1. Cancer Biology
  2. Cell Biology

Inhibitors of Rho kinases (ROCK) induce multiple mitotic defects and synthetic lethality in BRCA2-deficient cells

  1. Julieta Martino
  2. Sebastián Omar Siri
  3. Nicolás Luis Calzetta
  4. Natalia Soledad Paviolo
  5. Cintia Garro
  6. Maria F Pansa
  7. Sofía Carbajosa
  8. Aaron C Brown
  9. José Luis Bocco
  10. Israel Gloger
  11. Gerard Drewes
  12. Kevin P Madauss
  13. Gastón Soria  Is a corresponding author
  14. Vanesa Gottifredi  Is a corresponding author
  1. Fundación Instituto Leloir-CONICET, Argentina
  2. Centro de Investigaciones en Bioquímica Clínica e Inmunología, CIBICI-CONICET, Departamento de Bioquímica Clínica, Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, Argentina
  3. OncoPrecision, Argentina
  4. Center for Molecular Medicine, Maine Medical Center Research Institute, United States
  5. GlaxoSmithKline-Trust in Science, Global Health R&D, United Kingdom
  6. GlaxoSmithKline-Trust in Science, Global Health R&D, United States
https://doi.org/10.7554/eLife.80254
Share this article
Cite this article
  1. Julieta Martino
  2. Sebastián Omar Siri
  3. Nicolás Luis Calzetta
  4. Natalia Soledad Paviolo
  5. Cintia Garro
  6. Maria F Pansa
  7. Sofía Carbajosa
  8. Aaron C Brown
  9. José Luis Bocco
  10. Israel Gloger
  11. Gerard Drewes
  12. Kevin P Madauss
  13. Gastón Soria
  14. Vanesa Gottifredi
(2023)
Inhibitors of Rho kinases (ROCK) induce multiple mitotic defects and synthetic lethality in BRCA2-deficient cells
eLife 12:e80254.
https://doi.org/10.7554/eLife.80254
  2. Download BibTeX
  3. Download .RIS

Abstract

The trapping of Poly-ADP-ribose polymerase (PARP) on DNA caused by PARP inhibitors (PARPi) triggers acute DNA replication stress and synthetic lethality (SL) in BRCA2-deficient cells. Hence, DNA damage is accepted as a prerequisite for SL in BRCA2-deficient cells. In contrast, here we show that inhibiting ROCK in BRCA2-deficient cells triggers SL independently from acute replication stress. Such SL is preceded by polyploidy and binucleation resulting from cytokinesis failure. Such initial mitosis abnormalities are followed by other M phase defects, including anaphase bridges and abnormal mitotic figures associated with multipolar spindles, supernumerary centrosomes and multinucleation. SL was also triggered by inhibiting Citron Rho-interacting kinase, another enzyme that, similarly to ROCK, regulates cytokinesis. Together, these observations demonstrate that cytokinesis failure triggers mitotic abnormalities and SL in BRCA2-deficient cells. Furthermore, the prevention of mitotic entry by depletion of Early mitotic inhibitor 1 (EMI1) augmented the survival of BRCA2-deficient cells treated with ROCK inhibitors, thus reinforcing the association between M phase and cell death in BRCA2-deficient cells. This novel SL differs from the one triggered by PARPi and uncovers mitosis as an Achilles heel of BRCA2-deficient cells.

Editor's evaluation

This paper reports the fundamental discovery that BRCA2-deficient cells are highly sensitive to the inhibition or depletion of Rho-kinases (ROCK), known to regulate actin cytoskeleton dynamics. This observed synthetic lethality between ROCK and BRCA2 is suggested to be independent of acute replication stress, is outside of the cellular S phase and may represent a promising new synthetic lethality target for the treatment of BRCA2-deficient tumors.

https://doi.org/10.7554/eLife.80254.sa0

Introduction

Hereditary breast and ovarian cancer (HBOC) is an autosomal dominant disease that accounts for 5–10% of breast (Krainer et al., 1997; Langston et al., 1996) and 15% of ovarian cancer cases (Pal et al., 2005; Zhang et al., 2011). HBOC is primarily caused by mutations in the breast cancer susceptibility genes BRCA1 and BRCA2 (Futreal et al., 1994; Miki et al., 1994; Wooster et al., 1995). BRCA1 and BRCA2 are DNA repair genes, and their protein products regulate homologous recombination (HR), a repair pathway that is recruited to highly toxic DNA double-strand breaks (DSBs; Prakash et al., 2015). Additionally, BRCA1 and BRCA2 are essential for DNA replication events, including replication fork protection, reversal, restart and gap-filling (Cong and Cantor, 2022; Cong et al., 2021; Panzarino et al., 2021; Ray Chaudhuri et al., 2016b; Schlacher et al., 2012). BRCA1- and BRCA2-deficient cells exhibit structural chromosome abnormalities and are highly sensitive to DNA-damaging agents (Moynahan et al., 2001; Patel et al., 1998; Yu et al., 2000). Additionally, BRCA-deficient cells exhibit translocations, large deletions and chromosome fusions (Moynahan et al., 2001; Yu et al., 2000). This chromosome instability underlies the tumorigenicity of BRCA-deficient tumors and underscores the critical tumor suppressor function of BRCA genes in cells.

Mutations in BRCA genes are highly penetrant, and their carriers have a high risk of developing early-onset breast and ovarian cancer (Antoniou et al., 2003; King et al., 2003). Carriers of BRCA mutations are also at an increased risk of developing other malignancies, including pancreatic and prostate cancers and melanoma (Cavanagh and Rogers, 2015; Gumaste et al., 2015). BRCA-mutation carriers whose mutations are detected before cancer onset are suggested to undergo highly invasive surgeries such as salpingo-oophorectomy and mastectomy. The standard of care for BRCA-mutation carriers with tumors is similar to the approach used for patients with sporadic tumors, except for some types of BRCA-deficient tumors, which might be more sensitive to platinum-based therapies (Vencken et al., 2011; Yang et al., 2011). Unfortunately, chemotherapy resistance to platinum agents is common and alternative therapies are most needed for these patients.

One group of alternative therapeutic agents that are clinically available is poly-ADP-ribose polymerase (PARP) inhibitors which are highly effective in killing BRCA-deficient cells (Bryant et al., 2005; Farmer et al., 2005; McCabe et al., 2006) and several PARP inhibitors (PARPi) have been approved for clinical use. The synthetic lethality (SL) observed between BRCA deficiency, and PARPi is due to the ability of PARPi to physically trap PARP on DNA (Murai et al., 2014; Murai et al., 2012). PARP trapping causes the accumulation of DNA replication intermediates, such as gaps, which must be handled by BRCA proteins to protect DNA integrity (Taglialatela et al., 2021; Tirman et al., 2021). Additionally, some DNA structures that derive from the encounter of replication forks with PARP-bound DNA may require HR-mediated repair, a mechanism impaired in BRCA1- and BRCA2-deficient cells (Prakash et al., 2015). While the impaired DNA damage response of BRCA-deficient cells to PARPi leads to cell death, resistance to PARPi is also observed in the clinic (Barber et al., 2013). Molecular mechanisms of resistance to PARPi include, but are not limited to, secondary mutations that restore HR function, increased drug efflux, and decreased PARP trapping (D’Andrea, 2018; Noordermeer and van Attikum, 2019).

As mentioned above, although BRCA proteins were initially studied based on their roles in HR, we currently know that BRCA1 and BRCA2 have pleiotropic functions, performing functions outside canonical HR (Petsalaki and Zachos, 2020). Thus, it is likely that multiple targets not restricted to HR could be exploited for SL therapeutic approaches. This concept has been corroborated for BRCA1 deficiency in a phenotypic screening in which we tested BRCA-deficient cells for SL against the kinase inhibitor library PKIS2 (Carbajosa et al., 2019). Our findings unveiled that BRCA1-deficient cells have increased sensitivity to Polo-like kinase 1 (PLK1) inhibitors and that this sensitivity does not require excess DNA damage caused by external agents.

In this study, we present findings indicating that BRCA2-deficient cells are highly sensitive to the inhibition or depletion of Rho-kinases (ROCK), which regulate actin cytoskeleton dynamics. Unlike PARPi, ROCK inhibitors (ROCKi) did not induce acute replication stress in BRCA2-deficient cells but instead triggered mitotic defects including cytokinesis failure, polyploidy, aberrant multipolar spindles and centrosome amplification. Remarkably, SL-induction was also observed after inhibition of Citron Rho-interacting kinase (CITK), an enzyme that regulates cytokinesis at the level of mitotic furrow cleavage, indicating that cytokinesis failure is the likely trigger of this novel SL interaction. Moreover, preventing mitotic entry via depletion of Early mitotic inhibitor 1 (EMI1), abrogated ROCKi-induced BRCA2-deficient cell death. In conclusion, while the accumulation of DNA damage in S phase is required for PARPi-mediated cell death (Ray Chaudhuri et al., 2016a; Schoonen et al., 2017), our findings highlight that BRCA2-deficient cells bear additional vulnerabilities outside S phase that could represent promising new SL targets.

Results

BRCA2-deficient cells are sensitive to ROCK inhibition

In a previous work (Carbajosa et al., 2019), we developed a phenotypic survival screening method to evaluate the differential sensitivity of BRCA1-deficient cells against 680 ATP-competitive kinase inhibitors provided by GlaxoSmithKline (Drewry et al., 2017; Elkins et al., 2016). Briefly, the screening was performed using HCT116p21-/- cell lines in which BRCA1 or BRCA2 were stably downregulated using shRNA (Figure 1A). This strategy allowed a comparison of BRCA-proficient vs BRCA-deficient cell lines on an isogenic background. In addition, HCT116p21-/- cells are easy to grow and tolerate low seeding densities compatible with long-term (i.e. 6 days) survival analysis. Additionally, we used a p21 knockout background, which attenuates the cell cycle arrest that otherwise would mask the cytotoxic phenotypes during the screening time frame.

Figure 1
Download asset Open asset
Phenotypic screening identifies ROCK kinases as potential targets for synthetic lethality in BRCA2 cells.

(A) The screening assay is based on the co-culture of isogeneic BRCA-proficient and BRCA-deficient cell lines in equal proportions on each well of 96-well plates. Such cell lines were generated as double stable cell lines tagged with different fluorescent proteins (CFP, iRFP, and mCherry) and expressing shRNAs for Scramble, BRCA1 or BRCA2 were generated as described in Carbajosa et al., 2019. (B) Quantitative real-time PCR of BRCA2 in shScramble and shBRCA2 HCT116p21-/- cells (N=2). Statistical analysis was performed with a two-way ANOVA test followed by a Bonferroni post-test (*p<0.05, **p<0.01, ***p<0.001). (C) Relative cell number (%) of HCT116p21-/- cells expressing shScramble and shBRCA2 and treated with the indicated concentrations of olaparib (N=2). (D) Representative results expressed as RL-3-BL-3 dot plots (log scale, RL-3 780/60 nm filter, and BL-3 695/40 nm filter). A tested compound can be ‘non-synthetic lethal’ (the ratio between the populations' percentage remains unchanged when compared to the ratio used for seeding ~33% for each cell line); or ‘synthetic lethal’ (the ratio between cell types is altered when compared to the ratio used for seeding, with selective depletion of cells within the BRCA1- and/or BRCA2-deficient populations). (E) Screening results of PKIS2 library compounds (0.1 μM) in shBRCA2 HCT116p21-/- cells. Compounds were plotted based on their fold of SL (y axis) and their survival difference (x axis). A compound was considered a ‘hit’ if it exhibited a >5 standard deviations on these two variables. Fold of SL (y-axis): the ratios of the different populations in each individual well. Survival difference (x-axis): compares treated cells with the untreated control in the same plate. ROCK inhibitors and other inhibitors are plotted in red and gray, respectively. Please refer to Carbajosa et al., 2019 for statistical analysis of the screening. (F) Relative cell number (%) of shScramble and shBRCA2 HCT116p21-/- cells at different ROCK inhibitors. Data are shown as the average of independent experiments with the standard error of the mean.

In this work, we analyzed the screening results of the BRCA2-deficient cell population. BRCA2 depletion by shRNA in HCT116p21-/- cells was sufficient to trigger increased sensitivity to olaparib (Figure 1B–C). For the analysis, we focused on compounds that induced SL exclusively in the BRCA2-deficient population and were not toxic to control samples or BRCA1-deficient cells (Figure 1D). Interestingly, BRCA2-deficient cells showed remarkable sensitivity to three inhibitors of ROCK kinases (ROCK) (Figure 1E and Table 1). The selective activity of all ROCK inhibitors was further validated at a higher dose (Table 1) and in a dose-response curve for the three most potent ones (Figure 1F).

Table 1
Phenotypic screening identifies ROCK kinases as potential targets for synthetic lethality in BRCA2 cells.

(A) Table listing all ROCK inhibitors from the PKIS2 library and their corresponding survival difference at 0.1 and 1 μM.

Survival difference
Inhibitor0.1 µM1 µM
GSK180736A08.15
GSK248233B47.5741.99
GSK269962B25.5828.49
GSK270822A038.12
GSK429286A0.2918.11
GSK466314A025.41
GSK534911A25.533.72
GSK534913A032.50
SB-772077-B067.80

To test the sensitivity of BRCA2-depleted cells to ROCK inhibition, we took advantage of three commercially available ROCK inhibitors (ROCKi). Two of them are fasudil and ripasudil, which are approved for diseases other than cancer (Garnock-Jones, 2014; Shi and Wei, 2013). Both are ATP-competitive inhibitors targeting ROCK1 and ROCK2 (Nakagawa et al., 1996). In addition, we used the inhibitor SR 3677 dihydrochloride, which is a newer ROCK inhibitor that has interesting advantages such as a low IC50 and high potency in biochemical and cell-based assays as well as high selectivity for ROCK (Feng et al., 2008). We performed survival assays with fasudil in several cellular models of BRCA2 deficiency, including the HCT116p21-/- cell line used in the screening (Figure 2A). We also tested survival in DLD-1/DLD-1BRCA2-/- paired cell lines, which are BRCA2 knockout (Figure 2B) and the PEO4/PEO1, V-C8 #13 /V-C8 paired cell lines (see description of cell lines in the methods section - Figure 2C–D). SL was observed in all BRCA2-deficient cell line models following fasudil treatment (Figure 2A–D). Cell death was confirmed using SYTOX green, a dye that only enters cells when cellular membranes have been compromised (Figure 2E) and in clonogenic survival assays (Figure 2—figure supplement 1). Similar differences between control and BRCA2-deficient counterparts were observed with ripasudil and SR 3677 dihydrochloride, two other ROCKi (Figure 2—figure supplement 2A–C). In contrast, the BRCA1-deficient cell line HCC1937 (Tomlinson et al., 1998), which is sensitive to olaparib (Figure 2—figure supplement 2D), did not exhibit increased sensitivity to fasudil or ripasudil compared to the complemented HCC1937BRCA1 cell line (Treszezamsky et al., 2007; Figure 2—figure supplement 1E–F). Similar results were observed using HCT116 cellular models depleted from BRCA1 (Figure 2—figure supplement 2G–I). The unique sensitivity of BRCA2-deficient cells to ROCKi suggests that the SL observed is likely independent of the homologous recombination function of BRCA2.

Figure 2 with 3 supplements see all
Download asset Open asset
BRCA2-deficient cells are selectively killed by the ROCK kinase inhibitor fasudil.

(A) Relative cell number (%) of shScramble and shBRCA2 HCT116p21-/- cells after 6 days of treatment with fasudil (N=3). (B) Relative cell number (%) of DLD-1WTand DLD-1BRCA2-/- after 6 days of treatment with fasudil (N=2). (C) Relative cell number (%) of V-C8#13 and V-C8 cells after 6 days of treatment with fasudil (N=3). (D) Relative cell number (%) of PEO4 and PEO1 cells after 6 days of treatment with fasudil (N=4). Panels A-D: the cell cartoon shows the BRCA2 status caused by the modification introduced at last in each pair of cell lines (see Materials and methods for further details). Black borders indicate that the modification generated a BRCA2 proficient status and blue borders aBRCA2 deficiency. (E) FACS analysis of SYTOX green-stained PEO4 and PEO1 cells 6 days after fasudil treatment (128 μM, N=2). Statistical analysis was performed with a two-way ANOVA test followed by a Bonferroni post-test (*p<0.05, **p<0.01, ***p<0.001). Data in A-D are shown as the average of independent experiments with the standard error of the mean.

Importantly, we observed strong SL by ROCKi in growing conditions that triggered only mild sensitivity to PARPi. While HCT116p21-/- shBRCA2, V-C8 and DLD-1BRCA2-/- were all sensitive to olaparib (Figure 2—figure supplement 3A), PEO1 showed only modest sensitivity to olaparib in our experimental conditions (Figure 2—figure supplement 3B), despite reports indicating sensitivity to PARPi (Sakai et al., 2009; Stukova et al., 2015; Whicker et al., 2016). We confirmed that PEO1 were BRCA2-deficient. The BRCA2 mutation in PEO1 (5193C>G) creates a premature stop codon and also a digestion site for the enzyme DrdI. In contrast, the reversion mutation in PEO4 (5193C>T) abolishes this site (Figure 2—figure supplement 3C). Consistent with their expected point mutation, following DrdI digestion PEO1 cells showed two DNA fragments (480 bp and 214 bp), which were not observed in PEO4 cell lines (Figure 2—figure supplement 3D). Additionally, as previously reported for BRCA2-deficient cell lines (Sakai et al., 2009; Stronach et al., 2011; Stukova et al., 2015; Whicker et al., 2016) PEO1 cells are sensitive to cisplatin (Figure 2—figure supplement 3E). Our results suggest that while clonogenic assays and other approaches may better expose the sensitivity of PEO1 to olaparib, strong SL induced by ROCKi is observed in growing conditions that reveal only mild sensitivity to PARPi. Hence, synthetic lethal avenues that diverge from PARPi could provide efficient therapeutic alternatives for treating BRCA2-deficient cancer cells.

Replication stress is not the major driver of SL between BRCA2 deficiency and ROCK inhibition

The SL observed between BRCA deficiency and PARPi is preceded by the accumulation of acute replication stress caused by PARP trapping on the DNA (Murai et al., 2012; Schoonen et al., 2017). As BRCA-deficient cells keep progressing across S phase in the presence of PARPi, PARP/DNA adducts exacerbate replication stress resulting from fork stalling, gap formation and fork collapse (Kolinjivadi et al., 2017; Lemaçon et al., 2017; Mijic et al., 2017; Panzarino et al., 2021; Schlacher et al., 2011; Taglialatela et al., 2017). Consistent with those reports, the treatment of HCT116p21-/- shBRCA2 cells with olaparib caused the acute accumulation of replication stress markers such as γH2AX and 53BP1 nuclear foci, which represent sites of DSB formation in S phase (Figure 3A–C and Figure 3—figure supplement 1A–B). In striking contrast to olaparib, no increase in 53BP1 or γH2AX foci was induced by fasudil treatment in HCT116p21-/- shBRCA2 cells (Figure 3A–C and Figure 3—figure supplement 1A–B) at this time. These results were also validated in PEO cells (Figure 3D and Figure 3—figure supplement 1C). In line with the lack of acute replication stress, we did not observe alterations in DNA replication parameters, such as nascent DNA track length or the frequency of origin firing after fasudil treatment (Figure 3F–G and positive controls in Figure 3—figure supplement 1D). We also did not observe differences in the percentage of BrdU+ cells after 3 or 6 days of fasudil treatment compared to untreated cells (Figure 3E). Additionally, the intensity of BrdU, a parameter than reveals subtle alterations of the DNA replication program undetectable by the DNA fiber assay (Calzetta et al., 2021), was also unaffected (Figure 3—figure supplement 1E). Given that the synthetic lethality of fasudil was more evident 6 days post-treatment, we evaluated whether fasudil causes replicative stress at that time, and observed no evidence of augmented γH2AX intensity or 53BP1 focal organization in HCT116p21-/- shBRCA2 and PEO1 (BRCA2-/-) 6 days post-treatment (Figure 3—figure supplement 2). These findings point toward a cell death mechanism independent from the accumulation of DNA damage in S phase.

Figure 3 with 2 supplements see all
Download asset Open asset
Fasudil does not induce acute replication stress in BRCA2-deficient cells.

(A) yH2AX intensity/cell of shScramble or shBRCA2 HCT116p21-/- cells (N=2). (B) Representative images of yH2AX intensity in single cells. (C) Number of 53BP1 foci/cell in shScramble and shBRCA2 HCT116p21-/- cells (N=2). (D) yH2AX intensity/cell in PEO1 or PEO4 cells (N=2). (E) Percentage of PEO4 and PEO1 cells stained with BrdU at 3 and 6 days after fasudil treatment (128 μM, N=2). A total of 500 cells were analyzed for each sample. Representative images of PEO1 cells after 3 days of fasudil treatment (BrdU shown in green, DAPI shown in blue). (F) Labelling scheme and IdU track lengths of shScramble and shBRCA2 HCT116p21-/- cells, treated with fasudil for 48 h (N=2). Representative images of individual DNA fibers are shown on the left side of the panel. (G) Origin firing frequency (percentage) of shScramble or shBRCA2 HCT116p21-/- cells in samples showed in E (N=2). Statistical analysis was performed using a two-way ANOVA test followed by a Bonferroni post-test (*p<0.05, **p<0.01, ***p<0.001). Data are shown as the average of independent experiments with the standard error of the mean.

ROCK inhibition induces mitotic defects in BRCA2-deficient cells

To further characterize such a replication stress-independent SL, we analyzed cell cycle profiles with propidium iodide staining. Consistent with reduced survival at 6 days (Figure 2), in BRCA2-deficient cells, we observed a sub-G1 peak after fasudil treatment indicative of apoptotic cell death (Figure 4A–B). In terms of cell cycle distribution, BRCA2-deficient cells treated with fasudil exhibited an accumulation of cells in G2/M indicative of a G2/M arrest (Figure 4A–B). Intriguingly, BRCA2-deficient cells also exhibited a peak of >4N polyploid cells (Figure 4A–B). By performing a detailed time course in which samples were collected in 24 h intervals, we observed that the polyploidy phenotype was cumulative (Figure 4C). While the G2/M arrest in BRCA2-deficient cells appeared as early as 24 h post-treatment, polyploidy became strongly evident after 72 h (i.e.: 3 days). The sub-G1 population was also evident as early as 24 h post-treatment but increased at longer time points after polyploidy detection (i.e.: after 3 days). These data suggest that the accumulation of cells in G2/M precedes both polyploidy and cell death.

Figure 4
Download asset Open asset
Fasudil treatment induces polyploidy and aberrant mitotic figures in BRCA2-deficient cells.

(A–B) Cell cycle analysis of PEO4 and PEO1 cells following 3 or 6 days of fasudil treatment (96 and 128 μM; N=3). Cells were stained with propidium iodide, and DNA content was analyzed via FACS (10,000 events per sample). (C) Cell cycle analysis of PEO4 and PEO1 cells following a time course with fasudil treatment (N=2; 1–5 days, 64 μM). Cells were stained with propidium iodide, and DNA content was analyzed via FACS (10,000 events per sample). (D) Representative images of DAPI-stained normal and aberrant metaphases. Aberrant metaphases include metaphases with DNA being pulled in multiple directions or metaphases with misaligned chromosomes. (E) Percent of aberrant metaphases in PEO4 and PEO1 cells 3 or 6 days after fasudil treatment (128 μM; N=3). A total of 100 metaphases were analyzed for each sample. Statistical analysis was performed using a two-way ANOVA test followed by a Bonferroni post-test (*p<0.05, **p<0.01, ***p<0.001). Data are shown as the average of independent experiments with the standard error of the mean.

The concomitant accumulation of cells in G2/M (which could also include G1 cells with duplicated DNA content) and the DNA content >4N is highly suggestive of problems in the correct finalization of M phase, which leads to the accumulation of aberrant mitotic phenotypes. Consistent with this, after fasudil treatment, BRCA2-deficient cells exhibited an increase in metaphases in which the DNA was being pulled in multiple directions or in which the chromosomes were not aligned in the metaphase plate (Figure 4D–E). Altogether, these data pinpoint a dysregulated mitosis in BRCA2-deficient cells treated with ROCKi.

Aberrant metaphases can be triggered by unresolved DNA replication defects accumulated after DNA replication stress (Gelot et al., 2015), but can also be prompted within M phase as a consequence of aberrant mitotic spindle organization or disorganized chromosome alignment (Bakhoum et al., 2009; Shindo et al., 2021; Siri et al., 2021). Aberrant anaphases (bridges and lagging chromosomes; Figure 5A) can also be triggered either by replication defects not resolved before M phase entry or intrinsic mitotic defects dissociated from S phase (Bakhoum et al., 2009; Shindo et al., 2021). We documented an increase in chromosome bridges, but not in lagging chromosomes, after fasudil treatment of BRCA2-deficient cells (Figure 5B–C). To confirm the increase of chromosome bridges observed with fasudil, we used commercially available siRNAs against ROCK1 and ROCK2 (Figure 5D). Similarly to ROCKi, ROCK1 and ROCK2 (ROCK1/2) depletion promoted the accumulation of anaphase bridges in BRCA2-deficient cells (Figure 5E). Importantly, when resulting from unresolved replication defects, anaphase aberrations are typically accompanied by chromosome aberrations (i.e. breaks, exchanges) and micronuclei (Finardi et al., 2020; Utani et al., 2010). However, we did not find any indication of chromosome aberrations or micronuclei in fasudil-treated BRCA2-deficient cells (Figure 5—figure supplement 1A–B), suggesting that the trigger for anaphase bridge formation following fasudil treatment is a defect intrinsic to M phase.

Figure 5 with 1 supplement see all
Download asset Open asset
Mitotic DNA bridges accumulate in BRCA2-deficient cells following ROCK inhibition with fasudil.

(A) Representative images of normal and abnormal anaphases with bridges and lagging chromosomes. (B) Percentage of anaphases with chromosomes bridges and lagging chromosomes in PEO4 and PEO1 cells treated with fasudil (128 μM). Fifty to 70 anaphases per sample were analyzed in two independent experiments (N=2). (C) Percentage of anaphases with chromosomes bridges and lagging chromosomes in shScramble- or shBRCA2-transduced HCT116p21-/- cells treated with fasudil. 50–70 anaphases per sample were analyzed per independent experiment (N=3). (D) Quantitative real-time PCR of ROCK1 and ROCK2 in shBRCA2 HCT116p21-/- cells transfected with 150 μM of siROCK1 or siROCK2 (N=2). (E) Percentage of anaphases with chromosomes bridges and laggards in shBRCA2 HCT116p21-/- cells transfected with siROCK (1+2). A total of 50–70 anaphases per sample were analyzed in three independent experiments (N=2). The statistical analysis of the data was performed with a two-way ANOVA test followed by a Bonferroni post-test (*p<0.05, **p<0.01, ***p<0.001). Data are shown as the average of independent experiments with the standard error of the mean.

ROCK inhibition causes cytokinesis failure in BRCA2-deficient cells

Since BRCA2-deficient cells treated with ROCKi accumulate M phase defects, we explored the link between ROCK and mitosis. ROCK are crucial regulators of the actin cytoskeleton (Julian and Olson, 2014). ROCK have been implicated in regulating the contraction of the actin cytoskeleton towards the end of mitosis and its downregulation or absence induces multinucleation due to cytokinesis failure (Daniels et al., 2004; Jonsdottir et al., 2009; Mondal et al., 2012; Shive et al., 2010). On the other hand, BRCA2 localizes to the midbody during cytokinesis and its downregulation or absence was also reported to induce multinucleation (Lekomtsev et al., 2010). To explore whether a convergent defect triggers cytokinesis failure after ROCK inhibition in BRCA2-deficient cells, we stained the actin cytoskeleton with phalloidin to distinguish the cytoplasm of individual cells and analyzed the formation of binucleated as well as multinucleated cells after fasudil treatment (Figure 6A). We observed a marked increase of binucleation in BRCA2-deficient cells following fasudil treatment (Figure 6B–C). Also, we documented an increase of multinucleation in BRCA2-deficient cells transfected with siROCK (Figure 6—figure supplement 1A–B). Consistent with the polyploidy (>4N) observed with flow cytometry, fasudil treatment also increased the percentage of multinucleated cells with 3, 4, or 5+ nuclei (Figure 6B–C). Similar to the polyploidy observed in the cell cycle profiles, the proportion of multinucleated cells was more severe at later endpoints (Figure 6B–C), suggesting that despite cytokinesis failure, binucleated cells continue to cycle, thus further increasing their DNA content. Indeed, the percentage of BRCA2-deficient binucleated cells transiting S phase, as revealed by cyclin A staining, was between 30 and 40% irrespective of ROCKi. This result indicates that despite their diploid DNA content, BRCA2-deficient cells treated with fasudil were able to start a new cell cycle and transit through a second S phase (Figure 6—figure supplement 1C–D).

Figure 6 with 1 supplement see all
Download asset Open asset
BRCA2-deficient cells exhibit cytokinesis failure, centrosome amplification and multipolar mitotic spindles following fasudil treatment.

(A) Representative pictures of PEO1 cells after fasudil treatment. Nuclei are stained with DAPI (shown in blue), and the cytoplasm of individual cells is stained with phalloidin which stains the actin cytoskeleton (shown in green). (B) Percent of binucleated and multinucleated PEO4 and PEO1 cells after 3 days of fasudil treatment (N=3, 128 μM). (C) Percent of binucleated and multinucleated number of PEO4 and PEO1 cells after 6 days of fasudil treatment (N=3, 128 μM). A total of 200 cells were analyzed per sample. (D) Representative pictures of PEO1 metaphases showing cells with normal and abnormal mitotic spindles. DNA, centrosomes, and microtubules are shown in blue, red, and green, respectively. (E) Percent of metaphases in PEO4 and PEO1 cells with multipolar spindles after 3 days of fasudil treatment (N=3, 128 μM). (F) Percent of metaphase in PEO4 and PEO1 cells with multipolar spindles after 6 days of fasudil treatment (N=2, 128 μM). Mitotic spindles were visualized by staining centrosomes (γ-tubulin) and microtubules (α-tubulin) and DNA was stained with DAPI. Cells were classified as having multipolar spindle (3, 4, or 5 or more spindles). A total of 100 metaphases were analyzed per sample. Statistical analysis was performed using a two-way ANOVA test followed by a Bonferroni post-test (*p<0.05, **p<0.01, ***p<0.001). Data are shown as the average of independent experiments with the standard error of the mean.

One immediate consequence of cytokinesis failure is that the resulting cell contains two centrosomes instead of one (Ganem et al., 2007). Normal cells harbor one centrosome, which duplicates only once during S phase. During normal mitosis, duplicated centrosomes form a bipolar mitotic spindle ensuring equal chromosome distribution in daughter cells (Nigg, 2007). In contrast, multiple centrosomes can lead to multipolar mitosis and cell death (Ganem et al., 2009). We stained cells for gamma-tubulin and alpha-tubulin, central components of centrosomes and microtubules, respectively (Brinkley, 1997; Fuller et al., 1995) and focused the analysis on mitotic cells. BRCA2-deficient cells treated with fasudil exhibited increased numbers of multipolar mitosis that correlated with increased centrosome number (i.e.:>2; Figure 6D–F). Similar to previously observed phenotypes, such as aberrant metaphases, binucleated cells and polyploidy, the percentage of multipolar mitosis increased at later endpoints (Figure 6F). Together, these results suggest that the cytokinesis failure and altered centrosome numbers lead to multipolar mitosis, which could trigger cell death in fasudil-treated BRCA2-deficient cells.

Cytokinesis failure sensitizes BRCA2-deficient cells to cell death

The results described in Figure 4A–C, Figure 6—figure supplement 1C–D, and Figure 6 indicate that the treatment of BRCA2-deficient cells with ROCKi causes cytokinesis failure and triggers the accumulation of binucleated cells with proliferation capacity. The implications are that cells with >4N DNA content die when attempting to duplicate aberrantly duplicated DNA or when assembling aberrant mitotic spindles in the subsequent mitosis. Supporting such a model is the time course in Figure 7A–C. A significant change in the binucleation of BRCA2-deficient cells was observed as early as 24 h post-fasudil (Figure 7A), while a significant increase of aberrant anaphases and mitosis was detected later on, at 48 h (Figure 7B–C). Surprisingly, binucleation-related cell death is not triggered in control cells, even at doses of ROCKi that kill BRCA2-proficient cells (Figure 7—figure supplement 1A–B). Hence, these results support the likelihood of cytokinesis failure as the trigger for the SL caused by ROCK inhibition in BRCA2-deficient cells.

Figure 7 with 1 supplement see all
Download asset Open asset
Binucleation precedes anaphase and mitotic aberrations in BRCA2-deficient cells.

(A) Percent of binucleated PEO1 and PEO4 cells treated with fasudil at the indicated time points after treatment (N=2). (B) Percent of aberrant anaphases in PEO1 and PEO4 cells treated with fasudil at the indicated time points after treatment (N=2). (C) Percent of mitotic aberrations in PEO1 and PEO4 cells treated with fasudil at the indicated time points after treatment (N=2). For panels A to C, statistical analysis was performed using a two-way ANOVA test followed by a Bonferroni post-test (*p<0.05, **p<0.01, ***p<0.001). Data are shown as the average of independent experiments with the standard error of the mean. (D) Representative scheme of the results obtained in A-C.

If cytokinesis defects caused by ROCKi are the trigger of BRCA2-deficient SL, targeting other factors that regulate cytokinesis should also induce cell death. To test this hypothesis, we downregulated Citron Rho-interacting kinase (CITK), an enzyme that is highly enriched in the midbody during cytokinesis (Madaule et al., 1998; Sahin et al., 2019; Figure 8A). CITK is required for proper RhoA localization at the cleavage site during late cytokinesis (Sahin et al., 2019). Similar to the phenotypes of siROCK1/2, CITK downregulation reduced cell survival of BRCA2-deficient cells (Figure 8B and Figure 8—figure supplement 1A). In addition, and recapitulating the effect of ROCK inhibition or depletion, CITK downregulation increased the number of multinucleated cells in BRCA2-deficient cells (Figure 8C). Most remarkably, combined silencing of CITK and ROCK1/2 was not additive/synergic (Figure 8B), suggesting that ROCK and CITK depletion induce synthetic lethality in BRCA2-deficient cells. Together, these findings indicate that cytokinesis failure by multiple sources could induce death in BRCA2-deficient cells.

Figure 8 with 1 supplement see all
Download asset Open asset
Mitosis as an alternative synthetic lethality strategy for BRCA2-deficient cells.

(A) Quantitative real-time PCR of CITK in shBRCA2 HCT116p21-/- cells transfected with 150 μM of siCITK (N=2). (B) Relative cell number (%) of PEO4 and PEO1 after 6 days of being transfected with siROCK (1+2), siCITK or siROCK (1+2)/siCITK and representative images of the transfected cells (N=2). (C) Percent of binucleated PEO1 cells transfected with siROCK (1+2), CITK or siROCK (1+2)/siCITK (N=2). (D) Quantitative real-time PCR of EMI1 in shBRCA2 HCT116p21-/- cells transfected with 150 μM of siEMI1 (N=2). (E) Relative cell number (%) of PEO4 and PEO1 after 6 days of being transfected with siEMI1 and treated with fasudil (N=2). Representative images of the transfected and treated cells. (F) Cell cycle analysis of PEO1 cells following transfection with siEMI1 for 48hs (N=2). Cells were stained with propidium iodide and DNA content was analyzed via FACS (10,000 events per sample). (G) Model depicting the events leading to BRCA2-deficient cell death after fasudil treatment. The inhibition or depletion of ROCK in BRCA2-deficient cells leads to cytokinesis failure. As a result, the daughter cells are binucleated (4N) and have extra centrosomes (two instead of one). We speculate that after a subsequent DNA duplication, these cells can attempt mitosis. Mitosis entry with increased DNA content and extra centrosomes may frequently give rise to abnormal and multipolar spindles, leading to misaligned chromosomes and mitotic failure due to multipolar spindle formation. Alternatively, cytokinesis may fail again, and cells may temporarily survive as multinucleated cells, possibly facing cell death during subsequent mitotic attempts.

If aberrant transit through mitosis is the origin of the cell death triggered by ROCKi, then the bypass of mitosis should protect those cells from cell death. To this end, we downregulated Early mitotic inhibitor-1 (EMI1), an anaphase-promoting complex (APC) inhibitor that has a crucial role in the accumulation of mitosis activators, including B-type cyclins (Reimann et al., 2001). When transfecting siEMI1, we observed a 65% reduction in EMI1 expression (Figure 7D) and, as reported by others (Robu et al., 2012; Shimizu et al., 2013; Verschuren et al., 2007), accumulation of cells with G2/M DNA content or higher (Figure 8E). EMI1 depletion prevented the SL effect of ROCKi on different BRCA2-deficient cells (Figure 8F and Figure 8—figure supplement 1B). Therefore, these results indicate that BRCA2-deficient cells that die upon ROCK inhibition do so after transiting an aberrant mitosis.

Discussion

Targeting mitosis as an alternative SL strategy

In this work, we used a novel screening platform developed and validated by our group (Carbajosa et al., 2019; García et al., 2020) to identify ROCK as novel targets for SL induction in BRCA2-deficient cells. Using commercially available, and clinically relevant, ROCKi (i.e.: fasudil, ripasudil and SR 3677 dihydrochloride) (Feng et al., 2016; Feng et al., 2008; Lee et al., 2019), we observed a dose-dependent SL-induction in multiple BRCA2-deficient cell lines which showed no signs of DNA replication stress. In contrast, these cells exhibited strong mitotic defects due to the cytokinesis failure induced by ROCKi. Remarkably, cell death by ROCK inhibition or depletion was recapitulated by inhibiting another enzyme that facilitates cytokinesis, CITK, supporting a model in which binucleation precedes multinucleation and SL (Figure 8G). In fact, robust evidence in the literature indicates that highly abnormal metaphases/anaphases, such as the ones we observed, are incompatible with cell viability (Ganem et al., 2009) and are, therefore, the most plausible cause for the SL induced by ROCKi in BRCA2-deficient cells. While still viable, multinucleated cells are highly vulnerable. The presence of extra DNA content and centrosomes, increases the chances of abnormal spindle polarity, as well as the number of chromosomes that need to be properly aligned. In fact, attempts to trigger cell division in such states is incompatible with viability (Ganem et al., 2009; Dale Rein et al., 2015; Schoonen et al., 2017). We, therefore, postulate that the cytokinesis failure of a cell with 4N or more DNA content is the major driver for BRCA2-deficient cell death following ROCK inhibition. Because we have not identified the molecular target of ROCK which dysregulation triggers SL in BRCA2-deficient cells, further research on the mitotic functions of BRCA2 will be necessary to fully understand this SL pathway. However, we believe it is valuable to report that targeting mitosis alone in the absence of increased replication stress may suffice to kill BRCA2-deficient cells.

BRCA2-deficient cells can be killed in a manner that is independent from the induction of replication stress

In addition to the well-documented replication stress-mediated toxicity of PARPi in BRCA-deficient cells, a recent report indicates that BRCA2-deficient cells can also be killed by mild replication defects which do not cause γH2AX accumulation in S phase (Adam et al., 2021). This cell death depends on the transmission of under-replicated DNA from S to M phase triggered by BRCA1 or BRCA2 deficiency and the lack of CIP2A-TOPBP1 complex formation in M phase. In the absence of this complex, under-replicated DNA is aberrantly processed into acentric chromosomes and micronuclei, which are the source of SL (Adam et al., 2021). Our present work reveals yet another weakness of BRCA2-deficient cells: cytokinesis. Strikingly, this SL is not preceded by the accumulation of broken chromosomes or micronuclei and is independent of canonical players of the DDR, as it is observed after ROCK or CITK inhibition.

Intriguingly, while the triggers of SL by PARPi, CIP2A-TOPBP1 complex disruption and ROCKi are remarkably different, the three mechanisms converge at mitosis; see Adam et al., 2021; Schoonen et al., 2017; Schoonen and van Vugt, 2018; and this work. CDK1 inhibition blocks micronucleation which is the trigger for BRCA-deficient cell death by CIP2A-TOPBP1 complex disruption (Adam et al., 2021), while PARPi and ROCKi-mediated cell death is abrogated by EMI1-depletion- see Schoonen et al., 2017 and this work. Hence, the transit through M phase is required for all SL events triggered in BRCA2-deficient cells. Of note, the accumulation of viable multinucleated BRCA2-depleted cells capable of enabling DNA synthesis after ROCKi reveal that, at least for a few DNA replication cycles, a cytokinesis-free cell cycle progression does not affect survival of BRCA2-deficient cells. Interestingly, multinucleation was also reported after PARPi treatment (Schoonen et al., 2017) and anaphase bridges were detected both after ROCKi and PARPi as a potential source of either multinucleation or cell death- see Schoonen et al., 2017 and this work. In conclusion, despite the difference in the initial trigger of cell death, both after PARPi and ROCKi, BRCA2-deficient cells die at mitosis.

It should also be mentioned that our experimental analysis does not rule out that background levels of replication stress or increased levels of under-replicated DNA induced by BRCA2 deficiency could contribute to the cell death triggered by ROCK inhibition. As previously suggested by Adam et al., 2021, it is possible that BRCA2-deficient cells rely more on M phase due to their propensity to accumulate defects in DNA synthesis, making them more susceptible to a suboptimal M phase (e.g. triggered by ROCKi). However, if the source of SL was solely associated with DNA synthesis events, then it would also be present in BRCA1-deficient backgrounds, which we did not observe. Importantly, BRCA1-deficient backgrounds are also vulnerable during M phase, as we previously observed SL between BRCA1 and PLK1 inhibitors (Carbajosa et al., 2019). The identification of synthetic lethal interactions specific for BRCA1 or BRCA2, indicates that HR impairment is not the only possible trigger of SL in BRCA1- and BRCA2-deficient backgrounds that could be therapeutically exploited. In the future, M phase may provide a window of opportunity for novel treatments in patients that do not respond to PARPi therapy.

Cytokinesis failure as the trigger of the SL between BRCA2-deficiency and ROCK inhibition

We believe that DNA replication defects are not the main trigger for the SL observed with ROCKi, and that defects intrinsic to M phase are more likely to account for ROCKi-induced cell death of BRCA2-deficient cells. Intriguingly, BRCA2 and ROCK functions converge at cytokinesis. ROCK kinases accumulate at the cleavage furrow (Kosako et al., 2000; Yokoyama et al., 2005), regulate furrow ingression, and their knockdown induces multinucleation (Yokoyama et al., 2005). Similarly, CITK localizes to the cleavage furrow, and its downregulation or inhibition also causes multinucleation (Kosako et al., 2000; Sahin et al., 2019). Cytokinesis defects have also been reported for BRCA2-deficient cells (Daniels et al., 2004; Jonsdottir et al., 2009; Mondal et al., 2012; Rowley et al., 2011). However, BRCA2 localizes to a different cytokinesis structure than ROCK, the midbody (Daniels et al., 2004; Jonsdottir et al., 2009; Mondal et al., 2012; Rowley et al., 2011). Remarkably, previous reports suggest that the effect of BRCA2 downregulation on cytokinesis regulation may be very mild (Lekomtsev et al., 2010). Given ROCK and BRCA2 localize to cytokinesis structures that are also separated in time (furrow and midbody), the deficiency in both functions may potentiate cytokinesis failure and cell death. Supporting cytokinesis failure as the SL trigger between ROCK and BRCA2 deficiency backgrounds, we observed that binucleation significantly accumulates at 24 h of treatment, when other mitotic defects have not yet significantly increased.

SL can be enhanced by the formation of multipolar spindles due to centrosome amplification. BRCA2 contributes to the regulation of centriole splitting (Saladino et al., 2009) and centrosome number (Ehlén et al., 2020; Saladino et al., 2009; Tutt et al., 1999). BRCA2 also localizes to centrosomes and preventing such a localization causes centrosome amplification and multinucleation (Shailani et al., 2018). ROCK also localizes to the centrosome (Chevrier et al., 2002; Ma et al., 2006) and its activity is required for centrosome movement and positioning (Chevrier et al., 2002; Rosenblatt et al., 2004). Similar to BRCA2 deficiency, ROCK inhibition also induces centriole splitting and centrosome amplification (Aoki et al., 2009; Chevrier et al., 2002; Oku et al., 2014). Interestingly, both ROCK and BRCA2 bind nucleophosmin (NPM/B23), a protein involved in the timely initiation of centrosome duplication (Ma et al., 2006; Okuda et al., 2000) and disrupting the interaction between BRCA2 and NPM/B23 induces centrosome fragmentation and multinucleation (Wang et al., 2011). Hence, the SL observed after BRCA2 deficiency and ROCKi may be enhanced by centrosome dysregulation, leading to mitotic spindle defects, cytokinesis failure and cell death. Further work may shed additional light on this SL pathway and unravel other potential druggable targets that could provide therapeutic alternatives for treating BRCA2-deficient tumors.

Materials and methods

Screening

Request a detailed protocol

Stable HCT116p21-/- cell lines tagged with fluorescent proteins (CFP, iRFP or mCherry) and expressing Scramble, BRCA1, or BRCA2 shRNAs (Carbajosa et al., 2019) were co-cultured in equal proportions in 96-well plates for 6 days in the presence (0.1 μM) of each of the 680 compounds of the Protein Kinase Inhibitor Set 2 (PKIS2) library (Drewry et al., 2017; Elkins et al., 2016). At the end of treatment, the final cell number for each cell population was assessed with an automated flow cytometer Attune NxT acoustic focusing cytometer (Thermo Fisher). olaparib (#S1060, SelleckChem) at 100 nM was used as a positive control in each screening plate.

For each tested compound, two scenarios are possible: (A) non-selective effect, where the ratio of the populations remains unchanged. The non-selective compounds can either be non-toxic (the number of cells in all populations remains the same) or toxic (the number of cells from each population decreases similartestly); (B) synthetic lethal: selective toxicity against the BRCA2-deficient population, thus changing the relative abundance and ratio between the different populations. Additionally, a compound was considered a ‘hit’ if it exhibited a>5 standard deviation on two values: (1) Fold of SL induction, calculated from the ratios of the different populations in each well; and (2) Survival difference, calculated from the differential survival when comparing a given treatment to the untreated wells in the same plate. For more extensive details on the screening platform and calculations used for the analysis, please refer to Carbajosa et al., 2019.

Lentiviral production

Request a detailed protocol

Lentiviral shRNA vectors were generated by cloning shBRCA2 (sequence) or shScramble (5′-GTTAACTGCGTACCTTGAGTA) into the pLKO.1-TRC vector (Grotsky et al., 2013). HEK293T cells were transfected with pLKO.1 and packaging plasmids (psPAX, and pMD2.G) 24 h post-seeding using JetPrime transfection reagent (Polyplus). After another 24 hr, media was changed. Forty-eight h after, media was collected, centrifuged, and supernatants were aliquoted and stored at –80 °C. Optimal viral titers were tested by serial dilutions and selected based on the minimal toxicity observed in the target cells.

Generation of HCT116p21-/- shRNA stable cell lines

Request a detailed protocol

HCT116p21-/- cells (a kind gift from Bert Volgelstein, Johns Hopkins University) were used to generate stable shScramble or shBRCA2 HCT116p21-/- cells using lentiviral transduction. For viral transduction cells were seeded in 60 mm dishes, and 24 h post-seeding they were transduced using optimal viral titer and 8 μg/ml polybrene (#sc-134220, Santa Cruz Biotechnology). Transduced cells were selected with 1 μg/ml puromycin (#P8833, Sigma-Aldrich) 24 h post-transduction, and amplified for later freezing. Frozen stocks were not used for more than three weeks after thawing. BRCA2 knockdown was confirmed using quantitative real-time PCR.

Other cell lines and culture conditions

Request a detailed protocol

PEO1/PEO4: PEO1 is a BRCA2-deficient ovarian cell line derived from the ascites fluid of a patient (Langdon et al., 1988; Wolf et al., 1987). PEO4 derives from the same patient after the development of chemotherapy resistance and BRCA2 function recovery (Sakai et al., 2009; Wolf et al., 1987). V-C8 and V-C8#13: V-C8 (a kind gift from Bernard Lopez, Gustave Roussy Cancer Center) is a BRCA2-deficient Chinese hamster lung cell line, while V-C8#13 has restored BRCA2 function via one copy of human chromosome 13 harboring BRCA2 (Kraakman-van der Zwet et al., 2002). DLD-1/DLD-1BRCA2-/- cell lines (# HD PAR-008 and #HD 105–00, Horizon Discovery Ltd.): DLD-1 cell lines are human colorectal cancer cell lines, while the BRCA2-deficient DLD-1BRCA2-/- cell line has BRCA2 exon 11 disrupted with rAAV gene editing technology (Hucl et al., 2008).

PEO4/PEO1 and DLD-1/DLD-1BRCA2-/- cell lines were grown in RPMI (#31800–089, Gibco) supplemented with 10% fetal bovine serum (Natocor) and 1% penicillin/streptomycin. V-C8#13 /V-C8, HCC1937BRCA1/HCC1937 (ATCC) and HEK293T (a kind gift from Alejandro Schinder, Fundación Instituto Leloir) were grown in DMEM (#12800082, Gibco) supplemented with 10% fetal bovine serum (Natocor) and 1% penicillin/streptomycin. All cell lines were maintained in a humidified, 5% CO2 incubator and passaged as needed. Cell lines were regularly checked for mycoplasma contamination. The BRCA2 and BRCA1 status of all cell lines was checked, and none of the used cell lines is in the list of commonly misidentified cell lines maintained by the International Cell Line Authentication Committee.

Drugs and treatments

Request a detailed protocol

Cells were treated 24 h post-seeding. Treatment times for each experiment, ranging from 24 h to 6 days, are specified below or in the figure legends. olaparib (#S1060, SelleckChem) was resuspended in DMSO and stored at –20 °C. ROCK inhibitors, fasudil HCl (#A10381, Adooq), SR 3677 dihydrochloride (A12674) and ripasudil (#S7995, SelleckChem) were resuspended in water and stored at –80 °C. BrdU (Sigma-Aldrich) was resuspended in DMSO and stored at –20 °C. BrdU-containing media (10 μM) was added to cell cultures 15 min before harvest. Cisplatin was resuspended in 0.9% NaCl and stored at –20 °C (#P4394, Sigma-Aldrich). Cisplatin was added to cell cultures for 24 hr. All drug stocks were filter-sterilized (0.2 μM). Unless otherwise stated, all experiments were performed three times.

Survival assay

Request a detailed protocol

To perform a survival assay that can be directly compared with the phenotypic screening used in this report we plated in each single well from a 96-well plate, a number of cells that would reach 90% confluence at the time of finalization of the assay (6 days). HCT116p21-/- cell lines were seeded at 1500 cells/well, V-C8 at 500 cells/well, PEO at 2500 cells/well and DLD-1/DLD-1BRCA2-/- at 500 and 1500 cells/well, respectively. Cells were treated with the indicated reagents 24 h post-seeding. Each treatment had three technical replicates. The last day, cells were fixed with 4% paraformaldehyde/ 2% sucrose and stained with DAPI (#10236276001, Roche). Plates were photographed with the IN Cell Analyzer 2200 high content analyzer (GE Healthcare), using a ×10 objective. A total of nine pictures per individual well were taken, and all nuclei in the image were automatically counted to assess cell numbers for each well. Cell number (%) after each treatment was calculated relative to the total number of cells in untreated wells in the same plate. In this way and similarly to the phenotypic screening, cells were counted directly and no indirect metabolic parameter, sub G1 populations or other parameters were monitored. Also, variables such as extreme dilutions (e.g.: used in clonogenic survival) were not introduced by this assay.

Restriction enzyme digestion

Request a detailed protocol

Genomic DNA from PEO4 and PEO1 cell lines was extracted using phenol-chloroform-isoamyl alcohol (#P3803, Sigma-Aldrich). A fragment of 694 bp within the BRCA2 gene was PCR amplified using specific primers (Forward primer: AGATCACAGCTGCCCCAAAG, Reverse primer: TTGCGTTGAGGAACTTGTGAC). PCR fragments were gel purified, and equal amounts of DNA were subject to DrdI (New England Biolabs) enzyme digestion following the manufacturer’s instructions. Digestion products were run on an agarose gel and stained with ethidium bromide to visualize the band pattern.

Chromosome aberration analysis

Request a detailed protocol

Cells were seeded and treated 24 h post-seeding, and 0.08 μg/ml colcemid (KaryoMAX, Invitrogen) was added 20 h before harvest. Following trypsinization, cell pellets were incubated in hypotonic buffer (KCl 0.0075 M) at 37 °C for 4 min and fixed with Carnoy’s fixative solution (3:1 methanol: glacial acetic acid). Cells were dropped onto slides and air-dried before staining with 6% Giemsa in Sorensen’s buffer (2:1 67 mM KH2PO4:67 mM Na2HPO4, pH 6.8) for 2 min. Pictures of metaphases were taken using an automated Applied Imaging Cytovision microscope (Leica Biosystems). Fifty metaphase spreads per independent experiment were analyzed for chromosome gaps, breaks and exchanges.

Anaphase aberration assay

Request a detailed protocol

Cells were fixed with 2% paraformaldehyde/ 2% sucrose for 20 min and stained with DAPI (#10236276001, Roche) to visualize anaphases and quantify anaphase aberrations (bridges and lagging chromosomes). At least 50 anaphases/sample were analyzed. Z-stacks were acquired with a Zeiss LSM 510 Meta confocal microscope and were combined for image generation. Maximum intensity projections were generated using FIJI (ImageJ) Imaging Software.

Micronuclei assay

Request a detailed protocol

Micronuclei (MN) analyses were performed using protocols previously described by us (Federico et al., 2016). Briefly, cells were seeded at low density, treated and incubated with cytochalasin B (4.5 μg/ml, Sigma-Aldrich) for 40 hr. Cells were washed twice with PBS and fixed with PFA/sucrose 2% for 20 min. Phalloidin and DAPI staining were used to visualize whole cells and nuclei, respectively. A total of 300 binucleated cells were analyzed, and the frequency was calculated as MN/binucleated cells.

Immunofluorescence

Request a detailed protocol

Cells were seeded on coverslips, treated, fixed for 20 min with 2% paraformaldehyde/ 2% sucrose and permeabilized for 15 min with 0.1% Triton-X 100. Following 1 h blocking with 2.5% donkey serum in 0.05% PBS/Tween, coverslips were incubated as needed with primary antibodies: γH2AX S139 (1:1500, #05–636-I, Millipore), 53BP1 (1:1500, #sc-22760, Santa Cruz Biotechnology), cyclin A (1:1000, #GTX-634–420, GeneTex) or Phalloidin (1:50, #A12379, Invitrogen). For BrdU staining (1:500, #RPN20AB, GE Healthcare), cells were fixed with ice-cold methanol (40 s) and acetone (20 s), followed by DNA denaturing in 1.5 N HCl for 40 min. For staining of centrosomes (1:1000, #T6557, Sigma-Aldrich) and microtubules (1:1000, #T9026, Sigma-Aldrich), cells were fixed for 10 min with ice-cold methanol, followed by hydration with PBS. Following 1 h of incubation with primary antibodies, cells were washed (3 x/10 min each) with 0.05% PBS/Tween, incubated for 1 h with anti-donkey Alexa 488 or 546 (1:200, Invitrogen), washed, stained with DAPI (#10236276001, Roche) and mounted on slides with Mowiol (Sigma-Aldrich). Slides were analyzed with ×40 or100 x objectives using an Axio Observer microscope (Zeiss).

Number of 53BP1 foci

Request a detailed protocol

Cells were seeded on coverslips and treated as described in the immunofluorescence section above. The quantification of foci/cell was executed using the protocol used by Kilgas et al., 2021. For the experiment in Figure 3—figure supplement 2, in which cells were treated for 6 days, because of the presence of bi and multinucleation, the number of 53BP1 foci per cell was normalised according to their number of nuclei, resulting in the number of 53BP1 foci/nuclei informed.

Colony Assay

Request a detailed protocol

shScramble and shBRCA2 HCT116p21-/- cells were treated with fasudil for 24 hr. Samples were washed and the cells attached to the plate were trypsinized, counted and seeded at extremely low density in 24-well plates. After 10–12 days of culture, the media was removed, and crystal violet staining solution was added for colony visualization. The crystal violet staining solution was washed with ddH2O. The colony assay was performed utilizing for different cell dilutions. The cell colony number was determined as described in Joray et al., 2017.

Flow cytometry analysis

Request a detailed protocol

Cells were seeded, treated and harvested at different time points (24 hr-6 days). For propidium iodide staining, cells were trypsinized, fixed with ice-cold ethanol overnight, and stained with a solution of 100 μg/ml RNase (#10109142001, Roche) and 50 μg/ml propidium iodide (#P4170, Sigma-Aldrich). A total of 10,000 events were recorded using a FACSCalibur (BD Biosciences). Cell cycle distribution was analyzed with the Cytomation Summit software (Dako version 4.3). To assess cell death using SYTOX Green, cells were treated and harvested at different time points. Following trypsinization, samples were stained with SYTOX Green staining following manufacturer’s instructions (#S34860, Invitrogen). 10,000 events were recorded and analyzed using a FACSAria (BD Biosciences).

Quantitative real-time PCR

Request a detailed protocol

Total RNA was extracted with TRIzol reagent (Invitrogen), following the manufacturer’s instructions. A total of 2 μg of RNA was used as a template for cDNA synthesis using M-MLV reverse transcriptase (#28025, Invitrogen) and oligo-dT as primer. Quantitative real-time PCR was performed in a LightCycler 480 II (Roche) using the 5 X HOT FIREPol EvaGreen q PCR Mix Plus (#08-24-00001, Solis BioDyne).

To calculate relative expression levels, samples were normalized to GAPDH expression. Forward (FW) and reverse (RV) primers were as follows: BRCA2 (FW: AGGGCCACTTTCAAGAGACA, RV:TAGTTGGGGTGGACCACTTG), ROCK1 (FW: GATATGGCTGGAAGAAACAGTA, RV:TCAGCTCTATACACATCTCCTT), ROCK2 (FW:AGATTATAGCACCTTGCAAAGTA, RV:TATCTTTTTCACCAACCGACTAA), CITK (FW:CAGGCAAGATTGAGAACG, RV:GCACGATTGAGACAGGGA), EMI1 (FW:TGTTCAGAAATCAGCAGCCCAG, RV:CAGGTTGCCCGTTGTAAATAGC) and GAPDH (FW:AGCCTCCCGCTTCGCTCTCT, RV GAGCGATGTGGCTCGGCTGG).

siRNA transfection

Request a detailed protocol

siRNAs were transfected using JetPrime transfection reagent (Polyplus) following the manufacturer’s instructions. Unless otherwise stated, cells were transfected for a total of 48 hr. siROCK1 (#sc-29473 Santa Cruz Biotechnology) and siROCK2 (#sc-29474, Santa Cruz Biotechnology) were used at 100 nM. siEMI1 (#sc-37611 Santa Cruz Biotechnology) and siCITK (#sc-39214 Santa Cruz Biotechnology) were both used at 100 nM.

Statistical analysis

Request a detailed protocol

GraphPad Prism 5.0 was used for all statistical analyses. Regular two-way ANOVA, followed by a Bonferroni post-test or Student’s t-tests were used as appropriate. BrdU intensity was analyzed with a Kruskal-Wallis non-parametric test followed by a Dunn’s multiple comparison test. Statistical significance was set at p<0.05.

Data availability

All data generated or analysed during this study are included in the manuscript and supporting file.

The following data sets were generated

References

    1. Feng Y
    2. LoGrasso PV
    3. Defert O
    4. Li R
    (2016) Rho kinase (ROCK)
    Inhibitors and Their Therapeutic Potential. J Med Chem 59:2269–2300.
    https://doi.org/10.1021/acs.jmedchem.5b00683
    1. Langdon SP
    2. Lawrie SS
    3. Hay FG
    4. Hawkes MM
    5. McDonald A
    6. Hayward IP
    7. Schol DJ
    8. Hilgers J
    9. Leonard RC
    10. Smyth JF
    (1988)
    Characterization and properties of nine human ovarian adenocarcinoma cell lines
    Cancer Research 48:6166–6172.
    1. Moynahan ME
    2. Cui TY
    3. Jasin M
    (2001)
    Homology-Directed DNA repair, mitomycin-C resistance, and chromosome stability is restored with correction of a BRCA1 mutation
    Cancer Research 61:4842–4850.
    1. Tomlinson GE
    2. Chen TT
    3. Stastny VA
    4. Virmani AK
    5. Spillman MA
    6. Tonk V
    7. Blum JL
    8. Schneider NR
    9. Wistuba II
    10. Shay JW
    11. Minna JD
    12. Gazdar AF
    (1998)
    Characterization of a breast cancer cell line derived from a germ-line BRCA1 mutation carrier
    Cancer Research 58:3237–3242.

Decision letter

  1. Robert W Sobol
    Reviewing Editor; Brown University, United States
  2. Wafik S El-Deiry
    Senior Editor; Brown University, United States

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Inhibitors of ROCK kinases induce multiple mitotic defects and synthetic lethality in BRCA2-deficient cells" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen Wafik El-Deiry as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1) The evidence that this SL interaction is independent of replication defects is not solid. Please provide more data to support this conclusion.

2) The SL interaction is based on chemical inhibitors only, with 6 out of 9 ROCK inhibitors not demonstrating the SL interaction. Please explain and define why?

3) The mechanisms by which ROCKi specifically affects BRCA2-defective cells are elusive. Explain.

4) It remains unclear what the cause of the multiple mitotic defects is – best if the authors expand on this.

5) Missing is some discussion, at least, on how BRCA2 expression is related to actin cytoskeleton dynamics controlled by ROCK kinases.

6) The authors also show that ROCK kinase inhibitors are also quite toxic to WT cells. Is this via the same mechanism but just attenuated and if so, why?

7) The authors should provide more detail on the mechanism of BRCA2-specific cell death by ROCK kinase inhibitors

Reviewer #1 (Recommendations for the authors):

Authors focus all the PARPi effects in BRCA2-KO cells as due to HR defects, but recent reports have supported more clarity of that phenotype, related to the accumulation of replication gaps and not solely due to HR defects. Would be good to enhance the Intro and discussion accordingly.

Missing is some discussion, at least, on how BRCA2 expression is related to actin cytoskeleton dynamics controlled by ROCK kinases.

The authors point out how important these findings are and how it was dependent on their "novel phenotypic survival screening method". This should be explained in more detail. This is particularly important since the PE01 cells, using this method, are not very sensitive to PARPi's as expected.

The authors also show that ROCK kinase inhibitors are also quite toxic to WT cells. Is this via the same mechanism but just attenuated and if so, why?

The authors should provide more detail on the mechanism of BRCA2-specific cell death by ROCK kinase inhibitors.

Also, why do many of the ROCK kinase inhibitors evaluated show no phenotype?

Reviewer #2 (Recommendations for the authors):

– The SL interaction that the authors describe is potentially interesting and appears to be independent of replication stress. longer assays (ie clonogenic assays) and preferably in vivo work are required to demonstrate the value of this SL interaction.

– The evidence that this SL interaction is independent of replication defects is not solid. the observation depends on a few markers and could be strengthened by functional replication analysis (DNA fiber analysis). the cut-off for replication stress analysis is highly arbitrary (5 foci for 53BP1 and 35 foci for γH2AX). Absolute foci numbers should be plotted and statistically analyzed, especially because this is a key argument in the manuscript.

– The work depends solely on chemical inhibitors (of which 3 out of 9 score in the SL analysis). Why do the other ROCK inhibitors not show SL effects? Genetic approaches in long-term assays are required to strengthen the SL interaction.

– EME1 depletion allows cells to skip mitosis, but also restores HR in BRCA-deficient cells (work from Pagano lab). This effect should be considered.

– The observed G2/M arrest upon ROCKi could represent 4n G1 cells.

– The nature and cause of the observed anaphase bridges are enigmatic. A defect in cytokinesis does not explain the defects observed earlier in mitosis.

– It is unclear why ROCK or citron kinase selectively disrupt cytokinesis in BRCA2-defective cells? Also, depmap analysis shows that the viability of most cell lines is negatively affected by ROCK1 loss. How do the authors imagine that long-term loss of ROCK specifically affects BRCA2-defective cells? Is there evidence from public data that this interaction is BRCA2-specific?

– Aberrant multipolar spindles are quantified from DNA staining, but not spindle stainings. This requires analysis of spindle components.

– The chromosome bridge in figure 5A (left panel) does not seem to be a real bridge, but a chromosome that segregates late.

Reviewer #3 (Recommendations for the authors):

1. The timeline in the left panel of Figure 1D seems unnecessary. In addition, the axis should be abellin for the examples in the two right panels of Figure 1D.

2. Axis in Figure 2E is too small to read.

3. There are several formatting inconsistencies, typos, and grammatical errors throughout the manuscript and figures that should be carefully reviewed and corrected.

4. Figure 1 —figure supplement 1: The data shown is not overly convincing. Why wasn’t the [fasudil] taken out as far as it was in some of the BRCA2 deficient lines?

5. It would benefit readers less familiar with looking at metaphase spreads if the authors added arrows pointing out the aberrant phenotypic features.

6. In figure 6 the shades of gray and blue used are too similar. The black borders separating the colors get lost in the error bars especially when adjacent to the darkest blue/gray shades.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled “Inhibitors of Rho kinases (ROCK) induce multiple mitotic defects and synthetic lethality in BRCA2-deficient cells” for further consideration by eLife. Your revised article has been evaluated by Wafik El-Deiry (Senior Editor) and a Reviewing Editor.

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

In particular, I would ask that you work to address the concerns raised by Reviewer #2 below.

Reviewer #2 (Recommendations for the authors):

The authors have addressed my main comments, although some major shortcomings remain.

– The main shortcoming is that there is no mechanistic data that explains why ROCK kinase inhibition selectively affects BRCA2-defective cells. The model that the authors put forward is that ROCK inhibition leads to multinucleate cells, selectively in BRCA2-depleted cells and that this leads in the next cell cycle to structural mitotic defects (eg lagging chromosomes, multipolar spindles). Although the analysis of early and late time points supports this model, any mechanistic underpinning is lacking. Why does ROCK inhibition selectively induce binucleated cells in BRCA2-defective cells in the first cell cycle? There is some literature cited, but that only provides a shallow explanation. Time-lapse imaging would be very informative here to confirm the model that is put forward.

– The analysis of numbers/intensity of H2AX/53BP1 foci in absolute measures are much more insightful than the arbitrary cut-off. I would suggest showing these data in the main figure not supplement.

– Line 123: one ->ones.

– Supplement Figure 3E: please check abelling: should be HU with the images of fibers?

– Clonogenic assays: this is substandard. Please also include the controls and dose response.

Reviewer #3 (Recommendations for the authors):

The manuscript is much improved by the revisions and I feel all of the essential revisions have been adequately addressed.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled “Inhibitors of Rho kinases (ROCK) induce multiple mitotic defects and synthetic lethality in BRCA2-deficient cells” for further consideration by eLife. Your revised article has been evaluated by Wafik El-Deiry (Senior Editor) and a Reviewing Editor.

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

Reviewer #2 (Recommendations for the authors):

I appreciate that the authors have changed the manuscript, concerning data visualisation. Also, I value how the authors have changed their model and writing in which a first binucleation event precedes further toxic events, which aligns much better with the presented data.

I realize that I suggested live cell microscopy analysis, and that this comment came after the first revision. However, it is not per se a request for a specific technique, it was merely a request to provide some mechanistic insight. It was just a suggestion of a commonly used technique in the cell cycle/mitosis research field to uncover what the order of events is that happens in BRCA2-defective cells in which ROCK is inactivated.

Personally, I still feel that the mechanistic insight is not substantial and that the main novelty is that ROCK inactivation is more toxic in BRCA2-depleted cells when compared to BRCA2-proficient cells. I realize that the authors do not agree with this standpoint. At least make clear in the discussion that it is unclear what happens mechanistically.

As the reviewers suggested, I have considered the 4 points put forward in the rebuttal letter (especially points 1,3 and 4):

1. ROCKi can kill BRCA2-deficient cells in a manner that is independent of replication stress in S phase. Such a concept has been reinforced in the R1 version by adding fiber assays and including the quantification of H2AX suggested by reviewer 2.

True, there is no overt replication stress induced by acute ROCK inhibition that can explain the SL effects. The mode of action of PARP inhibitor is clearly different than those of ROCK inhibitors in this respect. However, whether replication stress is not involved in this form of SL is difficult to conclude.

The effects of 53Bp1, H2AX and fibers were done after 48h of treatment, whereas the toxic effects of ROCK inhibition occur much later, when binucleated cells have I. Since whole genome duplication upon mitotic failure is known to cause replication stress (see for instance Gemble et al., Nature, 2022). To really conclude that replication stress is not involved, analysis at a late time points should be done. Of note. this point of the author does not indicate an underlying mechanism of action, but suggests a mechanism that is not involved.

2. The cytokinesis defects triggered by ROCKi have been documented by others, and so was the contribution of BRCA2-depletion to cytokinesis defects. By monitoring cells with DNA content higher than 4 N and binucleation, this manuscript shows that the treatment of BRCA2-deficient cells with ROCKi causes a significant augmentation of binucleation at early time points hence suggesting that the cytokinesis defects is a common defect that is amplified by the combined depletion of BRCA2 and the inhibition of ROCKi. Moreover, supporting the idea of cytokinesis defect as the trigger of other defects, including the DNAreplication independent SL, it temporally precedes all other mitotic defects.

The mechanism proposed here is that ROCKi and BRCA2 inactivation both give cytokinesis defects, and that the SL effect is based on these effects adding up. This is not entirely novel, and also, BRCA2 inactivation in this manuscript only marginally induces cytokinesis defects on its own. That cytokinesis failure leads to secondary defects I completely support, but claiming that secondarily effects are DNA replication independent is not shown and should not be claimed.

3. Reinforcing the link between cytokinesis failure and SL, the inhibition of another kinase that has also been associated with cytokinesis defect, Citron Rho-interacting kinase, also triggers SL in BRCA-2 deficient cells.

True, the observation that Rho kinase and ROCK behave the same is strong, and points towards cytokinesis, in which both kinases are clearly implicated.

4. The prevention of mitosis by EMI1 depletion prevents the SL caused by ROCKi in BRCA2- deficient cells, thus reinforcing the association between M-phase transit and cell death in BRCA2-deficient cells.

True, the EMI experiment connects mitosis to the SL effect of ROCK inhibition in these cells, reinforcing that a cytokinesis failure is involved

https://doi.org/10.7554/eLife.80254.sa1

Author response

Essential revisions:

1) The evidence that this SL interaction is independent of replication defects is not solid. Please provide more data to support this conclusion.

In the original version of the manuscript, we included the analysis of BrdU+ cells (%) and BrdU intensity in cells transiting S phase (current Figure 3D and Figure 3—figure supplement 1F). We believe that such evidence is crucial since BrdU intensity reveals global changes in the replication choreography that might be overlooked in DNA spreading assays (Calzetta, Gonzalez Besteiro, and Gottifredi, 2021). Moreover, such a negligible effect of ROCKi in the S phase progression of BRCA2 deficient cells is not associated with changes in γH2AX or 53BP1 nuclear foci accumulation (current Figure 3A-C).

Nonetheless, we addressed the reviewers' request to provide additional data by performing two sets of experiments:

1) Reviewer 2 specifically highlighted the need for an alternative quantification method of γH2AX or 53BP1 nuclear foci accumulation independent of arbitrary thresholds. Hence, we quantified absolute γH2AX intensity and the number of 53BP1 foci/ cell (current Supplementary Figure 3 A-D). Such analyses are shown in Figure 3—figure supplement 1AD, and the conclusions (replication stress levels are not upregulated in BRCA2-deficient cells treated with ROCKi) are aligned with the data reported in Figure 3A-C.

2) We have also carried out a fiber assay analysis to monitor nascent DNA elongation tracks and origin firing frequency. ROCKi causes no changes in such DNA replication parameters in both BRCA2-WT and BRCA2-deficient cells (current Figure 3E-F). In contrast, hydroxyurea (HU) caused changes in the nascent DNA track length of BRCA2-deficient cells, as previously shown elsewhere (Schlacher et al., 2011) (current Figure 3—figure supplement 1E). Together, these results demonstrate that while BRCA2 deficiency alters DNA replication parameters after HU, it does not trigger obvious replication defects after the ROCK inhibitor, fasudil.

Hence, the data in Figure 3E-F and Figure 3—figure supplement 1A-E further support our claim of a SL interaction independent of replication defects.

2) The SL interaction is based on chemical inhibitors only, with 6 out of 9 ROCK inhibitors not demonstrating the SL interaction. Please explain and define why?

The SL interaction identified herein is not based on chemical inhibitors only. In fact, in the original version of this manuscript, we have provided genetic evidence (using siRNAs against ROCK1 and ROCK2) to confirm this SL interaction (current Figure 8B and Figure 8—figure supplement 1A).

When it comes to the reason for obtaining SL-positive results for only 3 out of 9 ROCK inhibitors in the screening, it is related to the low dosage at which the screening was performed. Given that ATP competitive kinase inhibitors are well known to hit multiple off-target kinases with similar ATP binding pockets, we decided to perform this screening at a low dose: 0.1 μM. As such, we aim to reduce false positives and the number of hits to validate, yet concomitantly risk increasing the rate of false negatives. This occurrence of false negatives was precisely the case with the ROCK inhibitors. While the 3 hits were within their activity range at the sub-micro-molar range (Figure 1 E and F), the 6 remaining inhibitors were out of their activity range. When re-testing the remaining 6 inhibitors a 1μM, they all showed SL activity on BRCA2-deficient cells. This information is now in Figure 1figure supplement 1.

In this revised version, we also included another potent and selective ROCK inhibitor (SR 3677 dihydrochloride) revealing SL activity in BRCA2-deficient samples (current Figure 2—figure supplement 2D).

3) The mechanisms by which ROCKi specifically affects BRCA2-defective cells are elusive. Explain.

The first type of mitotic defects we observe in BRCA2-deficient cells after fasudil treatment is a significant increase in polyploidy and high levels of binucleated cells (current Figure 4 A-C and Figure 7), which indicates that cytokinesis failure may be the trigger to the SL caused by ROCKi in BRCA2-deficient cells. As detailed below, both ROCK and BRCA2 (but not BRCA1) have been implicated in regulating cytokinesis during mitosis.

Evidence linking ROCK with cytokinesis regulation: ROCK accumulate at the cleavage furrow and phosphorylate myosin II, a component of the actomyosin ring whose contraction gives rise to the cleavage furrow which divides the cytoplasm (Kosako et al., 2000; Yokoyama, Goto, Izawa, Mizutani, and Inagaki, 2005). Depending on the cell line, inhibiting ROCK can impair or delay furrow ingression (Kosako et al., 2000; Lordier et al., 2008). If furrowing is impaired, cytokinesis fails, and there is polyploidization (Lordier et al., 2008). Knockdown of ROCK1 or ROCK2 using siRNAs also induces multinucleation (Yokoyama et al., 2005).

Evidence linking BRCA2 with cytokinesis regulation: BRCA2 localizes to a different cytokinesis structure than ROCK, the midbody, where it mediates the assembly of factors involved in abscission (Daniels, Wang, Lee, and Venkitaraman, 2004; Jonsdottir et al., 2009); (Mondal et al., 2012; Rowley et al., 2011). Cytokinesis defects have also been reported for BRCA2-deficient cell models (Daniels et al., 2004; Jonsdottir et al., 2009; Lee, Daniels, Garnett, and Venkitaraman, 2011; Rowley et al., 2011; Vinciguerra, Godinho, Parmar, Pellman,’and D'Andrea, 2010) (Mondal et al., 2012). Importantly, this bona fide role of BRCA2 in cytokinesis is independent of its role in HR (Mondal et al., 2012).

The evidence above indicates that both BRCA2 and ROCK regulate cytokinesis, albeit at different levels of the process. Such contributions must be mild or easy to compensate for, as individual deficiencies in our experimental settings are insufficient to impair cytokinesis. On this note, previous reports suggest that the individual effects may be very mild. For example, (Lekomtsev, Guizetti, Pozniakovsky, Gerlich, and Petronczki, 2010) compared BRCA2 knockdown against the knockdown of MgcRacGAP, a well-known mediator of cytokinesis, and the multinucleation levels were comparably so low that they concluded that BRCA2 had no significant contribution to cytokinesis.

Our results indicate that the initial trigger that precedes the mitotic defects observed when ROCK inhibition is combined with BRCA2-deficient backgrounds is binucleation (detected as early as 24 hours post-treatment-Figure 7 A). Given that ROCK and BRCA2 localize to cytokinesis structures (furrow and midbody) at different time points, we propose that the combined deficiencies may cause defects in the early and late steps of cytokinesis which augments the probability of cytokinesis failure.

4) It remains unclear what the cause of the multiple mitotic defe–ts is – best if the authors expand on this.

In BRCA2-deficient cells treated with ROCKi, the cytokinesis failure is accompanied by other mitotic defects that, we believe, derive from the DNA replication proficiency maintained by binucleated BRCA2-deficient cells (current Figure 6—figure supplement 1C-D). Because multipolar mitosis events accumulate at later time points than binucleation (48 and 24 hours of treatment respectively current Figure 7), we speculate that the attempt to separate DNA after multipolar spindle assembly could ultimately trigger cell death in BRCA2-deficient cells treated with fasudil. We also detected increased centrosome numbers in such samples (current Figure 6D). Centrosomes, the organelles responsible for nucleating the mitotic spindle, must be in the correct number (two) for successful chromosome segregation during mitosis (Fukasawa, 2007). Increased centrosome numbers (i.e., centrosome amplification) lead to abnormal spindle poles, chromosome instability or cell death (Ganem, Godinho, and Pellman, 2009). Centrosome amplification can result from multiple mechanisms, including cytokinesis failure (Fukasawa, 2007; Lens and Medema, 2019). Centrosome duplication is also ruled by timely centriole splitting (Tsou and Stearns, 2006), and in BRCA2.deficient tumors high levels of centriole splitting have been observed (Saladino, Bourke, Conroy, and Morrison, 2009). BRCA2 also localizes to centrosomes, and preventing this localization causes centrosome amplification and multinucleation (Han, Saito, Miki, and Nakanishi, 2008; Nakanishi et al., 2007; Zhang et al., 2016). Consistent with in vitro results, BRCA2-deficient tumors display increased centrosome numbers (Hou, Li, Ren, and Liu, 2016; Saladino et al., 2009; Tutt et al., 1999; Watanabe et al., 2018; Wiegant, Overmeer, Godthelp, van Buul, and Zdzienicka, 2006). Additionally, BRCA2 has also been implicated in facilitating chromosome alignment and as part of the spindle assembly checkpoint (SAC), which regulates the metaphase to anaphase transition by monitoring proper kinetochoremicrotubules attachment (Choi et al., 2012; Ehlén et al., 2020). The loss of M phase functions in BRCA2-deficient cells, including the regulation of centrosome numbers, may synergistically interact with ROCKi to kill cells. ROCK also localizes to the centrosome (Chevrier et al., 2002; Ma et al., 2006), and its activity is required for centrosome movement and positioning (Chevrier et al., 2002; Rosenblatt, Cramer, Baum, and McGee, 2004), which are crucial to proper spindle formation as well as mitotic exit (Piel, Nordberg, Euteneur, and Bornens, 2001). Inhibition of ROCK causes mitotic spindle misorientation, diffuse pericentrin and increases the nucleating capacity of astral microtubules (Heng et al., 2012; Rosenblatt et al., 2004). Like BRCA2 deficiency, ROCK inhibition also induces centriole splitting and centrosome amplification (Aoki, Ueda, Kataoka, and Satoh, 2009; Chevrier et al., 2002; Oku et al., 2014). Interestingly, both ROCK and BRCA2 bind nucleophosmin (NPM/B23), a protein involved in the timely initiation of centrosome duplication (Ma et al., 2006; Okuda et al., 2000) and disrupting the interaction between BRCA2 and NPM/B23 induces centrosome fragmentation and multinucleation (Wang, Takenaka, Nakanishi, and Miki, 2011).

Together, these findings may explain why BRCA2 deficiency and ROCKi potentiate the accumulation of multiple mitotic defects.

5) Missing is some discussion, at least, on how BRCA2 expression is related to actin cytoskeleton dynamics controlled by ROCK kinases.

As discussed in points 3 and 4, cytokinesis is affected by BRCA2 loss and ROCKi at different levels; while BRCA2 loss dysregulates the midbody organization, ROCKi impair the mitotic furrow formation. Additionally, ROCK and BRCA2 localize to the centrosomes and prevent their amplification. While BRCA2 prevents dysregulation of centriole number and participates in the SAC, ROCK regulates the movement and positioning of centrosomes. Hence, the SL does not result from BRCA2 controlling the ROCK-mediated cytoskeleton regulation. Instead, ROCK and BRCA2 have complementing functions that favor cell fitness in mitosis.

6) The authors also show that ROCK kinase inhibitors are also quite toxic to WT cells. Is this via the same mechanism but just attenuated and if so, why?

We designed an experiment to test whether WT cells die via the same mechanism by which ROCKi kills BRCA2-deficient cells. To test this, we increased fasudil doses in WT cells to achieve levels of cell death similar to those obtained after lower doses of fasudil in BRCA2-deficient cells. Surprisingly, despite the massive cell death observed in WT cells at high fasudil concentrations, increased binucleation was not observed (current Figure 7—figure supplement 1). Hence, the treatment of BRCA2-deficient cells with ROCKi triggers a cell death preceded by cytokinesis failure, while the treatment of WT cells with higher doses of ROCKi triggers a cell death independent of cytokinesis failure.

7) The authors should provide more detail on the mechanism of BRCA2-specific cell death by ROCK kinase inhibitors.

The evidence discussed in points 3 and 4 and the experiment reported in point 6 have already provided additional details on the BRCA2-specific cell death by ROCKi. Furthermore, based on one of the comments from reviewer 2: "The nature and cause of the observed anaphase bridges are enigmatic; a defect in cytokinesis does not explain the defects observed earlier in mitosis", we designed an experiment that provides further detail to the SL mechanism. We monitored binucleation, anaphase and mitotic aberrations at early times: 24, 48 and 72 hours after fasudil treatment. As shown in Figure 7, binucleated cells but not anaphase and mitotic aberrations, were detected at 24 hours in BRCA2-deficient cells treated with ROCKi. These data support the model in Figure 8G, indicating that the earliest alterations that accumulate in BRCA2-deficient cells treated with ROCKi are binucleated cells. Given that DNA duplication is not inhibited in binucleated cells (current Figure 6—figure supplement 1C-D), it is very likely that the anaphase bridges do not precede cytokinesis but accumulate later when binucleated or multinucleated cells attempt mitosis. From these data, we infer that the attempt to achieve separation of polyploid nuclei into daughter cells is the most likely trigger of cell death in BRCA2-deficient cells treated with ROCKi.

Reviewer #1 (Recommendations for the authors):

Authors focus all the PARPi effects in BRCA2-KO cells as due to HR defects, but recent reports have supported more clarity of that phenotype, related to the accumulation of replication gaps and not solely due to HR defects. Would be good to enhance the Intro and discussion accordingly.

Thank you very much for this comment. Reviewer 1 is correct; the introduction has been improved, and the bibliography has been updated. Please see changes in Page 4 of the introduction section.

Missing is some discussion, at least, on how BRCA2 expression is related to actin cytoskeleton dynamics controlled by ROCK kinases.

This comment is part of the "essential revision" section. Please read responses to points 3-5 above.

The authors point out how important these findings are and how it was dependent on their "novel phenotypic survival screening method". This should be explained in more detail. This is particularly important since the PE01 cells, using this method, are not very sensitive to PARPi's as expected.

The novel method we refer to is the screening method described by Carbajosa et al., 2019. Because of reviewer 1's comment, we added details to the automatized cell counting methodology used for PEO1/PEO4 analysis (page 17; Materials and methods).

The authors also show that ROCK kinase inhibitors are also quite toxic to WT cells. Is this via the same mechanism but just attenuated and if so, why?

This comment is part of the "essential revision" section. Please read point 6 from that section.

The authors should provide more detail on the mechanism of BRCA2-specific cell death by ROCK kinase inhibitors

This comment is part of the "essential revision" section. Please read point 7 from that section.

Also, why do many of the ROCK kinase inhibitors evaluated show no phenotype?

This comment is part of the "essential revision" section. Please read point 2 from that section.

Reviewer #2 (Recommendations for the authors):

– The SL interaction that the authors describe is potentially interesting and appears to be independent of replication stress. longer assays (ie clonogenic assays) and preferably in vivo work are required to demonstrate the value of this SL interaction.

This comment is part of the "essential revision" section. Please read point 1 from that section. We have performed clonogenic assays (see current Figure 2—figure supplement 1). We could not set up an in vivo model in the time dedicated to this revision.

– The evidence that this SL interaction is independent of replication defects is not solid. the observation depends on a few markers and could be strengthened by functional replication analysis (DNA fiber analysis). the cut-off for replication stress analysis is highly arbitrary (5 foci for 53BP1 and 35 foci for γH2AX). Absolute foci numbers should be plotted and statistically analyzed, especially because this is a key argument in the manuscript.

The arbitrary cuts have been established based on the replication stress induced by PARP inhibitors in BRCA2-deficient cells, but the comments from reviewer 2 are appropriate, as the replication stress induced by another agent could differ from the one caused by PARP inhibitors. As requested by reviewer 2, absolute γH2AX intensity and 53BP1 foci number have been analyzed and reported in Figure 3—figure supplement 1A-D. Our conclusion has not changed as both markers increased in PARPi-treated but not ROCKi-treated samples.

– The work depends solely on chemical inhibitors (of which 3 out of 9 score in the SL analysis). Why do the other ROCK inhibitors not show SL effects? Genetic approaches in long-term assays are required to strengthen the SL interaction.

This comment is part of the "essential revision" section. Please refer to point 2 from that section.

– EME1 depletion allows cells to skip mitosis, but also restores HR in BRCA-deficient cells (work from Pagano lab). This effect should be considered.

Thank you for bringing up this manuscript about the role of EMI1 in controlling RAD51 protein levels in human breast cancer samples, which reveals the role of EMI1 in HR. While the regulation of RAD51 by EMI1 is undoubtedly relevant when interpreting the effect of EMI1 on PARPi-induced synthetic lethality, some observations from our manuscript do not support a potential role of HR in the SL we describe. First, there is no replication stress in BRCA2-deficient cells treated with ROCKi, so the accumulation of double-strand breaks is unlikely. Second, ROCKi is synthetic lethal in BRCA2- but not BRCA1-deficient cells, and a SL involving HR modulation should equally affect both backgrounds. Third, ROCK overlap with BRCA2 in controlling mitotic events at the level of cytokinesis and centriole regulation (see points 3 and 4 in the "essential revision" section), but ROCK have no role in HR.

– The observed G2/M arrest upon ROCKi could represent 4n G1 cells.

Thank you for spotting this possibility that could explain the early accumulation of cells in the 4n region of the flow cytometry peaks. We have indicated it in the manuscript (see page 8, in the Results section).

– The nature and cause of the observed anaphase bridges are enigmatic. a defect in cytokinesis does not explain the defects observed earlier in mitosis.

This comment is part of the "essential revision" section. Please see point 6 from that section.

– It is unclear why ROCK or citron kinase selectively disrupt cytokinesis in BRCA2-defective cells? Also, depmap analysis shows that the viability of most cell lines is negatively affected by ROCK1 loss. How do the authors imagine that long-term loss of ROCK specifically affects BRCA2-defective cells? Is there evidence from public data that this interaction is BRCA2-specific?

Citron Kinase is intimately related to the function of ROCK, and how ROCK and BRCA2 defects can enhance cytokinesis failure have been described in the "essential revision" section (point 3). Regarding the DepMap analysis, we are not sure which data the reviewer is referring to, as ROCK1 knockout by CRISPR was essential only in 15 of the 1086 cell lines studied, and knockdown by RNAi did not reveal essentiality phenotypes in any of the 710 cell lines tested (https://depmap.org/portal/gene/ROCK1?tab=overview).

Regarding evidence from public databases regarding a specific interaction between BRCA2 and ROCK1, we couldn't find any previous link reported.

– Aberrant multipolar spindles are quantified from DNA staining, but not spindle stainings. This requires analysis of spindle components.

Our first indication of multipolar spindles were aberrant metaphases observed in experiments done only with DAPI staining (Figure 4 D-E). This was followed up by adding staining against α-tubulin (microtubules) and γ-tubulin (centrosomes) (Figure 6 D-F). Such an information is reported in the Methods section and in the legend of Figure 6. The quantifications in Figure 6E and 6F relied on spindle staining (revealed by γ- tubulin), centrosomes (revealed by alfa- tubulin) and DNA (revealed by DAPI). As shown in Figure 6 D, only anaphases revealing all three immunofluorescent marks were considered in the analysis (and all metaphases without exception were positive for the three marks). We realize that the label “percent of cells” may be misleading and we changed it to “Percent of metaphases”.

– The chromosome bridge in figure 5A (left panel) does not seem to be a real bridge, but a chromosome that segregates late.

We have selected a better representative image which is now part of Figure 6 A.

Reviewer #3 (Recommendations for the authors):

1. The timeline in the left panel of Figure 1D seems unnecessary. In addition, the axis should be labeled for the examples in the two right panels of Figure 1D.

We have introduced the changes suggested by the reviewer. Thank you.

2. Axis in Figure 2E is too small to read.

The size of the font was enlarged, as suggested by reviewer 3.

3. There are several formatting inconsistencies, typos, and grammatical errors throughout the manuscript and figures that should be carefully reviewed and corrected.

We have carefully proofread the manuscript.

4. Figure 1 —figure supplement 1: The data shown is not overly convincing. Why wasn't the [Fasudil] taken out as far as it was in some of the BRCA2 deficient lines?

Fasudil curves were the same in BRCA1- and BRCA2-deficient samples such as HCT116 (Figure 2 A for BRCA2-deficient vs WT samples and Supplementary Figure 1K for BRCA1-deficient vs WTt samples). We could not reach higher doses in the BRCA1-deficient sample because the control sample (present in both curves) was already affected by the highest dose. Further support to the absence of SL by fasudil in BRCA1-deficient samples is in Figure 2—figure supplement 2E. Please note that at higher doses, the control samples are even more sensitive than the BRCA1-deficient samples to fasudil (the opposite of what we observed in BRCA2-deficient samples).

5. It would benefit readers less familiar with looking at metaphase spreads if the authors added arrows pointing out the aberrant phenotypic features.

We have added arrows to the aberrant phenotypic feature in Figure 5 A.

6. In figure 6 the shades of gray and blue used are too similar. The black borders separating the colors get lost in the error bars especially when adjacent to the darkest blue/gray shades.

We have modified the colors to make the figure easier to understand. Thank you.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Reviewer #2 (Recommendations for the authors):

The authors have addressed my main comments, although some major shortcomings remain.

– The main shortcoming is that there is no mechanistic data that explains why ROCK kinase inhibition selectively affects BRCA2-defective cells. the model that the authors put forward is that ROCK inhibition leads to multinucleate cells, selectively in BRCA2-depleted cells and that this leads in the next cell cycle to structural mitotic defects (eg lagging chromosomes, multipolar spindles). although the analysis of early and late time points supports this model, any mechanistic underpinning is lacking. Why does ROCK inhibition selectively induce binucleated cells in BRCA2-defective cells in the first cell cycle? there is some literature cited, but that only provides a shallow explanation. time-lapse imaging would be very informative here to confirm the model that is put forward.

We submitted the original version of this manuscript on June 2022 and provided an R1 version of the manuscript attending all reviewer comments in November 2022. The live cell microscopy experiment has been suggested by reviewer 2 when evaluating the R1 version of our manuscript. We do not deny that such an experiment could be very interesting to perform as many others but we cannot provide this experiment at this time. It is technically challenging and impossible to perform without the building up of expertise that can only be gained after several months.

Moreover, we believe that the comment indicating that “no mechanistic data that explains why ROCK kinase inhibition selectively affects BRCA2-defective cells” is a bit harsh. As discussed below, there are many critical mechanistic insides which are already provided by our manuscript (when assessing the mechanistic contribution reviewer 2 focuses on point 2 but not 1, 3 and 4). We ask reviewer 2 and the editors to please consider the following mechanistic insights provided in the manuscript:

1) ROCKi can kill BRCA2-deficient cells in a manner that is independent of replication stress in S phase. Such a concept has been reinforced in the R1 version by adding fiber assays and including the quantification of H2AX suggested by reviewer 2.

2) The cytokinesis defects triggered by ROCKi have been documented by others, and so was the contribution of BRCA2-depletion to cytokinesis defects. By monitoring cells with DNA content higher than 4 N and binucleation, this manuscript shows that the treatment of BRCA2-deficient cells with ROCKi causes a significant augmentation of binucleation at early time points hence suggesting that the cytokinesis defects is a common defect that is amplified by the combined depletion of BRCA2 and the inhibition of ROCKi. Moreover, supporting the idea of cytokinesis defect as the trigger of other defects, including the DNA replication independent SL, it temporally precedes all other mitotic defects.

3) Reinforcing the link between cytokinesis failure and SL, the inhibition of another kinase that has also been associated with cytokinesis defect, Citron Rho-interacting kinase, also triggers SL in BRCA-2 deficient cells.

4) The prevention of mitosis by EMI1 depletion prevents the SL caused by ROCKi in BRCA2deficient cells, thus reinforcing the association between M-phase transit and cell death in BRCA2-deficient cells.

Finally, reviewer 2 also considers that the analysis of early and late time points supports a model indicating that ROCKi first leads BRCA2-depleted cells to binucleation and later to structural mitotic defects (e.g. lagging chromosomes, multipolar spindles). In the manuscript, we have made it clear that our findings indicate that a significant increase in binucleation is detected earlier than other mitotic defects without mentioning the first or second cycle. In addition, in the current R2 version of the manuscript, we have changed a sentence “we speculate that after a subsequent DNA duplication, these cells can attempt mitosis.” (see lane 1075).

– The analysis of numbers/intensity of H2AX/53BP1 foci in absolute measures are much more insightful than the arbitrary cut-off. I would suggest showing these data in the main figure not supplement.

We have modified the text and Figures as suggested by reviewer 2.

– Line 123: one ->ones.

We have corrected such a typo.

– Supplement Figure 3E: please check labeling: should be HU with the images of fibers?

Reviewer 2 is right. Thank you for spotting the mistake.

– Clonogenic assays: this is substandard. please also include the controls and dose response.

We did not perform the experiment as a dose curve because the fasudil dose was selected from the survival assays. However, we performed the colony assay using different dilutions of cells. We are including all the analysis as well as the controls as requested by reviewer 2.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Reviewer #2 (Recommendations for the authors):

I appreciate that the authors have changed the manuscript, concerning data visualisation. Also, I value how the authors have changed their model and writing in which a first binucleation event precedes further toxic events, which aligns much better with the presented data.

We thank reviewer 2 for this positive assessment.

I realize that I suggested live cell microscopy analysis, and that this comment came after the first revision. However, it is not per se a request for a specific technique, it was merely a request to provide some mechanistic insight. It was just a suggestion of a commonly used technique in the cell cycle/mitosis research field to uncover what the order of events is that happens in BRCA2-defective cells in which ROCK is inactivated.

Personally, I still feel that the mechanistic insight is not substantial and that the main novelty is that ROCK inactivation is more toxic in BRCA2-depleted cells when compared to BRCA2-proficient cells. I realize that the authors do not agree with this standpoint. At least make clear in the discussion that it is unclear what happens mechanistically.

We still believe that finding a SL association between a non-DDR kinase and BRCA2, which is independent from the induction of acute replication stress, is initiated by augmentation of binucleation, is prevented by EMI1 depletion and recapitulated by Citron Rho-Interacting kinase (another non-DDR kinase), and provides a certain level of mechanistic insight. We understand that reviewer 2 wants us to acknowledge that we have not identified the molecular target of ROCKi that triggers the excess cytokinesis failure in BRCA2-deficient cells. We have made this limitation explicit in the Discussion section of the R3 version of this manuscript (lines 328- 332).

As the reviewers suggested, I have considered the 4 points put forward in the rebuttal letter (especially points 1,3 and 4):

1. ROCKi can kill BRCA2-deficient cells in a manner that is independent of replication stress in S phase. Such a concept has been reinforced in the R1 version by adding fiber assays and including the quantification of H2AX suggested by reviewer 2.

True, there is no overt replication stress induced by acute ROCK inhibition that can explain the SL effects. The mode of action of PARP inhibitor is clearly different than those of ROCK inhibitors in this respect. However, whether replication stress is not involved in this form of SL is difficult to conclude.

The effects of 53Bp1, H2AX and fibers were done after 48h of treatment, whereas the toxic effects of ROCK inhibition occur much later, when binucleated cells have occured. Since whole genome duplication upon mitotic failure is known to cause replication stress (see for instance Gemble et al., Nature, 2022). To really conclude that replication stress is not involved, analysis at a late time points should be done. Of note. this point of the author does not indicate an underlying mechanism of action, but suggests a mechanism that is not involved.

Because this section starts with “As the reviewers suggested…” we assume that this part of the letter has been written by the reviewing editor or the editor. We would like to emphasize that we never claimed that the replication stress was absent. We focused on acute replication stress which is the type of stress triggered by PARP inhibitors. Our intention was to highlight the mechanistic differences between ROCK and PARP inhibitors. However, and due to this request, we have performed the experiment requested herein by reviewer 2 and the editors. Surprisingly, there is no replication stress at later time points for mono, bi and multinucleated cells, which reinforce the replication stress-independent nature of this SL between ROCKi and BRCA2deficiency.

2. The cytokinesis defects triggered by ROCKi have been documented by others, and so was the contribution of BRCA2-depletion to cytokinesis defects. By monitoring cells with DNA content higher than 4 N and binucleation, this manuscript shows that the treatment of BRCA2-deficient cells with ROCKi causes a significant augmentation of binucleation at early time points hence suggesting that the cytokinesis defects is a common defect that is amplified by the combined depletion of BRCA2 and the inhibition of ROCKi. Moreover, supporting the idea of cytokinesis defect as the trigger of other defects, including the DNAreplication independent SL, it temporally precedes all other mitotic defects.

The mechanism proposed here is that ROCKi and BRCA2 inactivation both give cytokinesis defects, and that the SL effect is based on these effects adding up. This is not entirely novel, and also, BRCA2 inactivation in this manuscript only marginally induces cytokinesis defects on its own. That cytokinesis failure leads to secondary defects I completely support, but claiming that secondarily effects are DNA replication independent is not shown and should not be claimed.

We did not claim independence from replication stress. Throughout the whole manuscript (original, R1 and R2) we did purposely limit such a conclusion to acute replication stress which is what we tested. Because of the experiments requested by the editors and reviewer 2 in this R3 round, we have monitored replication stress at the same time we measure cell death (6 days) by means of H2AX intensity and 53BP1 foci formation, in two cell lines (shBRCA2 HCT116p21-/- and PEO1 BRCA2-/-) (see new Figure 3—figure supplement 2). Surprisingly, there was no replication stress even at those late time points for mono, bi and multinucleated cells. Such observation reinforces the dissociation of this SL from replication stress.

3. Reinforcing the link between cytokinesis failure and SL, the inhibition of another kinase that has also been associated with cytokinesis defect, Citron Rho-interacting kinase, also triggers SL in BRCA-2 deficient cells.

True, the observation that Rho kinase and ROCK behave the same is strong, and points towards cytokinesis, in which both kinases are clearly implicated.

Thank you for agreeing with point 3.

4. The prevention of mitosis by EMI1 depletion prevents the SL caused by ROCKi in BRCA2- deficient cells, thus reinforcing the association between M-phase transit and cell death in BRCA2-deficient cells.

True, the EMI experiment connects mitosis to the SL effect of ROCK inhibition in these cells, reinforcing that a cytokinesis failure is involved.

Thank you for agreeing with point 4.

https://doi.org/10.7554/eLife.80254.sa2

Article and author information

Author details

  1. Julieta Martino

    Fundación Instituto Leloir-CONICET, Buenos Aires, Argentina
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – original draft, Writing – review and editing
    Contributed equally with
    Sebastián Omar Siri
    Competing interests
    No competing interests declared

  2. Sebastián Omar Siri

    Fundación Instituto Leloir-CONICET, Buenos Aires, Argentina
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing – review and editing
    Contributed equally with
    Julieta Martino
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0945-5605

  3. Nicolás Luis Calzetta

    Fundación Instituto Leloir-CONICET, Buenos Aires, Argentina
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared

  4. Natalia Soledad Paviolo

    Fundación Instituto Leloir-CONICET, Buenos Aires, Argentina
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization
    Competing interests
    No competing interests declared

  5. Cintia Garro

    1. Centro de Investigaciones en Bioquímica Clínica e Inmunología, CIBICI-CONICET, Departamento de Bioquímica Clínica, Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, Córdoba, Argentina
    2. OncoPrecision, Córdoba, Argentina
    Contribution
    Data curation, Formal analysis, Investigation, Visualization
    Competing interests
    No competing interests declared

  6. Maria F Pansa

    Centro de Investigaciones en Bioquímica Clínica e Inmunología, CIBICI-CONICET, Departamento de Bioquímica Clínica, Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, Córdoba, Argentina
    Present address
    GlaxoSmithKline-Trust in Science, Global Health R&D, Upper Providence, United States
    Contribution
    Data curation, Formal analysis
    Competing interests
    is affiliated with GlaxoSmithKline and has no other competing interests to declare

  7. Sofía Carbajosa

    1. Centro de Investigaciones en Bioquímica Clínica e Inmunología, CIBICI-CONICET, Departamento de Bioquímica Clínica, Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, Córdoba, Argentina
    2. OncoPrecision, Córdoba, Argentina
    Contribution
    Data curation, Formal analysis
    Competing interests
    No competing interests declared

  8. Aaron C Brown

    Center for Molecular Medicine, Maine Medical Center Research Institute, Scarborough, United States
    Contribution
    Data curation, Formal analysis, Methodology
    Competing interests
    No competing interests declared

  9. José Luis Bocco

    Centro de Investigaciones en Bioquímica Clínica e Inmunología, CIBICI-CONICET, Departamento de Bioquímica Clínica, Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, Córdoba, Argentina
    Contribution
    Resources, Funding acquisition
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-9682-1270

  10. Israel Gloger

    GlaxoSmithKline-Trust in Science, Global Health R&D, Stevenage, United Kingdom
    Contribution
    Data curation, Formal analysis
    Competing interests
    is affiliated with GlaxoSmithKline and has no other competing interests to declare

  11. Gerard Drewes

    GlaxoSmithKline-Trust in Science, Global Health R&D, Stevenage, United Kingdom
    Contribution
    Data curation, Formal analysis
    Competing interests
    is affiliated with GlaxoSmithKline and has no other competing interests to declare

  12. Kevin P Madauss

    GlaxoSmithKline-Trust in Science, Global Health R&D, Upper Providence, United States
    Contribution
    Data curation, Formal analysis, Supervision, Methodology
    Competing interests
    is affiliated with GlaxoSmithKline and has no other competing interests to declare

  13. Gastón Soria

    1. Centro de Investigaciones en Bioquímica Clínica e Inmunología, CIBICI-CONICET, Departamento de Bioquímica Clínica, Facultad de Ciencias Químicas, Universidad Nacional de Córdoba, Córdoba, Argentina
    2. OncoPrecision, Córdoba, Argentina
    Contribution
    Conceptualization, Resources, Data curation, Supervision, Funding acquisition, Investigation, Methodology, Project administration, Writing – review and editing
    For correspondence
    gsoria29@gmail.com
    Competing interests
    No competing interests declared

  14. Vanesa Gottifredi

    Fundación Instituto Leloir-CONICET, Buenos Aires, Argentina
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Investigation, Methodology, Writing – original draft, Project administration, Writing – review and editing
    For correspondence
    vgottifredi@leloir.org.ar
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9656-5951

Funding

L'Oréal (National Award 2019)

  • Vanesa Gottifredi

Agencia Nacional de Promoción Científica y Tecnológica (PAE-GLAXO 2014–0005)

  • José Luis Bocco

GlaxoSmithKline (PAE-GLAXO 2014–0005)

  • José Luis Bocco

Agencia Nacional de Promoción Científica y Tecnológica (PCE-GSK 2017–0032)

  • Gastón Soria

GlaxoSmithKline (PCE-GSK 2017–0032)

  • Gastón Soria

Agencia Nacional de Promoción Científica y Tecnológica (PICT 2018–0185 )

  • Vanesa Gottifredi

Consejo Nacional de Investigaciones Científicas y Técnicas (Researcher)

  • José Luis Bocco
  • Gastón Soria
  • Vanesa Gottifredi

Agencia Nacional de Promoción Científica y Tecnológica (Fellowship)

  • Julieta Martino
  • Sebastián Omar Siri
  • Natalia Soledad Paviolo
  • Sofía Carbajosa

Consejo Nacional de Investigaciones Científicas y Técnicas (Fellowship)

  • Sebastián Omar Siri
  • Nicolás Luis Calzetta
  • Natalia Soledad Paviolo
  • Cintia Garro
  • Maria F Pansa

Instituto Nacional del Cáncer (Fellowship)

  • Maria F Pansa

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Dr. Fernanda Ledda for providing critical reagents for this work. We would also like to thank all Gottifredi and Soria Laboratories members for their insightful comments and discussions. We thank Pamela Rodriguez, Esteban Miglietta, Andrés Hugo Rossi and Carla Pascuale for technical support with tissue culture, microscopy and flow cytometry. We also thank the flow cytometry, microscopy, and cell culture facilities of CIBICI-CONICET for technical support. This work was supported by a consortium grant of FONCyT and the Trust in Science Program (Global Health R&D) from GlaxoSmithKline (PAE-GLAXO 2014–0005) to JLB and (PCE-GSK 2017–0032) to GS and PICT 2018–01857 and L'Oréal-UNESCO National Award 2019 to VG. JLB, GS and VG are researchers from the National Council of Scientific and Technological Research (CONICET). JM, SOS, NSP and SC were supported by fellowships from the National Agency for the Promotion of Science and Technology (ANPCyT). SOS, NLC, NSP, CG and MFP were supported by fellowships from CONICET. MFP was supported by a fellowship from the National Institute of Cancer (Argentina).

Senior Editor

  1. Wafik S El-Deiry, Brown University, United States

Reviewing Editor

  1. Robert W Sobol, Brown University, United States

Publication history

  1. Received: May 13, 2022
  2. Preprint posted: June 28, 2022 (view preprint)
  3. Accepted: April 18, 2023
  4. Accepted Manuscript published: April 19, 2023 (version 1)
  5. Version of Record published: May 15, 2023 (version 2)

Copyright

© 2023, Martino, Siri et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 770
    Page views
  • 201
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Julieta Martino
  2. Sebastián Omar Siri
  3. Nicolás Luis Calzetta
  4. Natalia Soledad Paviolo
  5. Cintia Garro
  6. Maria F Pansa
  7. Sofía Carbajosa
  8. Aaron C Brown
  9. José Luis Bocco
  10. Israel Gloger
  11. Gerard Drewes
  12. Kevin P Madauss
  13. Gastón Soria
  14. Vanesa Gottifredi
(2023)
Inhibitors of Rho kinases (ROCK) induce multiple mitotic defects and synthetic lethality in BRCA2-deficient cells
eLife 12:e80254.
https://doi.org/10.7554/eLife.80254

Categories and tags

Further reading

    1. Cancer Biology
    2. Computational and Systems Biology

    Comprehensive characterization of tumor microenvironment in colorectal cancer via molecular analysis

    Xiangkun Wu, Hong Yan ... Li Liang
    Research Article

    Colorectal cancer (CRC) remains a challenging and deadly disease with high tumor microenvironment (TME) heterogeneity. Using an integrative multi-omics analysis and artificial intelligence-enabled spatial analysis of whole-slide images, we performed a comprehensive characterization of TME in colorectal cancer (CCCRC). CRC samples were classified into four CCCRC subtypes with distinct TME features, namely, C1 as the proliferative subtype with low immunogenicity; C2 as the immunosuppressed subtype with the terminally exhausted immune characteristics; C3 as the immune-excluded subtype with the distinct upregulation of stromal components and a lack of T cell infiltration in the tumor core; and C4 as the immunomodulatory subtype with the remarkable upregulation of anti-tumor immune components. The four CCCRC subtypes had distinct histopathologic and molecular characteristics, therapeutic efficacy, and prognosis. We found that the C1 subtype may be suitable for chemotherapy and cetuximab, the C2 subtype may benefit from a combination of chemotherapy and bevacizumab, the C3 subtype has increased sensitivity to the WNT pathway inhibitor WIKI4, and the C4 subtype is a potential candidate for immune checkpoint blockade treatment. Importantly, we established a simple gene classifier for accurate identification of each CCCRC subtype. Collectively our integrative analysis ultimately established a holistic framework to thoroughly dissect the TME of CRC, and the CCCRC classification system with high biological interpretability may contribute to biomarker discovery and future clinical trial design.

    1. Cancer Biology

    Defining function of wild-type and three patient specific TP53 mutations in a zebrafish model of embryonal rhabdomyosarcoma

    Jiangfei Chen, Kunal Baxi ... Myron S Ignatius
    Research Article

    In embryonal rhabdomyosarcoma (ERMS) and generally in sarcomas, the role of wild-type and loss or gain-of-function TP53 mutations remains largely undefined. Eliminating mutant or restoring wild-type p53 is challenging; nevertheless, understanding p53 variant effects on tumorigenesis remains central to realizing better treatment outcomes. In ERMS, >70% of patients retain wild-type TP53, yet mutations when present are associated with worse prognosis. Employing a kRASG12D-driven ERMS tumor model and tp53 null (tp53-/-) zebrafish, we define wild-type and patient-specific TP53 mutant effects on tumorigenesis. We demonstrate that tp53 is a major suppressor of tumorigenesis, where tp53 loss expands tumor initiation from <35% to >97% of animals. Characterizing three patient-specific alleles reveals that TP53C176F partially retains wild-type p53 apoptotic activity that can be exploited, whereas TP53P153D and TP53Y220C encode two structurally related proteins with gain-of-function effects that predispose to head musculature ERMS. TP53P153D unexpectedly also predisposes to hedgehog expressing medulloblastomas in the kRASG12D-driven ERMS-model.

    1. Cancer Biology

    Proteomic characteristics reveal the signatures and the risks of T1 colorectal cancer metastasis to lymph nodes

    Aojia Zhuang, Aobo Zhuang ... Chen Ding
    Research Article Updated

    The presence of lymph node metastasis (LNM) affects treatment strategy decisions in T1NxM0 colorectal cancer (CRC), but the currently used clinicopathological-based risk stratification cannot predict LNM accurately. In this study, we detected proteins in formalin-fixed paraffin-embedded (FFPE) tumor samples from 143 LNM-negative and 78 LNM-positive patients with T1 CRC and revealed changes in molecular and biological pathways by label-free liquid chromatography tandem mass spectrometry (LC-MS/MS) and established classifiers for predicting LNM in T1 CRC. An effective 55-proteins prediction model was built by machine learning and validated in a training cohort (N=132) and two validation cohorts (VC1, N=42; VC2, N=47), achieved an impressive AUC of 1.00 in the training cohort, 0.96 in VC1 and 0.93 in VC2, respectively. We further built a simplified classifier with nine proteins, and achieved an AUC of 0.824. The simplified classifier was performed excellently in two external validation cohorts. The expression patterns of 13 proteins were confirmed by immunohistochemistry, and the IHC score of five proteins was used to build an IHC predict model with an AUC of 0.825. RHOT2 silence significantly enhanced migration and invasion of colon cancer cells. Our study explored the mechanism of metastasis in T1 CRC and can be used to facilitate the individualized prediction of LNM in patients with T1 CRC, which may provide a guidance for clinical practice in T1 CRC.