in vitro and in vivo anticancer activities of synthetic...

[CANCER RESEARCH 61, 1013–1021, February 1, 2001]

In Vitro and In Vivo Anticancer Activities of Synthetic Macrocyclic KetoneAnalogues of Halichondrin B1

Murray J. Towle, Kathleen A. Salvato, Jacqueline Budrow, Bruce F. Wels, Galina Kuznetsov, Kimberley K. Aalfs,Susan Welsh, Wanjun Zheng, Boris M. Seletsky, Monica H. Palme, Gregory J. Habgood, Lori A. Singer,Lucian V. DiPietro, Yuan Wang, Jack J. Chen, David A. Quincy, Ashley Davis, Kentaro Yoshimatsu, Yoshito Kishi,Melvin J. Yu, and Bruce A. Littlefield 2

Departments of Anticancer Research [M. J. T., K. A. S., J. B., B. F. W., G. K., K. K. A., S. W., B. A. L.] and Medicinal Chemistry [W. Z., B. M. S., M. H. P., G. J. H., L. A. S.,L. V. D., Y. W., J. J. C., D. A. Q., M. J. Y.] and Advisory Board [Y. K.], Eisai Research Institute, Andover, Massachusetts 01810; Cytoskeleton, Incorporated, Denver, Colorado80206 [A. D.]; and Discovery Research Laboratories II, Tsukuba Research Laboratories, Eisai Company, Limited, Tsukuba-shi, Ibaraki 300-26, Japan [K. Y.]

ABSTRACT

Halichondrin B is a highly potent anticancer agent originally found inmarine sponges. Although scarcity of the natural product has hamperedefforts to develop halichondrin B as a new anticancer drug, the existenceof a complete synthetic route has allowed synthesis of structurally simpleranalogues that retain the remarkable potency of the parent compound. Inthis study, we show that two macrocyclic ketone analogues of halichondrinB, ER-076349 and ER-086526, have sub-nM growth inhibitory activities invitro against numerous human cancer cell lines as well as markedin vivoactivities at 0.1–1 mg/kg against four human xenografts: MDA-MB-435breast cancer, COLO 205 colon cancer, LOX melanoma, and NIH:OVCAR-3 ovarian cancer. ER-076349 and ER-086526 induce G2-M cellcycle arrest and disruption of mitotic spindles, consistent with the tubulin-based antimitotic mechanism of halichondrin B. This is supported furtherby direct binding of the biotinylated analogue ER-040798 to tubulin andinhibition of tubulin polymerization in vitro by ER-076349 andER-086526. Retention of the extraordinaryin vitro and in vivo activity ofhalichondrin B in structurally simplified, fully synthetic analogues estab-lishes the feasibility of developing halichondrin B-based agents as highlyeffective, novel anticancer drugs.

INTRODUCTION

Halichondrin B is a large polyether macrolide (Fig. 1) found in avariety of marine sponges, includingHalichondria okadai (1, 2),Axinella sp. (3),Phakellia carteri(4), andLissondendryx sp. (5). In1986, halichondrin B was shown to have remarkablein vitro and invivo anticancer activities against murine melanomas and leukemias(2). Subsequent studies by several groups (6–9) confirmed the sub-nM

in vitro potency of halichondrin B and identified a tubulin-basedantimitotic mechanism. Although halichondrin B is classified as atubulin depolymerizer (similar toVinca alkaloids, dolastatins, cryp-tophycin, and so on), the patterns of its interaction with tubulin arehighly specific and place it in its own unique class within this group(6–9). Although speculative, unique interactions with tubulin mayhelp explain the markedin vivo anticancer potency of halichondrin B,now well documented in several human tumor models includingmelanoma, osteogenic sarcoma, lung cancers, and colon cancers (10).More importantly, unique tubulin interactions relative to existingtubulin-based drugs suggest the possibility of unique tumor specific-ities or other desirable clinical characteristics.

Although the extraordinary potency of halichondrin B generatedconsiderable interest for development, limited availability of the nat-ural product severely restricted such efforts. However, the existence

of a synthetic route for halichondrin B (11) and knowledge that itsactivity resides in the macrocyclic lactone C1-C38 moiety3 (see alsoRefs. 12–14) have permitted development of structurally simplifiedsynthetic analogues that retain the exceptional activity of the parent.In this study, we describe ER-076349 and ER-086526, two macrocy-clic ketone analogues of halichondrin B with highly potent activitiesin vitro and in vivo. We show that these agents exert their effects viaa tubulin depolymerizing antimitotic mechanism similar or identicalto that reported for parental halichondrin B.

MATERIALS AND METHODS

Chemical Synthesis.Halichondrin B macrocyclic ketone analogues ER-076349 (NSC 707390) and ER-086526 (NSC 707389) were prepared asdescribed previously (15). The biotinylated analogue ER-040798 and thebiotinylated negative control ER-040792 were synthesized based on appropri-ate modifications to the published synthetic route for halichondrin B (11). Thestructures of halichondrin B and all of the synthetic analogues are presented inFig. 1.

Cells and Cell Culture. The following human tumor cell lines wereobtained from the American Type Culture Collection (Rockville, Maryland)and grown under American Type Culture Collection-recommended conditions:COLO 205 and DLD-1 colon cancer, HL-60 promyelocytic leukemia, U937histiocytic lymphoma, and LNCaP and DU 145 prostate cancer. LOX humanmelanoma cells were obtained from the Division of Cancer Treatment-NCI4

Tumor Repository (Frederick, Maryland) and were grown in RPMI 1640medium supplemented with 10% heat-inactivated fetal bovine serum and 2 mM

L-glutamine. MDA-MB-435 human breast cancer cells were generously pro-vided by Dr. Mary J. C. Hendrix (University of Iowa College of Medicine,Iowa City, Iowa) and were grown in DMEM (high glucose) supplemented with10% heat-inactivated fetal bovine serum, 20 mM N-2-hydroxyethyl-piperazine-N9-2-ethanesulfonic acid and 1 mM sodium pyruvate.

All of the cell lines were grown at 37°C in a humidified atmospherecontaining 5% CO2. Seeding densities for each cell line were empiricallyoptimized for passaging twice weekly. Quantification of cell number was byhemocytometer counting, and viability determinations were made by standardtrypan blue exclusion techniques. Routine harvesting of monolayer cultureswas by standard trypsinization procedures.

Cell Growth Inhibition Assays. Cells were plated in 96-well plates at7,500 cells/well (except LNCaP cells, which were at 10,000 cells/well) andgrown in the continuous presence of test compounds for 4 days. For monolayercultures (DU 145, LNCaP, LOX, and MDA-MB-435), growth was assessedusing modifications (16) of a methylene blue-based microculture assay (17).For suspension cultures (HL-60, U937) or loosely adherent monolayers(COLO 205), growth was assessed using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide-based assay (18) modified as follows. After 4days of incubation with test compounds, sterile-filtered 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (Sigma Chemical Co., St. Louis, MO)was added to each well (final concentration, 0.5 mg/ml), and plates wereincubated at 37°C for 4 h. Acid-isopropanol (0.1N HCl in isopropanol, 150ml)

Received 4/12/00; accepted 11/30/00.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby markedadvertisementin accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1 These data were presented in part (abstract 1370) at the 91st Annual Meeting of theAACR held April 1–5, 2000 in San Francisco, CA.

2 To whom requests for reprints should be addressed, at Department of AnticancerResearch, Eisai Research Institute, 4 Corporate Drive, Andover, MA 01810-2441. Phone:(978) 837-4638; Fax: (978) 794-4910; E-mail: [email protected].

3 M. J. Towle, Y. Kishi, and B. A. Littlefield, unpublished observations, 1992.4 The abbreviations used are: NCI, United States National Cancer Institute; DAPI,

49,6-diamidino-2-phenylindole; MTD, maximum tolerated dose; PEM, 80 mM PIPES (pH6.9), 1 mM EGTA, 1 mM MgCl2.

1013

Research. on June 30, 2018. © 2001 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

was then added to each well, and the resultant formazan crystals were dis-solved by gentle mixing. Absorbances at 540 nm were measured on a TitertekMultiscan MCC/340 plate reader.

Cell Cycle Analysis. DNA content analysis of U937 cells was performedby single channel flow cytometry as follows. U937 cells were exposed for0–14 h to 30 nM ER-076349 or 100 nM ER-086526; concentrations wereroughly 10-fold higher than those that minimally induce complete G2-Mblocks based on dose/response studies. Typically, 2–43 106 cells werecollected by centrifugation at 3003 g for 10 min and resuspended in 3 ml ofsaline, followed by the immediate addition of 7 ml of ice-cold 100% ethanolwhile vortexing. After fixing overnight at 4°C in the saline/ethanol solution,cells were recovered by centrifugation, washed in PBS, and resuspended in 0.5ml PBS containing 200mg/ml RNase A (Sigma; preboiled 5 min as 10 mg/mlstock in PBS; stored at280°C). RNase digestion was at 37°C for 30 min,followed by the addition of 0.5 ml of 10mg/ml propidium iodide (Sigma) inPBS at least 15 min before flow analysis. Single channel flow cytometry wasperformed on a Becton Dickinson FACScan with a Macintosh computer; thecollection and analysis of data were performed using Becton DickinsonCELLQuest software. Doublet events were eliminated from analyses by propergating on FL2-W/FL2-A primary plots before histogram analysis of DNAcontent (measured as FL2-A).

Fluorescence Microscopy.For simultaneous anti-tubulin and DAPI stain-ing of DU 145 cells, cells were plated in 4-well plastic Lab-Tek chamber slides(Nunc, Inc., Naperville, IL) at 7.53 104 cells/chamber in 0.9 ml of completemedium. After culturing overnight, 100ml of 10 3 concentrates of testcompounds in complete media were added to a final concentration equal tothree times the empirically determined IC50 of each compound; control wellsreceived 100ml of complete media only. After 20 h, cell monolayers werefixed with ice-cold anhydrous methanol for 6 min at220°C, followed by twowashes with 0.1% (v/v) Tween 20 in PBS and two washes with PBS alone.

Fixed cells were then incubated with 1:50 diluted (v/v in PBS) anti-b-tubulinTUB 2.1 monoclonal antibody (Sigma) for 1 h at room temperature withrocking, followed by three 5-min washes with PBS and incubation with 1:50diluted (v/v in PBS) FITC-conjugated goat F(ab9)2 antimouse immunoglobulin(BioSource International, Camarillo, CA) for 1 h at room temperature withrocking. Cells were washed 33with PBS, incubated with 2mg/ml DAPI inPBS for 10 min at room temperature, followed by one 5-min wash with PBSand two 5-min washes with 2 mM Tris, 0.1 mM EDTA (pH 7.5). Chamber wallsand gaskets were then removed, and the slides were air-dried. Coverslips weremounted with Gel/Mount (Biomeda, Foster City, CA), and fluorescence mi-croscopy was performed using a Nikon Labophot-2 microscope equipped with1003 oil immersion objective and B-2A and UV-2A filter cubes for FITC andDAPI staining, respectively. Images were acquired with a Photometrics NU200 CCD camera and digitized using IPLab Spectrum software (Scanalytics,Fairfax, VA).

Affinity Binding of Tubulin to Biotinylated Macrocyclic AnalogueER-040798.UltraLink™ Immobilized NeutrAvidin Plus (Pierce ChemicalCo., Rockford, IL) was charged with ER-040792, ER-040798 (135mM finalconcentrations), or buffer alone at room temperature for 60 min in 20 mM Tris,0.15M NaCl, 1 mM Na2EDTA (pH 7.5; wash-elution buffer). After five washesin wash-eluted buffer, matrices were incubated in the same buffer with 375mg/ml bovine brain tubulin (Molecular Probes, Eugene, OR) at room temper-ature for 30 min, followed by three washes in wash-elution buffer and elutionwith 20% acetonitrile (v/v in water). Eluted proteins were dried in a Speed-Vac, resuspended in SDS sample buffer, and subjected to modified discontin-uous SDS PAGE on 8% gels as described by Matthes (19) for separation ofa- and b-tubulin subunits. Gels were silver-stained (20) or transferred topolyvinylidene difluoride membranes (Sigma) for Western immunoblotting(21) with subtype-specific anti-a-tubulin or anti-b-tubulin monoclonal anti-bodies (clones B-5-1-2 and TUB 2.1, respectively; Sigma) and horseradishperoxidase-conjugated sheep antimouse immunoglobulin secondary antibodies(Amersham Corp., Arlington Heights, IL) using a chemiluminescence protocol(ECL Western Blotting Detection System; Amersham).

In Vitro Tubulin Polymerization Studies. ER-076349, ER-086526, andvinblastine in 10 mM anhydrous DMSO stocks were diluted into 10% DMSO,90% PEM. Substocks were made to final concentrations of 1, 10, 100, and1000mM; these were diluted into 100-ml volumes of 3.0 mg/ml bovine braintubulin (Cytoskeleton, Inc., Denver, CO) in PEM buffer plus 1 mM ATP, 3%(v/v) glycerol to achieve desired test compound concentrations. The amount of10% DMSO, 90% PEM was made up to 11ml in all of the samples to achieve1% final DMSO concentrations. Microtubule polymerization was initiated byraising the temperature from 4°C to 37°C over a 3-min period. Absorbance(A340 nm) was measured once/min for 60 min.

In Vivo Xenograft Studies. Anticancer effects of ER-076349 and ER-086526 were evaluated in the MDA-MB-435 human breast cancer, COLO 205human colon cancer, and LOX human melanoma xenograft models using5–6-week-old female Swiss nude mice and in the NIH:OVCAR-3 humanovarian cancer model using 7-week-old female BALB/c nude mice. All of thestudies using laboratory animals were approved either by the Eisai ResearchInstitute (Andover, MA) or Tsukuba Research Laboratories (Tsukuba, Japan)Institutional Animal Care and Use Committees and adhered to all of theapplicable institutional and governmental guidelines for the humane care anduse of laboratory animals.

On day 0 of each experiment, mice received injections s.c. with 13 106

cells (MDA-MB-435, COLO 205, and LOX) or 63 105 cells (NIH:OVCAR-3; previously adapted for enhanced penetrance via tumor passageinvivo). For COLO 205 and MDA-MB-435 models, mice received injectionswith 200 ml of test compound in saline on Monday/Wednesday/Friday i.p.(COLO 205) or i.v. (MDA-MB-435) schedules, respectively, beginning on day13 for four weekly cycles. Treatment of mice bearing LOX tumors was bydaily i.p. injection (200ml) in saline on days 3–7 and 10–14 inclusive. In theNIH:OVCAR-3 model, test compounds were injected i.v. (200ml/20g bodyweight) in 4% DMSO/saline beginning on day 40 and continuing Monday/Wednesday/Friday for three weekly cycles. Control groups in all of the studiesreceived appropriate vehicle injections. Paclitaxel at its empirically determinedMTD for each treatment regimen was included in all of the experiments andwas administered using Diluent 12 (1:1 Cremophor EL/ethanol diluted into 5%glucose in water) as described by others (22, 23). Group sizes were 10 mice forall of the tumor models except NIH:OVCAR-3, which used 6 mice/group.

Fig. 1. Structures of halichondrin B, macrocyclic ketone analogues ER-076349 andER-086526, biotinylated macrolactone ER-040798, and biotinylated negative controlER-040792.

1014

ANTICANCER ACTIVITIES OF SYNTHETIC HALICHONDRIN B ANALOGUES



Tumor volumes were determined bycaliper measurements and were calculatedaccording toV 5 0.67p[(L 1 W)/4]3 (i.e., volume [V] of a semisphere withdiameter equal to the mean of tumor length [L] and width [W]). Body weightsand water consumption were measured in all of the experiments to assesstoxicity.

RESULTS

Inhibition of Human Tumor Cell Growth in Vitro. More than180 analogues of the macrocyclic C1-C38 moiety of halichondrin Bwere synthesized and evaluated in search of suitably active andchemically stable drug development candidates. This culminated inthe identification of two simplified macrocyclic analogues,ER-076349 and ER-086526, in which the C1 lactone ester of hali-chondrin B is replaced by ketone, C31 methyl is replaced by methoxy,and the tricyclic C29-C38 system is replaced by a single five-membered ring (Fig. 1). The two analogues themselves differ only atC35, with hydroxyl and primary amine substituents for ER-076349and ER-086526, respectively.

On the basis of the known anti-tubulin mechanism of halichondrinB (6–9), in vitro growth inhibitory effects of ER-076349 and ER-

086526 were compared with those of the microtubule destabilizervinblastine and the microtubule stabilizer paclitaxel in MDA-MB-435human breast cancer cells (Fig. 2). In this experiment, ER-076349 andER-086526 inhibited cell growth at sub-nM levels (IC50 values of 0.15nM and 0.07 nM, respectively) with greater potency than vinblastine(0.54 nM) or paclitaxel (2.6 nM). Similar low- or sub-nM potencieswere observed in several other human cancer cell lines, includingCOLO 205 and DLD-1 colon cancers, DU 145 and LNCaP prostatecancers, LOX melanoma, HL-60 monocytic leukemia, and U937histiocytic lymphoma (Table 1). Interestingly, DLD-1 cells wereabout 10-fold less sensitive to ER-086526 compared with the otherseven lines (overall mean without DLD-1 data, 0.76 0.2 nM); thiswas not seen for ER-076349. The reasons for the relative insensitivityof DLD-1 cells to ER-086526 are unknown, although it is noted thatthese cells were also less sensitive to vinblastine and paclitaxel.

IC50 values for growth inhibition by ER-076349 and ER-086526agree well with values previously obtained using synthetic halichon-drin B, where an overall mean of 0.46 0.1 nM was obtained across 17human cancer cell lines (24). Thus, the structural simplifications ofER-076349 and ER-086526 relative to halichondrin B have not sig-nificantly altered growth inhibitory potency.

Despite sub-nM growth inhibition of several human cancer celltypes, neither ER-076349 nor ER-086526 showed cytotoxic effectsagainst quiescent IMR-90 human fibroblasts at concentrations up to 1mM, as assessed using an ATP-dependent luciferase-based viabilityassay (data not shown). Thus, growth inhibition by low- or sub-nM

levels of ER-076349 and ER-086526 is specific for proliferating cellsand not secondary to nonspecific cytotoxicity.

Cell Cycle Effects: G2-M Phase Block. Cell cycle analysis ofU937 human histiocytic lymphoma cells was performed after 0–14h-incubation with 30 nM ER-076349 or 100 nM ER-086526. As shownin Fig. 3, treatment with either ER-076349 or ER-086526 for as littleas 2 h led to increased numbers of G2-M phase cells. In both groups,G2-M cells accumulated at the expense of G1, which was almostcompletely depleted by 8 h. Progressive depletion of S phase thenfollowed beginning at 10 h because of cessation of new cells enteringS from G1. Depletion of S phase was essentially complete by 14 h, atwhich time almost all of the living cells were blocked in G2-M. Forboth compounds, increased numbers of hypodiploid events were seenafter 8–10 h, indicating apoptosis of cells after prolonged G2-Mblockage. The fact that both G1 and S phases became progressivelydepleted under continuous ER-076349 or ER-086526 exposure con-firms normal movement of cells through these phases even in thepresence of a drug. Thus, ER-076349 and ER-086526 are rapid andeffective G2-M phase blockers that do not affect cell cycle progressionthrough the G1 or S phases or the G1-S transition point. Similar G2-M

Fig. 2. Inhibition of MDA-MB-435 human breast cancer cell growth by ER-076349,ER-086526, paclitaxel, and vinblastine. Cells were grown for 4 days in the continuouspresence of the indicated concentrations of compounds, followed by measurement of cellgrowth as described in “Materials and Methods.” Results show means6 range ofduplicate determinations.

Table 1. Inhibition of human cancer cell growth by ER-076349, ER-086526, vinblastine, and paclitaxel

Cell lineb n

Growth inhibition, IC50 (nM) 6 SE/rangea

ER-076349 ER-086526 Vinblastine Pacitaxel

MDA-MB-435 4 0.146 0.1 0.096 0.01 0.596 0.02 2.56 0.34COLO 205 2 0.416 0.07 0.716 0.05 2.46 0.20 7.86 1.5DLD-1 3 0.756 0.08 9.56 1.0 7.36 0.74 196 2.3DU 145 3 0.706 0.06 0.916 0.06 3.66 0.92 9.46 3.0LNCaP 3 0.256 0.02 0.506 0.03 1.86 0.15 3.86 0.58LOX 3 0.766 0.11 1.46 0.31 3.26 0.39 7.36 0.43HL-60 2 0.416 0.01 0.906 0.24 2.66 0.64 4.36 1.3U937 2 0.226 0.04 0.436 0.02 4.06 2.0 3.96 2.0Average of eight cell linesc 8c 0.456 0.09 1.86 1.1d 3.26 0.70 7.36 1.9

a Growth inhibition determined under 4-day continuous exposure conditions as described in “Material and Methods.” Data shown are mean IC50 values in nM from n separateexperiments, each done with internal duplicates. Errors represent SE forn $ 3 or range forn 5 2.

b Human cancer cell lines used: MDA-MB-435 breast cancer, COLO 205 and DLD-1 colon cancer, DU 145 and LNCaP prostate cancer, LOX melanoma, HL-60 promyelocyticleukemia, and U937 histiocytic lymphoma.

c Overall mean6 SE of the eight individual means from each cell line.d ER-086526 mean excluding DLD-1 value (n5 7): 0.706 0.16 nM.

1015




blockage of U937 cells was observed previously using synthetichalichondrin B (24).

Disruption of Mitotic Spindles. The known anti-tubulin effects ofhalichondrin B (6–9) together with G2-M blockage by ER-076349and ER-086526 led us to assess effects on mitotic spindles in DU 145human prostate cancer cells (Fig. 4). In this study, cells in log phasegrowth were treated for 20 h with vehicle, 2.1 nM ER-076349, 2.7 nMER-086526, 11 nM vinblastine, or 28 nM paclitaxel, representingempirically determined 33 IC50 concentrations in this cell line(preliminary studies comparing 0.3, 1, 3, and 103 IC50 levelsestablished 33 IC50 as optimal; data not shown). After drug expo-sure, cells were fixed and dual stained with anti-b-tubulin monoclonalantibodies and DAPI to label microtubules and DNA, respectively. Asshown in Fig. 4, treatment with either ER-076349 (B) or ER-086526(C) led to marked disruption of normal mitotic spindle architecture(A). As expected, mitotic spindle disruption was also seen with theanti-tubulin agents vinblastine (D) and paclitaxel (E). With all of thefour drugs, cells with disrupted spindles showed the condensed butdisorganized chromosomal material characteristic of prometaphaseblockage. Occasional ER-076349- and ER-086526-treated cells were

seen that had not yet reached the mitotic block point during the 20-hdrug treatment period. These interphase (G1, S, or G2) cells, identifiedby intact DAPI-stained nuclei, appeared essentially normal with re-spect to filamentous cytoplasmic microtubules and intact, well-defined nuclear perimeters, consistent with the lack of effect ofER-076349 and ER-086526 on progression through G1 and S seen byflow cytometry (Fig. 3). Similar disruption of DU 145 mitotic spindleswas observed previously using synthetic halichondrin B (24).

Subtle drug-to-drug differences in mitotic spindle disruption pat-terns were noted at the 33 IC50 concentrations used in this study.Although such differences might derive from different tubulin-basedmechanisms (inhibitingversuspromoting tubulin polymerization, dif-ferences in tubulin binding, and so on), it is known that spindledisruption patterns byVincaalkaloids and paclitaxel at concentrationsclose to growth inhibitory IC50 levels look strikingly similar regard-less of mechanistic differences (25, 26). Indeed, we and others (datanot shown and Refs. 25–27) have noted that only small changes indrug concentration close to growth inhibitory IC50 levels can lead tomarked differences in spindle appearance. Thus, we cannot concludethat the subtle differences in spindle disruption patterns noted in Fig.4 are related to mechanistic differences.

Binding of Biotinylated Analogue ER-040798 to Tubulin. Todirectly address whether macrocyclic halichondrin B analogues werecapable of tubulin binding, the ability of the biotinylated analogueER-040798 (Fig. 1) to bind tubulin under affinity chromatographyconditions was tested. As shown in Fig. 5, immobilized ER-040798,but not the biotinylated negative control ER-040792, retained twoMr

50,000–55,000 protein bands from bovine brain tubulin (Lanes A–E).These bands were recognized by subtype-specific anti-a-tubulin(Lanes F–J) and anti-b-tubulin (Lanes K–O) monoclonal antibodies,confirming their identities asa- and b-tubulin, respectively. Thus,ER-040798 directly binds tubulin, suggesting that the biological ac-tivity of halichondrin B macrocyclic analogues derives from mecha-nisms similar or identical to parental halichondrin B. It should benoted that this study did not identify the tubulin subtype to whichER-040798 binds, because binding to either subtype ina/b-tubulinheterodimers would lead to retention of both. Finally, specific reten-tion of a- and b-tubulin by immobilized ER-040798 was also seenusing U937 whole cell lysates (data not shown), which demonstratedthat the retention ofa- andb-tubulin seen in Fig. 5 was not simplybecause of the presence of only these two proteins in the purifiedbovine brain tubulin preparations used.

Inhibition of Tubulin Polymerization in Vitro. To determinewhether the tubulin-depolymerizing mechanism of halichondrin B (6)had been conserved during structural simplification to macrocyclicketones, effects of ER-076349 and ER-086526 on polymerization ofbovine brain tubulinin vitro were assessed and compared with theknown tubulin depolymerizer vinblastine. As shown in Fig. 6, ER-076349, ER-086526, and vinblastine inhibited both rates and extentsof tubulin polymerization in dose-dependent fashions. The inhibitionof polymerization rates and extents was similar for each drug, withcalculated IC50s for ER-076349, ER-086526, and vinblastine of 5.7mM, 6.9 mM, and 1.8mM for rate inhibition, and 5.5mM, 6.0 mM, and2.1 mM for extent inhibition, respectively. By either criterion, ER-076349 and ER-086526 were about 3–4-fold less effective as inhib-itors of in vitro tubulin polymerization compared with vinblastine.

Although concentrations of all of the three drugs required to inhibittubulin polymerizationin vitro are several log orders higher than thosethat inhibit cell growth, such large discrepancies with tubulin-activeagents may result from inhibited microtubule dynamics at low drugconcentrations rather than net changes in cellular microtubule contentwhich occur only at higher drug levels (25, 26). In addition,Vinca

Fig. 3. Induction of G2-M cell cycle blocks by ER-076349 and ER-086526 in U937human histiocytic lymphoma cells. U937 cells in log phase growth were incubated with30 nM ER-076349 (top) or 100 nM ER-086526 (bottom) for 0–14 h; samples were takenat the times indicated in eachpanel (in h) for flow cytometric cell cycle analysis asdescribed in “Materials and Methods.” Data shown represent relative numbers of cells (Yaxis) as a function of fluorescence intensity (Xaxis), with G1 and G2-M DNA contentshaving been adjusted to 300 and 600 arbitrary units, respectively. Locations of G1, S, andG2-M populations are shown in the 0 hpanelsfor each drug.

1016




alkaloids and paclitaxel are known to accumulate inside cells to highlevels (25, 26); this may also occur with ER-076349 and ER-086526.

Inhibition of Human Tumor Xenograft Growth in Vivo. Theabilities of ER-076349 and ER-086526 to inhibit tumor growthin vivo

were tested in several human tumor xenograft models, includingMDA-MB-435 breast cancer, COLO 205 colon cancer, LOX mela-noma, and NIH:OVCAR-3 ovarian cancer (Fig. 7). In all of the cases,treatment with ER-076349 and ER-086526 in the 0.05–1 mg/kg rangeled to significant antitumor effects, with ER-086526 showing greaterefficacy in all of the models. Thus, in the MDA-MB-435 model (Fig.7, A and B), treatment with 0.25–1.0 mg/kg ER-076349 led to 60–70% inhibition at day 42, whereas ER-086526 at the same levels ledto .95% inhibition, including actual regression of measurable tumorspresent on day 14. For both compounds, tumor regrowth rates aftercessation of dosing allowed efficacy comparisons with paclitaxel at itsempirically determined MTD of 25 mg/kg. Again, ER-086526showed superiority over ER-076349; regrowth in all of the ER-076349 groups preceded the paclitaxel group by about 5 weeks,whereas all of the doses of ER-086526 were either equally efficacious(0.25, 0.5 mg/kg) or superior (1 mg/kg) to 25 mg/kg paclitaxel.Interestingly, the therapeutic window of ER-086526 seemed unusu-ally large for a cytotoxic drug;.95% tumor suppression occurredover the 4-fold dosing range of 0.25–1.0 mg/kg with no evidence oftoxicity based on body weight losses or decreased water consumption(data not shown). In contrast, complete tumor suppression by pacli-taxel with this dosing regimen is only seen between 15–25 mg/kg,with 10 mg/kg inducing only partial inhibition (data not shown); thetherapeutic window for paclitaxel in this model is thus just 1.7-fold.

The effects of 0.125 and 0.5 mg/kg ER-076349 and ER-086526

Fig. 4. Disruption of mitotic spindles by ER-076349, ER-086526, vinblastine, and paclitaxel inDU 145 human prostate cancer cells. DU 145 cellsin log phase growth were treated with test drugs for20 h at 3 3IC50 concentrations, empirically deter-mined for each drug, followed by fixation andimmunofluorescence staining for tubulin (green)and DNA (blue) as described in “Materials andMethods.”A, untreated cells;B, 2.1 nM ER-076349;C, 2.7 nM ER-086526;D, 11 nM vinblastine;E, 28nM paclitaxel.Scale barin A represents 20mm; thesame scale applies to all of the fivepanels.

Fig. 5. Binding of biotinylated affinity analogue ER-040798 to tubulin. The biotiny-lated macrocyclic analogue ER-040798 and a biotinylated negative control ER-040792were immobilized on an avidin-derived solid support and tested for their abilities to bindpurified bovine brain tubulin under affinity chromatography conditions. Proteins eluted by20% acetonitrile were analyzed by gel electrophoresis and Western blotting with anti-a-tubulin and anti-b-tubulin monoclonal antibodies. Details of all of the procedures arepresented in “Materials and Methods.”Lanes A-E, silver-stained gel;Lanes F–J,anti-a-tubulin Western blot;Lanes K–O,anti-b tubulin Western blot;Lanes A, E, F, J, K, O,bovine brain tubulin;Lane B, G, L,uncharged matrix;Lanes C, H, M,ER-040792-chargedmatrix;Lanes D, I, N,ER-040798-charged matrix.Left, sizes of molecular weight markersin thousands.

1017




were also examined in the COLO 205 colon cancer xenograft model(Fig. 7,C andD). As in the MDA-MB-435 study, ER-086526 showedsuperior efficacy, with frank regression and long-term suppression oftumor regrowth at 0.5 mg/kg and small but measurable decreases ingrowth rates at 0.125 mg/kg. Suppression with 0.5 mg/kg ER-086526was superior to that of paclitaxel at its empirically determined MTDof 20 mg/kg, both in rapidity of regression onset and in duration ofregrowth suppression. In contrast, inhibition by ER-076349 was lesspronounced at these dosing levels; no measurable effects were seen at0.125 mg/kg, and inhibition at 0.5 mg/kg was less than that with 20mg/kg paclitaxel. As in the MDA-MB-435 study, there was no evi-dence of toxicity based on body weights or water consumption (datanot shown).

Evaluation of 0.1–1.0 mg/kg ER-076349 and 0.05–0.5 mg/kg ER-086526 in the LOX melanoma model also showed potent antitumoreffects of both compounds (Fig. 7,E andF). With ER-076349, all ofthe doses led to.90% tumor suppression by day 17, three days aftercessation of dosing. The three top doses, 0.25, 0.5, and 1 mg/kg, ledto virtually complete tumor suppression up to day 17, with inhibitionof regrowth persisting well beyond that seen with 12.5 mg/kg pacli-taxel, its empirically determined MTD with this regimen. Again,complete tumor suppression by 0.25–1.0 mg/kg ER-076349 in theLOX model represents an unusually wide 4-fold therapeutic window.

In contrast, doubling the paclitaxel dose to 25 mg/kg in this model islethal, whereas halving it to 6.25 mg/kg leads to only minor inhibition(data not shown); the therapeutic window for paclitaxel with thisdosing regimen is thus,2-fold.

Inhibition of LOX tumor growth by ER-086526 was similar to thatwith ER-076349 but with about 2-fold greater potency. Thus, 0.05mg/kg ER-086526 inhibited tumor growth by 78% at day 17, withhigher doses of 0.1, 0.25, and 0.5 mg/kg ER-086526 leading tocomplete tumor suppression. Again, this represents an unusually wide5-fold therapeutic window. Tumor regrowth rates in the 0.5 mg/kgER-086526 group were delayed significantly beyond the 12.5 mg/kgpaclitaxel group; regrowth in the 0.25 mg/kg ER-086526 group wasslightly delayed relative to paclitaxel. Significantly, one and threemice in the 0.25 and 0.5 mg/kg ER-086526 groups (n5 10), respec-tively, became tumor-free by day 17; all of the mice had measurabletumors on day 10. These cured mice remained completely tumor-freefor an additional 7 months. Other than a 2% decrease in body weightsin the 0.5 mg/kg ER-086526 group during the first five daily injec-tions, there were no other indications of toxicity in any group witheither compound, as assessed by body weights or water consumption(data not shown). The slight decrease in body weights in the lattergroup resolved itself before the beginning of the second cycle of fivedaily injections and was not observed again.

Finally, effects of 0.125–1.0 mg/kg ER-076349 and ER-086526 onNIH:OVCAR-3 ovarian cancer xenograft growth were examined (Fig.7, G andH). During actual dosing periods, both compounds inhibitedtumor growth to about the same degree, with 0.5 and 1 mg/kg leadingto inhibition during dosing equivalent to paclitaxel at its empiricallydetermined MTD of 20 mg/kg. Two of six mice each in the 1 mg/kgER-076349 and ER-086526 groups and six of six mice in the pacli-taxel group became tumor-free by the end of dosing; these miceremained tumor-free through the end of the experiment (day 89). Aftercessation of dosing, increased efficacy of ER-086526 over ER-076349became evident in the 0.5 and 1.0 mg/kg groups; no dose of ER-076349 achieved the long-term complete regrowth suppression seenwith paclitaxel, whereas 1 mg/kg ER-086526 achieved completesuppression out to day 89, similar to paclitaxel. Significant bodyweight losses during dosing were only seen in the 20 mg/kg paclitaxeland 1 mg/kg ER-086526 groups (12 and 18% decreases, respectively);body weights in these groups recovered within about 10 days aftercessation of dosing (data not shown).

DISCUSSION

In the 15 years since its discovery, the extraordinaryin vitro andinvivo anticancer potency of halichondrin B (2, 10) has led to consid-erable interest in developing this agent as a new anticancer drug.Indeed, work on this compound by the NCI was ongoing as late as1999 (28). Nevertheless, the extremely limited supply of halichondrinB from marine sources has all but eliminated the possibility that thispromising agent might become a viable drug candidate. Fortunately,the complete synthesis of halichondrin B (11) and the subsequentdiscovery that its biological activity resides in its macrocyclic lactoneC1-C38 moiety3 (see also Refs. 12–14) created the possibility ofdeveloping structurally simpler, fully synthetic analogues that retainthe remarkable activity of halichondrin B. The resultant synthesis andevaluation of more than 180 macrocyclic analogues have culminatedin the identification of ER-076349 and ER-086526.

ER-076349 and ER-086526 differ only in their C35 alcohol andamine substituents, respectively. Both showed sub-nM in vitro growthinhibition against a variety of human cancer cell lines (Table 1), withthe alcohol ER-076349 being about 2-fold more potent than the amineER-086526 in most lines. Both were roughly 5–10-fold more potent

Fig. 6. Inhibition of polymerization of bovine brain tubulinin vitro by ER-076349,ER-086526, and vinblastine.In vitro polymerization of purified bovine brain in thepresence of 0mM (F), 1.5 mM (f), 3 mM (r), 5 mM (Œ), 7.5 mM (�), or 15 mM (3)concentrations of ER-076349 (A), ER-086526 (B), or vinblastine (C) was assessed asdescribed in “Materials and Methods.” Results are plotted as the extent of tubulinpolymerization (A340 nm) as a function of time and represent means of triplicate determi-nations at each time point. For graphic clarity, only data points from odd-numberedminutes are presented.

1018




Fig. 7. Inhibition of human tumor xenograft growthin vivo by ER-076349, ER-086526, and paclitaxel. Nude mice bearing MDA-MB-435 breast cancer (A, B), COLO 205 coloncancer (C, D), LOX melanoma (E, F), and NIH: OVCAR-3 ovarian cancer (G, H) xenografts were treated with the indicated concentrations of ER-076349 (A, C, E, G), ER-086526(B, D, F, H), and paclitaxel (all of the panels) as described in “Materials and Methods.” In the MDA-MB-435, COLO 205, and NIH: OVCAR-3 models, ER-076349 and ER-086526studies were run as side-by-side comparisons but are plotted separately for ease of viewing; the same control and paclitaxel groups are thus replicated in each pair of graphs. The twoLOX melanoma studies were run separately, each with its own control and paclitaxel groups. Periods of dosing are indicated bygray barsunderlying theX axes; specific details ofdosing schedules and routes of administration are presented in “Materials and Methods.” Plotted mean tumor volumes occurring after the onset of any complete remissions (see text)represent averages of only those animals continuing to bear measurable tumors.

1019




than vinblastine and paclitaxel, two tubulin-based antimitotic agentsrun as internal standards in each experiment.

Direct comparisons of ER-076349 and ER-086526 with halichon-drin B were not performed in the current study because of the lack ofavailability of either synthetic or natural halichondrin B. However, thelow- to sub-nM in vitro potencies measured for the two macrocyclicanalogues were similar to historical halichondrin B data from ourlaboratory and others, indicating that thein vitro potency of halichon-drin B had been retained during structural simplification. Thus, in anearly study with synthetic halichondrin B, we measured a mean IC50

of 0.4 nM against 17 human tumor cell lines representing eightdifferent cancer types (24), whereas other laboratories using naturalhalichondrin B have reported IC50 values of 0.08 nM, 0.3 nM, and 5 nMagainst murine B16 melanoma, murine L1210 leukemia, and PtK1normal kangaroo rat kidney cells, respectively (2, 6, 8). Moreover,performance of ER-076349 and ER-086526 in the 60-cell line screenof NCI was essentially identical to that of natural halichondrin B5

(data not shown). We conclude that virtually all of the growth inhib-itory potency was retained during structural simplification, includingremoving the entire C36-C54 polyether “left half” of halichondrin Band replacing the macrocyclic lactone ester of the latter with anonhydrolyzable ketone bridge.

Our results indicate that the growth inhibitory mechanisms ofER-076349 and ER-086526 are probably the same as those of hali-chondrin B. Thus, ER-076349 and ER-086526 induce G2-M cell cyclearrest and disrupt mitotic spindles, similar to previous results withsynthetic halichondrin B (24) and consistent with the tubulin-basedantimitotic mechanism of the latter (6–9). Moreover, our studies withthe biotinylated macrocyclic lactone analogue ER-040798 showeddirect binding to tubulin under affinity chromatography conditions(Fig. 5). Although ER-040798 was prepared before discovery of themacrocyclic ketone series, direct tubulin binding by ER-040798 isprobably representative of this entire class of drugs because simpli-fication from halichondrin B to the macrocyclic ketones involvednumerous incremental modifications made without significantchanges in growth inhibitory potency, G2-M arrest characteristics, ormitotic spindle disruption. Indeed, ER-076349 and ER-086526 di-rectly inhibit tubulin polymerizationin vitro with IC50 values (5.7mM

and 6.9mM, respectively) almost identical to the 7.2mM value reportedfor halichondrin B (6). We conclude that the anticancer activities ofER-076349 and ER-086526 derive from a tubulin-depolymerizingantimitotic mechanism similar or identical to that reported for hali-chondrin B.

ER-076349 and ER-086526 showed significantin vivo anticancerefficacy at doses well below 1 mg/kg in several human tumor xe-nograft models. Interestingly, despite somewhat lowerin vitro po-tency in cell growth and tubulin polymerization studies, ER-086526showed greaterin vivo efficacy, particularly in the MDA-MB-435breast and COLO 205 colon cancer models. One explanation for thismight be that ER-076349 is metabolizedin vivo more readily thanER-086526; pharmacokinetic studies are under way to investigate thispossibility. Another explanation might relate to our finding that mi-totic blocks induced by ER-086526 are much less reversible after drugwashout than those induced by ER-076349 (data not shown). Thiscould result in greaterin vivo tumor cell killing under the intermittentdosing schedules used in our studies. This observation, which corre-lates with potential forin vivo efficacy within the halichondrin Bmacrocyclic analogue series, will be described in a separate manu-script.6

In all of the four in vivo models, ER-076349 and ER-086526 were

effective at much lower doses compared with paclitaxel run at em-pirically determined MTD levels. For instance, potency differentialsfor ER-086526versuspaclitaxel based on complete tumor suppres-sion ranged from 20-fold in the NIH:OVCAR-3 model, to 40-, 50-,and 100-fold in the COLO 205, LOX, and MDA-MB-435 models,respectively. More importantly, ER-076349 and ER-086526 showedsignificantly wider in vivo therapeutic windows. For paclitaxel, em-pirically determined therapeutic windows were relatively small, 1.7 inthe MDA-MB-435 model and,2.0 in the LOX model. In markedcontrast, therapeutic windows for the macrocyclic ketone analogueswere much wider. In the MDA-MB-435 model, ER-086526 showed a4-fold therapeutic window, whereas in the LOX model, therapeuticwindows of 4- and 5-fold were observed for ER-076349 and ER-086526, respectively. We speculate that the wide therapeutic windowsseen with ER-076349 and ER-086526 contribute to their substantialinvivoefficacy, in that the ability to increase doses 4–5-fold above fullytumor-suppressive doses probably leads to more complete eradicationof residual tumor cells. This might explain the superiorin vivoefficacy of ER-086526 over paclitaxel MTD dosing in three of thefour models tested and the existence of complete cures in the LOXmodel by ER-086526. The reasons for the unusually wide therapeuticwindows of ER-076349 and ER-086526 are not currently known.

In summary, we have demonstrated highly potentin vitro and invivo anticancer activities of two fully synthetic, macrocyclic ketoneanalogues of halichondrin B, ER-076349, and ER-086526. Theseagents exert their anticancer effects by mechanisms currently indis-tinguishable from the microtubule-destabilizing effects of halichon-drin B. An unexpected but exciting finding was the existence ofunusually widein vivo therapeutic windows, which may contribute tothe remarkablein vivo efficacy of these new agents. Our encouragingearly results with ER-076349 and ER-086526 strongly support theircontinued development as novel anticancer agents for human use.

REFERENCES

1. Uemura, D., Takahashi, K., Yamamoto, T., Katayama, C., Tanaka, J., Okumura, Y.,and Hirata, Y. Norhalichondrin A: an antitumor polyether macrolide from a marinesponge. J. Am. Chem. Soc.,107: 4796–4798, 1985.

2. Hirata, Y., and Uemura, D. Halichondrins: antitumor polyether macrolides from amarine sponge. Pure Appl. Chem.,58: 701–710, 1986.

3. Pettit, G. R., Herald, C. L., Boyd, M. R., Leet, J. E., Dufresne, C., Doubek, D. L.,Schmidt, J. M., Cerny, R. L., Hooper, J. N., and Rutzler, K. C. Isolation and structureof the cell growth inhibitory constituents from the western pacific marine spongeAxinella sp. J. Med. Chem.,34: 3339–3340, 1991.

4. Pettit, G. R., Tan, R., Williams, M. D., Doubek, D. L., Boyd, M. R., Schmitt, J. M.,Chapuis, J-C., Hamel, E., Bai, R., Hooper, J. N. A., and Tackett, L. P. Isolation andstructure of halistatin 1 from the eastern Indian Ocean marine spongePhakelliacarteri. J. Org. Chem.,58: 2538–2543, 1993.

5. Gravelos, D. G., Lake, R. J., Blunt, J. W., Munro, M. H. G., and Litaudon, M. S. P.Halichondrins: cytotoxic polyether macrolides. European Patent Office, Munich,Switzerland: Publication number EP 0 572 109 A1, 1993.

6. Bai, R. L., Paull, K. D., Herald, C. L., Malspeis, L., Pettit, G. R., and Hamel, E.Halichondrin B and homohalichondrin B, marine natural products binding in thevincadomain of tubulin. Discovery of tubulin-based mechanism of action by analysisof differential cytotoxicity data. J. Biol. Chem.,266: 15882–15889, 1991.

7. Luduena, R. F., Roach, M. C., Prasad, V., and Pettit, G. R. Interaction of halichondrinB and homohalichondrin B with bovine brain tubulin. Biochem. Pharmacol.,45:421–427, 1993.

8. Bai, R., Cichacz, Z. A., Herald, C. L., Pettit, G. R., and Hamel, E. Spongistatin 1, ahighly cytotoxic, sponge-derived, marine natural product that inhibits mitosis, micro-tubule assembly, and the binding of vinblastine to tubulin. Mol. Pharmacol.,44:757–766, 1993.

9. Hamel, E. Natural products which interact with tubulin in thevinca domain: may-tansine, rhizoxin, phomopsin A, dolastatins 10 and 15 and halichondrin B. Pharmacol.Ther.,55: 31–51, 1992.

10. Fodstad, Ø., Breistøl, K., Pettit, G. R., Shoemaker, R. H., and Boyd, M. R. Compar-ative antitumor activities of halichondrins and vinblastine against human tumorxenografts. J. Exp. Ther. Oncol.,1: 119–125, 1996.

11. Aicher, T. D., Buszek, K. R., Fang, F. G., Forsyth, C. J., Jung, S. H., Kishi, Y.,Matelich, M. C., Scola, P. M., Spero, D. M., and Yoon, S. K. Total synthesis ofhalichondrin B and norhalichondrin B. J. Am. Chem. Soc.,114: 3162–3164, 1992.

12. Kishi, Y., Fang, F. G., Forsyth, C. J., Scola, P. M., and Yoon, S. K. Halichondrins andrelated compounds. U. S. Patent Office, Washington, D.C.: U. S. Patent 5,436,238,1995.

5 D. J. Newman, personal communication.6 M. J. Towle, Y. Kishi, M. J. Yu, and B. A. Littlefield, manuscript in preparation.

1020




13. Stamos, D. P., Sean, S. C., and Kishi, Y. New synthetic route to the C.14-C.38segment of halichondrins. J. Org. Chem.,62: 7552–7553, 1997.

14. Wang, Y., Habgood, G. J., Christ, W. J., Kishi, Y., Littlefield, B. A., and Yu, M. J.Structure-activity relationships of halichondrin B analogues: modifications at C.30-C.38. Bioorg. Med. Chem. Lett.,10: 1029–1032, 2000.

15. Littlefield, B. A., Palme, M. H., Seletsky, B. M., Towle, M. T., Yu, M. J., and Zheng,W. Macrocyclic analogs and methods of their use and preparation. U. S. PatentOffice, Washington, D.C.: U. S. Patent Application 09/334,488, 1999.

16. Amin, W., Karlan, B. Y., and Littlefield, B. A. Glucocorticoid sensitivity of OVCA433 human ovarian carcinoma cells: inhibition of plasminogen activators, cell growth,and morphological alterations. Cancer Res.,47: 6040–6045, 1987.

17. Finlay, G. J., Baguley, B. C., and Wilson, W. R. A semiautomated microculturemethod for investigating growth inhibitory effects of cytotoxic compounds on expo-nentially growing carcinoma cells. Anal. Biochem.,139: 272–277, 1984.

18. Mosmann, T. Rapid colorimetric assay for cellular growth and survival: applicationto proliferation and cytotoxicity assays. J. Immunol. Methods,65: 55–63, 1983.

19. Matthes, T. Antitubulin antibodies II. Natural autoantibodies and induced antibodiesrecognize different epitopes on the tubulin molecule. J. Immunol.,141: 3135–3141,1988.

20. Heukeshoven, J., and Dernick, R. Simplified method for silver staining of proteins inpolyacrylamide gels and the mechanism of silver staining. Electrophoresis,6: 103–112, 1985.

21. Burnette, W. N. “Western blotting”: electrophoretic transfer of proteins from sodiumdodecyl sulfate-polyacrylamide gels to unmodified nitro-cellulose and radiographic

detection with antibody and radiolabeled Protein A. Anal. Biochem.,112: 195–203,1981.

22. Riondel, J., Jacrot, M., Fessi, H., Puisieux, F., and Potier. Effects of free andliposome-encapsulated Taxol on two brain tumors xenografted into nude mice. InVivo (Athens),6: 23–28, 1992.

23. Sharma, A., Mayhew, E., and Straubinger, R. M. Antitumor effect of Taxol-contain-ing liposomes in a Taxol-resistant murine tumor model. Cancer Res.,53: 5877–5881,1993.

24. Towle, M. J., Aalfs, K. K., Budrow, J. A., Kishi, Y., and Littlefield, B. A.In vitro andin vivo anticancer properties and cell cycle effects of synthetic halichondrin B. Proc.Am. Assoc. Cancer Res.,36: 393 (Abstract #2342), 1995.

25. Jordan, M. A., Thrower, D., and Wilson, L. Mechanism of inhibition of cell prolif-eration byvinca alkaloids. Cancer Res.,51: 2212–2222, 1991.

26. Jordan, M. A., Toso, R. J., Thrower, D., and Wilson, L. Mechanism of mitotic blockand inhibition of cell proliferation by taxol at low concentrations. Proc. Natl. Acad.Sci. USA,90: 9552–9556, 1993.

27. Wood, D. L., Panda, D., Wiernicki, T. R., Wilson, L., Jordan, M. A., and Singh, J. P.Inhibition of mitosis and microtubule function through direct tubulin binding by anovel antiproliferative naphthopyran LY290181. Mol. Pharmacol.,52: 437–444,1997.

28. Mirsalis, J. C., Schindler-Horvat, J. E., Fairchild, D. G., Tomaszewski, J. E.,Donohue, S. J., and Tyson, C. A. Preclinical toxicology studies of halichondrin B inrats and dogs. Proc. Am. Assoc. Cancer Res.,40: 385 (Abstract #2550), 1999.

1021




2001;61:1013-1021. Cancer Res Murray J. Towle, Kathleen A. Salvato, Jacqueline Budrow, et al. Macrocyclic Ketone Analogues of Halichondrin B

Anticancer Activities of SyntheticIn Vivo and In Vitro

Updated version

http://cancerres.aacrjournals.org/content/61/3/1013

Access the most recent version of this article at:

Cited articles

http://cancerres.aacrjournals.org/content/61/3/1013.full#ref-list-1

This article cites 23 articles, 8 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/61/3/1013.full#related-urls

This article has been cited by 33 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected] at

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

Permissions

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

.http://cancerres.aacrjournals.org/content/61/3/1013To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/content/61/3/1013.full#ref-list-1

http://cancerres.aacrjournals.org/content/61/3/1013.full#related-urls

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

mailto:[email protected]



in vitro and in vivo anticancer activities of synthetic...

Documents