Mitotic DNA damage promotes chromokinesin-mediated missegregation of polar chromosomes in cancer cells

Abstract

DNA damage response (DDR) during interphase involves active signaling and repair to ensure genomic stability. However, how mitotic cells respond to DNA damage remains poorly understood. Supported by correlative live-/fixed-cell microscopy, it was found that mitotic cells exposed to several cancer chemotherapy compounds acquire and signal DNA damage, regardless of how they interact with DNA. In-depth analysis upon DNA damage during mitosis revealed a spindle assembly checkpoint (SAC)-dependent, but ataxia telangiectasia mutated–independent, mitotic delay. This delay was due to the presence of misaligned chromosomes that ultimately satisfy the SAC and missegregate, leading to micronuclei formation. Mechanistically, it is shown that mitotic DNA damage causes missegregation of polar chromosomes due to the action of arm-ejection forces by chromokinesins. Importantly, with the exception of DNA damage induced by etoposide—a topoisomerase II inhibitor—this outcome was independent of a general effect on kinetochore microtubule stability. Colony formation assays in pan–cancer cell line models revealed that mitotic DNA damage causes distinct cytotoxic effects, depending on the nature and extent of the damage. Overall, these findings unveil and raise awareness that therapeutic DNA damage regimens may contribute to genomic instability through a surprising link with chromokinesin-mediated missegregation of polar chromosomes in cancer cells.

INTRODUCTION

Genomic DNA is continuously damaged by endogenous and exogenous sources that can generate several types of lesions. To preserve genomic stability in response to DNA damage, cells have evolved a sophisticated system known as the DNA damage response (DDR), which controls DNA repair pathways in coordination with several cell cycle checkpoints (Jackson and Bartek, 2009; Hustedt and Durocher, 2017). Among DNA lesions, double-strand breaks (DSBs) are the most deleterious and represent the greatest threat to genomic integrity (Jackson, 2002). Mitotic cells are more vulnerable to DNA damage because most canonical DNA damage repair pathways are inactivated during mitosis. The presence of DSBs in mitosis leads to partial activation of the DDR through the recruitment of the MRE11-RAD50-NBS1 (MRN) complex and activation of ataxia telangiectasia mutated (ATM). ATM-mediated phosphorylation of the histone variant H2AX on serine 139 (γH2AX) promotes the recruitment of the mediator of DNA damage checkpoint 1 (MDC1) to damaged sites (Giunta et al., 2010; van Vugt et al., 2010; Benada et al., 2015). However, the mitotic kinases Cdk1 and Plk1 phosphorylate ring finger protein 8 (RNF8) and p53-binding protein 1 (53BP1) to inhibit their recruitment to DSBs and prevent DNA repair, which could otherwise trigger unintended genomic rearrangements, such as telomere fusions (Nelson et al., 2009; Lee et al., 2014; Orthwein et al., 2014; Benada et al., 2015).

Acute exposure to ionizing radiation during antephase (late G2 to midprophase) can reverse cell cycle progression, causing cells to return to G2 (Pines and Rieder, 2001). However, acute induction of DNA damage once cells commit to mitosis after nuclear envelope breakdown (NEBD) does not allow cell cycle reversal but delays mitotic progression in a spindle assembly checkpoint (SAC)-dependent manner (Carlson, 1950; Zirkle, 1970; Mikhailov et al., 2002; Hayashi et al., 2012; Silva et al., 2014; Gomez Godinez et al., 2020). In contrast, acute induction of DSBs with ionizing radiation or a topoisomerase II inhibitor during mitosis was found to cause an increase in whole-chromosome segregation errors due to the formation of anaphase lagging chromosomes, with no apparent delay in mitotic progression (Bakhoum et al., 2014). These differences were proposed to depend on the extent and localization of the DNA damage, as well as on the hyperstabilization of kinetochore-microtubule attachments due to Aurora A and Plk1 activation upon DNA damage (Bakhoum et al., 2014). Anaphase lagging chromosomes may result in the formation of micronuclei (Soto et al., 2019), which have been recently implicated in genomic instability as critical intermediates of massive chromosome rearrangements associated with chromothripsis (Janssen et al., 2011; Crasta et al., 2012; Zhang et al., 2015a; Shoshani et al., 2021).

The combination of DNA-damaging compounds with micro­tubule-targeting drugs that compromise mitosis is widely used in current cancer chemotherapy regimens (Poruchynsky et al., 2015). Thus clarifying the role of mitotic DNA damage has important clinical implications. Here, we investigated the mechanism by which a broad range of DNA-damaging compounds interfere with mitotic progression and fidelity using a combination of live- and fixed-cell assays, together with specific molecular perturbations. Surprisingly, we found that DNA damage specifically during mitosis interferes with chromosome congression to the spindle equator and induces a SAC-dependent, but ATM-independent, mitotic delay. Importantly, misaligned chromosomes eventually satisfy the SAC and lead to micronuclei formation. Mechanistically, it is shown that the activity of chromokinesins on chromosome arms leads to the missegregation of polar chromosomes specifically in cancer cells upon mitotic DNA damage. Taken together, our findings unveil a previously overlooked route by which DNA damage during mitosis contributes to whole-chromosome missegregation. Thus, by driving chromosomal instability, mitotic DNA damage might represent both a threat to and an opportunity for chemotherapeutic approaches involving the use of DNA-damaging drugs.

RESULTS

The extent of mitotic DNA damage depends on the type of DNA lesion

To investigate the impact of DNA damage during mitosis, we started by screening a chemical library of 448 cell cycle inhibitors/DNA-damaging compounds using high-content fluorescence microscopy. For this purpose, DNA damage was induced for 4 h in a mitotic-enriched population of HeLa cells obtained by previous treatment with the microtubule-depolymerizing drug nocodazole for 3 h. Cells were subsequently fixed and processed for immunofluorescence to determine the respective extent of DNA damage by quantifying the phosphorylation levels of γH2AX (Rogakou et al., 1998). To assist in the identification of mitotic cells, we simultaneously measured the phosphorylation of histone H3 at serine 10 (pH3S10 [Crosio et al., 2002]). We used the Z´-factor, a dimensionless, simple statistical coefficient that reflects both the signal dynamic range and variability (Zhang et al., 1999), as a measure of the robustness of our high-throughput assay in identifying mitotic cells with DNA damage (a Z´-factor between 0.5 and 1 is an excellent indicator of robustness). Accordingly, the Z´-factor between dimethyl sulfoxide (DMSO; negative control) and the topoisomerase II inhibitor etoposide (positive control) was 0.76 for mitotic cells (Figure 1A) and 0.60 for interphase cells (Supplemental Figure 1A). From the screened library, only the compounds that resulted in a percentage of cells with γH2AX higher than the median plus three SDs (3SD) of the negative control were considered as hits. On the basis of this criterion, we identified 73 hits that induced DNA damage in mitotic cells, 10 of which did not induce detectable DNA damage (or the damage was rapidly repaired) in interphase cells (Figure 1, A–C). In parallel, we identified 95 compounds that induced DNA damage in interphase cells (Supplemental Figure 1A), 32 of which did not cause detectable DNA damage in mitotic cells (Figure 1C).

FIGURE 1:

Extent of mitotic DNA damage depends on the type of DNA lesion. (A) Human HeLa cells were partially synchronized in mitosis by previous treatment with nocodazole for 3 h. Subsequently, cells were treated with a chemical library of 448 cell cycle inhibitors/DNA-damaging compounds for 4 h and then cell fixation and immunofluorescence. The percentages of mitotic HeLa cells with positive γH2AX were determined. DMEM and DMSO (green dots) and topoisomerase II inhibitor etoposide (blue dots) were used as negative and positive controls, respectively. Each dot represents a compound. The thresholds (dash lines) were defined as mean (Avg) ± 3SD from negative controls. Compounds above the threshold were considered hits (magenta dots). (B) Representative images of negative (DMSO) and positive (etoposide) control and hit compounds (mitoxantrone, clofarabine, and streptozocin). Scale bar = 10 μm. (C) Venn diagram of identified hits in interphase and mitotic HeLa cells from 448 tested compounds. (D) Human HeLa cells were tracked by time-lapse, bright-field microscopy after treatment with nocodazole for 5 h. Subsequently, cells were treated for 4 h with DNA-damaging compounds and then cell fixation and immunofluorescence. (Left) Selected time frames from bright-field microscopy of HeLa cells treated with nocodazole and etoposide. Images were acquired every 20 min. Time = h:min. (Right) Immunofluorescence images of the last time frame (09:00) from bright-field microscopy with the indicated antibodies. Arrows highlight examples of cells already committed to mitosis when exposed to the drug. Scale bar = 10 μm. (E) (Left) Histograms of selected DNA-damaging compounds (n = 14), showing fluorescence intensity of γH2AX in mitotic HeLa cells. The threshold (black dash line) was defined as the lowest fluorescence intensity detected for γH2AX foci. The different categories (low, medium, high; dashed gray squares) were defined based on intensity levels of γH2AX. (Right) Representative imagens of different intensity levels of γH2AX in mitotic HeLa cells. Scale bar = 10 μm.

Because the above setup did not fully eliminate the possibility that DNA damage was induced before a fraction of cells had entered mitosis, we implemented a correlative live-cell microscopy assay and quantitative immunofluorescence analysis of γH2AX to investigate whether a subset of the identified compounds induced DNA damage when cells were already in mitosis. Briefly, human HeLa cells were tracked by time-lapse, bright-field microscopy after treatment with nocodazole for 5 h. Subsequently, cells were treated for 4 h with 14 of the 73 mitotic DNA-damaging compounds (selection based on current clinical use). Upon fixation, the levels of γH2AX were analyzed by fluorescence microscopy and the history of each cell was inferred by backtracking using the corresponding live-cell data, confirming that DNA damage in all cells analyzed was inflicted when they were already committed to mitosis (Figure 1, D and E). We found that lesions caused by either DNA-alkylating compounds (cyclophosphamide, busulfan, lomustine, uramustine), DNA-cross-linking agents (mitomycin c, chlorambucil), or DNA adduct-inducing drugs (oxaliplatin, carboplatin) result in low levels of DNA damage during mitosis. Except for BMH-21 and irinotecan, which induced low levels of DNA damage in mitotic cells, treatment with DNA-intercalating compounds (actinomycin D and doxorubicin) or drugs that generate single- or double-strand breaks (etoposide and teniposide) resulted in higher levels of mitotic DNA damage (Figure 1E). Western blot analysis of mitotic cells treated with the selected DNA-damaging compounds confirmed these results (Supplemental Figure 1B). Taking the data together, we conclude that HeLa cells may acquire DNA damage specifically during mitosis and the extent of the damage is dependent on the type of DNA lesion.

Mitotic DNA damage causes an ATM-independent delay in anaphase onset due to the formation of polar chromosomes that result in micronuclei

To investigate the impact of mitotic DNA damage on the progression and exit of mitosis, we used live-cell fluorescence microscopy to directly determine the mitotic duration and respective cell fate. To ensure the induction of DNA damage during mitosis, we implemented a nocodazole treatment/washout protocol. Briefly, human HeLa cells were partially synchronized in mitosis upon treatment with nocodazole for 3 h. Subsequently, selected representative compounds that induce distinct types of DNA lesions (lomustine, mitomycin C, carboplatin, and etoposide) were added for 4 h at a concentration that prevented interphase cells to enter mitosis, as previously inferred by live-cell imaging (Figure 1D). So, all cells analyzed were already in mitosis when DNA damage was inflicted. After this period, nocodazole was washed out and cells allowed to progress through mitosis in the presence of the respective DNA-damaging compounds. Under these conditions, control DMSO-treated cells took 88 ± 34 min (mean ± SD) to enter anaphase after nocodazole washout and 85% of the cells divided without any detectable segregation defect (Figure 2, A–C, and Supplemental Figure 1C). In contrast, among other defects, treatment with any of the DNA-damaging compounds caused a significant mitotic delay due to the presence of misaligned chromosomes located between the cell cortex and a spindle pole (from here on, referred to as polar chromosomes) (Figure 2, A–C, and Supplemental Figure 1C). Despite the observed mitotic delay, most cells were able to enter anaphase, some of them (6–13%, depending on the condition) without ever completing congression of polar chromosome(s) that ultimately gave rise to micronuclei (Figure 2C). To evaluate the contribution of DDR to the mitotic delay observed in the presence of DNA-damaging compounds, we inhibited ATM activity with AZD0156 1 h before the induction of mitotic DNA damage and monitored mitotic duration and cell fate by live-cell fluorescence microscopy. Inhibition of ATM activity was confirmed by the decrease in γH2AX levels (Supplemental Figure 2A). We found that inhibition of ATM did not rescue the mitotic timing or segregation errors (Figure 2, A–C), indicating that the mitotic delay observed upon DNA damage is independent of DDR.

FIGURE 2:

Mitotic DNA damage causes an ATM-independent delay in anaphase onset due to the formation of polar chromosomes that result in micronuclei. (A) Human HeLa cells were partially synchronized in mitosis by previous treatment with nocodazole for 3 h. Subsequently, ATM activity was inhibited with AZD0156 (ATMi) for 1 h, and then cells were treated with DNA-damaging compounds for 4 h. After this period, nocodazole was washed out and cells allowed to progress through mitosis in the presence of ATMi and the respective DNA-damaging compounds. The mitotic duration and respective cell fate were determined using live-cell fluorescence microscopy. Representative images of time-lapse sequences illustrating the three main mitotic phenotypes observed. Arrows highlight chromosomes at the poles or micronuclei formation. Images were acquired every 10 min. Scale bar = 10 µm. Time = h:min. (B) Mitotic duration between nocodazole washout and anaphase onset in mitotic HeLa cells treated with DMSO or DNA-damaging compounds in the presence or the absence of ATMi. Each data point corresponds to one cell. Dash line indicates the point when recordings were stopped. Cells that were still in mitosis beyond this point are indicated. The results are expressed as mean ± SD; n.s.: not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001 relative to control, multiple comparison Kruskal–Wallis test. Two independent experiments per condition were performed. (C) Cell fate of mitotic HeLa cells treated with the same drugs as in B. Detailed description of cell fates listed as “other” is given in Supplemental Figure 1C.

To test whether the induction of DNA damage in prometaphase cells persists throughout mitosis, we fixed cells 70 min after nocodazole washout. We found that γH2AX foci remained on chromosomes in cells that have reached metaphase and entered anaphase upon nocodazole washout (Figure 3). These results indicate that damaged sites are recognized but not repaired during mitosis, in agreement with previous observations (Ciccia and Elledge, 2010; Giunta et al., 2010; Giunta and Jackson, 2011). Taken together, these experiments show that, regardless of the type of DNA lesion, mitotic DNA damage persists throughout mitosis and promotes micronuclei formation from polar chromosomes.

FIGURE 3:

Mitotic DNA damage persists throughout mitosis. Human HeLa cells were partially synchronized in mitosis by previous treatment with nocodazole for 3 h. Subsequently, cells were treated with DNA-damaging compounds for 4 h. After this period, nocodazole was washed out and, after 70 min, the cells were fixed and processed for immunofluorescence. Cells were immunostained with anti–CENP-C (gray), anti–β-tubulin (magenta), anti-γH2AX (green), and DAPI (blue). Representative immunofluorescence images of HeLa cells after nocodazole washout treated with either DMSO or DNA-damaging compounds are shown. Scale bar = 5 μm.

The mitotic delay observed upon DNA damage during mitosis is mediated by the SAC

Given that DDR was not required for the observed mitotic delay induced upon DNA damage during mitosis, we investigated the potential dependency on SAC activity. To do so, we performed live-cell imaging to monitor mitotic duration and cell fate upon mitotic DNA damage and subsequent SAC abrogation with the Mps1 inhibitor MPS1-IN-1 (Kwiatkowski et al., 2010). We found that treatment with either MPS1-IN-1 alone or in combination with DNA-damaging compounds caused a fast mitotic exit even in conditions of maximal SAC activity such as in nocodazole-arrested HeLa cells (Figure 4, A and B). This phenotype was fully reverted by preventing SAC-independent mitotic exit by treatment of cells with the proteasome inhibitor MG132 (Figure 4, A and B). These results indicate that the observed mitotic delay upon DNA damage during mitosis is exclusively mediated by the SAC.

FIGURE 4:

The mitotic delay observed upon DNA damage during mitosis is mediated by the SAC. (A) Human HeLa cells were tracked by time-lapse, bright-field microscopy after treatment with nocodazole for 5 h. Subsequently, cells were treated for 4 h with DNA-damaging compounds. After this period, the SAC was abrogated with the Mps1 inhibitor (MPS1-IN-1) in the presence or the absence of proteasome inhibitor MG132. The mitotic duration was determined. Each data point corresponds to one cell. The results are expressed as mean ± SD; n.s.: not significant relative to control, multiple comparison Kruskal–Wallis test. Two independent experiments per condition were performed. (B) Cell fate of mitotic HeLa cells treated with the same drugs as in A.

Mitotic DNA damage leads to the formation of polar chromosomes that eventually form stable kinetochore-microtubule attachments

Next, we focused on determining the status of kinetochore-micro­tubule attachments on polar chromosomes that form upon mitotic DNA damage. To do so, we investigated the localization of the SAC protein Mad1, which accumulates at unattached kinetochores, delaying anaphase onset (Musacchio, 2015). In this case, cells were fixed 70 min after nocodazole washout, a time window in which we detected similar levels of polar chromosomes with or without DNA damage. Of note, beyond this time frame, polar chromosomes were detectable only in cells treated with DNA-damaging agents (our unpublished observations), with an associated mitotic delay (Figure 2, A–C). Strikingly, in contrast to control DMSO-treated cells where the Mad1 signal was mostly detected on both kinetochores, mitotic DNA damage significantly increased the frequency of polar chromosomes in which only the most distal kinetochore remained Mad1 positive (Figure 5, A and B), indicating the establishment of stable monotelic attachments. Importantly, this effect was independent of ATM activity and was also observed after mitotic arrest/washout with the kinesin-5 inhibitor S-trityl-l-cysteine (STLC) upon prolonged DNA damage, ruling out a role of DDR and a nocodazole/microtubule-specific effect (Figure 5B and Supplemental Figure 2, B–F). The observed Mad1 asymmetry after mitotic DNA damage was not due to differences in the total levels of Mad1 protein (Figure 5C) and phenocopied the behavior observed upon microtubule stabilization with 20 nM taxol (Figure 5, D and E). These results suggest that prolonged mitotic DNA damage promotes the formation of polar chromosomes that eventually form stable kinetochore-microtubule attachments.

FIGURE 5:

Mitotic DNA damage promotes the formation of polar chromosomes that form stable monotelic attachments. (A) Human HeLa cells were partially synchronized in mitosis by previous treatment with nocodazole for 3 h. Subsequently, ATM activity was inhibited with AZD0156 (ATMi) for 1 h and then cells were treated with DNA-damaging compounds for 4 h. After this period, nocodazole was washed out and, after 70 min, the cells were fixed and processed for immunofluorescence. Cells were immunostained with anti–CENP-C (gray), anti–β-tubulin (magenta), anti-Mad1 (green), and DAPI (blue). Representative immunofluorescence images of HeLa cells after nocodazole washout treated with either DMSO or DNA-damaging compounds are shown. Scale bar = 5 μm. (B) Quantification of Mad1 on kinetochore (KT) pairs from cells with polar chromosomes under different conditions presented in A. Data pooled from three independent experiments and analysed using Fisher’s exact two-tailed test; n.s.: not significant, *p ≤ 0.05 relative to control. (C) Human mitotic HeLa cells were isolated by shake off after 3 h in nocodazole. Subsequently, mitotic cells were treated with DNA-damaging compounds for 4 h and processed for Western blot. Mad1 and MPM2 were immunodetected, and GAPDH was used as loading control. Approximate molecular weights are shown on the left. (D) HeLa cells were treated with Taxol for 2 h, followed by cell fixation and immunofluorescence. Cells were immunostained with anti–CENP-C (gray), anti–β-tubulin (magenta), anti-Mad1 (green), and DAPI (blue). Representative immunofluorescence images of HeLa cells treated with DMSO or Taxol are shown. Scale bar = 5 μm. (E) Quantification of Mad1 on kinetochore (KT) pairs from cells with polar chromosomes under different conditions presented in D. Data pooled from three independent experiments.

Etoposide, but not other mitotic DNA-damaging agents, increases kinetochore-microtubule attachment stability in cancer cells

To directly investigate the effect of the different DNA-damaging agents on kinetochore-microtubule attachment stability, we determined the respective kinetochore-microtubule half-life upon a nocodazole shock (Warren et al., 2020) in HeLa and U2OS cells treated with lomustine, mitomycin C, carboplatin, or etoposide. In agreement with previous measurements using photoactivation of GFP-α-tubulin where fluorescence dissipation after photoactivation was used as a readout for microtubule turnover in the mitotic spindle (Bakhoum et al., 2014), our measurements using a nocodazole shock revealed that kinetochore microtubules were more resistant to nocodazole-induced depolymerization upon prolonged mitotic DNA damage specifically with etoposide but not with lomustine, mitomycin C, or carboplatin, as determined by a significant increase in microtubule half-life from 2.22 min in HeLa cells and 2.10 min in U2OS cells without mitotic DNA damage to 4.02 min in HeLa cells and 4.61 min in U2OS cells upon mitotic DNA damage in the former (Figure 6, A and B, and Supplemental Figure 3, A and B). To determine how etoposide affects kinetochore-microtubule attachment stability, we investigated the localization of phosphorylated Plk1 and BubR1 at kinetochores (Godek et al., 2015). We also analyzed the kinetochore levels of Bub1, which mediates the kinetochore recruitment of Plk1 and BubR1 (Zhang et al., 2015b; Ikeda and Tanaka, 2017). We found that the addition of etoposide for 4 h to nocodazole-treated cells did not perturb the normal localization of Bub1 (Supplemental Figure 4, A and B) or the Cdk1-dependent phosphorylation of BubR1 at serine 670 (Supplemental Figure 4, C and D). Consistent with previous observations (Bakhoum et al., 2014), etoposide treatment for 4 h led to a 28% increase of active phosphorylated Plk1 (Plk1 Thr 210) (Carmena et al., 2012) (Supplemental Figure 4, E and F). These data suggest that etoposide increases kinetochore-microtubule attachment stability through activation of Plk1 at kinetochores. Finally, we investigated whether DNA-damaging compounds directly interfere with microtubule polymerization by measuring purified tubulin polymerization in vitro in the presence of lomustine, mitomycin C, carboplatin, or etoposide. We found no difference in the rates of tubulin polymerization relative to control, indicating that none of the DNA-damaging compounds directly interferes with microtubule polymerization (Figure 6C).

FIGURE 6:

Etoposide, but not other mitotic DNA-damaging agents, increases kinetochore-microtubule attachment stability in HeLa cells. (A) Human HeLa cells were partially synchronized in mitosis by previous treatment with nocodazole for 3 h. Subsequently, cells were treated with DNA-damaging compounds for 4 h. After this period, nocodazole was washed out and the cells allowed to progress through mitosis. After 70 min, nocodazole was added and the cells were fixed and processed for immunofluorescence after 1, 2.5, 5, 15, and 30 min. Cells were immunostained with anti–α-tubulin (inverted grayscale) and DAPI (blue). Representative immunofluorescence images of HeLa cells after nocodazole shock treated with either DMSO or DNA-damaging compounds are shown. Scale bar = 5 μm. Taxol was used as a positive control. (B) Normalized texture homogeneity of α-tubulin over time. The alterations on texture homogeneity of α-tubulin after nocodazole shock were determined as a readout of resistance to microtubule depolymerization. Whole lines show single exponential fitting curve. Each data point represents the mean ± SD; ***p ≤ 0.001 relative to control, extra sum-of-square F test. Two independent experiments per condition were performed. (C) In vitro tubulin polymerization was performed in the presence of DMSO or DNA-damaging compounds. Taxol and nocodazole were used as a known enhancer and inhibitor of tubulin polymerization, respectively. The kinetics of tubulin polymerization was monitored by measuring absorbance at 340 nm at 37°C every minute for 120 min. The results are expressed as mean ± SD. The V_max (mOD/min) was calculated for each condition. n.s.: not significant, ****p ≤ 0.0001 relative to control, multiple comparisons one-way ANOVA.

To test whether the observed stabilization of kinetochore-microtubule attachments upon prolonged mitotic DNA damage with etoposide is a common feature between cancer and noncancer cells, we extended our analysis to measure kinetochore-microtubule half-life in hTERT-immortalized, but not transformed, near-diploid RPE-1 cells. In sharp contrast with HeLa and U2OS cells, no detectable differences were found in kinetochore-microtubule half-life in RPE-1 cells relative to DMSO-treated controls (Supplemental Figure 5, A and B). Moreover, all polar chromosomes in RPE-1 cells completed congression after nocodazole washout upon prolonged mitotic DNA damage with etoposide for an equivalent time window (Supplemental Figure 5C). Overall, these results indicate that, with the exception of etoposide, the formation of polar chromosomes upon prolonged mitotic DNA damage is largely independent of any general effect on kinetochore-microtubule stability and appears to be a specific outcome in cancer cells.

Mitotic DNA damage causes missegregation of polar chromosomes due to the action of arm-ejection forces by chromokinesins

The kinetochore-microtubule attachment status of polar chromosomes depends on the coordinated activities of kinetochore- and arm-associated motor proteins that ultimately determine the proximity to the spindle pole where the microtubule-destabilizing activity of Aurora A is higher (Barisic et al., 2014; Barisic and Maiato, 2015; Ye et al., 2015). For instance, inhibition of the microtubule plus end–directed motor CENP-E at kinetochores generates polar chromosomes in the immediate vicinity of the spindle poles with both kinetochore pairs unattached due to the microtubule minus end–directed motor activity of dynein at kinetochores (McEwen et al., 2001; Putkey et al., 2002; Maia et al., 2010; Barisic et al., 2014). In contrast, simultaneous inhibition of both CENP-E and dynein motor activities at kinetochores results in the cortical ejection of polar chromosomes away from high Aurora A activity at the poles due to the action of the chromokinesins Kid/kinesin-10 and Kif4a/kinesin-4 on chromosome arms, thereby stabilizing monotelic kinetochore-microtubule attachments (Cane et al., 2013; Barisic et al., 2014; Barisic and Maiato, 2015; Drpic et al., 2015; Ye et al., 2015). Given that mitotic DNA damage results in the formation of polar chromosomes that establish stable monotelic attachments, we investigated whether this stabilizing effect was dependent on chromokinesin-mediated ejection of polar chromosomes away from high Aurora A activity at the poles. Interestingly, it had been previously shown that, upon ionizing radiation, Kif4a is recruited to damaged DNA sites in interphase cells (Wu et al., 2008). However, the total levels of Kif4a (and Kid) on chromosome arms of polar chromosomes relative to aligned chromosomes were not altered upon mitotic DNA damage (Supplemental Figure 6, A–D). Moreover, neither Aurora A activity at spindle poles nor Aurora B centromeric activity on polar chromosomes was affected upon mitotic DNA damage (Supplemental Figures 6, E–H, and 7, A–D). Nevertheless, codepletion of both Kid and Kif4a significantly rescued the formation of polar chromosomes with stable monotelic attachments that formed upon mitotic DNA damage (Figure 7, A–C). Moreover, we found that depletion of chromokinesins partially rescued the mitotic delay observed upon DNA damage, as well as the frequency of cells with micronuclei derived from polar chromosomes (Figure 8, A–C). Altogether, these results support a model in which mitotic DNA damage causes the missegregation of polar chromosomes due to the action of arm-ejection forces by chromokinesins specifically in cancer cells.

FIGURE 7:

Depletion of chromokinesins reverts the formation of stable monotelic attachments on polar chromosomes upon mitotic DNA damage. (A) Kid/Kif4a-depleted HeLa cells were partially synchronized in mitosis by previous treatment with nocodazole for 3 h. Subsequently, cells were treated with DNA-damaging compounds for 4 h. After this period, nocodazole was washed out and, after 70 min, the cells were fixed and processed for immunofluorescence. Cells were immunostained with anti–CENP-C (gray), anti–β-tubulin (magenta), anti-Mad1 (green), and DAPI (blue). Representative immunofluorescence images of Kid/Kif4a-depleted HeLa cells after nocodazole washout treated with either DMSO or DNA-damaging compounds are shown. Scale bar = 5 μm. (B) Western blot analysis of Kid/Kif4a-depleted HeLa cells with the indicated antibodies. Kid and Kif4a were immunodetected, and GAPDH was used as loading control. Approximate molecular weights are shown on the left. (C) Quantification of Mad1 on kinetochore (KT) pairs from cells with polar chromosomes under the different conditions presented in A. Data pooled from two independent experiments and analyzed using Fisher’s exact two-tailed test; n.s.: not significant, *p ≤ 0.05 relative to control.

FIGURE 8:

Mitotic DNA damage promotes missegregation of polar chromosomes due to chromokinesin activity. (A) Kid/Kif4a-depleted HeLa cells were partially synchronized in mitosis by previous treatment with nocodazole for 3 h. Subsequently, cells were treated with DNA-damaging compounds for 4 h. After this period, nocodazole was washed out and cells allowed to progress through mitosis in the presence of the respective DNA-damaging compounds. The mitotic duration between nocodazole washout and anaphase onset was determined. Each data point corresponds to one cell. Dash line indicates the point when recordings were stopped. Cells that were still in mitosis beyond this point are indicated. The results are expressed as mean ± SD; n.s.: not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001 relative to control, multiple comparison Kruskal–Wallis test. Two independent experiments per condition were performed. (B) Cell fate of Kid/Kif4a-depleted HeLa cells treated with the same drugs as in A. (C) Detailed description of cell fates listed as “other” in B.

The type and extent of mitotic DNA damage have distinct effects on cancer cell proliferation

To investigate the impact of mitotic DNA damage on cancer cell growth, we performed a colony formation assay using different cancer cell lines exposed to DNA-damaging compounds. For this purpose, we performed a mitotic shake off in cells previously exposed to nocodazole followed by lomustine, mitomycin C, carboplatin, or etoposide. We found that these treatments resulted in different extents of mitotic DNA damage (Figure 9A), with distinct repercussions on the capacity of these different cancer cells to survive and proliferate (Figure 9, B and C). It is noteworthy that, while all DNA-damaging compounds reduced the proliferative capacity of the different cancer cells, only mitomycin C and etoposide did so in an efficient manner. In contrast, lomustine and carboplatin were still permissive to a significant proliferation of the different cancer cells, leading to colony formation. These results suggest that the type and extent of mitotic DNA damage influence the proliferative capacity of cancer cells and might contribute to both enhanced cytotoxicity and cancer cell evolution, depending on the clinical context and the choice of chemotherapeutic regimen.

FIGURE 9:

The type and extent of mitotic DNA damage influence cancer cell proliferation and survival. (A) Cancer cells were partially synchronized in mitosis by previous treatment with nocodazole for 3–8 h, according to the cell line. Subsequently, cells were treated with DNA-damaging compounds for 4 h, and then they proceeded to cell fixation and immunofluorescence. Histograms of fluorescence intensity of γH2AX in mitotic U2OS, MDA-MB-231, OVCAR-8, HCT-116, and SF-295 cells are shown. The threshold (black dash line) was defined as the lowest fluorescence intensity detected for γH2AX foci. (B) Cancer mitotic cells were isolated by shake off after 3–8 h in nocodazole, according to the cell line. Subsequently, mitotic cells were treated with DNA-damaging compounds for 4 h. After this period, nocodazole was washed out and cells were seeded and allowed to grow for 14–28 d until colonies had formed. Representative images of HeLa colonies are shown. (C) Quantification of colony formation after treatment with DMSO or DNA-damaging compounds. The data are presented as mean ± SD; **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001 relative to control, multiple comparisons one-way ANOVA. Two independent experiments per condition were performed.

DISCUSSION

Several DNA-damaging compounds are currently used in combination with microtubule-targeting drugs that compromise mitosis in several cancer treatment regimens. Thus, understanding the consequences of DNA damage during mitosis remains fundamental in cell biology, with strong clinical implications. Previous studies suggested that DNA damage in mitosis is signaled but not repaired (Ciccia and Elledge, 2010; Giunta et al., 2010; Giunta and Jackson, 2011). Here we show that DNA damage induced after cells commit to mitosis (i.e., prometaphase) remains throughout metaphase and anaphase, supporting the idea of partial activation of DDR in mitosis. Interestingly, even without additional external DNA-damaging sources, a prolonged mitotic arrest induced by microtubule poisons (e.g., nocodazole) was shown to promote the accumulation of DNA damage and to delay mitotic exit after drug washout (Smits et al., 2000; Dalton et al., 2007). However, we found that partial activation of DDR in mitosis is insufficient to prevent human cancer cells from entering anaphase upon DNA damage. Instead, our results suggest that the SAC is both required and sufficient to delay mitosis in the presence of mitotic DNA damage. These findings are consistent with a previous work that showed that acute mitotic DNA damage induced by laser microsurgery and topoisomerase II inhibitors in human cells during late prophase results in the formation of few unattached kinetochores that delay cells in metaphase in a SAC-dependent manner (Mikhailov et al., 2002). However, much to our surprise, extended mitotic DNA damage resulted in a SAC-dependent delay of mitotic exit due to the formation of polar chromosomes. In many such cases, cells ended up satisfying the SAC, entering anaphase without ever completing congression of polar chromosomes to the spindle equator. We show that extended mitotic DNA damage causes the missegregation of polar chromosomes due to the action of arm-ejection forces by chromokinesins, directly leading to the formation of micronuclei (Figure 10).

FIGURE 10:

Proposed model for missegregation of polar chromosomes upon mitotic DNA damage. The induction of extended DNA damage during mitosis causes missegregation of polar chromosomes due to the action of arm-ejection forces by chromokinesins, leading to micronuclei formation.

The chromokinesin Kif4a had been shown to be recruited to DNA damage sites during interphase (Wu et al., 2008), and enhanced chromokinesin activity promotes the stabilization of kinetochore-microtubule attachments in Drosophila cells (Cane et al., 2013; Drpic et al., 2015). While we were unable to detect any significant enrichment of Kif4a or Kid on DNA-damaged sites on polar chromosomes, the possibility that mitotic DNA damage in cancer cells promotes chromokinesin activity on chromosome arms by a yet unclear mechanism cannot be excluded. Interestingly, both Kif4a and Kid are often overexpressed in human cancers (Hou et al., 2017; Li et al., 2020), which might help to explain why only cancer cells manifest this effect upon DNA damage. Acute DNA damage was also proposed to increase kinetochore-microtubule attachment stability through activation of Aurora A and Plk1 kinases (Bakhoum et al., 2014). However, in this case, increased kinetochore-microtubule attachment stability due to acute DNA damage induced the formation of lagging chromosomes during anaphase, which might result in the formation of micronuclei. Importantly, our data support that a general effect upon kinetochore-microtubule stability depends on the type of DNA damage, namely treatments that interfere with topoisomerase II function. Moreover, we found that Aurora A activity was not altered in the case of DNA damage in cells that have already been in mitosis for several hours. Because most anaphase lagging chromosomes in normal and cancer cells rarely form micronuclei and were recently shown to be actively corrected by an anaphase surveillance mechanism involving Aurora B activity at the spindle midzone (Orr et al., 2021; Sen et al., 2021), the presence of misaligned chromosomes that eventually satisfy the SAC upon mitotic DNA damage might represent a higher risk for micronuclei formation and consequent genomic instability due to chromothripsis. Indeed, we recently found that systematic perturbation of kinetochore-microtubule attachments caused human cancer cells to enter anaphase after a delay with chronically misaligned chromosomes that eventually satisfy the SAC and missegregate, leading to the formation of micronuclei (Gomes et al., 2022). Thus, chronically misaligned chromosomes may represent a previously overlooked mechanism driving chromosomal/genomic instability during cancer cell division, and this appears to be promoted by extended DNA damage during mitosis. This can be seen as both a threat to and an opportunity for chemotherapeutic approaches involving the combined use of DNA-damaging compounds with microtubule-targeting drugs that compromise mitosis. On one hand, these combinatorial therapies might result in enhanced cytotoxicity due to a synergistic effect caused by DNA damage and chromosome missegregation. On the other hand, this might contribute to cancer cell adaptation and rapid evolution, which might contribute to therapeutic resistance. The results provided in the present study justify further and careful scrutiny of the pros and cons of combinatorial therapies involving DNA-damaging and mitotic drugs in the context of specific cancers.

MATERIALS AND METHODS

Request a protocol through Bio-protocol.

Cell culture and siRNA transfections

HeLa parental and U2OS parental cells were kindly provided by J. Pines (Cancer Research Institute, London, UK) and S. Geley (Innsbruck Medical University, Innsbruck, Austria), respectively. HeLa cells stably expressing GFP-H2B/mRFP-α-tubulin were previously generated in Nunes et al. (2020). All cell lines, including hTERT RPE-1 parental (American Type Culture Collection), were grown in DMEM (Corning) supplemented with 10% fetal bovine serum (FBS; Life Technologies). All cell lines were grown with penicillin/streptomycin (100 IU/ml and 100 µg/ml; Life Technologies) at 37°C in a humidified incubator with 5% CO₂. To perform small interfering RNA (siRNA) experiments, HeLa cells were plated onto 22 × 22 mm no. 1.5 glass coverslips (Corning) in DMEM supplemented with 5% FBS. siRNA transfection was performed using Lipofectamine RNAiMax (Invitrogen) and 50 nM respective siRNA, diluted in serum-free medium (Opti-MEM; Life Technologies). Mock transfection was used as control. The cells were analyzed 72 h after depletion. Depletion efficiency was evaluated by Western blot. The following target sequences were used:

• Kid, 5´-CAAGCUCACUCGCCUAUUGTT-3´

• Kif4a, 5´-GCAAUUGAUUACCCAGUUATT-3´

Drug treatments

Microtubule depolymerization was induced by nocodazole (Sigma-Aldrich) at 3.3 μM for 3–8 h, according to the experiment. HeLa cells were also blocked in mitosis with 10 μM S-trityl-l-cysteine (STLC; Tocris). Proteasome inhibition was obtained using 5 μM MG132 (Calbiochem). To induce different types of DNA lesions, the following drugs were used for 4 h: 10 μM cyclophosphamide (MedChemExpress), 10 μM busulfan (MedChemExpress), 10 μM lomustine (Selleckchem), 10 μM Uramustine (MedChemExpress), 10 μM mitomycin C (Selleckchem), 10 μM chlorambucil (MedChemExpress), 10 μM oxaliplatin (Selleckchem), 10 μM carboplatin (Selleckchem), 10 or 1 μM etoposide (Selleckchem), 10 μM teniposide (MedChemExpress), 10 μM irinotecan (MedChemExpress), 10 μM BMH-21 (MedChemExpress), 8 μM actinomycin D (Selleckchem), and 0.5 μM doxorubicin (Sigma-Aldrich). For ATM inhibition, cells were treated with 1 μM AZD0156 (Selleckchem) 1 h before DNA damage induction. To stabilize microtubules, 20 nM Taxol (Sigma-Aldrich) was added to the cell culture media 2 h before immunofluorescence. For Aurora A and Aurora B inhibition, 250 nM MLN-8054 (Selleckchem) and 2 μM ZM447439 (Selleckchem) were added, respectively, to the cells 70 or 30 min before fixation. For Mps1 inhibition, mitotic cells were treated with 10 μM MPS1-IN-1 (Tocris). In the washout experiments, nocodazole or STLC was washed out by rinsing thrice with phosphate-buffered saline (PBS) followed by incubation with warm medium. Cells were fixed after 70 min. In RPE-1, to detect misaligned chromosomes, the cells were fixed 30 min after nocodazole washout. Microtubule stability assay was performed by triggering microtubule depolymerization with nocodazole treatment for 1, 2.5, 5, 15, and 30 min before fixation. For all live-cell experiments, drugs were added directly into the imaging medium and remained during the experiment.

High-throughput screening

HeLa cells were seeded in 96-well flat bottom plates (Capitol Scientific) at a density of 10,000 cells/100 μl/well using the Multidrop Combi Reagent Dispenser (Thermo Scientific) and allowed to attach for 24 h. Cells were then blocked in mitosis with nocodazole for 3 h. After this period, 448 compounds from the Cell Cycle/DNA-Damaging Compounds Library (MedChemExpress) were added to cells at a final concentration of 10 μM using a Janus Automated Workstation (PerkinElmer) equipped with a pin tool (V&P Scientific). On each plate, DMSO and 10 μM etoposide were used as negative and positive controls, respectively. After 4 h, cells were fixed and stained for phosphohistone H2AX (Ser-139), phosphohistone H3 (Ser-10), and 4′,6-diamidino-2-phenylindole (DAPI). The image acquisition was performed in an IN Cell Analyzer 2000 microscope (GE Healthcare) using a Nikon 20×/0.45 NA Plan Fluor objective, and nine fields of view per well were acquired. Image analysis was executed using IN Cell Investigator software (GE Healthcare). Briefly, the image analysis workflow consists of the identification of the cell nuclei from the DAPI channel. Then, the mean pixel intensities of the GFP and Texas-Red channels were measured for each cell. Spotfire software (TIBCO) was used to visualize and perform quality control of the data and to set the threshold to identify nuclei with positive GFP and Texas-Red signals. Cells with DNA damage were identified by the detected nuclei with positive GFP signal. Mitotic cells were defined by the cells with positive Texas-Red signal, and mitotic cells with DNA damage were identified by overlap between positive GFP and Texas-Red signals. Hits were identified using a threshold of 3SD above the mean of the negative control. The robustness of the screening assay was evaluated by the Z´-factor (Zhang et al., 1999):

where μ_p, μ_n and σ_p, σ_n represent mean and SD of the positive and the negative controls, respectively.

Single-cell backtracking

HeLa cells were imaged on a temperature-controlled IN Cell Analyzer 2000 microscope (Nikon 20×/0.45 NA Plan Fluor objective). Cells were plated in 96-well flat bottom plates and incubated overnight in complete medium. Before live-cell imaging, cells were treated with nocodazole. The DNA-damaging compounds were added directly into the medium during the imaging. Images were acquired every 20 min. After 4 h of treatment with DNA-damaging compounds, cells were fixed and stained for phosphohistone H2AX (Ser-139), phosphohistone H3 (Ser-10), and DAPI. The image acquisition was performed in an IN Cell Analyzer 2000 microscope using a Nikon 20×/0.45 NA Plan Fluor objective, and nine fields of view per well were acquired. A custom-written Fiji/ImageJ (Schindelin et al., 2012) script was developed to automatically quantify the fluorescence intensity of γH2AX on mitotic cells. Importantly, after exposure to DNA-damaging compounds, only cells already in mitosis were analyzed.

Time-lapse microscopy

HeLa cells were plated on a 12-well plate (Fisher Scientific) 24 h before imaging. Live-cell imaging experiments were performed at 37°C in an atmosphere of 5% CO₂ using a phase-contrast microscope (Leica DMI6000; Leica Microsystems; 20× objective lens HCX PL FLUOTAR L. CORR. Ph1, 0.40 NA) equipped with a charge-coupled device (CCD) camera (Hamamatsu FLASH4.0; Hamamatsu, Japan). To measure mitotic timing and to determine cell fate in mitotic arrested cells, nocodazole was added 1 h before acquisition. The other drugs were added directly into the medium during the imaging. The images were captured every 20 min for 72 h using LAS X software. In the washout experiment, HeLa cells stably expressing GFP-H2B/mRFP-α-tubulin were seeded 24 h before imaging on a µ-Plate 24-well Black (ibidi) and incubated overnight in phenol red–free DMEM supplemented with 10% FBS and 25 mM HEPES (Life Technologies). Images were acquired every 10 min for 18 h using an Operetta CLS (20×/1.0 NA Water Fluor objective; PerkinElmer) at 37°C with 5% CO₂. All images were analyzed using Fiji/ImageJ.

Immunofluorescence

Cells were fixed with 4% paraformaldehyde (Electron Microscopy Sciences) for 10 min and subsequently extracted in PBS–0.3% Triton (Sigma-Aldrich) for 10 min at room temperature (RT). After short washes in PBS–0.1% Triton, cells were incubated with blocking solution (10% FBS in PBS–0.1% Triton) for 1 h at RT and incubated with primary antibodies diluted in blocking solution overnight at 4°C. Then, cells were washed 3× with PBS–0.1% Triton and incubated with the respective secondary antibodies for 1 h at RT. DNA was counterstained with 1 µg/ml DAPI (Sigma-Aldrich). The coverslips were mounted using mounting medium (20 mM Tris, pH 8, 0.5 mM N-propyl gallate, 90% glycerol) on glass slides. Mouse anti-phosphohistone H2AX Ser-139 (1:2000; Merck), rabbit anti-phosphohistone H3 Ser-10 (1:5000; Cell Signaling Technology), mouse anti-Mad1 (1:500; Merck Millipore), guinea pig anti-CENP-C (1:1000; MBL International Corporation), mouse anti–α-tubulin clone B-512 (1:2000; Sigma-Aldrich), rabbit anti–β-tubulin (1:2000; Abcam), rabbit anti-Bub1 (1:500; Abcam), rabbit anti–phospho-BubR1 Ser-670 (1:1000; Abcam), mouse anti–phospho-Plk1 Thr-210 (1:1000; Abcam), rabbit anti–phospho-Aurora B Thr-232 (1:1000; Rockland Immunochemicals), rabbit anti–phospho-Aurora A Thr-288 (1:1000; Cell Signaling Technology), mouse anti-Kid (1:500; a gift from S. Geley), and rabbit anti-Kif4a (1500; Thermo Fisher Scientific) were used as primary antibodies, and Alexa Fluor 488, 568, and 647 (Invitrogen) were used as secondary antibodies (1:1000). Images were acquired with an AxioImager Z1 (63×; Plan oil differential interference contrast objective lens, 1.4 NA; all from Carl Zeiss) equipped with a CCD camera (ORCA-R2; Hamamatsu Photonics) using Zen software (Carl Zeiss). Forty-one z-planes with a 0.2 μm step covering the entire volume of the mitotic cell were collected. Nocodazole shock experiments in HeLa and U2OS cells were acquired with Opera Phenix (63×/1.15 NA Water Fluor objective; PerkinElmer) and 23 z-planes with a 1 μm step covering the entire volume of the mitotic cell were collected. Autoquant X software (Media Cybernetics) was used for blind deconvolution. Images were analyzed using Fiji/ImageJ and processed in Adobe Photoshop CS6 (Adobe Systems). Representative images were obtained through a maximum-intensity projection of a deconvolved z stack, and Adobe Illustrator CS6 (Adobe Systems) was used for panel assembly for publication.

Fluorescence quantification

For immunofluorescence quantitative measurements, all compared images were acquired using identical acquisition settings. Protein intensity was quantified using Fiji/ImageJ. Briefly, individual kinetochores were detected using CENP-C staining, and a region of interest (ROI) delimitating the kinetochore on the focused Z plan was drawn. The mean fluorescence intensities (pixel gray levels) of phospho-Aurora B at the inner centromere and Bub1, phospho-BubR1, and phospho-Plk1 at kinetochores were measured on the focused z plan. The background was determined with five ROIs drawn outside the kinetochore region, and the averages of these values were subtracted. Fluorescence intensity measurements were normalized to the CENP-C signals. Phospho-Aurora A levels at polar chromosomes was determined by drawing an elliptical ROI around the poles in sum-projected images. The background was determined with five ROIs drawn outside the spindle pole region, and the averages of these values were subtracted. The Kid and Kif4a levels were measured by drawing two elliptical ROIs, one containing the polar chromosomes and the other containing the aligned chromosomes in sum-projected images. The background was determined with five ROIs outside the chromosome region, and the averages of these values were subtracted. The proportions of Kid and Kif4a on polar chromosomes relative to aligned chromosomes were determined. The microtubule depolymerization rate after nocodazole shock in HeLa and U2OS cells was determined based on tubulin texture homogeneity, using Harmony software (PerkinElmer). Briefly, the image analysis workflow consists of the identification of the mitotic cell nuclei from the DAPI channel. Then, the intensity of microtubules on mitotic cells was analyzed based on the occurrence of typical patterns using the “SER Edge” method. In RPE-1 cells, the microtubule depolymerization rate after nocodazole was determined by the proportion of total and soluble α-tubulin levels. The total α-tubulin intensity was measured by drawing a larger, oval-shaped ROI that contained the entire cell in sum-projected images. The soluble α-tubulin levels were determined by drawing five smaller, oval-shaped ROIs outside the chromosome region; the averages of these values were calculated in sum-projected images. The fluorescence intensities were normalized to the level at time = 0 and are represented as a function of time. All values were normalized to the average fluorescence levels of control cells.

Western blot

HeLa cells were resuspended and lysed in ice-cold NP-40 lysis buffer (20 mM HEPES-KOH, pH 7.9, 1 mM EDTA, 1 mM EGTA, 150 mM NaCl, 0.5% NP-40, 10% glycerol, 2 mM dithiothreitol) supplemented with a cocktail of protease and phosphatase inhibitors (Roche). The samples were kept on ice for 30 min and flash frozen in liquid nitrogen. To complement and to increase lysis efficiency, cells were sonicated for 10 cycles of 30 s “on” and 30 s “off” at 4°C with Bioruptor (Diagenode). and the protein concentration was measured with the Bradford protein assay (Thermo Fisher Scientific). Equal protein concentrations were denatured in Laemmli buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.002% bromophenol blue) at 95°C for 5 min, separated into 8–12% SDS–PAGE gels, and transferred to nitrocellulose Hybond-C membranes using the Trans-Blot Turbo Transfer System (Bio-Rad). Membranes were blocked with 5% milk in PBS–0.1% Tween-20 (PBST; Sigma-Aldrich) at RT for 1 h, and all primary antibodies were incubated overnight at 4°C with agitation. After five washes in PBST, the membranes were incubated with the secondary antibodies for 1 h at RT. After several washes with PBST, the detection was performed with a Clarity Western ECL Blotting Substrate (Bio-Rad). Mouse anti-phosphohistone H2AX Ser-139 (1:2000; Merck), rabbit anti-phosphohistone H3 Ser-10 (1:5000; Cell Signaling Technology), mouse anti-Mad1 (1:1000; Merck), rabbit anti-Kif4a (1:1000; Thermo Fisher Scientific), mouse anti-Kid (1:1000; a gift from S. Geley), mouse anti-MPM2 (1:1000; Merck), and mouse anti-GAPDH (1:50,000; Proteintech) were used as primary antibodies, and anti-mouse-horseradish peroxidase (HRP) and anti-rabbit-HRP were used as secondary antibodies (1:5000; Jackson ImmunoResearch Laboratories).

Microtubule polymerization assay

Tubulin polymerization was monitored using a tubulin polymerization kit as specified by the manufacturer (Cytoskeleton). Briefly, purified tubulin (3 mg/ml) was polymerized in GTP buffer (80 mM PIPES, pH 6.9, 2 mM MgCl₂, 0.5 mM EGTA, 1 mM GTP, 10.2% glycerol) in the presence of 10 µM lomustine, mitomycin C, carboplatin, or etoposide at 37°C in a temperature-regulated spectrophotometer (VICTOR Nivo; PerkinElmer). The absorbance (at 340 nm) was measured every minute for 120 min. Taxol and nocodazole were used as a known enhancer and inhibitor of tubulin polymerization, respectively. The maximal slope values (V_max), which represent the maximal growth rates achieved during the polymerization reactions, were calculated for each condition.

Colony formation assay

Cells were plated on T25 flasks (Sarstedt) for 24 h in complete medium. Cells were then blocked in mitosis with nocodazole for 3–8 h, according to the cell line. Mitotic cells were collected by mechanical shake off. The isolated mitotic cells were exposed to selected DNA-damaging compounds for 4 h. After this period, cells were rinsing thrice with PBS and the cells were seeded in six-well plates (Fisher Scientific) at 100 cells/well and grown for 14–28 d until colonies had formed. The colonies were counted, and the surviving fraction was calculated as the ratio between the number of colonies and the number of cells seeded.

Statistical analysis

Statistical analysis was performed using GraphPad Prism, version 8. All results presented in this article were obtained by pooling data from at least two independent experiments, unless stated otherwise. All data represent mean ± SD, and values were obtained across experiments or cells as indicated in the figure legends. The D’Agostino–Pearson omnibus normality test was used to determine whether the data followed a normal distribution. p values were calculated with either a t test or multiple comparisons one-way ANOVA (for data that followed a normal distribution) and Mann–Whitney Rank Sum test or a multiple comparison Kruskal–Wallis test (for data that did not follow a normal distribution). Quantifications of Mad1 at kinetochore were analyzed using the Fisher’s exact two-tailed test. For the comparison of the single exponential fitting curve, an extra sum-of-squares F test was used. For each graph, n.s. = not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001.

Abbreviations used:

ATM

ataxia telangiectasia mutated

53BP1

p53-binding protein 1

CCD

cool charged device

DAPI

4’,6-diamidino-2-phenylindole

DDR

DNA damage response

DMSO

dimethyl sulfoxide

DSBs

double strand breaks

FBS

fetal bovine serum

γH2AX

phosphorylated histone variant H2AX on Ser-139

HRS

horseradish peroxidase

MDC1

mediator of the DNA damage checkpoint 1

MRN

MRE11-RAD50-NBS1

NEBD

nuclear envelope breakdown

PBS

phosphate-buffered saline

PBST

phosphate-buffered saline 0.1% Tween

pH3S10

phosphorylated histone 3 on Ser-10

RNF8

ring finger protein 8

ROI

region of interest

RT

room temperature

SAC

spindle assembly checkpoint

SD

standard deviation

siRNA

small interfering RNA

STLC

S-trityl- l-cysteine