Activation of the hypothalamic-pituitary-adrenal axis by exogenous and endogenous GDF15

Abstract

An acute increase in the circulating concentration of glucocorticoid hormones is essential for the survival of severe somatic stresses. Circulating concentrations of GDF15, a hormone that acts in the brain to reduce food intake, are frequently elevated in stressful states. We now report that GDF15 potently activates the hypothalamic-pituitary-adrenal (HPA) axis in mice and rats. A blocking antibody to the GDNF-family receptor α-like receptor completely prevented the corticosterone response to GDF15 administration. In wild-type mice exposed to a range of stressful stimuli, circulating levels of both corticosterone and GDF15 rose acutely. In the case of Escherichia coli or lipopolysaccharide injections, the vigorous proinflammatory cytokine response elicited was sufficient to produce a near-maximal HPA response, regardless of the presence or absence of GDF15. In contrast, the activation of the HPA axis seen in wild-type mice in response to the administration of genotoxic or endoplasmic reticulum toxins, which do not provoke a marked rise in cytokines, was absent in Gdf15^-/- mice. In conclusion, consistent with its proposed role as a sentinel hormone, endogenous GDF15 is required for the activation of the protective HPA response to toxins that do not induce a substantial cytokine response. In the context of efforts to develop GDF15 as an antiobesity therapeutic, these findings identify a biomarker of target engagement and a previously unrecognized pharmacodynamic effect, which will require monitoring in human studies.

Keywords: adrenal; corticosteroids; gdf15; stress; toxins.

Fig. 1.

Acute administration of human recombinant GDF15 activates the HPA axis in mice. (A and B) Mouse study 1 (MS1) at standard housing condition: acute effect of human recombinant GDF15 administration on (A) endogenous corticosterone and (B) human GDF15 plasma concentration at 1 h and 4 h post-GDF15 treatment. (C and D) MS2 at thermoneutral housing condition: acute effect of two doses of human recombinant GDF15 on (C) plasma corticosterone and (D) human GDF15 concentrations after 1-h treatment (0.03 and 0.1 mg/kg). (E and F) MS3: (E) Corticosterone serum level in anti-GFRAL- and IgG control-treated mice with or without human recombinant GDF15 administration. (F) Human GDF15 serum concentration in GFRAL blocking antibody (anti-GRFAL) and IgG groups treated with human recombinant GDF15. (G and H) MS4: (G) In situ hybridization analysis of Crh mRNA (red), c-Fos mRNA (blue-green), and hematoxylin counterstain for nuclei at the level of the PVN. Left, representative images of coronal sections of vehicle (Upper Left) and GDF15-treated mice (Lower Left). (Scale bar, 50 μm.) 3V, third ventricle. Middle, higher magnification of the PVN showing coexpression of Crh and c-Fos dots in vehicle (Upper Middle) and GDF15 treated (Lower Middle). (Scale bar, 100 μm.) Right, automated quantification of Crh- and c-Fos-positive cells in vehicle (Upper Right) and GDF15-treated mice (Lower Right) (Scale bar, 100 μm.) Black arrows indicate double-labeled cells. Cell nucleus color in the overlay (Right) represents the cell classification (muted red for Crh-positive cells, bright red for dual-positive cells), while nucleus color intensity correlates with spot counts per cell. Bright blue spots represent c-Fos spot assigned to a Crh-positive cell. (H) Percentage of Crh-positive cells that were also c-Fos positive. Data are expressed as mean ± SEM, n = 6 to 8 per group. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, for MS1, -2, and -3, data were analyzed by ANOVA and for MS4, by unpaired Student’s t test.

Fig. 2.

Chronic infusion of human GDF15 activates the HPA axis in rats. Rat study 1 (RS1): (A) Plasma concentration time course of human GDF15 in rats with continuous intravenous infusion of vehicle or human GDF15. Gray shade indicates the period with bolus infusion of 0.24 mg/kg/h followed by a period with maintenance infusion of 0.04 mg/kg/h. (B) Plasma concentration time course of endogenous corticosterone in rats in response to continuous intravenous infusion of vehicle buffer or human GDF15. Plasma corticosterone levels are expressed as the area under the curve (AUC) calculated from time point zero to the termination of the study. (C) Plasma concentration time course of endogenous ACTH in rats in response to continuous intravenous infusion of vehicle or human GDF15. Plasma corticosterone levels expressed as the AUC calculated from time point zero to the termination of the study. Data are expressed as mean ± SEM, n = 5 to 6. *P < 0.05 **P < 0.01, ****P < 0.0001 by repeated measurement model with baseline assessment as a covariate.

Fig. 3.

GDF15 is not necessary in mediating the LPS-induced rise in glucocorticoids in mice and humans. MS5: (A) Mouse GDF15 and (B) corticosterone serum concentrations at baseline (time = 0) and 2 h after LPS (0.5 mg/kg) or vehicle control injection in wild-type and Gdf15^−/− mice. (C) Corticosterone serum concentration 2 h after LPS (0.05 mg/kg) or vehicle control injection in wild-type and Gdf15^−/− mice. (D) MS6: Corticosterone serum concentration 4 h after E. coli infection in wild-type and Gdf15^−/− mice. (E and F) Human study 1 (HS1): Time course of (E) GDF15 and (F) cortisol serum levels at baseline (time = 0) and after 2 ng/kg bolus intravenous infusion of LPS in healthy human subjects. For MS5 and MS6 data are expressed as mean ± SEM, n = 4 to 6 per group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by ANOVA. For HS1 data are expressed as mean ± SEM, n = 11. *P < 0.05, ***P < 0.001 by one-way repeated measures with post hoc Dunnett’s test to compare each time point with baseline.

Fig. 4.

GDF15 is necessary to mediate corticosterone rising upon cisplatin administration. MS7 at standard housing temperature: (A) Mouse GDF15 and (B) corticosterone plasma concentration and (C) Gdf15 mRNA expression in the liver after 6-h cisplatin administration (10 mg/kg) in wild-type and Gdf15^−/− mice. (D–F) MS8 at thermoneutral housing condition: (D) Mouse GDF15 and (E) corticosterone plasma concentration and (F) Gdf15 mRNA expression in the liver after 6-h cisplatin injection (10 mg/kg) in wild-type and Gdf15^−/− mice. Data are expressed as mean ± SEM, For MS7, n = 8 to 9 per group, MS8, n = 13 to 16 per group, ****P < 0.0001 by ANOVA.

Fig. 5.

GDF15 is partially needed to mediate corticosterone rising upon tunicamycin-induced ER stress. (A–G) MS9 at standard housing temperature: (A) Mouse GDF15, (B) corticosterone, and (C) FGF21 serum levels after 6-h tunicamycin injection in wild-type and Gdf15^−/− mice. mRNA expression of (D) Gdf15, (E) Atf4, (F) Chop, and (G) Fgf21 in the liver after 6-h tunicamycin injection in wild-type and Gdf15^−/− mice. (H–N) MS10 at thermoneutral housing condition: (H) Mouse GDF15, (I) corticosterone, and (J) FGF21 plasma levels after 6-h tunicamycin injection in wild-type and Gdf15^−/− mice. (K) Gdf15, (L) Atf4, (M) Chop, and (N) Fgf21 mRNA expression in the liver after 6-h tunicamycin injection. Data are expressed as mean ± SEM, n = 5 to 9 per group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 by ANOVA.

Fig. 6.

GDF15 and FGF21 show a synergic action in mediating tunicamycin-induced corticosterone rising. MS11: Corticosterone serum levels in wild-type, Fgf21^−/−, and Gdf15^−/−:Fgf21^−/− mice 6 h after tunicamycin administration compared to vehicle control. Data are expressed as mean ± SEM, n = 3 to 7 per group. *P < 0.05 by ANOVA.