We first sought to determine the behavioral and metabolic consequences of exposure to calorie restriction (CR). Mice were exposed to a CR protocol in which they received 60% of ad lib calories daily for 10 days. During this time, mice lost ~15-20% of their original body weight (Figure S1 in the Supplement
). The mice were then given free access to regular chow. There was no significant difference in body weight between mice exposed to CR and ad lib fed mice within two days of re-feeding. Both groups were then allowed additional recovery, with behavioral and metabolic testing conducted the following week.
To test for motivation to obtain calorically dense food, mice were trained to nosepoke for higher in fat (HFD) pellets (22.7% fat) prior to exposure to CR. After the recovery period, the mice were moved to a progressive ratio schedule in which each successive reward required a greater number of nosepokes. The last reward earned within 30 min was used as our measure of instrumental responding for HFD. Mice with a history of CR earned a significantly greater number of rewards on the progressive ratio schedule compared to ad lib fed mice () in the week after regaining their lost weight. No difference was detected between the two groups after 2 weeks recovery (data not shown).
ΔFosB expression in NAc enhances motivation for HFD
Next we wanted to determine the effect of a history of CR on metabolic rate. A separate group of CR mice were analyzed for metabolic parameters using indirect calorimetry. One week after achieving stable weight, hourly measurements were collected for three consecutive days. Mice with a history of CR demonstrated reduced consumption of oxygen and production of carbon dioxide, suggesting a persistent decrease in energy expenditure (). Importantly, body weight and food intake did not differ between the two groups during this time (). Interestingly, mice with a history of CR displayed locomotor hyperactivity (), despite the reduced metabolic rate. Finally, we measured body composition at the end of the experiment. Animals with a history of CR displayed significantly increased levels of body fat () compared to ad lib fed mice, which indicates that a history of CR promotes a repartitioning of energy stores into adipose tissue. These findings demonstrate that the increased energy expenditure and reduced adiposity seen in transgenic mice that over-express ΔFosB are mediated via non-NAc mechanisms (11
Metabolic parameters after calorie restriction or ΔFosB over-expression
To test our hypothesis that accumulation of ΔFosB in NAc may be an important regulator of food intake and metabolism after CR, we first determined the effect of CR on ΔFosB levels. ΔFosB–positive neurons were quantified by immunohistochemistry (). Similar to published results (6
), CR significantly increased the number of neurons in NAc shell, but not NAc core, expressing ΔFosB (). No significant differences in ΔFosB levels were detected two weeks after re-feeding; this time frame is consistent with the observation, noted above, that operant responding does not differ between either group two weeks after re-feeding.
Pharmacologic inhibition of NAc neurons has previously been demonstrated to increase the intake of high fat food via an orexin (also known as hypocretin)-dependent mechanism (4
). Since CR increases motivation to obtain energy dense food (), the observed accumulation of ΔFosB in the NAc shell after CR may mediate the increased motivation to obtain highly palatable food observed after periods of CR. To directly test this hypothesis, we chose viral-mediated gene transfer (AAV-ΔFosB) to increase levels of ΔFosB in NAc, because this system allows for exact temporal and spatial control of ΔFosB expression in adult mice (Figure S2 in the Supplement
). Four weeks after viral injection, mice were trained to nosepoke for HFD pellets. Wild-type mice receiving the control AAV-GFP vector earned fewer rewards than wild-type mice receiving AAV-ΔFosB into the NAc (), indicating that over-expression ΔFosB in NAc was sufficient to increase instrumental responding for HFD. We next determined if this effect was dependent on the presence of orexin, a peptide previously implicated in food intake regulated by the reward circuitry (4
). Orexin-null mice received injection of AAV-GFP or AAV-ΔFosB into the NAc and the number of rewards earned on operant responding was determined. Unlike their wild-type littermates, mice expressing ΔFosB but lacking orexin failed to increase instrumental responding for HFD ().
Next we analyzed several metabolic parameters four weeks after viral injection using indirect calorimetry. Over-expression of ΔFosB decreased oxygen consumption and carbon dioxide production, indicating lower energy expenditure (). Similar to CR mice, there was no difference in body weight or food intake between the two groups during testing (). Interestingly, ΔFosB over-expression in NAc did not reproduce the locomotor hyperactivity phenotype () observed in CR mice. Finally, mice receiving AAV-ΔFosB into the NAc also demonstrated significantly elevated body fat compared to control mice ().