The human pancreatic cancer cell line BxPC-3 was obtained from the American Type Culture Collection (Rockville, MD). Monolayer cultures of BxPC-3 cells were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum, 1% penicillin-streptomycin (Invitrogen, Carlsbad, CA). Cultures were maintained at 37°C in a humidified atmosphere of 95% air and 5% CO₂.

Seven to eight-week old immunodeficient male mice (BALB/c nu/nu) were purchased from Shanghai SLAC Laboratory Animal Co., Ltd., China, and housed in laminar-flow cabinets under specific pathogen-free conditions. All animal procedures were approved by the institutional animal care committee, Fudan University, China.

The BxPC-3 cell subline was established by a serial in vivo selection process performed as previously described with some modifications (19,20) (Fig. 1). Each male nude mouse (n=6) was injected subcutaneously in the left hind footpad with 0.1 ml of BxPC-3 cells (1×10⁷/ml). The mice were sacrificed at 8 weeks after tumor implantation. The possible metastatic lymph nodes, including homolateral popliteal, inguinal, iliac, para-aortic, para-renal, and mesenteric lymph nodes, were dissected and dissociated sterilely. Metastatic lymph nodes were confirmed by hematoxylin and eosin staining. The suspension was filtered using a 200-metal mesh strainer (Xinrui Biotechnology, Shanghai, China) and centrifuged at 1000 rcf/ min for 5 min at 4°C. The cell pellet was resuspended in culture medium and cultivated until stable populations of human tumor cells were established under the same conditions as detailed above. Cultured tumor cells were then serially re-inoculated into new recipient mice. After five rounds of in vivo selection, a variant of the BxPC-3 human pancreatic cancer cell line (BxPC-3-LN) with high lymphatic metastasis capacity was established.

Cell morphology was viewed and photographed under a light microscope (Estativo Eclipse TE 2000S, Nikon, Spain). For ultrastructural observation, cells were fixed with 2.5% glutaraldehyde in 0.1 M PBS for 2 h at room temperature. For observation under scanning electron microscope (SEM), the specimens were immersed in isoamyl acetate, air-dried, and sputter-coated with gold before examination under the Hitachi S-520 microscope (Hitachi Ltd., Tokyo, Japan). For transmission electron microscopy (TEM), the specimens were immersed twice in absolute propylene oxide and embedded in Quetol 812. Following staining with uranyl acetate and lead citrate, the specimens were observed with a Jeol JEM-1230 electron microscope (Jeol, Tokyo, Japan) at 80 kV.

Cells were harvested, fixed by resuspension in 10 ml of 70% ethanol for 30 min, and washed in ice-cold PBS. They were incubated, washed and resuspended in DNA staining solution containing 20 μg/ml propidium iodide (Sigma-Aldrich) and 100 μg/ml RNase (Invitrogen). DNA content was determined by flow cytometry (BD FACS Aria, USA).

To evaluate in vitro growth kinetics, growth curves were prepared for each cell line. A total of 2×10³ cells/well were plated in triplicate in a 96-well plate and grown for 1–7 days. Cells were then incubated at 37°C for 4 h with 20 μl of MTS (Promega) per well, and absorbance was read at 492 nm. The number of viable cells was counted at the indicated time-points in triplicate.

The cell adhesion assay was performed as previously described with a slight change (21). Twenty-four-well plates were coated with 10 μg/ml of rat tail type IV collagen (BD Biosciences, Bedford, MA). The wells were blocked with 1% BSA for 1 h at 37°C. After the BSA solution was aspirated, 1×10⁶ cells in 0.1% BSA were added to the wells in triplicate. At 15 and 30 min, the wells were washed to remove unattached cells, incubated at 37°C for 4 h in 100 μl MTS and measured for absorbance at 492 nm. The assay was repeated at least 3 times.

The cell migration assay was performed using a 24-well transwell chamber (Corning Costar, Corning, NY) migration assay as previously described with some change (22). Briefly, exponentially growing cells were trypsinized and washed in RPMI-1640/0.1% BSA medium twice. Cells (∼1×10⁶) were seeded in the upper chamber of each transwell unit, which contained a polycarbonate 8-μm pore membrane pre-coated with 50 μl of 10 μg/ml fibronectin (BD Biosciences) overnight. The cells were allowed to migrate at 37°C and 5% CO₂ for 24 h. The unattached cells were rinsed off with PBS and the membranes containing attached cells were fixed in 10% formalin for 30 min and washed with PBS. The cells were stained with crystal violet for 20 min and rinsed with water. Cells on the unmigrated side were gently wiped off with a wet cotton tip applicator and the membrane was rinsed with water. The membranes containing the migrated cells were dried and mounted onto slides with permount. The number of migrated cells per high power field (hpf) was determined by averaging 20 randomly counted hpfs. The assays were performed in triplicate and were reproducible in at least three batches of independent experiments.

Transwell chamber inserts (Corning Costar) with filter membrane pore size of 8 μm were coated with 0.5 mg/ ml Matrigel (BD Biosciences) overnight. The membranes were rehydrated and 1×10⁵ cells in serum-free RPMI-1640 medium were placed onto the inner chamber of each insert unit. Media with 10% FBS was placed into the outer chamber as a chemoattractant. The cells were allowed to invade at 37°C and 5% CO₂ for 24 h. Cells were fixed in 10% formalin and washed with PBS. The cells were stained with crystal violet and rinsed with water. Cells on the unmigrated side were gently wiped off with a wet cotton tip applicator and the membrane was rinsed with water. The membranes containing the invaded cells were dried, and mounted onto slides with permount. The number of migrated cells per hpf was determined by averaging 20 randomly counted hpfs. The assays were performed in triplicate and were reproducible in at least three batches of independent experiments.

Cells were plated at a density of 10⁴ cells per well into 96-multiwell plates and were allowed to grow for 24 h before the experiment. The different concentrations (from 0 to 10⁴ nM) of gemcitabine resuspended in 100 μl RPMI-1640 medium were added to the cells, which were then incubated for 24 h at 37°C. After initial incubation, cells were incubated at 37°C for 4 h with 20 μl of MTS (Promega), and absorbance was read at 492 nm. Chemosensitivity was expressed as the drug concentration that inhibited cell proliferation by 50% (IC₅₀ values). The number of viable cells was counted at the indicated time-points in triplicate. Expression of chemoresistance genes related to gemcitabine, including ribonucleotide reductase M1 (RRM-1), ribonucleotide reductase M2 (RRM-2), human equilibrative nucleoside transporter-1 (hENT-1), and deoxycytidine kinase (dCK), were examined by reverse-transcription polymerase chain reaction (RT-PCR) as previously described (23).

Eight-week old mice were injected with 1×10⁶/ml BxPC-3 and BxPC-3-LN cells into the footpad (n=6). The primary tumors were allowed to form and the weight was measured every week. Tumor volume was calculated according to the formula: (length in centimeters x (width)²)/2. Six mice in each group were sacrificed separately at 8-week end-points to examine lymphatic involvement. The possible metastatic lymph nodes were divided into three groups: N1 (popliteal, inguinal), N2 (iliac, para-aortic), N3 (para-renal). At end-point, animals were euthanized and autopsies were performed. Organs (including lymph nodes, lungs, liver, spleen, pancreas, kidneys, ovaries, and any other tissues of abnormal appearance) were examined superficially for evidence of metastases using a dissecting microscope. Upon sacrifice, the tumors, metastatic lymph nodes, and other organs were fixed for hematoxylin and eosin staining.

Total RNA was extracted from BxPC-3 and BxPC-3-LN cells, respectively, with TRIzol (Life Technologies, USA) and dissolved in Milli-Q H₂O. The RNA from BxPC-3 was labeled with Cy3-dUTP and that from BxPC-3-LN with Cy5-dUTP. Fluorescent probe mixtures were denatured at 95°C for 5 min, and applied to the prehybridized chip (Affymetrix, Inc., USA) under a cover glass, which was hybridized at 42°C for 15–20 h. After washing, the microarray was scanned on a scanner (Agilent, G2565BA) and the image was analyzed using Feature Extraction software (Agilent). The signal intensity of each spot was calibrated by subtraction from the intensity of the negative control. The assay was performed twice and only genes showing the same tendency of differential expression were eligible for the final-result analysis.

Fisher’s exact test and Student’s t-test were used for comparisons of enumeration data and to measure mean t data, respectively. The statistical analysis software package Stata 10.0 was used for the tests, and P<0.05 was considered statistically significant.

Under a light microscope, the morphology of the BxPC-3-LN cell line was similar to the parental BxPC-3 cell line observed in culture. Both of them demonstrated polygonal epithelial morphology with large nuclei and 3–6 conspicuous nucleoli (Fig. 2A1 and A2). In addition, under SEM, in comparison to the parental cells, the BxPC-3-LN cells exhibited discernible aggressive morphological alterations such as filopodia and lamellipodia, protrusions emerging from the cellular membrane (Fig. 2B1 and B2). Under TEM, compared to the parental cells, the BxPC-3-LN cells contained more secretory granules (∼50–400 nm in diameter) (Fig. 2C1 and C2).

Cell cycle analysis by flow cytometry revealed that the percentage of BxPC-3-LN cells in S phase (13.3%) was similar to that of the parental cells (12.2%) (Fig. 3A). Growth curves were employed to evaluate the in vitro growth kinetics through MTS assay. BxPC-3-LN cells exhibited similar in vitro proliferation to BxPC-3 cells (Fig. 3B).

As shown in Fig. 3C, the highly metastatic BxPC-3-LN cells had decreased relative adherence compared to the parental BxPC-3 cells. At the 15- and 30-min time-points, the decrease in relative adherence ranged from 36.0 to 57.0% (P<0.05). Conversely, BxPC-3-LN cells demonstrated increased cell migration and invasion compared to the parental BxPC-3 cells (Fig. 4). In fact, when the migrated cells were counted, there was more than twice the number of cells migrated per high power field in BxPC-3-LN samples than in BxPC-3 samples (Fig. 4B1, B2 and B3).

We examined the cellular sensitivity of BxPC-3 cells and BxPC-3-LN cells to gemcitabine in vitro using the proliferation assay. Cells were treated with increasing concentrations of gemcitabine (50–300 nM) (Fig. 5A). BxPC-3-LN cells exhibited native resistance to gemcitabine and the IC₅₀ value of BxPC-3-LN cells (210.5 nM) was higher than that of BxPC-3 cells (125.3 nM) (P<0.05). In addition, expression of chemoresistance genes related to gemcitabine was examined by RT-PCR. RRM-1, RRM-2 and hENT-1 expression was higher in BxPC-3-LN cells compared to BxPC-3 cells (Fig. 5B).

After five rounds of consecutive in vivo selection, the lymphatic metastases of BxPC-3-LN cells were obvious. When injected into nude mice, BxPC-3-LN cells had a >12-fold increase in presence of lymphatic metastases compared to the parental BxPC-3 cells (Fig. 6 and Table I). BxPC-3-LN cells also formed larger primary tumors when compared to the parental BxPC-3 cells (Fig. 6C). Total lymph node and total animal metastatic ratio of the BxPC-3-LN cells was 71.1 and 100.0%, which was higher than that of the parental BxPC-3 cells (5.6 and 16.7%, respectively) (Table I). No metastases were found in the liver, other abdominal viscera or the lungs.

The differences in the gene expression levels between the two cell lines were assessed by measuring the ratios of their expression. Genes were identified as differentially expressed if the absolute value of the natural logarithm of the ratios was <0.3 or >2.0. Thirty-six genes were identified as being differentially expressed, including 6 with decreased expression and 30 with increased expression in BxPC-3 vs BxPC-3-LN (Fig. 7).