![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Departments of Chemistry (B.S.R., Y.S., Y.F., W.-H.J., D.P.C., B.W.D.), Pharmacology (A.V., R.P.S.), Drug Discovery Institute (A.V., K.M., B.W.D.), and Department of Pharmaceutical Sciences (C.M., R.B., B.W.D.), University of Pittsburgh, Pittsburgh, Pennsylvania
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
- and β-tubulin, which share significant sequence homology (Howard and Hyman, 2003
). The importance of MTs for mitosis and the potential to interfere with the delicate dynamics of MTs has led to the search for and development of many tubulin- and MT-interacting compounds. MT-perturbing agents can be roughly classified into two groups based upon the mechanism by which they disrupt normal microtubule function. These include agents that stabilize MT polymer formation and agents that destabilize MTs or inhibit MT polymer formation. The first discovered example of the former class is paclitaxel (Taxol) (Schiff et al., 1979), which has been used successfully in the treatment of breast, lung, and ovarian carcinomas (Jordan and Wilson, 2004).
Since the discovery of paclitaxel, many other MT stabilizers have been identified from various natural sources, including discodermolide (Gunasekera et al., 1990), dictyostatin (Pettit et al., 1994), the epothilones (Bollag et al., 1995), cyclostreptin (Sato et al., 2000), peloruside A (Hood et al., 2002), and laulimalide (Mooberry et al., 1999). Except for laulimalide and peruloside A (Pryor et al., 2002; Gaitanos et al., 2004), all of these agents share a common binding site on β-tubulin. Of these agents, dictyostatin and discodermolide have the greatest affinity for the taxoid binding site, followed by epothilone B (Madiraju et al., 2005) [epothilone B (EP0906; generic name patupilone) is currently in phase II/III development for a number of different indications]. In addition, all of these agents retain activity against paclitaxel-resistant cell lines (Kowalski et al., 1997; Edler et al., 2005).
The structure of dictyostatin remained unknown for a decade because of the scarcity of the compound. Finally in 2004, its structure was elucidated from high-field NMR and computational analyses (Paterson et al., 2004b) and was independently verified by two full syntheses (Paterson et al., 2004a; Shin et al., 2004). The chemical structure of dictyostatin closely resembles that of discodermolide, and 10 of the 11 stereocenters in dictyostatin are shared with discodermolide.
|
). Structure-activity relationship (SAR) studies to search for the minimum structural elements that contribute to cytotoxicity and retain affinity for the taxoid site were applied here to dictyostatin analogs. A few analogs of dictyostatin and their relative potencies against several cancer cell lines have been reported by others (Paterson and Gardner, 2007; Paterson et al., 2007). We have synthesized several analogs in attempts to study the SAR of the dictyostatin scaffold (Fukui et al., 2006; Jung et al., 2007; Shin et al., 2007). From this collection of existing dictyostatin analogs, we chose several of them based on the structural differences between dictyostatin and discodermolide for detailed biological SAR studies (Fig. 1). The C1-C9 chain is one of the main structural differences between the two natural products, and analogs designed for this purpose were epimers at C6, C7 and C9, and C2:C3 geometric isomers. From discodermolide SAR studies, it has been established that removal of the C14 methyl group does not decrease cytotoxicity in the cancer cell lines tested. However, when C14-normethyldiscodermolide was tested in paclitaxel-resistant cell lines, cytotoxicity was greatly diminished, with IC50 values in the micromolar range (Smith et al., 2005). Because discodermolide can be considered an open chain form of the macrolactone of dictyostatin, a 16-epi open chain methyl ester analog of the latter was included in our analysis. Finally, we made dictyostatin analogs containing a C15-C16 Z-alkene because discodermolide possesses a C13-C14 Z-alkene, which is missing in dictyostatin, and two 20-member macrolactone iso analogs (Shin et al., 2005, 2007; Fukui et al., 2006; Jung et al., 2007). To examine in detail the effects of these targeted structural modifications on biological activity and the interaction with the proposed target, we used quantitative multiparameter immunofluorescence assays to measure mitotic arrest and nuclear morphology in conjunction with in vitro tubulin assembly and displacement studies using the potent taxoid site binding agent epothilone B. Quantitative structure-activity relationship (QSAR) analyses were then employed to correlate the structures with cytotoxic effects.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
). The syntheses of the analogs are reported elsewhere (Shin et al., 2005, 2007; Fukui et al., 2006; Jung et al., 2007). Tubulin was isolated to electrophoretic homogeneity from fresh bovine brain (Hamel and Lin, 1984). Paclitaxel was obtained from the Drug Synthesis and Chemistry Branch, National Cancer Institute (Bethesda, MD). Discodermolide, epothilone B, and [14C]epothilone B (specific activity, 111 mCi/mmol) were generous gifts from Novartis Pharma AG (Basel, Switzerland). Rabbit anti-phosphohistone H3 antibody was purchased from Upstate Cell Signaling (Billerica, MA). Cy3-labeled donkey anti-rabbit and fluorescein-5-isothiocyanate (FITC)-labeled anti-mouse secondary antibodies were from Jackson Immunoresearch Laboratories (West Grove, PA). GTP and 2',3'-dideoxy-GTP (ddGTP) were purchased from U.S. Biochemical Corp. (Cleveland, OH). All other chemicals including mouse anti--tubulin antibody were purchased from Sigma (St. Louis, MO).
Methods
High-Content Analysis of Mitotic Arrest. Detailed conditions for the high-content analysis of microtubule stabilizers using a cell-based immunofluorescence assay have been described previously (Mínguez et al., 2003). In brief, HeLa cells were plated at a cell density of 8000 cells per well in collagen-1-coated 384-well microplates. Cells were allowed to attach and spread for 3 to 8 h and treated with either vehicle (DMSO) or 10 2-fold concentration gradients of test agents and incubated for an additional 20 h at 37 YC. After the incubation, cells were fixed with formaldehyde containing 10 µg/ml Hoechst 33342 to stain nuclei/chromatin, permeabilized with Triton X-100 and simultaneously immunostained with antibodies against -tubulin (mouse monoclonal, 1:2500 dilution; Sigma, St. Louis, MO) and phosphohistone H3 (rabbit polyclonal, 1:1000 dilution; Upstate) followed by FITC-labeled donkey antimouse IgG (1:250) and Cy3-labeled donkey anti-rabbit IgG (1:250) (both from Jackson Immunoresearch Laboratories). Each well was imaged on an ArrayScanII (Cellomics Inc., Pittsburgh, PA) using an Omega XF93 filter set at excitation/emission wavelengths of 350/461 nm (Hoechst), 494/519 nm (FITC), and 556/573 nm (Cy3). Images were analyzed with the Target Activation Bioapplication (Cellomics, Inc.). The algorithm provided quantitative measurements of nuclear FITC and Cy3 intensity, cell density, and mean and total nuclear Hoechst 33342 intensities. Microtubule mass was defined as the average FITC intensity in an area defined by the Hoechst-stained nuclei. To estimate the percentage of cells with elevated phosphohistone H3 levels and with condensed chromatin, thresholds were defined as the average intensity plus 1 S.D. of Hoechst or Cy3 intensities in anti-phosphohistone H3- or Hoechst 33342-stained cells, respectively, using a minimum of 1000 vehicle-treated cells per well. Cells were classified as positive if their average intensities in the Hoechst or Cy3 channel exceeded these thresholds. Cell density was defined as the number of nuclei per imaging field. The EC50 was the effective concentration of drug that decreased cell density by 50% compared with vehicle-treated control.
Tubulin Assembly in Vitro. Tubulin assembly was monitored by turbidity development at 350 nm as described previously (Madiraju et al., 2005). Reaction mixtures (final volume, 0.25 ml) contained tubulin (1 mg/ml; 10 µM), monosodium glutamate (MSG; 0.8 M from a stock solution adjusted to pH 6.6 with HCl), DMSO [final volume, 4% (v/v)], and test agent (10 µM). Reaction mixtures without test agent were cooled to 0°C and added to quartz cuvettes held at 0.25-0.5°C a temperature-controlled (via a Peltier unit), six-sample spectrophotometer (DU 7400; Beckman Coulter, Fullerton, CA). Test agent in DMSO was then rapidly mixed into the reaction mixture. Absorbance at 350 nm was monitored in each cuvette every 15 s. Each six-sample run contained one positive control (paclitaxel, 10 µM final concentration) and one negative control (DMSO only). Baselines were established at 0.25-2.5°C. The temperature was rapidly raised to 30°C (in approximately 1 min) and held there for 20 min. The temperature was then rapidly lowered back to 0.25-2.5°C. The change in absorbance 20 min after samples reached 30°C was used to calculate the extent of polymerization. The change in absorbance at 21 min in the paclitaxel sample was assigned as 100% assembly (positive control), whereas the change in absorbance in the DMSO-only sample (negative control) was assigned as 0% assembly.
Radioligand Displacement Studies. Determination of the equilibrium binding constant (Kd) for epothilone B was determined using methods described previously (Madiraju et al., 2005). An 80 µM stock solution of [14C]epothilone B was prepared in 50% DMSO. Samples of [14C]epothilone B at 0.005 to 2 µM were prepared in 4:1 (v/v) 0.75 M MSG/DMSO (total volume, 50 µl) and warmed to 37°C. Microtubules were preformed in separate tubes by incubating 2.5 µM tubulin and 25 µM ddGTP in 0.75 M MSG/DMSO for 30 min at 37°C. Equal volumes (50 µl each) of the radiolabeled compound mixtures and microtubule solutions were mixed and incubated for 30 min at 37°C. Reaction mixtures were then centrifuged in an Allegra TM 64R (Beckman-Coulter, Fullerton, CA) at 14,000 rpm for 30 min at room temperature. The amount of radiolabel in the supernatant was determined by scintillation spectrometry in a multipurpose scintillation counter (LS6500; Beckman-Coulter). Nonspecific radiolabel binding was determined from competition experiments with 8 µM discodermolide and processing as described above.
Inhibitors in 4:1 (v/v) 0.75 M MSG/DMSO were preincubated with radiolabeled compound in a final volume of 50 µl for 10 min at 37°C. To this mixture, 50 µl of microtubules that were formed as described above were added, and the resulting mixture was incubated for 30 min at 37°C. Reaction mixtures were centrifuged as described above, and a 50-µl aliquot of supernatant was analyzed by scintillation spectrometry.
Saturation binding data for [14C]epothilone B were analyzed using a modified version of Swillens' equation that accounts for ligand depletion (Swillens, 1995; Motulsky and Christopoulos, 2003):
 | (1) |
The equilibrium inhibition constant (Ki) values for dictyostatin and its analogs were determined using a modified Hanes analysis (Edler et al., 2005):
 | (2) |
is the fractional saturation of microtubule sites by epothilone B and inhibitor. GraphPad Prism was used for linear regression, providing the -Kd value as the x-intercept in the absence of inhibitor. The x-intercept in the presence of inhibitor, -Kd(1 + [I]/Ki), was used to calculate the Ki values.
QSAR. Molecular modeling and QSAR analyses were performed with the Cerius2 suite of programs (v. 4.5; Accelrys, Inc., San Diego, CA) using the Merck Molecular Force Field/CHARMM, the adopted-basis Newton-Raphson approach to molecular mechanics minimizations in a dielectric field = 1, rigid body superimpositions based on atom similarities, the MOPAC93 suite of semiempirical molecular orbital algorithms, and the genetic function approximation as described previously (Verdier-Pinard et al., 1998).
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
), to characterize the responses of tumor cells to MT-stabilizing agents (Mínguez et al., 2003; Madiraju et al., 2005). This assay provides information on the effects of such agents on MT mass, histone H3 phosphorylation, nuclear morphology, and cytotoxicity. HeLa cells were grown in 384-well microtiter plates and treated with each compound for 21 h. The cells were fixed and their nuclei stained with Hoechst 33342. They were then permeabilized, MTs and phosphohistone H3 were immunostained, and FITC- and Cy3-conjugated secondary antibodies were added. From Hoechst staining, a nuclear mask was generated defining the area used to determine MT mass and percentage of histone H3 phosphorylation. MT mass was determined from the cells' average nuclear FITC intensity. Histone H3 phosphorylation was determined from the number of cells with a higher pixel intensity for Cy3 than the average plus 1 S.D. seen within vehicle-treated cells. Using a range of concentration of test agents, the minimum detectable effective concentration (MDEC) was determined. The MDEC value is the minimum concentration at which each agent causes a response in excess of vehicle-treated control cells (Mínguez et al., 2003). The MDECs for nuclear condensation, histone H3 phosphorylation, and changes in MT morphology for dictyostatin and its analogs are shown in Table 1. For toxicity determinations, cell densities were measured as the number of Hoechst 33342-stained nuclei per imaging field. The toxicity measurements in HeLa cells were consistent with those reported for other human cancer cell lines, where the toxicity measure used was a decrease in absorbance at 490 nm after cells treated with drug were exposed to 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium and N-methylphenazine methylsulfate (Jung et al., 2007; Shin et al., 2007). Toxicity measurements and the MDEC values generated from this initial screen provided a rapid and economical assessment of the analogs' biological activities. As shown in Table 1, the 20-membered ring iso analogs of dictyostatin, 16-epi-isodictyostatin, 15Z,16-normethylisodictyostatin, and 16-epi-dictyostatin were inactive, yielding MDEC and EC50 values >5 µM. This led to the conclusions that the 22-membered ring of dictyostatin is essential for retention of biological activity and that if the C16 methyl group is present, it must be in the S configuration (Jung et al., 2007). 2E,15Z,16-Normethyldictyostatin analog was considerably less active than 15Z,16-normethyldictyostatin, proving that C2:3Z geometry is necessary for activity. From these data, the four most potent analogs were chosen for further evaluation: 6-epi-dictyostatin, 7-epi-dictyostatin, 16-normethyldictyostatin, and 15Z,16-normethyldictyostatin.
|
Figure 2 illustrates the cellular phenotype observed after treatment with the four test agents. MTs in cells treated with vehicle (A) were well organized, and the few cells undergoing mitosis showed normal spindle formation. Cells treated with dictyostatin (B) showed MT bundling and abnormal spindle formation, characteristic of cells treated with an MT stabilizer. Figure 2, C-F, represents the images obtained when HeLa cells were treated with the indicated concentrations of 6-epi-dictyostatin, 7-epi-dictyostatin, 15Z,16-normethyldictyostatin, or 16-normethyldictyostatin, respectively. All compounds produced morphological changes in the cells similar to those of dictyostatin, including MT bundling and abnormal spindles. Quantitative analysis showed that all compounds increased mitotic indices, nuclear condensation, MT mass, and a loss of cell density at nanomolar concentrations. 6-epi-Dictyostatin and 7-epi-dictyostatin affected all parameters at concentrations comparable with dictyostatin, whereas the 16-normethyl derivatives seemed to have lost some activity. We were surprised to find that 15Z,16-normethyldictyostatin was more potent in all aspects than 16-normethyldictyostatin, which was the least potent agent. Thus, high-content analysis provided a rapid and efficient means to generate cellular SAR data on mitotic arrest, microtubule morphology, and cytotoxicity. It demonstrated that the structural modifications made in the 6, 7, 15, and 16 positions did not significantly reduce compound activity in cells.
|
). To assess the hypernucleating ability of the dictyostatin epimers and olefin, we performed in vitro tubulin assembly assays under conditions in which polymerization does not occur in the absence of test agents and compared the analogs' activities with that of paclitaxel. Isolated tubulin was cooled to 2.5°C, and test agents were added to a final concentration of 10 µM. The test agent-containing tubulin solutions were rapidly warmed to 30°C to induce assembly, and after 20 min at 30°C, the temperature was lowered to 0°C to test the cold stability of the polymer induced by the agents. Figure 3 shows that three agents were more active than the positive control (paclitaxel). Dictyostatin was the most potent compound in terms of nucleating and polymerizing ability, closely followed by 7-epi-dictyostatin and 6-epi-dictyostatin. There were some slight differences between the C6- and C7-epimers. 7-epi-Dictyostatin induced polymer formation more rapidly than did 6-epi-dictyostatin, but the extent of polymerization caused by 6-epi-dictyostatin was greater. The polymer formed with 6-epi-dictyostatin seemed more stable against cold-induced depolymerization than the polymer from 7-epi-dictyostatin or dictyostatin itself. The nucleating and polymerizing activities of 15Z,16-normethyldictyostatin and 16-normethyldictyostatin were reduced but still comparable with those of paclitaxel. Polymer induced by these two compounds was cold-stable.
|
Characterization of [14C]Epothilone B Saturation Binding to Microtubules. We previously showed that dictyostatin and the four analogs have the ability to compete with radiolabeled paclitaxel for binding to preformed MTs in vitro (Madiraju et al., 2005; Jung et al., 2007). Here, we expanded on these preliminary observations by performing detailed binding kinetics using [14C]epothilone B. Epothilone B has an even higher affinity to the taxoid site on tubulin than paclitaxel (Kowalski et al., 1997b). Because the saturation binding kinetics of [14C]epothilone B have not been reported, we first determined its Kd value for the binding to MTs. Fig. 4, top, shows the extent of nonspecific binding of [14C]epothilone B bound to MTs in the presence and absence of 8 µM discodermolide. Fitting the data using Swillens' modified equation (eq. 1), which simultaneously fits total binding and nonspecific binding, and accounts for ligand depletion, yielded a Kd value for epothilone B of 36.04 ± 1.58 nM. Because the Kd value of paclitaxel is reported to be between 0.06 and 0.1 µM and because the relative affinities of agents for the taxoid site followed the order paclitaxel < epothilone B < discodermolide, the Kd value obtained in these experiments for epothilone B was considered reasonable and acceptable (Buey et al., 2004).
|
). The method uses the free concentration of radioligand instead of the total concentration and accounts for the free inhibitor concentrations. As shown in Fig. 4, bottom, plots of free concentration of [14C]epothilone B versus free substrate (S) divided by the fractional saturation (; used instead of enzyme velocity) for 6-epi-dictyostatin yielded plots of parallel lines with slopes close to 1. The Kd value for [14C]epothilone B, determined from the negative intercept on the x-axis, was 50 ± 20 nM, close to the value obtained from the saturation experiments in Fig. 4, top. As shown in Table 2, 16-normethyldictyostatin and 15Z,16-normethyldictyostatin gave the highest Ki values, reflecting their reduced binding affinity. The Ki value for 15Z,16-normethyldictyostatin was 4.47 ± 0.28 µM, and the value for 16-normethyldictyostatin was 4.55 ± 0.41 µM. These data correlated with the tubulin polymerization data, which showed that 16-normethyldictyostatin and 15Z,16-normethyldictyostatin were the least potent of the compounds in nucleation and MT stabilization experiments (Fig. 3). Dictyostatin showed the highest affinity for the taxoid site, with a Ki value of 70 ± 20 nM. Based on their effects on MT assembly, it was surprising to find that 6-epi-dictyostatin and 7-epi-dictyostatin had substantially less affinity under the conditions of the competition experiments, yielding Ki values of 480 ± 70 and 930 ± 310 nM, respectively.
|
QSAR Analyses. Using the NMR-derived solution structure of discodermolide as a template (Smith et al., 2001), molecular models of dictyostatin, its analogs, discodermolide, and 14-normethyldiscodermolide were built. The lowest energy conformers of the models were found by exhaustive molecular mechanics minimization, and then all models were superimposed onto that of discodermolide. The only structures that did not superimpose well were 2E,15Z,16-normethyldictyostatin and the 20-member iso analogs. A large number of thermodynamic, electronic, steric, and linear free-energy descriptors were then calculated for each model.
At this point we needed to decide upon the biological metric to use for the "activity" portion of the QSAR. We chose the 50% growth inhibitory concentrations (GI50) of the compounds against human ovarian carcinoma 1A9 cells, which express wild type β-tubulins. Because the number of descriptors far exceeded the number of compounds studied, the genetic function approximation (GFA) was used to fit the data to the smallest possible number of descriptors. In the GFA approach, a random population of equations is generated, their fitness is determined by a number of statistical metrics (in particular Friedman's lack-offit score and R2), and then the equations are "evolved" by inclusion of combinations (child) from two of the best equations (parents) (Verdier-Pinard et al., 1998). This process is reiterated (each iteration is termed a "generation"), with a given frequency of "mutation" (random descriptor loss or use) incorporated into each iteration until the best descriptors and QSAR equations are found. The five best equations (according to statistical metrics) from 100,000 generations were:
 | (3) |
 | (4) |
 | (5) |
 | (6) |
 | (7) |
|
The predicted activity values generated from these equations were plotted versus the actual activity in the 1A9 cells to determine correlations. As shown in Fig. 5, plotting of the -log molar predicted values generated from eq. 3 versus the -log molar actual values led to an R2 of 0.943. Values for R2 of the other four equations were also >0.93.
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
).
Although, qualitatively, the rank order of cellular activities correlated with the in vitro activities, there were some differences in the compounds' abilities to cause in vitro tubulin assembly and in the binding site affinity studies. It is noteworthy that the 16-normethyl derivatives 15Z,16-normethyldictyostatin and 16-normethyldictyostatin had almost 5-fold higher Ki values compared with the C6- and C7-epimers. This is consistent with the hypothesis that the C16 methyl group is important in binding to the taxoid site through a key interaction with Phe270 on β-tubulin. Our recent report shows that 16-normethyldictyostatin has decreased potency in the paclitaxel-resistant 1A9/PTX10 ovarian cancer cell line, which has a Phe270-to-Val point mutation in the taxoid binding site (Giannakakou et al., 1997; Shin et al., 2005). This is in agreement with our present results, where the lower affinity observed in competition experiments with 15Z,16-normethyldictyostatin and 16-normethyldictyostatin can be attributed to the loss of the favorable interaction of the C16 methyl group of dictyostatin with Phe270 in β-tubulin.
As observed from a similar loss in potency in the 1A9/PTX10 line, the interaction between the benzoyloxy group of paclitaxel and Phe270 of β-tubulin is important for paclitaxel's activity (Giannakakou et al., 1997). The importance of the interaction between Phe270 and methyl substituents in dictyostatin, epothilone B, and discodermolide is now becoming well established. Molecular modeling studies have suggested that the C14 methyl group of discodermolide is positioned in a favorable interaction with Phe270 in the hydrophobic binding pocket of the taxoid site (Xia et al., 2006). The importance of this interaction has been confirmed by the decrease of nucleating ability of 14-normethyldiscodermolide in tubulin polymerization assays (Smith et al., 2005). A two-dimensional NMR-derived binding model of epothilone A to the tubulin heterodimer suggests that the increased potency of epothilone B over epothilone A is due to the interaction of the C12 methyl with Phe270 (Reese et al., 2007).
As shown by our QSAR studies, the size of the molecule (area and molar refractivity) is the most significant contributor to cytostatic activity against 1A9 ovarian cancer cells. A loss of a methyl substituent would certainly lead to a decrease in the van der Waals surface area of the molecule and, according to our calculations, leads to a decrease in activity. Molar refractivity is a measure of the volume occupied by the molecule. It is dependent on a number of factors, such as molecular weight, refractive index, and density. This descriptor relates to the London dispersive forces that act in the drug-receptor interactions (Padron et al., 2002). Again, 16-normethyldictyostatin and 15Z,16-normethyldictyostatin would be expected to have molar refractivities differing from those of dictyostatin and its C6 epimer, whose molar refractivities are identical.
Changes to the C6 and C7 stereocenters of dictyostatin were explored because they are present in the structure of discodermolide as its C4 and C5 carbons, part of its delta lactone ring. The multiparameter immunofluorescence studies with 6-epi-dictyostatin and 7-epi-dictyostatin showed that these analogs produced MT bundling and abnormal spindle formation at nanomolar concentrations. Their ability to produce these various cellular responses was observed at lower concentrations than required with dictyostatin. Although there was a slight loss in nucleation ability in tubulin polymerization assays, these compounds retained their ability to induce assembly of microtubules and to form cold-stable polymer. In addition, both compounds still had a high affinity for the taxoid site on β-tubulin, with the Ki values determined in competition experiments with [14C]epothilone B of 480 and 930 nM for 6-epi-dictyostatin and 7-epi-dictyostatin, respectively. From these experiments it can be said that single stereochemical changes to this area are not only well tolerated but also lead to no loss in potency in vivo.
The saturation binding kinetics of epothilone B were characterized. One of the fundamental assumptions of equilibrium binding is that only a small amount of substrate binds to its target, leaving the relative amount of free substrate concentration unchanged. Ligand depletion occurs if more than 10% of the substrate binds to the protein (Motulsky and Christopoulous, 2003). When ligand depletion occurs, this assumption is invalidated and can, if not taken into account, cause errors in data analysis (Morton et al., 2002; Carter et al., 2007). This problem is commonly encountered in the centrifuge-based assay used to determine the kinetics of MT stabilizing agents. For this reason, the saturation binding data were analyzed using the modified Swillens' equation (eq. 1), which not only accounts for ligand depletion but also fits total binding and nonspecific binding simultaneously. Fitting the nonspecific and total binding data simultaneously allows for the calculation of free ligand, and defines nonspecific binding as a constant fraction of the free ligand concentration. The Kd value obtained using this analysis was reasonable as a measure of the affinity of epothilone B for the taxoid site in comparison to results from previous competitive inhibition studies (Kowalski et al., 1997; Madiraju et al., 2005).
The binding kinetics of dictyostatin and its analogs were determined with radiolabeled epothilone B. Epothilones are competitive inhibitors of paclitaxel binding, and the epothilones retain activity against almost all known paclitaxel-resistant cell lines (Kowalski et al., 1997); the exception is the 1A9/A8 line, which possess a mutation of Arg282->Gln in β-tubulin and is resistant to both paclitaxel and epothilone B (Giannakakou et al., 2000). All of the agents tested here were found to be competitive inhibitors of the binding of epothilone B to tubulin polymer, implying that dictyostatin and epothilone B share a common binding site on MTs. As in the saturation binding experiments with [14C]epothilone B, significant ligand depletion was observed in the competition experiments. For analysis of competition experiments with ligand depletion, it would be correct to assume that the inhibitor concentration would be depleted as well as substrate concentration. The modified Hanes analysis (eq. 2) used by Edler et al. (2005) and employed here accounts for the free concentration of radiolabeled tracer and that of the free inhibitor.
In conclusion, a number of potent analogs of the microtubule stabilizer dictyostatin have been tested to determine the cellular responses they evoke, their effects on tubulin polymerization, and their ability to inhibit binding of [14C]epothilone B to MTs. During this process, the saturation binding kinetics of epothilone B was characterized, and epothilone B demonstrated great affinity for the taxoid site. Dictyostatin and its analogs are all competitive inhibitors of [14C]epothilone B binding as demonstrated by Hanes analysis. These experiments have further led to the conclusion that dictyostatin, discodermolide, epothilone B, and paclitaxel have favorable interactions with Phe270 within the taxoid binding site on β-tubulin. A change to the configuration of the stereocenter at C6 or C7 of dictyostatin is well tolerated. From this work, it can be conjectured that for good efficacy against cancer cell lines, all future analogs of dictyostatin must retain the ability to interact with Phe270 on β-tubulin, and that this interaction is important to the efficacy of many other taxoid site-binding, MT-stabilizing compounds. The characterization of the 6-epi-dictyostatin analog as a potent MT-stabilizing agent has motivated the gram scale synthesis for use in preclinical studies. Our QSAR analyses led to a number of equations that will be useful in guiding the production of and predicting the cellular activity of new, more potent dictyostatin analogs. The four descriptors employed in these equations are all variables that can easily be manipulated by a medicinal chemist.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
ABBREVIATIONS: MT, microtubule; SAR, structure-activity relationship; QSAR, quantitative structure-activity relationship; FITC, fluorescein isothiocyanate; DMSO, dimethylsulfoxide; MSG, monosodium glutamate; MDEC, minimum detectable effective concentration; GFA, genetic function approximation.
Address correspondence to: Dr. Billy W. Day, 10017 Biomedical Science Tower 3, 3501 Fifth Avenue, Pittsburgh, PA 15213. E-mail: bday{at}pitt.edu
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Buey RM, Diaz F, Andreu JM, OÕBrate A, Giannakakou P, Nicolau KC, Samai PK, Ritzen A, and Namato K (2004) Interaction of epothilone analogs with the paclitaxel binding site: relationship between binding affinity, microtubule stabilization, and cytotoxicity. Chem Biol 11: 225-236.[CrossRef][Medline]
Carter CM, Leighton-Davies JR, and Charlton SJ (2007) Miniaturized receptor binding assays: Complications arising from ligand depletion. J Biomol Screen 12: 255-266.
Edler MC, Buey RM, Gussio R, Marcus AI, Vanderwal CD, Sorensen EJ, Diaz JF, Giannakakou P, and Hamel E (2005) Cyclostreptin (FR182877), an antitumor tubulin-polymerizing agent deficient in enhancing tubulin assembly despite its high affinity for the taxoid site. Biochemistry 44: 11525-11538.[CrossRef][Medline]
Fukui Y, Brückner AM, Shin Y, Balachandran R, Day BW, and Curran DP (2006) Fluorous mixture synthesis of (-)-dictyostatin and three stereoisomers. Org Lett 8: 301-304.[CrossRef][Medline]
Gaitanos TN, Buey RM, Diaz JF, Northcote PT, Teesdale-Spittle P, Andreu JM, and Miller JH (2004) Peloruside A does not bind to the taxoid site on beta-tubulin and retains its activity in multidrug-resistant cell lines. Cancer Res 64: 5063-5067.
Giannakakou P, Sackett DL, Kang YK, Zhan Z, Buters JT, Fojo T, and Poruchynsky MS (1997) Paclitaxel-resistant human ovarian cancer cells have mutant β-tubulins that exhibit impaired paclitaxel-driven polymerization. J Biol Chem 272: 17118-17125.
Giannakakou P, Gussio R, Nogales E, Downing KH, Zaharevitz D, Bollbuck B, Poy G, Sackett D, Nicolaou KC, and Fojo T (2000) A common pharmacophore for epothilone and taxanes: molecular basis for drug resistance conferred by tubulin mutations in human cancer cells. Proc Natl Acad Sci U S A 97: 2904-2909.
Giuliano KA, DeBiasio RL, Dunlay T, Gough A, Volosky JM, Zock J, Pavlakis GN, and Taylor DL (1997) High-content screening: a new approach to easing key bottlenecks in the drug discovery process. J Biomol Screen 2: 249-259.[Abstract]
Gunasekera SP, Gunasekera M, Longley RE, and Schulte GK (1990) Discodermolide: a new bioactive polyhydroxylated lactone from the marine sponge Discodermia dissoluta. J Org Chem 55: 4912-4915.
Hamel E and Lin CM (1984) Separation of active tubulin and microtubule-associated proteins by ultracentrifugation and isolation of a component causing the formation of microtubule bundles. Biochemistry 23: 4173-4184.[CrossRef][Medline]
Hood KA, West LM, Rouwe B, Northcote PT, Berridge MV, Wakefield SJ, and Miller, JH (2002) Peluroside A, a novel antimitotic agent with paclitaxel-like microtubule-stabilizing activity. Cancer Res 62: 3356-3360.
Howard J and Hyman AA (2003) Dynamics and mechanics of the microtubule plus end. Nature 422: 753-758.[CrossRef][Medline]
Jordan MA and Wilson L (2004) Microtubules as a target for anticancer drugs. Nat Rev Cancer 4: 253-265.[CrossRef][Medline]
Jung WH, Harrison C, Shin Y, Fournier JH, Balachandran R, Raccor BS, Sikorski RP, Vogt A, Curran DP, and Day BW (2007) Total synthesis and biological evaluation of C16 analogs of (-)-dictyostatin. J Med Chem 50: 2951-2966.[CrossRef][Medline]
Kowalski RJ, Giannakakou P, and Hamel E (1997b) Activities of the microtubule-stabilizing agents epothilones A and B with purified tubulin and in cells resistant to paclitaxel (Taxol). J Biol Chem 272: 2534-2541.
Madiraju C, Edler MC, Hamel E, Raccor BS, Balachandran R, Zhu G, Giuliano KA, Vogt A, Shin Y, Fournier JH, et al. (2005) Tubulin assembly, taxoid site binding, and cellular effects of the microtubule-stablilizing agent dictyostatin. Biochemistry 44: 15053-15063.[CrossRef][Medline]
Mínguez JM, Kim SY, Guillano KA, Balachandran R, Madiraju C, Day BW, and Curran DP (2003) Synthesis and biological assessment of simplified analogues of the potent microtubule stabilizer (+)-discodermolide. Bioorg Med Chem 11: 3335-3357.[CrossRef][Medline]
Mooberry SL, Tien G, Hernandez AH, Plubrukarn A and Davidson BS (1999) Laulimalide and isolaulimalide, new paclitaxel-like microtubule-stabilizing agents. Cancer Res 59: 653-660.
Morton MF, Harper EA, Tavares IA, and Shankley NP (2002) Pharmacological evidence for putative CCK1 receptor heterogeneity in human colon smooth muscle. Br J Pharmacol 136: 873-882.[CrossRef][Medline]
Motulsky H and Christopoulos A (2003) Fitting models to biological data using linear and nonlinear regression. A practical guide to curve fitting. GraphPad Software Inc., San Diego, CA.
Padron JA, Carrasco R, and Pellon RF (2002) Molecular descriptor based on a molar refractivity partition using Randic-type graph-theoretical invariant. J Pharm Sci 5: 258-265.
Paterson I, Britton R, Delgado O, Meyer A, and Poullennec KG (2004a) Total synthesis and configurational assignment of (-)-dictyostatin, a microtubule-stabilizing macrolide of marine sponge origin. Angew Chem Int Ed Engl 43: 4629-4633.[CrossRef]
Paterson I, Britton R, Delgado O, and Wright AE (2004b) Stereochemical determination of dictyostatin, a novel microtubule-stabilising macrolide from the marine sponge Corallistidae sp. Chem Commun (Camb) 632-633.
Paterson I and Gardner NM (2007) Design, synthesis and biological evaluation of a macrocyclic discodermolide/dictyostatin hybrid. Chem Commun (Camb) 49-51.
Paterson I, Gardner NM, Poullennec KG, and Wright AE (2007) Synthesis and biological evaluation of novel analogues of dictyostatin. Bioorg Med Chem Lett 17: 2443-2447.[CrossRef][Medline]
Pettit GR, Cichacz ZA, Gao F, Boyd MR, and Schmidt JM (1994) Isolation and structure of the cancer cell growth inhibitor dictyostatin 1. J Chem Soc Chem Commun 1111-1112.
Pryor DE, O'Brate A, Bilcer G, Diaz JF, Wang Y, Wang Y, Kabaki M, Jung MK, Andreu JM, Ghosh AK, et al. (2002) The microtubule stabilizing agent laulimalide does not bind in the taxoid site, kills cells resistant to paclitaxel and epothilones, and may not require its epoxide moiety for activity. Biochemistry 41: 9109-9115.[CrossRef][Medline]
Reese M, Sánchez-Pedregal VM, Kubicek K, Meiler JM, Blommers MJJ, Griesinger C, and Carlomagno T (2007) Structural basis of the activity of the microtubule-stabilizing agent epothilone A studied by NMR spectroscopy in solution. Angew Chem Int Ed Engl 46: 1864-1868.[CrossRef]
Sato B, Muramatsu H, Miyauchi M, Hori Y, Takase S, Hino M, Hashimoto S, and Terano H (2000) A new antimitotic substance, FR182877. I. Taxonomy, fermentation, isolation, physico-chemical properties and biological activities. J Antibiot (Tokyo) 53: 123-130.[Medline]
Schiff PB, Fant J, and Horwitz SB (1979) Promotion of microtubule assembly in vitro by taxol. Nature 277: 665-667.[CrossRef][Medline]
Shin Y, Fournier JH, Fukui Y, Bruckner AM, and Curran DP (2004) Total synthesis of (-)-dictyostatin: confirmation of relative and absolute configurations. Angew Chem Int Ed Engl 43: 4634-4637.[CrossRef]
Shin Y, Fournier JH, Balachandran R, Madiraju C, Raccor BS, Zhu G, Edler MC, Hamel E, Day BW, and Curran DP (2005) Synthesis and biological evaluation of (-)-16-normethyldictyostatin: a potent analogue of (-)-dictyostatin. Org Lett 7: 2873-2876.[CrossRef][Medline]
Shin Y, Fournier JH, Bruckner A, Madiraju C, Balachandran R, Raccor BS, Edler MC, Hamel E, Sikorski RP, Vogt A, et al. (2007) Synthesis and biological evaluation of (-)-dictyostatin and stereoisomers. Tetrahedron 63: 8537-8562.[CrossRef][Medline]
Smith AB 3rd, LaMarche MJ, and Falcone-Hindley M (2001) Solution structure of (+)-discodermolide. Org Lett 3: 695-698.[CrossRef][Medline]
Smith AB 3rd, Freeze BS, LaMarche MJ, Hirose T, Brouard I, Xian M, Sundermann KF, Shaw SJ, Burlingame MA, Horwitz SB, et al. (2005) Design, synthesis, and evaluation of analogues of (+)-14-normethyldiscodermolide. Org Lett 7: 315-318.[CrossRef][Medline]
Swillens S (1995) Interpretation of binding curves obtained with high receptor concentrations: Practical aid for computer analysis. Mol Pharmacol 47: 1197-1203.[Abstract]
Verdier-Pinard P, Lai JY, Yoo HD, Yu J, Marquez B, Nagle DG, Nambu M, White JD, Falck JR, Gerwick WH, et al. (1998) Structure-activity analysis of the interaction of curacin A, the potent colchicine site antimitotic agent, with tubulin and effects of analogs on the growth of MCF-7 breast cancer cells. Mol Pharmacol 53: 62-76.
Xia S, Kenesky CS, Rucker PV, Smith AB, Orr GA, and Horwitz SB (2006) A photoaffinity analogue of discodermolide specifically labels a peptide in betatubulin. Biochemistry 45: 11762-11775.[CrossRef][Medline]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
, 4% DMSO; 
, (-)-dictyostatin; 
, 16-normethyldictyostatin; x, 15Z,16-normethyldictyostatin; 
,6-epi-dictyostatin; *, 7-epi-dictyostatin; 
, paclitaxel.
). Nine concentrations in the range of 5 nM to 2 µM radiolabeled epothilone B (
). Each data point is the mean of three independent experiments ± S.D. Bottom, Hanes plot of the inhibition of binding of [14C]epothilone B to microtubules by 6-epi-dictyostatin. The y-axis is free substrate concentration (micromolar)/fractional saturation (S/
1. The x-axis is free concentration of [14C]epothilone B (micromolar). The fractional saturation was calculated from the total amount of tracer added minus the free concentration of tracer counted. Experiments were performed as described under Materials and Methods section. Ki values were determined from the x-intercept, equal to -(Kd(1+[I]/Ki). 