PARG inhibitor sensitivity correlates with accumulation of single-stranded DNA gaps in preclinical models of ovarian cancer

Abstract

Poly (ADP-ribose) glycohydrolase (PARG) is a dePARylating enzyme which promotes DNA repair by removal of poly (ADP-ribose) (PAR) from PARylated proteins. Loss or inhibition of PARG results in replication stress and sensitizes cancer cells to DNA-damaging agents. PARG inhibitors are now undergoing clinical development for patients having tumors with homologous recombination deficiency (HRD), such as cancer patients with germline or somatic BRCA1/2-mutations. PARP inhibitors kill BRCA-deficient cancer cells by increasing single-stranded DNA gaps (ssGAPs) during replication. Here, we report that, like PARP inhibitor (PARPi), PARG inhibitor (PARGi) treatment also causes an accumulation of ssGAPs in sensitive cells. PARGi exposure increased accumulation of S-phase-specific PAR, a marker for Okazaki fragment processing (OFP) defects on lagging strands and induced ssGAPs, in sensitive cells but not in resistant cells. PARGi also caused accumulation of PAR at the replication forks and at the ssDNA sites in sensitive cells. Additionally, PARGi exhibited monotherapy activity in specific HR-deficient, as well as HR-proficient, patient-derived, or patient-derived xenograft (PDX)-derived organoids of ovarian cancer, and drug sensitivity directly correlated with the accumulation of ssGAPs. Taken together, PARGi treatment results in toxic accumulation of PAR at replication forks resulting in ssGAPs due to OFP defects during replication. Regardless of the BRCA/HRD-status, the induction of ssGAPs in preclinical models of ovarian cancer cells correlates with PARGi sensitivity. Patient-derived organoids (PDOs) may be a useful model system for testing PARGi sensitivity and functional biomarkers.

Keywords: PARG inhibitor; PARylation; functional biomarkers; organoid models; single-stranded DNA (ssDNA) gaps.

Fig. 1.

PARG inhibition leads to toxic S phase PAR accumulation in sensitive cells. (A) Schematic of the method for generating ovarian cancer cells with acquired resistance to PARGi (PDD00017273). PARGi-sensitive cell line, RMUGS was exposed to increasing dose of PARGi over 4 mo and resistant cells were derived (RMUGS-R). (B) Clonogenic survival of RMUGS and RMUGS-R cells in the presence of increasing concentration of PARGi. Left: Graphical quantitation of the colonies from cells treated with PARGi compared to control cells treated with DMSO (IC50 RMUGS—0.07 μM and RMUGS-R—0.61 μM). Right: Representative images of the colonies. (C) Left: Representative images of EdU and PAR staining showing S phase PARylation after DMSO or PARGi (10 μM) treatment for 2 h in RMUGS and RMUGS-R cells. Scale 10 µm. Right: Quantification of S phase PARylation in RMUGS and RMUGS-R cells. Whiskers represent the minimum and maximum (****P < 0.0001; Mann–Whitney test). (D) Left: Representative images of EdU and PAR staining showing S phase PARylation after DMSO or PARGi (10 μM) treatment for 2 h in RPE TP53^−/− and RPE TP53^−/− BRCA1^−/− cells. Scale 10 µm. Right: Quantification of S phase PARylation in RPE TP53^−/− and RPE TP53^−/− BRCA1^−/− cells. Whiskers represent the minimum and maximum (****P < 0.0001; **P = 0.0044; Mann–Whitney test). (E) Western blot showing replication stress and DNA damage markers after PARGi treatment (10 μM) in RMUGS and RMUGS-R cells.

Fig. 2.

PARG inhibition or PARG loss results in accumulation of ssGAPs in drug-sensitive but not drug-resistant cells. (A) Top: Schematic of DNA fiber assay with CldU/IdU pulse-labeling protocol and drug treatment, followed by S1 nuclease treatment. Bottom: Representative DNA fiber images in RMUGS and RMUGS-R cells after PARG inhibition with or without S1 nuclease treatment. (B) Quantification of IdU tracts from RMUGS and RMUGS-R cells after PARGi (10 μM) treatment with or without S1 nuclease. Each dot represents one fiber. At least 200 fibers were analyzed per condition. Median values are represented by horizontal lines (n.s., not significant; ****P < 0.0001; Mann–Whitney test). (C) Top: Schematic of DNA fiber assay with CldU/IdU pulse-labeling protocol and drug treatment, followed by S1 nuclease treatment. Bottom: Representative DNA fiber images in RPE TP53^−/− and RPE TP53^−/− BRCA1^−/− cells after PARG inhibition with or without S1 nuclease treatment. (D) Quantification of IdU tracts from RPE TP53^−/− and RPE TP53^−/− BRCA1^−/− cells after PARGi (10 μM) treatment. Each dot represents one fiber. At least 200 fibers were analyzed per condition. Median values are represented by horizontal lines (n.s., not significant; ****P < 0.0001; Mann–Whitney test). (E) Top: Schematic of DNA fiber assay with CldU/IdU pulse-labeling protocol followed by S1 nuclease treatment. Bottom: Representative DNA fiber images in RPE TP53^−/− and RPE TP53^−/− PARG^−/− cells after BRCA1 knockdown with or without S1 nuclease treatment. (F) Quantification of IdU tracts in RPE TP53^−/− and RPE TP53^−/− PARG^−/− cells after BRCA1 knockdown with or without S1 nuclease. Each dot represents one fiber. At least 200 fibers were analyzed per condition. Median values are represented by horizontal lines (n.s., not significant; ****P < 0.0001; Mann–Whitney test). (G) Top: Schematic of DNA fiber assay with CldU/IdU pulse-labeling protocol followed by S1 nuclease treatment. Bottom: Representative DNA fiber images in RPE TP53^−/− PARG^−/− cells after transfection with siBRCA1 and cDNA encoding Empty vector (EV), PARG wildtype (PARG-WT), or catalytically dead PARG mutant (PARG-MUT) with or without S1 nuclease treatment. (H) Quantification of IdU tracts from RPE TP53^−/− PARG^−/− cells after transfection with siBRCA1 and EV, PARG-WT, or PARG-MUT with or without S1 nuclease. Each dot represents one fiber. At least 200 fibers were analyzed per condition. Median values are represented by horizontal lines (n.s., not significant; ****P < 0.0001; Mann–Whitney test).

Fig. 3.

PARG inhibition leads to accumulation of PAR at the ssDNA sites of replication forks in sensitive cells. (A) Left: Representative image of SIRF assay performed to detect PAR-EdU proximity ligation assay (PLA) foci at the replication forks after PARG inhibition in RMUGS and RMUGS-R cells. Scale 10 µm. Right: Quantification of PAR-EdU PLA foci at the replication forks after PARG inhibition (10 μM) for 2 h in RMUGS and RMUGS-R cells. PLA performed with either anti-Biotin (EdU) or anti-PAR antibody alone as negative controls are also shown. Mean and SEM values are represented (****P < 0.0001; ***P < 0.001; *P = 0.0189; Mann–Whitney test). (B) Left: Representative image of PAR-pRPA (pS33 RPA and pS4/S8 RPA) PLA foci after PARG inhibition in RMUGS cells. Scale 10 µm. Right: Quantification of PAR-pRPA PLA foci after PARG inhibition (10 μM) for 2 h in RMUGS cells. PLA performed with either anti-pRPA or anti-PAR antibody alone as negative controls are also shown. Mean and SEM values are represented (****P < 0.0001; Mann–Whitney test). (C) Volcano plot from aniPOND assay coupled with mass spectrometry (MS) showing enrichment of proteins at the replication forks after PARGi exposure in RMUGS cells. Samples treated with PARGi (10 μM) for 4 h were compared to the samples treated with DMSO. (D) Western blot from aniPOND assay showing enrichment of PAR, XRCC1, and MRE11 at replication forks after treatment with PARGi. (E) Left: Representative image of XRCC1-EdU PLA foci at the replication fork after PARG inhibition in RMUGS and RMUGS-R cells. Scale 10 µm. Right: Quantification of XRCC1-EdU PLA foci at the replication fork after PARG inhibition (10 μM) for 2 h in RMUGS and RMUGS-R cells. PLA performed with either anti-Biotin (EdU) or anti-XRCC1 antibody alone as negative controls are also shown. Mean and SEM values are represented (****P < 0.0001; Mann–Whitney test).

Fig. 4.

Increased accumulation of PAR and ssDNA occur at unligated Okazaki fragments after PARG inhibition. (A) Left: Representative images of PCNA and PAR staining showing S phase PARylation in the presence of indicated drugs in RMUGS cells. Scale 10 µm. Right: Quantification of S phase PARylation in the presence of indicated drugs in RMUGS cells. Median values are represented (****P < 0.0001; Mann–Whitney test). (B) Left: Representative image of PAR-EdU PLA foci at the replication fork in the presence of indicated drugs in RMUGS cells. Scale 10 µm. Right: Quantification of PAR-EdU PLA foci at the replication fork in the presence of indicated drugs in RMUGS cells. Mean and SEM values are represented (****P < 0.0001; ***P < 0.001; **P < 0.05; Mann–Whitney test). (C) Left: Representative images of PAR-pRPA PLA foci in the presence of indicated drugs in RMUGS cells. Scale 10 µm. Right: Quantification of PAR-pRPA PLA foci in the presence of indicated drugs in RMUGS cells. Mean and SEM values are represented (****P < 0.0001; *P = 0.018; Mann–Whitney test). (D) Model depicting increased unligated Okazaki fragments and ssGAP accumulation after PARGi treatment in PARGi-sensitive cells.

Fig. 5.

PARGi sensitivity correlates with ssGAP accumulation in ovarian cancer patient-derived organoids (PDOs). (A) Representative images of the morphology of ovarian cancer PDOs. (B) Representative images of H&E and immuno-histochemistry staining of PDOs. (C) Survival plots of PDOs exposed to increasing concentration of PARGi. (D) Schematic of DNA fiber assay with CldU/IdU pulse-labeling protocol and drug treatment, followed by S1 nuclease treatment. (E and F) Top: Representative DNA fiber images in PDOs after PARG inhibition with or without S1 nuclease treatment. Bottom: Quantification of IdU tracts from PDOs after PARGi (100 μM) treatment with or without S1 nuclease. Each dot represents one fiber. At least 90 fibers were analyzed per condition. Median values are represented by horizontal lines (n.s., not significant; ***P < 0.001; Mann–Whitney test).