Response of Human Prostate Cancer Cells and Tumors to Combining PARP Inhibition with Ionizing Radiation

Current treatment options for localized prostate cancer are radiation therapy or surgery. Both have shown similar survival outcomes in low- and intermediate-risk patients. Contemporary radiotherapy approaches such as intensity-modulated radiation therapy have permitted increased delivery of radiation to the prostate although sparing adjacent organs, reducing the potential for acute and chronic toxicity. However, proctitis, cystitis, and erectile dysfunction remain significant complications of high-dose radiotherapy. In turn, local failure after radiotherapy remains 20%–35% in intermediate- and high-risk patients (1, 2), leading to increased metastasis and lower survival. Hormone therapy has proven value when combined with localized radiotherapy in intermediate- and high-risk prostate cancer patients, but carries its own set of morbidities, including increased cardiovascular and thromboembolic risk (3, 4). Novel agents with more attractive side effect profiles that can be combined with radiotherapy to improve local control in high-risk patients and/or permit a dose reduction in lower-risk patients would be of great value.

An emerging strategy to improve efficacy at lower ionizing radiation (IR) doses is the use of radiosensitizers to target recognition and repair of DNA damage (5). PARPs are a family of enzymes that use NAD⁺ as a substrate to polymerize [poly(ADP-ribose); PAR] onto their cellular targets (6). The PARP1 and PARP2 isozymes are activated by DNA damage and participate in repair of single-strand breaks by activating XRCC1 and base-excision repair, and double-strand breaks (DSB) likely through influence on both the homologous recombination (HR) and nonhomologous end joining mechanisms. After a DSB, PARP is rapidly recruited and triggers poly-ADP ribosylation of PARP itself, histones, and other mediator proteins to stimulate chromatin loosening and DNA repair. PARP has long been considered a promising therapeutic target, and several small molecule PARP1 and PARP2 inhibitors are currently in preclinical and clinical trials, alone or in combination with DNA-damaging agents (7, 8). It has been observed that PARP is activated by IR and chemotherapy agents, and this has provided the rationale to examine the combined effects of PARP inhibitors and genotoxic therapy in tumor models and in clinical trials (9–11). Recent results have established an ability of PARP inhibitors to target cancers of specific genotypes via synthetic lethality (12–15), wherein PARP inhibition exposes the deficiency in tolerance for DNA damage created by defects in a DNA repair pathway such as HR by inhibiting the compensatory pathway, nonhomologous end joining. A specific example is the sensitization of BRCA mutant cancer cells to PARP inhibition, causing selective tumor cytotoxicity (16, 17). BRCA1 and BRCA2 mutations are not considered a major cause of familial or sporadic prostate cancer. However, a number of other mutations that decrease HR repair responses can also sensitize cells to PARP inhibitors, including defects in the inositide phosphatase PTEN, a gene commonly inactivated in prostate cancer (18). Downregulation of the HR pathway under hypoxic conditions can also lead to sensitization to PARP inhibition, an effect dubbed contextual synthetic lethality (19). Cells deficient in DNA DSB repair have been shown to be sensitized by PARP inhibitors to DNA damaging agents (20). Perhaps mediating all these effects, PARP may be continuously recruited to persistent damage in HR-defective cells, resulting in its modification by PAR and partial inactivation (21).

Nonetheless, the mechanisms by which PARP inhibitors mediate their beneficial effects in vivo remain poorly defined. Our previous work showed the combination of IR and the PARP inhibitor ABT-888 (veliparib; 2-[(R)-2-methylpyrrolidin-2-yl]-1H-benzimidazole-4-carboxamide) increased breast cancer cell senescence in vitro and in vivo, implicating persistent DNA damage as a mechanism (22). Senescence is an important tumor suppressive mechanism (23–25). The cellular equivalent of aging, replicative senescence is a DNA damage checkpoint response to telomere erosion mediated by activation of the p53/p21 and/or p16/ARF/Rb pathways and characterized by irreversible cell cycle arrest rather than cell death (26–28). Stress-induced premature senescence or accelerated senescence is an analogous persistent cell cycle arrest in cells with otherwise unlimited proliferative capacity, because of oncogene activation, oxidative stress, excessive mitogenic signals, chromatin perturbation, or accumulation of unrepaired DNA damage (25, 29–31). Current data suggest that the modified chromatin foci at sites of persistent DNA breaks serve a role in signaling to promote senescent arrest (31, 32). The defects in the p53 and/or Rb tumor suppressor pathways common in cancer may allow tumor cells to maintain genomic instability and tolerate persistent DNA damage, by blocking senescent signaling while promoting cell proliferation and survival.

Histone 2A, member X (γH2AX) and p53 binding protein 1 (53BP1) localization to IR-induced foci (IRIF) can serve as proxies for unrepaired DSBs and the DNA damage response (33, 34). Herein, by exploiting green fluorescent protein (GFP) fused to the chromatin-binding domain of 53BP1 as a live-cell reporter, and fluorescent immunocytochemistry for IRIF markers γH2AX and endogenous 53BP1, we monitored the effects of PARP inhibition on irradiated prostate cancer cells and tumor xenografts. We hypothesized that inhibition of PARP by using ABT-888 would enhance the antitumor effects of radiation in human prostate cancer cells in vitro and in experimental prostate cancer tumors in mice. Surprisingly, although PARP inhibition mediated radiosensitivity in both tumor cell lines in vitro, only PC-3 exhibited significant tumor regression in vivo. Our results suggest that in vitro assays of radiosensitivity may not predict in vivo efficacy of PARP inhibitors with radiation and that induction of senescence may be an important mechanism of PARP-induced radiosensitivity in vivo.

Two human androgen-unresponsive prostate cancer cell lines were used: PC-3, which is phosphatase and tensin homolog deleted on chromosome 10 (PTEN)-negative, p53-null; and DU-145, PTEN wild type, p53 mutated (35). Both cell lines were purchased from American Type Culture Collection (ATCC). They have been authenticated by short tandem repeat analysis at Johns Hopkins University on February of 2011, by using the Identifiler Kit (Applied Biosystems), and compared with known profiles (ATCC short tandem repeat profile database and NCI-60 cell line panel; ref. 36).

To examine the effect of PARP inhibition on IRIF persistence in living cells, we exploited our previously described IRIF reporter consisting of GFP fused to the 53BP1 IRIF binding domain, expressed under tetracycline-inducible control (GFP-IBD; ref. 22). GFP-IBD cloned into the pLVX-Tight-Puro vector (Clontech) was transfected along with pLVX-Tet-On Advanced (Vector) into the PC-3 cell line, by using FuGENE HD transfection reagent (Roche). Following G418 and puromycin selection, cells cultured in Dulbecco's Modified Eagle's Medium/F12 medium (Invitrogen) with 10% Tet system-approved FBS (Clontech) were induced with 1 μg/mL doxycycline and sorted to establish a PC-3 GFP-IBD cell line.

Antibodies used were rabbit anti-phospho-H2AX Ser139 diluted 1:500 (γH2AX; Cell Signaling Technology) and rabbit anti-53BP1 diluted 1:500 (Novus Biologicals), detected by Texas Red anti-rabbit IgG diluted 1:1,000 (Vector Laboratories). Cells were grown in cover slips and after treatment they were rinsed with PBS and fixed with 4% paraformaldehyde for 30 minutes at 4°C. Permeabilization was done with 0.3% Triton X-100 and blocking with 5% bovine serum albumin in phosphate buffered saline with Tween 20 for 30 minutes. Incubation with primary antibodies was carried out at room temperature for 1 hour. After washing with PBS, incubation with secondary antibodies was done at room temperature for 1 hour in dark. Cover slips were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and mounted by using ProLong Gold antifade reagent (Invitrogen). Harvested tumors were fixed in 10% formalin for 48 hours and embedded in paraffin. The sections (4 μm) were deparaffinized and then stained by using the same protocol. Images were captured on a Zeiss Axiovert 200M epifluorescence microscope with a ×63 PlanApo 1.4 NA objective by using a Hamamatsu ORCA ER digital camera and Improvision OpenLab software.

Five hundred cells were seeded to form colonies in p100 plates and treated the next day with 6 Gy alone, or with 10 μmol/L of ABT-888 followed by 6 Gy 30 minutes later. When sufficiently large colonies with at least 50 cells were visible (after 1 week for PC-3 and 2 weeks for DU-145 cells), the plates were fixed with methanol and stained with crystal violet. Colonies with more than 50 cells were counted.

For cell death analysis, cells were collected at 48 hours after treatment and stained with propidium iodide (PI). For cell cycle analysis, cells were fixed in ice-cold methanol while shaking to avoid agglutination and resuspended in PBS plus RNAse and PI. Stained cells were analyzed in an LSR-II flow cytometer (Becton Dickinson). Data were analyzed on FlowJo (Becton Dickinson) by using the cell cycle platform and calculating G₁, S, and G₂–M fractions by fit to the Watson Pragmatic model.

Total RNA was isolated by using Trizol (Invitrogen) and quantified by using the QuBit platform (Invitrogen). A total of 850 ng of total RNA was subjected to DNAse I digestion by using Amplification Grade DNAseI (Invitrogen) following the manufacturer's protocol. A total of 550 ng DNAsed RNA was subjected to cDNA synthesis in a 20 μL reaction volume by using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) and subsequently diluted 1:10 for use in quantitative reverse transcription (qRT)-PCR. qRT-PCR was carried out on the ABI7900HT in a 384-well plate in a 5 μL reaction volume containing 2.5 μL 2× Power SYBR Master Mix (Applied Biosystems), 0.5 μL of 10 μmol/L primer mixture, and 2 μL of the diluted cDNA. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the endogenous control, and fold change calculations were made by using the comparative C_t method. Primer sequences were as follows:

• CDKN1A(p21)-f: GCGAGGCCGGGATGAGTTGG

• CDKN1A(p21)-r: CAGCCGGCGTTTGGAGTGGT

• GAPDH-f: CTCTGCTCCTCCTGTTCGAC

• GAPDH-r: GTTAAAAGCAGCCCTGGTGA

The senescence-associated β-galactosidase (SA β-Gal) assay was conducted as described before (37). Images were captured on a Zeiss Axiovert 200M and Zeiss Axiocam color digital camera controlled by OpenLab software with a ×20 objective.

Female athymic nude mice underwent s.c. injection of 1 × 10⁷ PC-3 or DU-145 cells in 100 μL of PBS. Once tumors grew to 100 mm³, mice received a dose of 0 or 6 Gy and no ABT-888 or 25 mg/kg of ABT-888 in water twice daily by oral gavage 48 hours before IR and for 48 hours after IR.

To study the molecular effects of treatment with the PARP inhibitor ABT-888 and radiation in prostate cancer cell lines, we first investigated their effects on formation of DNA damage foci in PC-3 and DU-145 cells. We initially monitored the effect of ABT-888 alone in PC-3 cells expressing GFP-IBD (PC-3 GFP-IBD). Unirradiated PC-3 cells displayed pan-nuclear GFP-IBD, with absent or sporadic nuclear foci. Addition of 10 μmol/L ABT-888 to the culture media caused the GFP-IBD reporter to relocalize to nuclear foci, noted at 2 hours and that did not resolve after 24 hours. These foci have been shown to represent accumulation of unrepaired endogenous DNA damage (38) and formed in the absence of genotoxic exposure (Fig. 1A). Formation of DNA damage foci in PC-3 cells was confirmed with immunofluorescence for γH2AX and endogenous 53BP1 (Fig. 1B). In comparison, DU-145 cells did not show increased GFP-IBD nuclear foci after treatment with 10 μmol/L ABT-888 alone, and immunofluorescence revealed only pan-nuclear fluorescence or sporadic foci for the same markers (Fig. 1B).

Following treatment with ABT-888 alone, we explored its effect in combination with IR. Nuclear GFP-IBD foci formed rapidly in PC-3 cells after 6 Gy and were prominent at 2 hours but then diminished over the next 24 hours (Fig. 1A). Combining IR and ABT-888 markedly slowed the resolution of GFP-IBD foci over 24 hours (Fig. 1A). Immunofluorescence staining for γH2AX and endogenous 53BP1 in PC-3 cells after 6 Gy alone or combined with ABT-888 revealed a similar pattern and kinetics of nuclear foci (Fig. 1B). ABT-888 induced persistent foci, IR alone induced nuclear foci within 2 hours that mostly resolved after 24 hours, and the addition of ABT-888 to 6 Gy of IR prevented the resolution of foci after 24 hours. As observed in PC-3 cells, DU-145 cells treated with 6 Gy displayed γH2AX and 53BP1 foci at 2 hours that partially resolved by 24 hours. Similarly, the foci persisted when 6 Gy was combined with 10 μmol/L ABT-888. We conclude that DU-145 cells are not sensitive to induction of DNA damage foci by PARP inhibition alone, but they are able to form foci after IR and show foci persistence after combined treatment with 6 Gy and ABT-888.

To determine the effect on cell survival, PC-3 and DU-145 cells were analyzed by clonogenic assays after treatment with ABT-888 and IR, alone or in combination. A total of 10 μmol/L ABT-888 alone induced a significant inhibition in colony formation in PC-3 cells (100% ± 5% for control vs. 77% ± 6% for ABT-888; P = 0.006; t test), but no significant effect was observed in DU-145 cells (100% ± 11% for control vs. 90% ± 10% for ABT-888; P = 0.37; t test; Fig. 2A). However, 10 μmol/L ABT-888 combined with IR reduced colony formation in both cell lines, compared with IR alone. In PC-3 cells, significant differences were noticeable from 1 Gy (89% ± 10% for IR alone vs. 44% ± 3% for IR + ABT-888; P = 0.002; t test), with similar fold effects at each IR dose up to 6 Gy (Fig. 2B). In DU-145 cells, the survival fractions began to differ from 2 Gy (58% ± 4% for IR alone vs. 47% ± 4% for IR + ABT-888; P = 0.038; t test; Fig. 2B).

We explored the effect of PARP inhibition and IR on cell cycle kinetics 48 hours after treatment by using permeabilization, PI staining, and flow cytometry. Consistent with the formation of IRIF after addition of ABT-888 alone, PC-3 cells shifted toward G₂–M DNA content, whereas DU-145 cells displayed no appreciable change in cell cycle distribution (Fig. 3). IR treatment increased the proportion of cells with G₂–M DNA content in each cell line. This effect was enhanced by ABT-888, with PC-3 cells displaying a greater G₂–M shift than DU-145 cells.

The persistence of DNA damage foci and G₂–M shift led us to investigate cellular consequences of treatment with ABT-888 and IR. No differences were noted in cell death at short times in either cell line after any treatment combination via analysis of PI uptake by nonpermeabilized cells; the percentage of cells with PI uptake ranged from 2.5% to 6.8% across all groups in both cell lines. When cells remain viable and display persistent DNA damage, one potential outcome is the induction of accelerated senescence. Thus, we evaluated the treated cells for characteristic markers of senescence including a large and flattened cell morphology, accumulation of SA-β-Gal and p21 overexpression. ABT-888 or IR treatment alone did not induce significant SA-β-Gal activity or p21 expression in either cell line. However, we observed markers of accelerated senescence in PC-3 cells by 4 days after treatment with ABT-888 and IR, including an enlarged flat morphology and positive staining for SA-β-Gal (Fig. 4A). There was also a 5-fold increase in p21 gene expression as determined by PCR (Fig. 4B), and increased protein expression detected by immunofluorescence (Fig. 4C). In DU-145 cells, only isolated cells showed SA-β-Gal activity, and no increase in p21 was detected (Fig. 4A–C). We did not observe overexpression of other senescence markers, including p16 and p27, in either cell line.

To investigate the effects of PARP inhibition and IR, alone and in combination, in tumor xenografts, PC-3 and DU-145 cells were injected into nude mice to form tumors. Mice were treated with ABT-888 and/or IR as described above. Tumors were harvested 4 days after treatment, and tissue sections were evaluated for γH2AX and 53BP1 foci. The findings in tumor tissues mirrored the results in vitro, showing DNA damage foci induction by ABT-888 alone in PC-3 xenografts, but not in DU-145 cells. In irradiated PC-3 and DU-145 tumors, the addition of ABT-888 caused foci persistence 4 days after treatment compared with IR alone (Fig. 5A). Additional tumors were harvested at 7 days after treatment, and frozen sections were analyzed for senescence markers. In PC-3 tumors, numerous cells stained positive for SA-β-Gal (Fig. 5B), and PCR showed increased expression of p21 (not shown) after combined treatment. In DU-145 tumors, there were only isolated cells with detectable SA-β-Gal activity (Fig. 5B), and no induction of p21 was apparent (not shown).

We also analyzed tumor growth in mice after treatment with 6 Gy and ABT-888, alone or in combination. ABT-888 had a moderate effect slowing tumor growth in PC-3 tumors (mean V/V₀ at 15 days, 13 in the ABT-888 group vs. 10 in controls; Fig. 6A). In combination with 6 Gy, the effect of PARP inhibition was more robust, and it significantly delayed regrowth in PC-3 tumors (mean V/V₀ at 32 days, 12 in IR group vs. 4.5 in combination group; P = 0.07; t test). DU-145 tumor growth was not slowed by ABT-888 treatment, alone or combined with IR. In fact, we note some protective effect of ABT-888 on DU-145 tumor growth after 6 Gy (Fig. 6B).

PARP inhibitors are a class of highly promising targeted therapy agents that are showing benefits alone and in combination with genotoxic therapy in a wide range of cancers in preclinical models and clinical trials (7, 8). Much of the focus has been on the synthetic lethality mechanism (12–15) that has brought these agents to the forefront in treatment of BRCA mutant and triple-negative breast cancers. Trials in prostate cancer patients remain at an early stage and have yet to show clear benefits. Among the few papers examining PARP inhibitors in prostate cancer cells, prominent examples have focused on the potential for exploiting synthetic lethality. These studies have examined PARP inhibition as a sensitizer to defects in HR resulting from loss of Rad51 expression because of PTEN deficiency (PC-3; ref. 18) or induced by hypoxia (DU-145; ref. 39). Other studies have examined the effects of combining a PARP inhibitor with temozolomide on prostate cancer cells (40). In prostate cancer, PARP inhibitors are most likely to find their use in combination with conventional treatments such as radiation therapy. Perhaps most significant to in vivo studies, although the molecular targets of PARP inhibitors are known and a mechanism of action seems to be established, how PARP inhibition synergizes with genotoxic agents, and how it mediates its growth inhibitory effects on tumors remains poorly defined.

When IR treatment was added to PARP inhibition in vitro, we observed radiosensitization in both the PC-3 and DU-145 cell lines, manifested by a decrease in colony formation, increased IRIF persistence, and induction of cell cycle arrest. The in vitro effects were greater in PC-3 cells compared with DU-145, and PC-3 cells developed hallmarks of senescence after combined treatment. The effectiveness of the combination was confirmed in vivo in PC-3 tumor xenografts, delaying tumor growth compared with IR and inducing a senescent phenotype. However, no improvement in the effect of IR was observed in DU-145 xenografts. We hypothesize that senescence in PC-3 cells can partially explain the difference in the response in vivo, because no significant senescence was observed in DU-145 cells. DU-145 cells have been found to be more radioresistant in vitro to IR treatment, whereas PC-3 tumors have been reported to have a higher hypoxic fraction compared with DU-145 tumors (52% vs. 7%, respectively), which could potentially increase the radioresistance of PC-3 in vivo (41). However, ABT-888 has also been reported to sensitize hypoxic cancer cells to a level similar to oxic radiosensitivity (39); this mechanism might overcome the potential radioresistance from a higher hypoxic fraction in PC-3 tumors.

Consistent with prior work from Mendes-Pereira and colleagues (18), we found that treatment with ABT-888 alone had some efficacy in vitro in PC-3 prostate cancer cells, which are defective in PTEN. It has been described that PTEN deficiency causes an HR defect in human tumor cells, making them a therapeutic target of PARP inhibitors (18, 42, 43). PTEN, besides inactivating the P13-K/AKT pathway, has a nuclear function controlling chromosomal integrity and regulating the expression of Rad51, which reduces the incidence of spontaneous DSBs (44). PARP inhibition alone induced DNA damage foci, inhibited cell proliferation, and promoted G₂–M cycle arrest in PC-3 cells. A similar trend in PC-3 tumor xenografts was observed with a moderate suppression of tumor growth. ABT-888 alone did not exert any of the above effects in DU-145 cells in vitro at a similar drug concentration. These results support the current model that the efficacy of PARP inhibitors in HR deficient BRCA1⁻, BRCA2⁻, or PTEN-negative cancer cells is the result of accumulation of unrepaired endogenous DNA damage. It could be hypothesized that if the efficacy of PARP inhibitors in PTEN-deficient prostate tumors translates into a significant clinical benefit, in selected cases, they could be used as monotherapy, allowing adequate tumor control although preventing the complications associated with radiation therapy and surgery.

Our results are consistent with other reports that have shown increased efficacy of IR with the addition of PARP inhibitors (9, 11), and further implicate IRIF persistence as a determinant of the increase in accelerated senescence (31, 32). Senescence can result from several inducers, including accumulation of unrepaired DNA damage. Therapy-induced senescence is increasingly being reported as an alternative mode of cell death in addition to apoptosis and necrosis and is proposed to contribute to tumor control following treatment with cytotoxic agents (29, 45). Along with prior studies (11, 39), our results in PC-3 cells and tumors suggest that PARP inhibitors can be an effective radiosensitizing strategy. Additionally, consistent with our previous work (22), accelerated senescence may be a factor in the therapeutic response of some human tumors to IR combined with PARP inhibition.

Liu and colleagues (39) previously observed that DU-145 cells treated with IR and ABT-888 displayed increased toxicity without undergoing apoptosis. In our experiments, ABT-888 sensitized DU-145 cells to IR as measured by clonogenic assay and induction of DNA damage foci, without inducing senescence or apoptosis, suggesting a mitotic death. This p53 mutant cell line also has a nonsense mutation in the Rb gene (35), impairing both the p53/p21 and p16/ARF/Rb senescence pathways (32). We ascribe the lack of benefit of ABT-888 on the DU-145 tumors to their distinct response to persistent DNA damage. We also acknowledge that there may be other unidentified host or tumor factors that may impinge upon the effectiveness of ABT-888 in DU-145 tumors. A favorable interpretation of our data is that a key determinant of the efficacy of PARP inhibitors in vivo may be their ability to drive cells toward senescent arrest. Insofar as senescent cells induced by genotoxic treatment can persist in tissue for weeks or months and can alter the microenvironment via paracrine signaling (46, 47), they may be able to limit the recovery of other tumor cells. These results also highlight the importance of tailoring therapy to each tumor.

We conclude that PARP inhibitors have therapeutic potential in specific types of prostate cancer in combination with radiation therapy, and even as monotherapy in DNA repair defective tumors. Induction of accelerated senescence is a novel therapeutic approach, and deserves consideration in clinical trials. Despite its increasing recognition as a potential alternative for tumor control, there is still a lack of reliable senescence-inducing agents, and this area remains an open field for further research. PARP inhibitors are strong candidates for this purpose, though potentially limited to specific tumor types.

No potential conflicts of interest were disclosed.

Ludwig Center for Metastasis Research, NCI SPORE in Prostate Cancer P50 CA090386 (R.R. Weichselbaum, S.J. Kron), NIH R21 CA138365 (S.J. Kron, R.R. Weichselbaum), NIH R01 GM60443 (S.J. Kron), Foglia Family Foundation, The Francis L. Lederer Foundation, Center for Radiation Therapy, American Society of Clinical Oncology Translational Research Professorship (E.E. Vokes).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.