Loss of hypothalamic Furin affects POMC to proACTH cleavage and feeding behavior in high-fat diet-fed mice

Abstract

Objective: The hypothalamus regulates feeding and glucose homeostasis through the balanced action of different neuropeptides, which are cleaved and activated by the proprotein convertases PC1/3 and PC2. However, the recent association of polymorphisms in the proprotein convertase FURIN with type 2 diabetes, metabolic syndrome, and obesity, prompted us to investigate the role of FURIN in hypothalamic neurons controlling glucose and feeding.

Methods: POMC-Cre^+/- mice were bred with Furin^fl/fl mice to generate conditional knockout mice with Furin-deletion in neurons expressing proopiomelanocortin (POMCFurKO), and Furin^fl/fl mice were used as controls. POMCFurKO and controls were periodically monitored on both normal chow diet and high fat diet (HFD) for body weight and glucose tolerance by established in-vivo procedures. Food intake was measured in HFD-fed FurKO and controls. Hypothalamic Pomc mRNA was measured by RT-qPCR. ELISAs quantified POMC protein and resulting peptides in the hypothalamic extracts of POMCFurKO mice and controls. The in-vitro processing of POMC was studied by biochemical techniques in HEK293T and CHO cell lines lacking FURIN.

Results: In control mice, Furin mRNA levels were significantly upregulated on HFD feeding, suggesting an increased demand for FURIN activity in obesogenic conditions. Under these conditions, the POMCFurKO mice were hyperphagic and had increased body weight compared to Furin^fl/fl mice. Moreover, protein levels of POMC were elevated and ACTH concentrations markedly reduced. Also, the ratio of α-MSH/POMC was decreased in POMCFurKO mice compared to controls. This indicates that POMC processing was significantly reduced in the hypothalami of POMCFurKO mice, highlighting for the first time the involvement of FURIN in the cleavage of POMC. Importantly, we found that in vitro, the first stage in processing where POMC is cleaved into proACTH was achieved by FURIN but not by PC1/3 or the other proprotein convertases in cell lines lacking a regulated secretory pathway.

Conclusions: These results suggest that FURIN processes POMC into proACTH before sorting into the regulated secretory pathway, challenging the dogma that PC1/3 and PC2 are the only convertases responsible for POMC cleavage. Furthermore, its deletion affects feeding behaviors under obesogenic conditions.

Keywords: ACTH; FURIN; Hypothalamus; Obesity; POMC; proACTH.

Figure 4

POMC is selectively cleaved by FURIN at the C-terminal site of ACTH (KR164↓) into proACTH and β-lipotropin (LPH) in ΔFurHEK293T cells. Western blot analysis of POMC and each co-transfected PCs in cell lysate (A–B) and proACTH and ACTH in medium (C) of Furin-deficient HEK293T cells co-transfected with POMC and different PCs. The blot in A (upper panel) is labeled with an anti-myc antibody. The lower panel shows a ponceau S-staining of the cell lysates as loading reference. (B) Western blot analysis of each transfected PC detected either in the medium or in the cell lysate. As negative control was used medium or cell lysate from ΔFurHEK293T transfected with a different PC. The blot in C is labeled with the A2A3 antibody directed against the free carboxy-terminus of ACTH. One minute-exposure of the entire blot (upper panel), and 5 min-exposure of the lower part of the blot (lower panel) to show the less abundant low MW ACTH forms. Mouse pituitary protein extract (25 μg) was used as positive control (first lane). (D) Metabolic labeling of Furin-deficient HEK293T cells transfected with POMC alone or together with FURIN (40 min pulse). (E) New model for POMC processing in hypothalamus. The antibody used to detect POMC-myc is depicted in blue, and in orange the antibody used to detect proACTH and ACTH (glycosylated: gACTH, and mature: mACTH). In ΔFurHEK293T cells lacking the secretory pathway, FURIN preferentially cleaves the first POMC cleavage site, most likely in the TGN. The same cleavages are mediated by PC1/3 in cells with a regulated secretory pathway probably requiring the pH conditions of the ISGs. Cleavage of proACTH to ACTH is performed by PC1/3, the enzymes providing redundancy are probably FURIN and/or PC2. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Figure 1

Furin and Pomc mRNA levels are increased by a chronic high-fat-diet in Fur^fl/flcontrol mice. Relative mRNA expression in hypothalami of 23-week-old male Fur^fl/fl mice fed either a NCD or a HFD for 15 weeks. The mice were in a fed state at the moment of the euthanasia (A) Furin, (B) Pomc, (C) Agrp, (D) Npy. n = 3–6 mice/group. ∗∗P < 0.01, ∗∗∗P < 0.001, determined by unpaired t test with Welch's correction. All the values were normalized to Gapdh expression and data are represented as mean ± SEM. NCD: normal chow diet; HFD: high fat diet.

Figure 2

The absence of Furin in POMC-neurons leads to hyperphagia and increased body weight after a short HFD period. (A) Schematic of the breeding strategy used to generate the POMCFurKO mice. Female mice with the exon 2 of the Furin gene flanked by two loxP sites (Fur^fl/fl) were bred with male mice expressing the Cre recombinase under the control of the POMC promoter (POMC-Cre). Fur^fl/fl are used as controls; POMC-Cre^+/−, Fur^fl/fl mice are referred to as POMCFurKO. (B) Body weight of random NCD-fed male POMCFurKO and Fur^fl/fl mice of 10, 25, 30 weeks of age (n = 5–9 mice/group). No significant differences were observed by one-way ANOVA with Sidak's multiple comparisons test. (C) Body weight of 10, 12, and 18-week-old mice after 2, 4, and 10 weeks of HFD respectively (n = 4–9 mice/group). ∗P < 0.01; ∗P < 0.05 determined by one-way ANOVA with Sidak's multiple comparisons test. (D) Daily food intake (g) of 11-week-old male POMCFurKO and Fur^fl/fl mice on HFD for 3 weeks. Data are represented as the average over three days (n = 4–8 mice/group). ∗P < 0.05 determined by unpaired t-test. (E) IPGTT on 25-week-old POMCFurKO and Fur^fl/fl male mice on NCD after overnight fasting (n = 5–6 mice/group). (F) The area under the curve in E was expressed as g/dL/120 min. (G) IPGTT on 30-week-old POMCFurKO and Fur^fl/fl male mice on NCD after overnight fasting (n = 7 mice/group). ∗P < 0.05 determined by two-way ANOVA with Sidak's multiple comparisons test. (G) The area under the curve in G was expressed as g/dL/120 min. (G) IPGTT on 12-week-old POMCFurKO and Fur^fl/fl male mice on HFD for 4 weeks after overnight fasting (n = 3 mice/group). (H) The area under the curve in G was expressed as g/dL/120 min. No significant differences were observed by unpaired t-test. NCD: normal chow diet; HFD: high fat diet.

Figure 3

POMC expression and processing is altered in HFD-fed POMCFurKO mice. (A) Relative mRNA expression of Npy, Agrp, and Pomc in the hypothalami of 10-week-old male POMCFurKO (pink bars) and Fur^fl/fl (blue bars) mice on HFD for 2 weeks (n = 4–5 mice/group). ∗∗P < 0.01 determined by unpaired t-test. All data are represented as mean ± SEM. The total content of POMC (B), ACTH (C), αMSH (D), and the ratio ACTH/POMC (E), αMSH/POMC (F), and ACTH/αMSH (G) in the hypothalami of 11-week-old male POMCFurKO and Fur^fl/fl mice fed on HFD for 3 weeks (n = 4–6 mice/group). ∗∗P < 0.01; ∗∗∗P < 0.001 determined by unpaired t-test. For RT-qPCR analyses, the values were normalized to Gapdh expression levels; all data are represented as mean ± SEM. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)