The Deubiquitinase OTUB1 Is a Key Regulator of Energy Metabolism

Dysregulated energy metabolism is a major contributor to a multitude of pathologies, including obesity and diabetes. Understanding the regulation of metabolic homeostasis is of utmost importance for the identification of therapeutic targets for the treatment of metabolically driven diseases. We previously identified the deubiquitinase OTUB1 as substrate for the cellular oxygen sensor factor-inhibiting HIF (FIH) with regulatory effects on cellular energy metabolism, but the physiological relevance of OTUB1 is unclear. Here, we report that the induced global deletion of OTUB1 in adult mice (Otub1 iKO) elevated energy expenditure, reduced age-dependent body weight gain, facilitated blood glucose clearance and lowered basal plasma insulin levels. The respiratory exchange ratio was maintained, indicating an unaltered nutrient oxidation. In addition, Otub1 deletion in cells enhanced AKT activity, leading to a larger cell size, higher ATP levels and reduced AMPK phosphorylation. AKT is an integral part of insulin-mediated signaling and Otub1 iKO mice presented with increased AKT phosphorylation following acute insulin administration combined with insulin hypersensitivity. We conclude that OTUB1 is an important regulator of metabolic homeostasis.


Introduction
Metabolic perturbations are associated with many common human diseases, including obesity, heart failure, diabetes and cancer [1,2]. Pharmaceutical targeting of the bioenergetic metabolism is a current focus in the development of novel treatment option for many of these diseases [2]. For the identification of possible unknown therapeutic targets, it is of key importance to improve our understanding of metabolic homeostasis as well as of the underlying signaling pathways and proteins.
The ubiquitin system conjugates ubiquitin to cellular target proteins via an elaborate functional interaction of Ub-conjugating enzymes (E1s, E2s, and E3s) [3]. Proteins of virtually all known signaling pathways are regulated by the ubiquitin system, which hence is an important contributor to cell and tissue homeostasis [3]. Poly-Ub chains can differ in the amino acid residue(s) used to interconnect Ub proteins within the chains (either via the Ub N-terminal methionine or through the seven different lysine residues), leading to distinct three-dimensional chain structures and different downstream effects [4]. K48-linked polyubiquitination is the most abundant Ub modification in cells, targeting modified proteins

Induced Otub1 Ablation Decreases Weight Gain and Fat Mass in Mice
Assessing possible gross phenotypes in adult mice with induced whole-body Otub1 deletion (Otub1 iKO), Otub1 was ablated by tamoxifen gavage over five consecutive days and body weight (BW) progression was monitored for 122 days. The achieved KO efficiency on mRNA level was 94% in pgWAT, 77% in iBAT, 92% in liver tissue, 98% in quadriceps, 90% in lung tissue, 67% in the spleen and 96% in heart tissue (each as average from three animals) [22]. Mice without tamoxifen treatment (expressing Otub1; control) gained approximately 20 g within this time frame, while Otub1 iKO mice only gained approximately 11.5 g ( Figure 1A). The body composition of the same mice was assessed by EchoMRI to analyze possible causes of the reduced weight gain. The fat/lean mass ratio was decreased 3 and 4.5 months after Otub1 KO, indicating that the amount of adipose tissue was less in Otub1 iKO mice compared to control ( Figure 1B). Analyzing various excised adipose tissues, we found that subcutaneous white adipose tissue (sWAT) weight was significantly lower in Otub1 iKO mice and perigonadal (pgWAT) WAT showed a strong tendency towards being decreased ( Figure 1C). Interscapular brown adipose tissue (iBAT) and liver weight ( Figure 1C) as well as quadriceps, spleen and kidney weight ( Figure S1A) were comparable between the groups, and liver tissue showed no gross histological difference ( Figure S1B).
In summary, these results suggest that the reduced BW gain in Otub1 iKO mice was mainly caused by decreased WAT accumulation during aging.  (control) or the induction of Otub1 KO in the whole body by tamoxifen treatment (Otub1 iKO) in mice during months 1-5 (n = 5). Vehicle or tamoxifen treatment by gavage was indicated by arrows. (B) Fat mass to lean mass ratio analyzed by EchoMRI in the mice shown in (A) at the indicated time after Otub1 KO (4.5 and 6 months of age; n = 5). (C) Weight of indicated organs (6 months after Otub1 KO; 8.5-month-old mice; 2.5-month-old mice were gavaged with tamoxifen; control, n = 10; Otub1 iKO, n = 9). sWAT, subcutaneous WAT; pgWAT, perigonadal WAT; iBAT, interscapular brown AT. All data are presented as the mean ± SEM. *, p < 0.05; **, p < 0.01; ***, p < 0.001 by two-way ANOVA with Bonferroni post-test (A) or two-tailed Student's t-test (B,C).

Loss of Otub1 Leads to a Hypermetabolic State In Vivo
Next, we analyzed possible underlying causes of the observed BW and fat mass decrease. Food and water intake showed a tendency to be increased in Otub1 iKO mice ( Figure 2A). In agreement with a higher food and water intake, feces weight and urine volume were also increased ( Figure 2A). The body temperature showed a tendency to be higher in mice lacking Otub1 ( Figure 2B). O 2 consumption, CO 2 production and thus energy expenditure were increased during the light cycle in Otub1 iKO mice ( Figure 2C). The RER was not altered ( Figure 2C), indicating that the oxidized nutrients were the same between the groups. In the dark cycle, corresponding (albeit not significant) results were observed to the light cycle ( Figure S1C). Overall, these data suggest a hypermetabolic state in Otub1 iKO mice.  (C) O2 consumption, CO2 production, energy expenditure (EE) and respiratory exchange ratio (RER) of 3 consecutive days during the light cycle (5 months after Otub1 deletion; 6.5-month-old mice; n = 5). All data are presented as the mean ± SEM. *, p < 0.05; **, p < 0.01 by two-tailed Student's t-test.

Otub1 Deletion Increases AKT-Dependent Signaling
To analyze possible molecular mechanisms underlying the observed changes in energy metabolism, mouse embryonic fibroblasts (MEFs) were generated from wild-type (Wt) mice and mice with constitutive Otub1 deletion (Otub1 −/− ; Figure S2A). Despite the high cellular expression of OTUB1, total levels of Ub chains, K48-linked and K63-linked Ub chains were not altered in MEFs ( Figure 3A). Therefore, we concluded that Otub1 deletion did not affect general levels of Ub chains and we focused on the analysis of specific signaling pathways.
OTUB1 has previously been linked to the regulation of IL-15-stimulated phosphorylation of the serine/threonine kinase AKT in CD8 + T cells and NK cells [24] as well as to mammalian target of rapamycin (mTOR) activity through regulation of DEPTOR in vitro [25]. AKT/mTOR signaling is a major determinant of cellular energy metabolism [26,27]. Hence, we analyzed whether OTUB1 affected AKT and/or mTOR in MEFs. In Otub1 −/− MEFs, both AKT Thr308 and Ser473 phosphorylation were increased under basal conditions ( Figure 3B,C). Interestingly, combined depletion of FCS and amino acids ( Figure 3B), or FCS depletion alone ( Figure 3C), strongly reduced or abolished AKT phosphorylation in Wt cells. In Otub1 −/− cells, AKT phosphorylation was detected in the absence of a stimulus ( Figure 3B,C) and it was further enhanced in comparison to Wt cells after re-addition of FCS to stimulate AKT phosphorylation ( Figure 3B). (C) O 2 consumption, CO 2 production, energy expenditure (EE) and respiratory exchange ratio (RER) of 3 consecutive days during the light cycle (5 months after Otub1 deletion; 6.5-month-old mice; n = 5). All data are presented as the mean ± SEM. *, p < 0.05; **, p < 0.01 by two-tailed Student's t-test.
Hypermetabolism can be caused by hyperthyroidism [23], and plasma levels of thyroid-stimulating hormone (TSH) are commonly used as primary indicator for thyroid dysfunction [23]. However, plasma TSH levels were not altered between the groups ( Figure S1C), indicating a normal thyroid function. Thus, we excluded hyperthyroidism as a possible cause of the observed hypermetabolic state.

Otub1 Deletion Increases AKT-Dependent Signaling
To analyze possible molecular mechanisms underlying the observed changes in energy metabolism, mouse embryonic fibroblasts (MEFs) were generated from wild-type (Wt) mice and mice with constitutive Otub1 deletion (Otub1 −/− ; Figure S2A). Despite the high cellular expression of OTUB1, total levels of Ub chains, K48-linked and K63-linked Ub chains were not altered in MEFs ( Figure 3A). Therefore, we concluded that Otub1 deletion did not affect general levels of Ub chains and we focused on the analysis of specific signaling pathways.  OTUB1 has previously been linked to the regulation of IL-15-stimulated phosphorylation of the serine/threonine kinase AKT in CD8 + T cells and NK cells [24] as well as to mammalian target of rapamycin (mTOR) activity through regulation of DEPTOR in vitro [25]. AKT/mTOR signaling is a major determinant of cellular energy metabolism [26,27]. Hence, we analyzed whether OTUB1 affected AKT and/or mTOR in MEFs. In Otub1 −/− MEFs, both AKT Thr308 and Ser473 phosphorylation were increased under basal conditions (Figure 3B,C). Interestingly, combined depletion of FCS and amino acids ( Figure 3B), or FCS depletion alone ( Figure 3C), strongly reduced or abolished AKT phosphorylation in Wt cells. In Otub1 −/− cells, AKT phosphorylation was detected in the absence of a stimulus ( Figure 3B,C) and it was further enhanced in comparison to Wt cells after re-addition of FCS to stimulate AKT phosphorylation ( Figure 3B).
To investigate mTOR activity, the phosphorylation status of the ribosomal protein S6 kinase (S6K) Thr389 was analyzed, a direct target of mTOR complex 1 (mTORC1) [26]. Under baseline and combined FCS and amino acids depletion conditions, the phosphorylation of S6K Thr389 was comparable between Wt and Otub1 −/− cells ( Figure 3B). However, following stimulation by FCS re-introduction, the phosphorylation of S6K was reduced in Otub1 −/− relative to Wt MEFs ( Figure 3B), suggesting a decreased mTOR activity. OTUB1 has been reported to directly regulate DEPTOR protein stability [25], a negative regulator of mTORC1 and mTORC2 [28]. In Otub1 −/− MEFs, DEPTOR levels were increased ( Figure  S2B), inconsistent with a direct regulation of the DEPTOR protein half-life by OTUB1, but consistent with the observed decrease in mTOR activity.
To analyze the relevance of the OTUB1-dependent AKT regulation, MEF cell size and metabolic energy status were determined. We observed an increase in cell size in Otub1 −/− MEFs ( Figure 3C,D), in agreement with increased AKT activity [29,30]. Furthermore, ATP levels were elevated ( Figure 3E) and AMPK phosphorylation was downregulated ( Figure 3F), consistent with increased AKT activation [31].

Otub1 Deletion Increases Insulin Sensitivity
AKT can be activated by insulin and AKT activity plays a key role in insulin-regulated metabolism [32]. Therefore, we next assessed insulin function in Otub1 iKO mice. In an intraperitoneal glucose tolerance test (IPGTT), Otub1 iKO mice showed a decreased blood glucose peak and faster glucose clearance compared with control mice ( Figure 4A). An intraperitoneal insulin tolerance test (IPITT) confirmed increased insulin sensitivity in Otub1 iKO mice ( Figure 4B). Furthermore, basal plasma insulin levels (after 4 h of fasting) were decreased in Otub1 iKO mice ( Figure 4C). This increased insulin sensitivity in Otub1 iKO mice is in agreement with the increased AKT activity found in MEFs. lated metabolism [32]. Therefore, we next assessed insulin function in Otub1 iKO mice. In an intraperitoneal glucose tolerance test (IPGTT), Otub1 iKO mice showed a decreased blood glucose peak and faster glucose clearance compared with control mice ( Figure 4A). An intraperitoneal insulin tolerance test (IPITT) confirmed increased insulin sensitivity in Otub1 iKO mice ( Figure 4B). Furthermore, basal plasma insulin levels (after 4 h of fasting) were decreased in Otub1 iKO mice ( Figure 4C). This increased insulin sensitivity in Otub1 iKO mice is in agreement with the increased AKT activity found in MEFs.

Otub1 Deletion Increases Insulin-Dependent Signaling
Next, we investigated insulin-dependent intracellular signaling in vivo. In the liver, no difference was observed in the phosphorylation of AKT or S6K under basal conditions (analyzed 4 h after fasting; Figures 5A and S6A). There was also no difference in AKT phosphorylation in the heart under basal conditions, whereas S6K phosphorylation was increased ( Figure S6B). However, the same mice had shown decreased basal insulin levels ( Figure 4C). The maintained or increased AKT/mTOR signaling in the presence of decreased basal insulin levels indicated an enhanced sensitivity of AKT/mTOR signaling to insulin-dependent stimulation when Otub1 was absent. Consequently, we next analyzed the response of AKT/mTOR signaling to acute insulin stimulation. Therefore, mice were fasted overnight followed by insulin injection. Fifteen minutes after injection, the phosphorylation of AKT on Ser473 was increased in livers of Otub1 iKO mice ( Figure 5B). The effect on AKT Thr308 was less clear (Figures 5B and S6C). The phosphorylation of S6K was not altered, but total levels of S6K showed a tendency to be decreased in both basal as well as stimulated conditions ( Figure 5), indicating that additional regulatory pathways may be involved. In the hearts of the same animals, phosphorylation of AKT Ser473 was at least in some animals increased in response to insulin stimulation, which could also be observed for the phosphorylation of S6K ( Figure S6D).
The phosphorylation of ERK1/2, another component of the insulin signaling pathway [33], was not altered in the liver under basal conditions ( Figure 5A). In response to insulin stimulation, ERK1/2 phosphorylation was increased in Otub1 iKO livers ( Figure 5B), indicating that the OTUB1-dependent regulation of insulin signaling was either not confined to AKT or occurred upstream of the bifurcation point of both signaling cascades. The phosphorylation of ERK1/2 showed a tendency to be decreased in Otub1 iKO mice hearts

Otub1 Deletion Increases Insulin-Dependent Signaling
Next, we investigated insulin-dependent intracellular signaling in vivo. In the liver, no difference was observed in the phosphorylation of AKT or S6K under basal conditions (analyzed 4 h after fasting; Figures 5A and S6A). There was also no difference in AKT phosphorylation in the heart under basal conditions, whereas S6K phosphorylation was increased ( Figure S6B). However, the same mice had shown decreased basal insulin levels ( Figure 4C). The maintained or increased AKT/mTOR signaling in the presence of decreased basal insulin levels indicated an enhanced sensitivity of AKT/mTOR signaling to insulin-dependent stimulation when Otub1 was absent. Consequently, we next analyzed the response of AKT/mTOR signaling to acute insulin stimulation. Therefore, mice were fasted overnight followed by insulin injection. Fifteen minutes after injection, the phosphorylation of AKT on Ser473 was increased in livers of Otub1 iKO mice ( Figure 5B). The effect on AKT Thr308 was less clear (Figures 5B and S6C). The phosphorylation of S6K was not altered, but total levels of S6K showed a tendency to be decreased in both basal as well as stimulated conditions ( Figure 5), indicating that additional regulatory pathways may be involved. In the hearts of the same animals, phosphorylation of AKT Ser473 was at least in some animals increased in response to insulin stimulation, which could also be observed for the phosphorylation of S6K ( Figure S6D). following insulin stimulation, supporting a tissue-specific effect of OTUB1 on insulin-dependent signaling ( Figure S6D). In summary, Otub1 deletion led to increased insulin-dependent signaling under basal and stimulated conditions, which likely contributed to the hypermetabolic phenotype.  The phosphorylation of ERK1/2, another component of the insulin signaling pathway [33], was not altered in the liver under basal conditions ( Figure 5A). In response to insulin stimulation, ERK1/2 phosphorylation was increased in Otub1 iKO livers ( Figure 5B), indicating that the OTUB1-dependent regulation of insulin signaling was either not confined to AKT or occurred upstream of the bifurcation point of both signal-ing cascades. The phosphorylation of ERK1/2 showed a tendency to be decreased in Otub1 iKO mice hearts following insulin stimulation, supporting a tissue-specific effect of OTUB1 on insulin-dependent signaling ( Figure S6D). In summary, Otub1 deletion led to increased insulin-dependent signaling under basal and stimulated conditions, which likely contributed to the hypermetabolic phenotype.

Discussion
Our results demonstrate that the DUB OTUB1 is essential for the homeostasis of energy metabolism and for the regulation of insulin-dependent signaling. Otub1 −/− cells showed increased ATP levels and decreased AMPK phosphorylation, which is in agreement with our previous in vitro observation that OTUB1 overexpression elevates AMPK phosphorylation [18]. The here presented results suggest that OTUB1 defines a set-point for whole-body glucose metabolism in mice, with Otub1 deletion leading to increased glucose utilization and a hypermetabolic state.
In this study, tamoxifen was used to induce the deletion of Otub1. Tamoxifen had previously been reported to transiently affect body weight, body composition and glucose metabolism in mice; however, such investigations were generally performed within short time frames after tamoxifen treatment (days to a few weeks) [34,35]. All of the here described in vivo results (except the analysis of insulin-mediated signaling) were obtained at least 2.5 months after tamoxifen treatment, at which time tamoxifen does not affect murine metabolism any longer (if such an effect occurs) [35,36]. Moreover, the animals in our investigations obtained tamoxifen via gavage and oral tamoxifen application does not affect glucose metabolism [34]. In addition, the observed in vivo effect of tamo-xifeninduced Otub1 deletion on cellular signaling was reproducible in cells in vitro, which were never treated with tamoxifen (constitutive Otub1 deletion). Thus, the here described alterations were not caused by tamoxifen treatment but by Otub1 deletion.
Otub1 iKO mice presented with an increased energy expenditure combined with a tendency towards a higher body temperature. This can be explained by a simultaneous increase in catabolic and anabolic metabolism and/or by mitochondrial uncoupling, which had previously been reported for Hif1an KO mice [19]. Future investigations will therefore need to analyze whether deletion of Otub1 affects protein levels of enzymes that are involved in mitochondrial uncoupling, such as uncoupling proteins (UCPs). Furthermore, ATP levels were increased in MEFs, indicating an increase in catabolic cellular energy metabolism, which is in agreement with the observed increase in food intake in mice with Otub1 deletion. Interestingly, the RER showed no difference, supporting an increased utilization of the same nutrients through the same metabolic pathways in Otub1 iKO mice compared to control.
Lack of Otub1 generally enhanced AKT signaling in MEF cells and in the investigated organs. AKT signaling is known to affect cellular ATP levels and AMPK phosphorylation [31], and to regulate cell size [29,30]. In adult mice, increased AKT signaling leads to decreased body weight and fat mass, increased oxygen consumption, improved glucose clearance and insulin sensitivity as well as decreased serum insulin levels [37,38]. Therefore, we conclude that OTUB1 is a negative regulator of basal and stimulated AKT signaling in cells and tissues, and that an increased AKT signaling at least partially contributes to the observed phenotypes in Otub1 iKO mice.
Following stimulation by IL-15, OTUB1 has been reported to directly regulate AKT K63-linked ubiquitination in CD8 + T cells and NK cells, which is necessary for AKT membrane recruitment and the subsequent phosphorylation of AKT [24]. However, it remained unclear how exactly OTUB1 regulated IL-15-mediated AKT ubiquitination. It has previously been reported that the E2 enzymes UBC13 and/or UBCH5c are necessary for AKT ubiquitination following stimulation with insulin or epidermal growth factor (EGF) [39,40]. Both E2s can be inhibited by OTUB1 [9,12]. Hence, OTUB1 may be a negative regulator of insulin-dependent AKT ubiquitination by inhibiting UBC13 and/or UBCH5c.
Interestingly, the protein WTG1/OsOTUB1, a homologue of mouse and human OTUB1, regulates grain size and weight in rice [41][42][43]. This effect was linked to the regulation of sucrose metabolism and starch biosynthesis [41], indicating that the regulation of carbohydrate metabolism by OTUB1 was conserved during evolution.
We previously reported that the oxygen sensor FIH regulates cellular energy metabolism through hydroxylation of OTUB1 [18]. Deletion of Hif1an in mice led to a metabolic phenotype, but the underlying molecular mechanisms remained to be elucidated [19]. Remarkably, the reported phenotype of the Hif1an KO mice is virtually identical with our observations in Otub1 iKO mice (decreased body weight gain, decreased fat mass, increased insulin sensitivity, increased food and water intake, increased O 2 consumption, CO 2 production and energy expenditure, a maintained RER, increased intracellular ATP levels, decreased pAMPKα levels), leading to the hypothesis that OTUB1 is a physiologically relevant target of FIH.
In summary, OTUB1 plays an essential role for the regulation of cellular and wholebody energy metabolism, at least in part through the regulation of AKT activity and insulin signaling. Our results highlight the physiological relevance of this DUB, indicate that OTUB1 is a relevant target protein of FIH in vivo and suggest that OTUB1 may be a novel therapeutic target for the treatment of type 2 diabetes, which needs to be investigated further in the future.

Mice
To obtain mice with induced global Otub1 deletion, mice were obtained from the EUCOMM consortium http://www.mousephenotype.org (accessed on 4 January 2022) [44] via the EMMA repository https://www.infrafrontier.eu/ (accessed on 4 January 2022) and provided by the Institut Clinique de la Souris, ICS, Illkirch, France) as well as from The Jackson Laboratory (www.jax.org/, accessed on 4 January 2022), and crossed as previously described [22]. For our analyses, female mice (littermates) were randomly selected and gavaged with vehicle or 200 µg tamoxifen/g (Sigma-Aldrich, St. Louis, MO, USA) at 1.

Cell Culture
Mouse embryonic fibroblasts (MEFs) were isolated from wild-type (C57BL/6N, Charles River, MA, USA) and Otub1 −/− mice (obtained by crossing of heterozygous C57BL/6N-Otub1 tm1b(EUCOMM)Hmgu/H mice; purchased via the EMMA repository, from the EUCOMM consortium) at day 12.5 of gestation. The Mary Lyon Centre at MRC Harwell (http: //www.har.mrc.ac.uk, accessed on 4 January 2022; UK) provided Mouse Contract Services. MEFs were kept in culture until they spontaneously immortalized. MEFs were maintained in high-glucose DMEM containing 10% heat-inactivated fetal calf serum (FCS; Gibco by Life Technologies, Carlsbad, CA, USA), 1% GlutaMAX (Gibco), 200 µM 2-mercaptoethanol (Gibco), 100 µg/mL streptomycin and 100 U/mL penicillin (Sigma-Aldrich) in a humidified atmosphere at 37 • C with 5% O 2 and 5% CO 2 . For FCS deprivation, MEFs were grown in DMEM as described above, but in the absence of FCS. For the reintroduction of FCS, the medium was exchanged with the same DMEM, but including 10% FCS. For combined FCS and amino acid deprivation, MEFs were kept in RPMI 1640 medium containing glucose but no FCS or amino acids for 4 h. To reintroduce FCS, the medium was exchanged with RPMI 1640 containing 10% FCS and the cells were harvested at the indicated time points. Cells were regularly tested for potential contamination with mycoplasma.

Protein Extraction
Lysis buffer for protein extraction contained 150 mM NaCl, 1 mM EDTA, 25 mM Tris-HCl pH 8.0, 1% NP-40, 1 mM Na 3 VO 4 , 1 mM PMSF and 1 mM NaF. For lysis of cells, protease inhibitor cocktail (Sigma-Aldrich) was included in the lysis buffer. For lysis of tissue, 1 µg/mL pepstatin A, 1 µg/mL aprotinin and 1 µg/mL leupeptin were added to the lysis buffer. A polytron homogenizer (VWR International, Amsterdam, The Netherlands) was used to homogenize tissue samples. For the analysis of poly-ubiquitin chains, the deubiquitinase inhibitor [45] N-ethylmaleimide (50 mM; Sigma-Aldrich) was added to the lysis buffer. The BCA assay (Thermo Fisher Scientific, Waltham, MA, USA) was used to determine protein concentrations.

ATP Assay
Intracellular ATP levels were measured with the CellTiter-Glo 2.0 Assay according to the manufacturer's description (Promega, Madison, WI, USA). Cells were lysed with Passive Lysis Buffer (Promega) and diluted 1:100 in cell culture media. Samples were mixed with equal amounts of CellTiter-Glo 2.0 reagent for 2 min on an orbital shaker followed by additional 10 min of incubation at room temperature. Luminescence was determined in a microplate luminometer (Berthold Technologies, Bad Wildbach, Germany) and normalized to protein concentration determined by the Bradford method [46].

Intraperitoneal Glucose and Insulin Tolerance Tests
After fasting for 16 h over night, 2 mg of glucose/g (20% w/v; Braun, Hessen, Germany) or 0.75 IU/kg insulin (Lilly Humalog, Indianapolis, IN, USA) were intraperito-neally injected. Blood samples from the tip of the tail were analyzed after the indicated time points using a StatStrip Xpress glucose meter (Nova Biomedical).

ELISAs
Basal insulin levels were determined in plasma taken from mice starved for 4 h during the light cycle using an anti-insulin enzyme-linked immunosorbent assay (ELISA) following the manufacturer's description (Mercodia, Uppsala, Sweden). Thyroid-stimulating hormone (TSH) plasma values were analyzed with an anti-TSH ELISA following the manufacturer's description (Cloud-clone Corporation, Katy, TX, USA). Absorbance was determined with a microplate reader (Infinite 200 Pro, Tecan, Maennedorf, Switzerland).

Body Composition Analysis
An EchoMRI whole-body composition analyzer was used to analyze lean and fat mass non-invasively according to the manufacturer's instructions (Echo Medical systems, Houston, TX, USA).

Indirect Calorimetry
Animals were single housed in acrylic glass airtight metabolic cages attached to an open calorimetry system (TSE Systems, Bad Homburg, Germany). Food and water were provided ad libitum. O 2 consumption and CO 2 production were measured during 3 days and the average was calculated. Energy expenditure was calculated according to the Weir equation and normalized to the corresponding lean mass + 0.2 times the fat mass, as determined by body composition analysis by EchoMRI [48,49].

Food/Water Intake, Urine/Feces Production
Mice were single housed in metabolic cages with grid surfaces (Techniplast, Schwerzenbach, Switzerland) and food, water, urine, feces and BW were monitored every 24 h for 72 h. Urine was collected under light mineral oil (Sigma-Aldrich) to avoid evaporation.

Statistical Analysis
Two-tailed Student's t-test or Mann-Whitney test were applied for the analysis of two different data points as indicated. Comparison of more than two data points were carried out by two-way ANOVA followed by Bonferroni post-test.