TH588 and Low-Dose Nocodazole Impair Chromosome Congression by Suppressing Microtubule Turnover within the Mitotic Spindle

Simple Summary A promising anti-cancer compound TH588 has been recently identified as a microtubule-targeting agent that inhibits tubulin polymerization in vitro and interferes with microtubule dynamics in interphase cells. Although it was shown to arrest cells in mitosis, its effect on microtubule dynamics in dividing cells remained unknown. By analyzing microtubule dynamics in living cells treated with either TH588 or low-dose nocodazole, we revealed that both of these drugs stabilize microtubules within the mitotic spindle, leading to premature formation of kinetochore-microtubule end-on attachments on uncongressed chromosomes. This causes mitotic arrest, ultimately resulting in cell death or cell division with uncongressed chromosomes. Both of these cell fates could contribute to the selective effect associated with the activity of TH588 in cancer cells. Abstract Microtubule-targeting agents (MTAs) have been used for decades to treat different hematologic and solid cancers. The mode of action of these drugs mainly relies on their ability to bind tubulin subunits and/or microtubules and interfere with microtubule dynamics. In addition to its MTH1-inhibiting activity, TH588 has been recently identified as an MTA, whose anticancer properties were shown to largely depend on its microtubule-targeting ability. Although TH588 inhibited tubulin polymerization in vitro and reduced microtubule plus-end mobility in interphase cells, its effect on microtubule dynamics within the mitotic spindle of dividing cells remained unknown. Here, we performed an in-depth analysis of the impact of TH588 on spindle-associated microtubules and compared it to the effect of low-dose nocodazole. We show that both treatments reduce microtubule turnover within the mitotic spindle. This microtubule-stabilizing effect leads to premature formation of kinetochore-microtubule end-on attachments on uncongressed chromosomes, which consequently cannot be transported to the cell equator, thereby delaying cell division and leading to cell death or division with uncongressed chromosomes.


Introduction
Microtubules (MTs) are filamentous tubular structures, typically consisting of 13 laterally associated protofilaments that are built from αand β-tubulin heterodimers. MTs are highly dynamic, owing to the so-called "dynamic instability" [1][2][3][4][5] that is characterized by constant switching between persistent phases of elongation and shortening. The switch from shortening to elongation is called "rescue" and depends on the formation of a protective "cap" resulting from incorporation of GTP-tubulin into the growing MT end. Within the MT lattice, GTP-tubulin is transformed into GDP-tubulin via hydrolysis. The switch

Cell Culture
U2OS and HeLa cell lines were grown in DMEM (Gibco) and RPE-1 in DMEM/F12 media (Gibco) supplemented with 10% FBS (Invitrogen, Waltham, MA, USA) at 37 • C with 5% CO 2 . Cells were immediately fixed with ice-cold methanol at indicated time points and MTs and centrin were stained using anti-alpha-tubulin and anti-centrin antibodies, respectively. Images were acquired using Zeiss AxioObserver Z1 wide-field microscope (63× Plan-Apochromatic oil differential interference contrast objective lens, 1.4 NA) equipped with Metal halide arc lamp and Axiocam 702 mono CMOS camera and Zen 3.0 blue edition software (Carl Zeiss, Inc., Oberkochen, Germany). Representative images were acquired using LSM700 confocal microscope (Carl Zeiss Inc., Oberkochen, Germany) mounted on a Zeiss-Axio imager Z1 equipped with plan-apochromat 63×/1.40 oil DIC M27 objective (Carl Zeiss, Inc., Oberkochen, Germany) and Zen 2008 software (Carl Zeiss, Inc., Oberkochen, Germany). For quantification of MT nucleation, alpha-tubulin intensity at 1 min time-point was selected by a circular ROI covering the centrosome, followed by subtraction of the cytoplasmic background. The resulting intensity values were normalized to the median of control. Centrin was used to define the centrosomal position.

Immunofluorescence Imaging
For quantification of MT populations in mitotic cells, U2OS parental cells were arrested in metaphase using APC/C inhibitors Apcin (Sigma-Aldrich) (20 µM) and proTAME (BostonBiochem, Cambridge, MA, USA) (10 µM) for 2 h followed by treatment with indicated drugs or DMSO for 1 h. Following the treatment, the cells were fixed with 4% paraformaldehyde in PHEM buffer for 20 min at 37 • C as described before [27,28] and MTs were stained using anti-alpha-tubulin antibody. Images were acquired using a Plan-Apochromat 63×/1.4 NA oil objective with differential interference contrast mounted on an inverted Zeiss Axio Observer Z1 microscope (Marianas Imaging Workstation, 3i-Intelligent Imaging Innovations, Inc., Denver, CO, USA) equipped with an iXon Ultra 888 EM-CCD camera (Andor Technology, Belfast, UK). Astral and spindle MT intensities were quantified using ImageJ (National Institute of Health, Bethesda, MD, USA) as described before [26].
To study the localization pattern of ZW10 and kinetochore-MT attachments, U2OS cells were treated with indicated drugs for 1 h, followed by fixation using ice-cold methanol and staining using target specific primary antibodies and Alexa Fluor conjugated secondary antibodies. Images were acquired using LSM700 confocal microscope described above. Line intensity profiles were obtained using plot profile function in ImageJ.

Live Cell Imaging
Live-cell time-lapse imaging was performed as described before [29,30]. Briefly, cells were cultured in 35 mm glass-bottomed dishes (14 mm, No. 1.5, MatTek Corporation, Ashland, MA, USA) and imaging was performed in an environment controlled chamber (37 • C with controlled humidity and 5% CO 2 supply), using a Plan-Apochromat DIC 63×/1.4 NA oil objective mounted on an inverted Zeiss Axio Observer Z1 microscope (Marianas Imaging Workstation from 3i-Intelligent Imaging and Innovations Inc., Denver, CO, USA), equipped with a CSU-X1 spinning-disk confocal head (Yokogawa Corporation of America, Sugar Land, TX, USA) and four laser lines (405 nm, 488 nm, 561 nm and 640 nm). Images were acquired using an iXon Ultra 888 EM-CCD camera (Andor Technology).
To observe MT dynamics during interphase, U2OS EB1-GFP cells [31] (gift from P. Draber, Institute of Molecular Genetics of the Czech Academy of Sciences-IMG ASCR, Prague, Czech Republic) were treated with indicated drugs 15 min before imaging. To study MT dynamics during mitosis, cells were arrested in metaphase using anaphasepromoting complex/cyclosome (APC/C) inhibitors Apcin (20 µM) and a cell permeable tosyl-L-arginine methyl ester (proTAME) (10 µM) for 2 h before the addition of the drugs. DNA was labelled by adding 20 nM SiR-DNA (Spirochrome) 1 h prior to live-cell imaging. Cells were imaged at 2 frames per second rate for 1-2 min. About 2-5 EB1 comets per cell were tracked manually using ImageJ.
MT flux and turnover rates in mitotic spindles were imaged and quantified using U2OS-PA-GFP/mCherry-α-tubulin cells [35] (gift from R. Medema, Netherlands Cancer Institute-NKI, Amsterdam, Netherlands) as described before [28,29]. The cells were arrested in metaphase using 5 µM MG132 for 1 h followed by treatment with indicated drugs for 15 min.

Image Processing and Statistical Analysis
All graphs and statistical analysis were generated in GraphPad Prism 8.0. Data points were tested for normality using D'Agostino & Pearson test. Accordingly, statistical significance was determined by Student's t-test (unpaired, two-tailed; normal distribution) or Mann-Whitney U-test (unpaired, two-tailed; no normal distribution). F-test was used to compare variances and Welch's correction was employed when variances were not equal. Details of the statistical significance and n values for each conditions can be found in the figures and figure legends.

TH588 Disrupts In Vitro MT Polymerization and in Cellulo MT Nucleation and Reduces MT Dynamics in Interphase Cells
In order to test the direct effect of TH588 on MTs, we first performed a turbiditybased in vitro MT polymerization assay in the presence of TH588 at varying concentrations ( Figure 1A). In agreement with previous reports [15,22], our data show a strong effect of TH588 on MT dynamics, wherein it severely affected MT polymerization in a concentrationdependent manner.

TH588 Disrupts In Vitro MT Polymerization and in Cellulo MT Nucleation and Reduces MT Dynamics in Interphase Cells
In order to test the direct effect of TH588 on MTs, we first performed a turbiditybased in vitro MT polymerization assay in the presence of TH588 at varying concentrations ( Figure 1A). In agreement with previous reports [15,22], our data show a strong effect of TH588 on MT dynamics, wherein it severely affected MT polymerization in a concentration-dependent manner. To analyze the effect of TH588 on MT dynamics in cellulo, we performed spinningdisk confocal microscopy-based live-cell imaging of human osteosarcoma U2OS cells stably expressing EB1-GFP, a MT-plus-end tracking protein that localizes to the tips of growing MTs. Because structural [22] and in vitro [15,22] data suggest a similar MTdepolymerizing mode of action between TH588 and nocodazole, we compared the effect of these two drugs on MT dynamics. Tracking of EB1-GFP comets in interphase cells upon individual treatments with either TH588 or low-dose nocodazole revealed a drastic effect of both drugs on the growth of MTs, represented by the distance traversed by EB1-GFP (4.98 ± 1.18 µm with 5 µM TH588 and 1.07 ± 0.53 µm with 100 nM nocodazole, compared to 13.81 ± 3.18 µm in DMSO control) and its lifetime (9.27 ± 1.56 s with 5 µM TH588 and 3.27 ± 0.99 s with 100 nM nocodazole, compared to 20.46 ± 4.25 s in DMSO control) ( Figure 1B-D, Video S1). In addition, we also observed a reduction in the EB1-GFP comet velocity (0.53 ± 0.06 µm/s with 5 µM TH588 and 0.32 ± 0.09 µm/s with 100 nM nocodazole, compared to 0.68 ± 0.1 µm/s in DMSO control), coupled with spending a significantly higher percentage of lifetime in the paused state (12.77 ± 3.8% with 5 µM TH588 and 28.19 ± 16.44% with 100 nM nocodazole, compared to 7.57 ± 2.73% in DMSO control) ( Figure 1E,F, Video S1).
Taken together, these data suggest that both TH588 and low dose of nocodazole reduce MT dynamics in cellulo, showing the features characteristic for the MT dynamicssuppressing MTAs. This is in agreement with earlier observation that TH588 reduces interphase MT dynamics in H460 human large-cell lung carcinoma cells [21], as well as with the fact that, besides being a well-established MT-depolymerizing agent, nocodazole at low concentrations has a MT-stabilizing effect [36,37].
Next, we tested the impact of TH588 on MT nucleation. Similar to low-dose nocodazole, TH588 substantially reduced centrosome-mediated re-nucleation after cold-induced depolymerization of MTs ( Figure 1G,H). Notably, 20 min after re-initiation of MT nucleation, the MT network was restored to control levels in both TH588 and low-dose nocodazole-treated cells ( Figure 1G).

TH588 Induces Severe Chromosome Congression Problems That Lead to Mitotic Arrest, Followed by the Cell Death or Erroneous Cell Division
TH588 impaired chromosome congression to the spindle equator in H460 cells, inducing mitotic arrest [21]. To examine whether TH588 has the same impact on mitosis in two other cancer cell lines (U2OS and HeLa), as well as in non-tumor, hTERT-immortalized, human retinal pigment epithelial (RPE-1) cells, we performed live cell imaging and monitored mitotic duration, chromosome movement, and cell fate.
First, we compared the effects of TH588 and low-dose nocodazole treatments in U2OS cells stably expressing histone 2B (H2B)-GFP and mCherry-α-tubulin. Whereas control cells formed a proper bipolar spindle, congressed the chromosomes to the metaphase plate and divided normally (Video S2), both TH588 and nocodazole treatments drastically affected mitotic progression. About 5 µM TH588 induced severe chromosome congression problems, represented by multiple chromosomes getting trapped around the spindle poles, unable to reach the metaphase plate (Figure 2A,B, Video S3). This effect of TH588 on chromosome congression and occasional spindle-collapsing strongly resembled the mitotic phenotype observed upon treating the U2OS cells with 20 nM and 100 nM nocodazole (Figure 2A,B, Video S4). We further analyzed the fate of mitotic cells during 6 h after addition of 5 µM TH588. While all DMSO-treated control cells divided normally, we observed that 73.47% of the treated cells were arrested in mitosis due to the problems in chromosome congression, whereas 14.26% of cells entered anaphase with uncongressed chromosomes and 8.16% of cells died ( Figure 2C). Similar cell fates were observed upon the treatment with nocodazole, where 20 nM concentration showed a milder phenotype (20.97% of cells arrested in mitosis with uncongressed chromosomes, 25.81% of cells divided with uncongressed chromosomes and 4.84% of cells died), compared to the cells treated with 100 nM (82.54% of cells arrested in mitosis with uncongressed chromosomes, 6.35% of cells divided with uncongressed chromosomes and 11.12% of cells died) ( Figure 2C).
A very strong mitotic effect was also observed in HeLa cells stably expressing H2B-mCherry treated with either TH588 or 100 nM nocodazole. During 6 h of imaging after the addition of each drug, 100% of HeLa cells stayed arrested in mitosis with uncongressed chromosomes ( Figure S1A-C, Video S5). Interestingly, although TH588-treated RPE-1 cells stably expressing H2B-GFP and RFP-α-tubulin showed similar problems in chromosome congression (68% of cells arrested in mitosis with uncongressed chromosomes, 8% of cells divided with uncongressed chromosomes and 4% of cells died) during 6 h of imaging after the addition of drug, a larger fraction of RPE-1 cells managed to accurately congress all chromosomes and divide normally (20%), compared to U2OS and HeLa cells, as well as to nocodazole-treated RPE-1 cells (100% of cells arrested in mitosis with uncongressed chromosomes) ( Figure S2A-C, Video S6). Although TH588 reduced MT plus-end mobility in interphase cells, its effect on MT dynamics within the mitotic spindle of dividing cells remains elusive. To address this, we performed live-cell imaging of U2OS cells stably expressing EB1-GFP and monitored the impact of TH588 and low-dose nocodazole on highly dynamic astral MTs. In order to bypass the TH588 and nocodazole-associated severe congression problems that could otherwise interfere with astral MTs, we used anaphase promoting complex/cyclosome (APC/C) inhibitors to arrest cells in metaphase prior to the drug treatments. By quantifying the number of EB1 comets, we observed a severe reduction in astral MT dynamics upon the treatment with TH588 or low-dose nocodazole ( Figure 3A,B, Video S7). In agreement with these data, the immunostaining of metaphase-arrested cells showed a near-total loss of astral MT population upon application of either of these two drugs ( Figure 3C,D). Interestingly, although TH588 and low-dose nocodazole fully disrupted astral MTs, the quantification of fluorescence intensity upon immunostaining revealed a significant increase in the spindle MTs signal ( Figure 3C,D), suggesting that TH588 and low-dose nocodazole have a MT-stabilizing effect on longer-lived MTs.  number of EB1 comets, we observed a severe reduction in astral MT dynamics upon the treatment with TH588 or low-dose nocodazole ( Figure 3A,B, Video S7). In agreement with these data, the immunostaining of metaphase-arrested cells showed a near-total loss of astral MT population upon application of either of these two drugs ( Figure 3C,D). Interestingly, although TH588 and low-dose nocodazole fully disrupted astral MTs, the quantification of fluorescence intensity upon immunostaining revealed a significant increase in the spindle MTs signal ( Figure 3C,D), suggesting that TH588 and low-dose nocodazole have a MT-stabilizing effect on longer-lived MTs.

TH588 and Low Concentration of Nocodazole Similarly Reduce MT Turnover within the Mitotic Spindle
In order to perform an in-depth analysis of MT dynamics within the mitotic spindle of TH588-and nocodazole-treated cells, we quantified MT turnover and poleward flux rates in metaphase-arrested U2OS cells stably expressing PA-GFP/mCherry-α-tubulin. MT turnover was measured by quantifying the fluorescence dissipation after photoactivation (FDAPA), while MT flux was determined as the velocity with which the photoactivated region traversed toward the spindle pole. FDAPA analysis revealed decreased spindle MT turnover both in TH588-and low-dose nocodazole-treated cells. This was represented by increase in both fast (15.44 ± 0.77 s with 5 µM TH588 and 15.39 ± 2.42 s with 100 nM nocodazole, compared to 11.34 ± 0.43 s in DMSO control) and slow fraction of MT half-life (308.2 ± 40.84 s with 5 µM TH588 and 352.8 ± 79.91 s with 100 nM nocodazole, compared to 196.4 ± 7.8 s in DMSO control), indicating stabilization of both non-kinetochore-(interpolar) and kinetochore-MTs (also known as k-fibers) ( Figure 4A-E, Video S8). Moreover, we also observed an increase in slow fraction of MT half-life, corresponding to longerlived kinetochore-MTs (68.36 ± 5.2% with 5 µM TH588 and 79.57 ± 7.92% with 100 nM nocodazole, compared to 45.57 ± 7.35% in DMSO control), additionally highlighting the MT-stabilizing effect of TH588-and low concentration of nocodazole ( Figure 4F). Correspondingly, both TH588 and low-dose nocodazole caused a significant reduction in MT flux rates (0.21 ± 0.11 µm/s with 5 µM TH588 and 0.13 ± 0.09 µm/s with 100 nM nocodazole, compared to 0.62 ± 0.17 µm/s in DMSO control), characteristic for less dynamic, more stable MTs ( Figure 4G, Video S8).
Taken together, these results demonstrate that TH588 and low-dose nocodazole stabilize spindle MTs by suppressing their dynamics.

TH588 Triggers Mitotic Arrest and Erroneous Cell Divisions via Premature Stabilization of Kinetochore-MT End-On Attachments
Suppression of MT-dynamics interferes with the establishment of correct kinetochore-MT attachments [38]. To examine whether the observed chromosome congression problems, as well as cell divisions with uncongressed chromosomes are a consequence of premature stabilization of end-on kinetochore-MT attachments, we imaged U2OS cells stably expressing GFP-Centromere Protein A (CENP-A) and mCherry-α-tubulin. Whereas DMSO-treated control cells divided normally, the addition of TH588 or low-dose nocodazole prematurely stabilized kinetochore-MTs of uncongressed chromosomes. This eventually led to the formation of erroneous syntelic attachments (having both sister kinetochores attached to the MTs emanating from the same spindle pole), further impairing the chromosome transport toward the metaphase plate ( Figures 5A and S3, Videos S9 and S10). To test whether the stable kinetochore-MTs that we observed by imaging U2OS-GFP-CENP-A/mCherry-αtubulin cells are indeed end-on attached to kinetochores, we immunostained TH588-and low-dose nocodazole-treated U2OS cells against ZW10, a spindle assembly checkpoint (SAC) protein that is removed from kinetochores upon the establishment of stable end-on kinetochore-MT attachment. Since ZW10 was absent from MT-bound kinetochores, this immunostaining confirmed that TH588 and nocodazole induce premature stabilization of MTs associated with the chromosomes surrounding the spindle poles, thereby preventing their congression ( Figure 5B,C). Cancers 2021, 13, x FOR PEER REVIEW 13 of 18

Discussion
Whereas at high concentrations MTAs either promote or prevent tubulin polymerization, at 10-100-fold lower concentrations they suppress MT dynamics, thereby stabilizing MTs without altering the equilibrium between MT polymer and soluble tubulin [39]. For instance, at high, micromolar concentrations nocodazole prevents tubulin polymerization in vitro [40] and rapidly depolymerizes MTs in cellulo [41]. However, low, nanomolar concentrations of nocodazole reduce MT-growth and shortening rates and increase the time that MTs spend in paused state, both in vitro and in living cells in interphase [36,37]. Since low concentrations of nocodazole also induced mitotic arrest without detectable MT depolymerization [36], these data suggest that the corresponding problems in mitosis result from MT stabilization, rather than depolymerization.
Similar to low-dose nocodazole, TH588 inhibited tubulin polymerization in vitro [15,22], as well as it arrested cells in mitosis and suppressed MT dynamics in interphase [21]. Moreover, the X-ray crystallography-based structural elucidation of the α/β-tubulin: TH588 complex revealed a close overlap in β-tubulin-binding between TH588 and nocodazole [22].
Although these data suggest that mitotic problems induced by either TH588 or low concentrations of nocodazole result from a reduction in spindle MT turnover, it remains unknown whether and how TH588 and low-dose nocodazole affect MT dynamics in mitosis.
Here, we performed a detailed analysis of the impact of TH588 and low-dose nocodazole on MT dynamics within the mitotic spindles of dividing human cancer cells. We show that both of these treatments indeed reduce the MT turnover within mitotic spindle. Interestingly, this suppression of MT dynamics has a destabilizing effect on astral MTs, while it stabilizes interpolar-and kinetochore-MTs. This is consistent with earlier observations that low concentrations of nocodazole selectively disrupts astral MTs, without interfering with the formation of bipolar spindle [42].
Even though TH588 and low-dose nocodazole reduced MT-growth and shortening rates, as well as increased the time that MTs spend in paused state, they also increased the MT catastrophe frequency and decreased the frequency of MT rescue [37]. The increase in MT catastrophe may explain why these drugs destabilize astral MTs, but stabilize interpolarand kinetochore-MTs. The MT catastrophe rate is suppressed by the tension applied on kinetochores [43,44] and both microneedle-based force application at kinetochores [45,46] and polar-ejection forces (PEFs) exerted on chromosome arms [47,48] stabilize kinetochore-MT attachments. Moreover, kinetochore-based proteins, such as CLASPs, stabilize MT plus-ends by suppressing MT catastrophes and promoting rescue [49][50][51][52][53]. Thus, it is appealing to hypothesize that the increase in catastrophe rate, induced by TH588 and lowdose nocodazole, selectively disrupts astral MTs, while kinetochore-MTs remain protected by the suppressive impact of chromosomes on MT catastrophe rate. In a similar way, the increase in catastrophe rate may be attenuated on interpolar MTs by the actions of MT-crosslinking motors kinesin-5/EG5 [54] and kinesin-12/KIF15 [55], both of which were recently reported to suppress MT catastrophes. With higher concentrations of nocodazole, the increase in MT catastrophes may prevail the MT-stabilizing effects and therefore induce MT depolymerization.
We also show that the spindle MTs-associated stabilizing effect of TH588 and nocodazole results in a premature formation of kinetochore-MT end-on attachments on uncongressed chromosomes, which leads to severe chromosome congression problems and mitotic arrest. At the pole, the uncongressed monotelic chromosomes, with a single kinetochore-MT attachment being stabilized, keep the SAC activated due to the presence of unattached sister kinetochore [56]. Subsequent stabilization of the sister kinetochore-MT attachment leads to the formation of stably attached syntelic chromosomes. These chro-mosomes cannot congress to the spindle equator most likely due to a lack of activity of dynein and kinesin-7/CENP-E, the two motor proteins required for congression of laterally attached peripheral chromosomes [33,57], as both dynein and CENP-E are largely removed from the end-on attached kinetochores. Eventually, the cells with stably attached syntelic chromosomes satisfy the SAC and divide. Under these circumstances, the SAC satisfaction could occur via intra-kinetochore stretching or structural changes at kinetochores that are triggered by stable end-on attachments [58][59][60].
In addition to stabilizing the spindle MTs, TH588 and low-dose nocodazole prevent the formation of astral MTs that connect mitotic spindle with the cell cortex and are required for proper spindle positioning and orientation of cell division. Thus, these treatments do not only increase the chance of cells dividing with erroneously attached chromosomes, but may also compromise the proper spindle positioning (as observed in Video S9).
Interestingly, whereas the mitotic arrest induced by TH588 or low-dose nocodazole in cancer cells mainly ends in cell death or in cell division with unresolved erroneous kinetochore-MT attachments, in non-tumor cells we observed a significant number of normal cell divisions. This difference is consistent with previous observations that TH588 more efficiently kills cancer cell lines [13]. In addition to its inhibitory effect on MTH1, the selectivity of TH588 toward cancer cells may at least partially rely on its MT-associated activity. For example, non-tumor cells used in this study (RPE-1) are diploid, compared to the two nearly triploid cancer cell lines (U2OS and HeLA). It remains to be addressed by future studies whether, because of having more chromosomes, the polyploid cancer cells are less efficient in correcting prematurely stabilized kinetochore-MT attachments and thereby more prone to cell death upon the treatment with TH588.

Conclusions
In conclusion, this study demonstrates that the individual actions of two MT-targeting drugs, TH588 and low concentration of nocodazole, reduce MT turnover within the mitotic spindle of dividing cells. This MT-stabilizing effect leads to premature formation of kinetochore-MT end-on attachments on uncongressed chromosomes, which consequently cannot be transported to the cell equator. The resulting mitotic arrest is mainly resolved either by cell death or by cell division with uncongressed chromosomes that may lead to the problems in the following cell cycle. Both of these cell fates could contribute to the selective effect associated with the TH588 activity in cancer cells.