Since feeding behavior varies under ad libitum feeding, we controlled behavior using a restricted

feeding paradigm (Gooley et al., 2006 and Mistlberger, 1994; Figure 1A) in which access to food pellets was limited to 4 hr per day (11:00 to 15:00). After habituation for 9 days, mice were this website mostly devoted to eating during the initial hour (11:00–12:00) after supply but showed various postprandial behaviors during the following hour (12:00–13:00), such as grooming, resting, and sleeping (see Figure 3A). Mice were sampled at various times (Figure 1A; day 10) and the number of caspase-3-activated GCs was counted (Figures 1B and 1E). The number at 2 hr after the start of supply (13:00) was an average 2.4-fold higher than that before supply (11:00), and then tended to decrease at 4 hr TGF-beta inhibitor after the start of supply (15:00). In a separate experiment, caspase-3-activated GC number showed no significant increase at 1 hr after the start of feeding (see Figure 3C). Number of caspase-3-activated GCs thus increased in the short time window between

1 and 2 hr after the start of feeding and declined by 4 hr. Outside this feeding time window, the caspase-3-activated GC number was similar to or less than that immediately before supply (Figure 1E). To examine whether an increase in caspase-3-activated GCs also occurs during feeding and postprandial period at a different circadian time, a different feeding time (21:00 to 1:00) was set in a second group

of mice (Figure 1F). Results showed an increase in the number of caspase-3-activated GCs during the shifted time window of feeding and postprandial behaviors. This increase in the number of caspase-3-activated GCs during the eating and postprandial periods suggests that the number of apoptotic GCs would be increased during this time. However, given that activation of caspase-3 is not always associated with cell death (D’Amelio et al., 2010), we also Suplatast tosilate examined the TUNEL method of detecting cell death, which detects DNA fragmentation. Results showed a remarkable increase in TUNEL-positive GCs during the feeding and postprandial period, confirming the increased death of GCs during the time window (Figures 1C and 1H). This result indicates that caspase-3-activation is an excellent indicator of GC death. In fact, most caspase-3-activated GCs were TUNEL-positive (before feeding, 85.5% ± 4.7%; 2 hr after the start of food supply, 83.5% ± 3.4%; n = 4 mice, average ± SEM) (Figure 1G). In addition, about 96% of caspase-3-activated GCs showed apoptotic chromatin deformities, such as marginalization, fragmentation, and condensation (Clarke, 1990), whereas activated caspase-3-negative GCs showed no nuclear deformity (Figure 1G and see Figure S1A available online). The increase in the number of caspase-3-activated apoptotic GCs during the feeding and postprandial period occurred throughout the OB, from the rostral to caudal regions (Figure S1B).