Survivin Suppression Strengthens BZML-induced Mitotic Catastrophe to Overcome Multidrug Resistance by Removing Senescent A549/Taxol Cells


 Background: Mitotic catastrophe (MC) of cancer cells induced by BZML, a novel colchicine-binding site inhibitor, exerts a significant advantage in overcoming multidrug resistance (MDR) in NSCLC. However, the long cellular death process resulting from MC is not beneficial for anticancer treatment. Here, we study the mechanisms underlying MC occurrence and development to promote the development of anticancer therapies based on drug-induced MC.Methods: Cellular senescence was confirmed by morphological features, SA-β-Gal and C12FDG staining. Cell cycle analysis and Hoechst 33342 staining were used to detect MC. Relevant signal transduction pathways and protein location were detected by qRT-PCR, westren blot and immunofluorescence. The half-life of proteins was evaluated using the protein synthesis inhibitor cycloheximide. Flow cytometry, MTT assay, crystal violet staining, Hoechst 33342 staining and cell division detection were performed to determine the effects of BZML and/or YM155 on cell fate. Results: We found that BZML induced p53-dependent cellular senescence in A549/Taxol cells, but not in A549, H1299 and MDA-MB-231 cells. Interestingly, BZML-induced senescence was a secondary effect of MC. In addition, the destruction of the protein-degradation system induced by BZML contributed not only to an increase in p53 protein but also to the accumulation of survivin in the nucleus of A549/Taxol cells. However, in A549 cells, the overexpression of survivin had no effect on apoptosis resistance against BZML and failed to promote BZML-induced MC. The inhibition of survivin did not prevent MC occurrence. Unexpectedly, targeting survivin with YM155 accelerated the death of the MC cells by eliminating senescent cells and strengthening the efficiency of BZML in overcoming the MDR of A549/Taxol cells.Conclusions: Our data suggest that nuclear accumulation of survivin can delay cellular death during MC by promoting the survival of senescent BZML-treated A549/Taxol cells. Further, depending on the dose sequence, combination therapy with YM155 to inhibit survivin might be a new strategy for potentiating BZML-induced MC to overcome MDR during cancer treatment.


Background
Non-small-cell lung cancer (NSCLC) is the leading cause of cancer-related deaths worldwide [1,2] .
Currently, chemotherapy is still an indispensable tool for millions of NSCLC patients worldwide [3,4] .
Unfortunately, the occurrence of multidrug resistance (MDR) is becoming increasingly common in clinical treatment, leading to the failure of possible treatments [5,6] . Therefore, optimizing strategies to enhance therapeutic e cacy and reduce or eliminate MDR, as well as to achieve better prognosis and recovery, are urgently needed.
Currently, most anticancer drugs function by inducing the apoptosis of cells; however, the dysfunction of apoptosis during cancer cell evolution can cause MDR and result in therapy failure [7] . Importantly, a growing body of literature shows that apoptosis is not the sole anticancer mechanism and that the activation of nonapoptotic cell death is a promising strategy for overcoming MDR [8,9] . Mitotic catastrophe (MC) is a newly identi ed type of anticancer mechanism in cancer treatment and MDR prevention and has received more attention in recent years [10,11] . Additionally, a number of studies have shown that MC and senescence are closely related and that cancer cell death by MC is often accompanied by a senescence-like phenotype [12][13][14] . In general, cellular senescence is considered a tumor-suppressive mechanism that complements apoptosis [14] . However, senescent cancer cells may not be e ciently eliminated by immune cells due to the impaired anticancer immune response resulting from the cancer microenvironment or cancer therapy [15,16] . This ineffective immune response makes therapyinduced senescence a double-edged sword, and senescent cancer cells that survive for a long time may cause in ammation and promote cancer recurrence, metastasis and MDR through the senescenceassociated secretory phenotype of cancer cells [15,17] . Therefore, it is essential to intensively study the molecular mechanisms underlying the occurrence and development of MC and to nd an effective way to remove these senescent cells immediately upon anticancer treatment to improve the outcome.
Survivin, encoded by the baculoviral IAP repeat-containing 5 (BIRC5) gene, is a multifunctional protein that typically exists in the cytoplasm and nucleus [18] . Cytoplasmic survivin predominantly mediates antiapoptotic functions, whereas nuclear survivin mediates mitotic function [19,20] . In addition, given its preferential expression in cancer cells and correlation with poor patient prognosis, survivin has been proposed as a biomedical target for cancer therapy [20] . Interestingly, under certain conditions, survivin can be upregulated by various anticancer drugs and exogenous stress conditions, such as UV rays, adriamycin, docetaxel and cisplatin [21][22][23] . Additionally, in addition to its association with apoptosis, the relationship between survivin and other anticancer mechanisms is still unclear. Therefore, it is necessary to clarify the signi cance of increased survivin in cancer treatment, which may provide new ideas and strategies for further cancer treatment. 5-(3, 4, 5-trimethoxybenzoyl)-4-methyl-2-(p-tolyl) imidazol (BZML) is a novel colchicine-binding site inhibitor (CBSI). Our recent studies have demonstrated that nanomolar concentrations of BZML exhibited anticancer activity against various chemosensitive and resistant cancer cells by targeting the colchicinebinding site to inhibit microtubule polymerization in vitro and in vivo [24][25][26] . Interestingly, BZML can force A549 cells to die through apoptosis, whereas it mainly drives A549/Taxol cells to die by p53-independent MC, which overcomes MDR [24] . However, the BZML induction of A549/Taxol cell death takes a long time because it does not cause instantaneous death while eliciting MC, and this long death process does not bene t anticancer treatment. Therefore, in this study, we aimed to focus on the molecular mechanisms underlying MC occurrence and development, which may provide a theoretical basis for optimizing therapeutic strategies to make BZML-induced MC a promising anticancer mechanism to overcome MDR in NSCLC.

Cell culture
Human cancer cell lines A549, H1299 and MDA-MB-231 were purchased from the Shanghai Institute of Cell Resource Center Life Science. A MDR cell line (A549/Taxol) was established from its isogenetic cell line A549 by stepwise selection, and 300 nM paclitaxel was used to maintain MDR in the culture medium [24] . These cells were tested by short tandem repeat analysis, validated to be free of mycoplasma, which were used within 6 months. In addition, the cells were cultured within 25 passages for all experiments. All cells were cultured in RPMI 1640 or DMEM supplemented with 10% fetal bovine serum, penicillin and streptomycin at 37 °C in a humidi ed atmosphere with 5% CO 2 and maintained in a logarithmic growth phase for all experiments.
2.3. Detection of senescence-associated β-galactosidase (SA-β-Gal) Cells were exposed to BZML for indicated time and detected using the cellular senescence assay kit according to the manufacturer's protocol. The percentages of SA-β-Gal positive cells were observed and counted in ve random elds under an optical microscope (Olympus, Tokyo, Japan). Alternatively, SA-β-Gal activity was also measured by C 12 FDG staining for 40 min at 37 °C in the dark, the levels of uorescent product were measured immediately by ow cytometric analysis (FACS) (Accuri Cytometers, Inc., Ann Arbor, MI).

Cell morphology changes
A549/Taxol cells seeded in a 6-well plate were incubated with BZML for indicated time and then photographed by phase contrast microscopy (Olympus, Tokyo, Japan). Cellular area was evaluated by Image J2 software.
Morphological changes of nucleus were detected by Hoechst 33342 staining for 30 min. Cells were photographed by a uorescence microscope (Olympus, Tokyo, Japan) after washing off the unbound dye.

Cellular size and granularity
A549/Taxol cells were seeded into six well plates and then treated with BZML for indicated time. Forward scatter (FSC) and side scatter (SSC) values were measured by FASC (Bectone Dickinson, NJ, USA).

Lysosomal staining
Changes of lysosomes in cells under the in uence of BZML were detected by Lyso-Tracker Red staining. Brie y, A549/Taxol cells were seeded into six well plates. After BZML treatment, living cells were incubated with Lyso-Tracker Red at a concentration of 50 nM for 30 min at 37 °C. Fluorescence images of cells were observed using a uorescence microscope.

MTT assay
Cells were treated with BZML for indicated time and then transferred to MTT solution for 4 h at 37 °C. Subsequently, the medium was discarded and DMSO was added to dissolve the formazan precipitate.
The values of optical density (OD) at 492 nm were measured using a microplate reader (Thermo, Germany).

Crystal violet staining
Cells were seeded into six-well plates at density of 5000 cells/well and treated with BZML for indicated time. After xing with 4% paraformaldehyde, cells were stained with crystal violet and photographed by a digital camera.

Cell division assay
CFDA-SE staining kit was used to track the division of senescent cells. A549/Taxol cells labeled with CFDA-SE for 24 h were stimulated by BZML for indicated time. Then, cells were collected or stained with DAPI. After washing with ice-cold PBS, the uorescence intensity was analyzed and observed by FACS or uorescence microscope.

Cell cycle analysis
Cells incubated with BZML at the indicated time periods were xed with 70% ice-cold ethanol, washed with PBS and stained with PI (50 µg/ml) for 30 min at 4 °C in the dark. Next, the samples were analyzed by FACS.

Apoptosis analysis
Cells apoptosis was detected by Annexin V-FITC/PI double-staining. Brie y, cells were treated with BZML or paclitaxel for indicated time, then collected and stained using Annexin V-FITC/PI staining kit for 30 min at room temperature. Samples were analyzed by FACS.

Immuno uorescence staining
Cells were xed with 4% paraformaldehyde, permeabilized in 0.1% (v/v) Triton X-100, and blocked with 5% BSA in PBS for 30 min. Primary antibodies were incubated overnight at 4 °C, followed by incubation with FITC-or/and TRITC-conjugated secondary antibodies for 2 h at room temperature. The nuclei were stained with DAPI and images were visualized with a confocal microscopy (Nikon C2, Japan).

Preparation of nuclear and cytoplasmic fractions
Nuclear and cytoplasmic proteins were extracted from BZML-treated A549/Taxol cells according to the manufacturer's protocol of the nuclear and cytoplasmic protein extraction kit.

Protein half-life assay
A549/Taxol cells were treated with BZML for 48 h and then followed by treatment with a protein synthesis inhibitor, CHX (100 µg/mL). The expression levels of p53 and survivin protein were detected by western blot at the indicated times after treatment with CHX.

Western blot
The protein samples were prepared in ice-cold RIPA buffer supplemented with protease inhibitor and phosphatase inhibitor. Protein concentrations were measured using BCA protein assay. Cell lysate proteins were separated by SDS-polyacrylamide gels and transferred to a polyvinylidene uoride (PVDF) membrane. Membranes were blocked with 5% BSA or non-fat milk, and probed with primary antibodies overnight at 4 °C, followed by incubation with HRP-conjugated secondary antibodies for 2 h at room temperature. The signals were visualized by enhanced chemiluminescence and densitometry analysis was assessed using Image J2 software.

Statistical analysis
Data are presented as mean ± SD and all experiments were done in triplicate. Comparison among groups was made by One-Way ANOVA Tukey Test of variance and Student t-test of unpaired data using SPSS 22.0 software. The p-value < 0.05 was de ned as signi cant differences.

BZML induced the senescence of the A549/Taxol cells
Cellular senescence is recognized as a potent anticancer mechanism, and it often occurs with MC [17,27] .
In our previous study, we found that BZML overcomes the MDR of A549/Taxol cells by inducing MC [24] . To determine whether the MC induced by BZML is accompanied by a senescence-like phenotype of A549/Taxol cells, we employed SA-β-gal staining to detect acidic β-gal activity at pH 6.0 because this activity is a known characteristic of senescent cells. As shown in Fig. 1a, BZML treatment increased the percentage of SA-β-gal-positive A549/Taxol cells in a time-dependent manner, as particularly evident at 72 and 96 h. In addition, an increase in SA-β-gal activity was also detected by FACS using its uorogenic substrate, C 12 FDG. The percentage of uorescent product-labeled cells, which indicates SA-β-gal activity, reached 70% in the A549/Taxol cells 72 h post-BZML treatment (Fig. 1b).
In addition, the typical morphological features of cellular senescence, such as an enlarged and attened morphology with enhanced cytoplasmic granularity, were observed in the BZML-treated A549/Taxol cells ( Fig. 1c and d). Furthermore, we also con rmed cell enlargement using a FASC dot plot analysis of cellular granularity (SSC) and size (FSC) by observing the cell population. As expected, the cells treated with BZML for 48 h were much larger with more complex granularity than the cells treated with the vehicle control (Fig. 1e). In addition to morphological changes, the lysosomal mass was also signi cantly increased in the BZML-treated A549/Taxol cells (Fig. 1f).

BZML-induced senescence was a secondary effect of MC in the A549/Taxol cells
To further investigate the possible link between cells undergoing MC and senescence, we compared the proportion of polyploid cells and SA-β-gal-positive cells through costaining of the BZML-treated A549/Taxol cells. As shown in Fig. 2a, after BZML treatment for 24 and 48 h, the percentage of polyploid cells was higher than that of the SA-β-gal-positive cells. Importantly, all the SA-β-gal-positive cells (100%) showed polyploid features, but not all the polyploid cells showed staining indicative of positive SA-β-gal activity. In addition, this nding was consistent with our previous studies [24] showing that the proportion of polyploid cells exhibited a gradual increase for 12 h after BZML treatment and reached a peak at 36 and 48 h (Fig. 2b). Interestingly, as shown by the data presented in Fig. 1a and b, the A549/Taxol cells did not undergo senescence until 48 h post-BZML treatment. A time course analysis demonstrated that the MC occurred relatively early in this process and approximately parallel with cellular senescence in later development.

BZML-induced senescence was an irreversible anticancer mechanism of A549/Taxol cells
To determine whether cellular senescence secondary to BZML-induced MC is an adaptive survival response or contributes to anticancer effects, we investigated the nal fate of BZML-treated A549/Taxol cells after treatment withdrawal (Fig. 3a). As shown in Fig. 3b and c, among the A549/Taxol cells pretreated with BZML for 48 h and then transferred to drug-free medium for another 24 or 48 h of incubation, the percentage of SA-β-gal-positive cells continued to increase in a time-dependent manner. Importantly, crystal violet staining showed that treatment with BZML for 24, 48 and 72 h did not further increase the number of A549/Taxol cells compared with the number of untreated cells when the culture time was extended (Fig. 3d). Furthermore, even when the A549/Taxol cells were washed to remove BZML after 48 h of coincubation, the cells did not continue to proliferate in the drug-free medium for the following 24 and 48 h of cultivation; in contrast, the number of cells decreased gradually (Fig. 3e). Importantly, the MTT assay indicated found that the OD value of the BZML-treated A549/Taxol cells failed to increase after treatment withdrawal and that the decrease continued for the indicated times that the cells were cultured in drug-free medium (Fig. 3f).
Additionally, CFDA-SE staining was performed to track the division of the BZML-treated senescent A549/Taxol cells because the uorescence intensity of the CFSE converted from CFDA-SE by an esterase is reduced by one-half with each division of tested cells [28] . As shown in Fig. 3g, BZML signi cantly inhibited the division of the senescent A549/Taxol cells, as indicated by the uorescence failing to shift further leftward in the FASC plot. Similarly, the uorescence intensity of CFSE in the BZML-treated A549/Taxol cells was much higher than that of the untreated cell groups both at 48 h and 72 h, as indicated by the uorescence microscopy analysis (Fig. 3h). This suggests that BZML-induced senescence exerts an irreversible anticancer effect that prevents the MDR of A549/Taxol cells, even after it is withdrawn from the medium during transient incubation.

BZML-induced senescence of the A549/Taxol cells was dependent on p53
The p53-p21 and p16 INK4α -Rb pathways are believed to have important contributions to the regulation of cellular senescence [29,30] . As shown in Fig. 4a, BZML induced an evident increase in p53 expression within 24 h of coincubation, and the expression of p53 in the A549/Taxol cells reached a peak after 48 h of treatment and was sustained for at least 72 h, ndings in line with previous ndings [24] . In addition, the expression of p21 and Rb was increased, but p-Rb expression was decreased in the BZML-treated A549/Taxol cells.
To further investigate the functional involvement of p53 in BZML-induced senescence, the p53 inhibitor pi thrin-α and p53-siRNAs were used to abrogate the expression of p53 in BZML-treated A549/Taxol cells (Supplementary Fig. S4a and b). As shown in Fig. 1a, 4b and S5a, both pi thrin-α and p53-siRNAs signi cantly decreased the percentage of SA-β-gal-positive A549/Taxol cells at different time points after BZML treatment. Similarly, the C 12 FDG staining results showed that the increase in SA-β-gal-positive cells caused by BZML was also signi cantly downregulated by pi thrin-α and p53-siRNAs pretreatment of the A549/Taxol cells (Fig. 4c and d). These data suggest that BZML-induced cellular senescence is dependent on p53 in A549/Taxol cells.

BZML activated p53 through multiple mechanisms in the A549/Taxol cells
Recent studies have proven that p53 function is tightly regulated by its cellular localization, and p53 is synthesized in the cytoplasm while exerting its transcriptional effect on downstream targets responding to cellular stress only after being transported to the nucleus [31] . The results from western blot analysis and immuno uorescence staining showed that in the BZML-treated A549/Taxol cells, the expression of p53 increased only slightly in the cytoplasm but was signi cantly upregulated in a time-dependent manner in the nucleus (Fig. 5a and Supplementary Fig. S5b).
In addition, BZML downregulated the expression of MDM2, which is a negative regulatory protein of p53 (Fig. 5b). Further, the expression of proteasome 20S core subunits and PMSA6, which are important components of the ubiquitin-proteasome system, was also signi cantly decreased in the BZML-treated A549/Taxol cells (Fig. 5b). The data above indicated that the protein degradation system might be destroyed by BZML in the A549/Taxol cells. Next, to further clarify how BZML affects p53 protein stability, we measured the half-life of the p53 protein using the protein synthesis inhibitor CHX. Notably, the half-life of p53 in the untreated cells was found to be less than 30 min, while it was extended approximately 2-fold by BZML in the A549/Taxol cells (Fig. 5c). Moreover, the expression of p53 at the mRNA level was also increased after BZML treatment (Fig. 5d).
3.6. Survivin mainly accumulated in the nucleus of the BZML-treated A549/Taxol cells It is generally known that survivin is also a short-lived protein [32] . Importantly, survivin may mediate MDR and modulate cellular mitosis; therefore, we further investigated the expression of survivin in apoptotic A549 cells and in MC A549/Taxol cells treated with BZML. As shown in Fig. 6a and b, the intrinsic expression of survivin was very low, and there was no signi cant difference between the untreated A549 and A549/Taxol cells at either the protein or mRNA level. Interestingly, BZML signi cantly decreased survivin expression in the A549 cells from 0 to 96 h of treatment. However, in contrast, the expression of survivin in the BZML-treated A549/Taxol cells was signi cantly increased from 0 to 48 h and then maintained at a high level for 72 h after the transient increase; subsequently, at 96 h, it rapidly decreased; see Fig. 6c. Furthermore, the immuno uorescence staining also showed that BZML caused a signi cant increase in survivin expression in the A549/Taxol cells but not in the A549 cells (Fig. 6d). Notably the increased survivin expression was mainly found in the nucleus in the BZML-treated A549/Taxol cells ( Fig. 6d and 5b). Interestingly, the qRT-PCR results showed that the expression of survivin was only slightly increased at the mRNA level (Fig. 6e). Furthermore, to con rm whether the increase in survivin induced by BZML was attributable to the increase in protein stability in the A549/Taxol cells, CHX was used again in this study. As expected, the half-life of survivin was also signi cantly prolonged by BZML, suggesting that BZML improves the stability of the survivin protein in A549/Taxol cells (Fig. 5d).

Overexpression of survivin failed to promote BZML to induce MC in A549 cells
Since survivin is an important anti-apoptotic protein and its speci cally increased expression was detected in cells showing BZML-induced MC, we further evaluated the function of survivin on cells with BZML-induced MC. Here, A549 cells overexpressing survivin protein were used (Fig. 7a). Notably, transfected exogenous survivin was expressed in both the cytoplasm and nucleus (Supplementary Fig.   S5c). In addition, the MTT assay showed that the A549 cells overexpressing survivin exhibited drug resistance against paclitaxel but not against BZML (Fig. 7b). Furthermore, at the normal anticancer dose (15 nM), pacitaxel failed to induce apoptosis in the A549 cells overexpressing survivin (Fig. 7c). In contrast, the rate of apoptosis induced by BZML was not changed in the A549 cells overexpressing survivin (Fig. 7d). Importantly, the overexpression of survivin signi cantly decreased the percentage of cells in the sub-G1 phase, which had been increased by paclitaxel, but failed to promote BZML induction of MC in the A549 cells (Fig. 7e). Interestingly, with increases in dose and time, paclitaxel at a high dose also induced MC in the A549/Taxol cells (Supplementary Fig. S6). Therefore, we conclude that the nuclear accumulation of survivin is the result of MC rather than the result of BZML-induced MC.
3.8. YM155 treatment potentiated the effect of BZML in overcoming the MDR of the A549/Taxol cells To further de ne the relationship between ultimate cell fate and BZML-induced survivin expression, we used a speci c inhibitor (YM155) or survivin-siRNAs to inhibit BZML-induced survivin in A549/Taxol cells ( Fig. 8a and Supplementary Fig. S7). Unexpectedly, pretreatment with YM155 did not affect the proportion of polyploid and sub-G1 A549/Taxol cells treated with BZML for 48 h (Fig. 8b). Similarly, survivin-siRNA transfection did not affect the cell cycle distribution of the A549/Taxol cells treated with BZML (Fig. 8c). Furthermore, Hoechst 33342 staining showed that pretreatment with YM155 or survivin-siRNAs had no effect on the number of multinucleated A549/Taxol cells treated with BZML for 48 h (Fig. 8d).
However, the proportion of polyploid cells was signi cantly decreased, a nding associated with an increase in sub-G1 A549/Taxol cells cotreated with BZML and YM155 for 72 h (Fig. 8e). In addition, after 48 h of incubation with BZML, the A549/Taxol cells were washed to remove BZML and then were added to YM155-containing or drug-free medium for another 24 h of incubation. Compared to that of the cells with BZML treatment alone for 48 h, the proportion of polyploid cells was signi cantly decreased, and the proportion of sub-G1 cells in the drug-free medium was increased; however, the percentage of sub-G1 cells was further increased after 24 h of culture in the presence of YM155 (Fig. 8f). Similarly, the MTT assay demonstrated that YM155, as a cotreatment or sequential treatment, enhanced the inhibitory effect of the 72-h BZML exposure on the proliferation of A549/Taxol cells ( Fig. 8g and h). Furthermore, the percentage of SA-β-gal-positive cells was signi cantly decreased by sequence-dependent combination therapy with BZML and YM155 ( Fig. 8i and g). Therefore, these results indicate that the inhibition of survivin cannot prevent MC but accelerates the death of cells undergoing MC by removing senescent cells in time.

Discussion
Despite considerable research efforts aimed at developing novel therapeutic approaches against NSCLC, including molecularly targeted therapy and immunotherapy, the frequency of MDR severely restricts the effectiveness of various anticancer drugs [3,4] . Currently, apoptosis tolerance-mediated MDR has been considered the most common and complex drug-resistant mechanism [5,33] . However, in addition to apoptosis, anticancer drugs can also trigger other anticancer mechanisms, such as autophagy, MC, ferroptosis, pyroptosis, necrosis and senescence, to inhibit the proliferation of cancer cells and kill them by avoiding the apoptosis-resistant pathway [11,17,34,35] . In our previous studies, we con rmed that BZML has potent anticancer activity and overcomes MDR by inducing MC in A549/Taxol cells, an MDR cell line [24] . However, a detailed and clear explanation of the mechanisms underlying MC occurrence and development in BZML-treated A549/Taxol cells was lacking.
Cellular senescence is a permanent state of cell cycle arrest and has been considered a novel anticancer mechanism [14,17] . In our study, BZML treatment induced the cellular senescence in A549/Taxol cells, showing a signi cantly increased proportion of SA-β-gal-positive cells and some other typical features of cellular senescence, such as attened morphology, increased size and granularity and an elevated lysosomal mass in the cells. To further investigate whether these senescent cancer cells were dormant cells that can contribute to cancer redevelopment with the decrease in the concentrations of therapeutic drugs, the BZML-containing medium was replaced with drug-free medium at 48 h post-BZML treatment, and the results from the SA-β-gal staining and microscopy assays of the size and shape of the cells demonstrated that the A549/Taxol cells underwent senescence even after BZML withdrawal. Importantly, the senescent cancer cells neither reentered the normal cell cycle nor divided; thus, they lost normal proliferative potential, and the increase in cell death was signi cant after the withdrawal of the BZML treatment. Therefore, given the important role of senescence in cancer treatment, our ndings support the possibility that BZML-induced senescence might act as an anticancer mechanism against MDR in NSCLC.
MC is a newly discovered form of tumor suppression that differs from other cell death modes and is characterized by unique nuclear alterations, such as multi and/or micronucleation [24,36,37] . Interestingly, a number of studies reported that senescence is closely related to MC and that cells undergoing senescence may undergo polyploidization and/or become multinucleated [37,38] . In contrast, some studies showed that C85 cells senescence induced by methotrexate did not undergo polyploidization or multinucleation [39] . Therefore, to elaborate the relationship between BZML-induced MC and the senescence-like phenotype of A549/Taxol cells, we rst performed a time kinetic study to assess the dynamic changes of key parameters during MC and senescence. Notably, BZML-induced MC, indicated by the appearance of polyploid and/or multinucleated cells, occurred as early as 12 h, while an increase in SA-β-gal-positive cells was detected 48 h after BZML treatment. Analysis of these 2 events over the experimental time course indicated that BZML-induced MC occurred relatively early in the A549/Taxol cells. Moreover, SA-β-gal and Hoechst 33342 double staining revealed that almost all the SA-β-galpositive cells were multi-and/or micronucleated, but not all the polyploid cells were stained positive for SA-β-gal activity. Interestingly, our previous study proved that BZML-induced MC was independent of p53 [24] . As expected, regardless of p53 status, BZML did not induce MC in A549, H1299 or MDA-MB-231 cells, which are p53-wild-type, p53-null and p53-mutant cells, respectively. Importantly, these BZML-treated cells did not undergo senescence. In addition, in contrast to traditional cellular senescence, which is often accompanied by G0/G1 phase arrest [37] , in our study, a substantial level of polyploidy was detected in the senescent A549/Taxol cells after BZML treatment. Together, these data strongly support the supposition that BZML-induced senescence is downstream of MC and acts as an important phenotype associated with MC occurrence and development in A549/Taxol cells.
Interestingly, elevated ROS levels have been widely accepted as a major trigger of cellular senescence [37,40] , and BZML was con rmed in our previous study to cause ROS generation in a time-dependent manner, but ROS are not inducers of MC occurrence or development in A549/Taxol cells [25] . In addition, in this study, we also found that NAC did not reverse the BZML-induced senescence-like phenotype of the A549/Taxol cells (Supplementary Fig. S8a and b), suggesting that the senescence-like phenotype is secondary to the BZML-induced MC in A549/Taxol cells and may be attributed to mechanisms other than ROS. BZML-induced MC is independent of p53, but the expression of p53 was increased in BZML-treated A549/Taxol cells in a time-dependent manner. Given these considerations, p53-p21 pathway activation represents the trigger of senescence, and the p16 INK4α -Rb pathway is involved in maintaining senescence [37,41] . As expected, the expression of p21 was gradually increased, followed by an increase in Rb expression and a decrease in Rb phosphorylation in the BZML-treated A549/Taxol cells. Moreover, our study also demonstrated that pi thrin-α and p53-siRNAs can reverse the BZML-induced senescence-like phenotype, indicating that p53 might play a decisive role at the beginning of senescence in BZML-treated A549/Taxol cells.
Under normal conditions, p53 is a short-lived protein that shuttles between the nucleus and the cytoplasm in a cell cycle-speci c manner [31] . Relocation of p53 to the nucleus in response to cellular stress is a contributor to the inhibition of the growth of cancer cells [42] . In this study, the expression of p53 was signi cantly increased in the BZML-treated A549/Taxol cells at both the mRNA and protein levels. Importantly, the results from the western blot analysis and uorescence microscopy assay all indicated that BZML treatment increased p53 levels, mainly in the nucleus. In addition, increasing attention has been focused on MDM2 because it acts as an oncogene that negatively regulates the functions of the tumor suppressor p53 by inhibiting transcriptional activity and accelerating the degradation of p53 [43] .
Interestingly, several reports have demonstrated that some CBSIs, such as combretastatin A-4 and SQ, exhibit a potential MDM2 inhibitory effect in breast cancer cells [44] . Furthermore, in this study, BZML also signi cantly decreased the expression of MDM2 in the A549/Taxol cells. Additionally, it cannot be ignored that the ubiquitin-proteasome system (UPS) functions as the primary route of degradation for thousands of short-lived proteins [45,46] . Here, its important components, proteasome 20S core subunits and PMSA6, were signi cantly downregulated in the BZML-treated A549/Taxol cells, indicating that BZML might cause the destruction of the UPS. In fact, the half-life of p53 was signi cantly prolonged by BZML in the A549/Taxol cells. Therefore, these results suggest that BZML can activate p53 via multiple mechanisms in the A549/Taxol cells and that the destruction of the UPS also contributes to the increase in p53 protein, at least in part.
In cancer cells, survivin located in the cytoplasm plays indispensable roles in cell proliferation and apoptosis inhibition [32] . Interestingly, in this study, BZML treatment increased survivin mainly in the nucleus of the A549/Taxol cells. Given these considerations, survivin is also a short-lived protein and can be polyubiquitylated and undergo proteasomal destruction [47] . Furthermore, nuclear survivin is signi cantly more unstable [20] . Interestingly, survivin was expressed at low levels under unstressed conditions; however, in response to BZML-induced MC, its stability in the nucleus was quickly increased, and its half-life was signi cantly prolonged at the same time. This outcome suggests that the increase in survivin in the nucleus may also be attributed to the destruction of the UPS in BZML-treated A549/Taxol cells. Notably, chemotherapy and radiation therapy can increase survivin expression, which is attributed to mitotic arrest at the G2/M phase and to the augmented stability of the survivin protein through its phosphorylation at Thr34 by a cdc2/cyclin B1 complex [21,22,48] . In the present study, the expression of survivin was also slightly increased at the mRNA level. In addition, our previous study showed that BZML treatment caused a transient decrease in cyclin B1 within 24 h, followed by a signi cant and stable increase in cyclin B1 expression during BZML-induced MC [24] . Remarkably, in this study, the expression of survivin and cyclin B1 changed in parallel during this process. Importantly, the overexpression of survivin neither mediated the apoptosis resistance against BZML nor promoted BZML to induce MC in the A549 cells, suggesting that survivin is not an inducer to the BZML induction of MC in A549/Taxol cells, at least in some settings. Therefore, we speculated that the increase in survivin in the nucleus may result from the increase in cyclin B1 and the destruction of the UPS, and the nuclear expression of survivin is an important biological phenotype associated with MC occurrence and development.
Additionally, some studies have demonstrated that MC and senescence play important and paradoxical roles in the process of cancer treatment [36,37] . In some circumstances, senescent cells cannot be cleared at the time of treatment in time, which may result in the resumed division and apoptosis of drug-resistant cancer cells, and eventually potentiate cancer progression [49,50] . Therefore, the development new strategies to remove cells undergoing MC and senescence in a timely manner after the initial anticancer response are urgent. Our data showed that survivin was not required for a cell to undergo MC, but once cellular senescence, secondary to MC, was induced, the upregulation of survivin may provide additional vulnerability to and critical opportunities for sequentially applied therapies. Interestingly, in this study, there was no synergistic effect of YM155 treatment in response to BZML-induced MC during short periods, but the inhibition of survivin by YM155 signi cantly enhanced the e ciency of BZML in overcoming the MDR of the A549/Taxol cells after 72 h of treatment. Importantly, after BZML-induced MC associated with senescence, dose-sequence-dependent combination therapy with YM155 exhibited a synergistic lethal effect. This suggests that the synergistic effectiveness of BZML and YM155 occurs in the context of the MC-associated senescence induced by BZML. Combinational treatment is typically employed to achieve a better response rate than monotherapy and is based on the use of drugs with different cytotoxicity-inducing mechanisms [22,23] . Herein, sequential monotherapy not only retained the synergistic effect but also further reduced the toxicity of the anticancer drugs in clinical application. Due to the short half-life of YM155 in the human body [23] , we proposed a ''one-two punch'' approach to cancer treatment based on our ndings. That is, at the beginning of cancer treatment, MC-associated senescence is selectively induced in cancer cells; subsequently, in consecutive therapy, these senescent cells are killed by dose-sequence-dependent molecularly targeted drugs. Thus, MC-associated irreversible senescence induced by BZML may provide a treatment window for the opportunistic cell elimination using synergistic YM155. In addition, considering that the aberrant activation of an oncogene can cause cellular senescence, survivin, as an oncogene, may also promote the phosphorylation of Rb to maintain the senescence phenotype [17,20,37] . Subsequently, we tried to clarify whether the nuclear accumulation of survivin was critical for the senescence caused by BZML-induced MC. Notably, YM155 signi cantly decreased the percentage of SA-β-gal-positive cells among the BZML-treated A549/Taxol cells, suggesting that the nuclear accumulation of survivin may mediate self-protection by inducing a senescence-like phenotype during MC. Our data provide strong evidence for targeting survivin as a strategy for enhancing the e ciency of BZML because it induces MC to overcome the MDR of A549/Taxol cells.

Conclusion
BZML-induced senescence is a secondary effect of MC in A549/Taxol cells, and it also exhibits a potent and irreversible anticancer effect against MDR in NSCLC. Furthermore, cellular senescence is dependent on p53, and the activation of p53 is involved in the regulatory effects on its transcription, translation, and posttranslational modi cation in BZML-treated A549/Taxol cells. Interestingly, the destruction of the protein degradation system not only contributes to the increase in p53 protein but also promotes survivin accumulation in the nucleus of the BZML-treated A549/Taxol cells. However, the overexpression of survivin cannot lead to apoptosis resistance against BZML and fails to promote BZML to induce MC in the A549 cells. Importantly, the inhibition of survivin cannot prevent MC occurrence but accelerates cell death during MC by promptly removing the senescent cells, potentiates the effect of BZML in overcoming the MDR of A549/Taxol cells (Fig. 9). Therefore, our ndings not only reveal the relationships between BZML-induced MC and senescence but also provide novel insights into the mechanisms of MC occurrence and development. Moreover, they will likely open a new area for drug discovery research through the induction of MC.

Availability of data and materials
All data generated or analyzed during this study are available from the corresponding author upon reasonable request.

Ethics approval and consent to participate
The study was conducted in accordance with the Declaration of Helsinki principles. It was approved by the Ethics Committee of the Jiangsu Cancer Hospital.

Consent for publication
Not applicable.   CFDA-SE staining of A549/Taxol cells treated with BZML (60 nM) for the indicated time was performed by FACS and uorescence microscope, respectively (scale bar = 50 μm). **p < 0.01 vs control.