Differential alphav integrin-mediated Ras-ERK signaling during two pathways of angiogenesis

Abstract

Antagonists of alphavbeta3 and alphavbeta5 disrupt angiogenesis in response to bFGF and VEGF, respectively. Here, we show that these alphav integrins differentially contribute to sustained Ras-extracellular signal-related kinase (Ras-ERK) signaling in blood vessels, a requirement for endothelial cell survival and angiogenesis. Inhibition of FAK or alphavbeta5 disrupted VEGF-mediated Ras and c-Raf activity on the chick chorioallantoic membrane, whereas blockade of FAK or integrin alphavbeta3 had no effect on bFGF-mediated Ras activity, but did suppress c-Raf activation. Furthermore, retroviral delivery of active Ras or c-Raf promoted ERK activity and angiogenesis, which anti-alphavbeta5 blocked upstream of Ras, whereas anti-alphavbeta3 blocked downstream of Ras, but upstream of c-Raf. The activation of c-Raf by bFGF/alphavbeta3 not only depended on FAK, but also required p21-activated kinase-dependent phosphorylation of serine 338 on c-Raf, whereas VEGF-mediated c-Raf phosphorylation/activation depended on Src, but not Pak. Thus, integrins alphavbeta3 and alphavbeta5 differentially regulate the Ras-ERK pathway, accounting for distinct vascular responses during two pathways of angiogenesis.

Figure 1.

FGF and VEGF require and stimulate Ras, Raf, and ERK activity during angiogenesis in cooperation with integrins αvβ3 and αvβ3. (A) 10-d-old chick CAMs were exposed to filter paper disks saturated with RCAS-Ras T_17N (inactive Ras), RCAS-Raf_ATPμ (inactive c-Raf), or PD98059 (MEK inhibitor) followed by stimulation with either 2 μg/ml bFGF or VEGF for 72 h. Blood vessels were enumerated by counting vessel branch points in a double-blinded manner. Each bar represents the mean ± SEM of 24 replicates. *, P < 0.05 relative to control; **, P < 0.05 relative to treatment. (B) 10-d-old chick CAMs were exposed to filter paper disks saturated with either bFGF or VEGF for 5 min, followed by excision and detergent extraction of the tissues. 1 h before excision, the embryos were i.v. injected with 30 μg function-blocking antibodies selective for either integrin αvβ3 (LM609) or αvβ5 (P1F6). Relative Ras, c-Raf, and ERK was determined as described in Materials and methods. (C) Chick CAMs were treated as above with the exception that CAM tissue was excised 20 h after initial exposure to bFGF and VEGF.

Figure 2.

FAK is required for angiogenesis and signaling during bFGF- and VEGF-induced angiogenesis. (A) 10-d-old chick CAMs were exposed to filter paper disks saturated with RCAS(A)-FRNK (inactive FAK) or i.v. injected with anti-αvβ3 or -αvβ5 followed by stimulation with either 2 μg/ml bFGF or VEGF for 72 h. Blood vessels were enumerated by counting vessel branch points in a double-blinded manner. Each bar represents the mean ± SEM of 20 replicates. *, P < 0.05 relative to control; **, P < 0.05 relative to treatment. (B) 10-d-old chick CAMs were exposed to filter paper disks saturated with RCAS-FRNK (inactive FAK) followed by stimulation with either 2 μg/ml bFGF or VEGF for 20 h. Tissues were then excised and subjected to detergent extraction. Relative Ras activity was determined using a pulldown assay with the Ras-binding domain of c-Raf followed by SDS-PAGE and immunoblotting for Ras as described in Materials and methods. (C) Chick CAMs were treated as above with the exception that c-Raf was immunoprecipitated from the tissue extracts and subjected to an in vitro kinase assay using kinase-dead MEK as a substrate as described in Materials and methods. The above blot was probed with an anti-c-Raf antibody as a loading control. (D) Chick CAMs were treated as above with the exception that total CAM lysates were electrophoresed and probed with antibodies directed against the active phosphorylated form of ERK or an anti-ERK antibody as a loading control.

Figure 3.

Ras and Raf are differentially regulated by integrins αvβ3 and αvβ5 during bFGF- and VEGF-induced angiogenesis. (A) 10-d-old chick CAMs were exposed to filter paper disks saturated with either bFGF, VEGF, RCAS-Ras G12V (active Ras), or RCAS-Raf-caax (active c-Raf). After 20 h, embryos were i.v. injected with function-blocking antibodies directed against integrins αvβ3 or αvβ5. After 72 h, blood vessels were enumerated by counting vessel branch points in a double-blinded manner. Each bar represents the mean ± SEM of 24 replicates. *, P < 0.05 relative to control; **, P < 0.05 relative to treatment. (B) Chick CAMs were treated as above with the exception that antibodies directed against integrin αvβ3 were injected after 19 h of growth-factor stimulation followed by excision and detergent extraction 1 h later. Lysates were then electrophoresed and probed with antibodies directed against the active, phosphorylated form of ERK or an anti-ERK antibody as a loading control. (C) Chick CAMs were treated as above with the exception that antibodies directed against integrin αvβ5 were injected after 19 h of growth-factor stimulation followed by excision and detergent extraction 1 h later. Lysates were then electrophoresed and probed with antibodies directed against the active, phosphorylated form of ERK or an anti-ERK antibody as a loading control.

Figure 4.

Src requirement for Raf-ERK activation during VEGF-induced (but not bFGF-induced) angiogenesis. (A) 10-d-old chick CAMs were exposed to filter paper disks saturated with either bFGF or VEGF for 20 h, followed by excision and detergent extraction of the tissues. 1 h before excision, the embryos were 1.v. injected with 30 μg function-blocking antibodies selective for either integrin αvβ3 or αvβ5 as indicated. Endogenous Src was immunoprecipitated and subjected to an in vitro kinase assay using a FAK-GST fusion protein as a substrate, electrophoresed, and anti-Src antibody was used as a loading control as described in Materials and methods. (B) 10-d-old chick CAMs were treated as described above with the exception that after excision, c-Raf was immunoprecipitated from the tissue extracts and probed with an antibody directed against phosphorylated tyrosine 340 on c-Raf. The above blot was then stripped and probed with an anti-c-Raf antibody as a loading control. (C) Chick CAMs were stimulated as described above with the exception that filter paper disks on the CAM were saturated with either the Src inhibitor PP1 or RCAS-Src251 (inactive Src), followed by blotting for phospho-Raf 340 or anti-c-Raf. (D) Chick CAMs were stimulated as described above with the exception that lysates were probed with antibodies directed against the active, phosphorylated form of ERK or an anti-ERK antibody as a loading control.

Figure 5.

Integrins αvβ3 and αvβ5 differentially influence PAK activity during angiogenesis. (A) 10-d-old chick CAMs were exposed to filter paper disks saturated with RCAS-FRNK (inactive FAK), followed by stimulation with either 2 μg/ml bFGF or VEGF for 20 h. 1 h before tissue excision, function-blocking antibodies directed against integrin αvβ3 or αvβ5 were i.v. injected. Endogenous PAK was immunoprecipitated from equivalent amounts of total protein and subjected to a kinase assay using myelin basic protein as a substrate, electrophoresed, and transferred to nitrocellulose as described in Materials and methods. The above blot was probed with an anti-PAK antibody as a loading control. (B) Chick CAMs were treated as above with the exception that total lysates were probed with an antibody directed specific to c-Raf phosphorylated at serine 338. The above blot was probed with an anti-c-Raf antibody as a loading control. (C) Chick CAMs were treated as above with the exception that after 20 h, the angiogenic tissue was resected and snap frozen. Tissue sections were probed with an antibody directed against c-Raf phosphorylated at serine 338. Bar, 50 μm.

Figure 6.

PAK activity is required for bFGF-mediated ERK activation and angiogenesis. (A) 10-d-old chick CAMs were exposed to filter paper disks saturated with RCAS-PAK_83–149 (inactive PAK), followed by stimulation with either 2 μg/ml bFGF or VEGF for 20 h. CAM tissue was excised, subjected to detergent extraction, electrophoresed, and probed with antibodies directed against the active phosphorylated form of ERK or an anti-ERK antibody as a loading control as described in Materials and methods. (B) Chick CAMs were treated as above with the exception that after 20 h, the angiogenic tissue was resected and snap frozen. Tissue sections were probed with an antibody directed against the active phosphorylated form of ERK. (C) 10-d-old chick CAMs were exposed to filter paper disks saturated with RCAS-PAK_83–149 (inactive PAK), followed by stimulation with either bFGF or VEGF for 72 h. Blood vessels were enumerated by counting vessel branch points in a double-blinded manner. Each bar represents the mean ± SEM of 36 replicates. *, P < 0.05 relative to control; **, P < 0.05 relative to treatment.

Figure 7.

bFGF/αvβ3 and VEGF/αvβ5 signaling pathways. A summary of the signaling pathways outlined in this report as it relates to EC cell survival as recently described in Alavi et al. (2003). Evidence presented here reveals that bFGF/αvβ3 and VEGF/αvβ5 differentially activate Ras-Raf-ERK signaling. This, together with our recent work (Alavi et al., 2003), allows us to propose a model whereby each of these signaling pathways accounts for protection of EC from distinct mediators of apoptosis. The αvβ3 pathway promotes an ERK-independent survival mechanism preventing stress-mediated death based on Raf coupling to the mitochondria, whereas the αvβ5 pathway prevents receptor-mediated death in an ERK-dependent manner. In addition, ERK is likely playing a general role in both pathways of angiogenesis because it regulates gene transcription, cell cycle progression, and cell migration, which are critical to the growth and differentiation of new blood vessels.