Gut microbiota drives the metabolic dysregulation in obesity-prone individuals by impairing GDCA-mediated activation of brown adipose thermogenesis and ileal GLP-1 secretion

2.1. Human participants

Our research group has previously established a body constitution scale and the National Standards of Classification and Determination of Constitution in Chinese Medicine (ZYYXH/T157-2009), issued by the China Association of Chinese Medicine, to identify individuals of obesity-prone (OP), which is termed the phlegm-dampness (PD) body constitution in traditional Chinese medicine^27,28. Detailed methodology for PD score calculation and the criteria used to identify OP subjects were provided in Supporting Information Table S1. Screened by the PD score, 46 OP subjects with no metabolic diseases were recruited in this study, and 35 healthy people were set as the healthy control (HC) group. Informed consent was obtained from all participants. This study was approved by the Beijing University of Chinese Medicine ethics committee (2022BZYLL0607) and performed following the Declaration of Helsinki. All participants were healthy individuals aged between 18 and 60 years with no diagnosed diseases, particularly metabolic disorders such as obesity, diabetes, metabolic syndrome, etc. Both groups maintained a regular diet and had not undergone antibiotic treatment in the preceding three months. Blood and fecal specimens were collected for subsequent assays.

2.2. Bacterial 16S rDNA gene sequencing

Fecal samples were procured immediately post-defecation and preserved at −80 °C. Bacterial DNA was extracted utilizing a DNeasy PowerSoil kit (Qiagen, Hilden, Germany) according to the manufacturer's protocols. The concentration and integrity of the DNA were confirmed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and agarose gel electrophoresis, respectively. The extracted DNA was diluted to a concentration of 1 ng/μL and stored at −20 °C until further processing. PCR (Polymerase Chain Reaction) amplification of the V3–V4 hypervariable regions of the bacterial 16S rRNA gene was performed in a 25 μL reaction using universal primer pairs (343 F: 5′-TACGGRAGGCAGCAG-3′; 798 R: 5′-AGGGTATCTAATCCT-3′). Amplicons were purified (Beckman Coulter Co., Brea, CA, USA) and quantified using the Qubit dsDNA assay kit (Thermo Fisher Scientific, Waltham, MA, USA), pooled at equimolar concentrations, and sequenced on an Illumina MiSeq platform (Illumina Inc., San Diego, CA, USA) by OE Biotech Company (Shanghai, China) following standard protocols. Paired-end reads were then preprocessed using Trimmomatic software, 19 to identify and trim ambiguous bases. After trimming, paired-end reads were assembled using FLASH software, 20. Clean reads were subjected to primer sequence removal and clustering to generate operational taxonomic units (OTUs) using Vsearch software with a 97% similarity cutoff.

2.3. Metabolomics analysis

After sample preparation and derivatization, BA profiles in fecal and plasma samples from OP and HC subjects were qualitatively and quantitatively analyzed using UPLC–ESI–MS/MS. The chromatographic separations were achieved on a Phenomenex Kinetex C18 column (2.1 mm × 100 mm, 2.6 μm; Phenomenex, Torrance, CA, USA). The detection parameters of the mass spectrometer were as follows. Curtain Gas: 35 psi; Collision-activated ionization (CAD) parameters: medium; Negative ion spray voltage: −4.5 kV; Ion source temperature, 450 °C, constant flow rate of 0.45 mL/min, and the column temperature was kept at 45 °C. Ion source: Gas1: 55 (psi); Gas2: 55 (psi). The metabolites were quantified by the Triple Quadrupole Mass Spectrometer's multiple reaction monitoring (MRM) detection mode.

2.4. Animal studies

All experimental protocols were approved by the Ethics Committee of the Beijing University of Chinese Medicine (BUCM-4-2021121001-4123). Male C57BL/6J mice aged six weeks were purchased from SPF (Beijing) Biotechnology Co., Ltd. (Beijing, China) and maintained under specific pathogen-free (SPF) conditions with free access to chow and water (temperature 20–22 °C, humidity 45 ± 5%, 12 h light/dark cycle). Food intake was monitored daily and used to calculate the energy intake of each mouse. The body weight was recorded weekly.

2.5. Antibiotic treatment and FMT

After one week of acclimatization, all mice received an antibiotic cocktail (ampicillin 1 mg/mL, neomycin 1 mg/mL, metronidazole 0.25 mg/mL, and vancomycin 0.5 mg/mL) treatment in drinking water for seven days to deplete the gut microbiota²⁷. For FMT, fecal samples from 12 subjects were randomly picked from each OP and HC group and then pooled within each group. 200 mg of feces were resuspended in 1.5 mL PBS and homogenized, then centrifuged at 800 rpm (Eppendorf 5424R centrifuge, Eppendorf, Hamburg, Germany) for 5 min, taking the supernatant as inoculum for transplantation. To minimize the transformations of microbial composition, the inoculum was prepared daily and then immediately administered by oral gavage to antibiotic-treated mice in a volume of 300 μL/day. During the FMT, all mice were fed a HFD (D12492, consisting of 60% fat, 20% carbohydrate, and 20% protein) purchased from SPF (Beijing) Biotechnology Co., Ltd. (Beijing, China).

2.6. GDCA administration

C57BL/6J mice were maintained under SPF conditions. An HFD-induced obesity model was employed. The treatment group received a daily oral gavage of GDCA dissolved in saline at 30 mg/kg/day²⁹, while the validation group was concurrently administered the same dose of GDCA and the TGR5 inhibitor SBI-115 (Bide Pharmatech Ltd., Shanghai, China) at a dose of 10 mg/kg/day³⁰. In contrast, the control group was administered an equivalent volume of saline and fed a standard chow diet for four weeks. Prior to sacrifice, all mice were subjected to a fasting period and underwent both an oral glucose tolerance test (OGTT) and an insulin tolerance test (ITT). The OGTT was conducted by administering 1.0 g/kg of glucose orally, while the ITT was performed through the intraperitoneal injection of 0.75 U/kg insulin, and then blood glucose was tested subsequently at specific time points.

2.7. Cell experiments

C3H10T1/2 mesenchymal stem cells were maintained in MEM medium supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin until reaching optimal confluence. For differentiation into brown adipocytes, C3H10T1/2 cells were treated for 2 days with induction medium containing 0.005 mmol/L dexamethasone (Bio-Rad Laboratories, Hercules, CA, USA), 0.05 μmol/L 3,3′,5-triiodo-l-thyronine (Yuanye Bio-Technology Co., Ltd., Shanghai, China; T3), 0.5 mmol/L 3-isobutyl-1-methylxanthine (Sigma–Aldrich, Darmstadt, Germany), 0.125 mmol/L indomethacin (Yuanye Bio-Technology Co., Ltd., Shanghai, China), 1 μmol/L rosiglitazone (Yuanye Bio-Technology Co., Ltd., Shanghai, China), and 0.005 μg/μL insulin (Solarbio Science & Technology Co., Ltd., Beijing, China). Subsequently, cells were transferred to maintenance medium containing 0.05 μmol/L T3, 1 μmol/L rosiglitazone, and 0.005 μg/μL insulin for 4 days, with medium changed every 48 h, resulting in fully differentiated brown adipocytes by Day 6.

STC-1 enteroendocrine cells were cultured in high-glucose DMEM supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. To investigate the cellular effects of GDCA, differentiated brown adipocytes and STC-1 cells were assigned to various treatment groups. Control groups were treated with DMSO alone. Brown adipocytes were exposed to GDCA at concentrations of 100 or 200 μmol/L, while STC-1 cells were treated with 80 or 160 μmol/L GDCA. Additional inhibitor groups received the same GDCA treatments along with 20 μmol/L SBI-115 for brown adipocytes or 80 μmol/L SBI-115 for STC-1 cells. Following a 24 h treatment period, both cell types were harvested and processed into appropriate sample formats for subsequent analysis.

2.8. Luciferase assays

C3H10T1/2 derived brown adipocytes and STC-1 enteroendocrine cells were seeded in 12-well plates, allowed to adhere overnight, before transfection using Lipofectamine 2000 reagent (Invitrogen, Waltham, MA, USA). Specifically, a mixture of 500 ng TGR5 luciferase reporter plasmid and 10 ng pGL-TK plasmid (Xianghong Biotechnology Co., Ltd., Beijing, China) was prepared in Opti-MEM medium, incubated for 20 min at room temperature, added to cells for 6 h, and then replaced with fresh complete medium. After an additional 24 h incubation, cells were administered: control groups received DMSO vehicle, while test groups were exposed to either 80 μmol/L GDCA (for STC-1 cells) or 100 μmol/L GDCA (for C3H10T1/2 cells), with a selective TGR5 agonist INT-777 serving as the positive control for 24 h to assess TGR5 activation³¹. Finally, cells were lysed with 150 μL passive lysis buffer per well, and luciferase activity was quantified using a dual-luciferase reporter assay kit (Beyotime Biotechnology Co., Ltd., Shanghai, China) on a SpectraMax Mini microplate reader (Molecular Devices, San Jose, CA, USA), where firefly luciferase signals (representing TGR5 activity) were normalized to Renilla luciferase (from pGL-TK) to account for transfection efficiency.

2.9. Histopathology staining

Paraffin-embedded sections of epididymal fat were stained with hematoxylin and eosin (H&E). Fat sections were deparaffined and dehydrated with xylenes and ethanol. Slides were thoroughly washed and dipped in hematoxylin, washed again with water and ethanol 95%, and incubated with eosin for 30 min. After incubation, sections were re-incubated with ethanol and xylenes and covered with a mounting medium. Parts of the left and medial lobes of the mouse liver were sampled for Oil Red O staining. Prepared with 4% paraformaldehyde fixative solution, liver sections (4 μm thick) were cut with a cryostat and rinsed with 60% isopropanol, stained with Oil Red O staining solution for 10 min, then rinsed with 60% isopropanol. Finally, sections were counterstained with hematoxylin. Images were analyzed using an inverted microscope system (Olympus, Tokyo, Japan).

2.10. RT-qPCR assays

Total RNA was extracted from different tissues using an RNA extraction reagent (Servicebio, Wuhan, China). cDNA was synthesized from 10 μg of total RNA utilizing Servicebio RT Enzyme Mix (Servicebio, Wuhan, China). Real-time PCR was conducted with 2 μL of the cDNA template added to 7.5 μL of 2 × SYBR Green qPCR Master Mix (None ROX) (Servicebio, Wuhan, China) and specific primers (2.5 μmol/L each). The expression of the target genes was quantified relative to that of an internal control gene (ACTIN) based on the comparison of the threshold cycle (CT) at constant fluorescence intensity. Relative expression (R) was calculated using Eq. (1):

2.11. Statistical analysis

All data are represented as mean ± standard deviation (SD). The differences between two groups were compared using Student's t-test (unpaired, two-tailed). The differences among multiple groups were compared by using one-way analysis of variance (ANOVA) with Dunnett's post hoc test, while those tests were conducted only if F achieved P < 0.05 and there was no significant variance heterogeneity. The correlation was evaluated using Spearman's rank correlation coefficient. Data were analyzed using GraphPad Prism 9 software (GraphPad, San Diego, CA, USA). The statistical significance threshold was set at P < 0.05 and indicated as ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, respectively.

3.1. Clinical characteristics of the obesity-prone and healthy control subjects

The characteristics of clinical participants are presented in Table 1. Despite all the OP subjects being free of metabolic diseases and most of the clinical indicators falling within normal ranges according to established diagnostic criteria, the body weight, BMI, waist-to-hip ratio, and multiple glucolipid indicators, including fasting insulin, HOMA-IR, TyG index, QUICKI-IS, and triglycerides, etc. were significantly higher in the OP group (P < 0.01), implying the predisposition to metabolic derangements. Notably, OP subjects exhibited significantly elevated GLP-1 concentrations, which aligned with previous reports, suggesting a physiological compensatory response in the early stage of obesity or metabolic syndrome^32,33. Furthermore, correlation analysis revealed that the PD score was significantly correlated with multiple clinical indicators. Specifically, a positive correlation was found between the PD score and body weight, BMI, waist circumference, waist-to-hip ratio, body fat percentage, fasting insulin, HOMA-IR, TyG index, and triglycerides, etc., while a negative correlation was observed with QUICKI-IS and HDL cholesterol levels (Supporting Information Fig. S1), which also demonstrated the reliability of the PD score in identifying OP individuals.

3.2. The distinct gut microbiota compositions of obesity-prone individuals were significantly correlated with multiple metabolic parameters

Over the past few decades, mounting evidence has revealed the crucial role of gut microbiota perturbations in the development of obesity and metabolic diseases³⁴. To uncover the underlying pathological basis of the differential metabolic phenotype of the HC and OP subjects, we performed 16S rDNA sequencing to detect the gut microbiota composition of the two groups. No significant difference in bacterial alpha diversity between the HC and OP groups was observed (Fig. 1A–C). Notably, the orthogonal partial least squares discriminant analysis (OPLS-DA) demonstrated a distinct clustering pattern between subjects from the OP and HC groups (Fig. 1D). The degree of taxonomic abundance at the phylum and genus level was examined to evaluate the overall composition of the bacterial community. Compared to the HC, the OP group showed significantly different relative abundance at both the phylum and genus levels (Fig. 1E and F).

Moreover, the significantly elevated Firmicutes/Bacteroidetes ratio suggested a gut microbiota dysbiosis in OP individuals (Fig. 1G). The linear discriminant analysis (LDA) effect size (LEfSe) was further used to identify differentially abundant taxa between the two groups. Notably, Prevotella was found to be the signature gut microbiota genus in the OP group, which was reported to be closely correlated with metabolic diseases³⁵ (Fig. 1H and I). Correlation analysis, as shown in the heatmap, identified the significant association between differently abundant gut microbiota genera and various metabolic clinical indicators (Fig. 1J), which implies that gut microbiota might be germane to the physiopathology of OP.

3.3. Fecal microbiota transplantation validated the pivotal role of gut microbiota in the dysmetabolism susceptibility of obese-prone subjects

To further explore the role of gut microbiota in predisposing OP subjects to metabolic disturbances, we conducted a FMT experiment. An antibiotic cocktail treatment was applied to C57BL/6J mice to deplete most of the gut microbiota²⁷. Microbial inoculum from OP and HC fecal samples was administered orally to antibiotic-treated recipient mice under a HFD (Fig. 2A). The FMT-OP mice exhibited elevated body weights by the end of the first week, and this trend was consistently maintained (Fig. 2B), resulting in a significantly higher body weight gain than the FMT-HC group (Fig. 2C). Pathological staining also identified significantly enhanced liver fat accumulation and hypertrophic white adipocytes in FMT-OP mice (Fig. 2D–F). Serum total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) increased significantly in the FMT-OP group (Fig. 2G–J), suggesting that OP gut microbiota induced abnormality in lipid metabolism.

Furthermore, FMT-OP mice also exhibited remarkable dysglycemia. Glucose and insulin tolerance were markedly impaired in the FMT-OP mice, as shown by the OGTT and ITT, respectively (Fig. 2K and L). Although fasting serum insulin was similar in the FMT-OP and FMT-HC groups, the FMT-OP mice had significantly elevated fasting blood glucose and HOMA-IR index, with declined QUICKI-IS index levels (Fig. 2M–P). In a nutshell, these data suggested that the gut microbiota from OP individuals might lead to dysregulated glycolipid metabolism and insulin resistance.

3.4. Obesity-prone subjects exhibited disturbed bile acid profiles, with GDCA identified as the most potent discriminator between the two groups

As a key class of microbiota-derived metabolites, BAs play a critical role in the development of obesity and metabolic diseases. A UPLC–ESI–MS/MS approach was applied to quantitate both the fecal and plasma BA profiles of OP and HC groups. Multiple BAs exhibited significantly different concentrations. Specifically, in the fecal BAs, lithocholic acid (LCA), deoxycholic acid (DCA), GDCA, taurodeoxycholic acid (TDCA), taurochenodeoxycholic acid (TCDCA), taurocholic acid (TCA), hyodeoxycholic acid (HDCA), and 12-ketolithocholic acid (12-KLCA) were significantly decreased in the OP group, and most of the differential BAs belonged to the secondary and conjugated BA categories. Conversely, the primary BAs, including chenodeoxycholic acid (CDCA), ursodeoxycholic acid (UDCA), cholic acid (CA), 3β-cholic acid (3β-CA), ursocholic acid (UCA), and hyocholic acid (HCA) demonstrated increased concentrations in the OP group (Fig. 3A). In the plasma BAs, CDCA, UDCA, DCA, GDCA et al., also demonstrated significant difference between OP and HC subjects (Fig. 3A). Moreover, a distinct clustering was observed in the OPLS-DA score plot in both fecal and plasma BAs (Fig. 3B). OP subjects had markedly reduced GDCA concentrations in both fecal and plasma samples. Furthermore, GDCA was identified as the most potent BA in discriminating the OP and HC groups based on the variable importance in projection (VIP) scores (Fig. 3C). The receiver operating characteristic (ROC) curve displayed excellent discrimination of GDCA (AUC = 0.76, Fig. 3D), which further verified the results of the VIP score.

The BA pool of OP subjects had notably increased total primary BA concentrations in both fecal and plasma BAs (Fig. 3E and F), and the primary-to-secondary BA ratio also increased significantly in the OP group (Fig. 3G). The OP group exhibited elevated levels of total deconjugated BAs alongside a reduction in conjugated BAs, resulting in a corresponding increase in the deconjugated-to-conjugated BA ratio (Fig. 3H–J). Spearman correlation analyses were performed to elucidate the association between plasma BA levels and clinical parameters. As revealed in the heatmap, BAs exhibited significant correlations with multiple metabolic indicators, and GDCA had significantly negative correlations with body weight, BMI, and waist circumference, while it showed a notable positive correlation with GLP-1 levels (Fig. 3K).

3.5. Gut microbiota of obesity-prone individuals drove the bile acid dysregulation and decreased the GDCA concentration

The production of secondary BAs is regulated by the gut microbiota through the action of bile salt hydrolase (BSH) and 7α-dehydroxylase. As displayed in Spearman's correlation analysis, gut microbiota had a significant association with BAs (Fig. 4A). The gut microbiota can modulate the host metabolism via various signaling pathways. To further elucidate the functional prediction and pathway inferences of the differential intestinal flora of OP subjects, the Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) approach was employed. The results indicated that multiple metabolic pathways were enriched, including the AMPK, glucagon, and insulin resistance signaling, etc. Notably, secondary bile acid biosynthesis pathway was also a critical differential signal pathway of gut microbiota between the OP and HC individuals (Fig. 4B). In the intestine, gut bacteria BSH deconjugates conjugated bile acids, then bacterial 7α- and 7β-dehydroxylases convert the primary BAs to secondary BAs³⁶. We examined the bacterial taxa that conduct the key steps of secondary BA biosynthesis^37,38, and found that the gut bacteria that express BSH were increased in OP individuals (Fig. 4C), while the gut bacteria known to express 7α-dehydroxylases were significantly decreased (Fig. 4D).

To further verify the effects of gut microbiota on BA profile, FMT validation was conducted. Interestingly, the FMT-OP and -HC mice exhibited similar plasma BA compositions to the OP and HC subjects (Fig. 3B), with the FMT-OP group having significantly elevated absolute BA abundances (Fig. 4E). The OPLS-DA score plot also revealed a distinct clustering pattern between samples from the FMT-OP and FMT-HC groups (Fig. 4F). A variety of BAs displayed different levels between the two groups. Notably, GDCA was significantly lowered in the FMT-OP group (Fig. 4G), implying that the gut microbiota effectively phenocopied the dysregulated BA profiles of OP subjects. The ROC curve also confirmed the excellent predictive ability of GDCA for the distinction between FMT-OP and FMT-HC mice (AUC = 0.96, Fig. 4H). Furthermore, the BA composition analysis revealed that the FMT-OP group was characterized by a markedly elevated primary-to-secondary ratio (Fig. 4I), in line with that of the clinical OP subjects. The ratio of deconjugated to conjugated BAs did not change significantly between the two groups (Fig. 4J). However, the ratio of 12OH to non-12OH BAs demonstrated a notable increase in the FMT-OP group (Fig. 4K), suggesting the gut microbiota from OP subjects had largely shifted the BA biosynthesis towards the classical BA synthesis pathway in the liver. Therefore, we further examined the hepatic expression of genes related to BA synthesis (Fig. 4L–O). Of note, Cyp7a1, the rate-limiting enzyme in BA synthesis, demonstrated decreased gene expression in the livers of the FMT-OP group (Fig. 4L). Furthermore, down-regulated expression of Cyp7b1 (Fig. 4M) and up-regulated expression of Cyp8a1 were observed in the FMT-OP group (Fig. 4N), which confirmed that the classical pathway is the predominant BA synthesis route of FMT-OP mice.

3.6. GDCA ameliorated obesity linked metabolic dysfunction and remodeled the bile acid pool in a TGR5-dependent manner

Both OP subjects and FMT-OP mice exhibited the characteristic of deficient GDCA, and consequently, we proceeded to investigate the potential metabolic modulation effects of GDCA in a high-fat diet-induced obesity (DIO) mouse model. Given that the receptor TGR5 has been reported to play a pivotal role in mediating BAs’ metabolic regulation^31,39, we set a validation group by co-administering GDCA with a selective TGR5 antagonist SBI-115³⁰. Administration of GDCA significantly counteracted diet-induced obesity and related metabolic disturbances. Following a four-week intervention, GDCA-treated mice exhibited a marked reduction in body weight compared to obese controls (Fig. 5A and B). GDCA treatment substantially improved systemic glucose homeostasis. Fasting blood glucose levels were significantly lowered (Fig. 5C). Furthermore, GDCA-enhanced insulin sensitivity and glucose tolerance were evidenced by improved performances during both OGTT (Fig. 5D) andITT (Fig. 5E). Consistent with these metabolic improvements, circulating insulin levels were reduced (Fig. 5F), leading to a significant decrease in HOMA-IR index (Fig. 5N) and an increase in the QUICKI index (Fig. 5O).

The anti-obesity effects of GDCA were further reflected in altered body composition. Lee's index, a measure of obesity, was significantly lower in the treatment group (Fig. 5G). This was accompanied by a decreased adiposity index (Fig. 5M) and smaller mass of major fat depots, including epididymal white adipose, subcutaneous white adipose, and beige adipose tissue (Fig. 5I–K). Liver weight was also reduced (Fig. 5H), whereas brown adipose tissue mass increased substantially (Fig. 5L). Histological analyses provided direct morphological evidence of GDCA's efficacy. Oil Red O staining of liver sections revealed a dramatic reduction in lipid droplet accumulation, indicating alleviation of hepatic steatosis (Fig. 5P). Concordantly, histological examination of epididymal white adipose tissue via H&E staining showed a noticeable decrease in adipocyte size, signifying a reduction in fat storage (Fig. 5Q).

GDCA administration also induced a profound remodeling of circulating BA pools (Fig. 5R). Plasma GDCA levels were markedly elevated following the treatment, with GDCA identified as the potent discriminator to separate from the obesity model group (Fig. 5S). A stacked plot confirmed the shift in the compositional landscape of plasma BA profiles (Fig. 5T), accompanied by decreased primary-to-secondary BA (Fig. 5U) and 12OH/non-12OH BA ratios (Fig. 5W), along with an increased deconjugated-to-conjugated BA ratio (Fig. 5V). Correlation analysis demonstrated that the circulating GDCA concentrations were significantly associated with critical indicators of enhanced metabolic health, including reduced body weight, lowered fasting glucose, and improved insulin sensitivity (Fig. 5X). Moreover, GDCA levels correlated positively with Tgr5 gene expression in both BAT and ileum. They were also associated with genes involved in BAT thermogenesis and ileal GLP-1 secretion, as illustrated in Fig. 5X. These findings suggest a pivotal role for TGR5 in mediating the therapeutic effects of GDCA.

Notably, the metabolic benefits conferred by GDCA were abolished upon co-administration with the TGR5 inhibitor SBI-115, confirming that the observed effects are mediated specifically through the TGR5 signaling.

3.7. GDCA activated the TGR5 signaling, thereby enhancing brown adipose tissue thermogenesis and ileal GLP-1 secretion in mice with diet-induced obesity

Administration of GDCA significantly activated TGR5 signaling in BAT of diet-induced obesity mice, as demonstrated by elevated immunofluorescence staining and quantitative fluorescence intensity of TGR5 (Fig. 6A and B). Consistent with these observations, Tgr5 gene expression and cAMP concentration were markedly increased (Fig. 6C and D), confirming receptor activation. Consequently, key thermogenic markers including Ucp1, Pgc-1α, and Prdm16 were substantially upregulated (Fig. 6E–G), indicating enhanced thermogenic capacity of BAT. Further morphological analysis revealed that GDCA treatment promoted mitochondrial biogenesis and lipid utilization. Immunofluorescence staining of BAT sections revealed smaller lipid droplets and a more multilocular morphology, and pronounced reduction in lipid droplet (LD) number per cell (Fig. 6O and P). Accompanied by an increase in mitochondrial fluorescence intensity (Fig. 6Q), and a higher proportion of LD–mitochondria contacts, exhibiting a perilipin-encircled distribution and forming tight clusters (Fig. 6R). These structural changes enhance lipid mobilization, enable direct fatty acid transfer to mitochondria, and bypass cytosolic diffusion, thereby ensuring efficient substrate channeling for β-oxidation.

Parallel activation of TGR5 was observed in the ileum, with increased immunofluorescence signal and quantitative intensity of TGR5 (Fig. 6H and I), and accompanied by elevated Tgr5 expression and cAMP levels (Fig. 6J and K), confirming functional receptor engagement. Accordingly, genes critical for GLP-1 biosynthesis and secretion, e.g., Gcg, pcsk1, and Slc5a1 were significantly upregulated (Fig. 6L–N). This molecular response translated into enhanced GLP-1 production, as visualized by immunofluorescence staining of ileal tissues (Fig. 6S and T) and quantified via ELISA in both plasma and ileum samples (Fig. 6U and V).

Collectively, these findings indicate that GDCA elicits its metabolic effects via tissue-specific activation of TGR5, by enhancing thermogenesis in BAT and stimulating GLP-1 secretion from the ileum. The abrogation of all phenotypic and biochemical outcomes by the selective antagonist SBI-115 underscores the essential role of TGR5 signaling in mediating these responses. Further in vitro validation of these findings will be detailed in the following section.

3.8. In vitro models confirmed that GDCA augmented thermogenesis in brown adipocytes and GLP-1 secretion from enteroendocrine cells in a TGR5-dependent manner

Cell experiments were performed to verify the critical role of TGR5 in mediating GDCA's therapeutic effects. First, we assessed the impact of GDCA on thermogenesis using brown adipocytes differentiated from C3H10T1/2 mesenchymal stem cells. The dual luciferase assay demonstrated robust TGR5 activation by GDCA, using the selective agonist INT-777 as a positive control (Fig. 7A); this finding was supported by Tgr5 gene expression analysis (Fig. 7B). The cAMP signaling pathway is a critical pathway regulating the UCP-1-mediated heat production in brown adipose tissue⁴⁰. GDCA administration increased cAMP concentration dose-dependently in brown adipocytes, which could be reversed by SBI-115 (Fig. 7C), and a similar trend was observed in the TGR5 protein expression (Fig. 7G and H). GDCA improved the gene expression of Ucp-1 and Prdm-16, and this effect disappeared when exposed to the TGR5-specific inhibitor SBI-115 (Fig. 7D–F). In addition, we also determined the protein expression of UCP1, PGC1α, and PRDM16 via Western blot and observed similar trends to those of the mRNA results (Fig. 7G–K). In short, these findings provide compelling evidence that GDCA enhances energy expenditure in brown adipocytes via the TGR5–cAMP pathway. This elucidates the cellular mechanism by which GDCA combats obesity and improves metabolism.

We also examined how GDCA supplementation affects the secretion of GLP-1 using enteroendocrine STC-1 cells. GDCA activated the TGR5 receptor in STC-1 cells, as demonstrated by relative luciferase activity of TGR5 (Fig. 8A), up-regulated TGR5 mRNA and protein expressions (Fig. 8B–J and K), as well as increased cAMP levels (Fig. 8C). The GLP-1 production was promoted consequently by GDCA, as indicated by ELISA (Fig. 8G) and immunofluorescent staining (Fig. 8H and I). Subsequently, we further analyzed the expression of GLP-1 modulators. As expected, GDCA up-regulated the gene expression of Gcg, Pcsk1, and Slc5a1 (Fig. 8D–F). This result was confirmed by the elevated protein levels of Proglucagon (encoded by Gcg), SGLT1 (encoded by Slc5a1), and PC1/3 (encoded by Pcsk1) in GDCA groups (Fig. 8J, L–N). However, all the effects were abolished by the SBI-115 intervention, which substantiated the indispensable role of TGR5 in the bioactivity of GDCA.

Collectively, the data demonstrate that GDCA effectively improves thermogenesis in brown adipocytes and boostes GLP-1 secretion in enteroendocrine STC-1 cells through the activation of TGR5–cAMP signaling, thereby exerting potent effects on the improvement of glycolipid metabolism.