https://orcid.org/0000-0001-8019-3794
Abstract
Berberine (BBR), which is a compound derived from the Chinese medicinal plant Coptis chinensis, promotes weight loss, but the molecular mechanisms are not well understood. Here, we show that BBR increases the serum level of growth differentiation factor 15 (GDF15), which is a stress response cytokine that can reduce food intake and lower body weight in diet-induced obese (DIO) mice. The body weight and food intake of DIO mice were decreased after BBR treatment, and the weight change was negatively correlated with the serum GDF15 level. Further studies show that BBR induced GDF15 mRNA expression and secretion in the brown adipose tissue (BAT) of DIO mice and primary mouse brown adipocytes. In addition, we found that BBR upregulates GDF15 mRNA expression and secretion by activating the integrated stress response (ISR) in primary mouse brown adipocytes. Overall, our findings show that BBR lowers body weight by inducing GDF15 secretion via the activation of the ISR in BAT.
Obesity is caused by lipid deposition, which occurs due to an imbalance of energy intake and consumption. Obesity is closely associated with the occurrence and development of cardiovascular disease and respiratory system, nervous system, and other multisystem diseases (1). Obesity is also associated with high rates of all-cause mortality (2). According to the World Health Organization, 600 million people were obese in 2016. It is urgent to identify low-risk therapies. Nevertheless, the treatment of obesity with clinical drugs currently has limited effects.
Berberine (BBR), a natural plant isoquinoline alkaloid from the Chinese medicine Coptis chinensis, is widely used in the treatment of gastrointestinal infections (3). With further research on the pharmacological mechanisms of BBR, it has been shown that the beneficial effects of BBR on metabolism are pleiotropic (4). BBR can ameliorate obesity in many ways, such as reducing appetite (5), promoting thermogenesis of beige adipose tissue and brown adipose tissue (6, 7), improving glucose metabolism (6, 7), regulating short-chain fatty acid production (8), upregulating glucagon-like peptide expression (9), and modulating the gut microbiota (10). To date, molecular mechanisms underlying the effect of BBR on weight loss, including the activation of 5’ AMP-activated protein kinase, glucagon-like peptide-1, neuropeptide Y, and uncoupling protein 1 (UCP1) (5, 6, 11) and the inhibition of peroxisome proliferator-activated receptor-gamma an C/EBPα function (12, 13), have been proposed. However, the targets and mechanisms associated with the effects of BBR on weight loss remain still unclear.
Growth differentiation factor 15 (GDF15) is a distant member of the transforming growth factor-β superfamily (14) and is an established biomarker of cellular stress (15, 16). GDF15 has emerged as a new anti-obesity target (15). With the discovery of the high-affinity receptor of GDF15, glial cell-derived neurotrophic factor family receptor alpha-like (GFRAL), the mechanism underlying the function of GDF15 in suppressing appetite to lower weight has been revealed. GDF15 forms a ternary complex with GFRAL and receptor tyrosine kinase in the hypothalamus to reduce energy intake by suppressing appetite (17–19). In addition, peripheral GDF15 can also promote thermogenesis (20) and lipolysis (21) and reduce adipose tissue mass (21). GDF15 is an important target for weight loss. Some anti-obesity drugs, such as metformin, have been found to lower body weight by elevating the circulating levels of GDF15 (19). In a previous study, we also found that BBR upregulates the expression of GDF15 via RNA sequencing analysis of mature adipocytes (Supplementary Fig. S1) (22). However, to date, research has not yet determined the association between BBR and GDF15.
In this study, we found that BBR increases the serum levels of GDF15 in diet-induced obese (DIO) mice and that the GDF15 levels were negatively correlated with the degree of weight gain. In a further study, we explored the tissue source and mechanism underlying BBR-induced GDF15 production. We characterized brown adipose tissue (BAT) as one of the main sources of the increased GDF15 levels in DIO mice exposed to BBR, and we demonstrated that GDF15 production is regulated by the integrated stress response (ISR). In addition, we confirmed that BBR-induced the thermogenesis in brown adipocytes was independent of GDF15 in vitro. This study elucidated the effect of BBR-induced GDF15 secretion and the underlying mechanism, which provided a new theoretical basis for the use of BBR to promote weight loss.
Materials and Methods
Animal Studies
Male C57BL/6J mice were obtained from Chinese Academy of Sciences Shanghai SLAC Laboratory Animal Co. (SLACCAS, Shanghai, China) at 6 weeks of age and housed in a room with constant temperature (22 ± 2 °C), humidity (55 ± 5%), and a 12-hour light/dark cycle for 2 weeks. C57BL/6J mice were given free access to food and water throughout the experiment. The male C57BL/6J mice were adaptively fed until 8 weeks of age, and then the mice were randomly divided into the chow diet model group (4% fat, SLACOM P1101F) and high-fat diet (HFD) model group (60% fat, research diets D12492). BBR (0, 50, 100 mg/kg) and metformin (600 mg/kg) were administered to the mice by gavage for the designated time, and the body weight and food intake were measured every week. None of the mice were fasted during the whole experiment, and blood was collected from the retro-orbital plexus at the designated time. The whole blood, liver, kidney, subcutaneous adipose tissue, epididymal adipose tissue (EAT), BAT, soleus muscle, gastrocnemius muscle, colon, ileum, and jejunum of the mice were collected 4 hours after the final dose of BBR was administered.
Preadipocyte Isolation and Differentiation
Interscapular BAT was harvested from C57BL/6J male mice (5 weeks old) and digested at 37 °C for 20 minutes with medium supplemented with 2 g/L type II collagenase (Solarbio, A8020) and 15 g/L BSA (Sigma, C6885). Then the cells were placed on ice for 10 minutes, filtered, deoiled, and seeded in a 6-well plate with 10% fetal bovine serum DMEM. After the cells grew to fusion, the cells were incubated with 0.5 μM IBMX (Sigma, I7018), 1 μM dexamethasone (Sigma, D4902), 1.7 μM insulin (Lily, HI0240), 0.05 μM T3 (Sigma, T2877), and 5 μM rosiglitazone (Sigma, R2408) for 2 days and then treated with insulin, T3, and rosiglitazone for 4 days. The culture medium was changed to 2% BSA-DMEM 6 hours before treatment with 2% BSA-DMEM supplemented with or without BBR (Sigma), metformin (Sigma, D150959), tunicamycin (MCE, HY-A0098), cAMP (MCE, HY-B1511), and integrated stress response inhibitor (ISRIB) (Sigma, SML0843) for different times.
Oil Red O Staining of Mature Brown Adipocytes
Oil Red O (Sigma, O0625) (0.5% in isopropanol) was diluted with distilled water (3:2) and filtered twice through a 0.45 mm filter. The mature brown adipocytes were washed twice with PBS, frozen at 4 °C, and fixed with 4% paraformaldehyde (Servicebio, BL539A) for 5 minutes. After washing the cells with PBS two more times, the cells were incubated with filtered Oil Red O for 15 minutes at RT.
Real-time Quantitative PCR (RT–qPCR)
The total RNA was extracted from cells and tissues by TRIzol reagent (Invitrogen, 15596018) and reverse transcribed into cDNA with PrimeScript RT Master Mix (Takara, RR047B). The mRNA expression of related genes was measured by the SYBR green PCR kit (Vazyme, Q511-02). The relative genes expression levels were normalized to the expression of the internal control (36B4, GAPDH, β-actin, 18S), and the primer sequences are shown in Table 1.
Primers used for quantitative real-time PCR
| Gene | Sequence |
|---|---|
| 36B4 | F- AAGCGCGTCCTGGCATTGTCT |
| R- CCGCAGGGGCAGCAGTGGT | |
| GDF15 | F-CAATGCCTGAACAGCGACC |
| R-CTGAGTTCGAGTCCTCTCGG | |
| Actin | F- GGCTGTATTCCCCTCCATCG |
| R- CCAGTTGGTAACAATGCCATGT | |
| GAPDH | F- CCTGTTGCTGTAGCCGTAT |
| R-ACTCTTCCACCTTCGATGC | |
| 18S | F- GTAACCCGTTGAACCCCATT |
| R- CCATCCAATCGGTAGTAGCG | |
| ATF4 | F- CCTTCGACCAGTCGGGTTTG |
| R- CTGTCCCGGAAAAGGCATCC | |
| CHOP | F- CTGGAAGCCTGGTATGAGGAT |
| R- CAGGGTCAAGAGTAGTGAAGGT |
| Gene | Sequence |
|---|---|
| 36B4 | F- AAGCGCGTCCTGGCATTGTCT |
| R- CCGCAGGGGCAGCAGTGGT | |
| GDF15 | F-CAATGCCTGAACAGCGACC |
| R-CTGAGTTCGAGTCCTCTCGG | |
| Actin | F- GGCTGTATTCCCCTCCATCG |
| R- CCAGTTGGTAACAATGCCATGT | |
| GAPDH | F- CCTGTTGCTGTAGCCGTAT |
| R-ACTCTTCCACCTTCGATGC | |
| 18S | F- GTAACCCGTTGAACCCCATT |
| R- CCATCCAATCGGTAGTAGCG | |
| ATF4 | F- CCTTCGACCAGTCGGGTTTG |
| R- CTGTCCCGGAAAAGGCATCC | |
| CHOP | F- CTGGAAGCCTGGTATGAGGAT |
| R- CAGGGTCAAGAGTAGTGAAGGT |
Abbreviations: ATF4, activating transcription factor 4; CHOP, C/EBP homologous protein; GADPH, glyceraldehyde-3-phosphate dehydrogenase; GDF15, growth differentiation factor 15.
Primers used for quantitative real-time PCR
| Gene | Sequence |
|---|---|
| 36B4 | F- AAGCGCGTCCTGGCATTGTCT |
| R- CCGCAGGGGCAGCAGTGGT | |
| GDF15 | F-CAATGCCTGAACAGCGACC |
| R-CTGAGTTCGAGTCCTCTCGG | |
| Actin | F- GGCTGTATTCCCCTCCATCG |
| R- CCAGTTGGTAACAATGCCATGT | |
| GAPDH | F- CCTGTTGCTGTAGCCGTAT |
| R-ACTCTTCCACCTTCGATGC | |
| 18S | F- GTAACCCGTTGAACCCCATT |
| R- CCATCCAATCGGTAGTAGCG | |
| ATF4 | F- CCTTCGACCAGTCGGGTTTG |
| R- CTGTCCCGGAAAAGGCATCC | |
| CHOP | F- CTGGAAGCCTGGTATGAGGAT |
| R- CAGGGTCAAGAGTAGTGAAGGT |
| Gene | Sequence |
|---|---|
| 36B4 | F- AAGCGCGTCCTGGCATTGTCT |
| R- CCGCAGGGGCAGCAGTGGT | |
| GDF15 | F-CAATGCCTGAACAGCGACC |
| R-CTGAGTTCGAGTCCTCTCGG | |
| Actin | F- GGCTGTATTCCCCTCCATCG |
| R- CCAGTTGGTAACAATGCCATGT | |
| GAPDH | F- CCTGTTGCTGTAGCCGTAT |
| R-ACTCTTCCACCTTCGATGC | |
| 18S | F- GTAACCCGTTGAACCCCATT |
| R- CCATCCAATCGGTAGTAGCG | |
| ATF4 | F- CCTTCGACCAGTCGGGTTTG |
| R- CTGTCCCGGAAAAGGCATCC | |
| CHOP | F- CTGGAAGCCTGGTATGAGGAT |
| R- CAGGGTCAAGAGTAGTGAAGGT |
Abbreviations: ATF4, activating transcription factor 4; CHOP, C/EBP homologous protein; GADPH, glyceraldehyde-3-phosphate dehydrogenase; GDF15, growth differentiation factor 15.
Western Blotting
Protein samples were extracted with RIPA lysis buffer (Beyotime, P0013B) supplemented with phosphatase inhibitor (Roche, 4906845001) and protease inhibitor (Roche, 04693132001), and then, the samples were stored at −80 °C after the protein concentration was determined by a BCA assay kit. SDS–PAGE gels (8-12%) were selected according to the molecular weights of the different proteins, and the proteins were separated with 80 V for 2 hours. The proteins were transferred to NC membranes, which were then blocked with 5% milk for 1 hour. The primary antibodies used were specific for the following proteins: heat shock protein 90 (Cell Signaling Technology (CST), Cat#ab4874, RRID: AB_2121214), binding immunoglobulin (CST, Cat#ab3177, RRID: AB_2119845), endoplasmic reticulum oxidoreductase 1 alpha (CST, Cat#ab3264, RRID: AB_823684), PKR-like endoplasmic reticulum kinase (PERK) (CST, Cat#ab5683, RRID: AB_10841299), IRE1α (CST, Cat#3294, RRID: AB_823545), C/EBP homologous protein (CHOP) (CST, Cat#ab2895, RRID: AB_2089254) (ER Stress Antibody Sampler Kit) (CST, 9956), ATF-4 (D4B8) (CST, Cat#ab11815, RRID: AB_2616025), eukaryotic translation initiation factor 2A (eIF2α) (CST, Cat#ab9722, RRID: AB_2230924), and phospho-eIF2α (Ser51) (CST, Cat#ab9721, RRID:AB_330951). The membranes were incubated with these antibodies overnight at 4 °C. The secondary antibodies were HRP-conjugated goat anti-rabbit IgG (H + L) (BioTNT, Cat# A20120A0704, RRID:AB_2924777) and HRP-conjugated goat anti-mouse IgG (H + L) (Bio TNT, Cat# A20120A0703, RRID:AB_2924776). Immunoblotting images were captured with Image Quant LAS4000 Imaging Systems (GE Healthcare, USA).
Quantification of Soluble Protein Levels (ELISA)
Culture medium of cells and mouse serum were centrifuged at 500 r for 5 minutes, the impurities were removed, and the samples were diluted at different ratios. The GDF15 levels were measured with a Mouse/Rat GDF-15 Immunoassay (ELISA) kit (R&D Systems, Cat#MGD150, RRID:AB_2909411) according to the manufacturer's instruction manual.
Statistical Analysis
The individual replicates in the independently repeated experiments are indicated in the figures. Each group of animal experiments included at least eight independent mice of comparable ages and weights. The statistical analyses were performed using one-way ANOVA, two-way ANOVA or Student's test in GraphPad Prism 8.0. All the results are expressed as the mean ± s.e.m. For all the statistical analyses, P < .05 was considered statistically significant.
Results
BBR Reduces Body Weight and Elevates Circulating GDF15 Levels
To test the effect of BBR on serum GDF15 levels in vivo, both lean mice and DIO mice were orally treated with an acute single dose of vehicle or BBR, and serum GDF15 levels were measured 4 hours after dosing. Metformin, which could induce the secretion of GDF15 in both lean and DIO mice (19), was used as a positive control. Consistent with our hypothesis, BBR was shown to have an obvious effect in increasing serum GDF15 concentration in both lean and DIO mice (Fig. 1A, Supplementary Fig. S2A and 2B) (22). As shown in Fig. 1A, the serum GDF15 level was increased by 1.4-fold in DIO mice 4 hours after treatment with BBR. To further determine whether long-term BBR treatment could sustain high circulating levels of GDF15 and thereby lower body weight in DIO mice, we performed a longitudinal study in DIO mice that received 100 mg/kg/day BBR orally for 2 weeks, and we measured the GDF15 serum level, food intake, and body weight. The results showed that the circulating level of GDF15 increased by 26% (Fig. 1B) and the body weight decreased by 6% (Fig. 1C) in DIO mice after BBR treatment, and the serum GDF15 level was associated with weight loss in DIO mice treated with BBR (R2 = 0.3694, P = .0125) (Fig. 1D). BBR also inhibited the food intake of DIO mice (Fig. 1E). Taken together, these results suggest that elevated GDF15 levels might be one of the main causes of weight loss in BBR-treated DIO mice, and it might function by suppressing appetite.
Berberine promotes GDF15 secretion, suppresses appetite, and promotes weight loss in obese mice. (A) Serum levels of GDF15 in HFD-induced obese mice treated with berberine (0, 100 mg/kg) or metformin (600 mg/kg) for 4 hours (n = 8 per group). The data are expressed as the mean ± SEM. *P < .05, *P < .01, *P < .001. P-values (95% CI) were determined by one-way ANOVA. (B, C, E) Serum levels of GDF15 (B), weight change (%) (C), and weekly food intake change (g) (D) in HFD-induced obese mice treated with berberine (0, 100 mg/kg) for 14 days (n = 8 per group). The data are expressed as the mean ± SEM. *P < .05, *P < .01, *P < .001. P-values (95% CI) were determined by 2-tailed t test. (D) Change in body mass vs serum levels of GDF15 in HFD-induced obese mice treated with vehicle or berberine for 14 days (n = 8 per group). A linear model was constructed with serum levels of GDF15 after 2 weeks of treatment as the independent variable and weight change as the dependent variable. R2 = −0.3694, 2-sided P = .0125.
Abbreviations: GDF15, growth differentiation factor 15; HFD, high-fat diet.
BBR Induces Brown Adipose Tissue to Secrete GDF15
GDF15 is a secreted protein that can potentially be produced by multiple tissues under different conditions (15). To identify which tissues that are targeted by BBR secrete GDF15, we examined GDF15 mRNA expression in panel of tissues, including liver, kidney, subcutaneous adipose tissue, EAT, BAT, soleus muscle, gastrocnemius muscle, colon, ileum, and jejunum; these tissues have been reported to be potential sources of circulating GDF15. Our data and other studies (15, 19, 23) showed that in DIO mice, GDF15 was predominantly expressed in the kidney, liver, BAT, and EAT (Fig. 2A). After a 4-h treatment with BBR, GDF15 mRNA expression in BAT, colon, ileum, and jejunum increased by 2-, 5.6-, 4.0-, and 2.2-fold, respectively (Fig. 2B). This indicates that the major tissue sources of the elevated serum GDF15 levels induced by BBR are BAT, as well as intestinal tissue.
Berberine increases GDF15 expression in brown adipose and intestinal tissues in obese mice. (A) Multitissue transcriptomic profiling of GDF15 gene expression in HFD-induced obese mice treated with berberine (0, 100 mg/kg) for 4 hours shown as raw ct expression values (n = 6 per group). (B) Quantification of GDF15 mRNA expression in HFD-induced obese mice treated with berberine (100 mg/kg) for 4 hours is shown as the fold change compared to vehicle treatment group (n = 8 per group). The data are expressed as the mean ± SEM. *P < .05, *P < .01, *P < .001. P-values (95% CI) were determined by 2-tailed t test.
Abbreviations: GDF15, growth differentiation factor 15; HFD, high-fat diet.
BBR Promotes Brown Adipocytes to Secrete GDF15
To further determine whether brown adipocytes are capable of responding to BBR by increasing GDF15 production, we utilized mature brown adipocytes differentiated from the isolated BAT- Semliki Forest viruses of male C57BL/6J mice (Fig. 3A). We incubated the brown adipocytes with BBR and used metformin (24) and tunicamycin (19, 23) as positive controls. The results showed that within 24 hours, the basal protein levels of GDF15 in the supernatants of brown adipocytes were only 57 pg/mL, but BBR increased the concentrations of GDF15 in the supernatants of brown adipocytes by up to 8.8-fold (513 pg/mL) (Fig. 3B). To thoroughly investigate the details by which BBR upregulated GDF15 production, we incubated brown adipocytes with BBR, and further analysis showed that the mRNA expression of GDF15 was regulated by BBR at the transcriptional level in a time- and dose-dependent manner (Fig. 3C). BBR (1 μM) promoted the increase in GDF15 protein expression by the highest degree (17-fold) and increased GDF15 protein levels in the cell supernatants in a time-dependent manner (Fig. 3D and 3E). In summary, these results suggest that the BBR-induced increase in the serum levels of GDF15 in DIO mice at least partly originated from brown adipocytes. Furthermore, we transfected the brown adipocytes with GDF15 siRNA to confirm whether BBR activates the expression of UCP1 in brown adipocytes is dependent on GDF15. Inconsistent with our prediction, the expression of UCP1 in BBR-treated brown adipocytes was not different from vehicle controls upon GDF15 siRNA (Supplementary Fig. S3) (22), which suggests that BBR promotes the expression of UCP1 independent of GDF15, at least in vitro.
Berberine increases the expression and secretion of GDF15 in brown adipocytes. (A) Phase-contrast images (left) and representative Oil Red O staining (right) of mature primary brown adipocytes induced from brown adipose Semliki Forest virus cells. (B) GDF15 protein levels in the supernatants of mature primary brown adipocytes treated with berberine (0, 1 μM), Tn (5 μg/mL), or metformin (0.5, 1 mM) for 24 hours (n = 8 per group). (C) GDF15 mRNA expression in mature primary brown adipocytes treated with berberine (0, 0.2, 1, 5 μM) for 6, 12, and 24 hours. (D) GDF15 protein levels in the supernatants of mature primary brown adipocytes treated with berberine (0, 0.2, 1, 5 μM) for 24 hours. (E) GDF15 protein levels in the supernatants of mature primary brown adipocytes treated with berberine (0, 1 μM) for 6, 12, and 24 hours (n = 8 per group). The data are expressed as the mean ± SEM. *P < .05, *P < .01, *P < .001. P-values (95% CI) were determined by 1-way ANOVA (B, C, D). P-values were determined by 2-way ANOVA (E).
Abbreviations: GDF15, growth differentiation factor 15.
BBR Induces GDF15 Secretion by Activating the ISR in Brown Adipocytes
Our previous studies showed that in mature adipocytes, BBR can upregulate the expression of activating transcription factor 4 (ATF4) and CHOP, which are related to the ISR; the ISR is a classical pathway that promotes the expression and secretion of GDF15 (19, 24, 25) (Supplementary Fig. S1) (22). Therefore, we hypothesized that the ISR might be involved in the process by which BBR promotes GDF15 protein secretion by brown adipocytes. Consistent with this, BBR-treated brown adipocytes exhibited significant activation of the ISR pathway. As shown in Fig. 4A, 4B and 4C, the eIF2α phosphorylation and expression of ISR markers (PERK, binding immunoglobulin, endoplasmic reticulum oxidoreductase 1 alpha, ATF4, and CHOP) were significantly increased in brown adipocytes after BBR treatment for 1 hour. Compared with group with BBR for 0 hours, the protein expression of phospho-eIF2α/eIF2α in BBR treated group for 1, 2, 3 hours increased 1.7 times, 22 times and 2.2 times, respectively. To further confirm whether BBR-induced GDF15 secretion was dependent on the ISR, we pretreated the cells with an eIF2α inhibitor (ISRIB) to block eIF2α phosphorylation and then treated the cells with BBR and ISRIB at the appropriate times. The BBR-induced expression of ATF4 and CHOP and phosphorylation of eIF2α were markedly reduced by cotreatment with ISRIB (Fig. 4D-4F). This suggests that ISRIB inhibits the ISR induced by BBR. Furthermore, GDF15 secretion in response to BBR was also significantly reduced by cotreatment with ISRIB (Fig. 4G and 4H). These data suggest that the effects of BBR on GDF15 expression and secretion are at least partly dependent on the ISR pathway. Moreover, fibroblast growth factor 21 (FGF21) is also one of the factors that upregulate the secretion of GDF15 (26). In vitro experiments, BBR increased the expression of GDF15while promoting the expression of FGF21 mRNA, and the FGF21 mRNA expression was weakened upon the ISRIB (Supplementary Fig. S4) (22). Therefore, we speculated that BBR may also increase the expression of GDF15 by promoting FGF21, but the specific mechanisms need to be further explored.
Berberine increases GDF15 secretion by activating the ISR in brown adipocytes. (A) The expression of ATF4 and CHOP in mature primary brown adipocytes treated with berberine (0, 1 μM) for 6, 12, and 24 hours (n = 4). (B, C) Representative Western blots showing protein levels of BIP, ERO1L, PERK, p-eIF2α, eIF2α, ATF4, CHOP, and HSP90 in mature primary brown adipocytes treated with berberine (1 μM) for 0, 0.5, 1, and 3 hours. HSP90 was used as a loading control. Densitometric analysis of band intensity relative to HSP90 and normalized to the group of 0 hours. (A, C) The data are expressed as the mean ± SEM. *P < .05, *P < .01, *P < .001. P-values (95% CI) were determined by 1-way ANOVA. (D) The expression of ATF4 and CHOP in mature primary brown adipocytes treated with berberine (1 μM) in the absence or presence of ISRIB (0.2 μM) for 12 hours (n = 5). (E, F) The protein levels of p-eIF2α, eIF2α, ATF4, CHOP, and HSP90 in mature primary brown adipocytes treated with berberine (1 μM) for 0, 0.5, 1, and 3 hours in the absence or presence of ISRIB (0.2 μM). HSP90 was used as a loading control. Densitometric analysis of band intensity relative to HSP90 and normalized to control samples. (G) GDF15 mRNA expression in mature primary brown adipocytes treated with berberine (1 μM) for 12 hours in the absence or presence of ISRIB (5 μM) (n = 4). (H) GDF15 protein levels in the supernatants of mature primary brown adipocytes treated with berberine (1 μM) for 24 hours in the absence or presence of ISRIB (5 μM) (n = 6). (D, F, G, H) The data are expressed as the mean ± SEM. *P < .05, *P < .01, *P < .001. P-values (95% CI) were determined by 1-way ANOVA relative to control samples. #P < .05, ##P < .01, ###P < .001. P-values (95% CI) by 2-tailed t test relative to berberine-treated samples.
Abbreviations: ATF4, activating transcription factor 4; BIP, binding immunoglobulin; CHOP, C/EBP homologous protein; eIF2α, eukaryotic translation initiation factor 2A; ERO1A, endoplasmic reticulum oxidoreductase 1 alpha; GDF15, growth differentiation factor 15; HSP90, heat shock protein 90; ISR, integrated stress response; ISIRB, integrated stress response inhibitor; PERK, PKR-like endoplasmic reticulum kinase; p-eIF2α, phospho-eIF2α.
Discussion
In this study, we discovered that BBR acts as an inducer of GDF15 production. The vivo experiments show that in DIO mice administered BBR via gavage, changes in the serum levels of GDF15 were significantly correlated with changes in weight. In this study, we utilized brown adipocytes and tissues to confirm that the increased serum levels of GDF15 are partly derived from BAT in DIO mice and increased GDF15 production is at least partly dependent on the ISR pathway in mature brown adipocytes. Although many mechanisms have been suggested to be involved the reduction of body weight by BBR, there are still many unknown targets. Our discoveries provide a new, compelling explanation for this important aspect of the effects of BBR.
We provide a target to elucidate the benefits of BBR in weight loss. The most striking finding is that BBR increases the circulating levels of the peptide hormone GDF15, which has been shown to lower body weight by reducing food intake (27). The correlation analysis showed that changes in the serum GDF15 levels of DIO mice were negatively correlated with changes in body weight, which provides further support for the hypothesis that BBR can ameliorate obesity by promoting GDF15 secretion. In addition, our study found that the mRNA level and circulating level of GDF15 were increased, when the obese mice with BBR treatment for 4 hours, before the weight loss was caused by BBR in obese mice. It ruled out that BBR promoted the GDF15 secretion not due to its weight loss effect.
As we know, appetite suppression is one of the important ways by which BBR causes weight loss. Previous studies have proposed that BBR suppresses appetite, implicating several possible targets, such as neuropeptide Y (5), pro-opiomelanocortin (6), and the metabolic sensor 5’ AMP-activated protein kinase (6), in the hypothalamus. In this study, we found that BBR not only inhibited food intake but also increased the serum levels of GDF15 in DIO mice. GDF15 is a classic factor that suppresses appetite, and it can inhibit food intake by targeting hypothalamic GFRAL, change dietary preferences, delay gastric emptying, and enhance satiety (18, 28). Therefore, we infer that BBR may also inhibit appetite by promoting GDF15 secretion. Our study provides a new mechanism underlying the effect of BBR on appetite suppression for further study.
Another way by which BBR promotes weight loss is by promoting energy consumption (6) and reducing fat mass (21), which is also one of the effects of GDF15 (18). Hence, we hypothesize that GDF15 may be one of the targets through which BBR promotes energy consumption. How GDF15 increases energy consumption has always been a hot topic. Chrysovergis et al found that overexpression of GDF15 could increase the expression of UCP1 in brown adipocytes (29). Inconsistent with Chrysovergis, our data show that BBR increases the expression of UCP1 independently of GDF15. Consistent with our study, Choi et al found that in a mouse model with adipocyte-specific Crif1 gene knockout, which decreased oxidative phosphorylation in adipocytes, resulted in an increase in GDF15 expression and a downregulation of UCP1 expression in brown adipocytes; however, energy consumption in mice was increased (30). Our in vitro experiments also found that BBR promoted UCP1 mRNA expression independently of GDF15 in brown adipocytes. Therefore, we hypothesize that the GDF15-mediated increase in energy consumption may not be directly related to the thermogenesis of brown adipocytes. Of course, we cannot rule out whether the GDF15, from brown adipocytes induced by BBR, targets other tissues or cells or whether it reacts with BAT through paracrine or endocrine effects of other proteins to enhance the UCP1-dependent thermogenesis function of BAT. In addition, it cannot be ruled out whether BBR can exert UCP1-independent thermogenesis production through GDF15. Similarly, although the basal expression of GDF15 in white adipose tissue itself was low (19), we cannot judge that BBR reduces white adipose tissue mass and promotes white adipose tissue thermogenesis (6) related to GDF15-dependent paracrine and autocrine. For example, data from a study (21) on reversing cancer cachexia by antagonizing GDF15 receptors in mice showed that GDF15 can target GFRAL to affect sympathetic activity, promote the expression of hormone-sensitive lipase and adipose triglyceride lipase, increase lipolysis, reduce fat mass, and thereby cause weight loss; these results suggest that GDF15 may increase energy consumption by promoting lipolysis. We speculate that BBR promotes lipolysis to increase energy consumption (6) via GDF15. In future investigations, the brown adipocyte specific GDF15 knockout mice model and GDF15 neutralizing antibody will be needed to elucidate whether the effect and the specific mechanism of suppressing appetite and gain weight induced by BBR through GDF15.
After elucidating the possible role of BBR in weight loss via GDF15, this study attempted to identify the sources of BBR-induced GDF15 secretion in DIO mice. Our data revealed that BBR induced the secretion of GFD15 from the BAT and intestines of DIO mice. Although the liver is important in the metabolic action of BBR (31), notably, GDF15 mRNA expression of the liver was not changed in DIO mice exposed to BBR. To determine whether hepatocytes are capable of responding to BBR by increasing GDF15 expression, we incubated freshly isolated mouse hepatocytes with different concentrations of BBR and found that treatment with the high concentration BBR induced a clear elevation of GDF15 mRNA expression in hepatocytes (Supplementary Fig. S5) (22). We conclude that BBR can induce GDF15 expression in many cell types, but, at least when BBR is given orally to mice, GDF15 mRNA expression is mostly induced in BAT and intestine. Similar to metformin (19), they all have distinct functions to induce the expression of GDF15 in vitro and in vivo.
Therefore, in an in vitro experiment, we further determined whether BBR induced the secretion of GDF15 from cells derived from BAT and intestine. Previously, researchers agreed that adipocytes secrete a small amount of GDF15, and no in-depth study of adipocytes has been conducted (15, 19). However, our data show that BBR can stimulate the potential of brown adipocytes to secrete substantial amounts of GDF15, although brown adipocytes themselves secrete small amounts of GDF15. These data provide a basis for further research on the function of GDF15 secretion by brown adipocytes. We also preliminarily verified the effect of BBR on intestinal epithelial cells, which were reported to be the main site of GDF15 expression in the intestine (19). MODE-K cells were used in vitro experiments, and the results suggested BBR can increase the mRNA expression of GDF15 in MODE-K cells (Supplementary Fig. S6) (22). The in vivo and in vitro experiments suggested that the main tissue sources of BBR-indued GDF15 secretion were BAT and intestine. However, how much GDF15 is secreted by the intestine and BAT and released into circulation as well as the role of this secreted GDF15 in ameliorating obesity need to be further studied in tissue-specific knockout mice.
In this study, we found that the effects of BBR on GDF15 expression are at least partly dependent on the ISR pathway. Our data suggest that BBR can activate PERK and phosphorylate eIF2α to increase the expression of ATF4 and CHOP, thus increasing the transcription and translation of GDF15, which can be directly bound by CHOP (32). In addition, we also found that BBR promoted the expression of FGF21 through the ISR in brown adipocytes. Campderros et al found that knockout of the FGF21 receptor KLB hinders GDF15 secretion and that recombinant FGF21 can promote GDF15 secretion by brown adipocytes (26). These findings suggest that BBR promotes GDF15 transcription and may be related to both CHOP and FGF21 in the ISR, which requires further the establishment of the corresponding animal and cell models.
In conclusion, this study shows that BBR induces the secretion of GDF15 by activating the ISR in brown adipocytes, ultimately inhibiting appetite and promoting weight loss. We provide a new theoretical basis for further studying the mechanism by which BBR ameliorates obesity as a new inducer of GDF15 expression in brown adipocytes and a breakthrough for further study of the function of brown adipocytes. This study will support further studies on the specific role of BBR-induced GDF15 secretion by adipocytes in ameliorating obesity as well as the underlying mechanism.
Acknowledgments
We thank Ying Yang from Shanghai Jiao Tong University Affiliated Sixth People's Hospital for kindly providing experimental technical guidance.
Funding
This work was supported by the National Natural Science Foundation of China (82074178), Natural Science Foundation of Shanghai (20ZR1442700), Pudong New Area Health Commission clinical characteristic discipline construction project (PWYts2021-13), Pudong New Area Science and Technology Development Fund (PKJ2022-Y10), and Shanghai University of TCM 20th curriculum construction project (SHUTCM2021KCO21).
Author Contributions
C.L. performed experiments, analyzed the data and wrote the manuscript. Q.L. and L.L. performed experiments and analyzed the data. F.H., S.G., and Y.X. assisted the experiments. Y.Y. designed the experiments, oversaw experiments, and analyzed the data. H.Z. and X.L. conceived the project, supervised research, reviewed/edited the manuscript. All authors read and approved the final manuscript.
Disclosures
The article’s conclusions, implications, or opinions are the consensus of all the authors.
Data Availability
The original contributions presented in the study are included in the article/supplementary material (22). Further inquiries can be available from the corresponding author on reasonable request.
Ethics Statement
The animal study was reviewed and approved by the Animal Ethics Committee of the Shanghai University of TCM Affiliated Seventh People's Hospital.
References
Author notes
Chang Li, Qingyang Leng and Lihua Li contributed equally to this work.
