Acquired tumor cell resistance to sunitinib by increased invasion and epithelial-mesenchymal transition in LL/2 murine lung cancer

Objective This study aims to investigate biological behavior changes in a murine lung cancer cell characterized by acquired resistance to sunitinib, a potent inhibitor of multiple-targeted receptor tyrosine kinase. Methods A lung cancer cell line resistant to sunitinib (LL/2-R) was developed from its parental cell line (LL/2-P). Differences in biological characteristics and associated molecular profiles between these two cells were compared in vitro and in vivo. Results LL/2-R cells showed an approximately 5-fold higher IC50 of sunitinib than LL/2-P cells and exhibited a reduced growth inhibition following sunitinib treatment compared with LL/2-P. In LL/2-R cells and tumors, increased migration, invasion and metastasis were observed, along with upregulation of MMP-2 and MMP-9. We also analyzed the molecular profiles involved in EMT, and found that E-cadherin was downregulated in LL/2-R tumors, and vimentin was upregulated in LL/2-R cells and tumors, along with β-catenin translocating to the nuclei in LL/2-R cells. Furthermore, transcriptional factors mediated EMT, snail and twist, and the secretion of TGFβ1 also increased in LL/2-R cells and tumors. Conclusions We established a sunitinib-resistant lung cancer cell line and confirmed its drug-resistance to sunitinib in vivo. Our results implied that increased invasion and EMT may associate with the acquisition of resistant phenotype to sunitinib in cancer cells.

However, future of sunitinib is facing a major challenge: an emergent resistance to sunitinib will eventually develop. Generally, two groups of underlying mechanisms of resistance have been elucidated: first, tumor cells themselves mediate resistance to sunitinib. For instance, a study uncovered that continued exposure www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 40), pp: 68270-68279 Research Paper www.impactjournals.com/oncotarget to sunitinib could cause an increased lysosomal capacity in tumor cells, resulting in resistance to sunitinib [11]. Second, the other mechanisms mainly attribute to tumor microenvironment. Finke et al elaborated that myeloid derived suppressor cells (MDSCs) induced sunitinibresistance via providing enduring angiogenesis and immune suppression [12]. In this study, we focused on the biological behavior changes happened to sunitinibresistant tumor cells and its possible explanation for sunitinib-resistance.
In the attempt of addressing these questions, we developed a sunitinib-resistant lung cancer cell line in vitro, and then confirmed its drug-resistance to sunitnib in vivo. Moreover, we further investigated the changes in biological behavior and molecular profiles both in vitro and in vivo.

Establishment of acquired resistant LL/2 cell line to sunitinib
LL/2-P cells were treated with sunitinib for more than 6 months at gradually increasing concentrations. Finally, sunitinib-resistant LL/2 cell line, LL/2-R, was established. The values of IC 50 for the LL/2 cell sublines were determined by MTT. The IC 50 of LL/2-R cell line showed about 5-fold higher concentrations compared to that of LL/2-P cell line: 10.03μM vs.1.94μM ( Figure 1A).
In addition, we found that LL/2-R cells had a reduced cell proliferation inhibition and an increased clonogenic capacity compared to LL/2-P cells, when treated with sunitinib. In cell proliferation assay, the doubling time (DT) of LL/2-P cell line was 36.2h and 19.7h in sunitinib-and vehicle-treated cells, respectively, while it was 26.7h and 25.5h for LL/2-R cell line correspondingly. The proliferation inhibition rate was 85.2% in LL/2-P cell line, but only 13.8% in LL/2-R cell line ( Figure 1B). In clonogenic assay, the number of clonies in LL/2-R cell line was 49.0±10.0 and 116.0±4.0 in sunitinib-and vehicle-treated wells, respectively, while it was 19±5 and 228±14 in LL/2-P cell line. The percentage of clony formation in LL/2-R cell line was 41.8%, but only 8.3% in LL/2-P cell line ( Figure 1C). In conclusion, LL/2-R cells had a higher resistance to sunitinib treatment in contrast to LL/2-P cells in vitro.

Decreased growth-inhibitory effect of sunitinib on LL/2-R cells in vivo
To verify whether LL/2-R cells also had sunitinibresistance in vivo, LL/2-R cells and LL/2-P cells were injected subcutaneously in the right flank of mice, respectively. Ten days after injection, tumors were established (about 100mm 3 in size) and treatment with sunitinib (80mg/kg/day) or vehicle was performed.
Sunitinib resulted in a growth inhibition of LL/2-R tumors by 28.2% and LL/2-P tumors by 51.3% (Figure 2A). We confirmed that LL/2-R cells were also more resistant to sunitinib treatment in vivo, compared to LL/2-P cells.

Increased migratory, invasive and metastatic potential of LL/2-R cell subline
The cell migration was analyzed by wound healing assay. By 48h post-wounding, LL/2-R cells significantly improved closure of wound compared to LL/2-P cells, showing a faster migration ( Figure 3A). The invasive capacity of LL/2 cell sublines was examined by transwell assay. The numbers of LL/2-R and LL/2-P cells that invaded through the basement membrane were 376.0 ± 90.0 and 71.5 ± 16.2, respectively ( Figure 3B). To further investigate the metastatic potential of LL/2 cell sublines in vivo, LL/2-R cells and LL/2-P cells were injected subcutaneously in the right flank of mice, respectively. 40 days later, lung metastases could be detected on mice bearing LL/2-R tumors, but not on mice bearing LL/2-P tumors ( Figure 2B). In addition, MMPs has been reported to play crucial roles in invasion and metastasis of tumor cells, thus protein levels of two of most important MMPs, MMP-2 and MMP-9, in LL/2-R and LL/2-P cells and tumors were measured by western blotting assay. The expression of MMP-2 and MMP-9 were both higher in LL/2-R than LL/2-P cells ( Figure 3C), also higher in tumor tissues from LL/2-R than LL/2-P in mouse models. The data indicated that LL/2-R cells possessed increased migratory, invasive and metastatic potential.

EMT characteristics of LL/2-R cell
It has been reported that epithelial-mesenchymal transition (EMT) is associated with increased invasiveness of tumor cells. Firstly, we found that LL/2-R cells changed to flat and spindle-sharped, showing a mesenchymal morphology ( Figure 4A). Therefore, we investigate expression of vimentin and snail, and location of β-catenin in LL/2 cell sublines in vitro. The immunofluorescence staining showed that the expression level of vimentin and snail increased in LL/2-R cells relative to LL/2-P cells, while the immunofluorescence staining and western blotting assay both confirmed that β-catenin translocated to the nucleus ( Figure 4B and 4C). Additionally, we further tested whether LL/2-R cells also undergone EMT in vivo using western blotting assay. The expression of E-cadherin, an epithelial cell marker, significantly decreased, conversely, the protein level of vimentin, a mesenchymal cell marker, increased in LL/2-R tumors. Next we further analyzed the expression of EMT-related transcriptional factors like snail and twist. They were both upregulated in LL/2-R tumors ( Figure 4D). As one of the main player in inducing EMT, the TGFβ1 secretion of LL/2-R cells was also remarkably increased, compared to www.impactjournals.com/oncotarget LL/2-P cells ( Figure 4E). Our results demonstrated that LL/2-R cells undergone EMT.

DISCUSSION
Tyrosine kinase inhibitors (TKIs), such as sunitinib, are the major targeted therapies for many cancers. Unfortunately, resistance to the TKIs will eventually develop. To study the mechanism, we developed a resistant lung cancer cell and animal model to sunitinib, which have been tested in several clinical trials and preclinical studies in lung cancer [13][14][15], and subsequently validated the resistance to sunitinib both in vitro and in vivo. Our results that acquired tumor cell resistance to sunitinib causes resistance in vivo support the concept that tumor cells themselves might play a crucial role in resistance to sunitinib.
Thus, we focused our investigation on the changes in biological characteristics of tumor cells, as several studies also pointed that a significant role for tumor cells Figure 1: Acquired resistance to sunitinib in LL/2-R cells. To induce resistance, LL/2 cell line was continuously exposed for more than 6 months to gradually increasing concentrations of sunitinib. (A) Value of IC 50 to sunitinib increased in LL/2-R cells. Points, mean of three independent experiments; bars, SD. (B) LL/2-R cells had a reduced cell growth inhibition, when treated with sunitnib (C) LL/2-R cells had an increased clonogenic capacity, when treated with sunitinib. Cells were exposed to 1μm sunitinib for 10 days, and each group is made in triplicate. Columns, mean of three independent experiments; bars, SD; *, P<0.05. themselves in development of drug-resistance [16][17][18]. We observed that sunitinib-resistant LL/2 cancer cells exhibited increased migration and invasion in vitro, and enhanced metastatic potential in vivo, companied by MMP2 and MMP9 overexpression, which play crucial roles in tumor invasion and metastasis [19,20]. Similarly, there have been previous links reported between drugresistance and increased invasion [21][22][23]. Du et al [24] and Paez-Ribes et al [25] also showed that more invasive behavior was observed in cancer cells which are resistant to antiangiogenic therapy. Our data and studies implicate that invading more aggressively into normal tissue is possibly another latent adaption mode of tumor cells.
To invade into normal tissue, tumor cells have to dissolve cell-cell junction and cell-ECM adhesion. As a process during which the polarized, stationary epithelial cells break down their cell-cell and cell-ECM contacts and convert intosingle dissociated, non-polarized, motile mesenchymal cells [26], the EMT has been reported that associated with cancer cell invasion and distant metastasis [27,28]. When cancer cells undergo EMT, some molecular changes occur. For example, E-cadherin, a typical marker of epithelial cells, is lost, on the other hand, vimentin, a typical maker of mesenchymal cells, is induced [29].
Besides, E-cadherin is anchored to β-catenin at the epithelial cell membrane to maintain integrity of cell-cell adhesion junctions (AJs). Therefore, loss of E-cadherin also results in nuclear translocation of β-catenin, inducing activation transcriptional factor LEF/TCF4 which facilitates EMT by activating Wnt signaling pathway [30]. Moreover, EMT is driven by transcriptional factors that repress the expression of E-cadherin, such as snail and twist [31,32]. In this study, we observed morphology changes, downregulation of E-cadherin, on the contrary, upregulation of vimentin, snail and twist, along with nuclear translocation of β-catenin, in LL/2-R cells. Based on these results, we believe that sunitinib-resistant LL/2 cancer cells may experience EMT, which results in their increased activity to invade. Besides, we detected that the TGFβ1 level of LL/2-R cells was dramatically increased,

Figure 4: EMT of LL/2-R cell subline both in vitro and in vivo. (A) The morphology of LL/2-P and LL/-R cells. Scale bars,
100μm(×100). (B) Immunofluorescence images of LL/2 cell sublines. First row, vimentin imunolocalisation (green colour) was detectable in LL/2-R cells, but absent from LL/2-P cells. Second row, snail imunolocalisation (red colour) in nucleus was stronger in LL/2-R cells. Cell nuclei are coloured blue after staining with DAPI. Scale bars, 100μm (×100). (C) Nuclear translocation of β-cateninin LL/2-R cells. Immunofluorescence images displayed that β-catenin (green colour) gave nuclear foci pattern of imunolocalisation in LL/2-R cells, but diffuse cytoplasmic pattern in LL/2-P cells. Western blotting also showed that the β-catenin level of nucleus was upregulation, but that of cytoplasm was downregulation in LL/2-R cells. Cell nuclei are coloured blue after staining with DAPI. Scale bars, 50μm(×400). β-tubulin and Histone H3 were used as loading controls respective for cytoplasmic and nuclear protein. (D) Alterations in EMT marker in LL/2-R tumors. Western blotting revealed that E-cadherin was downregulation, and vimentin, snail and twist were all upregulation, with GAPDH as a loading control in LL/2-R tumors. Three independent experiments were conducted. (E) The TGFβ1 level was significantly increased in LL/2-R cells, which is measured by ELISA assays. and previous studies had reported that TGFβ1 could induce EMT. So the EMT process occurred in LL/2-R cells may be TGFβ1-dependent.
In the present study, we described that LL/2-R cancer cells undergone EMT, along with their resistant phenotype. To date, there is growing acceptance that sunitinib-resistance and EMT have strictly connection, as we displayed in our work as well. The EMT phenomena have been observed in some other sunitinibresistant cancers, such as prostate cancer, hepatocellular carcinoma, renal cell carcinoma [33,34]. Hammers et al. also described that EMT may be correlated with acquired resistance to sunitinib in patients with clear cell renal carcinoma [35]. Yet, the underlying mechanism between sunitinib-resistance and EMT remains unclear. Previous studies deemed that the EMT is triggered by hypoxic tumor microenvironment, which is induced by anti-angiogenesis therapy, in sunitinib-resistant cancers. However, we found that EMT characteristics also could be observed in sunitinib-resistant cancer cells in vitro in normoxia, meaning that sunitinib has a direct impact on tumor cells, eventually, EMT occurs. A previous study had reported that sunitinib could remarkly induced the expression of TGFβ [36], which is in accordance with our finding. These results suggested that sunitinib could induce the upregulation of TGFβ, which might be associated with EMT. Further research is needed to figure out the underlying mechanisms.
In conclusion, we successfully established LL/2-R cell line, which exhibits decreased sensitivity to sunitinib as opposed to its parental cell line, both in vitro and in vivo. We subsequently showed that resistant cells are possessed of increased invasive capacity and enriched EMT properties, which may be involved in acquisition of a phenotype resistant to sunitinib in LL/2 cells and this EMT maybe TGFβ1dependent. Thus, our results warrant further studies to investigate the mechanism of resistance and promising therapeutic strategies based on circumvention of EMT during sunitinib treatment.

Cell culture and reagents
The lewis lung carcinoma cell LL/2 was obtained from ATCC (American Tissue Culture Collection) and cultured in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% FBS and maintained in a humidified incubator containing 5% CO2 at 37 °C. Sunitinib malate was purchased from Selleckchem. For the development of sunitinib-resistant subline LL/2-R, the parental cell line LL/2-P was continuously exposed for more than 6 months to gradually increasing concentrations of sunitinib, which is increased by 0.2μm every 48h until the IC 50 to 2μm, and then by 0.5μm until to 20μm.

MTT assay
Cells (2×10 3 per well) were seeded in 96-well plates, and then incubated with different concentrations of sunitinib on the next day. 48hours later, the metabolically active cells were quantified using MTT (5mg/ml, Sigma-Aldrich) by measuring the Optical Density (OD) value at 570nm in ELISA reader. The proliferation inhibition due to different concentration of sunitinib was calculated by the following formula: proliferation inhibition (%)=(OD treated -OD control )/ (OD control -OD blank )×100. The value of IC 50 (the concentration required for a 50% proliferation inhibition) was determined by Graphpad prism 5.0.

Cell proliferation assay
Cells (5×10 3 per well) were seeded in 24-well plates. After 24 hours, culture medium was replaced with fresh medium or that containing 2μM sunitinib. The cells were

Clonogenic assay
Cells (1×10 3 per well) were exposed to 1μm sunitinib for 10 days to allow clony formation. Clonies were fixed, stained with crystal violet and counted manually, with a minimal clony cells number of 50 for required counting. The percentage of clony formation is calculated as the number of clonies in treated wells divided by those obtained in untreated wells.

Wound healing assay
Cells (2×10 6 per well) were seeded in 6-well plates. Scrape the cell monolayer in a straight line to create a scratch with a p200 pipet tip. To obtain the same field during the image acquisition, reference points was made on the outer bottom of the dish. Sequentially, the images were captured with a Nikon Eclipse Ti-U inverted microscope at 0h and 48h, respectively.

Transwell invasion assay
Cells were starved in serum-free DMEM for 48h. Transwell chambers (Corning Costar) consisting of 8μm pore size membrane filter inserts were coated with matrigel (BD Bioscience) overnight in incubator. 1×10 5 cells resuspended in serum-free medium were placed in the upper chamber, and medium supplemented with 10% FBS were added into the lower chamber. After 48 hours, cells on the upper surface of the membrane were removed with a cotton swab, and cells on the lower surface were stained with crystal violet, photographed and counted using an Olympus BX51 microscope.

Immunofluorescence assay
Cells on coverslips were fixed with 4% paraformaldehyde, permeabilized in 1% (V/V) Triton X-100 in PBS, blocked with 5% BSA and incubated with primary antibodies at 1:100 dilution at 4°C overnight. Secondary FITC-conjugated anti-rabbit or Texas Red-conjugated anti-mouse antibody (Santa Cruz Biotechnology) was used at 1: 50 dilution, and then DAPI (Beyotime) was included in this incubation for last 10 min. Coverslips were photographed with a Nikon Eclipse Ti-U inverted fluorescence microscope. Primary antibodies used are as follows: mouse anti-snail (Merck Millipore), rabbit anti-vimentin and rabbit anti-β-catenin (Cell Signaling Technology).

Measurement of TGFβ1 level
Cells (2×10 6 per well) were seeded in 6-well plates. 48 hours later, the TGFβ1 level in cultured medium was quantified using the Mouse TGFβ1 ELISA kit (Neobioscience) in accordance with manufacturer's protocol.

Tumor growth in vivo
All animal experiments were performed according to procedures approved by the Institutional Animal Use and Care Committee of Sichuan University, China. C57BL/6 mice were purchased from HuaFukang Biological Technology Co. Ltd. (Beijing, China) and housed under SPF (specific pathogen-free) conditions. Tumor cells (5×10 5 cells per mouse for LL/2-P or LL/2-R) were injected subcutaneously in the right flank of mice. When tumor size reached about 100mm 3 , the mice were randomly divided 2 groups of 10 mice each and treated daily by gavage with sunitinib at a dose of 80mg/kg or vehicle (control). The tumor size was assessed every 3 days with caliper measurement, and expressed in mm 3 using the formula: length×width 2 ×0.52. After 15 days of treatment, mice were sacrificed and tumors were snapfrozen in liquid nitrogen. The growth inhibition was measured by the following formula: growth inhibition (%)=1-V/V 0 ×100 (V 0 =volume of tumors in control group; V=volume of tumors in sunitinib-treated group).

Tumor metastasis
Three mice were injected subcutaneously in the right flank with LL/2-P or LL/2-R cells, respectively. After 40 days, mice were sacrificed and lungs were examined, embedded in paraffin, stained for hematoxylin and eosin and photographed using an Olympus BX51 microscope.

Statistical analysis
Data are presented as means ± SD. Statistical differences were examined by Student's t test. A P value less than 0.05 is considered to be statistically significant.