HMG-CoA Reductase Inhibition Delays DNA Repair and Promotes Senescence After Tumor Irradiation

Cancer cells are rarely intrinsically radiosensitive, but advances in image-guided radiotherapy allow radiation to be focused on the tumor, providing a significant therapeutic advantage. Nonetheless, radiation is often insufficient on its own. Concomitant genotoxic chemotherapy is highly effective as a radiosensitizer, but chemoradiotherapy incurs dose-limiting and debilitating normal tissue toxicities while offering relatively small gains in the therapeutic ratio. This has provided an impetus to identify radiosensitizers that enhance radiation effects in the tumor with minimal impact outside the radiation field (1).

The beneficial effects of radiotherapy are considered to be mediated by formation of DNA double strand breaks (DSB). DSBs are potentially lethal damage insofar as cell division can result in loss of chromosome arms. When cells suffer DSBs that exceed capacity for repair, they may die via apoptosis, necrosis or mitotic catastrophe. Surviving damaged cells may undergo therapy-induced senescence (TIS), a persistent cell-cycle arrest characterized by enlarged cell size and activation of a gene expression program leading to secretion of inflammatory mediators (2–5). TIS has been documented in diverse human tumors after treatment with radiation and/or chemotherapy. While considerable disagreement remains whether TIS is a desirable or detrimental outcome of treatment (6–10), several studies indicate that senescent cells can suppress proliferation of surviving tumor cells and/or promote antitumor immune response (5, 11, 12).

Agents targeting DSB repair (reviewed in refs. 13–15) are considered promising therapeutics based on their potential to exploit intrinsic features of cancer cells but may also be particularly well-suited to application as radiosensitizers. Although two PARP inhibitors have been approved and others are in late development, other promising DNA repair inhibitors have been abandoned due to low efficacy and/or high toxicity. Given practical constraints, an attractive source of new radiosensitizers might be agents that already have well-established safety profiles, such as known drugs approved for other indications. Indeed, repurposing or repositioning has broad value for oncology (16, 17) and approved drugs have proven to be a rich source of candidate radiosensitizers (e.g., refs. 18–20). Our strategy was to screen in vitro for persistence of ionizing radiation–induced foci (IRIF), the modified chromatin domains that mark DSBs (18). Among hits that were validated by assaying enhanced growth delay after irradiation (21) was the 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) reductase inhibitor pitavastatin (ref. 22; NK-104; Livalo, Kowa Pharmaceuticals).

HMG-CoA reductase is the rate-limiting enzyme in the mevalonate pathway, regulating synthesis of cholesterol and its isoprenoid intermediates, geranylgeranyl pyrophosphate (GGPP) and farnesyl pyrophosphate (FPP; refs. 23, 24). Modification by GGPP and/or FPP is essential for function of small GTPases, lamins, and other proteins with key roles in proliferation, survival, invasion, metastasis, inflammation, and immune response (25–27). Thus, by attenuating mevalonate synthesis, statins not only decrease cholesterol biosynthesis but have pleiotropic effects impinging on multiple cancer pathways. In particular, diverse prenylated proteins including Ras, Rho, Rac, Ran, and Rap GTPases and lamins regulate DNA repair, apoptosis, senescence, and/or mitosis, suggesting that effects on multiple targets might mediate statin interactions with radiation.

Taking advantage of the high percentage of adults treated with HMG-CoA reductase inhibitors (statins) to treat dyslipidemia or lower cardiovascular risk, multiple studies have associated use with lower incidence and/or mortality for diverse cancers (e.g., ref. 28). Several studies have linked statin use to improved outcomes after radiotherapy for prostate (29, 30), rectal (31), and breast (32) cancers, although others report conflicting results. In turn, multiple studies have demonstrated statins can protect against normal tissue damage after radiation or chemotherapy (33, 34).

Given the lack of consensus on interactions between HMG-CoA reductase inhibitors and radiotherapy, we examined the effects of statins on radiation response. Pitavastatin delayed DSB repair and increased cell senescence after irradiation, leading to prolonged tumor growth delay. Confirming "on-target" effects, blocking protein farnesylation by RNA interference recapitulated pitavastatin's effects. Taken together with prior work, our data support repurposing pitavastatin or other lipophilic statins as radiosensitizers, with particular applications to image-guided radiotherapy.

The MCF7^Tet-On advanced human mammary carcinoma cell line was obtained from Clontech in 2007 and was frozen in liquid nitrogen after 3–5 passages as a stock. We used the previously reported MCF7^GFP-IBD (35) where MCF7^Tet-On advanced was modified to express GFP fused to the 53BP1-IRIF–binding domain (IBD) under tetracycline-inducible control. MCF7^GFP-IBD cells were maintained in DMEM containing 4.5 g/L glucose (Gibco), 1% penicillin/streptomycin (Gibco) supplemented with 10% Tet-approved FBS from (Clontech), and cultivated less than 10 passages prior to use. The cells were tested for mycoplasma and authenticated by short tandem repeat profile (IDEXX BioResearch) prior to and after performing experiments. All experiments were performed from 3 to 10 passages after thawing cells. Mouse melanoma cell line B16.SIY was maintained in complete RPMI medium (Gibco) containing 1% penicillin/streptomycin supplemented with 10% FBS. The B16.SIY cells were a kind gift from T. Gajewski and used with authentication by IDEXX.

Statin drugs were obtained from commercial sources as follows: pitavastatin calcium from Atomole, lovastatin, pravastatin sodium, and atorvastatin calcium from Toronto Research Chemicals, simvastatin from Cayman Chemical, and rosuvastatin calcium from Biotang. Mevalonate pathway metabolites R-mevalonic acid sodium salt and farnesyl pyrophosphate ammonium salt were from Sigma and geranyl pyrophosphate ammonium salt was from Axon MedChem. PARP1/2 inhibitor veliparib (ABT-888) was obtained from ChemieTek.

Pairs of Sigma MISSION short hairpin RNA (shRNA) targeting expression of HMG-CoA reductase (HMGCR), farnesyl diphosphate synthase (FDPS), geranylgeranyl diphosphate synthase 1 (GGPS1), farnesyl diphosphate transferase beta (FNTB), and nontargeting scrambled (Scr) negative control plasmids were obtained in pLKO.1-puro vectors and used according to manufacturer's instructions. Lentivirus-containing supernatant was produced by transfection of the 293T Lenti-X cell line (Clontech) and applied to MCF7^Tet-On cells. Following selection in the presence of puromycin, pairs of stable MCF7^Tet-On cell lines with silencing of HMGCR, FDPS, GGPS1, and FNTB protein expression were established. Cells from the third passage postselection were frozen in liquid nitrogen as a stock. Most experiments were done on 4–6 passages of established cell lines. At least 2 shRNA constructs targeting different sequences of corresponding mRNA were evaluated for each gene. Silencing of targeted genes was validated by Western blot analysis. shRNAs and antibodies used in this study are described in Supplementary Tables S1 and S2.

Mice were maintained according to guidelines of the Institutional Animal Care and Use Committee and irradiated using a RadSource RS-2000 X-Ray generator operating at 160 kV and 25 mA at a dose rate of 2.20 Gy/minute, calibrated by NIST traceable dosimetry. To establish MCF7^GFP-IBD tumors, 17β-estradiol pellets (1.7 mg, Innovative Research of America) were implanted in 6-week-old athymic nude mice (Harlan Laboratories) 7 days before subcutaneous injection of 1 × 10⁷ MCF7^GFP-IBD cells in 100 μL of PBS. Once tumors grew to 300 mm³, 2 mg/mL doxycycline with 1% sucrose was added to the drinking water for 72 hours before irradiation. Mice were treated with pitavastatin (Atomole) by oral gavage (10 mg/kg or 30 mg/kg as indicated) 2 days before, the day of, and 2 days after irradiation. Mice were treated with veliparib (ChemieTek) twice daily by oral gavage (25 mg/kg). IRIF formation in MCF7^GFP-IBD tumors was analyzed at 3 and 24 hours after irradiation with 6 Gy. For B16.SIY tumors, C57BL/6 female mice (Harlan) were injected in the hind limb with 1 × 10⁶ cells suspended in 100-μL PBS. After 8 to 12 days, mice were placed into treatment groups: control, 15 Gy, drug alone, or drug + 15 Gy and treated with pitavastatin or veliparib as described above. Mice were maintained according to guidelines of the Institutional Animal Care and Use Committee.

MCF7^GFP-IBD and B16.SIY cells were plated at 100 cells per well in 6-well plates in triplicate in corresponding medium. Twenty-four hours later, drugs were added 1 hour prior to irradiation. Radiation was delivered using a GammaCell ⁶⁰Co source (MDS Nordion) with dose rate ranging from 10.5 to 9.4 cGy/second depending on the date of the experiment. Cells remained in culture for 9–14 days and colonies of at least 50 cells were counted.

For neutral comet assays, cells were seeded at 1 × 10⁵ per well in 6-well plates and treated as above. After 24 hours, cells were mixed with Comet LM agarose and single-cell electrophoresis was performed on CometSlides (Trevigen). Slides were fixed, dried, stained with SYBR green (Thermo Fisher Scientific), and imaged on an Axiovert 40 with a 20× Plan-NeoFluar objective and AxioCam camera. Two or more replicates were performed. Images were analyzed using an ImageJ comet assay macro (http://www.med.unc.edu/microscopy/resources/imagejplugins-and-macros/comet-assay).

For IRIF imaging, MCF7^GFP-IBD cells were seeded on cover glass at 2.5 × 10⁴ per well in 24-well plates. GFP-IBD expression was induced with 1 μg/mL doxycycline for 48 hours. Pitavastatin or other statins were added for 1 hour prior to IR. After 24 hours, cells were fixed, stained with 5 μg/mL Hoechst 33342, mounted using ProLong Gold (Invitrogen). Immunocytochemistry for γH2AX foci was performed using clone JBW301 (EMD Millipore). IRIF images were captured on an Axiovert 40CFL (Zeiss). Two or more replicates were performed.

Formaldehyde-fixed paraffin-embedded (FFPE) tumor sections were stained with hematoxylin and eosin. IHC for Ki-67 was performed using antibody clone SP6 from rabbit (Thermo Scientific), anti-rabbit-HRP polymer (Vector Laboratories), and HRP-reactive AMEC Red colorimetric staining kit (Vector Laboratories), using hematoxylin (Polysciences, Inc.) for tissue counterstaining. H&E and Ki-67 imaging was conducted using an Axioskop upright microscope with a 0x/0.3NA objective, Axiocam color CCD camera, and Zen imaging software (Zeiss). Immunofluorescence for γH2AX was performed using antibody clone JBW301 from mouse, anti-mouse-biotin (Vector Laboratories), and streptavidin-DyLight 594 (Vector Laboratories), using DAPI (Sigma-Aldrich) for tissue counterstaining. gH2AX was imaged using an Axiovert 200M inverted fluorescence microscope with a 40X/1.3NA objective (Zeiss), a monochrome CCD camera (Hamamatsu), and SlideBook imaging software (3i). ImageJ software (NIH) was used to prepare images for publication. A representative tumor sample from each treatment group was selected for analysis. Additional protocol details are provided in Supplementary Methods.

Senescence-associated beta-galactosidase (SA-β-Gal) activity was determined on cell culture as well as frozen sections of excised tumors 7 days after IR. Cells were seeded at 3 × 10⁴ per well in 6-well plates. Next day, cells were treated with drug for 1 hour prior to irradiation. Cells were fixed after 5 days and assayed for SA-β-Gal as described previously (35). SA-β-Gal–positive and negative cells were counted in multiple fields, yielding an average percent SA-β-Gal–positive staining, indicated on each SA-β-Gal image as mean ± SEM. Two or more replicates were performed. To evaluate senescence in vivo, 10–12 μm cryosections of optimal cutting temperature (OCT)-embedded tumors were fixed in 2% paraformaldehyde, stained for SA-β-Gal activity, counterstained with nuclear fast red, dehydrated, mounted, and imaged. A representative tumor sample from each group was selected for analysis.

Statistical significance for IRIF counting and comet assays was determined using the nonparametric Mann–Whitney test. Calculations were performed using Prism software (GraphPad). P value ≤ 0.05 was considered statistically significant.

In prior work (18), we identified compounds that target DNA DSB repair by screening drug repurposing libraries against MCF7^GFP-IBD human breast cancer cells irradiated with 6 Gy, using the GFP-IBD 53BP1 reporter as a live-cell imaging reporter for ionizing radiation-induced foci (IRIF) formation and resolution. Among diverse compounds found to promote IRIF persistence 24 hours after irradiation, we had identified two HMG-CoA reductase inhibitors in the NIH Clinical Collection library of agents with a history of use in phase I, II or III clinical trials, the withdrawn agent cerivastatin (Baycol, Bayer) and the clinically used drug pitavastatin (Livalo, Kowa, Supplementary Fig. S1A–S1C). In secondary screens, pitavastatin significantly increased IRIF persistence at 24 hours as compared with cells treated only with radiation without increasing background IRIF over that of vehicle control (P < 0.0001; Fig. 1A and B).

Pitavastatin combined with ionizing radiation (IR) suppressed colony formation in irradiated MCF7^GFP-IBD cells compared with IR alone, similar to the activity of the PARP1/2 inhibitor veliparib, a well-established radiosensitizer (Fig. 1C). Radiosensitization of MCF7^GFP-IBD cells by pitavastatin was dose-dependent (Supplementary Fig. S1D). While MCF7^GFP-IBD cells displayed an SF₂ (surviving fraction at 2 Gy) of 0.54 for IR alone, addition of 1.25, 2.5, or 5 μmol/L pitavastatin reduced SF₂ to 0.4 (P < 0.0001), 0.29 (P < 0.001), and 0.18 (P < 0.0001) respectively.

Pitavastatin also potentiated effects of radiation on cellular senescence in vitro (Fig. 1D). When combined with 6 Gy, 10 μmol/L pitavastatin increased the number of large, flat cells expressing SA-β-Gal at 7 days after treatment to 84% ± 2% from 24% ± 3% with IR alone (P < 0.0001).

IRIF persistence and senescence induction were verified in vivo by treating mice bearing MCF7^GFP-IBD tumor xenografts with pitavastatin and then irradiating the tumors with 6 Gy (Fig. 1E; Supplementary Fig. S1E and S1F). Pitavastatin at 10 mg/kg markedly increased tissue destruction and SA-β-Gal activity after 6 Gy, a pattern similar to treatment with 10 mg/kg of the PARP inhibitor veliparib. While increasing pitavastatin to 20 mg/kg resulted in tumor tissue damage and increased SA-β-Gal activity on its own, the higher dose further potentiated the effects of radiation. While statins are taken continuously for lipid lowering, radiosensitization may require only transient treatment. Here, effects were observed with treatment limited to 2 days before, the day of irradiation and 2 days after.

To confirm that the effects of pitavastatin on IRIF persistence were mediated by inhibition of mevalonate biosynthesis, cells were treated with mevalonic acid, the metabolic intermediate produced by HMG-CoA reductase. Mevalonic acid (5 mmol/L) was sufficient to restore IRIF resolution to MCF7^GFP-IBD cells treated with pitavastatin (Supplementary Fig. S2). In the absence of pitavastatin, while 5 mmol/L mevalonic acid had no effect, 10 mmol/L induced a small but significant acceleration of IRIF resolution.

Prior reports have ascribed greater anticancer activity to lipophilic statins (e.g., ref. 36), reflecting their greater effects on peripheral tissue versus the hepatoselectivity of hydrophilic statins. Thus, we compared the lipophilic statins pitavastatin, lovastatin, and simvastatin to the hydrophilic statins pravastatin, atorvastatin, and rosuvastatin for effects on DNA damage and DNA repair. None of the statins increased DNA damage on their own at concentrations up to 10 μmol/L. Compared with DMSO control, the three lipophilic statins, pitavastatin, lovastatin, and simvastatin, each induced IRIF persistence (from 11.1 ± 0.6 foci per nucleus to 16.6 ± 0.9, P < 0.0001, 17.1 ± 0.9, P < 0.0001, and 14.5 ± 0.8, P < 0.001, respectively) and increased % SA-β-Gal–positive cells (from 21 ± 1.1 to 39 ± 5, P < 0.0001, 44 ± 1.5, P < 0.0001 and 45 ± 1.7, P < 0.0001). Hydrophilic statins pravastatin, atorvastatin, and rosuvastatin each failed to potentiate IR with respect to IRIF persistence or senescence induction (Fig. 2A–C).

To confirm the on-target activity of statins and identify the critical step(s) downstream of HMG-CoA reductase (HMGCR) that affect DNA repair, we applied shRNA to silence mevalonate pathway enzymes involved in biosynthesis of both cholesterol and isoprenoid intermediates (Fig. 3A). Stable MCF7^GFP-IBD cell lines were obtained for two shRNAs for each target and scrambled control. Silencing of targeted protein was validated by Western blot analysis (Supplementary Fig. S3). Much like treatment with pitavastatin, silencing HMGCR resulted in persistent DNA DSBs at 24 hours after IR as determined by neutral comet assay (Fig. 3B and C). These cells also displayed increased IRIF persistence and senescence induction (Fig. 3D and G). MCF7^GFP-IBD cells expressing shRNA targeting FDPS, the enzyme required for synthesis of FPP and GGPP and thus both types of prenylation, displayed a similar phenotype to HMGCR silencing. The FDPS knockdown cells displayed elevated DNA damage, IRIF persistence, and accelerated senescence after 6-Gy irradiation (Fig. 3B–D and G).

To distinguish which form of prenylation most impacts radiation sensitivity, we compared MCF7^GFP-IBD cells expressing shRNAs silencing FNTB, farnesyl diphosphate transferase beta, or GGPS1, geranylgeranyl diphosphate synthase 1. To confirm inhibition of prenylation inhibition in pitavastatin-treated or shRNA-targeted cells, we monitored the modification states of Hdj2 and Rap1, proteins known to be farnesylated and geranylgeranylated, respectively (37, 38). Pitavastatin treatment or HMGCR silencing increased levels of unprenylated forms of both the Hdj2 and Rap1 proteins (Supplementary Methods; Supplementary Fig. S4). As expected, FNTB targeting increased unprenylated Hdj2, whereas GGPS1 silencing increased unprenylated Rap1.

Cells expressing shRNA targeting FNTB displayed delayed DSB repair, persistent IRIF, and senescence induction after 6 Gy (Fig. 3E–G). Further implicating farnesylation as critical for DNA repair, farnesyl pyrophosphate restored DSB resolution in cells treated with pitavastatin (Supplementary Fig. S5). In contrast to FTNB, silencing GGPS1 to block protein geranylgeranylation failed to delay DSB repair. Effects on IRIF persistence were equivocal, but GGPS1 silencing increased SA-β-Gal–positive senescent cells over scrambled control (Supplementary Fig. S6), suggesting distinct roles for farnesylated and geranylgeranylated proteins in DNA damage response.

To extend our results to a more physiologic model, we examined radiosensitization by pitavastatin in the highly radioresistant B16.SIY transplantable murine melanoma cell line. Much like with the MCF7^GFP-IBD human tumor cell line, we observed a dose-dependent effect of pitavastatin on colony formation in B16.SIY cells (Fig. 4A and B). The SF2 value was 0.88 for IR alone and 0.7 when combined with 2.5 μmol/L pitavastatin (P < 0.005), yielding a dose modification factor (DMF) of 1.4, similar to that of the positive control veliparib (Supplementary Fig. S7). Clonogenic assays take into account different forms of cell death or proliferative arrest, including apoptosis, necrosis, mitotic catastrophe, and senescence. To determine the contribution of cell death in pitavastatin-mediated radiosensitization, we performed cell viability assays at 48 hours after treatment. At 2.5 μmol/L, pitavastatin did not induce necrosis or apoptosis in B16.SIY cells nor increase cell death after irradiation (Supplementary Methods; Supplementary Fig. S8).

Pitavastatin significantly delayed DSB repair in B16.SIY cells after 6 Gy (P < 0.0001), with persistent damage reaching that of 12 Gy or 6 Gy plus 10 μmol/L veliparib (Fig. 4C and D). Similarly, pitavastatin enhanced B16.SIY senescence over radiation alone (Fig. 4E). Compared with 17 ± 1% SA-β-Gal–positive cells formed by IR alone, combining 6 Gy with 1.0, 2.5, or 5.0 μmol/L pitavastatin yielded 66% ± 2.3%, P < 0.0001, 67% ± 2.2%, P < 0.0001, and 71% ± 2%, P < 0.0001 respectively.

To examine radiosensitization in vivo, B16.SIY tumors were established and treated with pitavastatin, veliparib, or a single dose of 15 Gy, or either drug in combination with IR. Pitavastatin and veliparib were applied 2 days before, the day of irradiation and the 2 days after. Tumors were excised at day 7 after the 15-Gy irradiation and analyzed by H&E for tissue integrity, immunofluorescence for γH2AX to detect persistent DNA damage, IHC for Ki-67 to assess cell proliferation, and SA-β-Gal to evaluate therapy-induced senescence. Imaging revealed treatment with either pitavastatin or veliparib along with 15 Gy displayed increased tumor tissue destruction, increased DNA damage, decreased cellular proliferation, and enhanced senescence (Fig. 5A). Neither pitavastatin nor veliparib at a dose of 10 mg/kg appeared to affect tumor growth on their own. Consistent with the low expression of the proliferation marker Ki-67 at day 7, treatment with either pitavastatin or veliparib along with 15 Gy also conferred a significant delay in tumor growth determined at day 12, compared with untreated control, radiation, or either drug alone (P < 0.01; Fig. 5B and C).

The first applications of ionizing radiation to cancer therapy began within months of Roentgen's report of X-ray imaging in 1895. While efforts to control and focus the delivered dose began early on, until recent decades, significant toxicities due to exposure of normal tissue could not be avoided. The prevailing strategy to limit normal tissue toxicity has been to “fractionate” radiotherapy by delivering small daily doses of 1.8–2.5 Gy/day. However, dramatic progress in computer-controlled radiation delivery has allowed tumors deep in the body to be treated with one or a few ablative doses of up to 25 Gy that produce dense DNA damage in the tumor while sparing surrounding normal tissue. Nonetheless, even with the most advanced tools, while tumor growth can often be delayed by radiation, local recurrence remains common. Furthermore, high doses can only safely be delivered to small tumors. That precise delivery may not be sufficient on its own provides a rationale for developing radiosensitizers that can enhance genotoxic effects in the targeted field without increasing off-target toxicities. Given the positive relationship between radiation density and formation of DNA DSBs, targeting repair of these potentially lethal lesions offers an attractive strategy.

Although molecularly targeted small-molecule DSB repair inhibitors have been pursued for several decades, progress to the clinic has been slow. A proven strategy to accelerate the path from discovery to clinical use is repurposing/repositioning of existing drugs (16, 17, 39, 40). This may have particular value for discovery of radiosensitizers, based on identifying agents with well-established safety profiles that display previously unrecognized capacity to block DSB repair. To screen small molecules for effects on DSB persistence, we have used a GFP fusion to the IRIF-binding domain (IBD) of 53BP1 protein. GFP-IBD binds sites of H2AX phosphorylation to mark the chromatin domains that form around DSBs and then disperses upon DSB repair and H2AX dephosphorylation. Screening libraries of approved and investigational drugs, natural products, neutraceuticals, and other small molecules identified numerous hits, including the HMG-CoA reductase inhibitors pitavastatin and cerivastatin (18).

Patients typically receive HMG-CoA reductase inhibitors to lower risks of cardiovascular disease. The cardiovascular benefits of statins are not ascribed solely to blocking cholesterol biosynthesis. Via inhibition of isoprenoid biosynthesis, statins also display pleiotropic effects mediated by decreased protein prenylation, primarily affecting small GTPases such as Rac and Rho (41). The prevalence of statin use among cancer patients has faciliated retrospective studies examining the potential impacts of HMG-CoA reductase inhibition on cancer incidence, progression, and response to therapy. The convincing data for longer disease-free and/or overall survival after radiotherapy reported in prostate cancer (29, 30) may reflect the role for dysregulated cholesterol metabolism in this disease (42, 43) rather than a specific interaction with the DNA damage response. However, laboratory and clinical studies have reported radio- and/or chemosensitization in multiple cancers. Here, statin's effects may be linked to the DNA damage response, potentially via affecting activity of small GTPases and other prenylated proteins such as lamins.

Our data support direct effects on DSB repair mediated by reduced prenylation. We observed persistent DNA damage, loss of clonogenic survival, and increased senescence when cells were treated with pitavastatin or other lipophilic statins along with radiation. Confirming an on-target effect, silencing the HMGCR HMG-CoA reductase or FDPS farnesyl diphosphate synthase recapitulated the effects of statins. In turn, the effects of pitavastatin could be suppressed by feeding cells mevalonic acid or farnesyl-pyrophosphate. Conversely, silencing the FNTB farnesyl transferase delayed DSB repair, while knockdown of the GGPS1 geranylgeranyl phosphate synthase had no such effect. Taken together these observations suggest, at least in part, that decreased farnesylation of protein(s) may contribute to the effect of statin treatment on DNA repair. These data suggest new strategies and targets to develop radiosensitizers. A more complex pattern was observed with respect to the increased cellular senescence after radiation upon HMG-CoA reductase inhibition. Here, both FNTB and GGPS1 knockdown potentiated therapy-induced senescence after irradiation, suggesting roles for both farnesylated and geranylgeranylated proteins in this process.

To evaluate the impacts of HMG-CoA reductase inhibition on DNA damage response in vivo, we examined pitavastatin's effects on radiation response of tumors formed by the radioresistant melanoma cell line B16.SIY in syngeneic C57BL/6 mice. Pitavastatin had minimal effects on its own but strongly potentiated the effects of radiation. Combination therapy enhanced persistent DNA damage, decreased proliferation, increased senescence, and promoted tumor tissue destruction, leading to significant tumor growth delay compared with radiation alone.

Therapy-induced senescence has been proposed as a beneficial outcome of cancer therapy (3, 4). Damaged cells that survive genotoxic stress may enter senescence, thereby preventing a return to proliferation via undergoing terminal growth arrest (44, 45). Senescence may also spread to nearby cells via a bystander effect (46). However, an offsetting concern is that senescent tumor cells may mediate adverse effects via paracrine signaling from the senescence-associated secretory phenotype (SASP) that can promote inflammation and drive tumor cell proliferation, invasion, and metastasis (6, 9, 10). In prior work, we found that the PARP inhibitor veliparib delays DSB repair and promotes senescence after irradiation but modulates the SASP to decrease expression of inflammatory cytokines (12, 35). Strikingly, our results both in vitro and in vivo demonstrated that nontoxic doses of pitavastatin display radiosensitizing properties that are at least equivalent to those of veliparib. Cardiovascular benefits of statins have long been ascribed to their anti-inflammatory properties along with their impact on lipid levels (47). Recently, it was shown that simvastatin can block senescent fibroblast secretion of SASP factors including inflammatory cytokines, an effect mediated by inhibition of protein prenylation (48). Thereby, simvastatin also suppressed the ability of the senescent cells to drive MCF7 breast cancer cell proliferation. The apparent similarities between effects of PARP inhibition and HMG-CoA reductase inhibition when combined with radiation are striking. Both block DSB repair and promote therapy-induced senescence but suppress secretion of inflammatory SASP factors. Nonetheless, given their distinct molecular targets and limited evidence of crosstalk between statins and PARP inhibitors, their common activities may not derive from a shared mechanism.

Taken together, our studies identify pitavastatin and other lipophilic statins as promising candidates for repurposing as nontoxic radiosensitizers. Our results are consistent with retrospective studies in cancer patients (e.g., refs. 29–32) and provide a mechanistic basis for these prior observations. Our data also suggest that the type of statin and its dose may be critical, potentially explaining the wide range of effects observed between patients and across studies. Supporting feasibility of translation to the clinic, the pitavastatin dose required to demonstrate radiosensitization in vitro was over 100-fold lower than peak plasma concentrations of pitavastatin reached by humans taking standard doses. Interestingly, both lipophilic and hydrophilic statins have been shown to decrease normal tissue damage after radiotherapy without protecting tumors (e.g., refs. 49–51). Thus, together with prior studies, our results argue for repurposing HMG-CoA reductase inhibitors, such as pitavastatin or other lipophilic statins, as agents to enhance the therapeutic ratio of image-guided radiotherapy. Translation of these observations via a clinical trial in a favorable situation such as treatment of prostate cancer with curative intent appears justified.

No potential conflicts of interest were disclosed.

Conception and design: E.V. Efimova, E. Labay, S.J. Kron

Development of methodology: E.V. Efimova, N. Ricco, E. Labay, S.J. Kron

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E.V. Efimova, N. Ricco, E. Labay, H.J. Mauceri, A.C. Flor, A. Ramamurthy, H.G. Sutton, S.J. Kron

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E.V. Efimova, N. Ricco, E. Labay, H.J. Mauceri, A.C. Flor, A. Ramamurthy, S.J. Kron

Writing, review, and/or revision of the manuscript: E.V. Efimova, N. Ricco, E. Labay, H.J. Mauceri, A.C. Flor, A. Ramamurthy, R.R. Weichselbaum, S.J. Kron

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.C. Flor, A. Ramamurthy, R.R. Weichselbaum, S.J. Kron

Study supervision: R.R. Weichselbaum, S.J. Kron

E.V. Efimova, N. Ricco, E. Labay, H.J. Mauceri, A.C. Flor, A. Ramamurthy, H.G. Sutton, R.R. Weichselbaum, and S.J. Kron were funded by grants R01CA164492 and R01CA176843, and support from the University of Chicago Ludwig Center for Metastasis Research (to R.R Weichselbaum).

We thank Rolando Torres for technical support and Shirley Bond in the Integrated Light Microscopy Core for her advice and assistance.

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.