Fig 2: Culture requirements for astrocyte tiling. (a) Neuron and astrocytes in rat cerebrum neuron–astrocyte co-culture immunostained on DIV15. n = 4. Top, cellular proliferation was suppressed by 5 µM AraC. Middle, proliferation was suppressed by 5 µM AraC, and action potential generation was suppressed by 10 µM TTX supplemented every two days. Bottom, without proliferation suppression. Top, GFAP (red), GLT1 (green), and GLAST (blue). Bottom, MAP2 (green), and GFAP (magenta). Scale bar, 100 µm. Similar results were obtained for 3 independently prepared cultures. (b) Astrocytes in rat cerebrum primary glial culture treated with glutamate receptor ligands for 24 h, immunostained on DIV17. GLT1 (green), GFAP (red), and GLAST (blue). Neuronal conditioned medium was obtained from neuron–astrocyte co-culture prepared in parallel. Cells were prepared in parallel with Figure 1a top and imaged under the same conditions with the same linear signal enhancement. Scale bar, 100 µm. Similar results were obtained for 3 independently prepared cultures. Images of each channel are provided in Figure S4.

Fig 3: Neuronal damage and astrocyte deformation caused by glutamate. (a) Effect of incubation with glutamate receptor ligands for 24 h on rat cerebrum neuron–astrocyte co-culture. Immunostained by astrocytic markers of GFAP (green), GLAST (red), and MAP2 (blue). Similar results were obtained for 3 independently prepared cultures. Scale bar, 100 µm. Images of each channel are provided in Figure S5. (b) The time-course of the area covered by neurons. The percentages of neuronal dendritic marker MAP2 positive area were calculated by image thresholding. Mean ± s.e.m. of 8 images. Similar results were obtained for 3 independently prepared cultures. (c) Top left, the total length of branches of each astrocyte. Top right, the total number of branches of each astrocyte. Middle left, the area covered by each astrocyte, estimated by the area of a rectangle that covers all branches of each astrocyte. Middle right, the average number of branches that touch or override branches of other astrocytes to violate the tiling rule. Bottom left, the percentages of coverage area by neuronal dendritic marker MAP2. Bottom right, topological branch order analysis for astrocytes. Error bars are shown for every 5 branch orders. Cont-Glu, Cont-NMDA, Cont-GluAPV, NMDA-NMDAAPV, Glu-NMDAAPV pairs were p < 0.05 by Tukey–Kramer test of order 10 branch numbers. Mean ± s.e.m. of 92–104 cells. Asterisks, p > 0.05 with Tukey–Kramer test. Similar results were obtained for 3 independently prepared cultures. (d) The time-course of morphological changes in astrocytes expressing EGFP after application of 20 µM glutamate. Scale bar, 100 µm. Similar results were obtained for 3 independently prepared cultures with more than 10 cells observed each time.

Fig 4: Morphology of astrocytes and neurons in co-culture. (a) Astrocytes and neurons in rat cerebrum neuron–astrocyte co-culture on DIV15, immunostained by neuronal or astrocytic marker antibodies. Left, astrocytic markers of GLT1 (green), GFAP (red), and GLAST (blue). Middle, a neuronal dendritic marker of MAP2 (green) and astrocytic markers of GAT3 (red) and AQP4 (blue). Right, GLT1 (green), GFAP (red), and phospho-ezrin (blue). Top, scale bar, 100 µm. Bottom, an enlarged view of the area indicated with a square in the left image. Scale bar, 25 µm. Similar results were obtained for 3–7 independently prepared cultures. Images of each channel are provided in Figure S1. (b) High-magnification images of astrocytes and neurons in rat cerebrum neuron–astrocyte co-culture on DIV19. Astrocytic marker GLT1 (green), MAP2 (red), and phospho-ezrin (blue). Scale bar, 10 µm. Similar results were obtained for 3 independently prepared cultures.