VE-cadherin cleavage by ovarian cancer microparticles induces β-catenin phosphorylation in endothelial cells

Abstract

Microparticles (MPs) are increasingly recognized as important mediators of cell-cell communication in tumour growth and metastasis by facilitating angiogenesis-related processes. While the effects of the MPs on recipient cells are usually well described in the literature, the leading process remains unclear. Here we isolated MPs from ovarian cancer cells and investigated their effect on endothelial cells. First, we demonstrated that ovarian cancer MPs trigger β-catenin activation in endothelial cells, inducing the upregulation of Wnt/β-catenin target genes and an increase of angiogenic properties. We showed that this MPs mediated activation of β-catenin in ECs was Wnt/Frizzled independent; but dependent on VE-cadherin localization disruption, αVβ3 integrin activation and MMP activity. Finally, we revealed that Rac1 and AKT were responsible for β-catenin phosphorylation and translocation to the nucleus. Overall, our results indicate that MPs released from cancer cells could play a major role in neo-angiogenesis through activation of beta catenin pathway in endothelial cells.

Keywords: angiogenesis; microparticles; ovarian cancer; tumor microenvironment; β-catenin.

Figure 1. Ovarian cancer cells MPs trigger β-catenin phosphorylation in endothelial cells

A. Flow cytometry cell sorting chart. eGFP-E4+ECs (black) and APOCC or Skov3 (red) were gated through eGFP fluorescence intensity and an APC conjugated EpCam staining. eGFP-E4+ECs were sorted as eGFP⁺/EpCam⁻. B. E4+ECs after co-culture with OCC. Phosphorylation of β-catenin at the site Ser 675 and Ser 552 were quantified by western blot. E4+ECs display a phosphorylation of β-catenin for the sites studied after co-culture with Skov3 or APOCC. C–D. OCC-MPs uptake by E4+ECs. MPs from APOCC or Skov3 were extracted from 80% confluent cells and labelled with Alexa Fluor 594 conjugated-wheat germ agglutinin (WGA). MPs uptake by eGFP-E4+ECs was quantified by confocal microscopy (screenshot from time-lapse recording, C) or flow cytometry (D). E4+ECs are able to uptake Skov3-MPs or APOCC-MPs in less than 15 minutes. Scale bar: 10 μm. E. Phosphorylation of β-catenin. E4+ECs, serum-starved for 24 h, were treated with MPs extracted from APOCC (left panel) or Skov3 (right panel). Western blots for the phosphorylation of β-catenin at sites Ser675 and Ser552 were performed. The bar graphs represent the pixel density of each band normalized using actin band and the control of the experiment. OCC-MPs phosphorylate β-catenin at Ser675 and Ser552 in E4+ECs.

Figure 2. Ovarian Cancer cells MPs induce β-catenin translocation to the nucleus in ECs and trigger angiogenesis properties

A. Localization of β-catenin. E4+ECs treated or not (control) with MPs from Skov3 or APOCC for one hour were stained for β-catenin and analyzed by High-Content analysis microscopy. In the control condition, β-catenin is localized mainly at the membrane of the cells. After incubation with OCC-MPs, β-catenin could be found only in the nucleus of the E4+ECs. Scale bar: 20 μm. B. Quantification of β-catenin translocation. The experiments presented in A were quantified with ImageXpress Micro XLS Widefield High-Content Analysis System using 5 wells with 50 images taken by well. The bar graphs represent the fluorescence intensity of the staining inside the nucleus (left panel) or the percentage of positive cells for staining in the nucleus (right panel). C. Real-time qPCR. The relative quantification of genes under control of β-catenin pathway was performed by real-time qPCR on E4+ECs after treatment with MPs from APOCC (purple) or Skov3 (green). Most of β-catenin downstream genes were upregulated compared to control. Relative transcript levels are represented as the log10 of ratios between the 2 subpopulations of their 2^−ΔΔCp real-time PCR values. D. Proliferation assay. E4+ECs were plated and counted every 2 days in presence or not of MPs from OCC (Control, top panel). The same experiment was repeated in presence of 10 μM of FH535, an inhibitor of β-catenin pathway. MPs from Skov3 (green) or APOCC (purple) did not increase the proliferation of E4+ECs when β-catenin is inhibited. E. Wound closure assay. Migration of E4+ECs with or without FH535 (10 μM) was evaluated in the presence or absence of OCC-MPs. Motility of E4+ECs was reduced by β-catenin inhibition. F. Tube formation assay. E4+ECs with or without FH535 (10 μM) were plated on matrigel layer in presence or not of OCC-MPs. MPs from Skov3 and APOCC were able to increase the number of tubes and their stability only when β-catenin was not inhibited. G. Evaluation of mesenchymal and endothelial markers in E4+ECs. E4+ECs were treated with MPs from Skov3 or APOCC every 2 days during 6 days. Western blot for mesenchymal (Vimentin and αSMA) and endothelial (CD31 and CD144) markers were performed. While endothelial markers were conserved, mesenchymal markers were expressed in E4+ECs after incubation with OCC-MPs. H. F-actin polymerization in E4+ECs after treatment with OCC-MPs. E4+ECs were grown on glass bottom slides and actin cytoskeleton was revealed by a phalloïdin-fluorescein (1 μg/mL) labeling. Fluorescence microscope series of adherent E4+ECs unstimulated or stimulated with OCC-MPs and stained with Alexafluor488. OCC-MPs induced stress fibers in E4+ECs. Scale bar: 20μm. p < 0.05 (*), p < 0.01 (**) or p < 0.001 (***).

Figure 3. Ovarian Cancer cells MPs uptake by ECs is dependent on integrin activation

A. Integrin expression in E4+ECs, APOCC and Skov3. PCR for integrins α1, 2, 3, 5, V and β1, 3, 4 were performed. Integrins β3, α1, 5 and V were the main integrins expressed. B. Integrin expression. E4+ECs, serum-starved for 24 h, were treated with OCC-MPs during 15 minutes. Western blots for the integrins β3 and αV were performed. OCC-MPs were able to increase the expression of the two studied integrin in the E4+ECs. C. αV and β3 integrins localization. E4+ECs, serum-starved for 24 h, were treated or not (control) with MPs from Skov3 or APOCC for one hour. The staining for integrin β1 (top panel) or αV (bottom panel) displayed integrins clustering upon OCC-MPs treatment (arrows). Scale bar: 20 μm. D. β-catenin phosphorylation in the presence of integrin inhibitors. E4+ECs, serum-starved for 24 h, were treated or not with a monoclonal antibody against integrins α1, α5, αV or β3 prior to the incubation with OCC-MPs during 15 minutes. Western blots for the phosphorylation of β-catenin at sites Ser675 and Ser552 were performed. The blockade of integrins αV and β3 inhibited the phosphorylation of β-catenin induced by OCC-MPs. E. Quantification of β-catenin phosphorylation under integrins blockade. The bar graphs represent the pixel density of each band normalized using actin band and the control of the experiment. F. OCC-MPs uptake by E4+ECs in presence of integrin αVβ3 monoclonal blocking antibody. MPs from APOCC or Skov3 were extracted from 80% confluent cells and labelled with Alexa Fluor 594 conjugated-wheat germ agglutinin (WGA). MPs uptake by eGFP-E4+ECs after 1 h was quantified by confocal microscopy. E4+ECs were not able to uptake Skov3-MPs or APOCC-MPs in presence of integrin inhibitors. Scale bar: 10 μm. G Lactadherin expression. E4+ECs, Skov3 and APOCC were analysed in western blots for the Lactadherin. The inhibition of integrins αV and β3 were able to inhibit the phosphorylation of β-catenin induced by OCC-MPs. H. Quantification of phosphatidyl serine in OCC-MPs. MPs extracted from Skov3 or APOCC were stained with FITC-annexin V. Quantification of positive MPs were performed by flow cytometry. A large number of OCC-MPs were positive for annexin V staining revealing the presence of phosphatidyl serine at the surface of MPs. I. OCC-MPs uptake by E4+ECs. MPs from APOCC or Skov3 were extracted from 80% confluent cells and labelled with Alexa Fluor 594 conjugated-wheat germ agglutinin (WGA). E4+ECs control (plain plot) or pre-treated with annexin V (grey plots) or with a cocktail containing annexin V and an antibody against lactadehrin (black plot) were exposed to the stained OCC-MPs for 1 h. The quantification was performed by flow cytometry. MPs uptake by E4+ECs was inhibited completely by the combination of annexin V and the antibody against lactadherin.

Figure 4. β-catenin activation in ECs is dependent on VE-cadherin localization disruption mediated by MMPs activity

A. Localization of VE-cadherin. E4+ECs treated or not (control) with MPs from Skov3 or APOCC for one hour were strained for VE-cadherin and analyzed by High-Content analysis microscopy. In the control condition, VE-cadherin is localized at the membrane of the cells. After incubation with OCC-MPs, VE-cadherin disappears from E4+ECs cell junctions Scale bar: 20 μm. B. Matrix metalloproteases activity. APOCC and Skov3 were treated with doxycycline (1μg/ml) for 24h (Dox-Skov3 and Dox-APOCC). MPs were extracted from 80% confluent cell culture of Skov3, APOCC, Dox-Skov3 and Dox-APOCC. A zymogram was performed to assess the matrix metalloproteases (MMP) activity in OCC-MPs. Both Skov3 and APOCC demonstrated a strong MMPs activity, while the treatment with doxycline significantly reduced the metalloprotease activity. C. VE-cadherin localization. MPs extracted from Skov3, APOCC, Dox-Skov3 and Dox-APOCC were labelled with Alexa Fluor 488 conjugated-wheat germ agglutinin (WGA). E4+ECs were treated for 1h with the MPs and a staining for VE-cadherin was performed. Confocal images revealed that even if the Dox-OCC-MPs were uptaken by E4+ECs, VE-cadherin was not disrupted from the membrane. Scale bar: 20μm. D. β-catenin localization. E4+ECs were treated for 1h with the MPs extracted from Skov3, APOCC or from the same cells pre-treated for 24 with 1μg/ml of doxycycline (Dox-Skov3 and Dox-APOCC). Staining for β-catenin was performed and localization of β-catenin was studied by confocal microscopy. When E4+ECs were treated with OCC-MPs, β-catenin was translocated to the nucleus. While Dox-OCC-MPs, failed to induce β-catenin translocation. Scale bar: 10 μm. E. Phosphorylation of β-catenin. E4+ECs, serum-starved for 24 h, were treated with MPs extracted from Dox-APOCC or Dox-Skov3 for 15 minutes. Western blots for the phosphorylation of β-catenin Ser675 and Ser552 were performed. Dox-OCC-MPs were not able to phosphorylate β-catenin at the two sites.

Figure 5. Rac1 and AKT are responsible for phosphorylation of β-catenin in ECs

A. Phosphorylation of β-catenin in presence of Akt inhibitor. E4+ECs, serum-starved for 24 h, were pre-treated or not with LY 294002 (10μM). Phosphorylation of β-catenin Ser675 and Ser552 and of Akt at Ser473 were analyzed by western blot. While LY 294002 was able to inhibit the phosphorylation of Akt, it was only able to reduce the phosphorylation of β-catenin Ser552 but not Ser675. B. Rac1 activation. E4+ECs, serum-starved for 24 h, were treated with OCC-MPs for 15 or 30 minutes. Western blot analysis revealed an increase of Rac1 after incubation with OCC-MPs. C. Phosphorylation of β-catenin in the presence of Rac1 inhibitor. E4+ECs, serum-starved for 24 h, were pre-treated or not with NSC23766 (10μM). Phosphorylation of β-catenin Ser675 and Ser552 were analyzed by western blot. NSC23766 was able to reduce the phosphorylation of β-catenin at Ser552 and at Ser675. D. Phosphorylation of β-catenin in the presence of PAK1 blocking antibody. E4+ECs, serum-starved for 24 h, were pre-treated or not with a blocking antibody for PAK1. Phosphorylation of β-catenin Ser675 and Ser552 were analyzed by western blot. The inhibition of PAK1 reduced the phosphorylation of β-catenin Ser552 and Ser675. E. OCC-MPs uptake by E4+ECs in presence of inhibitor. MPs from APOCC or Skov3 were extracted from 80% confluent cells and labelled with Alexa Fluor 594 conjugated-wheat germ agglutinin (WGA). MPs uptake by eGFP-E4+ECs was quantified by flow cytometry in absence (purple and green plot) or presence of LY294002 or NSC23766 (orange plot). The inhibitors of AKT and Rac1 were not able to modify OCC-MPs uptake by E4+ECs. F. β-catenin localization. E4+ECs pre-treated or not with LY294002 and NSC23766, were treated for 1h with the MPs extracted from Skov3, APOCC. Staining for β-catenin was performed and localization of β-catenin was studied by confocal microscopy. When E4+ECs were pre-treated with LY294002 and NSC23766, β-catenin was not translocated to the nucleus but accumulated in the cytoplasm. Scale bar: 10 μm.

Figure 6. Schematic representation of the interaction between OCC-MPs and ECs

1. In absence of OCC-MPs, integrins, αV and β3 subunits present on the EC membrane, remain inactivated 2. OCC-MPs interact with integrins through lactadherin, expressed by EC, and phosptidylserine (PS) present on OCC-MPs surface. 3. Internalization of OCC-MP into the EC creates 2 different activations, (i) Extracellularly, metallprotinase matrix of the OCC-MP destructs VE-Cadherin in the adherent junction and leads to the translocalization of β-catenin from the membrane to the cytoplasm. (ii) Intracellularly, OCC-MP activates Rac1, which induces polymerization of the stress fibers and the further phosphorylation of β-catenin 4. Free β-catenin accumulates in the cytoplasm 5. β-catenin is phosphorylated by PAK1/Rac1 pathway on 2 phosphorylation sites Ser552-Ser675 and by Akt on the phosphorylation site Ser552. 6. The phosphorylated β-catenin is translocated to the nucleus where it interacts with the transcription factors TCF/LEF1. 7. The interaction with TCF/LEF1 leads to the up-regulation of Wnt/β-catenin target genes and the mesenchymal markers. This activation promotes proliferation, migration and tube formation. In the off state of Wnt/β-catenin pathway, the destruction complex (APC, Axin, GSK3 and CK1) in the cytoplasm binds and phosphorylates β-catenin at 2 phosphorylation sites Ser33-37 and T41-S45. β-catenin is then ubiquinated and degraded by proteasomes.