Marine Spongecribrochalina Vasculumcompounds Activate

Published OnlineFirst October 15, 2014; DOI: 10.1158/1535-7163.MCT-14-0329

Molecular Cancer Small Molecule Therapeutics Therapeutics

Marine Sponge Cribrochalina vasculum Compounds Activate Intrinsic Apoptotic Signaling and Inhibit Growth Factor Signaling Cascades in Non–Small Cell Lung Carcinoma

Ana Zovko1, Kristina Viktorsson1, Petra Haag1, Dimitry Kovalerchick2, Katarina Farnega€ rdh3, Andrea Alimonti4,5, Micha Ilan6, Shmuel Carmeli2, and Rolf Lewensohn1

Abstract Marine-derived compounds have been explored and considered as possible antitumor agents. In this study, we analyzed extracts of the sponge Cribrochalina vasculum for their ability to inhibit tumor cell proliferation. Screening identified two acetylenic compounds of similar structure that showed strong tumor-specific toxicity in non–small cell lung carcinoma (NSCLC) cells and small-cell lung carcinoma cells, and less prominent toxicity in ovarian carcinoma, while having no effect on normal cells. These acetylenic compounds were found to cause a time-dependent increase in activation of apoptotic signaling involving cleavage of caspase-9, caspase-3, and PARP, as well as apoptotic cell morphology in NSCLC cells, but not in normal fibroblasts. Further analysis demonstrated that these compounds caused conformational change in Bak and Bax, and resulted in loss of mitochondrial potential and cytochrome c release in NSCLC cells. Moreover, a decreased phosphorylation of the growth factor signaling kinases Akt, mTOR, and ERK was evident and an increased phosphorylation of JNK was observed. Thus, these acetylenic compounds hold potential as novel therapeutic agents that should be further explored for NSCLC and other tumor malignancies. Mol Cancer Ther; 13(12); 1–14. 2014 AACR.

Introduction sponge-derived compounds are at different stages of Marine organisms have been shown to be a great source development (preclinical to phase III and preregistration) of pharmaceutical compounds that can be used for anti- and also to some extent in clinical use (6, 7). Successful cancer, antiviral or anti-inflammatory purposes. Since the examples are discodermolide, trabectedin, eribulin, and cytosar-U (8–12). Cytosar-U more commonly known as 1970s, there is a steady increase in the number of new Cryptotethia crypta compounds discovered and the number of registered Ara-C, identified in , is clinically patents is also growing (1–3). Among the different marine approved for the treatment of acute myelogenous leuke- organisms, sponges (Porifera) have been identified as the mia (12). Another example is eribulin, a truncated ana- logue of the natural product halichondrin-B isolated from most promising source of new compounds with antican- Halichondria okadai cer potential (2–5). The leading role of Porifera as a source the sponge . This agent was approved of new anticancer compounds is likely attributed to their by the FDA in 2010, to treat patients with metastatic breast long evolutionary history, extreme plasticity, and their cancer (11). Trabectedin is yet another example of a remarkably rich associated microbiota (4). Indeed many compound from the marine source, the Caribbean marine tunicate Ecteinascidia turbinata, which has passed all steps of development into a pharmaceutical agent and is approved for patients with advanced or metastatic soft- 1Department of Oncology and Pathology, Karolinska Biomics Center, Karolinska Institutet, Stockholm, Sweden. 2School of Chemistry, Raymond tissue sarcoma or ovarian carcinoma (10). and Beverly Sackler Faculty of Exact Sciences, Tel Aviv University, Tel Aviv, For non–small cell lung carcinoma (NSCLC), small-cell 3 Israel. Science for Life Laboratory, Drug Discovery and Development lung carcinoma (SCLC) as well as ovarian carcinoma, Platform, Department of Organic Chemistry, Stockholm University, Stock- holm, Sweden. 4Molecular Oncology Laboratory, Institute of Oncology prognosis is generally poor, which in part is explained Research (IOR), Bellinzona, Switzerland. 5Atrahasis S.r.l. Rome, Italy. by the fact that these tumor cells are highly resistant to 6Department of Zoology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv, Israel. known chemotherapeutic agents used in the disseminated stage of these diseases. For all these tumor entities failure Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/). to elicit apoptotic response as well as increased growth factor signaling both contribute to the chemotherapy Corresponding Author: Kristina Viktorsson, Karolinska Biomics Center, Z5:01, Department of Oncology-Pathology, Karolinska Institutet, Stock- refractory phenotype (13, 14). Hence, novel agents, which holm S-171 76, Sweden. Phone: 46-8-517-701 77; Fax: 46-8-517-710 00; can circumvent these signaling aberrations, are needed. E-mail: [email protected] In the course of our screening for antitumor agent from doi: 10.1158/1535-7163.MCT-14-0329 marine sponges (3S)-icos-4E-en-1-yn-3-ol (1) and (3S)-14- 2014 American Association for Cancer Research. methyldocos-4E-en-1-yn-3-ol (2) were isolated from

www.aacrjournals.org OF1

Zovko et al.

Caribbean sponge Cribrochalina vasculum (family Nipha- phase. In the last step of purification, the semi-pure frac- tidae, order Haplosclerida). Structure elucidation and tions from the second Sephadex LH-20 column (fractions literature search of these compounds revealed that they 8–11) were combined and separated on a reversed phase both were previously described (15–17) and reported to semipreparative YMC C-8 HPLC (high-performance liq- have antitumor activity, yet their mechanism of action for uid chromatography) column with a mixture 40:51:9 of induction of tumor cell death has not been described. acetonitrile:methanol:water as the isocratic mobile phase . Beside these compounds, several other related acetylenic Several products were obtained from this last purification alcohols were previously reported from the sponge C. and pure compounds (Supplementary Fig. S1B) were vasculum. Such related acetylenic alcohols were shown to tested for tumor-specific toxicity. Two pure substances exhibit immunosuppressive activity, toxicity toward (3S)-icos-4E-en-1-yn-3-ol (1)(tR 41.8 minutes; 70.1 mg, 95% brine shrimp, and also to possess antitumor activity purity) and (3S)-14-methyldocos-4E-en-1-yn-3-ol (2) (8, 15, 18, 19). Here, we show that compound 1 causes (tR 67.0 minutes, 109.0 mg, 99% purity) were found to strong cytotoxic activity in tumor cells from NSCLC, have the best therapeutic window, and were therefore SCLC, and ovarian carcinoma but not in normal and chosen for further analysis. primary cells tested. Yet we see differences also among The following characteristics were noted for com- the tumor cells with NSCLC being more sensitive than pounds 1 and 2: SCLC and ovarian carcinoma. For compound 2,acyto- (3S)-icos-4E-en-1-yn-3-ol (1): colorless oil, [a]D 1.5 (c e toxic effect was evident in NSCLC and SCLC whereas in 0.35, CHCl3); UV (MeOH) ( ) 204 (1400) nm; IR (ATR) ovarian carcinoma the difference compared with nor- 3,292; 2,917; 2,850 cm/L; for NMR data see Supplementary mal cells was less prominent. Nevertheless, in NSCLC, Table S1; GCMS SMB EI m/z 292.3 Mþ, C20H38O (19). we can demonstrate that both these compounds, isolat- (3S)-14-methyldocos-4E-en-1-yn-3-ol (2): Colorless oil; C. vasculum a e ed from sponge , activate intrinsic apoptotic [ ]D 6.4 (c 5.8, CHCl3); UV (MeOH) ( ) 204 (1300) nm; IR signaling cascades, that is, activation of Bak/Bax, (ATR) 3,303; 2,922; 2,852; 2098 cm/L; For NMR data see decreased phosphorylation of Bad, depolarization of Supplementary Table S2; GCMS SMB EI m/z 334.3 Mþ, mitochondria, cytochrome c release from mitochondria, C23H42O (19). and activation of caspases 9 and 3. Moreover, a decreased phosphorylation of Akt, mTOR, and ERK Cell culture and treatments C. vasculum growth factor signaling kinases was evident and an parental extract (fraction 10) and com- increased phosphorylation of JNK was observed. Thus, pounds 1 and 2 were diluted in DMSO to make 10 mg/ compound 1, and to some extent compound 2,hold mL stock solutions that were kept at 20 C and diluted in potential as therapeutic agents that should be further cell culture media before use. Data on stability of com- explored for different tumor malignancies with specific pound 2 after long-term storage in DMSO and in cell emphasis on NSCLC and SCLC. culture media after 72 hours incubation time at concentrations used in experiments can be found in Supplemen- tary Materials and Methods, sections stability of com- Materials and Methods pound 2 after long-term storage and stability of com- Collection and purification of extracts pound 2 in cell culture media. Samples of the sponge C. vasculum were collected in Key The human SCLC U-1285 (20) and NSCLC U-1810 (21) Largo, Florida in June 2011. Further details about the were kind gifts from Uppsala University where they were collection of the material can be found in Supplementary established and characterized (22). SCLC H69 and H82, Materials and Methods, section collection of sponges. The ovarian cancer cell lines A2780 and SKOV-3, foreskin freeze-dried sample (206 g) was extracted with 45:45:10 fibroblasts immortalized with hTERT BJ-5ta, bronchial mixture of ethyl acetate:methanol:water to yield crude epithelial cells BEAS-2B and hTERT immortalized retina extract (43 g). Fractionations of crude extract were guided epithelial cells RPE-1 were all purchased from the ATCC. by MTT cell viability assay (see cell viability assay below) The lung fibroblast cell line WI-38 (23) was obtained from and each fraction was tested for cell cytotoxicity in NSCLC Coriell Cell Line Repository. Human cardiomyocyte pri- U-1810 cells or in normal diploid fibroblasts WI-38 after 72 mary cell culture was acquired from Celprogen Inc. BJ-5ta hours of continuous exposure. The fraction that gave the was obtained in 2014, BEAS-2B and cardiomyocytes in best therapeutic window was chosen for further purifi- 2012, RPE-1 and SKOV-3 in 2011, A2780 in 2008, and U- cation. Thus, the crude extract was separated sequentially 1810, U-1285, H69, and H82 in 1996. Cell lines were by different chromatographic methods starting with authenticated by cell banks using the short tandem repeat reversed phase vacuum flash chromatography using a profiling. No authentication for these cell lines was done gradient of solvents, from 100% water through 100% by the authors. Upon receipt cultures were passaged for methanol and then 100% ethyl acetate (Supplementary not more than 2 months and aliquots were frozen. For Fig. S1A). The fraction 10 (90:10 mixture of methanol: experiments each cell line was thawed according to the water) obtained from the initial separation was further protocol and grown out for no more than 6 weeks for separated twice on a size exclusion Sephadex LH-20 tumor and 3 for normal cells before use in an experiment. column using chloroform:methanol (1:1 v/v) as mobile Human peripheral blood mononuclear cells (PBMC) were

OF2 Mol Cancer Ther; 13(12) December 2014 Molecular Cancer Therapeutics

C. vasculum Compounds Possess Antitumor Potential

isolated from buffy coats of healthy donors (provided by relative to the absorbance in untreated cells, which was set the Department of Clinical Immunology and Transfusion to 100%. Values shown are the mean of three replicates in Medicine, Karolinska Hospital, Stockholm, Sweden) by three independent experiments SEM. using a standard Ficoll-Hypaque gradient. The U-1810, H69, H82, U-1285, A2780 cells, and PBMCs Analysis of apoptotic morphology and cytochrome c were cultured at 37 C and 5% CO2 in RPMI-1640 medium release (Sigma-Aldrich) containing 2 mmol/L L-glutamine (Invi- NSCLC U-1810 and diploid fibroblasts WI-38 were trogen) and 10% heat-inactivated FBS (HyClone). The treated with 3 mmol/L of 1 and 2,20mmol/L cisplatin ovarian cancer cell line SKOV-3 was grown in McCoy’s or with equal volume of DMSO and harvested after 24, 5A medium (Sigma-Aldrich) containing 2 mmol/L L-glu- 48, or 72 hours. Approximately 40,000 cells were resus- tamine and 10% FBS. Lung fibroblasts WI-38 were main- pended in 100 mL PBS, cytospinned onto glass sides at 500 tained in Eagle’s Minimum Essential Medium (Sigma- RPM for 4 minutes and fixed in 4% paraformaldehyde for Aldrich) supplemented with 15% FBS and 2 mmol/L L- 5 minutes followed by ice-cold acetone for 10 seconds. glutamine. hTERT immortalized retina epithelial cells Glassslidesweredriedandkeptat20C until staining. (RPE-1) were cultured in DMEM/F12 (Lonza) with 10% To detect and quantify apoptotic morphology, U-1810 of FBS. Cardiomyocytes were grown in human cardio- and WI-38 were stained with 4,60diamidino-2-phenylin- myocytes cell culture complete growth media with serum dole dihydrochloride (DAPI; Vector Laboratories). Apo- and antibiotics using Celprogen’s human cardiomyocyte ptotic morphology was examined in a fluorescence cell culture extracellular matrix precoated flasks (Celpro- microscope Zeiss Axioplan 2 imaging microscope using gen Inc.). BJ-5ta was cultured in DMEM (HyClone) sup- Zeiss 403 lens and processed using the Axiovision soft- plemented with 2 mmol/L L-glutamine and 10% FBS. ware (Carl Zeiss MicroImaging). The presence of frag- BEAS-2B was grown in BEGM kit media (Lonza) and mented nuclei was defined as apoptotic feature and the flasks were precoated with a mixture of 0.01 mg/mL percentage of such cells in total of 400 cells examined are fibronectin (Sigma-Aldrich), 0.03 mg/mL bovine collagen presented. Cytochrome c release was examined in U- type I glutamine (Invitrogen), and 0.01 mg/mL BSA 1810 cells treated with either DMSO or 3 mmol/L of 1 or 2 (Sigma-Aldrich) dissolved in BEBM (Lonza) to allow for 16 hours. Blocking was carried out with buffer (3% proper growth. BSA, 0.2% TritonX100, 10 mmol/L Hepes, pH 7.4) during 60 minutes at room temperature and the slides were Cell viability assay subsequently incubated with primary antibody against The cytotoxicity induced by extracts or pure com- cytochrome c (1:200; Cell Signaling Technology) for 16 pounds was determined with the previously described hours at 4C, followed by three steps of washing in PBS 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazolium bro- and incubation with goat-anti-rabbit Alexa 488 second- mide (MTT) assay (24) with minor modifications. Cell ary antibody (Invitrogen). To visualize nuclei, slides viability experiments on tumor cells were carried out 24 were counterstained with DAPI (Vector Laboratories). hours after seeding, at a confluence of about 80%, whereas Staining was examined using a Zeiss Axioplan 2 imaging normal cells were completely confluent to mimic a non- microscope with Zeiss 403 lens and processed using the dividing state, which is their natural state in the body. For Axiovision software (Carl Zeiss MicroImaging). all cells, the amount of cells required were titrated out in preparatory experiment and the following amounts were Mitochondrial permeability seeded in each well of a 96-well plate in 90 mL of complete The loss of mitochondrial membrane potential upon growth media: 5,000 cells (U-1810, H82, SKOV-3 cell lines, treatment with DMSO or 3 mmol/L of 1 or 2 for 24 hours and cardiomyocytes), 7,500 cells (H69, A2780, and BJ-5ta), was assessed by staining the cells with tetramethylrho- 10,000 cells (U-1285), 18,000 cells (WI-38), 20,000 cells damine ethyl ester perchlorate (MitoPT TMRE; Immu- (BEAS-2B and RPE-1), and 400,000 (PBMCs). All technical nochemistry Technologies Bloomington) as previously replicates were done in triplicates and for each experiment been described (25). Briefly, cells were treated with the three biologic replicates were carried out. Different DMSO or 3 mmol/L of 1 or 2 for 24 hours. As a positive concentrations of parental extract and 1, 2 or equal control for depolarization of mitochondria, the proton volumes of DMSO (v/v; negative control) diluted in fresh gradient uncoupling agent carbonyl cyanide 3-chloro- medium were added into cells followed by incubation for phenylhydrazone (CCCP; 50 mmol/L, 45 minutes) was 24, 48, or 72 hours. At the end of the experiment, MTT used. After harvesting, cells were stained with TMRE solution (Sigma-Aldrich) was added (0.5 mg/mL) and the (150 nmol/L) for 20 minutes in the dark. Cells were plates were incubated for 4 hours at 37 C and 5% CO2. centrifuged and resuspended in 1X Assay buffer pro- Thereafter, formed formazan crystals were dissolved in vided by the MitoPT TMRE Staining Kit. Cells (10,000/ stop solution (10% SDS and 0.01 mol/L HCl) and absor- sample) were analyzed on the FL-2 channel of Calibur bance was measured at 595 nm. The resulting absorbance, flow cytometer (BD Biosciences). The percentage of cells which is proportional to the number of viable cells in each that lost their TMRE staining was gated and quantified well, was recorded and the percentage of cell survival was and values shown are the mean of three independent determined by comparing the absorbance in treated cells experiments SD.

www.aacrjournals.org Mol Cancer Ther; 13(12) December 2014 OF3

Zovko et al.

Analysis of proapoptotic conformational changes of methanol. After blocking nonspecific binding with the Bak and Bax Odyssey blocking buffer (Li-Cor Biosciences; Bad Hom- During apoptosis, the proapoptotic Bcl-2 family pro- burg Germany) diluted in PBS (1:1), membranes were teins Bax and Bak undergo N-terminal conformational probed with primary antibodies overnight at 4C and changes leading to a proapoptotic state required for for- with secondary antibody for 1 hour at room temperature. mation of pores in the mitochondria and subsequent Following primary antibodies diluted in blocking cytochrome c release (26, 27). To analyze N-terminal buffer were used: phospho-Akt (Thr308), phospho-Akt changes in Bak or Bax in response to treatment, NSCLC (Ser473), Akt, phospho-Bad (Ser112), Bad, Bak, Bax, Bcl- U-1810 cells were treated with DMSO or 3 mmol/L of 1 or 2 xL, caspase-3, caspase-9, phospho-mTOR (Ser2448), for 16 hours. Cells were harvested and fixed in 4% para- phospho-p44/42MAPK(Erk1/2; Thr202/Tyr204), p44/ formaldehyde for 10 minutes and washed in PBS. Fixed 42MAPK (Erk1/2; Cell Signaling Technology); JNK cells were stained with antibodies recognizing N-terminal (FL), PARP (H250; Santa Cruz Biotechnology), and activation-related conformational changes in either Bak phospho-JNK1/JNK2 (Thr183/Tyr185; Abcam). Where (AM03, clone TC100; Oncogene Research Products) or Bax indicated, antibodies against b-tubulin, b-actin (Abcam), (clone 6A7; BD Biosciences Pharmingen). For staining or GAPDH (Trevigen) were used as control of equal cells were resuspended in 100 mL of primary antibody loading. IR-Dye–linked secondary antibodies (LI-COR diluted in PBS (Bak 1:50 and Bax 1:250) containing digi- Biosciences) were used to image bands on the Odyssey tonin (100 mg/mL) as permeabilizing agent. After 1 hour platform. incubation at 4C, samples were washed in PBS and incubated with secondary antibody Alexa 488–labeled Cell-cycle distribution anti-mouse antibody for 30 minutes at 4C in the dark to To assess cell-cycle distribution, U-1810 and WI-38 cells m enable monitoring by flow cytometry. Before FACS anal- were treated with 1.3 mol/L (IC50 in U-1810 at 24 hours), m ysis, samples were washed in PBS and resuspended in 200 3.4, 6.8, and 16.2 mol/L (IC50 in WI-38 at 24 hours) of mL of PBS. The Bak- or Bax-associated IFL signal was compound 1 or an equal volume of DMSO for 24 hours measured in the FL-1 channel of the FACS Calibur flow and fixed in 70% ethanol. Cells were stained with propi- cytometer (BD Biosciences). Data were processed using dium iodide as described previously (28). Signals were Cell Quest software (BD Biosciences) and change in fluo- recorded using FACSCalibur (BD Biosciences) and ana- rescence intensity was determined by comparing the lyzed with ModFit LT (Verity Software House). percentage of gated cells after treatment with that of untreated cells, which was set to 100%. Data shown are Results the mean values of three independent experiments SD. Isolation of antitumor extracts from C. vasculum Several compounds with antitumor activity have been Western blot analysis isolatedfrommarine-derivedorganismsandsomeofthem NSCLC U-1810 cells and WI-38 fibroblasts were seeded are already in clinical use (8–12). Here, we set out to isolate in 10-cm cell culture dishes at density of 1.2 million cells andcharacterizeantitumorcompoundsfromthespongeC. for U-1810 and 2 million for WI-38. After 24 hours, cells vasculum with aim to reveal their antitumor mechanism of were treated for 4, 16, 24, 48, and 72 hours with 3 mmol/L 1 action. For that purpose material was collected and or 2 or with equal volume of DMSO. In experiment in extracted as described in Materials and Methods. Result- which apoptotic induction was analyzed by PARP-1 and ing fractions were tested for antitumor activity against caspase cleavage, cells were treated for 24 hours with NSCLC U-1810 cells and for normal cell cytotoxicity using concentration of compounds that induce concentration of diploid fibroblasts WI-38 during 72 hours continuous compounds that induce 50% or 70% of growth inhibition exposure. Fraction 10 (named parental extract) from the in U-1810 or in WI-38 cells ( IC50 or IC70, respectively). C. vasculum extract induced significant cytotoxicity of Cells were harvested with cell dissociation solution NSCLC cells and almost completely inhibited cell growth (Sigma-Aldrich). Whole-cell lysates were prepared in at a concentration of 0.5 mg/mL, whereas only minor RIPA buffer (50 mmol/L Tris-HCl, pH 7.4, 150 mmol/L cytotoxicity was evident in a normal diploid fibroblasts NaCl, 0.5% Igepal, 5 mmol/L EDTA, and 0.1% SDS) WI-38 up to a concentration of 25 mg/mL (Fig. 1A and B). supplemented with protease and phosphatase inhibitor Given the observed NSCLC cell–specific cytotoxicity, we cocktail tablets (Roche Diagnostics AB). Protein concen- continued the purification of this fraction as outlined in tration was determined by BCA Protein Assay Reagent Materials and Methods and Supplementary Fig. S1A. (Interchim; Montiucon Cedex). Total protein (30–40 mg) of each extract was loaded to Bis Tris 4% to 12% or Tris Chemical structure elucidation of C. vasculum Acetate 3% to 8% gels (NuPAGE; Invitrogen) and electro- antitumor active fractions identifies (3S)-icos-4E-en- phoresed at 200 V for 60 minutes in MES or MOPS running 1-yn-3-ol (1) and (3S)-14-methyldocos-4E-en-1-yn- buffer (NuPAGE; Invitrogen). Transfer to polyvinylidene 3-ol (2) fluoride membranes (Hybond-C Extra; Amersham Bios- HPLC fractionation of the parental extract from C. ciences) was performed at 30 V for 90 minutes in transfer vasculum (Supplementary Fig. S1B) yielded two known buffer (NuPAGE; Invitrogen) supplemented with 10% alkyl-4-en-1-yn-3-ol derivatives; 70.1 mg (0.035% of

OF4 Mol Cancer Ther; 13(12) December 2014 Molecular Cancer Therapeutics

C. vasculum Compounds Possess Antitumor Potential

Figure 1. C. vasculum semi-pure fraction shows specific tumor cell cytotoxicity and generates two pure compounds. A, fraction 10 [crude mixture of (3S)-alkyl- 4E-en-1-yn-3-ol, named parental] after three fractionation steps of sponge C. vasculum extract was tested for capacity to induce tumor-specific cytotoxicity in NSCLC U-1810 cells or in diploid fibroblasts WI-38 after 72 hours of continuous exposure to indicated concentrations (phase-contrast images, 10 magnification; scale bar, 0.5 mm). B, cells were treated as in A and cell viability was examined using MTT cell viability assay. Cell viability is given as the percentage of cell survival as compared with DMSO solvent–treated cells. Data, mean of three independent experiments SEM. C, structure of (3S)-icos-4E- en-1-yn-3-ol (1) and (3S)-14-methyldocos-4E-en-1-yn-3-ol (2). sponge dry weight) of 95% pure (3S)-icos-4E-en-1-yn-3- in NSCLC U-1810 cells but not in normal diploid fibro- ol (1) and 109 mg (0.055% of sponge dry weight) of 99% blasts WI-38 until 3 mmol/L was used (Fig. 2A). A clear pure (3S)-14-methyldocos-4E-en-1-yn-3-ol (2) (ref. 19). cytotoxicity was also evident when treating the NSCLC The structures of 1 and 2 were confirmed by 1D and 2D U-1810 cells with different concentrations of 1 and 2 for NMR techniques. EIMS and MS-MS techniques were 72 hours and examining cell viability with MTT assay used to elucidate the position of the branched methyl (Fig.2B).Thus,atconcentrationof3mmol/L1and2 substituent in compound 2 (Supplementary Fig. S1C; reduced U-1810 cell viability for about 90% and 60%, ref. 19). The absolute stereochemistry of 1 and 2 was respectively. Importantly, applying the same concen- determined by the Mosher method (29) because their tration of these compounds to WI-38 fibroblasts caused optical rotations were small. Analysis showed that little or no cytoxicity, suggesting a tumor-selective win- compounds 1 and 2 are structurally related acetylenic dow for these compounds. Thus, the concentration of alcohols both containing a C4-C5 trans double bond and compounds 1 and 2 needed to inhibit 50% of cell thesameabsoluteconfigurationofthehydroxylgroup. viability (IC50) was more than 10 times higher for WI- The structure determination also showed that the com- 38 than for the U-1810 cells (Supplementary Table S4). pounds differed in the length of their carbon chains and The effect of 1 and 2 on cell cytotoxicity was examined that 2 has a branched methyl substituent at position 14 also at 24 and 48 hours after treatment in U-1810 and (Fig. 1C). WI-38 (Supplementary Fig. S2). A clear time-dependent increase in the cytotoxic effect of both compounds C. vasculum compounds 1 and 2 have tumor-specific toward U-1810 cells was evident with a higher effect cytotoxicity seen on the later time point. In contrast, neither 1 nor 2 We next continued with the acetylene containing impaired survival of diploid fibroblasts WI-38 at these compounds 1 and 2 to reveal their antitumor effect and earlier time points at doses up to 15 and 30 mmol/L, delineate their mechanism of action. Similar to the respectively (Supplementary Fig. S2 and Supplementa- parental extract, compounds 1 and 2 induced toxicity ry Table S3).

www.aacrjournals.org Mol Cancer Ther; 13(12) December 2014 OF5

Zovko et al.

Figure 2. Pure compounds isolated NSCLC Diploid fibroblasts from C. vasculum extract A B U-1810 WI-38 demonstrate high tumor-selective Compound 1 Compound 1 Compound 2 Compound 2 cytotoxicity in NSCLC and SCLC DMSO Compound 1 Compound 2 120 cells. A, either of the two isolated 100 1 2 NSCLC compounds ( and )ata m U-1810 80 concentration of 3 mol/L or equal volume of DMSO solvent was 60 added to NSCLC U-1810 and normal diploid fibroblasts WI-38 for 40 Diploid 72 hours. Phase-contrast pictures % Cell survival Cell % fibroblasts 20 showing significant toxicity in WI-38 tumor but not normal cells 0 fi 0 1 2 3 4 6 8 10 12 1520 (magni cation 20; scale bar, 0.25 Compound [μmol/L] mm). B, the compounds 1 and 2 were profiled for their cytotoxic C capacity in the same cells as in A at SCLC U-1285 SCLC H69 SCLC H82 OC A2780 OC SKOV-3 72 hours of continuous exposure Diploid fibroblasts WI-38 using MTT cell viability assay. Cell 120 120 viability is given as the percentage of cell survival as compared with 100 100 DMSO solvent–treated cells. Data, mean of three independent 80 80 experiments SEM. The IC30,IC50, 60 60 andIC70 concentrationsaregivenin Supplementary Table S3. C, the 40 40 cytotoxicityof1 and2 after72 hours % Cell% survival % Cell survival of treatment was examined in 20 20 additional tumor cell lines: SCLC (U-1285,H69,and H82) and ovarian 0 0 0 1 2 3 46 8 17 0 1 2 3 4 5 6 8 10 15 cell lines (A2780 and SKOV-3). Compound 1 [μmol/L] Compound 2 [μmol/L] Normal diploid fibroblasts (WI-38) were included for comparison. Cell viability was examined as in B. The D PBMC BEAS-2B Cardiomyocytes IC30,IC50, and IC70 concentrations RPE-1 BJ-5ta Diploid fibroblasts WI-38 are given in Supplementary Table 120 120 S4. D, the cytotoxicity of 1 and 2 after 72 hours of treatment was 100 100 examined in normal cells: human PBMCs, bronchial epithelial 80 80 cells (BEAS-2B), human cardiomyocytes, hTERT 60 60 immortalized retina epithelial cells 40 40 (RPE-1), and foreskin fibroblasts % Cell% survival % Cell survival Cell % immortalized with hTERT (BJ-5ta). 20 20 Normal diploid fibroblasts (WI-38) were included for comparison. Cell 0 0 viability is given as in B. The IC , 0 1 2 3 4 5 10 15 20 0 1 2 3 4 5 10 15 20 30 Compound 1 [μmol/L] Compound 2 [μmol/L] IC50, and IC70 concentrations are given in Supplementary Table S4.

C. vasculum compounds 1 and 2 induce prominent compound 1 was slightly lower for U-1285 and H69, yet cytotoxicity in SCLC but not in normal cells at 3 mmol/L both these cell lines showed an 80% Next, we extended our analyses onto three SCLC cell decrease in cell viability and the corresponding IC50 lines (U-1285, H69, and H82) and two ovarian cell lines values were 1.6 and 2.2 mmol/L, respectively (Supple- (SKOV-3 and A2780; Fig. 2C; Supplementary Table S4). mentary Table S4). For compound 2 U-1285, H69, and To asses normal cell toxicity human cardiomyocytes, H82 responded with about 80% decrease in cell viability human PBMCs, foreskin ﬁbroblasts immortalized with upon treatment with 3 mmol/L concentration (Fig. 2C, m hTERT (BJ-5ta), bronchial epithelial cells (BEAS-2B), right). IC50 values were 1.8 mol/L for U-1285, 1.3 andhTERTimmortalizedretinaepithelialcells(RPE- mmol/L for H69, and 1.1 mmol/L for H82 (Supplemen- 1) were used in the analyses (Fig. 2D; Supplementary tary Table S4). In ovarian carcinoma cells, compound Table S4). 1 caused a similar 80% reduction in cell viability at Of the SCLC cell lines, H82 showed the highest 3 mmol/LasseeninSCLCwhereasforcompound2 sensitivity toward compound 1 with an IC50 value of both A2780 and SKOV-3 cells were less responsive (Fig. m m 1.1 mol/L and with an almost complete inhibition of 2C). Compound 1 reached IC50 in A2780 at 1.8 mol/L cell survival at 1.5 mmol/L (Fig. 2C, left). The toxicity of and in SKOV-3 at 2.1 mmol/L. Toxicity of compound

OF6 Mol Cancer Ther; 13(12) December 2014 Molecular Cancer Therapeutics

C. vasculum Compounds Possess Antitumor Potential

A DMSO Compound 1 Compound 2 100 * 80

60 * * * 40 *

% Apoptotic cells 20 *

0 24 48 72 Hours after treatment

NSCLC Diploid fibroblasts B U-1810 WI-38

G –M 100 **** 2 100 S

75 G1 75

50 50 * % Cells % Cells % 25 * * * 25

0 0 DMSO 1.3 3.4 6.8 16.2 μmol/L DMSO 1.3 3.4 6.8 16.2 μmol/L

Figure 3. Compounds 1 and 2 induce apoptosis and cell-cycle arrest in NSCLC cells but not in diploid fibroblasts. A, NSCLC U-1810 cells or diploid fibroblasts WI-38 were treated with DMSO, 1 or 2 (3 mmol/L) for 24, 48, and 72 hours. Cisplatin (20 mmol/L, 72 hours)-treated U-1810 cells were used as a positive control for induction of apoptotic morphology. Nuclear morphology of cells was examined after fixation in 4% paraformaldehyde and staining with DAPI. Pictures showing nuclear morphology are given in Supplementary Fig. S3A. The percentage of cells displaying apoptotic morphology was quantified in 400 cells in three independent experiments using fluorescence microscopy. Data, mean; bars, SD; , P < 0.05. B, flow cytometric examination of the cell-cycle distribution in U-1810 (left) and WI-38 (right) cells treated with 1.3, 3.4, 6.8, and 16.2 mmol/L concentration of 1 or equal volume of DMSO for 24 hours. The percentage of cells in G1 phase, S phase, and G2–M phase are given. Data, mean of three independent experiments SD; , P < 0.001.

m 2 toward ovarian carcinoma was not so prominent, IC50 PBCM and RPE-1 (8.5 mol/LforPBMCand10.4for was reached only at 3.4 mmol/L in A2780 and 4.9 mmol/ RPE-1; Supplementary Table S4). L in SKOV-3 cell lines (Supplementary Table S4). Taken In summary, the cytotoxic effect of compound 1 was together, compounds 1 and 2 both demonstrate toxicity significantly lower in all normal cells than in tumor cells, toward SCLC, whereas for ovarian carcinoma com- whereas for compound 2 the tumor-selective effect was pound 1 has a greater potency than compound 2. less prominent but evident in NSCLC and SCLC. Both compounds were tested against several normal cells. Treatment with 3 mmol/L concentration of com- C. vasculum compounds 1 and 2 induce apoptosis in pound 1, which caused high cytotoxicity in all tumor NSCLC tumor cells but not in diploid fibroblasts cell lines tested, induced only 20% reduction in cell To further understand mechanism of action of these C. viability in cardiomyocytes and BEAS-2B whereas in vasculum compounds, their effect on apoptosis induction PBCM and BJ-5ta about 40% cytotoxicity was evident was examined in NSCLC cells (Supplementary Fig. S3A (Fig. 2D, left). With the applied concentrations, the IC50 and Fig. 3A). Treatment of NSCLC U-1810 cells with either value could not be reached for cardiomyocytes and 1 or 2 (3 mmol/L, 48 hours) or cisplatin (20 umol/L, 72 BJ-5ta whereas for RPE-1 it was 6.6 mmol/L, PBMC hours) caused a prominent induction of apoptotic mor- 7.8 mmol/L, and for BEAS-2B 12.9 mmol/L (Supplemen- phology of the cell nuclei, whereas no major alteration was tary Table S4). Compound 2 at 3 mmol/L, a concentra- observed in diploid fibroblasts WI-38 upon treatment tion that caused at least a 50% cell cytotoxicity in LC (Supplementary Fig. S3A). Induction of apoptotic mor- cells induced less than 15% cytotoxicity in cardiomyo- phology in U-1810 cells was also quantified at 24 to 72 cytes, BJ-5ta, BEAS-2B, and RPE-1 whereas in PBMC hours after treatment with 1 or 2 (3 mmol/L; Fig. 3A). abouta35%cytotoxiceffectwasevident(Fig.2D,right). Already after 24 hours, 1 induced apoptotic morphology With compound 2,theIC50 value was reached only for in about 30% of the cells with a further increase after 48

www.aacrjournals.org Mol Cancer Ther; 13(12) December 2014 OF7

Zovko et al.

hours and with 80% of the cells showing apoptotic nuclear Next, the release of cytochrome c after treatment of morphology at 72 hours after treatment (Fig. 3A). Com- NSCLC U-1810 cells with 1 and 2 (3 mmol/L) for 16 hours pound 2–induced apoptosis was delayed as compared was analyzed by immunofluorescence staining (Fig. 4B). with 1 but nevertheless 48 hours after treatment, about A punctate staining pattern of cytochrome c indicative of 40% of the U-1810 cells had fragmented nuclei and at 72 mitochondrial localization was found in untreated cells, hours almost 60% showed a prominent apoptotic whereas upon treatment with 1 or 2 diffuse staining response (Fig. 3A). Both compounds induced less than pattern indicating release of cytochrome from mitochon- 2% apoptosis in fibroblasts treated for 72 hours (data not dria to cytosol was evident (Fig. 4B). shown). Cisplatin, which was used as a positive control, To study whether depolarization of mitochondria and induced apoptotic morphology in 26% of U-1810 cells release of cytochrome c in response to 1 or 2 also resulted showing that these cells are capable to undergo apoptosis in the downstream activation of caspase 9/3 and subse- (data not shown). quent apoptotic-associated cleavage of the caspase-3 sub- strate PARP-1 NSCLC U-1810 cells were treated with 1 or 2 vasculum C. compound 1 induces G2 arrest in NSCLC for 24 hours at a concentration that caused 50% and 70% of but not in normal diploid fibroblasts cell death, respectively (Fig. 4C). As can be seen in U-1810 Loss of proliferation capacity and induction of cyto- cells both caspases 9 and 3 were cleaved into active forms toxicity may also be attributed to inhibition of cell-cycle (35–37 and 15–17 kDa, respectively) upon treatment with progression. The effect of 1 on cell-cycle distribution IC50 or IC70 of either compound. Treatment also resulted was, therefore, also evaluated (Fig. 3B). NSCLC U-1810 in clear cleavage of PARP-1 into the proapoptotic-associ- or WI-38 fibroblasts were treated with different concen- ated 89 kDa fragment confirming that these compounds trations of 1 for 24 hours. In NSCLC U-1810 cells, indeed triggered an apoptotic response (Fig. 4C). Normal treatment with 1 was found to induce a significant fibroblasts WI-38 were used for comparison and treated dose-dependent increase in the number of cells arrested with same concentrations as U-1810 cells. Importantly, in the G2–M phase (Fig. 3B, left graph). In contrast, cell- under these conditions no caspase-3 or PARP-1 cleavage cycle distribution of normal fibroblasts showed no alter- was evident in normal WI-38 fibroblasts (Fig. 4C). How- ation after treatment with compound 1 even at IC50 ever, the lower proapoptotic activity of compounds concentration (Fig. 3B, right graph). Thus, the results observed in fibroblasts could be due to cell permeability showed that compound 1 clearly induces cell-cycle rather than to a difference in actual mode of cell kill. To see arrest in G2–M in NSCLC tumor cells and this is a whether these compounds at all were able to induce tumor-specific effect. apoptosis in these normal lung fibroblasts, concentrations that induced 50% and 70% kill of fibroblasts at 24 hours C. vasculum compounds 1 and 2 induce mitochondrial were applied. Even at these concentrations, no PARP-1 or depolarization, activate caspases 9 and 3, and cause caspases cleavage was observed in the fibroblast further PARP cleavage in NSCLC cells but not normal diploid demonstrating a NSCLC cell–specific proapoptotic activ- fibroblasts ity of these compounds (Fig. 4D). One important gateway to apoptosis is mitochondrial We also examined the kinetic of caspases and PARP-1 depolarization allowing cytochrome c to be released to cleavage after treatment with 3 mmol/L of 2 for 24, 48, and the cytosol where it subsequently causes the pro–cas- 72 hours (Supplementary Fig. S3B). In line with MTT and pase-9 to be activated within the apoptosome complex. nuclear apoptotic morphology data, cleaved versus full- Caspase-9 in turn triggers cleavage of pro–caspase 3/7 length caspase-3, caspase-9, and PARP-1 all clearly into active proteases, thereby giving rise to apoptotic showed an increase over time in response to 2. Thus, there morphology (30). The effect of 1 and 2 on depolarization is a clear NSCLC-specific activation of apoptosis by both of mitochondria and release of cytochrome c were, compounds. therefore, examined. NSCLC U-1810 cells were treated with compound 1 or 2 (3 mmol/L) for 24 hours and C. vasculum compounds 1 and 2 trigger activation of mitochondrial depolarization was examined with Bak and Bax and JNK signaling and inhibit PI3K/Akt TMRE (Fig. 4A). Upon membrane depolarization, the and MAPK ERK survival signaling in NSCLC cells potential of mitochondria is lost and the TMRE dye Activation of the Bcl-2 proteins Bak and Bax is one leaks into cytosol. Thus, cells that have a decrease in important gateway for the intrinsic mitochondrial path- TMRE staining represent cells with depolarized mito- way of apoptosis (26, 27). The effect of compounds 1 and chondria and these cells were quantified. The mitochon- 2 on Bax and Bak conformational changes associated with drial uncoupler CCCP (50 mmol/L, 45 minutes) was activation was, therefore, analyzed (Fig. 5A and B). used as positive control showing that TMRE is capable Indeed, treatment of NSCLC U-1810 cells with 3 mmol/ of detecting alterations in mitochondrial potential. Lof1 or 2 resulted in prominent activation of Bak indi- Importantly, treatment with either compounds resulted cated by a shift of histogram peak to the right (Fig. 5A and in prominent mitochondrial depolarization. Hence, B, right). An activation of Bax was also observed after both compounds were capable of disrupting mitochon- treatment of NSCLC U-1810 cells with either 3 mmol/L of 1 drial membrane potential. or 2, yet less pronounced than for Bak (Fig. 5A and B, left).

OF8 Mol Cancer Ther; 13(12) December 2014 Molecular Cancer Therapeutics

C. vasculum Compounds Possess Antitumor Potential

A TMRE: CCCP 1.0

* 0.5

TMRE: Compound 1 * (3 μmol/L) * Fold TMRE staining 0.0

Counts DMSO CCCP Comp 1 Comp 2 (3 μmol/L)

B (3 μmol/L) TMRE: Compound 2 (3 μmol/L) DMSO Compound 1 Compound 2

FL 2 TMRE

C NSCLC Diploid fibroblasts D Diploid fibroblasts U-1810 WI-38 WI-38 Comp. 1 Comp. 2 Comp. 1 Comp. 2 Comp. 1 Comp. 2 DMSO IC IC IC IC DMSO IC IC IC IC (In U-1810) DMSO IC IC IC IC (In WI-38) 35 kDa Caspase-3 (FL) 35 kDa Caspase-3 (FL)

17 kDa 17 kDa 15 kDa Caspase-3 (CL) 15 kDa Caspase-3 (CL)

50 kDa β-Tubulin 50 kDa β-Tubulin 47 kDa Caspase-9 (FL) 47 kDa Caspase-9 (FL) 37 kDa 37 kDa 35 kDa Caspase-9 (CL) 35 kDa Caspase-9 (CL)

116 kDa PARP-1 (FL) 116 kDa PARP-1 (FL)

89 kDa PARP-1 (CL) 89 kDa PARP-1 (CL) 50 kDa β-Tubulin 50 kDa β-Tubulin

Figure 4. Compounds 1 and 2 induce depolarization of mitochondria, cytochrome c release, and activation of caspases in NSCLC cells but not in diploid fibroblasts. A, NSCLC U-1810 cells were treated with 3 mmol/L 1 or 2 for 24 hours. CCCP (50 mmol/L) was added to cells 45 minutes before staining as positive control. Depolarization of mitochondria was examined by staining with 150 nmol/L TMRE for 20 minutes in media. Left, the signal was recorded in the FL-2 channel on FACSCalibur. Histogram showing TMRE-associated fluorescence. Filled gray, DMSO; solid line, CCCP, 1 or 2 treated cells. Right, quantification of cells that lost their TMRE signals (i.e., depolarized mitochondria). Data, mean with bars representing SD; , P < 0.05. B, cytochrome c release was examined by immunofluorescence staining of U-1810 cells after treatment with either DMSO or 3 mmol/L of 1 or 2 for 16 hours (magnification 100; scale bar, 15 mm). Blue, DAPI staining of nucleus; green, cytochrome c. C, NSCLC U-1810 cells or diploid fibroblasts WI-38 were treated with compound 1 or 2 at concentrations that induced IC50 and IC70 in U-1810 cells at 24 hours. Cleavage of caspase-3, caspase-9, or PARP-1 was examined after 24 hours. b-Tubulin was used to visualize equal loading. D, diploid fibroblasts WI-38 were treated with 1 or 2 at concentrations that at 24 hours inhibited growth to 50% or 70% (i.e., IC50 and IC70), respectively. Cleavage of caspase-3, caspase-9, or PARP-1 was examined after 24 hours. b-Tubulin was used to visualize equal loading.

The total expression of Bax and Bak was also examined c release (31). The expression level of Bcl-xL was, there- using conformation-independent antibodies in Western fore, examined after treatment with compound 1 or 2 and blot analysis (Fig. 5C). However, no changes in their a clear decrease in Bcl-xL expression was evident after expression levels were evident upon treatment, indicating treatment with 1, whereas no change was observed after that the observed Bak/Bax activation upon 1 or 2 relates to treatment with 2 (Fig. 5C). All in all, these results show N-terminal conformational changes of these proteins. that both 1 and 2 can activate the Bak/Bax rheostat in The antiapoptotic Bcl-xL protein is reported to inhibit NSCLC cells, which is in line with the observed depolar- Bak/Bax complex formation and subsequent cytochrome ization of mitochondria, release of cytochrome c, and

www.aacrjournals.org Mol Cancer Ther; 13(12) December 2014 OF9

Zovko et al.

A B BAX BAK Bax: Compound 1 (3 μmol/L) P = 0.00002 6.0 P = 0.02 28 P = 0.009 5.5 P = 0.002 26 24 5.0 22 4.5 20 4.0 18 3.5 16 Figure 5. Compounds 1 and 2 Bax: Compound 2 14 3.0 (3 μmol/L) 12 activate proapoptotic Bcl-2 family 2.5 10 members and JNK signaling, 2.0 8 1.5 6 decrease Bad and Akt 1.0 4 phosphorylation, and inhibit Fold fluorescence intensity Bak Fold fluorescence intensity Bax 2 0.5 MAPK-ERK survival signaling in Counts DMSO Comp.1 Comp.2 Bak: Compound 1 DMSO Comp.1 Comp.2 NSCLC cells. A, conformational μ (3 mol/L) C Compound 1 Compound 2 changes in Bax and Bak in 1 2 DMSO 4 16 24 4 16 24 HOURS response to or were examined in NSCLC U-1810. Histograms Bcl-xL showing Bax and Bak activation. GAPDH Filled gray, DMSO; solid line, 1 or 2 Bak: Compound 2 Phospho-Bad (Ser112) treated cells. B, quantiﬁcation of (3 μmol/L) GAPDH Bax- (left) and Bak- (right) associated IFL. Data, mean of three Bad independent experiments SD. C, β -Tubulin NSCLC U-1810 cells were treated Bax with 3 mmol/L of either 1 or 2 for 4, FL Bax/Bak β-Tubulin 16, and 24 hours or with equal volume of DMSO for 24 hours and Bak Bcl-xL, Bax, Bak, and Bad β D -Tubulin expression and Bad phosphorylation at Ser112 were Compound 1 Compound 2 examined by Western blot analysis. b DMSO 4 16 24 4 16 24 HOURS GAPDH or -tubulin served as Phospho-Akt (Ser473) loading controls. D, deactivation of β-Actin AKT on phosphoresidues Ser473, Thr308, mTOR Ser2448, p42/ Phospho-Akt (Thr308) 44MAPK(Erk1/2) on Thr202/ β-Actin Tyr204, and JNK1/JNK2 on Thr183/Tyr185 and total forms of Akt AKT, p42/44MAPK(Erk1/2) ERK, β-Tubulin and JNK1/JNK2 were examined at Phospho-mTOR (Ser2448) 4, 16, and 24 hours after addition of β-Actin 1 or 2 or 24 hours treatment with equal volume of DMSO. GAPDH or Phospho-p42/44MAPK(Erk1/2) (Thr202/Tyr204) b-actin were used as loading β-Actin controls. p42/44MAPK(Erk1/2) β-Actin

Phospho-JNK1/JNK2 (Thr183/Tyr185) GAPDH JNK GAPDH

activation of caspases via the intrinsic route after treat- with 2, a moderate decrease in Bad phosphorylation was ment with these compounds. observed only at 24 hours (Fig. 5C). The observed alteration Bak and Bax complex formation is in part regulated by on Bad Ser112 in response to 1 was conﬁrmed to be a result the BH3-only protein Bad, which upon phosphorylation is of altered signaling as no difference in total Bad protein sequestered by binding to 14-3-3 protein preventing its expression was evident upon treatment (Fig. 5C). Thus, proapoptotic action (32). The effect of 1 or 2 in causing both compounds partly inhibit Bad phosphorylation, indi- dephosphorylation of Bad was, therefore, examined. cating a putative proapoptotic action of Bad at mitochon- Indeed a decreased phosphorylation of Bad at Ser112 was dria-mediated apoptotic signaling in NSCLC cells. clearly evident in response to treatment with 1 already at Several growth factor signaling cascades are reported to 16 hours after treatment and still evident at 24 hours in control Bad phosphorylation, thereby inhibiting its proa- NSCLC U-1810 cells (Fig. 5C). In response to treatment poptotic signaling capacity. The effect of 1 and 2 on the

OF10 Mol Cancer Ther; 13(12) December 2014 Molecular Cancer Therapeutics

C. vasculum Compounds Possess Antitumor Potential

phosphorylation status of the Bad-regulating kinases Akt, which tumor but not normal cells rely upon on for pro- Erk1/2 as well as on JNK, a kinase shown to promote tection against apoptotic cell death. apoptotic signaling (13, 33) was, therefore, evaluated. A At an early stage of the intrinsic apoptotic pathway, the decreased phosphorylation of Akt at both Ser473 and proapoptotic Bcl-2 family proteins Bak and Bax are acti- Thr308 was evident in NSCLC U-1810 cells after treatment vated leading to permeabilization of the mitochondrial with compound 1 or 2 with an earlier response seen with 1 membrane, release of cytochrome c, and activation of (Fig. 5D). Deactivation of mTOR was also evident after 1 as caspases within the apoptosome (34, 35). We demonstrate the phosphorylation on Ser2448 declined (Fig. 5D). These that both 1 and 2 cause all these events in NSCLC cells. data suggest that 1 and 2 impair PI3K/AKT pathway However, a more prominent activation of Bak over Bax signaling and may in this way inhibit proliferation and was evident in these NSCLC cells. This is likely attributed survival of NSCLC cells. A decreased phosphorylation on to the basic signaling propensity of these NSCLC cells Thr202/Tyr304 of ERK1/2 was also evident after treat- rather than specific action of these compounds to activa- ment with both 1 and 2 again with a more prominent tion of Bak preferentially over Bax. This conclusion is deactivation caused by compound 1 (Fig. 5D). Hence, a based on the fact that we showed similar less pronounced decreased prosurvival signaling via the MAPK–ERK may activation of Bax also after treatment with other agents, for also contribute to the observed cytotoxic activity of these example, cisplatin and radiation in these NSCLC cells (25). two compounds in NSCLC cells. A sustained increase in When comparing 1 and 2 with respect to Bak activation JNK phosphorylation is reported to be critical for efficient and mitochondria-mediated apoptotic signaling, com- induction of apoptotic signaling (33). In line with this 1 pound 2 was found to be less efficient. Compound 1 is, caused increased phosphorylation of JNK1/JNK2 at 16 at the applied concentration 3 mmol/L, more toxic toward and 24 hours, whereas in response to 2 a similar increase in U-1810 cells than is equimolar concentration of compound phosphorylation was seen only at 24 hours in these 2, likely explaining the more pronounced Bak and Bax NSCLC cells. Thus, this increased JNK phosphorylation activation and cytochrome c release. There was also a less may in part contribute to the observed apoptotic signaling efficient decrease of Bcl-xL protein expression upon expo- in response to these compounds in this tumor type. sure to 3 mmol/L of compound 2. The presence of Bcl-xL could sequester Bak in an inactive state, thereby prevent- Discussion ing cytochrome c release and proper intrinsic route to caspase activation (34). We demonstrate that both 1 and 2 C. vasculum Here, we analyzed the ability of –derived subsequently trigger the cleavage and activation of cas- metabolites to inhibit proliferation, induce cytotoxicity, pase-9 and thereafter also caspase-3 showing that a full and trigger apoptotic signaling in NSCLC, SCLC, and apoptotic signaling cascade is active in these NSCLC ovarian carcinoma. It has previously been shown that tumor cells but not in normal diploid fibroblasts. At their compounds isolated from this sponge can kill tumor cells IC50 and IC70 values, both compounds are equally efficient of different origin albeit through unknown mechanism of in causing of PARP and caspases cleavage, suggesting that action (19). We show here that the two acetylenic com- they have similar potency of apoptosis induction. Alto- pounds 1 and 2 as well as the parental mixture of these gether, our results demonstrate that both C. vasculum 1 compounds from which they were separated, indeed and 2 induce cell death by activation of the intrinsic cause tumor-specific cytotoxicity in NSCLC and SCLC. apoptotic pathway and this effect is tumor cell specific. Importantly, we demonstrate that for both compounds One of the regulators of Bax and Bak activation is the BH3 this cytotoxicity is tumor specific, as all normal cells tested only protein Bad that antagonize the antiapoptotic pro- representing heart tissue, bronchial and retina epithelium, teins Bcl-2 and Bcl-xl enabling Bak/Bax complex forma- normal PBMCs, and foreskin or lung fibroblasts remained tion and cytochrome c release (36). Bad function as an unaffected at concentrations that were toxic for NSCLC integrator of multiple growth factor signaling cascades, and SCLC. In ovarian carcinoma cells compound 1 but not including the PI3K/Akt and the Raf/MEK/ERK path- compound 2 caused tumor-specific cell death. For com- way, which antagonize Bad function by phosphorylation pound 2, a cytotoxic effect was evident in NSCLC and causing its binding and sequestration to 14-3-3 protein (37, SCLC whereas in ovarian carcinoma the difference com- 38). Akt inactivates Bad by phosphorylation at Ser136 (39) pared with normal cells was less prominent. whereas MAPK signaling cascade inactivates Bad protein With respect to NSCLC cytotoxicity, we for the first time at Ser112 and Ser155 (40–43). With respect to 1 and 2, Bad C. vasculum show that compounds from can induce a phosphorylation at Ser112 was markedly inhibited, indi- prominent apoptotic response whereas in normal fibro- cating a role for the MAPK signaling pathway whereas the blasts no induction of apoptosis was seen after exposure to basal level of phosphorylation of the other sites was weak the same concentration of 1 or 2. Moreover, no apoptosis in untreated cells making effect of treatment unsecure was seen in normal fibroblasts even when concentrations (data not shown). Given this altered phosphorylation of that caused 50% or 70% cytotoxic response in cell viability Bad Ser112 by compounds 1 and 2, we next examined how assay were used. This indicates that the induction of multiple growth and apoptosis-promoting kinases in apoptosis is a tumor-selective path that might be a con- NSCLC cells were affected. A decrease in ERK phosphor- sequence of inhibition of cell survival signaling pathways, ylation was observed, which is in line with the decrease in

www.aacrjournals.org Mol Cancer Ther; 13(12) December 2014 OF11

Zovko et al.

Bad protein phosphorylated at Ser112. Although results at this point, we can only say that both compounds are for inactivation of Bad protein at site Ser136 were incon- capable of inducing apoptosis in tumor cells. clusive, a major impairment of phosphorylation of Akt at We show that 1 is also capable of enhancing cytotoxicity Ser473 and Thr308 upon treatment with these compounds through induction of G2–M cell-cycle arrest in tumor cells. was evident. Deactivation of Akt is reported to cause It has been shown that Akt regulates cell cycle by activation of proapoptotic proteins, cause alterations in controlling of cyclins and CDK inhibitors (49–52). In mitochondrial membrane potential, release of cytochrome addition, JNK is involved in cell-cycle progression and c, and caspase activation leading to activation of apoptosis mediates G2–M phase cell-cycle arrest in different (37, 43, 44). Hence, the impaired Akt phosphorylation tumor cells (53–58). As both, decreased phosphorylation upon treatment with these compounds fits well with of Akt and increased phosphorylation of JNK are evi- observed apoptotic changes in these NSCLC cells. dent after treatment with 1, these signaling cascades WeandothershavelinkedNSCLCapoptoticresponseto may be responsible for the observed G2–M arrest upon a sustained JNK activation (25, 33, 45–48). In line with this treatment. Further studies are, however, required to and the observed apoptotic response, phosphorylation of uncover whether this is the case. JNK was increased by these C. vasculum–derived com- Albeit compounds 1 and 2 only show structural var- pounds. All in all, our data clearly demonstrate that 1 and iance in the length of their carbon chains and the methyl 2 inhibit proliferation and induce apoptosis through reg- substituent at position 14, a great difference in their ulation of the Ras/Raf/MAP kinase pathway as well as potency is evident when it comes to induction of cyto- deactivation of Akt signaling in NSCLC cells. As both Akt toxicity in NSCLC, SCLC, and more importantly in ovar- and MAPK kinase signaling is controlled by growth factor ian carcinoma where 1 but not 2 shows tumor-specific receptors, these results may indicate that a common cytotoxicity. The cytotoxic response by 1 was much more growth factor receptor instrumental in controlling their prominent in NSCLC and also appeared at an earlier time activitycouldindeedbeatargetofC.vasculumcompounds. point. One possible reason for the less prominent effects Further studies are, however, needed to reveal whether observed in tumor cells, especially in ovarian carcinoma, that is the case. We observed that Bak and Bax activation, by compound 2 could be lower affinity of drug to target a cytochrome c release, downregulation of Bcl-xL, Bad, Akt, growth factor receptor upstream of these pathways. Yet mTor, p43/p44MAPK (ERK1/2), JNK1/JNK2 phosphor- another possibility is that their target upstream of these ylation alteration were delayed or less affected by 2 than by kinases are different and with the compound 2 targeting 1 in these NSCLC cells when used at equimolar concentra- a signaling component, which play a less prominent role tions. Yet given that 1 induces a higher toxicity than 2, one in ovarian carcinoma. Permeability of different cell types cannot rule out that this is the underlying effect. Therefore, for the different compounds could also clearly have a

Figure 6. Schematic illustration of the proposed molecular mechanism of C. vasculum compounds in NSCLC cells. Experimental data from NSCLC cells support that compounds 1 and 2 indirectly or directly target growth factor receptors, thereby interfering with downstream proliferative signaling pathways Akt and MAPK, that is, decrease phosphorylation of ERK and Akt. Both compounds were found to deactivate Akt signaling leading to impaired Bad phosphorylation and activation of proapoptotic proteins Bak and Bax. Compounds were also found to trigger depolarization of mitochondria, release of cytochrome c, and activation of caspases. Compound 1 was also found to inhibit progression of cell cycle by arresting NSCLC cells in G2–M phase.

OF12 Mol Cancer Ther; 13(12) December 2014 Molecular Cancer Therapeutics

C. vasculum Compounds Possess Antitumor Potential

role. Hence, further research is required to put an answer Authors' Contributions to these issues. Conception and design: A. Zovko, K. Viktorsson, M. Ilan, S. Carmeli, R. Lewensohn In conclusion, our data reveal that 1 isolated from Development of methodology: A. Zovko, M. Ilan marine sponge C. vasculum causes tumor-speciﬁc cell Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Zovko, D. Kovalerchick, A. Alimonti, M. Ilan, death in NSCLC, SCLC, and ovarian carcinoma. For S. Carmeli, R. Lewensohn compound 2, we found such an activity to be evident in Analysis and interpretation of data (e.g., statistical analysis, biostatis- tics, computational analysis): A. Zovko, K. Viktorsson, D. Kovalerchick, NSCLC and SCLC but not in ovarian carcinoma. Our K. Farnega€ rdh, M. Ilan, R. Lewensohn ﬁndings are summarized in Fig. 6. We show that in Writing, review, and/or revision of the manuscript: A. Zovko, K. NSCLC these compounds trigger a prominent response Viktorsson, P. Haag, K. Farnega€ rdh, M. Ilan, S. Carmeli, R. Lewensohn of the intrinsic apoptotic pathway (activation of Bak/Bax, Administrative, technical, or material support (i.e., reporting or orga- nizing data, constructing databases): R. Lewensohn decreased phosphorylation of Bad, depolarization of Study supervision: K. Viktorsson, S. Carmeli, R. Lewensohn mitochondria, release of cytochrome c, and activation of caspases 9 and 3) and cell-cycle arrest in G2–M phase with Acknowledgments a concomitant inhibition of both Akt and ERK/MAPK The authors thank Birgitta Mork€ and Therese Juntti for excellent technical assistance. growth–promoting pathways. Moreover, we show that in NSCLC these compounds trigger a sustained JNK activation, which may contribute to the observed proapoptotic Grant Support This study was supported by grants from the Swedish Cancer Society capacity. Taken together, our results, therefore, implicate (to R. Lewensohn grant agreement 130550), Stockholm Cancer Society (to that one or several growth factor receptors could be K. Viktorsson grant agreement 131253; to P. Haag 131102; and to R. 1 2 Lewensohn 111213), Swedish Research Foundation (to R. Lewensohn targets of and either directly or indirectly and in this grant agreement 90266701/2009), the Swedish National Board of Health way trigger cell death through the intrinsic apoptotic and Welfare grant (to R. Lewensohn), the Stockholm County Council grant pathway and cell-cycle arrest. Further studies of the (to R. Lewensohn), the Karolinska Institutet Research Fund and the European Union (FP7/2007-2013) under grant agreement no KBBE- effects of compounds on growth factor receptors are, 2010-266033. A. Alimonti was supported by European Research Council therefore, warranted to reveal mechanism of action, starting grant (ERCsg; GA N261342). develop these compounds into suitable pharmacologic The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked leads as well as to understand their role in treatment of advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate NSCLC and SCLC but also other tumor malignancies. this fact.

Disclosure of Potential Conﬂicts of Interest Received April 14, 2014; revised September 25, 2014; accepted October 8, No potential conﬂicts of interest were disclosed. 2014; published OnlineFirst October 15, 2014.

References 1. Frenz JL, Kohl AC, Kerr RG. Marine natural products as therapeutic 13. Herbst RS, Heymach JV, Lippman SM. Lung cancer. N Engl J Med agents: part 2. Expert Opin Ther Pat 2004;14:17–33. 2008;359:1367–80. 2. Blunt JW, Copp BR, Hu WP, Munro MHG, Northcote PT, Prinsep MR. 14. Viktorsson K, Lewensohn R, Zhivotovsky B. Apoptotic pathways and Marine natural products. Nat Prod Rep 2007;24:31–86. therapy resistance in human malignancies. Adv Cancer Res 3. Blunt JW, Copp BR, Hu WP, Munro MHG, Northcote PT, Prinsep MR. 2005;94:143–96. Marine natural products. Nat Prod Rep 2009;26:170–244. 15. Aiello A, Fattorusso E, Menna M, Pansini M. Further bioactive acet- 4. Schwartsmann G. Marine organisms and other novel natural sources ylenic compounds from the Caribbean sponge Cribrochalina vascu- of new cancer drugs. Ann Oncol 2000;11:235–43. lum. J Nat Prod 1992;55:1275–80. 5. Sipkema D, Franssen MC, Osinga R, Tramper J, Wijffels RH. Marine 16. Gung BW, Dickson HD, Seggerson S, Bluhm K. A short synthesis of an sponges as pharmacy. Mar Biotechnol 2005;7:142–62. acetylenic alcohol from the sponge Cribrochalina vasculum. Synth 6. Harvey AL. Natural products in drug discovery. Drug Discov Today Commun 2002:32:2733–40. 2008;13:894–901. 17. Lu W, Zheng GR, Cai JC. Synthesis of two enantiomers of Eicos-(4E)- 7. Indumathy S, Dass CS. Finding chemo: the search for marine-based en-l-yn-3-ol, a bioactive components of the marine sponge Cribro- pharmaceutical drugs active against cancer. J Pharm Pharmacol chalina vasculum. Tetrahedron 1999;55:4649–54. 2013;65:1280–301. 18. Gunasekera SP, Faircloth GT. New acetylenic alcohols from the 8. Gunasekera SP, Gunasekera M, Longley RE, Schulte GKJ. Disco- sponge Cribrochalina vasculum. J Org Chem 1990;55:6223–5. dermolide:anewbioactivepolyhydroxylatedlactonefromthe 19. Hallock YF, Cardellina JH, Balaschak MS, Alexander MR, Prather TR, marine sponge Discodermia dissolute. Org Chem 1990;55:4912–5. Shoemaker RH, et al. Antitumor activity and stereochemistry of acet- 9. ter Haar E, Kowalski RJ, Hamel E, Lin CM, Longley RE, Gunasekera ylenic alcohols from the sponge Cribrochalina vasculum. J Nat Prod SP, et al. Discodermolide, a cytotoxic marine agent that stabilizes 1995;58:1801–7. microtubules more potently than taxol. Biochemistry 1996;35: 20. Bergh J, Larsson E, Zech L, Nilsson K. Establishment and character- 243–50. ization of two neoplastic cell lines (U-1285 and U-1568) derived from 10. Carter NJ, Keam SJ. Trabectedin: a review of its use in the management small cell carcinoma of the lung. Acta Pathol Microbiol Immunol Scand of soft tissue sarcoma and ovarian cancer. Drugs 2007;67:2257–76. A 1982;90:149–58. 11. Cortes J, O'Shaughnessy J, Loesch D, Blum JL, Vahdat LT, Petrakova 21. Bergh J, Nilsson K, Ekman R, Giovanella B. Establishment and K, et al. Eribulin monotherapy versus treatment of physician's choice in characterization of cell lines from human small cell and large cell patients with metastatic breast cancer (EMBRACE): a phase 3 open- carcinomas of the lung. Acta Path Microbial Immunol Scand label randomised study. Lancet 2011;377:914–23. 1985;93:133–47. 12. Lowenberg€ B, Downing JR, Burnet A. Acute myeloid leukemia. N Engl J 22. Sirzen F, Zhivotovsky B, Nilsson A, Bergh J. Lewensohn R. Spon- Med 1999;341:1051–62. taneous and radiation-induced apoptosis in lung carcinoma cells

www.aacrjournals.org Mol Cancer Ther; 13(12) December 2014 OF13

Zovko et al.

with different intrinsic radiosensitivities. Anticancer Res 1998;18: 43. Hayakawa J, Ohmichi M, Kurachi H, Kanda Y, Hisamoto K, Nishio Y, 695–9. et al. Inhibition of BAD phosphorylation either at serine 112 via 23. Hayflick L, Moorehead PS. The serial cultivation of human diploid cell extracellular signal-regulated protein kinase cascade or at serine strains. Exp Cell Res 1961;25:585–621. 136 via Akt cascade sensitizes human ovarian cancer cells to cisplatin. 24. Mosmann T. Rapid colorimetric assay for cellular growth and survival: Cancer Res 2000;60:5988–94. application to proliferation and cytotoxicity assays. J Immunol Meth- 44. Yamaguchi H, Wang HG. The protein kinase PKB/Akt regulates cell ods 1983;65:55–63. survival and apoptosis by inhibiting Bax conformational change. 25. Viktorsson K, Ekedahl J, Lindebro MC, Lewensohn R, Zhivotovsky B, Oncogene 2001;20:7779–86. Linder S, et al. Defective stress kinase and Bak activation in response 45. Xia Z, Dickens M, Raingeaud J, Davis RJ, Greenberg ME. Opposing to ionizing radiation but not cisplatin in a non–small cell lung carcinoma effects of ERK and JNK-p38 MAP kinases on apoptosis. Science cell line. Exp Cell Res 2003;289:256–64. 1995;270:1326–31. 26. Griffiths GJ, Dubrez L, Morgan CP, Jones NA, Whitehouse J, Corfe BM, 46. Dhanasekaran DN, Reddy EP. JNK signaling in apoptosis. Oncogene et al. Cell damage-induced conformational changes of the pro-apo- 2008;27:6245–51. ptotic protein Bak in vivo precede the onset of apoptosis. J Cell Biol 47. Wagner EF, Nebreda AR. Signal integration by JNK and p38 1999;144:903–14. MAPK pathways in cancer development. Nat Rev Cancer 2009; 27. Mandic A, Viktorsson K, Strandberg L, Heiden T, Hansson J, Linder S, 9:537–49. et al. Calpain-mediated Bid cleavage and calpain-independent Bak 48. Stahl S, Fung E, Adams C, Lengqvist J, Mork€ B, Stenerlow€ B, et al. modulation: two separate pathways in cisplatin-induced apoptosis. Proteomics and pathway analysis identifies JNK signaling as critical for Mol Cell Biol 2002;22:3003–13. high linear energy transfer radiation-induced apoptosis in non–small 28. Darzynkiewicz Z, Huang X. Analysis of cellular DNA content by flow lung cancer cells. Mol Cell Proteomics 2009;8:1117–29. cytometry. Curr Protoc Immunol 2004;5:5–7. 49. Roy SK, Srivastava RK, Shanka S. Inhibition of PI3K/AKT and MAPK/ 29. Dale JA, Mosher HS. Nuclear magnetic resonance enantiomer regents. ERK pathways causes activation of FOXO transcription factor, leading Configurational correlations via nuclear magnetic resonance chemical to cell cycle arrest and apoptosis in pancreatic cancer. J Mol Signal shifts of diastereomeric mandelate, O-methylmandelate, and a-meth- 2010;5:10. oxy-a-trifluoromethylphenylacetate (MTPA) esters. J Am Chem Soc 50. Nakamura N, Ramaswamy S, Vazquez F, Signoretti S, Loda M, Sellers 1973;95:512–9. WR. Forkhead transcription factors are critical effectors of cell death 30. Zou H, Henzel WJ, Liu X, Lutschg A, Wang X. Apaf-1, a human protein and cell cycle arrest downstream of PTEN. Mol Cell Biol 2000;20: homologous to C. elegans CED-4, participates in cytochrome c– 8969–82. dependent activation of caspase-3. Cell 1997;90:405–13. 51. Thaler S, Hahnel€ PS, Schad A, Dammann R, Schuler M. RASSF1A 31. Hengartner MO. The biochemistry of apoptosis. Nature 2000;470:770–6. mediates p21Cip1/Waf1-dependent cell cycle arrest and senescence 32. Strasser A, O'Connor L, Dixit VM. Apoptosis signaling. Annu Rev through modulation of the Raf-MEK-ERK pathway and inhibition of Biochem 2000;69:217–45. Akt. Cancer Res 2009;69:1748–57. 33. Tournier C. The 2 faces of JNK signaling in cancer. Genes Cancer 52. Shtivelman E, Sussman J, Stokoe D. A role for PI 3-kinase and PKB 2013;4:397–400. activity in the G2/M phase of the cell cycle. Curr Biol 2002;12: 34. Reed JC. Proapoptotic multidomain bcl-2/bax-family proteins: 919–24. mechanisms, physiological roles and therapeutic opportunities. Cell 53. Cho SD, Li G, Hu H, Jiang C, Kang KS, Lee YS, et al. Involvement of c- Death Differ 2006;13:1378–86. Jun N-terminal kinase in G2/M arrest and caspase-mediated apopto- 35. Kroemer G, Galluzzi L, Brenner C. Mitochondrial membrane permea- sis induced by sulforaphane in DU145 prostate cancer cells. Nutr bilization in cell death. Physiol Rev 2007;87:99–163. Cancer 2005;52:213–24 36. Willis SN, Adams JM. Life in the balance: how BH3-only proteins 54. Teraishi F, Wu S, Sasaki J, Zhang L, Davis JJ, Guo W, et al. JNK1- induce apoptosis. Curr Opin Cell Biol 2005;17:617–25. dependent antimitotic activity of thiazolidin compounds in human 37. Datta SR, Dudek H, Tao X, Masters S, Fu H, Gotoh Y, et al. Akt b nonsmall-cell lung and colon cancer cells. Cell Mol Life Sci 2005; phosphorylation of BAD couples survival signals to the cell-intrinsicn- 62:2382–9. death machinery. Cell 1997;91:231–41. 55. Fukuda T, Asai N, Enomoto A, Takahashi M. Activation of c-Jun 38. Huang DC, Strasser A. BH3-Only proteins-essential initiators of apo- aminoterminal kinase by GDNF induces G2/M cell cycle delay linked ptotic cell death. Cell 2000;103:839–42. with actin reorganization. Genes Cells 2005;10:655–63. 39. Harada H, Andersen JS, Mann M, Terada N, Korsmeyer SJ. p70S6 56. Kang KA, Kim YW, Kim SU, Chae S, Koh YS, Kim HS, et al. G1 phase kinase signals cell survival as well as growth, inactivating the pro- arrest of the cell cycle by a ginseng metabolite, compound K, in U937 apoptotic molecule BAD. Proc Natl Acad Sci U S A 2001;98:9666–70. human monocytic leukamia cells. Arch Pharmacol Res 2005;28: 40. Bonni A, Brunet A, West AE, Datta SR, Takasu MA, Greenberg ME. Cell 685–90. survival promoted by the Ras-MAPK signaling pathway by transcrip- 57. Antosiewicz J, Herman-Antosiewicz A, Marynowski SW, Singh SV. tion-dependent and -independent mechanisms. Science 1999; c-Jun NH2-terminal kinase signaling axis regulates diallyl trisul- 286:1358–62. fide-induced generation of reactive oxygen species and cell cycle 41. Eisenmann KM, VanBrocklin MW, Staffend NA, Kitchen SM, Koo HM. arrest in human prostate cancer cells. Cancer Res 2006;66: Mitogen-activated protein kinase pathway-dependent tumor-specific 5379–86. survival signaling in melanoma cells through inactivation of the proa- 58. Munoz-Alonso~ MJ, Gonzalez-Santiago L, Zarich N, Martínez T, poptotic protein bad. Cancer Res 2003;63:8330–7. AlvarezE,RojasJM,etal.Plitidepsin has a dual effect inhibiting 42. Tan Y, Demeter MR, Ruan H, Comb MJ. BAD Ser-155 phosphorylation cell cycle and inducing apoptosis via Rac1/c-Jun NH2-terminal regulates BAD/Bcl-XL interaction and cell survival. J Biol Chem kinase activation in human melanoma cells. J Pharmacol Exp Ther 2000;275:25865–9. 2008;324:1093–101.

OF14 Mol Cancer Ther; 13(12) December 2014 Molecular Cancer Therapeutics

Marine Sponge Cribrochalina vasculum Compounds Activate Intrinsic Apoptotic Signaling and Inhibit Growth Factor Signaling Cascades in Non−Small Cell Lung Carcinoma

Ana Zovko, Kristina Viktorsson, Petra Hååg, et al.

Mol Cancer Ther Published OnlineFirst October 15, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-14-0329

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2014/10/15/1535-7163.MCT-14-0329.DC1

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/early/2014/11/19/1535-7163.MCT-14-0329. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.