GLOBAL LOSS OF SELENOCYSTEINE LYASE IN MICE DRIVES LIPID ACCUMULATION IN BROWN ADIPOCYTES

Abstract

The enzyme selenocysteine (Sec) lyase (SCLY) decomposes Sec releasing selenide for the synthesis of selenoproteins, which contain Sec in their primary structure and participate in strong redox reactions, maintaining redox balance. We previously showed global disruption of the Scly gene (Scly KO) in mice leads to obesity. Targeted deletion of Scly in Agrp neurons enhances energy expenditure and brown adipose tissue (BAT) activation, augmenting leanness. We hypothesized that Scly KO mice develop obesity due to failure of BAT-controlled mechanisms of energy expenditure due to redirection of Sec to an alternative pathway. We analyzed BAT from male Scly KO mice on Se-adequate (0.25 ppm) and Se-deficient (0.08 ppm) diets for morphology, Se content, selenoprotein expression, thyroid hormones, and additional Sec-utilizing pathways. We found that BAT of Scly KO mice was enlarged, with lower Se levels, and substantial whitening on a Se-adequate diet. This phenotype worsened on low Se and coincided with a mild impairment in adapting to cold exposure. BAT whitening coincided with an increase in triglycerides and reduced 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) and cholesterol. BAT selenoproteins regulating energy metabolism DIO2, GPX1, and GPX4 were significantly decreased. DIO2 reduction corresponded with an increase in thyroxine (T4), thyroid stimulating hormone (TSH), and reduction in heat-producer uncoupling protein 1 (UCP1). Downregulation of GPX4 did not affect ferroptosis in the BAT. Therefore, the whitened BAT of the Scly KO mouse is a multifactorial process involving the disruption of BAT function through changes to selenoproteins involved in energy metabolism.

NEW AND NOTEWORTHY

• Global loss of the selenocysteine-decomposing enzyme selenocysteine lyase in mice leads to lipid accumulation and whitening of the brown adipose tissue, with consequent obesity development.

• Selenocysteine lyase modulates selenium levels and selenoprotein expression, specifically GPX1, GPX4 and DIO2, in brown adipocytes.

• Selenocysteine metabolic fate hinges on the actions of selenocysteine lyase.

INTRODUCTION

Selenium (Se) is an essential trace element participating in several energy homeostasis mechanisms in the liver, adipose tissue, and skeletal muscle^1–9. When Se levels are inadequate, it is postulated that selenocysteine (Sec) lyase (SCLY) recycles Se to be used in selenoprotein synthesis by decomposing Sec to selenide. Previously, we showed that whole body Scly knockout (KO) mice develop obesity with hepatic impairment of energy metabolism when fed a Se-deficient diet or a high-fat diet in a sexually dimorphic manner^10,11. Strikingly, targeted deletion of Scly in mouse agouti-related peptide (Agrp) neurons, the hypothalamic cells regulating energy homeostasis, exhibited reduced weight gain and lipid deposition with stronger activation of brown adipose tissue (BAT) thermogenic mechanisms¹² indicating the obesity phenotype previously observed in the global Scly KO was not due to central metabolic regulation and instead a result of an impairment in peripheral energy expenditure control. Targeted brown adipocyte KO of Trsp, the gene encoding the Sec-specific tRNA (tRNA^[Ser]Sec) required for selenoprotein synthesis, resulted in whitening of brown adipocytes with enhanced triglycerides and mild thyroid hormone dysfunction, indicating the potential requirement for selenoproteins to maintain energy homeostasis in the BAT. Energy homeostasis is controlled by BAT, which functions as a metabolic sink, consuming excess glucose, lipids, and branched-chain amino acids (BCAA), burning calories through uncoupling respiration to generate heat^13,14. Studies in both rodent models and humans have shown disruptions in BAT function due to genetic ablation or obesity reduce BAT activity and the ability to metabolize fatty acids, glucose and BCAA’s^15–18. Yet, whether energy homeostasis is compromised by the loss of SCLY through BAT impairment leading to the obesity developed by the global Scly KO mouse model has not yet been investigated.

Multiple selenoproteins are involved in energy homeostasis. Key selenoproteins contributing to BAT energy metabolism include the type 2 iodothyronine deiodinase (DIO2) and the glutathione (GSH) peroxidases 1 and 4 (GPX1,4), with DIO2 being essential for the brown adipocyte response to thermogenesis^4,19. By supplying the active thyroid hormone triiodothyronine (T3), DIO2 actions upregulate pro-thermogenic genes including uncoupling protein 1 (UCP1) that will promote the generation of heat. GPX1 and GPX4 are involved in energy homeostasis as GPX activity depletes cells of GSH, with subsequent activation of thermogenesis through elevated reactive oxygen species (ROS)^20,21. Additionally, GPX4 is a master regulator of ferroptosis, a form of iron-dependent cell death²². Decreased levels of GPX4 increase the sensitivity of brown adipocytes to ferroptotic death and also potentially contribute to BAT whitening as has been previously reported²³. It is yet unknown whether the loss of SCLY impacts the synthesis of the selenoproteins involved in maintaining energy metabolism and redox balance in the BAT and what the fate of the Sec intermediate is when it cannot be decomposed by Scly.

Notably, the direct substrate of SCLY, Sec, is an analogue of cysteine with a Se in place of sulfur (S)²⁴. Cysteine and Sec can be decomposed by cysteine desulfurases (NFS1 in mammalians) to produce hydrogen sulfide (H₂S) or hydrogen selenide (H₂Se). These molecules can potentially serve as precursors for the biosynthesis of taurine or selenotaurine, glutathione (GSH) or selenoglutathione (GSeH)^25–28, processed for iron-sulfur (Fe-S) or Fe-Se cluster assembly²⁹, or metabolized for selenoprotein synthesis³⁰, a process that can also involve cysteine³¹. Finally, both cysteine and Sec can be metabolized to yield pyruvate^32,33, which in turn undergoes oxidative decarboxylation by the pyruvate dehydrogenase complex and produces acetyl-CoA^11,33. Acetyl-CoA is then carboxylated by acetyl-CoA carboxylase (ACC) to generate malonyl-CoA, the metabolic intermediate in fatty acid synthesis, which is further converted by fatty acid synthase (FASN) to form long-chain fatty acids³⁴ and combined with glycerol to form triglycerides³⁵. Acetyl-CoA is also directly involved in peroxisome proliferator activated receptor gamma (PPARγ)-driven lipid synthesis^36,37, or alternatively, acetyl-CoA can also be converted to HMG-CoA to generate cholesterol³⁸. Intriguingly, increased plasma levels of cysteine have been associated with obesity in humans^39,40. In animal models, a high cystine diet decreases energy expenditure, enhances visceral adiposity, and reduces glucose tolerance⁴¹ while supplementation with cysteine on methionine-restricted rats abrogated the effects of methionine restriction on weight gain, serum lipids, and adipokines⁴². Thus, multiple lines of evidence suggest that increased levels of cysteine may contribute to obesity. As Sec is an analog of cysteine, it is conceivable that loss of Scly may also shift Sec away from selenoproteins towards alternative, analogous sulfur pathways that contribute to obesity.

In this study, we tested the hypothesis that mice with global disruption of Scly develops obesity under Se deficiency due to failure of energy expenditure mechanisms centered on BAT. Since SCLY is not available to deliver Se for selenoprotein synthesis, Sec can be used instead in an alternative manner in the BAT affecting its thermogenic function. We found substantially altered interscapular BAT (iBAT) morphology in Scly KO mice resulting in a scenario where key selenoproteins involved in energy metabolism were significantly decreased and triglycerides were elevated, leading to BAT whitening and a mild impairment in adaptive thermogenesis, and potentially contributing to the obese phenotype in mice lacking Scly.

METHODS

Chemicals and antibodies.

All chemicals used in experiments were purchased from Fisher Scientific (Waltham, MA, USA) or Sigma-Aldrich (St. Louis, MO, USA), unless specified. Antibodies used in this study, their vendors and concentrations in various assays are presented in Supplemental Table 1.

Animals and Diets.

All animal procedures were approved by the University of Hawaii Institutional Animal Care and Use Committee (IACUC) protocols #17–2521 and #17–2616. Mice were group housed in the vivarium of each institution and used for experiments in minimal numbers needed to provide significant results. Room temperature in the University of Hawaii Vivarium was maintained at 23°C. For some experiments, starting at weaning age, mice were housed at 30°C in a temperature-controlled incubator (Model RIS52SD, Powers Scientific, Doylestown, PA, USA) to keep mice at thermoneutrality. Mouse with a whole-body deletion of the Scly gene (Scly KO) were on a C57BL6/N background and their development was previously described⁴³. Male mice were fed customized Se diets containing either 0.08 ppm (low) or 0.25 ppm (adequate/medium) sodium selenite from Envigo/Inotiv (Indianapolis, IN, USA) for at least 8 weeks. Euthanasia was carried out by a mixture of Avertin (250 mg/kg, Sigma-Aldrich) injection followed by CO₂ asphyxiation, and extracted tissues were snap-frozen, dipped in RNALater solution or perfused with 12% formalin for histological analysis.

Survival surgery for thermal probe insertion and cold exposure of mice.

After seven weeks on the experimental diet, mice were surgically implanted with thermal probes (G2 E-mitter; Starr Life Sciences Corp, Oakmont, PA, USA). Anesthesia was induced with 5% isoflurane in oxygen (1 L/min). Adequate depth of anesthesia was confirmed by lack of response to toe pinch and stable respiratory rate. Mice were placed on a heated surgical pad, and the abdominal region was sterilized with alternating applications of betadine and 70% isopropanol (three cycles total), followed by placement of a sterile drape. A single vertical incision was made to access the abdominal cavity, and a sterilized thermal probe was secured to the back skin using a 7–0 silk suture. The incision was closed with interrupted sutures. Postoperatively, mice were single-housed and monitored daily during a one-week recovery period. Following recovery, mice were subjected to a six-hour cold exposure at 4 °C inside a cold room. During exposure, mice remained single-housed with minimal food and bedding in their home cages, which were placed on a receiver platform. Core body temperature was continuously recorded at one-minute intervals using the VitalView system (version 5.1; Starr Life Sciences Corp). Per UH institutional animal care policy, mice were removed from cold exposure if the core body temperature dropped below 30°C.

Mouse energy expenditure assessment.

Mice for the metabolic cage assessment at 4°C were given a 0.08 ppm sodium selenite diet for 7 weeks and housed at room temperature. After 7 weeks, mice were then allowed to acclimate in the Oxymax-CLAMS system (Columbus Instruments, Columbus, OH) at the Brigham and Women’s hospital (Boston, MA) at thermoneutrality (30°C for 7 days). They were then exposed to 4°C for 30 hours. Core body temperature, total energy expenditure, respiratory exchange rate, and food consumption was then assessed for 72 hours. Food was provided ad libitum. Mice were housed at thermoneutrality for 7 weeks on a 0.08 ppm sodium selenite diet. After 7 weeks, mice were then single-housed in TSE System’s Phenomaster metabolic cage system (Berlin, Germany) at room temperature at the University of Hawaii. Mice were allowed to acclimate for 24 hours. Energy expenditure, respiratory exchange rate, O₂ consumption, and CO₂ production were then acquired for 72 hours. Food was provided ad libitum. Data was analyzed with the open source CalR application⁴⁴.

CL 316,243 treatment.

Mice were randomly assigned to one of four experimental groups: 1) WT + low selenite diet, 2) Scly KO + low selenite diet, 3) WT + medium selenite diet, and 4) Scly KO + medium selenite diet. Mice in each group were daily administered CL 316,243, a β3-adrenergic agonist that activates BAT and induces lipolysis, for 6 consecutive days. The CL 316,243 was dissolved in a sterile water vehicle and injected intraperitoneally (IP) at a dose of 1 mg/kg/day, which was determined based on established protocols for inducing therapeutic effects in mice without causing overt toxicity⁴⁵. Mice were injected with CL 316,243 an hour before being humanely euthanized after isoflurane exposure. The BAT was promptly harvested and snap frozen in liquid nitrogen for downstream analysis.

Serum Collection and Analysis.

WT and Scly KO mice were anesthetized with an intraperitoneal injection of Avertin (250 mg/kg, Sigma-Aldrich), and blood was collected via cardiac puncture immediately before euthanasia. Samples were allowed to clot at room temperature for 15–30 minutes and then centrifuged at 2,000 × g for 10 minutes at 4°C. The isolated serum was stored at −80°C until use. Serum analytes were quantified using commercial kits per the manufacturers’ protocols for creatine kinase (CK) activity (Sigma-Aldrich, Catalog No: MAK116), triglycerides (Cayman Chemicals, Ann Arbor, MI, Catalog No: #10010303, USA), and glucose (Crystal Chem, Elk Grove Village, IL, Catalog No: #81692, USA). For the triglyceride assay, serum was first diluted 1:2 in assay buffer per manufacturer’s protocol. Results were expressed as U/L for CK activity and mg/dL for triglycerides and glucose.

RNA extraction.

For qPCR and NanoString analysis, mouse BAT was snap-frozen after removal from mice. Total RNA was extracted using a TissueRuptor (Qiagen, Germantown, MD, USA) with disposable probes, and isolated using the E.Z.N.A. total RNA kit I (Omega Biotek, Norcross, GA, USA; Catalog No: R6834).

Nanostring analysis.

Total RNA isolated from BAT of apropriate quality (RIN > 7) as assessed in an Agilent 2100 BioAnalyzer (RRID: SCR_018043, Agilent Technologies, Santa Clara, CA, USA) was loaded into a nCounter CodeSet custom-designed for the 24 mouse selenoprotein transcripts and assayed in a NanoString nCounter MAX Analysis System (NanoString Technologies, Seattle, WA, USA) in the Genomics and Bioinformatics Shared Resource facility at the University of Hawaii Cancer Center. Transcript counts were analyzed using the nSolver software (NanoString Technologies).

Real-time qPCR.

One μg of total RNA extracted from the interscapular BAT was reverse transcribed using the High-Capacity kit (Applied Biosystems - ThermoFisher Scientific, Waltham, MA, USA, Catalog No: 4368814). 10 ng of cDNA were used in real-time qPCR reactions for genes using specific primers, listed in Supplemental Table 2. qPCR reactions were carried out using the PerfeCTa SYBR Green SuperMix (Quantabio; Beverly, MA, Catalog No: 95054) and following MiQE guidelines⁴⁶. Calculations were performed using the Δ^Ct method after appropriate melting curves were generated, normalized by the expression of housekeeping genes 18s, β-actin or Gapdh, and plotted as fold change.

Immunohistochemistry.

Animals acclimated to room temperature had their BAT extracted in 12% formalin, paraffin-embedded and processed into histological slides for staining using hematoxylin-eosin or for immunohistochemistry [3,3’-diaminobenzidine (DAB) staining] to detect UCP1, as described previously⁴⁷. For the UCP1 quantification analysis, samples were blinded to the experimenter.

Western Blot.

Protein was extracted using RIPA lysis buffer (ThermoFisher Scientific, Catalog No: 89900) containing protease and phosphatase inhibitors (Cell Signaling Technology, Danvers, MA, USA, Catalog No: 5872). 10–20 μg of total protein was loaded into 4–20% SDS-PAGE (TGX Criterion, BioRad, Carlsbad, CA, USA) and transferred to an Immobilon-FL^® membrane (Millipore Sigma, Burlington, MA, USA) either in a wet assembly of tris-glycine buffer with 9% methanol overnight or using the semi-dry Trans-Blot Turbo Transfer System (BioRad) for 7 minutes. Primary antibodies were incubated either for 1 hour at room temperature or overnight at 4°C with rotation while fluorescently labeled secondary antibodies were incubated for 45 minutes at room temperature with rotation. Blots were imaged using a Li-Cor Odyssey Fc image infrared photo imager (RRID: SCR_023227, Li-Cor, Lincoln, NE, USA). Primary antibodies used in this study are listed in Supplemental Table 1.

Assessment of circulating thyroid hormones.

Thyroid hormone status was inferred from analyses of total thyroxine (T4), triiodothyronine (T3), and thyroid stimulating hormone (TSH) levels in the serum. 20 μl of serum were used to assay for TSH with a Milliplex MAP Mouse Pituitary Magnetic Bead Panel (Millipore Sigma) in a Luminex 200 system (Luminex Corporation, Austin, TX, USA), while 25 μl of serum were used to assay for total T4 and T3 using the AccuDiag^™ ELISA – T4 kit (Diagnostic Automation; Woodland Hills, CA, USA, Catalog No: 3149–15) and the T3 (Total) ELISA Kit (AbNova; Taipei, Taiwan, Catalog No: KA0925), respectively, assayed according to the manufacturer’s protocol⁴⁸.

Enzyme activity assays.

DIO2 activity was assessed by a previously described method⁴⁹. Total GPX activity was performed using a commercially available kit (Abcam, Waltham, MA, USA; Catalog No: ab102530). BAT was prepared by washing the tissue in ice cold PBS, followed by homogenization in ice cold assay buffer and centrifugation for 15 minutes at 4°C at 10,000 × g. Supernatant was collected and used in the assay according to the manufacturer’s specifications.

Lipid peroxidation assay

The lipid peroxidation assay was performed using a commercially available kit (Sigma-Aldrich; Lipid Peroxidation (MDA) Assay Kit, Catalog No: MAK568). Briefly, BAT was homogenized on ice in 300 μL malondialdehyde (MDA) lysis buffer containing 3 μL of butylated hydroxytoluene (BHT). Samples were centrifuged at 13,000 × g for 10 minutes to remove insoluble material. Assay was then performed using the supernatant and according to manufacturer’s specifications.

Total Se content

Total Se content was measured in the tissues and blood by previously described methods^50–53. Se content in the BAT was assessed using instrumental neutron activation analysis (INAA) after lyophilization in the University of Missouri-Columbia Research Reactor (MURR) Center. Samples for Se analysis were packaged in a polyethylene vial and irradiated by neutrons at the MURR. The isotope ⁷⁶Se captures a neutron to become ⁷⁷mSe. The ⁷⁷mSe released a 162 KeV gamma ray when it decays through isomeric transition. The 162 KeV gamma ray is measured using a high purity germanium detector. Three samples of NIST SRM 1577 bovine liver were analyzed with the samples. Se content in other tissues and plasma was measured by a basic fluorometric assay using diaminonaphtalene with modifications^50,51.

GSH assay

The GSH assay was performed using a commercially available kit from Cayman Chemicals (Ann Arbor, MI, USA; Catalog No: 703002). BAT was homogenized on ice in the supplied MES buffer (2X) diluted with to 1X with HPLC-grade water. The samples were then centrifuged at 10,000 × g for 15 minutes at 4°C. Samples were deproteinated using an equal volume of metaphosphoric acid (MPA) and centrifuged at 2000 × g for 2 minutes. The supernatant was collected and mixed with 4M triethanolamine (TEAM) reagent. The GSH assay was then performed according to the manufacturer’s protocol.

Triglyceride and cholesterol assays

Triglycerides in the BAT and serum were assessed using a commercially available kit from Cayman Chemicals (Catalog No: 10010303). 5 mg of minced BAT tissue was sonicated in 200 μL of diluted NP40 Substitute Assay Reagent containing 1 mM of EDTA. Samples were then centrifuged at 10,000 × g for 10 minutes at 4°C and the supernatant was transferred to a fresh tube. The assay was then performed according to the manufacturer’s protocol. For serum triglycerides, 5 ul of sample were used from mice raised at thermoneutrality. Cholesterol in the BAT was assessed using a commercially available assay kit from RayBiotech (Peachtree Corners, GA, USA, Catalog No: MA-TC). Samples for the cholesterol assay were prepared with the following protocol: 5 mg of BAT was homogenized in 200 μL of lysis buffer (250 mM sucrose, 50 mM Tris Cl, pH 7.0). The homogenate was mixed with 800 μL of chloroform:methanol (3:2 v/v) and extracted by 2 hours of vigorous shaking. Next, homogenate was centrifuged at 3,000 × g at room temperature for 10 min, and 300 μL of lower organic phase was collected and dried by exposure to N2 gas. To dry the lipids, 100 μL of 0.1% Triton X-100 was added and the solution was sonicated using ultrasonic homogenizer for 10 seconds. The assay was then performed according to the manufacturer’s protocol.

Liquid Coupled-Mass Spectrometry (LC-MS) analysis of CoA compounds

BAT samples were weighed into 1.5-mL microcentrifuge tubes. Tris buffer at pH 7 was added to each tube at 2 μL/mg of raw tissue. The samples were homogenized with the aid of two 3-mm metal beads at 30 Hz for 30 s. Methanol/chloroform (3:1) at 8 μL/mg of raw tissue was then added. The samples were homogenized for 1 min twice, followed by centrifugation at 21,000 × g, 5 °C for 10 min. 100 μL of the supernatant of each sample was mixed with 50 μL of the internal standard solution and then dried under a gentle nitrogen gas flow. The dried residue was reconstituted in 50 μL of 80% aqueous methanol and centrifuged for clarification. 10-μL aliquots of each resultant sample solution and each standard solution were injected to run UPLC-MRM/MS in the negative-ion mode on a Waters UPLC system coupled to a Sciex QTRAP 6500 Plus MS instrument. LC separation was carried out on a C18 column (2.1 × 100 mm, 1.8 μm) with the use of an ammonium bicarbonate buffer (A) and mixed acetonitrile/isopropanol (B) as the mobile phase for binary-solvent elution with an efficient gradient of 1% to 90% B over 18 min, at 50 °C and 0.3 mL/min. Concentrations of the detected compounds were calculated by interpolating the constructed linear-regression curves of individual compounds with the analyte-to-internal standard peak area ratios measured from each sample solution.

Total iron assay

Total iron was assessed using a commercially available kit (Sigma-Aldrich, Catalog No: MAK025). Tissue was homogenized in the supplied iron assay buffer and centrifuged at 16,000 × g for 10 minutes at 4°C to remove insoluble materials. The assay was then performed according to the manufacturer’s protocol.

Statistical Analysis.

Statistical analysis of collected data was performed using GraphPad Prism 10 software (RRID: SCR_002798, GraphPad, La Jolla, CA). Unless stated otherwise, the distributions of the continuous variables were expressed as the mean ± SEM. D’Agostino and Pearson (single comparison) or Kruskal-Wallis (multiple comparisons) normality tests were used to evaluate the normality of the data. If normal, the statistical significance of the difference between groups was measured by Student’s t-test, one-way ANOVA with Tukey’s post hoc test or two-way ANOVA with Bonferroni’s post hoc test for two variables. If the data did not pass the normality test, non-parametric tests (Mann-Whitney) for single comparisons were used to derive the P values. The null hypothesis was rejected for P < 0.05 with the two-tailed t-test. Survival curve was analyzed with the Gehan-Breslow-Wilcoxon test.

RESULTS

Loss of Scly reduces total Se in BAT and increases BAT whitening in mice

To determine the impact of the loss of Scly on the morphology of the BAT, we placed WT and Scly KO mice on either a 0.08 or 0.25 ppm selenite diet for 8 weeks starting at weaning as shown by the experimental timeline (Fig. 1A). Interscapular BAT (iBAT) was found to be significantly heavier in Scly KO mice on a 0.08 ppm (0.325 ± 0.066 vs. 0.653 ± 0.108 g, WT vs KO, n = 7/group, P < 0.01, Fig. 1B) than in WT or Scly KO mice fed Se-sufficient diets (0.316 ± 0.020 vs. 0.560 ± 0.073 g, n = 6/group, WT vs KO, Fig. 1B). Histological analysis revealed that Scly KO iBAT exhibited increased accumulation and enlargement of lipid droplets compared with WT iBAT on both a 0.25 and 0.08 ppm Se diet with low Se exacerbating the lipid droplet deposition (Fig. 1C). This corresponded with lower total Se levels overall in the BAT from Scly KO mice on an adequate Se diet (0.4198 ± 0.03844 vs. 0.1800 ± 0.01460 μg/g, n = 5/group, WT vs KO, P < 0.01, Fig. 1D). We also observed lower Se content in the liver, kidney, and testes of Scly KO mice (Fig. 1D). Then, as UCP1 is a key marker of brown fat, we next assessed UCP1 protein expression using immunohistochemistry and discovered that UCP1 protein levels were significantly decreased as evaluated by DAB staining (Fig. 1E–F). Correspondingly, mRNA expression of Ucp1 was also significantly decreased in the BAT (Fig. 1G). Overall, these data indicate that loss of Scly leads to an increase in lipid deposition in the BAT and a decrease in the BAT thermogenic marker, UCP1.

Figure 1. Loss of SCLY in mice leads to BAT whitening.

WT and whole body Scly KO were placed on a 0.08 or 0.25 ppm Se diet for 8 weeks. After 8 weeks, the BAT was harvested for either histology or analyzed for total Se and qPCR. A. Diagram of experimental timeline. B. Body weight of mice on a 0.08 ppm Se diet over the 8 weeks of diet. C. iBAT weights of WT and Scly KO mice on a 0.08 or 0.25 ppm Se diet. D. Representative histology of WT and Scly KO BAT on a 0.08 and 0.25 ppm Se diet. Scale bar = 100 μm. E. Left. Total Se levels in liver, kidneys, BAT, brain, testes, muscle, and blood on a 0.25 ppm Se diet. Right. Total Se levels in the BAT only. F. Representative immunohistochemistry DAB of UCP1 in WT and Scly KO mice on a 0.25 ppm diet. G. Quantification of E. H. qPCR of Ucp1 in WT and Scly KO BAT. Data in E, G, and H were analyzed with two-tailed Student’s t-test with Welch’s correction. Data in B was analyzed with two-way ANOVA with Bonferroni’s post hoc test, *P < 0.05, **P < 0.01. RT, room temperature; iBAT, interscapular brown adipose tissue; UCP1, uncoupling protein 1; DAB, 3,3’-diaminobenzidine.

Thyroid hormone homeostasis is perturbed by the loss of Scly

Due to the decrease in UCP1, we next investigated how the loss of Scly impacted thyroid hormone levels and thyroid receptor expression in the serum (Table 1). T4 levels were significantly increased in Scly KO mice on a low Se diet (Table 1), although triiodothyronine (T3) and the ratio of T3/T4 were unchanged on both a low and medium Se diet. Despite the ratio of T3/T4 not significantly different, circulating hypophyseal thyroid stimulating hormone (TSH) levels were ~45% higher in Scly KO mice regardless of Se intake, another potential indicator that the effect of genotype was more predominant than the dietary effect of Se, and that the elevated TSH was leading to the increased adiposity as high TSH levels are associated with obesity⁵⁴. Correspondingly, the transcript expression of the TSH receptor (Tshr) was also significantly increased in Scly KO mice (1.003 ± 0.072 vs. 2.496 ± 0.471, n = 3–5/group, WT vs. Scly KO, P < 0.05, Table 1), while the expression of thyroid hormone receptors alpha (Thra1, Thra2) and beta (Thrb) were unchanged.

Table 1.

Thyroid hormones T4 and T3 and TSH levels in the serum and relative thyroid hormone and TSH receptors mRNA expression in BAT.

	0.08 ppm Se				0.25 ppm Se
	WT	n	KO	n	WT	n	KO	n	P genotype	P diet	P interaction
T3 (ng/mL)	1.149 ± 0.054	3	1.318 ± 0.128	3	0.890 ± 0.094	7	1.073 ± 0.079	8	0.118	0.031	0.946
T4 (ng/mL)	1.933 ± 0.237	3	3.895 ± 0.272 **	3	5.184 ± 0.635	6	3.129 ± 0.521 *	8	0.943	0.074	0.007
T3/T4 (a.u.)	0.623 ± 0.117	3	0.347 ± 0.061	3	0.337 ± 0.141	7	0.379 ± 0.036	8	0.335	0.298	0.196
TSH (pg/mL)	153.67 ± 6.15	4	221.90 ± 18.97 *	6	153.37 ± 13.19	10	224.67 ± 22.87 *	8	0.001	0.950	0.938
Tshr (a.u.)	1.003 ± 0.072	3	2.496 ± 0.471 *	5	1.146 ± 0.289	5	1.183 ± 0.244	4	0.064	0.128	0.053
Thra1 (a.u.)	1.067 ± 0.268	3	1.014 ± 0.209	5	1.070 ± 0.187	5	0.933 ± 0.154	4	0.655	0.854	0.842
Thra2 (a.u.)	1.013 ± 0.107	3	0.996 ± 0.236	5	1.082 ± 0.217	5	1.135 ± 0.280	4	0.942	0.674	0.886
Thrb (a.u.)	1.020 ± 0.140	3	0.762 ± 0.132	5	1.054 ± 0.160	5	1.190 ± 0.247	4	0.740	0.221	0.293

Results were analyzed with two-way ANOVA with Bonferroni’s post hoc test.

^*P < 0.05

^**P < 0.01.

TSH, thyroid stimulating hormone; Tshr, TSH receptor; Thra, thyroid hormone receptor alpha; Thrb, thyroid hormone receptor beta.

Loss of Scly reduces thermogenesis-related selenoproteins in the BAT

As SCLY supplies selenide for maintaining selenoprotein synthesis, we then assessed how Scly loss affected the expression of all 24 selenoproteins in mice. Using Nanostring, we observed that Dio2 was significantly downregulated by 69% in Scly KO mice on a 0.08 ppm Se diet (Table 2). Other downregulated genes included Selenop, Selenow and Selenoh (Table 2). Only one selenoprotein transcript, Gpx4, was significantly increased compared with WT on a low Se diet (Table 2). Notably, qPCR of Dio2 confirmed the Nanostring finding (Fig. 2B) and also corresponded with a significant decrease in DIO2 enzyme activity on both a low (3.300 ± 0.802 vs. 1.033 ± 0.491 fmol/mg/hr, n = 3–5/group, WT vs. Scly KO, P < 0.01) and an adequate Se diet (2.800 ± 0.565 vs. 1.260 ± 0.093 fmol/mg/hr, n = 3–5/group, WT vs. Scly KO, P < 0.05, Fig. 2C). We also verified the protein levels of GPX1 and GPX4, as together with DIO2, GPX1 and GPX4 are involved in the control of energy metabolism in the BAT⁵⁵. Both GPX1 and GPX4 protein levels were significantly reduced in the BAT of Scly KO mice on a low Se diet, which corresponded with a trend towards a decrease in overall GPX activity (Fig. 2D–F, full blots Supplemental Figs. 4–5). While selenoprotein P (Selenop) mRNA levels were found to be downregulated, this was not reflected in SELENOP protein levels, which were not changed between WT and Scly KO mice (Fig. 2G, full blot, Supplemental Fig. 17). However, total GSH levels, while significantly lower in the BAT of Scly KO mice on a low Se compared to an adequate Se diet, there was no difference in GSH between genotypes (Fig. 2H). These data suggest that loss of Scly only impacts the expression of selenoproteins involved in energy homeostasis in the BAT, including DIO2, GPX1, and GPX4.

Table 2.

Relative expression of selenoproteins using Nanostring technology. n = 4 for all groups.

	0.08 ppm Se (a.u.)		0.25 ppm Se (a.u.)
	WT	KO	WT	KO
Dio1	50.12 ± 5.56	57.39 ± 2.26	58.45 ± 2.80	52.93 ± 3.39
Dio2	945.70 ± 16.32 ****	371.48 ± 80.61	448.58 ± 104.04	553.96 ± 154.48
Dio3	50.46 ± 4.16	43.61 ± 3.40	44.66 ± 4.62	45.50 ± 3.30
Gpx1	6806.64 ± 122.99	4719.94 ± 327.30	6990.65 ± 126.02	6735.85 ± 117.30
Gpx2	107.43 ± 5.98	123.92 ± 6.29	101.11 ± 6.40	116.65 ± 5.41
Gpx3	209.70 ± 9.86	159.97 ± 7.99	194.06 ± 19.34	117.9 ± 19.9
Gpx4	6.54 × 104 ± 2.40 × 103	7.10 × 104 ± 6.10 × 103 ****	6.30 × 104 ± 2.10 × 103	6.92 × 104 ± 2.40 × 103***
Msrb1	2112.17 ± 74.98	1815.39 ± 249.47	1631.08 ± 100.97	1468.35 ± 75.18
Selenof	2609.89 ± 51.47	2637.34 ± 73.54	2568.93 ± 91.96	2508.84 ± 87.60
Selenoh	737.68 ± 32.37	465.41 ± 9.76 ****	706.28 ± 29.51	687.55 ± 46.85
Selenoi	323.65 ± 12.78	313.76 ± 15.67	332.68 ± 6.86	315.78 ± 10.54
Selenok	4557.15 ± 144.94	4368.95 ± 44.83	4354.05 ± 137.69	4505.65 ± 193.48
Selenom	17.79 ± 3.13	15.50 ± 3.25	16.78 ± 4.03	16.37 ± 2.15
Selenon	131.62 ± 8.74	127.20 ± 6.20	130.11 ± 5.70	134.59 ± 2.98
Selenoo	1010.64 ± 45.36	871.30 ± 51.74	950.93 ± 40.84	948.31 ± 31.29
Selenop	1.72 × 104 ± 4.00 × 102	1.35 × 104 ± 3.76 × 102 *	1.41 × 104 ± 3.77 × 102	1.46 × 104 ± 3.70 × 102
Selenos	446.76 ± 22.81	433.52 ± 10.14	410.07 ± 24.87	425.02 ± 18.78
Selenot	3768.67 ± 148.99	4061.84 ± 106.79	3854.65 ± 72.40	3998.75 ± 132.48
Selenov	26.85 ± 0.84	22.01 ± 1.77	22.37 ± 2.65	23.69 ± 1.95
Selenow	8453.73 ± 693.19	5623.87 ± 436.71 **	1.02 × 104 ± 668.93	6266.14 ± 138.54 *
Sephs2	6780.32 ± 206.26	7500.21 ± 176.46	5700.51 ± 139.57	6958.81 ± 457.89
Txnrd1	2888.51 ± 65.75	2414.58 ± 80.59	2788.05 ± 58.48	2428.85 ± 30.21
Txnrd2	1195.42 ± 22.70	1042.24 ± 24.95	1301.68 ± 87.22	1079.21 ± 43.
Txnrd3	760.18 ± 26.74	726.75 ± 30.35	655.24 ± 24.23	738.40 ± 22.78

WT on a 0.25 ppm diet was set as control and data was analyzed with two-way ANOVA with Bonferroni’s post hoc test

^*P < 0.05

^**P < 0.01

^****P < 0.0001.

Figure 2. Selenoprotein changes in Scly KO BAT on a Se deficient diet.

A. Diagram of the potential pathways for Sec. Pathways focused on are highlighted by the colored box. B. qPCR of Dio2 normalized to an average of the following housekeeping genes: Gapdh, 18S, and Actb. C. DIO2 activity in WT and Scly KO BAT on 0.08 and 0.25 ppm Se diets. D. Representative western blot of GPX1 and quantification normalized to β-actin. E. Representative western blot and quantification of GPX4 normalized to β-actin. F. GPX activity assay. G. Representative western blot and quantification of SELENOP. H. GSH levels in WT and Scly KO BAT on 0.08 and 0.25 ppm Se diets. Each error bar represents mean ± SEM, n = 3–5. Data in all graphs except E were analyzed with two-way ANOVA with Bonferroni’s post hoc test. Graphs in E were two-tailed Student’s t-test with Welch’s correction. *P < 0.05, **P < 0.01. DIO, iodothyronine deiodinase; GPX, glutathione peroxidase; seleno, selenoprotein; GSH, glutathione.

Scly KO mice display a mild impairment in acute adaptive thermogenesis

As DIO2 activity and GPX1 and GPX4 protein levels were significantly decreased in Scly KO mice on a low and adequate Se diet along with UCP1, we next tested the response of Scly KO mice to acute cold exposure. As the phenotype of the Scly KO mice was observed to be stronger with the low selenite diet, we placed WT and Scly KO mice on a 0.08 ppm selenite diet for 8 weeks starting at weaning age and housed the mice at thermoneutrality. After 7 weeks of diet, mice were implanted with a thermal probe, allowed to recover for one week, and at 8 weeks of diet exposed to cold (4 °C) for 6 hours as diagrammed in Fig. 3A. At thermoneutrality, body weight was unchanged between WT and Scly KO mice after 8 weeks on the Se-deficient diet (Fig. 3B) in contrast to the Scly KO mice housed at room temperature (Fig. 1B), indicating a possible impairment in BAT function. Exposure to cold temperature (4°C) revealed Scly KO mice to have a mild impairment in acute adaptive thermogenesis. Six out of 8 Scly KO mice and 3 out of 7 WT mice were removed from the cold early as shown by the individual traces of each mouse in Fig. 3C. Additionally, averaged core body temperature was also reduced in the Scly KO mouse (P_genotype = 0.0192). Finally, survival curve analysis showed a trend towards an increased probability of early removal (P = 0.0937, Fig. 3E). We then checked CK activity and triglycerides in the serum of these mice exposed to the cold environment to determine if there was increased shivering or lipid usage in the Scly KO, and found no difference in both of these parameters between WT and Scly KO mice (Fig. 3F–G). However, serum glucose levels were significantly increased (192.6 ± 22.72 vs 271.3 ± 16.67 mg/dL, n = 6/group, WT vs. KO, *P <0.05, Fig. 3H) which could indicate a switch to increased carbohydrate input to supply BAT activation in the Scly KO mice.

Figure 3. Scly disruption combined with a Se-deficient diet mildly impairs acute adaptive thermogenesis.

WT and Scly KO mice were placed in the cold (4°C) for up to 6 hours. Mice were removed early if their core body temperature dropped below 30°C per institutional protocol. A. Experimental timeline. B. Body weight of the mice at thermoneutrality over 8 weeks. C. Individual body temperature of each mouse over the duration of the cold exposure. Blue lines indicate WT, red lines indicate Scly KO mice. Mice were removed early from the cold environment if their core body temperature dropped below 30°C per our institutional animal protocol guidelines for humane experimentation. D. Averaged core body temperature of WT and Scly KO mice throughout the cold exposure. E. Survival curve showing the time and number of mice removed early. 3/7 WT mice and 6/8 of Scly KO were removed early. F-H. Creatine kinase, triglyceride, and glucose levels in the serum of mice following cold exposure. I. Experimental timeline of mice for the CLAMS metabolic cage assessment. WT and Scly KO mice originally housed at room temperature were acclimated at thermoneutrality (30°C) for 7 days before 30 hours of cold exposure (4°C) in the CLAMS system. J-M. Body temperature, respiratory exchange ratio, total energy expenditure and total food consumption during the 30 hours in the CLAMS metabolic cages. Each error bar represents mean ± SEM, n = 9 (A), n = 7–8 (C-E), n = 6 (F-G), and n = 6 (I-L). Data in graphs B, D and I-L were analyzed with two-way ANOVA with Bonferroni’s post hoc test. Data in E was analyzed with Gehan-Breslow-Wilcoxon test. Data in F-G was analyzed with two-tailed test with Welch’s correction. *P < 0.05.

Since we observed mild impairment of adaptive thermogenesis, we assessed metabolic parameters of the mice during cold exposure using the Oxymax-CLAMS metabolic cage system. The experimental design is shown in Fig. 3I. We found no difference in core body temperature, respiratory exchange ratio (RER), total energy expenditure, and total food consumption between WT and Scly KO mice (Fig. 3J–M). Then, to obtain a comprehensive assessment of the metabolism of the mice, we also measured metabolic parameters of these mice at room temperature following housing at thermoneutrality for 7 weeks (Supplemental Fig 1A). We found respiratory exchange ratio and CO₂ production were significantly increased in Scly KO mice on a low selenite diet (P = 0.0060 and P = 0.0174, respectively) while O₂ consumption and energy expenditure were unchanged (Supplemental Fig. 1B-E). CK and serum glucose were also measured and serum glucose trended toward being increased (P = 0.0757) in Scly KO mice raised at thermoneutrality (Supplemental Fig. 1F-G). Increased RER and serum glucose may indicate a higher reliance on glucose for the BAT in mice lacking Scly.

Finally, we also evaluated proteins involved in adaptive thermogenesis. We found significantly decreased GPX1, GPX4, and UCP1 protein levels while the Se carrier protein, SELENOP, was unchanged in Scly KO BAT after cold exposure (Fig. 4A–C and Fig. 4F, full blots, Supplemental Fig. 13). Lipolysis markers adipose triglyceride lipase (ATGL) and hormone sensitive lipase (HSL) were unchanged (Fig. 4D–E, full blots, Supplemental Fig. 14) indicating there was no impairment of the mice lacking Scly to mobilize lipids. Together, these data suggest that Scly KO mice have a mild impairment in adaptive thermogenesis, likely due to a decrease in the selenoproteins directly participating in the control of thermogenesis responses.

Figure 4. Selenoproteins and lipolysis markers after cold exposure.

A-F. GPX1, GPX4, SELENOP, HSL, ATGL, and UCP1 western blots and quantification in the BAT of WT and Scly KO mice on a 0.08 ppm selenite diet following cold exposure. Each error bar represents mean ± SEM, n = 7 (WT) and 8 (KO). Data in all graphs were analyzed with two-tailed t-test with Welch’s correction. *P < 0.05, **P < 0.01. GPX, glutathione peroxidase; SELENOP, selenoprotein P; HSL, hormone sensitive lipase; ATGL, adipose triglyceride lipase; UCP1, uncoupling protein 1.

Pharmacological activation of lipolysis increases creatine cycling in mice with a disruption of Scly

To assure a comprehensive evaluation of the thermogenic response, particularly lipolysis, in mice lacking Scly, we investigated the impact of pharmacologic activation of thermogenic responses by treating mice with the beta-3 adrenergic receptor agonist, CL-316,243. The timeline of this experiment is diagrammed in Fig. 5A. After 6 days of injection, Scly KO mice on a low Se diet was significantly heavier than WT mice (26.644 ± 0.657 vs. 30.490 ± 1.042, n = 5/group, WT vs. Scly KO, P < 0.05, Fig. 5B). There was no difference in body weight between Scly KO mice on an adequate Se diet after CL injection compared to WT (25.732 ± 0.583 vs. 28.094, n = 5/group, WT vs. Scly KO, Fig. 5B). There was also no difference in inguinal white adipose tissue (iWAT), epidydimal WAT (eWAT), and BAT weight (Fig. 5C–E) after CL 314,243 treatment. SELENOP was increased in the BAT on a medium Se diet, potentially indicating a compensatory response for the loss of Scly (Fig. 5F, full blot, Supplemental Fig. 10), while assessment of thermogenic proteins in the BAT again revealed GPX1 and GPX4 to be significantly decreased (Fig. 5G–H, full blots, Supplemental Figs. 6–7) in both the low and adequate Se diet. Unlike the acute cold exposure, UCP1 was not significantly changed between WT and Scly KO (Fig. 5I, full blot, Supplemental Fig. 12). Analysis of the lipolysis proteins ATGL and HSL also again displayed no differences between WT and Scly KO BAT on either a low or adequate Se diet (Fig. 5I–J, full blots, Supplemental Figs. 8–9). However, there was a trend towards increased serum glucose in Scly KO on an adequate Se diet (P = 0.0688, Fig. 5L), and an increase in CK activity in the serum of Scly KO mice on a low Se diet (36.83 ± 17.80 vs 112.2 ± 21.30 U/L, n = 4–5/group, WT vs KO, P < 0.05, Fig. 5L). As the CL injections were given over multiple days, a chronic stimuli, while the cold exposure was acute, this result suggests that Scly KO mice may trigger creatine cycling⁵⁶ to compensate for the decrease in the selenoproteins GPX1 and GPX4.

Figure 5. Treatment with the beta 3-adrenergic agonist CL 316,243 impacts BAT selenoproteins and creatine kinase in mice with disruption of Scly.

A. Experimental design for the CL 316,243 treatment. B-E. Body weight, iWAT to body weight ratio, eWAT to body weight ratio, and BAT to body weight ratio after CL injections. F-H. Western blot and quantification of selenoproteins SELENOP, GPX1, and GPX4 in the BAT. I. Western blot and graph quantification of UCP1 levels. J-K. Western blot and quantification of levels of lipolysis proteins ATGL and HSL. K-M. Serum glucose, creatine kinase, and triglyceride levels in WT and Scly KO mice following CL injections. Each error bar represents mean ± SEM, n = 5 (WT) and 4 (KO). Data in graphs B-E were analyzed using two-way ANOVA with Bonferroni’s post hoc test. Graphs in F-M were analyzed with two-tailed t-test with Welch’s correction. *P < 0.05, ****P < 0.0001. iWAT, inguinal white adipose tissue; eWAT, epididymal white adipose tissue; BAT, brown adipose tissue; GPX, glutathione peroxidase; SELENOP, selenoprotein P; ATGL, adipose triglyceride lipase; HSL, hormone sensitive lipase; UCP1, uncoupling protein 1.

Reduced GPX4 does not increase ferroptosis sensitivity in the BAT of Scly KO mice

Since GPX4 is critical to curb ferroptosis, and we observed consistent decrease of GPX4 in Scly KO mice, we hypothesized that a potential mechanism for the increased lipid deposition of the BAT may be activation of ferroptosis of brown adipocytes as it has been previously reported that BAT whitening leads to brown adipocyte death by ferroptosis⁵⁷. We measured the expression of several genes related to ferroptosis sensitivity including Slc7a11, Ptgs2, Fth1, Fsp1, and Ascl4. The roles of the proteins that these genes express are diagrammed in a simplified scheme of ferroptosis (Supplemental Fig. 2A). Most genes except for Ascl4 were upregulated in the BAT of Scly KO on a 0.08 ppm Se diet (Supplemental Fig. 2B). As several of the ferroptotic genes were increased in the Scly KO BAT, we also determined the protein levels of both ferroptosis suppressor protein 1 (FSP1) and ferritin (FTH1). FSP1 is a GPX4-independent suppressor of ferroptosis while FTH1 stores iron and is typically decreased after ferroptotic cell death. FSP1 was checked to determine if FSP1 was increased to compensate for the decrease in GPX4 while FTH1 was assessed as an additional marker of the ferroptosis response. FSP1 was significantly reduced on both a 0.08 and 0.25 ppm Se diet (Supplemental Fig. 2C, full blot, Supplemental Fig. 18) while FTH1 levels were unchanged (Supplemental Fig. 2D, full blot, Supplemental Fig. 18). Functional markers of ferroptosis including lipid peroxidation and total iron levels were analyzed as increases in iron and lipid peroxidation are ferroptotic inducers. We used malondialdehyde (MDA) as a marker for lipid peroxidation and found that, while MDA levels were increased in WT BAT on a 0.08 ppm Se diet compared to WT mice on a Se-sufficient diet, this elevation did not occur in the Scly KO BAT (Supplemental Fig. 2E). Furthermore, total iron levels were unchanged in Scly KO BAT regardless of dietary Se levels (Supplemental Fig. 2F). Combined, these data indicate that despite the reduced levels of both GPX4 and FSP1, ferroptosis pathways are intact in Scly KO BAT. Therefore, BAT whitening is not due to ferroptotic mechanisms in brown adipocytes.

Evidence of lipogenic programming is not detected in mice lacking Scly

Without evidence of ferroptosis induction in Scly KO BAT, we then tested if the loss of Scly was instead diverting remaining Sec towards a pathway that activates lipogenesis and lipid deposition, as observed in the BAT histology at both an adequate and Se-deficient diet. The diagram in Fig. 6A illustrates the different pathways, some of sulfur metabolism, that may reroute the Sec amino acid when not being used for selenoprotein production, and these pathways include taurine synthesis, lipogenesis, iron-Se cluster formation, and GSH production. We had already confirmed total GSH to not be affected by loss of Scly in Fig. 2H and therefore, aimed to confirm that the other pathways that may compete for Sec usage were also not favored in the BAT of Scly KO mice. We therefore checked the transcript and protein levels of NFS1 which is involved in Fe-S cluster, and possibly Fe-Se cluster, formation. Although transcript levels of NFS1 were increased in BAT from Scly KO mice on an adequate diet, protein levels of NFS1 remained similar when mice were on a low Se diet (Supplemental Fig. 3B-C, full blot, Supplemental Fig. 19). Levels of the taurine synthesis enzyme CDO1 in the BAT were also similar between WT and Scly KO mice, indicating that there was no potential diversion of Sec to the synthesis of a Se-substitute taurine compound (Supplemental Fig. 3D, full blot, Supplemental Fig. 19).

Figure 6. Lipogenic programming in the BAT of Scly KO mouse.

A. Diagram of the pathways that can utilize Sec as a substrate, highlighting the lipogenic route. B. qPCR of lipid metabolism markers from BAT of WT and Scly KO mice on a 0.08 and 0.25 ppm Se diet. C. qPCR of PPARγ on a 0.08 and 0.25 ppm Se diet. D-E. Western blot and quantification analysis of PPARγ in WT and Scly KO mice on a 0.08 and 0.25 ppm Se diet with or without CL injection. F. Western blot and quantification of CHREBP. H-I. Western blot and quantification of HSL and ATGL. Each error bar represents mean ± SEM, n = 3–6 (B-C) and n = 3–5 (D-H). Data in all graphs except E were analyzed with two-way ANOVA with Bonferroni’s post hoc test. Data in E was analyzed with two-tailed t-test with Welch’s correction. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. PPARγ, peroxisome proliferator activated receptor gamma; HK, housekeeping; Ucp1, uncoupling protein 1; hormone sensitive lipase E, Lipe; PPAR coactivator 1 alpha, Ppargc1a; peptidyl-prolyl cis-trans isomerase B, Ppib; glycerol-3-phosphate dehydrogenase 1, Gpd1; and acetyl-coA carboxylase alpha, Acaca; CHREBP, carbohydrate-responsive element-binding protein; HSL, hormone sensitive lipase; ATGL, adipose triglyceride lipase; FC, fold change.

After confirming that three Sec-dependent pathways were not activated by the loss of Scly, we then investigated whether the remaining possibility, lipogenesis, was activated in the BAT of Scly KO mice. First, we assessed the expression of several genes involved in lipogenesis and mitochondrial function including peroxisome proliferator-activated receptor gamma coactivator 1-alpha (Ppargc1a), hormone sensitive lipase E (Lipe), peptidyl-prolyl cis-trans isomerase B (Ppib), glycerol-3-phosphate dehydrogenase (Gpd1), and acetyl-coA carboxylase alpha (Acaca), sterol regulatory element-binding protein 1(Srebp1 and Srebp1c), Srebp2, and carbohydrate-responsive element binding protein (Chrebp). We found that only Acaca was significantly increased in the Scly KO mice on an adequate Se diet (Fig. 6B). Pparg transcripts were significantly decreased (Fig. 6C) while Ppargc1a expression was significantly increased in both WT and Scly KO on a Se-deficient diet, although for both genes, there was no difference between genotypes (Fig. 6D). The Se effect, but not genotype, on BAT PPARγ was confirmed at the protein level as well (Fig. 6E, full blot, Supplemental Fig. 17). Strikingly, PPARγ was increased in the BAT of the Scly KO mice injected with CL-316,243 on both a low and adequate Se diet, and the Se effect was lost (Fig. 6F, full blot, Supplemental Fig. 11). We also checked for the lipogenesis regulatory protein, the carbohydrate responsive element binding protein (ChREBP), and found no difference in ChREBP levels in the BAT of WT and Scly KO (Fig. 6G, full blot, Supplemental Fig. 15). Then to determine if Scly KO mice had deficiencies in lipolysis and lipid mobilization, we verified lipolysis regulators ATGL and HSL and again found similar levels of these markers in the BAT (Fig. 6H–I, full blots, Supplemental Fig. 16).

Finally, as CoA intermediates are involved in fatty acid synthesis, we also checked the levels of CoA intermediates HMG-CoA, acetoacetyl-CoA, acetyl-CoA, coenzyme A (HS-CoA), malonyl-CoA, and succinyl-CoA along with triglyceride and cholesterol levels. We found that the, while most CoA intermediates were unchanged, HMG-CoA levels were decreased in Scly KO on a low Se diet (Table 3). Correspondingly, cholesterol levels were also significantly reduced by approximately a third (Table 3), signifying the inhibition of the mevalonate pathway which is crucial for BAT development and function⁵⁸. Triglycerides in the BAT were significantly increased (78.22 ± 6.61 vs. 110.2 ± 8.61 [(mg/dL)/mg], n = 5–8/group, WT vs KO, P < 0.01, Table 3), while triglycerides in the serum of mice lacking Scly were decreased (26.59 ± 2.381 vs. 19.09 ± 1.496 mg/dL, n = 4/group WT vs. KO, P < 0.05, Table 3). Therefore, this result suggests that the remaining Sec possibly accumulated by the lack of decomposition via Scly, may be shuttled towards another pathway.

Table 3.

Levels of coA intermediates, triglyceride, and cholesterol levels in the BAT and serum of mice on a low selenite diet.

	0.08 ppm Se
BAT	WT (nmol/g)	n	KO (nmol/g)	n
Acetyl-CoA	32.95 ± 10.26	5	22.96 ± 5.046	5
Acetoacetyl-CoA	1.974 ± 0.207	5	1.912 ± 0.0640	5
HMG-CoA	0.0425 ± 0.028	5	0.0286 ± 0.0033 *	5
HS-CoA	55.64 ± 14.120	5	45.63 ± 8.136	5
Malonyl-CoA	1.993 ± 0.172	5	1.863 ± 0.094	5
Succinyl-CoA	25.86 ± 9.797	5	25.28 ± 5.658	5
Triglycerides (mg/dL/mg) at RT	76.08 ± 13.43	5	99.12 ± 5.728 **	8
Cholesterol (μmol/mg) at RT	1639 ± 220.2	3	541.4 ± 104.7 *	8
SERUM	WT	n	KO	n
Triglycerides mg/dL	26.59 ± 2.381	4	19.09 ± 1.496 *	4

Data was analyzed with two-tailed t-test with Welch’s correction; n values are listed on the table.

^*P < 0.05

^**P < 0.01.

RT, room temperature. HMG-CoA, 3-hydroxy-3-methylglutaryl coenzyme A; HS-CoA, sulfhydryl-CoA.

DISCUSSION

We hypothesized that the loss of Scly diverts selenocysteine (Sec) utilization away from selenoprotein synthesis toward alternative metabolic pathways, such as lipogenesis, thereby perturbing lipid balance and impairing thermogenic BAT function. Our findings partially support this hypothesis. Scly KO mice exhibit marked alterations in BAT morphology, consistent with BAT whitening and mild impairment of adaptive thermogenesis. Despite increased triglyceride content, lipogenesis was not activated, suggesting that lipid accumulation reflects impaired thermogenic utilization rather than enhanced lipid synthesis. Impaired thermogenesis is substantiated by the lower Se content that was found in the BAT of Scly KO mice, resulting in decreased DIO2 activity and lower GPX1 and GPX4 protein levels. UCP1 was also significantly decreased, possibly as a consequence of lower thyroid hormone activation due to reduced DIO2 activity. Figure 7 summarizes our key findings, showing that when Sec is not utilized by Scly, it may be diverted into another fate. This shift is accompanied by downregulation of mevalonate pathway enzymes and thermogenic selenoproteins, including GPX1, GPX4, and DIO2, ultimately leading to BAT whitening and reduced UCP1 expression. Collectively, these results indicate that loss of SCLY disrupts multiple pathways critical for brown adipose tissue (BAT) function, promoting triglyceride accumulation, particularly under Se-deficient conditions, by redirecting Sec away from the production of key selenoproteins, thereby mildly impairing thermogenic capacity.

Figure 7. Schematic representation of how the loss of Scly leads to BAT whitening and mild impairment of adaptive thermogenesis.

Loss of Scly decreases selenoprotein synthesis of the ones participating in adaptive thermogenesis (DIO2, GPX1, and GP4), which directly contribute to lower active thyroid hormone input to synthesize UCP1 and redox regulation. Inhibition of the mevalonate pathway also contributes to BAT whitening. Dashed lines indicate pathways in which the direct effects of GPXs are still unclear.

Our results suggest that the predominant impact of Scly disruption lies in limiting the Se supply necessary for synthesizing key selenoproteins that regulate BAT energy metabolism and adaptive thermogenesis. Impairment in thyroid hormone activation due to lower selenoprotein DIO2 levels and activity can contribute to the enhanced lipid accumulation observed in the BAT of Scly KO mice as DIO2 is crucial to normal BAT function. Lower thyroid hormone availability has two major molecular effects, the downregulation of Ucp1 and reduced sensitivity of BAT to norepinephrine and β-adrenergic stimulation⁵⁹. Consistently, we observed a decrease in UCP1 gene expression and protein levels, important as UCP1 is the protein responsible for uncoupling mitochondrial respiration towards heat production⁶⁰. UCP1 has also shown to undergo facultative selenation, a post-translational modification that enhances energy expenditure in the BAT⁶¹, and this process might also be reducing UCP1-dependent heat production in our model due to the localized lower Se content in the BAT. As an additional support for the impairment in DIO2 activity, Scly KO mice also elevated circulating T4 and TSH hormone, with concomitant upregulation of the TSH receptor gene, suggesting a larger input being delivered to BAT in an attempt to compensate for the local impairment in conversion from the prohormone T4 to the active hormone T3. DIO2 is required for thermogenic function¹⁹, and reduced DIO2 activity also decreases activation of thyroid hormone receptors⁶². Once in the brown adipocyte, the larger pool of T4 is not fully converted to T3, due to the reduction in DIO2 activity. Reduced DIO2 activity limits the ability to use stored lipids in the BAT for thermogenesis, thus impairing BAT function⁶³. Normal T3 levels and a decrease in UCP1 is consistent with studies in DIO2 knockout mice that show impaired thermogenesis and a greater susceptibility to diet-induced obesity^64–67. Our data suggests that loss of Scly leads to a similar phenotype. Additionally, elevated serum T4 and TSH levels, and TSH receptor expression along with reduced BAT DIO2 activity, are also connected with high adiposity and obesity^54,68. High circulating TSH decreases energy expenditure, promotes adiposity, and disrupts glucose and lipid metabolism in mice, while loss of the TSH receptor decreases adiposity and increases energy expenditure⁶⁸; the elevation of the TSH receptor may have the opposite effect, and could be contributing here for the BAT whitening phenotype of the Scly KO mice.

Supporting the impact of Scly KO in BAT function in the Scly KO mouse, is that cold exposure resulted in several of the Scly KO mice being removed early from the cold. Together with the decrease in DIO2 activity, this suggests Scly KO mouse exhibit reduced sensitivity to β-adrenergic stimulation, which leads to increased lipid accumulation inside the adipocyte⁵⁷. Disruption of the β3 adrenergic receptor in mice increases susceptibility to diet-induced obesity⁶⁹. Intriguingly, PPARγ protein levels were also elevated in Scly KO mice given a β3-adrenergic agonist compared with WT indicating that despite the loss of DIO2 activity, there is still some residual BAT function. This could partially explain the variability in the response of the Scly KO mice to the cold exposure as mice with better BAT function may have improved adaptation to the cold, rendering them able to tolerate the low temperature for a longer period of time. Also notable was the discrepancy between unchanged PPARγ in mice lacking Scly without β3-adrenergic activation. This discrepancy may be due to the difference in PPARγ regulation. Although room temperature is a mild cold, the more extreme cold exposure at 4°C leads to transient PPARγ dissociation from DNA binding sites instead of a direct upregulation⁷⁰. In contrast, with CL treatment, PPARγ undergoes potent activation via PGC-1α co-activation, which was also shown to have its expression unaffected by the deletion if Scly.

In addition to impairments in BAT function as indicated by reduced DIO2 activity, we also uncovered reduced GPX1 and GPX4 levels in the BAT of Scly KO mice on low Se diet when compared to WT mice on the same diet. GPX1 and GPX4 are important for energy homeostasis as elevated ROS activates thermogenesis and energy consumption^20,21. We did not observe an increase in MDA levels nor GSH levels, which are proxies for increased peroxides and lipid peroxidation, respectively. With decreased levels of both GPX1 and GPX4, we would expect an increase in ROS, however, at least by MDA or GSH, we did not observe an increase in ROS. It may be that decreased GPX1 and GPX4 in the BAT results in lower activity since ROS is not being produced. Potentially, this contributes to the obesity observed in the Scly KO mice as BAT activity is lower in individuals with a higher body mass index⁷¹. Additionally, as GPX4 has a dual role as a suppressor of ferroptosis, we anticipated that a decrease in GPX4 may amplify the susceptibility of brown adipocytes to ferroptosis especially as Se levels were also lower in the BAT of Scly KO mice, and ferroptosis is inversely associated with Se levels⁷². Intriguingly, there were no differences in ferroptotic key players total iron, GSH, or lipid peroxidation between WT and Scly KO BAT. Total GPX activity levels were also unchanged in these mice suggesting that another GPX enzyme may be compensating for the decrease in GPX1 and GPX4 and helping maintaining, at least partially, GSH-dependent redox homeostasis. A previous report shows that BAT specific GPX4-deficient mice resist ferroptosis by reducing NFE2L1-associated proteasomal activity ²³. As GPX4 levels were significantly reduced when Se levels were deficient and not compensated by the secondary protective response to ferroptosis involving FSP1, reduced NFE2L1-associated proteasomal activity could similarly explain the insensitivity to ferroptosis in our model. Another possibility is that BAT is normally resistant to ferroptosis due to either greater exposure to iron availability, or greater capacity for Fe uptake⁷³. Overall, our evidence suggests that, despite the loss of GPX4, the whitening in the BAT of Scly KO mice does not seem a result of changes in ferroptosis sensitivity.

The lack of Scly may consequently buildup Sec and shift this amino acid from the synthesis of critical BAT selenoproteins such as DIO2, GPX1, and GPX4 into other pathways. The lower levels of Se in the BAT supports the observed reduction in synthesis of these selenoproteins and also potentially, the diversion of Sec into some of these other pathways. However, we did not observe activation of other pathways that can use Sec such as taurine synthesis or iron-Se cluster formation as levels of NFS1 and CDO1, respectively key enzymes for these pathways, remained unchanged in the BAT of Scly KO mice. Notably, these pathways are primarily geared to use cysteine, with Sec may be piggybacking on them, a phenomena known to occur in Se metabolism with the transsulfuration/trans-selenation pathway ⁷⁴. In silico analysis of CDO binding capacity to Sec is conflicting^75,76 and no in vivo data exists to dissect this possibility. Yet, Se deficiency affects urinary excretion of taurine⁷⁷, suggesting that these two pathways could be somehow interconnected. Moreover, while sulfur is 10⁵ times more abundant in the human body than Se, Se is significantly more reactive than sulfur⁷⁸, which could favor Se being utilized by one of these alternative pathways to reduce the buildup of Sec. As even a small amount of Se can have a noteworthy biological impact, the lack of SCLY could consequently shift the metabolism of Sec from synthesis of DIO2, GPX1, and GPX4 in the BAT to other pathways, such as lipogenesis.

Nevertheless, we did not detect a shift of Sec towards neither one of these potential pathways, including lipogenesis, and did not detect effects in the proteins that control lipid mobilization and usage, as PPARγ, CHREBP, ATGL and HSL were unaltered in the BAT of Scly KO mice. It may be that an additional, unexplored fate to Sec remains to be uncovered. It also could be that the impairment in BAT function in mice with a disruption in SCLY leads to a slow but steady increase in lipid accumulation, as indicated by the increase in BAT triglycerides, that inhibits the ability of the mice to respond to β3 activation and contributes to obesity development even in individuals consuming adequate levels of Se. Moreover, it could be a combined result of subtle changes in Se metabolic balance that leads overtime to lipid accumulation, and these subtle changes become augmented after puberty, given the sex differences component that these mice have been reported to show previously³. Therefore, the lipid accumulation driven by a disturbance in Scly activity raises specific nutritional implications, as populations with a mutation in Scly may be more susceptible to metabolic phenotypes. In fact, in a cohort of Mexican-Americans, it was observed that those with an intronic indel variant of the ubiquitin-conjugating enzyme E2F and SCLY (UBE2F-SCLY) read-through operon was associated with metabolic phenotypes including an increase in apolipoprotein B (ApoB) that plays a crucial role in the transport of triglycerides⁷⁹.

A notable secondary effect of Scly deficiency was reduced HMG-CoA and cholesterol levels, reflecting suppression of the mevalonate pathway. Inhibition of the mevalonate pathway leads to BAT whitening in both male mice and men, reducing the thermogenic function of brown adipocytes⁵⁸. Therefore, the concurrent decrease of both HMG-CoA and cholesterol from the mevalonate pathway in the BAT of our Scly KO mice combined with elevated triglycerides corroborates that alterations in lipid metabolism and adaptive thermogenesis proteins with the loss of Scly are leading to the BAT whitening and contributing to obesity development in this mouse model. Although we did not identify any alternative CoA intermediates accumulating in BAT, and we have not pinpointed the direct cause for the reduction in HMG-CoA and cholesterol, the results suggest that Scly disruption redirects Sec away from HMG-CoA and the mevalonate pathway.

In conclusion, loss of Scly leads to brown adipose tissue (BAT) with increased lipid accumulation and impaired thermogenic capacity, driven by reductions in key selenoproteins that participate in adaptive thermogenesis responses. The observed rise in triglyceride content and BAT whitening likely reflects a multifactorial disruption involving high TSH, T4 in circulation, reduced DIO2 activity and subsequent downregulation of UCP1, diminished redox control through decreased GPX1 and GPX4, and inhibition of the mevalonate pathway resulting in lower HMG-CoA and cholesterol levels. It should be highlighted that each alteration is potentially modest on its own, yet their combined effects from impaired Sec utilization and reduced selenoprotein synthesis, particularly under Se deficiency, can result in substantial functional deficits in BAT, contributing to the obese phenotype upon global loss of Scly in the organism.

Supplementary Material

Supplemental Figs. S1-S19: DOI: https://doi.org/10.6084/m9.figshare.30511706.v1

Supplemental Tables S1-S2: DOI: https://doi.org/10.6084/m9.figshare.30511706.v1

Supplemental Material for this article can be found online at: https://doi.org/10.6084/m9.figshare.30511706.v1