Evidence for brain serotonin-mediated control of carbohydrate consumption in normal weight and obese humans

A plasma insulin and amino acid-mediated mechanism is thought to modulate brain serotonin concentration, thereby regulating carbohydrate consumption on a meal to meal basis. It has been suggested that obesity is associated with a defect in the appetite control system. Furthermore, post-absorptive plasma levels of several amino acids are increased in obese subjects, which is ascribed to obesity-associated insulin resistance and/or hyperinsulinemia. We studied breakfast-induced changes in plasma ratios of tryptophan to other large neutral amino acids and associated differences in macronutrient composition of lunch food in normal weight and obese human subjects. The study was randomized, double blind and cross-over with a 2 X 2 factorial design with drug/placebo and type of breakfast as factors. Nineteen healthy, non-obese (body mass index (BMI) 22.5 +/- 1.9 kg/m2, mean +/- s.d.) and 19 obese (BMI 34.7 +/- 6.2 kg/m2) female volunteers were treated with either 60 mg fluoxetine (FXT), a serotonin re-uptake blocker specifically acting in the brain, or placebo for four days with a wash-out period between treatments of four weeks. The subjects received either a carbohydrate (CHO) breakfast (80 g maltodextrin, 300 kcal) or a protein-rich (PROT) breakfast (60% milk protein and 40% CHO, 300 kcal) on two consecutive days (days 4 and 5 of each treatment period). Plasma glucose, insulin and amino acids were measured at several time points after breakfast. Three hours after breakfast, subjects were able to choose from 29 different food items. Total energy content and weight of lunch food and energy percentage of each macronutrient were calculated. Data were analysed using repeated measures analysis of variance. The plasma Trp/LNAA ratio was significantly higher at 120 min after a CHO breakfast than at 120 min after a PROT breakfast. The energy percentage of carbohydrate consumed for lunch after a CHO breakfast was significantly lower than after a PROT breakfast. This effect of breakfast type on macronutrient composition of lunch food was abolished by the use of FXT. The drug in itself induced a decrease of total caloric intake for lunch, but did not significantly affect macronutrient composition of the meal. We did not find differences between obese and normal weight subjects concerning their response to the nutritional and serotoninergic challenges. However, the plasma Trp/LNAA ratio, irrespective of the challenge, was higher in obese subjects both in fasting conditions and at 120 min after breakfast, which was associated with a slightly (but not significantly) lower carbohydrate content and higher fat content of lunch food in obese subjects. This study supports the results of animal experiments indicating that the consumption of a protein-free, carbohydrate-rich meal suppresses the intake of carbohydrate in proportion to other macronutrients during a subsequent meal. This effect might be mediated by enhanced central serotoninergic neurotransmission, brought about by the consumption of carbohydrates. The small differences in macronutrient composition of lunch food between normal weight and obese subjects could be the result of differences in plasma amino acid concentrations and serotonin-mediated neurotransmission in the brain.