A novel synthetic microtubule inhibitor exerts antiproliferative effects in multidrug resistant cancer cells and cancer stem cells

The success of cancer chemotherapy is limited by multidrug resistance (MDR), which is mainly caused by P-glycoprotein (P-gp) overexpression. In the present study, we describe a novel microtubule inhibitor, 5-(N-methylmaleimid-3-yl)-chromone (SPC-160002), that can be used to overcome MDR. A synthetic chromone derivative, SPC-160002, showed a broad spectrum of anti-proliferative effects on various human cancer cells without affecting P-gp expression and its drug efflux function. Treatment with SPC-160002 arrested the cell cycle at the M phase, as evidenced using fluorescence-activated cell sorting analysis, and increased the levels of mitotic marker proteins, including cyclin B, pS10-H3, and chromosomal passenger complex. This mitotic arrest by SPC-160002 was mediated by promoting and stabilizing microtubule polymerization, similar to the mechanism observed in case of taxane-based drugs. Furthermore, SPC-160002 suppressed the growth and sphere-forming activity of cancer stem cells. Our data herein strongly suggest that SPC-160002, a novel microtubule inhibitor, can be used to overcome MDR and can serve as an attractive candidate for anticancer drugs.


Maleimide-containing chromone derivatives show anti-proliferative effects in MDR cancer cells.
To develop a novel anticancer drug that can overcome MDR, we examined the anti-proliferative effects of several chromone and xanthone derivatives (Fig. 1a) using P-gp-overexpressing KBV20C cells. The KBV20C cells, derived from KB 21,22 were evaluated to have more than 300 times resistance to paclitaxel compared to its parental KB cells (Table 1). Among them, maleimide-containing chromones and xanthones (SPC-160002, SPC-160003, and SPC-160004) showed anti-proliferative effects in both KBV20C cells and their parental KB cells, while succinimide-containing compounds (SPC-160001 and SPC-160005) had no growth inhibition effect until  www.nature.com/scientificreports/ a concentration of 100 μM in both the cells (Table 1 and Fig. 1b), indicating that the maleimide substituent is essential for the growth-inhibitory effect. Upon comparing the effects of SPC-160002, SPC-160003 and SPC-160004, we found that the structure of chromone had stronger activity than that of xanthone (Table 1). However, there was no difference in the anti-proliferative effects of N-methyl-substituted maleimide (SPC-160002) and (NH)-free-maleimide (SPC-160004) ( Table 1). Altogether these results suggest that the chromones containing maleimide or 5-N-methylmaleimide exert anti-proliferative effects against MDR cancer cells. However, their potencies are much weaker than that of paclitaxel (Table 1).
In this study, we also investigated the molecular mechanism by which SPC-160002 [5-(N-methylmaleimid-3-yl)-chromone] exerts anti-proliferative activity in MDR cancer cells. As shown in Table 2, SPC-160002 showed a broad spectrum of anti-proliferative effects on various human cancer cells. Based on the cytotoxicity of SPC-160002, we investigated whether SPC-160002 could induce apoptosis. Treatment of SPC-160002 led to an increase in the Annexin V-positive population in both KB and KBV20C cells (Fig. 1c). The apoptotic potential of SPC-160002 was further confirmed by measuring PARP cleavage, as a marker of apoptosis (Fig. 1d). As expected, KBV20C cells were found to be highly resistant to vincristine, thus showing the MDR phenotype of KBV20C cells (Fig. 1d). Taken together, our data show that SPC-160002 has a broad spectrum of anti-proliferative effects on various human cancer cells as well as MDR cancer cells.

SPC-160002 does not affect P-gp expression and its drug efflux function.
To determine whether SPC-160002 affects the expression or function of P-gp, we first examined the expression levels of P-gp mRNA and protein. As shown in Fig. 2a,b, there were no differences in both mRNA and protein levels between the control and SPC-160002-treated cells. Following that, to investigate whether the efflux function of P-gp is affected by SPC-160002, we measured the intracellular accumulation of rhodamine 123, a well-established substrate of P-gp. Rhodamine 123 was found to be extensively accumulated in KB cells, but not in KBV20C cells (Fig. 2c), re-validating the MDR phenotype of KBV20C cells. There was no change in the accumulation of rhodamine 123 in KBV20C cells upon SPC-160002 treatment, while there was a profound increase upon treatment with a P-gp inhibitor, verapamil (Fig. 2c), which was further confirmed using fluorescence microscopy (Fig. 2d). These results indicate that SPC-160002 does not affect both P-gp expression and its drug efflux function.

SPC-160002 blocks cell cycle progression at the G 2 /M phase.
To investigate the molecular mechanisms by which SPC-160002 exerts anti-proliferative activity on cancer cells, we analyzed the cell cycle using flow cytometry. As shown in Fig. 3a,b, treatment with SPC-160002 dose-dependently decreased the proportion of cells in G 1 phase and increased the proportion of cells in G 2 /M phase in both KB and KBV20C cells. As expected, vincristine induced G 2 /M arrest in KB cells, but not in KBV20C cells. To confirm G 2 /M arrest by SPC-160002, we examined the levels of cell cycle-related proteins. SPC-160002 dramatically increased cyclin B1, a key regulator of the G 2 /M transition, but decreased G 1 /S-specific cyclin D3 (Fig. 3c). In addition, there was a drastic decrease in the level of pY15-cdk1, an inactive form, indicating that SPC-160002 promotes M phase transition from G 2 phase. There was a clear decrease in the levels of the cdk inhibitors, p21 and p27 at 10 µM of SPC-160002, similar to the effect seen in case of nocodazole, a well-known mitotic inhibitor. Furthermore, SPC-160002 treatment dramatically increased both pS10-H3 and pT3-H3, which are representative histone marks for M phase (Fig. 3c). It has been well-established that pS10-H3 is mainly generated by Aurora B kinase in chromosomal passenger complex (CPC) during mitosis 23 , and the CPC complex, composed of Aurora B, INCENP, Borealin, and Survivin, is a major regulator of chromosome segregation during mitosis and cytokinesis [24][25][26] . We thus examined the expression levels of CPC components in the presence of SPC-160002. As expected, SPC-160002 treatment dramatically increased the level of all CPC components, Aurora B, Borealin, and Survivin, with only the exception of INCENP. These expression profiles upon SPC-160002 treatment were consistent with those of nocodazole. Altogether, our data support that SPC-160002 is a mitotic inhibitor and causes mitotic arrest of cancer cells.  There was a transient increase in the levels of pS10-H3, pT3-H3, and CPC in the control cells at the M phase, followed by a rapid decrease at the G1 phase, which depends on cell cycle progression (Fig. 4c). However, the levels in SPC-160002-treated cells were sustained after 12 h of release (Fig. 4c). Furthermore, similar results were obtained from the TN-release experiment; the levels of pS10-H3, pT3-H3, and CPC in the SPC-160002 treated cells were higher than those in the control cells (Fig. 4d). These results were in good accordance with the previous results of cell cycle progression ( Supplementary Fig. 1a,b and 4a,b), strongly supporting that SPC-160002 blocks mitotic exit in the cell cycle, and subsequently, causes mitotic arrest.

SPC-160002 promotes and stabilizes tubulin polymerization.
Given the mitotic arrest by SPC-160002, we performed immunofluorescence staining of microtubules to analyze the effect of SPC-160002 on mitotic morphology. Remarkably, 39% of SPC-160002-treated cells had a mitotic spindle structure with strong microtubule fluorescence intensity, which was comparable to paclitaxel treated cells (Fig. 5a,b). We then subdivided the cells displaying mitotic morphology according to the mitotic substages ( Supplementary Fig. 2). As shown in Fig. 5c, in the presence of SPC-160002, there was a profound increase in the mitotic cell populations in pro-metaphase, but not in anaphase or telophase, similar to the phenotypes of taxane class drugs 27,28 . The association of taxanes with β-tubulin stabilizes the microtubule structure and consequently suppresses the microtu- www.nature.com/scientificreports/ bule dynamics, resulting in mitotic arrest 27,28 . To examine the effects of SPC-160002 on tubulin polymerization dynamics, we performed an in vitro microtubule polymerization assay. Purified tubulin was incubated with SPC-160002 and the absorbance at 340 nm was measured every 30 s during incubation. Remarkably, we found that SPC-160002 dramatically promoted tubulin polymerization (Fig. 5d), like paclitaxel. Indeed, there was polymerization and stabilization of intracellular microtubule upon SPC-160002 treatment (Fig. 5e). However, the microtubule stabilizing effect of SPC-16002 appears to be much weaker than that of paclitaxel. Altogether, these results strongly suggest that SPC-160002 is a novel microtubule inhibitor that promotes and stabilizes microtubule polymerization.

SPC-160002 inhibits survival and sphere-formation of cancer stem cells. It has been well estab-
lished that CSCs represent a small population of cells with tumor-initiation potential 29,30 . Because CSCs are resistant to conventional chemotherapeutic agents as well as radiation therapy, and play a role in tumor metastasis and relapse 29,30 , they serve as important targets for the development of novel anticancer drugs. The resistance of CSCs to chemotherapy is mainly due to the overexpression of ABC transporters 8,31 .
We cultured KB cells in the attachment free condition with CSC culture medium. It generated typical CSC spheres (Fig. 6a) and were maintained in multiple passages. The P-gp expression of CSC spheres is higher than that of 2D monolayer cultured cells ( Supplementary Fig. 3), which is in good accordance with the observation that the KB-derived CSCs are at least tenfold more resistant to paclitaxel than the bulk culture KB cells (Supplementary Fig. 4). In the presence of SPC-160002, both the proliferation of 2D monolayer cultured KB cells and CSC sphere-forming activity were dramatically inhibited with the similar potency (Fig. 6b,c). To further confirm the inhibitory effect of SPC-160002 on the survival of CSCs, specifically, we performed Aldefluor assay which quantify aldehyde dehydrogenase (ALDH, a CSC-predominant enzyme)-positive CSCs. In the presence of SPC-160002, there was a significant reduction in the ALDH-positive cells in the KB-derived CSC spheres (Fig. 6d), suggesting that SPC-160002 effectively eliminates the CSCs in KB cells. To further strengthen the anti-CSC-sphere forming activity of SPC-160002, we used U87 glioblastoma CSC model in which the role of P-gp in CSC was well characterized 32 . As expected, SPC-160002 treatment effectively inhibits the population of CSCs in U87 cells ( Supplementary Fig. 5).

Discussion
Several MTIs, including paclitaxel and vincristine, have been used as potent and effective chemotherapeutics for various cancer treatments; however, acquired resistance to these molecules due to MDR serves as a major impediment for successful treatment. Thus, development of microtubule-targeting drugs that can overcome MDR is a major research problem. It has been reported that several chromone compounds have the ability to target microtubules 33,34 ; however, there is no evidence whether these anti-microtubule chromones have   29,30 . Although extensive research has been devoted to the development of anticancer drugs that target CSCs 29,31 , the effective treatment of CSCs remains a challenge that needs to be addressed. SPC-160002 inhibits the sphere-forming activity of both KB and U87 glioblastoma cells and decreases the ALDH-positive cells in CSC-like spheres, suggesting the potential of SPC-160002 as a CSC-targeting drug. Chromones and xanthones are classes of heterocyclic compounds that commonly exist in nature [17][18][19] . They consist of diverse functional substituents that are implicated in a variety of pharmacological activities, including anti-inflammatory, anti-bacterial, anti-oxidant, and anti-tumor activities. Their numerous biological and pharmacological activities depend on the structural composition of the substituents and several of these derivatives are indeed used as therapeutic agents. Our data show that SPC-160002 stabilizes microtubule structure and induces mitotic arrest, subsequently leading to apoptosis. There are the obvious differences between SPC-160002 and paclitaxel on microtubule stabilization effect (Fig. 5c,d), which seems to be due to their potency differences. The potency of paclitaxel is over 300 times higher than that of SPC-160002 (Table 1), which is good agreement with www.nature.com/scientificreports/ the observations that microtubule stabilizing effect of paclitaxel is higher than that of SPC-16002 (Fig. 5c,d). The microtubule-stabilizing effect of SPC-160002 appears to be due to the maleimide group, because the succinimidecontaining compounds (SPC-160001 and SPC-160005) did not show any such effect, while similar effects were seen in maleimide-containing compounds (SPC-160003 and SPC-160004). Although we cannot currently explain how the maleimide group contributes to microtubule stabilization, it is possible that the maleimide group interacts with microtubules by the Michael-type addition reaction. Tubulin is a sulfhydryl-rich heterodimer with 20 cysteine residues, which reacts with sulfhydryl-directed reagents, thus regulating microtubule polymerization or depolymerization 35,36 . This explanation is further supported by our preliminary observation that SPC-160002 blocks irreversibly cell proliferation ( Supplementary Fig. 6). Since our SPC compounds are less potent or effective than paclitaxel, their pharmacological properties are needed to be improved for further developing these compounds as anti-cancer drugs. Further research, including the number, position, or variation of the maleimide substituent, is needed to improve their potencies or efficacies. Altogether, we suggest that SPC-160002 might be an attractive lead molecule for further development of chemotherapeutics with the ability to overcome MDR.   Cell synchronization. To arrest cells in the G 1 /S phase, KB and KBV20C cells at 30% confluence were maintained in growth media supplemented with 2 mM thymidine for 18 h. After washing twice with PBS, the cells were switched to fresh growth media for 9 h. Next, 2 mM thymidine was re-added to the culture and cells were treated with it for 15 h. To arrest cells in the mitotic phase, cells at 40% confluence were incubated with 2 mM thymidine for 24 h. After 24 h, cells were washed twice with PBS and incubated with fresh media for 3 h, followed by addition of and incubation with 100 ng mL -l nocodazole for 12 h. Immunofluorescence. Cells seeded on coverslips were fixed by incubation with pre-chilled 99% methanol at − 20 °C for 30 min. After fixation, cover slips were blocked with PBS-BT (1% BSA, 0.1% Triton X-100 in PBS) for 30 min at room temperature. Subsequently, coverslips were incubated with anti-tubulin antibody diluted in PBS-BT overnight at 4 °C and then with Alexa Fluor 594-conjugated secondary antibodies for 1 h at room temperature. After 4,6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific, Waltham, MA, USA) staining for 3-5 min, the cells were washed three times with PBS and mounted onto glass slides. Staining was evaluated using a Zeiss 710 immunofluorescence microscope (Carl Zeiss, Oberkochen, Germany) and ZEN software.

Western blot analysis.
Rhodamine 123 uptake assay. Rhodamine 123 uptake assay was performed as previously described 22 .
KB and KBV20C cells were pretreated with verapamil or SPC-160002 for 1 h and then incubated with 10 μM rhodamine 123 (Sigma-Aldrich) for 3 h under normal culture conditions. After washing with cold PBS, cells were resuspended in PBS just before analysis. Fluorescence intensity of intracellular rhodamine 123 was determined using flow cytometry. www.nature.com/scientificreports/ glycerol, 2% SDS, 5% 2-mercaptoethanol, 0.1% bromophenol blue). Supernatant and pellet fractions were then loaded for SDS-PAGE, followed with analysis by western blotting.
CSC sphere culture. The surfaces of culture plates were covered using poly-HEMA (poly-2-hydroxyethyl methacrylate, Sigma-Aldrich) dissolved at a concentration of 5 mg mL -1 in 95% ethanol. After drying completely, the cells were seeded and incubated with neurosphere-culture media [DMEM/F12 medium (Invitrogen) containing 20 ng mL -1 epidermal growth factor, basic fibroblast growth factor, and B27 neural supplement (Invitrogen), in addition to 100 units mL -1 penicillin/streptomycin]. After 4 days, the enriched primary CSC spheres were dissociated to single cells with 0.5 mg mL -1 Dispase (Thermo Scientific) to generate secondary spheres for measuring drug effect.
CSC sphere-formation inhibition assay. To assess the effect on CCS sphere-forming ability, the cells (1000 cells per well) derived from the primary CSC cultures were re-seeded into poly-HEMA-coated 96-well plate and treated with or without drug. After 4 days, the spheres were fixed in 1% formalin and the spheres (≥ 100 μm in diameter) were counted under microscope (Nikon, Japan). For comparison, the cytotoxicity to the cells cultured in the attached condition on the plates in same duration were determined using MTS assay. Culture medium supplemented with 10% or 1% FBS was used to minimize the effect of serum proteins in the action of drugs because the defined CSC medium contains minimal proteins. ALDH assay was performed to quantify putative CSC cells that express high ALDH. After the secondary spheres were cultured with or without SPC-160002 for 48 h, the cells were dissociated and analyzed with Aldeflour assay kit (Stem Cell Technology, Cambridge, MA) followed by flow cytometry. A well-defined ALDH inhibitor, N,N-diethylaminobenzaldehyde (DEAB), was used for negative control.