Regulation of the Hippo/YAP axis by CXCR7 in the tumorigenesis of gastric cancer

Abstract

Background: The Hippo pathway is crucial in organ size control and tumorigenesis. Dysregulation of the Hippo/YAP axis is commonly observed in gastric cancer, while effective therapeutic targets for the Hippo/YAP axis are lacking. Identification of reliable drug targets and the underlying mechanisms that could inhibit the activity of the Hippo/YAP axis and gastric cancer progression is urgently needed.

Methods: We used several gastric cancer cell lines and xenograft models and performed immunoblotting, qPCR, and in vivo studies to investigate the function of CXCR7 in gastric cancer progression.

Results: In our current study, we demonstrate that the membrane receptor CXCR7 (C-X-C chemokine receptor 7) is an important modulator of the Hippo/YAP axis. The activation of CXCR7 could stimulate gastric cancer cell progression through the Hippo/YAP axis in vitro and in vivo, while pharmaceutical inhibition of CXCR7 via ACT-1004-1239 could block tumorigenesis in gastric cancer. Molecular studies revealed that the activation of CXCR7 could dephosphorylate YAP and facilitate YAP nuclear accumulation and transcriptional activation in gastric cancer. CXCR7 functions via G-protein Gα_q/11 and Rho GTPase to activate YAP activity. Interestingly, ChIP assays showed that YAP could bind to the promoter region of CXCR7 and facilitate its gene transcription, which indicates that CXCR7 is both the upstream signalling and downstream target of the Hippo/YAP axis in gastric cancer.

Conclusion: In general, we identified a novel positive feedback loop between CXCR7 and the Hippo/YAP axis, and blockade of CXCR7 could be a plausible strategy for gastric cancer.

Keywords: CXCR7; Gastric cancer; Hippo signalling; LATS; Tumorigenesis.

Fig. 1

CXCR7 correlates with the gene signature of the Hippo pathway in gastric cancer. A TCGA data analysis showed that the CXCR7 expression level correlated with poor survival in gastric cancer patients. B GSEA of TCGA data showed a significant positive correlation between CXCR7 and the YAP target gene signature (https://tcga-data.nci.nih.gov). C TCGA database analysis showed a significant positive correlation between the expression levels of CXCR7 and Hippo pathway target genes in gastric cancer samples. D RNA-seq data analysis indicated the top 10 KEGG pathways enriched by CXCR7 depletion in MGC803 cells. Threshold P < 0.05. E GSEA showed that CXCR7 depletion in MGC803 cells decreased the expression of the YAP target gene signature. F Volcano plot showing that CXCR7 depletion inhibits the expression of Hippo pathway signature genes (red) in MCG803 cells. Threshold criteria: P < 0.05 and fold change > 1.5. G Heatmap of YAP target genes in RNA-seq data from MGC803 cell lines treated with siControl or siCXCR7. H Immunohistochemical analysis of gastric cancer samples showed a positive correlation between CXCR7 and YAP (P < 0.001). I The correlation analysis revealed that CXCR7 expression correlated with lymph node metastasis and advanced tumour stage (P = 0.0434 and P = 0.0109, respectively)

Fig. 2

CXCR7 is required for gastric cancer cell progression. A, B Depletion of CXCR7 inhibited the proliferation of gastric cancer cells. MGC803 and Hs746T cells were transfected with siControl or siCXCR7. Two different siRNAs were used. After 24 h, CCK-8 assays were used to determine the metabolic activity of the cells at the indicated time points after transfection. Experiments were performed in triplicate. Comparison of cell growth, *P < 0.05, **P < 0.01, ***P < 0.001.C, D Wound healing assay of MGC803 and Hs746T cells transfected with siCXCR7 or siControl. Wound closure was quantified for the indicated time points. Data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). E, F Depletion of CXCR7 inhibited the migration of MGC803 and Hs746T gastric cancer cells. MGC803 and Hs746T cells were transfected with siControl or siCXCR7. After 24 h, the migration was assessed by Transwell assays. Cell numbers were determined, and data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). G, H Depletion of CXCR7 promoted apoptosis of MGC803 and Hs746T cells. MGC803 and Hs746T cells were transfected with siCXCR7 and siControl. After 24 h, the cells were stained with PI and Annexin V, and then, FACS analysis was performed on the cells to determine the proportion of apoptotic cells. Each group was analysed in triplicate. *P < 0.05; **P < 0.01; ***P < 0.001 for comparison. I, J Depletion of CXCR7 inhibited the colony-forming ability of MGC803 and Hs746T gastric cancer cells. MGC803 and Hs746T cells were transfected with siCXCR7 and siControl. Quantification of colony formation is shown at the indicated time points. Data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). K, L Cell cycle analysis was performed to assess the effect of CXCR7 silencing on MGC803 cells and Hs746T cells. MGC803 and Hs746T cells were transfected with siCXCR7 or siControl. After 24 h, the cells were harvested, fixed in 70% ethanol and stained with propidium iodide. Cells were subjected to FACS analysis. Experiments were performed in triplicate. Comparison of cell proportions, *P < 0.05, **P < 0.01, ***P < 0.001. Representative histograms and cell cycle phase distribution plots are shown in Fig. 2 K and L, respectively. M–O Depletion of CXCR7 inhibited gastric tumour growth in vivo. MGC803 cells were stably transduced by a lentiviral vector expressing either control shRNA or CXCR7 shRNA. These MGC803 cells (2 × 10⁶) were injected into the right dorsal side of 4-week-old female BALB/c nude mice. Tumour formation in nude mice was monitored over a period of 4 weeks. Tumour volume was calculated using the following formula: tumour volume = 0.5 × length × width². Five weeks after tumour cell injection, the mice were sacrificed. Tumour growth curves, weights and photographs are shown in Panels M, N and O, respectively. P Immunohistochemical analysis showed that CXCR7 depletion decreased the expression of Ki67 in xenograft tumours

Fig. 3

Drugs targeting CXCR7 restrain gastric cancer progression. A, B ACT antagonist treatment against CXCR7 inhibited the proliferation of gastric cancer cells. MGC803 and Hs746T cells were treated with different doses of ACT. After 24 h, CCK-8 assays were used to determine the metabolic activity of the cells at the indicated time points after transfection. Experiments were performed in triplicate. Comparison of cell growth, *P < 0.05, **P < 0.01, ***P < 0.001. C, D Wound healing assay of MGC803 and Hs746T cells treated with an antagonist against CXCR7. Wound closure was quantified for the indicated time points. Data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). E, F ACT antagonist treatment against CXCR7 inhibited the migration of MGC803 and Hs746T gastric cancer cells. MGC803 and Hs746T cells were transfected with different doses of ACT. After 24 h, the migration was assessed by Transwell assays. Cell numbers were determined, and data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). G, H ACT antagonist treatment against CXCR7 promoted apoptosis of MGC803 and Hs746T cells. MGC803 and Hs746T cells were transfected with different doses of ACT. After 24 h, the cells were stained with PI and Annexin V, and then, FACS analysis was performed on the cells to determine the proportion of apoptotic cells. Each group was analysed in triplicate. *P < 0.05; **P < 0.01; ***P < 0.001 for comparison. I, J ACT antagonist treatment against CXCR7 inhibited the colony-forming ability of MGC803 and Hs746T gastric cancer cells. MGC803 and Hs746T cells were transfected with different doses of ACT. Quantification of colony formation is shown at the indicated time points. Data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). K, L Cell cycle analysis was performed to assess the effect of treatment with the CXCR7 antagonist ACT on MGC803 cells and Hs746T cells. MGC803 and Hs746T cells were transfected with different doses of ACT. After 24 h, the cells were harvested, fixed in 70% ethanol, and stained with propidium iodide. Cells were subjected to FACS analysis. Experiments were performed in triplicate. Comparison of cell proportions, *P < 0.05, **P < 0.01, ***P < 0.001. Representative histograms and cell cycle phase distribution plots are shown in Fig. 3K and L, respectively. M–O ACT antagonist treatment against CXCR7 inhibited gastric tumour growth in vivo. MGC803 cells (2 × 10⁶) were injected into the right dorsal side of 4-week-old female BALB/c nude mice. Tumour formation in nude mice treated with vehicle or ACT at the indicated concentrations was monitored over a period of 4 weeks. Tumour volume was calculated using the following formula: tumour volume = 0.5 × length × width^2. Five weeks after tumour cell injection, mice were sacrificed. Tumour growth curves, weights and photographs are shown in Panels M, N and O, respectively. P, Q In the patient-derived explant (PDEx) assay, ACT treatment inhibited the proliferation potential of gastric tumours. The gastric tumour samples were cultured ex vivo on sponges for 48 h with 10% FBS medium. The gastric tumour explants were treated with vehicle or 4 µM ACT. The samples were fixed and stained with YAP, CXCR7 and Ki67 via IHC. The Ki67-positive cells were counted for analysis

Fig. 4

Drugs activating CXCR7 promote gastric cancer progression. A, B CXCR7 activation via TC promoted the proliferation of gastric cancer cells. MGC803 and Hs746T cells were treated with different doses of TC. After 24 h, CCK-8 assays were used to determine the metabolic activity of the cells at the indicated time points after transfection. Experiments were performed in triplicate. Comparison of cell growth, *P < 0.05, **P < 0.01, ***P < 0.001. C, D Wound healing assay of MGC803 and Hs746T cells with CXCR7 activation via TC. Wound closure was quantified for the indicated time points. Data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). E, F CXCR7 activation via TC promotes the migration of MGC803 and Hs746T gastric cancer cells. MGC803 and Hs746T cells were transfected with different doses of TC. After 24 h, the migration was assessed by Transwell assays. Cell numbers were determined, and data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). G, H CXCR7 activation via TC inhibited apoptosis of MGC803 and Hs746T cells. MGC803 and Hs746T cells were transfected with different doses of TC. After 24 h, the cells were stained with PI and Annexin V, and then, FACS analysis was performed on the cells to determine the proportion of apoptotic cells. Each group was analysed in triplicate. *P < 0.05; **P < 0.01; ***P < 0.001 for comparison. I, J CXCR7 activation via TC promotes the colony-forming ability of MGC803 and Hs746T gastric cancer cells. MGC803 and Hs746T cells were transfected with different doses of TC. Quantification of colony formation is shown at the indicated time points. Data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). K, L Cell cycle analysis was performed to assess the effect of CXCR7 activation via TC on MGC803 cells and Hs746T cells. MGC803 and Hs746T cells were transfected with different doses of TC. After 24 h, the cells were harvested, fixed in 70% ethanol and stained with propidium iodide. Cells were subjected to FACS analysis. Experiments were performed in triplicate. Comparison of cell proportions, *P < 0.05, **P < 0.01, ***P < 0.001. Representative histograms and cell cycle phase distribution plots are shown in Fig. 4 K and L, respectively

Fig. 5

CXCR7 activates the YAP axis by inducing YAP dephosphorylation. A, B CXCR7 depletion in MGC803 and Hs746T cells induced YAP phosphorylation as determined by immunoblotting with S127 phospho-antibody and phospho-tag assay. MGC803 and Hs746T cells were transfected with different siCXCR7s. Immunoblotting was performed with the indicated antibodies. Phos-tagged gels were used to assess the phosphorylation status of YAP. C, D CXCR7 depletion decreased Hippo target gene expression in MGC803 and Hs746T cells. MGC803 and Hs746T cells were transfected with siControl or siCXCR7. After 48 h, total RNA was extracted for gene expression analysis. Each group was tested in triplicate. *P < 0.05, **P < 0.01, ***P < 0.001 for comparisons of target gene expression. E, F Antagonist ACT treatment against CXCR7 induced YAP phosphorylation as determined by immunoblotting with S127 phospho-antibody and phospho-tag assay. MGC803 and Hs746T cells were transfected with different doses of ACT as indicated. Immunoblotting was performed with the indicated antibodies. Phos-tagged gels were used to assess the phosphorylation status of YAP. G, H ACT antagonist treatment against CXCR7 decreased Hippo target gene expression in MGC803 and Hs746T cells. MGC803 and Hs746T cells were treated with different doses of ACT as indicated. Total RNA was extracted for gene expression analysis. Each group was tested in triplicate. *P < 0.05, **P < 0.01, ***P < 0.001 for comparisons of target gene expression. I, J CXCR7 activation via TC induced YAP dephosphorylation as determined by immunoblotting with S127 phospho-antibody and phospho-tag assay. MGC803 and Hs746T cells were transfected with different doses of TC as indicated. Immunoblotting was performed with the indicated antibodies. Phos-tagged gels were used to assess the phosphorylation status of YAP. K, L CXCR7 activation via TC increased Hippo target gene expression in MGC803 and Hs746T cells. MGC803 and Hs746T cells were treated with different doses of TC as indicated. Total RNA was extracted for gene expression analysis. Each group was tested in triplicate. *P < 0.05, **P < 0.01, ***P < 0.001 for comparisons of target gene expression. M, N CXCR7 depletion in MGC803 and Hs746T cells decreased TEAD response element activity. MGC803 and Hs746T cells were transfected with siControl or siCXCR7. After 24 h, the cells were transfected with TEAD luciferase reporter plasmids. After another 24 h, the cells were harvested for luciferase activity analysis. O, P CXCR7 blockage via ACT in MGC803 and Hs746T cells decreased TEAD response element activity. MGC803 and Hs746T cells were transfected with different doses of ACT as indicated. After 24 h, the cells were transfected with TEAD luciferase reporter plasmids. After another 24 h, the cells were harvested for luciferase activity analysis. Q, R CXCR7 activation via TC increased TEAD response element activity. MGC803 and Hs746T cells were transfected with different doses of TC as indicated. After 24 h, the cells were transfected with TEAD luciferase reporter plasmids. After another 24 h, the cells were harvested for luciferase activity analysis. S, T TC promotes YAP translocation from the cytoplasm to the nucleus. MGC803 and Hs746T cells were stimulated with different doses of TC as indicated. Endogenous YAP (green) and nuclei (blue) were stained with specific antibodies and DAPI, respectively; scale bar, 20 mm. Quantifications of YAP subcellular localization from at least 100. C, cytoplasm; N, nucleus. U, V Nucleoplasm separation experiments by immunoblotting confirmed that TC induced YAP protein translocation from the cytoplasm to the nucleus in MGC803 and Hs746T cells

Fig. 6

CXCR7 facilitates gastric cancer progression via the Hippo/YAP axis. A, B Depletion of CXCR7 inhibited the proliferation of gastric cancer cells, which was partially rescued by YAP overexpression. MGC803 and Hs746T cells were transfected with siControl or siCXCR7. After 24 h, the cells were transfected with YAP plasmid or empty vector. After 48 h, CCK-8 assays were used to determine the metabolic activity of the cells at the indicated time points after transfection. Experiments were performed in triplicate. Comparison of cell growth, *P < 0.05, **P < 0.01, ***P < 0.001. C, D Wound healing assays of MGC803 and Hs746T cells transfected with siCXCR7 or siControl, and this effect was reversed by YAP overexpression. Wound closure was quantified for the indicated time points. Data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). E, F Depletion of CXCR7 inhibited the migration of MGC803 and Hs746T gastric cancer cells, and this effect was reversed by YAP overexpression. MGC803 and Hs746T cells were transfected with siControl or siCXCR7. After 24 h, cells were transfected with YAP plasmid or empty vector. After 48 h, the migration was assessed by Transwell assays. Cell numbers were determined, and data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). G, H Depletion of CXCR7 promoted apoptosis of MGC803 and Hs746T cells, and this effect was reversed by YAP overexpression. MGC803 and Hs746T cells were transfected with siCXCR7 and siControl. After 24 h, the cells were transfected with YAP plasmid or empty vector. After 48 h, cells were stained with PI and Annexin V, and then, FACS analysis was performed on the cells to determine the proportion of apoptotic cells. Each group was analysed in triplicate. *P < 0.05; **P < 0.01; ***P < 0.001 for comparison. I, J Depletion of CXCR7 inhibited the colony-forming ability of MGC803 and Hs746T gastric cancer cells, and this effect was reversed by YAP overexpression. MGC803 and Hs746T cells were transfected with siCXCR7 and siControl. After 24 h, the cells were transfected with YAP plasmid or empty vector. Quantification of colony formation is shown at the indicated time points. Data are expressed as the mean ± SD. **P < 0.01, ***P < 0.001 (Student's t test). K, L Cell cycle analysis was performed to assess the effect of CXCR7 silencing on MGC803 and Hs746T cells, and this effect was reversed by YAP overexpression. MGC803 and Hs746T cells were transfected with siCXCR7 or siControl. After 24 h, the cells were transfected with YAP plasmid or empty vector. After 48 h, cells were harvested, fixed in 70% ethanol, and stained with propidium iodide. Cells were subjected to FACS analysis. Experiments were performed in triplicate. Comparison of cell proportions, *P < 0.05, **P < 0.01, ***P < 0.001. Representative histograms and cell cycle phase distribution plots are shown in Fig. 6 K and L, respectively. M–O Representative image of tumours derived from BALB/c nude mice injected with the indicated stably transfected MGC803 cells. These MGC803 cells (2 × 10⁶) were injected into the right dorsal side of 4-week-old female BALB/c nude mice. Tumour formation in nude mice was monitored over a period of 4 weeks. Tumour volume was calculated using the following formula: tumour volume = 0.5 × length × width². Five weeks after tumour cell injection, mice were sacrificed. Tumour growth curves, weights and photographs are shown in Panels M, N and O, respectively

Fig. 7

CXCR7 activates YAP through the Gα_q/11-ROCK-LATS axis in gastric cancer. A, B CXCR7 activation induced YAP dephosphorylation through Gaq/11. MGC803 and Hs746T cells were transiently transfected with control, Ga_q/11, or Ga_s siRNAs. MGC803 and Hs746T cells were treated with TC. YAP, phosphorylated YAP, Ga_q/11 and Ga_s were determined by immunoblotting. C CXCR7 activation induces YAP nuclear localization through Ga_q/11. MGC803 cells were transiently transfected with control, Ga_q/11 or Ga_s siRNAs. MGC803 cells were treated with TC. Endogenous YAP (green) and nuclei (blue) were stained with specific antibodies and DAPI, respectively; scale bar, 20 mm. Quantifications of YAP subcellular localization from at least 100 randomly selected cells. C, cytoplasm; N, nucleus. D Nucleoplasm separation experiments by immunoblotting confirmed that Gα_q/11 silencing could block the YAP nuclear accumulation caused by CXCR7 activation in MGC 803 cells. E CXCR7 activation via TC decreases LATS1 activity. MGC803 and Hs746T cells were stimulated with TC. LATS1 was immunoprecipitated. Phosphorylation of YAP by LATS1 was determined by a phospho-YAP antibody. F Ectopic expression of LATS1 blocks YAP dephosphorylation induced by CXCR activation. MGC803 and Hs746T cells were transiently transfected with control, LATS1 wild type (WT), or kinase dead mutant (K/R). MGC803 and Hs746T cells were treated with TC for 1 h. Phosphorylation and protein levels of YAP were determined by immunoblotting. G Rho GTPase is involved in YAP dephosphorylation induced by CXCR activation. MGC803 and Hs746T cells were transiently transfected with control, Myc-Rho-L63, or C3. MGC803 and Hs746T cells were treated with TC. Total YAP and phosphorylated YAP protein levels were determined by immunoblotting. H ROCK is required for CXCR7-induced YAP activation. Serum-starved MGC803 and Hs746T cells were pretreated with GSK429286 (1 mmol/L) or Y27632 (1 mmol/L) for 4 h, followed by treatment with TC. Total YAP and phosphorylated YAP protein levels were determined by immunoblotting

Fig. 8

YAP transcriptionally regulates CXCR7 expression, which forms a forward regulatory loop between Hippo/YAP and CXCR7. A CXCR7 genome schematic and database analysis of the binding region of YAP to the CXCR7 promoter. B, C YAP depletion decreased CXCR7 protein levels in gastric cancer cells. MGC803 and Hs746T cells were transfected with YAP siRNA. Immunoblotting was performed with the indicated antibodies. D, E YAP depletion in MGC803 and Hs746T cells inhibited CXCR7 mRNA. MGC803 and Hs746T cells were transfected with siControl or siYAP. After 48 h, total RNA was extracted for gene expression analysis. Each group was tested in triplicate. *P < 0.05, **P < 0.01, ***P < 0.001 for comparisons of target gene expression. F, G VP treatment in MGC803 and Hs746T cells inhibited CXCR7 protein expression. MGC803 and Hs746T cells were treated with vehicle and VP. Immunoblotting was performed with the indicated antibodies. H, I VP treatment in MGC803 and Hs746T cells inhibited CXCR7 mRNA. MGC803 and Hs746T cells were treated with vehicle or VP. After 48 h, total RNA was extracted for gene expression analysis. Each group was tested in triplicate. *P < 0.05, **P < 0.01, ***P < 0.001 for comparisons of target gene expression. J ChIP assays showed that YAP could bind to the promoter region of CXCR7. MGC803 cells were fixed for 30 min. Rabbit IgG was used as the negative control. The primer sequences are shown in the Methods section. The enriched DNA fragments were subjected to PCR and DNA gel electrophoresis. K, L YAP silencing decreased binding to the promoter region of the CXCR7 gene. YAP was depleted in MGC803 and Hs746T cells, and ChIP-qPCR assays showed that YAP reduced binding to classical Hippo target genes, such as CTGF and CYR61, and reduced its binding to the CXCR gene. M-P YAP silencing or inhibition decreased CXCR7 expression in the membrane. MGC803 and Hs746T cells were transiently transfected with control or YAP siRNAs or were treated with VP. Endogenous YAP (green) and nuclei (blue) were stained with specific antibodies and DAPI, respectively; scale bar, 20 mm. Quantifications of YAP subcellular localization from at least 100 randomly selected cells. C, cytoplasm; N, nucleus. Q CXCR7 formed a regulatory loop with the Hippo/YAP axis in gastric cancer. The activation of CXCR7 could facilitate the Hippo/YAP axis and gastric tumour progression via the Gaq/11-ROCK-LATS axis. In turn, YAP could bind to the promoter region of the CXCR7 gene to promote its transcription