig. 5c)and all the above responses were partially or completely preventedby NAC (Fig. 5a–c).

Further verification of the vital role of GSH depletion in CuB’scytotoxic effects was obtained when a significant enhancementin CuB’s cytotoxicity was seen following a 4 h pre-treatment ofH1299 cells with 5 mM concentration of the glutathione synthesisinhibitor BSO (Fig. 5c). These findings thus suggest a critical role forcellular thiols in the development of CuB’s cytotoxicity. To rule outthe possibility that the observed effects of CuB on the thiols andsubsequent protection of its cytotoxicity by NAC was a cell line-specific effect, both A549 and HCC-827 cells were also treated withCuB in the presence and absence of NAC and/or BSO which essen-tially resulted in a similar trend in effects (Fig. 5d and e), thus sug-gesting the universality of CuB’s effect.

3.6. NAC and GSH interact with CuB in a cell-free system

Since the concurrent presence of both NAC and CuB in the cellculture medium was necessary to prevent CuB’s cytotoxic effects,the UV absorption characteristics of CuB in the presence and ab-sence of NAC was analyzed to determine a possible interaction be-tween the two. NAC exerted both a hypochromic effect on CuBabsorption and a bathochromic shift in kmax of CuB. A dose-depen-dent decrease in the absorption intensity of CuB from 2.1 to 1.4, 1.1and 0.65 with 0.98, 1.9 and 3.9 mM of NAC, respectively, and a shiftin kmax of CuB from 212 nm to 220 nm in the presence of NAC wereobserved thus suggesting that NAC does interact and bind withCuB (Fig. 6a). Similar effects were observed with GSH; howevervitamin E and ascorbic acid did not influence the absorption char-acteristics of CuB (data not shown).

To further confirm the CuB-thiol interaction, CuB was incubatedwith NAC and GSH respectively, and the masses of the individualcompounds and the resulting adducts were examined using FTICRmass spectrometry (Fig. 6b and c). A major peak was obtained forCuB with the addition of 1Na+ (m/z 581.31), NAC with the additionof 1Na+ (m/z 186.02) and GSH with the addition of 1H+(m/z 308.09)respectively. Furthermore, one major adduct with NAC at m/z 744.339 [CuB + NAC + Na]+ and 2 probable adducts with GSH atm/z 866.413 [CuB + GSH + H]+ and m/z 888.396 [CuB + GSH + Na]+,were also detected. While the reaction between CuB with NACwas very fast and gave one abundant adduct peak, its reaction withGSH was slow and resulted in less enriched adduct peaks. It isworthwhile to mention here that the complex of CuB with NACwas however not detectable by HPLC-UV, presumably due toeither the low sensitivity of HPLC or due to disintegration of thehydrophobic interaction between CuB and NAC under the chro-matographic conditions.

4. Discussion

In the current study, we demonstrate that CuB is a potent sup-pressor of human NSCLC cell growth both in vitro and in vivothrough its effects on actin-cytoskeleton, MAPK, STAT3, cell-cycleregulatory and apoptotic signaling. A direct role of p38 MAPK/Hsp27 signaling in mediating CuB-induced actin cytoskeletal alter-ations is further demonstrated. We also show for the first time thatCuB disrupts cellular redox status by depleting cellular GSH andprotein thiols levels and cellular GSH/GSSG ratios and that this de-crease in GSH levels is probably due to direct reaction of CuBwith GSH as evidenced by UV and FTICR mass spectrometricapproaches.

A plethora of recent reports suggest pivotal role for cellular re-dox status in the regulation of cell growth and function [10,43]. Inaddition to ROS, intracellular GSH/GSSH ratio and protein thiols arekey redox regulators that are crucial for multiple biologicalfunctions, including actin cytoskeletal remodeling [37], cell cycleprogression [44], mitochondrial apoptosis [15] and regulation of

Fig. 3. Effect of CuB on cell-cycle arrest and apoptosis. (a and b) Flow cytometric evaluation of (a) cell cycle distribution and (b) apoptosis following vehicle- and CuB-treatment of H1299 cells for 8, 24 and 48 h. Bars represent SD. �(p < 0.05), ��(p < 0.005) and ���(p < 0.0005) compared with vehicle- treated cells. (c) Apoptosis was alsodetermined by fluorescence staining with AO/EB. Insets and arrowheads indicate cells with normal nuclear morphology in control cells and membrane blebbing, condensedchromatin and nuclear fragmentation in CuB-treated cells. (d) Effects of CuB on the activation of ERK, STAT3 and other cell cycle-related proteins p21 (WAF1) cyclin B1,cdc2p34, cyclin D1 and PCNA. b-actin was used loading control. (e) Western-blot analysis of CuB- and vehicle-treated cells for protein expression of mitochondrial apoptosisregulators. Arrowheads indicates cleaved product of each protein. (f) (i and ii) Representative immunoblots from tumor lysates of vehicle and CuB-treated animals formarkers of cell cycle arrest and apoptosis. (g) (i and ii) Tables represent densitometric quantitation (mean arbitrary units ± SD) of cell cycle and apoptotic proteins fromvehicle and CuB treatment groups. Actin loading control was utilized to normalize the values. Statistical significance was determined by student’s t-test (n = 8 for control andn = 8 for CuB treatment group), �p < 0.05.

40 H. Kausar et al. / Cancer Letters 332 (2013) 35–45

redox sensitive signal transduction proteins and transcriptionfactors [9,40]. Furthermore, Franco et al. has also demonstratedthat alterations in cellular thiol levels can trigger apoptosis

independently of ROS [15]. In view of these reports and coupledwith (i) the lack of ROS involvement in this study, (ii) significantdepletion of cellular thiol levels, and (iii) exaggeration of CuB’s

Fig. 4. Effect of NAC on CuB-induced cellular response. (a) Cell viability of H1299 cells incubated concurrently with 5 mM NAC and (0.125–1 mM) CuB for 72 h, a (p < 0.005)compared with 0.125 lM of CuB, b (p < 0.005) compared with 0.25 lM of CuB, c (p < 0.005) compared with 0.5 lM of CuB and d (p < 0.005) compared with 1.0 lM of CuB. (b)Effects of NAC (5 mM) on CuB-induced cytotoxity was studied by application of NAC to cells for 2 h before or after, CuB (0.35 lM) a (p < 0.005) when compared with control.(c–e) H1299 cells were treated with CuB (0.35 lM) or NAC (5 mM) alone and in combination and observed for (1) morphologic changes (2) phosphorylation status of MAPKsmediating the cytoskeletal alterations and (3) proportion of cells in G2/M phase and apoptosis a (p < 0.005) compared with control b, (p < 0.005) compared with 0.35 lM ofCuB.

H. Kausar et al. / Cancer Letters 332 (2013) 35–45 41

cytotoxicity by the glutathione synthesis inhibitor BSO, it is there-fore reasonable to postulate that CuB-induced modulation of intra-cellular thiols might be amenable for its downstream effects. Whilethe observed effects of CuB on actin cytoskeleton, cell cycle andapoptosis is in concordance with previously published reports[8,42,49], the inability of CuB to induce intracellular ROS levels,however, is in disagreement with the study by Yasuda et al. [48]which probably could be attributed to differences in cell typesand the time points analyzed. While Yasuda et al. measured intra-cellular ROS levels following 24 h of CuB exposure, we measuredearly ROS levels following 0–3 h exposure to CuB, to preferentiallyaccount for its role in the initiation of CuB’s cellular effects.

An intriguing observation that remains central to our study andother studies by Yasuda et al. [48] and Zhang et al. [50] though, is

the complete protection of CuB’s biological effects by the thiol anti-oxidant NAC. While the protective effects of NAC in those studiesare attributed to its antioxidant effects either by decreasing CuB-induced cellular ROS levels [48] or basal cellular ROS levels [50],we postulate that NAC being a thiol nucleophile, (1) can interactwith CuB in the extracellular medium to form a CuB–NAC adductor (2) compete with cellular thiols targeted by CuB, thus leadingto the subsequent loss of CuB’s biological effects.

A fundamental question, which remains unclear though, is howCuB interacts with the cellular thiols and modifies their levels?According to previous reports, certain drugs can bind intracellularGSH and protein thiols and thereby decrease their levels[2,5,33,39]. Additionally, some cyclopentenone prostaglandinsand sesquiterpene lactones have also been shown to react with a

Fig. 5. Effect of NAC and CuB on intracellular thiol levels. (a and b) Effects of 0.35 lM CuB treatment in the presence and absence of 5 mM NAC on intracellular (a) GSH andGSSG levels and (b) GSH/GSSG ratios. Bars represent ± SD from 3 replicates. a (p < 0.005) compared with control, b (p < 0.005) compared with CuB. (c) Protein thiol content ofH1299 cells following 1 h treatment with 0.35 lM CuB in the presence and absence of NAC. a (p < 0.005) compared with control, b (p < 0.005) compared with CuB. (d) Cellviabilities of H1299, cells pretreated for 4 h with 5 mM BSO, followed by removal of BSO and then treated with either vehicle or CuB in the presence and absence of 5 mM NACfor 72 h. a (p < 0.005) compared with control, b,c (p < 0.005) compared with 0.1 and 0.35 lM of CuB, respectively, d,e (p < 0.005) compared with 0.1 and 0.35 lM of CuB in thepresence of BSO, respectively. (e and f) Cell viabilities of A549 and HCC-827 cells when treated with CuB in the absence or presence of 5 mM each of NAC and/or BSO. a(p < 0.005) compared with control b,c (p < 0.005) compared with 0.025 lM and 0.05 lM of CuB in A549 cells and 0.05 lM and 0.25 lM of CuB in HCC-827 cells, respectively.

42 H. Kausar et al. / Cancer Letters 332 (2013) 35–45

thiol nucleophile group via Michael-type addition due to the pres-ence of a-b-unsaturated ketone or carbonyl moieties in their struc-ture [2,5]. Since CuB also posses a similar a-b-unsaturated ketonemoiety in its structure, the presence of such functional groupmight render it susceptible to binding with thiols. Although similarspeculations has been made by Whitehouse and Doskotch [47] and

Kupchan and Tsou [28], conclusive evidence for this notion comesfrom our UV and mass spectrometric studies where we demon-strate a direct interaction of CuB with thiol nucleophiles, such asGSH and NAC, through the identification of the respective adducts.

Many drugs and chemicals that form reactive electrophiles alsomodify protein structure by binding to the functional –SH groups

Fig. 6. CuB interacts with NAC and GSH. (a) Absorption spectra of CuB in the presence and absence of NAC in a cell-free system. Arrows point to absorbance of CuB and thedotted lines shows the kmax of CuB in the absence and presence of NAC respectively. (b and c) CuB was incubated with NAC for 10 min and GSH for 30 min at room temprespectively and the mixture was resolved by FTICR mass spectrometry. Masses were recorded in positive ion mode (to save space and for clarity of the results, patternspresented have been truncated). (d) Structure of Cucurbitacin B (shaded region shows the site for nucleophillic attack as proposed by Kupchan and Tsou [28]. Adaptedpartially with permission from Kupchan S.M. and Tsou G. Tumor Inhibitors. LXXXI. Structure and partial synthesis of fabacein. J. Org. Chem. 1973; 38: 1055–1056. Copyright(1973), American Chemical Society.

H. Kausar et al. / Cancer Letters 332 (2013) 35–45 43

44 H. Kausar et al. / Cancer Letters 332 (2013) 35–45

in their cysteine moieties [6]. Additionally, changes in cellular re-dox status can also lead to the modification of protein structure[19,32]. Considering that fact that the vast array of protein targetsof CuB such as JAK2, STAT3, NF-kB, p53, cyclinD1, p21 and cdc2 aresusceptible to electrophile-induced structural modifications[5,23,25,31,35,46], we speculate that CuB’s effects on its variouscellular protein targets could also be facilitated either (i) primarilyby direct interaction of CuB with the functional nucleophilicgroups of these proteins or (ii) secondarily through its effects oncellular GSH which subsequently might lead to the structural andfunctional modification of the client proteins. Even though furtherstudies are warranted to explore these possibilities and identifythe putative protein targets for covalent modification by CuB, nev-ertheless data presented in this paper demonstrate for the firsttime that CuB-induced cytotoxicity are due to alterations in cellu-lar redox balance which, in turn, might be a consequence of CuB’sreactivity with the cellular nucleophiles such as GSH and the pro-tein thiols.

Our study thus demonstrates the therapeutic potential of CuB inNSCLC and advances our understanding into the working mecha-nisms of this potential future drug.

Acknowledgements

This work was supported from Agnes Brown Duggan Endow-ment, James Graham Brown Cancer Center funds and KentuckyLung Cancer Research Program cycles 7 and 10. R.C.G. holds theAgnes Brown Duggan Chair in Oncological Research. Drs. Tariq Ha-mid, Srivani Ravoori and Madhavi Rane are acknowledged for use-ful discussions during the course of the work. Finally, we alsothank Dr. Michelle T. Barati and Dr. Rick Higashi for their help withthe confocal and FTICR mass-spectrometry studies, respectively.

References

[1] J. Alexandre, F. Batteux, C. Nicco, C. Chereau, A. Laurent, L. Guillevin, B. Weill, F.Goldwasser, Accumulation of hydrogen peroxide is an early and crucial step forpaclitaxel-induced cancer cell death both in vitro and in vivo, Int. J. Cancer 119(2006) 41–48.

[2] J. Atsmon, M.L. Freeman, M.J. Meredith, B.J. Sweetman, L.J. Roberts 2nd,Conjugation of 9-deoxy-delta 9, delta 12(E)-prostaglandin D2 withintracellular glutathione and enhancement of its antiproliferative activity byglutathione depletion, Cancer Res. 50 (1990) 1879–1885.

[3] G. Bellomo, F. Mirabelli, Oxidative stress and cytoskeletal alterations, Ann. N. Y.Acad. Sci. 663 (1992) 97–109.

[4] H.A. Burris 3rd, Shortcomings of current therapies for non-small-cell lungcancer: unmet medical needs, Oncogene 28 (Suppl. 1) (2009) S4–13.

[5] E. Butturini, E. Cavalieri, A. Carcereri de Prati, E. Darra, A. Rigo, K. Shoji, N.Murayama, H. Yamazaki, Y. Watanabe, H. Suzuki, S. Mariotto, Two naturallyoccurring terpenes, dehydrocostuslactone and costunolide, decreaseintracellular gsh content and inhibit STAT3 activation, PLoS ONE 6 (2011)e20174.

[6] A. Casini, A. Scozzafava, C.T. Supuran, Cysteine-modifying agents: a possibleapproach for effective anticancer and antiviral drugs, Environ. Health Perspect.110 (Suppl. 5) (2002) 801–806.

[7] J.C. Chen, M.H. Chiu, R.L. Nie, G.A. Cordell, S.X. Qiu, Cucurbitacins andcucurbitane glycosides: structures and biological activities, Nat. Prod. Rep.22 (2005) 386–399.

[8] W. Chen, A. Leiter, D. Yin, M. Meiring, V.J. Louw, H.P. Koeffler, Cucurbitacin Binhibits growth, arrests the cell cycle, and potentiates antiproliferative efficacyof cisplatin in cutaneous squamous cell carcinoma cell lines, Int. J. Oncol. 37(2010) 737–743.

[9] B.Y. Chiang, T.C. Chen, C.H. Pai, C.C. Chou, H.H. Chen, T.P. Ko, W.H. Hsu, C.Y.Chang, W.F. Wu, A.H. Wang, C.H. Lin, Protein S-thiolation byglutathionylspermidine (Gsp): the role of Escherichia coli Gsp synthetASE/amidase in redox regulation, J. Biol. Chem. 285 (2010) 25345–25353.

[10] M.L. Circu, M.P. Moyer, L. Harrison, T.Y. Aw, Contribution of glutathione statusto oxidant-induced mitochondrial DNA damage in colonic epithelial cells, FreeRadic. Biol. Med. 47 (2009) 1190–1198.

[11] S. Clutton, The importance of oxidative stress in apoptosis, Br. Med. Bull. 53(1997) 662–668.

[12] I. Dalle-Donne, R. Rossi, A. Milzani, P. Di Simplicio, R. Colombo, The actincytoskeleton response to oxidants: from small heat shock proteinphosphorylation to changes in the redox state of actin itself, Free Radic. Biol.Med. 31 (2001) 1624–1632.

[13] C. Fiorentini, G. Donelli, P. Nicotera, M. Thelestam, Clostridium difficile toxin Aelicits Ca(2+)-independent cytotoxic effects in cultured normal rat intestinalcrypt cells, Infect. Immun. 61 (1993) 3988–3993.

[14] R. Franco, J.A. Cidlowski, Apoptosis and glutathione: beyond an antioxidant,Cell Death Differ. 16 (2009) 1303–1314.

[15] R. Franco, M.I. Panayiotidis, J.A. Cidlowski, Glutathione depletion is necessaryfor apoptosis in lymphoid cells independent of reactive oxygen speciesformation, J. Biol. Chem. 282 (2007) 30452–30465.

[16] S. Fulda, K.M. Debatin, Apoptosis signaling in tumor therapy, Ann. N. Y. Acad.Sci. 1028 (2004) 150–156.

[17] A. Hall, Rho GTPases and the actin cytoskeleton, Science 279 (1998) 509–514.[18] T. Haritunians, S. Gueller, L. Zhang, R. Badr, D. Yin, H. Xing, M.C. Fung, H.P.

Koeffler, Cucurbitacin B induces differentiation, cell cycle arrest, and actincytoskeletal alterations in myeloid leukemia cells, Leuk. Res. 32 (2008) 1366–1373.

[19] B.G. Hill, A. Bhatnagar, Role of glutathiolation in preservation, restoration andregulation of protein function, IUBMB Life 59 (2007) 21–26.

[20] H.S. Hsu, P.I. Huang, Y.L. Chang, C. Tzao, Y.W. Chen, H.C. Shih, S.C. Hung, Y.C.Chen, L.M. Tseng, S.H. Chiou, Cucurbitacin I inhibits tumorigenic ability andenhances radiochemosensitivity in nonsmall cell lung cancer-derived CD133-positive cells, Cancer 117 (2011) 2970–2985.

[21] B. Jayaprakasam, N.P. Seeram, M.G. Nair, Anticancer and antiinflammatoryactivities of cucurbitacins from Cucurbita andreana, Cancer Lett. 189 (2003)11–16.

[22] H.R. Jin, X. Jin, N.T. Dat, J.J. Lee, Cucurbitacin B suppresses the transactivationactivity of RelA/p65, J. Cell. Biochem. 112 (2011) 1643–1650.

[23] C.C. Juan, F.Y. Wu, Vitamin K3 inhibits growth of human hepatoma HepG2 cellsby decreasing activities of both p34cdc2 kinase and phosphatase, Biochem.Biophys. Res. Commun. 190 (1993) 907–913.

[24] H. Kausar, J. Jeyabalan, F. Aqil, D. Chabba, J. Sidana, I.P. Singh, R.C. Gupta, Berryanthocyanidins synergistically suppresses growth and invasive potential ofhuman non-small-cell lung cancer cells, Cancer Lett. 325 (1) (2012) 54–62.

[25] D.H. Kim, J.K. Kundu, Y.J. Surh, Redox modulation of p53: mechanisms andfunctional significance, Mol. Carcinog. 50 (2011) 222–234.

[26] F.E. Knock, R.M. Galt, Y.T. Oester, Protein-sulfhydryl groups in cellular controlmechanisms and cancer, J. Am. Geriatr. Soc. 15 (1967) 882–899.

[27] S. Kongtun, W. Jiratchariyakul, T. Kummalue, P. Tan-ariya, S. Kunnachak, A.W.Frahm, Cytotoxic properties of root extract and fruit juice of Trichosanthescucumerina, Planta Med. 75 (2009) 839–842.

[28] S.M. Kupchan, G. Tsou, Tumor Inhibitors. LXXXI. Structure and partialsynthesis of fabacein., J. Org. Chem. 38 (1973) 1055–1056.

[29] J.N. Lavoie, G. Gingras-Breton, R.M. Tanguay, J. Landry, Induction of Chinesehamster HSP27 gene expression in mouse cells confers resistance to heatshock. HSP27 stabilization of the microfilament organization, J. Biol. Chem.268 (1993) 3420–3429.

[30] D.H. Lee, G.B. Iwanski, N.H. Thoennissen, Cucurbitacin: ancient compoundshedding new light on cancer treatment, Scientific World J. 10 (2010) 413–418.

[31] J.H. Lee, T.H. Koo, H. Yoon, H.S. Jung, H.Z. Jin, K. Lee, Y.S. Hong, J.J. Lee, Inhibitionof NF-kappa B activation through targeting I kappa B kinase by celastrol, aquinone methide triterpenoid, Biochem. Pharmacol. 72 (2006) 1311–1321.

[32] H. Liu, R. Lightfoot, J.L. Stevens, Activation of heat shock factor by alkylatingagents is triggered by glutathione depletion and oxidation of protein thiols, J.Biol. Chem. 271 (1996) 4805–4812.

[33] P. Melius, C.A. McAuliffe, I. Photaki, M. Sakarellou-Daitsiotou, Interactions ofplatinum complexes, peptides, methionine and dehydrogenases, Bioinorg.Chem. 7 (1977) 203–210.

[34] M. Mendez, A. Custodio, M. Provencio, New molecular targeted therapies foradvanced non-small-cell lung cancer, J. Thorac. Dis. 3 (2011) 30–56.

[35] S.G. Menon, P.C. Goswami, A redox cycle within the cell cycle: ring in the oldwith the new, Oncogene 26 (2007) 1101–1109.

[36] R. Munagala, H. Kausar, C. Munjal, R.C. Gupta, Withaferin a induces p53-dependent apoptosis by repression of HPV oncogenes and upregulation oftumor suppressor proteins in human cervical cancer cells, Carcinogenesis 32(2011) 1697–1705.

[37] A.J. Phimister, K.J. Williams, L.S. Van Winkle, C.G. Plopper, Consequences ofabrupt glutathione depletion in murine Clara cells: ultrastructural andbiochemical investigations into the role of glutathione loss in naphthalenecytotoxicity, J. Pharmacol. Exp. Ther. 314 (2005) 506–513.

[38] J.L. Rios, I. Andujar, J.M. Escandell, R.M. Giner, M.C. Recio, Cucurbitacins asinducers of cell death and a rich source of potential anticancer compounds,Curr. Pharm. Des. 18 (2012) 1663–1676.

[39] N. Scott, K.M. Hatlelid, N.E. MacKenzie, D.E. Carter, Reactions of arsenic(III) andarsenic(V) species with glutathione, Chem. Res. Toxicol. 6 (1993) 102–106.

[40] M.D. Shelton, P.B. Chock, J.J. Mieyal, Glutaredoxin: role in reversible protein s-glutathionylation and regulation of redox signal transduction and proteintranslocation, Antioxid. Redox Signal. 7 (2005) 348–366.

[41] T. Tannin-Spitz, S. Grossman, S. Dovrat, H.E. Gottlieb, M. Bergman, Growthinhibitory activity of cucurbitacin glucosides isolated from Citrulluscolocynthis on human breast cancer cells, Biochem. Pharmacol. 73 (2007)56–67.

[42] N.H. Thoennissen, G.B. Iwanski, N.B. Doan, R. Okamoto, P. Lin, S. Abbassi, J.H.Song, D. Yin, M. Toh, W.D. Xie, J.W. Said, H.P. Koeffler, Cucurbitacin B inducesapoptosis by inhibition of the JAK/STAT pathway and potentiatesantiproliferative effects of gemcitabine on pancreatic cancer cells, CancerRes. 69 (2009) 5876–5884.

H. Kausar et al. / Cancer Letters 332 (2013) 35–45 45

[43] S. Ueda, H. Masutani, H. Nakamura, T. Tanaka, M. Ueno, J. Yodoi, Redox controlof cell death, Antioxid. Redox Signal. 4 (2002) 405–414.

[44] A.L. Vahrmeijer, J.H. van Dierendonck, J. Schutrups, C.J. van de Velde, G.J.Mulder, Effect of glutathione depletion on inhibition of cell cycle progressionand induction of apoptosis by melphalan (L-phenylalanine mustard) in humancolorectal cancer cells, Biochem. Pharmacol. 58 (1999) 655–664.

[45] N. Wakimoto, D. Yin, J. O’Kelly, T. Haritunians, B. Karlan, J. Said, H. Xing, H.P.Koeffler, Cucurbitacin B has a potent antiproliferative effect on breast cancercells in vitro and in vivo, Cancer Sci. 99 (2008) 1793–1797.

[46] Y. Wang, X. Ma, S. Yan, S. Shen, H. Zhu, Y. Gu, H. Wang, G. Qin, Q. Yu, 17-Hydroxy-jolkinolide B inhibits signal transducers and activators oftranscription 3 signaling by covalently cross-linking Janus kinases andinduces apoptosis of human cancer cells, Cancer Res. 69 (2009) 7302–7310.

[47] M.W. Whitehouse, R.W. Doskotch, Selective inhibition of thymidineincorporation into lymphocytes by cucurbitacins B and D, Biochem.Pharmacol. 18 (1969) 1790–1793.

[48] S. Yasuda, S. Yogosawa, Y. Izutani, Y. Nakamura, H. Watanabe, T. Sakai,Cucurbitacin B induces G2 arrest and apoptosis via a reactive oxygen species-dependent mechanism in human colon adenocarcinoma SW480 cells, Mol.Nutr. Food Res. 54 (2010) 559–565.

[49] D. Yin, N. Wakimoto, H. Xing, D. Lu, T. Huynh, X. Wang, K.L. Black, H.P. Koeffler,Cucurbitacin B markedly inhibits growth and rapidly affects the cytoskeletonin glioblastoma multiforme, Int. J. Cancer 123 (2008) 1364–1375.

[50] Y. Zhang, D. Ouyang, L. Xu, Y. Ji, Q. Zha, J. Cai, X. He, Cucurbitacin B inducesrapid depletion of the G-actin pool through reactive oxygen species-dependentactin aggregation in melanoma cells, Acta Biochim. Biophys. Sin. (Shanghai) 43(2011) 556–567.