Establishment and characterization of clear cell renal cell carcinoma cell lines with different metastatic potential from Chinese patients

We succeeded in maintaining the cultured cells from clinical specimens of the two patients and culturing them for more than 100 generations in DMEM with 10% FCS. The cells from both patients were proven to have strong tumorigenicity in nude mice. Primary cultures were also made using the subcutaneous and orthotropic tumors derived from the two cell lines. The successive ccRCC cell lines derived from the metastatic site of the patient (No.375771) and from the primary ccRCC tissue of the patient (No.378570) were termed as MRCC and NRCC, respectively. MRCC and NRCC cells were reserved in Chinese Center for Type Culture Collection (CCTCC) with store numbers CCTCC-C200909 and CCTCC-C200910, respectively. The following assays were carried out using the two cell lines of 50–60 generations.

Figure 1 shows the morphology of the cell lines and parental tissues. The cell lines had clear cytoplasm, round to oval nuclei with one or two nucleoli, and high nuclear-to-cytoplasmic ratio. NRCC cells were typically epithelial-like and spindle-shaped and proliferated in a pavement-like cell arrangement with distinct border. As compared to NRCC, MRCC cells were relatively larger in size with irregular shapes and a lack of distinct border. MRCC proliferated in a pavement-like cell arrangement with a lack of contact inhibition. The entire cells of MRCC and NRCC were shown as Figure 2a and 2e, respectively. The two cell lines were of epithelial cell origin as microvilli detected on their cell surfaces (Figure 2b and 2g). Ultrastructurally, the cell lines displayed characteristic filmy cytoplasm containing abundant glycogen granules, lipid droplets, and poorly developed mitochondria (Figure 2c and 2f) and nuclei with prominent nucleoli, which was consistent with the previous EM findings of ccRCC [8, 9]. Interestingly, the glycogen granules were more clustered in MRCC than in NRCC (Figure 2d and 2h).

The doubling times of NRCC cells and MRCC cells were 35.5 h and 55.2 h, respectively. However, MRCC cells showed stronger anchorage-independent growth than NRCC cells. Colony formation rates of MRCC cells and NRCC cells in the agarose were 45.3% ± 4.9% and 31.7% ± 4.5% (p < 0.05), respectively. In the invasion assay, it was found that the invasive capacity of MRCC cells surpassed that of NRCC cells (average invasion rates: 43.6% ± 5.8% vs. 30.2% ± 4.6%, p < 0.001). These results indicated that MRCC cells could be more invasive than NRCC cells.

The number of chromosomes in MRCC cells ranged from 45 to 68 with a modal number of 66; while that in NRCC cells ranged from 53 to 86 with a modal number of 82. Gains of chromosomes, rather than loss of chromosomes, were the major abnormalities in chromosome number. Table 1 shows the gain of chromosomes in the two cell lines. Two significant chromosomal translocations, der(6)t(3qter → q12::6q11 → pter) and der(11)t(11pter → q23::1q23 → qter) were identified in 80.0% and 96.7% of MRCC cells, respectively. Furthermore, a marker chromosome (mar1) was observed in 93.3% of MRCC cells. One deletion, del(4)(pter → q25:) and a translocation, der(1)t(1qter → q11::13q11 → qter) were found in 93.3% and 36.7% of NRCC cells, respectively. Figure 3 presents typical karyotype of each cell line for chromosomal abnormality in both number and structure.

Taking into account a previous report about stem-like cell markers of ccRCC [10], we examined the expression of CD105, CD133, CD44, CD24, CD56, CD99, and CD74 on MRCC and NRCC cells by cytometry. It was found that both cell lines were positive for CD44 but negative for CD133, CD105, and CD74. The positive rate of CD24 was higher in MRCC cells (Figure 4a) than in NRCC cells (Figure 4b) and the same was true for CD56. However, the positive rate of CD99 was higher in NRCC cells than in MRCC cells. We also examined the expression of epithelial and mesenchymal markers. In contrast to N-cadherin expression, E-cadherin expression was less frequent in MRCC cells than in NRCC cells. The expression of vimentin was a little higher in MRCC cells than in NRCC cells. These data indicated that NRCC cells displayed more epithelial characteristics, while MRCC cells tended to be more mesenchymal-like.

Cell cycle analysis showed that 18.9% of MRCC cells were in S phase; while 13.8% of NRCC were in S phase (Figure 5). Interestingly, a typical sub-G1 peak was frequently detected in NRCC cells, rather than in MRCC cells, indicating a possible apoptosis undertaken in NRCC cells.

Since the doubling times of the two cell lines were different, the transcription of interleukin-6 (IL-6), tumor necrosis factor-α (TNFα), vascular endothelial growth factor (VEGF), matrix metalloproteinase-2 (MMP2), hypoxia-inducible factor (HIF) 1α, HIF2α, glycogen synthase kinase 3 beta (GSK3ß), RhoC, Ubiquitin specific peptidase 6(USP6), AHNAK nucleoprotein (AHNAK), Leucine-rich repeat kinase 2 (LRRK2), SLIT-ROBO Rho GTPase activating protein 3 (SRGAP3) in MRCC cells and NRCC cells were examined at 24 h, 48 h and 72 h, respectively, after cell split. The expression patterns of the examined genes are shown in Figure 6. In general, the expression of the most genes at 24 h culture wasn’t significant different between MRCC and NRCC cells, except that the expression of TNFa, IL-6, and AHNAK was significantly higher in MRCC cells than in NRCC cells (p < 0.05 for each). Together with cell invasion assay performed at 24 h after cell split, the three genes were supposed to be involved in cell mobility and invasion. The expression of LRRK2 and USP6 was not significantly different between MRCC and NRCC cells. The expression of SRGAP3 was significantly higher in NRCC than in MRCC cells at 48 h and 72 h after cell spilt (p < 0.05). The levels of TNFα, IL-6, VEGF, and MMP2 were higher in MRCC cells than in NRCC cells at 72 h after cell split (p < 0.05 for each). The same trend was observed for HIF2α expression at 48 h after cell split (p < 0.05). However, the expression of HIF1α was higher in NRCC cells than in MRCC cells 72 h after cell split but the difference did not reach a significant level. The expression of GSK3β was lower in MRCC cells than in NRCC cells at 72 h after cell split (p < 0.05). Furthermore, the RhoC expression was higher in MRCC cells than NRCC cells at 48 h after cell split (p < 0.05). Western blot indicated that cytosolic IκBα protein was more degraded in MRCC cells than in NRCC cells within 60 min after the in vitro treatment with TNFα (Figure 7), indicating nuclear factor-kappa B (NF-κB) signaling pathway is more active in MRCC than in NRCC cells.

Subcutaneous transplantation of MRCC cells or NRCC cells generated tumors in nude mice within two weeks. No lung metastasis was detected following the first round of surgical orthotopic implantation (SOI) with MRCC and NRCC tumors. However, ccRCC metastasized to lung was frequently detected in the mice since the second cycle of SOI with MRCC tumors, and the incidence of lung metastasis was nearly 100%. Furthermore, the metastasis to lymph nodes near ventral aorta and hemorrhagic ascites were frequently evident in the SOI mice transplanted with MRCC tumors. With the increase of the cycles of the transplantation with MRCC, the duration of orthotopic tumor formation to about 10 mm in diameter became shorter and the incidences of hemorrhagic ascites and cachexy became higher (Table 2). We isolated and established the metastatic cell strain from pulmonary tumor mass, named MRCC-L. MRCC-L looked smaller in size and grew faster in vitro than their parental MRCC. Orthotopic transplantation of NRCC also generated tumor in all mice. However, metastasis was not observed at the first three rounds of the transplantation with NRCC cells. These data suggested that MRCC had higher malignant and metastatic potential than NRCC.

A 62-year-old male patient (No.375771) underwent surgical resection for a growing lesion in the spine at the 2nd affiliated hospital on February 6, 2006. This patient had received nephrectomy to excise ccRCC 10 years ago. He was pathologically diagnosed as ccRCC metastasized to bone and tumor nuclear grade was Fuhrman III. A 49-year-old male patient (No.378570) underwent nephrectomy at the same hospital on April 3, 2006. This patient was histopathologically diagnosed as ccRCC. The tumor nuclear grade was Fuhrman II. The fresh surgical specimens were immediately transported to our laboratory in ice-cold PBS, and processed for cell culture within 60 min after surgery. Primary cell culture was performed as previously described [32]. The experimental protocol was approved by the Institutional Ethical Review Board of Second Minitary Medical University conformed to the ethical guidlines of the 1975 Declaration of Helsinki. An informed consent was obtained from each patient. The two patients were followed since receiving the surgery in our hospital. The patient with metastatic ccRCC died of ccRCC six months after the last surgery. The patient with primary ccRCC was radiographically diagnosed as having a tumor in pancreas in 2008 and died in 2009. However, we were not sure if the tumor was pancreas-originated cancer or metastasis from original ccRCC because this patient didn’t receive further surgery or biopsy.

Morphology of the cells and their parental tissues was observed using inverted phase-contrast microscope (Leica DMI 3000B, Germany) following routine H&E staining as previously described [32]. Cells (5 × 10⁶) were processed as previously described [9] and examined using an electron microscope (Hitachi H-7650, Tokyo, Japan).

A total number of 3 × 10⁴ cells for each cell line suspended in 12 ml DMEM (GIBCO, Grand Island, NY) with 10% FCS (GIBCO) were plated in 24-well plates. Cells in every three wells were counted once a day. The average numbers were used to generate the growth curve. Anchorage independent growth potential was evaluated by double-layered soft agarose culture system. After cultured for 15d, the cells were stained with crystal violet and colony formation was counted under a light microscope (Leica). Cell invasion assay was performed using 24-well tissue culture plates (8-μm pore size, Transwell, Corning, NY). The bottom of the culture inserts was coated with 20 μg of Matrigel (BD Biosciences, Bedford, MA). The cells (5 × 10⁴) in 0.1 ml medium with 1% FCS were placed in the upper chamber and the lower chamber was loaded with 0.2 ml medium containing 10% FCS. After cultured for 24 h in 37°C, 5% CO₂, the cells that migrated to the lower surface of filters was quantified by counting 10 independent symmetrical visual fields under the microscope to determine the invasion rate. Each assay was performed in triplicate.

Chromosomal preparation and R-banding were performed as previously described [33]. A total of 100 metaphase spreads were observed under a microscope (Leica, DM6000B) and 30 complete karyotypes were prepared to determine the chromosome number of each cell line using CW4000 software (Leica).

Cell markers were determined using the following monoclonal antibodies: anti-CD44-Phycoerythrin (PE), anti-CD74-PE, anti-CD105-FITC, anti-CD56-FITC, anti-CD24-FITC, and anti-CD99-FITC (Biolegend, UK); anti-CD133-PE (Miltenyi, Germany); anti-vimentin (Santa Cruz, CA); anti-N-Cadherin (BD Biosciences); anti-E-cadherin, anti-EpCAM (Cell Signaling, MA). Briefly, both cell lines were cultured at the same condition. Then 5 × 10⁵ cells were washed twice and resuspended in PBS with 1% FCS, and incubated either in the primary antibody (anti-CD133, CD44, CD74, CD105, CD56, CD24, and CD99) conjugated with FITC or PE for 30 min on ice, or incubated in primary antibodies to vimentin, E-cadherin, N-cadherin, or EpCAM and then washed and incubated with secondary antibody conjugated with FITC. Cells were then washed twice with PBS containing 1% FCS. Cell fluorescence was analyzed within 1 h using a flow cytometer (FACSCalibur, BD Biosciences). Cell debris and fixation artifacts were excluded by appropriate gating. The acquisition process was stopped when 10,000 events were collected in the population gate. CellQuest software (BD Biosciences) was used for data acquisition and analysis.

The cells for cell cycle analysis were grown at the same pace and fixed with 70% ethanol at 4°C for more than 2 h and then washed twice. Fixed cells were stained with 100 mg/ml propidium iodide containing 100 mg/ml RNase. Samples on ice were immediately analyzed on the flow cytometer with CellQuest software to separate G0/G1, S, and G2/M phases.

The cells were cultured under the same condition in 6-well plates. Total RNA was isolated and reverse transcribed to cDNA, and subjected for quantitative PCR as previously described [34]. The assay for each gene was repeated for 4–5 times. The primers for the amplification of IL-6, TNFα, VEGF, MMP2, HIF1α, HIF2α, GSK3β, RhoC, USP6, AHNAK, LRRK2, and SRGAP3 as well as their corresponding amplified conditions are summarized in Table 3, GAPDH, and β ₂ -microglobulin were used as internal control.

The cells were in vitro treated with 10 ng/mL TNFα (R & D Systems, MN) at different time points and then harvested. Cytosolic protein extracts were prepared as previously described [35]. Cytosolic IκBα was determined by immunoblotting with an anti-IκBα antibody (Cell Signaling). β-actin was detected by immunoblotting with antibody against β-actin (Cell Signaling). Genetools software (version 4.02, Synoptics, Cambridge, England) was used to quantify the signal strength of the bands.

Six-week-old male BALB/c nude mice were purchased from Shanghai Experimental Animal Centre, Chinese Academy of Science (Shanghai, China) and treated in accordance with the American Association for the Accreditation of Laboratory Animal Care guidelines. The cells were washed and re-suspended in 200 μl PBS, and subcutaneously injected into the flanks of the mice (2 × 10⁶/mouse). The mice were sacrificed when tumors grown up to 10 mm in diameter. The tumors were excised and mechanically minced with scissors in a sterile manner. Half of the minced tissues were subjected for primary culture, another half of the minced tumors of 1-2 mm in diameter were transplanted into renal subcapsules of anaesthetized mice to establish SOI model as previously described [36]. The mice were sacrificed before dying. The tumor tissues were transplanted for the next round of SOI. All visceral organs were fixed in 10% formalin. Metastasis was confirmed using gross and histological examination.

Student’s t test was used to determine the differences in the colony-formation rates, invasion rates, and gene expression levels of the two cell lines. All statistical tests were two-sided and performed using the Statistical Program for Social Sciences (SPSS16.0 for Windows, Chicago, IL). A p value of <0.05 was considered as statistically significant.