Astroglial networks control visual responses of superior collicular neurons and sensory-motor behavior

Abstract

Astroglial networks closely interact with neuronal populations, but their functional contribution to neuronal representation of sensory information remains unexplored. The superior colliculus (SC) integrates multi-sensory information by generating distinct spatial patterns of neuronal functional responses to specific sensory stimulation. Here, we report that astrocytes from the mouse SC form extensive networks in the retinorecipient layer compared to visual cortex. This strong astroglial connectivity relies on high expression of gap-junction proteins. Genetic disruption of this connectivity functionally impairs SC retinotopic and orientation preference responses. These alterations are region specific, absent in primary visual cortex, and associated at the circuit level with a specific impairment of collicular neurons synaptic transmission. This has implications for SC-related visually induced innate behavior, as disrupting astroglial networks impairs light-evoked temporary arrest. Our results indicate that astroglial networks shape synaptic circuit activity underlying SC functional visual responses and play a crucial role in integrating visual cues to drive sensory-motor behavior.

Keywords: CP: Neuroscience; astrocytes; astroglial networks; gap junctions; sensory-motor behavior; superior colliculus; synaptic circuits; visual maps.

Graphical abstract

Figure 1

Astrocytes form extensive and non-compartmentalized networks in the visual layers of the SC (A and B) Schematic representation of the experimental approach (A), where GFP-labeled astrocytes from GFAP-EGFP mice were patched with a pipette filled with biocytin to visualize the astroglial network in the visual layers of the SC (B, left) or in V1 (B, right). The SC layers were visualized with myelin basic protein (MBP) stain. The white lines indicate the surface of the SC and V1, respectively, whereas the dotted line shows the borders of the visual layers in the SC. The white arrowheads indicate the injection site. The coupling (median distance of coupled cells from the injection site) for all 4 directions are represented with white arrows. Scale bars, 250 μm (A) and 100 μm (B). (C) Quantification of the size of the astroglial networks in the SC (n = 20) and in V1 (n = 6; p < 0.0001, U = 2, Mann-Whitney test). (D) Dorsoventral coupling in astrocytic networks. Injection site (black line) ± maximal extent (color) of coupled cells into dorsal and ventral directions. The astroglial dorsoventral coupling is plotted against increasing injection site distance from the SC surface. The dotted line represents the visual layers border. (E) Medio-lateral coupling of astroglial networks. Injection site ± maximal extent (color) of coupled cells into medial and lateral directions. The astroglial medio-lateral coupling is plotted against increasing injection site distance from the SC surface. (F and G) Astroglial network coupling in the dorsal (F) and ventral (G) directions with respect to the injection site distance to the SC surface. Linear regression with 95% confidence intervals (dorsal: R² = 0.9352, F(1, 18) = 259.7, p < 0.0001, n = 20; ventral: R² = 0.047, F(1, 18) = 0.8882, p = 0.3584, n = 20). (H) Ratio between the ventral coupling and the medio-lateral coupling for small (<800 coupled cells; n = 15) and large (>800; n = 5) astrocytic networks (p = 0.0009, U = 3, Mann-Whitney test). Data are shown as mean ± SEM.

Figure 2

The extensive astroglial network in the SC is mediated by high levels of Cx30 and Cx43 expression (A) Mice model used for control (GFAP-creERT2 + tamoxifen) and disconnected astrocytes (GFAP-creERT2 Cx30^fl/flCx43^fl/fl [cKD] + tamoxifen) conditions. Tamoxifen was injected in control and cKD mice during 4 consecutive days, and experiments were performed at least 3 weeks after the last injection. (B) Representative examples of astroglial networks in the visual layers of the SC (top) and in V1 (bottom) for control (left) and cKD (right) mice. (C and D) Quantification of the size of the astroglial networks in the SC (C) and in V1 (D) between control and cKD mice. SC: control, n = 5; cKD, n = 6; p < 0.0043, U = 0, Mann-Whitney test. V1: control, n = 5; cKD, n = 5; p = 0.0079, U = 0, Mann-Whitney test. (E) Representative images of brain tissues immunostained for MBP (top row, scale bar, 300 μm), Cx30 and Cx43 (center and bottom rows, scale bar, 50 μm) in the SC (left column) and in V1 (right column). Two to three fields of view were imaged per slice (white squares enlarged in center and bottom rows) from 4 mice to quantify Cx30 and Cx43 expression levels. (F and G) Quantification of Cx30 (F) and Cx43 (G) levels in the visual layers of the SC and in V1. Unpaired t test (Cx30 [t = 13.04, n = 11 for SC and V1, ^∗∗∗p < 0.0001]; Cx43 [t = 9.768, n = 12 for SC and V1, ^∗∗∗p < 0.0001]). Data are shown as mean ± SEM.

Figure 3

Astroglial networks are required for functional retinotopic maps in the SC (A) Schematic representation of the SC imaged with intrinsic optical imaging. (B and C) Averaged retinotopic maps of elevation (top) and azimuth (bottom) in the SC for (B) control (n = 5) and (C) cKD (n = 8) mice. Reproducibility in functional organization across animals was tested with the Moore-Rayleigh test (right). Boundaries of the SC is depicted with dotted line. Scale bar, 1 mm. (D) Quantification of collicular domains selective for elevation and azimuth in control (black) and cKD mice (blue). (E) Representation of visual locations selective for both elevation and azimuth in control (gray area) and cKD (blue area) mice. (F–J) Same as (A)–(E) when retinotopic maps are recorded from the visual cortex of control (n = 5) and cKD (n = 4) mice.

Figure 4

Activity-dependent control of orientation maps and light-induced temporary arrest behavior by astroglial networks (A and B) Averaged map of orientation in control (n = 5) (A) and cKD (n = 8) (B) mice. Reproducibility in functional organization across animals was tested with the Moore-Rayleigh test (right). Boundaries of the SC is depicted with the dotted line. Scale bar, 1 mm. (C) Representation of preferred orientations in control mice. (D) Quantification of collicular domains selective for orientation in control (gray) and cKD (blue) mice. (E and F) Left: schematics showing the location of the stimulation and recording pipette in the SC (E) and in V1 (F) to record field potentials. SGS, stratum griseum superficiale; SO, stratum opticum. Right: quantification of field potential peak amplitude in the SC (D) and in V1 (E) (two-way ANOVA, F(1,30) = 37.35, p < 0.0001) and (G) V1 (two-way ANOVA, F(1,32) = 1.085, p = 0.3053). (G) Schematic representation of the light-induced temporary arrest behavioral test. The mouse is placed in a closed arena. While the mouse is running, 3 flashes of white light (1 s total duration) is presented above the animal, causing the mouse to stop temporarily. (H) Averaged running speed following flashes presentation in control (n = 6) and cKD (n = 6) mice. (I) Quantification of the running speed modulation index. Normalized difference in running speed after light flashes relative to before light flashes (index = 1 indicates full stop) (control, n = 6; cKD, n = 6; p = 0.0022, U = 0, Mann-Whitney test). Data are shown as mean ± SEM.