Irisin and FGF21 are cold-induced endocrine activators of brown fat function in humans

Abstract

Rediscovery of cold-activated brown adipose tissue (BAT) in humans has boosted research interest in identifying BAT activators for metabolic benefits. Of particular interest are cytokines capable of fat browning. Irisin, derived from FNDC5, is an exercise-induced myokine that drives brown-fat-like thermogenesis in murine white fat. Here we explored whether cold exposure is an afferent signal for irisin secretion in humans and compared it with FGF21, a brown adipokine in rodents. Cold exposure increased circulating irisin and FGF21. We found an induction of irisin secretion proportional to shivering intensity, in magnitude similar to exercise-stimulated secretion. FNDC5 and/or FGF21 treatment upregulated human adipocyte brown fat gene/protein expression and thermogenesis in a depot-specific manner. These results suggest exercise-induced irisin secretion could have evolved from shivering-related muscle contraction, serving to augment brown fat thermogenesis in concert with FGF21. Irisin-mediated muscle-adipose crosstalk may represent a thermogenic, cold-activated endocrine axis that is exploitable in obesity therapeutics development.

Copyright © 2014 Elsevier Inc. All rights reserved.

Figures

Figure 1. Validation of immunoblot-detected irisin by mass spectrometry

Immunoblot of paired serum samples following albumin/immunoglobulin depletion, against anti-FNDC5 antibody revealed multiple distinct bands (a-f) [Panel A]. PNGase treatment reduced size of band e (~32 kDa) to band f (~24 kDa). Panel B showed the amino acid sequence of full length FNDC5 with the secreted irisin segment underlined. Mass spectrometry analysis of all bands (a-f) identified a specific peptide (in red), unique to irisin, only in band e and band f, with molecular weights matching those of glycosylated and deglycosylated irisin, respectively. Panel C showed representative immunoblots of serum irisin for fold change quantification from 2 subjects during cold exposure, maximal exercise and sub-maximal exercise. Subject 1 shivered during cold exposure while subject 2 did not. Accordingly, deglycosylated irisin band (~24 kDa) was stronger at the end of cold exposure only in Subject 1. In contrast, irisin band was stronger after submaximal exercise in both subjects. Full sized blots are shown in Figure S1. Panels D-F are graphical representation of serum irisin fold changes during cold exposure, maximal and sub-maximal exercise tests, respectively, of all 10 subjects. “Post” indicates the average band intensity of irisin extracted from the mid- and final blood samples of each clinical test. Similar results were obtained when analysis was conducted comparing irisin band intensity between baseline and final sample alone. Irisin level rose significantly following sub-maximal exercise [Panel F], and trended higher (p=0.07) after maximal exercise [Panel E]. Irisin levels increased only in the 7 subjects who shivered [closed circles, Panel D], but not those who did not [open circles, Panel D]. Panel G demonstrates neutralization of anti-FNDC5 antibody by FNDC5 recombinant protein. FNDC5 antibody mixture in increasing ratio resulted in quenching of western signal in a dose-dependent manner by excess FNDC5 recombinant protein. *P

Figure 2. Relationship between irisin, FGF21, BAT and temperature

Cold exposure resulted in reduction in skin temperature [open circles, Panel A], accompanied by a rise in energy expenditure (EE) [closed circles], both reaching significance from 16-12°C. Panels B (by immunoblot) and E (by ELISA) showed positive associations between irisin and EMG fold changes during cold exposure. Panel C compared irisin and EE fold changes during cold exposure with maximal exercise test. Changes in FGF21 levels correlated negatively with shivering [Panel D] but positively with thermogenic response (TR; difference between supraclavicular skin and chest T°C) [Panel F]. Panels G-H showed representative PET-CT images of BAT positive (N=3) and negative (N=2) individuals, respectively (BAT in red). FGF21 diurnal reduction was more markedly blunted in BAT positive (solid lines) compared to negative (dashed lines) individuals at 19°C [Panel L] vs. 24°C [Panel K]. Panels I-J compared FGF21 changes (N=5) measured between 8-10 am at either warm (27°C) or shivering (12°C) conditions. FGF21 reduction was significantly blunted in the cold. *P

Figure 3. Effects of FNDC5 and/or FGF21 treatment on gene/protein expression and bioenergenetics of neck adipocytes

Panel A showed effects of FGF21 and/or FNDC5 treatment on BAT/beige/white gene markers in neck adipocytes (N=6). UCP1 protein was absent in PBS-treated adipocytes, but was detected following FGF21 and/or FNDC5 treatment [Panel B]. UCP1 protein was highest in adipocytes treated with dual FGF21/FNDC5. Neck adipocytes displayed multi-lobulated lipid droplets (40x), similar before and after treatment [Panel C, F+F=FNDC5+FGF21 treatment]. Expression of FABP4, a general adipogenic gene, was not different following treatment [Panel D]. Induction of UCP1 was accompanied by up-regulation of basal [Panel F], oligomycin-insensitive, maximal uncoupled [Panel G] and norepinephrine-induced [Panel E] oxygen consumption, most robust in dual FGF21/FNDC5-treated adipocytes (N=4). Panel H showed infrared thermographic images of adipocytes in microplates treated with PBS, FGF21 and/or FNDC5 (N=4). The temperature scale showed color representation of temperature variation. Heat production was increased in the basal state by FNDC5- but not FGF21. Addition of norepinephrine (NE) increased heat production in increasing magnitude in FGF21-, FNDC5- and dual FGF21/FNDC5-treated adipocytes. These results are displayed in graphical format in Panel I. *p<0.05 compared to PBS, #p<0.05 compared to FNDC5- or FGF21-treated adipocytes and ψp<0.05 compared to FGF21-treated adipocytes. Data are presented as mean ± SD.