Connexin 30 controls astroglial polarization during postnatal brain development

Abstract

Astrocytes undergo intense morphological maturation during development, changing from individual sparsely branched cells to polarized and tremendously ramified cells. Connexin 30, an astroglial gap-junction channel-forming protein expressed postnatally, regulates in situ the extension and ramification of astroglial processes. However, the involvement of connexin 30 in astroglial polarization, which is known to control cell morphology, remains unexplored. We found that connexin 30, independently of gap-junction-mediated intercellular biochemical coupling, alters the orientation of astrocyte protrusion, centrosome and Golgi apparatus during polarized migration in an in vitro wound-healing assay. Connexin 30 sets the orientation of astroglial motile protrusions via modulation of the laminin/β1 integrin/Cdc42 polarity pathway. Connexin 30 indeed reduces laminin levels, inhibits the redistribution of the β1-integrin extracellular matrix receptors, and inhibits the recruitment and activation of the small Rho GTPase Cdc42 at the leading edge of migrating astrocytes. In vivo, connexin 30, the expression of which is developmentally regulated, also contributes to the establishment of hippocampal astrocyte polarity during postnatal maturation. This study thus reveals that connexin 30 controls astroglial polarity during development.

Keywords: Astrocytes; Connexin; Development; Hippocampus; Mouse; Polarity.

Fig. 1.

Cx30 regulates the orientation of migrating astrocytes. Primary astrocyte cultures were transfected with the Cx30-Venus plasmid. (A) 8 h after wounding, control and Cx30-expressing cells were immunolabeled for tubulin and the orientation of their tubulin fibers were measured. Scale bar: 20 µm. (B) The orientation of tubulin fibers was quantified in transfected and non-transfected cells as the mean angular variation to the perpendicular in relation to the wound [control (Ct), n=48 cells; Cx30, n=30 cells]. Cx30-expressing cells showed a greater angular deviation than control cells, suggesting that their cytoskeletal orientation was impaired. (C) Schematic depicting centrosome (red) reorientation of wound edge astrocytes (green) in front of the nucleus (blue) in the control condition. (D) 8 h after wounding, cells were immunolabeled for γ-tubulin (red) to mark the centrosome. The orientation of the nucleus-centrosome axis was measured. Arrows indicate the direction of the nucleus-centrosome axis. (E) Percentage of cells with their centrosome oriented perpendicularly (±45°) to the wound (Ct, n=313 cells; Cx30, n=187 cells; Cx30T5M, n=418 cells). Cx30 and Cx30T5M decreased the proportion of astrocytes with their centrosome properly reoriented. (F) 8 h after wounding, cells were immunolabeled for GM130 and the orientation of the nucleus-Golgi axis was measured. Scale bar: 20 µm. Arrows indicate the direction of the nucleus-Golgi axis. (G) Percentage of cells with their Golgi apparatus oriented perpendicularly (±60°) to the wound (Ct, n=352 cells; Cx30, n=153 cells; Cx30T5M, n=503 cells). Cx30 and Cx30T5M decreased the proportion of astrocytes with their Golgi properly orientated. Asterisks indicate statistical significance (***P<0.001; **P<0.01).

Fig. 2.

Cx30 regulates the secretion of laminin in migrating astrocytes. (A) Immunoblot detection of laminin expression in control (Ct) or Cx30-expressing (Cx30) astrocytes before (no treatment, NT) or after wounding (wound, 6 h). GAPDH was used as a loading control. (B) Quantitative analysis of laminin expression (n=6 cultures). Laminin expression values after wounding were normalized to levels before wounding (NT) in Ct and Cx30-expressing astrocytes. Cx30 prevented the increase in relative laminin levels after wounding. (C) Primary astrocytes were transfected with GFP (control) or Cx30-venus (Cx30). 24 h after wounding, cells were immunolabeled for laminin. Scale bar: 20 μm. (D) Quantitative analysis of laminin expression (area occupied by laminin puncta normalized to total cell area) and axial distribution (distance of laminin puncta to leading edge normalized to cell length). Cx30 impaired the proper recruitment of laminin toward the leading edge (Ct, n=14 cells; Cx30, n=15 cells). Asterisks indicate statistical significance (*P<0.05, **P<0.01).

Fig. 3.

Cx30 controls the polarized distribution of β1 integrin receptors in migrating astrocytes. (A) Primary astrocytes were transfected with β1 integrin-GFP and Cx30. 8 h after wounding, cells were immunolabeled for Cx30 and the intensity of the β1 integrin-GFP signal at the leading edge was measured. Scale bar: 20 µm. (B) Quantitative analysis of the β1 integrin-GFP signal intensity at the leading edge. Cx30 sharply reduced the recruitment of β1 integrin-GFP at the leading edge (Ct, n=15 cells; Cx30, n=9 cells). Asterisks indicate statistical significance (**P<0.01).

Fig. 4.

Cx30 controls the recruitment and activation of Cdc42 in migrating astrocytes. (A) Primary astrocyte cultures were transfected with Cx30 and GFP-Cdc42. 8 h after wounding, cells were immunostained for Cx30 and the intensity of the GFP-Cdc42 signal at the leading edge was measured. Scale bar: 20 µm. (B) Quantitative analysis of the linear intensity profile of GFP-Cdc42 at the leading edge. Cx30 reduced significantly the recruitment of GFP-Cdc42 at the leading edge (Ct, n=19 cells; Cx30, n=15 cells). (C) Cdc42 pull-down activation assay in astrocytes transfected with Cx30 or GFP 30 min after wounding. Immunoblots showing Cdc42 protein levels in the total (Input) and GTP-bound fractions (GST-PAK-PBD). GAPDH (total) and Ponceau staining (GTP) were used as loading controls. (D) Quantitative analysis of Cdc42 activation (n=3 cultures). Cdc42-GTP values were normalized to Cdc42 levels in the total fractions. Cx30 inhibited Cdc42 activation. Asterisk indicates statistical significance (*P<0.05, paired t-test).

Fig. 5.

Cx30 regulates hippocampal synaptic levels of laminin proteins. (A) Western blot detection of pre- (synapsin I, SYN1) and post- (PSD95) synaptic proteins showed an enrichment of plasma membrane synaptic proteins in crude synaptosomal membrane fraction (S) compared with total hippocampal fraction (T), whereas actin protein levels were clearly similar in both fractions. (B) Immunoblot detection of laminin expression in synaptosomal membrane fractions from mouse hippocampi. (C) Histogram showing the quantitative analysis of normalized laminin expression levels (+/+, n=3 mice; −/−, n=3 mice). The loss of Cx30 expression increased significantly the synaptic levels of laminin. (D) Immunoblot detection of laminin β2 protein levels in crude hippocampal membrane fractions. (E) Quantitative analysis of laminin β2 normalized expressions. Hippocampal synapses lacking Cx30 upregulated laminin β2 expression. In both experiments, actin was used as a loading control. Asterisks indicate statistical significance (*P<0.05). S+/+, wild-type mice; S−/−, Cx30^−/− mice.

Fig. 6.

Cx30 sets the polarity of hippocampal astrocytes in vivo. (A) Schematic representation of grid-baseline analysis for orientation quantification of GFAP-labeled CA1 s.r. astrocytes. Scale bar: 20 µm. (B) Quantitative analysis of astrocytes polarity with respect to the pyramidal cell layer. Cx30 was found to be required for the proper perpendicular orientation of developing astrocytes (+/+ P10, n=43 cells; +/+ 1 month, n=44 cells; −/− 1 month, n=49 cells; Cx30T5M 1 month, n=47 cells). (C,D) Cx30 expression in astrocytes from the hippocampal CA1 s.r. region of Cx30^−/− mice (1-month-old) infected with AAV-GFP (C) or AAV-GFP-Cx30 (D), as shown by GFAP, GFP and Cx30 labeling in hippocampal sections. Scale bar: 15 µm. (E) Quantitative analysis of astrocytes polarity in −/− mice showing that wild-type polarity is rescued in s.r. astrocytes by restoring Cx30 expression (n=32 cells), but not by expressing GFP alone (n=34 cells). (F) Immunohistochemical labeling of GM130 (green) and GFAP (red) in the CA1 area of the hippocampus. The diagram illustrates the quantification used to estimate the angular orientation of astrocyte Golgi with regards to the pyramidal layer. Scale bar: 30 µm. (G) Percentage of cells with their Golgi apparatus oriented perpendicularly (±30°) to the pyramidal layer (+/+, n=216 cells; −/−, n=262 cells). Cx30 deficiency decreased the proportion of CA1 astrocytes with their Golgi oriented perpendicularly to the pyramidal layer. Asterisks indicate statistical significance (***P<0.001; **P<0.01).

Fig. 7.

Cx30 sets the spatial properties of intercellular communication in astroglial networks. (A) Representative images of GJ-mediated coupling between CA1 astrocytes visualized by diffusion of sulforhodamine B via patch clamp whole-cell recording. Schematic showing the method used for estimating the spatial properties of CA1 astroglial networks. Scale bar: 100 µm. (B) Graph summarizing the spatial properties of CA1 astroglial networks in the form of polarity indexes. Cx30 is required for the anisotropic coupling of CA1 s.r. astrocytes (+/+, n=8 slices; −/−, n=10 slices). Polarity index ratio larger than one indicates preferential perpendicular orientation towards the pyramidal layer. Asterisk indicates statistical significance (*P<0.05).