FGF19, FGF21, and an FGFR1/β-Klotho-Activating Antibody Act on the Nervous System to Regulate Body Weight and Glycemia

Abstract

Despite the different physiologic functions of FGF19 and FGF21 as hormonal regulators of fed and fasted metabolism, their pharmacologic administration causes similar increases in energy expenditure, weight loss, and enhanced insulin sensitivity in obese animals. Here, in genetic loss-of-function studies of the shared co-receptor β-Klotho, we show that these pharmacologic effects are mediated through a common, tissue-specific pathway. Surprisingly, FGF19 and FGF21 actions in liver and adipose tissue are not required for their longer-term weight loss and glycemic effects. In contrast, β-Klotho in neurons is essential for both FGF19 and FGF21 to cause weight loss and lower glucose and insulin levels. We further show an FGF21 mimetic antibody that activates the FGF receptor 1/β-Klotho complex also requires neuronal β-Klotho for its metabolic effects. These studies highlight the importance of the nervous system in mediating the beneficial weight loss and glycemic effects of endocrine FGF drugs.

Keywords: FGF19; FGF19/FGF21 mimetic antibody; FGF21; adipose; diabetes; insulin sensitizer; nervous system; obesity; weight loss; β-Klotho.

Figure 1. FGF19 but not FGF21 has robust direct effects on hepatocytes under acute treatment conditions

(A) DIO Klb^fl/fl and Klb^Alb littermates were i.p. injected with 1 mg/kg FGF19, FGF21 or vehicle and sacrificed 6 hr later. Cyp7a1 mRNA in liver was measured by QPCR. n = 5/group. (B and C) DIO Klb^fl/fl and Klb^Alb littermates were i.p. injected with 1 mg/kg FGF19, FGF21 or vehicle and sacrificed 30 min later. (B) Phosphorylated ERK1/2 normalized to total ERK1/2 protein in liver was measured by western blot. (C) Egr1 mRNA levels in liver was measured by QPCR. n = 6/group. QPCR cycle time values are shown for the vehicle-treated Klb^fl/fl group. Data are shown as the mean ± SEM. **p < 0.01, ***p<0.001 compared to control. N.S., not significant. Significant interactions between genotype (Klb^fl/fl vs. Klb^Alb) and treatment (Vehicle vs. FGF19) were detected for p-ERK1/2/T-ERK/1/2 (p=0.0007) and Egr1 mRNA (p=0.004). See also Figure S1 and Table S1.

Figure 2. β-Klotho in liver is not required for FGF19 and FGF21 to cause weight loss

(A and B) DIO Klb^fl/fl and Klb^Alb littermates were treated with FGF19, FGF21 or vehicle for 2 wk. (A) Body weight change, plasma insulin and glucose levels and (B) hepatic triglyceride and cholesterol levels and plasma adiponectin levels were measured. n = 5–6/group. Data are shown as the mean ± SEM. *p < 0.05, **p < 0.01, ***p<0.001 compared to control. See also Figures S1 and S2, and Table S1.

Figure 3. β-Klotho in adipose tissue is required for the acute insulin-sensitizing actions of FGF21 and FGF19

(A) DIO Klb^fl/fl and Klb^Adipoq littermates were treated with FGF21 or vehicle during a hyperinsulinemic-euglycemic clamp and glucose infusion rate, endogenous glucose production and whole-body glucose uptake during the clamp and basal endogenous glucose production measured. n = 5–6/group. (B) DIO Klb^fl/fl and Klb^Adipoq littermates were treated with FGF19 or vehicle during a hyperinsulinemic-euglycemic clamp and glucose infusion rate, endogenous glucose production and whole-body glucose uptake during the clamp and basal endogenous glucose production measured. n = 6–8/group. Data are shown as the mean ± SEM. *p < 0.05, **p < 0.01 compared to control. Significant interactions between genotype (Klb^fl/fl vs. Klb^Adipoq) and treatment (Vehicle vs. FGF21) were detected for glucose infusion rate (p=0.03) and endogenous glucose production (p=0.02). See also Table S1.

Figure 4. β-Klotho in adipose tissue is not required for FGF19 and FGF21 to cause weight loss

(A and B) DIO Klb^fl/fl and Klb^Adipoq littermates were treated with FGF19, FGF21 or vehicle for 2 wk. Body weight change, plasma insulin and glucose levels, hepatic triglyceride levels, plasma adiponectin levels and hepatic ceramide concentrations were measured. n = 5–6/group. Data are shown as the mean ± SEM. *p < 0.05, **p < 0.01, ***p<0.001 compared to control. See also Figure S1 and S3, and Table S1.

Figure 5. FGF19 acts on the nervous system to cause weight loss

(A) DIO Klb^fl/fl and Klb^Camk2a littermates were administered FGF19 or vehicle for 2 wk. Body weight, plasma insulin and glucose levels, hepatic triglyceride levels and plasma adiponectin levels were measured. n = 5–6/group. (B) Percent change in sympathetic nerve activity subserving BAT was recorded in mice following i.c.v. (top panel) or i.v. (bottom panel) injection of FGF19 at the indicated doses or vehicle. n = 7–8/group. (C) Representative 24 hr energy expenditure in DIO mice administered FGF19 or vehicle (left panel). Quantification of 24 hr energy expenditure for the same mice (right panel). n = 5/group. (D) 24 hr food intake (left and middle panels) and activity (right panel) in DIO mice treated with FGF19 or vehicle as in (C). n = 5/group. Data are shown as the mean ± SEM. *p < 0.05, **p < 0.01, ***p<0.001 compared to control. Significant interactions between genotype (Klb^fl/fl vs. Klb^Camk2a) and treatment (Vehicle vs. FGF19) were detected for body weight change at week 2 (p=0.04) and plasma glucose (p=0.02). See also Figure S1 and S4, and Table S1.

Figure 6. The FGFR1c/β-Klotho activating antibody acts directly on the nervous system to regulate body weight and circulating glucose and insulin levels

(A) DIO Klb^fl/fl and Klb^Adipoq littermates were administered either bFKB1 or control antibody (trastuzumab) for 4 wk, with dosing every two weeks. Body weight change, plasma insulin and glucose levels, hepatic triglyceride levels and plasma adiponectin levels were measured. n = 6/group. (B) DIO Klb^fl/fl and Klb^Camk2a littermates were administered bFKB1 or control antibody (trastuzumab) for 4 wk. Body weight change, plasma insulin and glucose levels, hepatic triglyceride levels and plasma adiponectin levels were measured. n = 6/group. (C) bFKB1 levels in the cerebrospinal fluid (CSF) and plasma were measured in DIO mice 10 days after i.p. injection of bFKB1 or PBS (left and middle panels). CSF/plasma ratio for bFKB1 (right panel). n = 4–5/group. Data are shown as the mean ± SEM. *p < 0.05, **p < 0.01, ***p<0.001 compared to control. ND, not detected. Significant interactions between genotype (Klb^fl/fl vs. Klb^Camk2a) and treatment (control IgG vs. bFKB1) were detected for body weight change at week 4 (p=0.04) and plasma glucose (p=0.004). See also Figure S5 and Table S1.