Reward and punishment alterations are observed in depressed patients (Taylor Tavares et al., Neuroimage, 2008). Reward hyposensitivity and punishment hyper-sensitivity can be studied in humans and animals using operant tasks, such as the probabilistic reversal learning (PRL) task (Ineichen et al., Neuropharmacology, 2012).

To determine the effects of the new antidepressant agomelatine (AGO), a MT1/MT2 melatonergic receptor agonist and 5-HT2C receptor antagonist on reward and punishment processing as well as to evaluate its mechanism of action.

Mice learned spatial discrimination for sucrose-pellet reward in an operant apparatus. In the PRL task, mice were assessed on rewardstay, negative feedback sensitivity (NFS) and reversals completed. Using latin-square designs, agomelatine (AGO 10 or 25 mg/kg), melatonin (MT 10 or 25 mg/kg), the 5-HT2C antagonist S 32006 (2.5 or 5 mg/kg) or vehicle (VEH) were administered acutely i.p. An additional group of mice was treated acutely i.p. with the MT1/2 antagonist S 22153 (20 mg/kg) + VEH or S 22153 (20 mg/kg) + AGO (25 mg/kg).

Data underwent a median-split (high vs low performers) according to PRL behaviour under VEH. In low-performing mice specifically, AGO at 25 mg/kg increased reward-stay, decreased NFS and increased reversals completed, relative to VEH. There were no effects of MT or S 32006. Agomelatine effects were counteracted by prior administration of the MT1/2 antagonist S 22153.

Agomelatine acutely is able to increase reward sensitivity and decrease punishment sensitivity in mice. The results suggest a potential synergy between its melatonergic agonist and 5-HT2C antagonist properties.