Celastrol targets CKB-mediated futile creatine cycle in human brown adipocytes thermogenesis

Abstract

Celastrol is widely recognized as one of the most potent anti-obesity agents, and its ability to promote adipocyte thermogenesis is thought to be a key mechanism. However, the precise molecular targets through which celastrol modulates thermogenesis in human adipocytes remain unclear. In this study, we synthesized a celastrol-based small molecular probe and employed a combination of photoaffinity labeling (PAL), click chemistry, and Surface Plasmon Resonance (SPR) to identify its direct binding targets. Our results reveal that celastrol directly interacts with creatine kinase B-type (CKB), leading to an increase in CKB protein stability. This suggests that celastrol modulates the futile creatine cycle within human brown adipocytes, thereby contributing to thermogenesis. Collectively, our findings provide new insights into the molecular mechanisms by which celastrol promotes thermogenesis in human brown adipocytes. Notably, we demonstrated that celastrol targets CKB-mediated futile creatine cycle for the first time. These findings not only deepen our understanding of celastrol's role in weight loss but also provides a potential strategy for obesity treatment.

1. Introduction

Obesity and its associated complications, such as type 2 diabetes (T2DM), cardiovascular diseases and non-alcoholic fatty liver (NAFLD), has become a world-wide health challenge [[1], [2], [3], [4]]. Excess energy intake over energy expenditure leads to a persistent energy imbalance and excessive accumulation of adipose tissue, which eventually leads to obesity [5]. The well-known characteristic of brown adipose tissue (BAT) to dissipate energy makes it a promising therapeutic target for the treatment of obesity and related metabolic diseases [6,7].

Celastrol is a pentacyclic triterpenoid derived from the traditional Chinese herb Tripterygium wilfordii, which has been known for its antioxidative and anti-inflammatory properties [8], and reported to exert a spectrum of other pharmacological properties including anti-tumor [9], myocardial remodeling [10] and neuroprotective [11] effects. Of note, a diversity of discoveries demonstrated that celastrol suppresses food intake, improves energy expenditure, and induces up to 45 % weight loss in obese mice [12,13]. In 2015, Celastrol was first identified as a potent leptin sensitizer that alleviates obesity and reduces food intake in mice by inhibiting ER stress and increasing STAT3-dependent leptin signaling [14]. Additionally, celastrol protects against obesity and metabolic dysfunction via increasing energy expenditure and mitochondrial function in adipose and muscle through activation of a HSF1-PGC1α transcriptional axis [15]. Despite these promising therapeutic effects, the medical application of celastrol is limited by its off-target cytotoxicity [16]. Several drug delivery methods have been explored to reduce the toxicity of celastrol [17,18]. While these findings highlight celastrol's potential to promote adipose thermogenesis, the precise molecular targets through which it modulates adipocyte metabolism remains unclear.

Therefore, we employed the photoaffinity labeling (PAL) approach, a powerful technique for systematically detecting small molecule-protein interactions [19]. By inducing covalent bond between small molecules and noncovalent binding proteins, PAL stabilizes the reversible interaction, allowing for a more comprehensive identification of small molecule binding proteins. Besides, celastrol is known to combine with protein thiols and generate a reversible and covalent connection with some proteins [20,21]. In this study, we applied this approach to uncover the specific targets of celastrol in human brown adipocytes.

2. Materials and methods

2.1. Synthesis of cel-p probe

Celastrol probe was synthesized by ChomiX Biotech Co.Ltd (Nanjing, China). In detail, Celastrol (112.7 mg, 0.25 mmoL, 1.0 equiv) and 2-(3-(But-3-yn-1-yl)-3H-diazirin-3-yl) ethan-1-amine (102.9 mg, 0.75 mmoL, 3.0 equiv) were dissolved in THF (3 mL) and DMF (1.5 mL). Then Dicyclohexylcarbodiimide (DCC, 63.5 mg, 0.3075 mmoL, 1.25 equiv), 1-Hydroxybenzotriazole (HOBt, 48.2 mg, 0.3567 mmoL, 1.45 equiv) were added, and the reaction mixture was stirred overnight at 25 °C. After completion of the reaction, the solvent was removed under reduced pressure, and the residue was purified by flash silica gel column chromatography to obtain the Cel-p probe.

2.2. Human primary adipocytes differentiation

Human primary brown adipocytes were isolated and differentiated as previously reported [22]. Briefly, human primary cells were cultured in preadipocyte medium (PAM, ScienCell) in 5 % CO₂ at 37 °C. For differentiation of human brown adipocytes, confluent cells were induced with DMEM/F12 medium supplemented with 430 nM insulin, 1 μM rosiglitazone, 0.5 mM 3-isobutyl-1-methylxanthine, 1 μM dexamethasone, 10 μg/mL apo-transferrin and 1 nM T3. The above medium was changed 2 days later with maintenance medium containing only 100 nM insulin and 1 nM T3 until successful induction of brown adipocytes.

2.3. Cytotoxicity

The cytotoxicity of Cel-p towards human brown adipocytes was determined by CCK8 assay (Vazyme). In detail, human brown primary adipocytes were plated in a 96-well plate (100 μL/well) and fully differentiated. Cells were then treated with different concentrations of Cel-p for 2 h and irradiated with a 365 nm light array (10 W, 1 J/cm²) for 15 min and further incubated for 24 h. 10 μLof CCK8 solution was then added to each well for additional 2 h. The absorbance at 450 nm of cells was detected by a microplate reader (Synergy H4, BioTek).

2.4. Fluorescence labeling in situ

The gel-based fluorescence experiment was performed as previously descripted [23]. Briefly, human brown adipocytes were cultured in a 6-well plate and treated with Cel-p (20 μM) for 2 h. The competition group was treated with excess Celastrol (400 μM) for 2 h before the co-incubation with Cel-p (20 μM) for another 2 h. The cells were exposed to a 365 nm UV light array (10 W, 1 J/cm²) for 15 min on ice, and lysed by RIPA buffer (Beyotime), quantified by the BCA Protein Assay Kit (Thermo Fisher Scientific). Then an equal amount of proteins (0.1 mg) was subjected to the click chemistry reaction. Each protein sample was added with 1 μL of 5-TAMRA-azide (10 mM stock in DMSO), 2 μL BTTAA/CuSO4 premix (25 mM/12.5 mM stock in DMSO), 2 μL sodium ascorbate (250 mM stock in ddH₂O). The samples were incubated at room temperature for 1 h with gentle shaking. Subsequently, the proteins were denatured and separated by SDS-PAGE electrophoresis. The fluorescence gel images were captured by ChemiDoc MP (Biorad) and Coomassie Brilliant Blue was used as a loading control stain.

2.5. Pull down and proteomics analysis

All protein analyses, identification, and quality control (QC) were performed using standard procedures by the ChomiX Biotech Co., Ltd. (Nanjing, China) as follows. The probe incubation and cell lysis were same as those of labeling experiments described above. For click reaction, each of 500 μL protein samples were added with 2 μL Biotin-PEG3-azide (50 mM stock in DMSO), 20 μL BTTAA/CuSO₄ premix (25 mM/12.5 mM stock in DMSO) and 20 μL sodium ascorbate (250 mM stock in ddH₂O). The samples were incubated for 1 h at room temperature with gentle shaking and proteins were precipitated by chloroform-methanol, then solubilized in 500 μL 0.2 % SDS/PBS. The Cel-p probe-labeled lysates were incubated with 100 μL of streptavidin-sepharose beads at room temperature with continuous rotation for 3 h. After washed with PBS containing 0.2 % SDS (w/v) three times, PBS three times and 100 mM triethylammonium bicarbonate (TEAB) three times, the beads were resuspended in 500 μL of 6M Urea/100 mM TEAB, then incubated with 25 μL of 200 mM DTT at 37 °C for 30 min and 25 μL of 400 mM iodoacetamide (IAA) for another 30 min. The beads were further resuspended in 100 μL 1M Urea/100 mM TEAB and were firstly digested with 1 μg LysC overnight and with 1 μg Trypsin for 6 h at 37 °C in the next day. The digested protein peptides were subjected to the reductive dimethylation. The DMSO-treated and probe-treated samples were added with 4 μL of 4 % HCHO or D¹³CDO, respectively, and reduced with 4 μL of 0.6M NaCNBH₃ for 1 h at room temperature. The reaction was quenched with 16 μL of 1 % ammonia solution and 8 μL of formic acid. Then the stable isotope labeled peptides were desalted and spin-dried for LC-MS/MS analysis.

LC-MS/MS was performed on the Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific) coupled with EASY-nLY 1200 UPLC system. Mobile phases consisted of 0.1 % (v/v) ammonium formate in water (A) and 0.1 % ammonium formate in acetonitrile (B). Flow rate was 3 μL/min for loading and 0.3 μL/min for eluting. LC gradient is as follows: 0 s-5s, 3 %–5 % B; 5s-53.6min, 5 %–15 % B; 53.6 min-100.25min, 15 %–28 % B; 100.25 min–116.9 min; 28 %–38 % B; 110.9 min–111 min 38 %–100 % B; 111 min–119 min 100 % B.

Under the positive-ion mode, full-scan mass spectra were acquired over the m/z range from 350 to 1800 using the Orbitrap mass analyzer with mass resolution of 60000. MS/MS fragmentation is performed in a data-dependent mode, of which dependent scan events (MS2) occurred at top20 are selected for MS/MS analysis a resolution of 15000 using collision mode of HCD. Isolation window is set as 1.6 m/z, normalized collision energy is set as 28 %, maximum IT is set as 50 ms and dynamic exclusion is set as 30 s.

2.6. Western blotting

The cells were harvested and extracted with RIPA Lysis buffer (Beyotime) including complete protease inhibitor (Roche). The lysates were centrifuged at 14,000 g for 10 min and denatured in 5x SDS loading buffer at 95 °C for 10 min. Equal amount of protein samples were separated on SDS-PAGE gel, followed by electro-transferring to PVDF membranes and blocking with 5 % milk. Proteins blots were incubated with the corresponding primary antibodies (Supplement Table 1) at 4 °C overnight. After washing 3 times with TBST, blots were incubated with secondary antibody at room temperature for 1 h. Protein bands were visualized using ECL luminescence reagent (Tanon) by ChemiDoc MP (Biorad).

For the Pull-down Western blot analysis, the method of collecting target proteins after treatment is the same as the proteomics analysis labeling. The enriched proteins after streptavidin-sepharose beads binding were released by heating at 95 °C for 10 min and separated by SDS-PAGE electrophoresis for immunoblotting.

2.7. RNA preparation and quantitative RT-PCR

Total RNA was isolated with TRIzol reagent (Invitrogen) as described by the manufacturer. The concentration of RNA was evaluated using a NanoDrop 2000 system (Thermo Fisher Scientific). A total of 1 μg of RNA from each sample was reverse transcribed into cDNA using HisyGo RT Red SuperMix (Vazyme). The mRNA expression levels were determined using the 2^−△△Ct method. The results were normalized to peptidylprolyl isomerase A (PPIA) and are presented as the fold change of each gene. The primers used for RT-PCR were listed in Supplement Table 1.

2.8. Preparation of recombinant CKB protein and surface plasmon resonance (SPR)

Recombinant human creatine kinase B-type/CKB protein was obtained from Sino Biological (Cat: 506-CKB0020). CKB was diluted to 50 μg/mL and covalently immobilized on CM5 sensor chips via amine coupling (Biacore AB, China). Immobilization of soluble celastrol generated resonance units (RU) of 600. Celastrol were serially diluted from 0 to 200 μM in PBS and injected at a 100 μL/min flow rate for 90 s at 25 °C. The dissociation phase was monitored for 600 s by allowing buffer to flow over the chip. Kinetic parameters were calculated using Biacore X100 Control software (General Electric, USA).

2.9. Molecular docking

The structure of Cel was downloaded from PubChem (CID:122724). The protein 3D structure of CKB was downloaded from PDB (3b6R). The docking simulation was performed by MOE (2019.0102) [24]. The docking results were selected as the potential binding mode.

2.10. Lentiviral transfection

The lentivirus-mediated CKB-specific shRNA (shCKB) was synthesized by Shanghai Genechem Co., Ltd. The shRNA sequence targeting human CKB was CCCTGCTGCTTCCTAACTTAT. The lentivirus containing shCKB or vectors was transfected into human brown adipocytes according to the manufacturer's instruction before differentiation. Then, cells were differentiated, treated with celastrol (0.5 μM) and analyzed in assays of OCR, Western blotting and qRT-PCR.

2.11. Oxygen consumption measurement

Human brown preadipocytes were cultured in XF-24 cell culture plate and differentiated for 4 days. Then cells were treated with different concentrations of celastrol for 3 days. After the treatment, the medium was placed with XF base medium (Agilent Technologies) supplemented with 7.5 mM glucose, 0.5 mM sodium pyruvate and 2.5 mM L-glutamine (pH 7.4). The basal respiration, respiration uncoupled from ATP synthesis and maximal respiration of cells were detected following addition of oligomycin (1 μM), FCCP (1 μM) and Rotenone/AntimycinA (0.5 μM each). Then, Seahorse XF24 wave software was used to calculate the oxygen consumption rate (OCR).

2.12. Statistical analysis

All data in this paper were showed as mean ± (SEM). Statistical differences analysis was performed using one-way ANOVA for multiple groups or unpaired two-tailed t-test for two groups. P < 0.05 was considered statistically significant. Statistical analysis was performed using GraphPad Prism 8.0 software.

3. Result

3.1. Fluorescence labelling of cel targets in human brown adipocytes

To identify the potential target proteins of Celastrol (Cel), we synthesized an alkynylated PAL-celastrol probe on the basis of the reported structure [23]. Considering the unstable connection between celastrol and its possible target proteins, we introduced a diazirine-containing photoaffinity clickable tag to the C29-carboxyl group (named Cel-p), thereby enhancing the stability of celastrol probe binding to its target protein (Fig. 1A). Then we evaluated the cytotoxicity of Cel-p towards human brown adipocytes by CCK8 assay. Cel-p compound exhibited comparable effect towards cell viability as celastrol in human brown adipocytes and exhibited no significant cytotoxicity in human brown adipocytes between 0 and 20 μM, indicating that the Cel-p compound retained the original biological activity of the probe. However, at higher concentrations, the Cel-p compound demonstrated reduced cell inhibitory effects compared to celastrol (Fig. 1B). Then, Cel-p was used to label the proteins in situ by gel fluorescence experiments. The mature human brown adipocytes were incubated with various concentrations of Cel-p. After irradiation with a 365 nm UV on ice, cell lysates were subjected to the copper catalyzed alkyne–azide cycloaddition reaction (CuAAC) with 5-TAMRA–azide. The in-gel fluorescence scanning by Western blot showed concentration-dependent labeling of multiple protein bands, indicating a significant labeling with 20 μM Cel-p (Fig. 1C). In addition, pre-incubation of the excess Celastrol markedly attenuated the labeling intensity of Cel-p in cells, suggesting that Cel-p can specifically bind the same target proteins as Celastrol does (Fig. 1D). The target-binding specificity of the Cel-P rasies the possibility of its further usage to identify the direct targets of Celastrol in thermogenic adipocytes.

Fig. 1.

Celastrol-based chemical probes for target identification. (A) Chemical structures of celastrol and Cel-p; (B) Cell proliferation inhibitory effect of different concentrations of Cel-p and celastrol towards human brown adipocytes after 2 h treatment, ∗∗p < 0.01, ∗∗∗p < 0.001 as compared to the Cel group; (C) In-gel fluorescence scan of proteins labeled by different concentrations of Cel-p probe in human brown adipocytes; (D) In-gel fluorescence scan of proteins labeled by 20 μM of Cel-p probe after pretreated with different-fold of celastrol for 2 h in human brown adipocytes; (E) Overall workflow for ABPP profiling of potential celastrol targets. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

3.2. Analysis of Cel-binding proteins based on LC-MS/MS

To profile the binding proteins of Cel-p in human brown adipocytes, we performed pull-down experiments and the Cel-p labeled proteins were eluted and identified by LC-MS/MS. The results revealed that Cel-p could directly bind with a total of 139 labeled proteins as compared with the DMSO group based on stringent screening criteria (Fig. 2A). The competition group (Cel-p versus Cel-p + Cel) represented proteins significantly competed by the original Celastrol compound, and 235 targets were identified (Fig. 2A). Venn analysis showed that 51 target proteins could be markedly labeled by Cel-p and competed by the Celastrol compound (Fig. 2B–Supplement Table 1). In order to comprehensively understand the function and biological processes of these proteins, we used Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) to preform bioinformatics analysis. From the significantly enriched biological processes, we found that the proteins were related to cytoplasmic translation, ribosome biogenesis, amino acid activation (Fig. 2C). Besides, KEGG pathway analysis showed that these targets were involved in various biological processes, such as pyruvate metabolism, AMPK signaling pathway and so on (Fig. 2D).

Fig. 2.

Analysis of Cel-binding proteins based on LC-MS/MS. (A) Target profiles of celastrol in human brown adipocytes after Cel-p and Celatrrol treatment as shown by spot chart analysis; See supplement table S1 for full list of identidied proteins; (B) Venn diagrams showing the overlap of differential target proteins identifed by Cel-p based LC-MS/MS; (C) GO enrichment analysis and (D) KEGG functional pathway analysis for the identified celastrol target proteins. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

3.3. Cel directly targets to CKB

We then performed pull-down assays to verify the direct interactions between celastrol and four potential proteins that has been reported to be involved in thermogenesis, including Glutathione S-transferase P (GSPT1), Creatine kinase B-type (CKB), Creatine kinase U-type mitochondrial (CKMT1A), Serine/threonine-protein phosphatase 2A (PP2A) [[25], [26], [27], [28], [29]]. Pull down-WB assays demonstrated that Cel-p directly and specifically bind with these four potential targets. In addition, co-incubation with excess celastrol could compete with the labeling of Cel-p in live cells as well as in cell lysate (Fig. 3A). We also synthesized another celastrol-based probe (named as Cel-a) based on previous studies [30,31]. Pull down assays also demonstrated that Cel-a could directly and specifically bind with potential targets identified by Cel-P (Fig. 3A). Among these potential protein targets, CKB significantly caught our attention. Uncoupling protein 1 (UCP1) is regarded as the canonical means of heat production, which is required for proton leak in adipocytes [32,33]. However, recent reports showed the creatine-driven substrate cycle is one of the key UCP1-independent thermogenic mechanisms, while CKB is indispensable for thermogenesis resulting from the futile creatine cycle (FCC) in both mouse and human adipocytes and promote thermogenesis [29]. To further validate the binding specificity between celastrol and CKB, we performed a surface plasmon resonance (SPR) assay. We observed that celastrol interacted with rhCKB (recombinant human CKB) protein with a KD (the equilibrium dissociation constant) of 21.41 μmol/L (Fig. 3B). To further characterize the celastrol-CKB interaction, we conducted a molecular docking and simulation study using the 3D crystal structure of CKB and celastrol. Molecular docking simulation identified Arginine 236 and 341 as potential binding sites for Cel on CKB proteins (Fig. 3C). These results together demonstrated that celastrol can directly target to CKB protein in human brown adipocytes.

Fig. 3.

Cel directly targets to CKB. (A) Validation of the representative celastrol binding proteins(CKB, CKMT1A, PP2A, GSTP1) using pull-down and western blotting assay after living cells or cell lysates were treated with cel-p/cel-a and celastrol; (B) The measurement of binding affinity of Cel with CKB using the SPR assay; (C) Docking simulation of Cel binding to CKB.

3.4. Cel increases CKB protein stability

Previous studies have established that CKB protein abundance is increased in BAT and purified brown adipocytes following cold exposure [29]. Similarly, we found that human brown adipocytes presented increased CKB expression in response to differentiation in vitro (Fig. 4A). Moreover, celastrol treatment during brown adipogenesis upregulated the CKB protein expression (Fig. 4B). Surprisingly, the mRNA level of CKB was not affected by celastrol treatment (Fig. 4C). To further explore whether increased CKB protein expression was caused by increased protein synthesis or inhibited protein degradation, cycloheximide (Chx), a translation inhibitor was used 48 h after celastrol treatment. As shown in Fig. 4D, CKB degradation decreased in human brown adipocytes with celastrol treatment, suggesting that celastrol affects the stability of CKB. In addition, celastrol treatment further increased the protein level of CKB in the presence of MG132 (proteasome inhibitor) (Fig. 4E). These above data suggested that celastrol can increase the CKB protein stability in human brown adipocytes.

Fig. 4.

Cel increases CKB protein stability. (A) Western blotting assay of UCP1 and CKB expressions during the differentiation of human brown adipocytes; (B) Protein level and (C) mRNA level of CKB after treatment with Cel in different concentrations in differentiated human brown adipocytes; (D) Protein translation inhibitor cycloheximide (CHX, 20 μg/mL) was added after human brown adipocytes were treated with or without Cel (0.5 μM, 3d). Protein levels of CKB at indicated time (0, 6, 12 and 24 h after chx treatment) were detected by western blotting, ∗p < 0.05 as compared to the control group; (E) Proteasome inhibitor MG132 (25 μM) was added after human brown adipocytes were treated with or without Cel (0.5 μM, 3d). Protein levels of CKB were detected by western blotting after 24 h treatment with MG132.

3.5. CKB is responsible for cel-mediated thermogenesis

Next, we questioned whether CKB mediated Cel-induced activation of thermogenesis in our human brown adipocytes model. The knockdown of CKB was achieved by lentivirus-mediated CKB-specific shRNA (shCKB) transfection to human brown pre-adipocytes (Fig. 5A and B). We found by seahorse mitochondrial stress test that celastrol treatment led to a significant increase in ATP production but a substantial decrease in proton leakage, suggesting that celastrol promotes the mitochondrial OCR through UCP1-independent pathway. In addition, the elevation of ATP-production stimulated by celastrol was attenuated in CKB-knockdown adipocytes (Fig. 5C and D). Taken together, these data strongly suggest that celastrol induced thermogenesis in human brown adipocytes was mediated via CKB.

Fig. 5.

CKB is responsible for Cel-mediated thermogenesis. (A, B) RT‒qPCR and Western blot analyses of CKB expression after human brown adipocytes were transfected with control or shRNA against CKB; (C) Basal OCR, ATP production and proton leak in cel-treated human brown adipocytes (0.5 μM for 3 d), ∗∗p < 0.01, ∗∗∗p < 0.001 as compared to the control group; (D) Basal OCR, ATP production and proton leak in cel-treated shGFP or shCKB human brown adipocytes (0.5 μM for 3 d), ∗p < 0.05, ∗∗∗p < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

4. Discussion

Celastrol is a promising natural compound with therapeutic potential in combating obesity [34]. It has been shown to act as a leptin sensitizer, suppressing food intake and promoting energy expenditure, yielding a 45 % body weight loss in diet-induced obese mice [10,14,15,35,36]. However, the direct targets of celastol have not been fully elucidated. In this study, we identified that celastrol could directly bind to CKB protein, an effector of UCP1-independent thermogenesis. Our results showed celastrol targets CKB to reduce its degradation and increase the mitochondria ATP production, thereby improving energy metabolism in human brown adipocytes. We uncover a novel mechanism by which celastrol promotes adipocyte thermogenesis through the UCP1-independent futile creatine cycle. Previous studies have linked celastrol's weight loss effects to UCP1-dependent mechanisms, including the seminal work by Ma et al. [15], which highlighted the role of the HSF1-PGC1α transcriptional axis in mediating celastrol's effects. Their work highlighted HSF1 as a key target of celastrol, showing that its activation leads to the upregulation of PGC1α, a master regulator of mitochondrial biogenesis and oxidative metabolism, and our study reveals a distinct yet complementary mechanism that celastrol directly binds to CKB, increasing CKB protein stability and enhancing the futile creatine cycle in human brown adipocytes. We speculate that these mechanisms may operate synergistically, with the HSF1-PGC1α axis supporting long-term mitochondrial adaptation and the CKB-mediated futile creatine cycle providing rapid energy expenditure.

Brown adipocytes dissipates energy as heat through uncoupling protein 1 (UCP1) by increasing the permeability of the inner mitochondrial membrane and dissipating the proton gradient [37]. However, recent research has revealed several UCP1-independent pathways that participate in the regulation of thermogenesis, including creatine-dependent substrate cycling, calcium-dependent ATP hydrolysis and lipid cycling [32,33,38]. Exciting evidence demonstrates that CKB runs parrel to UCP1 to promote cold-induced energy dissipation in adipocytes [39]. CKB is indispensable in the creatine futile cycling, and adipose tissue-specific knockout of Ckb can lead to the loss of most creatine kinase activity in BAT [29]. Besides, it is suggested that almost 30 %–40 % of thermogenesis in acutely isolated brown adipocytes is ATP-linked [40,41]. In this study, we found that celastrol treatment did not affect the protein expression of UCP1, regardless of whether CKB was knocked down, further confirming that celastrol could directly target CKB to mediate creatine futile cycling. Based on these findings, we believe we may have identified the first natural small molecule targeting CKB in human brown adipocytes, which has not been reported to date. Identifying novel compounds for activating CKB in thermogenic adipocytes is of great significance. While the clinical application of celastrol is limited by challenges such as poor bioavailability and potential side effects, it remains a promising lead for obesity treatment. Further investigations into the binding interaction between celastrol and CKB could provide a foundation for developing celastrol or its derivatives as anti-obesity drugs.

To identify target proteins and clarify the underlying mechanisms of celastrol in human brown adipocytes, we synthesized the Cel-P probe by introducing a bis-acridine photo-crosslinking group and an alkynyl group at the C29-carboxylic position of celastrol. These modifications do not affect the original activity of celastrol molecule due to their small modification structure. Upon incubation with cells and UV irradiation, the photo-crosslinking group generates a radical that covalently binds to neighboring proteins. Then, following copper-catalyzed alkyne-azide click chemistry reaction (CuAAC) and LC/MS proteomics we could profile and verify the direct targets of celastrol. In this study, we synthesized another celastrol-based probe (named as Cel-a) building on previous studies [30,31]. Pull down assays also demonstrated that Cel-a could directly and specifically bind with potential targets identified by Cel-P, validating the reliability of our findings. Previous study confirmed the interaction between celastrol and glutathione-S- ransferases M1(GSTM1) in vitro through affinity purification assays and thermal shift assay, and successfully obtained a co-crystal structure of celastrol with mouse glutathione-S- transferases P1(GSTP1) [42]. In our study, we also found that celastrol (Cel-p) can directly bind to GSTP1, which further suggested the reliability of the Cel-p based proteomics results. However, no significant changes in the expression of GSTP1 was observed during the differentiation process of human brown adipocytes, nor did celastrol treatment affect its expression (data not shown). Given that GSTP1 is a high-abundance protein, its binding of celastrol may have limited biological significance. These findings suggest that celastrol likely does not regulate thermogenesis in human brown adipocytes by targeting GSTP1.

Previous studies revealed that control of CKB abundance appears to be downstream of cAMP-mediated thermogenic signaling, e.g., cold exposure and β-adrenergic signaling upregulate CKB mRNA and protein expression (2 ∼ 3-fold) in BAT [29]. In this study, we found that the CKB protein is increased by approximately 1.5-fold by celastrol, since celastrol treatment led to the increased stability of CKB protein. Although the expression level of CKB protein induced by celastrol was slightly lower than that elicited by cAMP-mediated thermogenic signaling, this finding suggests a potential strategy of targeting protein stability to enhance CKB abundance. While our data suggest that celastrol treatment led to the increased stability of CKB protein, the underlying mechanisms responsible for the enhanced stability of CKB protein require further elucidation. A limitation of this study is the lack of comparative analysis of Cel-p's biological activity in human brown adipocytes relative to native Celastrol, as well as the in vivo validation for the potential functional interaction between celastrol and CKB. Currently, there is a portion of the population in China receiving traditional celastrol herb treatment. In the future studies, we hope to further validate our findings based on these group of individuals.

Despite the limitations mentioned above, our data reveal a novel mechanism by which celastrol promotes adipocyte thermogenesis through the UCP1-independent futile creatine cycle in human brown adipocyte. We demonstrate that the natural molecule celastrol can directly target CKB, leading to enhanced mitochondrial ATP production. These findings highlight the therapeutic potential of targeting the Celastrol/CKB interaction to improve brown adipose thermogenesis and combat against obesity and associated metabolic diseases.

CRediT authorship contribution statement

Jingyi Ni: Writing – review & editing, Validation, Methodology, Investigation, Formal analysis, Data curation. Baicheng Wang: Validation, Methodology, Investigation, Formal analysis, Data curation. Xinyue Liu: Validation, Methodology, Investigation, Formal analysis, Data curation. Rui Yin: Validation, Methodology, Investigation. Jinlin Tang: Validation, Methodology, Investigation. Siyu Hua: Validation, Methodology, Investigation. Xiaoxiao Zhang: Validation, Methodology, Investigation. Yangyang Wu: Writing – review & editing, Visualization, Validation, Investigation, Funding acquisition, Formal analysis, Conceptualization. Shihu Zhang: Writing – original draft, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Chenbo Ji: Writing – original draft, Validation, Supervision, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Data availability

The authors confirmed that all data underlying this article are available to the corresponding authors for reasonable request.

Declaration of competing interest

The authors declare no competing interest.

Appendix A. Supplementary data

The following is the Supplementary data to this article.