Pristimerin (Fig. 1A, purity >99%, HPLC) was purchased from Paypaytech Inc. (Shenzhen, China), and dissolved in dimethyl sulphoxide (DMSO, Sigma-Aldrich, Shanghai, China) at a stock concentration of 20 mM, and stored at −20°C. Vinblastine was obtained from Selleck (Shanghai, China). Antibodies against survivin, Bcl-X_L and PCNA were from Santa Cruz Biotech (Santa Cruz, CA, USA). Antibodies against p65, IκBα, phospho-IκBα (S32), Tubulin, matrix metalloproteinase 2 (MMP2), matrix metalloproteinase 9 (MMP9), Slug, Sox2, Nanog and KLF4 were purchased from Cell Signaling Technology (Beverly, MA, USA). Antibodies against PARP (clone 4C10–5), caspase-3, active caspase-3 (CM1), cytochrome c (clone 6H2.B4), X-linked inhibitor of apoptosis protein (XIAP), Bcl-2, and c-Myc were purchased from BD Biosciences (San Jose, CA, USA). Anti-mouse and anti-rabbit immunoglobulin G fluorescent-conjugated secondary antibodies were purchased from LI-COR Biotechnology (NE, USA). Dual-luciferase assay kit was provided by Promega (Madison, WI, USA). The Aldeflour kit was from StemCell Technologies (Vancouver, Canada). Gelatin (G1890-100G) was obtained from Sigma-Aldrich). Coomassie Brilliant Blue R-250 (44329-7C) was from Farco Chemical Supplies (Hong Kong, China).

The UM cells 92.1, Mel 270, Omm 1 and Omm 2.3 were cultured in RPMI-1640 supplemented with 10% fetal bovine serum (FBS), penicillin (100 U/ml), streptomycin (100 µg/ml) and L-glutamine (2 mM), and incubated at 37°C and 5% CO₂, as described previously (23). 293T cells were cultured in DMEM supplemented with 10% FBS, penicillin (100 U/ml), streptomycin (100 µg/ml) and incubated at 37°C and 5% CO₂.

Omm 1, 92.1 and Omm 2.3 cells seeded in 24-well plates were co-transfected with NF-κB-TATA-Luc reporter plasmid (0.5 µg) and Renilla luciferase reporter plasmid (10 ng), respectively. After 24 h, these UM cells were exposed to the indicated concentrations of pristimerin for 16 h, then added with TNFα (10 ng/ml) for 8 h and measured with dual-luciferase assay kit as described previously (9,24). The luciferase activities were normalized with NF-κB-dependent firefly luciferase to Renilla luciferase activity.

The prepared cell lysates were quantified by Pierce™ BCA Protein assay kit (Thermo Scientific, USA) according to the manufacturer's instructions. Protein samples were subjected to SDS-PAGE and then blotted to nitrocellulose membranes. After blocked by 5% skim milk, membranes were subsequently incubated with the primary antibodies overnight at 4°C before incubation with appropriate secondary antibodies. The immunoblots were recorded with the Odyssey infrared imaging system (LI-COR).

After cultured overnight, 92.1 and Omm 1 cells were incubated with or without pristimerin for 24 h, and then the cells were stimulated with TNFα (20 ng/ml) for 15 min before fixed by 4% paraformaldehyde. After permeabilization by 0.2% (v/v) Triton X-100 at room temperature for 15 min, the samples were washed with 500 µl of PBS for 5 min three times. Subsequently, PBS buffer containing 5% bovine serum albumin was used to block non-specifical binding for 1 h and then the cells were incubated with 200 µl of PBS containing the primary antibody for 1 h at room temperature. After washing with PBS containing 0.1% (v/v) NP-40, the cells were incubated with 200 µl of PBS containing the secondary antibody for 1 h at room temperature. Nuclear staining was performed with DAPI for 20 min. After washing with PBS for 5 min, all samples were added with a half drop of anti-fade regent and then sealed by nail polish (9).

After stimulation with TNFα (10 ng/ml), cells treated with or without pristimerin were collected by centrifugation and washed with PBS. Pellets resuspended in 150 µl cold lysis buffer (10 mM HEPES, 10 mM KCl, 0.1 mM EDTA, 0.4% NP-40, 1 mM DTT, 0.5 mM PMSF, 1 mM sodium orthovanadate and Complete Protease Inhibitor Mix) by pipetting up and down ~10 times were incubated on ice for 10 min, the lysates were centrifuged at 20,000 g (9,25). The supernatants were transferred to a fresh tube as cytoplasmic extracts. After rinsed with the lysis buffer, the pellets were vigorously suspended in nuclear protein extraction buffer (20 mM HEPES, 0.4 mM NaCl, 1 mM EDTA, 1 mM DTT, 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM sodium orthovanadate and Complete Protease Inhibitor Mix) and incubated for 15 min. After centrifugation with 10 min at 20,000 g, the supernatant was referred to as nuclear fractions (9,25).

Control or treated cells were pelleted by centrifugation and then rinsed with PBS. The whole cell lysates were prepared with RIPA buffer (1X PBS, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate) supplemented with 1X protease inhibitor cocktail, 10 mM β-glycerophosphate, 1 mM sodium orthovanadate, 10 mM sodium fluoride, and 1 mM PMSF (26). The cytosolic fraction was extracted with digitonin extraction buffer (10 mM PIPES, 0.015% digitonin, 300 mM sucrose, 100 mM NaCl, 3 mM MgCl₂, 5 mM EDTA, and 1 mM PMSF) on ice for 10 min and then centrifuged at 20,000 g for 10 min at 4°C. The supernatants were collected as cytosolic fractions (27).

The UM cells (5,000 cells/well) seeded into 96-well plates overnight were then exposed to serially diluted pristimerin for 72 h. Optical intensity was measured by microplate reader after addition with 20 µl mixture of MTS and PMS (28).

The UM cells were treated with various concentrations of pristimerin for 24 h and rinsed with PBS, then 8,000 cells were seeded to a modified drug-free double layer soft agar system (29). After 10–14 days, the number of colonies composed of >50 cells was counted under a microscope.

After pretreated with or without pristimerin for 24 h and washed with PBS, UM cells were fixed with 66% ethanol overnight. Cells rinsed with PBS were pelleted by centrifugation. Subsequently, the pellets were resuspended in PBS and stained with propidium iodide (PI, 50 µg/ml) and RNase (2.5 µg/ml) for 30 min at room temperature. DNA content was analyzed by flow cytometry (9).

Annexin V-FITC/PI double staining was used to detect apoptosis by flow cytometry. Cells pre-incubated with or without pristimerin were collected by centrifugation and washed with PBS, and then stained with 0.3% Annexin V-FITC in binding buffer for 15 min at room temperature in the dark. After centrifuged at 250 g and resuspended in 1X binding buffer, the cells were added with PI before flow cytometry analysis (9,29).

UM cells were incubated with 2.0 µM pristimerin for different durations, and then stained with MitoTracker probes (CMXRos and MTGreen) for 1 h at 37°C. After being centrifuged at 250 g for 10 min, the pellets were raised in PBS and subjected to flow cytometry to detect the changes in mitochondrial transmembrane potential (Δψ) (25,29).

Lentiviruses were produced by transient transfection in 293T cells with control shRNA (Scramble) or specific shRNA together with the pCMV-dR8.2 packing construct and the pCMV-VSVG envelope construct. 92.1 and Omm 1 cells (1×10⁵ cells/well) were infected twice with virus-containing supernatants. The cells were then incubated in the presence of puromycin (1 µg/ml) for ~5 days. Scramble and specific shRNAs were purchased from Sigma-Aldrich, and the sequences were as follows: PLKO.1-Non-target shRNA: CCG GGC GCG ATA GCG CTA ATA ATT TCT CGA GAA ATT ATT AGC GCT ATC GCG CTT TTT; shsurvivin: CCG GGA AGA ATT AAC CCT TGG TGA ACT CGA GTT CAC CAA GGG TTA ATT CTT CTT TTTG. The Omm 1 cells overexpressed survivin were established with the same methods. Constructs bearing full length human survivin cDNA in pTSB-CMV-MCS-SBP-tRFP-F2A-PuroR and empty vector were provided by Transheep (Shanghai, China). Human baculoviral inhibitor of apoptosis repeat-containing 5 (BIRC5, NCBI Reference Sequence ID: NG_029069.1) was cloned into the pTSB-CMV-MCS-SBP-tRFP-F2A-PuroR (lentivirus) plasmid using ClonExpress MultiS One Step Cloning kit, which was provided by Vazyme (Nanjing, China) and clone sites were EcoRI and BamHI. The efficiency of knockdown or overexpression was examined by western blot analysis.

After pre-incubated with various concentrations of pristimerin, total RNA was extracted by TRIzol reagent, and then reverse-transcribed into first-strand complementary DNA with maxima first strand cDNA synthesis kit. GAPDH was used as an endogenous reference. The PCR primers were as follows: survivin forward, 5′-CAT CTC TAC ATT CAA GAA CTG G-3′; reverse, 5′-GGT TAA TTC TTC AAA CTG CTT C-3′ (30). GAPDH forward, 5′-GAT CGA ATT AAA CCT TAT CGT CGT-3′; reverse, 5′-AGC AGC AGA ACT TCC ACT CGG T-3′.

The qRT-PCR reaction was performed in SYBR Premix EX Taq with Bio-Rad CFX96 Real-Time Thermocycler according to the manufacturer's instructions (26,29). In relative quantification, ΔΔCq method was used, as described previously (31).

92.1 and Omm 1 cells (5×10⁵/well) were cultured in a 6-well plate supplemented with 10% FBS. The monolayer was gently and slowly scratched with a 200-µl pipette tip across the center of the well at ~90% confluence. The cells were treated with or without pristimerin (0.25 or 0.5 µM), the same wounded area was recorded under a microscope at different time periods (29,32).

After exposed to vehicle, 0.25 or 0.5 µM pristimerin for 24 h, equal amounts of 92.1 and Omm 1 cells were cultured in the upper of Transwell chamber covered with or without Matrigel and contained with FBS-free medium, while to the lower chamber 20% FBS was added as chemoattractant. After 48 h, the cells on the surface of chamber were removed by a cotton swab, and the migrated or invaded cells were fixed with 4% paraformaldehyde and stained by crystal violet. The cells in three randomly selected low power fields were counted (29).

92.1 and Omm 1 cells were seeded into 6-well culture plates. When cells grew to ~80% confluency, were washed by serum-free medium three times. Cells were incubated with serum-free medium with or without pristimerin (0.25 and 0.5 µM) for 24 h. MDA-MB-231 cells were used as positive control (33). Serum-free medium without cells was used as negative control. After incubation, the supernatants were collected and were subjected to SDS-PAGE using 10% acrylamide gels containing 0.1% gelatin. After electrophoresis, the gels were washed for 15 min at room temperature in a buffer (50 mM Tris-Cl pH 7.6, 10 mM CaCl₂, 20 mM NaCl, 1 µM ZnCl₂ and 2.5% Triton X-100) three times. After incubation with activation buffer (50 mM Tris-Cl pH 7.6, 10 mM CaCl₂, 20 mM NaCl, 1 µM ZnCl₂) for 48 h at 37°C, gels were stained with 0.25% Coomassie Brilliant Blue R-250 in 40% methanol and 10% acetic acid and then briefly destained in 10% acetic acid and 30% methanol. The locations of gelatinolytic enzymes were visualized as clear bands on the background (34). The experiment was repeated three times and the optical density of each bands was quantitated by image pro plus.

The Aldefluor kit was used to identify a population of cells with high ALDH enzymatic activity (35). In brief, UM cells (92.1 and Omm 1) pre-incubated with DMSO or 1.0 µM pristimerin for 24 h, and then incubated with 5 µl ALDH reagent in the absence or presence of 5 µl DEAB for 1 h at 37°C. After washed with ALDH assay buffer, the cells were resuspended in ALDH assay buffer. The proportion of ALDH⁺ cells was defined by flow cytometry (29,35).

92.1 and Omm 1 cells pretreated with DMSO or 1.0 µM pristimerin for 24 h were seeded to 24-well low attachment plates containing 500 µl DMEM/F12 medium supplemented with 1X B27, 10 ng/ml basic fibroblast growth factor (bFGF), 20 ng/ml epidermal growth factor (EGF) each well. After ~14 days, melanospheres with >50 cells were counted as per the universal standard (29,36). The secondary and tertiary rounds of melanosphere cultures were implemented in drug-free culture after collecting the first or second round of tumor sphere culture cells. Representative images were taken by a microscope.

All experiments were performed at least thrice, and the results are expressed as the means ± standard error (SE), unless otherwise stated. GraphPad Prism 5.0 software was used for statistical analysis. Comparison between 2 groups used two-tailed Student's t-test, and comparison among multiple groups involved one-way ANOVA with post hoc intergroup comparisons using Tukey's test. P<0.05 was regarded as statistically significant.

Our previous report indicated that pristimerin potently inhibits activation of canonical NF-κB pathway in leukemia cells (9). Here, we tested the effect of pristimerin in UM cells. We first examined whether pristimerin exerted the inhibitory effect on TNFα-induced NF-κB-dependent reporter gene transcription. The results showed that the NF-κB dependent luciferase activity was increased after TNFα stimulation in UM cells. This effect was inhibited by pristimerin in a dose-dependent manner (Fig. 1B). Because the ubiquitination-dependent degradation of IκBα protein triggers p65 nuclear translocation, which is a critical step in the activation of the canonical NF-κB pathway (37), we determined the influence of pristimerin on IκBα and p65. Western blot analysis of cytoplasmic fractionations showed that the levels of phosphorylated IκBα in the cytoplasm were appreciably increased after TNFα induction compared to the untreated control, while the total IκBα was coincidently decreased (Fig. 1C). However, the phosphorylation of IκBα was blocked, and the level of total IκBα remained constant in the cells that were pretreated with pristimerin (Fig. 1C). Consistently, the level of p65 in nuclear fractionation was remarkably escalated after TNFα stimulation (Fig. 1C), which was diminished by the presence of pristimerin (Fig. 1C). On the other hand, immunofluorescence staining analysis similarly revealed that nuclear translocation of p65 was prominently increased after TNFα stimulation; whereas pristimerin abrogated its translocation in both 92.1 and Omm 1 cells (Fig. 1D).

We next ascertained the effect of pristimerin on the expression of encoding products of NF-κB-dependent genes involved in cell survival by western blotting (10). The expression of MMP9, c-Myc, Bcl-X_L and survivin was increased after stimulation with TNFα (Fig. 1E). Nevertheless, pristimerin blocked such an increase (Fig. 1E).

Taken together, our results suggested that pristimerin represses the activation of the canonical NF-κB pathway.

Seventy-two-hour cell-based MTS assay revealed that pristimerin significantly inhibited the cell viability of 92.1, Mel 270, Omm 1 and Omm 2.3 cells, and the IC₅₀ values were 1.1, 1.7, 2.2 and 2.9 µM, respectively (Fig. 2A). Given the advantages of colony formation assay in reflecting malignant behavior of tumor cells, we assessed the impact of pristimerin on anchorage-independent growth of UM cells in soft agar. The results indicated that pristimerin significantly inhibited the formation of colony in a dose-dependent manner (Fig. 2B).

To further study the role of pristimerin in the growth of UM cells, cell cycle distribution analysis was performed after UM cells were exposed to different concentrations of pristimerin for 24 h. The results showed that pristimerin caused G₁-phase arrest in UM cells (Fig. 2C).

We carried out flow cytometry analysis based on Annexin V-FITC/PI double staining and found that pristimerin remarkably induced cell death in UM cells in a dose- and time-dependent manner (Fig. 3A). Western blot analysis revealed that pristimerin induced cleavage of PARP and caspase-3 in UM cells (Fig. 3B), which demonstrated the occurrence of apoptosis induced by pristimerin.

Next, we evaluated the effect of pristimerin on mitochondrial depolarization in UM cells. After being treated with pristimerin, the cell populations with loss of mitochondrial transmembrane potential were significantly increased (Fig. 3C). In parallel, western blot analysis showed that the levels of cytochrome c in the cytosolic fractionation were increased in a time-dependent manner (Fig. 3D). These results suggested that pristimerin might cause damage of mitochondria and trigger the intrinsic pathway of apoptosis in UM cells.

To dissect the mechanism by which pristimerin induced apoptosis in UM cells, we next measured apoptosis-related proteins. The results showed that pristimerin decreased the expression of XIAP and survivin, but not Bcl-2 and Bcl-X_L (Fig. 4A). Given that survivin was obviously declined and the anti-apoptotic role of survivin in pristimerin-inducing apoptosis in prostate cancer cells (38), we hypothesized that survivin might be critical in pristimerin-inducing apoptosis in UM cells. The Omm 1 cell ectopic overexpression of survivin by lentiviral construct encoding survivin were more resistant to pristimerin than cells transfected with empty vector (Fig. 4B). On the other hand, 92.1 cells were transfected with scramble or survivin shRNA, and the results showed that knockdown of survivin facilitated the sensitivity of UM cells to pristimerin (Fig. 4C). These data imply that survivin indeed plays a critical role in pristimerin-inducing apoptosis.

We next studied the regulation of survivin by pristimerin. qRT-PCR analysis showed that pristimerin treatment in 92.1 and Omm 1 cells led to a decrease in the mRNA level of BIRC5 (encoding survivin) (Fig. 4D). The results suggested that downregulation of survivin by pristimerin occurs at the level of transcription in UM cells.

Although multiple therapies are available for primary UM patients, therapeutic options are limited for the metastatic patients with no apparent reduced mortality in the past decades (1,2,4). In order to investigate the effect of pristimerin on the motility of UM cells, wound healing ability was tested in 92.1 and Omm 1 cells. After 72 h of pristimerin treatment, the occlusion of UM cells was remarkably slowed (Fig. 5A). We next examined the inhibitory effect of pristimerin on Transwell migration and invasion and found that pristimerin profoundly decreased the number of migrated (Fig. 5B) and invaded (Fig. 5C) cells. Local invasion is an initial step of metastasis with involvement of matrix metalloproteinases (MMPs) (29,39). We therefore assumed that pristimerin suppressed the invasion of UM cells by decreasing the expression or activity of MMP2 and MMP9. Western blotting results showed that pristimerin decreased the protein levels of MMP2 and MMP9 in a dose-dependent manner (Fig. 5D). Moreover, gelatin zymography assay results indicated that pristimerin inhibited the activity of MMP9 in a dose-dependent manner (Fig. 5E).

CSC theory holds a view that CSC is the main cause for metastasis and therapeutic resistant (40). Several in vitro assays such as ALDH⁺ assay and tumor sphere formation allow for the study of CSCs (10). Flow cytometry analysis showed that the percentage of ALDH⁺ cells was significantly decreased in the cells treated with pristimerin compared to that in control group (Fig. 6A). On the other hand, three rounds of serially re-plating culture of melanoshpere showed that the numbers of melanosphere in the pristimerin-treated cells were significantly lower than those in control group (Fig. 6B), implying that pristimerin exerts an inhibitory activity on the self-renewal capacity of CSCs in UM cells. We also examined the universal stemness-related proteins which were also regulated by NF-κB pathway such as Slug, Sox2, Nanog and KLF4 (10), the results showed that the expression of Slug and Sox2 was decreased in the cells treated with the highest concentration (2 µM) of pristimerin (Fig. 6C). The protein levels of Nanog and KLF4 remained constant (Fig. 6C).

Vinblastine, a conventional chemotherapeutic agent, was used in clinical treatment of UM patients (3). We thus exploited whether pristimerin increased sensitivity of UM cells to vinblastine. The results of 72 h of combinational treatment showed a synergistic effect in 92.1 and Omm 1 cells (combinational index, CI<1) (Fig. 7A). Furthermore, detection of PARP cleavage and trypan blue exclusion assay confirmed an enhanced apoptosis-inducing effect in UM cells (Fig. 7B).