Adtrp regulates thermogenic activity of adipose tissue via mediating the secretion of S100b

Abstract

Brown and beige adipose tissues dissipate chemical energy in the form of heat to maintain your body temperature in cold conditions. The impaired function of these tissues results in various metabolic diseases in humans and mice. By bioinformatical analyses, we identified a functional thermogenic regulator of adipose tissue, Androgen-dependent tissue factor pathway inhibitor [TFPI]-regulating protein (Adtrp), which was significantly overexpressed in and functionally activated the mature brown/beige adipocytes. Hereby, we knocked out Adtrp in mice which led to multiple abnormalities in thermogenesis, metabolism, and maturation of brown/beige adipocytes causing excess lipid accumulation in brown adipose tissue (BAT) and cold intolerance. The capability of thermogenesis in brown/beige adipose tissues could be recovered in Adtrp KO mice upon direct β3-adrenergic receptor (β3-AR) stimulation by CL316,243 treatment. Our mechanistic studies revealed that Adtrp by binding to S100 calcium-binding protein b (S100b) indirectly mediated the secretion of S100b, which in turn promoted the β3-AR mediated thermogenesis via sympathetic innervation. These results may provide a novel insight into Adtrp in metabolism via regulating the differentiation and thermogenesis of adipose tissues in mice.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-022-04441-9.

Introduction

Adipose tissues play critical roles in controlling the energy balance and metabolism, which makes them natural therapeutic targets to deal with obesity and metabolically disordered diseases [1, 2]. Adipose tissues are composed of brown adipose tissue (BAT) and white adipose tissue (WAT). Brown adipocytes are muscle-like cell lineages with multiple small lipid droplets and with abundant mitochondria [3, 4]. BAT highly expresses uncoupling protein-1 (UCP1) to uncouple respiration for thermogenesis in cold environments [5, 6]. Small mammals, especially rodents, have copious BAT and recent studies have shown that BAT existed in the neck and supraclavicular regions of adult human [7–9]. White adipocytes have large lipid droplets and their main function is to store excess energy [4]. However, in response to various stimuli, such as cold exposure or activator of the β3 adrenergic receptor (β3-AR), WAT can express UCP-1 for thermogenesis and UCP1-expressing WAT is called as beige adipocytes [10–12]. The sympathetic nervous system (SNS) controls the activation and thermogenesis of brown or beige adipocytes [13]. Upon cold exposure, SNS-released catecholamine or noradrenaline binds the β3-AR and activates its downstream pathways to dissipate energy as heat in adipose tissues [14–16].

Androgen-dependent tissue factor pathway inhibitor [TFPI]-regulating protein (ADTRP), also known as C6ORF105, has been reported as a coronary artery disease (CAD) susceptibility gene in human [17]. The deficiency of ADTRP also causes oral cleft syndrome and craniosynostosis in human [18, 19]. Functional studies have shown that androgens may induce ADTRP, and increased ADTRP could regulate the anticoagulant protection in endothelial cells (ECs) in vitro and vascular development, integrity and stability in vivo [20, 21]. ADTRP is further identified as an atypical hydrolytic enzyme to specifically hydrolyze fatty acid esters of hydroxy fatty acids (FAHFAs), a new class of bioactive lipids shown in vitro or in vivo [22–24].

Adtrp is highly expressed in liver, BAT, inguinal WAT (iWAT), epididymis WAT (eWAT), kidney and duodenum in mice [23]. Adtrp has been shown to participate in the regulatory functions of liver and adipose tissues [23], although a recent study has reported that Adtrp did not regulate the lipid and glucose metabolism in liver of mice [25]. However, the underlying functions and molecular mechanisms of Adtrp in adipose tissues are unknown.

Here, we hypothesized that Adtrp might play a crucial role in metabolism and thermogenesis of BAT and iWAT in mice. To test our hypothesis, we generated an Adtrp knockout (KO) mouse model to prove that Adtrp would drive the differentiation of BAT and iWAT stromal vascular fraction (SVF) cells and participate in the metabolism and thermogenesis of BAT and beige iWAT in vitro and in vivo. Additionally, we did mechanistic studies to show whether Adtrp was involved in the secretion of S100 calcium-binding protein b (S100b) to regulate the metabolic homeostasis and thermogenesis in mice.

Materials and methods

Animals

The Adtrp KO mouse model was generated by Nanjing biomedical research institute of Nanjing University via Clustered regularly interspaced short palindromic repeats/ CRISPR associated 9 (CRISPR/Cas9) technology. Genotyping primers are listed in Table S1. 6-week-old male C57BL/6 J mice were purchased from the Si Pei Fu (SPF biotechnology Co., Ltd. China). Mice were housed under specific pathogen-free (SFP) conditions at 25 °C. All the studies were conducted with 8-week-old male mice with chow diet. For acute cold exposure, mouse was individually caged at 4 °C for 8 h, and the core body temperature of WT or Adtrp KO mice were measured every 2 h with a portable intelligent digital thermometer (CCCC Jianyi Technology Development Co., Ltd. China). The thermal images were captured by a thermal camera (FLIR, USA) at 4 °C. For a long-term 4 °C treatment, WT mouse was individually caged at 4 °C for 7 days, and WT mouse individually caged at 25 °C was used as control. For a long-term 16 °C treatment, WT or Adtrp KO mice were individually caged at 16 °C for 7 days. For β3-AR agonist treatment, mice were intraperitoneally injected with 1 mg/Kg CL316,243 (Sigma-Aldrich, USA), every day for 7 days, whereas intraperitoneal injection of 0.9% NaCl solution was used as control. All animals were sacrificed by inhalation of carbon dioxide.

Bioinformatics analysis

Gene microarray data of GSE13432 were re-analyzed by the limma package for biological statistics. RNA-seq data were acquired from the GSE86338, GSE104285, GSE129083, where edgeR or DESeq2 package was used for analysis of differentially expressed genes. The threshold of the differentially expressed genes (DEGs) in GSE86338, GSE104285, GSE13432 and GSE129083 was set as log₂ fold change > 0.5 or log₂ fold change < − 0.5, p value < 0.05. All up-regulated genes are listed in the Supplementary Table S2–S6. Heatmaps for differential genes were analyzed by pheatmap package of R.

Cell culture

Human embryonic kidney cells (HEK) 293T (293T) cell line was purchased from the Cell Bank of the Peking Union Medical College Hospital (China. Licensed by ATCC). The human telomerase reverse transcriptase (hTERT) immortalized cell lines—hTERT A41hBAT-SVF and hTERT A41hWAT-SVF were purchased from American Type Culture Collection (ATCC, USA). Primary BAT SVF cells were isolated from the BAT of the newborn male mice and primary iWAT SVF cells were isolated from the iWAT of 4-week-old male mice as described previously [25–27]. Isolated adipose tissues were digested with 1.5 mg/mL type I collagenase (Sigma-Aldrich) at 37 °C for 40 min. After filtering and centrifuging, BAT and iWAT SVF cells were plated in the culture medium. Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, USA) with 10% fetal bovine serum (FBS, Gibco) at 37 °C with 5% CO₂. For differentiation of the primary adipocytes, cells were treated with the induction medium [DMEM with 10% FBS, 20 nM Insulin (Sigma-Aldrich), 1 nM T3 (Sigma-Aldrich), 0.5 mM Isobutylmethylxanthine (Sigma-Aldrich), 2 μg/mL Dexamethasone (Sigma-Aldrich), 1 μM Rosiglitazone (Sigma-Aldrich) and 125 nM Indomethacin (Sigma-Aldrich)] for 2 days and then the medium was replaced with the maintenance medium (DMEM with 10% FBS, 20 nM Insulin and 1 nM T3) for connective 4 days.

Serum analyses

Blood samples were collected by cardiac puncture after anesthesia by isoflurane inhalation. Serum was obtained from the blood samples after centrifuging at 3000 RPM for 5 min. Serum triglyceride (TG) determination kit (Sigma-Aldrich) and serum-free fatty acid (FFA) assay kit (Abcam, UK) were used to measure the serum TG and FFA of mice, separately.

RNA extraction, reverse transcription PCR (RT-PCR) and quantitative reverse transcription PCR (RT-qPCR)

Trizol reagent (Invitrogen, USA) was used to extract the total RNA from tissues and cells based on the manufacturer’s instructions. Moloney Murine Leukemia Virus (M-MLV) Reverse Transcriptase (Takara, Japan) was used to synthesize cDNAs from the total RNA according to the manufacturer's instructions. RT-qPCR was performed using SYBR Green PCR Master Mix (Invitrogen) with the StepOnePlus System (Invitrogen) according to manufacturer’s instructions. For analyzing the gene expressions of hTERT A41hBAT-SVF cells and hTERT A41hWAT-SVF cells, β-ACTIN was used as the house keeping internal control gene to normalize ADTRP; UCP1; Fatty Acid-Binding Protein 4 (FABP4). For analyzing the gene expressions of mice samples, β-Actin was used as the internal control to normalize Adtrp; Ucp1; Deiodinase, iodothyronine, type II (Dio2); Cell death-inducing DFFA-like effector A (Cidea); Cytochrome c oxidase subunit 8b (Cox8b); Peroxisome proliferative activated receptor, gamma, coactivator 1 alpha (Pgc1α), Peroxisome proliferator-activated receptor gamma (Pparγ); Fabp4; PR domain containing 16 (Prdm16) and S100b. The relative expressions of these genes were calculated by 2^−ΔΔCt method. Primers are listed in Table S1.

Glucose Tolerance Test (GTT) and Insulin Tolerance Test (ITT)

For the GTT assay, 8-week-old male mice were fasted for 16 h and the glucose concentrations from the tail vein blood were measured by a glucometer (Roche, Switzerland) at 0, 15, 30, 45, 60, 90, 120 min after intraperitoneal injections with glucose (Sigma-Aldrich) (2 g/kg body weight). Two weeks later, an ITT assay was conducted with the same mice. Mice were fasted for 6 h and the glucose concentrations from the tail vein blood were measured at 0, 15, 30, 60, 90 min after intraperitoneally injecting with the insulin (Sigma-Aldrich) (0.75 U/kg body weight).

Indirect calorimetry

Mouse was individually caged and maintained with a 12 h light (07:00–19:00) and 12 h dark (19:00–07:00) cycle under 25 °C or 16 °C. Food and water intake, carbon dioxide (VCO₂) generation, oxygen consumption (VO₂) and calorimetry were measured by the Oxylet System (PANLAB, Spain) and total energy expenditure (EE) was calculated by the VO₂ and the resting energy requirements. The data were collected and analyzed by the METABOLISM 3.0 software (PANLAB).

Cellular metabolism assay

The mice BAT and iWAT SVF cells were plated and differentiated for 6 days on the Seahorse 24-well microplates (Agilent, USA). The oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) of the cells were measured using a Seahorse XFe24 Analyzer (Agilent) according to the manufacturer’s protocol and as per earlier publications [28, 29]. For the mitochondrial stress test, cells were treated with 25 mM glucose (Sigma-Aldrich), 1 mM pyruvate (Sigma-Aldrich) and 2 mM l-glutamine (Sigma-Aldrich) contained Seahorse XF Base Medium (Agilent) for 1 h at 37 °C without CO₂, and then measured with 3 times assay cycles for baseline and every treatment, including the injections with following order, 1.5 μM Oligomycin (Agilent), 1 μM Carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP, Agilent), 0.5 μM of Antimycin A (AA, Agilent) and 0.5 μM Rotenone (Rot, Agilent), respectively, with the programs of (mixing for 3 min, waiting for 2 min and measurement lasting for 3 min). For the glycolysis stress test, cells were treated with 1 mM l-glutamine (Sigma-Aldrich) contained Seahorse XF Base Medium (Agilent) for 1 h at 37 °C without CO₂, and then cells were measured with 3 times assay cycles for baseline and every treatment, including the injections with following order, 10 mM Glucose (Agilent), 1 μM Oligomycin (Agilent), and 50 mM 2-Deoxy-d-glucose (2-DG, Agilent) respectively, with the programs of (mixing for 3 min, waiting for 2 min and measurement lasting for 3 min). After these assays, the total cell proteins/each well were extracted and quantified with a Bicinchoninic acid (BCA) protein quantification kit (Thermo Scientific, USA) for normalization. Seahorse Wave Desktop Software (Agilent) was used to analyze the results.

Hematoxylin and Eosin (HE), Immunohistochemistry (IHC) and Immunofluorescence (IF) analyses

The tissues were fixed with 4% paraformaldehyde, embedded in paraffin, and sectioned at 4 μm. After dewaxing, sections were stained with an HE staining kit (Solarbio, China) according to the manufacturer's instructions. For IHC and IF, sections were pretreated with hydrogen peroxide (3%) for 10 min to remove the endogenous peroxidase, followed by antigen retrieval in a microwave for 15 min in 10 mM citrate buffer (pH 6.0). For IHC, anti-UCP1 Rabbit Polyclonal (ab10983; Abcam) was used to incubate the sections for 1 h at 37 °C with 1:1000 dilution and then incubated with the biotinylated Goat anti-Rabbit IgG (31820; Invitrogen) with 1:200 dilution for 30 min at 37 °C. Afterward, the sections were incubated with the Horseradish Peroxidase (HRP) Conjugated Streptavidin (N100; Thermo Scientific) at a dilution of 1:200 for 30 min at room temperature and followed by staining with 3,3′-diaminobenzidine tetra-hydrochloride (DAB) staining kits (Boster, China) according to the manufacturer’s instructions. The nucleus was stained with Hematoxylin (Solarbio) for 2 min at 37 °C. The slides were imaged with VENTANA SYSTEM (Roche). The anti-Adtrp antibody was made by ABclonal (China). To generate the anti-Adtrp polyclonal antibody, the peptide fragment of Adtrp (176-193 amino acids, NP_780626.1) was used as antigen to immunize rabbits for producing the anti-Adtrp polyclonal antibody. For IF, anti-Adtrp Rabbit Polyclonal antibody (ABclonal) and anti-S100b Mouse Monoclonal antibody (66616-1-Ig-1-AP; Proteintech, China) were used to incubate the sections with a dilution of 1:200 for 12 h at 4 °C, and then incubated with Alexa Fluor Plus 594-conjugated Goat anti-Rabbit IgG (A32740; Invitrogen) and Fluorescein isothiocyanate (FITC)-conjugated Goat anti-Mouse IgG (F-2761; Invitrogen) with a dilution of 1:500 for 1 h at 37 °C. The nucleus was stained with DAPI 10 μg/mL (Solarbio) for 10 min at 37 °C. The slides were imaged with the A1 confocal laser microscope (Nikon, Japan).

Plasmid construction and transfection

The pCNDA3.1 plasmid was used to express the mouse Adtrp-flag and Adtrp. For S100b, the pEGFP-C3 and pCNDA3.1 plasmids were used to express S100b-GFP and S100b, respectively. For cAMP-responsive element-binding protein 3 (Creb3), the pEGFP-C3 plasmid was used to express the Creb3-GFP. Plasmids were transfected into cells using a Lipofectamine 3000 transfection reagent kit (Invitrogen) based on the manufacturer's instructions.

Western blot and immunoprecipitation (IP) assays

For western blot assay, total protein was extracted by the protein lysis buffer and quantified with a BCA protein quantification kit (Thermo Scientific). Twenty μg total proteins of cell samples or 40 μg total proteins of tissue samples were separated by Sodium Dodecyl Sulfate–Polyacrylamide Gel Electrophoresis (SDS-PAGE), and transferred onto the Polyvinylidene fluoride (PVDF) membrane (Millipore, USA). After incubating with primary antibodies and the corresponding HRP-conjugated secondary antibodies, the substrate of HRP, an enhanced chemiluminescence (ECL) reagent (Thermo Scientific) was used to detect the signal of the membranes. For IP assay, lysis was incubated with the GFP–antibody bond A/G magnetic beads (Sigma-Aldrich) at 4 °C with gentle rotation for 3 h, and the precipitated complex was subjected to western blot analysis as described previously [30]. In this study, the following antibodies were used in western blot and IP assays: anti-β-Actin antibody (AC026, ABclonal), anti-S100b antibody (66616-1-Ig-1-AP; Proteintech), anti-Adtrp antibody (ABclonal), anti-Heat Shock Protein 90 (HSP90) antibody (60318-1-Ig; Proteintech), anti-Ucp1 antibody (ab10983; Abcam), anti-Pparγ antibody (ab178860; Abcam), anti-Flag antibody (AE005, ABclonal), anti-GFP–antibody (66002-1-Ig Proteintech), HRP-conjugated goat anti‐rabbit IgG (SA00001-1; Proteintech), and HRP-conjugated goat anti-mouse IgG (SA00001-2; Proteintech).

Oil Red O staining

Cells were fixed with 4% paraformaldehyde for 15 min at 25 °C. Then the cells were stained using the Oil Red O Staining Kit (containing with 0.3% Oil Red O) (Beyotime, China) according to the manufacturer's instructions. After staining, the cells were pictured with a camera (650D, Nikon).

Statistical analysis

All data were expressed as the mean ± SEM (standard error of the mean). Two-tail unpaired or paired Student’s t test was applied to analyze the differences between the two groups. The values of *p < 0.05, **p < 0.01, and ***p < 0.001 were indicative of statistical significance and ns was indicative of non-statistical significance. The statistical analysis was performed by GraphPad Prism 8.0. (USA).

Result

Cold exposure or treatment with CL316,243 induced Adtrp expression in mice BAT and iWAT

Cold, as a dominant activator of BAT, via sympathetic nerve controls β3-AR activation and, in turn, induces Ucp1 mediated non-shivering thermogenesis [13, 31, 32]. CL316,243 as an activator of β3-AR can partially mimic the effects of cold stimulation in mice [33, 34]. By mining the data from the GEO dataset (GEO: GSE86338, GSE104285, GSE13432 and GSE129083), we found five highly expressed genes both in BAT and iWAT after 4℃ or CL316,243 treatment including Adtrp, Glycogen synthase 2 (Gys2), Dio2, Ucp1and Elongase of very long chain fatty acids-3 (Elovl3) (Fig. 1A and S1A-D; Table S2-S5). Due to Gys2, Dio2, Ucp1 and Elovl3 have been reported to participate in lipid metabolism [31, 35–37]. We then paid more attention to Adtrp. We treated mice at 4℃ or given them intraperitoneal injection of CL316,243, respectively. Adtrp was found to be significantly overexpressed in BAT after 7 days of treatment at 4 ℃ (vs. 25 ℃ treated mice as control) or with CL316,243 injections (vs. mice with the injection of 0.9% NaCl solution as control) (Fig. 1B–D). Similarly, Adtrp was also overexpressed in iWAT after 4℃ or CL316,243 treatment (Fig. 1E–G). Taken together, these results suggested that the Adtrp expression could be induced in mice BAT and iWAT by either cold or CL316,243 treatment. The similar expression patterns between Adtrp and Ucp1 indicated that Adtrp might be involved in the thermogenesis of BAT and beige adipocytes in mice.

Fig. 1.

Effects of cold exposure or CL316,243 treatment on Adtrp expression in mice BAT and iWAT. A Schematic illustration showing genes expression profile at 4 °C or with CL316,243 treatment of mice BAT and iWAT predicted by overlapping GEO: GSE86338, GSE104285, GSE13432 and GSE129083 data. B RT-qPCR analysis of Adtrp in mice BAT with 4 °C treatment for 7 days (n = 5) (25 °C treated mice as control) or CL316,243 injection for 7 days (n = 5) (mice with the injection of 0.9% NaCl solution as control). C and D Western blot analyses of Adtrp and Ucp1 in mice BAT with 4 °C or CL316,243 treatment for 7 days, Hsp90 as the internal control (n = 4). E RT-qPCR analysis of Adtrp in mice iWAT with 4 °C or CL316,243 treatment for 7 days (n = 5). F and G Western blot analyses of Adtrp and Ucp1 in mice iWAT with 4 °C or CL316,243 treatment for 7 days, Hsp90 as the internal control (n = 3). Error bars represent the means ± SEM of three independent experiments, **p < 0.01

Upregulated ADTRP expression in the differentiation process of brown/beige adipocytes

We re-analyzed the data (GEO, GSE80614) about the differentiated process of human mesenchymal stromal cells (hMSCs) to adipocytes [38, 39]. Along with the expressions of the following biomarkers UCP1, Adiponectin C1Q and collagen domain containing (ADIPOQ), PPARγ, the expression of ADTRP was also up-regulated in the differentiated process of hMSCs (Fig. 2A and Table S6).

Fig. 2.

Expression of ADTRP in the differentiation process of brown/beige adipocytes. A Heat map showing the expressions of Adtrp and the marker genes of the adipogenic differentiation process (GEO: GSE80614). B and C RT-qPCR analyses of ADTRP, UCP1 and FABP4 in hBAT SVF and hWAT SVF cells after adipogenic differentiation. D and E Western blot analyses of ADTRP and UCP1 in hBAT SVF and hWAT SVF cells after adipogenic differentiation, β-Actin as the internal control. F RT-qPCR analysis of Adtrp in mice BAT SVF cells with the adipogenic differentiation process. G Western blot analyses of Adtrp and Ucp1 in mice BAT SVF cells with the adipogenic differentiation process, β-Actin as the internal control. H RT-qPCR analysis of Adtrp in mice iWAT SVF cells with the adipogenic differentiation process. I Western blot analyses of Adtrp and Ucp1 in mice iWAT SVF cells with the adipogenic differentiation process, β-Actin as the internal control. Error bars represent the means ± SEM of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ns: non-statistical significance

To test whether the expression of ADTRP could be up-regulated in the differentiation process, we took advantage of the human hBAT SVF and hWAT SVF cell lines in this study. We found up-regulated ADTRP, both at mRNA and protein levels in mature hBAT and beige hWAT adipocytes (Fig. 2B–E). UCP1 and FABP4 as differentiation markers were significantly overexpressed in both mature hBAT and beige hWAT adipocytes (Fig. 2B–E).

We further differentiated the isolated mice BAT SVF and iWAT SVF cells in vitro and found gradual significant increase of Adtrp in both brown or beige adipocytes during the process of differentiation, in parallel with the increasing expressions of Ucp1 (Fig. 2F–I). Overall, these results provided the evidence that ADTRP was induced in the differentiated process of human and mice brown/beige adipocytes, which indicated that ADTRP might participate in the maturation process of brown/beige adipocytes.

KO of Adtrp suppressed the maturation of BAT and beige of iWAT in vitro

To further investigate the functions of Adtrp in mice, we successfully generated an Adtrp KO mouse with a deletion in exon 2 to exon 4 of Adtrp genomic loci via using the CRISPR/Cas9 system (Fig. S2A and B). Adtrp KO mice BAT and liver tissues were completely devoid of Adtrp expression (Fig. 3A and B; S2C and D).

Fig. 3.

Adtrp KO effects on the maturation of BAT and beige of iWAT in vitro. A and B Western blot analysis of Adtrp in BAT and Liver of Adtrp KO and WT mice, Hsp90 as the internal control. C–E RT-qPCR analyses of Ucp1, Pparγ and Fabp4 in Adtrp KO and WT mice BAT SVF cells with the adipogenic differentiation process. F Western blot analysis of Ucp1 in Adtrp KO or WT mice BAT SVF cells after adipogenic differentiation, β-Actin as the internal control. G–I RT-qPCR analyses of Ucp1, Pparγ and Fabp4 in Adtrp KO and WT mice iWAT SVF cells with the adipogenic differentiation process. J Western blot analysis of Ucp1 in Adtrp KO or WT mice iWAT SVF cells after adipogenic differentiation, β-Actin as the internal control. K and L Oil Red O staining of BAT or iWAT SVF cells in Adtrp KO or WT mice before and after adipogenic differentiation (Scale bars, 1 cm). Error bars represent the means ± SEM of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001; ns, non-statistical significance

To confirm whether Adtrp was essential in the differentiation process of BAT and iWAT in vitro, we analyzed the BAT SVF and iWAT SVF cells isolated from Adtrp KO mice. Ucp1, Pparγ and Fabp4 in Adtrp KO BAT SVF cells were significantly decreased compared with the WT cells in the differentiation process (Fig. 3C–E). The protein levels of Ucp1 in the mature Adtrp KO BAT SVF adipocytes was also markedly lower than that in the WT cells (Fig. 3F). Consistently, the expressions of Ucp1, Pparγ and Fabp4 were also significantly reduced in Adtrp KO iWAT SVF cells compared with the WT cells during the differentiation process of the iWAT SVF cells, where the protein levels of Ucp1 in the Adtrp KO iWAT adipocytes were also lower than that of the WT cells (Fig. 3G–J). Accordingly, the lipid accumulations of the mature Adtrp KO BAT SVF adipocytes were decreased compared with the WT adipocytes (Fig. 3K–L). Overall, we demonstrated that the maturation of BAT and beige iWAT in Adtrp KO mice were suppressed, which indicated the fundamental role of Adtrp in the maturation process of BAT and beige iWAT in mice.

KO of Adtrp caused accumulation of lipid excess of BAT and metabolic disorder in mice

We used further Adtrp KO mice to demonstrate the role of Adtrp in the adipose tissues in vivo. There was no significant difference in the body weights between Adtrp KO and WT mice with normal chow diet (Fig. S3A). Next, we examined the histopathology of the adipose and liver tissues from Adtrp KO and WT mice. There were abundant lipid accumulated areas in BAT in Adtrp KO mice, compared with the WT mice; as well as no morphological differences of iWAT, eWAT and liver in between WT and Adtrp KO mice (Fig. 4A and S3B). The thermogenic genes (Ucp1, Dio2, Cidea, Cox8b, Pgc1α and adipogenic genes (Pparγ and Fabp4) were significantly downregulated in BAT of Adtrp KO mice (Fig. 4B). Consistently, we also found notably decreased Ucp1 and Pparγ expressions in Adtrp KO mice, which suggested that the deletion of Adtrp may impair the thermogenesis and metabolism in mice (Fig. 4C). Accordingly, using the metabolic cages, we observed the significant decreases of VCO₂ generation, VO₂ consumption, and EE in Adtrp KO mice from nightfall to midnight at 25 °C compared with the WT mice (Fig. 4D–F). There was no difference in food or water intake between the Adtrp KO and WT mice at 25 °C (Fig. S3C and D). There was also no difference in the glucose tolerance and insulin sensitivity between the Adtrp KO and WT mice (Fig. S3E and F).

Fig. 4.

Adtrp KO effects on the accumulation of lipid excess in BAT and metabolic disorders in mice. A Histopathological images of BAT, iWAT and eWAT from Adtrp KO mice and WT mice at the age of 8 weeks, (n = 4, Scale bars, 50 μm). B RT-qPCR analyses of Ucp1, Dio2, Cidea, Cox8b, Pgc1a, Pparγ and Fabp4 in BAT of Adtrp KO and WT mice (n = 5). C Western blot analyses of Ucp1 and Pparγ in BAT of Adtrp KO or WT mice, Hsp90 as the internal control (n = 4). D–F VCO₂ generation, VO₂ consumption and EE of Adtrp KO and WT mice in metabolic cages at the age of 8 weeks (n = 6). G and H OCR analyses of the mitochondrial respiration in Adtrp KO and WT differentiated BAT or iWAT SVF adipocytes. I and J ECAR analyses of glycolytic functions in Adtrp KO and WT differentiated BAT or iWAT SVF adipocytes. Error bars represent the means ± SEM of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001; ns, non-statistical significance

Furthermore, the results of the mitochondrial stress OCR tests showed that the differentiated Adtrp KO BAT-SVF and iWAT-SVF adipocytes exhibited significant decline in basal respiration, maximal respiration and spare respiration capacity tests compared with the WT cells in vitro (Fig. 4G and H; S3G and H). Meanwhile, the ATP-linked respiration of the differentiated Adtrp KO iWAT-SVF cells was significantly decreased compared with the WT cells (Fig. S3H). The results of the glycolysis stress ECAR tests demonstrated that glycolysis, glycolytic capacity as well as glycolytic reserve were decreased in the differentiated Adtrp KO BAT-SVF and iWAT-SVF cells compared with the WT cells (Fig. 4I and J; S3I and J). In this context, our results demonstrated that KO of Adtrp resulted in lipid excess accumulation in BAT and impaired the metabolic homeostasis both in vitro and in vivo.

KO of Adtrp suppressed the thermogenesis of BAT and beige iWAT in mice

To further illustrate the influence of Adtrp on thermogenic functions in vivo, we treated mice with acute cold (4 °C for 8 h). The core temperatures of Adtrp KO mice decreased dramatically and thermal images also showed that the surface temperatures of Adtrp KO mice dropped much faster than WT mice (Fig. 5A and B).

Fig. 5.

Adtrp KO effects on the thermogenesis of BAT and beige iWAT in mice. A The core body temperatures of Adtrp KO and WT mice at different time points after exposure to 4 °C (n = 10). B Thermal images of Adtrp KO and WT mice at different time points after exposure to 4 °C (n = 6). C Histological images of BAT, iWAT and eWAT from Adtrp KO mice and WT mice at the age of 8 weeks with 16 °C exposure for 7 days (n = 4, Scale bars, 50 μm). D RT-qPCR analyses of Ucp1, Dio2, Cidea, Cox8b, Pgc1α, Pparγ and Prdm16 in BAT of Adtrp KO and WT mice with 16 °C exposure for 7 days (n = 5). E Western blot analyses of Ucp1 and Pparγ in BAT of Adtrp KO or WT mice with 16 °C exposure for 7 days, Hsp90 as the internal control (n = 4). F RT-qPCR analyses of Ucp1, Dio2, Cidea, Cox8b, Pgc1α, Pparγ and Prdm16 in iWAT of Adtrp KO and WT mice with 16 °C exposure for 7 days (n = 5). G Western blot analyses of Ucp1 and Pparγ in iWAT of Adtrp KO and WT mice with 16 °C exposure for 7 days, Hsp90 as the internal control (n = 3). H–I VCO₂ generation, VO₂ consumption of Adtrp KO and WT mice in metabolic cages at the age of 8 weeks with 16 °C exposure (n = 6). J–K Statistics of AUC data of VCO₂ and VO₂ about Adtrp KO and WT mice at 25 °C and 16 °C. Error bars represent the means ± SEM of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001; ns, non-statistical significance

Next, we then exposed the mice with 16 °C for 7 days. The histological images showed that the lipid droplet areas of BAT, iWAT and eWAT in Adtrp KO mice were more abundant than those of WT mice exposed to 16 °C for 7 days (Fig. 5C). The expressions of thermogenic-related genes (such as Ucp1, Dio2, Cidea, Cox8b, Pgc1α, Pparγ and Prdm16) were also at significantly lower levels in BAT and beige iWAT in Adtrp KO mice than those of WT mice after the treatment of 16 °C for 7 days (Fig. 5D–G). Consistently, using the metabolic cages at 16 °C, we observed that the metabolism of Adtrp KO mice was impaired with the dramatically decrease in VCO₂ generation, VO₂ consumption, and EE (Fig. 5H and I; Fig. S4A). There was no significant difference in food or water intake between the Adtrp KO and WT mice at 16 °C (Fig. S4B and C). The Area Under Curve (AUC) analyses showed that the metabolic capabilities (VCO₂ generation, VO₂ consumption, and EE) of Adtrp KO mice were not increased as much as in WT mice in 16 °C compared with 25 °C (Fig. 5J and K; Fig. S4D). Taken together, these results showed that the metabolism and thermogenesis of BAT and beige iWAT were impaired in Adtrp KO mice.

CL316,243 treatment recovered the thermogenesis of BAT and beige iWAT in Adtrp KO mice

To further detect whether Adtrp directly participated in thermogenesis through stimulation of β3-AR in BAT and beige iWAT, we intraperitoneally injected the β3-AR agonist CL316,243 to Adtrp KO and WT mice for 7 days. Unexpectedly, when we exposed these mice to acute cold (4 °C for 8 h), the core temperatures of Adtrp KO mice did not have any difference compared with the WT mice (Fig. S4E). Histopathological images showed no morphological differences in BAT, iWAT and eWAT between the WT and Adtrp KO mice after 7 days of CL 316,243 treatment (Fig. 6A).

Fig. 6.

CL316,243 treatment effects on the potential recovery of the thermogenesis in BAT and beige iWAT of Adtrp KO mice. A Histopathological images of BAT, iWAT and eWAT from Adtrp KO mice and WT mice at age of 8 weeks with CL316,243 injections for 7 days (n = 4, Scale bars, 50 μm). B RT-qPCR analyses of Ucp1, Dio2, Cidea, Cox8b, Pgc1α, Pparγ and Prdm16 in BAT of Adtrp KO and WT mice with CL316,243 treatment for 7 days (n = 5). C Western blot analyses of Ucp1 and Pparγ in BAT of Adtrp KO or WT mice with CL316,243 treatment for 7 days, Hsp90 as the internal control (n = 3). D RT-qPCR analysis of Ucp1, Dio2, Cidea, Cox8b, Pgc1α, Pparγ and Prdm16 in iWAT of Adtrp KO and WT mice with CL316,243 treatment for 7 days (n = 5). E Western blot analyses of the expression of Ucp1 and Pparγ in iWAT of Adtrp KO and WT mice with CL316,243 treatment for 7 days, Hsp90 as the internal control (n = 3). F and G The levels of TG and FFA in the serum of Adtrp KO and WT mice with 16 °C exposure for 7 days or CL316,243 treatment for 7 days (n = 5). H and I IHC images of Ucp1 in BAT and iWAT of Adtrp KO and WT mice with 16 °C exposure for 7 days or CL316,243 treatment for 7 days (n = 4, scale bars, 50 μm). Error bars represent the means ± SEM of three independent experiments. *p < 0.05, ns: non-statistical significance

There was no significant difference in the expression profile of thermogenic genes in BAT and beige iWAT between the Adtrp KO and WT mice treated with CL316,243 (Fig. 6B–E). This indicated that the thermogenic functions of Adtrp KO mice recovered after 7 days’ of CL316,243 treatment. Moreover, we found no significant differences in serum TG and FFA levels of Adtrp KO mice compared with the WT mice at 25 °C, whereas the TG and FFA levels of Adtrp KO mice were significantly higher than those of WT mice at 16 °C for 7 days. TG and FFA levels of Adtrp KO mice were significantly lower than those of the WT mice after 7 days’ of CL316,243 treatment (Fig. 6F and G), which indicated CL316,243 treatment activated the lipid metabolism and thermogenesis of Adtrp KO mice. In addition, the immunohistological analysis images also showed that the expression of Ucp1 in BAT and iWAT of Adtrp KO mice was significantly less abundant than those of the WT mice, both at 25 °C and 16 °C. However, there was no significant difference in the expression of Ucp1 between the Adtrp KO and WT mice after 7 days’ CL316,243 treatment (Fig. 6H and I). These results demonstrated that the CL316,243 treatment could recover the thermogenesis of BAT and beige iWAT in Adtrp KO mice.

Adtrp mediated the secretion of S100b in BAT and iWAT of mice

To understand how Adtrp may impact the β3-AR mediating thermogenesis, we searched the NCBI database, and found 5 possible protein candidates (CKLF like MARVEL transmembrane domain containing 7 [CMTM7], [CREB3], S100B, Transmembrane p24 trafficking protein family member 8 [TMED8] and Vitronectin [VTN]), which may interact with the Adtrp [40–42] (Fig. S5A). There has been reports suggesting Creb3 and S100b to participate in the metabolism of adipose tissues [43, 44]. Using IP analysis, we confirmed that Adtrp could interact with S100b instead of Creb3 (Fig. 7A and Fig. S5B). We also observed that KO of Adtrp did not impact the expression of S100b in mice BAT and iWAT (Fig. 7B and C; Fig. S5C and D).

Fig. 7.

Adtrp mediated secretion of S100b in BAT and iWAT of mice. A IP by anti-GFP–antibody bond A/G magnetic beads, and Western blot analyses of S100b and Adtrp in 293T cells. B and C Western blot analysis of S100b in BAT and iWAT of Adtrp KO or WT mice, Hsp90 as the internal control (n = 3). D Western blot analysis of S100b in Adtrp KO or WT mice BAT SVF cells after adipogenic differentiation, β-Actin as the internal control. E Western blot analysis of S100b in a conditional medium of Adtrp KO or WT mice BAT SVF cells after adipogenic differentiation. F Western blot analysis of S100b in Adtrp KO or WT mice iWAT SVF cells after adipogenic differentiation, β-Actin as the internal control. G Western blot analysis of S100b in a conditional medium of in Adtrp KO or WT mice SVF iWAT SVF cells after adipogenic differentiation. H Western blot analyses of Adtrp and S100b in different Adtrp or S100b expressed 293T cells, β-Actin as the internal control. I Western blot analysis of S100b in a conditional medium in different Adtrp or S100b expressed 293T cells. J Immunofluorescence staining analyses of Adtrp (Red), S100b (Green), and nucleus (Blue) in mice BAT and iWAT (scale bars, 10 μm)

It has been demonstrated that S100b could be secreted from thermogenic adipocytes to process the neurotrophic effects on the sympathetic innervations of BAT and beige iWAT to stimulate the β3-AR mediated activation of thermogenesis [45]. Thereafter, we hypothesized that Adtrp might bind to the S100b and mediate the secretion of S100b. To verify this hypothesis, we collected the conditional medium from the differentiated BAT and iWAT SVF adipocytes. We found that the secretion of S100b in Adtrp KO adipocytes was significantly lower than that in the WT adipocytes with the similar expression patterns of S100b in BAT and iWAT adipocytes (Fig. 7D–G). Consistently, we also showed that Adtrp could mediate the secretion of S100b in 293T cells (Fig. 7H and I). Moreover, the immunofluorescence staining analyses of Adtrp (Red) and S100b (Green) showed the co-localization of Adtrp and S100b in BAT and iWAT (Fig. 7J). Together, these findings demonstrated that Adtrp could regulate the thermogenic functions of BAT and iWAT by mediating the secretion of S100b in mice.

Discussion

BAT and beige WAT plays key regulatory role in metabolism and non-shivering thermogenesis in small mammals and humans [46, 47]. In this study, we determined that Adtrp was significantly up-regulated in the differentiated brown/beige adipocytes and in the thermogenic process of BAT and beige iWAT in mice. Moreover, the Adtrp KO mice showed the metabolic disorders and cold intolerance with significantly lower expressions of thermogenic genes, and β3-AR agonist CL316,243 treatment could recover the thermogenesis of BAT and beige iWAT in Adtrp KO mice. Overall, we demonstrated that Adtrp might serve as a novel fundamental key regulator of metabolism, especially in the thermogenesis of adipose tissues.

The multiple functions of ADTRP have been reported by several studies as stated below. ADTRP acted as a CAD susceptibility gene [17], and ADTRP also regulated the anticoagulant protection of ECs and participated in the vascular development, integrity and stability [20, 21]. Furthermore, ADTRP served as an atypical hydrolytic enzyme to hydrolyze FAHFAs [22, 23]. Hereby, we found a novel unprecedented function of Adtrp, its participation in the metabolism of adipose tissues, especially in the regulation of the non-shivering thermogenesis in mice.

The noradrenaline, released by SNS, activates the β3-AR pathway, in turn to drive the non-shivering thermogenesis of BAT and beige iWAT in mice [48–50]. Researchers have reported that some genes could mediate the activations of BAT and beige iWAT via regulating the sympathetic innervations in mice. A study has shown that the adipose-specific ablation of Prdm16 resulted in a decrease in the sympathetic innervations of beige iWAT [51]. Recently, another study has shown that Calsyntenin 3β (Clstn3β) mediated the secretion of S100b, which posed a neurotrophic factor, enhancing the sympathetic innervations of BAT and beige iWAT [45]. Consistent with this study, we demonstrated that Adtrp could bind to S100b and mediate the secretion of S100b, which in turn promoted β3-AR mediated thermogenesis via enhancing the sympathetic innervations of BAT and beige iWAT in mice. The secretion pathway of S100b might be driven by the Clstn3β in the endoplasmic reticulum and Adtrp in the cytomembrane.

When mice were exposed to cold, the serum TG and FFA were transported to BAT and beige iWAT for non-shivering thermogenesis [52]. However, in our study, we showed that serum TG and FFA of Adtrp KO mice were not consumed as much as WT in a long-term 16 °C exposure, which indicated that the thermogenic function of Adtrp KO mice was restrained.

A previous study has suggested that ADTRP involves in the osteogenic differentiation of bone marrow MSC (BM-MSC) with the osteogenic medium. However, no any differences in bone/cartilage/oral cleft between Adtrp KO and WT mice were found [21]. In contrast to this study, we showed that Adtrp was essential in the differentiation process of mice brown/beige adipocytes. In addition, Adtrp KO caused the dysplasia of BAT with an accumulation of lipid excess in mice. The discrepancy between the osteogenic and adipogenic differentiation might be due to the complex roles of Adtrp in different tissues at different developmental periods.

Studies have demonstrated that androgen induced high expression of ADTRP in ECs could inhibit the atherosclerotic process [20, 53]. Another study showed that PPARγ transcriptionally activated the expression of ADTRP in monocytes/macrophages and human atherosclerotic plaques [54]. Different from these studies, we demonstrated that cold exposure or β3-AR agonist CL316,243 treatment could induce the expression of Adtrp. We did not find any phenotypic differences between female and male Adtrp KO mice, which indicated that the functions of Adtrp in thermogenic adipose tissues were androgen independent. Future studies by generating conditional adipose tissues Adtrp KO mice might extend our knowledge on the precise role of Adtrp in the metabolism of adipose tissues.

Earlier few studies have illustrated that suppressing the BAT thermogenic capacity might damage the glucose homeostasis. However, in our current study, we found that there were no significant changes in body weights, serum biochemical indices, glucose tolerances or insulin sensitivities between WT and KO mice at 25 °C with regular chow diet. Consistent with our results, a recent study also failed to find any significant differences in glucose tolerances or insulin sensitivities between Adtrp KO and WT mice with high-fat diet (HFD) [23]. We got similar results on the glucose homeostasis of Adtrp KO mice, which may indicate that Adtrp is not associated with the process of obesity and diabetes. Similar to Adtrp KO data, Estrogen-related receptor gamma (ERRγ) KO mice exhibited a pronounced whitening of BAT and a decrease of thermogenic capacity, without any changes in body weight, glucose tolerances, insulin sensitivities with chow diet or HFD [55]. We speculated that there might be a compensating effect of the glucose homeostasis in Adtrp KO mice. Adtrp KO inhibited the hydrolysis of FAHFAs in mice adipocytes and the high level of FAHFAs could enhance insulin-stimulated glucose uptake [23, 24]. On the other hand, deletion of Adtrp caused accumulation of lipid excess in BAT and metabolic disorders, which might lead to further glucose intolerance in mice. This dual-function of Adtrp might balance and maintain the normal glucose metabolism in Adtrp KO mice.

ADTRP is a CAD susceptibility gene and a single-nucleotide polymorphism (SNP) of rs6903956 is associated with the low expression of ADTRP in human [17]. Some studies have shown that perivascular adipose tissue (PVAT) was similar to classical BAT, and associated with CAD, especially, in the development of atherosclerosis in the aorta [56–58]. A study has shown that BAT-Specific PPARγ KO impaired the development of PVAT and aggravated the atherosclerosis in mice [59]. In line with the above studies, one could assume that ADTRP might affect the occurrence of CAD via regulating the functions of PVAT and BAT in human.

One of the limitations of our present study was that we did not study the functions of Adtrp in PVAT and CAD in mice. In future, further studies, such analyses of the PVAT and CAD in Adtrp KO mice, as well as to establish the mice model of CAD and atherosclerosis to study the relations between ADTRP and CAD should be done. Another limitation of this study could be that we did not investigate the clinical metabolic disorders related to CAD. We are going to collect the samples from the metabolic disorders related to CAD patients and record the corresponding follow-up information to investigate the relationship between ADTRP and the metabolic disorders related to CAD in our future project.

In summary, the present study demonstrated that Adtrp could be a novel fundamental regulator of the differentiation and thermogenesis of BAT and beige iWAT. Adtrp mediated the secretion of S100b to promote the β3-AR mediated thermogenesis via enhancing the sympathetic innervations of BAT and beige iWAT in mice (summarized in Fig. 8).

Fig. 8.

A proposed working model of Adtrp action in brown and beige adipose tissues. Adtrp mediates the secretion of S100b to promote the β3-AR mediated thermogenesis via enhancing the sympathetic innervation of BAT and beige iWAT in mice

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

Adtrp

Androgen-dependent tissue factor pathway inhibitor (TFPI)-regulating protein

BAT

Brown adipose tissue

β3-AR

β-Adrenergic receptor

S100b

S100 calcium-binding protein b

WAT

White adipose tissue

Ucp1

Uncoupling protein-1

SNS

Sympathetic nervous system

CAD

Coronary artery disease

ECs

Endothelial cells

FAHFAs

Fatty acid esters of hydroxy fatty acids

iWAT

Inguinal WAT

eWAT

Epididymis WAT

GEO

Gene Expression Omnibus

KO

Knockout

SVF

Stromal vascular fraction

CRISPR/Cas9

Clustered regularly interspaced short palindromic repeats/CRISPR associated 9

SFP

Specific pathogen-free

DEGs

Differentially expressed genes

hTERT

Human telomerase reverse transcriptase

DMEM

Dulbecco’s modified Eagle’s medium

FBS

Fetal bovine serum

TG

Triglyceride

FFA

Free fatty acid

RT-PCR

Reverse transcription PCR

RT-qPCR

Quantitative reverse transcription PCR

M-MLV

Moloney Murine Leukemia Virus

Fabp4

Fatty acid-binding protein 4

Dio2

Deiodinase, iodothyronine, type II

Cidea

Cell death-inducing DFFA-like effector A

Cox8b

Cytochrome c oxidase subunit 8b

Pgc1α

Peroxisome proliferative activated receptor, gamma, coactivator 1 alpha

Pparγ

Peroxisome proliferator-activated receptor gamma

Prdm16

PR domain containing 16

GTT

Glucose Tolerance Test

ITT

Insulin Tolerance Test

VCO₂

Carbon dioxide generation

VO₂

Oxygen consumption

EE

Energy expenditure

OCR

Oxygen Consumption Rate

ECAR

Extracellular Acidification Rate

FCCP

Carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone

AA

Antimycin A

Rot

Rotenone

2-DG

2-Deoxy-d-glucose

BCA

Bicinchoninic acid

HE

Hematoxylin and Eosin

IHC

Immunohistochemistry

IF

Immunofluorescence

HRP

Horseradish Peroxidase

DAB

3,3′-Diaminobenzidine tetra-hydrochloride

FITC

Fluorescein isothiocyanate

Creb3

cAMP-responsive element-binding protein 3

IP

Immunoprecipitation

SDS-PAGE

Sodium Dodecyl Sulfate–Polyacrylamide Gel Electrophoresis

PVDF

Polyvinylidene fluoride

ECL

Enhanced chemiluminescence

Hsp90

Heat shock protein 90

hMSCs

Human mesenchymal stromal cells

ADIPOQ

Adiponectin C1Q and collagen domain containing

Gys2

Glycogen synthase 2

Elovl3

Elongase of very long chain fatty acids-3

AUC

Area Under Curve

CMTM7

MARVEL transmembrane domain containing 7

TMED8

Transmembrane p24 trafficking protein family member 8

VTN

Vitronectin

Clstn3β

Calsyntenin 3β

BM-MSC

Bone marrow MSC

HFD

High-fat diet

ERRγ

Estrogen-related receptor gamma

SNP

Single-nucleotide polymorphism

Author’s contributions

XL designed the study concept and supervised the project and analyzed the data. PL, RS and XL interpreted the data. PL, RS and YD conducted the experiments. PL, RS and HL analyzed RNA-seq and microarray data. All authors approved the final content.

Funding

This study was supported by grants from the National Natural Science Foundation of China (82171854 and 31970802), Beijing Municipal Natural Science Foundation (7202099) and the Medical University of Bialystok, Poland (SUB/1/DN/20/006/1104, to Xiangdong Li).

Availability of data and material

The datasets or materials generated in the current study are available on reasonable request.

Declarations

Conflict of interests

The authors declare that they have no conflict of interest.

Ethics approval

All animal studies were approved by the ethical committee of the China Agricultural University (No.: AW32201202-3-2).

Consent to participate

Not applicable. No human subjects were recruited for this study.

Consent for publication

Not applicable.