ELMO2 association with Gαi2 regulates pancreatic cancer cell chemotaxis and metastasis

Cell culture

The pancreatic cancer cell lines, PANC-1 and AsPC-1, were purchased from American Type Culture Collection (Manassas, VA, USA). All pancreatic cancer cell lines were cultured in RPMI-1640 medium (Hyclone, Shanghai, China) supplemented with 10% fetal bovine serum (FBS; Gibco Invitrogen Corporation, Australia) and were incubated at 37 ° C in a humidified atmosphere containing 5% CO₂.

Western blotting

After indicated treatments, cells were collected and total proteins were extracted with PMSF (Beyotime Biotechnology, shanghai, China) containing RIPA lysis buffer (Beyotime Biotechnology, shanghai, China) as instructed. Protein samples were quantified by Pierce™ BCA Protein Assay Kit (Thermo Scientific) via spectrophotometer (Thermo Scientific, USA) at 562 nm. Protein loading buffer (Sigma) was applied to denature protein samples. Total protein samples (20 µg/lane) were separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to a polyvinylidene difluoride membrane (GE Healthcare Amersham™ Hybond™). The membranes were blocked with 5% skim milk for one hour at room temperature and incubated with the primary antibodies anti-ELMO2 (ab2240, Abcam), anti-Gαi2 (sc-13534, Santa Cruz Biotechnology) or anti-GAPDH (ab8245, Abcam) at 4 °C overnight, and then were washed three times with PBS containing 0.1% Tween-20 (PBST). The membranes were then incubated with the horseradish peroxidase-conjugated secondary antibodies anti-goat (ab6885, Abcam), anti-mouse (sc-516102, Santa Cruz Biotechnology) or anti-rabbit (ab6721, Abcam) for 1 h at room temperature. Finally, the bound proteins were visualized using the SuperSignal™ West Pico PLUS Chemiluminescent Substrate (Thermo Scientific) and analyzed with FluorChemHD2 system (ProteinSimple, USA).

Transient transfection

Pancreatic cancer cell lines were cultured until they reached 60–80% of confluence before transfection. To reduce the expression of ELMO2 and Gαi2, specific siRNAs were used for in vitro transfection. Cells were then incubated for 48 h, followed by protein expression analysis by western blotting. The sequences of ELMO2 siRNA were 5′-CCCAGAGUAUUAUACCCUCCGUUAU-3′, 5′-CCCACUACAGUGAGAUGCUGGCAUU-3′, and 5′-CACAUCAAUCCAGCCAUGGA- CUUUA-3′. The sequences of G αi2 siRNA were 5′-GAGGACCUGAAUAAGCGCAAAGACA-3′, 5′-ACGCCGUCACCGAUGUCAUCAA-3′, and 5′-CCGACACCAAGAACGUGCAG- UUCGU-3′. To enhance the expression of ELMO2 and Gαi2, the overexpression plasmids GV362 and GV141, respectively, were transfected into PANC-1 cells. Both plasmids were purchased from Genechem Co., Ltd (Shanghai, China). All transfections were performed using Lipofectamine 3,000 reagent (Invitrogen) in accordance with the manufacturer’s instructions. For the knockdown of ELMO2 throughout the studies, we chose siRNA_1 whose sequence was 5′-CCCAGAGUAUUAUACCCUCCGUUAU-3′, to perform the following experiments.

Exogenous Co-immunoprecipitation (Co-IP)

PANC-1 cells were plated in 10-cm culture dishes before transfection. To obtain a high level of exogenous ELMO2 and G αi2 expression, PANC-1 cells were transfected with GV362 Flag-ELMO2 and GV141 Flag-Gαi2, respectively. Cells were collected and total proteins were extracted with PMSF (Beyotime Biotechnology, shanghai, China) containing RIPA lysis buffer (Beyotime Biotechnology, shanghai, China). First, we incubated the cell lysates with monoclonal anti-flag antibodies (cat. F3165, Sigma) with continuous mixing overnight. Cell lysates were also immunoprecipitated with control rabbit IgG antibodies (CST-2729, Cell Signaling Technology). Next, PureProteomeTM protein A/G mix magnetic beads (Merck-Millipore) were added to the antibody-antigen complex and subjected to continuous mixing. Third, the precipitates were eluted from the magnetic beads by boiling in electrophoresis sample buffer, separated by SDS-PAGE, and detected with anti-Gαi2 (sc-13534, Santa Cruz Biotechnology) and anti-ELMO2 (ab2240, Abcam) antibodies.

Immunofluorescence

Briefly, PANC-1 cells were plated in 24-well plates containing round glass coverslips (one per well) and incubated for 24 h to obtain stable attachment to the glass coverslips. Before stimulation with CXCL12 (R&D Systems, Inc.), cells were serum-starved for 3 h, followed by incubation with CXCL12 (100 ng/ml) at 37 °C for 1 h. A solution containing 4% paraformaldehyde was used for cell fixation. Cell membranes were permeabilized with 0.1% Triton X-100. Donkey serum was used for blocking non-specific interactions, based on the species in which the secondary antibody was raised. After the blocking step, cells were incubated with diluted primary antibodies anti-ELMO2 (ab2240, Abcam) or anti-Gαi2 (sc-13534, Santa Cruz Biotechnology) overnight.

The cells were subsequently stained with Alexa Fluor 488-conjugated secondary antibody (cat. A32814, Invitrogen) or 546-conjugated secondary antibody (cat. A10036, Invitrogen) for 1 h at room temperature. Finally, cells on coverslips were mounted and visualized using a Leica TCS SP5 II confocal microscope (Leica Microsystems CMS GmbH).

Wound-healing assay

Pancreatic cancer cell lines were seeded in 6-well plates and divided into three groups: normal, control, and siELMO2. Cells were cultured until they reached an 80–90% density (ca. 24 h). The cell monolayer was gently and slowly scratched with a 10-µl pipette tip across the well. Then, the wells were gently washed twice with PBS to remove the detached cells. The medium was replaced by RPMI 1640 medium with 1% FBS. Finally, photographs of the monolayer were taken with a microscope at various time points (0, 3, 6, 9, 12, and 24 h).

Chemotaxis assay

The Chemotaxis assays were performed as described by the manufacturer (Neuro Probe) and Sun et al. (2005). In this assay, a 48-well microchemotaxis chamber (AP48 chemotaxis chamber, Neuro Probe) was used, and prepared 50 µl cell suspension (2 ×10⁵ cells per ml) were placed on the upper chamber and were allowed to migrate through the permeable filter membrane (PFB8, Polycarbonate membranes, 25 × 80 mm, Neuro Probe) into the lower chamber. A solution containing the chemokine (0, 10, 100, 1,000 ng/ml CXCL12) was placed below the cell-permeable membrane. After a 3-h incubation in 5% CO2 at 37 °C, the cells that had migrated through the 8-µm filter membrane were fixed, stained, and counted in three separate fields (×400) by a microscope.

Cell invasion assay

Invasion assays were performed using a transwell chamber inserted with a polycarbonic membrane (cat. 3422 Corning, USA). To reproduce appropriate in vivo environments for 2D and 3D cell movements, we added 80 µl of extracellular matrix (Corning 356234) into the upper compartment of the transwell cell culture inserts. CXCL12 (0, 10, 100, 1,000 ng/ml) was added to the lower well of the plates as an attractant. 2 × 10⁴ cells suspended in 100 µl serum-free medium were seeded into the upper chamber. The plates were incubated for 24 h at 37 °C. Then, the cells on the lower side of the insert membrane were fixed. Finally, the cells on the lower side of the filter were counted under a microscope.

Adhesion assay

Briefly, a fibronectin (Sigma-Aldrich Corporation) solution was previously prepared and stored at 4 ° C. Then, 96-well plates (Costar-3599, Corning, US) were coated with fibronectin (10 μg/ml in PBS) at room temperature for 1 h. After coating, the fibronectin solution was removed. Thermally denatured BSA was added to the plates, followed by an incubation of 1 h at 37 ° C. Then, the plates were washed twice with serum-free RPMI 1,640 medium. CXCL12 (0, 10, 100, and 1,000 ng/ml) was added to the wells of fibronectin-coated plates. After the addition of the prepared 100 µl cell suspension (5 ×10⁵ cells per ml), the plates were incubated at 37 ° C for 30 min and then washed thrice with PBS to remove non-adherent cells. Next, a 10 µl Cell Counting Kit-8 (CCK-8) solution (Dojindo, Japan) was added to each well and incubated for 1.5 h. Finally, the adherent cells were stained and quantified at OD450 using a Microplate Reader (Thermo) according to the manufacturer’s instructions. The cell adhesion ratio of the assay was calculated according to the following formula: cell adhesion (%) = OD of the adhered cells /OD of the total cells ×100%.

F-actin polymerization assay

Cellular F-actin measurement was done as described by Tsuboi (2006). Pancreatic cancer cells were stimulated with 100 ng/ml CXCL12 for the designated time points at 37 ° C. Then, the cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and stained with Alexa Fluor 568-phalloidin (Invitrogen) for 60 min at room temperature. The level of F-actin was measured by a microplate fluorescence reader. The results were expressed as relative F-actin values, as follows: ${\displaystyle \mathrm{F}-\mathrm{actin}\mathrm{t∕F}-\mathrm{actin}0=\left(\mathrm{fluorescence}\mathrm{t∕mg}ml-1\right)∕\left(\mathrm{fluorescence}t0∕\mathrm{mg}\mathrm{ml}-1\right).}$

Statistical analysis

The statistical analyses were performed with GraphPad Prism 8 (La Jolla, CA, USA). The experimental data are expressed as the mean ± SD. We designed each experiment with 3 replicates, which were done all at the same time. And each experiment was performed three times. One-way and two-way ANOVA were used to analyze the data. A p value below 0.05 was considered statistically significant.

Role of ELMO2 in the migration and chemotaxis of pancreatic cancer cells

To explore the role played by ELMO2 in the process of cell migration, we initially investigated its expression level in pancreatic cancer cell lines. The reasons why PANC-1 and AsPC-1 were chosen in this study were as follows: Firstly, information concerning the clinical course of the sites where cell lines were deprived from was important in defining the biologic and pathologic characteristics of the tumor cell lines. Both these two cell lines were derived from patients with an adenocarcinoma in the head of the pancreas and they shared similar phenotypic characteristics, such as adhesion, invasion and migration. Secondly, the cell population doubling times for PANC-1 and AsPC-1 were very close which made it more convenient for our experimental operation. Small interfering RNA (siRNA) was used to suppress ELMO2 expression (Fig. 1A). Then, a wound-healing assay was utilized to evaluate cell migration. The decreased expression of ELMO2 reduced the migration capacity of PANC-1 and AsPC-1 cells (Fig. 1C). Moreover, a chemotaxis assay indicated that CXCL12 could distinctly enhance the chemotactic ability of PANC-1 and AsPC-1 cells, while ELMO2 silencing inhibited the CXCL12-induced chemotaxis in these cell lines (Fig. 1B).

Knockdown of ELMO2 inhibited invasion, adhesion, and F-actin polymerization in PANC-1 and AsPC-1 cells

Next, a cell invasion assay was performed to monitor the movement of pancreatic cancer cells through an extracellular matrix. We found that ELMO2 knockdown suppressed CXCL12-induced invasiveness in both PANC-1 and AsPC-1 cells (Fig. 2A). Furthermore, cell adhesion assay suggested that ELMO2 downregulation by siRNA weakened the adhesion ability of PANC-1 and AsPC-1 cells (Fig. 2B). When CXCL12 combines with its receptor CXCR4, intracellular signaling events induce membrane protrusions because of actin polymerization. These events enhance the motility of cancer cells and promote chemotaxis and invasion. CXCL12 generated a transient F-actin accumulation in PANC-1 and AsPC-1 cells, in line with previous findings (Yan et al., 2012). Notably, F-actin filaments were clearly reduced in siELMO2 cells within 30 s (Fig. 2C). These results suggested that ELMO2 knockdown inhibited F-actin polymerization in pancreatic cancer cells. Thus, ELMO2 might participate in CXCL12-mediated invasion.

ELMO2 interacts with G αi2

Previous reports have shown that the association between ELMO1 and Gαi2 contributes to actin polymerization in human breast cancer cells. Thus, it was reasonable to expect that G αi2 interacts with ELMO2 in pancreatic cancer cells. To verify this possibility, a Co-IP assay was performed. First, exogenous overexpression of ELMO2 was successfully induced by transfecting PANC-1 cells with the GV362-ELMO2-Flag plasmid. Then, ELMO2-Flag, along with its endogenous interactors, was captured from PANC-1 cell lysates using specific anti-Flag antibody. Interestingly, Gαi2 was found among the ELMO2-interacting partners, as assessed by immunoblotting with G αi2 antibody (Fig. 3A). Moreover, when Gαi2-Flag was overexpressed following cell transfection with the GV141-GNAI2-Flag plasmid, endogenous ELMO2 was co-precipitated by the anti-Flag antibody along with Gαi2-Flag (Fig. 3B). Taken together, our results confirmed the physical association between ELMO2 and G αi2 in pancreatic cancer cells.

Cell stimulation with CXCL12 results in ELMO2 membrane translocation

To further investigate the interaction networks involving ELMO2 and Gαi2, immunofluorescence microscopy was used to examine the subcellular localization of the two proteins. In unstimulated cells, clear Gαi2-specific fluorescence, but not ELMO2 fluorescence, was detected at the plasma membrane. Interestingly, after CXCL12 stimulation, ELMO2 was also detected on the plasma membrane. According to the colocalization analysis performed on a pixel by pixel basis, Gαi2 and ELMO2 were clearly found to co-localize on the plasma membrane after CXCL12 stimulation of pancreatic cancer cells (Fig. 4A). Next, the impact of Gαi2 silencing on ELMO2 localization, and vice versa was explored in PANC-1 cells transfected with the appropriate siRNAs (Figs. 4B, 4C). Interestingly, the membrane translocation of ELMO2 was reduced in Gαi2-knockdown cells, even in the presence of CXCL12 stimulation, while ELMO2 knockdown had no significant impact on the level of plasma membrane Gαi2 (Figs. 4C, 4D). Thus, Gαi2 might be a key factor for chemokine-induced ELMO2 recruitment to the plasma membrane.