withanone …...ran gao1, navjot shah1, jung-sun lee2, shashank p. katiyar3, ling li1, eonju oh2,...

Small Molecule Therapeutics

Withanone-Rich Combination of AshwagandhaWithanolidesRestricts Metastasis and Angiogenesis through hnRNP-K

Ran Gao1, Navjot Shah1, Jung-Sun Lee2, Shashank P. Katiyar3, Ling Li1, Eonju Oh2, Durai Sundar3,Chae-Ok Yun2, Renu Wadhwa1, and Sunil C. Kaul1

AbstractAshwagandha is an important herb used in the Indian system of traditional home medicine, Ayurveda.

Alcoholic extract (i-Extract) from its leaves and its component, withanone, were previously shown to possess

anticancer activity. In the present study, we developed a combination of withanone and withaferin A, major

withanolides in the i-Extract, that retained the selective cancer cell killing activity and found that it also has

significant antimigratory, -invasive, and -angiogenic activities, in both in vitro and in vivo assays. Using

bioinformatics and biochemical approaches, we demonstrate that these phytochemicals caused downregula-

tion of migration-promoting proteins hnRNP-K, VEGF, andmetalloproteases and hence are candidate natural

drugs for metastatic cancer therapy. Mol Cancer Ther; 13(12); 2930–40. �2014 AACR.

IntroductionMetastasis is a complex andmultistep process of cancer

cell movement from their primary to a secondary sitewithin the body through bloodstream or the lymph sys-tem. Cancer itself is extremely complex to understand andtreat; metastasis poses additional hurdles in its therapy.Multiple factors that determine metastasis include origin,type, and stage of the cancer and, in turn, determine thetype and outcome of the treatment. The process of metas-tasis requires (i) an acquisition of migratory characteris-tics by tumor cells to leave the primary site, (ii) capacity toproteolytically digest the surrounding connective tissues,(iii) characteristics to enter the lymphatic or blood vesselsthroughwhich they travel to distant sites of the body, and(iv) ability to proliferate at the new site to establish into

tumor. The choice of cancer treatment (surgery, chemo-therapy, radiation therapy, hormone therapy, laser immu-notherapy, or a combination) depends on the type ofprimary cancer, size, and location of themetastasis. How-ever, current options to cure metastatic cancers are verylimited. There is a substantial need to understand thephenomenon of metastasis, uncover the factors that influ-ence cellular migration and adhesion characteristics, andformulate new strategies and reagents for safe and effec-tive cancer treatment.

Recently, there has been renewed interest in herbalmedicines because of the safety and economic issues onone hand and the traditional history of wide use on theother. According to the World Health Organization, 80%of theworld population does not have access to advancedmedical care and hence uses herbs for primary health andtherapeutic purposes. Ashwagandha (Withania somnifera)is often categorized amongst the world’s most renownedherbs. It is classified in GRAS (generally regarded as safe)family as a nontoxic edible herb and is also called Indianginseng andQueen ofAyurveda (Indian traditional homemedicine). The main active constituents of ashwagandhaare alkaloids and steroidal lactones, commonly known aswithanolides. Withaferin A is a major chemical constitu-ent ofW. somnifera. It has been shown topossess antitumorpotentials, induces apoptosis by activation of caspase-3,and inhibits JNK, Akt, pERK, and IL6 signal pathways(1–3). Compared with treatment with either withaferin Aor radiation alone, the combination of both resulted in asignificant enhancement of apoptosis in human renalcancer cells and hence was suggested as an effectiveradiosensitizer in cancer therapy (4). Withaferin A wasshown to induce depolymerization of vimentin (5) andcause reregulationofNotch1 signaling (6).Wehave earliershown that the high dose of withaferin A, but not with-anone, was cytotoxic to normal cells (7). Furthermore,

1Cell Proliferation Research Group and DBT-AIST International Laboratoryfor Advanced Biomedicine, National Institute of Advanced Industrial Sci-ence & Technology (AIST), Tsukuba, Ibaraki, Japan. 2Department of Bio-engineering, College of Engineering, Hanyang University, Seongdong-Gu,Seoul, Korea. 3Department of Biochemical Engineering & Biotechnology,Indian Institute of Technology (IIT) Delhi, Hauz Khas, New Delhi, India.

Note: R. Gao and N. Shah contributed equally to this article.

Current address forR.Gao: Instituteof LaboratoryAnimal Science,ChineseAcademy of Medical Science (CAMS) & Comparative Medicine Center,Peking Union Medical College (PUMC), China; and current address for N.Shah: Department of Pediatrics, Darby Children Research Institute, Med-ical University of South Carolina, Charleston, SC 29425.

Corresponding Authors: Renu Wadhwa, National Institute of AdvancedIndustrial Science & Technology (AIST), Central 4, 1-1-1 Higashi, Tsukuba,Ibaraki, 305 8562, Japan. Phone/Fax: 81-29-861-2900; E-mail:[email protected]; Sunil C. Kaul. Phone: 81-298-61-6713; E-mail:[email protected]; and Chae-Ok Yun, Department of Bioengineering, Col-lege of Engineering, Hanyang University, 222 Wangsimni-ro, Seongdong-gu, Seoul, Korea. Phone: 82-2-2220-0491; Fax: 82-2-2220-4850; E-mail:[email protected]

doi: 10.1158/1535-7163.MCT-14-0324

�2014 American Association for Cancer Research.

MolecularCancer

Therapeutics

Mol Cancer Ther; 13(12) December 20142930

on October 23, 2020. © 2014 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst September 18, 2014; DOI: 10.1158/1535-7163.MCT-14-0324

http://mct.aacrjournals.org/

bioinformatics and experimental data suggested the dif-ferential binding efficacies of withaferin A andwithanoneto cellular targets including mortalin, p53, p21, andNRF2(8). We had earlier reported that the low dose of alcoholicextract of ashwagandha leaves (i-Extract) and its compo-nents, withanone and withaferin A, were nontoxic andinstead induced differentiation in glioma cells (9). Hence,in low dose, they were proposed as useful in differenti-ation-based milder and effective glioma therapy. We alsofound that withanone, when added alongwithwithaferinA, was able to decrease the toxicity of the latter in normalcells (7). On the basis of these findings, we undertook thepresent study to formulate potent combination of with-anone and withaferin A that could have stronger anti-metastatic and antiangiogenic activities than the i-Extract.The combination, rich in withanone, was nontoxic tonormal cells and showed potent antimetastatic and anti-angiogenesis activities in in vitro and in vivo assays. Wedemonstrate that such activities are mediated throughinactivation ofmultifunctional RNA-binding protein, het-erogeneous nuclear ribonucleoprotein K (hnRNP-K), andits downstream effectors, matrix metalloproteinases(MMP), pERK, and VEGF.

Materials and MethodsCell culture and colony-forming assaysHuman glioblastoma (A172 and YKG1), osteosarcoma

(U20S), fibrosarcoma (HT1080), neuroblastoma(IMR32); rat glioblastoma (C6), and mouse immortalfibroblasts (NIH 3T3) were used for this study. Cellswere purchased from JCRB (Japanese Collection ofResearch Bioresources) Cell Bank, National Institute ofBiomedical Innovation, Japan, and were maintained inDMEM (Invitrogen) supplemented with 10% FBS in ahumidified incubator (37�C and 5% CO2). Cells wereused within 10 to 15 passages of the original stocks, andhence no additional authentications were performed.Human umbilical vein endothelial cells (HUVEC) werecultured in M199 medium (Invitrogen) containing 20%FBS, penicillin/streptomycin (100 IU/mL), 3 ng/mLbasic FGF (Upstate Biotechnology), and 5 U/mL hepa-rin. HUVECs were used for experiments in passages 2through 7. Cells (40%–60% confluency) were treatedwith withanone (i-Factor; 5 mg/mL; 10.6 mmol/L), with-aferin A (0.25 mg/mL; 0.531 mmol/L), and their combi-nation WiNA 20-1 for about 48 hours and were har-vested for molecular assays as described below. Forcolony-forming assay, 1,000 cells were plated in 10-cmdish in triplicates. Cells were allowed to expand andmake colonies during the next 2 weeks with regularreplacement of culture medium (either control or phy-tochemical-supplemented, as mentioned) with the freshmedium every alternate day. Colonies were fixed inmethanol:acetone (1:1), stained with Crystal violet,destained with running tap water, and then counted.The plates were scanned using Epson GT-9800F scan-ner. Statistical significance of the data was calculatedfrom 3 independent experiments.

Cell-cycle analysisFor cell-cycle analysis, cells (1 � 105) treated with

indicated drugs for 48 hours were harvestedwith trypsin,washed twicewith PBS, and fixedwith 70% ethanol at 4�Cfor 12 hours. The fixed cells were centrifuged (2,000 rpmfor 10min),washed twicewith coldPBS, and resuspendedin 0.25mLPBS. RNAwas removedbyRNAseA treatment(5 ml of 10 mg/mL RNAse was added to 250 mL of the cellsuspension and incubated at 37�C for 1 hour). The cellsuspension was stained with propidium iodide (PI; 10 ml,1 mg/mL) at 4�C in dark for 30 minutes. The cell-cycleanalysis was done using Guava cell cycle flow cytometer(Millipore) following the manufacturer’s protocol.

Wound scratch assayCell motility was examined using the wound scratch

assay. Cell monolayers were wounded by uniformlyscratching the surface with a needle (20 gauge). Move-ment of the control and treated cells in the scratched areawas seriallymonitoredunder aphase contrastmicroscopewith a 10� phase objective. Migration capacity was cal-culated by measuring the percentage of open area in 6 to10 randomly captured images.

Cell invasion assaysInvasion assays were carried out in Boyden chambers

(pore size of 8 mm: Corning Inc.) usingMatrigel followingthe manufacturer’s instructions. For fluorometric deter-mination of Cell Invasion, QCM Cell invasion assay kit(Millipore) was used. It was also performed using xCEL-Ligence System (Roche) that used dual chamber Matrigelplates equippedwith sensor. Automatic scan and the datawere acquired for 48 hours following the manufacturer’sinstructions.

ImmunoblottingCellswere lysed inRIPA lysis buffer. Theprotein (20mg)

was immunoblotted with anti-phospho Rb, -phosphoERK, -phospho p38 (Cell Signalling); -NCAM (AbCysSA); -MMP2 (Santa Cruz Biotech); and -actin (ChemiconInternational) antibodies by standardWestern blotting asdescribed earlier (7).

ImmunostainingCells were cultured and treated on glass coverslips

placed in 12-well culture dish. The cells were stainedwithanti-VEGF and -hnRNP-K antibodies (Santa Cruz Bio-tech), as described previously (7, 9).

Endothelial cell tube formation assayThe formation of HUVEC capillary-like structures on a

basement membrane matrix was used to assess the anti-angiogenic activity of i-Extract and its constituent phyto-chemicals. The 16-mm diameter tissue culture plates werecoated with 250 mL growth factor–reduced Matrigel (Col-laborative Biomedical Products) at 37�C for 30 minutes.HUVECs were seeded on the Matrigel bed (1.5 � 105

cells per well) and cultured in M199 medium containingeither Avastin or ashwagandha extract or phytochemical

Ashwagandha Inhibits Metastasis and Angiogenesis

www.aacrjournals.org Mol Cancer Ther; 13(12) December 2014 2931




combination in the presence of VEGF165 (10 ng/mL) for20 hours.M199medium containing VEGF165 alone servedas a control. Capillary networks were photographed, andthe area covered by the tube network was quantified byImage-Pro Plus software (Media Cybernetics).

Endothelial cell chemotactic migration assayThe effect of ashwagandha on the chemotactic motility

of HUVECs responding to VEGF165 was assessed usingTranswell migration chambers (Corning Costar) with 6.5-mm diameter polycarbonate filters (8-mmpore size). CellsinM199mediumcontaining 1%FBSwere stimulatedwith10 ng/mL VEGF165 and treated with Avastin or ashwa-gandha regents for 30 minutes at room temperature andthen seeded into the lower wells. HUVECs, incubated for4 hours in M199 medium containing 1% FBS, were har-vested by trypsinization and loaded into the upper wells(1� 105 cells perwell). The chamberwas then incubated at37�C for 4 hours. Chemotaxis was quantified by countingthe migrated cells under an optical microscope (OlympusI X 71; Olympus) in 10 random fields.

VEGF ELISAHuman VEGF-A was quantified in cell supernatants

using the human VEGF Quantikine Immunoassay Kit(R&D Systems), following the manufacturer’s protocol.

In vivo studyNude mice were obtained from Charles River. HT1080

(1 � 106) cells were injected subcutaneously into theabdomen. Intraperitoneal injections ofWiNA (withanone,1mg/kg andwithaferin A, 0.5mg/kg)were startedwhensmall tumor buds were formed in about 7 days. Theseconcentrations were determined on the basis of indepen-dent experiments that involved testing of different ratiosof withaferin A and withanone for intraperitoneal injec-tions. The concentrations higher than 1mg/kgwithanoneand 0.5 mg/kg withaferin A in 100 mL volume of 0.5%DMSO showed precipitation and hence were consideredinappropriate. Injections were continued every alternateday, and themiceweremonitored for tumor size until 3 to4 weeks. For metastasis assay, cells were injected into thetail vein. After 7 days, WiNA injections were performed.Mice were sacrificed and the lungs were examined for thepresence of tumors.

Molecular docking and dynamics simulationsVirtual molecular docking of KH3 domain of hnRNP-K

protein with withaferin A and withanone was executedusing Autodock suite 4.2 (10). KH3 domain of hnRNP-K(KH3_hnRNP-K) protein was downloaded from ProteinData Bank (PDBID: 1ZZI) and PubChemCompound data-base was used to retrieve the structures of withaferin A(PubChem: 265237) and withanone (PubChem: 21679027).Structure files of both the ligandswere prepared formolec-ular docking by defining the number of torsion angles,addition of hydrogen atoms, and conversion into software-specific file format (pdbqt). Similarly, KH3_hnRNP-K pro-

teinwas also prepared by removal of single-strandedDNA(ssDNA), removing bad contacts, addition of hydrogenatoms, removal of needless water molecules, and con-version of file format into pdbqt. ssDNA binding site onKH3_hnRNP-Kwas defined as the site of ligand bindingover KH3_hnRNP-K. First, prepared ligands were vir-tually docked against KH3_hnRNP-K protein blindly.Furthermore, both the ligands were docked at thedefined binding site using Lamarckian Genetic Algo-rithm of Autodock 4.2. Top scoring conformations wereinspected against binding at defined binding site.

The GPU accelerated AmberMolecular Dynamics suitewithAmber ff99SBprotein force fieldwasused toperformall atoms explicit molecular dynamics (MD) simulationsof protein–ligand complexes (http://ambermd.org/#Amber12; refs. 11–13). Protein–ligand complex mole-cules were solvated with TIP3P water model in a cubicperiodic boundary box to generate required systems forMD simulations and systems were neutralized usingappropriate number of counterions. Thedistance betweenboxwall and protein complexwas set to greater than 10 Ato avoid direct interaction with its own periodic image.Neutralized system was then minimized, heated up to300K temperature, and equilibrated until the pressure andenergies of systems were stabilized. Finally, equilibratedsystems were used to run 60 ns long MD simulations.

During the MD simulations, H-bond fluctuations ofligand with protein were calculated using VMD software(14). Molecular interaction diagrams were generatedusing Maestro, version 9.4, Schr€odinger LLC. All simula-tion studies were performed on Intel Core 2 Duo CPU@ 3GHz of HP origin with 1 GB DDR RAM and DELL T3600workstation with 8 GB DDR RAM and NVIDIA GeForceGTX TITAN 6 GB GDDR5 Graphics Card.

ResultsEffect of withanone andwithaferin A combination oncancer cell proliferation in vitro

Alcoholic extract of ashwagandha leaves was earliershown to possess withanone and withaferin A as majorand withanolide A, withanolide D, 12-deoxywitha-stra-monolide as minor withanolides. The ratio of these con-stituents in leaves varies depending on their source andstages. Whereas withanone and withaferin A were foundto be toxic to cancer cell, withanone-rich i-Extract wasnontoxic to normal human cells (7). On the basis of thesefindings, we first generated various combinations ofwith-anone and withaferin A and investigated their effect oncancer and normal cell viability in culture. We found thatthe combination of withanone (5 mg/mL; 10.6 mmol/L)and withaferin A (0.25 mg/mL; 0.53 mmol/L), at a ratio ofwithanone to withaferin A as 20:1, called WiNA 20-1,selectively killed cancer cells. Normal cells remainedunaffected (Fig. 1A). In contrary, the combination ofwith-anone and withaferin A either at 5:1 (WiNA 5-1) or 3:1(WiNA 3-1) were toxic to normal cells also (Fig. 1A).Furthermore, the combination WiNA 20-1 was cytotoxicto a variety of human cancer cells including osteosarcoma

Gao et al.

Mol Cancer Ther; 13(12) December 2014 Molecular Cancer Therapeutics2932




(U20S), breast carcinoma (MCF7), glioblastoma, (A172and YKG1), fibrosarcoma (HT1080), neuroblastoma(IMR32); rat glioblastoma (C6), and mouse immortalfibroblasts (NIH3T3; Fig. 1A and data not shown). There-fore, in the present study, 20:1 was determined as theoptimum ratio of withanone and withaferin A in themixture to selectively kill cancer cells in culture.We first examined the effect of withanone and with-

aferin A and their combination WiNA 20-1 on cancer cellproliferation by colony-forming assay using highly met-astatic human fibrosarcoma, HT1080. As shown in Fig. 1Band C, treatment with WiNA 20-1 caused reduction incolony number and size as compared with either thecontrol or the cells treated with individual purified com-ponents. The effect of WiNA 20-1 was stronger and sta-tistically significant; 35%, 55%, and 70% reduction incolony formation in the presence of withaferin A, with-anone, and WiNA 20-1, respectively (Fig. 1C). The datawere supported by cell-cycle analysis (Fig. 1D). Numberof cells in G2 increased from 51.33% to 67.38%, to 77.08%,and to 85.37% in control, withaferin A, withanone, andWiNA 20-1–treated cells, respectively. Cell number at S-phase decreased from control (27.64%) to withaferin A–treated (14.53%) to withanone-treated (11.06%), and toWiNA 20-1–treated (6.06%) cells.

Effect of withanone andwithaferin A combination onin vitro cancer cell migrationWound scratch assays revealed that the cells treated

with either withanone or withaferin A moved slowly

to the scratched area as compared with the control. Fur-thermore, the migration of cells treated with WiNA 20-1combination was highly reduced (Fig. 2A and B). Real-time measure of cell migration also showed that thetreatment of cells with (i) either Wi-A or Wi-N reducedtheir migration capacity and (ii) the effect of WiNA 20-1combination was more pronounced (Fig. 2C). Further-more, quantitative assays (Fig. 2D and E) revealed thatWiNA20-1–treated cells showed strongest reduction bothin their invasion and migration capacities.

Treatment with withanone and withaferin A limitsmigration, invasion, and angiogenic potential ofhuman endothelial cells

On the basis of the above data, we next investigated theeffect of WiNA 20-1 on migration, invasion, and tubeformation capacity of HUVECs. Cells were treated withVEGF (10 ng/mL) for induction ofmigration on basementmembrane matrix constituted of Matrigel. Avastin(humanized monoclonal antibody that inhibits VEGF-A), an approved inhibitor of angiogenesis, was used asa positive control. As shown in Fig. 3A, whereas VEGFinduced the tube formation, Avastin showed clear inhi-bition. Of note, both i-Extract and WiNA 20-1 stronglyinhibited the tube formation even at lower doses (one fifthof the concentration used for cytotoxicity assays inHT1080 and YKG-1 cells as described in Fig. 1). We foundthat the i-Extract and WiNA 20-1 also limited the migra-tion and invasion capacity of VEGF-stimulated HUVECcells, as shown in Fig. 3B and C. Most interestingly, the

Figure 1. Withanone and withaferinA combination (WiNA 20-1) isselectively toxic to cancer cells.Cell morphology (A), colony-forming efficacy (B andC), and cell-cycle distribution (D) of control andtreated cells showing that thecombination of withanone (Wi-N)and withaferin A (Wi-A) in the ratioof 20:1 was selectively cytotoxic tocancer cells. The combinations inthe ratio of 5:1 and3: 1were toxic tonormal cells. The differencesbetween withaferin A andWiNA 20-1 as well as withanoneand WiNA 20-1 in C and D werestatistically significant (C, asshown; and D, P < 0.05).






efficacy of inhibition of migration and invasion was com-parable with Avastin (50 mg/mL) suggesting that the i-Extract and WiNA 20-1 possess highly potent antimeta-static activities. We next performed VEGF ELISA in con-trol and treated cells to investigate whether this effect wasdue to the direct inhibition of VEGF by i-Extract andWiNA 20-1. As shown in Fig. 3D, we found that thetreatment with i-Extract and WiNA 20-1 both resulted insubstantial downregulation of VEGF in the conditionalmedium. Western blotting and immunostaining of VEGFalso showed strong decrease in WiNA 20-1–treated cells(Fig. 3E and F).

Molecular mechanism of antimigratory activity ofwithanone, withaferin A, and WiNA 20-1: hnRNP-Kas a target

As shown in Fig. 3, antimigratory and antiangiogenicactivities of WiNA 20-1 seemed to be mediated by down-regulation of VEGF, an established player of metastasisand angiogenesis, and is regulated by hnRNP-K (15). Wenext examined the expression of hnRNP-K in control andtreated cells. Two other proteins, mortalin (a member ofHSP70 stress chaperone family) and ezrin [a member ofEzrin-Radixin-Moesin (ERM) family], reported to be asso-ciatedwith cancer cellmetastasis (16–18), were also exam-ined. As shown in Fig. 4, the 3 proteins were downregu-lated in WiNA 20-1–treated cells, suggesting their rela-tionship with decreased migration of cells, as shownin Figs. 2 and 3. Mortalin was earlier shown to be a target

of withaferin A and withanone (8). Examination ofhnRNP-K by immunocytochemistry showed that thenumber of cells with bright nuclear staining was less intreated cells suggesting an inhibition of its transcriptionalactivation function (Fig. 4C and D).

On the basis of the above data, we hypothesized thatwithaferin A and withanone may bind to hnRNP-K andresult in an inhibition of its metastatic signaling, as shownin Fig. 5A. We, therefore, investigated the binding byMDsimulations. The crystal structure of KH3_hnRNP-K wasavailable as DNA–protein complex with ssDNA in PDB.The analysis of ssDNA and KH3_hnRNP-K complexrevealed the binding site of ssDNA/RNA overKH3_hnRNP-K. ssDNA was found to interact with resi-dues Lys22, Asp23, Ala25, Ile29, Lys31, Arg40, Lys47,Ile48, Asp50, Tyr84, and Ser85 of KH3_hnRNP-K protein(Fig. 5B). Before docking to ssDNA/RNA-binding site,withaferin A and withanone were docked randomly overKH3_hnRNP-K protein to identify most preferable bind-ing site of ligands over protein.We found that the bindingsites of both the ligands were coinciding with ssDNA/RNA-binding site of the protein suggesting that bothwithaferin A and withanone have tendency to hinder thebinding of ssDNA/RNA at KH3_hnRNP-K protein.

Docking of withaferin A and withanone specifically atssDNA/RNA-binding site generated the bound confor-mation of ligands within the protein with docking scoresof �8.92 and �9.19. Withaferin A was in contact withresidues Gly26, Ser27, Gly30, Lys31, Gln34, Gln83, and

Figure 2. Withanone and withaferinA combination (WiNA 20-1) hasantimigratory activity. A, woundscratch assay showingmigrationofcontrol and treated HT1080 cells.Open area of the wound revealedthat the treated cells moved slowlyas compared with the control cells.B, quantitation of the open areafrom 3 independent experiments.C, real-time measure of themigration capacity of cells. DandE,quantitative cell invasion assaysshowing strong inhibition by WiNA20-1.

Gao et al.





Figure 3. Withanone and withaferin A combination (WiNA 20-1) has antiangiogenic activity. Tube formation capacity (A), migration (B), and invasion (C) ofcontrol and treated HUVECs. Quantitation from three independent experiments is shown. �, P < 0.05; ��, P < 0.01; ��, P < 0.001 compared with VEGF165.VEGF expression, as determined by ELISA (D), Western blotting (E), and immunostaining (F), is shown. Quantitation from 3 independent experiments wasperformed using Imaging J. Actin was used as an internal control (E). An unpaired Student t test was used to determine the statistical significance of the data.






Ser85 of KH3_hnRNPK via hydrogen bonds (H-bonds)and hydrophobic interactions (Fig. 5C and D). MDof withanone with KH3_hnRNP-K resulted in 2 high-affinity conformations (docking score:�9.1 and�8.9) thatwere docking at minutely different positions within samebinding site. A comparison of withanone of docking score�9.1 (Withanone_c1) and withanone of docking score�8.92 (Withanone_c2) with ssDNA-bound conformationrevealed that Withanone_c1 was binding completely atssDNA-binding site but Withanone_c2 was only partiallyinterfering the binding of ssDNA (data not shown).Because both the conformations were interfering with thebinding of hnRNP-K to ssDNA,we confirmed the bindingcharacteristics of withanone, using its 2 conformations byMD simulations. Significance of binding of withaferin Aand withanone was revealed by the fact that the secondnucleotide Thymine (DT) of ssDNA also interacts withresidues Tyr84 and Ser85. Binding of any of the ligands atthese amino acid residues of KH3_hnRNP-K is expectedto hinder its binding to ssDNA/RNA (Fig. 5C and D). Itmay either decrease the binding affinity of RNA/ssDNAto hnRNP-K ormaymake the interaction of RNA/ssDNAand hnRNP-K completely nonfunctional because of the

involvement of first few nucleotides that play key role inits transcription activation/deactivation function.

Stability of the protein–ligand complex was furtherverified by longMD simulations.WithaferinAwas foundhighly stable at its placeduring 60nsMDsimulationswithlittle or no fluctuation. All the interactions of withaferin Awith KH3_hnRNP-Kwere conserved during and after the60 ns MD simulations and involved the ssDNA-bindingsite of hnRNP-K (Fig. 5C and D). Withanone_c1 showed aslight shift in its binding position within ssDNA-bindingsite, attaining a more stable conformation. StabilizedWithanone_c1was found interactingwith residuesGly26,Ser27, Ile29, Gly30, Lys31, Arg35, Ser84, Tyr85, and Lys87ofKH3_hnRNP-Kprotein, strongly hindering the bindingof the protein to ssDNA (Fig. 5C and D). Withanone_c2was also found stable at its binding site during MDsimulation. On the basis of higher binding efficacy ofWithanone_c1 and binding site on the ssDNA, Withano-ne_c1 conformation appeared as a highly efficacy ligand(Fig. 5D). hnRNP-K was earlier shown to regulateErkp44/42 and MMP2 signaling (19). Western blottingrevealed significant decrease in the expression of MMPsand phospho-Erkp44 in WiNA 20-1–treated cells, as

Figure 4. Treatment with withanone and withaferin A combination (WiNA 20-1) downregulates the level of ezrin, hnRNP-K, and mortalin. A, Western blottingshowing decrease in ezrin, hnRNP-K, andmortalin in the treated cells. Actin was used as an internal control. B, quantitation of the signals onWestern blottingwas performed from 3 independent experiments using Imaging J. C, immunostaining showing decrease in the number of cells with bright nuclear hnRNP-K intreated cells (D). An unpaired Student t test was used to determine the statistical significance of the data.

Gao et al.





Figure 5. Withanone and withaferin A target hnRNP-K. A, schematic diagram showing the interaction of hnRNP-K with DNA and its downstream effectorsinvolved in cancer cell migration. Model also shows the abrogation of hnRNP-K and DNA complex and inhibition of cell migration by withaferin (WA) orwithanone (WN). B, interaction diagram of KH3 domain of hnRNP-K protein with ssDNA is shown. Amino acid residues of KH3 domain of hnRNP-K protein,such as Gly26, Ser27, Gly30, Ile36, Lys37, Tyr84, and Ser85, were seen to interact with ssDNA. C, interactions of withaferin A with amino acid residues ofKH3_hnRNP-Kprotein are shown (left). Interactions of stabilizedwithanone after 60 nsMDsimulationwithKH3_hnRNP-Kprotein are shown (right). In both thecases, most of the residues were the ones involved in interaction of the protein with ssDNA. D, binding conformations of withaferin A and withanonewith KH3_hnRNP-K are shown. Hindrance in interaction of hnRNP-K with ssDNA is shown in the superimposed structures of withanone–KH3_hnRNP-Kcomplex over ssDNA-KH3_hnRNP-K complex. E, Western blotting showing decrease in the level of phospho-ERKp44/42 andMMP2 and increase in the celladhesion protein NCAM in WiNA 20-1–treated HT1080.






compared with the ones treated with either withanone orwithaferinA (Fig. 5E).On theother hand, the cell adhesionprotein, NCAM, showedmaximum increase inWiNA 20-1–treated cells (Fig. 5E)).

In vivo validation of antimetastatic activity of WiNA20-1 and hnRNP-K as a target

We next determined the effect of withaferin A, with-anone, and their combination on cancer cell proliferationand migration in in vivo using nude mice HT1080 subcu-taneous xenograft and lungmetastasismodels. Toxicity asa result of intraperitoneal injections of either withanone(1mg/kg),withaferinA (0.5mg/kg), or their combination(WiNA) in 100 mL of 0.5%DMSOwas first tested by visualobservations andbodyweightmeasurements ofmice. Thecombinationwith constituents higher than 1mg/kgwith-anone and 0.5 mg/kg withaferin A in 100 mL of injectionvolume showed precipitation and hence was consideredinappropriate for in vivo study. There was no significantdifference in the bodyweight of withaferin A, withanone,andWiNA-injectedmice as comparedwith the uninjectedcontrol (data not shown). Hence, the 3 reagents, at thedoses used, were considered nontoxic in vivo. Mice withsmall HT1080 tumor buds (7 days postinjection of cells)were given the intraperitoneal injections of either with-aferin A, withanone, or combination on every alternateday. As shown in Fig. 6A and B, WiNA-injected miceshowed strong suppression of subcutaneous HT1080tumor xenografts. In the lung metastasis model, bigtumorsweredetected only in controlmice (Fig. 6C). Taken

together, these data suggested that WiNA has significantanticancer and antimetastatic activities in vivo. We alsoperformed Western blotting of tumor excised from thecontrol and treated mice and found that hnRNP-K wasdecreased in WiNA-treated small tumors as comparedwith the control big tumors. Furthermore, in agreementwith the in vitro data (Figs. 4 and 5), small tumors showeddecrease in hnRNP-K downstream effectors VEGF,Erkp44/42, and MMP2.

DiscussionTumor metastasis involves dissociation of cancer cells

from the primary tumor site, followed by migration,invasion, adhesion, and proliferation at a distant site.MMPs, critical regulators of extracellular matrix, metas-tasis, and angiogenesis (20), are regulated by cytokines,growth factors via interlinked signaling pathways(19, 21, 22). hnRNP-K is a multifunctional protein thatregulates ERK1/2, MMPs, and VEGF and contributes tocell migration, invasion, and ascites formation (19, 23–25).ERK1/2 and VEGF have been connected by positiveautocrine feedback loop (26–28).

Several studies have shown that the alcoholic extract ofashwagandha leaves, and its components, withaferin Aand withanone, are cytotoxic to cancer cells and possessradiosensitizing, immunomodulatory, anti-inflammato-ry, antimetastatic, andantiangiogenic properties, suggest-ing their potential as anticancer drugs (1–6, 29–31).Mechanisms of these activities are only beginning to be

Figure 6. In vivo antimetastaticactivity of withaferin A, withanone,and WiNA 20-1. A, nude miceshowing the suppression of tumorgrowth in response to WiNA 20-1treatment. B, quantitation of tumorvolume from 6 mice each. C,suppression of lung tumors inWiNA 20-1–injected mice ascompared with control. D, Westernblotting showing decrease in thelevel of hnRNP-K, VEGF, ERKp44/42, and MMP2 expression in smalltumors excised from the treatedmice as compared with the largetumors excised from the controlmice.

Gao et al.





resolved. It was shown that withaferin A inhibits prome-tastatic intermediate filament protein—vimentin, an epi-thelial-to-mesenchymal transition (EMT) signaling pro-tein (5, 32), pAkt signaling pathway and MMP9 (33),STAT3 and its downstream effectors Bcl-xL, Bcl-2, cyclinD1 (34), pAkt and pERK signaling (2), oncogenic tran-scription factor STAT3 (35), andNF-kB (36). It was shownto induce extracellular proapoptotic tumor suppressorprotein, Par-4 (2), oxidative stress to cancer cells(37–39). Despite of these beneficial anticancer effects,cytotoxicity of withaferin A, in high dose, to normalhuman cells has been a concern (7, 8, 40). A closely relatedwithanolide, withanone, on the other hand, caused selec-tive cytotoxicity to human cancer cells (7, 40). In view ofthis, we investigated the cytotoxicity of withanone andwithaferin A in various combinations in normal andcancer cells. The combination WiNA 20-1 (withanone,10.6 mmol/L and withaferin A, 0.53 mmol/L) was selec-tively toxic to cancer cells and showed potent antimigra-tory and antiangiogenic activities in vitro. These activitieswere supported by molecular analysis of marker proteinsincluding MMPs that play crucial role in the process ofcancer invasionandmetastasis. In anearlier study,wehadreported that compared with withanone, withaferin Apossess stronger affinity for target proteins includingmortalin, p53, p21, and Nrf2 and that might account forits toxicity to normal cells (8). In the present study, wedemonstrate that withaferin A targets hnRNP-K, anupstreamregulator ofMMPs, Erk-p44/42, andVEGF (Fig.5A), and inhibits its metastasis signaling. Immunofluo-rescent analysis revealed decrease in the number of cellswith bright nuclear hnRNP-K, suggesting its compro-mised transcriptional function resulting in decreasedlevels of MMPs, ERK, VEGF, as shown in Figs. 3–6. Thelatter might also be due to, at least in part, decreased levelof ezrin and mortalin proteins that are enriched in cancercells (refs. 16, 41–46; Fig. 5A and B), associatedwith tumormetastasis as discussed above.Whereas themechanism(s)of effect of these phytochemicals on ezrin warrant furtherstudies, mortalin/p53 complex has been shown to betargeted by withaferin A and withanone resulting inactivation of p53 function (41). Furthermore, withanone

was also shown to target TPX2 oncogene, a prime regu-lator of Aurora A kinase that plays a critical role duringmitosis and cytokinesis (47). These mechanisms are alsoexpected to contribute to the anticancer andantimetastaticactivities of withaferin A, withanone, and WiNA 20-1. Inin vivo tumor formation assays, injections of withanone-rich combination ofwithanone andwithaferinA (1 and0.5mg/kg body weight, respectively) showed significantinhibition of tumor growth andmetastasis. Taken togeth-er, we report that (i) withanone-rich combination of with-anone and withaferin A limits cancer cell growth, migra-tion, and angiogenesis in vitro and in vivo and (ii) theantimetastatic activity ismediatedby targetingmultifunc-tional RNA-binding protein, hnRNP-K.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: D. Sundar, C.-O. Yun, R. Wadhwa, S.C. KaulDevelopment of methodology: R. Gao, N. Shah, D. Sundar, R. WadhwaAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): N. Shah, J.-S. Lee, L. Li, E. Oh, S.C. KaulAnalysis and interpretation of data (e.g., statistical analysis, biostatis-tics, computational analysis): R. Gao, N. Shah, S.P. Katiyar, L. Li, E. Oh,D. Sundar, S.C. KaulWriting, review, and/or revision of the manuscript: R. Gao, N. Shah,S.P. Katiyar, D. Sundar, R. Wadhwa, S.C. KaulAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): S.P. Katiyar, R. WadhwaStudy supervision: D. Sundar, R. Wadhwa, S.C. Kaul

AcknowledgmentsThe authors thank Roche for providing the xCELLigence System and

Masumi Maruyama for technical help.

Grant SupportThis work was partly supported by grants from theMinistry of Knowl-

edge Economy, Korea (10030051) and the Korea Science and EngineeringFoundation (2013K1A1A2A02050188, 2013M3A9D3045879, 2010-0029220)to C.-O. Yun andAIST, Japan Special Budget to S. C. Kaul and R.Wadhwa.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received April 14, 2014; revised August 11, 2014; accepted August 22,2014; published OnlineFirst September 18, 2014.

References1. Oh JH, Lee TJ, Kim SH, Choi YH, Lee SH, Lee JM, et al. Induction of

apoptosis by withaferin A in human leukemia U937 cells throughdown-regulation of Akt phosphorylation. Apoptosis 2008;13:1494–504.

2. Rah B, Amin H, Yousuf K, Khan S, Jamwal G, Mukherjee D, et al. Anovel MMP-2 inhibitor 3-azidowithaferin A (3-azidoWA) abrogatescancer cell invasion and angiogenesis by modulating extracellularPar-4. PLoS One 2012;7:e44039.

3. Ndlovu MN, Van Lint C, Van Wesemael K, Callebert P, Chalbos D,Haegeman G, et al. Hyperactivated NF-{kappa}B and AP-1 transcrip-tion factors promote highly accessible chromatin and constitutivetranscription across the interleukin-6 gene promoter in metastaticbreast cancer cells. Mol Cell Biol 2009;29:5488–504.

4. Yang ES, Choi MJ, Kim JH, Choi KS, Kwon TK. Withaferin A enhancesradiation-induced apoptosis in Caki cells through induction of reactive

oxygen species, Bcl-2 downregulation and Akt inhibition. Chem BiolInteract 2011;190:9–15.

5. Thaiparambil JT, Bender L, Ganesh T, Kline E, Patel P, Liu Y, et al.Withaferin A inhibits breast cancer invasion and metastasis at sub-cytotoxic doses by inducing vimentin disassembly and serine 56phosphorylation. Int J Cancer 2011;129:2744–55.

6. Lee J, Sehrawat A, Singh SV.Withaferin A causes activation of Notch2and Notch4 in human breast cancer cells. Breast Cancer Res Treat2012;136:45–56.

7. Widodo N, Kaur K, Shrestha BG, Takagi Y, Ishii T, Wadhwa R, et al.Selective killing of cancer cells by leaf extract of Ashwagandha:identification of a tumor-inhibitory factor and the first molecularinsights to its effect. Clin Cancer Res 2007;13:2298–306.

8. Vaishnavi K, Saxena N, Shah N, Singh R, Manjunath K, UthayakumarM, et al. Differential activities of the two closely related withanolides,






withaferin a and withanone: bioinformatics and experimental evi-dences. PLoS One 2012;7:e44419.

9. Shah N, Kataria H, Kaul SC, Ishii T, Kaur G, Wadhwa R. Effect of thealcoholic extract of Ashwagandha leaves and its components onproliferation, migration, and differentiation of glioblastoma cells: com-binational approach for enhanced differentiation. Cancer Sci2009;100:1740–7.

10. Morris GM,Huey R, LindstromW, SannerMF, BelewRK, Goodsell DS,et al. AutoDock4 and AutoDockTools4: automated docking withselective receptor flexibility. J Comput Chem 2009;30:2785–91.

11. Case DA, Darden TA, Cheatham TE III, Simmerling CL, Wang J, DukeRE, et al. AMBER12. San Francisco, CA:University of California; 2012.

12. Hornak V, Abel R, Okur A, Strockbine B, Roitberg A, Simmerling C.Comparison of multiple Amber force fields and development ofimproved protein backbone parameters. Proteins 2006;65:712–25.

13. Gotz AW, Williamson MJ, Xu D, Poole D, Le Grand S, Walker RC.Routinemicrosecondmolecular dynamics simulationswithAMBERonGPUs. 1. Generalized Born. J Chem Theory Comput 2012;8:1542–55.

14. HumphreyW, Dalke A, Schulten K. VMD: visual molecular dynamics. JMol Graph 1996;14:33–8, 27–8.

15. Sataranatarajan K, Lee MJ, Mariappan MM, Feliers D. PKCdeltaregulates the stimulation of vascular endothelial factor mRNA trans-lation by angiotensin II through hnRNP K. Cell Signal 2008;20:969–77.

16. WadhwaR, TakanoS, KaurK,Deocaris CC,Pereira-SmithOM,ReddelRR, et al. Upregulation of mortalin/mthsp70/Grp75 contributes tohuman carcinogenesis. Int J Cancer 2006;118:2973–80.

17. Konstantinovsky S, Davidson B, Reich R. Ezrin and BCAR1/p130Casmediate breast cancer growth as 3-D spheroids. Clin Exp Metastasis2012;29:527–40.

18. Li J, Tu Y, Wen J, Yao F, Wei W, Sun S. Role for ezrin in breast cancercell chemotaxis to CCL5. Oncol Rep 2010;24:965–71.

19. Gao R, Yu Y, Inoue A,Widodo N, Kaul SC,Wadhwa R. Heterogeneousnuclear ribonucleoprotein K (hnRNP-K) promotes tumormetastasis byinduction of genes involved in extracellular matrix, cell movement, andangiogenesis. J Biol Chem 2013;288:15046–56.

20. Westermarck J, Kahari VM. Regulation of matrix metalloproteinaseexpression in tumor invasion. FASEB J 1999;13:781–92.

21. Coussens LM, Fingleton B, Matrisian LM. Matrix metalloproteinaseinhibitors and cancer: trials and tribulations. Science 2002;295:2387–92.

22. Bachmeier BE, Iancu CM, Jochum M, Nerlich AG. Matrix metallopro-teinases in cancer: comparison of known and novel aspects of theirinhibition as a therapeutic approach. Expert Rev Anticancer Ther2005;5:149–63.

23. Uribe DJ, Guo K, Shin YJ, Sun D. Heterogeneous nuclear ribonucleo-protein K and nucleolin as transcriptional activators of the vascularendothelial growth factor promoter through interactionwith secondaryDNA structures. Biochemistry 2011;50:3796–806.

24. Salven P, Ruotsalainen T, Mattson K, Joensuu H. High pre-treatmentserum level of vascular endothelial growth factor (VEGF) is associatedwith poor outcome in small-cell lung cancer. Int J Cancer 1998;79:144–6.

25. Wang FQ, So J, Reierstad S, Fishman DA. Vascular endothelial growthfactor-regulated ovarian cancer invasion and migration involvesexpression and activation of matrix metalloproteinases. Int J Cancer2006;118:879–88.

26. Lee CC, Chen SC, Tsai SC, Wang BW, Liu YC, Lee HM, et al.Hyperbaric oxygen induces VEGF expression through ERK, JNK andc-Jun/AP-1 activation in human umbilical vein endothelial cells.J Biomed Sci 2006;13:143–56.

27. Narasimhan P, Liu J, Song YS, Massengale JL, Chan PH. VEGFStimulates the ERK 1/2 signaling pathway and apoptosis in cerebralendothelial cells after ischemic conditions. Stroke 2009;40:1467–73.

28. Belotti D, Paganoni P,Manenti L, Garofalo A,Marchini S, Taraboletti G,et al. Matrix metalloproteinases (MMP9 andMMP2) induce the release

of vascular endothelial growth factor (VEGF) by ovarian carcinomacells: implications for ascites formation. Cancer Res 2003;63:5224–9.

29. Sinha P, Ostrand-Rosenberg S.Myeloid-derived suppressor cell func-tion is reduced by Withaferin A, a potent and abundant component ofWithania somnifera root extract. Cancer Immunol Immunother2013;62:1663–73.

30. Roy RV, Suman S, Das TP, Luevano JE, Damodaran C. Withaferin A, asteroidal lactone fromWithania somnifera, inducesmitotic catastropheand growth arrest in prostate cancer cells. J Nat Prod 2013;76:1909–15.

31. Hahm ER, Singh SV. Withaferin A-induced apoptosis in human breastcancer cells is associated with suppression of inhibitor of apoptosisfamily protein expression. Cancer Lett 2012;334:101–8.

32. Grin B,Mahammad S,Wedig T, ClelandMM, Tsai L, HerrmannH, et al.Withaferin a alters intermediate filament organization, cell shape andbehavior. PLoS One 2012;7:e39065.

33. Lee DH, Lim IH, Sung EG, Kim JY, Song IH, Park YK, et al. Withaferin Ainhibits matrix metalloproteinase-9 activity by suppressing the Aktsignaling pathway. Oncol Rep 2013;30:933–8.

34. Um HJ, Min KJ, Kim DE, Kwon TK. Withaferin A inhibits JAK/STAT3signaling and induces apoptosis of human renal carcinoma Caki cells.Biochem Biophys Res Commun 2012;427:24–9.

35. Lee J, Hahm ER, Singh SV. Withaferin A inhibits activation of signaltransducer and activator of transcription 3 in human breast cancercells. Carcinogenesis 2010;31:1991–8.

36. Mohan R, Hammers HJ, Bargagna-Mohan P, Zhan XH, Herbstritt CJ,Ruiz A, et al. Withaferin A is a potent inhibitor of angiogenesis.Angiogenesis 2004;7:115–22.

37. Yang Z, Garcia A, Xu S, Powell DR, Vertino PM, Singh S, et al.Withaniasomnifera root extract inhibits mammary cancer metastasis and epi-thelial to mesenchymal transition. PLoS One 2013;8:e75069.

38. Hahm ER, Moura MB, Kelley EE, Van Houten B, Shiva S, Singh SV.Withaferin A-induced apoptosis in human breast cancer cells is medi-ated by reactive oxygen species. PLoS One 2011;6:e23354.

39. WidodoN,PriyandokoD,ShahN,WadhwaR,Kaul SC.Selective killingof cancer cells by Ashwagandha leaf extract and its componentWithanone involves ROS signaling. PLoS One 2010;5:e13536.

40. Widodo N, Takagi Y, Shrestha BG, Ishii T, Kaul SC, Wadhwa R.Selective killing of cancer cells by leaf extract of Ashwagandha:components, activity and pathway analyses. Cancer Lett 2008;262:37–47.

41. Grover A, Priyandoko D, Gao R, Shandilya A, Widodo N, Bisaria VS,et al. Withanone binds to mortalin and abrogates mortalin-p53 com-plex: computational and experimental evidence. Int J Biochem CellBiol 2011;44:496–504.

42. Yu Y, Khan J, Khanna C, Helman L, Meltzer PS, Merlino G. Expressionprofiling identifies the cytoskeletal organizer ezrin and the develop-mental homeoprotein Six-1 as key metastatic regulators. Nat Med2004;10:175–81.

43. Wei YC, Li CF, Yu SC, Chou FF, Fang FM, Eng HL, et al. Ezrinoverexpression in gastrointestinal stromal tumors: an independentadverse prognosticator associated with the non-gastric location. ModPathol 2009;22:1351–60.

44. Meng Y, Lu Z, Yu S, Zhang Q, Ma Y, Chen J. Ezrin promotes invasionand metastasis of pancreatic cancer cells. J Transl Med 2010;8:61.

45. Deocaris CC, Lu WJ, Kaul SC, Wadhwa R. Druggability of mortalin forcancer and neuro-degenerative disorders. Curr Pharm Des 2013;19:418–29.

46. LuWJ, LeeNP, Kaul SC, Lan F, Poon RT,WadhwaR, et al. Induction ofmutant p53-dependent apoptosis in human hepatocellular carcinomaby targeting stress protein mortalin. Int J Cancer 2011;129:1806–14.

47. Grover A, Singh R, Shandilya A, Priyandoko D, Agrawal V, Bisaria VS,et al. Ashwagandha derived withanone targets TPX2-Aurora A com-plex: computational and experimental evidence to its anticancer activ-ity. PLoS One 2012;7:e30890.


Gao et al.




2014;13:2930-2940. Published OnlineFirst September 18, 2014.Mol Cancer Ther Ran Gao, Navjot Shah, Jung-Sun Lee, et al. Restricts Metastasis and Angiogenesis through hnRNP-KWithanone-Rich Combination of Ashwagandha Withanolides

Updated version

10.1158/1535-7163.MCT-14-0324doi:

Access the most recent version of this article at:

Cited articles

http://mct.aacrjournals.org/content/13/12/2930.full#ref-list-1

This article cites 46 articles, 6 of which you can access for free at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/13/12/2930To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-14-0324

http://mct.aacrjournals.org/content/13/12/2930.full#ref-list-1

http://mct.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mct.aacrjournals.org/content/13/12/2930


withanone …...ran gao1, navjot shah1, jung-sun lee2, shashank p. katiyar3, ling li1, eonju oh2,...

Documents