PARP Inhibitors Sensitize Ewing Sarcoma Cells to Temozolomide

Published OnlineFirst October 5, 2015; DOI: 10.1158/1535-7163.MCT-15-0587

Cancer Biology and Signal Transduction Molecular Cancer Therapeutics PARP Inhibitors Sensitize Ewing Sarcoma Cells to Temozolomide-Induced Apoptosis via the Mitochondrial Pathway Florian Engert1, Cornelius Schneider1, Lilly Magdalena Weib1,2,3, Marie Probst1,and Simone Fulda1,2,3

Abstract

Ewing sarcoma has recently been reported to be sensitive to that subsequent to DNA damage-imposed checkpoint activa- poly(ADP)-ribose polymerase (PARP) inhibitors. Searching for tion and G2 cell-cycle arrest, olaparib/TMZ cotreatment causes synergistic drug combinations, we tested several PARP inhibi- downregulation of the antiapoptotic protein MCL-1, followed by tors (talazoparib, niraparib, olaparib, veliparib) together with activation of the proapoptotic proteins BAX and BAK, mitochon- chemotherapeutics. Here, we report that PARP inhibitors syner- drial outer membrane permeabilization (MOMP), activation of gize with temozolomide (TMZ) or SN-38 to induce apoptosis caspases, and caspase-dependent cell death. Overexpression of a and also somewhat enhance the cytotoxicity of doxorubicin, nondegradable MCL-1 mutant or BCL-2, knockdown of NOXA or etoposide, or ifosfamide, whereas actinomycin D and vincris- BAX and BAK, or the caspase inhibitor N-benzyloxycarbonyl-Val- tine show little synergism. Furthermore, triple therapy of ola- Ala-Asp-fluoromethylketone (zVAD.fmk) all significantly reduce parib, TMZ, and SN-38 is significantly more effective compared olaparib/TMZ-mediated apoptosis. These findings emphasize with double or monotherapy. Mechanistic studies revealed that the role of PARP inhibitors for chemosensitization of Ewing the mitochondrial pathway of apoptosis plays a critical role in sarcoma with important implications for further (pre)clinical mediating the synergy of PARP inhibition and TMZ. We show studies. Mol Cancer Ther; 14(12); 2818–30. 2015 AACR.

Introduction accompanied by the release of mitochondrial intermembrane space proteins such as cytochrome c into the cytosol, which in Ewing sarcoma is the second most common pediatric primary turn results in caspase activation and apoptosis (5). MOMP is bone tumor in children and young adults (1). State-of-the-art tightly controlled by various factors including proteins of the BCL- treatment options for Ewing sarcoma patients include chemo- 2 family. BCL-2 family proteins consist of both antiapoptotic therapy consisting of vincristine, ifosfamide, doxorubicin, temo- members, for example BCL-2, BCL-x , and MCL-1, and proapop- zolomide (TMZ), actinomycin D, irinotecan, etoposide, and L totic molecules such as the multidomain proteins BAK and BAX cyclophosphamide (2). Despite aggressive treatment protocols, and BH3-only domain proteins, for example, BIM, BMF, PUMA, patients especially with advanced or relapsed diseases still harbor and NOXA (6). an overall poor survival (2). This highlights the unmet medical Characteristic for Ewing sarcoma is the existence of a chimeric need to develop innovative treatment strategies. fusion protein which is in >85% of cases generated by the Tumor response to cytotoxic therapies including chemotherapy translocation of the EWS gene on chromosome 22 and the FLI1 critically relies on intact cell death programs in cancer cells, as gene on chromosome 11 (11;22)(q24;q12) (7). This fusion chemotherapeutics exert their anticancer effects to a large extent protein encodes the chimeric transcription factor EWS-FLI1 that by engaging programmed cell death (3). Apoptosis is mediated causes upregulation of a range of target genes including PARP1 via two well-deﬁned pathways, that is the extrinsic (death recep- (7, 8), which promotes transcriptional activation by EWS-FLI1 in tor) pathway and the intrinsic (mitochondrial) pathway, which a positive feedback loop (9). Elevated PARP levels have been both eventually lead to activation of caspases as cell death effector detected in Ewing sarcoma specimens (8, 10), indicating that molecules (4). Engagement of the mitochondrial pathway results PARP may represent a promising target for therapeutic exploita- in mitochondrial outer membrane permeabilization (MOMP) tion in Ewing sarcoma. Indeed, Ewing sarcoma cells have been shown to be particularly sensitive to PARP inhibition (9, 10). 1Institute for Experimental Cancer Research in Pediatrics, Goethe- Several PARP inhibitors including talazoparib (BMN-673), nir- University, Frankfurt, Germany. 2German Cancer Consortium (DKTK), aparib (MK-4827), olaparib (AZD-2281), and veliparib (ABT- 3 Heidelberg, Germany. German Cancer Research Center (DKFZ), Hei- 888) are currently under investigation (11). delberg, Germany. Despite the sensitivity of Ewing sarcoma cells to PARP inhibi- Note: Supplementary data for this article are available at Molecular Cancer tors in vitro (9, 10, 12), PARP inhibitors as single therapy displayed Therapeutics Online (http://mct.aacrjournals.org/). limited efﬁcacy in xenograft models of Ewing sarcoma and in Corresponding Author: Simone Fulda, Goethe-University, Komturstr. 3a, 60528 initial clinical trials (12–15), highlighting the need for combina- Frankfurt, Germany. Phone: 49-69-67866557; Fax: 49-69-6786659157; E-mail: tion therapies. PARP inhibitors have been documented to poten- [email protected] tiate DNA damage-mediated cytotoxicity caused in particular by doi: 10.1158/1535-7163.MCT-15-0587 topoisomerase I poisons, DNA methylating agents, or radiation in 2015 American Association for Cancer Research. several cancers including Ewing sarcoma, which has been related

2818 Mol Cancer Ther; 14(12) December 2015

Synergistic Apoptosis by PARP Inhibitors and Temozolomide

A A4573 SK-ES-1

Talazoparib Talazoparib Niraparib Niraparib Olaparib Olaparib Veliparib Veliparib 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30

Cell viability (% (% of viability control) Cell 20 Cell viability (% (% of viability Cell control) 20 10 10 0 0 0.1 1 10 100 1,000 10,000 100,000 0.1 1 10 100 1,000 10,000 100,000 Drug concentration (nmol/L) Drug concentration (nmol/L) B

Talazoparib (nmol/L) Niraparib (nmol/L) Olaparib (nmol/L) Veliparib (nmol/L) A4573 CI values 0.5 1 2.5 5 7.5 25 50 75 100 300 600 1,600 2,200 2,800 25 < 0.1 Very strong synergism TMZ (µmol/L) 50 0.1–0.3 Strong synergism 100 0.3–0.7 Synergism 0.2 0.7–0.85 Moderate synergism SN38 (nmol/L) 0.4 0.85–0.9 Slight synergism 0.6 0.9–1.1 Nearly additive 25 1.1–1.2 Slightly antagonistic Etoposide (ng/mL) 50 1.2–1.45 Moderate antagonistic 75 >1.45 Antagonism 0.5 Not calculated Ifosfamide (µmol/L) 1 2 1 Doxo (ng/µL) 2.5 4 0.1 Vincristine (nmol/L) 0.25 0.4 0.1 AMD 0.25 (ng/mL) 0.4

Talazoparib (nmol/L) Niraparib (nmol/L) Olaparib (nmol/L) Veliparib (nmol/L) SK-ES-1 CI values 0.5 1 2.5 5 7.5 25 50 75 100 300 600 1,600 2,200 2,800 25 < 0.1 Very strong synergism TMZ (µmol/L) 50 0.1–0.3 Strong synergism 100 0.3–0.7 Synergism 0.2 0.7–0.85 Moderate synergism SN38 (nmol/L) 0.4 0.85–0.9 Slight synergism 0.6 0.9–1.1 Nearly additive 25 1.1–1.2 Slightly antagonistic Etoposide (ng/mL) 50 1.2–1.45 Moderate antagonistic 75 >1.45 Antagonism 0.5 Not calculated Ifosfamide (µmol/L) 1 2 1 Doxo (ng/µL) 2.5 4 0.1 Vincristine (nmol/L) 0.25 0.4 0.1 AMD 0.25 (ng/mL) 0.4

Figure 1. Screening for synergistic drug interactions of PARP inhibitors and chemotherapeutic drugs. A, A4573 and SK-ES-1 cells were treated for 72 hours with indicated concentrations of talazoparib, niraparib, olaparib, or veliparib. Cell viability was assessed by crystal violet assay and is shown as percentage of untreated control. B, A4573 and SK-ES-1 cells were treated for 72 hours with different PARP inhibitors (i.e., talazoparib, niraparib, olaparib, and veliparib) in combination with different cytostatic drugs (i.e., TMZ, SN-38, etoposide, ifosfamide, doxorubicin, vincristine, and actinomycin D). Cell viability was assessed by crystal violet assay. Synergistic, additive, or antagonistic drug interactions were calculated by combination index (CI). CI values were then used to generate a heatmap of drug interactions according to the subclassiﬁcation provided by the software's manual.

www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2819

Engert et al.

A4573 SK-ES-1 A 100 100 *** 100 *** 100 *** * 80 80 80 80

60 60 60 60

40 40 40 40

20 20 20 20 DNA fragmentation (%) fragmentation DNA DNA fragmentation (%) fragmentation DNA DNA fragmentation (%) fragmentation DNA DNA fragmentation (%) fragmentation DNA 0 0 0 0 Olaparib +–+– Olaparib +–+– Olaparib +–+– Olaparib +–+– TMZ ++–– SN-38 ++–– TMZ ++–– SN-38 ++–– B ** ** * ** 100 100 100 100 80 80 80 80 60 60 60 60

40 40 40 40

20 20 20 20 Cell viability (% of Cell viability control) Cell viability (% of Cell viability control) Cell viability (% of Cell viability control) 0 0 0 (% of Cell viability control) 0 Olaparib +–+– Olaparib +–+– Olaparib +–+– Olaparib +–+– TMZ ++–– SN-38 ++–– TMZ ++–– SN-38 ++––

C ****** 120 120 120 100

100 100 100 80 80 80 80 60 60 60 60 40 40 40 40

20 20 20 20 Colonies (% Colonies (% of control) Colonies (% Colonies (% of control) Colonies (% Colonies (% of control) Colonies (% Colonies (% of control) 00 0 0 Olaparib +–+– Olaparib +–+– Olaparib +–+– Olaparib +–+– TMZ ++–– SN-38 ++–– TMZ ++–– SN-38 ++––

– Olaparib + Olaparib – Olaparib + Olaparib – Olaparib + Olaparib – Olaparib + Olaparib – TMZ – TMZ – SN-38 – SN-38 + TMZ + TMZ + SN-38 + SN-38

D *** ***

100 100 ** * 80 80

60 60

40 40

20 20 Cell viability (% of Cell viability control) Cell viability (% of Cell viability control)

0 0 Olaparib +++–––+– Olaparib +++–––+– TMZ +–++–+–– TMZ +–++–+–– SN-38 ++–++––– SN-38 ++–++–––

2820 Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics

Synergistic Apoptosis by PARP Inhibitors and Temozolomide

to the role of PARP in the repair of the DNA lesions caused by these violet incorporated by the cells was resolubilized in a solution cytotoxic therapies (9, 13, 16, 17). In addition to impairing single- containing 1% SDS. Absorbance at 550 nm was measured using a strand break (SSB) repair, the anticancer effects of PARP inhibitors microplate reader (Infinite M100; Tecan). Results are expressed as have recently also been attributed to their ability to trap PARP– percentage of cell viability relative to untreated controls. For DNA complexes, thereby generating cytotoxic lesions (18). colony formation assay, cells were treated as indicated for 24 Although different PARP inhibitors were reported to vary in hours. Subsequently, living cells were counted, 100 cells were their potency of PARP-DNA trapping (18, 19), they have not yet reseeded, and cultured in drug-free medium for additional 12 been systematically tested in combination with a range of anti- days before fixation and staining with 0.5% crystal violet, 30% cancer drugs in Ewing sarcoma. In addition, little is yet known on ethanol, and 3% formaldehyde. Colonies were counted cell death signaling pathways that mediate the synergistic inter- macroscopically. action of PARP inhibitors and anticancer drugs. Therefore, in this study, we aimed (i) at examining the effects of different PARP Western blot analysis inhibitors in combination with clinically used chemotherapeutics Western blot analysis was performed as described previously in Ewing sarcoma cells to identify the most potent synergistic drug (20) using the following antibodies: BAK, BCL-2, BCL-xL (BD combinations and (ii) at elucidating the molecular mechanisms Biosciences), pChk1, pChk2, Chk1, Chk2, caspase-3, caspase-9, of synergy with a specific focus on cell death pathways. PARP, BIM, PUMA (Cell Signaling), BAX NT, pH 3, a-tubulin (Millipore), BMF (Novus Biologicals) MCL-1, NOXA, caspase-8 Materials and Methods (Enzo Life Science), GAPDH (HyTest), murine BCL-2 (Life Tech- nologies, Inc.). Goat antimouse IgG and goat anti-rabbit IgG Cell culture and chemicals conjugated to horseradish peroxidase (Santa Cruz Biotechnology) Ewing sarcoma cell lines were kindly provided by C. Roessig were used as secondary antibodies and enhanced chemilumines- (Muenster, Germany) or obtained from German Collection of cence (Amersham Bioscience) or infrared dye-labeled secondary Microorganisms and Cell Cultures (Braunschweig, Germany) or antibodies and infrared imaging were used for detection (Odyssey American Type Culture Collection (Manassas, VA, USA) in 2013. Imaging System; LI-COR Bioscience). Representative blots of at Cells were maintained in DMEM GlutaMAX-l or RPMI medium least two independent experiments are shown. (Life Technologies, Inc.), supplemented with 10% fetal calf serum (FCS), 1% penicillin/streptomycin, 1 mmol/L sodium pyruvate (all from Life Technologies, Inc.). Cell lines were authenticated by Cell-cycle analysis STR profiles and regularly tested for mycoplasma contamination. Cells were stained with PI as described previously (20). Cell- Talazoparib, olaparib, and veliparib were obtained from Sell- cycle analysis was performed using FlowJo Software (TreeStar eckchem and niraparib from ChemieTek. actinomycin D, doxo- Inc.) following the manufacturer's instructions. rubicin, etoposide, TMZ, SN-38, and vincristine were purchased from Sigma, zVAD.fmk from Bachem, bortezomib from Jansen- RNA interference and overexpression Cilag. 4-Hydroperoxy-ifosfamide (active metabolite of ifosfa- Cells were reversely transfected with 10 nmol/L Silencer Select mide and further designated as ifosfamide) was kindly provided (Life Technologies, Inc.) control siRNA (10 nmol/L, 4390844), by Baxter and TRAIL receptor-2 agonistic antibody lexatumumab NOXA siRNA (10 nmol/L s10708 and s10709), or a combination (ETR2) Human Genome Sciences. Chemicals were purchased of targeting siRNAs (5 nmol/L for BAK, s1880 and s1881; 5 nmol/L from Sigma or Carl Roth unless otherwise indicated. for BAX, s1889 and s1890) using Lipofectamine RNAiMAX reagent (Life Technologies, Inc.) and Opti-MEM medium (Life Technolo- gies, Inc.). After 6 hours of incubation with transfection solution, Determination of apoptosis, cell viability, colony formation, the medium was changed and cells recovered for 48 hours before and mitochondrial membrane potential (MMP) drug treatment. For transient overexpression, Ewing sarcoma cells Apoptosis was determined by flow cytometric analysis (FACS- were transfected with 4 mg of pCMV-Tag 3B plasmid [empty vector; Canto II; BD Biosciences) of DNA fragmentation of propidium MCL-1 4A (S64A/S121A/S159A/T163A)], supplied with Lipofec- iodide (PI)-stained nuclei as described previously (20). Cell tamine 2000 (Life Technologies, Inc.) and selected with 500 mg/mL viability was assessed by 3-(4,5-dimethylthiazol-2-yl)- 2,5-diphe- G418 (Carl Roth). Murine BCL-2 was stably overexpressed by using nyltetrazolium bromide (MTT) assay according to the manufac- lentiviral vectors. Briefly, Phoenix cells were transfected with 20 mg turer's instructions (Roche Diagnostics) or by crystal violet assay of pMSCV plasmid (empty vector, BCL-2) using calcium phosphate using crystal violet solution (0.5% crystal violet, 30% ethanol, 3% transfection. Virus-containing supernatant was collected, sterile- formaldehyde). Plates were then rinsed with water and crystal filtered, and used for spin transduction at 37C in the presence of 8

Figure 2. Olaparib synergizes with TMZ and SN-38 to induce cell death in Ewing sarcoma cells. A and B, A4573 and SK-ES-1 cells were treated for 48 hours with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ and/or 0.6 nmol/L SN-38. Apoptosis was determined by analysis of DNA fragmentation of PI-stained nuclei using ﬂow cytometry (A). Cell viability was measured by MTT assay and is expressed as percentage of untreated control (B). Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.05; , P < 0.01; , P < 0.001. C, A4573 and SK-ES-1 cells were treated with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ and/or 0.6 nmol/L SN-38 for 24 hours, living cells were counted and subsequently 100 cells/well were reseeded in drug-free medium in a six-well plate for additional 12 days. Colony formation was assessed by crystal violet staining and colonies were counted macroscopically. The number of colonies is expressed as percentage of untreated control (top) and representative images are shown (bottom). Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.05; , P < 0.01. D, A4573 and SK-ES-1 cells were treated for 72 hours with 0.3 mmol/L olaparib and/or 10 mmol/L TMZ and/or 0.2 nmol/L SN-38. Cell viability was measured by crystal violet staining and is expressed as percentage of untreated control. Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.05; , P < 0.01; , P < 0.001.

www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2821

Engert et al.

A4573 SK-ES-1

A 100 Control 100 Control Olaparib Olaparib 90 TMZ 90 TMZ Olaparib/TMZ Olaparib/TMZ 80 80 70 70 60 *** 60 *** 50 50 *** *** 40 40 30 *** 30 **

DNA fragmentation (%) 20 *** 20 *** * DNA fragmentation (%) 10 10 0 0 12 15 18 21 24 48 72 12 15 18 21 24 48 72 Time (h) Time (h)

A4573 SK-ES-1 B Olaparib +–+– Olaparib +–+– TMZ ++–– TMZ ++––

pChk1 56 kDa- pChk1 56 kDa-

Chk1 56 kDa- Chk1 - 56 kDa

α-Tubulin - 55 kDa α-Tubulin 55 kDa-

62 kDa- - 62 kDa pChk2 pChk2

Chk2 62 kDa- Chk2 - 62 kDa

GAPDH 36 kDa- α-Tubulin - 55 kDa

** (G2–M) * (G2–M)

C 100 100

80 80

G –M 60 2 60 G2–M

S of cells (%) S G 40 1 40 G1

20 20 Frequency Frequency of cells (%)

0 0 Olaparib – + – + Olaparib – + – + TMZ – – + + TMZ – – + +

A4573 SK-ES-1 D Olaparib –+–+– Olaparib –+–+– TMZ –++–– TMZ –++–– VCR +–––– VCR +––––

pH3 pH3 17 kDa- - 17 kDa

GAPDH 36 kDa- GAPDH - 36 kDa

2822 Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics

Synergistic Apoptosis by PARP Inhibitors and Temozolomide

mg/mL polybrene. Transduced Ewing sarcoma cells were selected PARP inhibitors and anticancer drugs that caused up to 20% with 10 mg/mL blasticidin (Carl Roth). reduction of cell viability when used as single agents compared with untreated control (Supplementary Fig. S1). Synergistic, Determination of BAK and BAX activation or MMP additive, or antagonistic drug interactions were calculated by CI BAK and BAX activation was determined by immunoprecipi- (Supplementary Table S2). CI values were then used to generate a fi fl tation of active conformation by speci c antibodies. Brie y, cells heatmap of drug interactions according to the subclassification were lysed in CHAPS lysis buffer (10 mmol/L HEPES, pH 7.4; 150 provided by the software's manual (Fig. 1B). The strongest syn- mmol/L NaCl; 1% CHAPS). Five hundred micrograms of protein ergism was observed for PARP inhibitors in combination with m were incubated overnight at 4 C with 8 g mouse anti-BAX TMZ; the second best synergistic interaction was found for cotreat- m antibody (clone 6A7; Sigma) or 0.5 g mouse anti-BAK antibody ment with SN-38 (Fig. 1B, Supplementary Fig. S2). We also noted m (AB-1; Calbiochem) and 10 L pan mouse IgG Dynabeads that, among the tested anticancer drugs, talazoparib preferentially (Dako), washed with CHAPS lysis buffer and analyzed by Western synergized with TMZ (Fig. 1B, Supplementary Fig. S2). In addi- blotting using rabbit anti-BAX NT antibody (Millipore) or rabbit tion, PARP inhibitors enhanced doxorubicin-, etoposide-, or anti-BAK antibody (BD Biosciences). Loss of MMP was assessed ifosfamide-induced cytotoxicity in Ewing sarcoma cells, although þ by tetramethylrhodamine methyl ester (TMRM ) staining less consistently (Fig. 1B, Supplementary Fig. S2). By comparison, according to the manufacturer's instructions (Life Technologies the DNA-binding drug actinomycin D and the vinca alkaloid Inc.). vincristine exerted little synergistic effects together with PARP inhibitors (Fig. 1B, Supplementary Fig. S2). Statistical analysis fi t Statistical signi cance was assessed by Student -test (two-tailed Olaparib synergizes with TMZ and SN-38 to induce cell death in distribution, two-sample, equal variance) using Microsoft Excel Ewing sarcoma cells P < P < P < (Microsoft Deutschland GmbH); , 0.05; , 0.01; , Because in our screening approach we identified cotreatment of 0.001. Drug interactions were analyzed by the combination index olaparib with TMZ or SN-38 as the most potent combinations, we (CI) method based on that described by Chou (21) using Calcu- focused our validation experiments on these combinations. ola- fi Syn software (Biosoft). Subclassi cation of CI values according to parib acted together with TMZ or SN-38 to increase DNA frag- < – CalcuSyn's manual was used (CI 0.1 very strong synergism, 0.1 mentation, a typical marker of apoptosis, and to reduce cell – – 0.3 strong synergism, 0.3 0.7 synergism, 0.7 0.85 moderate viability compared to treatment with either agent alone (Fig. – – synergism, 0.85 0.9 slight synergism, 0.9 1.1 nearly additive, 2A and B). Moreover, we extended our studies to two additional – – 1.1 1.2 slightly antagonistic, 1.2 1.45 moderate antagonistic, Ewing sarcoma cell lines. Similarly, olaparib cooperated with TMZ > and CI 1.45 antagonism. or SN-38 to induce apoptosis and to decrease cell viability in TC- 32 and TC-71 cells (Supplementary Fig. S3). To explore whether Results combined treatment with olaparib and TMZ or SN-38 also affects Screening for synergistic drug interactions of PARP inhibitors long-term survival of Ewing sarcoma cells, we performed colony and chemotherapeutic drugs assays. Of note, cotreatment of olaparib together with TMZ or SN- To investigate the sensitivity of Ewing sarcoma against PARP 38 acted in concert to significantly suppress colony formation of inhibitors, we initially tested the dose response to four different Ewing sarcoma cells compared to treatment with olaparib, TMZ, PARP inhibitors using the Ewing sarcoma cell lines A4573 and SK- or SN-38 alone (Fig. 2C). Because TMZ and SN-38 are adminis- ES-1 and subsequently extended our studies also to two addi- tered together in clinical protocols as second-line treatment of tional Ewing sarcoma cell lines. Talazoparib, niraparib, olaparib, Ewing sarcoma (22), we also tested whether the addition of and veliparib showed a differential ability to reduce cell viability olaparib further potentiates the antitumor activity of this chemo- of Ewing sarcoma cells with talazoparib being the most active and therapy regimen. Of note, triple therapy of olaparib, TMZ, and SN- veliparib being the least active compound (talazoparib IC50 10 38 was significantly more effective to reduce cell viability of Ewing nmol/L > niraparib IC50 300 nmol/L > olaparib IC50 1000 sarcoma cells compared to treatment with single agents or to nmol/L > veliparib IC50 10000 nmol/L; Fig. 1A, Supplementary cotreatment with TMZ/SN-38 (Fig. 2D). Together, this set of Table S1). experiments shows that olaparib cooperates together with TMZ To investigate the question whether PARP inhibitors modulate or SN-38 to induce apoptosis, to reduce cell viability, and to chemosensitivity of Ewing sarcoma cells, we screened the efficacy suppress long-term clonogenic survival of Ewing sarcoma cells. To of several anticancer drugs that are used in the clinic for the investigate in more detail the molecular mechanisms underlying treatment of Ewing sarcoma (11) in the presence and absence of the synergism of PARP inhibitors and chemotherapeutic drugs in PARP inhibitors. We used suboptimal drug concentrations of Ewing sarcoma, we focused the subsequent experiments on TMZ,

Figure 3. Olaparib/TMZ cotreatment causes G2 arrest prior to cell death. A, A4573 and SK-ES-1 cells were treated with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ for indicated time. Apoptosis was determined by quantiﬁcation of DNA fragmentation of PI-stained nuclei using ﬂow cytometry. Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.05; , P < 0.01; , P < 0.001. B, A4573 and SK-ES-1 cells were treated with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ for 18 hours. Phosphorylation of checkpoint kinases was assessed by Western blotting. Expression of GAPDH or a-tubulin served as loading controls. Representative blots of two independent experiments are shown. C, A4573 and SK-ES-1 cells were treated for 18 hours with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ. DNA was stained with PI and cell-cycle analysis was performed using FlowJo software. Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.05; , P < 0.01 comparing olaparib/TMZ cotreated to single treated or untreated cells in G2–M phase. D, Ewing sarcoma cells were treated with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ or 2.5 nmol/L vincristine for 18 hours. Expression of mitotic marker pH3 was analyzed by Western blotting. Expression of GAPDH served as loading control. Representative blots of two independent experiments are shown.

www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2823

Engert et al.

A A4573 SK-ES-1 B

Olaparib –+–+–+–+– A4573 TMZ –++––++–– 60 ETR2 Ctrl +––+––––– zVAD.fmk 50 *** Caspase-9 - 47 kDa 40

◄ 37 kDa 30 ◄ 35 kDa 20

Caspase-3 - 32 kDa DNA fragmentation (%)

0 Olaparib/TMZ –– + + ◄ 19 kDa ◄ 17 kDa ◄ 12 kDa SK-ES-1

PARP - 116 kDa 60 Ctrl ◄89 kDa 50 zVAD.fmk *** 40 Caspase-8 - 55 kDa - 53 kDa 30

20 ◄ 43 kDa ◄ 41 kDa 10 * DNA fragmentation (%)

0 GAPDH - 36 kDa Olaparib/TMZ –– + +

C D A4573 SK-ES-1 Olaparib +–+– +–+–

TMZ ++–– ++–– A4573 MCL-1 - 43 kDa 30 Control Olaparib 25 TMZ Olaparib/TMZ BCL-2 - 26 kDa * 20

15 BCL-XL - 28 kDa 10 ** Lossof MMP (%) 5 BIMEL - 23 kDa

0 18 21 24 Time (h)

BIML - 16 kDa BIMS - 13 kDa SK-ES-1

30 Control Olaparib BMF - 25 kDa 25 TMZ -- + + Olaparib/TMZ 20 - 20 kDa ** 15 PUMA 10 * - 23 kDa Loss of MMP (%) 5 NOXA 0 - 11 kDa 18 21 24 Time (h) GAPDH - 36 kDa

2824 Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics

Synergistic Apoptosis by PARP Inhibitors and Temozolomide

because this drug yielded the most pronounced synergistic activation and caspase-dependent apoptosis in Ewing sarcoma interaction. cells.

Olaparib/TMZ cotreatment causes G2 arrest prior to cell death Olaparib/TMZ cotreatment downregulates MCL-1 levels To define the onset of olaparib/TMZ-induced apoptosis we To investigate whether olaparib/TMZ combination treatment performed a kinetic analysis. Olaparib and TMZ cooperated to engages the mitochondrial pathway of apoptosis, we assessed trigger apoptosis in a time-dependent manner starting at about 18 the MMP. Olaparib and TMZ cooperated to trigger loss of MMP hours (Fig. 3A). inatime-dependentmannerattheonsetofapoptosis(Fig.4C). Next, we investigated whether the combination treatment Because BCL-2 family proteins are key regulators of the mito- leads to a DNA damage-related stress response. Olaparib and chondrial pathway, we then asked whether olaparib/TMZ TMZ acted together to trigger phosphorylation of the check- cotreatment causes changes in their expression levels. Of note, point kinases Chk1 and Chk2 at 18 hours pointing to activation cotreatment with olaparib and TMZ resulted in downregulation of DNA damage response pathways prior to cell death (Fig. 3B). of MCL-1, whereas expression levels of BCL-2, BCL-xL,BIM, Analysis of cell-cycle distribution using flow cytometry revealed BMF, and PUMA remained largely unchanged (Fig. 4D). We that olaparib/TMZ cotreatment caused a significant increase of also noted that NOXA levels decreased upon olaparib/TMZ cells in G2–M phase compared with cells treated with either cotreatment (Fig. 4D). drug alone or untreated cells (Fig. 3C, Supplementary Fig. S4). As flow cytometric analysis of cell-cycle distribution does not Olaparib/TMZ cotreatment promotes proteasomal degradation allow to distinguish between G2 or M phase arrest, we addi- of MCL-1 tionally analyzed phosphorylation of histone H3 (pH3) as a To investigate whether the observed downregulation of MCL-1 specific M-phase marker. Olaparib/TMZ cotreatment did not upon olaparib/TMZ cotreatment is due to proteasomal degrada- cause phosphorylation of histone H3 in contrast to the micro- tion and/or caspase-mediated cleavage, we tested the effects of the tubule-interfering drug vincristine that was used as a positive proteasome inhibitor bortezomib and/or the pan-caspase inhib- control for the induction of M-phase arrest (Fig. 3D). Together, itor zVAD.fmk. Although addition of bortezomib significantly these experiments show that olaparib and TMZ induce cell- rescued the olaparib/TMZ-imposed downregulation of MCL-1 cycle arrest in the G2 phase prior to induction of apoptosis in protein levels, addition of zVAD.fmk largely failed to significantly Ewing sarcoma cells. protect against MCL-1 downregulation upon olaparib/TMZ cotreatment (Fig. 5A). This indicates that MCL-1 is degraded via Olaparib/TMZ-induced apoptosis is executed via caspase- the proteasome upon treatment with olaparib/TMZ. dependent effector pathways To elucidate the role of MCL-1 in olaparib/TMZ-mediated cell To explore whether induction of apoptosis involved activation of death we overexpressed a phosphomutant variant of MCL-1 caspases, we performed Western blotting. Olaparib and TMZ acted (MCL-1 4A), which is resistant to phosphorylation of a phos- together to trigger cleavage of caspase-9 into active p37/p35 frag- pho-degron and subsequent proteasomal degradation (Fig. 5B). ments, cleavage of caspase-3 into active p17/p12 fragments, and Notably, ectopic expression of nondegradable MCL-1 mutant cleavage of PARP into p89 fragment. By comparison, little cleavage significantly decreased olaparib/TMZ-induced apoptosis (Fig. of caspase-8 was observed upon olaparib/TMZ combination treat- 5C). Because NOXA is known to specifically antagonize MCL- ment in contrast to treatment with the tumor necrosis factor-related 1, we also knocked down NOXA by RNAi to further investigate the apoptosis-inducing ligand (TRAIL) receptor 2 agonistic antibody involvement of the MCL-1/NOXA axis. NOXA silencing signifi- lexatumumab that served as positive control (Fig. 4A). Expression cantly reduced olaparib/TMZ-induced apoptosis in Ewing sarco- levels of caspase-8 were very low in A4573 cells consistent with ma cells (Fig. 5D and E). Together, these results indicate that previous studies demonstrating that caspase-8 is frequently olaparib/TMZ-triggered degradation of MCL-1 contributes to silenced by epigenetic mechanisms in Ewing sarcoma (23, 24). olaparib/TMZ-induced apoptosis. To test whether caspase activity is required for the induction of apoptosis, we used the broad-range caspase inhibitor zVAD.fmk. Olaparib/TMZ cotreatment promotes BAX/BAK activation and Addition of zVAD.fmk significantly diminished olaparib/TMZ- MOMP induced apoptosis compared with cells that were treated with Next, we explored whether degradation of MCL-1 leads to olaparib/TMZ in the absence of zVAD.fmk (Fig. 4B). These experi- activation of BAK and BAX, two multidomain proapoptotic ments show that olaparib and TMZ cooperate to trigger caspase BCL-2 proteins that control MOMP. Since activation of BAK and

Figure 4. Olaparib and TMZ cooperate to trigger caspase activation and mitochondrial perturbations. A, A4573 and SK-ES-1 cells were treated with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ for 24 hours. Cleavage of caspase-9, caspase-3, caspase-8, and PARP was analyzed by Western blotting. Arrowheads indicate active cleavage fragments, expression of GAPDH served as loading control; asterisk denotes unspecified protein bands. SK-ES-1 cells treated for 2 hours with 2 mg/mL lexatumumab (ETR-2) served as positive control for caspase activation. Representative blots of two independent experiments are shown. B, A4573 and SK-ES-1 cells were treated for 48 hours with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ in the presence or absence of 50 mmol/L zVAD.fmk. Apoptosis was determined by quantification of DNA fragmentation of PI-stained nuclei using flow cytometry. Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.001. C, A4573 and SK-ES-1 cells were treated with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ for indicated times and loss of MMP in the living cell population was determined by flow cytometry using TMRM fluorescent dye. Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.05; , P < 0.01. D, A4573 and SK-ES-1 cells were treated with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ for 18 hours. Expression of BCL-2 family proteins was analyzed by Western blot, and expression of GAPDH served as loading control. Representative blots oftwo independent experiments are shown.

www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2825

Engert et al.

A A4573 SK-ES-1 Olaparib ++++–++++–

TMZ ++++–++++–

zVAD.fmk +–+––+–+–– Bortezomib ++–––++–––

MCL-1 - 43 kDa

GAPDH - 36 kDa

A4573 SK-ES-1 3.0 * 3.0 *** *** * * 2.5 2.5 ** n.s. 2.0 n.s. 2.0

1.5 1.5

1.0 1.0

0.5 0.5 MCL-1 expression (fold change) MCL-1 expression MCL-1 expression (fold change) MCL-1 expression 0.0 0.0 Olaparib – + + + + Olaparib – + + + + TMZ – + + + + TMZ – + + + +

zVAD.fmk – – + – + zVAD.fmk – – + – + Bortezomib – – – + + Bortezomib – – – + +

A4573 SK-ES-1 B A4573 SK-ES-1 D EV MCL-1 4A EV MCL-1 4A

MCL-1 - 43 kDa NOXA - 11 kDa - 55 kDa GAPDH - 36 kDa α-Tubulin

C A4573 E A4573 ** 70 EV MCL-1 4A 50 siCtrl ** siNOXA #1 60 siNOXA #2 40 50

40 30

30 20 20 10 DNA fragmentation (%) fragmentation DNA

10 (%) fragmentation DNA

0 0 Olaparib/TMZ Olaparib/TMZ –– + + ––– + + +

SK-ES-1 SK-ES-1 70 ** EV MCL-1 4A 50 siCtrl siNOXA #1 60 * siNOXA #2 40 50

40 30

30 20 20

DNA fragmentation (%) fragmentation DNA 10

10 (%) fragmentation DNA

0 0 Olaparib/TMZ –– + + Olaparib/TMZ ––– + + +

2826 Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics

Synergistic Apoptosis by PARP Inhibitors and Temozolomide

BAX is accompanied by a conformational change that can be enhanced normal tissue toxicity upon cotreatment with PARP detected by specific antibodies, we immunoprecipitated BAK and inhibitors and TMZ or topoisomerase I inhibitors (12, 16, 25). BAX using conformation-specific antibodies. Combination treat- PARP inhibitors also enhanced doxorubicin-, etoposide-, or ifos- ment with olaparib and TMZ stimulated activation of BAK and famide-induced cytotoxicity in Ewing sarcoma cells, although less BAX (Fig. 6A). To explore the functional relevance of BAK and BAX consistently. By comparison, the cytotoxic antibiotic actinomycin in olaparib/TMZ-induced apoptosis we used a combined siRNA D and the vinca alkaloid vincristine did hardly act in a synergistic approach to concomitantly silence BAK and BAX (Fig. 6B). or additive manner together with PARP inhibitors and also Knockdown of BAK and BAX significantly rescued olaparib/ showed some antagonistic effects, depending on the drug con- TMZ-mediated apoptosis in all different siRNA construct combi- centrations. Antagonistic interactions of olaparib together with nations we used (Fig. 6C), emphasizing the importance of BAK vincristine have previously been described in Ewing sarcoma cells and BAX in olaparib/TMZ-induced apoptosis. and have been linked to vincristine's mechanism of action which is independent of DNA damage (13). In line with the results of our Overexpression of BCL-2 protects against Olaparib/TMZ- systematic testing approach that identified TMZ and SN-38 as the induced apoptosis most suitable anticancer drugs for combinations with PARP To further examine the requirement of the mitochondrial inhibitors in Ewing sarcoma, cooperative interaction of PARP pathway for the synergistic induction of apoptosis by olaparib inhibitors together with TMZ or SN-38 has previously been and TMZ we overexpressed the antiapoptotic protein BCL-2 that is described for Ewing sarcoma cells in vitro and in preclinical Ewing known to block mitochondrial apoptosis (Fig. 6D). Importantly, sarcoma models in vivo (9, 12, 13, 16, 26). fi BCL-2 overexpression signi cantly decreased olaparib/TMZ- The differential ability of chemotherapeutic drugs to synergize induced apoptosis (Fig. 6E). This underscores that the synergistic with PARP inhibitors has been attributed to the type of DNA induction of apoptosis by olaparib and TMZ is mediated via the lesions generated by the anticancer agents that differentially mitochondrial pathway of apoptosis. require PARP1 for DNA repair or affect binding of PARP1 to the DNA (27). In combination with PARP-trapping PARP inhibitors Discussion such as talazoparib, the DNA-methylating agent TMZ elicits its PARP is currently considered as a promising target for thera- cytotoxicity through N3 and N7 methyl adducts that induce base peutic exploitation in Ewing sarcoma. Therefore, the two aims of excision repair. Single-strand breaks generated during base exci- this study were (i) to examine the effects of different PARP sion repair are cytotoxic via PARP trapping, as PARP1 binds to and inhibitors in combination with a range of commonly used che- is activated by these base excision repair intermediates, and poly motherapeutics to identify the best synergistic combinations in (ADP-ribos)ylation is important for efficient repair (27). Consis- Ewing sarcoma cells and (ii) to elucidate the molecular mechan- tently, we show that, among the tested anticancer drugs, talazo- isms of synergy with a specific focus on cell death pathways. parib, a potent PARP-trapping agent, preferentially synergizes Here, we report that different PARP inhibitors synergized in with TMZ, confirming that PARP inhibitors with high PARP- particular together with TMZ and also with SN-38 to reduce cell DNA-trapping activity are especially effective together with TMZ viability and to trigger cell death in Ewing sarcoma cells. Calcu- (28). By comparison, TMZ as single agent induces cytotoxicity lation of CI values confirmed that the combinations of PARP mainly through O6 methylation followed by futile cycles of inhibitors together with TMZ or SN-38 were the best and second mismatch repair, pointing to different mechanisms of cytotoxicity best combinations among the tested chemotherapeutics. Further- for TMZ monotherapy or in combination with PARP inhibitors. more, triple therapy of olaparib, TMZ, and SN-38 proved to be Topoisomerase I inhibitors such as SN-38 cause DNA SSBs and significantly more effective to reduce cell viability of Ewing covalent topoisomerase I–DNA complexes, which trigger PARP sarcoma cells compared to cotreatment with TMZ/SN-38 or activation and require PARP1 for repair (29). Accordingly, cata- single-agent treatment. These findings may have important clin- lytic PARP inhibitors have shown highly synergistic effects in ical implications, because topoisomerase I poisons and DNA combination with topoisomerase I inhibitors (28). Consistently, methylating agents such as TMZ are used in second-line chemo- we demonstrate that potent catalytic PARP inhibitors such as therapy regimens for Ewing sarcoma (22). However, increased niraparib and olaparib cause synergistic cytotoxicity together with host toxicity may limit the therapeutic window for such combina- SN-38. Olaparib has previously been reported to act in concert tions, because there is preclinical and clinical evidence for with doxorubicin, melphalan, and carboplatin in Ewing sarcoma

Figure 5. Olaparib/TMZ cotreatment promotes proteasomal degradation of MCL-1. A, A4573 and SK-ES-1 cells were treated for 18 hours with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ and/or 50 mmol/L zVAD.fmk and/or 5 ng/mL bortezomib. Expression of MCL-1 was analyzed by Western blotting, and expression of GAPDH served as loading control. For further evaluation, Western blots were quantified using ImageJ software and changes in MCL-1 protein levels are given as fold change in comparison to untreated control. Data are shown as mean SD of three independent Western blots; , P < 0.05; , P < 0.01; , P < 0.001; n.s., not significant. B and C, A4573 and SK-ES-1 cells were transfected with nondegradable phospho-defective mutant of MCL-1 (MCL-1 4A) or empty vector (EV). Expression of MCL-1 was analyzed by Western blotting, and expression of GAPDH served as loading control; arrow indicates exogenously expressed MCL-1 (B; representative blots of two independent experiments are shown). Cells were treated for 48 hours with 0.3 mmol/L olaparib and 50 mmol/L TMZ and apoptosis was determined by quantification of DNA fragmentation of PI-stained nuclei using flow cytometry (C). Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.05; , P < 0.01. D and E, A4573 and SK-ES-1 cells were transiently transfected with 10 nmol/L nonsilencing siRNA or two different constructs targeting NOXA. Expression of NOXA was analyzed by Western blotting, and a-tubulin served as loading control; representative blots of two independent experiments are shown (D). Transiently transfected Ewing sarcoma cells were treated for 24 hours with 0.3 mmol/L olaparib and 50 mmol/L TMZ and apoptosis was determined by quantification of DNA fragmentation of PI-stained nuclei using flow cytometry. Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.01.

www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2827

Engert et al.

A A4573 SK-ES-1

Olaparib +–+– Olaparib +–+–

TMZ ++–– TMZ ++––

BAKactive - 24 kDa BAKactive - 24 kDa

BAK - 24 kDa BAK - 24 kDa

GAPDH - 36 kDa GAPDH - 36 kDa

BAXactive - 21 kDa BAXactive - 21 kDa

BAX - 21 kDa BAX - 21 kDa

GAPDH - 36 kDa GAPDH - 36 kDa

B A4573 SK-ES-1

BAK BAK - 24 kDa - 24 kDa

BAX - 21 kDa BAX - 21 kDa

GAPDH - kDa36 GAPDH - kDa36

C A4573 SK-ES-1

siCtrl siCtrl 50 50 siBak #1 + siBax #2 * siBak #1 + siBax #2 45 siBak #1 + siBax #3 45 siBak #1 + siBax #3 * siBak #2 + siBax #2 siBak #2 + siBax #2 40 siBak #2 + siBax #3 40 siBak #2 + siBax #3 35 35 30 30 25 25 20 20 15 15 10 10 DNA fragmentation (%) fragmentation DNA (%) fragmentation DNA 5 5 0 0 Olaparib – + – + Olaparib – + – + TMZ – – + + TMZ – – + +

D A4573 E A4573 SK-ES-1 EV BCL-2 70 70 *** EV BCL-2 EV BCL-2 BCL-2 - 24 kDa 60 60 50 50 α-Tubulin - 55 kDa 40 *** 40 SK-ES-1 30 30

EV BCL-2 20 20 10 10 - 24 kDa BCL-2 (%) fragmentation DNA (%) fragmentation DNA 0 0 – + – + – – + + - 36 kDa GAPDH – – + + – + – +

2828 Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics

Synergistic Apoptosis by PARP Inhibitors and Temozolomide

cells (13), whereas no cooperative interaction was reported in the i.e., Mule, SCF/b-TrCP, SCF/Fbw7, and Trim17 (32). Mule is past for the combination of etoposide with the first-generation considered to be responsible for constitutive MCL-1 degradation, PARP inhibitor NU1025 (30). whereas ubiquitination of MCL-1 by SCF/b-TrCP, SCF/Fbw7, and In addition to the identification of synergistic PARP inhibitor- Trim17 is mediated via phosphorylation of a phosphodegron site based combination therapies, our study demonstrates marked by several kinases, for example by JNK and GSK3 in interphase or differences in the single-agent cytotoxicity of the four PARP inhi- postmitotic cells (32). Furthermore, binding of NOXA to MCL-1 bitors in Ewing sarcoma cells with decreasing potency in the order has been reported to induce MCL-1 degradation by the protea- of talazoparib (IC50 10 nmol/L) > niraparib (IC50 300 nmol/ some (33), which may also explain the downregulation of NOXA L) > olaparib (IC50 1,000 nmol/L) > veliparib (IC50 10,000 that we observed upon prolonged treatment with olaparib/TMZ. nmol/L). These results are in line with previous studies showing During the initial induction phase of apoptosis, constitutive differential antitumor activity of PARP inhibitors as single agents expression of NOXA turned out to be necessary for olaparib/ in various tumors including Ewing sarcoma (18, 28). The antitu- TMZ-induced cell death, because knockdown of NOXA signifi- mor activity of PARP inhibitors has been shown to depend on both cantly reduced olaparib/TMZ-mediated cell death. Moreover, we catalytic PARP inhibition and PARP-DNA trapping (18). Consis- show that expression of a nondegradable phosphomutant variant tent with our findings showing that talazoparib exerts the highest of MCL-1 significantly reduces olaparib/TMZ-mediated apopto- single-agent cytotoxicity of the tested PARP inhibitors, the potency sis, emphasizing that degradation of MCL-1 contributes to ola- of talazoparib has been attributed to its high efficiency at trapping parib/TMZ-induced apoptosis. PARP–DNA complexes (19). Third, rescue experiments demonstrate that activation of BAX Our molecular studies revealed that the mitochondrial path- and BAK upon olaparib/TMZ cotreatment is required for the way of apoptosis plays a critical role in mediating the synergy synergistic induction of apoptosis, because genetic silencing of of concomitant PARP inhibition (using olaparib as a prototypic BAX and BAK significantly protects Ewing sarcoma cells from PARP inhibitor) and TMZ used as a prototypic DNA-damaging olaparib/TMZ-induced apoptosis. Fourth, the importance of the agent. This is of particular importance, as to our knowledge mitochondrial pathway of apoptosis is underscored by overex- the downstream signaling events mediating synergistic cell pression of the antiapoptotic protein BCL-2, which is known to death induction by PARP inhibition together with TMZ have block mitochondria-mediated apoptosis, because BCL-2 over- so far remained largely elusive. We show that subsequent to expression inhibits olaparib/TMZ-induced apoptosis. Fifth, ola- DNA damage-imposed checkpoint activation and G2 cell-cycle parib and TMZ act together to trigger cleavage of caspases into arrest, olaparib/TMZ cotreatment causes downregulation of the their active fragments. The requirement of caspases for the induc- antiapoptotic protein MCL-1, activation of the proapoptotic tion of cell death is demonstrated by using the broad-range proteins BAX and BAK, MOMP, proteolytic activation of cas- caspase inhibitor zVAD.fmk, which significantly rescues ola- pases, and caspase-dependent cell death. This conclusion is parib/TMZ-induced apoptosis. underscored by the following independent pieces of experi- Our study has important implications for the development of mental evidence. PARP inhibitor-based therapies in Ewing sarcoma. Despite the fact First, olaparib/TMZ cotreatment triggers activation of the DNA that PARP inhibitors as single agents exerted promising cytotoxicity damage response, indicated by increased phosphorylation of the against Ewing sarcoma cells in vitro (10), they exerted limited checkpoint kinases Chk1 and Chk2. In response to DNA damage, efficacy in xenograft models of Ewing sarcoma and did not elicit ATM and ATR phosphorylate and thereby activate Chk1 and objective responses in patients with Ewing sarcoma in early clinical Chk2, which control the G2–M checkpoint (31). Consistently, trials (12–15). This underscores the need for combination strate- olaparib/TMZ cotreatment results in cell-cycle arrest in G2 prior to gies with PARP inhibitors, for example together with chemo- cell death induction. Second, our rescue experiments showing that or radiotherapy. Phase I clinical trials combining PARP inhibitors addition of the proteasome inhibitor bortezomib rescues cells (i.e., olaparib, niraparib, or talazoparib) together with TMZ for from olaparib/TMZ-imposed downregulation of MCL-1 protein the treatment of Ewing sarcoma are currently under way (i.e., indicate that MCL-1 is degraded via the proteasome upon treat- NCT01858168, NCT02044120, NCT02116777). Our present ment with olaparib/TMZ. MCL-1 expression is tightly regulated at study emphasizes the role of PARP inhibitors for chemosensitiza- multiple levels, involving ubiquitination of MCL-1 that targets it tion of Ewing sarcoma. By showing that PARP inhibitors synergis- for proteasomal degradation (32). Four different E3 ubiquitin- tically induce apoptosis together with several DNA-damaging ligases that mediate MCL-1 ubiquitination have been identified, agents, most notably TMZ and also with SN38, it provides new

Figure 6. Olaparib/TMZ cotreatment promotes BAX/BAK activation and MOMP. A, A4573 and SK-ES-1 cells were treated with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ for 21 hours. Active conformations of BAK or BAX were immunoprecipitated using active conformation-specific antibodies and were analyzed by Western blotting. Expression of total BAK or BAX and GAPDH served as loading controls. Representative blots of two independent experiments are shown. B and C, A4573 and SK-ES-1 cells were transiently transfected with 10 nmol/L nonsilencing siRNA or 5 nmol/L each of different combinations of constructs targeting BAK or BAX and expression of BAK and BAX was analyzed by Western blotting (B). GAPDH served as loading control. Representative blots of two independent experiments are shown. Transiently transfected Ewing sarcoma cells were treated for 24 hours with 0.3 mmol/L olaparib and/or 50 mmol/L TMZ and apoptosis was determined by quantification of DNA fragmentation of PI-stained nuclei using flow cytometry (C). Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.05. D and E, A4573 and SK-ES-1 cells were transfected with a murine BCL-2 construct or empty vector and BCL-2 expression was analyzed by Western blotting (D). Expression of GAPDH or a-tubulin served as loading controls. Representative blots of two independent experiments are shown. BCL-2-overexpressing Ewing sarcoma cells were treated for 48 hours with 0.3 mmol/L olaparib and 50 mmol/L TMZ and apoptosis was determined by quantification of DNA fragmentation of PI-stained nuclei using flow cytometry (E). Data are shown as mean SD of three independent experiments performed in triplicate; , P < 0.001.

www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2829

Engert et al.

insights into cotreatment rationales for distinct PARP inhibitors Writing, review, and/or revision of the manuscript: F. Engert, C. Schneider, and chemotherapeutic drugs and sets the ground for further (pre) L.M. Weib, S. Fulda clinical evaluation of PARP inhibitor-based combination regimens Study supervision: S. Fulda with cytotoxic chemotherapy. Acknowledgments ﬂ The authors thank Human Genome Sciences (Rockville, MD) for kindly Disclosure of Potential Con icts of Interest providing TRAIL-R2 agonistic antibody and C. Hugenberg for expert secretarial ﬂ No potential con icts of interest were disclosed. assistance.

Authors' Contributions Grant Support Conception and design: F. Engert, C. Schneider, S. Fulda This work has been partially supported by a grant from the BMBF (to S. Development of methodology: F. Engert, L.M. Weib Fulda). Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): F. Engert, C. Schneider, M. Probst Analysis and interpretation of data (e.g., statistical analysis, biostatistics, Received July 17, 2015; revised September 22, 2015; accepted September 23, computational analysis): F. Engert, C. Schneider, M. Probst, S. Fulda 2015; published OnlineFirst October 5, 2015.

References 1. Balamuth NJ, Womer RB. Ewing's sarcoma. Lancet Oncol 2010;11:184–92. 18. Murai J, Huang SY, Das BB, Renaud A, Zhang Y, Doroshow JH, et al. 2. Potratz J, Dirksen U, Jurgens H, Craft A. Ewing sarcoma: clinical state-of- Trapping of PARP1 and PARP2 by clinical PARP inhibitors. Cancer Res the-art. Pediatr Hematol Oncol 2012;29:1–11. 2012;72:5588–99. 3. Fulda S. Tumor resistance to apoptosis. Int J Cancer 2009;124:511–5. 19. Murai J, Huang SY, Renaud A, Zhang Y, Ji J, Takeda S, et al. Stereospecific 4. Fulda S, Debatin KM. Extrinsic versus intrinsic apoptosis pathways in PARP trapping by BMN 673 and comparison with olaparib and rucaparib. anticancer chemotherapy. Oncogene 2006;25:4798–811. Mol Cancer Ther 2014;13:433–43. 5. Fulda S, Galluzzi L, Kroemer G. Targeting mitochondria for cancer therapy. 20. Fulda S, Sieverts H, Friesen C, Herr I, Debatin KM. The CD95 (APO-1/Fas) Nat Rev Drug Discov 2010;9:447–64. system mediates drug-induced apoptosis in neuroblastoma cells. Cancer 6. Adams JM, Cory S. The Bcl-2 apoptotic switch in cancer development and Res 1997;57:3823–9. therapy. Oncogene 2007;26:1324–37. 21. Chou TC. The median-effect principle and the combination index for quan- 7. Delattre O, Zucman J, Plougastel B, Desmaze C, Melot T, Peter M, et al. titation of synergism and antagonism. In:Chou TC, editor. Synergism and Gene fusion with an ETS DNA-binding domain caused by chromosome antagonism in chemotherapy. San Diego: Academic Press; 1991. p. 61–102. translocation in human tumours. Nature 1992;359:162–5. 22. Vormoor B, Curtin NJ. Poly(ADP-ribose) polymerase inhibitors in Ewing 8. Soldatenkov VA, Albor A, Patel BK, Dreszer R, Dritschilo A, Notario V. sarcoma. Curr Opin Oncol 2014;26:428–33. Regulation of the human poly(ADP-ribose) polymerase promoter by the 23. Fulda S, Kufer MU, Meyer E, van Valen F, Dockhorn-Dworniczak B, Debatin ETS transcription factor. Oncogene 1999;18:3954–62. KM. Sensitization for death receptor- or drug-induced apoptosis by re- 9. Brenner JC, Feng FY, Han S, Patel S, Goyal SV, Bou-Maroun LM, et al. PARP- expression of caspase-8 through demethylation or gene transfer. Oncogene 1 inhibition as a targeted strategy to treat Ewing's sarcoma. Cancer Res 2001;20:5865–77. 2012;72:1608–13. 24. Kontny HU, Hammerle K, Klein R, Shayan P, Mackall CL, Niemeyer CM. 10. Garnett MJ, Edelman EJ, Heidorn SJ, Greenman CD, Dastur A, Lau KW, Sensitivity of Ewing's sarcoma to TRAIL-induced apoptosis. Cell Death et al. Systematic identification of genomic markers of drug sensitivity in Differ 2001;8:506–14. cancer cells. Nature 2012;483:570–5. 25. Kummar S, Chen A, Ji J, Zhang Y, Reid JM, Ames M, et al. Phase I study of 11. Sonnenblick A, de Azambuja E, Azim HA Jr., Piccart M. An update on PARP PARP inhibitor ABT-888 in combination with topotecan in adults with inhibitors–moving to the adjuvant setting. Nat Rev Clin Oncol 2015; refractory solid tumors and lymphomas. Cancer Res 2011;71:5626–34. 12:27–41. 26. Wilcoxen K, Brooks D, Tiruchinapalli D, Anderson N, Donaldson R, Nivens 12. Smith MA, Reynolds CP, Kang MH, Kolb EA, Gorlick R, Carol H, et al. M, et al. The PARP inhibitor niraparib demonstrates synergy with chemo- Synergistic activity of PARP inhibition by talazoparib (BMN 673) with therapy in treatment of patient derived Ewing's sarcoma tumor graft temozolomide in pediatric cancer models in the pediatric preclinical models. Mol Cancer Ther 2013;12(suppl; abstr A258):11s. testing program. Clin Cancer Res 2015;21:819–32. 27. Horton JK, Wilson SH. Predicting enhanced cell killing through PARP 13. Norris RE, Adamson PC, Nguyen VT, Fox E. Preclinical evaluation of the inhibition. Mol Cancer Res 2013;11:13–8. PARP inhibitor, olaparib, in combination with cytotoxic chemotherapy in 28. Murai J, Zhang Y, Morris J, Ji J, Takeda S, Doroshow JH, et al. Rationale for pediatric solid tumors. Pediatr Blood Cancer 2014;61:145–50. poly(ADP-ribose) polymerase (PARP) inhibitors in combination therapy 14. Choy E, Butrynski J, Harmon D, Morgan J, George S, Wagner A, et al. with camptothecins or temozolomide based on PARP trapping versus Translation of preclinical predictive sensitivity of Ewing sarcoma to PARP catalytic inhibition. J Pharmacol Exp Ther 2014;349:408–16. inhibition: phase II study of olaparib in adult patients with recurrent/ 29. Pommier Y. Drugging topoisomerases: lessons and challenges. ACS Chem metastatic Ewing sarcoma following failure of prior chemotherapy. Cancer Biol 2013;8:82–95. Res 2013;73(suppl; abstr LB-174):8s. 30. Bowman KJ, Newell DR, Calvert AH, Curtin NJ. Differential effects of the 15. Wainberg Z, de Bono J, Mina L, Sachdev J, Averett Byers L, Chugh R, et al. poly (ADP-ribose) polymerase (PARP) inhibitor NU1025 on topoisom- Update on first-in-man trial of novel oral PARP inhibitor BMN 673 in erase I and II inhibitor cytotoxicity in L1210 cells in vitro. Br J Cancer patients with solid tumors. Mol Cancer Ther 2013;12(suppl; abstr 2001;84:106–12. C295):11s. 31. Bartek J, Lukas J. Chk1 and Chk2 kinases in checkpoint control and cancer. 16. Stewart E, Goshorn R, Bradley C, Griffiths LM, Benavente C, Twarog NR, Cancer Cell 2003;3:421–9. et al. Targeting the DNA repair pathway in Ewing sarcoma. Cell Rep 32. Mojsa B, Lassot I, Desagher S. Mcl-1 ubiquitination: unique regulation of 2014;9:829–41. an essential survival protein. Cells 2014;3:418–37. 17. Lee HJ, Yoon C, Schmidt B, Park do J, Zhang AY, Erkizan HV, et al. 33. Czabotar PE, Lee EF, van Delft MF, Day CL, Smith BJ, Huang DC, et al. Combining PARP-1 inhibition and radiation in Ewing sarcoma results in Structural insights into the degradation of Mcl-1 induced by BH3 domains. lethal DNA damage. Mol Cancer Ther 2013;12:2591–600. Proc Natl Acad Sci U S A 2007;104:6217–22.

2830 Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics

PARP Inhibitors Sensitize Ewing Sarcoma Cells to Temozolomide-Induced Apoptosis via the Mitochondrial Pathway

Florian Engert, Cornelius Schneider, Lilly Magdalena Weib, et al.

Mol Cancer Ther 2015;14:2818-2830. Published OnlineFirst October 5, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-15-0587

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2015/10/02/1535-7163.MCT-15-0587.DC1

Cited articles This article cites 32 articles, 10 of which you can access for free at: http://mct.aacrjournals.org/content/14/12/2818.full#ref-list-1

Citing articles This article has been cited by 7 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/14/12/2818.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/14/12/2818. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.