Fibroblast growth factor 21 (FGF21) is a stress hormone that is released from the liver in response to nutritional and metabolic challenges. In addition to its well-described effects on systemic metabolism, a growing body of literature now supports the notion that FGF21 also acts via the central nervous system to control feeding behavior. Here we review the current understanding of FGF21 as a hormone regulating feeding behavior in rodents, non-human primates, and humans. First, we examine the nutritional contexts that induce FGF21 secretion. Initial reports describing FGF21 as a ‘starvation hormone’ have now been further refined. FGF21 is now better understood as an endocrine mediator of the intracellular stress response to various nutritional manipulations, including excess sugars and alcohol, caloric deficits, a ketogenic diet, and amino acid restriction. We discuss FGF21’s effects on energy intake and macronutrient choice, together with our current understanding of the underlying neural mechanisms. We argue that the behavioral effects of FGF21 function primarily to maintain systemic macronutrient homeostasis, and in particular to maintain an adequate supply of protein and amino acids for use by the cells.

The liver regulates energy partitioning and utilization in a sex-dependent manner, coupling hepatic substrate availability to female reproductive status. Fibroblast growth factor-21 (FGF21) is a hepatokine produced in response to metabolic stress that adaptively directs systemic metabolism and substrate utilization to reduce hepatic lipid storage. Here we report that FGF21 alters hepatic transcriptional and metabolic responses, and reduces liver triglycerides, in a sex-dependent manner. FGF21 decreased hepatic triglycerides in obese male mice in a weight loss-independent manner; this was abrogated among female littermates. The effect of FGF21 on hepatosteatosis is thought to derive, in part, from increased adiponectin secretion.Accordingly, plasma adiponectin and its upstream adrenergic receptor à cAMP à EPAC1 signaling pathway was stimulated by FGF21 in males and inhibited in females. Both ovariectomized and reproductively senescent, old females responded to FGF21 treatment by decreasing body weight, but liver triglycerides and adiponectin remained unchanged. Thus, the benefit of FGF21 treatment for improving hepatosteatosis depends on sex, but not on a functional female reproductive system. Because FGF21 provides a downstream mechanism contributing to several metabolic interventions, and given its direct clinical importance, these findings may have broad implications for the targeted application of nutritional and pharmacological treatments for metabolic disease.

No abstract

Pharmacological administration of Fibroblast growth factor 21 (FGF21) alters food choice, including that it decreases the consumption of sucrose and other sweet tastants. Conversely, endogenous secretion of FGF21 by the liver is modulated by diet, such that plasma FGF21 is increased after eating foods that have a low dietary protein: total energy (P: E) ratio. Together, these findings suggest a strategy to promote healthy eating, in which the macronutrient content of a pre-load meal could reduce the later consumption of sweet desserts. Here, we tested the prediction that individuals eating a low P: E pre-load meal, and next offered a highly palatable sweet 'dessert', would eat less of the sugary snack compared to controls, due to increased FGF21 signaling. In addition to decreasing sweet intake, FGF21 increases the consumption of dietary protein. Thus, we predicted that individuals eating a low protein pre-load meal, and subsequently offered a very high-protein pellet as 'dessert' or snack, would eat more of the high protein pellet compared to controls, and that this depends on FGF21. We tested this in C57Bl/6J, and liver-specific FGF21-null (FGF21ΔL) null male and female mice and littermate controls. Contrary to expectation, eating a low protein pre-load did not reduce the later consumption of a sweet solution in either males or females, despite robustly increasing plasma FGF21. Rather, eating the low protein pre-load increased later consumption of a high protein pellet. This was more apparent among males and was abrogated in the FGF21ΔL mice. We conclude that physiologic induction of hepatic FGF21 by a low protein pre-load is not sufficient to reduce later consumption of sweet dessert, though it effectively increases the subsequent intake of dietary protein in male mice.