The following monoclonal antibodies (mAbs) and polyclonal antibodies (pAbs) were used: anti-KLF4 (pAb, ab106629), and anti-ZEB1 (pAb, ab155249) were purchased from Abcam (Cambridge, MA, USA); anti-GAPDH (pAb) was purchased from Anbo Biotechnology Co. (San Francisco, CA, USA); Cell counting kit-8 was purchased from Dojindo Laboratories (Tokyo, Japan). Gemcitabine was purchased from Eli Lilly (Indianapolis, IN, USA). ZEB1 siRNA, miR-200b and miR-183 were synthesized by Biosune Biotechnology (Shanghai, China).

The human PDAC cell lines BxPC-3, Panc-1 and MIApaca-2 were purchased from the American Type Culture Collection (Manassas, VA, USA). Cells were cultured in RPMI-1640 medium (Hyclone, USA) supplemented with 10% fetal bovine serum (Hyclone), 100 U/ml penicillin, 100 U/ml streptomycin and 0.03% L-glutamine at 37°C in 5% CO₂.

PDAC cells were seeded at a density of 1.0×10⁴ cells/well in 96-well microplates with 100 µl RPMI-1640 medium. After grown to ~80% confluence, the cells were incubated with the indicated concentration of gemcitabine. After treatment for 72 h, the cells were incubated with 10% WST-8 dye for 2 h at 37°C. The absorbance was detected at 450 nm using a SpectraMax M2. IC₅₀ values were determined using GraphPad Prism 6.0.

PDAC cells were seeded at a density of 2×10⁵ cells/well in 6-well microplates. After treatment, flow cytometric analysis was performed to determine cell apoptosis. In brief, after treatment with gemcitabine, the cells were digested using trypsin lacking EDTA and phenol red, and were incubated in binding buffer containing Annexin V-FITC (2.5 mg/ml) and propidiumiodide (5 mg/ml) for 10 min in the dark at room temperature. Subsequently, the labeled cells were detected by a flow cytometer (Beckman Coulter, Chicago, IL, USA).

For KLF4 knockdown, the short hairpin RNA (shRNA) targeting the coding sequence of KLF4 (TACCCATCCTTCCTGCCCGAT) was digested using the restriction enzymes AgeI and EcoRI (New England Biolabs UK Ltd., UK), and was then cloned into pGCSIL-GFP plasmid. For KLF4 overexpression, the full coding region of KLF4 was amplified using specific primers (forward, 5′-CGCGGATCCGCGATGGCTGTCAGCGACGCG-3′; reverse, 5′-GGGTACCGGTCGCCACCTTCTCTTCTGGCAGTGTGGG-3′). The product was digested using the restriction enzymes BamHI and AgeI, and was then linked to p-GC365-EGFP plasmid. The plasmids were transformed into E. coli DH5a. Positive clones were harvested and identified by sequencing. The recombinant plasmids were transfected to 293T cells by Lipofectamine 2000 (Invitrogen). The infection efficiency was determined by measuring fluorescence intensity under microscopy (Olympus 3.3RTV, Japan). After 48 h of transfection, the supernatant was collected by centrifugation at 4,000 g for 10 min at 4°C, followed by ultrafiltration centrifugation at 4000 g for 12 min. The concentrated solution was stored at −80°C for future experiments.

BxPC-3, Panc-1 and MIApaca-2 cells were seeded at a density of 2×10⁵ cells/well in 6-well microplates. After growing overnight, the cells were transfected with lentivirus carrying shRNAs against KLF4 or KLF4 cDNA (GeneChem, Shanghai, China). For ZEB1 knockdown, PDAC cells were transfected with siRNA against ZEB1 (GGTAGA TGGTAATGTAATA) using Lipofectamine 3000 (Invitrogen).

Western blot analysis was performed as previously described (11). Equal amounts of protein were loaded to sodium dodecyl sulfate polyacrylamide gels. The protein was subsequently transferred onto PVDF membranes. After blocked with 5% BSA, membranes were incubated with the indicated primary antibodies overnight at 4°C, followed by treatment with horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. The protein signals were detected using the ECL kit and quantified using an Alpha Imager 2200 (Alpha Innotech Corp., USA).

The total RNA was extracted using TRIzol reagent (Invitrogen). RNA (20 µg) was subjected to 15% urea-polyacrylamide gels, and was then transferred to nylon membrane. The expression of miR-200b, miR-183 and U6 were detected using biotin-labeled oligonucleotide probes (5′-TCATCATTACCAGGCAGTATTA-3′, miR-200b; 5′-CAGTGAATTCTACCAGTGCCAT-3′, miR-183 and 5′-ATATGGAACGCTTCACGAATT-3′, U6). The signals of blots were measured using the Fujifilm LAS-4000 imaging system.

Panc-1 and MIApaca-2 cells were co-transfected with increasing concentrations of pcDNA3.1 carrying KLF4 and 0.2 µg of luciferase reporter plasmid containing the miR-200b or miR-183 promoter using Lipofectamine 3000 (Roche); the pRL-SV40 plasmid (Promega) was transfected as a normalization control. After 24-h incubation, cell lysates were collected to determine luciferase activity using the Dual-Luciferase assay (Promega) according to the manufacturer's instructions. The primers were used as follows: 5′-ATATGCGGCCGCGTTTTGGCTTCGTTTCTTCT-3′ and 5′-GCTTGGTACCGACCCCATCTGTTCTTTGATT-3′ for KLF4; 5′-AGAAAGGGTGGGAAGGAGGACA-3′ and 5′-GGACTCGCTGGGAAGCTCAGTA-3′ for miR-200b; 5′-ATTGTAGTAAGGGAAACTGAGGC-3′ and 5′-GCAGAAGTGGGTAAGGTGCT-3′ for miR-183. The sequence of ZEB1-3′UTR containing the miR-200b seed regions were amplified using specific primers (forward, 5′-CGAGCTCATTGTTTTATCTTATCAGTATTATC-3′; reverse, 5′-CCGCTCGAGAACTAAAAGAAATAAAATAATACTG-3′) SacI and XhoI restriction sites. The PCR product was digested using the restriction enzymes SacI and XhoI, and was subsequently linked to pHSA-MIR-Report (Ambion). Mutations in the miR-200b seed regions of the ZEB1-3′UTR were generated using the QuikChange Multi-site-directed mutagenesis kit (Stratagene). For miR-183, two targeting sequences locating at position 880-886, 1801-1807 of ZEB1-3′UTR were amplified using forward (5′-CGAGCTCAGTGCCATTTCTCAGTATTTTCAAG-3′) and reverse (5′-CCGCTCGAGAGTGCCATTTCTCAGTATTTTCAAG-3′) primers containing SacI and XhoI restriction sites. The corresponding mutagenesis was generated using specific primers (forward, 5′-CGAGCTCACACGGATTTCTCAGTATTTTCAAG-3′; reverse, 5′-CCGCTCGAGAGACGGATTTCTCAGTATTTTCAAG-3′).

ChIP was performed using a ChIP assay kit (Millipore) according to the manufacturer's instructions. In brief, Panc-1 cells were crosslinked with fresh 1% formaldehyde, followed by incubation in SDS lysis buffer containing 1% protease inhibitors. The cell lysates were sonicated to shear crosslinked DNA to 200-1,000 bp in length. Protein/DNA complexes were immunoprecipitated with 3 µg anti-ZEB1 antibody or control IgG. The complexes were eluted from the antibodies, and were dissociated with 5 M NaCl. ChIP samples were analyzed with real-time quantitative PCR using primers specific for the promoters of miR-200b and miR-183. The primers are listed in Tables I and II.

Total RNA was extracted using TRIzol reagent (Invitrogen). The expression of miR-200b and miR-183 were analyzed using TaqMan miRNA assays (Applied Biosystems, CA, USA) according to the manufacturer's instructions. U6 snRNA was used as an internal control.

Immunohistochemistry was performed as previously described (12). Briefly, tumor sections were fixed in formalin, embedded with paraffin, deparaffinized in xylene, and hydrated with ethanol. After antigen retrieval using microwave, the tissue sections were treated with 3% hydrogen peroxide and blocked with 10% goat serum. The sections were then stained with anti-KLF4 and anti-ZEB1 antibodies, followed by treatment with a biotinylated secondary antibody. The images were captured using a fluorescence microscope (Leica DM IRE2).

The orthotopic PDAC nude mouse model was established as previously described by us (12), with a minor modification. Briefly, 4- to 6-week-old BALB/c nude mice of both sexes were obtained from Weitonglihua Animal Center (Beijing, China) and maintained under specific pathogen-free conditions in the animal facility. All procedures were approved by the Institutional Laboratory Animal Care and Use Committee at Shandong Provincial Hospital. Panc-1 cells (1×10⁷) in 100 µl PBS/Matrigel (1:1, v/v, BD Biosciences, USA) transfected with control lentivirus or lentivirus carrying KLF4 cDNA were injected into both flanks of the mice. One week after injection, gemcitabine (80 mg/kg; q3 days) was administered via intraperitoneal injection for six weeks. The long (L) and short axes (S) of the tumors were measured weekly using Vernier calipers, and the tumor volumes were calculated as follows: V = (L x S²) x π/6. The mice were sacrificed 24 h after the last injection, and the tumors were harvested for subsequent analysis.

Statistical analysis was carried out using the SPSS 18.0 software package. All data are presented as the mean ± standard deviation from at least three independent experiments. Student's t-test was used for comparisons between two groups. One-way ANOVA was used for comparisons among multiple groups. p<0.05 was considered statistically significant.

First, we measured the expression of KLF4 and ZEB1 in BxPC-3, Panc-1 and MIApaca-2 cells. Fig. 1A showed a higher level of KLF4 in BxPC-3 cells than those in Panc-1 and MIApaca-2 cells while lower levels of ZEB1 in BxPC-3 cells than those in Panc-1 and MIApaca-2 cells. Exposure of BxPC-3 cells to gemcitabine caused more apparent decline in cell viability (IC₅₀=3.942 µM) compared to Panc-1 (IC₅₀=9.383 µM) and MIApaca-2 cells (IC₅₀=7.904 µM) (Fig. 1B). Flow cytometry analysis demonstrated that treatment with 10 µM gemcitabine induced the death of 68.34% of the BxPC-3 cells, 50.45% of the Panc-1 cells and 51.02% of the MIApaca-2 cells (Fig. 1C), suggesting that BxPC-3 cells were more sensitive to gemcitabine than Panc-1 and MIApaca-2 cells. Further investigation found that gemcitabine treatment reduced the expression of KLF4, miR-200b and miR-183 in a dose-dependent manner, whereas led to a dose-dependent increase in ZEB1 expression (Fig. 1D and E).

Next, we examined the correlation between KLF4 and ZEB1. KLF4 knockdown significantly promoted ZEB1 expression while KLF4 overexpression inhibited ZEB1 expression (Fig. 2A). KLF4 depletion inhibited gemcitabine-induced activation of caspase 3 while KLF4 overexpression promoted the activation of caspase 3 even in the absence of gemcitabine (Fig. 2B), suggesting that KLF4 downregulation conferred gemcitabine resistance of PDAC cells. KLF4 knockdown restrained the expression of miR-200b and miR-183 (Fig. 2C). In contrast, KLF4 overexpression elevated the levels of miR-200b and miR-183 (Fig. 2C). In KLF4-depleted cells, forced expression of miR-200b and miR-183 repressed the ZEB1 expression again (Fig. 2D). In KLF4-overexpressing cells, forced expression of anti-miR-200b and anti-miR-183 restored the ZEB1 expression (Fig. 2E). Restoration of miR-200b and miR-183 activated caspase 3 in KLF4-depleted cells whereas anti-miR-200b and anti-miR-183 significantly reduced caspase 3 activation in KLF-overexpressing cells (Fig. 2F). Overexpression of miR-200b and miR-183 in KLF-4-depleted cells promoted cell apoptosis in the presence of gemcitabine, but inhibition of miR-200b and miR-183 in KLF-4-overexpressing cells blocked cell apoptosis induced by gemcitabine (Fig. 2G). These results suggested that KLF4 knockdown contributed to ZEB1 expression and gemcitabine resistance by suppressing miR-200b and miR-183.

Since KLF4 is a transcriptional factor that regulates the expression of multiple genes, we hypothesized that KLF4 directly regulated the expression of miR-200b and miR-183. The putative binding sites of KLF4 located on the promoters of miR-200b and miR-183 were shown in Fig. 3A and D. Chromatin immunoprecipitation (ChIP) assays revealed that KLF4 could interact with site 4 and 5 in the miR-200b promoter and the site 4 in the miR-183 promoter (Fig. 3B and E). We subsequently constructed the pGL3-200-luc and pGL3-183-luc plasmids by inserting these sequences into the pGL3 luciferase reporter. Panc-1 and MIApaca-2 cells were co-transfected with 0.2 µg pGL3-200-luc or pGL3-183-luc and the indicated concentration of pcDNA3.1 carrying KLF4 together with 100 ng pRL-SV40. As shown in Fig. 3C and F, with the increasing dose, KLF4 substantially enhanced luciferase activities. These data suggested that KLF4 positively regulated the expression of miR-200b and miR-183 by directly binding to their promoters.

We further explored whether miR-200b and miR-183 directly targeted the-3′UTR region of ZEB1. Targetscan (http://www.targetscan.org/) was used to predict the recognition sites of miR-200b and miR-183. As shown in Fig. 4A, four putative sites for miR-200b and two putative sites for miR-183 were found. The sequences containing these sites were cloned into the luciferase reporter plasmid to obtain the ZEB1-3′UTR-WT-luc plasmid (wild-type), and the sequence containing the mutant sites was cloned to construct ZEB1-3′UTR-MT-luc plasmid (mutant type). Subsequently, the Panc-1 cells were co-transfected with ZEB1-3′UTR-WT or ZEB1-3′UTR-MT and miR-200b or miR-183. Both miR-200b and miR-183 significantly inhibited luciferase activities in the cells carrying ZEB1-3′UTR-WT compared to those carrying negative control (NC) and ZEB1-3′UTR-MT (Fig. 4B). Furthermore, both miR-200b and miR-183 notably reduced mRNA and protein levels of ZEB1 in BxPC-3, Panc-1 and MIApaca-2 cells (Fig. 4C and D). Conversely, both anti-miR-200b and anti-miR-183 elevated mRNA and protein levels of ZEB1 (Fig. 4C and D). These results indicated that miR-200b and miR-183 directly targeted ZEB1.

As shown in Fig. 5A, gemcitabine treatment promoted ZEB1 expression, which was consistent with the results in Fig. 1D. ZEB1 knockdown mildly activated caspase 3. A combination of ZEB1 knockdown and gemcitabine significantly increased cleaved caspase 3, suggesting that ZEB1 knockdown facilitated gemcitabine sensitivity of PDAC cells. Fig. 5B illustrated that gemcitabine treatment induced cell apoptosis in ZEB1-depleted cells. These data indicated that ZEB1 knockdown contributed to gemcitabine sensitivity of PDAC cells.

We established xenograft tumors of PDAC to determine the effect of KLF4 overexpression on the ZEB1 expression and gemcitabine resistance of PDAC in vivo. The results showed that KLF4 overexpression reduced ZEB1 expression, consistent with the data in vitro (Fig. 6A–C). Linear regression analysis showed that KLF4 expression was negatively correlated to ZEB1 expression (Fig. 6D). Compared to the control group, KLF4 overexpression did not significantly affect tumor growth, whereas it enhanced gemcitabine sensitivity (Fig. 6E and F). These results suggested that KLF4 overexpression attenuated gemcitabine resistance of PDAC in vivo by negatively regulating ZEB1 expression.