Enhanced Killing of Cancer Cells by Poly(ADP-ribose) Polymerase Inhibitors and Topoisomerase I Inhibitors Reflects Poisoning of Both Enzymes*

Background: PARP inhibitors and topoisomerase I poisons (Top1p) synergize by an unknown mechanism. Results: Although Parp1 deletion fails to increase Top1p sensitivity, transfection with catalytically inactive PARP1 or its isolated DNA binding domain does sensitize. Conclusion: PARP inhibitors poison PARP1 to diminish repair of topoisomerase I-triggered DNA damage. Significance: These results predict that tumors with elevated PARP1 will be particularly sensitive to Top1p/PARP inhibitor combinations. Poly(ADP-ribose) polymerase-1 (PARP1) plays critical roles in the regulation of DNA repair. Accordingly, small molecule inhibitors of PARP are being developed as agents that could modulate the activity of genotoxic chemotherapy, such as topoisomerase I poisons. In this study we evaluated the ability of the PARP inhibitor veliparib to enhance the cytotoxicity of the topoisomerase I poisons topotecan and camptothecin (CPT). Veliparib increased the cell cycle and cytotoxic effects of topotecan in multiple cell line models. Importantly, this sensitization occurred at veliparib concentrations far below those required to substantially inhibit poly(ADP-ribose) polymer synthesis and at least an order of magnitude lower than those involved in selective killing of homologous recombination-deficient cells. Further studies demonstrated that veliparib enhanced the effects of CPT in wild-type mouse embryonic fibroblasts (MEFs) but not Parp1−/− MEFs, confirming that PARP1 is the critical target for this sensitization. Importantly, parental and Parp1−/− MEFs had indistinguishable CPT sensitivities, ruling out models in which PARP1 catalytic activity plays a role in protecting cells from topoisomerase I poisons. To the contrary, cells were sensitized to CPT in a veliparib-independent manner upon transfection with PARP1 E988K, which lacks catalytic activity, or the isolated PARP1 DNA binding domain. These results are consistent with a model in which small molecule inhibitors convert PARP1 into a protein that potentiates the effects of topoisomerase I poisons by binding to damaged DNA and preventing its normal repair.

A number of previous studies have demonstrated that small molecule PARP inhibitors selectively sensitize cells to topo I poisons in vitro and in vivo (23)(24)(25)(26)(27)(28)(29)(30)(31)(32). At least three explanations have been advanced to explain these observations. First, studies with purified enzymes have shown that PARP1 can covalently attach pADPr to topo I. The presence of this pADPr polymer alters the affinity of topo I for DNA, shifting the cleavage/religation equilibrium of the enzyme toward sealed DNA (33)(34)(35)(36)(37).
Second, a series of studies suggests a requirement for PARP1 to help resolve stalled replication forks (38,39), which are produced upon treatment with topo I poisons (16,17,40,41). Whether PARP1 acts by modulating WRN helicase (42,43) or recruiting MRE11 (39,44) or both is unclear. Nonetheless, PARP1 deletion has been reported to inhibit the restarting of stalled replication forks (45), providing an alternative explanation for the observed synergy between topo I poisons and PARP inhibitors.
Finally, a series of studies have identified tyrosyl-DNA phosphodiesterase 1 (TDP1) as an enzyme capable of cleaving the phosphotyrosine linkage between topo I and the DNA backbone (46,47). TDP1 interacts with several components of the base excision repair pathway, including XRCC1, polynucleotide kinase phosphatase, and DNA ligase III (48,49). Other studies have shown that cells lacking functional base excision repair components such as XRCC1 are also hypersensitive to topo I poisons (30,50,51). Moreover, XRCC1 and DNA ligase III are typically recruited to sites of DNA damage by PARP1 and pADPr (52,53). These studies have led to proposed models in which PARP1 contributes to repair of topo I-mediated damage by recruiting a multiprotein complex consisting of TDP1, XRCC1, DNA ligase III, and polynucleotide kinase phosphatase to sites of trapped Top1cc or the subsequent non-proteinlinked strand breaks (9,46,48).
In each of the preceding models, cells lacking PARP1 would be expected to be hypersensitive to topo I poisons compared with parental cells. Here we show that PARP inhibitors sensitize cells to topo I poisons at concentrations that result in very little inhibition of PARP catalytic activity. Moreover, we report that Parp1 Ϫ/Ϫ cells, which are resistant to the sensitizing effects of PARP inhibitors, exhibit sensitivity to topo I poisons that is indistinguishable from parental cells. Finally, we show that the effects of PARP inhibitors on CPT sensitivity can be recapitulated by transfection with catalytically inactive PARP1 or the PARP1 DNA binding domain (DBD). Collectively, these observations suggest that PARP inhibition converts PARP1 into a dominant-negative molecule that appears to poison the ability of DNA repair machinery to participate in either PARP1dependent or PARP1-independent repair of CPT-induced DNA damage.
Transfection of PARP1 expression constructs (20 g DNA) into Parp1 Ϫ/Ϫ MEFs was performed by electroporation using a 280-V, 10-ms pulse delivered by a square-wave electroporator. Forty-eight hours after transfection, cells were utilized for clonogenic survival assays or harvested for immunoblotting.
Clonogenic Survival Assays-Colony formation assays were performed on A549 cells by trypsinizing cells and plating at 300 cells/plate in triplicate 35-mm dishes containing 2 ml of medium A. Cells were allowed to adhere for 4 -14 h, then treated as indicated. After treatment, plates were washed with drug-free RPMI 1640 and cultured in medium A for 5-8 days. After staining with Coomassie Brilliant Blue, colonies containing Ն50 cells were scored manually. A similar protocol was used for other adherent cell lines with the following changes; Ovcar5 cells were plated at 500 cells/plate in medium B, A2780 cells were plated at 1000 cells/plate in medium C, SKOV3 cells were plated at 500 cells/plate in 60-mm dishes in 3 ml medium D, MEFs and 3T3 cells were plated at 2000 cells/plate in 60-mm dishes in 3 ml medium E, and V79 and V-C8 Chinese hamster cells were plated at 300 cells/plate in 60-mm dishes in 3 ml of medium F. HL-60 cells, which are nonadherent, were treated with the indicated concentrations of topotecan with or without veliparib for 24 h, sedimented at 100 ϫ g for 5 min, washed in drug-free medium, and plated in 0.3% (w/v) agar in the medium of Pike and Robinson (58). After 10 days, colonies containing Ն50 cells were counted on an inverted microscope.
Flow Cytometry-Propidium iodide staining was performed as described previously (59). Logarithmically growing cells were incubated with drugs as indicated in the figures, washed with drug-free RPMI 1640, trypsinized, and pelleted by centrifugation at 100 ϫ g for 5 min. After a wash with ice-cold PBS, cells were fixed at 4°C in 50% (v/v) ethanol, digested with RNase A, stained with propidium iodide, and subjected to flow microfluorimetry. Results were analyzed using ModFit software (Verity Software; Topsham, ME).
The induction of apoptosis was analyzed in HL-60 cells, which (like many other leukemia lines) are particularly sensitive to topotecan-induced apoptosis (60). Cells were treated for 24 h with the indicated concentrations of topotecan without and with veliparib, sedimented at 100 ϫ g for 5 min, and resuspended in ice-cold buffer consisting of 0.1% (w/v) sodium citrate containing 50 g/ml propidium iodide and 0.1% Triton X-100. After incubation at 4°C overnight, samples were subjected to flow microfluorimetry as described (61,62). Results were analyzed using BD Biosciences CellQuest software.
siRNA and shRNA-PARP1 siRNA oligonucleotides (63, 64) were synthesized by Ambion (Austin TX). A2780 cells were transfected by electroporation. On day 1, 1 ϫ 10 7 were sedimented at 50 ϫ g, washed twice with PBS, resuspended in antibiotic-free medium supplemented with 400 pmol of PARP1 or luciferase siRNA, treated with 320 V for 10 ms using a BTX 830 square wave electroporator, and seeded in 100-mm tissue culture plates. On day 2 the electroporation was repeated. On day 4 cells were harvested for immunoblotting or plated (1000 cells/60 mm dish), allowed to adhere for 4 h, treated for 24 h with the indicated concentrations of topotecan, washed, and allowed to form colonies for 7 days in drug-free medium.
For short hairpin-mediated knockdown in of TDP1 in A549 cells, lentivirus encoding TDP1 shRNA (Open Biosystems catalogue # V2LHS_219357) or control vector pGIPZ was packaged in HEK293T cells as previously described (65). A549 cells were exposed to virus for 48 h, washed, and selected with 2 g/ml puromycin. After knockdown was verified by immunoblotting in pooled transductants, cells were treated for 24 h with topotecan in the absence or presence of veliparib, washed, and allowed to form colonies in drug-free medium.
Drug Accumulation-Topotecan accumulation assays were performed by flow microfluorimetry as described previously (66). Briefly, adherent cells were trypsinized from plates, pelleted at 100 ϫ g for 5 min, and resuspended in RPMI 1640 buffered with 10 mM HEPES (pH 7.4 at 21°C). Cells were exposed to 20 M topotecan alone or in combination with veliparib or canertinib (1 M) for 10 min and immediately analyzed on a FACScan flow cytometer (BD Biosystems) using an excitation wavelength of 488 nm and an emission wavelength of 585 nm. CPT accumulation was similarly assayed on a BD Biosciences LSRII flow cytometer using an excitation wavelength of 355 nm and a 450/25-nm bandpass emission filter.
Alkaline Elution-Alkaline elution studies were performed as described (67) with several modifications. Logarithmically growing A549 cells were labeled for 24 h in medium A supplemented with 0.1 Ci/mmol [ 14 C-methyl]thymidine (PerkinElmer Life Sciences, Waltham, MA). After labeling, cells were washed with RPMI 1640 and allowed to grow in label-free medium A for another 24 h. After cells were released from their plates by trypsinization, aliquots were incubated at 37°C as indicated in Fig. 6. Cells were then deposited by gentle suction on Nucleopore phosphocellulose filters (1-m pore size; Millipore) and lysed by allowing 5 ml of buffer consisting of 1% (w/v) SDS, 100 mM glycine, 25 mM EDTA (pH 10), and 0.5 mg/ml proteinase K to drip through the filters. After filters were washed with 20 mM EDTA (pH 10), DNA was eluted with 20 mM EDTA (adjusted to pH 12.1 with tetrapropylammonium hydroxide). Fractions consisting of eluate, the filter, and the tubing were analyzed as described (68,69). Cells that received 0 -1500 centiGray of ␥-irradiation at 4 -8°C from a 137 Cs source were included in each experiment as a standard curve. For each drug treatment, the fraction of DNA remaining bound to the filter after the 6-h elution period was compared with this standard curve to designate damage in "rad equivalents." Immunoblotting-Cells were washed twice with calciumand magnesium-free Dulbecco's phosphate-buffered saline (PBS) and solubilized in 6 M guanidine hydrochloride containing 250 mM Tris-HCl (pH 8.5 at 20°C), 10 mM EDTA, 1% (v/v) 2-mercaptoethanol, and 1 mM freshly added phenylmethylsulfonyl fluoride. After preparation for electrophoresis as described previously (70), aliquots containing 50 g of protein (determined by the bicinchoninic acid method; Ref. 71) were separated on SDS-polyacrylamide gels containing 8% (w/v) acrylamide, electrophoretically transferred to nitrocellulose, and probed as described (72).
Estimation of Relative pADPr Levels-Effects of veliparib on relative pADPr levels were assessed by quantitative fluorescence microscopy. Staining was performed as described by Haince et al. (73) and Segovis et al. (74). In brief, cells grown on ethanol-washed coverslips were treated with the indicated concentration of veliparib for 4 h before the addition of 1 mM MMS for 30 min or 4 M topotecan for 1 h. After treatment, drugcontaining medium was aspirated, and the cells were fixed in methanol/acetone (70:30) for 20 min at Ϫ20°C. After fixation, cells were washed 3 times in washing buffer consisting of PBS, 0.1% (v/v) Triton X-100, and 0.1% (w/v) bovine serum albumin for 5 min each. Coverslips were then incubated in blocking buffer consisting of 1% (v/v) glycerol, 0.1% (w/v) gelatin from cold-water fish, 0.1% (w/v) bovine serum albumin, 5% (v/v) goat serum, and 0.4% (w/v) sodium azide in PBS for 1 h at 21°C. Coverslips were then exposed to 96-10 rabbit polyclonal anti-pADPr (1:500 dilution; Ref. 75) overnight at 4°C, after which they were washed 3 times with washing buffer, incubated for 1 h with Alexa Fluor-568 conjugated goat anti-rabbit secondary antibody (Invitrogen) diluted 1:1000 in blocking buffer, washed 3 times with washing buffer, counterstained with 1 g/ml Hoechst 33258 in PBS, and mounted with UltraLong antifade reagent (Invitrogen). Confocal images were captured on a Zeiss LSM 710 scanning confocal microscope with a 100ϫ/1.4 NA oil-immersion objective. Quantitation and image processing were performed with the Zeiss Zen software package and Adobe Photoshop CS3. All images from a single experiment were handled identically. Where indicated, mouse monoclonal anti-C-II-10 anti-PARP1 antibody and Alex Fluor-488-conjugated goat anti-mouse IgG were added to allow simultaneous visualization of PARP1 (76,77) and pADPr.

PARP Inhibition Sensitizes Cells to Topo I Poisons-Previous
studies demonstrated that a variety of PARP inhibitors enhance the antiproliferative and cytotoxic effects of topo I poisons (23)(24)(25)(26)(27)(28)(29)(30)(31). Consistent with those earlier studies, we observed that treating various solid tumor cell lines, including A2780, SKOV3, Ovcar5, and A549 cells, with the PARP inhibitor veliparib (78) sensitized cells to the topo I poison topotecan (Fig. 1,  A-D), which is used for the treatment of lung, ovarian, and cervical cancers, or the parent drug CPT (Fig. 1E). Similarly, the acute myelogenous leukemia cell line HL-60 was sensitized to topotecan (Fig. 1F), reflecting enhanced induction of apoptosis (supplemental Fig. S1). The enhanced effects of topotecan were observed at veliparib concentrations as low as 100 nM (Fig. 1A), which are readily achievable in the clinical setting (79). In contrast, even higher veliparib concentrations failed to sensitize cells to cisplatin, etoposide, gemcitabine, or paclitaxel (supplemental Figs. S1C and S2), excluding the possibility that veliparib sensitizes cells to all chemotherapeutic agents.
To place these results in context, effects of veliparib on pADPr formation in intact cells were examined. Using an assay that assesses levels of pADPr, which in our studies is detected exclusively in nuclei, the same compartment where PARP1 is located (Fig. 1G), we observed that concentrations above 10 M veliparib significantly inhibited MMS-induced pADPr formation, whereas 0.1 and 1 M had little effect ( Fig. 1H and supplemental Fig. S3). Similar results were obtained when MMS was replaced with topotecan, although the amount of polymer synthesized was smaller (Fig. 1I).
Veliparib Effects on Topo I-poisoned Cells and BRCA2-deficient Cells Are Distinguishable-Previous studies have shown that homologous recombination (HR)-deficient cells are hypersensitive to PARP inhibition (80,81), most likely due to disinhibition of error-prone nonhomologous end-joining and the resulting induction of chromosomal rearrangements and mutations (64). Additional experiments have shown that this cytotoxicity in HR-deficient cells occurs only at high PARP inhibitor concentrations (typically 2.5-40 M veliparib) that result in substantial inhibition of pADPr formation (64). To further investigate whether extensive inhibition of pADPr formation is likewise responsible for sensitization to topo I poisons, we have examined V-C8 (BRCA2-deficient) Chinese hamster cells because these cells, unlike slower growing HR-deficient human cell lines, can be readily killed by a 24-h exposure to PARP inhibitor.
When compared with HR-proficient V79 cells, V-C8 cells were hypersensitive to cisplatin (supplemental Fig. S4A) or CPT ( Fig. 2A), consistent with previous reports in other BRCA2-deficient lines (82,83). The hypersensitivity to CPT was not attributable to increased topo I levels (supplemental Fig. S4A), altered pretreatment cell cycle distribution (supplemental Fig. S4B), or increased CPT uptake (supplemental Fig.  S5, A-C). Moreover, V-C8 cells were not hypersensitive to gemcitabine (supplemental Fig. S4C), demonstrating that these cells do not exhibit hypersensitivity to all DNA-damaging agents.
We then compared the concentrations of veliparib required to inhibit V-C8 colony formation and the concentrations required for CPT sensitization. V-C8 cells exhibited veliparib hypersensitivity (Fig. 2, B and C) with a 50% reduction in colony formation after incubation with ϳ1.25 M veliparib for 48 h. Importantly, however, both V79 (BRCA2 wild-type) and V-C8 cells (BRCA2-deficient) were sensitized to CPT (Fig. 2, D and E) by 100 nM veliparib, a concentration that has minimal effect on V-C8 cells by itself (Fig. 2C). Similarly, V-C8 cells were sensitized to the topo I poison SN-38 at concentrations of the PARP inhibitor MK-4827 (50 nM) that are far below the concentration required to inhibit colony formation as a single agent (supplemental Fig. S6), ruling out the possibility that the observations are unique to any particular topo I poison or PARP inhibitor. The consistent observation that sensitization to topo I poisons occurred at PARP inhibitor concentrations much lower than those required for killing of BRCA2-mutant cancer cells suggested that these two phenomena might occur via distinct mechanisms.
PARP Inhibition Does Not Directly Alter the Cell Cycle, Topotecan Accumulation, or Stabilization of Topo I-DNA Covalent Complexes-The ability of veliparib to sensitize to topotecan at concentrations far below the IC 50 for inhibition of PARP activity ( Fig. 1) or killing of HR-deficient cells (Fig. 2) was striking and prompted further investigation. Because topo I poisons preferentially induce toxicity in S phase cells (13)(14)(15), we examined the cell cycle effects of veliparib administered alone or in combination with topotecan. Veliparib alone had no effect on cell cycle distribution in A549 cells (Fig. 3, A and B). Notably, however, when A549 cells were exposed to the combination of topotecan and veliparib, the previously reported G 2 /M-phase arrest (16,84) observed at any particular topotecan concentration was enhanced (Fig. 3B). Similar results were observed in A2780 cells (supplemental Fig. S7) and Ovcar5 cells. 4 The ability of veliparib to enhance topotecan-induced G 2 /Mphase arrest could potentially be explained by increased topotecan accumulation, increased topotecan-induced stabilization of Topo1cc, or impaired cellular response to DNA after forma-tion of Topo1cc. To test the first of these possibilities, a flow cytometry-based assay (66) was used to examine topotecan uptake in the absence and presence of veliparib. Canertinib, which inhibits ATP binding cassette transporters that affect topotecan accumulation (66), served as a positive control. These experiments failed to demonstrate any veliparib-induced alteration of topotecan accumulation in A549 cells (Fig. 3C) or other lines examined (Fig. 3D). Veliparib likewise had no effect on CPT accumulation in cell lines used in the present study (supplemental Fig. S5).
Earlier in vitro experiments suggested a model in which active PARP1, possibly acting by poly(ADP-ribosyl)ating topo I, directly alters the kinetics of the topo I enzyme and shifts the cleavage-religation equilibrium away from Topo1cc toward intact DNA (Fig. 3E, third panel) (33)(34)(35)(36)(37). This model would predict that increased Topo1cc covalent complexes would be formed in the presence of PARP inhibitors. To test this prediction, protein-linked DNA single-strand breaks were analyzed by alkaline elution in the presence of proteinase K. Topotecan induced DNA protein-linked single-strand breaks as previously reported (85), but co-treatment with veliparib failed to induce a further increase (Fig. 3F), arguing against a shift in the cleavagereligation equilibrium as an explanation for the observed sensitization by veliparib.
PARP1 Is Required for the Sensitizing Effects of Veliparib-Veliparib, like many PARP inhibitors in development, diminishes the catalytic activity of several PARP enzymes (78,86,87). To examine the importance of PARP1 in the mechanism of veliparib-induced sensitization, we initially utilized MEF lines developed by Wang et al. (56) from Parp1 Ϫ/Ϫ mice. These cells had similar cell cycle profiles to their wild-type counterparts (Fig. 4A) and expressed similar amounts of topo I (Fig. 4B). In agreement with results from Fig. 1, veliparib sensitized wildtype MEFs to CPT (Fig. 4C, circles). In contrast, Parp1 Ϫ/Ϫ MEFs were not sensitized by veliparib (Fig. 4C, triangles), indicating that PARP1 protein is necessary for the sensitization. Similar results were obtained with two additional PARP inhibitors (supplemental Fig. S8).
If PARP1 were playing a critical role in reversal of Top1cc, e.g. by helping recruit TDP1, Parp1 ablation would be expected to sensitize cells to CPT. Contrary to this prediction, Parp1 Ϫ/Ϫ MEFs were not hypersensitive to CPT (Fig. 4C, open symbols). Instead, they displayed a CPT sensitivity that is indistinguishable from wild-type cells, indicating that PARP1 catalytic inhibition and Parp1 deletion are not equivalent.
To ensure that these results were not a result of clonal variability between MEF lines, we repeated the experiment in an independently derived pair of 3T3 fibroblast lines provided by de Murcia et al. (57). As observed in the first pair of MEF lines, Parp1 Ϫ/Ϫ 3T3 fibroblasts were not intrinsically more sensitive to CPT than wild-type cells and were not sensitized to CPT by the addition of veliparib (Fig. 4D, triangles).
As an added control, we examined the effects of two different PARP1 siRNAs on topotecan sensitivity in A2780 human ovarian cancer cells. As indicated in Fig. 4E, these siRNAs failed to sensitize cells to topotecan even though the same siRNAs were able to selectively kill BRCA2-deficient PEO1 ovarian cancer cells in our previous studies (64). In contrast, TDP1 knockdown sensitized human cells to topotecan (supplemental Fig. S9), indicating that sensitization by PARP1 knockdown should have been observed using the current methodology if it had occurred.
Veliparib Converts PARP1 into Toxic Protein-The preceding results suggest that PARP1 must be present and must be inhibited by a PARP inhibitor to sensitize cells to topo I poisons.
To confirm this, we expressed wild-type human PARP1 in Parp1 Ϫ/Ϫ MEFs (Fig. 5, A and B), then treated with CPT in the absence and presence of veliparib. Expression of wild-type PARP1 was sufficient to restore the effects of veliparib on CPT sensitivity (Fig. 5C, triangles). Importantly, however, transfection with PARP1 did not sensitize to CPT in the absence of veliparib (Fig. 5C), suggesting that inhibited PARP1 is responsible for sensitization.
To further test this model, Parp1 Ϫ/Ϫ MEFs were transfected with PARP1(E988K), which is catalytically inactive (88). As indicated in Fig. 5D, this catalytically inactive PARP1 sensitized cells to CPT, and they were not further sensitized by the addition of veliparib. This veliparib-independent sensitization was also completely recapitulated by transfecting cells with the PARP1 DBD (Fig. 5, A and E).
Further experiments examined the effect of veliparib on topotecan-induced DNA damage signaling. This analysis demonstrated that activating phosphorylation of Chk1 was enhanced at each topotecan concentration by the presence of veliparib (Fig. 6, A and B). These observations raised the possibility that veliparib enhances topotecan-induced DNA damage.
Previous studies under cell-free conditions suggested a model in which PARP1 binds to damaged DNA and, in the absence of pADPr formation, inhibits repair (89). To assess whether this occurs in the presence of veliparib, Topo1cc were stabilized by topotecan, then allowed to reseal after dilution of the topo I poison (Fig. 6C). This analysis indicated that, as expected, the vast majority of DNA single-strand breaks were reversed upon dilution of cells treated with topotecan alone (Fig. 6D, open circles). In contrast, there was persistent DNA damage in cells treated with topotecan ϩ veliparib (Fig. 6, D and  E). In additional experiments, alkaline elution performed in the absence of proteinase K also detected the veliparib-induced increase in DNA strand breaks (Fig. 6F). Because Top1cc requires proteinase K for detection (85), these observations suggest that veliparib increases DNA breaks that occur downstream of Top1cc complexes.  FEBRUARY (23)(24)(25)(26)(27)(28)(29)(30)(31)(32). Despite the widespread reproducibility of this observation, the mechanistic basis for this sensitization has remained unclear. In the present study we report that this sensitization occurs at veliparib concentrations that inhibit PARP1 activity only by a small amount. Moreover, we demonstrate that this sensitization to topo I poisons, which depends on the presence of PARP1, is not recapitulated by Parp1 gene deletion. Instead, it is reproduced by transfection with catalytically inactive PARP1. These results have important implications for the ongoing clinical development of combination chemotherapy involving PARP inhibitors and topo I poisons (www.clinicaltrials.gov).  (66). Error bars, Ϯ S.E. of three independent assays. E, predicted shifts in cleavage-religation equilibrium based on previous reports that PARP1 binds to and facilitates topo I-mediated resealing are depicted. In this model inhibition of PARP1 would be predicted to result in accumulation of topo I-DNA covalent complexes (Top1cc). F, alkaline elution was performed to measure protein-linked DNA single-strand breaks in A549 cells exposed for 45 min at 37°C to the indicated topotecan concentrations in the presence or absence of 1 M veliparib. Lower inset, shown is a radiation standard curve from this experiment. Upper inset, shown are the summarized results of four independent experiments comparing the DNA single-strand breaks detected by alkaline elution in the presence of proteinase K after treatment with 250 nM topotecan in the absence and presence of 1 M veliparib. Gray horizontal line, mean of four determinations.

PARP1 Poisoning Sensitizes to Topoisomerase I Inhibitors
Previous experiments using purified proteins under cell-free conditions demonstrated a potential role for pADPr generated by PARP1 in the modulation of topo I activity (33)(34)(35)(36)(37). These experiments predicted that PARP inhibition would alter the kinetics of topo I-mediated strand religation, thereby shifting the cleavage-religation equilibrium (Fig. 3E). Contrary to these predictions, alkaline elution presented here (Fig. 3F) as well as recent experiments reported by Zhang et al. (32) failed to detect an alteration of the cleavage-religation equilibrium at veliparib concentrations that sensitized cells to topotecan and CPT. Although the present results cannot rule out a subtle effect of PARP inhibitors on the topo I cleavage-religation equilibrium, e.g. at higher PARP inhibitor concentrations that more completely inhibit pADPr synthesis, this does not appear to be the predominant mechanism by which submicromolar veliparib concentrations sensitize cells to topo I poisons.
An alternative model suggests that PARP inhibitors sensitize cells to topo I poisons by diminishing synthesis of pADPr, a polymer that plays a critical role in recruiting components of the base excision repair pathway and regulating the balance between HR and NHEJ. If this model were correct, then deletion of the Parp1 gene, which encodes the PARP responsible for the bulk of pADPr synthesis, would also be expected to sensitize cells to topo I poisons. Contrary to this prediction, we observed  that Parp1 gene deletion failed to sensitize two independently derived sets of mouse fibroblasts cells to CPT (Fig. 4, C and D). On the other hand, Parp1 deletion did abrogate the sensitizing effect of PARP inhibitors (Fig. 4, C and D, and supplemental Fig.  S8), thereby establishing a critical role for PARP1 in this sensitization.
Our observation that Parp1 Ϫ/Ϫ and Parp1 ϩ/ϩ fibroblasts exhibit similar camptothecin sensitivity appears to be at odds with a prior report that Parp1 Ϫ/Ϫ fibroblasts are hypersensitive to camptothecin (30). In that study Smith et al. (30)   of control cells (Fig. 4, C and D). It is certainly possible that the difference results reflect subtle differences in the origin of the cell lines or the assays used. However, we have also observed that knocking down PARP1 in A2780 cells (Fig. 4E) or transiently reexpressing PARP1 in Parp1 Ϫ/Ϫ MEFs (Fig. 5C) also fails to affect topotecan or camptothecin sensitivity, respectively, suggesting that PARP1 levels per se do not affect sensitivity to topo I poisons.
Additional experiments demonstrated that BRCA2-deficient V-C8 cells could be sensitized to CPT at veliparib concentrations at least an order of magnitude lower than those required to kill these PARP inhibitor-sensitive cells (Fig. 2, C and E). As is the case with Parp1 gene deletion, these observations suggest that PARP inhibitors are sensitizing by some mechanism other than extensively depleting cells of the PARP1 product pADPr.
Results of the present study suggest that veliparib concentrations above 10 M are required to inhibit pADPr synthesis in A2780 cells ( Fig. 1 and supplemental Fig. S3). In contrast, while this work was in progress, Zhang et al. (32) reported that camptothecin-induced polymer synthesis in HT29 colon cancer cells was robustly inhibited by 0.5 M veliparib. In view of recent results showing that veliparib uptake can be affected by certain drug transporters (90), part of this disparity might reflect the use of different cell lines. It is also possible that differences in the assays contribute. The pADPr immunofluorescence method reported by Zhang et al. (32) appears to detect an antigen that is present in both cytoplasm and nucleus (91). In contrast, the pADPr method used in the present work, which employs a different antibody, detects an antigen that localizes exclusively in the nucleus ( Fig. 1G and supplemental Fig. S3), where the vast majority of PARP1 is known to be located (76,77,92,93). Moreover, the need for 10 -40 M veliparib to substantially suppress PAR synthesis is consistent with recent results examining the effects of veliparib on protein poly(ADPribosyl)ation by immunoblotting (64).
The observations that (i) PARP inhibitors exert their effects at concentrations far below those required to inhibit the majority of pADPr synthesis and that (ii) PARP1 protein is required for the cells to be sensitized are highly reminiscent of the action of topo I poisons such as CPT or topotecan, which convert topo I into an agent that causes DNA damage (9,12). Indeed, prior studies performed under cell-free conditions raised the possibility that PARP1 might act as poison, binding to damaged DNA and preventing its repair when the PARP1 substrate NAD ϩ is absent (89). To our knowledge, the present observations provide the first evidence that PARP inhibitors, acting in intact cells, sensitize to topo I poisons in this fashion. In particular, we have observed that veliparib, although failing to alter the cleavage-religation equilibrium (Fig. 3F and Ref. 32), diminishes the repair of DNA damage that is not rapidly reversed by topo I itself (Fig. 6, C-F). Moreover, we have demonstrated that Parp1 deletion diminishes sensitivity of cells to the CPT/veliparib combination (Fig. 4, C and D, closed triangles) and transfection with catalytically inactive PARP1 recapitulates the effects of treating cells with PARP inhibitors (Fig. 5D). Finally, we have shown that sensitization to topo I poisons can be accomplished by transfection with the PARP1 DNA binding domain (Fig. 5E), which also sensitized cells to alkylating agents and ionizing radiation in prior studies (94 -97). All of these observations are consistent with a model in which PARP inhibitors poison the enzyme, i.e. convert PARP into a protein that diminishes repair of topo I-induced DNA damage arising downstream of Top1cc stabilization.
Studies are currently in progress in a number of laboratories to determine which aspects of repair are modulated by PARP inhibitors (32,64,98,99). Our results along with those of Zhang et al. (32) indicate that TDP1 knockdown sensitizes cells to topotecan and abrogates further sensitization by veliparib (supplemental Fig. S9). These results, which place TDP1 downstream of poisoned PARP1, raise the possibility that TDP1 cannot be recruited to trapped Top1cc when enzymatically inactive PARP1 is bound to the DNA.
PARP inhibitors are in the process of being tested as potential agents that might enhance the therapeutic effects of various DNA damaging agents (20). Thus, the mechanism by which PARP inhibitors sensitize cells has potentially important implications for this clinical development. For example, if PARP inhibitors sensitize cells solely by diminishing the synthesis of pADPr, then cells with lower levels of PARP1 might require lower PARP inhibitor concentrations to diminish pADPr synthesis to a sufficiently low level to inhibit pADPr-dependent repair. In contrast, if PARP inhibitors sensitize cells by converting PARP1 into an agent that potentiates the DNA damage of other agents, then cells with lower PARP1 expression would be expected to exhibit less sensitization as illustrated by the effects in Parp1 Ϫ/Ϫ cells (Fig. 4, C and D, and supplemental Fig. S8). Previous studies showing that Parp1 knock-out cells were intrinsically hypersensitive to alkylating agents and ionizing radiation (57, 94 -97) suggest that PARP inhibitors enhance the effects of alkylating agents and ionizing radiation at least in part by diminishing the synthesis of pADPr that is involved in repair. These observations suggest that, all other factors being equal, tumors with lower PARP1 expression might be better sensitized to alkylating agents and ionizing radiation by PARP inhibitors. In contrast, the present data suggest that tumors with low PARP1 might be less effectively sensitized to topo I poisons by PARP inhibitors.
In summary, this study provides evidence that PARP inhibitors can sensitize to topo I poisons at concentrations far below the IC 50 for pADPr synthesis. Further experiments have not only shown that PARP1 is responsible for these sensitizing effects, but also demonstrated that this sensitization reflects the ability of catalytically inactive PARP1 to serve as a protein that diminishes the repair of topo I-initiated DNA damage in intact cells. Further study is required to determine whether these observations can be successfully developed into a combination therapy that enhances the efficacy of topo I poisons.