VEGF Triggers the Activation of Cofilin and the Arp2/3 Complex within the Growth Cone

Abstract

A crucial neuronal structure for the development and regeneration of neuronal networks is the axonal growth cone. Affected by different guidance cues, it grows in a predetermined direction to reach its final destination. One of those cues is the vascular endothelial growth factor (VEGF), which was identified as a positive effector for growth cone movement. These positive effects are mainly mediated by a reorganization of the actin network. This study shows that VEGF triggers a tight colocalization of cofilin and the Arp2/3 complex to the actin cytoskeleton within chicken dorsal root ganglia (DRG). Live cell imaging after microinjection of GFP (green fluorescent protein)-cofilin and RFP (red fluorescent protein)-LifeAct revealed that both labeled proteins rapidly redistributed within growth cones, and showed a congruent distribution pattern after VEGF supplementation. Disruption of signaling upstream of cofilin via blocking LIM-kinase (LIMK) activity resulted in growth cones displaying regressive growth behavior. Microinjection of GFP-p16b (a subunit of the Arp2/3 complex) and RFP-LifeAct revealed that both proteins redistributed into lamellipodia of the growth cone within minutes after VEGF stimulation. Disruption of the signaling to the Arp2/3 complex in the presence of VEGF by inhibition of N-WASP (neuronal Wiskott-Aldrich-Scott protein) caused retraction of growth cones. Hence, cofilin and the Arp2/3 complex appear to be downstream effector proteins of VEGF signaling to the actin cytoskeleton of DRG growth cones. Our data suggest that VEGF simultaneously affects different pathways for signaling to the actin cytoskeleton, since activation of cofilin occurs via inhibition of LIMK, whereas activation of Arp2/3 is achieved by stimulation of N-WASP.

Keywords: Arp2/3; LIMK; N-WASP; VEGF; actin; cofilin; cytoskeleton; growth cone; time-lapse imaging.

Figure 1

Distributional changes of anti-cofilin and phalloidin rhodamine in DRG growth cones subsequent to application of NGF, VEGF, or BMS-5. Controls (a) show a congruent distribution pattern of cofilin (green) and actin (red) (n = 51). Stimulation with NGF (b) (n = 53) or VEGF (c) (n = 69) lead to a highly significant increase in their mean circumference and area for both actin and cofilin, compared to control conditions (*** = p < 0.0001). The number of filopodia increased. Inhibition of cofilin signaling by application of the LIMK-inhibitor BMS-5 (d) under control conditions (n = 55) and (e) under stimulation with VEGF plus NGF (n = 57) did not lead to significant alterations of growth cone morphology as compared to controls. The table (f) shows statistical data of anti-cofilin-stained growth cones. The graphs show the mean circumference (g) and the mean area (h) of the analyzed growth cones. Scale bars = 10 µm. Error bars = SEM.

Figure 2

Live cell imaging of the axonal growth cones (n = 5). DRG cells were microinjected with a combination of RFP-LifeAct and eGFP-cofilin plasmids, and observed for approximately 2 h. Distributional behavior of cofilin and actin was observed and analyzed after VEGF application. Both proteins displayed a fast response to the VEGF application, and steadily changed their distributional pattern (arrows). The distribution of both RFP-LifeAct and eGFP-cofilin showed colocalization at every time point of observation. Scale bar: 10 µm.

Figure 3

DRG-cells microinjected with eGFP-cofilin-S3D and RFP-LifeAct were stimulated with VEGF and observed for at least one hour (n = 10). Stimulated growth cones did not show any significant reaction after stimulation. Other than some ruffling of actin near the edge of the growth cone, there was no actin activity observable. Activity of cofilin-S3D was not observed. Scale bar: 10 µm.

Figure 4

When the activity of LIMK was inhibited by the application of BMS-5, growth cones with clear alterations in their shape were observed (n = 10). The number of filopodia increased (arrows), but there was no noticeable protrusion. The base of the growth cones retracted towards the axonal shaft, which is a sign of repellent growth. Scale bar: 10 µm.

Figure 5

Distributional changes of anti-Arp2 and phalloidin rhodamine, subsequent to the application of NGF, VEGF, or wiskostatin. (a) Controls (n = 52) show a congruent distribution pattern of Arp2 (green) and actin (red). Arp2 showed a stronger fluorescent level in lamellipodia under every condition examined. Stimulation with NGF (b) (n = 50) and additional VEGF (c) (n = 50) lead to a highly significant increase in the mean circumference and area in both Arp2 and actin compared to standard conditions (*** = p < 0.001). After inhibition of Arp2/3-signaling by application of the N-WASP-inhibitor (d) wiskostatin under control conditions (n = 55) and after stimulation with VEGF plus NGF (e) (n = 57), the growth cones did not differ significantly from controls. The table (f) shows the statistical data for growth cones, stained with anti-Arp2. The graphs (g) show the mean circumference and the mean area (h) of the analyzed growth cones. Scale bars = 10 µm. Error bars = SEM.

Figure 5

Distributional changes of anti-Arp2 and phalloidin rhodamine, subsequent to the application of NGF, VEGF, or wiskostatin. (a) Controls (n = 52) show a congruent distribution pattern of Arp2 (green) and actin (red). Arp2 showed a stronger fluorescent level in lamellipodia under every condition examined. Stimulation with NGF (b) (n = 50) and additional VEGF (c) (n = 50) lead to a highly significant increase in the mean circumference and area in both Arp2 and actin compared to standard conditions (*** = p < 0.001). After inhibition of Arp2/3-signaling by application of the N-WASP-inhibitor (d) wiskostatin under control conditions (n = 55) and after stimulation with VEGF plus NGF (e) (n = 57), the growth cones did not differ significantly from controls. The table (f) shows the statistical data for growth cones, stained with anti-Arp2. The graphs (g) show the mean circumference and the mean area (h) of the analyzed growth cones. Scale bars = 10 µm. Error bars = SEM.

Figure 6

Live cell imaging of the axonal growth cones. DRG-cells were microinjected with a combination of RFP-LifeAct and eGFP-p16b plasmids and observed for approximately 2 h (n = 10). Distributional behavior of p16-Arc and actin was observed and analyzed after VEGF application. Both proteins showed a fast response to the VEGF application, and steadily changed their distributional pattern. P16-Arc predominantly aggregated in lamellipodia, but it was also detected in filopodia (arrows). The distributional changes of both proteins were congruent, with a tight colocalization. Scale bar: 10 µm.

Figure 7

VEGF increases growth cone motility and the growth rate by signaling through VEGFR-2. Cdc42 phosphorylates N-WASP, resulting in the activation of ARP2/3. Activated Arp2/3 can bind actin, initiating polymerization and branching. Inhibition of N-WASP via wiskostatin prevents the activation of Arp2/3, and thus also G-actin polymerization to F-actin. VEGF signaling via Cdc42 also activates LIMK, which inactivates cofilin, preventing the fragmentation of F-actin into G-actin. BMS-5 inhibits LIMK, shifting the equilibrium away from Ssh-1, resulting in more active cofilin. Arrows in blue indicate a positive impact on axonal growth, whereas red arrows indicate a repulsive effect.