A haploid genetic screen identifies the G1/S regulatory machinery as a determinant of Wee1 inhibitor sensitivity

Abstract

The Wee1 cell cycle checkpoint kinase prevents premature mitotic entry by inhibiting cyclin-dependent kinases. Chemical inhibitors of Wee1 are currently being tested clinically as targeted anticancer drugs. Wee1 inhibition is thought to be preferentially cytotoxic in p53-defective cancer cells. However, TP53 mutant cancers do not respond consistently to Wee1 inhibitor treatment, indicating the existence of genetic determinants of Wee1 inhibitor sensitivity other than TP53 status. To optimally facilitate patient selection for Wee1 inhibition and uncover potential resistance mechanisms, identification of these currently unknown genes is necessary. The aim of this study was therefore to identify gene mutations that determine Wee1 inhibitor sensitivity. We performed a genome-wide unbiased functional genetic screen in TP53 mutant near-haploid KBM-7 cells using gene-trap insertional mutagenesis. Insertion site mapping of cells that survived long-term Wee1 inhibition revealed enrichment of G1/S regulatory genes, including SKP2, CUL1, and CDK2. Stable depletion of SKP2, CUL1, or CDK2 or chemical Cdk2 inhibition rescued the γ-H2AX induction and abrogation of G2 phase as induced by Wee1 inhibition in breast and ovarian cancer cell lines. Remarkably, live cell imaging showed that depletion of SKP2, CUL1, or CDK2 did not rescue the Wee1 inhibition-induced karyokinesis and cytokinesis defects. These data indicate that the activity of the DNA replication machinery, beyond TP53 mutation status, determines Wee1 inhibitor sensitivity, and could serve as a selection criterion for Wee1-inhibitor eligible patients. Conversely, loss of the identified S-phase genes could serve as a mechanism of acquired resistance, which goes along with development of severe genomic instability.

Keywords: AZD-1775; MK-1775; cell cycle; checkpoint; polyploidy.

Fig. S1.

Mutations in G₁/S regulatory genes are enriched in KBM-7 cells that survive MK-1775 treatment. (A) KBM-7, MDA-MB-231, and SKOV3 cells were treated with increasing MK-1775 concentrations and viability was assessed 72 h after treatment using methyl thiazol tetrazolium (MTT). Averages and error bars of at least three replicates are shown. (B) KBM-7 cells were treated with MK-1775 (300 nM) for indicated time points and analyzed by Western blotting for indicated proteins. (C) KBM-7 cells were treated with MK-1775 (300 nM) and cell cycle profiles were determined at indicated time points by flow cytometric analysis after staining with propidium iodide. Plotted are the average percentages and SDs of sub-G₁ cells from three experiments. (D) Schematic representation of the genetic screen with the Wee1 inhibitor MK-1775 in KBM-7 cells.

Fig. 1.

Haploid genetic screen identifies S phase genes as determinants of Wee1 inhibitor sensitivity. (A) Identification of gene-trap insertions enriched in sense orientation in MK-1775-selected KBM-7 cells. Y axis indicates fraction of gene-traps in sense orientation compared with total insertions. X axis indicates number of gene-trap insertions. (B) Network modeling with 142 genes enriched in sense orientation. Most significant module is shown. Red and yellow proteins were identified in the screen. Indirect (dashed lines) and direct interactions (solid lines) are indicated. Arrowheads indicate interaction direction. (C) Flow cytometric analysis of pLKO.mCherry transduced KBM-7 cells 10 d after MK-1775 (150 nM) or DMSO treatment. (D) Ratio of mCherry-positive cells of Wee1-inhibited vs. DMSO-treated KBM-7 cells (150 nM MK-1775) and MDA-MB-231/SKOV3 cells (1 µM MK-1775). (E) Nontransformed Tp53^−/−,Cdk2^−/−, or Tp53^−/− MEFs were treated for 4 d with 500 nM MK1775 or DMSO, and stained with crystal violet.

Fig. S2.

Canonical pathways of mutated genes, enriched in MK-1775–resistant KBM-7 cells. Canonical pathway analysis was performed with the 142 selected genes using Ingenuity Pathway Analysis (IPA) software (Qiagen). Presented are the canonical pathways that have a −log(P value) score greater than 1.5.

Fig. S3.

Knockdown efficiencies of S-phase genes and long-term survival assay with chemical Cdk2 inhibition. (A) KBM-7, MDA-MB-231 and SKOV3 cells were transduced with indicated shRNA vectors. Knockdown efficiencies of the various shRNAs were assessed by Western blotting. (B) Flow cytometric analysis of pLKO.mCherry transduced KBM-7 cells treated with Wee1 inhibitor PD-166285 (2.5 nM). The ratio of mCherry-positive KBM-7 cells in Wee1-inhibited versus DMSO-treated cells is presented. (C) Flow cytometric analysis of KBM-7 cells transduced with pLKO.mCherry-SCR or pLKO.mCherry-CDK2 in combination with pLKO.puro-WEE1 or pLKO.puro-SCR. The ratio of mCherry-positive KBM-7 cells in Wee1-depleted versus control-depleted cells at indicated time points is shown in the graph. WEE1 shRNA knockdown efficiency was assessed by Western blotting and is shown on the left. (D) Flow cytometric analysis of pLKO.mCherry transduced MDA-MB-231 cells at day 0 and 10 d after MK-1775 (1 µM) or DMSO treatment is indicated. (E) Long-term clonogenic survival assay of MDA-MB-231 and SKOV3 cells treated with increasing concentrations of Cdk2 inhibitor SU-9516 with or without MK-1775 (500 nM).

Fig. S4.

Nontransformed and Ras-V12-transformed Tp53^−/−, Cdk2^−/− MEFs are resistant to Wee1 inhibition. (A) Four-day survival assay of Ras-V12-transformed Tp53^−/−, Cdk2^−/−, or Tp53^−/− MEFs treated with MK-1775 (500 nM) or DMSO. After 4 d of treatment, cells were fixed, stained with crystal violet solution, and photographed. (B) Nontransformed Tp53^−/−, Cdk2^−/−, or Tp53^−/− MEFs were grown in the presence of MK-1775 (500 nM) or DMSO. Relative cell viability was measured after 4 d by crystal violet quantification. Averages and SDs of two independent experiments are indicated.

Fig. 2.

Wee1 inhibition exerts its cytotoxic effects during S phase. (A) MDA-MB-231 cells were synchronized in G₁/S by thymidine or in G₂/M by nocodazole. Cells were treated with MK-1775 (4 μM) or DMSO upon wash-out, fixed at indicated time points and stained for propidium iodide. Shown is the percentage of sub-G₁ cells, calculated as the mean of three experiments (*P < 0.05, **P < 0.01, Student’s t test). (B) Thymidine synchronized MDA-MB-231 cells were treated with or without MK-1775 (4 μM) directly or 6 h after wash out. Cells were fixed at indicated time points and stained for γ-H2AX (DNA damage) and propidium iodide. The mean of the percentage of γ-H2AX–positive cells of two experiments is shown (**P < 0.01, ***P < 0.001, Student’s t test). (C) pLKO.mCherry transduced MDA-MB-231 cells were treated with MK-1775 (4 μM). Cells were fixed at indicated time points and stained as in B. The mean of the percentage of γ-H2AX–positive cells of two experiments is shown (*P < 0.05, **P < 0.01, Student’s t test). In all figures error bars indicate SD.

Fig. S5.

Reduced γ-H2AX formation after combined inhibition of Wee1 and Cdk2. (A) MDA-MB-231 cells were synchronized at the G₁/S transition using thymidine, and were treated with MK-1775 (4 µM) or DMSO upon release. At indicated time points, cells were fixed and stained with propidium iodide. Cell cycle profiles were determined by flow cytometry. (B) MDA-MB-231 cells were synchronized in mitosis using nocodazole. Upon nocodazole wash-out, cells were treated with MK-1775 (4 μM) or DMSO and analyzed as in A. (C) MDA-MB-231 cells were synchronized at the G₁/S transition using thymidine and treated with or without MK-1775 directly or at 6 h after thymidine wash-out. Cells were fixed at indicated time points and stained for γ-H2AX/Alexa-488 and propidium iodide and analyzed by flow cytometry. (D) Flow cytometric analysis of pLKO.mCherry transduced MDA-MB-231 cells treated with MK-1775 (4 μM). Cells were fixed at indicated time points and stained for γ-H2AX and propidium iodide. (E and F) MDA-MB-231 cells were synchronized at the G₁/S transition with thymidine. Upon thymidine wash-out, cells were treated with MK-1775 (4 μM) and/or with SU-9516 (1 μM). Cells were fixed at indicated time points and subjected to γ-H2AX/Alexa-488 staining and propidium iodide staining and analyzed by flow cytometry. Positive staining for γ-H2AX in G₂-cells was quantified with FlowJo, using gates as indicated in the figure panels. Representative images of MK-1775–treated and MK-1775/SU-9516–treated cells are indicated in E. The averages and SDs of γ-H2AX–positive cells of two experiments are presented in F.

Fig. 3.

Wee1 inhibition abrogates G₂ phase. (A) FUCCI-MDA-MB-231 or FUCCI-SKOV3 cells were treated with 4 μM MK-1775 and/or 1 μM Cdk2 inhibitor SU-9516 and were subsequently imaged every 7 min for 62 h. The duration of G₂ phase was analyzed for >28 cells per condition. (B) pLKO.puro transduced FUCCI-MDA-MB-231 cells were treated with or without MK-1775 (4 μM) and were imaged as described in A. The duration of G₂ phase was analyzed for >20 cells per condition.

Fig. S6.

Inhibition or siRNA-mediated depletion of Wee1 induces cytokinesis failure. (A) FUCCI-MDA-MB-231 cells were left untreated or were treated with MK-1775 (4 μM) alone or in combination with SU-9516 (1 μM). Subsequently, cells were imaged every 7 min for 62 h. Progression from S via G₂ into M phase was monitored using FUCCI reporters to score the duration of G₂ phase. Representative experiments are shown, with t = 0 denoting the onset of G₂ phase. Zoom-in images of right-most panels indicate cell boundaries (dashed lines) and nuclear content (arrows) after cell division. (B) MDA-MB-231 and SKOV3 cells were transfected with SCR or Wee1 siRNA. After 4 d, cells were fixed and stained as described in (Fig. 4B). The percentage of multinucleated cells was quantified and averages and SDs of two experiments are indicated. (C) Quantification of the percentages of multinucleated MDA-MB-231 cells, treated and imaged as in (Fig. 4B). (D) MDA-MB-231 cells were left untreated (asynchronous, AS) or were treated with nocodazole for 16 h. Mitotic cells after nocodazole treatment were obtained by mitotic shake-off and replated in the presence or absence of MK-1775 (4 μM). At indicated time points, cells were harvested and processed for immunoblotting. (E) MDA-MB-231 cells were transiently transfected with CDK1-AF-CFP or CDK1-WT-CFP. Subsequently, cells were treated with MK-1775 (4 μM) and live cell images for DIC and CFP were obtained every 7 min for 62 h. Mitotic progression was monitored and scored for normal or aberrant mitotic exit. Representative images are shown, with t = 0 denoting the onset of mitosis.

Fig. S7.

G₂/M checkpoint function and DNA damage formation during replication is not altered by Cdk2 inactivation. (A and B) MDA-MB-231 cells were synchronized at the G₁/S transition with thymidine. Upon thymidine wash-out, cells were treated with MK-1775 (4 μM) and/or SU-9516 (1 μM). Cells were fixed at indicated time points and stained for MPM2/Alexa-488, γ-H2AX/Alexa-647 and propidium iodide and analyzed by flow cytometry. For analysis of γ-H2AX in mitotic cells, MPM2-positive cells were gated, and MFI of γ-H2AX was quantified. Representative images of MK-1775–treated and MK-1775/SU-9516–treated cells are shown in A. The averages and SDs of γ-H2AX MFI of two experiments are presented in B. (C and D) MDA-MB-231 cells were synchronized and treated as in A. Cells were fixed at indicated time points and stained for γ-H2AX/Alexa-647 and propidium iodide and analyzed by flow cytometry. Mean fluorescence intensity (MFI) of γ-H2AX in S phase was quantified. Representative images of MK-1775–treated and MK-1775/SU-9516–treated cells are indicated in C. The averages and SDs of γ-H2AX MFI of two experiments are presented in D. (E) Control-depleted or Cdk2-depleted MDA-MB-231 cells were treated with doxorubicin (dox, 1μM) for 16 h. Thirty minutes after doxorubicin treatment, nocodazole was added to cultures (250 ng/mL). If indicated MK-1775 (4 µM; Wee1i) was added at the time of doxorubicin treatment. Mitotic cells were identified based on MPM2 reactivity. Representative DNA profiles MPM2 plots are indicated for Cdk2-depleted cells. (F). Quantification of data from E. Averages and SDs of two independent experiments are shown.

Fig. 4.

Wee1 inhibition induces Cdk1-dependent cytokinesis failure. (A) FUCCI-MDA-MB-231 cells were treated and imaged as in Fig. 3 and Fig. S6A. The percentages of multinucleated MDA-MB-231 and SKOV3 cells were quantified at indicated time points. (B) Representative immunofluorescence images of untreated MDA-MB-231 colonies or colonies that survived combined MK-1775 (500 nM) and SU-9516 (1 μM) treatment. Cells were fixed at 14 d after treatment and stained for CD44 (red) and DNA (white). (C) pLKO.puro transduced MDA-MB-231 and SKOV3 cells were treated with MK-1775 (500 nM) for 14 d and stained as in B. Percentages of multinucleated cells were quantified and presented. (D) MDA-MB-231 cells transfected with CDK1-AF or CDK1-WT were treated with MK-1775 (4 μM) and subsequently imaged every 7 min for 62 h. Mitotic exit was analyzed for >15 CFP-positive cells. (E) A model for Wee1 inhibition causing cytotoxicity and genomic instability.

Fig. S8.

Association between Wee1 inhibitor sensitivity and mRNA expression of CDK2, SKP2, and CUL1. mRNA expression levels and Wee1 inhibitor sensitivity data of 33 breast and ovarian cancer cell lines was retrieved from publicly available datasets (see SI Materials and Methods and Dataset S3). (A) CDK2 mRNA expression values and corresponding IC50, IC75 and IC90 values for Wee1 inhibitor 681640 values are plotted. R² values define the goodness-of-fit of a linear regression. (B and C) Linear regression analysis as performed in A with SKP2 or CUL1 mRNA levels, respectively.