A Genome-wide CRISPR Screen Identifies CDC25A as a Determinant of Sensitivity to ATR Inhibitors

Abstract

One recurring theme in drug development is to exploit synthetic lethal properties as means to preferentially damage the DNA of cancer cells. We and others have previously developed inhibitors of the ATR kinase, shown to be particularly genotoxic for cells expressing certain oncogenes. In contrast, the mechanisms of resistance to ATR inhibitors remain unexplored. We report here on a genome-wide CRISPR-Cas9 screen that identified CDC25A as a major determinant of sensitivity to ATR inhibition. CDC25A-deficient cells resist high doses of ATR inhibitors, which we show is due to their failure to prematurely enter mitosis in response to the drugs. Forcing mitotic entry with WEE1 inhibitors restores the toxicity of ATR inhibitors in CDC25A-deficient cells. With ATR inhibitors now entering the clinic, our work provides a better understanding of the mechanisms by which these compounds kill cells and reveals genetic interactions that could be used for their rational use.

Figure 1. Efficient and Dox-inducible gene knockouts in ES^Cas9 cells.

(A) Scheme illustrating the two-allele system used for the generation of ES^Cas9 cells. In this previously described system(Beard et al., 2006), the Cas9 cDNA is placed under the control of a tet-responsible sequence (tetO) at the Col1a1 locus. At the same time, the reverse tetracycline-controlled transactivator (rtTA) is expressed from the Rosa26 locus, providing Dox-inducible-activation of Cas9 expression. (B) Levels of Cas9 mRNA evaluated by RT-PCR (normalized to levels of GAPDH mRNA) in the 2 clones of ES^Cas9 cells used in this study. The high stringency of the system prevents cleavage in the absence of Dox. Data are represented as mean ± s.d. (n=3). See also Figure S1A,B. (C) FACS analysis illustrating the loss of GFP signal in ES^Cas9 cells that were made GFP positive by infection with a lentiviral construct expressing GFP, and simultaneously infected with a lentivirus expressing a Gfp-targeting sgRNA together with BFP. Doubly infected cells are thus BFP and GFP positive, and gradually lose GFP expression after the addition of Dox. See also Figure S1C. (D) Western blot illustrating the Dox-dependent loss of P53 expression observed in ES^Cas9 cells that were infected with lentiviruses expressing 3 independent P53-targeting sgRNAs. TUBULIN levels are shown as a loading control. Dox was used at 2 μg/ml.

Figure 2. Resistance to ATR inhibition in CDC25A-deficient ES^Cas9 cells.

(A) Western blot illustrating the loss of CDC25A expression in 3 independent ATRi-resistant clones. (B) Examples of the mutations in Cdc25a identified in ATRi-resistant ES^Cas9 cells shown in (A). Note that all mutations include small insertions or deletions that change the reading frame in exon 2, precisely at the 3’ of the sgRNA sequence (in red). (C) Representative pictures of the resistance to ATRi observed in CDC25A-deficient ES^Cas9 cells exposed to the compound for 72 hr. ES cells are grown on top of a feeder layer of growth-arrested MEF, which are unaffected by the treatment. Scale bar (white) indicates 25 μm. (D) XTT viability assay in wild type and CDC25A-deficient ES^Cas9 cells exposed to ATRi for 24 hr at the indicated doses. Data are representative of 3 independent experiments. Error bars indicate s.d. See also Figure S2.

Figure 3. ATR inhibition generates RS but not DSB in CDC25A-deficient cells.

(A) HTM analysis of the EdU signal per individual nucleus in wild type (WT) and CDC25A-deficient ES^Cas9 cells exposed to ATRi for 1 hr (900 nM). (B) HTM analysis of the BrdU signal accumulated in ssDNA per individual nucleus in wild type (WT) and CDC25A-deficient ES^Cas9 cells exposed to ATRi for 4 hr (900 nM). (C) WB illustrating the levels of H2AX (γH2AX) and RPA (S4/8) phosphorylation in WT and CDC25A-deficient ES^Cas9 cells exposed to ATRi for 4 hr (0, 0.3, 0.9, 2 and 3 μM). TUBULIN and total RPA levels are shown as loading controls. (D) HTM analysis of the γH2AX signal per individual nucleus in WT and CDC25A-deficient ES^Cas9 cells treated (or not) with ATRi (0.3, 0.9 and 2 μM; 4 hr). See also Figure S3.

Figure 4. ATRi induces premature mitotic entry and DNA breakage in a CDC25A-dependent manner.

(A) Representative FACS profiles showing the distribution of γH2AX-positive cells (x, DNA content; y, γH2AX) of WT and CDC25A-deficient ES^Cas9 cells exposed to ATRi (900 nM) for the indicated times. Numbers indicate the percentage of γH2AX-positive cells (red) in each case. (B) Representative image of the presence of DSBs in mitotic ES^Cas9 cells exposed to ATRi for 4 hr (300 nM). Mitotic cells were identified by antibodies against pH3S10 (red). DSBs were identified by γH2AX (green) foci. DAPI was used to stain DNA. Numbers in white indicate the percentage of pH3S10-positive cells presenting γH2AX foci. Scale bar (white) indicates 2.5 μm. (C) Representative FACS profiles showing the distribution of pH3S10-positive cells (x, DNA content; y, pH3S10) of WT and CDC25A-deficient ES^Cas9 cells exposed to ATRi (900 nM) for the indicated times. Numbers indicate the percentage of pH3S10-positive cells (red) in each case. (D) WB illustrating the levels of H2AX (γH2AX), RPA (S4/8) and H3 (S10) phosphorylation in WT and CDC25A-deficient ES^Cas9 cells exposed to ATRi (300 nM) and/or WEE1i (100 or 300 nM) for 4 hr. An antibody detecting phosphorylated CDK substrates was also used as measure of overall CDK activity. TUBULIN and total RPA levels are shown as loading controls. (E) Representative pictures of cultures of WT and CDC25A-deficient ES^Cas9 cells exposed to ATRi (300 nM) and/or WEE1i (100 nM) for 72 hr. Scale bar (white) indicates 25 μm. See also Figure S4.