Thyroid hormone status defines brown adipose tissue activity and browning of white adipose tissues in mice

The present study aimed to determine the effect of thyroid hormone dysfunction on brown adipose tissue activity and white adipose tissue browning in mice. Twenty randomized female C57BL/6NTac mice per treatment group housed at room temperature were rendered hypothyroid or hyperthyroid. In-vivo small animal 18F-FDG PET/MRI was performed to determine the effects of hypo- and hyperthyroidism on BAT mass and BAT activity. Ex-vivo14C-acetate loading assay and assessment of thermogenic gene and protein expression permitted analysis of oxidative and thermogenic capacities of WAT and BAT of eu-, hyper and hypothyroid mice. 18F-FDG PET/MRI revealed a lack of brown adipose tissue activity in hypothyroid mice, whereas hyperthyroid mice displayed increased BAT mass alongside enhanced 18F-FDG uptake. In white adipose tissue of both, hyper- and hypothyroid mice, we found a significant induction of thermogenic genes together with multilocular adipocytes expressing UCP1. Taken together, these results suggest that both the hyperthyroid and hypothyroid state stimulate WAT thermogenesis most likely as a consequence of enhanced adrenergic signaling or compensation for impaired BAT function, respectively.

The present study aimed to determine the effect of thyroid hormone dysfunction on brown adipose tissue activity and white adipose tissue browning in mice. Twenty randomized female C57BL/6NTac mice per treatment group housed at room temperature were rendered hypothyroid or hyperthyroid. 14 C-acetate loading assay and assessment of thermogenic gene and protein expression permitted analysis of oxidative and thermogenic capacities of WAT and BAT of eu-, hyper and hypothyroid mice. 18 F-FDG PET/MRI revealed a lack of brown adipose tissue activity in hypothyroid mice, whereas hyperthyroid mice displayed increased BAT mass alongside enhanced 18 F-FDG uptake. In white adipose tissue of both, hyper-and hypothyroid mice, we found a significant induction of thermogenic genes together with multilocular adipocytes expressing UCP1. Taken together, these results suggest that both the hyperthyroid and hypothyroid state stimulate WAT thermogenesis most likely as a consequence of enhanced adrenergic signaling or compensation for impaired BAT function, respectively.

In-vivo small animal 18 F-FDG PET/MRI was performed to determine the effects of hypo-and hyperthyroidism on BAT mass and BAT activity. Ex-vivo
Thyroid hormones (TH) are intimately involved in the regulation of energy metabolism. Administration of TH to mammals leads to an increase in basal metabolic rate and thermogenesis 1 . The important role of TH in the regulation of body temperature homeostasis is perhaps best reflected by the cold or heat intolerance of animals and humans with hypothyroidism or hyperthyroidism, respectively 2 . Under conditions of thermoneutrality, core body temperature can be maintained in the absence of any thermoregulatory process (e.g. shivering). However, under conditions of cold exposure adaptive thermogenesis occurs through a compensatory increase in sympathetic nervous system tone (SNS) which, in concert with TH, accelerates energy expenditure (EE) and increases body temperature 3 .
Numerous studies primarily performed on animal models have investigated the mechanism of TH for activating brown adipose tissue (BAT). Findings from different TR knockout mice (e.g. TRα , TRβ , TRα 1+ m) and the use of isoform-selective agonists indicate specific roles for TH-adrenergic interactions in BAT and white adipose tissue(WAT) [4][5][6][7][8][9][10] . Early studies in thyroidectomized and hypothyroid rats demonstrated the relevance of BAT specific type II iodothyronine deiodinase Dio2 as a local and systemic source of 3,3′ ,5-triiodothyronine (T 3 ) under cold exposure and for the first time provided evidence for T 3 in the enhancement of the thermogenic response to sympathetic stimulation 11,12 . This suggests that the subsequent heat liberation is actually a result of synergism between norepinephrine (NE) and T 3 signaling 13 . The significance of the central nervous system in TH activation of BAT was demonstrated in rats in which central administration of T 3 , via the inhibition of hypothalamic AMP-activated protein kinase (AMPK), induces thermogenic gene expression in BAT through stimulation of sympathetic nervous system (SNS) activity 14 . Additionally, SNS stimulation results in activation of β -adrenergic receptors (β -AR) by NE which induces both the enhanced thermogenic activity of existing brown adipocytes and the recruitment of new cells to BAT depots 1 . Energy release is accomplished by the activity of UCP1, a BAT-specific mitochondrial uncoupling protein 1 . With regards to WAT, a distinct population of UCP1-positive adipocytes arises within subcutaneous and visceral depots of mice in response to cold exposure 15,16 and β 3 adrenergic receptor agonist treatment 15,17,18 . To date it remains a matter of controversy as to whether these cells, entitled "brite" or "beige" adipocytes, are produced from the conversion of existing white adipocytes and/or the differentiation of WAT resident progenitor cells, leading to a process called "browning" [19][20][21] .
There is preliminary evidence that suggests a role for TH in the induction of beige adipocytes in WAT. In human multipotent adipose-derived stem cells, 3-10 day treatment with T 3 induced UCP1 expression which was most pronounced during the differentiation phase and was dependent on the presence of thyroid hormone receptor β (TRβ ) 22 . Similarly, treatment of mice with the TRβ specific agonist GC-1 leads to induction of adaptive thermogenesis in subcutaneous WAT, increased energy expenditure, and restoration of cold tolerance in cold-intolerant mice independent from BAT activity 23,24 . Furthermore, in mice centrally administered T3 leads to the recruitment of BAT in inguinal WAT which is accompanied by increased energy expenditure and body temperature. Since the effect was abolished in UCP1 KO mice, the findings highlight the significance of UCP1 as a central relais modulating the central effects of TH 25 . Furthermore, a novel circuit for the control of TH feedback in the brain via the Liver X receptor α and β has recently been published 26 . LXRα β −/− mice display increased TH serum level combined with changes in expression of genes associated with TH synthesis and TH transport as well as prominent UCP1 staining in the subcutaneous adipose depot 26 . In a positron emission tomography (PET) study of a patient with a history of thyroid cancer, systemic administration of TH led to BAT activation in periumbilical WAT indicating that browning had occurred as a result of T 4 supplementation for 14 days 27 .
Despite the significance of TH for adaptive thermogenesis, there is currently limited understanding of whether and how different thyroid states, i.e. hypo-and hyperthyroidism, regulate BAT thermogenesis as well as the thermogenic activity of WAT. Therefore, the present study aimed to determine TH induced effects on (I) BAT activity

Results
Both hypo-and hyperthyroidism induce brown fat biogenesis. Comparison of total T 4 and free T 3 serum concentrations between eu-, hyper-and hypothyroid mice confirmed the hyperthyroid and hypothyroid state of the animals (Fig. 1A). This finding is further substantiated by the differential hepatic expression of the TH-responsive genes deiodinase 1 (Dio1), thyroxine-binding globulin (Tbg) and pyruvate dehydrogenase kinase 4 (Pdk4) (Fig. 1B). The mRNA level of Dio1 and Pdk4 in liver increased significantly upon T 4 -treatment and decreased significantly in hypothyroid animals. In addition, induction of Tbg, negatively regulated by TH, confirmed the hypothyroid state (Fig. 1B).
In view of the striking differences in core body temperature and body fat distribution between hyper-and hypothyroid mice, we next asked to which extent WAT depot contributes to TH induced thermogenesis. Consistent with the finding of increased fat mass, adipocytes from hypothyroid mice were larger in size with increased lipid accumulation in iWAT and gWAT when compared with hyper-and euthyroid mice ( Fig. 2A). In addition we observed moderate increases in thermogenic gene program in the white fat depots of both hypo-and hyperthyroid mice as compared to euthyroid littermates. Strikingly, Ucp1 gene expression was upregulated by about 40-fold in iWAT of hypo-and hyperthyroid mice compared to euthyroid controls. Moreover, we observed a significant increase in Ucp1 mRNA in gWAT, which is considered to have a much lower capacity to induce thermogenic gene programs (Fig. 2B). Of interest, the zinc finger protein Zfp423, which is an essential determinant of preadipocyte commitment 28,29 , was predominantly upregulated in WAT of hypothyroid mice (Fig. 2B).
Histological examination of the white fat depots revealed the presence of abundant islets with a distinctive multilocular appearance of adipocytes which were positively stained for UCP1 in both, hyper-and hypothyroid states (Fig. 3A). We then investigated whether the presence of brown fat-like cells has consequences for fatty acid oxidation. As demonstrated in Fig. 3B, the highest uptake of 14 C-acetate was found in both iWAT and gWAT of hyperthyroid mice in comparison with hypo-and euthyroid mice. This was accompanied by a significantly higher level of Ardb1 expression in iWAT of hyperthyroid mice, whereas the expression Ardb1 and Ardb3 in gWAT of hypo-and euthyroid mice did not change (Fig. 3C).
Hypothyroidism is associated with enhanced β 3 -adrenergic tone in BAT. Given the differences in core body temperature despite the activation of the thermogenic program in WAT at room temperature, we next asked whether this is due to changes in the thyroid-adrenergic axis induced by thyroidal dysfunction. Immunohistochemical staining of tyrosine hydroxylase, which is the rate-limiting enzyme for catecholamine synthesis 30 , revealed no differences in abundance in the BAT of hyperthyroid and hypothyroid mice (data not shown). Furthermore, while we did not observe changes in gene expression of the β -adrenergic receptor 1 (Ardb1) in BAT, there was a significant increase in gene expression of Ardb3 in BAT of hypo-vs hyperthyroid mice (Fig. 4A).
BAT is profusely innervated by sympathetic nerve terminals with norepinephrine (NE) acting via β -ARs 31 . Therefore, we next determined the concentrations of circulating NE and epinephrine in the experimental groups. Intriguingly, we found that concentrations of NE only increased in hypothyroid mice whereas epinephrine increased in both hyperthyroid and hypothyroid mice compared to euthyroid mice (Fig. 4B). The activity of the   Dio2 in hypothyroid BAT was 20-fold increased compared to hyperthyroid mice (p < 0.05). In hyperthyroid BAT Dio2 activity was significantly lower than in euthyroid controls (p < 0.01; Fig. 4C).

Different morphology and gene expression in BAT of hypo-and hyperthyroid mice.
In order to further unravel the discrepancies between hypothermia despite increased adrenergic signalling and WAT browning in hypothyroid mice we next investigated BAT metabolism.
Histological examination of interscapular BAT revealed gross differences in cell morphology. Whereas BAT of euthyroid mice contained mixed regions of white and brown adipocytes, the BAT of hypothyroid mice contained predominantly adipocytes with unilocular lipid droplets of intermediate size between WAT and BAT. Hyperthyroid BAT displayed a distinct morphology with a decreased cell size of the mainly multilocular adipocytes (Fig. 5A+B).
Gene expression analysis of thermogenic markers, including Ucp1, Fgf21, Cidea, Dio2 and Elovl3, revealed a remarkable collective overexpression in the hypothyroid BAT compared with hyperthyroid mice (Fig. 5C). However, the high induction of Ucp1 mRNA in hypothyroid mice was not reflected on the level of UCP1 protein expression. Furthermore, there was no difference in UCP1 protein expression between hyperthyroid and euthyroid mice (Fig. 5D). In contrast, in hyperthyroid BAT we found an increased activation of β -adrenergic signaling as demonstrated by higher gene expression of the hormone-sensitive lipase (Hsl) and adipose triglyceride lipase (Atgl; Fig. 5C) together with an increased Ser660-phosphorylation of HSL (Fig. 5D).
Volume and activity of BAT is distinctively regulated in hypo-and hyperthyroid mice. Given the magnitude of thermogenic gene expression in BAT of hypothyroid mice together with the increased stimulation of the TH-adrenergic axis, hypothyroid mice cannot compensate for hypothermia. Therefore we next addressed whether this is due to defects in BAT activity. First, to determine the oxidative capacity in BAT, we performed ex-vivo 14 C-acetate uptake assay and observed a significant increase in fatty acid oxidation in BAT of hyperthyroid mice (74.0 ± 14.0 dpm/count) as compared to hypothyroid mice (18.9 ± 1.95 dpm/count; p < 0.01). There was no significant difference between hypo-and euthyroid mice (47.2 ± 14.5 dpm/count; Fig. 6A). Second, we used a combined PET-and MR-based delineation of active BAT to investigate whether there are differences in BAT formation between the experimental mouse groups. Hyperthyroid mice contained a higher volume of active iBAT (0.06 ccm ± 0.01 ccm) when compared to euthyroid (0.04 ccm ± 0.01 ccm; p < 0.05) and hypothyroid mice (0.03 ccm ± 0.01 ccm; Fig. 6B). We then asked whether the increase in amount of BAT also correlates with increased BAT activity. Metabolically active BAT depots with symmetrical uptake of 18 F-FDG were observed in all three groups being most marked in the hyperthyroid group (Fig. 6C). The SUVR of 18 F-FDG accumulation in iBAT was significantly decreased in hypothyroid mice vs. euthyroid mice (8.9 ± 1.3 vs. 18.9 ± 3.8; p < 0.05; Fig. 6D). Furthermore, there was a trend towards an increased 18 F-FDG accumulation in hyperthyroid compared to euthyroid and hypothyroid mice mice (21.99 ± 5.0 vs. 18.9 ± 3.8 and 8.9 ± 1.3, p = 0.06, respectively).
Finally, to determine to which extent the MR based iBAT volume is metabolically activated, the respective 18 F-FDG avid iBAT with a SUV threshold of > 2 was adjusted to the MR based iBAT volume. Hyperthyroid mice presented with an active iBAT fraction of 94% the highest BAT activity. In contrast, hypothyroid mice showed only in 81% of the interscapular adipose tissue relevant metabolic activity. In iBAT of euthyroid mice we found that 92% is activated (Fig. 6E).

Discussion
The present study provides new evidence of how TH affects thermogenesis in both brown and white adipose depots. For the first time with regard to metabolic activity of BAT, we were able to identify a distinct profile in hyper-and hypothyroid mice at room temperature. First, regarding the amount of BAT, the in vivo 18 F-FDG PET/ MRI-estimated BAT volumes gave first evidence for an increased total amount of BAT in hyperthyroid mice compared to hypo-and euthyroid animals (Fig. 6B). Second, although hypo-and euthyroid animals contained detectable BAT, hyperthyroid mice displayed an increased metabolic activity as determined by 18 F-FDG uptake (Fig. 6C+D). In line with this observation, fatty acid oxidation activity in BAT was increased in hyperthyroid mice as determined by the uptake of 14 C-acetate (Fig. 6A). The opposite metabolic profile was found in hypothyroid animals, which display significantly lower uptake of 18 F-FDG and of 14 C-acetate in BAT when compared to hyperthyroid mice (Fig. 6A-D).
The observation of active BAT in all three experimental groups emphasizes the significance of thermal conditions for metabolic control irrespective of thyroidal state. As discussed by Nedergaard et al. under sub-thermal conditions (for mice 18-22 °C) BAT derived heat will be used to maintain body temperature 32 . Paradoxically, the low BAT activity in hypothyroid animals was associated with a strong increase of thermogenic genes (Ucp1, Dio2, Fgf21, Cidea, and Elovl3; Fig. 5) and with a significant upregulation of the adrenergic system, as evidenced by significant increases of Ardb3 and NE concentrations in hypothyroid animals (Fig. 4). Although NE per se is not an index of NE release or sympathetic tone, these data suggest an increase in norepinephrine outflow to the periphery as a compensatory response to maintain body temperature. This finding is principally in agreement with observations in cold exposed hypothyroid rodents 6,33 .
Interestingly, in iWAT and gWAT of hypothyroid mice we detected features of adipose tissue browning, evidenced by an increased expression of brown specific genes (Ucp1, Cidea, and Elovl3; Fig. 2) and a multilocular UCP1-positive phenotype (Fig. 3). Collectively, these data suggest a compensatory WAT browning as a response of decreased heat production due to BAT inactivity in hypothyroid mice. Indeed, decreased BAT thermogenesis in mice lacking bone morphogenetic protein (BMPR1A) is associated with increased circulating NE and a compensatory browning of iWAT and eWAT in a very similar fashion to the hypothyroid mice in the present study 34 . The idea of heat compensation is further substantiated by the observation of enriched expression of the zinc finger protein 423 (Zfp423) in WAT and BAT selectively in hypothyroid mice (Figs 2+ 5). Gupta et al. demonstrated that Zfp423+ endothelial cells undergo an endothelial-mesenchymal transition into Zfp423+ Pparg+ pericytes that serve as a pool of committed pre-adipocytes 29 . Thus, the increased Zfp423 expression in the adipose tissues of hypothyroid mice likely suggests the formation of beige adipocytes by recruitment and de-novo differentiation of progenitor cells as has been demonstrated by cold exposure or adrenergic β 3 -AR agonist treatment 28 .
The observation of browning of white adipose tissue was also made in white adipose tissues of hyperthyroid mice, where in particular, established markers for adipose tissue browning such as Ucp1, Fgf21, Cidea, and Elovl3 were upregulated (Fig. 2). In addition, hyperthyroid mice were characterized by a significant increase in Ardb1 expression in iWAT (Fig. 3C). Recent studies demonstrated that the absence of the β 1 -AR receptor impairs NE-induced de-novo brown adipogenesis in BAT 35 . Conversely, β 1 -AR transgenic mice are resistant to diet-induced obesity and display a high abundance of adipocytes expressing Ucp1 in WAT 36 . With the results gained in the present study we cannot conclude whether or not central effects of T3 contribute to the observed WAT browning. However, evidences from a recent study by Alvarez-Crespo et al. suggest a coordinated action of central hyperthyroidism on WAT browning. Thus, the authors for the first time demonstrated that central T3 administration in mice induces UCP1 dependent thermogenesis in BAT and browning of WAT 25 . Also, similar to the observed increased BAT amount in our cohort of hyperthyroid mice, Alvarez-Crespo reported BAT recruitment upon central T3 infusion even at thermoneutrality 25 . Notably, keeping in mind that the hyperthyroid state of our mouse cohort is rather mild it is even more interesting to observe the significant effects on BAT recruitment and WAT browning. Further work will be necessary to address the extent of centrally induced T3 effects, i.e. by the determination of hypothalamic AMPK and ACC in our model of hyperthyroidism 37 .
Collectively, our data provide first evidence for differential mechanisms contributing to WAT browning dependent upon altered thyroidal state. This is most apparent in hypothyroid mice that exhibit increased adipogenesis and de-novo differentiation of brown adipocytes potentially as a compensatory mechanism to hypothermia resulting from BAT inactivity. In hyperthyroid mice, it can be hypothesized that increased β -adrenergic activation contributes to WAT browning most likely by central effects of TH. However, it has to be emphasized that with the current data we cannot exclude potential non cell-autonomous contributions from the SNS to adipose tissue browning irrespective of thyroidal state. Thus, in order to rule out that temperature as opposed to thyroidal state is the underlying reason for WAT browning, further studies should be performed under temperatures where sympathetic stimulation of facultative thermogenesis is depressed i.e. thermoneutrality (30 °C) 32 .
Scientific RepoRts | 6:38124 | DOI: 10.1038/srep38124 Methods Mice. Nine week-old female C57BL/6NTac mice were purchased from Taconic Europe (Lille Skensved, Denmark). The mice were housed in pathogen-free facilities in groups of three at 22 °C with a 12:12-h dark-light cycle (lights on at 06:00 h). Guidelines were approved by the local authorities of the State of Saxony, Germany as recommended by the responsible local animal ethics review board (Regierungspräsidium Leipzig, TVV04/12, Germany). All experiments with mice were carried out according to the approved guidelines.
After one week of adaptation, induction of hyper-or hypothyroidism started at the age of ten weeks with n = 20 mice per experimental group. Mice were rendered hyperthyroid by application of L-thyroxine at a dose of 2 μ g/ml diluted in drinking water for 4 weeks 38 . Hypothyroidism was induced by feeding an iodine-free chow diet supplemented with 0.15% propylthiouracil (PTU, catalog TD 97061; Harlan Teklan, Madison, WI, USA) for 4 weeks. Euthyroid mice fed standard chow diet (Altromin GmbH, Lage, Germany) served as controls.
Body composition. Body weight was recorded once a week and after 4 weeks of treatment. Whole body composition (fat mass, lean mass and total body water) was determined in conscious mice by using nuclear magnetic resonance technology with an EchoMRI700 ™ instrument (Echo Medical Systems, Houston, TX, USA).
Ten animals per experimental group were measured. Rectal body temperature was measured in fed mice between 8:00-9:00 am. Mice were sacrificed by CO 2 inhalation at the age of 14 weeks.

F-FDG PET/MRI of BAT activation.
In order to investigate the influence of thyroidal status on BAT activity, small animal PET/magnetic resonance (MR) imaging studies were performed using a dedicated high-resolution scanner (nanoScan, Mediso Medical Imaging Systems, Hungary). Anaesthetized (induction at 4%, maintenance at 1.8% isoflurane in 60%/40% oxygen/air) mice were injected intraperitoneally with 14.5 ± 1.3 MBq 18 F-FDG followed by a PET/MR scan from 30-60 minute post-injection in list mode. The data were reconstructed dynamically into 5-min time frames (OSEM, 4 iterations, 6 subsets, MR-based attenuation and scatter correction).
A T1-weighted MR image (Gradient Echo sequence (GRE), TE = 3.2 ms (out phase), TE = 6.4 ms (in phase), TR = 15.0 ms, Flip Angle = 25°; NEX = 2) served for anatomical orientation with individual PET datasets as well as for threshold based delineation of BAT in the intrascapular region. The 18 F-FDG uptake was determined using mean standardized uptake values (SUV) in PET/MR based volumes of interest (VOIs) in the intrascapular BAT (iBAT) and liver. To rule out an impact of alterations in the physiological 18 F-FDG distribution associated with a hyperthyroid or hypothyroid state, 18 F-FDG uptake was accessed semi-quantitatively. Therefore, SUV ratios (SUVR) were calculated to normalize iBAT uptake to unspecific liver uptake 39,40 . Furthermore, iBAT volume was retrieved by a MR-based delineation of fatty tissue in the intrascapular region. Active iBAT volume was defined as iBAT with an increased 18 F-FDG uptake of SUV > 2. Subsequently, the fraction of active iBAT was calculated as percentage of active iBAT volume out of total adipose tissue volume delineated by MR in the interscapular region. The SUVR and active iBAT volume between the groups were compared using 1-way ANOVA, followed by Bonferroni's multiple comparison test, using Prism 6.0 (GraphPad, San Diego, Ca, USA). Values of p < 0.05 were considered significant. 14 C-acetate loading assay. To quantify the oxidative capacity of WAT and BAT in hyper-, hypo-, and euthyroid mice, we performed an ex-vivo 14 C-acetate loading assay that measures the incorporation of radioactively labeled acetate into triacyl glycerides (TAG) as described elsewhere 41  Deiodinase type 2 activity. Activity of type 2 deiodinase were determined in duplicates in BAT homogenates (100 μ g protein per 100 μ l reaction) as described 43,44 . Western Blot analysis. Western Blot analysis was performed on samples of inguinal WAT (iWAT), gonadal WAT (gWAT) and BAT as described previously 45 . The following antibodies and dilutions were used: mouse polyclonal anti-Ucp1 antibody (1:500; Merck Millipore; Darmstadt, Germany), pHSL and HSL(1:1000; Cell Signaling Technology, Danvers, USA). Blots were visualized by enhanced chemiluminescence (Pierce, Milwaukee, USA).

Ex-vivo
Quantitative real-time-PCR (qPCR). For quantification of gene expression qPCR was performed using the LightCycler System LC480 and LightCycler-DNA Master SYBR Green I Kit (Roche, Mannheim, Germany) as described previously 46 . Primer sequences are available on request. Gene expression was calculated by the delta-delta CT method using 36B4 as a reference gene 47 . Relative gene expression was calculated by setting the mean of the euthyroid control group to 1 and then calculating each individual value of the three groups of mice studied.

Statistical Analyses.
Data are shown as mean ± SEM. Statistical significance was determined by Student´s t-test for comparison of two experimental groups or by ANOVA for comparison of multiple conditions followed by Bonferroni's multiple comparison test, using Prism 6.0 (GraphPad, San Diego, Ca, USA). P values less or equal to 0.05 were considered to be significant.