The human pancreatic resistant cancer cell lines MIA PaCa-2 and PANC-1 were kindly provided by Shanghai Institute of Materia Medica. The human pancreatic resistant cancer cell line Capan-1 was kindly provided by Department of Pharmacology, Sun Yat-Sen University. The paclitaxel-resistant endometrial cell line HEC-1A was purchased from Cell Bank of Chinese Academy of Sciences. The platinum-resistant ovarian cancer cell line OVCAR-3 was purchased from Experimental Animal Center of Sun Yat-Sen University. The cells were routinely cultured in DMEM (Invitrogen, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS) (Invitrogen), 100 U/ml penicillin, and 100 mg/ml streptomycin in 5% CO₂ at 37°C. Ursolic acid (UA) (National Institutes for Food and Drug Control, China) was dissolved in DMSO (Sigma-Aldrich, St. Louis, MO) as a 100 mM stock solution and stored at −20°C. Further dilution was done in serum-free cell culture medium. PI3K inhibitor LY294002 (Cell Signaling Technology, Beverly, MA) was dissolved in DMSO (Sigma) as a 50 mM stock solution and stored at −20°C. Similarly, JNK inhibitor SP600125 (Sigma) was dissolved and stored as a 10 mM stock solution by DMSO at −20°C.

Cells were seeded into 96-well plates at a density of 2×10³/well and allowed to attach overnight before being treated by different concentrations of UA for 24 h. Then cell viability and proliferation were evaluated. Cell viability was tested by the 3-(4,5-dimethylthiaziazol-2-yl)-2, 5-diphenyl tetrazolium bromide (MTT, Sigma) assay as described in the literature (21). Detection of cell proliferation was through the colorimetric 5′-bromo-2′-deoxy-uridine (BrdU) incorporation assay (Roche Diagnostic Corp., Indianapolis, IN, USA) according to the instruction of the manufacturer. Briefly, cells were labeled by addition of BrdU for 4 h in 5% CO₂ at 37°C. Then, FixDenat solution was used for cells fixation. The anti-BrdU-POD antibody was added to locate the BrdU label in the DNA. After the incubation of the peroxidase substrate, the immune complex was detected by a multi-well spectrophotometer reader using an iMark™ Microplate Reader (Bio-Rad, Richmond, CA).

The three types of cells at the density of 2×10⁶/well in a 6-well plate were fixed by 4% paraform for 10 min. Then Hoechst 33258 (Sigma) at 5 μg/ml was added to the cells for 15 min at room temperature in the dark. The nuclei morphology was observed by fluorescence microscopy (Olympus, Melvie, NY) with a 340-nm excitation filter.

Caspase-3/7, −8 and −9 activity were assessed by Caspase-Glo luminescent-based assays (Promega, Madison, WI). Accordance with the instructions, cells (1×10⁴/well) were seeded into a 96-well plate overnight to ensure adherence. After various concentrations of UA treatment for 24 h in a 5% CO₂-humidified atmosphere at 37°C, 100 μl of 3/7, −8 and −9 Caspase-Glo reagents were added to each well, respectively. Mixed and incubated for 1 h, the mixtures were transferred to a fluorescence microtiter plate and quantified fluorescence intensity by a luminometer. The corresponding OD values from MTT assay described above were considered as the viable cell number to normalize the fluorescence intensity.

Cytoplasmic cytochrome C levels were determined by a cytochrome C ELISA kit (R&D Systems, Minneapolis, MN). The cytoplasmic protein was collected by a nuclear and cytoplasmic extraction reagents (Pierce, Rockford, IL) from adherent cells (2×10⁶/well) in a 6-well plate after UA incubation for 24 h. The BCA protein assay kit (Pierce) was used to normalize the cytoplasmic protein. The protein was stored in −80°C until it was analyzed by cytochrome C ELISA kit according to the manufacturer's instructions.

Western blotting was performed as described previously (22). The following antibodies were used: antibodies against PI3K, Akt, phospho-Akt, GSK-3β, phospho-GSK-3β, NF-κB phospho-p65, JNK, phospho-SAPK/JNK, phospho-MDM2, IκB (1:1000, Cell Signaling Technology); NF-κB p65, phospho-IκB (1:1000, Santa Cruz Biotechnology, Santa Cruz, CA), and actin (1:1000, Thermo Scientific IHC, Fremont, CA).

The immunofluorescent analysis was performed as previous protocol (23). After UA treatments, the MIA PaCa-2 cells were fixed in 4% paraformaldehyde for 30 min and perforated by PBS containing 0.1% Triton X-100. The cells were blocked with 5% bovine serum albumin (BSA) for 1 h at 37°C and incubated the NF-κB p65 antibody (1:200) at 4°C overnight. After washing by PBS, Cy3-conjugated goat anti-rabbit IgG (1:50, Proteintech, Chicago, IL) was used as the secondary antibody for 1 h at 37°C. Nuclear staining was achieved by incubating the cells in DAPI for 10 min. The fluorescence was photographed with fluorescence microscopy (Olympus).

To establish a subcutaneous xenograft model, MIA PaCa-2 cells (5×10⁶) were inoculated subcutaneously on the flanks of 6-week-old nude mice (female, purchased from Experimental Animal Center of Southern Medical University). When the average tumor size reached ~100 mm³, mice with tumor xenografts were randomized into three groups with six mice in each group including: i) vehicle alone (normal saline with 1% DMSO i.p.), ii) low dose UA (100 mg/kg twice a week i.p.), iii) high dose UA (200 mg/kg twice a week i.p.). All of the treatments were administered at a dose of 0.1 ml/10 g bodyweight. The formula, (length × width²)/2, was used to estimate tumor sizes. Both the body weight and tumor sizes were recorded twice a week. At the end of the experiment, the animals were sacrificed and the tumors were dissected and weighed to calculate the inhibitory rate. Animal studies were performed according to the institutional guidelines of Experimental Animal Center in Sun Yat-sen University.

The paraffin tissue sections were obtained from the tumor tissues from the nude mice, IHC were performed as described previously (24). Briefly, the sections were deparaffinizated with xylene and microwaved for 10 min in the presence of 10 mM citric acid buffer. After quenching the endogenous peroxidases by hydrogen peroxide, the sections were blocked by 5% BSA/0.3% Triton X-100 for 20 min. Then the sections were incubated with cleaved caspase-3 (1:200) antibody overnight at 4°C and secondary antibody, goat anti-rabbit IgG (1:50, Proteintech) for 30 min at 37°C, and then signal was generated by using the DAB substrate kit (Boster, Wuhan, China) for peroxidase.

Data are expressed as mean ± standard deviation (SD) of three separate experiments and analyzed by SPSS. P<0.05 was considered significant and determined by the two-sample Student's t-test or one-factor ANOVA analysis.

In order to appraise the cytotoxicity of UA on pancreatic-resistant carcinoma, three different malignancy grades of pancreatic resistant cancer cell lines including MIA PaCa-2 (grade 3), PANC-1 (grade 2) and Capan-1 (grade 1) were exposed to increasing concentrations of UA (15, 30, 45 and 60 μM) for 24 h (6,25–27). Firstly, the morphology of three cell lines under microscopic observation revealed the corresponding reduction of surviving cell numbers and morphological changes after UA treatments (Fig. 1A). In addition, cell viability was analyzed by MTT assay. As shown in Fig. 1B, UA dose-dependently decreased the viability of three cell lines with the half maximal inhibitory concentration (IC₅₀ value) at 40.8, 45.3 and 41.3 μM, respectively. Gemcitabine as the positive control at the concentration of 45 μM did not induce significant cytotoxicity to PANC-1 cells comparing with the vehicle for 24 h. However, UA decreased the viability by 62.5% at the same concentration in the identical cell lines. Similarly, UA achieved stronger cytotoxic effect than gemcitabine in MIA PaCa-2 and Capan-1 cells. The proliferation inhibition of UA was further investigated by 5′-bromo-2′-deoxy-uridine (BrdU) incorporation assay. As observed in Fig. 1C, growth inhibition effect of UA in MIA PaCa-2 cells started at 15 μM and increased up to 60 μM. The results were in agreement with the outcomes of MTT assay. These data suggested that UA was capable of causing growth inhibition to pancreatic-resistant carcinoma in vitro.

UA has been shown to induce apoptosis in multiple cancer cell lines (16–20). To identify whether UA has the same effect in pancreatic-resistant cancer cell lines, the morphologic analysis was performed by Hoechst 33258 staining. MIA PaCa-2, PANC-1 and Capan-1 cells were exposed to various concentrations (15, 30, 45 and 60 μM) of UA for 24 h, the morphological characteristics of apoptotic cells such as condensed nuclei and cell shrinkage were discovered and apoptotic proportion grew with upward dose (Fig. 2A). To confirm the UA-triggered apoptosis in three cell lines, another apoptotic marker, caspase-3/7 activity, was examined. Clearly, the caspase-3/7 activity was promoted gradually along with the rising concentrations of UA. Fig. 2B shows that the caspase-3/7 activity was enhanced 20-, 16- and 8-fold, respectively, in MIA PaCa-2, PANC-1 and Capan-1 cells at 60 μM UA.

The caspase-dependent cascade is mediated by endogenous and exogenous signal transductions (22). To study by which UA induces caspase-3/7 activation, several markers belonging to the two pathways were inspected. Caspase-8 is a downstream apoptotic molecule of the death receptor, which mediates the death receptor signaling pathway (extrinsic pathway). As shown in Fig. 2C, caspase-8 activity increased by ~15-fold with 60 μM UA treatment. Cytochrome C released from mitochondria into cytoplasm was the initial step in the intrinsic apoptotic pathway. Fig. 2D exhibits the cytochrome C relative release rate increase by ~4.5-fold at 60 μM UA. Besides, caspase-9 activity, another marker of the intrinsic apoptotic pathway, was also increased by ~14-fold after 60 μM UA intervention (Fig. 2E).

The mitogen-activated protein kinase (MAPK) family is closely related to cell apoptosis (28). There are some reports revealing that UA is able to activate JNK, which is one of crucial components in MAPK signaling pathway (16,29). Here, we detected JNK and P38 expression levels, both of which were reported as vital molecules of MAPK family in the programmed cell death, to explore more potential signal channels (30). We found that JNK activity, measured by phosphorylated-JNK level, was significantly up-regulated at the presence of 30 and 45 μM UA. But, UA had no obvious effect on JNK, P38 and phosphorylated-P38 protein levels (Fig. 3A). This result indicated that JNK may contribute to UA-induced apoptosis. To further study the apoptotic pathway in which JNK was involved, we used the JNK inhibitor SP600125 to block the activity of JNK pathway. MIA PaCa-2 cells were pretreated with 10 μM SP600125 for 1 h, and then treated with 45 μM UA for 12 h. As shown in Fig. 3B and C, pretreatment by SP600125 partially inhibited the caspase-9 activity by 16%, while caspase-8 activity was not altered significantly. These data indicated that JNK is at least partly involved in UA-induced endogenous apoptotic signaling transduction.

To ascertain the potential mechanism by which UA caused growth suppression and apoptosis to pancreatic cancer, we also examined the activity of PI3K/Akt pathway in MIA PaCa-2 cell line. When cells were in the presence of increasing concentrations of UA (15, 30 and 45 μM) for 12 h, the results by western blotting demonstrated the total PI3K and the phosphorylated Akt levels were attenuated gradually in a dose-dependent manner, while the total Akt protein expressions did not alter (Fig. 4A). PI3K/Akt pathway is constitutively activated in most tumor cell lines and regulates a series of biological characteristics of tumors such as growth and apoptosis (31). Therefore, we checked the possible Akt targeting effector molecules downstream. Similar changes to Akt, were recorded for the phosphorylation level of GSK-3β, MDM2, P65, IκB that also decreased gradually, whereas the total proteins of GSK-3β, P65, IκB did not show any corresponding alteration (Fig. 4B).

Aberrant activation of NF-κB greatly contributes to chemoresistance in pancreatic cancer, and suppression of NF-κB sensitizes the treatment outcome of gemcitabine (32). To assess whether the UA-induced NF-κB activation is a downstream event of PI3K/Akt signaling pathway, IκB and its phosphorylated form were further investigated. We inhibited the phosphorylation of Akt by 10, 20, 40 μM LY294002 for 12 h. As seen in Fig. 4C, 20 μM LY294002 and above abolished more than half of the phosphorylated level of Akt protein. Phosphorylated IκB was evidently modulated downward in a dose-dependent pattern. However, there was no alteration of total IκB within 12 h. In addition, the nuclear translocation of the cytoplasmic NF-κB p65 subunit was also checked by immunofluorescence. UA (45 μM) and/or 40 μM LY294002 for 12 h repressed the NF-κB p65 nuclear translocation compared with vehicle (Fig. 4D). These results suggested there was crosstalk between PI3K/Akt and NF-κB signaling pathways in MIA PaCa-2 cells.

We next examined the involvement of PI3K/Akt/NF-κB pathway in the apoptotic cascade. The MIA PaCa-2 cells were incubated with 40 μM LY294002 for 12 h. As seen in Fig. 4E and F, inhibiting the activation of Akt can not activate caspase-8 and −9. Hence, PI3K/Akt/NF-κB pathway is not the main participant in UA-induced intrinsic and extrinsic apoptotic signaling transductions. However, the cell viability was decreased by ~25% after exposure to 40 μM LY294002 for 24 h. These findings indicated that PI3K/Akt/NF-κB pathway was involved in UA-induced cytotoxicity to MIA PaCa-2 cells, but the mechanism was not caspase-8- and −9-dependent.

We also tested whether UA possessed the potential to induce other resistant cancer cells death in the similar fashion. A paclitaxel-resistant endometrial cell line (HEC-1A) and a platinum-resistant ovarian cancer cell line (OVCAR-3) were used to examine the cell viability after UA exposure (33,34). As seen in Fig. 5A and B, UA inhibited the cell viability in a time and dose-dependent manner. Moreover, UA also increased the phosphorylated JNK, and decreased the expression of phosphorylated Akt (Fig. 5C and D). These results further confirmed that the mechanism of action of UA was associated with JNK and Akt pathways.

The data described above clearly showed that UA was a potent compound to inhibit proliferation and caused deaths of various chemotherapy-resistant cancer cell lines. Thus we tried to apply UA in a nude mouse xenograft model to provide evidence for clinical application. Until the tumor volume reached ~100 mm³, the mice were treated i.p. with either vehicle or UA at 100/200 mg/kg twice a week. Significant inhibition of tumor growth by UA was observed after UA treatment for 20 days (Fig. 6A and B). Tumor weights in UA-treated groups were also lower than the vehicle group (30% less in 100 mg/kg group and 50% less in 200 mg/kg group) (Fig. 6C). In order to measure the safety of UA, body weights of mice were regularly detected. As seen in Fig. 6D, there were no significant differences compared with control group after the UA treatment. Furthermore, we checked the cleavage caspase-3 expression by immunohistochemistry to reveal that the tumor in UA-exposed group had more visible apoptosis. These data elucidated that UA caused resistant pancreatic cancer death also in vivo.