[CANCER RESEARCH 55. 2325-2333, June 1 1995]

Epothilones, a New Class of Microtubule-stabilizing Agents with a Taxol-like

Mechanism of Action

Daniel M. Bollag, Patricia A. McQueney, Jian Zhu, Otto Hensens, Lawrence Koupal, Jerrold Liesch, Michael Goetz,Elias Lazarides,1 and Catherine M. Woods2

Department of Pharmacology, Merck Research Laboratories, West Point, Pennsylvania 19486 ID. M. B., P. A. M., J. Z, E. L, C. WJ and Department of Natural ProductsChemistry, Merck Research Laboratories, Kahway, New Jersey 07065 ¡O.H., L K., J. L, M. C.¡

ABSTRACT

Tubulin polymerization into microtubules is a dynamic process, withthe equilibrium between growth and shrinkage being essential formany cellular processes. The antineoplastic agent taxol hyperstabilizespolymerized microtubules, leading to mitotic arrest and cytotoxicity inproliferating cells. Using a sensitive filtration-calorimetric assay todetect microtubule nucleating activity, we have identified epothilonesA and B as compounds that possess all the biological effects of taxolboth in vitro and in cultured cells. The epothilones are equipotent andexhibit kinetics similar to taxol in inducing tubulin polymerization intomicrotubules in vitro (filtration, light scattering, sedimentation, andelectron microscopy) and in producing enhanced microtubule stabilityand bundling in cultured cells. Furthermore, these 16-membered ma-crolides are competitive inhibitors of [3H]taxol binding, exhibiting a

50% inhibitory concentration almost identical to that of taxol in displacement competition assays. Epothilones also cause cell cycle arrestat the G2-M transition leading to cytotoxicity, similar to taxol. Incontrast to taxol, epothilones retain a much greater toxicity againstP-glycoprotein-expressing multiple drug resistant cells. Epothilones,therefore, represent a novel structural class of compounds, the first tobe described since the original discovery of taxol, which not only mimicthe biological effects of taxol but also appear to bind to the samemicrotubule-binding site as taxol.

INTRODUCTION

Taxol and taxotere are novel cancer chemotherapeutic agents thatstabilize cellular MTs.3 Taxol, a complex diterpene with a taxane ring

system, was discovered in 1971 in a screen for anticancer activity (1).Following the elucidation of the mechanism of action of taxol (2),clinical trials have established taxol as an anticancer agent withsignificant activity against various human solid tumors (3). Withearlier drug supply problems resolved, vigorous synthetic effortsunder way to modify the structure of taxol, and trials of taxol efficacyin combination therapy in progress, further improvements in taxane-

based chemotherapy are likely. Nevertheless, the complexity of thetaxol structure presents a major obstacle to facile chemical modification aimed at improving the solubility characteristics and side effectprofile of taxol. A novel class of drugs which stabilizes MTs mightstimulate the development of more effective cancer chemotherapeu-

tics with this mechanism of action.Microtubules are one of the fundamental structures comprising the

cytoskeleton of eukaryotic cells and are involved in such diverse

Received 1/11/95; accepted 4/5/95.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicale this fact.

1 Present address: Astral, Inc., San Diego, CA 92121.2 Present address: Alliance Pharmaceutical Corp., San Diego, CA 92121.3The abbreviations used are: MT, microtubule(s); MDR, multiple drug resistance;

MEM buffer, 100 mw 2-(/V-morpholiro)ethanesulfonic acid, pH 6.75-1 rriMethyleneglycolbis(ß-aminoethyl ether)-/V,JV,/V',A''-tetraacetic acid-1 mM MgCI2; CNS agar, cellulose-

nitrate salts agar; GNS3 agar, glucose-nitrate salts agar; NMR, nuclear magnetic resonance; EC50, 50% effective concentration; TÚNEL, biotin-terminal deoxytransferase endlabeling of nicked DNA; HRMS, high resolution mass spectrometry; HMQC, helero-nuclear multiple quantum coherence; HMBC, heteronuclear multiple bond connectivity;HR-EIMS, high resolution electron impact mass spectrometry.

cellular processes as cell division, locomotion, and intracellular transport (4). MTs are highly dynamic polar structures, with growthfavored at the plus end and shrinkage more prevalent at the minus end.However, both MT ends experience periods of growth and shrinkage,a phenomenon described as dynamic instability (5). In vertebratecells, the centrosome acts as the major site of MT nucleation (micro-tubule-organizing center) by lowering the critical concentration of

tubulin required for polymerization and anchoring the minus ends ofthe resultant MTs (6, 7).

Taxol preferentially binds the polymeric MT form of tubulin ina 1:1 stoichiometry with the aß-tubulin heterodimer subunits, witha Ap of ~1 /AM(8). Taxol binding markedly reduces the rate of

aß-tubulin dissociation, hence augmenting and stabilizing the MT

pool (9). In vitro taxol has also been shown to nucleate tubulinpolymerization into MTs, eliminating the requirement for GTP innormal polymerization (2, 10). Within cells this effect is manifested by micromolar taxol levels overriding the centrosomalmicrotubule-organizing center function and inducing the appearanceof many short non-centrosomally linked bundles throughout the cyto

plasm (11). In rapidly cycling cultured human cells, taxol induces a blockat the transition between G, and M phase (12). Recent studies haveclearly indicated that it is this mitotic arrest rather than the disruption ofinterphase MT function which is the actual mechanism behind the antineoplastic activity of taxol (13, 14).

Although taxol has shown efficacy against refractive ovarian (15),metastatic breast (16), head and neck (17), melanoma (18), and lung(19) cancer, its clinical usefulness is limited by its side effect profile(neutropenia, peripheral neuropathy, and alopecia) (20) and the factthat its low solubility necessitates that taxol be delivered in Cremo-

phor. Cremophor delivery in itself can affect cardiac function andcause severe hypersensitivity responses (20). Furthermore, taxol is asubstrate for P-glycoprotein which pumps many cytotoxic compounds

out of MDR cells (21). Multiple drug resistance represents a majorlimitation of many cancer interventions. The complex taxane ringstructure of taxol is not readily amenable to chemical manipulation toimprove the therapeutic index and solubility properties of this class ofcompounds. We therefore sought to screen for a different structuralclass of compounds that would mimic the MT-stabilizing properties

of taxol.Here we report that epothilones A and B, 16-membered macrolides,

mimic all the biological effects of taxol, both in vitro and in culturedcells. Independently, Hoefle et al. (22) have described epothilones Aand B as having antifungal and cytotoxic activity. Competition studiesreveal that epothilones act as competitive inhibitors of taxol binding toMTs, consistent with the interpretation that they share the sameMT-binding site and posses a similar MT affinity as taxol. Epothi

lones appear to possess one advantage over taxol; namely, they exhibita much lower drop in potency compared to taxol against a multipledrug-resistant cell line. The epothilones represent the first class of

compounds to be described in the two decades following the originaldiscovery of taxol which mimic the MT-stabilizing effect of the

taxane ring structure.

2325

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EPOTHILONES. A NEW CLASS OF MT-STABILIZING AGENTS

MATERIALS AND METHODS

Materials

Taxol and GTP were purchased from Sigma. Uranyl acetate was obtainedfrom Ernest F. Fullam, Inc. For immunofluorescence microscopy, anti-chickß-tubulinmonoclonal (Amersham) and FITC-conjugated anti-mouse (Cappel)

antibodies were used. HeLa cervical epithelioid carcinoma cells and Hs578Tand Hs578Bst breast carcinoma and Rati cells were obtained from the American Type Culture Collection. The multiple drug-resistant human carcinomacell line KBV-1 and its parental line, KB3-1, were provided by I. Pastan (23).

Tubulin Purification

Microtubule protein was purified from bovine brains essentially as described by Asnes and Wilson (24). Brains were rinsed in cold saline solutionwithin l h of sacrifice. Whole brains were minced and homogenized with 500ml lysis buffer per brain [100 mM 2-(N-morpholino)ethanesulfonic acid. pH6.75-1 HIM ethylcneglycol bis(ß-aminoethyl ether)-/V,AW,Ar-tetraacetic ac-

id-1 mM MgCU-1 mM DTT-100 fig/ml phenylmethylsulfonyl fluoride] for 60

s in an Osterizer blender and then incubated on ice for 30 min. The extract wascentrifuged at 4°Cfor 45 min at 30,000 X g. GTP was added to the supernatantat a final concentration of I mM and the mixture was incubated at 37°Cfor 30min. This extract was centrifuged at 37°Cfor 30 min at 10,000 X g and thepellet was resuspended in one-tenth the 4°Csupernatant volume of lysis buffer.This solution was incubated on ice for 30 min and then centrifuged at 4°Cfor

40 min at 39,000 X i>.The supernatant was adjusted to 2.5 mM GTP, incubatedat 37°Cfor 15 min, and centrifuged at 100,000 x g for 20 min at 37°C.The

supernatant was removed and the pellets were frozen in liquid nitrogen andstored at -80°C. On the day of the assay, the microtubule pellets wereresuspended in MEM buffer at 4°Cand spun at 100,000 X g for 20 min. The

supernatant protein concentration was determined by the bicinchoninic acidassay (Pierce).

Turbidity and Sedimentation Assays

To monitor the change in turbidity in the microtubule protein solution, thetest compound was added to a cuvet containing I mg/ml microtubule protein.The cuvet was incubated at 37°Cfor 30 min in a Beckman D640 spectropho-

tometer equipped with a Peltier heating unit, and the absorbance was measuredat 340 nm. The change in absorbance over the 30-min incubation was meas

ured and compared to the absorbance change induced by 1 mM GTP. Followingpolymerization, microtubules were separated from unpolymerized tubulin bycentrifugation at 37°Cfor 5 min at 100,000 X g in a TLA100 rotor using a

Beckman TLX benchtop ultracentrifuge. The protein concentration of thesupernatant before and after centrifugation was determined by the bicinchoninic acid method. The quantity of sedimcnted microtubule protein was determined by comparing the supernatant protein concentration before and aftercentrifugation and compared to that sedimented by I mM GTP.

Filtration-Colorimetric Assay

Microtubule protein (0.25 ml of I mg/ml) was placed into an assay tube and2.5 fil of the test compound were added. The sample was mixed and incubatedat 37°Cfor 30 min. Sample (150 /j.1) was transferred to a well in a 96-well

Millipore Multiscreen Durapore hydrophilic 0.22-jj.m pore size filtration plate

which had previously been washed with 200 /j.1of MEM buffer under vacuum.The well was then washed with 200 /j.1of MEM buffer.

To stain the trapped protein on the plate, 50 ¿ilamido black solution [0.1%naphthol blue black (Sigma)/45% methanol/10% acetic acid] were added to thefilter for 2 min; then the vacuum was reapplied. Two additions of 200 ¿ilamidoblack destain solution (90% methanol/2% acetic acid) were added to removeunbound dye. The signal was quantitated by the method of Schaffner andWeissmann (25). Briefly, 200 jj,l of elution solution (25 mM NaOH-0.05 mMEDTA-50%) ethanol) were added to the well and the solution was mixed witha pipet after 5 min. Following a 10-min incubation at room temperature, 150/J.1of the elution solution were transferred to the well of a 96-well plate and the

absorbance was measured on a Molecular Devices Microplate Reader.With this assay, we were able to test over 7000 samples in less than 6

months. Nearly 6% of all plant extracts tested gave positive results, whereasless than 0.5% of samples from other sources (marine, insect, microbial)generated positive results.

Purification and Identification of Epothilones

Bacterial Strains and Growth Media. The cellulose-degrading myxobac-

terium, Sorangium cellitlosum strain SMP 44, was obtained from the culturecollection of J. E. Peterson. Emporia State University, Emporia, KS, and wasfound to form bright orange fruiting body structures on media containingcellulose. The fruiting bodies of this strain were routinely preserved as air-dried preparations on sterile 6-mm paper disks. Primary cultivation was ac

complished by inoculation of the dried fruiting bodies onto CNS agar andincubating the medium at 27°Cfor 10-15 days. CNS agar contained (g/liter):

KNO,, 0.5; Na2HPO4, 0.25; MgSO4-7H2O, 1.0; HEPES, 2.4; FeCl2, 0.01;

Difco agar, 15.0. Before the medium was autoclaved, the pH was adjusted to7.0-7.2 with dilute HCI. After solidification of the medium in sterile 90-mm

Petri dishes, sterile Whatman No. 1 filter paper was applied to the surface ofthe agar. Dried fruiting bodies characteristically required an incubation periodof about 10-12 days before germination and the emergence of cell swarms was

seen. Growing cells were transferred from the edges of the colony onto freshmedium before 4 weeks of incubation.

Sufficient biomass for inoculation of production medium was producedusing a GNS3 agar medium composed of (g/liter): glucose, 15.0; KNO3, 1.5;Na,HPO4, 0.625; MgSO4-7H20, 0.25; HEPES, 2.4; CaCU, 0.05; Difco agar,15.0. The pH was adjusted to 7.0-7.2 with dilute HCI and 1 ml each of iron

molybdate and A5 trace element solution was added. The iron molybdatesolution contained (g/liter): ferric citrate 6.0; Na2MoO4, 0.5; and the A5 traceelement solution contained (g/liter): H,BO4, 2.86; MnCI2-4H2O, 1.81; ZnSO4,0.222; CuSO4-5H,O, 0.079; and Co(NO,)2-6H,O, 0.049. The production

medium, So8, was composed of (g/liter): KNO,, 0.5; Na2HPO4, 0.1;MgSO4-7H2O, 0.25; CaCl,, 0.05; Difco agar, 15.0. The pH was adjusted to

7.0-7.2 with dilute HCI. and 1 ml each of iron molybdate and A5 trace element

solutions was added. After sterilization 50 ml of the agar medium weredispensed into sterile 150 x 15-mm Petri dishes.

A standard cell inoculum was prepared by scraping growth from CNS agarand point inoculating GNS3 agar with small clumps of cells. After 10 days ofgrowth on GNS3 agar the entire cell mass was scraped off, diluted 3-fold with

sterile water, and homogenized with a sterile glass Dounce tissue grinder(Wheaton Glass Co., Millville, NJ). The resulting cell suspension was dilutedwith an equal volume of a sterile cellulose solution (ball-milled Whatmanchromatography paper; 50 g/200 ml distilled water) and 2-ml portions of the

resulting cell suspension were evenly spread over the surface of each 50 ml ofproduction medium.

The production medium was incubated at 28°Cand 85% relative humidity

for 8 days in an Adolf Kuehner model 15F-4-V environmental cabinet (Sulzer,

Woodbury. NY).Purification. Extraction of the epothilones from a batch of 80 plates of the

agar-based fermentation of 5. cellulosuni was accomplished by blending the

agar and mycelium with methyl ethyl ketone (40 ml/plate), followed byovernight stirring. After filtration, the solvent was evaporated to dryness undervacuum, resulting in 450 mg of oily residue. This was readily dissolved in 15ml méthylènechloride and fractionated on an 80 cm' E. Merck Silica Gel 60

column, eluting with méthylènechloride containing increasing amounts ofmethanol. Epothilone A and B were found in fractions corresponding to the 3%methanol wash; these afforded 42 mg of residue upon evaporation. Furtherpurification was achieved by gel filtration on a 70-car Pharmacia SephadexLH-20 column in methanol; the target compounds eluted after 0.7-0.8 columnvolume (15 mg residue). Semipreparative HPLC followed on a 9 x 250-mm

Zorbax RxC8 column clutcd at 4 ml/min with a gradient of acetonitrile in water(40-60% acetonitrile). This afforded pure preparations of A and B (2.7 and 0.9

mg, respectively).The homogeneity of the isolated materials was ascertained in several TLC

systems [e.g., E. Merck Silica Gel 60F2M, méthylènechloride:methanol (95:5),A,(A) 0.40 and R¿B)0.45; méthylènechloride:ethyl acetate (66:33), /?,{A)0.30 and R,<B) 0.36], by HPLC (Zorbax RxCK column maintained at 50°C,1ml/min 45% aqueous acetonitrile, k' 7.0 for A and 8.2 for B). and by NMR.

Time courses of fermentations could be readily monitored for compoundproduction by HPLC (column as above, 55% aqueous acctonitrile. k' 2.6 and

3.1), analyzing 10-/¿Ialiquots of 25-fold concentrated methyl ethyl ketone

extracts prepared as above and reconstituted in methanol.Structure Determination. I3C NMR analysis of compound A in CD2C12,

indicated the presence of 20 carbons (6 in CH,, 5 in CH2, 8 in CH, 1 in COX).

2326

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

KPOTHIl.ONhS. A NEW CLASS OF MT-STABIUZING AGENTS

6 of which were extensively exchange-broadened and well short of the 26carbons required by the HRMS-derived molecular formula C\,,H,,,NO,,S (m/-

on a Jeol SX-I02A mass spectrometer using perfluorokerosene as internal

standard: found. 493.2495: calculated, 493.2498). HMQC data suggested anadditional carbon at 116.6 ppm whereas HMBC analysis indicated another fivequaternary carbons (53.3, 138.3, 151.9, 165.7, ~219 ppm) which were completely exchange-broadened in the !1C NMR spectrum. With the exception of

the ketone carbonyl carbon all 26 carbons were observed as sharp signals onspiking the solution with CD,OD (~10%). On the basis of extensive twodimensional NMR ['H]-['H] (correlated spectroscopy and total correlatedspectroscopy) and ['H]-["C] (HMQC, HMBC) correlation studies the 2-meth-

ylthiazole 16-membered macrolide A was proposed (Fig. 1). The structure wasconfirmed by HR-EIMS analysis.

Similarly, component B was assigned as the methyl homologue on the basis ofthe HRMS-derived molecular formula C27H4INO,,S (in/-: found. 507.2654; calculated 507.2654) and the extra methyl singlet resonance at 81.26 in the 'H NMR

spectrum. The position of the methyl group on the epoxide ring followed clearlyfrom the COSY 'H NMR data. But for these modified features the spectra were

almost identical to those of A. HR-EIMS fragmentation data also clearly allowedthe methyl group to be located at C-12. Comparison of the structures and spec-

troscopic data with those of epothilone A and B recently reported in the patentliterature (22) suggests that they are the same compounds.

Cold and Calcium Stability Studies

For the cold stability studies, microtubulc protein was treated with the testcompound and incubated at 37°Cfor 30 min. One portion of the sample was

transferred to a test tube on ice for 30 min while the remainder was maintainedat 37°C.Both portions of the sample were then centrifuged as described above

for the sedimentation assay. The percentage of maximal polymerization wascalculated relative to a 1 mM GTP control incubated at 37°Cfor 60 min. To

assay calcium stability, microtubule protein was prepared in MEM bufferlacking ethyleneglycol bis(/3-aminoethyl etherJ-iV.iV./V'./V'-tetraacetic acid.

The microtubule protein solution was adjusted to 3 mM CaCI; and the testcompound was added. The turbidity of the solution during a 30-min 37°C

incubation was monitored and compared to the turbidity produced by 1 mMGTP addition to microtubule protein in MEM buffer.

Competition Studies

Microtuhulc protein (0.4 mg/ml) in MEM containing 0.1% bovine serumalbumin and 0.05% Triton X-100 was preincubated for 20 min with 100 /J.MGTP and 7.5 nM taxol at 37°C.Preliminary experiments had defined these

conditions as inducing essentially maximal microtubule polymerization, thusproviding a roughly constant number of taxol binding sites over the range ofdrug concentrations used in the displacement studies. Tritiated taxol (100 nM:Moravek Biochemicals) and the competing compound were then added and thesamples were incubated for an additional 30 min. MTs were then pelleted byccntrifugation for 10 min at 100,000 X g in a TLA 100 rotor in a Beckman

OH

'OH

Epothilone A : k-11Epothilone B : R=CH3

Fig. 1. Molecular structure of epolhilones.

TLX benchtop ultracentrifuge. The pellets were carefully washed twice with100 /xl MEM containing 0.1% Triton X-100, resuspended in 100 /xl 0.1 N

NaOH. and transferred to a scintillation vial. After neutralization with 100 /¿Iof 0.1 N HC1. 5 ml of ReadyProtein*™ scintillation fluid (Beckman) were

added and the sample was counted in an LKB model 1209 Rackbeta scintillation counter. Competitor taxol concentrations above IO /J.M resulted innonspecific sample precipitation.

Electron Microscopy

Microtubule protein solution (0.3 mg/ml) was incubated at 37°Cfor 30 min

after addition of test compound. The protein sample was stained with 0.5%uranyl acetate as described by McEwen and Edelstein (26) and the grids wereviewed on a Philips CM 12 electron microscope.

Immunofluorescence

Cells were incubated for 4 h in 25 or 1 /AMtaxol. Cells were processed forimmunofluorescence as described previously (27) using a monoclonal antibody against chicken brain ß-tubulin.

Mitotic Block and Cytotoxicity Studies

Mitotic block, aberrant mitosis, and cytotoxicity were evaluated as described elsewhere.4 Briefly, cells were plated either in 48-well plates (for

trypan blue and cell counting) or onto No. 1 coverslips. After 24 h. the cellswere treated with drug and were scored at regular intervals. For the cytotoxicity analysis, both attached cells and cells floating in the media were countedand scored as trypan blue positive or negative. Concurrently, coverslips andaliquots of cells in the culture supernatant were fixed and stained with Hoechst33342 (Boehringer Mannheim) in PBS. These cells were scored for cellsblocked at the G,-M transition and aberrant mitosis.

DNA Labeling and Agarose Gel Electrophoresis

Isolation of genomic cellular DNA is described by Sambrook el ul. (28).End labeling of DNA using terminal transferase was carried out as describedby Tilly el al. (29) except that the transferase reaction was incubated for 30

RESULTS

A Novel Filtration-Colorimetric Assay Can Detect Microtubule

Polymerization Induced by Low Concentrations of Taxol. Wedeveloped a novel sensitive microtubule filtration assay to enablelarge scale screening for compounds that would mimic the ability oftaxol to induce tubulin polymerization. Tubulin solutions comprisedof aßheterodimers incubated in the presence of either GTP or taxolyield microtubule polymer. Tubulin solution filtration through a0.22-ju.m pore size filter was found to separate tubulin heterodimer

subunits effectively from MT polymer, the latter being readily visualized in a quantifiable manner by amido black staining. With theconditions defined in "Materials and Methods," this filtration-colori-

metric assay could readily detect the MT-nucleating activity of taxol

down to 100 nM (Fig. 2/1). which is markedly more sensitive thanpreviously described assays. Standard turbidity or sedimentation assays require at least 1 /AMtaxol in the absence of GTP for detection ofpolymerization (Fig. 2B: Ref. 9). To distinguish between nonspecificprotein aggregation or precipitation versus bona fide MT induction,follow-up electron microscopy analysis of incubation mixes wascarried out (see below). The filtration-colorimetric assay was readily

adapted to large scale screening with the use of 96 well plate technology. One worker could easily test over 300 samples in 1 day. Thus,the filtration-colorimetric assay provides a more sensitive method for

detecting tubulin polymerization as a high throughput screen.

4 C. Woods et ai, manuscript in preparation.

2327

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EPOTHILONES, A NEW CLASS OF MT-STABILIZING AGENTS

110

100

90

3 80

«70

O 60

J 50

7 40

:»20

10

0

- TaiolGTP

"""1 '

10"' IO"8 IO'7 IO"6 IO"5 IO"4 IO"3 IO"2 IO"1

Concentration (M)

110 -

100 -

90 -

80

70 -

60

50

40 -

30 -

20 -

10 -

0

10"'

Concentration (M)

Fig. 2. Effect of GTP and taxol on tubulin polymerization. A, filtration-colorimetricassay as described in "Materials and Methods." After measurement of the absorbance, the

percentage signal was calculated relative to the maximal absorbance produced by GTPincubation. B, turbidity assay as described in "Materials and Methods."

Epothilones, a New Structural Class of Compounds with Tu-bulin-polymerizing Properties. Subsequent screening of variousplant, marine, insect, and fermentation extracts for tubulin polymerization activity revealed an extract from S. celluhsum to be an effective promoter of GTP-independent tubulin polymerization into MTs.

This extract was fractionated with an initial silica gel column followedby a gel filtration step. Two pure peaks of activity eluted from a finalreverse phase chromatography step. Subsequent analysis by massspectrometry and NMR resulted in the identification of epothilones Aand B (Fig. 1; Mr 493.25 and 507.27, respectively). The epothilonesinduced tubulin polymerization in the filtration assay at concentrations paralleling those of taxol (Fig. 3). As with taxol, the dose-

response curve for the epothilones is very steep in this assay, reflecting extensive MT polymerization occurring above the minimalnucleating drug concentration. Turbidity and sedimentation assays ofthe epothilones also showed induction of tubulin polymerization atconcentrations similar to those of taxol (data not shown). To determine whether the epothilones induced bona fide MT formation ratherthan nonspecific protein precipitation, samples of tubulin proteinincubated in the presence of GTP, taxol, or epothilone A or B wereanalyzed by electron microscopy. All four treated samples displayedfilamentous arrays while untreated protein samples showed no proteinorganization (Fig. 4). At higher magnification (Fig. 5), the tubulinfilaments induced by treatment with epothilones A and B were shownto have a diameter of approximately 25 nm and possessed 8-9protofilament-like striations, a structure identical to that of MTs

induced by taxol.A search for compounds with homology to the epothilones revealed

a family of 16-membered macrolide antibiotics. Erythromycin, chal-comycin, carbomycin, and rosamycin were tested with the filtration-

colorimetric assay, but none of these compounds displayed activity(results not shown). Interestingly, a different 16-membered macrolide,the MT-depolymerizing drug rhizoxin, possesses tubulin bindingproperties (KD 1.7 X 10~7 M; Ref. 30).

Epothilones Confer Resistance against Cold- or Calcium-in

duced MT Destabilization. Taxol has been shown to stabilize MTsagainst cold-induced and calcium-induced depolymerization (2). To

determine whether epothilones could also mimic these effects, wetested for the ability to confer stability to microtubules exposed tocold or calcium. Microtubules treated with l /UMepothilone A, epothilone B, or taxol were resistant to depolymerization after incubation at

4°Cfor 30 min whereas samples treated with 1 HIMGTP depolymer-

ized (Table 1). Similarly, 10 JU.Mtaxol or epothilone A or B inducedtubulin polymerization in the presence of 3 mM calcium, whereas 1mM GTP did not. Thus, the epothilones share with taxol the ability toconfer resistance to calcium or cold treatment.

Epothilones Are Competitive Inhibitors of Taxol Binding toMTs. Competition assays of [3H]taxol binding to preformed MTs

revealed that both epothilone A and epothilone B displaced 100 nM[3H]taxol with a similar 50% inhibitory concentration and slope to

unlabeled taxol (Fig. 6). Preliminary experiments had established thatunder the conditions of the assay (100 /XMGTP and a minimum of10~8 Mup to 3 X 10~6 Mtaxol), the amount of MT polymer (therefore,

the number of taxol-binding sites) remained essentially constant in aconsistent and reproducible manner. At 10~5 M taxol, the amount of

MT polymer increased dramatically, and consequently the maximumcompeting taxol concentration that could be used was 10~5 M. Satu

ration studies of [3H]taxol binding to preformed MTs confirmed the

KD for taxol binding to MTs to be 0.55 /UM,similar to the originalobservations of Parness and Horwitz (Ref. 8; data not shown). Underconditions of 100 nM [3H]taxol, the 50% inhibitory concentration

values for displacement by taxol, epothilone A, and epothilone B were3.6, 2.3, and 3.3 JAM,respectively. The fact that all three compoundsgave very similar slopes is consistent with the interpretation that theepothilones compete for the same binding site as taxol.

Epothilones Induce All the Classical Effects of Taxol on MTArrays in Cultured Cells. Taxol-treated cells develop characteristic,

extensive MT bundles that arise randomly throughout the cytoplasm.At lower concentrations (10~8 to IO"7 M), the interphase MT arrays

remain generally unaffected, and taxol selectively affects mitoticspindle MT formation and function, leading to mitotic arrest andcytotoxicity (13). As illustrated in Fig. 7, epothilones A and B inducedidentical changes in intracellular MT arrays as taxol. Rati cells,chosen here for their large flattened morphology, treated for 4 h with25 /XMtaxol (Fig. IB), epothilone A (Fig. 1C), and epothilone B (Fig.ID), all developed extensive MT bundles. Note that these bundlesarise independently from the centrosome. Thus, as reported previouslyfor taxol (11), this new class of MT-stabilizing agent overrides the

100

90

-soliS 70o.0 60

1 50

*40

*30

20

10 -

* Taxol* Epothilone A* Epothiloae B

250 500 750

Concentration (nM)

1000

Fig. 3. Effect of taxol, epothilone A, and epothilone B on tubulin polymerizationdetected by filtration-colorimetric assay as described in "Materials and Methods" and

legend to Fig. 2.

2328

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EPOTHILONES, A NEW CLASS OF MT-STABIL1ZING AGENTS

1

Epothilone A Epothilone B

Fig. 4. Electron micrographs of MTs assembled with no addition, 1 mM GTP, 500 nM taxol, 500 nM epothilone A, or 500 nM epothilone B. Negatively stained samples, X 18,(KK).

MT-nucleating activity of the centrosome in interphase cells. Similar

bundling activity was observed in HeLa, Hs578T, and Hs578Bst cells(data not shown). At 5 X 10~9 to 10~6 M taxol (Fig. IE), epothilone

A (Fig. IF), and epothilone B (Fig. 1C), the primary effect was seenin cells entering mitosis. All three agents induce similar multipolarspindles in HeLa cells, eventually resulting in arrest at the G2-M

transition. As noted previously for taxol (13), it is striking that even atconcentrations that override centrosomal MT nucleation (1(T8-10~7

M), these taxol- or epothilone-stabilized MTs still undergo massive

reorganization to form spindle arrays, albeit abnormal ones. In addition to the gross rearrangements of cellular MT arrays described

above, epothilones also stabilized cellular MTs from cold-induced

depolymerization similarly to the action of taxol as shown in Fig. 8.The cytotoxic effects of taxol in human cells have been shown to

correlate with arrest at the G2-M transition (Refs. 12, 13, 14, and 31;

see also Footnote 4). Therefore, to determine whether epothilonescould also mediate similar effects, we analyzed G-.-M block by

Hoechst staining and cell death by trypan blue exclusion followingexposure of 100 nM taxol, epothilone A, or epothilone B to HeLa orHs578T cells over a 72-h period. Care was taken to score unattached

cells in the supernatant also, since cells detach from the substratefollowing a few h of mitotic arrest. As shown in Fig. 9, all three agents

WTaxol

Bar = 0.298 pm

Epothilone A Epothilone B

Fig. 5. Electron micrographs of MTs assembled with 500 nM taxol, 500 nM epothilone A, or 500 nM epothilone B. Negatively stained samples, X 168,000.

2329

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EPOTHILONES. A NEW CLASS OF MT-STAB1LIZ1NG AGENTS

induced identical kinetics of G2-M block and identical lag and kinetics

of cytotoxicity. A small percentage of cells, more evident in theHs578T cells, appeared to passage through aberrant mitosis, as manifested by multimininucleation in the subsequent G, stage. Aberrantalignment of chromosomes at the metaphase plate preceded and

Table 1 Effects of drug treatment on MT polymerization following incuhalion underdestabilizing conditions

Values represen! Ihe percentage of MT polymerization relative to maximal polymerization with GTP incubated under nondestabilizing conditions as described in "Materialsand Methods." The controls for the cold incubation experiment were GTP-treated MTsincubated at 37°C.The controls for the calcium incubation experiment were GTP-treated

MTs incubated in the absence of calcium.

After 4°Cincubation1

HIMGTP1

/AMtaxol1 /AMepothilone A1 /AMepothilone B%36

9211192With

calciumincubation1

mMGTP10/AMtaxol

10 JAMepothilone A10 (AMepothilone B%0

109119115

250000n

•O

I

Q.Q

200000-

150000-

100000-

50000-

0

TaxedEpothilone A

Epothilone B

l 1 1—o 10-'° io-8 106 io-4 io-2

Competitor Concentration (M)

Fig. 6. Competition for [3H]taxol binding by taxol, epothilone A, or epothilone B.Steady state MTs were treated with 100 nM [3H]taxol and competing drug for 30 min and

sedimented. Tritialed taxol sedimenting with MTs is plotted as a function of competingtaxol, epothilone A, and epothilone B. Displacement 50% inhibitory concentrations were3.6 (AM(taxol), 2.3 (AM(epothilone A), and 3.3 (AM(epothilone B). The data were plottedwith a Windows-based Prism program.

Fig. 8. Immunofluorescence images of Rail cells treated with ß-tubulin antibodyfollowing l h of incubation at 4°C.Cells were previously untreated (A) or incubated for

4 h with 25 /AMtaxol (B), 25 /AMepothilone A (C), or 25 /AMepothilone B (D). No tubulinstaining remained in the untreated cells whereas MT bundles in the drug-treated samples

resisted depolymerization.

seemed to underlie multimininucleation. Multimininucleated cellsalso died but with much slower kinetics. Careful analysis of the timecourse of these changes showed that multimininucleation becameprevalent only at later time points and appeared to represent thefraction of cells entering mitosis after —24-30 h following taxol

addition. Epothilone B showed a slight difference in that a largerpercentage of cells showed multimininucleation (up to 38% comparedwith 10-15% for epothilone A or taxol with Hst578T cells) around

the EC5I) in both cell lines. As described by Kung et al. (32) for taxol,the majority of rodent Rati cells were tolerant of the prolongedabnormal mitosis induced by all three drugs, becoming arrested in thesubsequent G, stage as multimininucleated cells.

The EC50 for these effects was determined by scoring these parameters for a range of drug concentrations. Table 2 outlines the EC5U

Fig. 7. Immunofluorescence images of Rati fi-broblasts labeled with ß-tubulin antibody. A-D,induction of MT bundles by high drug concentrations after 4 h incubation. A, no addition; B, 25 /AMtaxol; C 25 JAMepothilone A; D, 25 (AMepothiloneB. £-//, induction of multipolar spindles in cells

entering mitosis by low drug concentrations after 4h incubation. £,no addition, normal mitotic spindle: F, 1 (AMtaxol; C, 1 JAMepothilone A; H, l (AMepothilone B. Fluorcscein-labeled cells, X 40.

2330

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EPOTHILONES, A NEW CLASS OF MT-STABILIZING AGENTS

0 12 24 36 48 60 72

Hours Exposed to Drug

Fig. 9. Analysis of mitotic arrest and cytotoxicity time course induced by taxol,epothilone A, or epothilone B in Hs578T (A) and HeLa (fl) cells. , mitotic arrest aspercentage of all cells; , cytoloxicity as percentage of all cells. Q, control; O, 100nMtaxol; O, 100 nMepothilone A; A, 100 nMepothilone B. The same drug concentrationscause cytotoxicity ~24 h after blocking the cells in mitosis. Micrographs of fluorescein-

labeled cells, X 63.

isolated DNA confirmed that epothilone induced the DNA ladders consistent with internucleosomal nicking typical of apoptosis as also seenwith taxol (Fig. 10). In situ TÚNEL labeling revealed that DNA nickingoccurred specifically in the G2-M-blocked cells (data not shown).

Epothilones Are Cytotoxic to Multiple Drug-resistant Cells.

Since one major form of acquired resistance to taxol is by increasedexpression of the P-glycoprotein pump which serves to lower the

intracellular drug concentration in MDR cells (21), we compared thepotency of taxol versus the epothilones in cell lines lacking or expressing P-glycoprotein (Table 3). In the parental KB3-1 cell linewhich does not express elevated P-glycoprotein, the sensitivity to all

three drugs was very similar, with EC50 values of 1.2 nM (taxol), 13HM(epothilone A), and 15 nM(epothilone B). However, in the K.BV-1P-glycoprotein-expressing cells treated with epothilone A or B, theepothilones showed only a 4-12-fold drop in potency (EC50 values of160 and 58 nM, respectively) compared to an approximately 20,000-

24 hr

1 2 3

Table 2 Effects of taxol and epothilones on mitolic arrest and c\totoxicil\The cells were plated either in 48-well plates (for trypan blue and cell counting) or onto

No. 1 coverslips and were scored after 24 h (mitotic arrest) or 72 h (cylotoxicity). For thecytotoxicity analysis, both attached cells and cells floating in the media were counted andscored as trypan blue positive or negative. Concurrently, coverslips and aliquots of cellsin the culture supernatant were fixed and stained with Hoechst 33342 (BoehringerMannheim) in PBS. These cells were scored for cells blocked at the G2-M transition andaberrant mitosis. The EC5O values shown below were derived using a Windows-basedPrism program (GraphPad Software, Inc.) with data generated from triplicate experiments.The R for curve fit is above 0.985 for each estimation.

CelltypeHeLaHs578TTreatmentTaxolEpothilone

AEpothiloneBTaxolEpothilone

AEpothilone BMitotic

arrest(nM)730

321073Cytotoxicity

(nM)2039

401513

6

values for mitotic arrest and cytotoxicity. The results demonstrate thatthe potencies of all three drugs are essentially identical in inducingmitotic arrest as well as cytotoxicity and are well below concentrations required for gross changes in interphase MT arrays (i.e., abovel /J.M).

Taxol-induced cytotoxicity appears to proceed via an apoptoticpathway (33, 34).* Since DNA fragmentation is a commonly used

hallmark of apoptosis, we used DNA gel electrophoresis and in situTÚNEL labeling of human HeLa cells treated with taxol, epothiloneA, or epothilone B. Agarose gel electrophoresis of 32P-end-labeled

Fig. 10. Autoradiogram of agarose gel containing genomic cellular DNA from cellstreated with taxol or epothilone A. HeLa cells were treated with drugs for 24 h andprocessed as described in "Materials and Methods." Lane I, untreated control; Lane 2, 1

(IM taxol; Lane 3, 1 (¿Mepothilone A. Arrowheads, a 1-kb DNA ladder.

2331

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EPOTHILONES, A NEW CLASS OF MT-STABILIZING AGENTS

Table 3 Comparison of effects of laxol and epothilones on milotic arrest andcytotoxicity in an MDR and parental cell line

The cells were treated as described in Table 2 and the EC50 values are shown. The R~

for curve fit is above 0.89 for each estimation.

CelltypeKB3-IKBV-ITreatmentTaxolEpothilone

AEpothiloneBTaxolEpothilone

AEpothiloneBMitotic

arrest6nMHhlM16

mi17(iMIVOnM92

nMCytotoxicity1.2nM13

nM15nM23

/IMKidnM58

nM

fold drop in potency for taxol (23 JXMEC50). This suggests thatepothilones are less effective substrates than taxol for the P-glycopro-tein. As expected, the EC5(, values for mitotic arrest by taxol, epothi-

lone A, and epothilone B (17 /AM,170 nM, and 92 nM, respectively)were very similar to the cytotoxicity values.

DISCUSSION

Taxol and taxotere, related complex taxanes, have emerged asimportant new cancer therapeutics. Much of the recent enthusiasm forthese agents is due to the fact that taxol has shown considerableefficacy against drug-refractory ovarian tumors (15) indicating that

the mechanism of action of taxol can override at least certain aspectsof acquired resistance to commonly used DNA-damaging antineoplas-

tic agents. In addition, taxol has shown activity against breast cancer,also in patients who have undergone prior chemotherapy (16). Sincethe majority of clinical trials to date have been reduced in scope dueto supply limitations and also have included a high percentage ofpatients who had undergone previous therapeutic regimens, there ishope that taxanes could have a broader application than the currentlyapproved indications for breast and ovarian cancer.

Extended use of taxol suffers from a number of limitations. Thefirst has been a limited taxol availability compounded by the inabilityto perform a large-scale synthetic reconstruction of the taxane ring.Fortunately, improved forestry of high-taxane yews and semisynthesis

of taxotere from yew needles partially address this issue. Secondly,the therapeutic index of taxol is narrow due to mechanism-based sideeffects of neutropenia, peripheral neuropathy, alopecia, and hypersen-

sitivity reactions (20). To date, neutropenia and neuropathy haveproved to be dose limiting in the clinic (35), although adjuvantcytokine therapy has been useful in reducing hematopoietic sideeffects. In addition, taxol is highly hydrophobic, necessitating deliveryin Cremophor which itself can induce cardiac arrhythmias and extensive hypersensitivity reactions (3). Together, these features have resulted in the recommendation that taxol be administered in the 170-mg/m2 dose range over a 24-h schedule to minimize hypersensitivity

reactions. At this dosage, taxol levels in plasma reach steady stateconcentrations of more than 1 X 10~7 M (36). One complication of

taxol therapy is that the plasma half-life is less than 5 h (37),

indicating a relatively narrow window of treatment (see below). Invitro studies have highlighted additional limitations in the mechanismof cytotoxicity of taxol. Taxol cytotoxicity correlates with mitoticarrest but, as with most chemotherapeutics, taxol is effective onlyagainst rapidly dividing cells (14). Our studies have indicated thateffective intracellular taxol concentrations can drop below thresholdvalues required for arrest at the G2-M transition 48-72 h after additionof taxol to the culture medium.4 Thus, cells entering mitosis after this

time are unaffected. Taken together, the in vitro data suggest that taxolwill be effective only against cells with rapid doubling times, and amore effective dosing schedule might utilize lower doses over longer

periods of time. However, the profile of taxol side effects mightpreclude such a regimen.

To address some of these issues, we sought to identify an alternative structural class of compounds which would mimic all the biological effects of taxol but with a structure more readily amenable tochemical manipulation. Using a novel, high throughput filtration-colorimetric assay for detecting MT-nucleating activity with a 10-fold

greater sensitivity than traditional methods, we identified an activecompound from an extract of the myxobacterium S. cellulosum.Electron microscope analysis confirmed that this extract promoted theassembly of bona fide microtubules. Purification by silica gel, gelfiltration, and reverse phase chromatography revealed two relatedactive components, epothilone A and B. Both epothilone A andepothilone B appeared to be equipotent to taxol in all of our in vitroassays. Most significantly, competition of [3H]taxol from preformed

MTs revealed that taxol, epothilone A, and epothilone B possesssimilar half-maximal displacement values. This observation together

with the fact that epothilones A and B had similar biological activityin our assays was consistent with the interpretation that epothilones Aand B compete for the same MT binding site as taxol.

Many agents identified as potent agonists or antagonists of in vitroactivity in drug screening can prove to have secondary activities thatlimit their usefulness in vivo. To ascertain whether epothilone A or Balso exactly mimics the activity of taxol in ceils, we compared theirability to induce MT bundling, to perturb mitosis, and to stabilize MTsagainst depolymerizing treatments. At 10~6-10~4 M, epothilones in

duced formation of extensive short MT bundles throughout the cytoplasm of interphase cells over the course of 2-3 h, as has been shown

for taxol. As noted by DeBrabander et al. (11) for taxol, these bundleswere clearly not linked to the centriole, suggesting that epothilones,like taxol, not only stabilize ¡ntracellular MTs but can also overridecentrosome nucleating activity at high concentrations. At lower concentrations (5 X 10"9-10"6 M), mitotic MT arrays were preferentially

affected. Cells entering M phase developed multipolar spindles in thepresence of either taxol or epothilones. Treatment with all three agentsresulted in early prophase arrest in human cells and aberrant mitosisleading to multimininucleation in rodent cells, as described previouslyfor taxol by Kung et al. (32). Finally, epothilones and taxol wereequally effective at stabilizing the cellular MT network when exposedto prolonged cold (4°C)treatment. These activities confirm that the

MT-stabilizing properties of epothilones and taxol affect cellular MT

form and function in an indistinguishable manner.Taxol-induced mitotic arrest has been shown to correlate with

cytotoxicity by an apoptotic pathway (33). We have analyzed thekinetics of cell death relative to G2-M block and have demonstratedthat DNA nicking is initiated rapidly after arrest at the G2-M transition.4 As shown in Fig. 9, epothilone- and taxol-induced cell death

follows G2-M block with a similar time lag. Agarose gel electrophore-

sis of DNA isolated from these cells 24 h after drug treatment revealedthe DNA ladder that is the hallmark of apoptosis. In situ TÚNELlabeling of DNA nicks in individual cells confirmed that this processoccurs selectively in G2-M blocked cells soon after mitotic arrest and,subsequently, in those multimininucleated G,-arrested cells that have

succeeded in surviving an aberrant mitosis, as had been seen withtaxol. From these observations, we conclude that epothilones exertequivalent biological effects to taxol on cultured cells. Since taxol hasbeen shown to bind MTs in cells (38), we conclude that epothiloneslikewise mediate all these downstream effects by their MT-stabilizing

activity observed in vitro and in cultured cells.Epothilones A and B were found to differ from taxol in one

important respect. Namely, the drop in potency exhibited by epothilones in MDR cells expressing P-glycoprotein was three orders of

magnitude less than the drop in potency exhibited by taxol. Since

2332

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EPOTHILONES. A NEW CLASS OF MT-STABILIZING AGENTS

MDR-mediated resistance accounts for some aspects of acquired taxolresistance (21), this observation suggests that epothilones as antineo-

plastic agents could provide an important advantage over taxol.Epothilones A and B have been patented recently as antifungal

agents with possible cytotoxic or immunosuppressive properties (22).The epothilones are structurally homologous to a family of 16-mem-

ber macrolide antibiotics including erythromycin. Interestingly, noneof these agents possessed the MT-stabilizing properties of the epothilones, although rhizoxin is a known MT-destabilizing drug (30). The

patent for epothilones A and B indicates that these compounds possesscytotoxic activity against a human T-24 bladder carcinoma line. Here

we describe the mechanism of action underlying epothilone cytotoxicactivity and reveal that epothilones represent an alternative structuralclass of MT-stabilizing agents to the taxanes. The epothilones furthermore appear to be competitive ligands for the taxol-binding site.

Both taxol and epothilones appear to bind tubulin aßheterodimerswithin MTs with a 1:1 stoichiometry. Whereas taxol EC50 values formitotic arrest and cytotoxicity of 5-20 nM have been observed here

and elsewhere for human cell lines, these values refer to extracellulardrug concentrations. However, the apparent differences in taxol sensitivity between human and rodent cell lines are caused by differencesin intracellular accumulation, with the actual intracellular drug concentrations effecting these changes in human cell lines reaching muchhigher concentrations (39). Such findings predict that a high degree ofoccupancy of the MT polymer may be required for mitotic perturbation. Given that both epothilones and taxol effect MT stabilizationthrough stoichiometric binding and that tubulin represents an abundant cellular protein, the biology of the system may preclude findinga drug which can improve on the stoichiometric potency of taxol forthis mechanism of action. The simpler chemical structure of epothilones may instead provide a useful lead compound in the quest for adrug operating by the same MT-stabilizing mechanism as taxol but

with an improved solubility profile and therapeutic index.

ACKNOWLEDGMENTS

We thank Phillip Hertzog for assistance with electron microscopy.

REFERENCES

1. Wani, M. C, Taylor, H. L., Wall, M. E., Coggon, P., and McPhail, A. T. Plan!antitumor agents. VI. The isolation and structure of taxol, a novel antileukemic andantitumor agent from Taxus brevi/olia. J. Am. Chem. Soc., 93: 2325-2327, 1971.

2. Schiff, P. B., Fant, J., and Horwitz, S. B. Promotion of microtubule assembly in vitroby taxol. Nature (Lond.), 277: 665-667, 1979.

3. Rowinsky, E. K., and Donehower, R. C. Taxol: twenty years later, the story unfolds.J. Nati. Cancer Inst.. 83: 1778-1781, 1991.

4. Dustin, P. Microtubules, Ed. 2, 482 pp. Berlin: Springer-Verlag. 1984.5. Kirschner. M., and Mitchison, T. Beyond self-assembly: from microtubules to mor

phogenesis. Cell, 45: 329-342, 1986.6. Brinkley. B. R. Microtubule organizing centers. Annu. Rev. Cell Biol., /: 145-172,

1985.7. De Brabander. M. A model for the microtubule organizing activity of the centrosomes

and kinetochores in mammalian cells. Cell Biol. Int. Rep., 6: 901-915, 1982.8. Parness, J., and Horwitz, S. B. Taxol binds to polymerized tubulin in vitro. J. Cell

Biol., 91: 479-487, 1981.9. Schiff. P. B., and Horwitz, S. B. Taxol assembles tubulin in the absence of exogenous

guanosine 5'-triphosphate or microtubule-associated proteins. Biochemistry, 20:

3247-3252, 1981.10. Hamel, E., del Campo, A. A., Lowe, M. C., and Lin, C. M. Interactions of taxol,

microtubule-associated proteins, and guanine nucleotidcs in tubulin polymerization.J. Biol. Chem., 256: 11887-11894, 1981.

11. DeBrabander, M., Geuens, G., Nuydens, R., Willebrords, R., and DeMey, J. Taxolinduces the assembly of free microtubules in living cells and blocks the organizingcapacity of the centrosomes and kinetochores. Proc. Nati. Acad. Sci. USA, 78:5608-5612, 1981.

12. Schiff, P. B., and Horwitz. S. B. Taxol stabilizes microtubules in mouse fibroblastcells. Proc. Nati. Acad. Sci. USA. 77: 1561-1565, 1980.

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

14. Liebmann. J., Cook, J. A., Lipschultz, C., Teague. D., Fisher, J., and Mitchell, J. B.The influence of Cremophor EL on the cell cycle effects of paclitaxcl (Taxol*) inhuman tumor cell lines. Cancer Chemother. Pharmacol., 33: 331-339, 1994.

15. McGuire, W. P., Rowinsky, E. K., Rosenshein, N. B., Grumbine, F. C., Ettinger,D. S., Armstrong, D. K., and Donehower. R. C. Taxol: a unique antineoplastic agentwith significant activity in advanced ovarian epithelial neoplasms. Ann. Intern. Med..///: 273-279, 1989.

16. Holmes, F. A., Walters, R. S., Theriault, R. L., Forman, A. D., Newton, L. K., Raber,M. N., Buzdar, A. U., Frye, D. K., and Hortobagyi, G. N. Phase 11trial of taxol, anactive drug in the treatment of metastatic breast cancer. J. Nail. Cancer Inst., 83:1797-1805, 1991.

17. Forastiere, A. A. Use of paclitaxel (TAXOL8) in squamous cell carcinoma of thehead and neck. Semin. Oncol., 20: 56-60, 1993.

18. Einzig, A. I., Höchster, H., Wiernik, P. H., Trump, D. L., Dutcher, J. P., Garowski,E., Sasloff, J., and Smith, T. J. A phase II study of taxol in patients with malignantmelanoma. Invest. New Drugs, 9: 59-64, 1991.

19. Ettinger, D. S. Overview of paclitaxel (TAXOL®)in advanced lung cancer. Semin.Oncol., 20: 46-49, 1993.

20. Rowinsky, E. K., Eisenhauer. E. A., Chaudhry, V., Arbuck, S. G., and Donehower,R. C. Clinical toxicities encountered with paclitaxel (TAXOL«). Semin. Oncol., 20:1-15, 1993.

21. Benber, R. A., Hamel. E., and Hande, K. R. Plant alkaloids. In: B. A. Chabner, andJ. M. Collins (eds.), Cancer Chemotherapy: Principles and Practice, pp. 253-274.Philadelphia: J. B. Lippincott Co., 1990.

22. Hoefle, G., Bedorf, N., Gerth, K., and Reichenback, H. Patent DE 4138042, 1993.23. Shen, D.. Cardarelli, C., Hwang, J.. Cornwell, M., Richer!. N., Ishii. S., Pastan, I., and

Gottesman, M. M. Multiple drug-resistant human KB carcinoma cells independentlyselected for high-level resistance to colchicine, Adriamycin, or vinblastine showchanges in expression of specific proteins. J. Biol. Chem., 261: 7762-7770, 1986.

24. Asnes, C. F., and Wilson. L. Isolation of bovine brain microtubule protein withoutglycerol: polymerization kinetics change during purification cycles. Anal. Biochem.,98: 64-73, 1979.

25. Schaffner, W., and Weissmann, C. A rapid, sensitive, and specific method for thedetermination of protein in dilute solution. Anal. Biochem.. 56: 502-514, 1973.

26. McEwen, B. and Edelstein. S. J. Evidence for a mixed lattice in microtubulesreassembled in vitro, i. Mol. Biol., 139: 123-145, 1980.

27. Zhu, J., Lazarides, E., and Woods, C. Identification of a novel Ca2+-regulated protein

that is associated with the marginal band and centrosomes of chicken erythrocytes.J. Cell Sci., 108: 685-698, 1995.

28. Sambrook, J., Fritsch. E. F., and Maniatis, T. Molecular Cloning. A LaboratoryManual, Ed. 2. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. 1989.

29. Tilly, J. L., Kowalski, K. I., Johnson, A. L., and Hsueh, A. J. W. Involvement ofapoptosis in ovarian follicular atresia and postovulatory regression. Endocrinology,129: 2799-2801. 1991.

30. Takahashi, M.. Iwasaki, S., Kobayashi. H., Okuda, S., Murai. T., and Sato, Y.Rhizoxin binding to tubulin at the maytansine-binding site. Biochim. Biophys. Acta,926: 215-223, 1987.

31. Fuchs, D. A., and Johnson, R. K. Cytological evidence that taxol. an antineoplasticagent from TÕÕ.VW.Vbrevifoltii. acts as a mitotic spindle poison. Cancer Treat. Rep., 62:1219-1222, 1978.

32. Kung, A. L.. Sherwood, S. W., and Schimke, R. T. Cell line-specific differences in thecontrol of cell cycle progression in the absence of mitosis. Proc. Nati. Acad. Sci.USA, 87: 9553-9557, 1990.

33. Bhalla, K.. Ibrado, A. M., Tourkina, E., Tang. C., Mahoney. M. E., and Huang. Y.Taxol induces ¡nternucleosomal DNA fragmentation associated with programmed celldeath in human mycloid leukemia cells. Leukemia (Baltimore), 7: 563-568, 1993.

34. Willingham, M. C., and Bhalla, K. Transient mitotic phase localization of Bcl-2oncoprotein in human carcinoma cells and its possible role in prevention of apoptosis.J. Histochem. Cytochem., 42: 441-450, 1994.

35. Chaudhry, V., Rowinsky, E. K.. Sartorius, S. E., Donehower, R. C., and Cornblath,D. R. Peripheral neuropathy from taxol and cisplatin combination chemotherapy:clinical and electrophysiological studies. Ann. Neurol., 35: 304-311, 1994.

36. Berg, S. L., Cowan, K. H., Balis, F. M., Fisherman, J. S., Denicoff, A. M., Hillig. M..Poplack, D. G., and O'Shaughnessy. J. A. Pharmacokinetics of taxol and doxorubicin

administered alone and in combination by continuous 72-hour infusion. J. Nati.Cancer Inst., «6:143-145, 1994.

37. Rowinsky, E. K., Cazenave, L. A., and Donehower, R. C. Taxol: a novel investiga-tional antimicrotubule drug. J. Nati. Cancer Inst., 82: 1247-1259, 1990.

38. Manfredi, J. J., Parness, J., and Horwitz. S. B. Taxol binds to cellular microtubules.J. Cell Biol.. 94: 688-696. 1982.

39. Lopes, N. M., Adams, E. G.. Pitts, T. W., and Bhuyan, B. K. Cell kill kinetics and cellcycle effects of taxol on human and hamster ovarian cell lines. Cancer Chemother.Pharmacol., 32: 235-242, 1993.

2333

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1995;55:2325-2333. Cancer Res   Daniel M. Bollag, Patricia A. McQueney, Jian Zhu, et al.   a Taxol-like Mechanism of ActionEpothilones, a New Class of Microtubule-stabilizing Agents with

  Updated version

  http://cancerres.aacrjournals.org/content/55/11/2325

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment 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/55/11/2325To request permission to re-use all or part of this article, use this link

on July 7, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from