The PANC-1, AsPC-1, HPAF-II, BxPC-3 and MiA PaCa-2 human pancreatic cancer cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and maintained in the laboratory. The MRC-5 immortalized human lung epithelial cell line was provided by Dr Sitta Sittampalam at the National Center for Advancing Translational Sciences, NIH (Bethesda, MD, USA), and was used as a comparison to the cancer cells. All cells were cultured at 37°C in 5% CO₂/95% air in recommended growth media: PANC-1 and Mia PaCa-2 in DMEM (cat no. 10-013-CV; Corning, Inc., Corning, NY, USA); AsPC-1 and BxPc-3 in RPMI-1640 (cat. no. 10-040-CV; Corning, Inc.) and HPAF-II in EMEM (cat. no. 10-010-CV; Corning Inc.), containing 10% fetal bovine serum (FBS; Sigma-Aldrich, St. Louis, MO, USA; cat. no. F0926) and 1% antibiotics (cat. no. 30-001-C; Corning, Inc.). The Rau extract was provided by Natural Source International, Ltd. (New York, NY, USA) and was prepared in sterile phosphate-buffered saline (PBS) in 10 mg/ml stock solutions and stored at −20°C.

The cells were assessed for viability using a 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay at 48 h of treatment. Cells in the exponential growth phase were exposed to serial dilutions of Rau for 48 h. The medium was then replaced with fresh media containing MTT and cells were incubated for 4 h at 37°C. The colorimetric MTT assay assesses relative proliferation, based on the ability of living, but not dead cells, to reduce MTT to formazan. The cells did not reach a plateau phase during the incubation period. The 50% inhibitory concentration (IC₅₀) was defined as the concentration of drug that inhibited cell growth by 50% relative to the untreated control. Pilot experiments for each cell line were performed to optimize cell density and assay duration, and to center drug dilution series approximately on the IC₅₀.

For the PANC-1 cells, a single-cell suspension was plated into 24-well ultra-low attachment plates (Corning Inc.) at a density of 5,000 cells/well in stem cell media and incubated at 37°C in a humidified atmosphere of 95% air and 5% CO₂. For the MIA PaCa-2 cells, a single-cell suspension was plated into 96-well ultra-low attachment plates (Corning Inc.) at a density of 100 cells/well in stem cell media and incubated under the same conditions. The stem cell media consisted of DMEM (Corning Inc.) supplemented with 1X B27 Supplement, 20 ng/ml human basic fibroblast growth factor, 20 ng/ml epidermal growth factor, 100 U/ml penicillin/streptomycin (Invitrogen; Thermo Fisher Scientific, Inc.) and 4 µg/ml heparin calcium salt (Thermo Fisher Scientific, Inc.). The PANC-1 spheroids were counted following 4 weeks of culture and the MIA PaCa-2 spheroids were counted following 2 weeks of culture under the microscope. Spheroid diameter was measured using ImageJ software v1.48 (NIH, Bethesda, MD, USA).

Rau has marked autofluorescence in two ranges of emission wavelength, at 400–600 nm and 800–900 nm, overlapping the emission wavelength of several fluorescent labeling molecules. Therefore, PE-Cy7-conjugated CD24 and APC-conjugated EpCam antibodies were used as indicative markers for pancreatic CSCs (CD24⁺EpCam⁺) to avoid overlapping with Rau autofluorescence. The cells were exposed to various concentrations of Rau for 24 or 48 h. The cells were then washed with PBS three times, and resuspended in binding buffer (PBS supplemented with 0.1% bovine serum albumin (BSA; Fisher BioReagents, Waltham, MA, USA; cat. no. BP1605-100) for 15 min. PE-Cy-7-conjugated anti-CD24 antibody (dilution 1:100; cat. no. 311119; BioLegend, Inc., San Diego, CA, USA) and APC-conjugated anti-EpCam antibody (dilution 1:100; cat. no. 324207; BioLegend, Inc.) were added into the cell suspension and incubated for 15 min according to the manufacturer's protocol. The cells were washed in PBS three times following staining and then analyzed using a BD LSR II flow cytometer. The data were normalized to cell death, as follows: Normalized CSC population = original CSC population detected with flow cytometry × % cell viability detected with the MTT assay.

Dye Cycle Violet (DCV; Invitrogen; Thermo Fisher Scientific, Inc.) was used for staining of the non-CSC population. Cells that efficiently exclude DVC from the cytoplasm are considered CSC-like population (DCV⁻ cells). The MIA PaCa-2 cells were suspended at a density of 1×10⁶ cells/ml in DEME supplemented with 10% FBS and 10 mM HEPES. DCV (10 µM) was added and incubated for 30 min at room temperature. The cells were then washed twice with PBS, and resuspended in DMEM supplemented with 10% FBS and 10 mM HEPES for 1 h. The cells were transferred to ice-cold HBSS/2% FBS/10 mM HEPES buffer immediately prior to flow cytometric sorting. The DCV⁻ and DCV⁺ cells were separately collected for further analysis. Gate setting was performed using cells treated with a pump inhibitor (verapamil; 200 µM) prior to DCV staining.

The cells were lysed with RIPA buffer containing protease inhibitors and phosphatase inhibitors (Sigma-Aldrich; EMD Millipore) followed by sonication for 10 sec. Either whole cell lysate or supernatant was used for further experiments, depending on the proteins of interest. The BCA method was used for protein quantification (Pierce BCA protein assay kit; Thermo Fisher Scientific, Inc.). SDS-PAGE and western blot analyses were performed as routine: 10 µg total protein or 2 µg nuclear or cytoplasmic fractions were loaded on a 10% SDS-PAGE gel, and electrophoresis was performed at 60 V for 35 min followed by 90 V for 90 min. Proteins were transferred to PVDF membrane (cat. no. ISEQ00010; ED Millipore, Burlington, USA) overnight. Membrane was then blocked with 5% blocking grade blocker (cat. no. 170-6404; Bio-Rad Laboratories, Inc., Hercules, CA, USA) in 1X TBS-T (Tween-20, 0.1%) for 2 h at room temperature with constant shaking. Primary and secondary antibodies were from Cell Signaling Technology Inc. (Danvers, MA, USA): rabbit anti-β-catenin (dilution 1:1,000; cat. no. 9582), rabbit anti-vinculin (dilution 1:1,000; cat. no. 4650), rabbit anti-Histone H3 (dilution 1:2,000; cat. no. 4499), rabbit anti-Nanog (dilution 1:2,000; cat. no. 4903), mouse anti-β-actin (dilution 1:2,000; cat. no. 3700), and goat anti-rabbit (dilution 1:5,000; cat. no. 7074) or anti-mouse (dilution 1:5,000; cat. no. 7076) IgG. Primary antibodies were incubated overnight at 4°C and secondary antibodies were incubated for 2 h at room temperature. The blots were established using a chemiluminescence detection kit (Pierce ECL; cat. no. 32106) or ECL Plus Western Blotting Substrate (cat. no. 32132) which were both from Thermo Fisher Scientific, Inc.

Total RNA was extracted from cells or tissue samples using TRIzol reagent according to the protocol of the manufacturer (Invitrogen; Thermo Fisher Scientific, Inc.). cDNA synthesis was performed with 1 µg of total RNA using an Omniscript RT kit according to the manufacturer's protocol (Qiagen, Inc., Valencia, CA, USA). cDNA was diluted 1:5 in DEPC-treated nanopure water and used for further analysis. RT-qPCR analysis was performed using a Bio-Rad iQ iCycler detection system with iQ SYBR green supermix (Bio-Rad Laboratories, Inc.). The reactions were performed in a total volume of 10 µl, including 5 µl of 2X iQ SYBR Green supermix, 0.4 µl of primers at 20 pmol/µl and 0.4 µl of cDNA template. Primer sequences are as follows: BCL2L2 (F, GCGGAGTTCACAGCTCTATAC and R, AAAAGGCCCCTACAGTTACCA); Cox-2 (F, CTGGCGCTCAGCCATACAG and R, CGCACTTATACTGGTCAAATCCC); MMP14 (F, GGCTACAGCAATATGGCTACC and R, GATGGCCGCTGAGAGTGAC); MYC (F, TCCCTCCACTCGGAAGGAC and R, CTGGTGCATTTTCGGTTGTTG); hBim (F, TACTCCAGTGCAGTCTCCTC and R, TCCCATCTTTCCTAACACAG); hDppa4 (F, AAAAGCAAGAAGGGAGAGTGA and R, CGGAGATTGCACTGAACTGA); hEsrrb (F, TCAGAGAGCAGCCCATACCT and R, GCGTCACAAACTCCTCCTTC); hOct4 (F, GAGAATTTGTTCCTGCAGTGC and R, GTTCCCAATTCCTTCCTTAGTG); hSox2 (F, ATGGGTTCGGTGGTCAAGTC and R, GTGGATGGGATTGGTGTTCTC); hTbx3 (F, GAAGAAGAGGTGGAGGACGA and R, ATTCAGTTTCGGGGAACAAG); hTcl1 (F, GATACCGATCCTCAGACTCCA and R, GAGGGACAGAAGGGACAGAA); GAPDH (F, CCAGGTGGTCTCCTCTGACTTCAACA and R, AGGGTCTCTCTCTTCCTCTTGTGCTC). All qPCR were run according to the following thermocyclers: Initial denaturation and enzyme activation at 95°C for 3 min, followed by 40 cycles of denaturing (95°C for 15 sec), annealing (55–60°C for 30 sec), and extension (72°C for 30 sec). Melt curve was carried out at 55-95oC (in 0.5oC increments) for 30 sec.

All reactions were performed in four repeats for every sample and with three independent experiments for each group. GAPDH was used as housekeeping gene for normalization. Gene expression was quantified using ∆∆CT method and 2^−∆∆Ct was used as the relative expression changes for each gene (30).

All animal experiments followed a protocol approved by the Institutional Animal Care and Use Committee of the University of Kansas Medical Center (Kansas City, KS, USA). Single treatment and repeated treatments were each used for the measurement of tumorigenecity. In the single treatment model, the PANC-1 pancreatic cancer cells at three densities were used for tumor inoculation (2×10⁴ cells per injection, 2×10⁵ cells per injection, or 1×10⁶ cells per injection). The PANC-1 cells were suspended in PBS as single cell suspension and then mixed with either 200 mg/ml Rau or PBS. At each cell injection number, cells mixed with Rau were injected subcutaneously into the left flank of the mouse, and cells mixed with PBS were injected into the right flank of the same mouse. A total of 10 mice (Athymic Ncr nu/nu, female, 4–6 weeks, 15–20 g) were used for each cell density. The formation of tumors were monitored daily, and longitudinal tumor growth was measured using calipers. Mice were housed in 5 mice/cage in a sterile rodent room with 12-h/12 h-light/dark cycle. Housing was handled by the University of Kansas Medical Center Laboratory Animal Resources following standard protocol.

In the repeated treatment model, a single cell suspension of PANC-1 cells was mixed with 200 mg/ml Rau, and then inoculated into 10 mice at 2×10⁵ cells per injection, in the left and right flanks. Treatment was started the following day with oral gavage of 20 mg/kg Rau, five times per week for 3 weeks. In the control group (10 mice), the mice were inoculated with the same number of cells in PBS, and were then gavaged with an equivalent volume of saline solution. Tumor formation was monitored daily, and longitudinal tumor growth was measured using calipers.

Statistical analysis was performed using IBM SPSS statistical software, version 23 (IBM PSS, Armonk, NY, USA). Student's t-test and a log-rank test were used. P<0.05 was considered to indicate a statistically significant difference.

Five human pancreatic cancer cell lines (PANC-1, MiA PaCa-2, AsPC-1, HPAF-II and BxPC-3) and an immortalized epithelial cell line (MRC-5) were treated with various concentrations of Rau, and cell viability was detected 48 h later. Rau inhibited the proliferation of all five cancer cells (Fig. 1A), with IC₅₀ values ranging between 140 and 317 µg/ml. The non-cancerous MRC-5 epithelial cell line was less affected by Rau treatment, with a higher IC₅₀ value of 567 µg/ml (Fig. 1B). These results are consistent with our previous findings that Rau inhibited the overall proliferation of pancreatic cancer cells (29).

To investigate the inhibitory effect of Rau in CSCs, a tumor spheroid formation assay was performed. The ability to form tumor spheroids is an in vitro indication of the tumorigenic capacity and self-renew ability of CSCs. When cancer cells are cultured in non-adherent, serum-free conditions, non-CSC populations die by anoikis, whereas CSCs overcome anoikis and go through division leading to the formation of tumor spheroids (32,33). Single cell suspensions were treated with Rau and tumor spheroids were counted 4 weeks later. Data showed that Rau significantly reduced the number of the PANC-1 tumor spheroids at the concentrations of 50 and 100 µg/ml, and completely eliminated the tumor spheroids at 200 µg/ml (Fig. 2A and B). The estimated IC₅₀ value for tumor spheroids inhibition is 39.44 µg/ml. By contrast, the IC₅₀ value of Rau to the bulk of PANC-1 cells was 317 µg/ml (Fig. 1B). The MIA PaCa-2 pancreatic cancer cells were also treated by Rau for the detection of tumor spheroids. Similar results were obtained. Rau reduced the number of the MIA PaCa-2 spheroids at 50 µg/ml, and completely inhibited spheroid formation at ≥100 µg/ml (Fig. 2C and D). The estimated IC₅₀ value of 34 µg/ml (Fig. 2D) was lower than the IC₅₀ value for the bulk of the MIA PaCa-2 cells (Fig. 1A).

Cells with stemness features are reported to exclude dyes as side populations (34,35). In order to separate the CSC-like population, MIA PaCa-2 cells were sorted using flow cytometry with DCV staining. The DCV⁻ cells (CSC-like) and DCV⁺ (non CSC-like) cells were collected and treated with Rau. Rau inhibited the viability in all unsorted, DCV⁺ and DCV⁻ cells, preferentially inhibiting DCV⁻ cells (Fig. 3A). The estimated IC₅₀ values were 162 µg/ml in unsorted cells, 177 µg/ml in DCV⁺ cells and 122 µg/ml in DCV⁻ cells. This result suggested that Rau preferentially inhibited CSC-like cells.

Tumor spheroid formation was detected. Although cell spheroids were also formed in the DCV⁺ cell culture, they were significantly smaller (Fig. 3B). By contrast, DCV⁻ cells formed large spheroids, as expected. As there is currently no exclusive way to pin-point pancreatic CSCs, the formation of spheroids in DCV⁺ cells may due to the remaining CSC-like cells in the DCV⁺ population. However, the DCV staining and sorting provided a side population enriched with ‘stemness’. Rau at 50 µg/ml inhibited spheroids in the DCV⁻ and DCV⁺ populations (Fig. 3B), a result consistent with those in unsorted cells.

The effects of Rau on CSCs in a shorter time period were also examined. The PANC-1 cells were treated with Rau for 24 or 48 h at concentrations of 50, 100 or 200 µg/ml. The pancreatic CSC markers CD24 and EpCAM were examined by immune staining followed by flow cytometric analysis. Rau reduced the CD24⁺EpCam⁺ cell population following 24 and 48 h treatment (Fig. 4A and B). In the control groups, CD24⁺EpCam⁺ cells consisted of 6.50–10.72% of the whole population. At the concentration of 200 µg/ml, Rau significantly reduced CD24⁺EpCam⁺ cells to 2.52–2.95% following 24 and 48 h of treatment (Fig. 4A and B). At a lower concentration of 100 µg/ml, Rau also significantly reduced the CD24⁺EpCam⁺ cells to 2.99–5.22% following 24 and 48 h of treatment (Fig. 4B). It was estimated that the IC₅₀ value at 24-h treatment was 142.04±12.40 µg/ml, and at 48 h treatment was 126.09±12.51 µg/ml (Fig. 4A and B), and these values were lower than the IC₅₀ values for the bulk tumor cells. These data are consistent with the results above showing that Rau preferentially inhibited pancreatic CSCs.

One of the essential pathways in maintaining the self-renewal and spheroid formation capacities of CSCs is activation of the canonical Wnt/β-catenin signaling pathway (18,36). When there is active Wnt signaling, the β-catenin degradation complex in the cytosol dissociates, and β-catenin accumulates in the nucleus and functions as a transcriptional factor to upregulate genes that promote CSC stemness, including Nanog (37). In the present study, the cytoplasmic and nuclear fractions of the PANC-1 cells were each examined for β-catenin levels with or without Rau treatment. Treatment with Rau (100 µg/ml) for 24 and 48 h reduced the levels of β-catenin in the nucleus (Fig. 5A), whereas the cytoplasmic β-catenin levels were not changed (Fig. 5A). A panel of β-catenin downstream target genes, including B-cell lymphoma 2-like 2 (BCL2L2), cyclooxygenase-2 (COX-2), matrix metalloproteinase (MMP)14 and MYC, were examined by RT-qPCR analysis (Fig. 5B). Following treatment for 48 h, the expression of MYC was significantly decreased by Rau treatment, which is consistent with Wnt/β-catenin signaling pathway inhibition. Studies have shown that the stem cell-related gene Nanog has the ability to induce β-catenin phosphorylation and enhances its degradation (38). Therefore, the present study examined the expression of Nanog by western blot analysis. Nanog was increased following 24 h of Rau treatment, and was then decreased following 48 h of Rau treatment (Fig. 5C). It was hypothesized that the increase in Nanog at the earlier time point enhanced β-catenin degradation and therefore suppressed nuclear levels of β-catenin. The suppressed β-catenin levels subsequently resulted in inhibition of the expression of Nanog at a later time-point (39,40). As a result, the Nanog and Wnt signaling pathway was suppressed by Rau.

A panel of other CSC-related genes were also examined by RT-qPCR analysis (31). Data showed that the expression levels of developmental pluripotency associated 4 (Dppa4), estrogen related receptor β (Esrrb) and SRY-box 2 (Sox2) were inhibited following 48 h of Rau treatment (Fig. 5D).

Taken together, Rau treatment had an early effect in increasing the expression of Nanog, which led to the phosphorylation and degradation of β-catenin, and repressed the nuclear level of β-catenin. The decreasing level of nuclear β-catenin negatively affected the expression of Nanog. Rau treatment also appeared to directly inhibit the nuclear accumulation of β-catenin and other CSC-related genes, including MYC, resulting in the overall suppression of Nanog and nuclear β-catenin (Fig. 5E). The full mechanism of Rau-induced CSC inhibition warrants further investigation.

The inhibitory effects of Rau against pancreatic CSCs were examined in vivo by tumorigenicity in immunocompromised mice. Single treatment was performed first using inoculation of different numbers of PANC-1 cells at limited dilutions. The cells (2×10⁴, 2×10⁵ and 1×10⁶) were mixed with 200 mg/ml Rau and injected subcutaneously into the left flanks of nude mice (n=10), respectively. For the control, the same number of cells were mixed with PBS and inoculated into the right flanks of the same mouse. The results are shown in Fig. 6A-F. At the lowest inoculation number (2×10⁴ cells), the tumor formation rate was low and no difference was observed between the treated and untreated groups. At the highest inoculation number (1×10⁶ cells per injection), the untreated group reached a maximum of 90% tumor formation and the treated group reached 80%, with no significant difference between the two. There was also no difference in the growth of the formed tumors between the treated and untreated groups. At 2×10⁵ cells per injection, the single Rau treatment significantly inhibited the tumor formation rate. The growth of the formed tumors was also inhibited.

As single Rau treatment showed limited effect on the inhibition of tumor formation rate and tumor size, repeated treatment was performed with oral administration of Rau. The optimal cell number for injection was selected as 2×10⁵ per injection. The mice (n=10) were injected subcutaneously in the left and right flanks with PANC-1 cells mixed with 200 mg/ml of Rau. Treatment started the following day and lasted for 3 weeks with oral gavage of 20 mg/kg Rau, five times per week. The control mice (n=10) were inoculated with the same number of cells mixed with PBS, and were administered with equivalent volumes of saline.

The rate of tumor formation and time of tumor formation were significantly different between the control and treated groups (Fig. 7A). At day 6, the tumor formation rate in the control group reached 80%, whereas that in the Rau-treated group was only 35%. At day 20 when the treatment had stopped, all mice in the control group were bearing tumors on both flanks (100% tumor formation), whereas the Rau-treated group only had 65% tumor formation. All mice were kept for 2 months following the end of treatment. At the end of the experiment, the Rau-treated group had a maximum of 85% tumor formation, compared with 100% tumor formation in the control group. These data indicated that Rau administration at 20 mg/kg orally eliminated CSCs in 15% of the injection sites.

The growth of the formed tumors was not significantly inhibited by Rau treatment compared with the control group, which indicated the lack of a long-term inhibitory effect on tumor growth following the end of treatment (Fig. 7B). No adverse effects were observed in either group during the treatment (Fig. 7C).