Thyroid hormones induce browning of white fat

The canonical view about the effect of thyroid hormones (THs) on thermogenesis assumes that the hypothalamus acts merely as a modulator of the sympathetic outflow on brown adipose tissue (BAT). Recent data have challenged that vision by demonstrating that THs act on the ventromedial nucleus of the hypothalamus (VMH) to inhibit AMP-activated protein kinase (AMPK), which regulates the thermogenic program in BAT, leading to increased thermogenesis and weight loss. Current data have shown that in addition to activation of brown fat, the browning of white adipose tissue (WAT) might also be an important thermogenic mechanism. However, the possible central effects of THs on the browning of white fat remain unclear. Here, we show that 3,3′,5,5′ tetraiodothyroxyne (T4)-induced hyperthyroidism promotes a marked browning of WAT. Of note, central or VMH-specific administration of 3,3′,5-triiodothyronine (T3) recapitulates that effect. The specific genetic activation of hypothalamic AMPK in the VMH reversed the central effect of T3 on browning. Finally, we also showed that the expression of browning genes in human WAT correlates with serum T4. Overall, these data indicate that THs induce browning of WAT and that this mechanism is mediated via the central effects of THs on energy balance.


Introduction
Thyroid hormones (THs; 3,3′,5,5′ tetraiodothyroxyne or T 4 and 3,3′,5-triiodothyronine or T 3 ) exert important biological actions, not only modulating the development and growth but also regulating metabolism and energy balance (Brent 2012, Warner & Mittag 2012, Lopez et al. 2013. Impaired function of the thyroid gland, by either hyperthyroidism or hypothyroidism, leads to alterations in metabolism and energy homeostasis. Hyperthyroidism is associated with an increase in the metabolic rate and the patients suffering from this condition undergo body weight loss, despite increased food intake; quite the opposite, hypothyroid patients show lowered metabolic rate and reduced food intake (Brenta et al. 2007, Kaptein et al. 2009, Pearce 2012.
THs are key regulators of thermogenesis, which represents a major component of the energy expenditure in homeothermic ('warm-blooded') animals (Cannon & Nedergaard 2004, Silva 2006. In mammals, including humans, thermogenesis occurs mainly in the brown adipose tissue (BAT) (Cannon & Nedergaard 2004, Silva 2006, von Ballmoos et al. 2009). THs act on brown adipocyte thermogenesis by increasing the stimulatory action of norepinephrine (NE), as well as enhancing the cAMP-mediated acute rise in ucp1 gene expression (Bianco et al. 1988, Silva 2006, Ribeiro et al. 2010. The existence of central effects of THs in the regulation of BAT thermogenesis was proposed long time ago (Nedergaard et al. 1997). Recent evidence from our group has also shown a homeostatic link between the central effects of THs on hypothalamic AMP-activated protein kinase (AMPK), sympathetic tone and UCP1 expression in BAT (Lopez et al. 2010, Alvarez-Crespo et al. 2016.
Over the last years, accumulating evidence has demonstrated that activation of beige/brite ('brown in white') adipocytes in the white adipose tissue (WAT), a process known as browning (Fisher et al. 2012, Cohen et al. 2014, Nedergaard & Cannon 2014, Contreras et al. 2016b, is responsible for a significant increase in total energy expenditure (Shabalina et al. 2013). Thus, stimulation of browning has therapeutic potential to promote body fat reduction (Yoneshiro et al. 2013, Beiroa et al. 2014. Several mechanisms have been proposed for WAT browning (Villarroya & Vidal-Puig 2013, Nedergaard & Cannon 2014, including prolonged cold exposure (Loncar et al. 1986), adrenergic activation (Cousin et al. 1992, Ghorbani et al. 1997, Cao et al. 2011) and also thyroid hormone receptor (TR) agonism (Lin et al. 2015, Alvarez-Crespo et al. 2016. However, the role of central THs in the control of WAT browning remains unclear.
The aim of this study was to investigate the role of central THs on the browning of WAT and the mechanisms behind this action. Our data show that peripherally induced hyperthyroidism promoted browning of white fat and that this effect is recapitulated by central and specific administration of T 3 in the ventromedial nucleus of hypothalamus (VMH), via a mechanism dependent of AMPK. Notably, we also demonstrate that the expression of browning markers in WAT correlates with serum T 4 levels in humans. Thus, in addition to the well-known effects of central THs on BAT (Lopez et al. 2010, Alvarez-Crespo et al. 2016, our data indicate an additional mechanism by which central THs influence energy expenditure, namely browning of WAT.

Animals
Male Sprague-Dawley rats (200-250 g; Animalario General USC, Santiago de Compostela, Spain) were housed on a 12-h light (08:00-20:00), 12-h darkness cycle, in a temperature and humidity controlled room and maintained with chow (STD, SAFE A04: 3.1% fat, 59.9% carbohydrates, 16.1% proteins, 2.791 kcal/g; Scientific Animal Food & Engineering; Nantes, France) and water ad libitum. For all the procedures, the animals were individually caged and used for experimentation 7 days later. During all experimental approaches, animals and their respective food intake and body weight were monitored every day. The experiments were performed in agreement with the International Law on Animal Experimentation and were approved by the USC Ethical Committee (Project ID 15010/14/006).

Patients
A group of 163 (80 visceral, vWAT and 83 subcutaneous, sWAT) white adipose tissues from participants were analyzed (Table 1). These participants were recruited at the Endocrinology Service of the Hospital of Girona 'Dr Josep Trueta'. All subjects were of Caucasian origin and reported that their body weight had been stable for at least three months before the study. Subjects were studied in their post-absorptive state. They had no systemic disease other than obesity and all were free of any infections in the previous month before the study. Liver diseases (specifically tumoral disease and HCV infection) and thyroid dysfunction were exclusion criteria. All subjects gave written informed consent, validated and approved by the Ethics committee of the Hospital of Girona 'Dr Josep Trueta', after the purpose of the study was explained to them.

Induction of hyperthyroidism
Hyperthyroidism in rats was induced by chronic subcutaneous (SC) administration of l-thyroxine 232:2 (T 4 , 100 µg/day, dissolved in 200 µL of saline; Sigma) for a period of three weeks (21 days), as previously described (Lopez et al. 2010, Gonzalez et al. 2012. Euthyroid (control) rats were treated with vehicle (saline).

Stereotaxic microinjection of T 3 and viral vectors
Rats were placed in a stereotaxic frame (David Kopf Instruments; Tujunga, CA, USA) under ketamine/xylazine anesthesia. Nuclei-specific injections were delivered via a permanent 28-gauge stainless steel cannula (Plastics One, Roanoke, VA, USA) inserted bilaterally either in the VMH or the arcuate nucleus of the hypothalamus (ARC), following stereotaxic coordinates: (a) for the VMH: −2.8 mm posterior to the bregma, ±0.6 mm lateral to bregma and 10.1 mm deep from the skull; (b) for the ARC: −2.8 mm posterior to the bregma, ±0.3 mm lateral to bregma and 10.2 mm deep from the skull. A catheter tube was connected from each infusion cannula to an osmotic minipump flow moderator (Model 1007D; Alzet Osmotic Pumps, Cupertino, CA, USA). These pumps had a flow rate of 0.5 µL/h during 7 days of treatment. The osmotic minipumps were inserted in a subcutaneous pocket on the dorsal surface created using blunt dissection (Imbernon et al. 2013, Contreras et al. 2014, Martins et al. 2016. Adenoviral GFP or constitutive active AMPKα isoforms (AMPKα-CA; Viraquest; North Liberty, IA, USA) vectors (Woods et al. 2000, Minokoshi et al. 2004, Lopez et al. 2008 were delivered in the VMH of rats using a 25-gauge needle (Hamilton; Reno, NV, USA) and the stereotaxic coordinates: −2.4 mm and −3.2 mm posterior to the bregma, ±0.6 mm lateral to bregma and 10.1 mm deep at a rate of 200 nL/min for 5 min for rat (1 µL/injection site) as previously reported (Lopez et al. 2008, Beiroa et al. 2014, Contreras et al. 2014, Martins et al. 2016. Animals were treated for 6 days.

Blood biochemistry
For the rat samples, plasma levels of T 3 and T 4 were measured using rat ELISA kits (Crystal Chem Inc; Downers Grove, IL, USA) (Lopez et al. 2010, Gonzalez et al. 2012. For the human samples, serum glucose concentrations were measured in duplicate by the glucose oxidase method using a Beckman Glucose Analyser II (Beckman Instruments; Brea, CA, USA). Roche Hitachi Cobas c711 instrument (Roche) was used to perform HDL cholesterol and total serum triglycerides determinations. HDL cholesterol was quantified by a homogeneous enzymatic colorimetric assay through the cholesterol esterase/cholesterol oxidase/peroxidase reaction (Cobas HDLC3; Roche). Serum fasting triglycerides were measured by an enzymatic, colorimetric method with glycerol phosphate oxidase and peroxidase (Cobas TRIGL; Roche). LDL cholesterol was calculated using the Friedewald formula. Serum free T 4 was measured by electrochemiluminescence (Roche Diagnostics) with intra-and inter-assay coefficients of variation less than 5%. Methods have been previously reported (Ortega et al. 2015, Gavalda-Navarro et al. 2016.

Sample processing
Rats were killed by cervical dislocation. From each animal, gonadal WAT (gWAT), subcutaneous inguinal WAT (sWAT) or both (only for the euthyroid and hyperthyroid animals) were harvested and immediately frozen in dry ice. Samples were stored at −80°C until further processing. Human adipose tissue samples were obtained from sWAT and vWAT depots during elective surgical procedures (cholecystectomy, surgery of abdominal hernia and gastric bypass surgery) (Ortega et al. 2015, Gavalda-Navarro et al. 2016. Samples of adipose tissue were immediately transported to the laboratory (5-10 min). Tissue handling was carried out under strictly aseptic conditions. Adipose tissue samples were washed in PBS, cut off with forceps and scalpel into small pieces (100 mg), and immediately flash-frozen in liquid nitrogen before storage at −80°C.

Real-time PCR
We performed real-time PCR (TaqMan; Applied Biosystems) as previously described (Lopez et al. 2010, Contreras et al. 2014, Alvarez-Crespo et al. 2016, Martins et al. 2016, using specific sets of primers and probes for rat (Supplementary Table 1, see section on 232:2 supplementary data given at the end of this article). Values were expressed relative to hypoxanthine-guanine phosphoribosyltransferase (HPRT) levels. For the analysis of the human WAT samples, we used commercially available and pre-validated TaqMan primer/probe sets (Applied Biosystems) as follows: endogenous control peptidylprolyl isomerase A (cyclophilin A) (PPIA, 4333763), PR domain containing 16 (PRDM16, Hs00223161_m1), uncoupling protein 1 (UCP1, Hs00222453_m1) and cell deathinducing DFFA-like effector a (CIDEA, Hs00154455_m1). Gene expression values were expressed relative to PPIA levels.

Histology and immunohistochemistry
Adipose tissue depots were fixed in 10% buffered formaldehyde. For the hematoxylin-eosin processing, the WAT sections were first stained with hematoxylin for 5 min, washed and stained again with eosin for 1 min. The detection of UCP1 in WAT was performed using anti-UCP1 (1:500; ab10983; Abcam) as previously reported (Alvarez-Crespo et al. 2016, Martins et al. 2016. The specificity of the UCP1 antibody has been previously validated by using WAT samples from UCP1 KO mice (Alvarez-Crespo et al. 2016). Images were taken with a digital camera Olympus XC50 (Olympus Corporation) at 20×. Digital images from WAT for immunohistochemistry were quantified with FRIDA image analysis software (FRIDA Software; The Johns Hopkins University; MD, USA); briefly, a color mask (pixel threshold masks) was set to define the UCP1 staining. This color mask was applied to all photographs, and the software obtained a numeric value proportional to the color level in each image. These values are represented with respect to control (100%). For the adipocyte area, images were analyzed with ImageJ Software (National Institutes of Health; MD, USA). Direct detection of GFP fluorescence was performed after perfusion of the animals and detected with a fluorescence microscope Olympus IX51, at 4×.

Statistical analysis
For the rat experiments, data are expressed as mean ± s.e.m. (error bars represent s.e.m.), mRNA and protein data were expressed in relation (%) to control (euthyroid, vehicletreated or GFP) rats. Statistical significance was determined by Student's t-test when two groups were compared or ANOVA followed by two-tailed Bonferroni post hoc test when more than two groups were compared. P < 0.05 was considered statistically significant. For the human studies, statistical analyses were performed using SPSS 12.0 software (IBM). Descriptive results of continuous variables are expressed as mean and s.d. for Gaussian variables or median and interquartile range unless otherwise stated. Parameters that did not fulfill normal distribution were mathematically Log transformed to improve symmetry for subsequent analyses. The relation between variables was analyzed by simple correlation (Pearson's test and Spearman's test) and by multivariate regression analysis. Levels of statistical significance were set at P < 0.05.

Hyperthyroidism induces browning of WAT in rats
T 4 -treated rats exhibited decreased weight gain (Fig. 1A) despite hyperphagia (Fig. 1B). Increased circulating levels of T 4 (Fig. 1C) and T 3 (Fig. 1D), confirmed their hyperthyroid status. Next, we analyzed whether hyperthyroidism induced browning of WAT in these animals. Our mRNA data showed that the mRNA expression of browning markers, such as UCP1, peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC1α), CIDEA, PRDM16 and also of uncoupling protein 3 (UCP3) was significantly increased in the gWAT (Fig. 1E) and sWAT (Fig. 1H) of hyperthyroid rats. Histological analysis of WAT showed that hyperthyroid rats exhibited a 'brown-like' multilocular pattern, associated with decreased adipocyte area ( Fig. 1F and I) and increased UCP1 immunostaining ( Fig. 1G and J) in both gWAT and sWAT.

Central T 3 induces browning of WAT in rats
Recent data have shown that the effect of THs on thermogenesis is centrally mediated (Lopez et al. 2010, Alvarez-Crespo et al. 2016. Therefore, we hypothesized that central chronic exposure of T 3 may stimulate browning of WAT. ICV T 3 administration induced a feeding-independent decrease in body weight ( Fig. 2A  and B). mRNA analysis of gWAT showed tendencies even though statistically non-significant for browning markers to be increased (Fig. 2C). Nevertheless, and more relevant, when histological analyses were assessed, our results were much clearer, indicating that ICV T 3 -treated rats displayed a 'brown-like' multilocular pattern, associated to decreased adipocyte area (Fig. 2D) and increased UCP1 immunostaining (Fig. 2E).

T 3 in the VMH, but not in the ARC, induces browning of WAT in rats
Next, we aimed to identify the hypothalamic nucleus where T 3 exerted its action on WAT. Therefore, we

232:2
brains (data not shown). When given into the VMH, T 3 promoted a feeding-independent weight loss ( Fig. 3A and B). On the other hand, when T 3 was administered into the ARC, there was a tendency to increase body weight at the end of the treatment, which was associated with hyperphagia ( Supplementary Fig. 1A and B). mRNA analysis of gWAT showed significantly increased or clear trends toward increased levels of browning markers when T 3 was delivered within the VMH (Fig. 3C), but not the ARC (Supplementary Fig. 1C). Again, histological analyses confirmed that VMH T 3 -treated rats displayed decreased adipocyte area (Fig. 3D) and increased UCP1 immunostaining (Fig. 3E) in gWAT, confirming browning.

Central effects of T 3 on browning of WAT depend on AMPK in the VMH
Next, we investigated the molecular mechanisms within the VMH leading to modulation of browning after central T 3 administration. Recent evidence has linked the inhibition of hypothalamic AMPK, and more specifically within the VMH, as a mechanism for the central regulation of BAT thermogenesis by THs (Lopez et al. 2010, Alvarez-Crespo et al. 2016). Based on this evidence, we hypothesized that the central effect of T 3 on browning might be mediated by specific inhibition of AMPK in the VMH. To test this, adenoviruses encoding either a constitutively active isoform of AMPKα (AMPKα-CA) or a GFP control vector were injected stereotaxically into the VMH of ICV T 3 -treated rats. The AMPKα-CA adenovirus was previously validated (Lopez et al. 2010, Beiroa et al. 2014, Martins et al. 2016).
Overexpression of AMPKα-CA in the VMH, confirmed by GFP immunofluorescence (Fig. 4A), blunted the weight loss caused by central T 3 injection, without alteration in feeding ( Fig. 4B and C). Of note, this effect was associated with the reversal of the T 3 -induced browning of gWAT, as demonstrated by increased adipocyte area (Fig. 4D) and decreased UCP1 staining (Fig. 4E) in T 3 -treated rats receiving AMPKα-CA adenoviruses in the VMH compared with those treated with control GFP adenoviruses. Together, these results are consistent with the observation that AMPK activity in the VMH mediates the central effects of T 3 on browning of WAT.

Browning markers in WAT are positively correlated with circulating T 4 in humans
Finally, we analyzed the relationship between T 4 serum concentrations and the mRNA expression levels of browning marker in sWAT and vWAT, in samples derived from a large cohort of patients. Our data showed that the mRNA levels of UCP1 and CIDEA in sWAT ( Fig. 5A and B, Supplementary Fig. 2) and PRDM16 in sWAT and vWAT (Fig. 5C and D), correlated with circulating free T 4 , showing a positive association between THs and browning in humans. Multivariate regression analysis indicated that serum free T 4 levels contributed significantly to browning-related (PRDM16, CIDEA and UCP1) mRNA levels variation after controlling for age, gender and BMI (Table 2).

Discussion
In this study, we show that THs induce browning of WAT in rodents and circulating T 4 levels correlate with the expression of browning markers in the WAT of humans. The effect of THs is centrally mediated involving specifically the VMH, a key nucleus modulating energy balance (Morrison et al. 2014, Contreras et al. 2015. Notably, this hypothalamic mechanism is mediated through AMPK, which has been described as a key factor regulating the actions of THs at the central level (Lopez et al. 2010, Alvarez-Crespo et al. 2016.
It has been known for more than a century that THs increase the basal metabolic rate (Magnus-Levy 1895). Typically, most of these effects have been related to the direct actions of THs on metabolically active tissues, such as the liver (Yen 2001), BAT (Bianco et al. 2005, Lopez et al. 2010, Ribeiro et al. 2010, Alvarez-Crespo et al. 2016, heart (Klein & Ojamaa 2001, Kahaly & Dillmann 2005 and skeletal muscle (Short et al. 2001). In those tissues, THs increase metabolic rate and thermogenesis by promoting the generation of energy and also by reducing the thermodynamic efficiency, which lead to heat production and increased temperature (Hulbert & Else 1981, Silva 2006, Lopez et al. 2013. The process in which precursor cells placed in WAT become beige/brite cells, instead of white adipocytes, is called browning (Fisher et al. 2012, Shabalina et al. 2013. Consequently, certain WAT depots significantly increase gene expression for UCP1 and their thermogenic capacity (Shabalina et al. 2013). Although the sWAT from the inguinal area is the most classical fat pad where browning studies have been performed, it has also been described in other depots, such as gonadal (Plum et al. 2007, Tews et al. 2013, Neinast et al. 2015, Contreras et al. 2016a, Fulzele et al. 2016, Jia et al. 2016, Lee et al. 2016, Martins et al. 2016, Shao et al. 2016. In this sense, it has been recently demonstrated that when centrally induced, browning affects gWAT in a similar extent to inguinal sWAT (Contreras et al. 2016a, Martins et al. 2016. However, despite the main thermogenic role of central THs, whether they are able to modulate the browning of WAT remains unclear. Here, we show that hyperthyroidism induces browning of WAT (sWAT from the inguinal area and gWAT) in rats. In our hyperthyroid model, T 4 was administered peripherally, which might imply the existence of direct effects of THs on white adipocytes, known to express TRs (Brent 2012, Lopez et al. 2013. Alternatively, THs may exert a central action after crossing the blood-brain barrier (BBB), which would be in agreement with recent evidence from our group, demonstrating that the metabolic effects of THs on brown fat are centrally mediated (Lopez et al. 2010, Alvarez-Crespo et al. 2016. Therefore, we investigated the contribution of the central effects of THs on browning of WAT. Our data show that, when administered centrally, T 3 promotes a similar pattern of browning of WAT as observed in the hyperthyroid model. Remarkably, the central action of T 3 targets one particular hypothalamic nucleus, the VMH. Indeed, stereotaxic administration of the hormone into the ARC (a neighboring nucleus) did not recapitulate the effects on the browning program induced by T 3 within the VMH. Considering that the VMH also plays a major role in the modulation of BAT function via THs (Lopez et al. 2010, Alvarez-Crespo et al. 2016, our data suggest that this hypothalamic site is a key modulator of both white and brown fat activity. Current evidence has demonstrated that inhibition of AMPK in the VMH plays a major role in mediating either the actions of THs on BAT (Lopez et al. 2010, Alvarez-Crespo et al. 2016 or the browning of WAT, for example by glucagon-like peptide 1 (GLP-1) agonism (Beiroa et al. 2014). To elucidate the contribution of hypothalamic AMPK activity on the browning of WAT by THs, we genetically activated AMPK in the VMH of rats centrally treated with T 3 . Our data showed that activation of AMPK totally blunted the effects of T 3 on WAT browning. Interestingly, this action was associated with a feeding-independent recovery of body mass, which was reduced by central T 3 . This evidence suggests that the augmented thermogenic capacity of brite adipocytes  (Shabalina et al. 2013) participates together with the BAT-mediated action in the weight-reducing effects of central T 3 .
Finally, we aimed to investigate whether browning markers correlate with circulating THs levels in humans. Remarkably, our results indicate that serum levels of T 4 are positively associated with UCP1, CIDEA and PRDM16 in WAT. To our knowledge, this is the first demonstration that THs modulate WAT browning in humans. Whether increased WAT browning is observed in hyperthyroid patients is not reported, but considering that THs also stimulate BAT in humans (Lahesmaa et al. 2014) and that most of the human BAT is actually beige fat (Jespersen et al. 2013, Shinoda et al. 2015, it is tempting to speculate that browning of WAT may account for the increased energy expenditure that characterizes hyperthyroidism (Warner & Mittag 2012, Lopez et al. 2013. In this sense, we have performed some preliminary studies in hyperthyroid patients and detected a trend in the correlation between UCP1 mRNA expression and T 4 ; however, further work will be necessary to properly investigate this association. If that is the case, strategies to modulate browning might be of therapeutic benefit in controlling the effects of thyrotoxicosis. This latter possibility is particularly relevant in the context of life-threatening conditions, such as thyroid storm, for which current treatments are not satisfactory. In addition, the induction of browning by TR agonism might be a suitable strategy for the treatment of obesity. In this regard, recent data have shown that treatment with the TR agonist GC-1 promotes browning of WAT and ameliorates obesity and diabetes in mice (Lin et al. 2015).
In summary, our results make evident the importance of THs in the browning of WAT in rodent and humans. This observation provides new insights into the physiological effects of THs and also in the pathogenesis of hyperthyroidism-induced effects on energy balance; it also suggests potential therapeutic strategies to counteract this disorder or other catabolic states.

Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1530/ JOE-16-0425.

Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

Funding
The research leading to these results has received funding from the European Community's Seventh Framework Programme (