Combination of AZD2281 (Olaparib) and GX15-070 (Obatoclax) results in synergistic antitumor activities in preclinical models of pancreatic cancer

Cancer Letters 348 (2014) 20–28

Contents lists available at ScienceDirect

Cancer Letters journal homepage: www.elsevier.com/locate/canlet

Combination of AZD2281 (Olaparib) and GX15-070 (Obatoclax) results in synergistic antitumor activities in preclinical models of pancreatic cancer Shaohua Chen a, Guan Wang a, Xiaojia Niu a, Jianyun Zhao a, Wenxi Tan b, Hebin Wang b, Lijing Zhao b,⇑, Yubin Ge a,c,d,⇑ a

The State Engineering Laboratory of AIDS Vaccine, College of Life Sciences, Jilin University, Changchun, China Department of Pathophysiology College of Basic Medical Sciences, Jilin University, Changchun, China c Department of Oncology, Wayne State University School of Medicine, Detroit, MI, USA d Molecular Therapeutics Program, Barbara Ann Karmanos Cancer Institute, Wayne State University School of Medicine, Detroit, MI, USA b

a r t i c l e

i n f o

Article history: Received 6 December 2013 Received in revised form 21 January 2014 Accepted 10 February 2014

Keywords: Pancreatic cancer Obatoclax Olaparib Drug combination

a b s t r a c t In this study, we explored the antitumor activities of the PARP inhibitor AZD2281 (Olaparib) and the panBcl-2 inhibitor GX15-070 (Obatoclax) in six pancreatic cancer cell lines. While both agents were able to cause growth arrest and limited apoptosis, the combination of the two was able to synergistically cause growth arrest and non-apoptotic cell death. Furthermore, in an in vivo xenograft model, the combination caused substantially increased tumor necrosis compared to either treatment alone. Our results support further investigation of the combination of Bcl-2 and PARP inhibitors for the treatment of pancreatic cancer. Ó 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Pancreatic cancer represents only 2% of all cancers in the US, but it has the highest mortality rate at 99% and the lowest 5-year survival rate of less than 5% [30,39]. There has been little improvement in the prognosis over the past 20 years partially due to the delay of diagnosis, with many cases being diagnosed at an advanced stage [10]. Gemcitabine (20 ,20 -difluorodeoxycytidine, dFdC) is the standard first-line drug for treating advanced pancreatic cancer [3]. However, its efficacy remains low, with a median survival of 5.7 months and 1-year survival rate of 18% [18,21]. Therefore, new therapies for this extremely aggressive disease are urgently needed. PARP-1 is a DNA binding protein involved in apoptosis as well as DNA single- and double-strand break repair, and has been becoming a popular therapeutic target for many different malignancies [1,17,41]. PARP inhibitors have been demonstrated to have a strong synthetic lethal relationship with BRCA1/2 deficiency. The ⇑ Corresponding authors. Address: Department of Pathophysiology, College of Basic Medical Sciences, Jilin University, Changchun 130021, Jilin Province, China (L. Zhao). Address: Department of Oncology, Wayne State University School of Medicine, 110 East Warren Ave., Detroit, MI 48201, USA. Tel.: +1 (313) 578 4285; fax: +1 (313) 578 4287 (Y. Ge). E-mail addresses: [email protected] (L. Zhao), [email protected] (Y. Ge). http://dx.doi.org/10.1016/j.canlet.2014.02.010 0304-3835/Ó 2014 Elsevier Ireland Ltd. All rights reserved.

loss of PARP activity leads to more lesions needing to be repaired by the homologous recombination (HR) pathway, which can be lethal in a BRCA1/2 deficient background [24]. Based on this finding, PARP inhibitors have been widely used in recent clinical trials for treating BRAC1/2-deficient tumors including pancreatic cancer and have shown promising clinical activities [12,25,32,41]. The use of PARP inhibitors has recently been extended to tumors with other defects in the HR DNA repair pathway, as well as in combination with chemotherapy drugs and chemoradiotherapy [7,24,32]. Combining PARP inhibitors with agents that impair DNA damage repair to treat BRCA1/2 wild-type tumors could broaden the clinical use of these promising PARP inhibitors. A common cause for chemotherapy resistance arises from the overexpression of the anti-apoptotic Bcl-2 family members. BclxL and Mcl-1 are expressed in 88% whereas Bcl-2 is expressed in 23% of invasive ductal carcinomas [26]. Therefore targeting these proteins could be an effective treatment for pancreatic cancer. GX15-070 (Obatoclax) is a small molecule that binds to the BH3binding site of Bcl-2, Bcl-xL, and Mcl-1 [9,20,28]. This pan-Bcl-2 inhibitor has been reported to directly induce apoptosis in a variety of cultured cancer cell lines and primary patient samples through the mitochondrial apoptotic pathway as well as non-apoptotic cell death [16,28,33]. Recent studies have found compelling evidence to suggest that the anti-apoptotic Bcl-2 family proteins can regu-

S. Chen et al. / Cancer Letters 348 (2014) 20–28

late DNA double-strand break repair independent of their pro-survival functions [2,11,14,17,19,31,36–38,42,44]. We hypothesize that using a pan-Bcl-2 inhibitor may sensitize pancreatic cancer cells to PARP inhibitors. In this study, we investigated the combination of the PARP inhibitor AZD2281 and the pan-Bcl-2 inhibitor GX15-070 in six pancreatic cancer cell lines harboring wild-type BRCA1 and BRCA2 genes. When combined simultaneously, the two agents caused additive to synergistic growth arrest and cooperatively induced non-apoptotic cell death in the pancreatic cancer cell lines. Our in vivo xenograft model studies revealed that the two drugs cooperate to induce substantial tumor necrosis, resulting in a cavity in the center of the tumors. These results provide support for the combination of PARP and Bcl-2 inhibitors in the treatment of pancreatic cancer.

21

2.6. Western blot analysis Western blotting was performed using polyvinylidene difluoride (PVDF) membranes (Thermo Fisher Inc., Rockford, IL, USA) and immunoblotted with anti-PARP1, -Bcl-2, -Mcl-1, -Bcl-xL, -cleaved caspase3, -CDK1, -CDK2, cyclin B1, -caspase 3 (Cell Signaling Technology, Beverly, MA, USA), or -b-actin (Sigma–Aldrich) antibody, as described previously [6]. Immunoreactive proteins were visualized using the Odyssey Infrared Imaging System (Li-Cor, Lincoln, NE, USA), as described by the manufacturer. 2.7. Cell death, apoptosis, and cell cycle progression Cells were treated as indicated and cell death was determined by trypan blue exclusion. The remaining cells were fixed with ice-cold 80% (v/v) ethanol for 24 h. After centrifugation at 200g for 5 min, the cell pellets were washed with PBS (pH 7.4) and resuspended in PBS containing propidium iodide (PI, 50 lg/mL), Triton X-100 (0.1%, v/v), and DNase-free RNase (1 lg/mL). DNA contents were determined by flow cytometry analysis, as previously described [43]. Cell cycle analysis was performed with Multicycle software (Phoenix Flow Systems, Inc., San Diego, CA, USA). Apoptotic events were recorded as PI+ events (Sub-G1 population).

2. Materials and methods 2.1. Drugs GX15-070 (Obatoclax) and AZD2281 (Olaparib) were purchased from Selleck Chemicals LLC (Houston, TX, USA). Both agents were dissolved in DMSO and stored at 80 °C, as recommended by the supplier.

2.2. Cell culture The BxPC-3, HPAC, MIAPaCa-2, PANC-1, AsPC-1 and CFPAC-1 human pancreatic cancer cell lines were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). All of these cell lines harbor wild-type BRCA1 and BRCA2 genes [8,40]. The cell lines were cultured in Dulbecco’s Modified Eagle medium (DMEM, Invitrogen, Carlsbad, CA, USA, for HPAC, MIAPaCa-2 and PANC-1), RPMI1640 medium (Invitrogen, for AsPC-1 and BxPC-3), or Iscove’s Modified Dulbecco’s medium (IMDM, Invitrogen, for CFPAC-1) with 10% heat-inactivated fetal bovine serum (FBS; Hyclone Labs, Logan, UT, USA) plus 100 U/mL penicillin and 100 lg/mL streptomycin in a 37 °C humidified atmosphere containing 5% CO2/95% air. Cell lines were authenticated by the University of Arizona Genetics Core Facility (Tucson, AZ, USA).

2.3. In vitro cytotoxicity assays In vitro AZD2281 or GX15-070 cytotoxicities of pancreatic cancer cell lines were measured by using MTT (3-[4,5-dimethyl-thiazol-2-yl]-2,5-diphenyltetrazoliumbromide, Sigma–Aldrich, St Louis, MO, USA) reagent, as previously described [45,46]. The extent and direction of GX15-070 and AZD2281 antitumor interactions were evaluated using CompuSyn software (ComboSyn, Inc., Paramus, NJ, USA). Drug interactions were quantified by determining the combination index (CI), where CI < 1, CI = 1, and CI > 1 indicate synergistic, additive, and antagonistic effects, respectively [4].

2.4. Colony formation assays Colony formation assays were carried out as previously described [43]. Briefly, 500 BxPC-3 cells were seeded into 100 mm dishes in complete RPMI1640, cultured for 24 h, and then treated with the indicated concentrations of GX15-070, AZD2281, or AZD2281 plus GX15-070 for 48 h. The cells were then washed twice with drugfree RPMI1640 and cultured in complete RPMI1640 for up to 3 weeks. Colonies were visualized by coomassie blue staining and counted. The extent and direction of interactions between GX15-070 and AZD2281 in suppressing colony formation were determined using CompuSyn software (ComboSyn, Inc.).

2.5. shRNA knockdown of PARP-1 in PANC-1 cells The pMD-VSV-G and delta 8.2 plasmids were gifts from Dr. Dong at Tulane University. The transfection was carried out by using Lipofectamine and Plus reagents (Life Technologies, Carlsbad, CA, USA), as previously described [6,45,46]. Briefly, lentivirus vector (PARP-1 or non-target control shRNA construct from the RNAi Consortium, Sigma–Aldrich), pMD-VSV-G and delta 8.2 were co-transfected into TLAHEK293T cells and the culture medium was harvested 48 h post-transfection. PANC-1 cells were transduced by adding 1 mL of virus supernatant and 4 lg of polybrene. After selection with puromycin, a pool of infected cells was expanded and tested for PARP-1 expression by Western blotting. A pool of cells from the non-target control transduction was used as the negative control.

2.8. Establishment of a mouse pancreatic cancer xenograft model Female BALB/c nude mice (18–22 g) were purchased from Vital River Laboratories (Beijing, China). The animal study was conducted following internationally recognized guidelines and was approved by the Animal Research Committee of Norman Bethune College of Medicine, Jilin University. Log phase BxPC-3 cells were digested with trypsin, adjusted to 2  107 cells/ml with matrigel (BD Biosciences, San Jose, CA, USA), and inoculated subcutaneously in the right side axillae of BALB/c mice to generate a xenograft (100 ll per mouse). When the xenografts reached an average volume of 75.6 ± 36.5 mm3, the mice were randomized into four groups (8 animals per group) and injected intraperitoneally with 200 ll of (i) phosphate-buffered saline (PBS) with 10% Hydroxypropyl-b-Cyclodextrin (HP-b-CD) (control group), (ii) AZD2281 (50 mg/kg) dissolved in PBS with 10% HP-b-CD once daily Monday through Friday (AZD2281 group), (iii) GX15-070 (3 mg/kg) dissolved in PBS with HP-b-CD once daily Monday through Wednesday (GX15-070 group), or (iv) AZD2281 five times a week and GX15-070 three times a week dissolved in PBS with HP-b-CD (combination group), for 3 weeks. Tumor diameters were measured with a caliper every 3–4 days. Blood samples were taken via retro-orbital bleed under anesthesia on day 28 post drug treatment initiation. The whole blood was collected into sterile Ep tubes and centrifuged at 3000 rpm for 10 min, and serum was collected for detection of carbohydrate antigen 19-9 (CA19-9). The mice were then sacrificed and the tumors were removed, weighed and fixed in 10% formalin for hematoxylin and eosin (H&E) and immunohistochemical staining. 2.9. Detection of CA 19-9 in blood serum Serum CA 19-9 was measured by using a CA19-9 [I125] IRMA kit (3 V Biosciences, Weifang, China), according to the manufacturer’s instructions. 2.10. H&E and immunohistochemical staining Tumors from 3 mice in each treatment group were analyzed by H&E and immunohistochemical staining. All specimens were fixed in 10% formalin, embedded in paraffin, and cut into 4 lm-thick slides. The slides were dewaxed and stained with H&E for histological assessments. Center necrosis areas were quantified by using the BI-2000 Medical Image Analysis System (TME Technology Co., Chengdu, China). For immunohistochemical staining, the endogenous peroxidase activity was blocked by treatment with 3% hydrogen peroxide solution in methanol for 20 min. Epitope retrieval was performed by treating the slides with 10 mM sodium citrate buffer (pH 6.0) and heating twice in a microwave oven at high power for 6 min each. Non-specific binding was prevented by blocking with goat serum (Dingguo Biosciences, Beijing, China) (1:10 in PBS) for 10 min. Immunostaining of proliferating cell nuclear antigen (PCNA) and CD34 was performed using mouse anti-human monoclonal antibodies (Abgent Inc., San Diego, CA, USA). After incubation with the primary antibody (1:100 in PBS) for 60 min, the slides were incubated with a biotinylated goat anti-mouse IgG (H + L) (Beyotime, Shanghai, China) at 37 °C for 30 min, followed by incubation with a 1:200 streptavidin–biotin–peroxidase complex (Sigma–Aldrich) for 30 min. Reactive products were visualized with 3,30 diaminobenzidene (DAB) as the chromogen, and the slides were counterstained with hematoxylin and coverslipped. Sections previously known to express PCNA or CD34 were included in each run, receiving either the primary antibody as the positive control, or a mouse IgG as the negative control. The stained slides were analyzed with a microscope, and brown staining was scored using Image-Pro Plus 6.0 (Media Cybernetics, Inc., Bethesda, MD, USA). The microvessel density (MVD) was calculated using the ‘‘Hot spot’’ method: CD34 positive vascular dense areas, or ‘‘hot spots’’, were identified at 40 magnification. The discrete microvessels in a 200 field were then counted. Three separate hotspots were assessed to give a mean MVD value for each tumor [22].

22

S. Chen et al. / Cancer Letters 348 (2014) 20–28

3. Results

070 ranged from 219.98 nM (PANC-1) to 589.7 nM (AsPC-1, Fig. 1B and C and Table 1). The individual drug sensitivities did not appear to correlate with PARP-1, Bcl-2, Bcl-xL or Mcl-1 protein levels. Next, we treated BxPC-3 and PANC-1 cells with varying concentrations of AZD2281 (0–20 lM) or GX15-070 (0– 2.5 lM) and assessed drug-induced apoptosis by PI staining and flow cytometry analyses and measuring the population with DNA fragmentation (sub-G1). Even at the highest concentration there was at most approximately 6% sub-G1 cells, indicating a minimal induction of apoptosis by either of the two agents (Fig. 1D and E).

3.1. PARP-1 and Bcl-2 family protein expression and inhibitor sensitivities in pancreatic cancer cell lines

3.2. AZD2281 and GX15-070 synergized to cause growth arrest and cooperated to induce non-apoptotic cell death

To begin to determine if AZD2281 and GX15-070 cooperate to kill pancreatic cancer cells, we first determined PARP-1, Bcl-2, Bcl-xL, and Mcl-1 protein levels and IC50 values for each drug in six pancreatic cancer cell lines. The proteins were expressed at variable levels (Fig. 1A). The IC50 values for AZD2281 ranged from 16.18 lM (CFPAC-1) to 52.5 lM (MIAPaCa-2) and those for GX15-

To determine if AZD2281 and GX15-070 synergize to induce pancreatic cancer cell growth arrest, the pancreatic cancer cell lines were treated simultaneously with varying concentrations of AZD2281 and 0 nM, 85 nM, or 170 nM GX15-070, clinically achievable concentrations [13], for 96 h and the AZD2281 IC50 values were determined using MTT assays. The AZD2281 IC50s were

2.11. Statistical analysis The in vitro data were expressed as mean values ± standard errors, while the in vivo data were expressed as mean values ± standard deviations. For the in vitro data, differences in AZD2281 IC50s between GbX15-070 treated and vehicle control treated cells and differences in cell death/apoptosis between GX15-070 and AZD2281 treated (individually or combined) and vehicle control treated cells were compared using the paired two-sample t-test. Statistical analyses were performed with GraphPad Prism 5.0. For the in vivo data, statistical comparisons were performed with SPSS17.0 using analysis of variance (ANOVA) test, p < 0.05 was deemed to be statistically significant.

Fig. 1. Cytotoxic effects of AZD2281 and GX15-070 on pancreatic cancer cell lines. Panel A: Protein extracts from log phase AsPC-1, BxPC-3, CFPAC-1, HPAC, MIAPaCa-2 and PANC-1 cells were subjected to Western blotting and probed by anti-Bcl-2, -Bcl-xL, -Mcl-1, -PARP-1, or -b-actin antibody. Panels B and C: AsPC-1, BxPC-3, CFPAC-1, HPAC, MIAPaCa-2 and PANC-1 cells were treated with a range of concentrations of AZD2281 (0–40 lM, panel B) or GX15-070 (0–680 nM, panel C) in complete medium in 96-well plates at 37 °C for 96 h. Viable cells were determined using the MTT reagent and a visible light microplate reader. Panels D and E: BxPC-3 or PANC-1 cells were treated with a range of concentrations of AZD2281 (0–20 lM) or GX15-070 (0–2500 nM) for 72 h. Apoptosis was then determined by PI staining and flow cytometry analysis. The data are presented as mean values ± standard errors from at least 3 independent experiments. Western blots were repeated at least three times and one representative image is shown.

23

S. Chen et al. / Cancer Letters 348 (2014) 20–28 Table 1 Effects of GX15-070 on AZD2281 cytotoxicities in pancreatic cancer cell lines. Cell line

IC50 of GX15-070 (nM)

AsPC-1 BxPC-3 CFPAC-1 HPAC MIAPaCa-2 PANC-1

589.7 ± 15.4 245.5 ± 9.7 527.7 ± 51.6 322.8 ± 51.4 322.3 ± 26.6 220.0 ± 12.9

IC50 of AZD2281 (lM) in the absence or presence of GX15-070 (nM) 0

85

170

29.0 ± 1.8a 25.4 ± 2.6 16.2 ± 1.1 51.5 ± 3.0a 52.5 ± 4.7a 24.9 ± 4.2a

13.7 ± 2.1 (0.62) 15.1 ± 1.0 (0.94) 10.8 ± 1.0 (0.83) 17.9 ± 1.9 (0.61) 38.0 ± 3.3 (0.99)a 6.6 ± 2.8 (0.66)

5.8 ± 1.5 2.1 ± 0.3 5.9 ± 1.3 1.6 ± 0.1 7.6 ± 1.0 1.7 ± 0.4

p Value

(0.49) (0.78) (0.69) (0.56) (0.67) (0.84)

<0.005 <0.05 <0.05 <0.001 <0.05 <0.05

Note: The pancreatic cancer cell lines were treated with variable concentrations of AZD2281 for 96 h in the absence or presence of variable concentrations of GX15-070 administered simultaneously. Viable cells were measured by MTT assays and the extent and direction of antitumor interactions between AZD2281 and GX15-070 were determined by using CompuSyn software. The numbers in the parentheses represent the combination index (CI) values, where CI < 1, CI = 1, and CI > 1 indicate synergistic, additive, and antagonistic effects, respectively. The data are presented as mean values ± standard errors from at least 3 independent experiments. a IC50s were estimated using GraphPad Prism 5.0.

Fig. 2. GX15-070 and AZD2281 cooperate to induce growth arrest in pancreatic cancer cell lines. AsPC-1 (panel A), BxPC-3 (panel B), CFPAC-1 (panel C), HPAC (panel D), MIAPaCa-2 (panel E), or PANC-1 (panel F) cells were treated with variable concentrations of AZD2281 (0–40 lM) for 96 h in the absence or presence of variable concentrations of GX15-070 (85 or 170 nM) administered simultaneously. Viable cells were measured by MTT assays. The data are presented as means ± standard errors from at least 3 independent experiments.

significantly reduced by 2.7- (CFPAC-1) to 32.8-fold (HPAC). Additive to synergistic antitumor interactions were detected by calculating the combination index values for the combinations of AZD2281 with 85 nM GX15-070, while uniform synergistic antitu-

mor interactions were detected for the combinations of AZD2281 with 170 nM GX15-070 (Fig. 2 and Table 1). Next, we investigated whether or not AZD2281 and GX15-070 would cooperate to induce cell death. BxPC-3 and PANC-1 cells

24

S. Chen et al. / Cancer Letters 348 (2014) 20–28

Fig. 3. GX15-070 and AZD2281 Cooperate to induce non-apoptotic cell death and cell cycle arrest in BxPC-3 and PANC-1 cells. Panels A–D: BxPC-3 or PANC-1 cells were treated with AZD2281 or GX15-070 alone or combined for 72 h. Cell death was determined by trypan blue exclusion (panels A and B). Full-length caspase-3 (FL-casp3), cleaved form of caspase-3 (CF-casp3), PARP-1 (cleaved form designated as CF-PARP-1), Bcl-2, Bcl-xL, and Mcl-1 protein levels were determined by Western blots analyzing whole cell lysates (panel C). Protein lysates from SK-N-DZ cells treated with 500 nM MK1775 for 48 h were used as the positive control. Effects of these drug treatments on cell cycle progression were determined by PI staining and flow cytometry analyses (panels D and E). Protein levels of Cyclin B1, CDK1 and CDK2 were determined by Western blotting (panel F). Panels G and H: Engineered PANC-1 cells stably expressing a negative control (NTC)- or PARP-shRNA were established as described in the Materials and Methods and PARP-1 expression was determined by Western blotting (panel G). Death of the NTC-shRNA or PARP-shRNA cells treated with 85 nM GX15-070 for 72 h was determined by the trypan blue exclusion assay (panel H). The data are presented as mean percentages ± standard errors from at least 3 independent experiments. ⁄ Indicates p < 0.05, while ⁄⁄ indicates p < 0.005. Western blots were repeated at least three times and one representative image is shown.

were treated with 5 or 1.25 lM AZD2281, 85 nM GX15-070, or in combination for 72 h and percentages of dead cells were determined with the trypan blue exclusion assay. Simultaneous treatment resulted in a significantly increased percentage of dead

cells compared to the individual drug treatments (Fig. 3A and B). We did not detect cleavage of caspase-3 or PARP-1, indicating that the increase in cell death was due to non-apoptotic cell death (Fig. 3C). Bcl-xL and Mcl-1 protein levels were largely unchanged

S. Chen et al. / Cancer Letters 348 (2014) 20–28

25

Fig. 4. AZD2281 and GX15-070 cooperate in suppressing colony formation of BxPC-3 cells. Five hundred BxPC-3 cells were plated in 100 mm dishes 24 h prior to the treatments with variable concentrations of AZD2281 or GX15-070 alone (panels A and B) or in combination for 48 h (panels C and D). The drugs were then washed out and the cells were cultured in drug-free complete medium for up to 3 weeks. Colonies were visualized by coomassie blue staining and then counted. The data are presented as means ± standard errors from at least 3 independent experiments. ⁄⁄ Indicates p < 0.005, while ⁄⁄⁄ indicates p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

and a small decrease in Bcl-2 was detected after the combined drug treatment compared to vehicle control treated cells. Cell cycle analysis in BxPC-3 cells revealed a small significant increase in G2/M phase after treatment with AZD2281 alone, which was further significantly increased by the addition of GX15-070 and accompanied by a decrease in CDK1 protein levels (Fig. 3D and F, p < 0.05). In the PANC-1 cells, we detected a small significant increase in S phase after individual drug treatments, which was further significantly increased when the drugs were administered simultaneously and accompanied by a minor decrease in CDK2 protein levels (Fig. 3E and F, p < 0.05). Cyclin B1 protein levels remained unchanged in the BxPC-3 cells, while there was an increase in the AZD2281 treated PANC-1 cells. Consistent with the lack of cleavage of caspase-3 and PARP-1 in the cells treated with AZD2281 or GX15-070, alone or combined, no increase of sub-G1 population in the combined treatments could be detected by PI staining and flow cytometry analyses (data not shown). To provide further evidence that PARP inhibition is required for the enhanced non-apoptotic cell death for the combination of AZD2281 and GX15-070, we performed lentiviral shRNA knockdown of PARP-1 in PANC-1 cells (designated PARP-shRNA), which was confirmed by Western blots (non-target control cells are designated NTC-shRNA, Fig. 3G). The cells were treated with or without 85 nM GX15-070 for 72 h and percentage of dead cells were determined using the trypan blue exclusion assay. After treatment with GX15-070, the PARP-shRNA cells had a significantly higher percent of dead cells compared to the NTC-shRNA control cells (Fig. 3H). Cell death was further confirmed by colony formation assays in BxPC-3 cells. AZD2281 or GX15-070 treatment alone caused dosedependent decrease in colony formation (Fig. 4A and B). Combined AZD2281 and GX15-070 treatment showed a synergistic and significant decrease in colony formation compared to the individual drug treatments (CI = 0.69, Fig. 4C and D).

3.3. In vivo antitumor efficacy of AZD2281 and GX15-070 in a xenograft model To evaluate the effects of AZD2281 and GX15-070 on pancreatic tumor growth in vivo, we used a mouse BxPC-3 xenograft model. Individual and combined drug treatments were well tolerated, indicated by the lack of a significant loss of body weight for the drug treatment groups compared to the vehicle control group (Fig. 5A). Surprisingly, neither the individual nor the combined drug treatments resulted in significant delay of externally measurable tumor growth during the three week drug treatment period (Fig. 5B). Based on these results, the mice were sacrificed on day 28 post drug treatment initiation. On that day, body weight and tumor weight were essentially the same among the four treatment groups (Fig. 5C and D), however, serum CA19-9 levels (a pancreatic cancer biomarker that is used to monitor chemotherapy response [23,27,29,47]) in the combination group were significantly lower compared to that for the other treatment groups, indicating treatment response (Fig. 5E). To investigate the in vivo effects of these drug treatments, tumors from 3 mice in each group were analyzed with H&E and immunohistochemical staining. Interestingly, H&E staining revealed substantially increased necrosis in the tumors for the drug treated groups compared to the vehicle control group, especially for the combination group (Fig. 5F). Quantification showed mean 5.4 ± 2.3%, 10.6 ± 2.5%, 34.0 ± 16.8%, and 55.2 ± 10.6% center necrosis areas for the vehicle control, AZD2281, GX15-070, and the combination groups, respectively (Fig. 5G). The difference in the center necrosis areas between the combination group and the vehicle control group or AZD2281 group was statistically significant, while that between the combination group and the GX15-070 group was not, probably due to the large variation of center necrosis areas within the GX15-070 group and the small sample size. There was no obvious invasion of the tumor into other tissues (data not shown).

26

S. Chen et al. / Cancer Letters 348 (2014) 20–28

Fig. 5. In vivo antitumor efficacies of AZD2281 and GX15-070 alone or in combination in a BxPC-3 xenograft model. Panels A–C: Body weights (panels A and C) and tumor volumes (panel B) were measured every 3–4 days. Tumor volume was calculated according to the following formula: m12  m2  0.5236 (m1: short diameter; m2: long diameter). Panel D: The mice were sacrificed on day 28 post drug treatment initiation and the tumors were removed and weighed. Panel E: The serum CA19-9 levels were detected using a CA19-9 IRMA kit. Panel F: Tumor specimens were fixed in 10% formalin, embedded in paraffin, and cut into 4 lm-thick slides for H&E, PCNA, and CD34 staining. Panel G: Center necrosis areas were quantified by using the BI-2000 Medical Image Analysis System. Necrosis was defined as tumor area with loss of clearly defined cell boundaries and without pyknotic nuclei. Panel H: The proliferation index was calculated as proliferation index = PCNA positive cells/observed cells X 100% and graphed. Panel I: Microvessel density was calculated and graphed as described in the Materials and Methods. The data are presented as mean values ± standard deviations. ⁄ Indicates p < 0.05, while ⁄⁄ indicates p < 0.01.

Analogous to the results from H&E staining, immunohistochemical staining showed substantially decreased expression of PCNA and CD34 in the combination group compared to the other groups (Fig. 5H and I). The proliferation index values of the combination group (18.3 ± 6.6%) were significantly lower than that in the vehicle control group (56.5 ± 10.5%, Fig. 5H). Further, the MVDs of the combination group (25.6 ± 8.8%) were significantly lower than the vehicle control group (60.3 ± 17.1, Fig. 5I). 4. Discussion Emerging evidence has demonstrated that Bcl-2, Bcl-xL and Mcl-1 possess non-apoptotic functions, in addition to their tradi-

tional roles in the mitochondrial apoptosis pathway, and are involved in the DNA damage response [5,19,38,44]. PARP inhibitors have demonstrated efficacy in cells with impaired HR due to BRCA1/2 mutations, as well as mutations in other HR-related genes [7,24,32]. Thus, targeting Bcl-2, Bcl-xL, and/or Mcl-1 may impair the DNA damage response and potentiate BRCA1/2 wild-type pancreatic cancer cells to PARP inhibitors. In this study, we combined GX15-070 and AZD2281 in vitro and in vivo to determine if they cooperate to induce growth arrest and cell death. Based on the traditional roles Bcl-2, Bcl-xL, and Mcl-1 play in the mitochondrial apoptosis pathway, GX15-070 would be expected to cause apoptotic cell death [16,33]. However, at clinically achievable concentrations of GX15-070 we detected minimal

S. Chen et al. / Cancer Letters 348 (2014) 20–28

apoptosis, but significant growth arrest, indicating a non-classical mechanism underlying the cytotoxic effects of GX15-070 (Fig. 1C and E). Our results demonstrate that these two drugs induce pancreatic cancer cell growth arrest in an additive to synergistic manner in all of the cell lines tested (Fig. 2 and Table 1). Interestingly, the combination also caused a significant increase in non-apoptotic cell death (Figs. 3 and 4), potentially due to decrease in Bcl-2 protein levels. This was accompanied by significantly enhanced S or G2/M cell cycle arrest compared to the individual drug treatments (Fig. 3), possibly due to the decrease in CDK1 or CDK2 protein levels. These results further indicate that GX15-070 potentiates the cytotoxic effects of AZD2281 through non-classical mechanisms. Surprisingly, in our in vivo studies, externally measuring tumor volume of pancreatic cancer xenografts in nude mice indicated a lack of response to the drug treatments. However, the combined drug treatment caused a significant decrease in serum CA19-9 levels compared to the other treatment groups, indicating treatment response. Upon further investigation, H&E staining revealed that the tumors from the combined drug treatment had larger center necrosis areas compared to the individual drug treatments, while only minimal was detected in the vehicle control group. Furthermore, immunohistochemical examination of PCNA showed that the combined drug treatment group had lower PCNA expression than the other treatment groups, indicating that the two drugs cooperated in suppressing proliferation of pancreatic cancer cells in the tumors. Analogous results were obtained with CD34 immunohistochemical staining, indicating that GX15-070 and AZD2281 cooperated in suppressing angiogenesis in the tumors, resulting in the appearance of a necrotic cavity in the center of tumor. Indeed, there are data to suggest that Bcl-2, Bcl-xL and PARP-1 are involved with the process of angiogenesis [15,34,35]. Taken together, our results demonstrate cooperative antitumor effects of GX15-070 and AZD2281 on preclinical models of pancreatic cancer in vitro and in vivo. Although we do not have direct evidence, we suspect that GX15-070 predominantly targets the nonapoptotic functions of Bcl-2, Bcl-xL and/or Mcl-1 to potentiate the cytotoxic effects of the PARP inhibitor AZD2281 in pancreatic cancer cells. Further studies are necessary to determine the molecular mechanisms of non-apoptotic cell death induced by the combination. Our novel findings suggest that there might be a clinical benefit for using the combination of GX15-070 and AZD2281 to treat pancreatic cancer. Conflict of Interest The authors declare no competing financial interests. Acknowledgments This study was support by a Start-up Fund from Jilin University, Changchun, China, and grants from the National Natural Science Foundation of China (NSFC 31271477 and 81200363). References [1] M.W. Audeh, J. Carmichael, R.T. Penson, M. Friedlander, B. Powell, K.M. BellMcGuinn, C. Scott, J.N. Weitzel, A. Oaknin, N. Loman, K. Lu, R.K. Schmutzler, U. Matulonis, M. Wickens, A. Tutt, Oral poly(ADP-ribose) polymerase inhibitor olaparib in patients with BRCA1 or BRCA2 mutations and recurrent ovarian cancer: a proof-of-concept trial, Lancet 376 (2010) 245–251. [2] E. Bolderson, D.J. Richard, B.B. Zhou, K.K. Khanna, Recent advances in cancer therapy targeting proteins involved in DNA double-strand break repair, Clin. Cancer Res.: Off. J. Am. Assoc. Cancer Res. 15 (2009) 6314–6320. [3] H.A. Burris 3rd, M.J. Moore, J. Andersen, M.R. Green, M.L. Rothenberg, M.R. Modiano, M.C. Cripps, R.K. Portenoy, A.M. Storniolo, P. Tarassoff, R. Nelson, F.A. Dorr, C.D. Stephens, D.D. Von Hoff, Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial, J. Clin. Oncol.: Off. J. Am. Soc. Clin. Oncol. 15 (1997) 2403–2413.

27

[4] T.C. Chou, Theoretical basis, experimental design, and computerized simulation of synergism and antagonism in drug combination studies, Pharmacol. Rev. 58 (2006) 621–681. [5] A. Dumay, C. Laulier, P. Bertrand, Y. Saintigny, F. Lebrun, J.L. Vayssiere, B.S. Lopez, Bax and Bid, two proapoptotic Bcl-2 family members, inhibit homologous recombination, independently of apoptosis regulation, Oncogene 25 (2006) 3196–3205. [6] H. Edwards, C. Xie, K.M. LaFiura, A.A. Dombkowski, S.A. Buck, J.L. Boerner, J.W. Taub, L.H. Matherly, Y. Ge, RUNX1 regulates phosphoinositide 3-kinase/AKT pathway: role in chemotherapy sensitivity in acute megakaryocytic leukemia, Blood 114 (2009) 2744–2752. [7] H. Farmer, N. McCabe, C.J. Lord, A.N. Tutt, D.A. Johnson, T.B. Richardson, M. Santarosa, K.J. Dillon, I. Hickson, C. Knights, N.M. Martin, S.P. Jackson, G.C. Smith, A. Ashworth, Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy, Nature 434 (2005) 917–921. [8] S.A. Forbes, N. Bindal, S. Bamford, C. Cole, C.Y. Kok, D. Beare, M. Jia, R. Shepherd, K. Leung, A. Menzies, J.W. Teague, P.J. Campbell, M.R. Stratton, P.A. Futreal, COSMIC: mining complete cancer genomes in the Catalogue of Somatic Mutations in Cancer, Nucleic Acids Res. 39 (2011) D945–950. [9] S. Fulda, L. Galluzzi, G. Kroemer, Targeting mitochondria for cancer therapy, Nat. Rev. Drug Discovery 9 (2010) 447–464. [10] E. Giovannetti, V. Mey, S. Nannizzi, G. Pasqualetti, M. Del Tacca, R. Danesi, Pharmacogenetics of anticancer drug sensitivity in pancreatic cancer, Mol. Cancer Ther. 5 (2006) 1387–1395. [11] T. Helleday, E. Petermann, C. Lundin, B. Hodgson, R.A. Sharma, DNA repair pathways as targets for cancer therapy, Nat. Rev. Cancer 8 (2008) 193–204. [12] S. Huang, K. Okumura, F.A. Sinicrope, BH3 mimetic obatoclax enhances TRAILmediated apoptosis in human pancreatic cancer cells, Clin. Cancer Res.: Off. J. Am. Assoc. Cancer Res. 15 (2009) 150–159. [13] J.J. Hwang, J. Kuruvilla, D. Mendelson, M.J. Pishvaian, J.F. Deeken, L.L. Siu, M.S. Berger, J. Viallet, J.L. Marshall, Phase I dose finding studies of obatoclax (GX15070), a small molecule pan-BCL-2 family antagonist, in patients with advanced solid tumors or lymphoma, Clin. Cancer Res.: Off. J. Am. Assoc. Cancer Res. 16 (2010) 4038–4045. [14] S.P. Jackson, J. Bartek, The DNA-damage response in human biology and disease, Nature 461 (2009) 1071–1078. [15] E. Karl, Z. Zhang, Z. Dong, K.G. Neiva, M.S. Soengas, A.E. Koch, P.J. Polverini, G. Nunez, J.E. Nor, Unidirectional crosstalk between Bcl-xL and Bcl-2 enhances the angiogenic phenotype of endothelial cells, Cell Death Differ. 14 (2007) 1657–1666. [16] M. Konopleva, J. Watt, R. Contractor, T. Tsao, D. Harris, Z. Estrov, W. Bornmann, H. Kantarjian, J. Viallet, I. Samudio, M. Andreeff, Mechanisms of antileukemic activity of the novel Bcl-2 homology domain-3 mimetic GX15-070 (obatoclax), Cancer Res. 68 (2008) 3413–3420. [17] R. Krishnakumar, W.L. Kraus, The PARP side of the nucleus: molecular actions, physiological outcomes, and clinical targets, Mol. Cell 39 (2010) 8–24. [18] F. Kullmann, S. Hollerbach, M.M. Dollinger, J. Harder, M. Fuchs, H. Messmann, J. Trojan, E. Gabele, A. Hinke, C. Hollerbach, E. Endlicher, Cetuximab plus gemcitabine/oxaliplatin (GEMOXCET) in first-line metastatic pancreatic cancer: a multicentre phase II study, Br. J. Cancer 100 (2009) 1032–1036. [19] C. Laulier, A. Barascu, J. Guirouilh-Barbat, G. Pennarun, C. Le Chalony, F. Chevalier, G. Palierne, P. Bertrand, J.M. Verbavatz, B.S. Lopez, Bcl-2 inhibits nuclear homologous recombination by localizing BRCA1 to the endomembranes, Cancer Res. 71 (2011) 3590–3602. [20] B. Leber, F. Geng, J. Kale, D.W. Andrews, Drugs targeting Bcl-2 family members as an emerging strategy in cancer, Expert Rev. Mol. Med. 12 (2010) e28. [21] J. Li, M.W. Saif, Advancements in the management of pancreatic cancer, JOP: J. Pancreas 10 (2009) 109–117. [22] J.T. Liang, K.C. Huang, Y.M. Jeng, P.H. Lee, H.S. Lai, H.C. Hsu, Microvessel density, cyclo-oxygenase 2 expression, K-ras mutation and p53 overexpression in colonic cancer, Br. J. surgery 91 (2004) 355–361. [23] N.R. Maisey, A.R. Norman, A. Hill, A. Massey, J. Oates, D. Cunningham, CA19-9 as a prognostic factor in inoperable pancreatic cancer: the implication for clinical trials, Br. J. Cancer 93 (2005) 740–743. [24] N. McCabe, N.C. Turner, C.J. Lord, K. Kluzek, A. Bialkowska, S. Swift, S. Giavara, M.J. O’Connor, A.N. Tutt, M.Z. Zdzienicka, G.C. Smith, A. Ashworth, Deficiency in the repair of DNA damage by homologous recombination and sensitivity to poly(ADP-ribose) polymerase inhibition, Cancer Res. 66 (2006) 8109–8115. [25] P. Michl, T.M. Gress, Current concepts and novel targets in advanced pancreatic cancer, Gut 62 (2013) 317–326. [26] Y. Miyamoto, R. Hosotani, M. Wada, J.U. Lee, T. Koshiba, K. Fujimoto, S. Tsuji, S. Nakajima, R. Doi, M. Kato, Y. Shimada, M. Imamura, Immunohistochemical analysis of Bcl-2, Bax, Bcl-X, and Mcl-1 expression in pancreatic cancers, Oncology 56 (1999) 73–82. [27] Y. Nakai, T. Kawabe, H. Isayama, T. Sasaki, H. Yagioka, Y. Yashima, H. Kogure, T. Arizumi, O. Togawa, Y. Ito, S. Matsubara, K. Hirano, N. Sasahira, T. Tsujino, M. Tada, M. Omata, CA 19-9 response as an early indicator of the effectiveness of gemcitabine in patients with advanced pancreatic cancer, Oncology 75 (2008) 120–126. [28] M. Nguyen, R.C. Marcellus, A. Roulston, M. Watson, L. Serfass, S.R. Murthy Madiraju, D. Goulet, J. Viallet, L. Belec, X. Billot, S. Acoca, E. Purisima, A. Wiegmans, L. Cluse, R.W. Johnstone, P. Beauparlant, G.C. Shore, Small molecule obatoclax (GX15-070) antagonizes MCL-1 and overcomes MCL-1-mediated resistance to apoptosis, Proc. Natl. Acad. Sci. USA 104 (2007) 19512–19517. [29] J.K. Park, W.H. Paik, J.K. Ryu, Y.T. Kim, Y.J. Kim, J. Kim, B.J. Song, J.M. Park, Y.B. Yoon, Clinical significance and revisiting the meaning of CA 19-9 blood level

28

[30]

[31]

[32]

[33]

[34]

[35]

[36] [37]

[38]

[39]

S. Chen et al. / Cancer Letters 348 (2014) 20–28 before and after the treatment of pancreatic ductal adenocarcinoma: analysis of 1446 patients from the pancreatic cancer cohort in a single institution, PLoS ONE 8 (2013) e78977. V.R. Penchev, Z.A. Rasheed, A. Maitra, W. Matsui, Heterogeneity and targeting of pancreatic cancer stem cells, Clin. Cancer Res.: Off. J. Am. Assoc. Cancer Res. 18 (2012) 4277–4284. R. Plummer, Perspective on the pipeline of drugs being developed with modulation of DNA damage as a target, Clin. Cancer Res.: Off. J. Am. Assoc. Cancer Res. 16 (2010) 4527–4531. L. Porcelli, A.E. Quatrale, P. Mantuano, M.G. Leo, N. Silvestris, J.F. Rolland, E. Carioggia, M. Lioce, A. Paradiso, A. Azzariti, Optimize radiochemotherapy in pancreatic cancer: PARP inhibitors a new therapeutic opportunity, Mol. Oncol. 7 (2013) 308–322. M. Rahmani, M.M. Aust, E. Attkisson, D.C. Williams Jr., A. Ferreira-Gonzalez, S. Grant, Inhibition of Bcl-2 antiapoptotic members by obatoclax potently enhances sorafenib-induced apoptosis in human myeloid leukemia cells through a Bim-dependent process, Blood 119 (2012) 6089–6098. M. Rajesh, P. Mukhopadhyay, S. Batkai, G. Godlewski, G. Hasko, L. Liaudet, P. Pacher, Pharmacological inhibition of poly(ADP-ribose) polymerase inhibits angiogenesis, Biochem. Biophys. Res. Commun. 350 (2006) 352–357. M. Rajesh, P. Mukhopadhyay, G. Godlewski, S. Batkai, G. Hasko, L. Liaudet, P. Pacher, Poly(ADP-ribose)polymerase inhibition decreases angiogenesis, Biochem. Biophys. Res. Commun. 350 (2006) 1056–1062. E. Reed, DNA damage and repair in translational oncology: an overview, Clin. Cancer Res.: Off. J. Am. Assoc. Cancer Res. 16 (2010) 4511–4516. D. Rossetto, A.W. Truman, S.J. Kron, J. Cote, Epigenetic modifications in doublestrand break DNA damage signaling and repair, Clin. Cancer Res.: Off. J. Am. Assoc. Cancer Res. 16 (2010) 4543–4552. Y. Saintigny, A. Dumay, S. Lambert, B.S. Lopez, A novel role for the Bcl-2 protein family: specific suppression of the RAD51 recombination pathway, EMBO J. 20 (2001) 2596–2607. A. Strimpakos, M.W. Saif, K.N. Syrigos, Pancreatic cancer: from molecular pathogenesis to targeted therapy, Cancer Metastasis Rev. 27 (2008) 495–522.

[40] R. Sultana, D.R. McNeill, R. Abbotts, M.Z. Mohammed, M.Z. Zdzienicka, H. Qutob, C. Seedhouse, C.A. Laughton, P.M. Fischer, P.M. Patel, D.M. Wilson 3rd, S. Madhusudan, Synthetic lethal targeting of DNA double-strand break repair deficient cells by human apurinic/apyrimidinic endonuclease inhibitors, Int. J. Cancer 131 (2012) 2433–2444. [41] A. Tutt, M. Robson, J.E. Garber, S.M. Domchek, M.W. Audeh, J.N. Weitzel, M. Friedlander, B. Arun, N. Loman, R.K. Schmutzler, A. Wardley, G. Mitchell, H. Earl, M. Wickens, J. Carmichael, Oral poly(ADP-ribose) polymerase inhibitor olaparib in patients with BRCA1 or BRCA2 mutations and advanced breast cancer: a proof-of-concept trial, Lancet 376 (2010) 235–244. [42] K. Valerie, L.F. Povirk, Regulation and mechanisms of mammalian doublestrand break repair, Oncogene 22 (2003) 5792–5812. [43] G. Wang, J. He, J. Zhao, W. Yun, C. Xie, J.W. Taub, A. Azmi, R.M. Mohammad, Y. Dong, W. Kong, Y. Guo, Y. Ge, Class I and class II histone deacetylases are potential therapeutic targets for treating pancreatic cancer, PLoS ONE 7 (2012) e52095. [44] Q. Wang, F. Gao, W.S. May, Y. Zhang, T. Flagg, X. Deng, Bcl2 negatively regulates DNA double-strand-break repair through a nonhomologous endjoining pathway, Mol. Cell 29 (2008) 488–498. [45] C. Xie, H. Edwards, X. Xu, H. Zhou, S.A. Buck, M.L. Stout, Q. Yu, J.E. Rubnitz, L.H. Matherly, J.W. Taub, Y. Ge, Mechanisms of synergistic antileukemic interactions between valproic acid and cytarabine in pediatric acute myeloid leukemia, Clin. Cancer Res.: Off. J. Am. Assoc. Cancer Res. 16 (2010) 5499– 5510. [46] X. Xu, C. Xie, H. Edwards, H. Zhou, S.A. Buck, Y. Ge, Inhibition of histone deacetylases 1 and 6 enhances cytarabine-induced apoptosis in pediatric acute myeloid leukemia cells, PLoS ONE 6 (2011) e17138. [47] C. Ziske, C. Schlie, M. Gorschluter, A. Glasmacher, U. Mey, J. Strehl, T. Sauerbruch, I.G. Schmidt-Wolf, Prognostic value of CA 19-9 levels in patients with inoperable adenocarcinoma of the pancreas treated with gemcitabine, Br. J. Cancer 89 (2003) 1413–1417.