SOD1 is a synthetic lethal vulnerability of PPM1D-mutant leukemia cells.

(A) Schematic of whole-genome CRISPR dropout screen. WT Cas9-expressing OCI-AML2 and two isogenic PPM1D-mutant lines with three technical replicates were transduced with the Human Improved Whole Genome Knockout CRISPR library V1 containing 90,709 guide RNAs (gRNAs) targeting 18,010 human genes at low multiplicity of infection (MOI∼0.3). Three days post-transduction, cells underwent puromycin selection for three days. Cells were harvested at day 10 as the initial timepoint and then harvested every three days. sgRNA sequencing was performed on cells collected on day 28. (B) Top biological processes based on gene ontology analysis of the top 37 genes essential for PPM1D-mutant cell survival. Enrichment and depletion of guides and genes were analyzed using MAGeCK-VISPR by comparing read counts from each PPM1D-mutant cell line replicate with counts from the initial starting population at day ten. (C) Volcano plot of synthetic lethal hits ranked by fitness score with a negative score indicating genes for which their knockout leads decreased growth or survival. SOD1 (highlighted) was the top hit from the screen. (D) Left: Schematic of competitive proliferation assays used for validation of CRISPR targets. Right: WT and PPM1D-mutant Cas9-OCI-AML2 and Cas9-OCI-AML3 cells were transduced with lentivirus containing a single SOD1-gRNA with a blue fluorescent protein (BFP) reporter. Cells were assayed by flow cytometry every 3-4 days post-transduction and normalized to the BFP percentage at day 3. Two unique gRNAs against SOD1 were used per cell line and each condition was performed in technical duplicates; multiple unpaired t-tests, **p<0.01, ***p<0.001.

PPM1D-mutant cells are sensitive to SOD1 inhibition and have increased oxidative stress.

(A,B) Dose response curves for cell viability with SOD1-inhibitor (LCS-1) (A) or LCS-1 in combination with 0.25 uM NAC (B) in WT and PPM1D-mutant leukemia cell lines after 24-hours. Mean + SD (n=3) is shown with a non-linear regression curve. All values are normalized to the baseline cell viability with vehicle, as measured by MTT assay. (C) Endogenous mitochondrial superoxide levels of WT and PPM1D-mutant leukemia cell lines were measured using MitoSox Green staining (1 uM). The mean fluorescence intensity (MFI) of MitoSox Green was measured by flow cytometry. Mean + SD (n=3) is shown. (D) Lipid peroxidation measured using BODIPY 581/591 staining (2.5 uM) of WT and PPM1D-mutant OCI-AML2 cells. The MFI was measured by flow cytometry. Mean + SD (n=3) is shown. (E-F) Measure of total reactive oxygen species using 2’,7’–dichlorofluorescin diacetate (DCFDA) staining (10 uM) measured by flow cytometry. WT and PPM1D-mutant OCI-AML2 cells were measured at baseline and 24-hrs after SOD1 inhibition (ATN-224 12.5 uM, LCS-1 0.625 uM) (E) or 24-hrs after pharmacologic PPM1D inhibition (GSK2830371, 5 uM) (F); unpaired t-tests were used for statistical analyses, ns=non-significant (p>0.05), **p<0.01, ***p<0.001, ****p<0.0001.

PPM1D-mutant cells have altered mitochondrial function.

(A) Mitochondrial mass of WT and PPM1D-mutant leukemia cells was determined using MitoTracker Green (100 nM) and analzyed by flow cytometry. Data represents mean + SD of triplicates. At least three independent experiments were conducted with similar findings; unpaired t-tests. (B) Immunoblot of WT and PPM1D-mutant cell lysates probed with the human OXPHOS antibody cocktail (1:1000) and vinculin (1:2000). (C) Measurement of mitochondrial oxygen consumption ratio (OCR) by seahorse assay in WT vs. PPM1D-mutant OCI-AML2 cells after treatment with oligomycin (1.5 uM), FCCP (0.5 uM), and rot/AA (0.5 uM). Quantification of basal, maximal, and ATP-linked respiration are shown. Data shown are the mean + SD of technical triplicates. (D) Mitochondrial membrane potential of WT and PPM1D-mutant OCI-AML2 cells was measured using MitoTracker CMXRos (400 nM) and analyzed by flow cytometry. Data represents mean + SD of triplicates, unpaired t-test, ns=non-significant (p>0.05), *p<0.05, **p<0.01.

PPM1D-mutant cells have a reduced oxidative stress response.

(A) RNA-seq GSEA analysis of PPM1D-mutant cells compared to WT Cas9-OCI-AML2 cells. Significantly up-and downregulated pathways are indicated by the blue and red bars, respectively. Normalized enrichment scores (NES) are shown with FDR <0.25. (B) RPPA profiling of WT and PPM1D-mutant OCI-AML2 cells. Proteins from the “Response to Oxidative Stress” pathway have been selected for the heatmap. Each column represents a technical replicate. See Figure 4-source data 2 for the raw data. (C) Total-and small molecule antioxidant capacity of WT and PPM1D-mutant cells performed in technical duplicates. Student t-test was used for statistical analysis, **p<0.01. (D) Immunoblot of WT and PPM1D-mutant OCI-AML2 cells in the untreated state (NT) or treated for 30 minutes with H2O2 (25 mM). Lysates were probed with an anti-oxidative stress defense cocktail (1:250) and vinculin (1:2000).

PPM1D mutations increase genomic instability and impair non-homologous end-joining.

(A) Left: Representative images of comet assays of mouse embryonic fibroblasts (MEFs). Two biological replicates were assessed for each genotype. Right: Quantification of n≥150 comets per experimental group with the Comet IV software; 2way ANOVA. (B) Left: Representative images of metaphase spreads of WT and Ppm1d-mutant mouse primary B-cells treated with low (0.5 uM) or high (5 uM) doses of cisplatin. Right: n≥50 metaphase cells were quantified in each experimental condition for chromosomal aberrations (white arrows). n=2 biological replicates used for each genotype. Student’s t-test was used for statistical analysis. (C–D) Left: Schematic of the homologous recombination (C) or non-homologous end-joining (D) U2OS DNA damage repair cassettes. Right: Quantification of GFP% analyzed by flow cytometry 48-hours after induction of DNA damage by I-SceI transduction; student’s t-test. (E) Left: Representative images of comet assays of OCI-AML2 cells at baseline and treated with ATN-224 (12.5 uM) for 24 hours. Right: Quantification of n>150 comets per experimental group using the Comet IV software; 2way ANOVA. (F) MFI of 8-oxo-dG lesions within OCI-AML2 cells 48-hours after ATN-224 (12.5 uM) treatment compared to control (NT) as measured by flow cytometry. Data are mean + SD (n=3), ns=non-significant (p>0.05), *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Genome-wide CRISPR screen identifies SOD1 as a synthetic lethal partner of PPM1D-mutant leukemia cells.

(A) Immunoblot validation of PPM1D-mutant Cas9-expressing OCI-AML2 cells generated and used for CRISPR screening. Blots were probed with anti-PPM1D (1:1000) and GAPDH (1:1000). Clones 2102 and 2113 were selected for the dropout screen. (B) Venn diagram of genes that were depleted from the two PPM1D-mutant clones (#2102, 2113) used in the dropout screen, but not depleted in the WT control lines. 37 genes were found to be depleted in both mutant clones. For a full list of genes, see Figure 1-source data 1. (C) Volcano plot of synthetic lethal hits ranked by fitness score with the Fanconi Anemia pathway genes highlighted in blue. (D) Cas9-OCI-AML2 and Cas9-OCI-AML3 WT or PPM1D-mutant cells were transduced with the empty vector control backbone tagged with a blue fluorescent protein (BFP) reporter. Cells were assayed by flow cytometry between 3- and 24-days post-transduction and normalized to the BFP percentage at day 3. Data shown are mean + SD (n=2 per condition).

PPM1D-mutant cells have increased oxidative stress.

(A) Left: Representative flow cytometry plots of WT and PPM1D-mutant cells treated with ATN-224 (25 uM for 24 hours) and stained for Annexin V-APC and PI for apoptosis; multiple unpaired t-tests. (B–C) Dose response curves for cell viability with SOD1-inhibitor (ATN-224) (B) or ATN-224 in combination with 0.25 uM NAC (C) in WT and PPM1D-mutant leukemia cell lines after 24-hours. Mean + SD (n=3) is shown along with a non-linear regression curve. All values are normalized to the baseline cell viability with vehicle, as measured by MTT assay. (D) Total reactive oxygen species (ROS)of WT and Ppm1d-mutant MEFs measured by DCFDA (10 uM) staining. MFI was determined by flow cytometry. n=6 biological replicates were used for each genotype. Data shown are the mean of each biological replicate; unpaired t-test. (E) Total ROS of WT GM12878) and PPM1D-mutant patient lymphoblastic cell lines (LCLs) at baseline, and after 24-hrs of SOD1 inhibition measured by DCFDA (10 uM) staining. MFI was determined by flow cytometry; multiple unpaired t-tests, (F) Dose response curve of WT and PPM1D-mutant LCLs after ATN-224 treatment. IC50s of WT and PPM1D-mutant LCLs were 48.8 uM and 20.51 uM, respectively as measured by MTT assay; non-linear regression analysis, ns=non-significant (p>0.05), **p<0.01, ***p<0.001, ****p<0.0001.

PPM1D-mutant cells have altered mitochondrial function.

(A,B) Measurement of mitochondrial oxygen consumption ratio (OCR) by seahorse assay in WT vs. PPM1D-mutant MOLM-13 (A) and OCI-AML3 (B) cells after treatment with oligomycin (1.5 uM), FCCP (0.5 uM), and rot/AA (0.5 uM). Quantification of basal, maximal, and ATP-linked respiration shown. Each cell line was performed in technical triplicates, student’s t-test. (C) Growth curves of WT and PPM1D-mutant leukemia cell lines at 24-, 48-, and 72-hours. Cell counts were normalized to day 0. ns=non-significant (p>0.05), *p<0.05, ***p<0.001.

PPM1D-mutant cells have reduced oxidative stress response.

(A) Schematic of the experimental setup for the bulk RNA-sequencing and reverse-phase protein array. WT and PPM1D-mutant Cas9 OCI-AML2 cells were transduced with either empty vector (EV)-BFP or SOD1-sgRNA-BFP. Cells were passaged for ten days and then sorted for BFP expression for downstream analysis. (B, D) GSEA enrichment plots for PPM1D-mutant cells compared to WT after transduction with EV (B) or after SOD1-knockout (D) for the “Regulation of Response to Oxidative Stress” (GO:1902882) and “Response to Oxidative Stress” (GO:0006979). NES are shown with FDR<0.25. (C) GSEA analysis of RNA-sequencing of SOD1-deleted cells compared to EV control in WT and PPM1D-mutant cells. Blue and red bars indicate significantly up- and downregulated pathways, respectively. Normalized enrichment scores (NES) are indicated. All pathways filtered for FDR<0.25.

PPM1D-mutant cells have reduced oxidative stress response.

(A) Volcano plot of the differentially expressed proteins from the RPPA in PPM1D-mutant OCI-AML2 cells compared to WT. Red and blue dots indicate significantly up- or downregulated proteins, respectively, with a cutoff FDR<0.2 and linear fold change >|1.2|. (B) RPPA profiling of WT and PPM1D-mutant cells after SOD1 deletion. Proteins from the “Response to Oxidative Stress” pathway have been selected for the heatmap. Each column represents a technical replicate. See Figure 4-source data 2 for the raw data.

PPM1D-mutations increase genomic instability and impairs non-homologous end-joining repair.

(A) Left: Sanger sequencing traces of the parental U2OS cell line harboring a c.1372 C>T mutation in PPM1D and the CRISPR-edited U2OS cell line with mutation corrected to WT PPM1D. Right: Immunoblot validation of these clones are shown. Lysates were probe with anti-PPM1D (1:1000) and anti-GAPDH (1:1000). (B,C) Left: Representative images of Rad51 and 53BP1 immunofluorescence microscopy. Mouse embryonic fibroblasts were treated with 10 Gy irradiation, harvested 1-hour post-irradiation and stained for the indicated markers. Right: Quantification of the number of foci per cell is shown. Analysis was performed using CellProfiler. n>100 cells for each condition; students t-test. (D) Comet assay quantification of mouse embryonic fibroblasts at baseline and after 1-hour post-irradiation (10 Gy). Quantification and analysis of tail moments were performed using the Comet IV software. n≥150 comets were scored per experimental group; 2way ANOVA, ns=non-significant (p>0.05), **p<0.01, ***p<0.001, ****p<0.0001.