ISIR-042 was prepared as previously described (5). Gemcitabine, tamoxifen, 5FU, α-tocopherol and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyltetrazolium bromide (MTT) were purchased from Sigma-Aldrich Japan (Tokyo, Japan). Taurine, L-nitroarginine methyl ester (L-NAME), N-acetyl cysteine (NAC), and protease inhibitor cocktail were obtained from Wako Pure Chemical (Tokyo, Japan).

Human pancreatic cancer cell lines CFPAC-1, BxPC-3 and Capan-2 were purchased from American Tissue Culture Collection (Manassas, VA, USA). Cells were cultured in RPMI-1640 medium (Sigma-Aldrich Japan) supplemented with 10% fetal bovine serum and 80 μg/ml gentamycin at 37°C in a humidified atmosphere of 5% CO₂ in air.

MIAPaCa-2 cells (1×10⁴ per dish) were plated into 1.1 ml of a semisolid methylcellulose medium with 0.8 % methylcellulose and 20% fetal bovine serum in triplicate for 14 days. A solution of 0.2 ml of PBS containing various concentrations of drugs was added to the semisolid medium. Colonies were photographed under an inverted microscope. Colonies in enlarged photographs were measured and counted. Colonies >0.4 mm in diameter were counted.

Bone marrow plugs from a single femur of three Balb/c mice were pooled in RPMI-1640 medium. A single cell suspension was prepared by vigorous pipetting. Nucleated cells (1,000 cells/ml/dish) were placed into a semi-solid medium containing hematopoietic growth factors (MethoCult^® GF M3434, Stem Cell Technology Inc., Vancouver, BC, Canada) and incubated for 10 days.

Human cancer cells were seeded into 24-well multidishes at 1.0×10⁴ cells/ml/well. The cells were cultured with various concentrations of drugs for 4–6 days. After culture with or without drugs, viable cells were examined by the MTT assay (5).

An isobologram analysis was used to determine the effects of combinations of drugs on Panc-1 cells. Dose-dependent effects were determined for each compound and for one compound with fixed concentrations of another. The interaction of two compounds was quantified by determining the combination index (CI), in accordance with the following classic isobologram (8).

The cell density of the drug-treated cells was kept at 2–8×10⁴/ml to maintain the growth phase in a 24-well multidish (Falcon). The medium of treated cultures was replaced by fresh medium with drugs at least every 4 or 5 days to remove cell debris of dead cells. The viable cell number was measured by MTT assay. The cumulative cell number was calculated from the MTT values and the dilution used when feeding the culture.

Cells were packed after being washed with cold PBS and then lysed at 1.5×10⁷ cells/ml in sample buffer. The resultant lysate was resolved on SDS-polyacrylamide gel, as described elsewhere (5). Primary antibodies were obtained from the following sources: K-ras (Abnova, Taiwan); acid ceramidase (BD Biosciences Japan, Tokyo, Japan); and cytochrome c and actin (Cell Signaling Technology Japan, Tokyo, Japan).

After cells were washed twice with PBS, protease inhibitor cocktail was added and cells were homogenized by 30 strokes with a glass homogenizer. The homogenate was centrifuged at 1,000 × g and the pellet was discarded. The supernatant was centrifuged at 10,000 × g for 30 min and the cytoplasmic (supernatant) and mitochondrial (pellet) fractions were separated.

Cytochrome c was determined by western blot analysis using monoclonal anti-cytochrome c antibody. Lipid peroxidation (LPO) was determined by measuring thiobarbituric acid-reactive substances (TBARS) at 535 nm, as described elsewhere (9).

Total RNA was extracted from cells using TRI reagent (Sigma). Total RNA was converted to first-strand cDNA primed with random hexamer in a reaction volume of 20 μl using an RNA PCR kit (qPCR RT Master Mix, Toyobo Co. Ltd., Osaka, Japan), and 2 μl of this reaction was used as a template in real-time PCR. The primers used were previously described (10).

Four-week-old female athymic nude mice with a BALB/c genetic background were obtained from CLEA Japan (Tokyo). They were housed under specific pathogen-free conditions. Mice were subcutaneously inoculated with 2×10⁶ Panc-1 cells. The adjusted 2×10⁷ cells/ml were mixed with an equal volume of Matrigel (BD Sciences) and 0.2 ml of the cell solution was injected subcutaneously (11). Mice were given daily intraperitoneal injections of 3 mg/kg ISIR-042 in 0.2 ml PBS and/or 2 mg/kg tamoxifen in 0.1 ml corn oil, with the first treatment given 2 days after the inoculation of tumor cells (11). Treatment with 5FU was initiated when tumors in each group achieved an average volume of 80 mm³. Treatment with 5 mg/kg 5FU was performed three times per week. Tumor volume was measured with vernier calipers. Our protocol was approved by the animal ethics committee at Shimane University.

The results are expressed as means ± standard deviation (SD). Pairs of data were compared using Student’s t-test. Significant differences were considered to exist for probabilities <5% (P<0.05). For the in vivo experiment, an F-test was performed to demonstrate statistical significance. Again, significant differences were considered to exist for probabilities <5%.

Tamoxifen inhibited the growth of Panc-1 cells in a concentration-dependent manner, and the combined effects of tamoxifen and ISIR-042 are shown in Fig. 1. ISIR-042 produced synergistic effects with tamoxifen (Fig. 1A) and the results were confirmed by isobologram analysis. Fig. 1B shows isoboles for the combination of tamoxifen with ISIR-042 that were isoeffective (IC₅₀: the concentration of the drug required for 50% inhibition of cell growth) for inhibition of the proliferation of Panc-1 cells. These isoboles indicate that the combination of these drugs had synergistic effects. ISIR-042 and tamoxifen also had synergistic effects in other pancreatic cancer cell lines, although the sensitivity of pancreatic cancer cell lines to tamoxifen varied among the cell lines (Fig. 1C). Similar results were obtained when lung carcinoma and leukemia cells were treated with ISIR-042 and tamoxifen.

Multiple non-estrogen receptor-mediated mechanisms have been implicated in cell death induced by tamoxifen (6). These mechanisms include changes in intra-cellular calcium, modulation of protein kinase C, changes in calmodulin activity, signaling though mitogen-activated protein kinases including ERK, K-ras degradation (12) and inhibition of acid ceramidase activity (13). To understand the growth-inhibitory effect of tamoxifen on pancreatic cancer cells, we examined the effects of various compounds including estrogen and inhibitors of signaling on the growth of MIAPaCa-2 cells in the presence of tamoxifen and/or ISIR-042. Estrogen did not affect the growth-inhibitory effect of tamoxifen and/or ISIR-042, suggesting a non-estrogen receptor-mediated mechanism.

Several inhibitors of protein kinases also had no effect, whereas α-tocopherol effectively reduced tamoxifen-induced growth inhibition but not ISIR-042-induced growth inhibition (Fig. 2A). Taurine, another membrane stabilizer and antioxidant, also reduced the growth inhibition induced by tamoxifen, although this effect was less than that of α-tocopherol (Fig. 2B). When MIAPaCa-2 cells were incubated with tamoxifen in the presence of the antioxidant NAC, tamoxifen-induced growth inhibition was not affected by NAC (Fig. 2C). However, treatment with NAC effectively inhibited the growth inhibition induced by the Japanese apricot extract MK615 (11) (Fig. 2C). Although the inhibitor of nitric oxide synthase L-NAME alone did not significantly affect tamoxifen-induced growth inhibition (Fig. 2D), LNAME effectively reduced tamoxifen-induced growth inhibition in the presence of a suboptimal concentration of α-tocopherol (Fig. 2E). These results suggest that tamoxifen-induced changes in membrane dynamics are associated with growth inhibition in pancreatic cancer cells.

A recent study indicated that tamoxifen induced autophagic K-Ras degradation in many cancer cells including pancreatic cancer cells (12). Since pancreatic cancers exhibit an extremely high mutation rate of K-Ras (>90%) (14), we examined the effect of α-tocopherol on the degradation of K-Ras protein induced by tamoxifen in the presence or absence of ISIR-042. Tamoxifen concentration-dependently induced K-Ras degradation, and α-tocopherol effectively counteracted this tamoxifen-induced degradation (Fig. 3A). ISIR-042 did not essentially affect the effects of α-tocopherol and tamoxifen. These results suggest that K-ras degradation is closely associated with growth inhibition in tamoxifen-treated cells. However, no significant K-Ras degradation was observed until 8 h after treatment with tamoxifen, suggesting that K-Ras degradation is not the cause of growth inhibition, but rather the result of growth inhibition, in tamoxifen-treated cells (Fig. 3B).

Acid ceramidase (AC) is a central player in ceramide metabolism. Since AC regulates the levels of pro-apoptotic ceramide and mitogenic sphingosine-1-phosphate, it is considered to be an apt target in cancer therapy. A recent study indicated that tamoxifen induces the downregulation of AC protein in cancer cells including pancreatic cancer cells (13). Therefore, we examined the effect of tamoxifen on AC protein in MIAPaCa-2 cells (Fig. 3B and C). A significant down-regulation of AC protein was observed within 1 h, suggesting that this is an early event in the action of tamoxifen (Fig. 3B). However, α-tocopherol did not affect the tamoxifen-induced degradation of AC protein (Fig. 3C), suggesting that the down-regulation of AC protein was not associated with the growth inhibition induced by tamoxifen. Tamoxifen induces apoptosis in cancer cells via nitric oxide (NO)-dependent pathways (15). Exogenously added NO potentiated tamoxifen-induced apoptosis and the inhibition of NO synthase (NOS) activity prevented the apoptosis induced by tamoxifen (15). Tamoxifen also induces oxidative stress and the release of cytochrome c from mitochondria by mechanisms that resemble those of mitochondrial NOS-derived peroxynitrite formation (16).

α-tocopherol is a membrane stabilizer and a lipophilic antioxidant (17). Therefore, we examined the effects of tamoxifen on lipid peroxidation and cytochrome c release in the presence or absence of α-tocopherol. Stimulation of mitochondrial NOS increases mitochondrial peroxynitrite, which releases cytochrome c from the mitochondria (18). The stimulation of mitochondrial NOS also increases lipid peroxidation (LPO), which is a widely used biomarker of peroxynitrite (18,19). Fig. 4A shows that tamoxifen increased LPO within 4 h of treatment in MIAPaCa-2 cells and that the effects of tamoxifen were prevented when samples were co-treated with α-tocopherol. The increase in LPO in mitochondria induced by tamoxifen was more evident than that in whole cells (Fig. 4B and C). Furthermore, tamoxifen released cytochrome c from the particulate fraction which included mitochondria in MIAPaCa-2 cells (Fig. 4D).

Pancreatic cancer stem cells express the cell surface markers CD24 and CD44 (20,21). Gemcitabine showed an increased percentage of CD24⁺CD44⁺ cells in MIAPaCa-2 cells, whereas ISIR-042 significantly reduced the percentage of CD24⁺CD44⁺ cells, suggesting that ISIR-042 preferentially inhibits stem/progenitor cells in pancreatic cancer cells (5). Therefore, we examined the effects of ISIR-042 on colony formation by MIAPaCa-2 cells in semi-solid culture and the expression of stemness-related genes. ISIR-042 effectively inhibited colony formation by MIAPaCa-2 cells in semi-solid culture, and the formation of large colonies was inhibited more evidently than that of small colonies (Fig. 5A). Colony formation in semi-solid culture was more efficiently inhibited by ISIR-042 than cell proliferation in liquid culture, whereas the inhibition of colony formation by tamoxifen or 5FU was similar to the inhibition of cell proliferation (Fig. 5B).

The growth of pancreatic cancer cell lines was effectively inhibited by ISIR-042, whereas only weak growth-inhibition was seen in normal endothelial HUVEC cells (5). Cotylenin A, the parent compound of ISIR-042, preferentially induced apoptosis in human cancer cells, while sparing normal cells (22). We examined the effects of ISIR-042 on colony formation by normal mouse bone marrow cells in the presence or absence of tamoxifen. Although ISIR-042 alone hardly affected colony formation by bone marrow cells, the combination of ISIR-042 and 4 μM tamoxifen only weakly inhibited colony formation by normal bone marrow cells (Fig. 5C). These results suggest that pancreatic cancer cells were more sensitive to combined treatment with ISIR-042 and tamoxifen than normal bone marrow cells.

Since the expression of pluripotency-associated transcription factors has been implicated in the maintenance of cancer stem cells, we evaluated mRNA expression of Sox-2, Oct3/4 and Nanog (Fig. 5D). The expression of Sox-2 and Nanog mRNAs was significantly suppressed by treatment with ISIR-042. This reduction was observed within 4 days. ISIR-042 did not significantly affect the expression of Oct3/4 mRNA under these conditions. Tamoxifen also reduced the expression level of Nanog mRNA, although the effects were less than those of ISIR-042 (Fig. 5D).

The combined effects of ISIR-042, tamoxifen and anticancer drugs were examined. A triple combination (ISIR-042, tamoxifen and 5FU) cooperatively inhibited the growth of MIAPaCa-2 cells (Fig. 6A). 5FU at 1,000 ng/ml did not completely inhibit cell viability, and the combinations of 5FU and 6 μg/ml of ISIR-042 and the combination of 5FU and 4 μM tamoxifen did not produce 100% killing either. However, treatment with the triple combination led to no viable cells in the culture (as measured by MTT assay). Similar results were obtained when cells were treated with gemcitabine. Next, we examined the growth-inhibitory effects of triple combinations in long-term culture. The growth of Panc-1 cells was markedly reduced for 12 days by treatment with 500 ng/ml of 5FU alone. However, after ~16 days, growing cells reappeared and the proliferation rate of the treated cells became similar to that of untreated cells (Fig. 6B). Drug-resistant cells were not observed when cells were cultured for 28 days with 500 ng/ml 5FU, 3 μM tamoxifen and 3 μg/ml ISIR-042, and dual combinations (5FU+tamoxifen or 5FU+ISIR042) were less effective. Similar results were obtained in the long-term culture of cells treated with gemcitabine (Fig. 6C). These results suggest that the triple combination therapy is effective for overcoming the emergence of drug-resistant cells in pancreatic cancers.

The in vitro studies described above suggested that treatment with the triple combination should be more effective therapeutically than treatment with 5FU alone. Treatments were started at day 2 after the inoculation of Panc-1 cells. We administered both tamoxifen and ISIR-042 daily by intraperitoneal injection. The combined treatment had no appreciable adverse effects in mice, including weight loss. Although neither ISIR-042 nor tamoxifen alone retarded tumor growth, combined treatment was more effective (Fig. 6D). Next, we examined the antitumor effects of the combination of 5FU with tamoxifen and ISIR-042. Treatment was initiated when tumors in each group achieved an average volume of 80 mm³. Mice were treated with 5 mg/kg 5FU three times per week.

While treatment with 5FU alone induced a slight loss of body weight (6–8% loss), treatment with the triple combination did not induce a further reduction of body weight. Although 5FU alone retarded tumor growth, combined treatment with 5FU, tamoxifen and ISIR-042 was more effective (Fig. 6E). Statistical analysis (F-test) revealed that the difference was significant (P<0.05 at 6 weeks). There were no significant differences between treatment with 5FU and combination treatment with 5FU+tamoxifen or 5FU+ISIR042. These results indicate that the combination of tamoxifen and ISIR-042 is effective therapeutically, consistent with the in vitro findings.