Ubiquitinated PCNA Drives USP1 Synthetic Lethality in Cancer

Abstract

CRISPR Cas9-based screening is a powerful approach for identifying and characterizing novel drug targets. Here, we elucidate the synthetic lethal mechanism of deubiquitinating enzyme USP1 in cancers with underlying DNA damage vulnerabilities, specifically BRCA1/2 mutant tumors and a subset of BRCA1/2 wild-type (WT) tumors. In sensitive cells, pharmacologic inhibition of USP1 leads to decreased DNA synthesis concomitant with S-phase-specific DNA damage. Genome-wide CRISPR-Cas9 screens identify RAD18 and UBE2K, which promote PCNA mono- and polyubiquitination respectively, as mediators of USP1 dependency. The accumulation of mono- and polyubiquitinated PCNA following USP1 inhibition is associated with reduced PCNA protein levels. Ectopic expression of WT or ubiquitin-dead K164R PCNA reverses USP1 inhibitor sensitivity. Our results show, for the first time, that USP1 dependency hinges on the aberrant processing of mono- and polyubiquitinated PCNA. Moreover, this mechanism of USP1 dependency extends beyond BRCA1/2 mutant tumors to selected BRCA1/2 WT cancer cell lines enriched in ovarian and lung lineages. We further show PARP and USP1 inhibition are strongly synergistic in BRCA1/2 mutant tumors. We postulate USP1 dependency unveils a previously uncharacterized vulnerability linked to posttranslational modifications of PCNA. Taken together, USP1 inhibition may represent a novel therapeutic strategy for BRCA1/2 mutant tumors and a subset of BRCA1/2 WT tumors.

Figure 1.

USP1 is a synthetic lethal target in BRCA1/2 mutant tumors amenable to small molecule inhibition. A, Volcano plot showing MAGeCK analysis comparing CRISPR-Cas9 screens performed in 11 BRCA1/2 WT versus two BRCA1 mutant breast cancer cell lines. Genes on the left are preferentially depleted in the BRCA1 mutant compared with WT cells, whereas genes on the right are preferentially depleted in BRCA1/2 WT compared to mutant cells. B,USP1 Chronos from CRISPR screen datasets in projects Achilles and Score (DepMap Public, Q122). C, Chemical structures of literature USP1-UAF1 inhibitors ML323 (top) and I-138 (bottom). D, Activity of I-138 (IC₅₀ 4.1 nmol/L) and ML323 (IC₅₀ 569 nmol/L) in an USP1-UAF1 ubiquitin-Rhodamine 110 cleavage assay. E, Top panel shows mutual exclusive test of I-138 and ubiquitin (top). The result was fitted to modified Yontenani and Theorell equation to obtain K_I-138 (5.4 nmol/L), K_Ubiquitin (4.5 μmol/L), and α (0.16). Bottom panel shows mutual exclusive test of I-138 and ML323. The result was fitted to modified Yontenani and Theorell equation to obtain K_I-138 (3.2 nmol/L), K_Ubiquitin (341 nmol/L), and α (12.1). F, I-138 profiling in a biochemical assay panel of 45 deubiquitinating enzymes at the fixed concentration of 10 μmol/L (Ubiquigent DUBprofiler). G, Immunoblotting analysis of HAP-1 USP1 WT and knockout cells treated with DMSO mock or 500 nmol/L I-138 for 4 hours. H, Immunblotting analysis of MDA-MB-436 whole cell lysates treated in indicated concentrations of I-138 for 24 hours. DMSO concentration was normalized across all conditions. I, Quantitation of ub-PCNA intensity in MDA-MB-436 cells treated with I-138 at indicated concentrations and time points and stained with immunofluorescence. Statistical significance was evaluated by ANOVA followed by Dunett post hoc test comparing treatment conditions against DMSO control (****, P < 0.0001).

Figure 2.

USP1 inhibition causes loss of viability and replication stress in BRCA1/2 mutant cells. A, Dose–response curves of MDA-MB-436 and HCC1954 cells treated with I-138 (left) or niraparib (right). Cells were seeded in 96-well plates and incubated in drug under normalized DMSO conditions for 10 days. Cell viability was estimated using CellTiter-Glo. B, Clonogenic assays performed using isogenic UWB1.289 cells with and without BRCA1 re-expression. Cells were treated with indicated compounds for 14 days with media change every 3 to 4 days and visualized using crystal violet. C, Quantitation of EdU intensity in MDA-MB-436 (left) and UWB1.289 (right) cells pulsed-labelled with EdU after mock or compound treatment for indicated durations. Statistical significance was evaluated by ANOVA followed by Dunett post hoc test, comparing treatment conditions against DMSO control (****, P < 0.0001). D, Flow cytometry-based cell-cycle analysis of MDA-MB-436 cells treated with USP1 inhibitor and stained with propidium iodide. E, Immunoblotting analysis of MDA-MB-436 whole cell lysates, collected after treatment with 500 nmol/L I-138 at indicated time points under normalized DMSO conditions. F, Replication fork speed of MDA-MB-436 cells treated with DMSO, hydroxyurea, or I-138, measured with DNA fiber assay following the labeling schematic as shown. Statistical significance was evaluated by ANOVA followed by Dunett post hoc test, comparing treatment conditions against DMSO control (****, P < 0.0001; n.s., not significant). G, Quantitation of γ2HAX intensity in MDA-MB-436 cells stained with immunofluorescence treated with I-138 across multiple time points. Statistical significance was evaluated by ANOVA followed by Dunett post hoc test, comparing treatment conditions against DMSO control (****, P < 0.0001). H, Quantitation of micronuclei formation in MDA-MB-436 cells based on fluorescence imaging, where the percentage of cells with micronuclei were calculated from cells imaged in an entire well of a 96-well plate. Results are shown as mean ± SD, N = 3. Statistical significance was evaluated by ANOVA followed by Dunett post hoc test, comparing treatment conditions against DMSO control (****, P < 0.0001). I, Cell death in MDA-MB-436 treated with DMSO mock or USP1 inhibitor for 168 hours, measured using a flow cytometry–based Annexin V assay.

Figure 3.

PCNA ubiquitination and protein loss mediate USP1 dependency in BRCA1/2 mutant cells. A, CRISPR-Cas9 UMI screen workflow. MDA-MB-436 cells stably expressing Cas9 were transduced with a genome-wide CRISPR-Cas9 UMI library and cultured for 21 days in the presence of or absence of USP1 inhibitor. B, Number of statistically significant hits from MAGeCK and UMIBB analyses (P < 0.001) comparing DMSO control and 50 nmol/L treatment groups. C, Volcano plot of UMIBB analysis comparing CRISPR-Cas9 UMI screens performed under DMSO control versus 50 nmol/L I-138 conditions in MDA-MB-436 cells. D, Isogenic RAD18 knockout or UBE2K knockout cell lines were constructed by transducing targeting sgRNAs into MDA-MB-436 cells stably expressing Cas9. Clonogenic assays and (E) immunoblotting analyses were performed in these cell lines in the presence or absence of USP1 inhibition. F, Quantitation of total PCNA protein levels using tandem mass tag MS (Ubiquigent PROTEOMEprofiler) in WT control, RAD18 knockout, or UBE2K knockout MDA-MB-436 cells following 200 nmol/L I-138 or DMSO mock treatment. Values are represented as log2-fold change from DMSO. G, Seven-day CellTiter-Glo assay showing I-138 cell viability dose response (left) and immunoblotting analysis (right) of MDA-MB-436 cells stably expressing empty vector, WT, or K164R ub-dead PCNA. Data are presented as mean ± SD, N = 3.

Figure 4.

USP1 and PARP inhibition are synergistic in BRCA1/2 mutant tumors. A, Cell lines were profiled in a 9-point I-138 + olaparib combination dose–response matrix in 384-well plates using a 7-day CellTiter-Glo assay. Synergy score was calculated based on the Bliss model using Combenefit v2.021. B, Cell viability dose response of indicated PARP inhibitors in SUM149PT, with and without USP1 inhibitor cotreatment. Cell viability was estimated using 7-day CellTiter-Glo assay, and data are represented as mean ± SD. C, Olaparib dose-response in UWB1.289 cells with and without BRCA1 re-expression, in the presence or absence of 120 nmol/L I-138 co-treatment. Cell viability was estimated using 7-day CellTiter-Glo assay, and data are represented as mean ± SD. D, Quantitation of γH2AX intensity in UWB1.289 cells treated with DMSO control or indicated compounds for 72 hours and stained using immunofluorescence. Data are represented as mean ± SD of the total intensity in entire well of a 96-well plate, N = 3. Statistical significance was evaluated by ANOVA followed by Dunett post hoc test, comparing treatment conditions against DMSO control (****, P < 0.0001; *, P < 0.5). E, Quantitation of single-strand breaks as measured by number of sSTRIDE (intoDNA) foci per cell in UWB1.289 cells treated with indicated compounds for 8 hours. Statistical significance was evaluated by ANOVA followed by Dunett post hoc test, comparing treatment conditions against DMSO control (****, P < 0.0001). F, Tumor volume of MDA-MB-436 xenograft tumors inoculated in NOD SCID mice, treated with vehicle, niraparib (15 mg/kg once daily), I-138 (50 mg/kg once daily), or niraparib + I-138 by oral administration. Drug administration for all groups was stopped on day 41. Data are represented as mean ± SEM, N = 8 animals per group.

Figure 5.

Ubiquitinated PCNA mediates USP1 dependency in selected BRCA1/2 WT cells. A,USP1 dependency estimated using USP1 Chronos across multiple cancer lineages, extracted from DepMap Public (Q1 2022) CRISPR screening datasets. B, Clonogenic assays in cell lines treated with DMSO, olaparib, or I-138 for 10 to 17 days with media change every 3 to 4 days, followed by visualization using crystal violet. C, RAD51 foci formation in cell lines treated with DMSO or 100 nmol/L gemcitabine. Data are represented as mean ± SD of averages across three replicate wells. Statistical significance was evaluated using two-tailed Student t test (*, P < 0.05; **, P < 0.01; n.s., not significant). D, Volcano plot showing results from UMIBB analysis comparing genome-wide CRISPR-Cas9 UMI screens performed NCI-H1693 cells treated with DMSO or 50 nmol/L I-138. Targeting of genes on the right conferred survival advantage under drug treatment. E, I-138 dose–response curves (top) and immunoblotting analysis (bottom) of NCI-H1693 cells stably expressing empty vector, WT PCNA, or K164R ubiquitin-dead PCNA. Data are represented as mean ± SD.

Figure 6.

Mechanism of USP1 dependency in BRCA1/2 mutant and selected BRCA1/2 WT tumors.