European Journal of Pharmacology 650 (2011) 64–71

Contents lists available at ScienceDirect

European Journal of Pharmacology

j ourna l homepage: www.e lsev ie r.com/ locate /e jphar

Molecular and Cellular Pharmacology

Novel hexahydrocannabinol analogs as potential anti-cancer agents inhibit cellproliferation and tumor angiogenesis

Dinesh Thapa a, Jong Suk Lee a, Se-Woong Heo a, Yong Rok Lee b,⁎, Keon Wook Kang c, Mi-Kyoung Kwak a,Han Gon Choi a,1, Jung-Ae Kim a,⁎a College of Pharmacy, Yeungnam University, Gyeongsan 712-749, South Koreab School of Chemical Engineering and Technology, Yeungnam University, Gyeongsan 712-749, South Koreac College of Pharmacy, Chosun University, Gwangju 501-759, South Korea

⁎ Corresponding authors. Tel.: +82 53 810 2816; faxE-mail addresses: yrlee@yu.ac.kr (Y.R. Lee), jakim@y

1 Current address: College of Pharmacy, Hanyang Univ426-791, South Korea.

0014-2999/$ – see front matter © 2010 Elsevier B.V. Adoi:10.1016/j.ejphar.2010.09.073

a b s t r a c t

a r t i c l e i n f o

Article history:Received 25 May 2010Received in revised form 16 September 2010Accepted 23 September 2010Available online 13 October 2010

Keywords:Synthetic hexahydrocannabinolAnti-angiogenesisCell proliferationNF-κBTamoxifen-resistant MCF-7 breast cancer

Both natural and synthetic cannabinoids have been shown to suppress the growth of tumor cells in culture and inanimal models by affecting key signaling pathways including angiogenesis, a pivotal step in tumor growth,invasion, and metastasis. In our search for cannabinoid-like anticancer agents devoid of psychoactive side effects,we synthesized and evaluated the anti-angiogenic effects of a novel series of hexahydrocannabinol analogs.Among these, two analogs LYR-7 [(9S)-3,6,6,9-tetramethyl-6a,7,8,9,10,10a-hexahydro-6H-benzo[c]chromen-1-ol]and LYR-8 [(1-((9S)-1-hydroxy-6,6,9-trimethyl-6a,7,8,9,10,10a-hexahydro-6H-benzo[c]chromen-2-yl)etha-none)] were selected based on their anti-angiogenic activity and lack of binding affinity for cannabinoid receptors.Both LYR-7 and LYR-8 inhibited VEGF-induced proliferation, migration, and capillary-like tube formation ofHUVECs in a concentration-dependent manner. The inhibitory effect of the compounds on cell proliferation wasmore selective in endothelial cells than in breast cancer cells (MCF-7 and tamoxifen-resistant MCF-7). We alsonoted effective inhibition of VEGF-induced new blood vessel formation by the compounds in the in vivo chickchorioallantoic membrane (CAM) assay. Furthermore, both LYR analogs potently inhibited VEGF production andNF-κB transcriptional activity in cancer cells. Additionally, LYR-7 or LYR-8 strongly inhibited breast cancer cell-inducedangiogenesis and tumorgrowth. Together, these results suggest that novel synthetic hexahydrocannabinolanalogs, LYR-7 and LYR-8, inhibit tumor growth by targeting VEGF-mediated angiogenesis signaling in endothelialcells and suppressing VEGF production and cancer cell growth.

: +82 53 810 4654.u.ac.kr (J.-A. Kim).ersity, 1271, Sa-3-Dong, Ansan

ll rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Cannabinoids exert anti-proliferative actions on a wide spectrumof tumor cells in culture and in animal models by inducing directgrowth arrest and death of tumor cells and by inhibiting tumorangiogenesis and metastasis (Galve-Roperh et al., 2000; Guzmanet al., 2002; Guzman, 2003). However, the potential development ofcannabinoids as anti-cancer drugs has been severely restrictedbecause of their undesired psychoactive properties. In particular,activation of the central cannabinoid receptors (cannabinoid CB1

receptors), which are primarily found in the brain, is linked topsychoactivity. On the other hand, peripheral cannabinoid receptors(cannabinoid CB2 receptors) are almost exclusively found in theimmune system. Selective CB2 agonists without psychoactivity exhibitother side-effects such as immune suppression (Pertwee, 2005; Zhuet al., 2000). Therefore, the alternative use of such cannabinoids in

cancer treatment is best supported by the use of cannabinoids withvery weak binding or no affinity to CB receptors.

Angiogenesis is a crucial regulator of tumor growth andmetastases(Folkman, 1995). Tumor angiogenesis is regulated by the productionof angiogenic stimulators including vascular endothelial growth factor(VEGF), which is a key regulatory factor in the prognosis of variouscancers. Therefore, inhibition of VEGF production is a promisingtherapeutic approach for cancer treatment. One of the recent majorclinical advances in cancer treatment is the use of antiangiogenicdrugs such as bevacizumab, sorafenib, and sunitinib. Bevacizumab,the monoclonal anti-VEGF antibody, combined with taxane has beenapproved for the first-line treatment of metastatic breast cancer(Kerbel, 2009). Sunitinib, VEGF receptor tyrosine kinase inhibitor, isanother approach for anti-angiogenic therapy and acts directly onendothelial cells. Several class of cannabinoids have been shown tosuppress tumor growth either by inhibiting proangiogenic factorproduction (Casanova et al., 2003; Blazquez et al., 2004; Preet et al.,2008) or by directly inducing apoptosis of vascular endothelial cells(Kogan et al. 2006).

In the present study, we examined whether novel synthetichexahydrocannabinol analogs could inhibit tumor-angiogenesis

65D. Thapa et al. / European Journal of Pharmacology 650 (2011) 64–71

through the suppression of VEGF in cancer cells and VEGF-mediatedsignaling in endothelial cells. Since tamoxifen-resistant MCF-7(TAMR-MCF-7) cells have shown a strong association betweenenhanced VEGF production and more aggressive phenotype (Kimet al., 2008; Kim et al., 2009), we used this cell line as themodel cancercell line for the tumor-angiogenesis study. The cancer cell-inoculatedCAM assay was used to mimic the tumor microenvironment; thisassay has been widely used to evaluate not only angiogenesis, but alsotumor-induced angiogenesis and metastasis (Ribatti et al., 2001;Tufan and Satiroglu-Tufan, 2005).

2. Materials and methods

2.1. Reagents and antibodies

Recombinant human vascular endothelial growth factor (VEGF)was purchased from R&D Systems (Minneapolis, MN, USA). Endothe-lial growth medium (EGM)-2 bullet kit, which contains endothelialcell basal medium (EBM)-2 and EGM-2 SingleQuots (hydrocortisone,hFGF, VEGF, R3-IGF-1, ascorbic acid, hEGF, heparin, gentamicin andFBS), was purchased from Cambrex (San Diego, CA, USA). Matrigelwas obtained from BD Biosciences (San Jose, CA, USA). HEPES-buffered saline solution, Trypsin/EDTA, and Trypsin Neutralizingsolution (TNS) were purchased from Clonetics, Inc. (Walkersville,MD, USA). Triton-X-100, bovine serum albumin (BSA), 3-(4,5-dimethylthiazol-2-yl)-2,5-di-phenyl tetrazolium bromide (MTT), mi-tomycin C, sodium pyruvate, dimethyl sulfoxide (DMSO), proteaseinhibitor cocktail, sodium dodecyl sulfate (SDS) were obtained fromSigma-Aldrich (St. Louis, MO, USA). Cannabinoid CB1 receptorantagonist AM281 [N-(morpholin-4-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide] and canna-binoid CB2 receptor antagonist AM630 (6-iodo-pravadoline) werepurchased from Tocris Bioscience (Bristol, UK). The bicinchoninic acid(BCA) protein assay reagents and the chemiluminescent substrate(ECL kit) for horseradish peroxidase were from Pierce Biotechnology(Rockford, IL, USA). Primary antibodies specific for VEGF and actinwere from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

2.2. Chemical synthesis and treatment of LYR analogs

LYR analogs were synthesized according to the method describedpreviously (Lee and Xia, 2008). The chemical structures are shown inFig. 1A. A 50 mM stock solution of LYRs were prepared in DMSO,stored at−20 °C, and then diluted as needed. For in vitro incubations,LYRs were directly applied at a final DMSO concentration of 0.1–0.2%(v/v). For in vivo experiments (CAM tumor implantation), LYRs wereprepared at 1% (v/v) DMSO in phosphate buffered saline (PBS)supplemented with 0.1% BSA. No significant influence of the vehiclewas observed on any of the parameters assessed.

2.3. Cell culture

Human umbilical vein endothelial cells (HUVECs) were purchasedfrom Lonza Walkersville, Inc. HUVECs were maintained in cell cultureflasks coated with 0.2% gelatin and cultured with EBM-2 containinghydrocortisone, human basic fibroblast growth factor, VEGF, R3-IGF-1,ascorbic acid, human epidermal growth factor, heparin, gentamicin,and 2% FBS. Confluent cultures of HUVECs were serially passaged andwere used between passages 2 and 6.

MCF-7 (human breast cancer cell line) cells were obtained fromAmerican Type Culture Collection (Manassas, VA, USA). Tamoxifen(TAM)-resistant breast cancer cell line (MCF-7/TAMR cells) wasdeveloped as described earlier (Kim et al., 2009). Both of these celllines were grown in DMEM supplemented with 10% FBS and 1%penicillin/streptomycin. HT29 human colon cancer cells were main-tained in RPMI 1640 medium. All cells were maintained at 37 °C in a

5% CO2-humidified atmosphere. The culture medium was replacedevery other day.

2.4. Tube formation assay

Tube formation assays were performed on 48-well plates coatedwith 100 μl of Matrigel per well and polymerized at 37 °C for 30 min.HUVECs were suspended in EBM-2 containing 1% FBS, plated onMatrigel at a density of 1×104 cells per well, and then co-treated withVEGF (20 ng/ml) and indicated concentrations of LYR analogs. After14 h, four fields of each culture well were randomly selected andphotographedwith amicroscope attached to a CCD camera (TE2000-U;Nikon). Digital images were analyzed with an image analysis system(ImageInside Ver 3.32) for the quantitation of tube length.

2.5. Migration assays

Two types of cell migration assays were performed with HUVECs.First, cell migration was examined in scratch assays as describedpreviously (Park et al., 2007). Confluent HUVECswere pretreated withmitomycin C (25 μg/ml) for 30 min and then a scratch line was madein the cell monolayer. After being rinsed with Dulbecco's phosphatebuffered saline (DPBS), the cells were further incubated in EBM-2medium (1% FBS) supplemented with 20 ng/ml VEGF in the presenceof LYR-7 or LYR-8 for 10 h. Pictures of the scratches were taken using adigital camera system (Nikon, Japan) connected to a light microscope(Olympus).

For modified Boyden chamber migration assay, HUVECs werecultured onto gelatin-coated 8-μm pore size chambers (Corning, NY,USA), and the bottom well was filled with EBM-2 medium containingVEGF (20 ng/ml) as the chemoattractant. 8 h after incubatingchambers, cells were fixed with methanol and stained with H&E.Migrating cells were imaged (100×) and counted using a microscopeconnected to a digital camera.

2.6. Cell viability assay

The cytotoxicity of LYR analogs in MCF-7 and TAMR-MCF-7 breastcancer cells was measured using the MTT assay (Carmichael et al.,1987). Briefly, cells (1×104 cells/well) were seeded in 96-wellmicrotiter plates (Nunc, Denmark). After 24 h, fresh medium(DMEM supplemented with 10% FBS) containing indicated concen-trations of LYR analogs or DMSO vehicle was replaced and incubatedfor 48 h. Relative cell viability was determined by the amount of MTTconverted to formazan salt and expressed as a percent of the controlculture.

2.7. HUVEC proliferation assay

HUVECs plated at a density of 1×104 cells/cm2 were incubated inserum-reduced (0.2% FBS) EBM-2 for 24 h and then co-treated withVEGF (20 ng/ml) in the absence or presence of LYR-7 or LYR-8 for48 h. After incubation, the viable cell number was determined by MTTassay.

2.8. VEGF enzyme-linked immunosorbent assay

Secreted VEGF levelswere determined by using aQuantikine humanVEGF ELISA kit (R&D Systems, Minneapolis, MN, USA) as describedpreviously (Kim et al., 2008; Park et al., 2009). In brief, MCF-7 cells wereseeded in 24-well plates and grown to 80–90% confluence. The cellswere switched to fresh serum-free medium in the presence or absenceof LYR analogs and incubated for 18 h. After the treatment, thesupernatants were collected and the cells were subjected to the MTTassay tomeasure relative cell viability. The concentration of VEGF in theunknown samples was then determined by comparing the optical

Fig. 1. LYR analogs inhibit tube formation of HUVECs. (A) LYR analogs (LYR-1 to LYR-8) are structurally related to classical cannabinoid Δ9-tetrahydrocannabinol (THC). Chemicalstructure of Δ9-THC is shown in the box. (B) HUVECs (1×104) were plated in a well coated with Matrigel basement membrane matrix. Cells were treated with 5 μM LYR analogs inthe presence of VEGF (20 ng/ml). After 14 h, cells were photographed with a digital camera under a phase contrast microscope at 100× magnification. (C) Experiment was carriedout as described above with the indicated concentrations of LYR-7 or LYR-8 for 14 h. (D) Total tube length per field was measured by ImageInside software. The bar graph shows themeans±S.E.M. of the experiment carried out in triplicate. *Pb0.05, compared to VEGF-stimulated control group.

66 D. Thapa et al. / European Journal of Pharmacology 650 (2011) 64–71

density of the samples to the standard curve and normalized to the cellviability in each well.

2.9. Protein extraction and Western blotting

Whole cell lysates were prepared using RIPA buffer. Proteincontent was measured with BCA protein assay reagent (Pierce,Rockford, IL, USA). Equal amounts of total protein were separatedby SDS-PAGE and transferred onto Hybond ECL nitrocellulosemembranes (Amersham Life Science, Buckinghamshire, UK) at200 mA for an hour. The membranes were blocked in 5% skim milkin Tris-buffered saline (TBS)-Tween 20 (TBS-T) at room temperaturefor 1 h followed by incubation with specific antibodies in skim milk-TBS at 4 °C overnight. Then, the membrane was washed three times

with TBS-T and incubated with horseradish peroxidase-conjugatedsecondary antibody in skimmilk-TBS for 1 h at room temperature. Theimmunoreactive proteins were visualized using an ECL kit (Pierce)and digitally processed using LAS-4000mini (Fuji, Japan). Membraneswere stripped and reprobed with an actin antibody for loadingcontrol. Densitometric analysis of the blots was performed with MultiGauge Ver 3.2 imaging software in a Fuji Image Station.

2.10. Transient transfection and luciferase assay

Transient transfections were performed using the GeneJammertransfection reagent (Stratagene, La Jolla, CA, USA) as describedpreviously (Thapa et al., 2008). Briefly, HT29 cells were transfectedwith plasmid mixtures containing 0.6 μg of NF-κB-luc (Panomics,

67D. Thapa et al. / European Journal of Pharmacology 650 (2011) 64–71

Fremont, CA, USA) and 0.05 μg of pRL-TK. The transfected cells werepretreated with LYR analogs for 1 h, followed by tumor necrosis factor(TNF)-α for 3 h. The cell lysates were used for luciferase assay using adual luciferase reporter assay kit (Promega, Madison, WI, USA), andthe emitted light was measured with a luminometer (TurnerBioSystems, Sunnyvale, CA, USA).

2.11. In vivo chicken chorioallantoic membrane (CAM) assay

Fertilized eggs were purchased from Baek-ja Farm (Cheongsong,Korea) and the CAMwas prepared as described previously (Park et al.,2009). VEGF was used to stimulate vessel growth on the CAMs of 9-day-old chick embryos. Sterile filter disks absorbed with VEGF (20 ng/disk) dissolved in PBS containing 0.1% BSA were placed on thegrowing CAMs. Test compounds or a vehicle was then added directlyto CAMs topically. After 72 h, CAM tissue directly beneath the diskwere resected from the embryo and harvested under light microscopy(Leica, Wetzlar, Germany). The number of vessel branch pointscontained in a circular region equal to the area of filter disk was thencounted for each section.

In the tumor angiogenesis experiments, all the procedures werethe same as above except that TAMR-MCF-7 or MCF-7 human breastcancer cells (1.5×106 cells/CAM) were inoculated onto the CAMinstead of VEGF (Kim et al., 2008, 2009). The number of vessel branchpoints contained in a tumor region was counted by two observers in adouble-blind manner.

2.12. Statistical analysis

The data are presented as means±S.E.M. Statistical analysis wasdone with a Student's t test or one-way ANOVA followed by theStudent–Newman–Keuls comparison (GraphPad Prism4.0 software,

Fig. 2. LYR-7 and LYR-8 inhibit migration and invasion of HUVECs. (A) Scratch woundmigratInactivated HUVECs were subjected to wound-healing migration assays and the migrating c(B) Chemotactic migration in Transwell: effect of LYR analogs on VEGF-induced HUVECmigraon the outside surface of the top chamber. (A–B) The bar graphs in the right panel show the*Pb0.05, compared to VEGF-stimulated control group.

San Diego, CA, USA) to calculate differences between groups. P valuesof b0.05 were considered statistically significant.

3. Results

3.1. Anti-angiogenic activities of hexahydrocannabinol analogs:inhibition of HUVEC tube formation, migration, and invasion

From a novel series of synthetic hexahydrocannabinols that arestructurally similar to natural cannabinoid THC, eight analogs namedLYR-1 to LYR-8 (Fig. 1A) were screened for anti-angiogenic properties.The tube formation assay was employed for preliminary screeningbecause it is one of themost popular in vitro angiogenesis tests (Vernonet al., 1992). Endothelial cells were plated on a Matrigel-coated surfacein the presence of VEGF, a well-known factor that induces endothelialcell growth and iswidely expressed inmost cancers (Ferrara andKerbel,2005). As shown in Fig. 1B, the cells differentiate and align to form anetwork of tubes, i.e., capillary-like structures. All the LYR analogs(5 μM) tested caused a blockage of the in vitro endothelial tubeformation. Among these, LYR-7 and LYR-8 were chosen for furtherexperiments because they strongly inhibited angiogenesiswhile havingvery little or no affinity for cannabinoid receptor CB1 or CB2

(Supplementary Fig. S1; Thapa et al., 2010), thus eliminating theconcern about potential psychoactive side effects. Further study withLYR-7 and LYR-8 showed that such inhibitory effects of the two analogswere concentration-dependent (Fig. 1C and D).

As endothelial cell migration is another important step ofangiogenesis (Park et al., 2006), wound-healing migration assay wasperformed to determine the effects of LYR-7 and LYR-8 on HUVECmigration. LYR analogs significantly delayed VEGF-induced migrationat 10 h after HUVEC wound injury (Fig. 2A). These effects wereconcentration-dependent, and significant inhibitory activity was seenat 1 μM (Fig. 2A, right panel). Similarly, in a subsequent Transwell

ion of HUVECs: inhibitory effect of LYR-7 and LYR-8 on VEGF-induced HUVECmigration.ells were counted using ImageInside software. Experiment was performed in triplicate.tion in the Transwell assay. Red cells with irregular shape were migrating cells attachedsummary of quantitative results of the average number of migrating HUVECs±S.E.M.

68 D. Thapa et al. / European Journal of Pharmacology 650 (2011) 64–71

assay (Fig. 2B), both analogs significantly inhibited chemotacticmigration of HUVECs.

3.2. Anti-proliferative activities of LYR-7 and LYR-8 against endothelialand cancer cells

We also examined the inhibitory action of the analogs against theproliferation of endothelial cells. LYR-7 and LYR-8 significantlyinhibited VEGF-induced HUVEC proliferation with an IC50 of 8 μM(Fig. 3A). Since primary HUVECs are reported to express CB receptors(Blazquez et al., 2003) and the analog LYR-7 shows little affinity for CBreceptors (Supplementary Fig. S1), we examined the involvement of

Fig. 3. Inhibitory effects of LYR-7 and LYR-8 on growth factor-induced cell proliferation and Tplate) were starved with 0.2% FBS medium and then treated with VEGF (20 ng/ml) and diffeValues are means±S.E.M. of eight measurements. *Pb0.05 versus VEGF alone. (B) HUVECs w(1 μM) were pre-treated for 1 h before the LYRs (5 μM) and VEGF co-treatment. Cell viab#Pb0.05 versus vehicle-treated control. *Pb0.05 versus VEGF alone. (C–D) The effects of LYRtreatment is the same as in A and B. The data points represent three experiments perfoGeneJammer reagent. Cells were then pre-treated with or without LYR analogs (5 μM) for 1 has relative luciferase units (RLU) (firefly luciferase/Renilla luciferase). Bar graphs show theversus TNF-α-stimulated group.

CB receptors in such action of the two analogs. Preincubation ofHUVECs with AM281, a selective cannabinoid CB1 receptor antagonist(1 μM), or AM630, a selective cannabinoid CB2 receptor antagonist(1 μM), did not alter the inhibitory effects of LYR analogs on VEGF-stimulated HUVEC proliferation (Fig. 3B). Next, we examined theselectivity of the LYR analogs toward endothelial cells by comparingcell viability changes in the analog-treated endothelial cells andcancer cells. Under normal HUVEC culture conditions (2% serum andother growth supplements), LYR analogs inhibited endothelialproliferation in a concentration-dependent manner (Fig. 3C and D).Much higher concentrations of the LYR analogs were needed to inhibitthe proliferation of MCF-7 breast cancer cells than of endothelial cells.

NF-α-induced NF-κB transcriptional activity. (A) HUVECs (1×104 per well in a 48 well-rent concentrations of LYR analogs for 48 h. Cell viability was quantified by MTT assay.ere cultured and stimulated with VEGF as described above. AM281 (1 μM) and AM630ility was quantified by MTT assay. Values are means±S.E.M. of eight measurements.analogs on serum-treated proliferation of HUVECs, MCF-7 and TAMR-MCF-7 cells: Thermed in triplicate. (E) NF-κB-luc and pRL-TK were introduced into HT29 cells using, followed by TNF-α (10 ng/ml) for 3 h. Luciferase activity was measured and expressedmeans±S.E.M. of three experiments. #Pb0.05 versus vehicle-treated control. *Pb0.05

Fig. 4. LYR-7andLYR-8 inhibit theVEGF-inducedangiogenesis invivo. (A)VEGF(20 ng/CAM)or vehicle (0.1% BSA in PBS)was loaded onto a dried cortisone-saturated filter disk placed onan avascular area of the CAM. The diskswere treatedwith various concentrations of LYR-7 orLYR-8. After 72 h incubation, the CAM membrane was resected and imaged undermicroscope. (B) Quantitation of new branches formed from existing blood vessels:Photographs were imported into imaging software to quantitate the number of newbranches formed. The bar graph represents the mean number of branch points±S.E.M. of atleast six chick embryos. #Pb0.05, compared to PBS-treated control. *Pb0.05, compared toVEGF-treated group.

69D. Thapa et al. / European Journal of Pharmacology 650 (2011) 64–71

More importantly, the effects of the two compounds on chemoresis-tant TAMR-MCF-7 cell proliferation were similar to those observed inparent MCF-7 cells (Fig. 3C and D).

Since there aremany reports highlighting the potential role of NF-κBin angiogenesis and its blockade resulted in impaired tumor angiogen-esis (Huang et al., 2001; Rius et al., 2008; Schmidt et al., 2007), we alsoexamined whether the compounds had NF-κB inhibitory activity usingNF-κB-luciferase reporter assay in HT29 cancer cells. NF-κB transcrip-tional activity, which was stimulated by TNF-α, was strongly inhibitedby LYR-7 and LYR-8 among the other analogs (Fig. 3E).

3.3. LYR analogs inhibit angiogenesis in vivo

To confirm the anti-angiogenic effects of LYR-7 and LYR-8 in vivo,CAM assays were performed at different concentrations. As shown inFig. 4A and B, LYR-7 or LYR-8 at 0.5 μg/CAM strongly inhibited VEGF-induced angiogenesis, and 5 μg/CAM LYR-7 or LYR-8 almost com-pletely abolished VEGF-induced angiogenesis in the CAM assays,indicating that these analogs effectively inhibited angiogenesis in vivo.

3.4. LYR analogs inhibit VEGF expression in breast cancer cells

VEGF production by tumor cells under hypoxic conditions isbelieved to be one of the most specific and critical regulators ofangiogenic signaling cascades (Ferrara, 2002). In breast cancer cells,increased levels of VEGF are correlated with aggressive, metastatic,and drug-resistant phenotypes (Marson et al., 2001). TAMR-MCF-7 isone of such phenotypes (Kim et al., 2009). In the cells, LYR analogssignificantly suppressed VEGF secretion (Fig. 5A) and proteinexpression (Fig. 5B and C) in a concentration-dependent manner.

3.5. LYR analogs inhibit tumor angiogenesis and breast cancer growth

To investigate whether LYR analogs inhibit tumor-inducedangiogenesis in vivo, the cancer-implanted CAM model was used.TAMR-MCF-7 and MCF-7 cancer cell suspensions with or without LYRanalogs were seeded onto the chorioallantoic membrane of 9-day-oldchick embryos. New blood vessel formation and tumor growth on themembrane after 5 days of incubation were analyzed. As shown inFig. 6A and B, the cancer cell-implanted tumors induced profusevasculature radiating from cell masses. Notably, TAMR-MCF-7 cell-induced angiogenesis was about 5-fold higher than that in the controlgroup treated with PBS and about 2-fold higher than that in therecombinant human VEGF-treated group (Fig. 6C). Consistent withthe previous finding (Kim et al., 2009), tumor vasculature density wassignificantly higher in TAMR-MCF-7 cell-implanted CAMs than inMCF-7 cell-implanted CAMs. Such cancer cell-induced angiogenesiswas significantly suppressed by LYR-7 or LYR-8 treatment. Althoughwe did not dissect out the tumor mass from the membranes, thevisible tumor masses were notably diminished by LYRs treatment(Fig. 6A and B; indicated by dot circles).

4. Discussion

In this study, hexahydrocannabinols, novel synthetic analogs thatare structurally related to the cannabinoids, were identified as potentangiogenesis inhibitors. We have demonstrated that, among thehexahydrocannabinol analogs, LYR-7 and LYR-8 most effectively(a) inhibit endothelial and tumor cell growth; (b) inhibit VEGF-induced angiogenesis in vitro and in vivo, and the growth of tumorsinoculated on chick chorioallantoic membrane; and (c) block thesecretion of VEGF in cancer cells.

Angiogenesis is a complex multistep process involving endothelialcell proliferation, invasion, chemotactic migration, differentiation intotube-like structures, and the production of a basement membranearound the vessel (Folkman, 1995; Kesisis et al., 2007). Using various

in vitro and in vivo experiments, we clearly showed that LYR analogscould inhibit these multistep processes of angiogenesis. Unlike otherconventional cannabinoids, which have been linked to potentialpsychoactive properties via activation of cannabinoid CB1 receptor,our previous (Thapa et al., 2010) and the present study (Supplemen-tary Fig. S1) revealed that these two compounds have little or noaffinity for CB receptors in a receptor binding assay. The CB receptorinvolvement in the anti-cancer action of cannabinoids depends on thenature of cannabinoid. Cannabidiol (CBD), one of the bioactiveconstituents of marijuana, exhibits anti-tumor activities through acannabinoid receptor-independent mechanism (Massi et al., 2004;Vaccani et al., 2005), whereas the anti-tumor activities of THC areassociated with both cannabinoid CB1 and CB2 receptors (Sanchezet al., 2001; Ramer and Hinz, 2008; Sarfaraz et al., 2008). In thepresent study, neither the CB1 antagonist AM281 nor the CB2

antagonist AM630 prevented the LYR-7- or LYR-8-induced inhibitionof cell proliferation (Fig. 3B) and tube formation (data not shown).These findings along with poor or non-receptor affinity of both LYRanalogs indicate that the anti-angiogenic activity of both LYR analogsseems to be mediated through a cannabinoid receptor-independentmechanism. Since LYR-7 and LYR-8 strongly inhibit NF-κB transcrip-tional activity and NF-κB plays a well-known role in VEGF regulation

Fig. 5. LYR-7 and LYR-8 inhibit VEGF expression in TAMR-MCF-7 cells. (A) TAMR-MCF-7cells were treated with the indicated concentrations of LYR analogs for 18 h, andsupernatants were collected. The secreted level of VEGF was quantified by ELISA asdescribed in Materials and methods. The bar graph represents the means±S.E.M. of sixanalyses. *Pb0.05, compared to vehicle-treated control. (B) The TAMR-MCF-7 cellswere treated as above and whole cell lysates were used to determine cellular VEGFlevels by western blot analysis. (C) The band densities were measured and presented ina graph showing relative density±S.E.M. of three independent experiments. *Pb0.05,compared to vehicle-treated control.

Fig. 6. LYR-7 and LYR-8 inhibit cancer-induced angiogenesis and tumor growth. (A) Fortumor implantation on CAM, 1.5×106 TAMR-MCF-7 cells were loaded onto each CAMand a single dose (20 μM) of LYR-7 or LYR-8 was given at the time of implantation. After5 days of incubation, the CAM tissues were resected and digital images were captured.(B) A parallel experiment was carried out with MCF-7 cells. As shown in thephotographs, LYR-7 and LYR-8 inhibited both angiogenesis and tumor growth (size).(C) New blood vessels were quantified and expressed as percentage of PBS-treatedcontrol. #Pb0.05, compared to PBS- or VEGF-treated groups. $Pb0.05 versus vehicle-treated MCF-7 cell-implanted group. *Pb0.05, compared to vehicle-treated cancer cell(TAMR-MCF-7 or MCF-7)-implanted group.

70 D. Thapa et al. / European Journal of Pharmacology 650 (2011) 64–71

and angiogenesis, we suggest that NF-κB is a possible target of LYRaction.

Althoughmost breast cancers are initially responsive to anti-estrogentherapy with tamoxifen, they ultimately develop resistance to tamoxifen(Clemons et al., 2002). In TAM-resistant breast tumor tissues, microvesselcounts are significantly higher than those in TAM-responsive tumortissues (Marson et al., 2001). Consistent with this, we have also shownthat angiogenicpotential andVEGFproductionarehigher inTAMR-MCF-7than in MCF-7 cells (Kim et al., 2008, 2009). VEGF is therefore animportant therapeutic target for tamoxifen-resistant breast cancerphenotypes. Our results showed that LYR-7 and LYR-8 inhibit VEGFproduction and suppress proliferation to a similar degree in bothchemoresistant TAMR-MCF-7 cells and responsive MCF-7 cells. Wetherefore hypothesize that LYR-7 and LYR-8 may have great potential asvaluable anti-cancer drugs in the treatment of drug-resistant breastcancers. This notion was further supported by the results from the breastcancer-inoculated CAMmodel in which cancer cell-induced neovascular-ization and tumor growth were suppressed by LYR-7 and LYR-8.

There are several reports showing cell density-dependent toxicityfor cannabinoids (Ramer and Hinz, 2008), and other therapeutic drugssuch as tamoxifen and doxorubicin (Brandt et al., 2004; Kobayashiet al., 1992). In our study, 1×104 cells were used for cell proliferationassay in vitrowhereas 1.5×106 cells (1500× higher) in mass was usedfor tumor growth onto CAM assay in vivo. Based on the reports thatincreasing cell density is associatedwith decreased cytotoxicity, in the

present study, the inhibition of tumor-induced angiogenesis in CAMassay was not probably due to cytotoxicity. In addition, the inhibitoryeffect of LYR analogs on TAMR-MCF-7 cell-induced angiogenesis wasalso noted at low non-toxic doses, 1–10 μM(data not shown). Besides,concomitant observations of chick embryo viability and morphologyof resected membranes confirmed that LYR analogs were non-toxic atthe test doses (1–20 μM). In summary, these findings suggest that LYRanalog-mediated anti-angiogenic effects are not likely due to celltoxicity, but largely due to modulation of angiogenic process andVEGF-production. On the other hand, based on the result that LYRanalogs at 20 μM caused significant inhibition of the tumor cellproliferation in vitro, the direct action of LYR analogs on tumor cellproliferation cannot be ruled out as one of the possible mechanisms inreduced tumor growth and tumor-induced angiogenesis.

71D. Thapa et al. / European Journal of Pharmacology 650 (2011) 64–71

In conclusion, our results indicate that at least two mechanismsmay contribute in this anti-cancer action of synthetic hexahydrocan-nabinols: direct action in cancer cells (inhibition of growth and VEGFsecretion) and indirect action through endothelial cells (inhibition ofendothelial cell proliferation, migration, invasion, and tube forma-tion). These results further suggest that hexahydrocannabinol andsimilar analogs may be excellent leads for the development of anti-angiogenic and anti-cancer drugs.

Acknowledgements

This research was supported by the Basic Science Research Programthrough theNational Research Foundation of Korea (NRF) funded by theMinistry of Education, Science and Technology [2010-0001661] and bythe Regional Technology Innovation Program of the Ministry ofCommerce, Industry, and Energy (MOCIE) [grant No. RTI04-01-04].

Appendix A. Supplementary data

Supplementary data to this article can be found online atdoi:10.1016/j.ejphar.2010.09.073.

References

Blazquez, C., Casanova, M.L., Planas, A., Gomez Del Pulgar, T., Villanueva, C., Fernandez-Acenero, M.J., Aragones, J., Huffman, J.W., Jorcano, J.L., Guzman, M., 2003. Inhibitionof tumor angiogenesis by cannabinoids. FASEB J. 17, 529–531.

Blazquez, C., Gonzalez-Feria, L., Alvarez, L., Haro, A., Casanova, M.L., Guzman, M., 2004.Cannabinoids inhibit the vascular endothelial growth factor pathway in gliomas.Cancer Res. 64, 5617–5623.

Brandt, S., Heller, H., Schuster, K.D., Grote, J., 2004. Tamoxifen induces suppression ofcell viability and apoptosis in the human hepatoblastoma cell line HepG2 via down-regulation of telomerase activity. Liver Int. 24, 46–54.

Carmichael, J., Degraff, W.G., Gazdar, A.F., Minna, J.D., Mitchell, J.B., 1987. Evaluation of atetrazolium-based semiautomated colorimetric assay: assessment of chemosensi-tivity testing. Cancer Res. 15, 936–942.

Casanova, M.L., Blazquez, C., Martinez-Palacio, J., Villanueva, C., Fernandez-Acenero, M.J.,Huffman, J.W., Jorcano, J.L., Guzman, M., 2003. Inhibition of skin tumor growth andangiogenesis in vivo by activation of cannabinoid receptors. J. Clin. Invest. 111, 43–50.

Clemons, M., Danson, S., Howell, A., 2002. Tamoxifen (“Nolvadex”): a review. CancerTreat. Rev. 28, 165–180.

Ferrara, N., 2002. VEGF and the quest for tumour angiogenesis factors. Nat. Rev. Cancer.2, 795–803.

Ferrara, N., Kerbel, R.S., 2005. Angiogenesis as a therapeutic target. Nature 438,967–974.

Folkman, J., 1995. Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat.Med. 1, 27–31.

Galve-Roperh, I., Sanchez, C., Cortes, M.L., Gomez del Pulgar, T., Izquierdo, M., Guzman,M., 2000. Anti-tumoral action of cannabinoids: involvement of sustained ceramideaccumulation and extracellular signal-regulated kinase activation. Nat. Med. 6,313–319.

Guzman, M., 2003. Cannabinoids: potential anticancer agents. Nat. Rev. Cancer 3,745–755.

Guzman, M., Sanchez, C., Galve-Roperh, I., 2002. Cannabinoids and cell fate. Pharmacol.Ther. 95, 175–184.

Huang, S., Pettaway, C.A., Uehara, H., Bucana, C.D., Fidler, I.J., 2001. Blockade of NF-kappaB activity in human prostate cancer cells is associated with suppression ofangiogenesis, invasion, and metastasis. Oncogene 20, 4188–4197.

Kerbel, R.S., 2009. Issues regarding improving the impact of antiangiogenic drugs forthe treatment of breast cancer. Breast 18, S41–S47.

Kesisis, G., Broxterman, H., Giaccone, G., 2007. Angiogenesis inhibitors. Drug selectivityand target specificity. Curr. Pharm. Des. 13, 2795–2809.

Kim, J.A., Cho, K.B., Kim, M.R., Park, B.C., Kim, S.K., Lee, M.Y., Kang, K.W., 2008. Decreasedproduction of vascular endothelial growth factor in adriamycin-resistant breastcancer cells. Cancer Lett. 268, 225–232.

Kim, M.R., Choi, H.S., Yang, J.W., Park, B.C., Kim, J.A., Kang, K.W., 2009. Enhancement ofvascular endothelial growth factor-mediated angiogenesis in tamoxifen-resistantbreast cancer cells: role of Pin1 overexpression. Mol. Cancer Ther. 8, 2163–2171.

Kobayashi, H., Takemura, Y., Ohnuma, T., 1992. Relationship between tumor cell densityand drug concentration and the cytotoxic effects of doxorubicin or vincristine:mechanism of inoculum effects. Cancer Chemother. Pharmacol. 31, 6–10.

Kogan, N.M., Blazquez, C., Alvarez, L., Gallily, R., Schlesinger, M., Guzman, M.,Mechoulam, R., 2006. A cannabinoid quinone inhibits angiogenesis by targetingvascular endothelial cells. Mol. Pharmacol. 70, 51–59.

Lee, Y.R., Xia, L., 2008. Efficient one-pot synthetic approaches for the cannabinoidanalogues and their application to biologically interesting (−)-hexahydrocanna-binol and (+)-hexahydrocannabinol. Tetrahedron Lett. 49, 3283–3287.

Marson, L.P., Kurian, K.M., Miller, W.R., Dixon, J.M., 2001. The effect of tamoxifen onbreast tumour vascularity. Breast Cancer Res. Treat. 66, 9–15.

Massi, P., Vaccani, A., Ceruti, S., Colombo, A., Abbracchio, M.P., Parolaro, D., 2004.Antitumor effects of cannabidiol, a nonpsychoactive cannabinoid, on human gliomacell lines. J. Pharmacol. Exp. Ther. 308, 838–845.

Park, H.J., Zhang, Y., Georgescu, S.P., Johnson, K.L., Kong, D., Galper, J.B., 2006. Humanumbilical vein endothelial cells and human dermal microvascular endothelial cellsoffer new insights into the relationship between lipid metabolism and angiogen-esis. Stem Cell Rev. 2, 93–102.

Park, B.C., Thapa, D., Lee, Y.S., Kwak, M.K., Lee, E.S., Choi, H.G., Yong, C.S., Kim, J.A., 2007.1-furan-2-yl-3-pyridin-2-yl-propenone inhibits the invasion and migration ofHT1080 human fibrosarcoma cells through the inhibition of proMMP-2 activationand down regulation of MMP-9 and MT1-MMP. Eur. J. Pharmacol. 567, 193–197.

Park, B.C., Park, S.Y., Lee, J.S., Mousa, S.A., Kim, J.T., Kwak, M.K., Kang, K.W., Lee, E.S., Choi,H.G., Yong, C.S., Kim, J.A., 2009. The anti-angiogenic effects of 1-furan-2-yl-3-pyridin-2-yl-propenone are mediated through the suppression of both VEGFproduction and VEGF-induced signaling. Vasc. Pharmacol. 50, 123–131.

Pertwee, R.G., 2005. Pharmacological actions of cannabinoids. Handb. Exp. Pharmacol.168, 1–51.

Preet, A., Ganju, R.K., Groopman, J.E., 2008. Delta9-Tetrahydrocannabinol inhibitsepithelial growth factor-induced lung cancer cell migration in vitro as well as itsgrowth and metastasis in vivo. Oncogene 27, 339–346.

Ramer, R., Hinz, B., 2008. Inhibition of cancer cell invasion by cannabinoids viaincreased expression of tissue inhibitor of matrix metalloproteinases-1. J. NatlCancer Inst. 100, 59–69.

Ribatti, D., Nico, B., Vacca, A., Roncali, L., Burri, P.H., Djonov, V., 2001. Chorioallantoicmembrane capillary bed: a useful target for studying angiogenesis and anti-angiogenesis in vivo. Anat. Rec. 264, 317–324.

Rius, J., Guma, M., Schachtrup, C., Akassoglou, K., Zinkernagel, A.S., Nizet, V., Johnson, R.S.,Haddad, G.G., Karin, M., 2008. NF-kappaB links innate immunity to the hypoxicresponse through transcriptional regulation of HIF-1alpha. Nature 453, 807–811.

Sanchez, C., de Ceballos, M.L., Gomez del Pulgar, T., Rueda, D., Corbacho, C., Velasco, G.,Galve-Roperh, I., Huffman, J.W., Ramon Y Cajal, S., Guzman, M., 2001. Inhibition ofglioma growth in vivo by selective activation of the CB2 cannabinoid receptor.Cancer Res. 61, 5784–5789.

Sarfaraz, S., Adhami, V.M., Syed, D.N., Afaq, F., Mukhtar, H., 2008. Cannabinoids forcancer treatment: progress and promise. Cancer Res. 68, 339–342.

Schmidt, D., Textor, B., Pein, O.T., Licht, A.H., Andrecht, S., Sator-Schmitt, Fusenig, N.E.,Angel, P., Schorpp-Kistner, M., 2007. Critical role for NF-kappaB-induced JunB inVEGF regulation and tumor angiogenesis. EMBO J. 26, 710–719.

Thapa, D., Lee, J.S., Park, S.Y., Bae, Y.H., Bae, S.K., Kwon, J.B., Kim, K.J., Kwak, M.K., Park, Y.J.,Choi, H.G., Kim, J.A., 2008. Clotrimazole ameliorates intestinal inflammation andabnormal angiogenesis by inhibiting interleukin-8 expression through a nuclearfactor-kappaB-dependent manner. J. Pharmacol. Exp. Ther. 327, 353–364.

Thapa, D., Babu, D., Park, M.A., Kwak, M.K., Lee, Y.R., Kim, J.M., Kwon, T.K., Kim, J.A., 2010.Induction of p53-independent apoptosis by a novel synthetic hexahydrocannabinolanalog is mediated via Sp1-dependent NSAID-activated gene-1 in colon cancercells. Biochem. Pharmacol. 80, 62–71.

Tufan, A.C., Satiroglu-Tufan, N.L., 2005. The chick embryo chorioallantoic membrane asa model system for the study of tumor angiogenesis, invasion and development ofanti-angiogenic agents. Curr. Cancer Drug Targets 5, 249–266.

Vaccani, A., Massi, P., Colombo, A., Rubino, T., Parolaro, D., 2005. Cannabidiol inhibitshuman glioma cell migration through a cannabinoid receptor-independentmechanism. Br. J. Pharmacol. 144, 1032–1036.

Vernon, R.B., Angello, J.C., Iruela-Arispe, M.L., Lane, T.F., Sage, E.H., 1992. Reorganizationof basement membrane matrices by cellular traction promotes the formation ofcellular networks in vitro. Lab. Invest. 66, 536–547.

Zhu, L.X., Sharma, S., Stolina, M., Gardner, B., Roth, M.D., Tashkin, D.P., Dubinett, S.M.,2000. Delta-9-tetrahydrocannabinol inhibits antitumor immunity by a CB2

receptor-mediated, cytokine-dependent pathway. J. Immunol. 165, 373–380.