Stereospecific PARP trapping by BMN 673 and comparison with olaparib and rucaparib

Abstract

Anti-PARP drugs were initially developed as catalytic inhibitors to block the repair of DNA single-strand breaks. We recently reported that several PARP inhibitors have an additional cytotoxic mechanism by trapping PARP-DNA complexes, and that both olaparib and niraparib act as PARP poisons at pharmacologic concentrations. Therefore, we have proposed that PARP inhibitors should be evaluated based both on catalytic PARP inhibition and PARP-DNA trapping. Here, we evaluated the novel PARP inhibitor, BMN 673, and compared its effects on PARP1 and PARP2 with two other clinical PARP inhibitors, olaparib and rucaparib, using biochemical and cellular assays in genetically modified chicken DT40 and human cancer cell lines. Although BMN 673, olaparib, and rucaparib are comparable at inhibiting PARP catalytic activity, BMN 673 is ∼100-fold more potent at trapping PARP-DNA complexes and more cytotoxic as single agent than olaparib, whereas olaparib and rucaparib show similar potencies in trapping PARP-DNA complexes. The high level of resistance of PARP1/2 knockout cells to BMN 673 demonstrates the selectivity of BMN 673 for PARP1/2. Moreover, we show that BMN 673 acts by stereospecific binding to PARP1 as its enantiomer, LT674, is several orders of magnitude less efficient. BMN 673 is also approximately 100-fold more cytotoxic than olaparib and rucaparib in combination with the DNA alkylating agents methyl methane sulfonate (MMS) and temozolomide. Our study demonstrates that BMN 673 is the most potent clinical PARP inhibitor tested to date with the highest efficiency at trapping PARP-DNA complexes.

Figure 1. Comparative PARP catalytic inhibition of BMN 673

(A) Chemical structures of olaparib (AZD2281), rucaparib (AG-014699), BMN 673 and NAD. The nicotinamide moiety is outlined in dotted lines. Arrows in the BMN 673 structure indicate chiral centers involved in drug activity (see Figure S1) (B) Catalytic PARP inhibition potency of BMN 673 in comparison with olaparib and rucaparib. Total poly(ADP-ribosyl)ation (PAR) levels in wild type DT40 cells were examined by Western blotting against PAR 30 min after the indicated drug treatments. The asterisk represents a nonspecific band. (C) PAR levels in drug-treated wild type DT40 and DU145 cells measured by ELISA. Cells were incubated with the indicated concentrations of PARP inhibitors for 2 hours. PAR levels without drug treatment were set as 100% in each cell line.

Figure 2. BMN 673 is markedly more cytotoxic than olaparib and rucaparib while requiring PARP-1/2 for activity

For all experiments, viability curves were derived after continuous treatment for 72 hours with the indicated PARP inhibitors in the indicated cell lines. Cellular ATP concentration was used to measure cell viability. The survival of untreated cells was set as 100%. Error bars represent standard deviation (SD) (n ≥ 3). (A) Survival curves of wild type, PARP1−/−, and BRCA2tr/− DT40 cells. Drug IC₉₀’s are tabulated at the bottom. (B) Survival curves of DU145 (human prostate cancer) and EW8 (Ewing’s sarcoma) cells. Drug IC₉₀’s are tabulated at the bottom. (C) Survival curves of PARP1−/− DT40 cell line to high concentrations of the PARP inhibitors. (D) Survival curves of MDB-MB231 (human breast cancer) cells.

Figure 3. Comparison of the sensitivity patterns in the three PARP inhibitors across the NCI60 cell lines

The IC₅₀ (drug concentration that reduces cell survival down to 50%) obtained from the NCI60 databases (35, 41) (http://discover.nci.nih.gov/cellminer) is plotted for each cell line. Cell lines are colored according to tissue of origin (41). NA: data not available.

Figure 4. Comparative trapping of PARP1- and PARP2-DNA complexes by BMN 673, olaparib and rucaparib

PARP-DNA complexes were determined by Western blotting analyses of chromatin bound fractions from drug-treated DT40 cells (A) and DU145 cells (B). DT40 and DU145 treatments were for 30 min and 4 hours, respectively. Blots were probed with the indicated antibodies. Histone H3 was used as positive markers for chromatin-bound fractions and as loading control. The blots of lanes 1–4 are identical to the blots of lanes 8–11 for PARP2 in (B). The blots are representative of multiple experiments.

Figure 5. Biochemical trapping of PARP1 by BMN 673

(A) Scheme of the fluorescence anisotropy (FA) binding assay. The star indicates the site labeled on the DNA substrate with Alexa Fluor488. Unbound nicked DNA substrate rotates fast and gives low FA. PARP1 binding to the substrate slows the rotation and gives high FA. Addition of NAD⁺ leads to PARP1 dissociation from DNA due to autoPARylation. (B) Concentration-dependent PARP1-DNA association in the presence of BMN 673 or olaparib. FA was measured 40 min after adding NAD⁺. (C) Time-course of PARP1-DNA dissociation in the presence of BMN 673 and olaparib (0.12 μM each). Addition of NAD⁺ in the absence of PARP inhibitor immediately reduces PARP1-DNA complexes (DMSO control). In the absence of NAD⁺, PARP1-DNA complexes remain stable for at least 120 min (no NAD⁺).

Figure 6. BMN 673 enhances the cytotoxicity of alkylating agents more efficiently than olaparib and rucaparib

(A) Survival curves of wild type DT40 cells treated with MMS alone (upper curves labeled “0”) or with the indicated concentrations of PARP inhibitors (at the concentration shown beside each curve in micromolar units). The survival of untreated cells was set as 100%. Data are mean ± SD (n ≥ 3). (B) PARP1−/−cells are hypersensitive to MMS (compare with upper curves in panel A) and resistant to the PARP inhibitors. (C) Survival curves of the indicated human cancer cells treated with MMS in combination with the indicated PARP inhibitors (the concentration of each PARP inhibitor is shown beside each curve). The survival of untreated cells was set as 100%. Data are mean ± SD (n ≥ 3). (D) Same as C but using temozolomide instead of MMS in prostate cancer DU145 cells.