Cachexia in mice blunts improved cognitive flexibility induced by calorie restriction.

Approximately 50-80 % of cancer patients suffer from cachexia, a metabolic syndrome involving inflammation, appetite loss, and muscle and fat wasting. Another common co-morbidity of cancer patients is cognitive impairment, and clinical evidence suggests the incidence of cachexia is linked to more severe cognitive symptoms. Given the difficulty of studying changes in cognitive function in human cancer patients, we set out to examine key aspects of cognitive performance in a mouse model of pancreatic cancer (pancreatic ductal adenocarcinoma; PDAC) cachexia, using an in-cage operant device (Feeding Experimental Device version 3; FED3) and a reversal learning task. Performance on the operant reversal task was compared to two control groups without cancer: ad libitum fed, sham injected with phosphate buffered saline (PBS), calorie restricted (CR) to 90-95 % of original body weight to control for reduced food intake and weight loss in cachexia mice. Our PDAC model recapitulated features of cachexia, including anorexia, weight loss, muscle wastage and inflammation. CR mice performed significantly better on the reversal task than both PDAC and PBS mice, achieving significantly more reversals and greater pellet retrieval. There was no difference between PBS and PDAC groups. These results suggest that the weight and appetite loss that occurs during cancer is processed by the brain differently to weight loss that occurs as a result of calorie restriction, with PDAC mice not experiencing an increase in motivational drive for food in line with their falling body weight. To mimic the malaise experienced by the PDAC group, we dosed CR mice with LiCl. Low dose (150 mM) LiCl did not affect responding, however, high dose (300 mM) LiCl significantly reduced both number of active pokes and pellet retrieval. This indicates a sickness-induced devaluation of reward, a factor that may impact poor performance of this task in the PDAC group. We additionally examined exploratory and anxiety-like behaviour in PBS and PDAC groups using a battery of maze-based tests. We saw no significant differences in performance between groups in the elevated plus maze, open field or light/dark box, suggesting no elevations in baseline anxiety-like symptoms in this cachexia model. These results occurred in the face of significantly elevated levels of the pro-cachexia factors GDF15, Activin A and Activin B, indicating that elevated levels of these TGF-β family peptides are not sufficient to produce behavioural changes in these tests. Our results provide evidence for a specific impact of sickness state on cognitive flexibility during pancreatic cancer.