PARP-1 Regulates Resistance of Pancreatic Cancer to TRAIL Therapy

Pancreatic cancer is an important therapeutic challenge with nearly similar incidence and mortality rates. It is the most lethal cancer of the digestive system with a 5-year survival rate of 5% (1). Because of the resistance of pancreatic cancer to chemotherapy and radiotherapy, current treatments for pancreatic cancer are generally not curative with modest benefit (2, 3). We and others have shown that modulating death receptor (DR)-activated extrinsic apoptosis pathways may provide potential novel avenues for cancer therapy (4, 5). Upon ligand stimulation, DRs (Fas or death receptor 4/5, DR4/5) recruit Fas-associated death domain (FADD) and the initiator caspases, caspase-8 or caspase-10, resulting in formation of the death-inducing signaling complex (DISC). The recruited caspase-8 or caspase-10 undergoes autocatalytic cleavage and activation to trigger the caspase cascade that ultimately leads to apoptosis (4, 5). We have previously reported that regulating the DISC formation promotes Fas DR–induced apoptosis of cancer cells, including pancreatic cancer cells (6–10).

Activating the DR4/5-mediated extrinsic apoptotic signaling pathway by TNF-related apoptosis-induced ligand (TRAIL) induces apoptosis of a variety of cancer cells including melanoma, colon, lung, breast, kidney, and pancreatic cancers (11, 12). With virtually no toxicity for normal cells, recombinant TRAIL or agonistic antibodies specifically targeting DR4/5 are currently being tested in several phase I–III clinical trials (12–14). Although TRAIL is a promising anticancer agent, many cancer cells, including pancreatic cancer cells, develop resistance to TRAIL-induced apoptosis. The resistance of cancer cells to TRAIL-induced apoptosis has been associated with mutations of DR5 receptor (15), increased expression of the decoy DRs DcR1 or DcR2 (16, 17), decreased expression of apoptotic protein caspase-8 (18), and increased expression of a number of antiapoptotic proteins, such as Bcl-2, Bcl-XL (19, 20) and the FLICE-like inhibitory protein (FLIP; refs. 21, 22).

Combination therapy with other anticancer agents has been shown to enhance the efficacy of TRAIL treatment for the resistant cancers (23–25). We and others have found that activating intrinsic apoptotic signaling pathways using gemcitabine (26), doxorubicin (27), or ionizing radiation (28) sensitizes resistant cancer cells to DR-induced apoptosis. These intrinsic apoptotic activators function via inducing mitochondria leakage or DNA damage (26–28).

The PARPs are key to repairing damaged DNA by modifying proteins associated with DNA repair (29). Inhibition of PARPs results in accumulation of irreparable damaged DNA and facilitates triggering of the intrinsic apoptosis signaling pathway in cancer therapy (29). PARP-1 is the most abundant of PARPs, accounting for 75% to 90% of the PARPs (29). Inhibition of PARP-1 through siRNA knockdown or chemical inhibitors for PARP has been used to treat cancers with DNA repair deficiencies, such as BRCA-deficient breast and ovarian cancer (30) as well as BRCA2-associated pancreatic cancer (31). In addition, combined use of PARP inhibitors with DNA-damaging reagents or radiation sensitizes glioma, ovarian, and pancreatic cancers to therapy (32–34).

The role of PARP-1 in DR-mediated extrinsic apoptotic signaling pathways is unknown. The present study provides the first evidence that PARP-1 inhibition sensitizes resistant pancreatic cancer cells to DR5-activated apoptosis and enhances the efficacy of DR5 antibody therapy for pancreatic cancers in a mouse tumor xenograft model. Furthermore, we have uncovered a novel mechanism of PARP-1 in regulating DR5-induced DISC recruitment and subsequent caspase-8 activation, which is independent of its known function in DNA repair. Our studies provide novel molecular insights into the function PARP-1 in the DR-activated apoptosis machinery and support the use of combined therapy with DR5 agonists and PARP inhibitors for resistant pancreatic tumors.

The human pancreatic cancer cell lines BxPc-3, MiaPaCa-2, PANC-1, and Suit-2 were from the American Type Culture Collection. BxPc-3 and Suit-2 cells were grown in RPMI-1640; MiaPaCa-2 and PANC-1 cells were grown in Dulbeccos' Modified Eagle's Media (DMEM; Invitrogen) supplemented with penicillin (5 units/mL), streptomycin (5 μg/mL), and 10% heat-inactivated FBS.

DR5 agonist antibody TRA-8 was generated as previously described (35). All antibodies used were commercially available, including FADD (Millipore), caspase-3 and pADPr (Enzo Life), caspase-8 (BD Bioscience and Cell Signaling Technology), PARP-1 and GAPDH (Santa Cruz Biotech), and DR5 (Prosci).

Protein G agarose was from Invitrogen. Recombinant human PARP-1 protein was purchased from Enzo. β-NAD was from Sigma-Aldrich. Active human recombinant caspase-8 was from Research & Diagnostics. Full-length mouse recombinant caspase-8 was from Abcam.

Lentiviral constructs expressing a 21-nucleotide short hairpin RNA (shRNA) targeting human PARP-1 gene (GenBank accession number NM_001618.3) or human FADD gene (GenBank accession number NM_003824.3) were purchased from Open Biosystems. The constructs were packaged into lentivirus-like particles pseudotyped with the vesicular stomatitis virus glycoprotein as we previously described (10). Transduction was conducted by incubating cancer cells with lentivirus, and stably transduced cells were selected with puromycin (2 μg/mL).

Apoptosis was determined by Annexin V-fluorescein isothiocyanate (FITC) and propidium iodide (PI) staining (BD Biosciences) and analyzed by flow cytometry (BD Biosciences).

Proteins were extracted, quantified with a BCA protein assay kit (Thermo Scientific), separated by SDS-PAGE, and transferred to Immobilon P membranes (Millipore) as described previously (10). Membranes were blocked in 5% non-fat milk and incubated with primary antibodies overnight at 4°C and then with horseradish peroxidase–conjugated secondary antibodies for 1 hour at room temperature. Signals were detected using the Immobilon Western Chemiluminescent Horseradish Peroxidase Substrate Detection Kit (Millipore).

Five hundred micrograms of extracted proteins was incubated with 1 μg of antibodies as indicated in figure legends for 1 hour and subsequently incubated with 50 μL of 1:1 slurry of protein G-agarose beads overnight at 4°C. Beads were then washed, and 20 μL 2×Laemmli sample buffer was added to the beads followed by heating at 95°C for 5 minutes and chilling on ice. After brief centrifugation, proteins in the supernatant were analyzed by Western blotting with specific antibodies.

Recombinant full-length mouse Caspase-8 (300 ng, Abcam) was immunoprecipitated using antibody for caspase-8. Immunoprecipitated beads were incubated in 100 μL reaction buffer [100 mmol/L Tris-HCl (pH 8), 10 mmol/L MgCl₂, 1 mmol/L dithiothreitol] with or without 100 ng of PARP-1 and 500 μmol/L β-NAD for 30 minutes at 37°C and then washed 3 times.

For caspase-8 activation analysis, immunoprecipitated beads were incubated in 100 μL reaction buffer [20 mmol/L HEPES (pH 7.5), 5 mmol/L dithiothreitol] with or without 100 ng of recombinant active human caspase-8 (Research & Diagnostic) at 37°C for 60 minutes. Beads were then washed and 20 μL 2×Laemmli sample buffer was added to the beads followed by heating at 95°C for 5 minutes and chilling on ice. After brief centrifugation, proteins in the supernatant were analyzed by Western blotting with specific antibodies.

The animal protocol was approved by the Institutional Animal Care and Use Committee at the University of Alabama at Birmingham (Birmingham, AL). Male athymic nu/nu mice (6 weeks old, NCI-Frederick) were used for tumor inoculation. Briefly, control PANC-1 pancreatic cancer cells or PANC-1 cells with PARP-1 knockdown were inoculated subcutaneously into the flanks of mice (2 × 10⁶ cells in 200 mL PBS/site). Three weeks following tumor inoculation, mice were divided into 2 groups (8 mice per group), mice in the treatment group were intraperitoneally injected with TRA-8 once a week for 6 weeks (200 μg/mice), and mice in the control group were injected with 0.9% sodium chloride. Tumor size and body weight were measured every week, and volumes were determined using the formula volume = length × width²/2. At the end of the experiment, tumors were removed from mice, half of each tumor was homogenized for Western blot analysis and the other half was fixed in 4% paraformaldehyde and embedded in paraffin.

TUNEL staining was conducted on tumor sections (7 μm; DeadEnd Fluorometric TUNEL System; Promega) to determine cell death and DAPI staining (4′,6-diamidine-2-phenylindole dihydrochloride) was used to identify nuclei. Stained specimens were examined microscopically (Leica M165 FC). For quantitative analysis, cell numbers were counted under a microscope (×200). Four fields in each slide were counted and the percentage of apoptotic cells was determined.

Results are expressed as means ± SD. Differences between 2 groups were identified with the Student t test. For multiple groups, one-way ANOVA and Student–Newman–Keuls tests were conducted to identify differences. Significance was defined as P < 0.05.

Using TRA-8, a monoclonal agonistic antibody specifically targeting DR5 (35), we determined DR-activated apoptosis in four pancreatic cell lines. BxPc-3 and MiaPaCa-2 cells were highly sensitive to TRA-8–induced apoptosis, whereas PANC-1 and Suit-2 cells are resistant to TRA-8 (Fig. 1A). To understand the underlying mechanisms, we analyzed the expression of the receptor DR5, as well as the expression of an antiapoptotic protein FLIP (Fig. 1B) that has been shown to inhibit DR5-induced apoptosis (36). The expression of DR5 is higher in MiaPaCa-2 and Suit2 cells compared with that in BxPc-3 and PANC-1 cells, which was not correlated with their sensitivity to TRA-8–induced apoptosis. Similarly, the expression level of FLIP was not correlated to the resistance of the cells to TRA-8–induced apoptosis, suggesting that additional mechanisms may contribute to TRA-8 resistance.

We found that the expression of PARP-1 was correlated with TRA-8 resistance of the pancreatic cancer cells. The expression of PARP-1 was markedly higher in the TRA-8–resistant cell PANC-1 and Suit-2 cells, compared with that in the TRA-8–sensitive BxPc-3 and MiaPaCa-2 cells (Fig. 1B), suggesting a role of PARP-1 in TRA-8 resistance. Using a pharmacologic inhibitor for PARPs, PJ-34, we found that inhibition of PARP activity enhanced TRA-8–induced apoptosis in PANC-1 and Suit-2 cells in a concentration-dependent manner (Fig. 1C). Such an effect of PJ-34 was not due to its toxicity, as PJ-34 did not induce apoptosis at all concentrations used (0–40 μmol/L). Consistent with its effects on TRA-8–induced apoptosis, PJ-34 dramatically enhanced TRA-8–induced activation of caspase-8 and its downstream apoptotic effector, caspase-3 (Fig. 1D).

The specific role of PARP-1 in mediating TRA-8 resistance of pancreatic cancer cells was further determined using shRNAs specific for PARP-1. Two lines of PANC-1 cells with PARP-1 knockdown were generated using shRNAs specifically targeting different regions of the PARP-1 gene. Knockdown of PARP-1 by the 2 shRNAs was confirmed by Western blot analysis (Fig. 2A, inserts). Knockdown of PARP-1 per se did not affect cell viability (Fig. 2A, white bars) but significantly increased the sensitivity of PANC-1 cells to TRA-8–induced apoptosis (Fig. 2A, black bars). Similar to the observations with PJ-34 (Fig. 1D), PARP-1 knockdown increased TRA-8–induced activation of caspase-8 and caspase-3 (Fig. 2B).

To determine the effects of PARP-1 knockdown on the efficacy of TRA-8 therapy on tumorigenesis in vivo, we characterized tumorigenesis and response to TRA-8 treatment of PARP-1 knockdown PANC-1 cells in a mouse xenograft model. PANC-1 cells with scrambled shRNAs were used as controls. In the control groups, TRA-8 treatment did not reduce tumor size, further showing resistance of these cancer cells to TRA-8 therapy (Fig. 3A). In contrast, TRA-8 treatment effectively reduced tumor size of the PARP-1–knockdown cells (Fig. 3A). Western blot analysis determined that the expression of PARP-1 was maintained at a lower level in xenograft tumors of PARP-1–knockdown cells, which was not affected by TRA-8 treatment (Fig. 3B). Activation of caspase-8 and its downstream apoptosis effector caspase-3, indicated by cleaved caspase-8 and caspase-3, was shown in PARP-1–knockdown tumors from the TRA-8–treated mice (Fig. 3B). Furthermore, TUNEL staining of tumor sections showed significant increases in cell death in the PARP-1–knockdown tumors from TRA-8–treated mice (Fig. 3C).

Because inhibition of PARP-1, by PJ-34 (Fig. 1) or by shRNAs (Fig. 2), increased TRA-8–induced activation of the upstream apoptosis initiator, caspase-8, we determined whether inhibition of caspase-8 may block the effects of PARP-1 inhibition. Z-IETD-FMK, a caspase-8 inhibitor, blocked TRA-8–induced apoptosis in the control PANC-1 cells. Moreover, inhibition of caspase-8 significantly reduced TRA-8–induced apoptosis in the PARP-1–knockdown cells (Fig. 4A). Consistently, caspase-8 inhibition resulted in marked decreases in TRA-8–induced cleavage/activation of caspase-8 and its downstream caspase-3 in the PARP-1–knockdown cells (Fig. 4B).

Furthermore, knockdown of PARP-1 did not affect the expression of the death receptor DR5, the adaptor protein FADD, or the procaspase-8 (Fig. 4C). As procaspase-8 is known to be proteolytically cleaved and activated in the DR5-induced DISC, we analyzed DR5-recrutied DISC proteins, including FADD and caspase-8, in response to TRA-8 stimulation. PARP-1 knockdown did not affect the recruitment of FADD into DR5-associated DISC in response to TRA-8 stimulation (Fig. 4C). As expected, TRA-8 induced the recruitment and activation of caspase-8 into the DR5-associated DISC in the control cells (Fig. 4C, shScr). TRA-8–induced DR5 recruitment of procaspase-8 appeared to be similar in the control and PARP-1–knockdown cells. However, increased activation of caspase-8 was shown in the PARP-1–knockdown cells (Fig. 4C, p41/43 and p24/26). Furthermore, PARP-1 was identified in the DR5-associated complex in the control cells under the basal condition, which was not affected by TRA-8 treatment (Fig. 4C, PARP-1).

A key function of PARP-1 is to modify proteins by adding poly(ADP-ribose) (pADPr) chain to target proteins. Accordingly, we analyzed pADPr modification profile of the DR5-associated DISC proteins. Using anti-pADPr antibody, we determined that knockdown of PARP-1 significantly reduced overall protein pADPr modification in the PANC-1 cells (Fig. 5A). TRA-8 treatment did not affect overall protein pADPr modification profile (Fig. 5A, pADPr). However, increased pADPr modification on DISC components, DR5, FADD, and caspase-8, was detected in the TRA-8–treated control cells (Fig. 5A, shScr). Furthermore, PARP-1 knockdown decreased TRA-8–induced pADPr modification on DR5, FADD, and caspase-8 (Fig. 5A, PARP-1 KD).

To further determine that pADPr modification of DR5, FADD, and caspase-8 occurred in the DR5-associated DISC, we knocked down FADD, the adaptor protein of DISC formation, with shRNA specifically targeting FADD (Fig. 5B). Knockdown of FADD did not affect the overall protein pADPr modification profile but decreased TRA-8–induced pADPr modification on DR5 and caspase-8 (Fig. 5B, pADPr), indicating that TRA-8–induced pADPr modification on DR5 and caspase-8 is dependent on FADD-mediated DISC formation (Fig. 5B).

Using purified recombinant PARP-1 and caspase-8 proteins, a direct effect of PARP-1 on pADPr modification of caspase-8 was shown (Fig. 6A). Furthermore, the effect of pADPr modification of caspase-8 on its activation was characterized. In the absence of PARP-1, active caspase-8 cleaved procaspase-8 into cleaved p10/18 fragments (Fig. 6B). However, decreased caspase-8 cleavage was observed when PARP-1 was added (Fig. 6B, last lane), indicating that pADPr modification of caspase-8 inhibited caspase-8 activation.

TRAIL therapies have shown low toxicity in patients in several clinical trials (12–14). However, therapeutic effects of TRAIL agonists as a monotherapy have been limited due to tumor resistance (12–14). Experimental studies have postulated several mechanisms that are responsible for TRAIL resistance, including low expression of the death receptors DR4/DR5, high expression of decoy receptor DcR1/DcR2, and increased expression of antiapoptotic molecular such as FLIP (12). However, a consistent correlation between the expression of DR4/DR5 and TRAIL sensitivity of cancer cells has not been determined (12). We showed in the present study that the expression of DR5 was not correlated with the sensitivity of pancreatic cancer cells to TRA-8–induced apoptosis. TRA-8 is a monoclonal antibody that specifically targets DR5 but not DR4 or the decoy receptors DcR1 and DcR2 (35). Thus resistance of the pancreatic cancer cells to TRA-8–induced apoptosis is independent of the expression of the decoy receptors. In addition, the expression level of the antiapoptotic protein FLIP was not correlated to TRA-8 resistance of the pancreatic cancer cells, although we and others have shown a protective effect of FLIP on DR-induced apoptosis in several cancer cells (6, 21, 22).

We found that the expression of PARP-1 contributed to the TRA-8 resistance of pancreatic cancer cells. As the predominant member of the PARP family that regulates protein pADPr modification, PARP-1 is an important regulator of the DNA base excision repair pathway and is essential for the maintenance of genomic integrity and cell survival in response to genotoxic insults (37, 38). PARP-1 elevation is associated with adverse outcomes of breast and ovarian cancers (39, 40). Consistently, we found that the expression of PARP-1 was increased in 2 TRA-8–resistant pancreatic cell lines, PANC-1 and Suit-2 cells, and inhibition of PARP-1 rendered the resistant cells more sensitive to TRA-8–induced apoptosis, showing the engagement of PARP-1 in TRA-8–resistant pancreatic cancer cells.

Consistent with its essential role in DNA repair, PARP-1 antagonists, used as a monotherapy for tumors with DNA repair deficiencies or in combination with DNA damage–inducing agents, increase accumulation of damaged DNA that triggers intrinsic apoptotic pathways (31–34). In pancreatic cancer, PARP inhibitors enhance the effects of chemotherapy with DNA-alkylating agents or radiation therapy, both via promoting DNA damage–induced intrinsic apoptotic signaling pathways. The present studies have provided the first evidence that PARP-1 inhibition alone did not induce apoptosis of pancreatic cancer cells but enhanced TRA-8–induced apoptosis, suggesting that PARP-1 regulates TRA-8–induced extrinsic apoptosis via a mechanism independent of its function in DNA repair.

The expression of PARP-1 was identified in both nucleus and cytoplasm of the PANC-1 and Suit-2 cells (data not shown). PARPs are mainly localized in the nucleus, where they execute the DNA repair function (29). Recently, PARP-1 has also been identified in the cytoplasm of neurons (41) and smooth muscle cells (42), as well as in the mitochondria of fibroblasts and HeLa cells (43). In breast cancer, expression of cytoplasmic PARPs is correlated with tumor aggressiveness, higher risk for relapse and death (44). However, the function of cytoplasmic PARP-1 in cancer is not clear. PARP-1 is capable of modifying proteins via pADPr (29, 32). Chang and colleagues have shown that pADPr modification of the spindle pole protein NuMA in the cytoplasm by PARP-1 is required for assembly, maintenance of bipolar spindle structure, as well as the function of the mitotic spindle (45). Our studies showed the presence of PARP-1 in a new cellular location in pancreatic cancer cells, the DR-associated DISC, which enhanced pADPr modification of DISC proteins. PARP-1 did not appear to modify DR5 directly, although an association between PARP-1 and DR5 was shown under basal conditions. In response to TRA-8 treatment, pADPr modification of the DISC complex, DR5/caspase-8/FADD, was identified, which was dependent on the DISC recruitment of FADD. Therefore, PARP-1 may not directly modify DR5 at basal conditions but regulates the pADPr modification of the DR5-associated DISC complex and thus modulating downstream apoptotic signaling.

Our studies revealed that PARP-1 modified caspase-8 and determined a novel function of PARP-1 in regulating pADPr modification of caspase-8, thereby inhibiting caspase-8 activation. Such an observation is consistent with previous studies showing that pADPr modification inhibits protein function. For instance, pADPr modification of NFκB(p65) decreases its interaction with the nuclear export factor and thus retains it in the nucleus (46). The activation of procaspase-8 is believed to occur in the DR-activated DISC, which provides a platform that allows 2 procaspase-8 homodimers to be in close proximity necessary to initiate cleavage and activation of procaspase-8 (47–49). By the same token, the presence of PARP-1 in the DR5-associated complex provides the proximity for PARP-1 to modify (pro)caspase-8 in the DISC. This concept is supported by several lines of evidence: (i) pADPr modification of caspase-8 was identified in the DR5-associated DISC induced by TRA-8; (ii) inhibition of DISC formation by FADD knockdown blocked pADPr modification on caspase-8; and (iii) recombinant PARP-1 induced pADPr modification of caspase-8 and inhibited caspase-8 cleavage/activation directly. We found PARP-1 inhibited the cleavage of procaspase-8 into the p18/p10 active form (Fig. 6B). The amino acids that are critical for the cleavage of caspase-8 into the p18/p10 active form are the aspartate residues at 216, 374, and 384 (47, 48). Because aspartate is one of the PARP-1 targeting modification residues (29), pADPr modification on these caspase-8 aspartate residues may contribute to the inhibition of caspase-8 cleavage and activation. In support of this concept, when PARP-1 was inhibited, TRA-8–induced activation of caspase-8 was increased, and TRA-8–induced apoptosis was enhanced. This is also supported by the observation that caspase-8 inhibitor blocked the effect of PARP-1 inhibition on TRA-8–induced apoptosis.

In summary, we have found that PARP-1 knockdown sensitized resistant pancreatic cancer cells to TRA-8–induced apoptosis in vitro and enhanced the efficacy of TRA-8 therapy of pancreatic tumors in vivo. PARP-1 was identified in the DR5-associated complex, which induced pADPr modification of caspase-8 in the TRA-8–activated DR5-associated DISC and modulated caspase-8 activation. Therefore, inhibition of PARP-1 blocks pADRr modification of caspase-8, which facilitates DR5-mediated activation of caspase-8 in the DISC, and thus sensitizing resistant pancreatic cancer cells to TRA-8–induced apoptosis (Fig. 7). These studies show a novel function of PARP-1 in the DR-activated apoptotic signaling machinery, which provides important molecular insights into the mechanisms underlying resistance of pancreatic cancer to TRAIL therapy. Results from the present studies support interventions combining PARP-1 inhibitors with DR agonists for pancreatic cancer therapy.

No potential conflicts of interest were disclosed.

Conception and design: K. Yuan, T. Zhou, J.M. McDonald, Y. Chen

Development of methodology: K. Yuan, Y. Sun, Y. Chen

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Yuan, Y. Sun

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K. Yuan, Y. Sun, J.M. McDonald, Y. Chen

Writing, review, and/or revision of the manuscript: K. Yuan, J.M. McDonald, Y. Chen

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Yuan, Y. Sun, T. Zhou, J.M. McDonald

Study supervision: J.M. McDonald, Y. Chen

This work was supported in part by UAB Department of Pathology start-up funds (to Y. Chen) and a VA Merit Review Award (to J.M. McDonald). T. Zhou was support by a NIH grant P50CA101955.

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