Vascular Endothelial Growth Factor (VEGF) Promotes Assembly of the p130Cas Interactome to Drive Endothelial Chemotactic Signaling and Angiogenesis

Abstract

p130Cas is a polyvalent adapter protein essential for cardiovascular development, and with a key role in cell movement. In order to identify the pathways by which p130Cas exerts its biological functions in endothelial cells we mapped the p130Cas interactome and its dynamic changes in response to VEGF using high-resolution mass spectrometry and reconstruction of protein interaction (PPI) networks with the aid of multiple PPI databases. VEGF enriched the p130Cas interactome in proteins involved in actin cytoskeletal dynamics and cell movement, including actin-binding proteins, small GTPases and regulators or binders of GTPases. Detailed studies showed that p130Cas association of the GTPase-binding scaffold protein, IQGAP1, plays a key role in VEGF chemotactic signaling, endothelial polarization, VEGF-induced cell migration, and endothelial tube formation. These findings indicate a cardinal role for assembly of the p130Cas interactome in mediating the cell migratory response to VEGF in angiogenesis, and provide a basis for further studies of p130Cas in cell movement.

Fig. 1.

Biochemical and computational workflow for determining the interactomes of mutated and wild type p130cas, before and after VEGF stimulation. HUVECs were infected with Ad.p130CasWT or Ad.p130Cas15F for 48h. Cells were then incubated for 18 h in EBM with 0.5% FBS followed by stimulation with or without VEGF for the times indicated. Cells were then lysed and p130Cas was immunoprecipitated. A, Representative blot demonstrating immunoprecipitation of a band corresponding to the p130Cas detected in whole cell lysates (WCL), but not with the IgG control. B, Representative blot of WCLs showing that VEGF treatment increases the amount of tyrosine phosphorylated (Tyr⁴¹⁰) p130Cas in Ad.p130CasWT infected cells but not in cells overexpressing p130Cas15F. C, Overview of the experimental procedure and downstream analysis. Briefly, confluent HUVECs were infected with either Ad.p130CasWT or Ad.p130Cas15F for 48h, incubated overnight in EBM/0.5% FBS followed by stimulation with VEGF for 0, 10, 30, or 60 min. Cells were lysed and p130Cas immunoprecipitated in three independent experiments. Immunoprecipitated proteins were prepared and identified by LC-MS/MS prior to reconstruction of the p130Cas interactome. See Materials and Methods & supplemental material for further details. D, Number of proteins showing increased association with p130Cas after VEGF stimulation for the times indicated of cells overexpressing Ad.p130CasWT (black bars) or Ad.p130Cas15F (open bars). Numbers shown indicate proteins identified in all 3 experiments at each treatment time.

Fig. 2.

Comparison of the dynamic interactomes of wildtype and mutant p130cas during VEGF stimulation. Pie charts showing the percentage of p130Cas-interacting proteins at each VEGF treatment time point and between times for the WT and mutant p130Cas overexpressing HUVECs. The diameter of WT and mutant pies is proportional to the interactome size.

Fig. 3.

Comparison of enriched GO terms in the VEGF induced WT and mutant p130Cas interactomes. Enriched GO terms for the WT (left hand panel) and mutant (right hand panel) p130Cas interactomes were determined using the Gene Ontology Consortium's enrichment analysis tool. There were no enriched GO terms in the mutant p130cas interactors at 60 min, therefore, only those enriched GO terms in mutant p130cas at 10 and 30 min are shown. The color scale indicates fewer proteins associated with GO term (blue) and increasing proteins associated with GO term denoted by red shading.

Fig. 4.

IQGAP1 associates with p130Cas in a VEGF-dependent manner and also requires p130Cas tyrosine phosphorylation. A, Fold change in IQGAP1 association with VEGF treatment in HUVECS infected with Ad.p130CasWT as determined by the LFQ intensities. B, C, Cells were incubated in EBM/0.5% serum for 18 h prior to treatment with VEGF for the times indicated. Cells were lysed and immunoprecipitated with anti-p130Cas antibody and immunoblotted with antibodies to p130Cas and IQGAP1. Coimmunopreciptated IQGAP1 in C from 3 independent experiments was quantified by densitometry and normalized to total immunopreciptated p130Cas using ImageJ; *p < 0.05 versus 0 min time point, n = 3. D, HUVECs were infected with either Ad.p130CasWT or Ad.p130Cas15F for 48 h. Adenovirus infected cells, in parallel with uninfected cells, incubated in EBM/0.5% serum for 18 h followed by stimulation with or without 25 ng/ml VEGF for 30 min. Cells were lysed and immunoprecipitated with anti-p130Cas antibody and immunoblotted with antibodies to p130Cas and IQGAP1. Blots shown are representative of at least 2 independent experiments. E, Cells were transfected with 200 nm siNRP1 or 200 nm siScr for 48h, incubated in EBM/0.5% serum for a further 18 h, then stimulated with or without VEGF for 30 min, lysed and immunoprecipitated with anti-p130Cas antibody and immunoblotted with antibodies to p130Cas and IQGAP1. Quantification of coimmunoprecipitated IQGAP1 is shown below. ^#p < 0.05 versus siScr minus VEGF, *p < 0.01 versus siScr plus VEGF, n = 3.

Fig. 5.

VEGF induces localization of p130Cas and IQGAP1 into focal adhesions. Focal adhesion (FA) isolates were prepared as detailed in Materials and Methods and validated by Western blot and immunofluorescent analysis. A, Representative blots of lysates from either the cell body (CB) or FA fractions immunoblotted with the indicated antibodies. Paxillin is expressed at higher levels in focal adhesions whereas Akt and Erk expression is largely limited to the cell body fraction. β-actin was used to confirm equal loading. B, Representative blots of isolated FAs after VEGF treatment. HUVECs were plated onto fibronectin and made quiescent by overnight incubation in EBM/0.5% FBS. Cells were stimulated for 0 to 60 mins with 25 ng/ml VEGF before FA isolation. C, D, FAs were prepared from HUVECs adhering to fibronectin-coated coverslips, and stained with antibodies to vinculin (red) and either p130Cas or IQGAP1 (C, D respectively; green). Images shown are representative of at least 3 different experiments.

Fig. 6.

IQGAP1 is required for VEGF-induced cell polarization, migration and angiogenesis. A, HUVECS were transfected with two different siRNAs targeting IQGAP1 at a concentration of 200 nm, or 200 nm of siScr, for 72h and plated onto coverslips to form a confluent monolayer. The center of the coverslip was scored with a pipette tip, removing cells and leaving a gap. Cells were incubated for 90 min to allow them to polarize toward the scratch, then fixed in 4% PFA. Cells were stained with antibodies to Actin (red) and GM130 (Golgi marker; green) and nuclei were visualized using DAPI (blue). Cells were determined to be polarized when the golgi were orientated to face the scratch. Arrowheads indicate examples of nonpolarized cells. Polarized and nonpolarized cells were counted up to 40 μm from the scratch (indicated by the white dotted line) and expressed as a percentage of cells orientated toward the scratch. One-way ANOVA indicates a significant reduction in the percentage of cells polarized toward the scratch in the IQGAP1 knockdown samples (***p < 0.0005 for both IQGAP1 siRNAs). B, HUVECS were transfected with two different siRNAs targeting IQGAP1 at a concentration of 200 nm, or 200 nm of siScr for 72h and cell migration determined in response to 25 ng/ml VEGF. Values (n ≥ 3) are means ± s.e.m, expressed as the number of cells migrating per field; *p < 0.05 compared with siScr. Inset panel shows representative blots of siRNA mediated knockdown of IQGAP1. C, HUVECS were transfected with two different siRNAs targeting IQGAP1 at a concentration of 200 nm, or 200 nm of siScr for 48 h. Cells were trypsinized and 1 × 10⁴ cells were applied to a confluent monolayer of human dermal fibroblasts in EBM supplemented with 1% FBS and with or without 25 ng/ml VEGF. Cocultures were incubated for 7 days, with a change of media after 4 days, before being fixed in ethanol. Endothelial cells were stained with anti-VWF antibody followed by an Alexafluor 488 conjugated secondary antibody. Tube formation was visualized by scanning in an Incucyte Zoom and parameters of tube formation were determined using the Incucyte Zoom Angiogenesis software package. Representative images of endothelial tube formation after 7 days coculture with human dermal fibroblasts either in the absence (upper panels) or presence of VEGF (lower panels) are shown on the left hand panel. Quantification of the network area as determined by the Incucyte Zoom Angiogenesis software is shown on the right hand panel. Values (n = 3) are means ± s.e.m, *p < 0.05 versus siScr minus VEGF, ^#p < 0.01 versus siScr plus VEGF.

Fig. 7.

Role of the p130Cas-IQGAP1 association in VEGF signaling. A, HUVECS were transfected for 48 h with two different siRNAs targeting IQGAP1 (#1 and #2), p130Cas or siScr, all at a concentration of 200 nm. After incubation in EBM/0.5% FBS for 18 h, cells were stimulated with or without VEGF for 10 min, lysed and immunoblotted for the proteins indicated. Blots shown are representative of at least 3 independent experiments. B, Quantification of AKT Ser⁴⁷³ and eNOS Ser¹¹⁷⁷ phosphorylation; values (n = 3) are the AKTSer⁴⁷³:AKT and eNOSSer¹¹⁷⁷:eNOS ratios expressed as mean ± s.e.m. *p < 0.05 versus siScr minus VEGF, ^#p < 0.05; ^##p < 0.01 versus siScr plus VEGF. C, Association of IQGAP1 and Rac1 increases with VEGF stimulation. After incubation in EBM/0.5% FBS for 18 h, HUVECs were stimulated with or without VEGF for 10 min. Cells were then lysed and immunoprecipitated with anti-IQGAP1 antibody and immunoblotted for IQGAP1 and Rac1. Quantification of the immunoblots are shown below; values (n = 3) are the IQGAP1:Rac1 ratios expressed as mean ± s.e.m. *p < 0.05 versus vehicle control minus VEGF, by Student's t-Test. D, After incubation in EBM/0.5% FBS for 18 h, cells were incubated with 0.1 mm Rac1 Inhibitor II or 0.1% DMSO (vehicle control) for 4 h, followed by 10 min stimulation with or without VEGF. Cells were lysed and immunoblotted with antibodies to total and Ser⁴⁷³ phosphorylated Akt. Blots shown are representative of at least 3 independent experiments. Quantification of the immunoblots are shown below; values (n = 3) are the AKTSer⁴⁷³:AKT ratios expressed as mean ± s.e.m. *p < 0.05 versus vehicle control minus VEGF, ^#p < 0.05 versus vehicle control plus VEGF.