Increased gene expression of FGF21 is associated with calorie restriction and extends life in mice when artificially induced, such as via gene therapy. Researchers here find that the FGF21 response to calorie restriction in humans is quite different, which is a start on the journey to explain exactly how it is that calorie restriction has a much larger effect on mouse life span than on human life span, while the short-term benefits observed to date in both species are very similar:
The adaptive response to starvation includes a series of key physiologic changes in fuel utilization. Certain aspects of the starvation response, such as the depletion of adipose lipid stores and serum triglycerides, could, in theory, provide metabolic benefits if recapitulated outside the context of starvation, such as in obese individuals. This concept underscores the intense excitement elicited by the discovery of FGF21, a novel fasting-induced hormone in murine models. Having been ascribed a central role in coordinating the ketogenic response to starvation, FGF21 also mediates additional metabolically beneficial functions in mice. Although such preclinical mouse data provided a rationale to develop FGF21 or FGF21 mimetics to treat human metabolic disease, the very question of whether FGF21 is a fasting-induced hormone in humans has remained unresolved.
In this study, in which healthy volunteers underwent a prolonged, medically supervised fast, we provide strong evidence for induction of FGF21 over a 10-day period. By documenting the serial dynamics of circulating FGF21, we provide insight into why it has been so challenging to establish FGF21 as a fasting hormone in humans: namely, the protracted time scale necessary to elucidate the response. Not only did it take the full 10 days to demonstrate a statistically meaningful induction of FGF21, but we also observed an initial decline in FGF21 levels in the majority of subjects.
In considering potential explanations for the differing dynamics of circulating FGF21 between mice and humans, it is tempting to invoke the divergent metabolic rates between the 2 species. Indeed, the mass-specific resting metabolic rate of mice exceeds that of humans by a factor of approximately 7, which approximates the time-scale difference in the FGF21 effect between mice and humans. Our data, however, suggest that the difference between mice and humans with respect to FGF21 regulation and function may go beyond the time course of its release into the circulation. In contrast to the pattern observed in fasting mice, the ketogenic response in our human subjects preceded the induction of FGF21. Importantly, in several patients, the onset of ketosis occurred at a time point when serum FGF21 levels had dropped below baseline values. While this study does not exclude the possibility that FGF21 promotes ketone production in some contexts these data argue strongly against a paradigmatic role for FGF21 as a determinant of the human ketogenic response to starvation.
The late increase in FGF21 levels in humans correlated with weight loss and markers of tissue stress, providing a rationale to focus future mechanistic studies on the role of FGF21 in regulating fuel production and trafficking during the latter phase of starvation, with a particular focus on human studies, given the apparent evolutionary divergence in some FGF21 functions between mice and humans. With regard to the potential therapeutic implications of this study, we cannot assume that any hypothesized functional roles of FGF21 in the starved state are relevant to supraphysiologic administration of FGF21 mimetics in patients with metabolic disease.