Cross‐talk between gastric cancer and hepatic stellate cells promotes invadopodia formation during liver metastasis

Abstract

In gastric cancer (GC), the liver is a common organ for distant metastasis, and patients with gastric cancer with liver metastasis (GCLM) generally have poor prognosis. The mechanism of GCLM is unclear. Invadopodia are special membrane protrusions formed by tumor cells that can degrade the basement membrane and ECM. Herein, we investigated the role of invadopodia in GCLM. We found that the levels of invadopodia‐associated proteins were significantly higher in liver metastasis than in the primary tumors of patients with GCLM. Furthermore, GC cells could activate hepatic stellate cells (HSCs) within the tumor microenvironment of liver metastases through the secretion of platelet‐derived growth factor subunit B (PDGFB). Activated HSCs secreted hepatocyte growth factor (HGF), which activated the MET proto‐oncogene, MET receptor of GC cells, thereby promoting invadopodia formation through the PI3K/AKT pathway and subsequently enhancing the invasion and metastasis of GC cells. Therefore, cross‐talk between GC cells and HSCs by PDGFB/platelet derived growth factor receptor beta (PDGFRβ) and the HGF/MET axis might represent potential therapeutic targets to treat GCLM.

Invadopodia could serve as a biomarker of gastric cancer with liver metastasis. Cross‐talk between gastric cancer cells and hepatic stellate cells promotes invadopodia formation during liver metastasis through HGF/MET axes, which might represent potential therapeutic targets for gastric cancer with liver metastasis.

Abbreviations

AR

anoikis resistance

CXCL11

C‐X‐C motif chemokine ligand 11

EdU

5‐ethynyl‐2′‐deoxyuridine

GC

gastric cancer

GC^AR

anoikis‐resistant gastric cancer

GCLM

gastric cancer with liver metastasis

HGF

hepatocyte growth factor

HSC

hepatic stellate cell

IHC

immunohistochemistry

MET

MET proto‐oncogene receptor tyrosine kinase receptor

N‐WASP

neural Wiskott‐Aldrich syndrome protein

PDGFB

platelet‐derived growth factor subunit B

PDGFRβ

platelet‐derived growth factor receptor beta

TME

tumor microenvironment

α‐SMA

alpha smooth muscle actin

1. INTRODUCTION

Gastric cancer is the fifth most common cancer in the world and is a very common cause of cancer‐related deaths. ¹ Many patients with GC are already in advanced stage or have distant metastases when diagnosed. ² The liver is one of the most common organs with distant metastasis of progressive GC. ³ The prognosis of patients with GCLM is extremely poor, with a 5‐year survival rate of less than 10%, regardless of surgery combined with chemotherapy or chemotherapy followed by step‐down surgery. ⁴ Nevertheless, the details surrounding the specific mechanism and development of GCLM remain unclear. Therefore, the current work aimed to assess the mechanisms of GCLM, with a view to finding effective treatments and new therapeutic strategies.

In the process of metastasis from the primary tumor to distant metastasis, tumor cells lose their nest environment and adhesive matrix support. ⁵ , ⁶ , ⁷ Therefore, tumor cells must develop AR to survive in the subsequent environment without the support of the adhesive matrix, thus laying the foundation for the formation of metastatic tumors. ⁸ Previously, we found a significant increase in pseudopodia in GC^AR cells compared to normal GC cells. ⁹ Pseudopodia are amorphous, finger‐like protrusions formed during cell migration and are morphological alterations necessary for tumor migration and invasion, which include filopodia, lamellipodia, and invadopodia. ¹⁰ , ¹¹ Invadopodia can degrade the basement membrane and ECM and are required for tumor cells to break through the vascular basement membrane and various ECM. ¹² Therefore, invadopodia might play an important role in the process of GCLM.

Tumor cell colonization and growth in distant metastatic organs are regulated by complex interactions between tumor cells and the TME. ¹³ , ¹⁴ Hepatic stellate cells are the most abundant nonhepatocyte resident cells in the liver, which are usually in a resting state; however, they are activated when stimulated by various factors. ¹⁵ Activation of HSCs is the most common biological process in hepatocellular carcinoma. ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ Activated HSCs can secrete cytokines, such as HGF and transforming growth factor‐β, which then regulate biological behaviors, including proliferation and metastasis of tumor cells. ²⁰ Among them, HGF is an important cytokine secreted by HSCs, being important for cell proliferation, migration, and cytoskeleton remodeling. ²¹ , ²² Hepatocyte growth factor can promote tumor cell migration and invasion by regulating cytoskeletal remodeling through binding to MET proto‐oncogene, MET receptors on the surface of tumor cells, and activating downstream tyrosine kinase signaling pathways. ²³ Studies have reported that the HGF/MET pathway could promote tumor cell metastasis by affecting tumor cell pseudopodia formation. ²⁴ , ²⁵ Therefore, exploring the interaction between GC cells and HSCs might help to determine the mechanisms of GCLM.

In addition, our previous results have shown that GC^AR cells secrete significantly more PDGFB compared with normal GC cells. ²⁶ Platelet‐derived growth factor subunit B is an angiogenic ligand that induces survival, proliferation, and migration of target cells by activating its membrane tyrosine kinase receptor, PDGFR. ²⁷ Studies have reported that the PDGFB/PDGFRβ pathway is important for HSC activation. ²⁸ , ²⁹ Other studies have shown that PDGFB is also involved in the interactions between GC cells and stromal cells to promote GC progression. ³⁰ Thus, PDGFB might also have an important function between GC cells and HSCs.

In this study, an important role of invadopodia was identified in the promotion of GCLM progression. Compared with GC cells in primary tumors, those in liver metastases had significantly more invadopodia, highlighting their stronger metastatic ability. In addition, we revealed that the PDGFB/PDGFRβ and HGF/MET pathways mediated the interaction between GC cells and HSCs, suggesting that these pathways could be used as therapeutic targets to inhibit or slow down the progression of GCLM.

2. MATERIALS AND METHODS

2.1. Clinical samples

Primary tumors, liver metastases, and matched paracancerous normal tissues from 54 patients with GCLM were acquired from the Department of General Surgery of Nanjing Drum Tower Hospital. Each patient was required to provide informed consent before implementing the study, which was carried out in accordance with the ethical guidelines of the Declaration of Helsinki of 1975. The ethics and animal welfare committees of Drum Tower Hospital of Nanjing University reviewed and approved all experiments using human samples.

2.2. siRNA transfection assay

Transfection reagents and MET siRNAs were obtained from Shanghai Gemma Pharmaceutical Co. Ltd. First, 5 pmol siRNA and 6 μL medium interferon reagent were diluted in 200 μL serum‐free culture medium. After mixing well, the resulting solution was allowed to stand at room temperature for 5 min. The solution was then mixed again and allowed to stand for 15 min. During this time, the medium of the cells to be transfected was removed, while the serum‐free medium was prewarmed to 37°C. The configured transfection mixture was then added to the serum‐free medium and then added to the cells for culture. The serum‐free medium was replaced with serum‐containing medium after 6 h to maintain the culture. After 48–72 h, the transfected cells were used for subsequent experiments. In this case, the siRNA targeting sequence was GCAACAGCUGAAUCUGCAATT.

2.3. Lentivirus transfection

The lentiviral vector encoding an shRNA against MET and the corresponding transfection reagents were acquired from Shanghai Heyuan Biotechnology Co. Ltd. After mixing the lentiviral vector and transfection reagent, the resulting mixture was diluted with serum‐free medium and added to the GC cells to be transfected. The medium was then changed to serum‐containing medium after 12–16 h, and selected using puromycin (10 μg/mL) (Thermo Fisher Scientific) for 2 weeks.

2.4. Detection of invadopodia formation

Gastric cancer cells (1 × 10⁵) were inoculated into 24‐well plates and incubated for 24 h at 37°C. The cells were then fixed for 15 min by 4% paraformaldehyde, followed by cell permeation using PBS supplemented with 0.3% Triton X‐100 for 15 min. The cells were subsequently blocked for 60 min by PBS containing 3% BSA, after which they were incubated for 1 h with cortactin primary Abs (ab81208; Abcam) at 37°C. This was followed by 1 h of incubation in the dark with the secondary Abs (ab150113; Abcam) at 37°C as well as a 1 h incubation in the dark with phalloidin (Solarbio) at room temperature. The cells were then incubated in the dark with DAPI staining solution (Beyotime) for 10 min at room temperature. Finally, they were transferred to slides prior to visualization and the capture of immunofluorescence images with an FV3000 confocal laser scanning microscope (Olympus).

2.5. Statistical analysis

SPSS 22.0 software (IBM Corp.) was used to analyze all the experimental data. The results of three or more independent experiments were expressed as the mean ± SD, with one‐way ANOVA or Student's t‐test used to compare differences between the experimental groups. In addition, count data were compared using the χ²‐test. Correlations between variables were also assessed using linear regression. Finally, the Kaplan–Meier method was used to generate survival curves, which were subsequently compared using the log‐rank test. For all results, differences were considered to be statistically significant at p values <0.05.

Additional materials and methods are provided in Appendix S1.

3. RESULTS

3.1. Invadopodia‐associated proteins upregulated during GCLM and related to poor prognosis

To assess the expression levels of invadopodia‐associated proteins in primary tumors and liver metastases of patients with GCLM, the levels of N‐WASP, MMP14, CDC42, and RAC1 ¹³ , ³¹ were detected in normal tissues, primary tumors, and corresponding liver metastases of 54 patients with GCLM. The results showed that the levels of invadopodia‐associated proteins in liver metastases were higher than those in primary tumors and normal tissues (Figure 1A). The formation of invadopodia depends on the activation of the actin‐related protein (ARP)2/3 complex caused by N‐WASP. ¹³ Compared with the other three invadopodia‐associated proteins, the level of N‐WASP can better reflect the number of invadopodia. Therefore, we chose N‐WASP to evaluate the relationship between invadopodia and the prognosis of patients with GCLM. Immunohistochemistry showed that the expression of N‐WASP was highest in liver metastases, followed by the primary tumors, and lowest in the normal tissues (Figure 1B,C). Furthermore, the levels of invadopodia‐associated proteins were examined, which indicated that these proteins were significantly elevated in liver metastases (Figure 1D). Additionally, to further evaluate how invadopodia are related to the prognosis of patients with GCLM, clinical data for 54 patients with GCLM were collected and analyzed. We observed that the level of N‐WASP correlated significantly with the size and number of liver metastases (Table 1). Kaplan–Meier survival analysis further indicated that higher N‐WASP expression in patients with GCLM predicted shorter overall survival (Figure 1E). The above results suggested that invadopodia are abundant in liver metastases of GC and predict poor prognosis.

FIGURE 1.

Invadopodia‐associated proteins are upregulated in liver metastases of patients with gastric cancer with liver metastasis (GCLM) and correlates with the prognosis of patients with GCLM. (A) Multiplexed fluorescent immunohistochemistry (IHC) of invadopodia‐related proteins N‐WASP, MMP14, CDC42, and RAC1 in normal tissues, primary tumors, and liver metastases from patients with GCLM (magnification, ×400). (B) IHC of N‐WASP in normal tissues, primary tumors, and corresponding liver metastases of patients with GCLM (magnification, ×100 and ×400). (C) IHC scores of N‐WASP in normal tissues, primary tumors, and corresponding liver metastases from patients with GCLM. (D) Western blot analysis of invadopodia‐associated protein (N‐WASP, MMP14, CDC42, and RAC1) levels in normal tissues (N), primary tumors (T), and corresponding liver metastases (M) of patients with GCLM. (E) Overall survival in patients with GCLM with different N‐WASP expression levels. Results are shown as mean ± SD of three independent experiments, each experiment was carried out in triplicate. ***p < 0.001; ****p < 0.0001.

TABLE 1.

Associations between N‐WASP expression and clinicopathologic characteristics in patients with gastric cancer with liver metastasis

Variable	N‐WASP expression		p value
	Strong (n = 28)	Weak (n = 26)
Age at diagnosis (years)	64.75 ± 9.06	63.08 ± 9.05	0.501
Gender
Male	25	21	0.379
Female	3	5
Lauren classification
Intestinal	23	18	0.464
Diffuse	2	2
Mixed	3	6
Size of liver metastasis (cm)
<3	18	23	0.038*
≥3	10	3
Number of liver metastases
1	11	20	0.018*
2	4	2
≥3	13	4
Liver metastasis differentiation
Low	18	11	0.127
Medium	9	15
High	1	0
Nervous invasion
Positive	25	20	0.223
Negative	3	6
Venous invasion
Positive	27	22	0.135
Negative	1	4

^* p < 0.05.

3.2. Hepatic stellate cells promote invadopodia formation in GC cells through the HGF/MET pathway

Studies have shown that the activation of HSCs is the most common biological process of hepatocellular carcinoma. ¹⁷ , ³² The activation of HSCs refers to the transformation of HSCs under static states into highly proliferative and active myofibroblasts after various stimuli. ²⁹ Activated HSCs could also affect the proliferation and invasion of tumor cells by remodeling the ECM. ³³ Therefore, we hypothesized that activated HSCs influence the biological behavior of GC cells in liver metastases. Cancer‐associated fibroblasts in the liver are widely believed to be derived from HSCs. ³³ , ³⁴ Some previous studies have also reported that α‐SMA‐positive fibroblasts in the liver are derived from activated HSCs. ³⁵ So we detected the activation of HSCs by undertaking α‐SMA staining of liver metastases in human samples. We found that liver metastasis showed strong staining for α‐SMA, indicating that HSCs in liver metastasis were indeed widely activated (Figure 2A).

FIGURE 2.

Hepatic stellate cells promote invadopodia formation in gastric cancer (GC) cells through the hepatocyte growth factor (HGF)/MET pathway. (A) Immunohistochemistry of α‐smooth muscle actin in normal liver tissue and liver metastases of patients with gastric cancer with liver metastasis (GCLM) (magnification, ×100 and ×400). (B) EdU assay to detect the proliferation of GC cells and GC cells cocultured with LX2. (C) Transwell assay to detect cell migration. (D) Colocalization of cortactin (green) and phalloidin (red) to observe the formation of invadopodia (white arrow) in Hs746t cells. Nuclei were stained with DAPI and observed under a confocal fluorescence microscope. (E) ELISA for the HGF concentration in supernatant of LX2 cells (NC) and LX2 cocultured with MKN‐45 and Hs746t cells. (F) ELISA for the HGF concentration in the serum of healthy controls, patients with GC, and patients with GCLM. (G) Western blot analysis of invadopodia‐associated protein levels in GC cells treated with control conditioned medium (Ctrl CM), conditioned medium from LX2 cells (LX2 CM), and coculture with LX2 in the presence or absence of recombinant HGF (40 ng/mL) or HGF‐Ab (neutralizing Ab against HGF; GTX10678, 0.5 μg/mL) for 48 h. (H) Western blot analysis of phospho‐ (p)‐MET, MET, and invadopodia‐associated proteins in GC cells treated with Ctrl CM and LX2 CM in the presence or absence of recombinant HGF (40 ng/mL) or foretinib (MET inhibitor; 1 μmol/mL) for 48 h. Results are shown as the mean ± SD of three independent experiments. Each experiment was performed in triplicate. **p < 0.01; ***p < 0.001; ****p < 0.0001. ns, not significant.

To further investigate whether HSCs affect the biological behavior of GC cells, the latter were cocultured with LX2 cells. The results of the EdU assay showed that GC cells cocultured with LX2 cells were more proliferative compared with the control cells (Figure 2B). Furthermore, Transwell migration assays showed that GC cells cocultured with LX2 cells had a significantly stronger migration ability than the control cells (Figure 2C). To further investigate whether HSCs affect invadopodia formation in GC cells, invadopodia were observed by colocalization of cortactin and phalloidin staining. The results showed that GC cells cocultured with LX2 cells had more invadopodia (Figure 2D). Hepatocyte growth factor, an important cytokine secreted by HSCs, ³⁶ can affect tumor cell migration and invasion by regulating cytoskeletal remodeling. ²³ The ELISA assays showed that LX2 cells cocultured with GC cells secreted more HGF than the control cells (Figure 2E). In addition, a significantly higher serum concentration of HGF was observed in patients with GCLM compared with that in patients with GC without liver metastases (Figure 2F).

To further investigate whether HSCs affect invadopodia formation of GC cells by secreting HGF, the levels of invadopodia‐associated proteins in MKN‐45 and Hs746t cells cultured with LX2 cells were detected using western blotting. The results indicated that the levels of invadopodia‐associated proteins were significantly higher in GC cells cultured with LX2 cell conditioned medium, in GC cells cocultured with LX2 cells, and in GC cells treated with HGF, compared with those in normal GC cells. By contrast, the levels of invadopodia‐associated proteins did not change in GC cells after the removal of HGF from the LX2 cell conditioned medium by HGF‐neutralizing Abs (Figure 2G). These findings suggested that HSCs promote invadopodia formation in GC cells by secreting HGF. MET is a major receptor for HGF. ²¹ To further verify whether HGF promotes invadopodia formation in GC cells through MET, the cells were treated with HGF and foretinib, an inhibitor of MET. Foretinib inhibited the LX2 cells and HGF induced upregulation of invadopodia‐associated protein levels in GC cells (Figure 2H), thereby indicating that HSCs and HGF promoted invadopodia formation in GC cells by acting on MET. In conclusion, our findings suggested that HSCs promote invadopodia formation in GC cells through the HGF/MET pathway.

3.3. Gastric cancer cells activate HSCs through the PDGFB/PDGFRβ pathway

Previous IHC results of liver metastases showed extensive activation of HSCs in the metastases, suggesting that GC cells metastasized to the liver are able to activate HSCs. This was further supported by the significantly higher α‐SMA levels in LX2 cells cocultured with GC cells compared with those in normal LX2 cells (Figure 3A,B). Then, we wondered how the HSCs were activated by GC cells. In our previous study, we found that, compared with normal GC cells, GC^AR cells could secrete more PDGFB and CXCL11. ²⁶ Hence, we speculated that GC cells might activate HSCs by secreting PDGFB or CXCL11. To test this conjecture, PDGFB or CXCL11 was inactivated using the corresponding neutralizing Abs. EdU assays showed that neutralizing Abs against PDGFB significantly reduced the promotion of LX2 cell proliferation by GC^AR cells (Figure 3C). In addition, wound healing experiments and cell migration assays indicated that immune‐depletion of PDGFB significantly decreased GC^AR cell promotion of LX2 cell migration (Figure 3D,E). Immunofluorescence and western blotting assays further suggested that neutralizing Abs against PDGFB significantly reversed the upregulated expression of α‐SMA and fibroblast activation protein alpha (FAP) (a marker of HSC activation) in LX2 induced by GC^AR cells (Figure 3F,G). However, neutralizing Abs against CXCL11 did not show these effects. These findings suggested that HSCs might be activated by PDGFB secreted by GC cells. Platelet‐derived growth factor receptor beta is the main receptor of PDGFB ²⁸ ; to verify whether GC cells could activate HSCs through the PDGFB/PDGFRβ pathway, LX2 cells, cocultured with GC cells, were treated with a specific inhibitor of PDGFRβ. Western blot analysis indicated that the PDGFRβ inhibitor reduce the promotion of conditioned medium from GC^AR cells and PDGFB on the activation of HSCs (Figure 3H), which further supported the view that GC^AR cells activated HSCs through the PDGFB/PDGFRβ pathway. In addition, we extracted primary HSCs from mouse liver (Figure S1A), and identified them by staining desmin (a marker of HSCs) with cellular immunofluorescence (Figure S1B). Then we cocultured GC cells with primary HSCs, and found that primary HSCs cocultured with GC cells were significantly activated compared with normal primary HSCs by cellular immunofluorescence and western blot (Figure S1C,D). Furthermore, primary HSCs could promote invadopodia formation in GC cells and enhance migration of GC cells (Figure S1E,F). Western blotting also suggested that, at the molecular level, primary HSCs could promote significant upregulation in the expression of invadopodia‐associated proteins in GC cells (Figure S1G). These results further indicated that GC cells can activate HSCs, and HSCs could promote invadopodia formation in GC cells.

FIGURE 3.

Gastric cancer (GC) cells activate hepatic stellate cells (HSCs) through the platelet‐derived growth factor subunit B (PDGFB)/platelet‐derived growth factor receptor beta (PDGFRβ) pathway. (A) Immunofluorescence detection of α‐smooth muscle actin (α‐SMA) expression in LX2 cells and in LX2 cocultured with GC cells. (B) Western blot analysis of α‐SMA levels in LX2 and LX2 cocultured with MKN‐45 and Hs746t cells for 48 h. (C) EdU assay. (D) Wound healing experiment. (E) Transwell assay. (F) Cellular immunofluorescence. (G) Western blot analysis of α‐SMA and fibroblast activation protein alpha (FAP) (markers of HSC activation) of control conditioned medium (CM), GC cells CM, and anoikis‐resistant GC (GC^AR) cells CM cultivating LX2 cells, and LX2 cells in the presence of neutralizing Ab against PDGFB (PDGFB‐Ab) or neutralizing antibody against C‐X‐C motif chemokine ligand 11 (CXCL11‐Ab) for 48 h. (H) Western blot analysis of proteins in LX2 cells treated with CM obtained from GC cells in the presence or absence of recombinant PDGFB or PDGFRβ inhibitor (SU‐16f), GC^AR cells in the presence or absence of PDGFB‐Ab or PDGFRβ inhibitor for 48 h. Results are shown as the mean ± SD of three independent experiments, each experiment was carried out in triplicate. *p < 0.05; **p < 0.01. ns, not significant.

3.4. MET expression is increased in liver metastases of patients with GCLM and related to poor prognosis

To further investigate the relationship between GCLM and MET expression in normal tissues, primary tumors, and corresponding liver metastases, MET was detected in GCLM patients by IHC. The highest MET expression was observed in liver metastasis, followed by primary tumors, while paracancerous tissues had the lowest expression (Figure 4A,B). This suggested that MET was associated with GCLM, which was also supported by the results of western blotting (Figure 4C). Furthermore, Kaplan–Meier survival analysis revealed significantly shorter survival for patients with GCLM with higher MET expression levels compared with those with lower MET expression (Figure 4D). Therefore, we hypothesized that MET expression is associated with GCLM and that patients with GCLM with high MET expression have a worse prognosis. In addition, we did a regression analysis on the expression levels of MET and N‐WASP in liver metastases and found that the expression levels of MET and N‐WASP were closely related. The higher the expression of MET, the higher the expression of N‐WASP in liver metastases (Figure S2). Thus, there is a correlation between the expression level of MET and invadopodia‐associated proteins in human samples.

FIGURE 4.

MET expression is increased in liver metastases of patients with gastric cancer with liver metastasis (GCLM) and is associated with poor prognosis. (A) Representative immunohistochemical (IHC) staining of MET in normal tissues, primary tumors, and liver metastases from patients with GCLM (magnification, ×100 and ×400). (B) Immunohistochemical score of MET in normal tissues, primary tumors, and liver metastases from patients with GCLM. (C) Western blot analysis of invadopodia‐associated protein levels in normal tissues (N), primary tumors (T), and corresponding liver metastases (M) of patients with GCLM. (D) Overall survival of 54 patients with GCLM with different MET expression levels. Results were shown as the mean ± SD of three independent experiments, each experiment was carried out in triplicate. **p < 0.01,***p < 0.001.

3.5. MET regulates migration, invasion, and invadopodia formation of GC cells

The detailed function of MET in GC cells was assessed by downloading mRNA expression data for MET from all GC cell lines from the Broad Institute Cancer Cell Line Encyclopedia (CCLE) before analysis using Morpheus (Figure 5A). Gastric cancer cell lines MKN‐45 and Hs746t, with high MET expression, were selected and validated by western blotting (Figure 5B). Next, siRNA was used to knockdown MET in the two cell lines, which was validated with western blotting and quantitative RT‐PCR (Figure 5C,D). The #2 siRNA was found to provide the most efficient knockdown and thus was selected for further research. Cell migration and wound healing assays revealed that, after knockdown of MET, MKN‐45 and Hs746t migration was markedly weaker compared with that of the control. However, the migratory ability returned after treatment with HGF (Figure 5E,F). In addition, the invasion ability of the siMET group was examined by a gelatin invasion assay. Black shadows appeared after green gelatin was degraded by tumor cells, and the darker and wider shadows indicated the stronger invasion ability of cells. ³⁷ The intensity and range of shadows formed by GC cells after the knockdown of MET were markedly reduced compared with those in the control. The invasion ability was again improved after adding HGF (Figure 5G). Invadopodia formation detection assays also showed that GC cells in the siMET group had fewer invadopodia compared with those in the control group, while HGF reversed this decrease (Figure 5H). Consistently, western blotting revealed that GC cells in the siMET group had lower levels of invadopodia‐associated proteins (Figure 5I). In addition, the number of pseudopodia, observed under electron microscopy in GC cells with MET knockdown, was significantly decreased compared with that in the control, while the number of pseudopodia in HGF‐treated GC cells was significantly higher (Figure 5J). These results suggested that MET regulates invadopodia formation in GC cells. In addition, we cocultured low MET‐expressing HGC‐27 GC cells with LX2 or treated them with HGF, and found that the migration ability of HGC‐27 cells cocultured with LX2 or treated with HGF was significantly stronger than that of the control group (Figure S3A,B), and the western blot analysis also suggested that the expression of invadopodia‐associated proteins in HGC‐27 cells cocultured with LX2 or treated with HGF were also significantly higher than that of the control group (Figure S3C). These results suggested that there is cross‐talk between LX2 cells and GC cells with low MET expression. We also cocultured normal GC cells and shMET GC cells with LX2 cells, and did not find significant differences in the activation of LX2 between normal and shMET GC cells (Figure S4).

FIGURE 5.

MET regulates the migration, invasion, and invadopodia formation of gastric cancer (GC) cells. (A) MET mRNA expression data of the GC cell line were downloaded from the Broad Institute Cancer Cell Line Encyclopedia and analyzed using Morpheus. (B) Western blot analysis of MET in six GC cell lines. (C) Western blot detection of the knockdown efficiency of siMET in GC cells at the protein level. (D) Quantitative RT‐PCR detection of the knockdown efficiency of siMET in GC cells at the mRNA level. (E) Transwell assay and (F) wound healing assay detection of GC cell migration. (G) Gelatin invasion assay to detect the invasion of Hs746t cells. (H) Colocalization of cortactin (green) and phalloidin (red) to observe the formation of invadopodia (white arrow) in Hs746t cells. Nuclei were stained with DAPI and observed under a confocal fluorescence microscope. (I) Western blot analysis of invadopodia‐associated protein expression in MKN‐45 and Hs746t cells with hepatocyte growth factor (HGF; 40 ng/mL) stimulation for 48 h or MET knockdown. (J) Observation of MKN‐45 and Hs746t cells with HGF stimulation or MET knockdown by scanning electron microscope. Results are shown as the mean ± SD of three independent experiments, each experiment was carried out in triplicate. *p < 0.05; **p < 0.01.***p < 0.001; ****p < 0.0001. NC, negative control.

3.6. MET regulates invadopodia formation in GC cells through the PI3K/AKT pathway

The process of invadopodia formation involves multiple factors. Gene Set Enrichment Analysis was undertaken to investigate the specific pathway through which MET regulates invadopodia formation in GC cells. The results suggested a significant positive correlation between MET expression and activation of the PI3K/AKT pathway (Figure 6A,B). To determine whether PI3K/AKT is associated with invadopodia formation, recovery experiments were also carried out using 740‐YP, an activator of PI3K, as well as LY‐294002, an inhibitor of PI3K. Cell migration assays revealed that the migration of MKN‐45 and Hs746t cells transfected with shMET was significantly weaker compared with that of normal GC cells. In addition, the migration of GC cells was enhanced after the addition of the PI3K activator 740‐YP and PI3K inhibitor LY‐294002, by inhibiting the facilitation effect of HGF on the migration of GC cells (Figure 6C). Similar changes were observed in the Transwell invasion assay (Figure 6D). Additional invadopodia formation detection assays showed that invadopodia formation in GC cells transfected with siMET was significantly reduced compared with that in the control group, while invadopodia formation in the siMET group inhibited the facilitation of invadopodia formation in GC cells by HGF (Figure 6E). These results suggested that MET regulates invadopodia formation of GC cells through the PI3K/AKT pathway. In addition, the same changes were observed for the levels of p‐PI3K, p‐AKT, p‐mTOR, and invadopodia‐associated proteins (Figure 7A,B), which further validated the proposed mechanism.

FIGURE 6.

MET regulates invadopodia formation of gastric cancer (GC) cells through the PI3K/AKT pathway. (A, B) Gene Set Enrichment Analysis evaluating MET expression and the PI3K/AKT_SIGNALING pathway in GC. (C) Transwell migration assay detection of the migration of GC cells with MET knockdown or 740‐YP (20 μg/mL), hepatocyte growth factor (HGF; 40 ng/mL) or LY294002 (10 mmol/mL) treatment for 48 h. (D) Transwell invasion assay detection of the invasion of GC cells with MET knockdown or 740‐YP, HGF, or LY294002 treatment for 48 h. (E) Colocalization of cortactin (green) and phalloidin (red) to observe the formation of invadopodia in Hs746t cells with MET knockdown or 740‐YP, HGF, or LY294002 treatment for 48 h. Nuclei were stained with DAPI and observed under a confocal fluorescence microscope. Results are shown as the mean ± SD of three independent experiments, each experiment was carried out in triplicate. *p < 0.05; **p < 0.01.***p < 0.001; ****p < 0.0001. FDR, false discovery rate; NES, normalized enrichment score.

FIGURE 7.

Knockdown of MET inhibits gastric cancer with liver metastasis (GCLM) in vivo. (A, B) Western blot analysis of phospho‐ (p)‐PI3K/PI3K, p‐AKT/AKT, p‐mTOR/mTOR, and invadopodia‐associated proteins in Hs746t and MKN‐45 cells with MET knockdown or 740‐YP (20 μg/mL), hepatocyte growth factor (HGF; 40 ng/mL), or LY294002 (10 mmol/mL) for 48 h. (C) Verification of the infection efficiency in MKN‐45 cells by fluorescence microscopy (magnification, ×200). (D) Western blotting was used to detect the infection efficiency of shMET in MKN‐45 cells. (E) Images of liver metastases in mice in the shMET and shNC (negative control) groups and the graph shows the percentage of liver metastases volume (liver metastases volume / liver volume). (F) H&E, MET, N‐WASP, MMP14, CDC42, and RAC1 staining in xenograft tissues of shNC and shMET groups (magnification, ×400). Results are shown as the mean ± SD of three independent experiments, each experiment was carried out in triplicate. **p < 0.01.

3.7. Knockdown of MET inhibits GCLM in vivo

To observe the effects of MET in vivo, MKN‐45 cells were infected with lentivirus expressing shNC or shMET, and the transfection efficiency was verified using fluorescence and western blot assays (Figure 7C,D). The lentiviral‐transfected GC cells were used to establish a trans‐splenic liver metastasis model of GC in nude mice. Nude mice were killed 6 weeks after injecting GC cells transfected with lentiviruses in the spleen and the size of liver metastases was assessed. The size of liver metastases formed by shMET GC cells was significantly smaller than that formed by shNC GC cells (Figure 7E). Levels of MET, N‐WASP, MMP14, CDC42, and RAC1 were also significantly lower in the liver metastasis tissues of nude mice in the shMET group compared with those in the shNC group (Figure 7F). The above results suggested that the knockdown of MET could inhibit GCLM in vivo. In addition, we further investigated whether PDGF inhibition could affect GCLM, and we found that the size of liver metastases was smaller in mice treated with the PDGFRβ inhibitor SU‐16f compared to the control group (Figure S5A), and immunohistochemistry results of α‐SMA also suggested that liver metastases of mice in the SU‐16f group had fewer activated HSCs compared to the control group (Figure S5B). These suggested that PDGF inhibition reduces the activation of HSCs and inhibit GCLM in vivo.

4. DISCUSSION

The liver is one of the common organs for distant metastasis of GC, and once liver metastasis has developed in patients with GC, the prognosis of patients becomes extremely poor. ² , ⁴ , ³⁸ However, the intrinsic mechanism of GCLM is unclear. Invadopodia play an important role in the invasion and metastasis of tumor cells by degrading the basement membrane and ECM. ¹³ However, no studies have investigated the role of invadopodia in GCLM. Therefore, in this study, we explored the specific role of invadopodia in GCLM. The main findings of our study are summarized in Figure 8. First, metastatic GC cells activate HSCs through the PDGFB/PDGFRβ pathway. Then, HGF, secreted by activated HSCs, acts on MET receptors on GC cells after metastasis and PI3K/AKT signaling induced by MET facilitates invadopodia formation of GC cells, which further promotes the development of GCLM.

FIGURE 8.

Schematic illustration of the potential mechanism of gastric cancer with liver metastasis (GCLM). First, metastatic gastric cancer (GC) cells activate hepatic stellate cells (HSCs) through the platelet‐derived growth factor subunit B (PDGFB)/ platelet‐derived growth factor receptor beta (PDGFRβ) pathway. Then, hepatocyte growth factor (HGF) secreted by activated HSCs acts on MET receptors of GC cells after metastasis. Finally, PI3K/AKT signaling induced by MET facilitates the invadopodia formation of GC cells, which is responsible for the development of GCLM.

The growth and development of tumor cells metastasized to the liver are closely related to the TME of liver metastases. ³⁹ , ⁴⁰ Hepatic stellate cells are important functional cells in the TME of liver metastases and studies have shown that hepatocellular carcinoma cells can promote liver cancer progression by activating HSCs. ¹⁶ , ⁴¹ In addition, colon cancer cells could activate HSCs by releasing exosomes, thereby promoting liver metastasis from colon cancer cells. ³³ These studies suggested that HSCs play an important role in the TME of liver metastases. In the present study, we found that metastatic GC cells could secret PDGFB to activate HSCs and activated HSCs, in reverse, could promote invadopodia formation in GC cells. It was not investigated in this study whether invadopodia could result in the upregulation of PDGFB, but it is worth exploring further in the future. Activated HSCs could promote the progression of GCLM by secreting HGF, as reported previously. ⁵ Our in vivo experiments also showed that PDGF inhibition reduced the activation of HSCs and inhibited progression of GCLM. However, there are various ways to activate HSCs, ²⁹ and our results illustrated that HSCs can be activated by PDGFB, which is consistent with the result of a previous study. ⁴² Although PDGFB secreted by GC^AR cells could strongly activate HSCs, many other factors might also be responsible for the activation of HSCs, such as exosomes, metabolites, and reactive oxygen species. Therefore, we believe that PDGFB might not be an optimal target for blocking the cross‐talk between GC and HSCs, and further research is needed in this regard.

Invadopodia are membrane protrusions formed by tumor cells based on a specific form of actin, which can degrade the ECM, playing a key role in tumor cell invasion and metastasis. ¹³ In the present study, we explored the role of invadopodia in GCLM. We found that the expression of invadopodia‐associated proteins in liver metastases was significantly higher than that in primary tumors, and the high expression of invadopodia‐associated proteins predicted poor prognosis. These findings suggested that invadopodia might play an important role in the process of GCLM, and treatments targeting invadopodia formation could represent a therapeutic approach for GCLM. Then, we further explored the specific mechanisms of invadopodia formation in GCLM. We found that MET could regulate invadopodia formation in GC cells through the PI3K/AKT pathway, which agreed with the results of previous studies. ⁴³ , ⁴⁴ Furthermore, we evaluated the role of MET in GCLM, and found that it was highly expressed in liver metastases of patients with GCLM and high MET expression predicted poor prognosis. Interestingly, there is a correlation between the expression level of MET and invadopodia‐associated proteins in human samples. In addition, knockdown of MET could inhibit GCLM in vivo. These results strongly supported the view that MET promotes the formation of invadopodia, thus playing an important role in GCLM.

In summary, the current study showed that GC cells metastasized to the liver can activate HSCs through the PDGFB/PDGFRβ pathway, and activated HSCs subsequently act on MET receptors of GC cells by secreting HGF, further promoting invadopodia formation of GC cells through the PI3K/AKT pathway. Eventually, invadopodia enhance the invasive and metastatic ability of GC cells and promote the progression of GCLM. Combined targeting of PDGFB/PDGFRβ and HGF/MET pathways might be developed as a new therapeutic strategy for GCLM. Therefore, the cross‐talk between GC and HSCs through the PDGFB/PDGFRβ and HGF/MET axes might represent potential therapeutic targets for GCLM treatment.

AUTHOR CONTRIBUTIONS

Chuanfu Ren: Investigation; methodology; writing – original draft. Zhi Yang: Investigation; methodology; writing – original draft. En Xu: Investigation; methodology; writing – original draft. Xing Kang: Investigation; methodology; writing – original draft. Xingzhou Wang: Formal analysis; software; validation. Qi Sun: Formal analysis; software; validation. Chao Wang: Formal analysis; software; validation. Liang Zhang: Formal analysis; software; validation. Ji Miao: Data curation; resources; supervision. Banxin Luo: Data curation; resources; supervision. Kai Chen: Data curation; resources; supervision. Song Liu: Data curation; resources; supervision. Xiaofei Shen: Data curation; resources; supervision. Xiaofeng Lu: Data curation; resources; supervision. Kai Yin: Conceptualization; funding acquisition; project administration; supervision; writing – review and editing. Xuefeng Xia: Conceptualization; funding acquisition; project administration; supervision; writing – review and editing. Wenxian Guan: Conceptualization; funding acquisition; project administration; supervision; writing – review and editing.

FUNDING INFORMATION

This work was supported by grants from the National Natural Science Foundation of China (grant number 82172645), National Natural Science Foundation of China (grant number 82203578), Key Research and Development Program of Jiangsu Province (grant number BE2022667), Key Research and Development Program of Jiangsu Province (grant number BE2022753), Key Project of Nanjing Health Commission (grant number ZKX21013), the Health Science and Technology Development Special Fund Project of Nanjing in China (grant number JQX19001), the Research Project of Jiangsu Health Committee (grant number H2019046), the Research Project of Jiangsu Health Committee (grant number M2022096), and Taikang Health Investment Youth Medical Research Start‐up Fund Project (grant number 2022002).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ETHICS STATEMENT

Approval of the research protocol by an institutional review board: This study was approved by the Ethics Committee of Nanjing Drum Tower Hospital, China.

Informed consent: Each patient was required to provide informed consent before implementing the study, which was performed in accordance with the ethical guidelines of the Declaration of Helsinki of 1975.

Registry and the registration no. of the study/trial: N/A.

Animal studies: The experimental animal ethics committee of Nanjing Hospital affiliated to Nanjing Medical University reviewed and approved the application for animal experiments (grant number DWSY‐22171298).

Supporting information

Appendix S1.

Figure S1–S5.

Ren C, Yang Z, Xu E, et al. Cross‐talk between gastric cancer and hepatic stellate cells promotes invadopodia formation during liver metastasis. Cancer Sci. 2024;115:369‐384. doi: 10.1111/cas.16023

DATA AVAILABILITY STATEMENT

The data supporting the findings of this study are available from the corresponding author upon reasonable request.