The antagonistic modulation of Arp2/3 activity by N-WASP, WAVE2 and PICK1 defines dynamic changes in astrocyte morphology

Abstract

Astrocytes exhibit a complex, branched morphology, allowing them to functionally interact with numerous blood vessels, neighboring glial processes and neuronal elements, including synapses. They also respond to central nervous system (CNS) injury by a process known as astrogliosis, which involves morphological changes, including cell body hypertrophy and thickening of major processes. Following severe injury, astrocytes exhibit drastically reduced morphological complexity and collectively form a glial scar. The mechanistic details behind these morphological changes are unknown. Here, we investigate the regulation of the actin-nucleating Arp2/3 complex in controlling dynamic changes in astrocyte morphology. In contrast to other cell types, Arp2/3 inhibition drives the rapid expansion of astrocyte cell bodies and major processes. This intervention results in a reduced morphological complexity of astrocytes in both dissociated culture and in brain slices. We show that this expansion requires functional myosin II downstream of ROCK and RhoA. Knockdown of the Arp2/3 subunit Arp3 or the Arp2/3 activator N-WASP by siRNA also results in cell body expansion and reduced morphological complexity, whereas depleting WAVE2 specifically reduces the branching complexity of astrocyte processes. By contrast, knockdown of the Arp2/3 inhibitor PICK1 increases astrocyte branching complexity. Furthermore, astrocyte expansion induced by ischemic conditions is delayed by PICK1 knockdown or N-WASP overexpression. Our findings identify a new morphological outcome for Arp2/3 activation in restricting rather than promoting outwards movement of the plasma membrane in astrocytes. The Arp2/3 regulators PICK1, and N-WASP and WAVE2 function antagonistically to control the complexity of astrocyte branched morphology, and this mechanism underlies the morphological changes seen in astrocytes during their response to pathological insult.

Keywords: Actin dynamics; Arp2/3; Astrocyte; Brain injury; CNS; Central nervous system; Morphology.

Fig. 1.

Inactivation of the Arp2/3 complex in astrocytes results in an expanded destellated morphology. (A) Phase-contrast live-cell imaging of serum-starved cultured astrocytes in the presence of forskolin with or without the Arp2/3 inhibitor CK-548. Scale bars: 20 µm. (B) Western blot analysis of astrocytes transfected either with a control (siControl) or Arp3 (siArp3)-specific siRNA. Arp3 expression was determined using an Arp3-specific antibody, and GAPDH immunoreactivity was used as loading control. (C) Confocal images of astrocytes transfected with Arp3 siRNA or control siRNA followed by serum starvation and forskolin treatment. Arp3 expression was visualized by immunostaining for Arp3 (red) and F-actin by phalloidin staining (green). Arp3-depleted cells (arrowheads) do not have a stellated morphology, compared with Arp3-positive cells (arrows), which do. Scale bars: 10 µm. (D) Schematic example to illustrate representative differences in cell outlines and cell areas of polygonal (left) and stellate (right) astrocytes. (E) Frequency analysis of astrocyte complexity in Arp3-knockdown and control cells after forskolin treatment. Cells were analyzed regarding the ratio of cell outline and cell area. Cells with a cell outline to cell area ≤0.2 are defined as polygonal. High values for the cell-outline:cell-area ratios correspond to high levels of astrocyte complexity (n = 300 cells per condition from three independent experiments). (F) Quantification of the proportion of polygonal cells as shown in C and E. ***P<0.0005 (unpaired Student's t-test). (G) Phase contrast live-cell imaging of serum-starved astrocytes previously treated with forskolin and kept in serum-free medium in the absence or presence of CK-548. Scale bars: 10 µm. (H) Frequency analysis of astrocyte complexity of CK-548- and DMSO-treated cells (300 cells per condition from three independent experiments). (I) Quantification of the proportion of polygonal cells shown in G and I. **P<0.005 (Student's unpaired t-test). (J) Confocal images of stellated astrocytes before (upper panels), after 5 min of CK-548 incubation. Arp2/3 localization and actin filaments were visualized by immunostaining for Arp3 (red) and phalloidin staining for F-actin (green). Note that Arp3 is enriched along processes and the plasma membrane of stellated astrocytes (arrows). Scale bars: 10 µm.

Fig. 2.

Acute inhibition of the Arp2/3 complex in astrocytes in brain slices. (A) Demonstration of a modified tissue clearance procedure, allowing deep-tissue antibody staining and imaging by confocal microscopy. Untreated cortical slice (top) in comparison to cleared tissue (bottom). (B) Z-projection of a control astrocyte 40 µm within the cortical slice, previously stained for DNA (Hoechst 33258), GFAP and S100β, as acquired by confocal microscopy after tissue clearance. Scale bars: 10 µm. (C) Filament tracing of GFAP- and S100β-positive processes in an individual astrocyte. Main processes were defined by GFAP immunoreactivity (top). Fine structures were determined by S100β immunoreactivity (center). Overlay of 3D models and confocal z-projections (bottom). Left panels: control astrocyte. Right panels: astrocyte from a slice treated with CK-548. Scale bars: 10 µm. (D) Quantification of longest processes in control and CK-548-treated astrocytes, based on GFAP immunoreactivity. n = 20 per condition, P>0.05 (unpaired Student's t-test). (E) Sholl analyzes on combined GFAP- and S100β-positive processes in control (blue) and CK-548-treated (red) astrocytes. n = 20 per condition, **P<0.005, ***P<0.0005 (unpaired Student's t-test and Sidak–Bonferroni method). (F) Quantification of soma volumes from control and CK-548-treated astrocytes. n = 20 per condition, ***P<0.0005 (unpaired Student's t-test). (G) Frequency of individual S100β-positive process volumes from control and CK-548-treated astrocytes. (90,000 S100β-positive processes from 20 cells per condition). Left graph: small processes up to 1.25 µm³. Right graph: larger processes greater than 1.5 µm³.

Fig. 3.

Inhibition of formins and Myosin II counteracts Arp2/3 inhibition and is associated with changes in small GTPase activation. Images (left) and frequency analysis (right) for astrocytes after serum starvation, forskolin and subsequent incubation with DMSO (A), CK-548 (B), CK-548 plus the formin inhibitor SMIFH2 (C), CK-548 plus blebbistatin (D) and CK-548 plus the ROCK inhibitor Y-27632 (E). Scale bars: 10 µm. Cell morphology is visualized by Alexa-546–phalloidin staining. Graphs show quantification of astrocyte morphology following the drug treatments (n = 300 cells per condition from three independent experiments). (F) Quantification of the proportion of polygonal astrocytes after the treatments shown in A–E. *P<0.05, **P<0.005 (ANOVA followed by Bonferroni's correction). (G) Determination of RhoA activation in astrocytes after forskolin treatment followed by CK-548 treatment for 1 h. Upper blots indicate total RhoA levels, lower blots show the GTP-bound fraction. The graph shows a quantification of the relative proportion of active RhoA, as shown in the western blots. n = 5, *P<0.05 (unpaired Student's t-test). (H) Determination of Rac1 activation in astrocytes after forskolin treatment, followed by CK-548 treatment for 1 h. Upper blots indicate total Rac1 levels, lower blots show the GTP-bound fraction. The graph shows a quantification of the relative proportion of active Rac1, as shown in the western blots. n = 4. *P>0.05 (unpaired Student's t-test).

Fig. 4.

Identification of Arp2/3 regulators in astrocytes. (A) Astrocytes were transfected with control siRNA (siControl) or WAVE2-specific siRNA (siWAVE2). WAVE2 and GAPDH expression were analyzed by western blotting. (B) Confocal images of siControl- and siWAVE2-transfected astrocytes, subjected to by serum starvation, forskolin treatment and phalloidin staining. Scale bars: 10 µm. (C) Frequency analysis of the astrocyte complexity in WAVE2-knockdown and control cells after forskolin treatment. For frequency analysis, n = 300 cells per condition from three independent experiments. (D) Quantification of the proportion of polygonal astrocytes of siControl- and siWAVE2-transfected cells, *P<0.05 (unpaired t-test). (E) Sholl analysis on processes of astrocytes transfected with either control siRNA (siControl, blue) or siWAVE2-specific siRNA (siWAVE2, red). n = 108 (siControl), n = 140 (siWAVE2), *P<0.05 (unpaired t-test and Sidak-Bonferroni method). (F) Astrocytes were transfected with control siRNA (siControl) or N-WASP-specific siRNA (siN-WASP). N-WASP and GAPDH expression were analyzed by western blotting. (G) Confocal images of astrocytes after transfection with control siRNA (siControl) or N-WASP-specific siRNA (siN-WASP), followed by serum starvation, forskolin treatment and phalloidin staining. Scale bars: 10 µm. (H) Frequency analysis of astrocyte complexity of N-WASP-knockdown and control cells after forskolin treatment. For frequency analysis, n = 300 cells per condition from three independent experiments. (I) Quantification of the proportion of polygonal astrocytes from control (siControl) and N-WASP-depleted cells (siN-WASP), as shown in H. **P<0.005 (unpaired Student's t-test). (J) Astrocytes were transfected with control siRNA (siControl) or PICK1-specific siRNAs (siPICK1). PICK1 and GAPDH expression were analyzed by western blotting. (K) Confocal images of astrocytes after transfection with control siRNA (left) and PICK1-specific siRNA (right), stained with actin and PICK1-specific antibodies. Before fixation and immunocytochemistry, cells were serum-starved and treated with forskolin. Scale bars: 10 µm. (L) Frequency analysis of astrocyte complexity of PICK1-knockdown and control cells after forskolin treatment (300 cells per condition from three independent experiments in each frequency analysis). (M) Sholl analysis on processes of astrocytes transfected with either control siRNA (siControl, blue) or PICK1-specific siRNA (siPICK1, red). n = 108 (siControl), n = 82 (siPICK1), *P<0.05 (unpaired Student's t-test and Sidak–Bonferroni method).

Fig. 5.

PICK1 knockdown inhibits morphological changes in astrocytes in response to OGD. (A) Confocal images of serum-starved and forskolin-treated astrocytes before and after 20 min of OGD. Cell morphology was visualized by F-actin staining with phalloidin–Alexa-546. Scale bars: 10 µm. (B) Frequency analysis on complexity of control astrocytes before and after 20 min OGD (n = 300 cells per condition from three independent experiments for each frequency analysis). (C) Frequency analysis on cell complexity of PICK1-deficient astrocytes before and after 20 min OGD (n = 300 cells per condition from three independent experiments for each frequency analysis). (D) Direct comparison of cell complexities of control and PICK1-deficient astrocytes after 20 min OGD. (E) Quantification of the proportion of polygonal astrocytes, as shown in B, C and D. ***P<0.0005 (ANOVA with Bonferroni's correction). (F) Confocal images of serum-starved and forskolin-treated siControl- and siPICK1-transfected astrocytes after 20 min OGD, followed by 3 h reperfusion with oxygenated and glucose-containing basal medium. Visualization of morphology by F-actin staining with phalloidin–Alexa-546. Scale bars: 10 µm. (G) Frequency analysis on cell complexity of control astrocytes after 20 min OGD, and after 20 min OGD and 3 h of reperfusion (OGD/RPF) (n = 300 cells per condition from three independent experiments for each frequency analysis). (H) Frequency analysis on cell complexity of PICK1-deficient astrocytes after 20 min OGD, and after 20 min OGD and 3 h of reperfusion of reperfusion (OGD/RPF) (n = 300 cells per condition from three independent experiments for each frequency analysis). (I) Frequency analysis on cell complexity of control and PICK1-deficient astrocytes after 20 min OGD and 3 h of reperfusion (OGD/RPF). (J) Quantification of the proportion of polygonal astrocytes, as shown in G, H and I. ***P<0.0005 (ANOVA with Bonferroni's correction). (K) Confocal images of serum-starved and forskolin-treated astrocytes after 5 min of OGD with either DMSO or CK-548. Cell morphology was visualized by F-actin staining with phalloidin–Alexa-546. Scale bars: 10 µm. (L) Frequency analysis of astrocytes after 5 min OGD and treated with either DMSO or CK-548 (n = 300 cells per condition from three independent experiments for each frequency analysis). (M) Quantification of the proportion of polygonal astrocytes, as shown in L. **P<0.0005 (unpaired Student's t-test).

Fig. 6.

Arp2/3 stimulation by N-WASP overexpression inhibits OGD-dependent changes in astrocyte morphology. Astrocytes were transfected with GFP, GFP–N-WASP WT or constitutively active GFP–N-WASP-Δ227–267. (A) Confocal images of transfected, serum-starved and forskolin-treated astrocytes before OGD, stained with phalloidin–Alexa-546. Scale bars: 10 µm. (B) Confocal images of transfected astrocytes after 20 min OGD and stained for F-actin. Scale bars: 10 µm. (C) Frequency analysis of complexity of astrocytes transfected with GFP, N-WASP WT and N-WASP-Δ227–267 after OGD. For this analysis only cells with exogenous N-WASP in the cytosol and nuclei were taken into account (300 cells per condition from three independent experiments were used in each frequency analysis). (D) Quantification of the proportion of polygonal astrocytes, as shown in C. *P<0.05 (unpaired Student's t-test).