Suppression of Microtubule Dynamics by Discodermolide by a Novel Mechanism Is Associated with Mitotic Arrest and Inhibition of Tumor Cell Proliferation

Molecular Cancer Therapeutics 1303

Suppression of microtubule dynamics by discodermolide by a novel mechanism is associated with mitotic arrest and inhibition of tumor cell proliferation

Ste´phane Honore,1 Kathy Kamath,2 Diane Braguer,1 tubule dynamics as compared with those of paclitaxel that Leslie Wilson,2 Claudette Briand,1 and may in turn reflect differences in their binding sites and Mary Ann Jordan2 their effects on tubulin conformation. (Mol Cancer Ther. 2003;2(12):1303–1311) 1 UMR-CNRS 6032, Universite´ de la Me´diterrane´e, Marseille, France and 2 Department of Molecular, Cellular and Developmen- tal Biology and Neuroscience Research Institute, University of Introduction California, Santa Barbara, CA Microtubule-targeted drugs, including the taxanes and Vinca alkaloids, are powerful chemotherapeutic agents Abstract used for treatment of many types of human tumors. However, the problem of the development of clinical Discodermolide is a new microtubule-targeted drug in resistance to taxanes and Vinca alkaloids, as well as the Phase I clinical trials that inhibits tumor growth and general clinical success of microtubules as a target for induces G2-M cell cycle arrest. It is effective against anticancer drugs, have highlighted the need for new paclitaxel-resistant cell lines and acts synergistically in microtubule-targeted chemotherapeutic drugs. Discoder- combination with paclitaxel. Suppression of microtubule molide (Fig. 1) is a new and structurally unique dynamics by microtubule-targeted drugs has been hypoth- compound isolated from a marine sponge, Discodermia esized to be responsible for their ability to inhibit mitotic dissoluta. It was first described as an immunosuppressive progression and cell proliferation. To determine whether agent (1) but more recently, it has been found to enhance discodermolide blocks mitosis by an effect on microtubule the nucleation of microtubules and their stability to dynamics, we analyzed the effects of discodermolide depolymerization by cold temperature or Ca2+ more on microtubule dynamics in living A549 human lung potently than paclitaxel (2). The drug is particularly cancer cells during interphase at concentrations that interesting because it is effective against paclitaxel- block mitosis and inhibit cell proliferation. We found resistant cell lines and acts synergistically in combination that discodermolide (7–166 nM) significantly suppressed with paclitaxel (2, 3). In the hollow fiber mouse model, microtubule dynamic instability. At the IC50 for prolifera- discodermolide produced significant inhibition at tolerat- tion (7 nM discodermolide, 72 h), overall dynamicity was ed doses against a panel of human tumor cells, including reduced by 23%. The principal parameters of dynamic paclitaxel-resistant tumor cells. In human tumor xenograft instability suppressed by discodermolide were the micro- models, it is efficacious at tolerated doses against non- tubule shortening rate and length shortened. In addition, small cell lung, breast, and colon tumor models, some of discodermolide markedly increased the frequency of which are refractory to paclitaxel. Growth inhibitory rescued catastrophes. At the discodermolide concentra- activity is dose dependent and sustained for 5 weeks (or tion that resulted in 50% of maximal mitotic block (83 nM, more) after administration of a single i.v. bolus dose of 20 h), most microtubules were completely non-dynamic, 15 mg/kg (4, 5). It is currently in Phase I clinical trials (6). no anaphases occurred, and all spindles were abnormal. Discodermolide induces microtubule bundling at high The dynamicity of the remaining dynamic microtubules concentrations, cell cycle arrest in the G -M phase, was reduced by 62%. The results indicate that a principal 2 aneuploidy, and apoptosis (7–9). It competitively and very mechanism of inhibition of cell proliferation and mitotic strongly inhibits the binding of paclitaxel to microtubules, block by discodermolide is suppression of microtubule suggesting that the two drugs may have common or dynamics. Importantly, the results indicate significant overlapping binding sites (2, 10), although it remains additional stabilizing effects of discodermolide on micro- possible that discodermolide binds to more than one site on microtubules. The finding that discodermolide cannot substitute for paclitaxel in paclitaxel-requiring cells, where- Received 8/5/03; revised 9/26/03; accepted 10/7/03. as another microtubule-stabilizing drug, epothilone B, can The costs of publication of this article were defrayed in part by the substitute for paclitaxel (3), suggests that there are payment of page charges. This article must therefore be hereby significant differences between the mechanisms of disco- marked advertisement in accordance with 18 U.S.C. Section 1734 dermolide, paclitaxel, and epothilones. solely to indicate this fact. Grant support: GEFLUC Marseille Provence, by the Association pour la Inhibition of microtubule dynamics is a fundamental Recherche sur le Cancer (ARC), from whom S. H. received a fellowship, mechanism of action of several microtubule-targeted and by NIH CA57291 and NS13560. drugs (11, 12). One form of microtubule dynamics, called Requests for Reprints: Mary Ann Jordan, Department of Molecular, dynamic instability, is a behavior in which the ends of Cellular and Developmental Biology, University of California, Santa Barbara, CA 93106. Phone: (805) 893-5317; Fax: (805) 893-4724. microtubules undergo frequent stochastic transitions E-mail: [email protected] between episodes of growing and shortening. Dynamic

96-well microtiter plates and allowed to attach for 24 h. After incubation with the indicated drug concentrations for 72 h, cell proliferation was determined using the sulforhodamine B assay (19). Fluorescence Microscopy After growth for 24 h in Labtek II chambers (Nalge Nunc, Rochester, NY), A549 cells were incubated with discoder- Figure 1. Structure of (+)-discodermolide. molide at the indicated concentrations for 20 h. Immuno- fluorescence microscopy of microtubules and DNA staining were performed as previously described (20). instability appears to be essential for progression through Briefly, cells were fixed with 3.7% formaldehyde in PBS mitosis. Suppression of microtubule dynamic instability and stained with an anti-tubulin antibody, tetramethyl by microtubule-targeted drugs is associated with aberrant rhodamine isothiocyanate (TRITC)-conjugated secondary organization of mitotic spindles and inhibition of progress antibody and 4V,6-diamidino-2-phenylindole (DAPI) (Sig- into anaphase (13–17). We recently found that microtu- ma, St. Louis, MO). bule dynamics must be finely regulated for progress Preparation and Microinjection of Rhodamine- through mitosis, and that both excessively rapid dynamics Labeled Tubulin as well as suppressed dynamics correlate with impaired mitotic spindle function and inhibition of cell proliferation Rhodamine-labeled tubulin was prepared using carbox- (18). Information on the effects of discodermolide on ymethylrhodamine (Molecular Probes, Eugene, OR) labeling microtubule dynamics in vitro or in cells is lacking. of microtubules assembled from phosphocellulose-purified Whether, and if so, how, discodermolide blocks mitosis bovine brain tubulin (18). Labeled tubulin was stored in 6-Al aliquots at À70jC until use. A549 cells were seeded and inhibits cell proliferation by an effect on microtubule 5 dynamics has not been determined. 48 h before use at 0.5 Â 10 cells/ml on glass coverslips A A Inthepresentwork,weanalyzedtheeffectsof coated with 10 g/ml laminin and 20 g/ml fibronectin discodermolide on the dynamic instability of individual (Life Technologies, Inc., Carlsbad, CA). Rhodamine-tubu- microtubules in living human A549 lung cancer cells. We lin (5 mg/ml) was centrifuged in a Beckman (Palo Alto, found that discodermolide strongly suppresses microtu- CA) tabletop ultracentrifuge using a TLA 100.3 rotor at Â j bule dynamic instability in a concentration-dependent 38,000 g for 14 min at 4 C to remove aggregates just manner. At the concentration of discodermolide that before microinjecting into A549 cells using a Transjector 5246 and Injectman (Eppendorf AG, Hamburg, Germany). inhibits cell proliferation by 50% at 72 h (7 nM), overall j dynamicity was reduced by 23%. The principal parameters Injected cells were incubated for 2 h at 37 C to allow for of dynamic instability that were altered include a decrease incorporation of rhodamine-tubulin into microtubules and in the mean shortening rate and length together with an further incubated with discodermolide for 4 h to allow increase in the rescue frequency. When the concentration of attainment of an equilibrium intracellular drug concen- discodermolide was increased to the concentration that tration. Time-Lapse Microscopy and Image Acquisition blocked mitosis by 50% (83 nM), dynamicity was reduced by 62%, and most of the dynamic instability parameters Microinjected cells were placed in RPMI culture medium were significantly affected. There are significant differences lacking sodium bicarbonate and supplemented with 25 mM between the effects of discodermolide, epothilone B, and HEPES, 4.5 g/l glucose, and 30 Al Oxyrase/ml (Oxyrase paclitaxel on microtubule dynamics that may reflect Inc., Mansfield, OH) to reduce photodamage, in a double differences in their binding sites and effects on tubulin coverslip chamber, and observed using a fluorescence conformation in microtubules. The results strongly suggest microscope (Nikon Eclipse E800) maintained at 37 F 1jC, that a principal mechanism of inhibition of cell prolifera- with a Nikon plan apochromat 1.4 N.A., 100Â objective tion and mitotic block by discodermolide is suppression of lens. Thirty to 41 images per cell were acquired at 3-s microtubule dynamics. intervals using a Hamamatsu Orca II digital camera (Middlesex, NJ) driven by Metamorph software (Universal Materials and Methods Imaging, Media, PA). Analysis of Microtubule Dynamics Materials The positions of the plus ends of individual micro- Discodermolide was obtained as a generous gift from F. tubules with time were recorded using Metamorph Kinder (Novartis Pharmaceuticals Corp., East Hanover, software, exported to Microsoft Excel, and analyzed by j NJ). It was dissolved in sterile DMSO and stored at À20 C. using Real Time Measurement software (21). The lengths Cell Culture and Proliferation of individual microtubules were graphed as a function of Human non-small cell lung carcinoma cells, A549, were time (life histories). Changes in length z0.5 Am were maintained in RPMI 1640 medium containing 10% fetal considered growth or shortening events. Changes in bovine serum. Cells were grown as monolayer cultures in length <0.5 Am were considered phases of attenuated 5–7% CO2 and passaged at intervals of 5 days. A549 cells dynamics or pause. The rates of growth and shortening were seeded at a density of 1 Â 104 cells/ml (200 Al/well) in events were determined by linear regression. Means and

SEs were calculated per event. The time-based catastrophe at low nanomolar concentrations. The drug inhibited frequency was calculated by dividing the number of proliferation half-maximally at a concentration of 7 nM, transitions from growth or pause to shortening by the whereas at concentrations z30 nM discodermolide, signif- total time growing and paused for each individual icant cell killing occurred (Fig. 2). microtubule. The length-based catastrophe frequency Suppression of Microtubule Dynamic Instability by was calculated by dividing the number of transitions DiscodermolideinLivingCells from growth or pause to shortening by the total distance To determine the effects of discodermolide on the grown for each individual microtubule. The rescue dynamic instability behavior of microtubules in living frequencies based on distance or time were calculated A549 cells, the cells were injected with rhodamine-tubulin, similarly, dividing the total number of transitions from and the labeled tubulin was allowed to incorporate into shortening to pause or growth by the length shortened or microtubules for 2 h before adding the drug. Cells were the time spent shortening, respectively, for each individ- observed by time-lapse fluorescence microscopy and ual microtubule. Means and SEs of transition frequencies images were collected (‘‘Materials and Methods’’). were calculated per microtubule (n = 30 for each experi- Fig. 3A shows a gallery of images of the lamellar region mental condition). Dynamicity is the total length grown of a living, untreated A549 cell (control, top row) and a cell and shortened divided by the life span of the population incubated with 83 nM discodermolide (4 h) (bottom row). of microtubules (22). In controls, the plus ends of microtubules (marked by arrowheads) alternated between prolonged phases of slow Results growing, rapid shortening, and a pause state (a state of attenuated dynamic instability). In contrast, 83 nM Mitotic Arrest and Inhibition of Cell Proliferation and discodermolide markedly suppressed microtubule dynam- by Discodermolide ic instability in cells as indicated by the unaltered We determined the concentration dependence for the positions of their plus ends during a 40-s time period effects of discodermolide on the mitotic index and on cell (Fig. 3A, bottom panel). Fig. 3B shows several life history proliferation in A549 cells (‘‘Materials and Methods’’). As traces of the changes in length of individual microtubules determined by counting cells in mitosis and interphase in the presence or absence of 83 nM discodermolide. The after fixation and staining for microscopy, discodermolide life history traces of control microtubules in Fig. 3B (top induced the accumulation of cells in mitosis after one cell panel) show extensive length changes in the microtubules, cycle (20 h) half-maximally at a concentration of 83 nM, and whereas the life history traces are nearly flat, indicating maximal mitotic accumulation (65%) occurred at concen- marked suppression of dynamics in the presence of 83 nM trations of 166 nM and above (Fig. 2). Proliferation was discodermolide (Fig. 3B, bottom row). The life history plots determined by the sulforhodamine B assay after incubation from 30 individual microtubules were used to determine with discodermolide for 72 h. Consistent with previous the parameters of dynamic instability (Tables 1 and 2). In reports (3), discodermolide inhibited A549 cell proliferation control cells, the plus ends of microtubules grew at a mean rate of 13.8 F 1.4 Am/min, slower than their mean shortening rate of 24.0 F 2.1 Am/min (P < 0.01). The mean lengths grown and shortened were 2.5 F 0.2 and 5.0 F 0.5 Am, respectively. Microtubules spent half of their time (52%) in a paused state, neither growing nor shortening to a detectable extent. Discodermolide (7–166 nM, 4 h) suppressed microtubule dynamics in a concentration-dependent manner. The effects are shown at the IC50’s for cell proliferation (7 nM) and for mitotic block (83 nM) in Tables 1 and 2, and they are presented graphically for the entire concentration range in Fig. 4. At the relatively low concentration of 7 nM, several dynamic instability parameters were significantly altered. Notably, the shortening rate and length were decreased by 21% and 42%, respectively, and the rescue frequency calculated on the basis of length was increased by 93%. The dynamicity, a parameter reflecting the overall exchange of tubulin with the microtubule end, was reduced by 23%. Figure 2. Effect of discodermolide on A549 cell proliferation at 72 h (., left-hand axis) and induction of mitotic arrest at 20 h (o, right-hand At the IC50 for mitotic accumulation (83 nM), signifi- axis). Cell number was determined at the time of discodermolide addition cantly fewer microtubules were dynamic as compared and 72 h later and mitotic arrest was determined after one cell cycle by with control cells and cells incubated with 7 nM counting 1000 cells after staining with 4,6-diamidino-2-phenylindole (‘‘Materials and Methods’’). Points, means of three independent experi- discodermolide. Counts of stable and dynamic micro- ments; bars, SD. tubules in three cells at each drug concentration during a

Figure 3. A, time-lapse sequence of video frames, showing the plus ends of several microtubules in a control cell (top row) and a cell treated with 83 nM discodermolide (bottom row). In the control cell, arrowheads indicate two microtubules that undergo shortening events. In the discodermolide-treated cell, the arrowhead indicates one microtubule end for which the position did not change over a period of 40 s. Time is indicated in seconds. B, life history plots of the changes in length of microtubules in living A549 cells in the absence of drug (Control) or after discodermolide incubation (83 nM) for 4 h. Graphs, lengths of excursions made by individual microtubules over time. Microtubules in the presence of discodermolide are characterized by longer pause events and less extensive growing and shortening events than those in control cells.

2-min recording period indicated that 31 F 6% of the result of the exclusion of completely stabilized micro- microtubules were stable in controls, 44 F 7% were stable tubules from determination of the parameters, these at 7 nM discodermolide, and 77 F 5% were stable at 83 nM values represent an underestimate of the suppressive discodermolide. To determine the mechanistic effects of effects of discodermolide. the drug on the individual parameters of dynamic Can Cells Complete Anaphase with Impaired instability, only dynamic microtubules were included in Dynamics? the measurements. All of the parameters affected by 7 nM To understand the effects of suppression of microtu- discodermolide (shortening rates, lengths, and length- bule dynamics on mitosis, we examined the arrange- based rescue frequency, as well as dynamicity, discussed ments of microtubules and chromosomes in A549 cells above) were more strongly altered by 83 nM discodermo- incubated with the concentrations of discodermolide that lide (À30%, À54%, +172%, and À62%, respectively). In suppressed microtubule dynamics (Figs. 5 and 6). In addition, the growth rate and length were decreased by controls, the mitotic index was 5 F 1.2%, and most 30% and 34%, respectively; the time-based catastrophe spindles were normal bipolar spindles. For example, frequency was decreased by 42%; and the time-based Fig. 5, A and B, shows a control metaphase spindle; it is rescue frequency was increased by 88%. All of these bipolar and the chromosomes have all congressed to the changes indicate extreme stabilization of microtubule metaphase plate. The ratio of cells in anaphase to cells in dynamics at the discodermolide concentration that metaphase is a reflection of the times required to pass resulted in half-maximal mitotic block at 20 h and also through metaphase and through anaphase; this ratio was resulted in significant cell killing at 72 h (Fig. 2). As a 0.55 in control cells.

Table 1. Discodermolide suppresses microtubule dynamics at concentrations that inhibit A549 cell proliferation and induce mitotic arrest

Parameters Control 7 nM Change 83 nM Change

Mean rates (Am/min F SE) Growth 13.8 F 1.4 12.8 F 1.0 9.6 F 0.8** À30% Shortening 24.0 F 2.1 19.0 F 1.8* À21% 16.98 F 2.34** À30% Mean duration (min F SE) Growth 0.18 F 0.02 0.19 F 0.01 0.17 F 0.01 Shortening 0.21 F 0.02 0.15 F 0.01 0.14 F 0.02 Pause 0.34 F 0.06 0.29 F 0.03 0.40 F 0.05 Mean length (Am F SE) Growth 2.5 F 0.2 2.5 F 0.2 1.6 F 0.1*** À34% Shortening 5.0 F 0.5 2.9 F 0.2*** À42% 2.3 F 0.2*** À54% % Time spent Growth 20 26 18 Shortening 28 20 11 Pause 52 54 71 Dynamicity (Am/min) 9.45 7.23 À23% 3.57 À62%

Note: ***, P < 0.001; **, P < 0.01; *, P < 0.05, estimates of significance of difference from controls by Student’s t test.

Table 2. Discodermolide increases the rescue frequency and suppresses the catastrophe frequency

Parameter Control 7 nM Change 83 nM Change

Catastrophe frequency (minÀ1 F SE) 1.68 F 0.12 1.67 F 0.09 0.97 F 0.06*** À42% Rescue frequency (minÀ1 F SE) 3.38 F 0.30 4.62 F 0.32 6.36 F 0.30*** +88% Catastrophe frequency (AmÀ1 F SE) 0.48 F 0.03 0.44 F 0.02 0.53 F 0.04 Rescue frequency (AmÀ1 F SE) 0.14 F 0.01 0.27 F 0.01** +93% 0.38 F 0.01*** +172%

Note: ***, P < 0.001; **, P < 0.01, estimates of difference from controls by Student’s t test.

At 7 nM discodermolide (a concentration that sup- (Figs. 3 and 4; Tables 1 and 2), many small compact pressed dynamics to a small but significant degree and bundles of microtubules were formed throughout the cells also was the IC50 for cell proliferation), there was a slight and the cells were no longer flat, but were rounded (Fig. 7, but significant increase in the mitotic index (to 9 F 2.9% E and F). from 5 F 1.2% in controls, P < 0.05). The anaphase/ metaphase ratio decreased from 0.55 to 0.09, indicating Discussion that the metaphase/anaphase transition was strongly slowed at 20 h of drug incubation. Thus, many cells We have found that discodermolide, a novel microtubule- appeared able to exit mitosis without proceeding through targeted drug in Phase I clinical trials, acts to block anaphase. At this concentration, significant numbers of mitosis and inhibit cell proliferation in concert with the spindles were abnormal in appearance. We deter- suppression of microtubule dynamics in living human mined the frequencies of abnormal spindles using the tumor cells. At low concentrations (7 nM), discodermolide classification previously developed by Jordan et al. (13, 14, primarily decreased the rate and extent of shortening and 16). In controls, nearly all spindles were normal. With 7 nM discodermolide, 18% were normal metaphase spindles; 36% were abnormal spindle types I and II, with from a few to many uncongressed chromosomes that remained at the spindle poles (Fig. 5, C, D and E, F, Fig. 6); 16% had many chromosomes that were arranged in a ball about the spindle (Type III) (Fig. 5, I and J, Fig. 6); and 20% were multipolar (Fig. 5, G and H, Fig. 6). Thus, although mitotic block was not prevalent, it appears that spindle morphol- ogy was distinctly abnormal which may have resulted in aberrant cytokinesis and either cell death or inability of cells to continue proliferating (8), thus leading to a decrease in cell number when examined at 72 h of drug incubation.

At 83 nM discodermolide, the IC50 for mitotic accumulation (20 h), no cells progressed to anaphase and severely abnormal type III ball-shaped spindles and asymmetric bipolar spindles (37% and 43% of all spindles, respectively) predominated (Fig. 5, I, J and K, L, Fig. 6). The asymmetric bipolar spindles were similar to those observed in kangaroo rat (PtK2) cells treated with an extremely high concentration of discodermolide (10 AM) (2). One half-spindle was bigger than the other one, and chromosomes accumulated to the side of the spindle. In some cases, the spindle poles appeared to be unconnected and independent of each other. Microtubules in control cells in interphase formed a fine meshwork emanating from the centrosome. At discodermolide concentrations ranging from 7 to 21 nM, there were Figure 4. Concentration dependence for the effects of discodermolide no detectable changes in the microtubule network on the shortening and growing rates (A); the lengths shortened and grown organization (Fig. 7, B and C). At 41.5 nM discodermolide, (B); the percentages of time growing, shortening, and paused (C); the microtubule bundles occurred around the nucleus but time-based catastrophe and rescue frequencies (D); the length-based o were absent from the cell periphery (Fig. 7D). At high catastrophe and rescue frequencies (E); and the overall dynamicity (F). , growing; n, shortening; E, pause; *, rescue; x, catastrophe. Points, concentrations (i.e., 83–166 nM discodermolide) at which means [except for overall dynamicity (mean only)] of 30 microtubules for microtubule dynamics were nearly completely suppressed each condition obtained from three independent experiments; bars, SE.

Figure 5. Mitotic spindle organization in A549 cells incubated F 41.5 nM discodermolide (20 h). Spindle microtubules are shown by anti-a-tubulin immunofluorescence (A, C, E, G, I, K) and chromosomes are shown by 4,6- diamidino-2-phenylindole stain (B, D, F, H, J, L) (‘‘Materials and Methods’’). A and B, controls; C – L, 41.5 nM discodermolide. C and D, type I spindle with a few uncongressed chromosomes; E and F, type II spindles with many uncongressed chromosomes remaining at one or both spindle poles; G and H, tripolar spindles; I and J, type III or ball-shaped spindle. Asymmetric spindles (K and L) with chromosomes located predomi- nantly at one end of the spindle or on both sides of the spindle are indicated with arrowheads and arrows, respectively. Scale bar, 10 Am.

increased the distance-based rescue frequency leading to difference between the effects of discodermolide, paclitaxel, 23% suppression of the overall microtubule dynamicity. and epothilone B on microtubule dynamics. At the IC50 for These effects occurred in the presence of significantly mitotic block (83 nM), discodermolide had no significant aberrant mitoses, but in the absence of detectable effect on the catastrophe frequency per length grown. In microtubule bundling. At this concentration, discodermo- contrast, in the same cell line, paclitaxel at the IC50 for lide induced a 1.8-fold increase in the mitotic index and mitotic block vastly increased the catastrophe frequency decreased the ratio of cells in anaphase to cells in per length grown (+80%).1 Whatthismeansisthat metaphase by 82%, indicating that the transition from discodermolide has an additional stabilizing effect on metaphase to anaphase was impeded. The slowed mitosis microtubule dynamics as compared with paclitaxel; micro- and the suppressed microtubule dynamics were associat- tubules in the presence of paclitaxel continued to undergo ed with a 50% inhibition of cell proliferation after 72 h significant, even enhanced, catastrophes. In agreement with drug incubation. this distinction, in another human tumor cell line (MCF7), Higher concentrations of discodermolide (83 nM) further paclitaxel and epothilone B at the IC50s for mitotic block suppressed dynamic instability parameters including also increased the length-based catastrophe frequency (by significant reduction in the growing rates and lengths 38% at 7.5 nM paclitaxel and by 78% for 3.5 nM epothilone (by 30% and 34%, respectively), a significant reduction in B) (24). Thus, there are significant differences between the the time-based catastrophe frequency (by 42%), and mechanisms of discodermolide, on the one hand, and significant increases in the time-based and length-based epothilone B and paclitaxel, on the other hand, which may rescue frequencies (by 88% and 172%. respectively). relate to differences in their binding (10), in the number of Microtubule dynamicity was decreased by 62%. binding sites on microtubules, and in their effects on the Discodermolide Suppresses Microtubule Dynamics resulting conformation of tubulin in the microtubule. These Differently from Paclitaxel and Epothilone B Many of the observed changes resemble the changes in microtubule dynamics induced by paclitaxel and epothilone B (13, 17, 23, 24). However, there is an interesting 1 S. Honore, K. Kamath, D. Braguer, L. Wilson, and M. A. Jordan, unpublished data.

Figure 6. Frequency of spindle abnormalities induced in A549 cells by discodermolide at concentrations ranging from 7 to 166 nM (20 h). Metaphase, anaphase, and telophase con- sisted of normal bipolar spindles; types I and II were bipolar spindles with one or more uncongressed chromosomes; type III were monopolar spindles with chromosomes arranged in a ball around the spindle; multipolar spindles were spindles with more than two poles; asymmetric spindles were bipolar spindles, described in Fig. 5. Columns, means of three independent experiments and evaluation of at least 600 mitotic cells per condition; bars, SD.

differences may be responsible for the inability of dis- lide). Fifty percent of all metaphase spindles were codermolide to substitute for paclitaxel in paclitaxel- abnormal (Types I–III), indicating that chromosome requiring cells, for its ability to act in synergy when congression was impaired. There are several mechanisms combined with paclitaxel, and to its effectiveness in by which mitotic slowing could result from suppression of paclitaxel-resistant cells lines (2, 3). In addition, these microtubule dynamics by discodermolide. During pro- results are consistent with the observation that discoder- metaphase, the plus ends of dynamic spindle micro- molide enhances nucleation of microtubules from purified tubules appear to probe the cytoplasm until linkage with a brain tubulin more potently than paclitaxel (2). This chromosomal kinetochore is established. Reduced micro- phenomenon may reflect discodermolide’s ability to en- tubule lengths shortened in cells treated with discoder- hance the rescue frequency of shortening polymers, or, in molide may result in impaired probing of the intracellular other words, to inhibit the off-reaction of tubulin from space by microtubules and to the attachment of a less than small nucleating tubulin polymers. normal number of microtubules to the kinetochores and/ Discodermolide also induced the formation of an incom- or a delayed attachment to the kinetochores, thus delaying plete or abnormal metaphase plate of chromosomes and an chromosome congression to the metaphase plate. In abnormal arrangement of spindle microtubules. At high addition, dynamic microtubules appear to be required to concentrations (> 83 nM), the spindles became asymmetric. produce the tension required for the metaphase/anaphase Interestingly, the asymmetric bipolar spindles were similar to those observed in kangaroo rat (PtK2) cells treated with an extremely high concentration of discodermolide (10 AM) (2). These asymmetric spindles have not been reported with other microtubule-stabilizing drugs, and may result from differences in the antimitotic mechanism of discodermolide as compared with paclitaxel or epothilones. The mitotic index was increased 6.6-fold (to 33 F 4.1% at 83 nM discodermolide) and no cells progressed to anaphase, suggesting a clear mitotic arrest at the metaphase/anaphase boundary. These effects were associated with complete inhibition of cell proliferation at 72 h. To What Degree Is Suppression of Microtubule Dynamics by Discodermolide Responsible for Its Ability to Inhibit Cell Proliferation? Figure 7. Microtubules of interphase A549 cells incubated with discodermolide (20 h). (A), control, irregular meshwork of microtubules. At the IC50 for inhibition of cell proliferation (7 nM), we At 7 nM discodermolide (B) and 21 nM discodermolide (C), microtubules observed a 23% reduction in overall microtubule dynam- were similar to those in control cells. At 41.5 nM discodermolide (D), icity, a slight increase in the mitotic index (from 5% to 9%) microtubule bundles occurred in the center of the cell around the nucleus and were absent from the periphery. At 83 nM discodermolide (E)and166nM and a significant decrease in the anaphase/metaphase discodermolide (F), compact bundles of microtubules formed throughout ratio (from 0.55 in controls to 0.09 in 7 nM discodermo- the cells and the cells were rounded. Scale bar, 10 Am.

transition (25–27), and for chromosome separation. Such tumor actions of the microtubule-targeted drugs appear to slowing of mitotic progress may be in part responsible for involve their suppression of microtubule dynamics to levels the inhibition of cell proliferation at 72 h. We also below those necessary to accurately form a mitotic spindle observed that 20% of spindles were multipolar at 7 nM and to transit from metaphase to anaphase. discodermolide, as compared with 2.5% in controls. It is In summary, we have shown that discodermolide conceivable that suppression of microtubule dynamics suppresses microtubule dynamics in living cells in a may be involved in the formation of multipolar spindles, concentration-dependent manner. Slight to moderate sup- but other mechanisms for their formation can also be pression of microtubule dynamics is associated with mitotic envisioned. Interference with the function of centrosomes slowing or block and inhibition of cell proliferation, and a may contribute to aberrant mitosis and aneuploidy (8). greater suppression of microtubule dynamics is required to Our results suggest that suppression of microtubule completely block mitosis. Our results confirm the impor- dynamics is one of the principal mechanisms for tance of the regulation of microtubule dynamics for normal inhibition of cell proliferation by low concentrations of mitosis (18). Altogether, the results indicate that a major discodermolide while other mechanisms may coexist. mechanism of mitotic block and inhibition of cell prolifer- We have measured microtubule dynamics in cells in ation by discodermolide is the suppression of microtubule interphase at concentrations of discodermolide that block dynamics by mechanisms that differ from those of mitosis. The question arises as to whether these measure- paclitaxel. Discodermolide has effects on some dynamic ments accurately reflect the drug effects on microtubules in instability parameters that are different from those of other the mitotic spindle. While this question has not yet been microtubule-stabilizing drugs. These effects may be in- directly addressed experimentally, studies of the effects of volved in the synergism observed between discodermolide taxol and Vinca alkaloids on the dynamics of centromeres, and paclitaxel (3), a hypothesis that is currently under kinetochores, and their attached microtubules in mitotic investigation. In addition, these effects may be relevant to spindles of human osteosarcoma cells indicate that mitotic the clinical effectiveness of discodermolide, particularly in block is closely associated with suppression of centromere taxane-resistant cells. dynamics (26, 27). Likewise, examination of suppression of interphase microtubule dynamics in cell cultures containing high percentages of mitotically blocked cells indicates that Acknowledgment for Vinca alkaloids, a halichondrin analog, taxol, and We thank Herb Miller for excellent technical assistance. epothilone B, suppression of interphase microtubule dynamics occurs at the same drug concentrations that block References mitosis (15, 17, 24).2 Thus, although dynamics of mitotic 1. Longley, R. E., Gunasekera, S. P., Faherty, D., McLane, J., and spindle microtubules are faster than those of interphase Dumont, F. Immunosuppression by discodermolide. Ann. NY Acad. Sci., microtubules, the drugs appear to suppress microtubule 696: 94 – 107, 1993. dynamics relatively similarly, perhaps proportionally, re- 2. Kowalski, R. J., Giannakakou, P., Gunasekera, S. P., Longley, R. E., gardless of the phase of the cell cycle. Day, B. W., and Hamel, E. The microtubule-stabilizing agent discodermolide competitively inhibits the bindings of paclitaxel (Taxol) to tubulin Dynamics parameters and overall dynamicity vary from polymers, enhances tubulin nucleation reactions more potently than cell type to cell type. Among the four human cancer cell paclitaxel, and inhibits the growth of paclitaxel-resistant cells. Mol. Pharmacol., 52: 613 – 622, 1997. types examined to date, measurements of overall dynam- 3. Martello, L. A., McDaid, H. M., Regl, D. L., Yang, C-P. H., Meng, D., icity in interphase have ranged from a low of 3 Am/min in Pettus, T. R. R., Kaufman, M. D., Arimoto, H., Danishefsky, S. J., Smith, A498 kidney cells and CaOv-3 ovarian cells to a high of 10.4 A. B. I., and Horwitz, S. B. Taxol and discodermolide represent a F 4.7 Am/min in MCF7 breast cells (17, 24). Values obtained synergistic drug combination in human carcinoma cell lines. Clin. Cancer Res., 6: 1978 – 1987, 2000. for microtubule dynamicity in non-human cell lines vary 4. Kinder, F. R., Jr., Bair, K. W., Chen, W., Florence, G., Francavilla, C., more widely, to both higher and lower values (28, 29). Geng, P., Gunasekera, S., Guo, Q., Lassota, P. T., Longley, R. E., Palermo, Microtubule polymerization and dynamics are regulated in M. G., Paterson, I., Pomponi, S., Ramsey, T. M., Rogers, L., Sabio, M., Sereinig, N., Sorensen, E., Wang, R. M., and Wright, A. Synthesis and cells by a number of factors including the tubulin isotype antitumor activity of analogues of the novel microtubule stabilizing agent composition, tubulin posttranslational modifications, and a discodermolide. Abstracts of Papers American Chemical Society, 224th large number of microtubule regulatory proteins.3 Each cell ACS National Meeting, Boston, MA, United States: MEDI-236, 2002. type expresses a balance of these factors that appear to 5. Kinder, F. R., Jr., Bair, K. W., Chen, W., Florence, G., Francavilla, C., Geng, P., Gunasekera, S., Lassota, P. T., Longley, R. E., Palermo, M. G., maintain its microtubule dynamics within a narrow range Paterson, I., Pomponi, S., Ramsey, T. M., Rogers, L., Sabio, M., Sereinig, that supports mitosis, cell signaling, transport, and migra- N., Sorensen, E., Wang, R. M., and Wright, A. Global SAR study of the tion. Dynamics that are too fast or too slow appear to lead to novel microtubule stabilizing agent discodermolide. Proc. AACR, 43: 3650, 2002. mitotic block and ensuing cell death (18, 30). Regardless 6. Francavilla, C., Chen, W., and Kinder, F. R., Jr. Formal synthesis of of the inherent cellular microtubule dynamicity, the anti- (+)-discodermolide. Org. Lett., 5: 1233 – 1236, 2003. 7. ter Haar, E., Kowalski, R., Hamel, E., Lin, C., Longley, R., Gunasekera, S., Rosenkranz, H., and Day, B. W. Discodermolide, a cytotoxic marine agent that stabilizes microtubules more potently than taxol. Biochemistry, 35: 243 – 250, 1996. 2 M. A. Jordan, K. Kamath, B. A. Littlefield, and L. Wilson, unpublished data. 8. Chen, J-G. and Horwitz, S. B. Differential mitotic responses to 3 For review, see M. A. Jordan and L. Wilson. Microtubules as a target for anti-cancer microtubule-stabilizing and -destabilizing drugs. Cancer Res., 62: 1935 – drugs, manuscript in preparation. 1938, 2002.

9. Broker, L. E., Huismann, C., Ferreria, C. G., Rodriguez, J. A., Kruyt, F. A., colorimetric cytotoxicity assay for anticancer-drug screening. J. Natl. and Giaccone, G. Late activation of apoptotic pathways plays a negligible Cancer Inst., 82: 1107 – 1112, 1990. role in mediating the cytotoxic effects of discodermolide and epothilone B in 20. Guise, S., Braguer, D., Carles, G., Delacourte, A., and Briand, C. non-small cell lung cancer cells. Cancer Res., 4081 – 4088, 2002. 62: Hyperphosphorylation of H is mediated by ERK activation during anticancer 10. Hamel, E., Sackett, D., Vourloumis, D., and Nicolaou, K. The coral- drug-induced apoptosis in neuroblastoma cells. J. Neurosci. Res., 63: derived natural products eleutherobin and sarcodictyins A and B: effects 257 – 267, 2001. on the assembly of purified tubulin with and without microtubule- 21. Walker, R. A., O’Brien, E. T., Pryer, N. K., Soboeiro, M. F., Voter, associated proteins and binding at the polymer taxoid site. Biochemistry, W. A., Erickson, H., and Salmon, E. D. Dynamic instability of individual 38: 5490 – 5498, 1999. microtubules analyzed by video light microscopy: rate constants and 11. Jordan, M. A. Mechanism of action of antitumor drugs that interact transition frequencies. J. Cell Biol., 107: 1437 – 1448, 1988. with microtubules and tubulin. Curr. Med. Chem.—Anti-Cancer Agents, 2: Toso, R. J., Jordan, A., Farrell, K. W., Matsumoto, B., and Wilson, L. 1 – 17, 2002. 22. Kinetic stabilization of microtubule dynamic instability in vitro by 12. Landen, J. W., Lang, R., McMahon, S. J., Rudan, N. M., Yvon, A-M., vinblastine. Biochemistry, 32: 1285 – 1293, 1993. Adams, A. W., Sorcinelli, M. D., Campbell, R., Bonaccorsie, P., Ansel, J. C., Archer, D. R., Wadsworth, P., Armstrong, C. A., and Joshi, H. C. 23. Derry, W. B., Wilson, L., and Jordan, M. A. Substoichiometric binding Noscapine alters microtubule dynamics in living cells and inhibits the of taxol suppresses microtubule dynamics. Biochemistry, 34: 2203 – progression of melanoma. Cancer Res., 62: 4109 – 4114, 2002. 2211, 1995. 13. Jordan, M. A., Toso, R. J., Thrower, D., and Wilson, L. Mechanism of 24. Kamath, K. and Jordan, M. A. Suppression of microtubule dynamics mitotic block and inhibition of cell proliferation by taxol at low concen- by epothilone B in living MCF7 cells. Cancer Res., 63: 6026 – 6031, 2003. trations. Proc. Natl. Acad. Sci. USA, 90: 9552 – 9556, 1993. 25. Li, X. and Nicklas, R. B. Mitotic forces control a cell-cycle checkpoint. 14. Jordan, M. A., Thrower, D., and Wilson, L. Mechanism of inhibition of Nature, 373: 630 – 632, 1995. cell proliferation by Vinca alkaloids. Cancer Res., 51: 2212 – 2222, 1991. 26. Okouneva, T., Hill, B. T., Wilson, L., and Jordan, M. A. The effects of 15. Dhamodharan, R. I., Jordan, M. A., Thrower, D., Wilson, L., and vinflunine, vinorelbine, and vinblastine on centromere dynamics. Mol. Wadsworth, P. Vinblastine suppresses dynamics of individual micro- Cancer Ther., 2: 427 – 436, 2003. tubules in living cells. Mol. Biol. Cell, 6: 1215 – 1229, 1995. 27. Kelling, J., Sullivan, K., Wilson, L., and Jordan, M. A. Suppression of 16. Ngan, V., Bellman, K., Hill, B., Wilson, L., and Jordan, M. Mechanism centromere dynamics by taxol in living osteosarcoma cells. Cancer Res., of mitotic block and inhibition of cell proliferation by the semisynthetic 63: 2794 – 2801, 2003. Vinca alkaloids vinorelbine and its newer derivative vinflunine. Mol. 28. Wadsworth, P. and Bottaro, D. P. Microtubule dynamic turnover is Pharmacol., 60: 225 – 232, 2001. suppressed during polarization and stimulated in hepatocyte growth factor 17. Yvon, A-M., Wadsworth, P., and Jordan, M. A. Taxol suppresses scattered Madin-Darby canine kidney epithelial cells. Cell Motil. Cytoske- dynamics of individual microtubules in living human tumor cells. Mol. Biol. let., 35: 225 – 236, 1996. Cell, 10: 947 – 949, 1999. 29. Shelden, E. and Wadsworth, P. Observation and quantification of 18. Goncalves, A., Braguer, D., Kamath, K., Martello, L., Briand, C., individual microtubule behavior in vivo: microtubule dynamics are cell- Horwitz, S., Wilson, L., and Jordan, M. A. Resistance to taxol in lung type specific. J. Cell Biol., 120: 935 – 945, 1993. cancer cells associated with increased microtubule dynamics. Proc. Natl. 30. Jordan, M. A., Wendell, K. L., Gardiner, S., Derry, W. B., Copp, H., Acad. Sci. USA, 98: 11737 – 11741, 2001. and Wilson, L. Mitotic block induced in HeLa cells by low concentrations of 19. Skehan, P., Storeng, R., Scudiero, D., Monks, A., Mcmahon, J., paclitaxel (Taxol) results in abnormal mitotic exit and apoptotic cell death. Vistica, D., Warren, J. T., Bokesh, H., Kenney, S., and Boyd, M. R. New Cancer Res., 56: 816 – 825, 1996.

Downloaded from mct.aacrjournals.org on September 24, 2021. © 2003 American Association for Cancer Research. Suppression of microtubule dynamics by discodermolide by a novel mechanism is associated with mitotic arrest and inhibition of tumor cell proliferation

Stéphane Honore, Kathy Kamath, Diane Braguer, et al.

Mol Cancer Ther 2003;2:1303-1311.

Updated version Access the most recent version of this article at: http://mct.aacrjournals.org/content/2/12/1303

Cited articles This article cites 28 articles, 17 of which you can access for free at: http://mct.aacrjournals.org/content/2/12/1303.full#ref-list-1

Citing articles This article has been cited by 8 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/2/12/1303.full#related-urls

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

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

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/2/12/1303. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.