The larger increase in apoptotic GCs at 2 hr was partially S3I-201 in vitro but significantly suppressed by disturbing grooming, resting and sleeping behavior during the 2 hr. Gentle handling in nostril-occluded mice did not reduce the amount of food pellet consumed (data not shown). These results indicate that enhanced GC apoptosis occurred in association with postprandial behaviors in sensory-deprived OB. Under unilateral sensory deprivation, enhanced GC apoptosis can occur in association with postprandial extended grooming even without apparent sleep. GC apoptosis in the open side of the OB of the nostril-occluded mice also

showed an increase in GC apoptosis at 1 hr, and this increase was also suppressed by gentle handling (Figure 5G). The presence of olfactory sensory input to the open side of the OB and its absence to the closed side during feeding time was

confirmed by examining the presence and MDV3100 mouse absence of induced arc expression in GCs (Figure S4G; Guthrie et al., 2000). The odor map of the OB shows domain and cluster organization (Mori et al., 2006). The survival rate of adult-born GCs is regulated in local OB areas by local activation with odor learning (Alonso et al., 2006). Does local sensory input regulate the extent of GC elimination during the postprandial period in local OB areas? To address this question, we utilized dorsal zone-depleted mice (ΔD mice), in which olfactory sensory neurons (OSNs) in the dorsal zone (D-zone) of the epithelium were selectively ablated (Kobayakawa et al., 2007). Glomerular structure was lacking in the D-domain of the ΔD mouse OB due to the depletion of OSNs targeting the D-domain (Figure S5A). Other layers were largely maintained, including the granule cell layer (GCL), the majority of cells in which were NeuN-expressing GCs (data not shown). As expected, the number of GCs expressing an immediate early gene c-fos with odor stimulation (Magavi et al., 2005) was drastically reduced in the D-domain (Figure S5B). The quantitative analyses in the paragraph below were over conducted in coronal sections at the central portion in the rostrocaudal axis of ΔD and wild-type mouse OBs, which include a considerable

volume of both the D-domain and ventral domain (V-domain) (Figure S5C). ΔD mice and wild-type mice were subjected to food restriction and examined for caspase-3-activated GCs in the D- and V-domains (Figures 6A and S5D). In the ΔD mouse OB, the density of caspase-3-activated GCs in the D-domain increased 3.2-fold during the postprandial period compared to that before feeding, while that in the V-domain increased 2.2-fold (Figures 6A and 6B). The ratio of caspase-3-activated GC density in the D-domain to that in the V-domain was greater in the postprandial period (2.0 ± 0.2; average ± SEM) than before food (1.3 ± 0.1) (Figure 6D; p = 0.009). In wild-type mouse OB, the density of caspase-3-activated GCs increased during the postprandial period by 2.3-fold in the D-domain and 2.