Central action of FGF19 reduces hypothalamic AGRP/NPY neuron activity and improves glucose metabolism

Abstract

Tight control of glucose excursions has been a long-standing goal of treatment for patients with type 2 diabetes mellitus in order to ameliorate the morbidity and mortality associated with hyperglycemia. Fibroblast growth factor (FGF) 19 is a hormone-like enterokine released postprandially that emerged as a potential therapeutic agent for metabolic disorders, including diabetes and obesity. Remarkably, FGF19 treatment has hypoglycemic actions that remain potent in models of genetic and acquired insulin resistance. Here, we provided evidence that the central nervous system responds to FGF19 administered in the periphery. Then, in two mouse models of insulin resistance, leptin-deficiency and high-fat diet feeding, third intra-cerebro-ventricular infusions of FGF19 improved glycemic status, reduced insulin resistance and potentiated insulin signaling in the periphery. In addition, our study highlights a new mechanism of central FGF19 action, involving the suppression of AGRP/NPY neuronal activity. Overall, our work unveils novel regulatory pathways induced by FGF19 that will be useful in the design of novel strategies to control diabetes in obesity.

Keywords: AGRP/NPY neurons; Diabetes; FGF19; Obesity.

Figure 1

Intraperitoneal delivery of recombinant FGF19 promotes ERK1/2 activation in the arcuate nucleus of lean and obese mice. Levels of pERK1/2 were analyzed in response to intraperitoneal injection of FGF19 (0.5 μg/g of body weight) in WT lean mice or ob/ob mice fasted for 24 h. Mice were treated twice (starting 16 h after fasting and 6.5 h apart) and tissues were collected 90 min after the last ip injection. (A) Representative immunoblots of pERK1/2 and total ERK1/2 in liver and in visceral WAT of lean mice injected with saline or FGF19 (n=4 per group). (B) Immunofluorescence analysis showing pERK1/2 staining in the ARC of lean mice injected with saline or with FGF19. Quantification of pERK1/2 positive cells in ARC (Bregma −1.7 to −1.9 mm) of (C) lean mice (12 sections from 3 mice were used in each group) and (D) ob/ob mice (20 sections from 6 mice were examined in each group). (E) Analysis of co-localization between Npy-hrGFP positive neurons (green) and p-ERK1/2 positive cells (red) in Npy-hrGFPob/ob mice fasted for 24 h. Mice were treated twice with ip injection of FGF19 (0.5 μg/g of BW, starting 16 h after fasting and 6.5 h apart) and tissues were collected 90 min after the last intraperitoneal injection. (F) Quantification of Npy-hrGFP positive neurons (green) that colocalized with pERK1/2 staining (red) in ARC of Npy-hrGFPob/ob mice (10 sections from 3 mice were examined in each group). (G) Representative sections showing absence of co-localization between Pomc-eGFP positive neurons (green) and p-ERK1/2 positive cells (red) in Pomc-eGFPob/ob mice fasted for 24 h and treated twice with FGF19 (0.5 μg/g of BW, starting 16 h after fasting and 6.5 h apart). Sections from 3 mice were examined in each group). ^*P<0.05 and ^***P<0.005. Scale bars: 100 μm.

Figure 2

Central FGF19 delivery improves whole body energy balance in HFD-fed and ob/ob mice. (A) Representative sections showing pERK1/2 (red) and co-localization between Npy-GFP positive neurons (green) and p-ERK1/2 positive cells (red) in ARC of HFD-fed Npy-GFP males 90 min after icv injection of aCSF and FGF19 (n=3). (B) Body weight (BW) change of HFD-fed males after 4 daily icv injections of aCSF or FGF19 (n=9 per group). (C) Food intake per mouse was measured 8 and 24 h post-icv injection of aCSF or 2 μg of FGF19 in HFD-fed males (n=5 per group). (D) GTT, with intraperitoneal (ip) injection of 1 g of glucose/kg of body weight (BW), was performed on HFD-fed males, after 6 h of fast (n=4–5 per group). The last icv injection of aCSF or FGF19 was performed 3 h before the GTT assay. (E) Serum insulin concentrations were assessed before and 15 min into the GTT in HFD-fed males treated with 4 daily icv injections of aCSF or FGF19 (n=5 per group). (F) ITT was performed with 0.5 U of insulin/kg of BW in HFD-fed males treated with 4 daily icv injections of aCSF or FGF19, after 6 h of fast (n=4 per group).The last icv injection of aCSF or FGF19 was performed 3 h before the ITT assay. (G) GTT, with ip injection of 0.5 g of glucose/kg of BW, was performed on ob/ob males fasted for 18 h and 3 h after the fourth icv injection (n=6 per group). (H) Insulinemia was assessed before and 15 min into the GTT (n=6 per group). (I) ITT was performed with 3 U of insulin/kg of BW in ob/ob males, after 6 h of fast and 3 h after the fourth icv injection (n=4 per group). ^*P<0.05, ^**P<0.005, and ^***P<0.0005. Scale bars: 100 μm.

Figure 3

Acute icv administration of FGF19 improved insulin-signaling in HFD-fed mice. (A) GTT with 2 g of glucose/kg of body weight (intraperitoneal injection) and (B) ITT (0.5 U of insulin/kg of BW) were performed on HFD-fed males fasted for 6 h. (C) Serum insulin concentrations were assessed before and 15 min into the GTT in HFD-fed males treated with icv injections of aCSF or FGF19. Mice were treated twice with icv injection of aCSF or 2 μg of FGF19 (starting when food was removed and 3 h apart). Analysis were performed 3 h after the last icv (n=5 per group). (D) Ser473 phosphorylation of AKT, pan AKT and GAPDH expression in liver, quadriceps, and epidymal fat (WAT) of HFD-fed males. (E) Quantification of S473pAKT normalized against pan AKT in liver, quadriceps and WAT of HFD-fed males treated with icv injection of aCSF or FGF19. Mice were fasted for 18 h, treated twice with icv injection of aCSF or 2 μg of FGF19 (starting 12 h after fasting and 3 h apart) and received 8 U of insulin/kg of BW, 3 h after the last icv injection. Animals were sacrificed 8 min after insulin injection and tissues were collected for western blotting, (n=4 per group). Data for saline injections were arbitrary set to 1 and data after insulin injections were expressed relative to the value for saline injection for aCSF and FGF19 conditions. ^*P<0.05, ^**P<0.005, and ^***P<0.0005.

Figure 4

Central ERK1/2 signaling is involved in FGF19-mediated improvement in glucose metabolism. (A) Representative sections showing pERK1/2 (red) and co-localization between Npy-hrGFP positive neurons (green) and p-ERK1/2 positive cells (red) in ARC of HFD-fed Npy-hrGFP males 90 min after icv injection of vehicle+FGF19 and U0126+FGF19 (8–10 sections from 4 to 5 mice were examined in each group). (B) GTT with 2 g of glucose/kg of body weight (BW) was performed on HFD-fed males after 6 h of fast. Mice were treated twice with icv injection of vehicle+aCSF or U0126+aCSF (starting when food was removed and 3 h apart). GTT was performed 2.5 h after the last icv (n=4 per group). (C) GTT with 2 g of glucose/kg of BW was performed on HFD-fed males after 6 h of fast. Mice were treated twice with icv injection of vehicle+FGF19 or U0126+FGF19 (starting when food was removed and 3 h apart). GTT was performed 2.5 h after the last icv injection (n=8 per group). ^*P<0.05 and ^***P<0.005. Scale bars: 100 μm.

Figure 5

Central FGF19 inhibits fasting-induced activation of AGRP/NPY neurons. c-FOS immunostaining in fasted (A) HFD-fed and (B) ob/ob males, 90 min after the fourth FGF19 administration in the third ventricle (12 sections in ARC from 4 mice in each group were examined). (C) Analysis of co-localization between Npy-hrGFP positive neurons (green) and c-FOS positive cells (red) in HFD-fed Npy-hrGFP transgenic mice treated with icv injection of aCSF or FGF19 for 4 days. Animals were perfused 90 min after the last icv injection. (D) Quantification of Npy-hrGFP positive neurons (green) that colocalized with c-FOS staining (red) in ARC of HFD-fed Npy-hrGFP mice (9 sections in ARC from 3 mice in each group were examined). (E) Effect of FGF19 icv injection on Agrp and Npy gene expression in MBH of 24 h fasted wt mice treated with aCSF or FGF19 at the onset of fast and 12 h later. Fed and fasted mice were sacrificed 12 h after the last icv injection and MBH were collected for quantitative RT-PCR analysis. Data were normalized with -ctin expression level and expressed relative to the value for fed condition, which was arbitrarily set at 100. (F) GTT with 1 g of glucose/kg of body weight (n=4 per group) and (G) ITT with 3 U of insulin/kg of body weight (n=5 per group) were performed on A^y males fasted for 6 h and that had received 4 daily icv injections of aCSF or FGF19. The last icv injection of aCSF or FGF19 was performed 3 h before the GTT and ITT assays. (H) Insulinemia was measured before and 15 min into the GTT in A^y males in aCSF and FGF19 treated group (n=4 per group). ^*P<0.05, ^**P<0.005, and ^***P<0.0005. Scale bars: 100 μm.