Skip to main content
NCBI home page
Endocrinology logoLink to Endocrinology
. 2014 Feb 24;155(5):2031–2040. doi: 10.1210/en.2013-1902

Fibroblast GroWTh Factor 21 and Thyroid Hormone Show Mutual Regulatory Dependency but Have Independent Actions In Vivo

Eleni M Domouzoglou 1, ffolliott Martin Fisher 1, Inna Astapova 1, Elliott C Fox 1, Alexei Kharitonenkov 1, Jeffrey S Flier 1, Anthony N Hollenberg 1, Eleftheria Maratos-Flier 1,✉
PMCID: PMC3990851  PMID: 24564398

Abstract

Thyroid hormone (TH) regulates fibroblast growth factor 21 (FGF21) levels in the liver and in the adipose tissue. In contrast, peripheral FGF21 administration leads to decreased circulating levels of TH. These data suggest that FGF21 and TH could interact to regulate metabolism. In the present study, we confirmed that TH regulates adipose and hepatic FGF21 expression and serum levels in mice. We next investigated the influence of TH administration on key serum metabolites, gene expression in the liver and brown adipose tissue, and energy expenditure in FGF21 knockout mice. Surprisingly, we did not observe any significant differences in the effects of TH on FGF21 knockout mice compared with those in wild-type animals, indicating that TH acts independently of FGF21 for the specific outcomes studied. Furthermore, exogenous FGF21 administration to hypothyroid mice led to similar changes in serum and liver lipid metabolites and gene expression in both hypothyroid and euthyroid mice. Thus, it appears that FGF21 and TH have similar actions to decrease serum and liver lipids despite having some divergent regulatory effects. Whereas TH leads to up-regulation in the liver and down-regulation in brown adipose tissue of genes involved in the lipid synthesis pathway (eg, fatty acid synthase (FASN) and SPOT14), FGF21 leads to the opposite changes in expression of these genes. In conclusion, TH and FGF21 act independently on the outcomes studied, despite their ability to regulate each other's circulating levels. Thus, TH and FGF21 may modulate the availability of each other in critical metabolic states.

Fibroblast growth factor 21 (FGF21) is a critical factor regulating multiple metabolic pathways. In the liver, FGF21 regulates fatty acid oxidation both during fasting and in mice consuming ketogenic diets, whereas in adipose tissue FGF21 mediates increased glucose uptake through glucose transporter 1 (1, 2). FGF21 acts directly on the liver to stimulate phosphorylation of fibroblast growth factor receptor substrate 2 and ERK1/2 in a peroxisome proliferator–activated receptor γ coactivator-1α (PGC1-α)–independent manner (3). More recently, FGF21 has been shown to play a role in brown adipose tissue (BAT)–induced thermogenesis and to also mimic the effects of cold exposure in brown inguinal adipose tissue (4–6). FGF21 overexpression and its systemic administration lead to weight loss and improved glucose tolerance (1, 7, 8). In contrast, animals lacking FGF21 demonstrate late-onset obesity. Furthermore, these mice have an abnormal response to consumption of a ketogenic diet, gaining rather than losing weight (9). In humans, increased FGF21 serum concentrations correlate with increased body mass index and obesity and may be a marker of nonalcoholic fatty liver disease, making FGF21 a likely biomarker for the progression of this condition (10–13).

We showed previously in mice that T3 markedly increases hepatic FGF21 mRNA expression (14), an effect that is dependent on thyroid hormone receptor β, retinoid X receptor, and peroxisome proliferator–activated receptor α. Conversely, FGF21 treatment is associated with a decrease in serum levels of T3 and T4 in diet-induced obese mice (7). Because T3 is well known to increase the metabolic rate and exogenous administration of FGF21 increases energy expenditure in diet-induced obese mice (7, 8, 15–17), we speculated that some of the effects of T3 may be mediated by the induction of FGF21 expression. Furthermore, we also hypothesized that the effects of FGF21 to increase energy expenditure could depend on normal thyroid function. Therefore, to better understand the potential interrelationship of FGF21 and thyroid hormone (TH) signaling we performed a series of parallel experiments and evaluated the effects of TH in mice lacking FGF21 and the effect of FGF21 in hypothyroid mice. Herein we demonstrate that TH exhibited the expected range of actions in mice lacking FGF21. In addition, FGF21 action was preserved in our model of methimazole-induced hypothyroidism. Thus, although both hormones promote increased energy expenditure and lipid utilization, their actions were not linked (7–9).

Materials and Methods

Mouse maintenance

All animal experimental protocols were approved by the Beth Israel Deaconess Medical Center Institutional Animal Care and Use Committee. All mice were maintained in a temperature-controlled environment at roughly 24°C under a 12-hour light (6:00 am–6:00 pm) and 12-hour dark cycle with ad libitum access to food and water.

Recombinant FGF21 protein

Recombinant FGF21 protein was generated by expression of human FGF21 in Escherichia coli and was subsequently refolded in vitro as described previously (1).

Mouse treatments

Mice were fed a standard chow diet (LabDiet 5008; LabDiet). Mice were made hypothyroid by adding 0.1% methimazole (Sigma-Aldrich) and 0.5% wt/vol Splenda (sucralose) to their drinking water. Control euthyroid mice were kept on Splenda. TH replacement was given and hyperthyroidism was induced by ip injections of 0.05 or 0.5 μg/g T4 (Sigma-Aldrich), respectively, for 3 consecutive days to hypothyroid mice. In the energy expenditure assessments, the hyperthyroid state was achieved by implantation of T3 slow-release pellets (0.5 mg/pellet; Innovative Research of America) in normal untreated animals for 4 days to minimize mouse handling. FGF21 (0.7 mg/h/mouse) was administered by sc implantation of 3-day-release mini-osmotic pumps (Durect; Alzet). All of our TH treatments were planned and chosen on the basis of previously published experiments that used similar models for studying TH states in rodents, taking into account results from our experiments and the treatment duration (15, 18, 19).

Experimental procedures

TH administration in wild-type (WT) and FGF21-KO mice

A total of 30 animals, 15 FGF21 knockout (FGF21-KO) and 15 WT were used for this experimental procedure, divided in subgroups as follows. Fifteen 16-week-old FGF21-KO female mice were treated with methimazole for 3 weeks to become hypothyroid. Subsequently, while all mice were receiving methimazole treatment, 5 of them received a high dose (0.5 μg/g) of T4 for 3 consecutive days to become hyperthyroid, 5 of them received a low dose (0.05 μg/g) of T4 for 3 days to become euthyroid, and the remaining 5 received vehicle (PBS) for 3 days to remain hypothyroid. Fifteen age-matched C57/BL6 females (WT) were held as controls and received the same treatments: 3 days of a high dose of T4 to create a hyperthyroid group, 3 days of a low dose of T4 to create a euthyroid group, and 3 days of vehicle (PBS) for the hypothyroid group. At the end of the experiments, all mice were killed, and tissues were collected, flash-frozen, and stored at −80°C until further analysis.

Energy expenditure assessment

To investigate energy expenditure, a total of 10 animals, 5 FGF21-KO and 5 WT female mice (12 weeks of age), were continuously monitored while singly housed in the comprehensive laboratory animal monitoring system (CLAMS) (Oxymax; Columbus Instruments). Measurements of locomotor activity, oxygen consumption (VO) and carbon dioxide emission as well as food intake were acquired in real time. The mice were acclimated 48 hours before the initiation of the experiment, and basal data were acquired during the first 2 days of the experiment. Subsequently, T3 slow-release pellets were implanted in the morning, and immediately thereafter, the mice were placed back in the CLAMS apparatus to continue monitoring. Data were collected for the next 4 days. At the end of the experiment, all mice were killed, and tissues were collected, flash-frozen, and stored at −80°C until further analysis.

FGF21 administration in hypothyroid mice

Thirty-two C57/BL6 12-week-old male mice were acquired from The Jackson Laboratory and fed the standard chow diet ad libitum for 4 weeks in our animal facility. Subsequently, half of the mice received methimazole treatment in their drinking water for 4 weeks to induce hypothyroidism. After this period, half of the mice in each subgroup, 8 of the 16 mice in each subgroup for a total of 16 mice, received FGF21 treatment for 3 days. At the end of the 3 days, mice were killed, and tissues were collected, flash-frozen, and stored at −80°C until further analysis.

Body composition analysis

Whole-body fat mass, lean tissue mass, free water, and total body water in our live ad libitum–fed animals were measured by EchoMRI 3-in-1 quantitative magnetic resonance (Echo Medical Systems, LLC).

Biochemical tissue analysis

Liver lipids were extracted by the method described by Folch et al (20) with modifications. In brief, 150 mg of frozen liver was homogenized in 1 mL of extracting solution containing chloroform and methanol (2:1). Then 3 mL more of extracting solution was added, and the mixture was incubated at room temperature overnight with shaking. The samples were washed with 0.9% NaCl and centrifuged at 1000 × g for 5 minutes. The lower phase was aspirated and dried overnight. The lipids were dissolved in a solution of a final volume of 5 times the initial weight of the livers containing 60% butanol, 26.67% Triton X-100, and 13.3% methanol. The lipids were measured with photometric assays (Stanbio Laboratory).

Serum analysis

Photometric assays were performed in duplicate for the measurement of serum glucose, triglycerides, and cholesterol (Stanbio Laboratory). A mouse ELISA (Quantikine; R&D Systems) was used for FGF21 serum levels and an RIA kit (TKT41 and TKT31; Siemens Health Care Diagnostics) for the assessment of serum T4 and T3.

RNA purification and quantitative real-time PCR

RNA from frozen liver, interscapular BAT, and muscle samples was purified using the RNeasy tissue mini kit with RNeasy spin columns and QIAzol lysis reagent (QIAGEN). Then 0.5 μg of the total RNA was reverse transcribed using oligo(dt) and random hexamer primers and Moloney murine leukemia virus reverse transcriptase (QuantiTech RT for PCR; QIAGEN). Quantitative PCR was performed in duplicate samples using the 7900HT PCR system (Applied Biosystems) and SYBR Green Master Mix (Applied Biosystems). The expression level of each gene was normalized to 36B4 expression (housekeeping control gene) and is represented as fold change relative to a reference group. Real-time PCR Miner software (http://www.miner.ewindup.info) (21) was used to calculate the mean PCR amplification efficiency for each gene. The primer sequences are presented in Supplemental Table 1 published on The Endocrine Society's Journals Online web site at http://end.endojournals.org.

Western blot analysis

Protein from frozen liver samples was extracted using lysis buffer for radioimmunoprecipitation assay with addition of a Complete Mini Protease Inhibitor Cocktail (Roche) and phosphatase inhibitors. Protein levels were measured with a Bradford assay (Bio-Rad Laboratories). Proteins were resolved by SDS-PAGE on a 4% to 15% Criterion Tris/HCl gel (Bio-Rad Laboratories) and transferred onto nitrocellulose (Protran; Schleicher & Schuell). Incubation with β-actin or fatty acid synthase (FASN) (Abcam) secondary antibody was used, and films were developed using SuperSignal West Pico chemiluminescent reagent (Pierce; Thermo Scientific).

Statistics

Data are presented as means ± SEM. The Anderson-Darling test for normality was performed on all observational groups. When sample normality was justified, statistical comparisons between 2 groups were performed using the unpaired t test, whereas for 3-group comparisons (PBS vs low-dose-treated vs high-dose-treated) one-way ANOVA was performed. When normality could not be supported, the nonparametric 2-sample Mann-Whitney test and the Kruskal-Wallis test for 3-group comparisons were used. When 3 groups were compared, post hoc tests were performed if the overall analysis (ANOVA or Kruskal-Wallis) revealed significance to determine which experimental groups differed from each other; P values were adjusted for multiple comparisons using the Bonferroni correction. To investigate the differential effect of (1) FGF21 administration in the hypothyroid vs the euthyroid state and (2) the thyroid dose (PBS vs low-dose-treated vs high-dose-treated) after T4 administration in WT vs FGF21-KO hypothyroid mice, we assessed the interaction between thyroid state and FGF21 administration or thyroid dose and mouse genotype, respectively, using two-way ANOVA, where appropriate. In the energy expenditure experiment, differences between WT and FGF21-KO mice in the measurements of the energy expenditure parameters at baseline and on each of the 4 days after T3 treatment initiation were investigated using repeated-measures ANOVA. All tests were two-tailed, and differences were considered to be statistically significant at P < .05. For the statistical analysis, GraphPad Prism 5.00 (GraphPad Software Inc) was used.

Results

FGF21 regulation during chronic T4 administration

The acute administration of TH in WT animals (ie, harvesting tissues a maximum of 6 hours after 1 injection of TH) has been shown previously to regulate FGF21 gene expression in the liver and in adipose tissue (14). Consequently, to further investigate the relationship between TH and FGF21, we used a model of chronic T4 administration (ie, harvesting tissues after 3 consecutive days of TH injections) at 2 doses (high of 0.5 μg/g and low of 0.05μg/g) in hypothyroid WT and FGF21-KO mice. Importantly, serum FGF21 concentrations were significantly elevated in the high T4 dose–treated hypothyroid mice (1305.84 pg/mL for the high T4 dose vs 192.90 pg/mL for the low T4 dose vs 141.14 pg/mL for PBS) (Figure 1A), demonstrating a significant increase in high-dose–treated mice compared with that in either PBS-treated or low-dose-treated mice (P = .027 for both comparisons) (Figure 1A). Chronic administration of T4 also led to trending up-regulation of hepatic FGF21 mRNA expression and down-regulation of BAT FGF21 mRNA expression, confirming the previously reported effect during acute administration (P = .094 and P = .018 for any difference in liver and BAT FGF21 expression among the 3 groups, respectively) (Figure 1, B and C).

Figure 1.

Figure 1.

Open in a new tab

FGF21 regulation under chronic T4 administration in WT mice. A, Increase in serum FGF21 is dependant on T4 dosage (overall difference among the 3 groups, P = .007). B, Up-regulation of liver FGF21 expression (overall difference among the 3 groups, P = .094). C, Dose-dependent down-regulation of BAT FGF21 expression (overall difference among the 3 groups, P = .018). Error bars indicate SEM. *, P < .05.

T4 effects on serum metabolites and the expression of genes in the liver and BAT are preserved in FGF21-KO mice

To demonstrate the effects of the 2 doses of T4, we measured serum T4 and T3 levels as well as liver deiodinase iodothyronine type 1 (DIO1) expression in WT mice treated with PBS or replaced with low and high T4 (hypothyroid, euthyroid, and hyperthyroid, respectively). As shown in Figure 2, there was an increase in serum TH levels and hepatic DIO1 gene expression in WT mice. Moreover, T4 administration decreased serum cholesterol as expected and produced a pattern of gene regulation in liver and BAT in WT mice typical for TH action (Figure 3, A–C, and Table 1). In FGF21-KO mice, we initially asked whether the conversion of T4 to T3 was altered, given the ability of FGF21 to regulate TH circulating levels. As shown in Figure 2B, T3 serum levels were increased in the FGF21-KO mice, demonstrating that T4 conversion to T3 was intact (Figure 2, A and B). T3 serum levels are in fact higher in the FGF21-KO mice than in the WT mice (P = .049 for the interaction between thyroid state and mouse genotype). Furthermore, a higher level of hepatic DIO1 up-regulation was observed in the high-dose T4-treated FGF21-KO mice than in the respective WT mice (9.29-fold vs 4.70-fold increase; P = .042 for the interaction between thyroid dose and mouse genotype) (Figure 2C), which is consistent with the observed higher level of serum T3 in the high-dose T4-treated FGF21-KO mice. Overall, we did not observe any differences between hypothyroid PBS-treated, euthyroid low-dose T4-treated, and hyperthyroid high-dose T4-treated mice at the regulatory level in serum lipids or hepatic and BAT genes. Serum cholesterol concentrations were decreased dose dependently in T4-treated FGF21-KO mice similar to those in WT animals (25.8% decrease for high-dose T4-treated WT vs 42.8% decrease for high-dose T4-treated FGF21-KO mice compared with that for the respective PBS group) (Figure 3C). Target gene regulation was also similar between the 2 genotypes (Table 1). Liver malic enzyme, which is typically up-regulated by T4, was similarly increased in both mouse genotypes (2.35-fold increase for high-dose T4-treated WT vs 2.80-fold increase for high-dose T4-treated FGF21-KO mice compared with that for the respective PBS group) (Table 1). In BAT, FASN expression was down-regulated after T4 administration in both WT and FGF21-KO mice (Table 1).

Figure 2.

Figure 2.

Open in a new tab

Conversion of serum T4 to serum T3 in FGF21-KO and WT Mice. A, Increase in serum T4 levels after exogenous T4 administration in both genotypes. B, Increase in T3 serum levels after conversion of the exogenously administered T4 in both WT and FGF21-KO mice. FGF21-KO mice show higher T3 serum levels than their WT counterparts (P = .049 for the interaction between thyroid state and mouse genotype). C, Up-regulation of liver DIO1 expression by T4 in WT and FGF21-KO mice. There was a higher degree of up-regulation in FGF21-KO than in WT mice (P = .042 for the interaction between thyroid state and mouse genotype), explaining the higher T3 levels in FGF21-KO mice compared with those in the respective WT group. Error bars indicate SEM. **, P ≤ .01; *, P < .05 for significant differences compared with the respective control group (Hypothyroid+PBS), #, P < .05 for the differential effect (interaction) between thyroid state and mouse genotype.

Figure 3.

Figure 3.

Open in a new tab

T4 effects in FGF21-KO and WT mice. A, Normalized liver weight (liver weight [milligrams]/body weight [grams]) remained unchanged after T4 treatment in both genotypes, B, Total liver cholesterol and triglycerides showed a mild increase in the WT mice. C, Serum cholesterol was decreased dose dependently by T4 in both WT and FGF21-KO mice, whereas serum triglycerides remained unchanged in both genotypes. Error bars indicate SEM. **, P ≤ .01; *, P < .05 for significant differences compared with the respective control group (Hypothyroid+PBS).

Table 1.

Effect of T4 Treatment in WT vs FGF21-KO Mice

Tissue Gene Fold Change in Low and High T4 Doses vs PBS
 P for Interaction of Thyroid Dose and Genotype
WT
 FGF21-KO
PBS Low T4-Treated Euthyroid High T4-Treated Hyperthyroid PBS Low T4-Treated Euthyroid High T4-Treated Hyperthyroid
Liver FASN 1 ± 0.10 3.28 ± 1.37 2.52 ± 0.67 1.95 ± 0.66 1.05 ± 0.22 1.73 ± 0.25 .284
PEPCK 1 ± 0.17 0.75 ± 0.22 1.15 ± 0.09 1.09 ± 0.28 1.26 ± 0.35 1.97 ± 0.53 .353
SPOT14 1 ± 0.11 5.45 ± 1.05 6.42 ± 1.44a 3.10 ± 1.02 1.79 ± 0.46 4.76 ± 1.93 .306
SREBP-1c 1 ± 0.08 0.85 ± 0.34 0.24 ± 0.09 1.02 ± 0.30 0.63 ± 0.14 0.41 ± 0.07 .642
ME 1 ± 0.13 2.10 ± 0.42 2.35 ± 0.12a 1.09 ± 0.36 1.14 ± 0.11 2.80 ± 0.29a .119
BAT DIO2 1 ± 0.47 0.07 ± 0.01 0.08 ± 0.02 0.64 ± 0.12 0.29 ± 0.05a 0.46 ± 0.24 .377
UCP1 1 ± 0.33 0.51 ± 0.04 0.61 ± 0.18 0.84 ± 0.12 0.62 ± 0.12 0.82 ± 0.09 .790
FASN 1 ± 0.29 0.41 ± 0.07 0.07 ± 0.01a 0.99 ± 0.44 0.26 ± 0.08a 0.20 ± 0.04a .679
Open in a new tab

Data are averages ± SEM.

a

Significant fold changes within each genotype (P < .05).

b

Significant fold changes within each genotype (P < .01).

FGF21-KO mice respond normally to T3 administration

Because increased energy expenditure is a hallmark of TH action (17), we next asked whether FGF21 is necessary for this T3 action. To do this, we implanted slow-release T3 pellets in FGF21-KO and WT female mice and monitored them in a CLAMS apparatus. The 2 groups were matched per weight, fat mass and lean mass before the experiment was started (mean body weight, fat mass, and lean mass values of 21.62 g, 10.24%, and 71.38% for the WT and 22.22 g 8.33%, and 73.25% for the FGF21-KO mice, respectively. P values for mean body weight, fat mass, and lean mass for the comparison between WT and FGF21-KO mice were P = .35, P = .33, and P = .27, respectively). Mean daily food intake was increased in WT mice after T3 administration (39% increase, P = .04). The T3 serum levels achieved with T3 pellets were 479 and 455 ng/dL for the WT and FGF21-KO groups, respectively. Spontaneous physical activity was similar in the WT and FGF21-KO mice during the light and dark cycles (Figure 4A). Furthermore, no difference was observed in oxygen consumption during T3 administration between FGF21-KO and WT animals (Figure 4B); both genotypes increased their oxygen consumption during the dark cycle after T3 pellet implantation (Figure 4C). PGC-1α protein content was shown in the past to be increased in muscles of T3-treated rats (22). The fibronectin type III domain containing 5 (Fndc5) gene was recently shown to be downstream to PGC-1α and encodes a newly identified hormone, irisin, which is thought to increase total body energy expenditure (23). The expression of irisin is enriched in red muscle such as soleus. Gene expression in soleus muscles of our T3-treated animals revealed an up-regulation of PGC-1α in both WT and FGF21-KO mice compared with that in untreated animals (Figure 4D). The expression of Fndc5 was also up-regulated to a similar extent in both WT and FGF21-KO mice after T3 treatment (Figure 4D).

Figure 4.

Figure 4.

Open in a new tab

Effect of chronic high-dose T3 administration on energy expenditure in WT and FGF21-KO mice. A, Monitoring of spontaneous physical activity during the 48 hours before initiation of T3 treatment (ie, basal activity) and during the 4 days after initiation of T3 treatment showed no significant difference between the 2 genotypes (light cycle, P = .17; dark cycle, P = .27). B, Monitoring of Vo2 48 hours before initiation of T3 treatment (ie, basal consumption) and 4 days after T3 treatment showed a similar response in both WT and FGF21-KO mice. C, WT and FGF21-KO animals respond to T3 treatment by increasing their Vo2 in the dark cycle compared with basal measurements. D, PGC-1α and Fndc5 are up-regulated after T3 treatment in soleus muscle. Error bars represent SEM. **, P ≤ .01; *, P < .05. Dark cycle of the light-dark cycle is indicated as the black horizontal bar.

Metabolic effects of FGF21 administration are preserved in a hypothyroid mouse model

To explore whether FGF21 action requires normal levels of TH, we assessed FGF21 action in hypothyroid mice. In this set of experiments, we used male mice because we have noticed throughout our studies that FGF21 and TH action are very similar in male and female animals. In all of our experiments, control animals and experimental animals were of the same sex. Peripheral administration of FGF21 led to a decrease in liver weight, cholesterol, and triglyceride content as well as in serum lipids in both euthyroid and hypothyroid mice (Figure 5, A–C). The hypothyroid saline-treated group had increased normalized liver weight and total liver cholesterol compared with those for the euthyroid saline-treated group (P = .0001 and P = .025, respectively) (Figure 5, A and B) and demonstrated a trend for a larger normalized liver weight decrease compared with that of euthyroid mice after FGF21 administration (decrease in normalized liver weight: 18.6% vs 13.8%, P = .060 for the interaction between thyroid state and FGF21 effect). FGF21 administration induced a decrease in serum cholesterol (15.4% decrease for hypothyroid [P = .056] vs 15.1% decrease for euthyroid [P = .018] compared with that for the respective saline-treated groups) (Figure 5C) and serum triglycerides to a similar extent in both groups (28.6% decrease for hypothyroid [P = .009] vs 28.2% decrease for euthyroid [P = .008] compared with those for the respective saline-treated groups) (Figure 5C). We also noticed that there was a decrease in TH concentrations after FGF21 administration in our euthyroid mice (Figure 6, A and B).

Figure 5.

Figure 5.

Open in a new tab

Effects of FGF21 treatment in methimazole-induced hypothyroid mice. A, Normalized liver weight (liver weight [milligrams]/body weight [grams]) was higher in hypothyroid saline-treated compared with euthyroid saline-treated mice (P < .001), and there was a trend toward a larger decrease in the hypothyroid group after FGF21 administration (decrease in liver weight, 18.6% in hypothyroid vs 13.8% in euthyroid mice [P = .060] for the interaction between thyroid state and FGF21 effect). B, Liver cholesterol content was increased in hypothyroid saline-treated vs euthyroid saline-treated mice (P = .025), and FGF21 administration led to a decrease in both groups. Liver triglycerides were decreased in both groups after FGF21 treatment. C, Serum cholesterol tended to be increased in the hypothyroid saline-treated group (P = .079) with a similar decrease in both euthyroid and hypothyroid mice after FGF21 administration. FGF21 led to an equal decrease in serum triglycerides in both groups. Error bars indicate SEM. **, P ≤ .01; *P < .05.

Figure 6.

Figure 6.

Open in a new tab

Serum TH decrease after FGF21 administration. A, Serum T4 decrease after FGF21 administration in euthyroid mice. B, Serum T3 after FGF21 administration in euthyroid mice. Error bars indicate SEM. *, P < .05.

The expression of most genes assayed in the liver and BAT was similarly regulated after FGF21 administration in hypothyroid and euthyroid mice. This result is displayed in Table 2, in which a significant interaction between FGF21 treatment and thyroid status was only shown for hepatic FASN (Table 2). Genes known to be induced by FGF21, such as uncoupling protein 1 (UCP1) in BAT, were up-regulated to a similar extent in both euthyroid and hypothyroid mice receiving FGF21 (for UCP1, a 1.63-fold increase in euthyroid vs 1.94-fold increase in hypothyroid mice) (Table 2). However, liver FASN expression was the only gene that showed a significant impairment during FGF21 regulation in the hypothyroid group, whereas there was also a partial impairment in liver SPOT14 expression that did not reach significance. There was a 0.45-fold change in hepatic FASN expression in the euthyroid FGF21 group, whereas the respective change in the hypothyroid FGF21 group was minimal (0.94-fold, P = .022 for the interaction between thyroid state and FGF21 effect) (Table 2). BAT FASN expression was 2.10-fold up-regulated in the euthyroid group, whereas there was only a 1.14-fold increase in the hypothyroid group without, however, this difference reaching statistical significance (P = .21 for the interaction). We analyzed hepatic FASN protein levels, which reproduced the results observed for gene expression without reaching statistical significance (0.30-fold change [P = .07] in the euthyroid FGF21 group vs 0.74-fold change [P = .34] in the hypothyroid FGF21 group) (Supplemental Figure 1).

Table 2.

Effect of FGF21 Administration on Liver and BAT Genes in Euthyroid vs Hypothyroid Mice

Tissue Gene FGF21 Effect in Euthyroid Mice
 FGF21 Effect in Hypothyroid Mice
 P forInteraction Between Thyroid State and FGF21 Effect
Saline FGF21 Fold Change Above Saline Saline FGF21 Fold Change Above Saline
Liver DIO1 1 ± 0.14 0.80 ± 0.17 0.80 ± 0.17 0.94 ± 0.12 0.82 ± 0.19 0.87 ± 0.21 .367
FASN 1 ± 0.10 0.45 ± 0.07 0.45 ± 0.07a 0.69 ± 0.14 0.65 ± 0.10 0.94 ± 0.15 .022
PEPCK 1 ± 0.20 1.70 ± 0.24 1.70 ± 0.24a 0.89 ± 0.13 1.58 ± 0.14 1.77 ± 0.16a .660
SPOT14 1 ± 0.25 0.22 ± 0.07 0.22 ± 0.07a 0.53 ± 0.13 0.18 ± 0.13 0.35 ± 0.08a .084
SREBP-1c 1 ± 0.12 0.25 ± 0.05 0.25 ± 0.05a 1.08 ± 0.20 0.60 ± 0.13 0.54 ± 0.13a .341
BAT DIO2 1 ± 0.06 2.13 ± 0.16 2.13 ± 0.16a 0.91 ± 0.05 1.27 ± 0.16 1.40 ± 0.17 .111
UCP1 1 ± 0.05 1.63 ± 0.12 1.63 ± 0.12a 0.99 ± 0.12 1.93 ± 0.12 1.94 ± 0.12 .436
FASN 1 ± 0.09 2.10 ± 0.24 2.10 ± 0.24a 3.30 ± 0.37 3.78 ± 0.16 1.14 ± 0.05 .210
Open in a new tab

Data are averages ± SEM. The effect of FGF21 on the regulation of liver and BAT genes is similar in the euthyroid and hypothyroid mice with the exception of FASN and liver SPOT14 presenting an interaction between the thyroid state and FGF21 effect (P = .022 and P = .084 for the interaction, respectively).

a

Significant fold changes within the euthyroid and hypothyroid group (P < .05).

Discussion

A relationship between FGF21 and TH is supported by evidence that acute T3 administration regulates FGF21 transcription in the liver and in adipose tissue, whereas peripherally administered FGF21 decreases serum TH concentrations (7, 14). In this study, we confirm that TH regulates FGF21 gene expression in liver and adipose tissue, and we also show that TH increases circulating FGF21 levels in vivo. Furthermore, we demonstrate that the decrease in adipose expression and the increased FGF21 serum levels induced by TH are sustained through the duration of treatment. There was a trend toward increased hepatic FGF21 expression. This trend derives from the assessment of a single sample and not from time course sampling as we did in our previous acute administration experiments, which could partially explain the reason we may have missed FGF21 expression. However, because serum FGF21 elevation may represent an accumulation over time, we speculate that the increase in serum FGF21 concentrations probably results from enhanced hepatic production of FGF21 as TH leads to decreased expression of FGF21 in BAT (14). This result demonstrates that the regulation of FGF21 by TH is tissue specific. Moreover, our data demonstrate that FGF21 is not required for TH effects on gene expression for the specific genes studied in either liver or BAT and that TH effects on energy expenditure are independent of FGF21.

Both the type 1 and type 2 iodothyronine deiodinases play a role in the conversion of T4 to T3 (24, 25). DIO1 has also been proven to be the enzyme responsible, at least in part, for high serum T3 concentrations in hyperthyroid patients (26). In our study, even though the key biological endpoints were the same, T3 concentrations were higher in FGF21-KO than in WT animals. FGF21-KO mice had higher serum T3 levels, consistent with the increased response in DIO1 expression in the livers of these mice. In agreement with the aforementioned findings, the higher concentrations of T3 in FGF21-KO mice suggest that the presence of FGF21 may limit T3 concentrations in the blood in hyperthyroidism. However, T3 plasma levels are not necessarily equivalent to tissue concentrations. Thus, the effect of FGF21 on serum T3 may not reflect its actual effects on tissue levels in the various specific tissues. In the absence of FGF21, TH had the same effect on gene expression in the liver and BAT, indicating that FGF21 is not required for the specific TH actions studied in these tissues. Based on the fact that the T3 levels achieved after T4 treatment differ in WT vs FGF21-KO animals, we administered T3 and not T4 to assess energy expenditure to achieve the same T3 levels in mice. Similarly, the increase in oxygen consumption with no change in spontaneous physical activity in both WT and FGF21-KO mice confirms that FGF21-KO mice have a normal response to chronic T3 administration in the context of energy expenditure at a systemic level. Moreover, we noticed an early increase in oxygen consumption in the FGF21-KO group of mice; however, there is no significant difference between the FGF21-KO and WT groups on the treatment days. At the regulatory level, we show for the first time that Fndc5 is a novel target of TH action in muscle, and this was effective in both mouse genotypes.

When we evaluated FGF21 treatment in WT euthyroid and hypothyroid male mice, we found that FGF21 was equally efficacious in euthyroid and hypothyroid mice with regard to decreased liver and serum lipids. The efficacy of FGF21 in hypothyroid animals was further confirmed by normal induction of genes known to be regulated by FGF21, such as UCP1 in BAT, that were similarly expressed in euthyroid and hypothyroid mice (3, 4). Interestingly, FASN gene expression in liver was the only example of a gene that demonstrated attenuated induction by FGF21 in hypothyroid mice. FASN is a lipogenic enzyme whose gene expression is down-regulated in the liver in states of hypothyroidism and is induced in hyperthyroidism, whereas, in contrast, the reverse regulation is observed in BAT (27). FGF21 decreased FASN gene expression in livers of the control group, but had almost no effect in hypothyroid mice, which was confirmed by a trend at the protein level. However, hypothyroid mice already have low expression of FASN, and FGF21 does not induce a further reduction. SPOT14 is another gene for which the action of FGF21 was trending to be attenuated in hypothyroid animals. Because both FASN and SPOT14 are known to be robustly down-regulated by low TH levels (27), it is likely that FGF21 could not produce any further effect in the hypothyroid mice.

Taken together, our results indicate that the effects of FGF21 and TH action in the specific target tissues studied led to the amelioration of a lipemic profile and up-regulation of energy expenditure. A closer look at the individual actions of these hormones at the liver and BAT regulatory levels reveals that divergent actions exist between FGF21 and TH. FGF21 administration leads to down-regulation of SPOT14, FASN, and sterol regulatory element-binding protein 1c (SREBP-1c) and up-regulation of phosphoenolpyruvate carboxykinase (PEPCK) in the liver, whereas TH has the opposite effect. Similarly, in BAT, FGF21 administration leads to an up-regulation in FASN, DIO2, and UCP1 gene expression, whereas the TH regulatory effect was opposite to that of FGF21.

In conclusion, we have demonstrated that the metabolic actions of TH and FGF21 are largely independent with regard to serum lipids, liver and BAT gene expression, and energy expenditure. This finding is surprising, considering the further demonstration here that TH stimulates FGF21 levels and that FGF21 peripherally administered leads to reduced TH levels. Therefore, we speculate that FGF21 may act as a biological rheostat in states of excess circulating TH to dial down the expression of metabolically adverse gene targets (ie, lipogenic gene expression), while promoting the beneficial effects of TH signaling. Further assessment the possible contribution of FGF21 to control the thyroid axis including local tissue metabolism of TH should be investigated in future studies.

Acknowledgments

This work was supported by the National Institutes of Health (Grant RO1 DK028082 to E.M.-F.) and the Hellenic Harvard Foundation (E.M.D.).

Disclosure Summary: The authors have nothing to disclose.

Footnotes

Abbreviations:
BAT
brown adipose tissue
CLAMS
comprehensive laboratory animal monitoring system
DIO1
deiodinase iodothyronine type 1
FASN
fatty acid synthase
FGF21
fibroblast groWTh factor 21
FGF21-KO
fibroblast groWTh factor 21 knockout
PGC-1α
peroxisome proliferator–activated receptor γ coactivator-1α
UCP1
uncoupling protein 1
Vo2
oxygen consumption.

References

  • 1. Kharitonenkov A, Shiyanova TL, Koester A, et al. FGF-21 as a novel metabolic regulator. J Clin Invest. 2005;115:1627–1635 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Badman MK, Pissios P, Kennedy AR, Koukos G, Flier JS, Maratos-Flier E. Hepatic fibroblast groWTh factor 21 is regulated by PPARα and is a key mediator of hepatic lipid metabolism in ketotic states. Cell Metab. 2007;5:426–437 [DOI] [PubMed] [Google Scholar]
  • 3. Fisher FM, Estall JL, Adams AC, et al. Integrated regulation of hepatic metabolism by fibroblast groWTh factor 21 (FGF21) in vivo. Endocrinology. 2011;152:2996–3004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Fisher FM, Kleiner S, Douris N, et al. FGF21 regulates PGC-1α and browning of white adipose tissues in adaptive thermogenesis. Genes Dev. 2012;26:271–281 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. van Marken Lichtenbelt W. Brown adipose tissue and the regulation of nonshivering thermogenesis. Curr Opin Clin Nutr Metab Care. 2012;15:547–552 [DOI] [PubMed] [Google Scholar]
  • 6. Chartoumpekis DV, Habeos IG, Ziros PG, Psyrogiannis AI, Kyriazopoulou VE, Papavassiliou AG. Brown adipose tissue responds to cold and adrenergic stimulation by induction of FGF21. Mol Med. 2011;17:736–740 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Coskun T, Bina HA, Schneider MA, Dunbar JD, Hu CC, Chen Y, Moller DE, Kharitonenkov A. Fibroblast groWTh factor 21 corrects obesity in mice. Endocrinology. 2008;149:6018–6027 [DOI] [PubMed] [Google Scholar]
  • 8. Xu J, Lloyd DJ, Hale C, et al. Fibroblast groWTh factor 21 reverses hepatic steatosis, increases energy expenditure, and improves insulin sensitivity in diet-induced obese mice. Diabetes. 2009;58:250–259 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Badman MK, Koester A, Flier JS, Kharitonenkov A, Maratos-Flier E. Fibroblast groWTh factor 21-deficient mice demonstrate impaired adaptation to ketosis. Endocrinology. 2009;150:4931–4940 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Dushay J, Chui PC, Gopalakrishnan GS, et al. Increased fibroblast groWTh factor 21 in obesity and nonalcoholic fatty liver disease. Gastroenterology. 2010;139:456–463 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Cuevas-Ramos D, Almeda-Valdes P, Gómez-Pérez FJ, et al. Daily physical activity, fasting glucose, uric acid, and body mass index are independent factors associated with serum fibroblast groWTh factor 21 levels. Eur J Endocrinol. 2010;163:469–477 [DOI] [PubMed] [Google Scholar]
  • 12. Zhang X, Yeung DC, Karpisek M, Stejskal D, et al. Serum FGF21 levels are increased in obesity and are independently associated with the metabolic syndrome in humans. Diabetes. 2008;57:1246–1253 [DOI] [PubMed] [Google Scholar]
  • 13. Mraz M, Bartlova M, Lacinova Z, et al. Serum concentrations and tissue expression of a novel endocrine regulator fibroblast groWTh factor-21 in patients with type 2 diabetes and obesity. Clin Endocrinol (Oxf). 2009;71:369–375 [DOI] [PubMed] [Google Scholar]
  • 14. Adams AC, Astapova I, Fisher FM, et al. Thyroid hormone regulates hepatic expression of fibroblast groWTh factor 21 in a PPARα-dependent manner. J Biol Chem. 2010;285:14078–14082 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Klieverik LP, Coomans CP, Endert E, et al. Thyroid hormone effects on whole-body energy homeostasis and tissue-specific fatty acid uptake in vivo. Endocrinology. 2009;150:5639–5648 [DOI] [PubMed] [Google Scholar]
  • 16. Oppenheimer JH, Schwartz HL, Lane JT, Thompson MP. Functional relationship of thyroid hormone-induced lipogenesis, lipolysis, and thermogenesis in the rat. J Clin Invest. 1991;87:125–132 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Freake HC, Oppenheimer JH. Thermogenesis and thyroid function. Annu Rev Nutr. 1995;15:263–291 [DOI] [PubMed] [Google Scholar]
  • 18. Harun-Or-Rashid M, Asai M, Sun XY, Hayashi Y, Sakamoto J, Murata Y. Effect of thyroid statuses on sodium/iodide symporter (NIS) gene expression in the extrathyroidal tissues in mice. Thyroid Res. 2010;3:3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Hayashi K, Akiba Y, Tomita Y, Matsumoto T. The concerted effects of thyroid function and dietary protein on groWTh and protein metabolism in mice at different groWTh stages. J Nutr Sci Vitaminol (Tokyo). 1984;30:235–244 [DOI] [PubMed] [Google Scholar]
  • 20. Folch J, Lees M, Sloane Stanley GH. A simple method for the isolation and purification of total lipides from animal tissues. J Biol Chem. 1957;226:497–509 [PubMed] [Google Scholar]
  • 21. Zhao S, Fernald RD. Comprehensive algorithm for quantitative real-time polymerase chain reaction. J Comput Biol. 2005;12:1047–1064 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Irrcher I, Adhihetty PJ, Sheehan T, Joseph AM, Hood DA. PPARγ coactivator-1α expression during thyroid hormone- and contractile activity-induced mitochondrial adaptations. Am J Physiol Cell Physiol. 2003;284:C1669–C1677 [DOI] [PubMed] [Google Scholar]
  • 23. Boström P, Wu J, Jedrychowski MP, et al. A PGC1-α-dependent myokine that drives brown-fat-like development of white fat and thermogenesis. Nature. 2012;481:463–468 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Maia AL, Goemann IM, Meyer EL, Wajner SM. Deiodinases: the balance of thyroid hormone: type 1 iodothyronine deiodinase in human physiology and disease. J Endocrinol. 2011;209:283–297 [DOI] [PubMed] [Google Scholar]
  • 25. Maia AL, Kim BW, Huang SA, Harney JW, Larsen PR. Type 2 iodothyronine deiodinase is the major source of plasma T3 in euthyroid humans. J Clin Invest. 2005;115:2524–2533 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Laurberg P, Vestergaard H, Nielsen S, et al. Sources of circulating 3,5,3′-triiodothyronine in hyperthyroidism estimated after blocking of type 1 and type 2 iodothyronine deiodinases. J Clin Endocrinol Metab. 2007;92:2149–2156 [DOI] [PubMed] [Google Scholar]
  • 27. Blennemann B, Leahy P, Kim TS, Freake HC. Tissue-specific regulation of lipogenic mRNAs by thyroid hormone. Mol Cell Endocrinol. 1995;110:1–8 [DOI] [PubMed] [Google Scholar]

Articles from Endocrinology are provided here courtesy of The Endocrine Society

RESOURCES