T3 and Glucose Coordinately Stimulate ChREBP-Mediated Ucp1 Expression in Brown Adipocytes From Male Mice.

Increasing brown adipose tissue (BAT) activity is regarded as a potential treatment of obese, hyperglycemic patients with metabolic syndrome. Triiodothyronine (T3) is known to stimulate BAT activity by increasing mitochondrial uncoupling protein 1 (Ucp1) gene transcription, leading to increased thermogenesis and decreased body weight. Here we report our studies on the effects of T3 and glucose in two mouse models and in mouse immortalized brown preadipocytes in culture. We identified carbohydrate response element binding protein (ChREBP) as a T3 target gene in BAT by RNA sequencing and studied its effects in brown adipocytes. We found that ChREBP was upregulated by T3 in BAT in both hyperglycemic mouse models. In brown preadipocytes, T3 and glucose synergistically and dose dependently upregulated Ucp1 messenger RNA 1000-fold compared with low glucose concentrations. Additionally, we observed increased ChREBP and Ucp1 protein 11.7- and 19.9-fold, respectively, along with concomitant induction of a hypermetabolic state. Moreover, downregulation of ChREBP inhibited T3 and glucose upregulation of Ucp1 100-fold, whereas overexpression of ChREBP upregulated Ucp1 5.2-fold. We conclude that T3 and glucose signaling pathways coordinately regulate the metabolic state of BAT and suggest that ChREBP is a target for therapeutic regulation of BAT activity.

Brown adipose tissue (BAT) is responsible for nonshivering heat generation (1, 2) by having abundant well-developed mitochondria containing uncoupling protein 1 (UCP1), a protein that uncouples mitochondrial respiration from adenosine triphosphate (ATP) synthesis (2). BAT is now regarded as a potential target to combat obesity (3, 4).

Triiodothyronine (T3), the active thyroid hormone, has been shown to have a direct effect on Ucp1 transcription and on the stabilization of Ucp1 messenger RNA (mRNA) in brown adipocytes (5–9). Several regions of the rat Ucp1 gene promoter have thyroid hormone response elements (10), and there is a high correlation between the occupancy of nuclear T3 receptors and increases in Ucp1 expression (11).

Metabolic syndrome is a cluster of conditions including high blood glucose levels and excess body fat around the waist that increases the risk of heart disease, stroke, and diabetes (12). Its prevalence in the United States is estimated to be 25% (13). Treatment with T3 has been shown to be beneficial for patients with metabolic syndrome by causing a hypermetabolic state characterized by increased energy expenditure and weight loss despite hyperphagia (14–16). T3 also increases BAT activity in mammals (17–19).

Carbohydrate response element binding protein (ChREBP, also known as Mlxipl) is a glucose-responsive transcription factor (20) activated by increased glucose flux into cells (21, 22). In white adipose tissue (WAT), it activates fatty acid synthesis and glycolysis (23). ChREBP is tightly coupled to glucose influx via glucose transporter 4 (Glut4) in adipose tissue, activates de novo lipogenesis in WAT, and is a key determinant of systemic insulin sensitivity and glucose homeostasis (24–27). In WAT, ChREBP expression strongly correlates with GLUT4 and fatty acid synthase (Fasn) expression (28). Importantly, it has been shown recently that T3 induces browning of WAT (29).

In this article we demonstrate that T3 and glucose coordinately upregulate Ucp1 and mitochondrial biogenesis in brown adipocytes and identify ChREBP as a main mediator of this response.

Materials and Methods

Mice experiments

All animal experiments were conducted according to the National Institutes of Health guidelines. Mice were housed at 22°C to 23°C with a 12 hour:12 hour light cycle. For the model of metabolic syndrome, male C57Bl/6J obese and control mice were purchased from the Jackson Laboratory at 12 weeks of age and maintained on the same diets: 10 kcal% as fat for controls and 60 kcal% as fat for obese mice (Research Diets, New Brunswick, NJ). Beginning at week 14, 3 µg/mL T3 (Sigma-Aldrich, St. Louis, MO) was added to the drinking water for 10 days. Body weight, core body temperature, and blood glucose levels (Contour®; Bayer, Mishawaka, IN) were measured every 48 hours. Terminal bleeds were performed to measure T3 and T4 levels using the Rat Thyroid Hormone Magnetic Bead Panel (Milliplex, Billerica, MA).

Glucose clamp procedures were performed by the Vanderbilt Mouse Metabolic Phenotyping Center. C57BL/6 male mice on regular chow diets at weeks 12 to 14 underwent jugular vein and carotid artery catheterization a week before the clamp procedure. Hyperglycemic clamps were performed for a 6-hour period as previously described (30). A total of 4 µg T3 was administered intravenously for the first 30 minutes of the clamp. All experiments were performed with prior approval of the National Institute of Diabetes and Digestive and Kidney Diseases Institutional Animal Care and Use Committee.

Cell culture and primary brown adipocyte isolation

Immortalized brown preadipocytes (31) were maintained at ≤70% confluence in Dulbecco’s modified Eagle medium (DMEM) containing 10% fetal calf serum (FCS), 25 mM glucose, 100U/mL penicillin, 100 mg/mL streptomycin, and 2 mM glutamine at 37°C in a 6% CO2 incubator.

To isolate primary brown adipocytes, we excised intrascapular BAT from 15-week-old male C57BL/6 mice on control or high-fat diets, minced it, and digested it with 1 mg/mL collagenase V (Sigma Aldrich) in Hanks balanced salt solution (Corning-Cellgro, Manassas, VA), supplemented with 100 U/mL penicillin and 100 mg/mL streptomycin, at 37°C for 30 minutes. After digestion, 20% FCS was added for 5 minutes, cells were centrifuged at 150g for 5 minutes, and floating cells were collected with a pipette. Cells were then cultured in DMEM medium containing 10% FCS, 100 U/mL penicillin, 100 mg/mL streptomycin, and 2 mM glutamine.

Adipogenesis induction

Cells were plated at ~30% confluency in growth medium (GM) [DMEM containing 10% resin-stripped FCS, to deplete thyroid hormones as described in Cao et al. (32), 0.1 μM insulin, 100 U/mL penicillin, 100 mg/mL streptomycin, and 2 mM glutamine] 4 days before induction of adipogenesis. At day 0, cells were fully confluent and were treated with minimal adipogenic medium (MAM) (GM supplemented with 2 μg/mL dexamethasone and 0.125 mM indomethacin) with or without 10 nM T3 (Sigma Aldrich). After 3 days, the medium was changed back to GM with stripped serum and replenished at 2-day intervals for 8 days. On the days indicated, RNA was collected with the RNeasy Mini Kit (Qiagen, Hilden, Germany).

Mitochondria labeling

Cells were incubated in Hanks balanced salt solution containing 250 nM MitoTracker® Orange CM-H2TMRos (Life Technologies, Eugene, OR) for 30 minutes at 37°C. Mitochondrial activity was analyzed based on fluorescence intensity in a BD FACSAria II Cell Sorter (Becton Dickinson, Sunnyvale, CA).

RNA sequencing and genome-wide expression profiling

RNA sequencing (RNA-seq) samples were sequenced on Illumina HiSEquation 2000 with a single-ended protocol. The sequence reads were 50 bp in length and aligned to the reference genome assembly NCBI37/mm9. The output of the Illumina Analysis was converted to browser extensible data files detailing the genomic coordinates of each mapped read. Using a utility script in SICER, we calculated RNA-seq read count and reads per kilobase per million mapped reads (RPKM) on each exon (reads on introns were disregarded). Individual gene expression levels were represented by aggregated RPKM over all exons of the gene, and genes with RPKM <0.75 in all samples were regarded as not expressed and excluded from subsequent study. The gene expression levels of preadipocytes without T3 treatment were used as base levels to compute fold changes for all expressed genes, and a fold threshold of 3 was used to identify whether a gene was significantly regulated by T3 during adipogenesis. Genes of interest were selected based on the following criteria: gene levels are unchanged by T3 in GM (ratio <1.5 and >0.75), and the ratio of averaged gene levels in T3-treated adipocytes on day 8 over the averaged levels in untreated adipocytes on day 8 of adipogenesis is >3.

Gene ontology and promoter prediction

Gene ontology was analyzed with ToppCluster (http://toppcluster.cchmc.org/). Prediction of transcription factor binding sites was made with MatInspector on Genomatix (http://www.genomatix.de/online_help/help_matinspector/matinspector_help.html) and TFSEARCH (http://www.cbrc.jp/research/db/TFSEARCH.html). A prediction was made based on sites that were identified by both ToppCluster and MatInspector.

Real-time reverse transcription polymerase chain reaction

Total RNA was purified with the RNeasy Mini Kit (Qiagen). Complementary DNA (cDNA) was prepared with a High Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA). Quantitative polymerase chain reaction (PCR) was performed on cDNAs. All reactions were conducted in 96-well plates in a total volume of 25 μL. Each reaction contained cDNA from 100 ng total RNA and appropriate amounts of Universal PCR Master Mix (Applied Biosystems) and primer/probe mix. All primers and probes were from Applied Biosystems Assay-on-Demand. The cycle threshold value >40 was considered undetectable, and the level calculated was a cycle threshold of 41.

Western blots

Cells were lysed in radioimmunoprecipitation assay buffer [50 mM tris(hydroxymethyl)aminomethane, pH 8.0/150 mM NaCl/1.0% octylphenoxypolyethoxyethanol/0.5% deoxycholate/0.1% sodium dodecyl sulfate/0.2 mM NaVO4/10 mM NaF/0.4 mM EDTA/10% glycerol with protease inhibitors] (Roche). Lysates were sonicated for 20 seconds on ice and centrifuged at 10,000g for 5 minutes. Lysates were then placed in a boiling water bath for 5 minutes with Laemmli loading buffer, followed by electrophoresis on 10% sodium dodecyl sulfate polyacrylamide gels. Western blot analyses were performed with antibodies directed at Ucp1 (Sigma Aldrich), ChREBP (Novus Biologicals, Littleton, CO), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Abcam, Cambridge, MA) at the recommended antibody dilutions. The blots were scanned with the LI-COR laser-based image detection method.

Chromatin immunoprecipitation

Chromatin immunoprecipitation (ChIP) assays were performed with 100 mg of cell chromatin extracts from 20 × 106 BAT immortalized cells. DNA was obtained with the Active Motif (Carlsbad, CA) chromatin shearing kit. Chromatin was precipitated by incubation with ChREBP antibody (Novus Biologicals) and thyroid hormone receptor antibody ab2743 (Abcam) at the recommended dilutions at the recommended concentrations or a 1:10,000 dilution of rabbit immunoglobulin G (Abcam) followed by separation with protein G magnetic beads (Active Motif). Binding was analyzed by real-time PCR. Primer sequences are available upon request.

Electrophoretic mobility shift assay

Nuclear extracts from HeLa cells overexpressing ChREBP, derived from adenovirus (kindly provided by Dr. Towle, University of Minnesota) (33), or adenovirus expressing green fluorescent protein (GFP) were obtained with the NE-Per kit (Thermo Scientific). A total of 10 μg nuclear extract was added to electrophoretic mobility shift assay reaction buffer [20 mmol/L HEPES (pH 7.9), 5 mmol/L MgCl2, 0.5 mmol/L EDTA, 50 mmol/L KCl, 1 mmol/L dithiothreitol, 6.25% glycerol, 1 μg bovine serum albumin, 2 μg salmon sperm DNA, and 2 μg poly(dI-dC)]. Probes (ChREBP on Ucp1 S: CAAAAGGCACCACGCTGCGGGGACGCGGGTGAAGC; ChREBP on Ucp1 AS: GCTTCACCCGCGTCCCCGCAGCGTGGTGCCTTTTG) were fluorescently labeled with 5IRD800 (Integrated DNA Technologies), or unlabeled probes (“cold probes”) were annealed for 5 minutes at 95°C. For supershift experiments, nuclear extracts (10 μg) were preincubated with the 1:500 dilution of ChREBP antibody (Novus Biologicals) for 10 minutes at room temperature before the addition of labeled DNA probes. Samples were loaded to 6% Tris-borate-EDTA gel (Life Technologies, Gaithersburg, MD) and scanned with the LI-COR laser-based image detection method.

Knockdown and overexpression of ChREBP in brown adipocytes

Knocking down ChREBP was accomplished with short hairpin RNA (shRNA) derived from lentivirus (Mission® shRNA, Sigma-Aldrich) or control scrambled sequences. Brown preadipocytes were infected at a multiplicity of infection of 50, followed by selection with puromycin (5 μg/mL). The selected cells were induced to differentiate into adipocytes.

Overexpression of ChREBP was accomplished with an adenovirus. Brown preadipocytes were infected at a multiplicity of infection of 100 and then induced to differentiate into adipocytes. We used adenovirus expressing GFP as a control for the infection.

Oil Red O staining

Immortalized BAT overexpressing or silenced for ChREBP expression was stained with Oil Red O as previously described (34). Briefly, confluent cells were washed twice with phosphate-buffered saline and incubated in 200 μL of Oil Red O solution (0.33% wt/vol in 60% isopropanol) for 20 minutes at room temperature. Images of Oil Red O–stained adipocytes were acquired with the EVOS (Life Technologies, Carlsbad, CA) microscope, ×40 objective. Photomicrographs of multiple fields of multiple fragments from five standard chow and five high-fat diet male mice were imported into Image J, and total red (representing neutral lipids) and lipid droplet size were quantified as previously described (35).

Glucose and lactate measurements

Glucose and lactate in the media from cells under the indicated conditions (after 48 hours without any media change) were measured with Contour® (Bayer) glucose test strips and Lactate Plus (Nova Biomedical, Waltham, MA) test strips.

Fluorescence-activated cell sorting

Flow cytometry was performed with a BD FACSAria II Cell Sorter (Becton Dickinson), with a 70-μm nozzle and a sheath pressure of 65 psi. Debris and clustered cells were excluded from gated populations.

Measurement of oxygen consumption rate

Brown adipocytes that were either silenced or overexpressing ChREBP underwent the differentiation protocol. Oxygen consumption rate (OCR) was determined in a Seahorse XF96 extracellular flux analyzer. OCR was recorded after the following additions: oligomycin (2.5 μM), carbonyl cyanide p-trifluoromethoxy-phenylhydrazone (1 μM), and a rotenone (1 μM) and antimycin A (1 μM). Results were normalized to total protein with the PierceTM BCA Protein Assay Kit (Thermo Scientific, Rockford, IL).

Statistics

All experiments were performed at least three times. A Student t test was used, and differences were considered significant when P < 0.05.

Results

T3 causes greater increases in BAT activity in obese compared with normal mice

We tested the effects of T3 on BAT activity in obese mice. Male C57Bl/6J mice kept on a high-fat diet (60% of calories as fat) became obese (body weight >42 g) after 14 weeks. Obese and control mice (diet with 10% calories as fat) were treated with 3 µg/mL T3 in their drinking water for 10 days. As expected, treatment with T3 led to increased serum T3 levels and suppressed T4 levels (Fig. 1A). T3 had no effect on body weight in control mice, but led to decreased body weight in obese mice (Fig. 1B). Hyperglycemia in obese mice was brought down to control levels with T3 treatment; in control mice, T3 also decreased blood glucose levels (Fig. 1C). Mice treated with T3 had increased food intake in both normal and high-fat diet groups. Core body temperature was higher in both obese and control mice treated with T3 (Fig. 1D). Increased body temperature is an indication of BAT generation (1, 36, 37). Mitochondrial activity in cells from intrascapular BAT was highest in cells from obese mice treated with T3 (Fig. 1E). When BAT was stained for active mitochondria with the MitoTracker® Orange CM-H2TMRos, the mean fluorescence level was 4.6- ± 2.7-fold (P < 0.02) greater than in mice fed with control diet without T3. Ucp1 mRNA levels in intrascapular BAT were also highest in T3-treated obese mice (Fig. 1F). Taken together, these data show that treatment with T3 increases BAT activity to higher levels in obese, hyperglycemic mice than in control mice. The increased BAT activity in these mice explains the decreased body weight and improved glycemic levels. To study the causative factors responsible for the higher BAT activity in obese mice treated with T3, we used immortalized brown preadipocytes and adipocytes in culture.

Figure 1.

T3 causes greater increases in intrascapular BAT activity in obese hyperglycemic mice than in control mice. Mice were fed a control diet in which 10% of Kcal was from fat (10% Kcal) and a high-fat diet in which 60% of Kcal was from fat (60% Kcal) without (ctrl) or with 3 µg/mL T3 in the drinking water (T3) for 10 days. (A) T3 and T4 levels after 10 days of T3 treatment. (B) Body weight at euthanasia. (C) Blood glucose levels. (D) Core body temperature. (E) Labeling for active mitochondria by MitoTracker Orange CM-H2TMRos (analyzed by fluorescence-activated cell sorting). Tracings represent primary brown adipocytes from mice, with the median fluorescent levels from each group. (F) Ucp1 mRNA levels in intrascapular BAT measured by quantitative reverse transcription PCR. Values are means ± standard error; n = 5 mice per group. *P < 0.05.

Optimization of culture conditions for the study of T3 effects in brown preadipocytes and adipocytes

To study the effects of T3 on brown adipocytes, we established culture conditions in which preadipocytes can differentiate, expression of key adipogenic genes (leptin, Pparg, and Tshr) was upregulated, and Ucp1, a known target of the T3 nuclear receptor (11, 17), was further induced by T3 addition. We tried all combinations of classic adipogenic media additives (insulin, isobutylmethylxanthine, dexamethasone, indomethacin, and T3) (38, 39). We found that DMEM supplemented with 10% resin-stripped FCS [to deplete thyroid hormones as described in Cao et al. (32)], insulin, dexamethasone, and indomethacin led to preadipocyte differentiation with minimal Ucp1 induction (Fig. 2A). Addition of 10 nM T3 led to a robust increase in Ucp1 mRNA expression (Fig. 2A). We refer to this medium as MAM.

Figure 2.

Effects of T3 on mRNA levels in brown preadipocytes and adipocytes. (A) Left panel: schematic representation of the differentiation protocol of immortalized preadipocytes. Right panel: mRNA levels measured by quantitative reverse transcription PCR of Ucp1, adiponectin (Adipoq), leptin, and peroxisome proliferator–activated receptor γ in the different culture conditions. T3 was added at a concentration of 10 nM for 3 days. Values are means ± standard error for at least three different experiments. *P < 0.05. (B) Genes upregulated by T3 in brown adipocytes but not in preadipocytes. Heat map represents the relative RPKM levels from RNA-seq for each mRNA under the different conditions tested. NS, no significant effect.

Genes upregulated by T3 in adipocytes but not in preadipocytes

RNA-seq was performed on lysates from brown preadipocytes and adipocytes. We identified genes that were affected by T3 in adipocytes but not by T3 in preadipocytes. Aside from Ucp1, 27 additional genes were found (Fig. 2B). A gene ontology search revealed that these genes are involved in metabolic sensing and processing of glucose and energy generation (Supplemental Table 1A) and that knockdown of these genes led to altered fatty acid levels and body temperature in mice (Supplemental Table 1B).

We chose to focus on Ucp1 and Fasn because these genes are known to play key roles in BAT function, and ChREBP and Glut4, which are important in glucose sensing and response, respectively, in other tissues. In WAT, ChREBP has two splice isoforms, α and β (28); however, RNA-seq and quantitative reverse transcription polymerase chain reaction (qRT-PCR) show that only the α isoform is expressed in mouse brown preadipocytes and adipocytes under the conditions tested.

Prediction of transcription factor binding sites on the proximal promoter (−1 kb) regions of these four genes based on two databases (see Materials and Methods) revealed that all four genes have both ChREBP and T3 nuclear receptor/retinoid X receptor binding sites (Supplemental Fig. 1a), suggesting that these four genes may be regulated by both T3 and glucose. We confirmed binding of T3 receptor and ChREBP on all four promoters by using ChIP (Supplemental Fig. 1b). Using electrophoretic mobility shift assay, we validated that ChREBP can bind within the proximal site of Ucp1 promoter between positions −256 and −272 (Supplemental Fig. 1c).

T3 upregulates mRNA levels of Ucp1, ChREBP, Glut4, and Fasn in intrascapular BAT of mice undergoing hyperglycemic clamp

We determined whether T3 would increase expression of ChREBP, Ucp1, Glut4, and Fasn mRNAs in the intrascapular BAT of hyperglycemic mice. C57BL/6 male mice (13 to 15 weeks old) were given intravenous infusions of T3 (4 µg for the first 30 minutes of a 6-hour hyperglycemic clamp). After 6 hours, plasma levels of T3 were two times higher in the T3-treated group than in controls (Fig. 3A). Glucose levels remained in the range of 260 ± 22 mg/dL throughout the experiment in both groups (Fig. 3B). Glucose infusion rates had to be adjusted to higher levels in the T3-treated group to achieve similar glucose levels (Fig. 3C) because insulin levels were higher in the group treated with T3 than in controls (Fig. 3D). These differences in the two mice groups are possibly caused by previously known effects of T3 on hepatic gluconeogenesis and insulin resistance (40–43). We explored whether transcriptional changes in BAT occur even after short exposure to T3 and glucose. To determine whether transcriptional changes in BAT occurred, after 6 hours intrascapular BAT was excised and mRNA levels of the four genes of interest were measured (Fig. 3E). Interestingly, ChREBP was upregulated (4.9-fold) in the hyperglycemic mice treated with T3 compared with controls. Ucp1, Glut4, and Fasn were also upregulated under these conditions. By contrast, there was no effect of T3 on ChREBP and Ucp1 mRNA levels in euglycemic mice (data not shown). This response in gene expression within a short time period of T3 treatment supports our conclusion that T3 in combination with glucose regulates brown adipose activity.

Figure 3.

T3 increases expression of ChREBP, Ucp1, Glut4, and Fasn in the intrascapular BAT of mice undergoing a hyperglycemic clamp. Normal weight C57BL/6 mice were clamped to become hyperglycemic for 6 hours. T3 was administered intravenously during the first 30 minutes of the clamp. (A) T3 plasma levels after 6 hours were increased. *P < 0.05. (B) Arterial glucose levels. (C) Glucose infusion rate. (D) Plasma insulin levels. (E) mRNA levels of ChREBP, Ucp1, Glut4, and Fasn in the intrascapular BAT at the end of the 6-hour hyperglycemic clamp. Values are means ± standard error; n = 5 mice per group.

Transcriptional regulation of Ucp1, ChREBP, Glut4, and Fasn mRNA levels depends on both glucose and T3 signaling pathways

We showed that Ucp1 and ChREBP mRNA levels were upregulated by T3 more in brown adipocytes than in preadipocytes (Fig. 4A). At low glucose concentrations (1.67 mM), mRNA levels for Ucp1 and ChREBP (Fig. 4A) and for Glut4 and Fasn (not shown) were low and were only minimally affected by T3. However, in high glucose (25 mM), T3 led to robust increases in ChREBP, Ucp1, Glut4, and Fasn mRNA levels in brown adipocytes. Moreover, we showed that Ucp1 and ChREBP protein levels were increased most robustly in T3-treated adipocytes (Fig. 4B). mRNA levels of adiponectin, leptin, and peroxisome proliferator–activated receptor γ (PPARγ) in adipocytes were significantly increased by high glucose concentrations but were not affected by addition of T3 (Supplemental Fig. 2a and 2b). Maximal induction of leptin and PPARγ is obtained after 8 days in MAM (Supplemental Fig. 2b). Addition of T3 to the media increased levels of Ucp1 mRNA to higher extent, reaching a maximum at day 8 of adipogenesis, whereas in the absence of T3 maximal levels were reached at day 4 of adipogenesis and were significantly lower than in the presence of T3 (Supplemental Fig. 2b). We confirmed these mRNA data in a second immortalized brown fat cell line from a different mouse background (C57/FVB/DBA) (data not shown). To assess whether fat deposition is different because of our treatments, we performed Oil Red O staining. Preadipocytes had very little fat deposition and adipocytes had significantly greater Oil Red O staining, which was independent of the glucose or T3 presence in the MAM (Supplemental Fig. 2C). Average lipid droplet size remained unchanged between the treatments (data not shown). OCR was assessed; basal respiration, proton leak, and maximal respiration capacity were increased by T3 and further increased to a higher extent in adipocytes in high glucose (25 mM) compared with low glucose (1.67 mM), indicative of a hypermetabolic state in brown adipocytes exposed to high glucose and T3 (Supplemental Fig. 2D).

Figure 4.

Effects of glucose and T3 on Ucp1 and ChREBP in brown preadipocytes and adipocytes. (A) qRT-PCR of mRNA levels of Ucp1 and ChREBP in brown preadipocytes and adipocytes in low or high glucose without or with 10 nM T3 for 3 days. Values are means ± standard error; each performed three times for three independent experiments. *P < 0.05. (B) Western blot analysis of Ucp1, ChREBP, and GAPDH protein expression. Low glucose = 1.67 mM glucose; high glucose = 25 mM glucose.

In high glucose (25 mM), Ucp1 mRNA levels were dose-dependently increased by increasing concentrations of T3 in brown adipocytes, whereas in low glucose, Ucp1 mRNA levels were 1000-fold lower and were less responsive to T3 (Fig. 5A). A dose response to glucose revealed that Ucp1 mRNA reached the highest mRNA levels (~50-fold increase) only in the presence of T3 (Fig. 5B). These data indicate that T3 and glucose act synergistically on Ucp1 mRNA levels.

Figure 5.

T3 and glucose have a synergistic effect on Ucp1 gene expression in brown adipocytes. (A) Dose response of Ucp1 mRNA levels to increasing T3 concentrations in cells incubated in low (1.67 mM) or high (25 mM) glucose for 3 days. (B) Dose response of Ucp1 mRNA levels to increasing glucose concentrations in cells incubated in the absence or presence of 10 nM T3 for 3 days. Data are the mean ± standard deviation of three independent experiments. *P < 0.05.

T3 and high glucose coordinately lead to increased mitochondrial activity and hypermetabolic brown adipocytes

Next, we measured mitochondrial activity of brown adipocytes by using MitoTracker, a fluorescent dye that labels active, reduced mitochondria in live cells. The highest fluorescence levels were found in adipocytes treated with T3 and high glucose. Mitochondrial activities were lower in adipocytes incubated in high glucose without T3, even lower in preadipocytes incubated in low glucose with T3, and lowest in preadipocytes incubated in low glucose without T3 (Fig. 6A). The levels of mitochondrial activity corresponded to Ucp1 mRNA and protein levels (Fig. 4A and 4B).

Figure 6.

T3 and glucose coordinately increase mitochondrial activity and metabolic rate in brown preadipocytes and adipocytes. (A) Brown preadipocytes and adipocytes were labeled for 30 minutes with MitoTracker Orange CM-H2TMRos to stain cells for active (reduced membrane potential) mitochondria and were analyzed on a fluorescence-activated cell sorter. (B) Glucose and lactate were measured in brown preadipocytes and adipocyte cultures in the indicated media. Data are the mean ± standard deviation of three independent experiments. *P < 0.05.

The metabolic states of the cells, monitored by glucose depletion from the medium and lactate production, corresponded to the levels of mitochondrial activity also (Fig. 6B). Brown adipocytes exhibited the highest metabolic levels when incubated in high glucose and T3. These data show that T3 and glucose coordinately regulate brown adipocytes to become hypermetabolic.

Changing ChREBP expression affects the levels of Ucp1, Glut4, and Fasn

To confirm a central role for ChREBP, we knocked down ChREBP by using shRNA expressed by lentivirus infection. ChREBP mRNA was decreased by 51% ± 2.5% in brown preadipocytes and 89% ± 3.7% (P < 0.04) in cells differentiated into adipocytes, and ChREBP protein was markedly decreased also (Supplemental Fig. 3A). Knockdown of ChREBP affected the ability of brown preadipocytes to differentiate and accumulate fat in low glucose in the absence of T3 but did not affect fat accumulation in high glucose or in the presence of T3 (Supplemental Fig. 3C). Decreasing ChREBP resulted in decreased levels of Ucp1, Glut4, and Fasn mRNAs in adipocytes cultured in medium containing T3 (Fig. 7A). The OCR was measured. Basal respiration, proton leak, and maximal respiration capacity were increased by T3 but decreased by ChREBP knockdown and abolished the T3 response (Fig. 7A).

Figure 7.

ChREBP mediates expression of Ucp1, Glut4, and Fasn in brown adipocytes. (A) Left panel: Knockdown of ChREBP with shRNA (shChREBP) lowers Ucp1, Glut4, and Fasn mRNA levels in brown adipocytes incubated in high-glucose (25 mM) medium containing T3. Right panel: The OCR of brown adipocytes cultured in 25 mM glucose was measured with a Seahorse XF96. (B) Left panel: Overexpression of ChREBP by adenovirus infection (Ad-ChREBP) increases Ucp1, Glut4, and Fasn mRNA levels in brown adipocytes incubated in low-glucose (1.67 mM) medium containing T3 for 3 days. Right panel: The OCR of brown adipocytes cultured in 1.67 mM glucose was measured with a Seahorse XF96. Data are the mean ± standard deviation of three independent experiments. *P < 0.05. AA, antimycin A; Ad-GFP, Adenovirus containing GFP; FCCP, carbonyl cyanide p-trifluoromethoxy-phenylhydrazone; OE, overexpression.

Next, we overexpressed ChREBP in brown preadipocytes and adipocytes. ChREBP mRNA was increased 202- ± 41-fold in brown preadipocytes and 134- ± 11-fold (P < 0.04) in cells differentiated into adipocytes, and ChREBP protein was markedly increased also (Supplemental Fig. 3B). The overexpression of ChREBP did not alter the ability of brown preadipocytes to differentiate, as measured by their abilities to accumulate fat by Oil Red O staining in all conditions tested (Supplemental Fig. 3C). In brown adipocytes incubated in medium with low glucose containing T3, Ucp1 levels increased 5.2-fold, Glut4 2.7-fold, and Fasn 2.5-fold (Fig. 7B). We measured OCR in the brown adipocytes overexpressing ChREBP and found that the overexpression of ChREBP in low glucose led to significantly enhanced basal and maximal respiratory capacity (Fig. 7B). In high glucose, overexpression of ChREBP had only minor effects on the levels of the mRNAs of these genes (data not shown). Taken together, these data indicate that ChREBP is a critical regulator of expression of Ucp1, Glut4, and Fasn in brown adipocytes.

Discussion

In this article we demonstrate that T3 and glucose synergistically upregulate Ucp1 expression in brown adipocytes and thereby increase mitochondrial activity. Moreover, this effect is mediated by ChREBP, which is coordinately upregulated by T3 and glucose. T3 and glucose also increased the expression of Glut4 and Fasn, key genes important in glucose sensing and glucose transport into the cell and in fatty acid synthesis, respectively. These four genes are important for BAT activity. A model describing the role of ChREBP in BAT is presented in Fig. 8.

Figure 8.

A model of the central role of ChREBP in T3 and glucose coordinative regulation of brown adipocyte function. Glucose, transported into the cell via GLUT4, is metabolized and signals for cytoplasmic ChREBP to be dephosphorylated and transported into the nucleus. In the nucleus, T3 binds to its nuclear receptor and coordinately with ChREBP regulates transcription of Ucp1, ChREBP, GLUT4, and Fasn. Glucose also serves as a substrate to generate FFAs after processing in the mitochondria to obtain acetyl coenzyme A (acetyl-CoA), via the enzyme FASN in lipid vesicles. Both FFAs and glucose serve as substrates for mitochondrial activity and thermogenesis via Ucp1 uncoupling of mitochondrial ATP synthesis.

Our data indicate that ChREBP mediates the coordinate effects of T3 and glucose to induce adipogenesis and mitochondrial biogenesis in BAT. It was previously demonstrated that ChREBP regulates mitochondrial activity and glycolysis in other cell systems and fatty acid synthesis, glycolysis, and mitochondrial activity in WAT and hepatocytes (23, 44–48). ChREBP was previously shown to be expressed at high levels in BAT, and knockdown of ChREBP results in decreases in BAT mass (23).

In brown adipocytes, T3 and glucose had a synergistic effect to upregulate the level of Ucp1 mRNA, a finding that we suggest is caused in part by coordinate regulation of Ucp1 transcription by T3 and glucose. This suggestion is based on the prediction of binding sites for ChREBP and the T3 nuclear receptor in the Ucp1 promoter region. The genes for ChREBP, Glut4, and Fasn are also predicted to contain binding sites for ChREBP and the T3 nuclear receptor, and we found additive but not synergistic upregulation of these genes by glucose and T3. We think that the concept of coordinate regulation of gene transcription by T3 and glucose signaling pathways may apply to other genes in other cell types.

It has been shown that T3 treatment of obese mammals results in loss of weight, increased thermogenesis, and improved glucose tolerance (6, 49, 50). We initially confirmed these findings in our test animals. We found that BAT activity in obese hyperglycemic mice (a mouse model of metabolic syndrome) treated with T3 was higher than in mice fed a control diet and treated with T3. However, we could not exclude the possibility that the presence of increased levels of free fatty acids (FFAs) (or other unknown factors) were the cause of increased BAT activity in the obese mice. Therefore, we performed a 6-hour glucose clamp in mice. Similar to our observations in the mouse model of metabolic syndrome, we found that mice clamped at an elevated glucose level and treated with T3 showed increased Ucp1 mRNA expression in intrascapular BAT after just 6 hours of T3 treatment. This finding strongly supports the idea that T3 was acting via a direct effect on BAT, perhaps in part via ChREBP. The T3 receptor has been previously shown to regulate and directly bind the ChREBP promoter in the liver (51), supporting our parallel findings in brown adipocytes. Therefore, it appears that patients with metabolic syndrome could be treated with T3. However, T3 administration was shown to have numerous side effects, including cardiac problems, muscle weakness, and excessive erosion of lean body mass [reviewed in (52)]. However, we suggest that a T3 analog with BAT-selective activity, like the liver-specific thyroid hormone analog being developed for the treatment of hyperlipidemia (53), could be used for this purpose.

Reported effects of insulin on BAT activity have not been consistent. In previous studies, insulin has been found to have no effect (54) or to upregulate Ucp1 levels (55). In the BAT cell system we used, insulin was a necessary component of the adipogenic media to obtain a T3 effect on Ucp1 expression (data not shown). Thus, the increase in insulin secretion observed in the hyperglycemic T3-treated mice probably increased glucose uptake by BAT and contributed to the increases in ChREBP, Ucp1, Glut4, and Fasn mRNA levels.

Our RNA-seq data showed that a total of 28 genes were upregulated by T3 in brown adipocytes. Interestingly, many genes involved in glycolysis were upregulated by T3, including phosphofructokinase (Pfkfb 1 and 3); PRKAA2, which phosphorylates Pfkfb1,3; and ATP citrate lyase, which positively regulates glycolysis. Therefore, T3 not only regulates glucose uptake (Glut4) and downstream signaling caused by glucose flux, in part, by ChREBP, it also regulates glucose metabolism. This finding is important in view of recent data indicating that Glut4 and therefore glucose uptake is substantially higher in BAT than in WAT in response to either insulin or cold exposure (56).

Inguinal fat was shown to contain “beige” fat cells that upon exposure to cold can function in a fashion similar to brown fat to produce heat by consumption of fat and glucose (57, 58). It is possible that in the context of hyperglycemia T3 can induce beige tissue to become more brown-like. Additional work is needed to address this question.

In summary, T3 and glucose signal coordinately to upregulate ChREBP, Ucp1, Glut4, and Fasn in BAT. We identified ChREBP as a central regulator of BAT activity. Lastly, we suggest that a BAT-selective T3 analog could be beneficial in the treatment of obesity, especially in hyperglycemic patients.

Appendix.

Antibody Table

Peptide/Protein Target	Antigen Sequence (if Known)	Name of Antibody	Manufacturer, Catalog No.	Species Raised in; Monoclonal or Polyclonal	Dilution Used	RRID
		Ucp1	Sigma-Aldrich, SAB1404511	Mouse; monoclonal	1:500	AB_10740377
		ChREBP	Novus Biologicals, NB400-135	Rabbit; polyclonal	1:1000 WB, 4 μg ChIP	AB_10002435
		GAPDH	Abcam, ab8245	Mouse; monoclonal	1:1000	AB_2107448
		THR	Abcam, ab2743	Mouse; monoclonal	4 μg	AB_2043044

Abbreviation: RRID, Research Resource Identifier.