For analysis, the cross-sectional areas of fluorescently labeled cell bodies in the ganglion cell or inner nuclear layer of retinal slices were measured (Zeiss LSM Image Examiner Version 3.2.0.70). Electrophysiological recordings were performed on Purkinje BMN 673 cell line cells in cerebellar slices and on acutely isolated Müller cells by the whole-cell patch-clamp technique. Spike activity in the ganglion cell layer of retinae was recorded by MEAs ex vivo. Light-evoked electrical responses of retinal layers were recorded by ERGs in vivo. Details

are described in Supplemental Information. SLO images were obtained from anesthetized mice immediately after ERG recordings as described previously using a Heidelberg Retina Angiograph (HRA I) (Seeliger et al.,

2005). Images were acquired under illumination with an argon laser for fundus autofluorescence and EGFP detection (488 nm) and red-free (RF) Autophagy Compound Library imaging of retinal structures (514 nm). OCT imaging was performed immediately after SLO using a Spectralis HRA+OCT device (Heidelberg Engineering) and a broadband superluminescent diode at λ = 880 nm as light source (Huber et al., 2009). Adaptation for the optical qualities of the mouse eye was achieved as described previously (Fischer et al., 2009). For behavioral tests the animals were kept in ventilated cages (Ehret) in 12/12 hr light/dark cycle with free access to food and water. The tests were performed between hr 2 and 6 of the light phase and registered and analyzed with the ANY-maze software (Stoelting). To assess visual perception of mice several behavioral tests were performed with some modifications (Arqué et al., 2008). For the NOR test, mice were placed at day 1 for 5 min in the empty open field apparatus (gray PVC box 40 × 40 × 34 cm, illumination 160 lux). At day 2, mice were exposed for 10 min to an object A placed 5 cm

from the wall. After 3 min, the animals were exposed for 10 min to two objects: the previous object A and a novel object B, positioned in two opposite Isotretinoin corners, 5 cm from the walls. Both objects presented similar textures, shapes, and sizes but distinctive colors (white versus deep blue plastic caps, 4.5 cm diameter, 2.5 cm height, randomly assigned as “old” or “novel”). The novel object recognition was assessed as the percentage of time the mice explored object B compared to the time of exploration of both objects during the second trial (NOR index = (time B/time A + B) ∗ 100). The Morris water maze consisted of a plastic cylindrical pool (120 cm diameter), which was filled with water (temperature controlled at 22°C ± 1°C, illumination 50 lux at the center of the maze). The water was opaque by the addition of white, nontoxic talcum powder (Pharma Cosmetic). Visual cues were positioned around the pool, 60 to 90 cm from its rim.