Mechanism of In vitro Pancreatic Cancer Cell Growth Inhibition by Melanoma Differentiation–Associated Gene-7/Interleukin-24 and Perillyl Alcohol

Pancreatic ductal adenocarcinoma is the predominant form of pancreatic cancer and one of the most lethal and aggressive human malignancies. Approximately 37,000 new cases are diagnosed annually in the United States alone, which is virtually the same number of deaths reported to occur every year due to metastatic complications (an estimate of the National Cancer Institute shows this number to be 33,370 for 2007). Meaningful therapy in the form of surgical resection is feasible only with early-stage detection, but this applies to just <20% of patients and results in a 5-year survival of <20% (1, 2). For the vast majority of patients, the overall 5-year survival is <5%, which is attributable to multiple factors, including the plethora of molecular genetic changes contributing to the cellular phenotype of pancreatic neoplasms that are believed to produce resistance to chemotherapy and radiation therapy. This medical problem is further compounded by the lack of therapeutic approaches targeting metastatic disease. Considering these terrible statistics, it is imperative to develop rational molecular target-based preventative and therapeutic strategies for this consistently fatal disease.

Using subtraction hybridization, we discovered genes, originally called melanoma differentiation associated (mda), which encode molecules that are now better defined in terms of their physiologic and pathologic importance and have been shown to play crucial roles in cell cycle control, metastasis, senescence, differentiation, innate immunity, and apoptosis (3). One of these originally novel genes, mda-7, displayed an inverse relationship with the pathogenesis and progression of melanoma (4). We showed that mda-7 manifests tumor suppressor–like functions in vitro and in vivo in athymic human tumor xenograft animal models (5). Ectopic expression of mda-7 (by transfection or by adenovirus transduction) exerts potent growth-suppressive and apoptosis-inducing effects not only in human melanoma cells but also in a wide spectrum of human cancer cells, including malignant glioma, osteosarcoma, mesothelioma, and carcinomas of the breast, cervix, colon, lung, ovary, and prostate (reviewed in refs. 6–8). Based on structure, chromosomal location, and biological/biochemical properties, mda-7 has now been classified as a novel member of the interleukin (IL)-10 gene family, IL-24 (6–8). Data from multiple laboratories including our own indicate that mda-7/IL-24, a secreted cytokine, which also manifests “bystander” antitumor activity, can retard tumor growth by impinging on several critical signaling pathways resulting in tumor apoptosis as well as by inhibiting tumor angiogenesis and modulating immune responses (8–10). Remarkably, in the context of normal cells/tissues, no such growth-suppressive or cytotoxic effects are evident (reviewed in refs. 6–9). Considering these intriguing differential properties in tumor versus normal cells, mda-7/IL-24 was evaluated for its in vivo efficacy using several human tumor xenograft murine models. These results confirmed potent selective antitumor activity of this cytokine and prompted testing [using a replication-incompetent adenovirus, Ad.mda-7 (INGN 241)] in patients with advanced carcinomas and melanoma (11–13). A recent phase I clinical trial was highly promising, confirming the retention of tumor-specific activity in patients. More impressively, mda-7/IL-24 was well tolerated and showed no adverse effects in these patient populations (11–13). Based on these data, investigations are now in progress to further evaluate the potential therapeutic applications of this novel cytokine for cancer gene therapy in phase II clinical trials.

Although effective in virtually all human tumor cells tested, pancreatic cancer cells represent an enigma, being inherently resistant to Ad.mda-7–based therapy (14, 15). Transfection or adenovirus transduction of mda-7/IL-24 under conditions in which growth suppression and apoptosis occur in various other human tumor cells was without activity (14, 15). An interesting mechanism underlies this resistance, limited conversion of mda-7/IL-24 mRNA into protein in pancreatic tumor cells due to a “protein translational block” targeting this mRNA (14, 15). This protein translational block in mutant K-ras pancreatic cancer cells could be reversed by ablating K-ras expression as well as by inhibiting its downstream target, extracellular signal-regulated kinase (ERK) 1/2, resulting in growth arrest and apoptosis in pancreatic cancer cells in an analogous manner as observed in other permissive cancer cells (14, 15). We also discovered that augmented generation of reactive oxygen species (ROS) could restore mda-7/IL-24 mRNA translation in pancreatic cancer cells irrespective of their K-ras status (16). Although of potential clinical interest, many agents that could be used with Ad.mda-7 for gene therapy of pancreatic carcinomas also manifest nonspecific toxicity, thereby limiting their utility.

We hypothesized that specific nontoxic dietary agent(s) with ROS-inducing properties might be agents of choice that would complement Ad.mda-7 action, providing a safe combination for human use with potential to abolish the pathogenesis of pancreatic cancer. We presently evaluated perillyl alcohol (POH), a dietary monoterpene present in a variety of plants, including citrus plants, for potential synergy of action with Ad.mda-7. Burke and colleagues (17, 18) showed that dietary monoterpenes, farnesol, geraniol, and POH, block growth of transplanted PC-1 hamster pancreatic adenocarcinomas in Syrian Golden hamsters potentially by their ability to inhibit the prenylation of growth-regulatory proteins other than K-ras, including H-ras. Karlson and colleagues (19) showed that POH is effective in inhibiting the growth of pancreatic tumor cells harboring activated ras oncogenes, acting in a Ras-independent manner. Stark and colleagues (20) reported chemotherapeutic effects of POH on pancreatic cancer. They found that POH reduces the growth of hamster pancreatic tumors to less than half that of controls. Moreover, 16% of POH-treated pancreatic tumors completely regressed, whereas no control tumors regressed (20). Published data suggest that POH blocks several important signaling pathways that include ras and ERK in addition to nuclear factor-κB (21–24). POH also prevents the isoprenylation of the Ras family of small GTPase proteins (25, 26). Additionally, the pharmacokinetics of POH has been defined in both murine models and humans (27, 28), and phase I and II clinical trials have been conducted, which reveal that this dietary agent is well tolerated (23, 29). Using a battery of human pancreatic carcinoma cells in culture, we document that the combination of POH and Ad.mda-7 efficiently reversed the MDA-7/IL-24 protein translational block in pancreatic cancer cells by generating ROS, via induction of xanthine oxidase (XO), resulting in pancreatic cancer growth suppression and apoptosis. These studies elucidate a new combinatorial approach for pancreatic cancer involving a dietary supplement and a viral gene therapy that may show promise for treating this devastating cancer.

Cell lines. AsPC-1, MIA PaCa-2, PANC-1, and BxPC-3 pancreatic carcinoma cells were obtained from the American Type Culture Collection. The human immortalized pancreatic mesenchymal cell line LT2 was purchased from Chemicon and cultured according to the manufacturer's instructions. Stable clones of PANC-1 and MIA PaCa-2 pancreatic carcinoma cells expressing mda-7/IL-24 mRNA were obtained by transfection of the corresponding cells with an mda-7/IL-24 expression vector and selection of clones in hygromycin as described previously (16). Northern and Western blotting, respectively, confirmed expression of mda-7/IL-24 mRNA in these clones without detectable protein.

Virus construction, purification, and infectivity assays. The recombinant replication-defective Ad.mda-7 virus was created in two steps as described previously (5) and plaque purified by standard procedures. Cells were infected with 100 plaque-forming units (pfu)/cell of Ad.vec or Ad.mda-7 viruses (50 pfu/cell of each virus) and analyzed as described.

RNA isolation and Northern blot assays. Total RNA was extracted from cells using the Qiagen RNeasy Mini kit according to the manufacturer's protocol and was used for Northern blotting as previously described (5, 15, 30).

Purification of polysomes, RNA extraction, and Northern blot analysis. Polysomes were purified essentially as described previously (31). Cells (2 × 10⁶) were infected with adenovirus, and 48 h later, the cells were harvested in 500 μL Buffer A [200 mmol/L Tris-HCl (pH 8.5), 50 mmol/L KCl, 25 mmol/L MgCl₂, 2 mmol/L EGTA, 100 μg/mL heparin, 2% polyoxyethylene 10-tridecyl ether, and 1% sodium deoxycholate supplemented with Complete Mini protease inhibitor cocktail and RNase inhibitor, Invitrogen] and centrifuged at 12,000 rpm for 10 min at 4°C to clear cell debris. The supernatant was loaded on top of a 10% to 50% sucrose gradient prepared in Buffer B (50 mmol/L Tris-HCl, 25 mmol/L KCl, and 10 mmol/L MgCl₂) and was centrifuged at 40,000 rpm for 1 h at 4°C. Fractions of 500 μL were collected, the absorbance at 260 nm was monitored, and polysome fractions were identified (typically fractions 10–20). RNA was extracted from each fraction with Qiagen RNeasy Mini kit according to the manufacturer's protocol and Northern blotting was performed as described using a radiolabeled mda-7/IL-24 cDNA probe.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide viability assays. Cell viability was assessed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays as described previously (32).

Annexin V binding assays. Cells were trypsinized, washed once with complete medium and PBS, resuspended in 0.5 mL of binding buffer containing 2.5 mmol/L CaCl₂, and stained with allophycocyanin-labeled Annexin V (BD Biosciences) and propidium iodide (PI) for 15 min at room temperature. Flow cytometry assays were performed immediately after staining using FACS Calibur (Becton Dickinson). Data were analyzed using CellQuest software, version 3.1 (Becton Dickinson).

Preparation of cell extracts and Western blotting analysis. Cells were washed twice with cold PBS and lysed on ice for 30 min in 100 μL of cold radioimmunoprecipitation assay buffer [50 mmol/L Tris-HCl (pH 8.0), 150 mmol/L NaCl, 0.1% SDS, 1% NP40, and 0.5% sodium deoxycholate] with freshly added 0.1 mg/mL phenylmethylsulfonyl fluoride, 1 mmol/L sodium orthovanadate, and 1 mg/mL aprotinin. Aliquots of cell extracts containing 20 to 50 μg of total protein were resolved in 12% SDS-PAGE and transferred to Immobilon-P polyvinylidene difluoride membranes (Millipore Corp.). Filters were blocked and stained with appropriate antibodies as described in Results. Enhanced chemiluminescence was performed according to the manufacturer's recommendation. For all Western blots, equal loading of protein was verified by reblotting of membranes for EF-1α protein.

Assessment of ROS production. To determine ROS production, cells were stained with 2.5 μmol/L hydroethidine (HE) or 5 μmol/L 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) in PBS for 30 min at 37°C in the dark (16, 33). Immediately after staining, cells were analyzed by flow cytometry (FACS Calibur), and data were analyzed using CellQuest software, version 3.1. For inhibition experiments, N-acetyl-l-cysteine (NAC) or allopurinol (AP; both from Sigma) was added 2 h before infection with Ad.mda-7. In all cases, cells were gated to exclude cell debris.

Small interfering RNA experiments. Adherent cells were transfected with small interfering RNA (siRNA) using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's instructions. Briefly, cationic lipid complexes were formed with 600 pmol of RNA and 30 μL of Lipofectamine 2000 reagent in 3 mL of Opti-MEM I Reduced Serum Medium (Invitrogen) and added to 10-cm dishes containing 15 mL of complete medium (final concentration of siRNA was 40 nmol/L). Cells were maintained in culture for 24 h and infected with Ad.vec or Ad.mda-7 (50 pfu/cell) followed with the addition of 200 μmol/L POH or 0.03% DMSO (vehicle) 2 h later. Next day, cells were trypsinized, counted, and replated for MTT, ROS, Annexin V binding, and Western blot assays. siRNA specific for XO (accession number NM_00379) and control scrambled siRNA were obtained from Santa Cruz Biotechnology, Inc.

Statistical analysis. All of the experiments were performed at least thrice. Results are expressed as mean ± SE. Statistical comparisons were made using an unpaired two-tailed Student's t test. A P < 0.05 was considered significant.

Combinatorial treatment with Ad.mda-7 and POH induces growth inhibition and MDA-7/IL-24 protein production in pancreatic cancer cell lines in vitro by generation of ROS. We used established pancreatic cancer cell lines, AsPC-1, MIA PaCa-2, and PANC-1 (mutant K-ras) and BxPC-3 (wild-type K-ras), as well as normal immortal pancreatic mesenchymal cells (LT2), to investigate cancer-specific growth-inhibitory properties of combinatorial treatment with Ad.mda-7 and POH. A nontoxic concentration of POH (200 μmol/L) was chosen, which might be clinically achievable in patients, to evaluate a combinatorial effect of POH and Ad.mda-7. Multiple pharmacokinetic studies indicate that 140 to 500 μmol/L of perillic acid and 10 to 40 μmol/L of dihydroperillic acid, the two main metabolites of POH, are found in plasma and urine after oral administration of POH-containing capsules (34). Because the growth inhibition effect of MDA-7/IL-24 is mediated by generation of ROS, we also included pretreatment with a nontoxic dose of antioxidant NAC in these studies. In a 6-day assay, POH or Ad.mda-7 alone had no discernible effect on any cell type, whereas their combination significantly inhibited growth of pancreatic cancer cells, PANC-1 and BxPC-3, irrespective of their K-ras status, with no growth-inhibitory effect on LT2 cells (Fig. 1A). Pretreatment with NAC significantly protected PANC-1 and BxPC-3 cells from growth inhibition by this combinatorial treatment. Similar findings were also observed in AsPC-1 and MIA PaCa-2 pancreatic carcinoma cells (data not shown).

In normal cells, such as LT2, MDA-7/IL-24 protein expression could be detected on Ad.mda-7 infection alone and simultaneous treatment with POH did not significantly alter the level of MDA-7/IL-24 protein (Fig. 1B). However, MDA-7/IL-24 protein expression was detected in PANC-1 and BxPC-3 pancreatic cancer cells receiving a combinatorial treatment of POH and Ad.mda-7, but not either agent alone, indicating that POH treatment overrides the intrinsic “translational block of mda-7/IL-24 mRNA” observed in pancreatic cancer cells (Fig. 1B; refs. 14, 15). These results were similar in pancreatic carcinoma cells irrespective of carrying wild-type or mutant K-ras and our treatment protocol did not alter the expression level of p21 K-RAS (data not shown). Pretreatment with NAC completely nullified MDA-7/IL-24 protein expression in both LT2 and pancreatic cancer cells, indicating that ROS production plays a fundamental role in generation of MDA-7/IL-24 protein (Fig. 1B).

MDA-7/IL-24 protein expression and apoptosis induction in pancreatic cancer cells after combinatorial treatment of Ad.mda-7 and POH are dependent on XO-mediated ROS production. As a corollary to our results obtained for cell viability, we found that a combinatorial treatment with POH and Ad.mda-7 induced significant apoptosis (30–40%) in pancreatic carcinoma cells, whereas single treatment with these agents alone did not manifest apoptosis (Fig. 2A; results are shown for PANC-1 cells and similar results were observed in AsPC-1, MIA PaCa-2, and BxPC-3 cells). Similar findings were obtained when apoptosis was evaluated by two different assays: Annexin V binding assay and fluorescence-activated cell sorting analysis of PI-stained cells counting sub-G₀ population (the data of the latter studies not shown). Apoptosis induction by Ad.mda-7 and POH treatment could be effectively inhibited by treatment with NAC and AP, a pharmacologic inhibitor of ROS-generating enzyme XO (Fig. 2A). No apoptosis was observed in LT2 cells treated with a single agent or their combinations (data not shown).

ROS production was measured after different treatment protocols using two fluorescent dyes, DCFH-DA and HE, which enabled us to distinguish between different types of ROS (35). Following POH + Ad.mda-7 treatment, HE-stained cells showed fluorescence, indicating that superoxides are generated (Fig. 2B). However, no significant increase in fluorescent staining was detected with DCFH-DA, which mainly detects peroxide/peroxynitrate generation (data not shown). NAC or AP pretreatment completely abolished ROS generation by the combinatorial treatment of POH and Ad.mda-7 (Fig. 2B). Similar to NAC, as shown in Fig. 1B, AP pretreatment completely abrogated generation of MDA-7/IL-24 protein after POH and Ad.mda-7 treatment (Fig. 2C), confirming the critical role of XO in ROS generation, MDA-7/IL-24 protein production, and apoptosis induction following the combinatorial treatment.

Involvement of XO was further confirmed by genetic approaches using siRNA targeting XO (Fig. 3). In this experiment, cells treated with scrambled siRNA were used as a control. Western blot assays confirmed an efficient down-regulation of XO protein following XO siRNA transfection but not following control scrambled siRNA transfection (Fig. 3A). XO siRNA-transfected PANC-1 cells, when treated with Ad.mda-7 and POH in combination, did not show MDA-7/IL-24 protein expression (Fig. 3A), decreased viability, ROS production, or apoptosis induction (Fig. 3B). Western blot analysis also showed that a combination treatment with Ad.mda-7 and POH down-regulated antiapoptotic Bcl-2 and Bcl-x_L proteins. These changes are eliminated by treatment with siRNA targeting XO (Fig. 3C). Similar results were obtained in AsPC-1 and BxPC-3 pancreatic carcinoma cell lines (data not shown).

Single treatment of stable clones of pancreatic cancer cells expressing mda-7/IL-24 mRNA with POH induces growth suppression and apoptosis that correlates with ROS production and MDA-7/IL-24 protein production and involves XO. Further support for the efficacy of the combination of mda-7/IL-24 and POH in suppressing growth and induction of apoptosis in pancreatic cancer cells was provided by studies using PANC-1 clones, PANC M8 and PANC M14, and MIA PaCa-2 clones, PaCa M1 and PaCa M2, stably expressing mda-7/IL-24 mRNA (Fig. 4). Despite stable expression of mda-7/IL-24 mRNA, these clones of pancreatic cancer cells do not express detectable MDA-7/IL-24 protein (16). In these mda-7/IL-24 mRNA-expressing clones, treatment with only POH at concentrations not affecting growth of parental PANC-1 or MIA PaCa-2 cells induced growth suppression and a loss of viability (Fig. 4A). A potential role of ROS production in mediating these effects in mda-7/IL-24 mRNA-expressing clones was confirmed by the fact that the antioxidant NAC inhibited apoptosis induction and ROS production in these POH-treated pancreatic tumor cells (Fig. 4B). Inhibition of XO by AP also abrogated POH-induced apoptosis induction, ROS generation, and MDA-7/IL-24 protein production in PANC M8 cells (Fig. 4C). Similarly, XO siRNA also inhibited MDA-7/IL-24 protein expression, decreased viability, ROS production, and apoptosis induction, and associated with apoptosis changed the levels of Bcl-2 and Bcl-x_L proteins (Figs. 3 and 5).

Combination treatment of pancreatic cancer cells with POH and Ad.mda-7 facilitates association of mda-7/IL-24 mRNA with polysomes. Translational regulation requires mRNA association with polysomes. We therefore tested whether treatment with POH and Ad.mda-7 facilitates association of mda-7/IL-24 mRNA with polysomes, resulting in the generation of MDA-7/IL-24 protein. AsPC-1 pancreatic carcinoma cells were infected with Ad.mda-7 (50 pfu/cell) and then treated with vehicle (DMSO) or POH (200 μmol/L). Polysomal fractions were isolated from the cell lysates and analyzed by Northern blotting using an mda-7/IL-24 cDNA probe. We confirmed that in the presence of POH, mda-7/IL-24 mRNA is associated with polysomes to a significantly greater extent when compared with vehicle treatment (Fig. 6A).

Despite aggressive therapeutic approaches, resistance of pancreatic cancer to established treatment regimens still constitutes a major problem in providing long-term survival benefit to these patients. In this investigation, we show a remarkable synergism between the action of a dietary agent, POH, and a gene therapy, mda-7/IL-24, in inhibiting pancreatic cancer growth in vitro. Our data show that nontoxic doses of POH sensitize resistant pancreatic carcinoma cells, but not normal immortal pancreatic mesenchymal cells, to mda-7/IL-24–mediated gene therapy. This novel strategy to selectively kill pancreatic carcinoma cells by mda-7/IL-24 remains equally effective for various types of pancreatic tumor cells, including those harboring both wild-type and mutant K-ras genotypes. We also elucidate a primary mechanism invoking this enhanced sensitivity to the combinatorial effect of Ad.mda-7/IL-24 + POH that involves elevated ROS production that reverses the protein translational block in mda-7/IL-24 mRNA conversion into protein by facilitating mRNA association with polysomes that ultimately culminates in apoptosis of pancreatic cancer cells. This combinatorial effect did not involve a change in K-ras protein levels. Interestingly, our results also showed that generation of oxidative stress is required for this process, as pretreatment of pancreatic carcinoma cells with antioxidants such as NAC blocked both MDA-7/IL-24 protein expression and apoptosis.

The precise mechanism by which POH can promote the production of ROS remains to be determined. Multiple sources of ROS production are activated in response to treatment with various chemoprevention and chemotherapeutic agents (36). However, in this investigation, we show that XO serves as a major source of ROS production in this system. Using a genetic approach using siRNA, we show that abrogating XO protein levels in these cells reduces ROS production, MDA-7/IL-24 protein expression, and apoptosis, and associated with apoptosis changes in expression levels of antiapoptotic Bcl-2 and Bcl-x_L proteins. Similar results were obtained using a pharmacologic approach in which pancreatic cancer cells were treated with a dose of AP, a small molecular weight inhibitor of XO, which functions as a specific scavenger of either superoxide or hydroxyl radicals (37). It is well established that XO is an important source of ROS, which generates superoxide and hydroxyl radicals in the cytoplasm (38). Our results also show that these ROS target mitochondria, leading to its membrane depolarization and a subsequent drop in membrane potential proceeded by down-regulation of the antiapoptotic proteins Bcl-2 and Bcl-x_L (Fig. 3). These results provide evidence that activation of XO by POH is a crucial step required for effective Ad.mda-7 gene therapy of pancreatic cancer. However, once MDA-7/IL-24 protein is generated in pancreatic cancer cells, the subsequent mechanism(s) of elimination of cancer cells may remain essentially identical to those observed in other cancer types such as melanoma and prostate cancer (30, 32, 33, 39), that is, direct apoptosis-inducing effects and indirect bystander antitumor effects, including inhibition of angiogenesis and immune modulation.

It is well known that treatments of pancreatic cancer cells with agents that increase generation of ROS induce oxidative stress, augment cell killing, and inhibit tumor growth (40–44). However, similar treatment regimens in normal control cells exhibit significantly less oxidative stress as well as associated physiologic and pathophysiologic responses. These results are substantiated by the observations in the present study where we did not observe comparable damage in immortalized control cells of different origins, including those of pancreas. This differential effect on cancer versus normal cells is widely reported for many other agents that only affect cancer cells sparing normal cell populations (45, 46). We hypothesize that POH at low doses induces short-term initial ROS production by XO, but this burst of ROS production is not sufficient by itself to induce apoptosis in pancreatic cancer cells. However, this initial release of ROS might be sufficient to abrogate the protein translational block, thereby resulting in association of increased amounts of mda-7/IL-24 mRNA with polysomes and consequently MDA-7/IL-24 protein production. This latter event may then promote a feedback loop resulting in a more significant and intense release of ROS that triggers mitochondrial depolarization, mitochondrial potential drop, and the apoptotic cascade (Fig. 6B). It is possible that treatment with POH could also inactivate the reduced glutathione (GSH) antioxidant system, which is involved in scavenging of superoxides (47, 48). This possibility is supported by the protective effects observed with NAC and Tiron in this study, which might restore the GSH levels in combination-treated cells. Additional studies are ongoing to experimentally validate this hypothesis.

In summary, we describe a novel approach of combining a dietary agent with gene therapy as a potent inhibitor of pancreatic cancer cell growth and survival. The safety of both POH and Ad.mda-7 has been confirmed by clinical trials and studies in animal models also reveal that this combination profoundly inhibits pancreatic cancer xenografts in nude mice (49). These exciting observations pave the way for future clinical evaluation of POH and Ad.mda-7 combination therapy in the near future to determine if this approach has efficacy in patients with pancreatic cancer.

No potential conflicts of interest were disclosed.

Grant support: NIH grants R01 CA097318 and R01 CA098712 and Samuel Waxman Cancer Research Foundation. D. Sarkar is the Harrison Endowed Scholar in Cancer Research. P.B. Fisher holds the Thelma Neumeyer Corman Chair in Cancer Research and is a Samuel Waxman Cancer Research Foundation Investigator.

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.

We thank Dr. Xiuwei Tang for assistance with and suggestions relative to K-ras biochemical studies.