lable at ScienceDirect

Food and Chemical Toxicology 96 (2016) 35e42

Contents lists avai

Food and Chemical Toxicology

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

The synthetic ajoene analog SPA3015 induces apoptotic cell deaththrough crosstalk between NF-kB and PPARg in multidrug-resistantcancer cells

Jee Won Hwang a, Hyewon Cho a, Jae Yeon Lee a, Youngsic Jeon a, b, Su-Nam Kim b,Sang Jin Lee a, Gyu-Un Bae a, Sungpil Yoon c, Raok Jeon a, Yong Kee Kim a, *

a Research Center for Cell Fate Control, College of Pharmacy, Sookmyung Women's University, Seoul 04310, Republic of Koreab Natural Products Research Center, Korea Institute of Science and Technology Gangneung Institute, Gangneung, Gangwon-do 25451, Republic of Koreac Research Institute, National Cancer Center, Ilsan-gu, Goyang-si, Gyeonggi-do 10408, Republic of Korea

a r t i c l e i n f o

Article history:Received 31 March 2016Received in revised form15 July 2016Accepted 19 July 2016Available online 20 July 2016

Keywords:SPA3015Multidrug resistanceApoptosisNF-kBPPARg

* Corresponding author. Research Center for Cell Fmacy, Sookmyung Women's University, 100 Cheongpa04310, Republic of Korea.

E-mail address: yksnbk@sookmyung.ac.kr (Y.K. Kim

http://dx.doi.org/10.1016/j.fct.2016.07.0200278-6915/© 2016 Published by Elsevier Ltd.

a b s t r a c t

Multidrug resistance (MDR) caused by P-glycoprotein (P-gp) overexpression impedes successful cancerchemotherapy. In this study, we investigated the anticancer effects of SPA3015, a synthetic ajoene analog,in P-gp-overexpressing MDR cancer cells (KBV20C and MES-SA/DX5). Treatment with SPA3015 caused adramatic decrease in the cell viabilities of both KBV20C and MES-SA/DX5 cells. This decrease wasaccompanied by apoptotic cell death without affecting the expression level or drug efflux function of P-gp. SPA3015 selectively suppressed NF-kB reporter gene activity, which led to decreased expression ofNF-kB target genes such as CIAP1, CIAP2, XIAP, and Bcl-XL. Surprisingly, nuclear localization and DNAbinding affinity of the p65 subunit were not affected by SPA3015, suggesting that SPA3015 inhibits thetranscriptional activity of NF-kB at the nucleus. Indeed, SPA3015 treatment led to an increase in thephysical interaction of p65 with PPARg, which resulted in the inhibition of NF-kB activity. Our findingssupport the hypothesis that SPA3015 inhibits NF-kB transcriptional activity by facilitating the physicalinteraction of the p65 subunit and PPARg, which leads to apoptotic cell death in MDR cancer cells.

© 2016 Published by Elsevier Ltd.

1. Introduction

Multidrug resistance (MDR), a significant impediment tochemotherapeutic cancer treatment, results from the elevatedexpression of proteins such as cell-membrane ATP-binding cassettetransporters. These transporters are involved in the efflux of cyto-toxic drugs from cells, thus reducing intracellular drug concentra-tions (Bradley and Ling, 1994; Szakacs et al., 2006). A 170 kDa ATP-dependent P-glycoprotein (P-gp) acts as a major efflux pump; it isclear that the overexpression of P-gp is linked to MDR in manyhuman cancer cells, including colon, kidney, adrenal, pancreas, andliver cancer (Ambudkar et al., 2003; Gottesman et al., 2002; Thomasand Coley, 2003). Abundant evidence links the failure of certainchemotherapeutic agents to the expression of P-gp (Thomas andColey, 2003). Thus, the development of a chemotherapeutic

ate Control, College of Phar--ro 47-gil, Yongsan-gu, Seoul

).

strategy against MDR remains a major challenge in the treatment ofcancer.

Since the first explanation of P-gp function in MDR, strong efforthas been made to use P-gp inhibitors to reverse MDR phenotype invarious cancers (Amin, 2013; Palmeira et al., 2012; Szakacs et al.,2006). First-generation P-gp inhibitors, such as verapamil, qui-nine, and cyclosporine, are less selective for and display lowbinding affinity to P-gp. This results in unwanted toxic side effectsand failure to show improved outcome due to high serum con-centrations of the inhibitor (Amin, 2013; Palmeira et al., 2012). Thedevelopment of second-generation P-gp inhibitors such as dex-verapamil and valspodar focused on improving their specificity andaffinity for inhibiting P-gp, as well as decreasing their toxicity.However, they have been shown to inhibit the P450 CYP3A4enzyme and other transport systems (Amin, 2013; Palmeira et al.,2012). Because of these limitations, the use of first and second-generation inhibitors is limited to clinical trials. The third-generation inhibitors such as zosuquidar and tariquidar, now un-der clinical development, are designed specifically for high

J.W. Hwang et al. / Food and Chemical Toxicology 96 (2016) 35e4236

transporter affinity, low pharmacokinetic interaction, and decreaseof toxicity. Although P-gp is clearly established as a prognosticmarker in several human tumors, the clinical benefit of modulatingP-gp-mediated MDR is still in question. In addition to traditionalpharmacological approaches, more creative searches for fourth-generation inhibitors are in progress, examining diverse mecha-nisms such as capturing, neutralizing, or evading of P-gp, anddownregulation of P-gp gene expression (Szakacs et al., 2006;Thomas and Coley, 2003).

Ajoene (4,5,9-trithiadodeca-1,6,11-triene 9-oxide: supplemen-tary data; sFig. 1A) is one of the major active compounds derivedfrom garlic (Allium sativum) (Apitz-Castro et al., 1983). It has beenreported that ajoene possesses a broad spectrum of medical andbiological activities including anti-thrombosis, antioxidant, anti-microbial, virucidal, and anticancer activity (Apitz-Castro et al.,1992; Dirsch et al., 1998; Kaschula et al., 2010; Naganawa et al.,1996; Weber et al., 1992). In particular, ajoene has anti-proliferative activities in several human cancer cells includingleukemia, esophageal, and basal skin tumor (Kaschula et al., 2012;Li et al., 2002; Scharfenberg et al., 1990; Taylor et al., 2006). On thebasis of this anticancer activity, efforts have been made to developmore active ajoene-based anticancer agents by substituting theterminal allyl groups with alkyl, aromatic, or heteroaromaticgroups (Kaschula et al., 2010). However, the anticancer effects ofajoene and its analogs in MDR cancer cells remain largelyundefined.

In this study, we examined the antiproliferative potential ofnewly synthesized ajoene analogs on P-gp-overexpressing MDRcancer cells. Since SPA3015 [(E)-1-phenyl-7-phenyl-2,3,7-trithiahepta-4-ene 7-oxide: supplementary data; sFig. 1B] showsthe most potent antiproliferative effect, we studied possible mo-lecular mechanisms. SPA3015 induced apoptotic cell death in MDRcancer cells without affecting the expression level and function ofP-gp. These effects seemed to be mediated by suppression of anti-apoptotic genes through the inhibition of NF-kB function. Inter-estingly, we found that SPA3015 inhibits NF-kB transcriptional ac-tivity by enhancing the physical interaction of the p65 subunit andPPARg at the nucleus. Our results suggest a novel mechanism forthe ajoene analog SPA3015 in inhibiting the survival and prolifer-ation of cancer cells and highlight the potential to develop a novelchemotherapeutic drug to reverse MDR phenotype in cancer.

2. Materials and methods

2.1. Cell culture

KB (Human oral squamous cell carcinoma), KBV20C (vincristine-resistant daughter cells), MES-SA (human uterine sarcoma), andMES-SA/DX5 (doxorubicin-resistant daughter cells) were obtainedfrom American Type Culture Collection (ATCC, Manassas, VA). KBand KBV20C cells were grown in RPMI 1640 supplemented with10% heat-inactivated fetal bovine serum (FBS; HyClone Labora-tories, Logan, UT), 100 units/mL penicillin, and 100 mg/mL strepto-mycin (HyClone Laboratories). KBV20C cells were maintained with20 nM vincristine in their growth media. MES-SA and MES-SA/DX5cells were grown in DMEM supplemented with equal amounts ofFBS and penicillin/streptomycin. Cells were maintained at 37 �Cunder 5% CO2 in a humidified chamber.

2.2. Reagents and cytotoxicity test

Synthetic ajoene analogs were obtained from Prof. R. Jeon(Sookmyung Women's University). Doxorubicin and vincristinewere purchased from Sigma-Aldrich (Sigma-Aldrich, St. Louis, MO).All chemicals were dissolved in dimethyl sulfoxide (DMSO, Sigma-

Aldrich) and stored at �80 �C until use. The cytotoxicity test wasdetermined by the MTT assay (DUCHEFA BIOCHEMIE, B.V,Netherlands) according to the manufacturer's instruction.

2.3. Cell cycle analysis

Cells were treatedwith the indicated concentrations of drugs for24 h. Cells were harvested by trypsinization, washed twice withPBS, fixed with 70% ethanol, and stored overnight at 4 �C. Forstaining, cells were washed with PBS, incubated with 100 mg/mLRNase A (BD Bioscience, Franklin Lakes, NJ) at 37 �C for 30 min,stained with 10 mg/mL propidium iodide (PI, BD Bioscience) solu-tion, and analyzed by FACSCalibur (BD Bioscience).

2.4. Immunoblotting

For preparation of whole cell extracts, cells were lysed in RIPAbuffer supplemented with 1� protease inhibitor cocktail (Roche,Basel, Switzerland) and 1� phosphatase inhibitor cocktail (Gen-DEPOT, Barker, TX) for 20 min in an ice bath. Cell lysates werecentrifuged at 16,000 � g for 15 min at 4 �C, and the supernatantswere collected in new tubes. After boiling in Laemmli sample bufferfor 5 min, 30 mg of protein was used for SDS-polyacrylamide gelelectrophoresis. Proteins were then transferred to polyvinylidenedifluoride membranes. The membranes were blocked with 5% skimmilk/0.1% Tween 20/TBS (Tris-buffered saline) for 1 h at roomtemperature and incubated overnight with the appropriate primaryantibody. b-actin (sc-47778), PARP (sc-7150), I-kBa (sc-847), andp65/RelA (sc-372) were purchased from Santa Cruz, Dallas, TX. SP-1(ab13370) was purchased from Abcam, Cambridge, UK. a-tubulin(#2144), cleaved active caspase-3 (#9661), phosphor-p38 (#9211),p38 (#9212), phosphor-ERK (#9101), and ERK (#9102) were pur-chased from Cell Signaling Technology, Danvers, MA. The mem-branes were then washed with 0.1% Tween 20/TBS, incubated withsecondary antibody (Jackson ImmunoResearch Laboratories, WestGrove, PA), and visualized with an enhanced chemiluminescencedetection kit (Advansta, Menlo Park, CA).

2.5. Reverse-transcriptase PCR

Total RNA was extracted using the easy-BLUE™ total RNAextraction kit (iNtRON Biotechnology, Seongnam, Korea), and theintegrity of the RNAwas checked by agarose gel electrophoresis andethidium bromide staining. One microgram of RNA was used as atemplate for each M-MLV reverse transcriptase-mediated PCR re-action. The primer sets for P-gp were 50-CCCATCATTGCAA-TAGCAGG-30 and 50-GTTCAAACTTCTGCTCCTGA-30; the primer setsfor GAPDH were 50-CTCATGACCACAGTCCATGCCATC-30 and 50-CTGCTTCACCACCTTCTTGATGTC-30.

2.6. Quantitative real-time PCR

Total cDNA synthesized by M-MLV reverse transcriptase(Promega, Fitchburg, WI) was used as a template for PCR performedwith the Eco Real-Time PCR System (Illumina, San Diego, CA) usingSensiFAST SYBR® No-ROX Kit (Bioline, London, UK). Reactionparameters were as follows: cDNA synthesis at 37 �C for 60 min,transcriptase inactivation at 95 �C for 5min, PCR cycling at 95 �C for10 s, 58 �C for 30 s, and 72 �C for 20 s for 40 cycles. The expressionlevels of mRNA were normalized to the expression of 28S rRNA asan internal control. The following primers were used for real-timePCR: CIAP1 upstream (50-ACTTGAACAGCTGCTATCCACATC-30) anddownstream (50-GCTAGGATTTTTCTCTGAACTGTC-30); CIAP2 up-stream (50-AGCTGTTGTCAACTTCAGATACCA-30) and downstream(50-TTTCACCAGGTCTCTATTAAAGCC-30); XIAP upstream (50-GAAG-

J.W. Hwang et al. / Food and Chemical Toxicology 96 (2016) 35e42 37

ACCCTTGGGAACAACA-30) and downstream (50-CGCCTTAGCTGCTC-TTCAGT-30); Bcl-xL upstream (50-CCCAGAAAGGATACAGCTGG-30)and downstream (50-GCGATCCGACTCACCAATAC-30); and 28S rRNAupstream (50-AACGAGATTCCCACTGTCCC-30) and downstream (50-CTTCACCGTGCCAGACTAGAG-30).

2.7. Rhodamine 123 assay

KBV20C and MES-SA/DX5 cells were incubated with 10 mMRhodamine 123 (Sigma-Aldrich) for 3 h under normal cultureconditions. Cells were harvested, washed twice with ice-cold PBS,and resuspended in PBS just before use. Untreated cells resus-pended in PBS were used as blanks. Fluorescence intensity ofintracellular Rhodamine 123 was determined by flow cytometry(FACSCalibur, BD Bioscience). Ten thousand cells per sample wereanalyzed.

2.8. Nucleus/cytosol fractionation

Cells were washed twice in PBS, harvested, and pelleted bycentrifugation at 500 � g for 3 min. Cell pellets were resuspendedwith hypotonic buffer A (10 mM HEPES, 10 mM KCl, 1.5 mM MgCl2,0.5 mM dithiothreitol, 1� protease inhibitor cocktail (Roche)). Cellswere pelleted by spinning at 1000 � g for 5 min and lysed in ice-cold 0.5% NP-40 containing buffer A on ice for 10 min. The cyto-plasmic fraction was obtained by spinning the cells for 10 min at900 x g. To obtain the nuclear fraction, the remaining pellet waslysed on ice in lysis buffer B (20 mM HEPES, 420 mM NaCl, 1.5 mMMgCl2, 0.2 mM EDTA, 0.5 mM DTT, 25% glycerol, 1� protease in-hibitor cocktail) for 30 min and vortexed periodically. Nuclearfractions were cleared by centrifugation for 10 min at 16,000 x g.

2.9. Luciferase reporter gene assay

KBV20C cells were transfected with NF-kB-Luc, AP-1-Luc, orPPRE-Luc firefly luciferase reporter constructs plus SV40-Renillaluciferase plasmid using Lipofectamine2000 (Invitrogen, Carlsbad,CA). After incubation for 24 h, cells were treated with SPA3015 foran additional 24 h. The dual luciferase reporter assay kit (Promega)was used for the assay with 10 ml lysate, 100 ml luciferase assayreagent II (Promega), and 100 ml Stop&Glow™ (Promega) accordingto the manufacturer's protocol and quantified in a GloMax® 20/20(Promega). The data are represented as the mean ± SD of threeindependent experiments.

2.10. Co-immunoprecipitation (Co-IP)

Cells were collected and lysed in nonionic lysis buffer supple-mented with protease inhibitors, incubated on ice for 15 min, andcleared by centrifugation at 16,000� g at 4 �C for 15min. The lysate(1000 mg of total protein) was subjected to immunoprecipitationwith the agarose-immobilized antibody (1 mg of anti-PPARg)overnight at 4 �C.

2.11. DNA pull down assay

Nuclear extracts were incubated with 5’-biotinylated double-strand NF-kB binding DNA probes (50-AGT TGA GGG GAC TTT CCCAGG C-30 and 50-GCC TGG GAA AGT CCC CTC AAC T-30, synthesizedby Integrated DNA Technologies, Coralville, IA) for overnight at 4 �C.Protein-DNA complexes were pulled down by streptavidin agaroseresin (Thermo Fisher Scientific, Waltham, MA). Precipitated pro-teins were detected by Western blotting using specific antibodies.

2.12. Statistics

The data are expressed as the mean ± standard deviation (S.D.)of three independent experiments. Differences between the meanvalues in the two groups were analyzed using one-way analysis ofvariance (ANOVA). P < 0.05 was considered statistically significant.

3. Results

3.1. The synthetic ajoene analog SPA3015 inhibits cellularproliferation and induces apoptosis in MDR cancer cells

First, we examined the antiproliferative effects of syntheticajoene analogs on both KBV20C cells and its parental KB cells. All ofthe compounds tested significantly decreased the cell viability inboth KB and KBV20C cells when treated at a concentration of 10 mM(data not shown). Among them, SPA3015 proved the most effectivein decreasing the cell viability with very similar levels of decrease inboth cell lines, suggesting the antiproliferative potential of SPA3015on P-gp-overexpressing MDR cancer cells. We thus selectedSPA3015 to investigate the molecular mechanism by which ajoenecompounds exert antiproliferative effects in MDR cancer cells.

We next confirmed the antiproliferative effects of SPA3015 intwo different MDR cancer cell lines, KBV20C and MED-SA DX5. Asshown in Fig. 1A, SPA3015 suppressed the proliferation of KBV20Ccells with the IC50 at about 3.5 mM, which was similar to the effecton parental KB cells. In addition, similar effects were observed inanother MDR cancer cell line MED-SA DX5 (Fig. 1B). Next, weexamined the effect of SPA3015 on cell cycle progression using flowcytometry. SPA3015 did not affect cell cycle progression in either KBor KBV20C cells (supplementary data; sFig. 2A). However, we foundthat the sub-G1 fraction, which often indicates DNA fragmentation,a marker of apoptotic cell death, was increased by SPA3015 in bothcells (supplementary data; sFig. 2A and 2B). Further, to confirm theapoptotic potential of SPA3015, we analyzed the intracellular levelsof active caspase-3 and PARP cleavage. As shown in Fig. 1C and D,KB and MES-SA cells are more sensitive to doxorubicin thanKBV20C and MES-SA/DX5 cells. However, SPA3015 treatment led tothe activation of caspase-3 and cleavage of PARP at similar levels inboth MDR cancer cells and their parent cells, which correlates withour previous observations that SPA3015 suppresses the prolifera-tion of MDR cancer cells.

3.2. SPA3015 does not influence the expression or drug effluxfunction of P-gp

Because most drugs targeting MDR affect transport function orthe expression level of MDR-related efflux transporters (Szakacset al., 2006), we examined the effect of SPA3015 on P-gp expres-sion inMDR cancer cells. As shown in Fig. 2A and B, SPA3015 did notalter the expression level of P-gp in KBV20C or MES-SA/DX5 cells.Next, we determined whether SPA3015 affects the drug effluxfunction of P-gp. The efflux function of P-gp was analyzed bymeasuring the intracellular accumulation of Rhodamine 123, awell-known fluorescent P-gp substrate, using flow cytometry.Treatment with verapamil, an inhibitor of P-gp, led to a dramaticincrease in the intracellular levels of Rhodamine 123 in both MDRcancer cells lines, whereas SPA3015 did not affect the intracellularaccumulation of Rhodamine 123 (Fig. 2C and D), suggesting thatSPA3015 has no effect on the efflux function of P-gp.

3.3. SPA3015 specifically inhibits NF-kB transcriptional activity atthe nucleus

To investigate the molecular mechanism for the apoptotic

Fig. 1. The synthetic ajoene analog SPA3015 inhibits cellular proliferation and induces apoptosis in MDR cancer cells. (A) KB/KBV20C or (B) MES-SA/MES-SA/DX-5 cells weretreated with serially diluted concentrations of SPA3015 from 30 mM for 48 h. Cell viability was determined using the MTT assay as described in Materials and Methods. Datarepresent the mean ± S.D. of three independent experiments. (C, D) Cells were treated with 1 or 10 mM SPA3015 or 1 mM doxorubicin for 24 h, and the cleavage of PARP and caspase-3 activation were assessed by immunoblot analysis.

Fig. 2. SPA3015 does not influence the expression or drug efflux function of P-gp. After treatment of 1 or 10 mM SPA3015 for 24 h, the expression level of MDR1 in KBV20C (A) orMES-SA/DX5 (B) cells was analyzed using RT-PCR. KBV20C (C) or MES-SA/DX5 (D) cells were pretreated with 10 mM verapamil or 10 mM SPA3015 for 1 h. Following treatment with10 mM Rhodamine 123, the fluorescence intensity of intracellular Rhodamine 123 was measured using flow cytometry.

J.W. Hwang et al. / Food and Chemical Toxicology 96 (2016) 35e4238

Fig. 3. SPA3015 specifically inhibits NF-kB transcriptional activity at the nucleus. (A) HEK293T cells were cotransfected with Renilla-SV-40-luc and firefly-NF-kB-luc plasmids for24 h, then treated with 10 mM SPA3015 combined with 50 nM phorbol-dibutyrate (PDBu) for 4 h. Luciferase activities were measured as described in the Materials and Methodssection. Data represent the mean ± S.D. of three independent experiments. *P < 0.05 vs. PDBU treatment. (B) After 24 h treatment with 10 ng/mL TNF-a combined with 1 or 10 mMSPA3015 in KBV20C cells, the mRNA levels of Bcl-xL, CIAP1, CIAP2, and XIAP were analyzed using quantitative real-time PCR. Data represent the mean ± S.D. of three independentexperiments. *P < 0.05 vs. TNF-a treatment. (C) KBV20C cells were transfected with GFP-RelA and then pre-treated with 10 mM SPA3015 for 1 day. Green fluorescent images werecaptured after treatment with 10 ng/mL TNF-a. (D) Cytosol/nucleus fractionation was performed for KBV20C cells, followed by IB with various antibodies, as indicated. (E) KBV20Ccells were treated with 10 ng/mL TNF-a combined with 10 mM SPA3015 for 30 min. Nuclear extracts were used for DNA pull down assay. (F) KBV20C or MES-SA/DX5 cells weretreated with 10 ng/mL TNF-a combined with 10 mM SPA3015 for the indicated times. Cell lysates were subjected to immunoblotting using the indicated antibodies. (For inter-pretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

J.W. Hwang et al. / Food and Chemical Toxicology 96 (2016) 35e42 39

Fig. 4. SPA3015 inhibits NF-kB through interaction with PPARg. (A) HEK293T cells were cotransfected with Renilla-SV-40-luciferase and firefly-luciferase reporter constructscontaining the PPARg-response element (PPARg-luc) for 24 h, and treated with 1 or 10 mM SPA3015 or 10 mM troglitazone (TRO) for 24 h, followed by luciferase reporter activitymeasurement. Data represent the mean ± S.D. of three independent experiments. *P < 0.05 vs. control. (B) KBV20C cells were transfected with GFP-p65/RelA for 24 h and thentreated with 1 or 10 mM SPA3015 for 6 h. Cell lysates were subjected to IP with control IgG or anti-PPARg antibody, followed by IB with various antibodies, as indicated. (C) HEK293Tcells were cotransfected with Renilla-SV-40-luc and firefly-NF-kB-luc plasmids for 24 h, and pretreated with 10 mM pioglitazone, 10 mM troglitazone, or 5 mM SPA3015 with orwithout addition of 20 mM GW9662 for 1 h, followed by treatment with 50 nM PBDU for 24 h. Luciferase activities were measured as described in Materials and Methods. Datarepresent the mean ± S.D. of three independent experiments. *P < 0.05 vs. PDBU treatment.

J.W. Hwang et al. / Food and Chemical Toxicology 96 (2016) 35e4240

potential of SPA3015, we examined whether SPA3015 affects theNF-kB signaling pathway, which functions in cell survival signaling(Karin et al., 2002). SPA3015 strongly inhibited NF-kB reporter geneactivity (Fig. 3A), which was further confirmed by the dramaticdecrease in the expression levels of NF-kB target genes (e.g., cIAP1,cIAP2, XIAP, and Bcl-xL) associated with cell survival followingSPA3015 treatment (Fig. 3B). However, SPA3015 did not alter AP-1reporter gene activity (supplementary data; sFig. 3), suggestingthat inhibition of NF-kB activity by SPA3015 is a specific response.We next attempted to identify the detailed mechanisms for NF-kBinhibition by SPA3015. First, we examined the effects of SPA3015 onthe nuclear translocation of the p65/RelA subunit. The nucleartranslocation of p65/RelA was measured by two different methods:fluorescence microscopy and nuclear fractionation (Fig. 3C and D).TNF-a treatment led to a dramatic increase in nuclear translocationof the p65/RelA subunit; however, this translocation was notaffected by SPA3015 treatment. In addition, SPA3015 does not affectthe DNA binding activity of NF-kB, which was evidenced by DNApull down assay (Fig. 3E). These results strongly indicate that in-hibition of NF-kB by SPA3015 might not be mediated by disruptingthe IKK/I-kBa cascade. This notionwas further strengthened by theobservation that I-kBa degradation was not affected by SPA3015treatment (Fig. 3F, see lanes 3 and 6). However, de novo synthesis ofI-kBa, which is an NF-kB-dependent process, was significantlyinhibited by SPA3015 (Fig. 3F, see lanes 5 and 8), indicating thatSPA3015 suppresses NF-kB transcriptional activity at the nucleus.

3.4. SPA3015 inhibits NF-kB by interacting with PPARg

To clarify our hypothesis that SPA3015 suppresses NF-kB tran-scriptional activity at the nucleus, we tried to identify anotherelement that influences the activity of NF-kB. Recently, it has beenreported that the transactivation of NF-kB could be regulated byphysical interaction with the PPARg transcription factor in thenucleus (Delerive et al., 1999; Jeon et al., 2014; Shih et al., 2014). Toinvestigate whether the crosstalk between NF-kB and PPARg isinvolved in the inhibition of NF-kB by SPA3015, we first performedthe PPARg reporter gene assay. As shown in Fig. 4A, PPARg trans-activation was significantly increased by SPA3015 treatment. Wenext examined the physical interaction between NF-kB and PPARgin KBV20C cells. Although PPARg slightly interacted with p65/RelA

in resting conditions, this interactionwas dramatically increased bySPA3015 (Fig. 4B), indicating that PPARg activation by SPA3015 fa-cilitates the physical interaction between PPARg and NF-kB, whichleads to an inhibition of NF-kB transactivation.

Further, to confirm that PPARg activation by SPA3015 is requiredfor the suppression of NF-kB transactivation, we examined the ef-fects of standard PPARg agonists, pioglitazone and troglitazone, onNF-kB transactivation. As shown in Fig. 4C, these two PPARg ago-nists strongly suppressed NF-kB reporter gene activity. The sup-pression was significantly reversed by treatment with GW9662, acommon PPARg antagonist, suggesting that the transcriptionalactivity of NF-kB might be regulated by crosstalk with PPARg.Interestingly, the effect of SPA3015 was quite similar to that oftroglitazone and pioglitazone (Fig. 4C). Taken together, these re-sults support our postulation that SPA3015 activation of PPARgenhances the physical interaction with NF-kB, resulting in a sup-pression of NF-kB activity.

4. Discussion

Ajoene, a natural garlic compound, has an attractive anticancereffect in many human cancers, including mammary, bladder,colorectal, hepatic, nasopharyngeal, gastric, prostrate, lung,pancreatic, lymphoma, leukemia, and skin, with an IC50 range of5e41 mM (Kaschula et al., 2010). Several groups have reported thatthe anticancer effects of ajoene are largely attributed to inhibitionof proliferation and induction of apoptosis (Li et al., 2002;Scharfenberg et al., 1990; Taylor et al., 2006). However, neitherthe molecular mechanisms of these anticancer effects nor the po-tency of ajoene inMDR cancer cells is well defined. In this study, ourdata indicate that the ajoene analog SPA3015 induces apoptosis inMDR cancer cells to a similar degree as in the parental cells, sug-gesting that SPA3015 could evade the drug efflux system andconsequently inhibit the proliferation of MDR cancer cells. Inter-estingly, it has been reported that ajoene sensitizes the apoptoticeffect of two chemotherapeutic drugs, cytarabine and fludarabine,in human KG1 (P-gp-expressing drug-resistant myeloid leukemia)cells through Bcl-2 inhibition and caspase-3 activation (Ahmedet al., 2001). Consistent with this study, we observed thatSPA3015 induces apoptosis without any effect on either expressionor efflux function of P-gp. Based on these observations, we propose

J.W. Hwang et al. / Food and Chemical Toxicology 96 (2016) 35e42 41

that ajoene and its analog SPA3015 can be developed as a novelstrategy against MDR in cancer chemotherapy.

Because of the pivotal role of NF-kB in cancer, it is a potentialtarget of chemotherapeutic anticancer drugs (Baud and Karin,2009; Hoesel and Schmid, 2013). Indeed, NF-kB-targeting drugsare increasingly being investigated for the treatment of humancancers (Baud and Karin, 2009). Regulation of the IKK/I-kBa cascadehas been thought to be a key step for controlling NF-kB activity.However, several studies demonstrate that the control of the NF-kBsignaling pathway is more complex than simply IKK-mediatedregulation of the I-kBa/NF-kB interaction. Therefore, it is veryimportant to develop an innovative NF-kB-targeting drug that ismore specific and effective. In this study, we found that SPA3015does not affect the steps of I-kBa degradation in cytosol, NF-kBnuclear translocation, or NF-kB DNA binding activity, whereas denovo synthesis of I-kBa is sustained by SPA3015 treatment. Thesedata suggest that SPA3015 could be a novel type of NF-kB inhibitor,which acts at the nucleus, not in the cytosol, to inhibit NF-kB ac-tivity. Several underlying mechanisms for this notion are of inter-est: the recruitment of other transcriptional suppressors, orinhibition of co-activator. Although we cannot explain its exactmechanism, we suggest that the crosstalk between NF-kB andPPARg might be associated with SPA3015 suppression of NF-kB.

PPARg belongs to the family of nuclear hormone receptors thatmainly governs lipid metabolism and insulin sensitization(Krishnan et al., 2007). The relationship between the over-expression of PPARg and its roles in tumorigenesis in many humancancers have been identified (Dicitore et al., 2013; Kotta-Loizouet al., 2012; Tseng and Tseng, 2012). PPARg activation by specificagonists leads to growth inhibition, apoptosis, and differentiationof tumor cells. Recently, it was reported that the crosstalk betweenNF-kB and PPARg could lead to a suppression of the NF-kB pathway(Delerive et al., 1999; Jeon et al., 2014; Shih et al., 2014). Our studyshows that SPA3015 functions as a PPARg partial agonist and fa-cilitates the physical interaction of NF-kB with PPARg, leading to asuppression of the NF-kB pathway. However, we observed thatthiazolidinedione PPARg agonists (pioglitazone, rosiglitazone ortroglitazone) did not suppress cell viability of KBV20C cells atphysiologically functional concentrations of these agonists, butsuppress the viability at high concentration (supplementary data;sFig. 4). These results suggest that PPARg agonists alone may not beenough to suppress cell growth of cancer cells, and PPARg activa-tion may be associated with, in part, the growth inhibition. Indeed,SPA3015 has no adipogenic potential in despite of its PPARgagonistic effects (data not shown). All these results indicate thatPPARg activation mechanisms by SPA3015 might be different fromthiazolidinediones, and either adipogenesis or NF-kB inhibition byPPARg may be regulated by different and independentmechanisms.

In summary, we have demonstrated that the synthetic ajoeneanalog SPA3015 has pharmacological potential to induce apoptoticcell death of P-gp-overexpressing MDR cancer cells. The anticancereffects of SPA3015 are due to the suppression of antiapoptotic geneexpression by inhibiting NF-kB activity through, in part, crosstalkbetween NF-kB and PPARg. Collectively, we suggest that SPA3015 isa strong candidate for further development as an anticancer drug toreverse MDR phenotype in cancer chemotherapy.

Conflict of interest

We declare that we have no conflict of interest.

Acknowledgements

This study was supported by the National Research Foundation

of Korea (NRF) grant funded by the Korea government (MSIP) (No.2011-0030074).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.fct.2016.07.020.

Transparency document

Transparency document related to this article can be foundonline at http://dx.doi.org/10.1016/j.fct.2016.07.020.

References

Ahmed, N., Laverick, L., Sammons, J., Zhang, H., Maslin, D.J., Hassan, H.T., 2001.Ajoene, a garlic-derived natural compound, enhances chemotherapy-inducedapoptosis in human myeloid leukaemia CD34-positive resistant cells. Anti-cancer Res. 21, 3519e3523.

Ambudkar, S.V., Kimchi-Sarfaty, C., Sauna, Z.E., Gottesman, M.M., 2003. P-glyco-protein: from genomics to mechanism. Oncogene 22, 7468e7485.

Amin, M.L., 2013. P-glycoprotein inhibition for optimal drug delivery. Drug TargetInsights 7, 27e34.

Apitz-Castro, R., Badimon, J.J., Badimon, L., 1992. Effect of ajoene, the major anti-platelet compound from garlic, on platelet thrombus formation. Thromb. Res.68, 145e155.

Apitz-Castro, R., Cabrera, S., Cruz, M.R., Ledezma, E., Jain, M.K., 1983. Effects of garlicextract and of three pure components isolated from it on human platelet ag-gregation, arachidonate metabolism, release reaction and platelet ultrastruc-ture. Thromb. Res. 32, 155e169.

Baud, V., Karin, M., 2009. Is NF-kappaB a good target for cancer therapy? Hopes andpitfalls. Nat. Rev. Drug Discov. 8, 33e40.

Bradley, G., Ling, V., 1994. P-glycoprotein, multidrug resistance and tumor pro-gression. Cancer Metastasis Rev. 13, 223e233.

Delerive, P., De Bosscher, K., Besnard, S., Vanden Berghe, W., Peters, J.M.,Gonzalez, F.J., Fruchart, J.C., Tedgui, A., Haegeman, G., Staels, B., 1999. Peroxi-some proliferator-activated receptor alpha negatively regulates the vascularinflammatory gene response by negative cross-talk with transcription factorsNF-kappaB and AP-1. J. Biol. Chem. 274, 32048e32054.

Dicitore, A., Caraglia, M., Colao, A., Zappavigna, S., Mari, D., Hofland, L.J., Persani, L.,Vitale, G., 2013. Combined treatment with PPAR-gamma agonists in pancreaticcancer: a glimmer of hope for cancer therapy? Curr. Cancer Drug Targets 13,460e471.

Dirsch, V.M., Gerbes, A.L., Vollmar, A.M., 1998. Ajoene, a compound of garlic, inducesapoptosis in human promyeloleukemic cells, accompanied by generation ofreactive oxygen species and activation of nuclear factor kappaB. Mol. Pharma-col. 53, 402e407.

Gottesman, M.M., Fojo, T., Bates, S.E., 2002. Multidrug resistance in cancer: role ofATP-dependent transporters. Nat. Rev. Cancer 2, 48e58.

Hoesel, B., Schmid, J.A., 2013. The complexity of NF-kappaB signaling in inflam-mation and cancer. Mol. Cancer 12, 86.

Jeon, Y., Jung, Y., Kim, M.C., Kwon, H.C., Kang, K.S., Kim, Y.K., Kim, S.N., 2014. Sar-gahydroquinoic acid inhibits TNFalpha-induced AP-1 and NF-kappaB signalingin HaCaT cells through PPARalpha activation. Biochem. Biophys. Res. Commun.450, 1553e1559.

Karin, M., Cao, Y., Greten, F.R., Li, Z.W., 2002. NF-kappaB in cancer: from innocentbystander to major culprit. Nat. Rev. Cancer 2, 301e310.

Kaschula, C.H., Hunter, R., Parker, M.I., 2010. Garlic-derived anticancer agents:structure and biological activity of ajoene. Biofactors 36, 78e85.

Kaschula, C.H., Hunter, R., Stellenboom, N., Caira, M.R., Winks, S., Ogunleye, T.,Richards, P., Cotton, J., Zilbeyaz, K., Wang, Y., Siyo, V., Ngarande, E., Parker, M.I.,2012. Structure-activity studies on the anti-proliferation activity of ajoene an-alogues in WHCO1 oesophageal cancer cells. Eur. J. Med. Chem. 50, 236e254.

Kotta-Loizou, I., Giaginis, C., Theocharis, S., 2012. The role of peroxisomeproliferator-activated receptor-gamma in breast cancer. Anticancer Agents Med.Chem. 12, 1025e1044.

Krishnan, A., Nair, S.A., Pillai, M.R., 2007. Biology of PPAR gamma in cancer: a criticalreview on existing lacunae. Curr. Mol. Med. 7, 532e540.

Li, M., Ciu, J.R., Ye, Y., Min, J.M., Zhang, L.H., Wang, K., Gares, M., Cros, J., Wright, M.,Leung-Tack, J., 2002. Antitumor activity of Z-ajoene, a natural compound pu-rified from garlic: antimitotic and microtubule-interaction properties. Carci-nogenesis 23, 573e579.

Naganawa, R., Iwata, N., Ishikawa, K., Fukuda, H., Fujino, T., Suzuki, A., 1996. Inhi-bition of microbial growth by ajoene, a sulfur-containing compound derivedfrom garlic. Appl. Environ. Microbiol. 62, 4238e4242.

Palmeira, A., Sousa, E., Vasconcelos, M.H., Pinto, M.M., 2012. Three decades of P-gpinhibitors: skimming through several generations and scaffolds. Curr. Med.Chem. 19, 1946e2025.

Scharfenberg, K., Wagner, R., Wagner, K.G., 1990. The cytotoxic effect of ajoene, anatural product from garlic, investigated with different cell lines. Cancer Lett.

J.W. Hwang et al. / Food and Chemical Toxicology 96 (2016) 35e4242

53, 103e108.Shih, C.Y., Lin, I.H., Ding, J.C., Chen, F.C., Chou, T.C., 2014. Antiplatelet activity of

nifedipine is mediated by inhibition of NF-kappaB activation caused byenhancement of PPAR-beta/-gamma activity. Br. J. Pharmacol. 171, 1490e1500.

Szakacs, G., Paterson, J.K., Ludwig, J.A., Booth-Genthe, C., Gottesman, M.M., 2006.Targeting multidrug resistance in cancer. Nat. Rev. Drug Discov. 5, 219e234.

Taylor, P., Noriega, R., Farah, C., Abad, M.J., Arsenak, M., Apitz, R., 2006. Ajoene in-hibits both primary tumor growth and metastasis of B16/BL6 melanoma cells inC57BL/6 mice. Cancer Lett. 239, 298e304.

Thomas, H., Coley, H.M., 2003. Overcoming multidrug resistance in cancer: an up-date on the clinical strategy of inhibiting p-glycoprotein. Cancer Control. 10,159e165.

Tseng, C.H., Tseng, F.H., 2012. Peroxisome proliferator-activated receptor agonistsand bladder cancer: lessons from animal studies. J. Environ. Sci. Health C En-viron. Carcinog. Ecotoxicol. Rev. 30, 368e402.

Weber, N.D., Andersen, D.O., North, J.A., Murray, B.K., Lawson, L.D., Hughes, B.G.,1992. In vitro virucidal effects of Allium sativum (garlic) extract and com-pounds. Planta Med. 58, 417e423.