Identification of BATSP1 as a regulator of adipocyte thermogenesis

Previously, we successfully isolated human primary brown adipocytes and optimized the differentiation conditions [13]. To identify extracellular peptides responsible for thermogenesis, we stimulated mature human brown adipocytes with forskolin (FSK) for 4 h. Serum-free conditioned medium was then collected to extract peptides and further sent for liquid chromatography-tandem mass spectrometry (LC‒MS/MS) analysis (Fig. 1A). A total of 4015 secreted peptides derived from 1322 protein precursors were identified in both groups (Table S1). Of these peptides, 357 showed a significant difference in abundance, with a fold change of > 1.3 and a p value of < 0.05 (t test), namely, 236 peptides with an increased abundance and 121 peptides with a decreased abundance upon FSK treatment (Table S2). We then analyzed the general features of these peptides. Most of the peptides had a molecular weight of between 0.4 and 1.6 kDa and had an acidic isoelectric point (pI) range (3.0–6.0) (Fig. S1A). A typical characteristic of this set of identified peptides was the presence of 15 peptides derived from the Mucin 16 (MUC16) protein. The top 20 precursor proteins from which the highest number of related peptides were derived are listed in Fig. S1B. Many of these precursors have known biological functions in pathways such as the notch signaling pathway, hedgehog signaling pathway, basal cell carcinoma, and regulation of lipolysis in adipocytes (Fig. 1B). The 25 peptides with the greatest increases and decreases in abundance, as determined by the highest fold changes, were visualized in heatmaps (Fig. S1C).

Fig. 1

Identification of BATSP1 as a regulator of adipocyte thermogenesis. A Flow chart showing the search strategies used to identify human brown adipocyte-secreted peptides. B Top KEGG pathways significantly enriched in the precursor proteins identified from the dysregulated peptides. C Volcano plot depicting differentially expressed extracellular peptides in brown adipocytes treated with either FSK or vehicle control. D Representative confocal images of adipocytes treated with FITC-labeled BATSP1. Cells were costained with LipidTox and DAPI. E Western blot analysis of UCP1 expression in fully differentiated brown and white adipocytes after BATSP1 stimulation for 6 h. F Basal OCR, proton leakage, ATP production and maximal respiration in BATSP1-treated brown adipocytes. G Thermogenic gene expression in brown adipocytes. H, I Cellular respiration and thermogenic gene expression in white adipocytes were examined by a Seahorse analyzer and real-time quantitative PCR (RT‒qPCR), respectively, after BATSP1 treatment. hBA, human brown adipocytes; hWA, human white adipocytes. The data are presented as the means ± SDs; *p < 0.05; **p < 0.01 by unpaired Student’s t test

We then tried to screen for peptides that may be involved in adipocyte thermogenesis regulation. Specifically, the secretion of BATSP1 was strikingly induced by FSK signaling (Fig. 1C); BATSP1 is composed of AA 160–178 of short-chain enoyl-CoA hydratase (ECHS1) and has the sequence FAQPEILIGTIPGAGGTQR (Fig. S2A). Multiple sequence alignment revealed that this domain is highly conserved between humans and mice (Fig. S2B). We next synthesized FITC-labeled BATSP1 and found that the peptide clearly entered both brown and white adipocytes and dispersed throughout the cytoplasm (Fig. 1D). These data indicate the physiological regulation of BATSP1 in adipose cells. BATSP1 is a bona fide regulator of adipose thermogenesis that was specifically screened from the upregulated secreted peptides. UCP1 protein expression was markedly induced by BATSP1 in brown adipocytes (Fig. 1E). The OCR was also increased by BATSP1, as indicated by measurements of the basal OCR, proton leakage, ATP production and maximal respiration (Fig. 1F and Fig. S2C). At the molecular level, the transcript levels of thermogenic genes (Ucp1, Pparα, Pgc1α, Cpt1α, Cideα and Cyt C) were also increased in BATSP1-treated brown adipocytes (Fig. 1G). To assess whether BATSP1 can potentiate thermogenesis in white adipocytes, fully differentiated human primary white adipocytes were treated with synthesized BATSP1. In concordance with the above results, the UCP1 expression level, the OCR and the levels of thermogenic markers were significantly elevated in response to BATSP1 (Fig. 1E , H, I and Fig. S2C). We finally performed a Cell Counting Kit-8 assay to exclude the possibility of cytotoxic effects of BATSP1, and we found that BATSP1 had no effect on cell viability (Fig. S2D). Together, these results strongly suggest that BATSP1 is a brown adipokine that is tightly linked to the activation of adipose thermogenesis.

BATSP1 promotes adipose thermogenesis in mice

We next revealed the tissue distribution of BATSP1 in mice by i.p. injection. The signal produced by FITC was higher in the liver, kidney, iWAT and eWAT—and to a lesser extent, in BAT and muscle—3 h after injection (Fig. S3A), suggesting that BATSP1 could be taken up by adipose tissues and perform its function. To further investigate the function of BATSP1 in vivo, C57BL/6 J mice fed a chow diet were treated with BATSP1 or vehicle control for 2 weeks. BATSP1 administration in vivo did not markedly affect body weight or food intake (Fig. 2A and Fig. S3B) or the serum concentrations of glucose and insulin under fasting conditions (Table S3). However, the surface temperature specifically at the interscapular region corresponding to the location of BAT was significantly increased after BATSP1 treatment (Fig. 2B), and the mice also exhibited marked resistance to cold (Fig. 2C). In addition, BATSP1-treated mice showed increased glucose uptake in BAT compared with vehicle control mice, as visualized by PET/CT imaging (Fig. 2D). However, no major difference was observed in mitochondrial biogenesis, as evaluated by TEM (Fig. 2E). Histological analysis revealed markedly increased expression of UCP1 in both BAT and iWAT; however, the volume of adipocytes was not affected by BATSP1 (Fig. 2F). The mRNA and protein expression of thermogenic genes was consistently induced by BATSP1 (Fig. 2G and Fig. S3C). These experiments suggest that BATSP1 is involved in regulating adipose thermogenesis in vivo, whereas these changes were attenuated when scrambled peptides were used (Fig. S3D–F). Furthermore, we observed no differences in the serum concentrations of alanine aminotransferase (ALT) and aspartate aminotransaminase (AST), biomarkers of liver damage (Table S3), between these two groups of mice. Additionally, no histological changes were observed in other tissues, such as brain, heart, kidney, liver, lung, spleen, intestine, pancreas and muscle tissue (Fig. S3G), indicating the lack of obvious side effects of BATSP1.

Fig. 2

BATSP1 promotes adipose thermogenesis in mice. C57BL/6J mice fed a chow diet were injected intraperitoneally with 5 mg/kg/day BATSP1 or vehicle (saline) for 14 days (n = 5 mice/group). A Body weight time course. B Representative thermal images (left) and calculated interscapular temperatures (right). C Measurements of rectal temperatures in mice housed at 4 °C. D PET/CT images of mice after injection of 18F-FDG. E Representative transmission electron micrographs of BAT and iWAT. F Representative histological data (H&E and immunochemical staining of UCP1) of BAT and iWAT. G Expression of marker genes related to thermogenesis. The data are presented as the means ± SDs; *p < 0.05; **p < 0.01 by unpaired Student’s t test

BATSP1 increases whole-body energy expenditure in vivo

To further investigate the effects of BATSP1 on whole-body metabolism, we performed metabolic cage studies with BATSP1-treated mice for 2 weeks. Interestingly, despite the tendency toward increased thermogenesis, whole-body energy expenditure under ambient conditions (25 °C) was not significantly different (Fig. S4A–E). We then asked whether thermogenically demanding conditions are required for the function of BATSP1. Mice were housed at 4 °C for 3 days after 2 weeks of BATSP1 injection. As shown in Fig. 3A–C, BATSP1 increased whole-body oxygen consumption (VO2) and CO2 production (VCO2) and elevated heat generation after cold exposure. However, the respiratory exchange ratio (RER) and locomotor activity were unchanged (Fig. S4F and G). We then confirmed that the increase in oxygen consumption was due to the upregulation of BAT activity in cold-exposed mice. Notably, the uptake of 18F-FDG, a glucose tracer, was significantly increased, as determined by PET/CT imaging (Fig. 3D). Accordingly, the mitochondrial content was substantially increased in mice injected with BATSP1 compared with those injected with vehicle control upon exposure to a cold environment (Fig. 3E).

Fig. 3

BATSP1 increases whole-body energy expenditure in vivo. Mice were treated with BATSP1 for 14 days and then exposed to 4 °C for 3 days (n = 5/group). Whole-body energy expenditure was evaluated by measuring oxygen consumption (VO2) (A), carbon dioxide release (VCO2) (B) and heat production (C). D 18F-FDG PET/CT imaging. E Representative electron micrographs of BAT and iWAT. The data are presented as the means ± SDs; *p < 0.05; **p < 0.01 by unpaired Student’s t test

BATSP1 ameliorates diet-induced obesity under mild cold exposure

BATSP1 facilitates a negative energy balance by increasing adipose thermogenesis and therefore may hold promise for the development of novel anti-obesity approaches. To determine the potential therapeutic uses of BATSP1, mice fed a HFD were injected with BATSP1 twice a week. We first conducted the experiment with mice housed at room temperature. After 16 weeks of treatment, no significant reduction in body weight gain or food intake was observed in the BATSP1 group (Fig. S5A and B) compared with the control group (injected with vehicle). However, we did observe a higher skin temperature in the BAT region in BATSP1-treated mice fed a HFD (Fig. S5C). Consistent with this finding, the expression of UCP1 was strongly induced in BAT and iWAT, as indicated by histological and Western blot analyses (Fig. S5D and E). The effects of BATSP1 on energy expenditure may be masked at room temperature, at which no appreciable cold stimulation occurs and heat production is only partially activated. Therefore, we conducted a second experiment to determine whether the increased thermogenesis induced by chronic BATSP1 injection leads to a reduction in the mass of mice upon cold stress. Notably, mice housed in a mildly cold environment (16 °C) exhibited marked resistance to HFD-induced body weight gain (Fig. 4A), although their food intake was similar to that of control mice (Fig. 4B). Lipid stores were accordingly decreased in iWAT and eWAT pads after 16 weeks of treatment, with no effect on lean mass (Fig. 4C–E). BATSP1 also induced pronounced decreases in fasting serum glucose and insulin (Fig. 4F), as well as in the concentrations of triglycerides (TG) and total cholesterol (TC) (Fig. 4G).

Fig. 4

BATSP1 ameliorates diet-induced obesity under mild cold exposure. Mice fed a HFD were housed at 16 °C and injected with 5 mg/kg BATSP1 or vehicle twice per week for a period of 16 weeks (n = 5 mice/group). A–E Body weight (A), food intake (B), fat mass ratio (C), morphology of fat tissues (D) and lean mass (E) of HFD-fed mice. F Serum glucose and insulin concentrations. G Lipid concentrations. H 18F-FDG PET/CT imaging. I Representative thermal images (left) and calculated interscapular temperatures (right). J, K A GTT and an ITT were performed in HFD-fed mice treated prophylactically with BATSP1 or vehicle control. The data are presented as the means ± SDs; *p < 0.05; **p < 0.01 by unpaired Student’s t test

To determine whether this body weight loss results from the stimulation of BAT thermogenesis, we evaluated the phenotypic effects of BATSP1 treatment on BAT and iWAT. 18F-FDG PET/CT scanning showed that glucose uptake was robustly increased in BATSP1-treated mice in the cold environment (Fig. 4H). Consistent with this finding, the surface temperature was significantly increased in the interscapular region by BATSP1 treatment (Fig. 4I). Hematoxylin and eosin (H&E) staining revealed that BAT and iWAT from BATSP1-treated mice contained smaller lipid droplets (Fig. S5F). Immunohistological staining indicated that BATSP1 treatment in a cold environment resulted in an increase in the UCP1 protein level (Fig. S5F). Consistent with this result, Western blot analysis showed 1.7-fold and 1.6-fold increases in UCP1 expression in BAT and iWAT, respectively, in BATSP1-treated animals (Fig. S5G). Along with a decreased body weight and reduced fat mass, BATSP1-treated mice also exhibited marked improvements in glucose handling and insulin action (Fig. 4J, K), as revealed by the GTT and ITT. These data demonstrate a new function for BATSP1 in mediating resistance to HFD-induced obesity and improving certain aspects of metabolic health under cold exposure.

BATSP1 regulates the subcellular localization of FOXO1

To delineate the molecular mechanism responsible for BATSP1-regulated thermogenesis, we performed RNA sequencing (RNA-seq) in BATSP1-treated brown adipocytes. Transcripts with a fold change of ≥ 2 and false discovery rate-adjusted p value (q value) of < 0.05 were considered differentially expressed genes. A total of 89 genes with significant differences in expression were identified (Table S4), of which 27 were upregulated and 62 were downregulated (Fig. S6A and B). Pathways associated with ABC transporters, pathogenic Escherichia coli infection and the FoxO signaling pathway were enriched in these genes (Fig. S6C). Specifically, FOXO1, a key player in the FOXO family, plays a crucial role in adipocyte metabolism as a common transcription factor [15]. To directly test the involvement of the FOXO1 signaling pathway, we first determined the expression level of FOXO1. However, BATSP1 treatment did not alter the mRNA level of FoxO1 in either brown or white adipocytes (Fig. S6D). Consistent with this finding, the protein expression level of FOXO1 was also unaffected in these cells upon BATSP1 treatment (Fig. S6E). Thus, the transcription and translation of FOXO1 cannot mechanistically explain the thermogenic reprogramming mediated by BATSP1.

Accumulating studies have highlighted the physiological role of nucleocytoplasmic shuttling of FOXO1 in regulating its activity [15], thus affecting UCP1 transcription. To precisely visualize the subcellular localization of FOXO1 in response to BATSP1 stimuli, we performed immunofluorescence staining in primary brown and white preadipocytes. As shown in Fig. 5A, FOXO1 remained in the nucleus under normal conditions in vehicle-treated cells, while BATSP1 stimulation decreased the content of FOXO1 in the nucleus and increased cytosolic FOXO1 accumulation, indicating that BATSP1 is involved in FOXO1 nuclear exclusion. Western blot analysis of subcellular fractions isolated from BATSP1-treated brown and white adipocytes showed similar trends (Fig. 5B). Since phosphorylation has been shown to be responsible for FOXO1 nuclear exclusion, we analyzed FOXO1 phosphorylation at S256, which directly affects its translocation. As expected, the level of phosphorylated FOXO1 was significantly elevated by BATSP1 in adipocytes (Fig. 5C, D). Therefore, BATSP1 controls the subcellular localization of FOXO1 and simultaneously affects its transcriptional function.

Fig. 5

BATSP1 regulates the subcellular localization of FOXO1. A Subcellular localization of FOXO1 in brown and white adipocytes treated with or without BATSP1 for 3 h. B Human brown and white adipocytes were treated as indicated in Panel A, and the cytosolic and nuclear fractions were isolated and analyzed by Western blotting. HSP90 and Lamin B1 were used as controls for the cytosolic and nuclear fractions, respectively. C, D Representative immunoblot showing the level of FoxO1 phosphorylated at S256 in brown and white adipocytes treated with or without BATSP1. The data are presented as the means ± SDs; **p < 0.01 by unpaired Student’s t test

BATSP1 releases the transcriptional inhibition of UCP1 by FOXO1

Because FOXO1 has been shown to inhibit UCP1 gene expression [16], we analyzed the effect of BATSP1 on UCP1 transcriptional activity. We then used a dual human Ucp1-luciferase reporter system to examine the role of FOXO1 translocation resulting from BATSP1 treatment in controlling Ucp1 gene transcription. As shown in Fig. 6A, the luciferase reporter assay showed that FoxO1 overexpression mediated by adenoviral transduction indeed suppressed Ucp1 transcription, whereas BATSP1 stimulation partially restored Ucp1 transcription, increasing it by 45%. In accordance with the released inhibition of UCP1 transcription after BATSP1 treatment, we observed increased UCP1 expression in brown and white adipocytes (Fig. 6B, C and Fig. S6F). To further confirm the importance of BATSP1-FOXO1 signaling in regulating adipose thermogenesis in vitro, we measured cellular respiration with a Seahorse extracellular flux analyzer. Our findings showed that BATSP1 antagonized the effect of FOXO1 and decreased oxygen consumption in both brown and white adipocytes, as indicated by measurements of the basal OCR, ATP production, proton leakage and maximal respiration (Fig. 6D, E and Fig. S6G). Thus, BATSP1-FOXO1 signaling is implicated in regulating thermogenesis in adipocytes through control of the subcellular localization of FOXO1 and restoration of Ucp1 expression.

Fig. 6

BATSP1 releases the transcriptional inhibition of UCP1 by FOXO1. A Reporter assays in HEK293T cells infected with adenoviruses expressing FOXO1 or mock viruses and treated with BATSP1 or vehicle control. B, C RT‒qPCR and Western blot analyses of UCP1 expression in vitro. Brown and white adipocytes were transduced with control or FOXO1-encoding adenoviral vectors, and differentiation was then initiated prior to stimulation with BATSP1 or treatment with vehicle control for 6 h. D, E Continuous measurement of oxygen consumption in adipocytes isolated from BAT and treated as described in Panels B and C. The data are presented as the means ± SDs; **p < 0.01; #p < 0.05; ##p < 0.01 by unpaired Student’s t test

Previous studies indicate that FOXO-binding proteins, including Zfp38, 14–3-3 Z and FCoR [17,18,19], are implicated in FOXO1-mediated transcriptional repression. We thus measured the expression levels of these potential targets in brown and white adipocytes upon BATSP1 treatment. First, we measured the transcript levels of these genes in adipocytes by RT‒qPCR analysis and found that they were not influenced by BATSP1 (Fig. S7A), suggesting that the regulation of BATSP1 does not occur at the transcriptional level. Second, we compared the protein levels of these genes in adipocytes treated with BATSP1. However, the protein expression levels of ZFP238 and 14-3-3 Z were comparable in BATSP1-treated and control adipocytes (Fig. S7B). Regrettably, no antibodies against FCoR are commercially available. Based on these findings, we conclude that these genes may not be involved in FOXO1-mediated repression of UCP1 transcription after BATSP1 stimulation. The target of BATSP1 that regulates the thermogenic program in cooperation with Foxo1 remains to be further clarified.