Meal sugar-protein balance determines postprandial FGF21 response in humans

Abstract

Biological mechanisms to promote dietary balance remain unclear. Fibroblast growth factor 21 (FGF21) has been suggested to contribute to such potential regulation considering that FGF21 1) is genetically associated with carbohydrate/sugar and protein intake in opposite directions, 2) is secreted after sugar ingestion and protein restriction, and 3) pharmacologically reduces sugar and increases protein intake in rodents. To gain insight of the nature of this potential regulation, we aimed to study macronutrient interactions in the secretory regulation of FGF21 in healthy humans. We conducted a randomized, double-blinded, crossover meal study (NCT05061485), wherein healthy volunteers consumed a sucrose drink, a sucrose + protein drink, and a sucrose + fat drink (matched sucrose content), and compared postprandial FGF21 responses between the three macronutrient combinations. Protein suppressed the sucrose-induced FGF21 secretion [incremental area under the curve (iAUC) for sucrose 484 ± 127 vs. sucrose + protein -35 ± 49 pg/mL × h, P < 0.001]. The same could not be demonstrated for fat (iAUC 319 ± 102 pg/mL × h, P = 203 for sucrose + fat vs. sucrose). We found no indications that regulators of glycemic homeostasis could explain this effect. This indicates that FGF21 responds to disproportionate intake of sucrose relative to protein acutely within a meal, and that protein outweighs sucrose in FGF21 regulation. Together with previous findings, our results suggests that FGF21 might act to promote macronutrient balance and sufficient protein intake.NEW & NOTEWORTHY Here we test the interactions between sugar, protein, and fat in human FGF21 regulation and demonstrate that protein, but not fat, suppresses sugar-induced FGF21 secretion. This indicates that protein outweighs the effects of sugar in the secretory regulation of FGF21, and could suggest that the nutrient-specific appetite-regulatory actions of FGF21 might prioritize ensuring sufficient protein intake over limiting sugar intake.

Keywords: fibroblast growth factor 21; macronutrient balance; meal study; protein; sucrose.

Graphical abstract

Figure 1.

Overview of the study design. A: flowchart of the process that resulted in the study sample. B: description of the three test drinks. C: overview of the crossover study design. D: timeline of the test day procedure. Created with BioRender.com.

Figure 2.

Postprandial plasma FGF21 response up to 4 h after the three test drinks (primary outcome). For each of the three test drinks [sucrose (S), sucrose + protein (SP), and sucrose + fat (SF)], A displays the means (SE) of absolute concentrations in FGF21 across the five time points, B displays the means (SE) of fold changes in FGF21 concentrations from fasting at time point 0, and C displays the means (SE) of FGF21 iAUCs. Graphs show crude means (SE) values, while all P values are determined using LMM adjusted for age and sex with individuals treated as random effects. Repeated measurements analyses are performed with log-transformed FGF21 values for both absolute concentrations and fold changes. In A and B, only significant P values for pairwise comparisons are presented. FGF21, fibroblast growth factor 21; iAUC, incremental area under the curve; LMM, linear mixed model.

Figure 3.

Postprandial glucose, insulin, and glucagon responses up to 4 hours after the three test drinks (secondary outcomes). For each of the three test drinks [sucrose (S), sucrose + protein (SP), and sucrose + fat (SF)], A–C displays the mean (SE) of absolute concentrations across the five time points: glucose (A), insulin (B), and glucagon (C). The means (SE) of the iAUCs calculated from the postprandial responses: glucose (D), insulin (E), and glucagon (F). Graphs show crude means (SE) values, while all P values are determined using LMM adjusted for age and sex with individuals treated as random effects. Analyses of repeated measurements are performed with log-transformed values of glycemic markers. Only significant P values for pairwise comparisons are presented.