Sinapic acid induces the expression of thermogenic signature genes and lipolysis through activation of PKA/CREB signaling in brown adipocytes

Abstract

Lipid accumulation in white adipose tissue is the key contributor to the obesity and orchestrates numerous metabolic health problems such as type 2 diabetes, hypertension, atherosclerosis, and cancer. Nonetheless, the prevention and treatment of obesity are still inadequate. Recently, scientists found that brown adipose tissue (BAT) in adult humans has functions that are diametrically opposite to those of white adipose tissue and that BAT holds promise for a new strategy to counteract obesity. In this study, we evaluated the potential of sinapic acid (SA) to promote the thermogenic program and lipolysis in BAT. SA treatment of brown adipocytes induced the expression of brown-adipocyte activation–related genes such as Ucp1, Pgc-1α, and Prdm16. Furthermore, structural analysis and western blot revealed that SA upregulates protein kinase A (PKA) phosphorylation with competitive inhibition by a pan-PKA inhibitor, H89. SA binds to the adenosine triphosphate (ATP) site on the PKA catalytic subunit where H89 binds specifically. PKA-cat-α1 gene–silencing experiments confirmed that SA activates the thermogenic program via a mechanism involving PKA and cyclic AMP response element–binding protein (CREB) signaling. Moreover, SA treatment promoted lipolysis via a PKA/p38-mediated pathway. Our findings may allow us to open a new avenue of strategies against obesity and need further investigation.

INTRODUCTION

Brown adipose tissue (BAT) in adult humans was discovered recently in imaging studies and has a function opposite to that of white adipose tissue (1). The unique feature of BAT is expression of UCP1, which helps to dissipate energy by diminishing the proton gradient in the inner membrane of mitochondria (2). Cold exposure of adult human obese and normal subjects has various physiologically beneficial effects such as increased glucose uptake and increased insulin sensitivity (3). In addition, cold exposure has been shown to cause BAT activation (4).

Thus, the well-known method for activation of BAT is cold exposure, which increases norepinephrine secretion and thermogenesis. In a study on a mouse model, it has been demonstrated that differentiation of brown fat is accompanied by mitochondrial biogenesis (5). Another study (on a brown preadipocyte cell line) has revealed that BMP7 treatment significantly increases the expression of genes involved in mitochondrial biogenesis and function, including Pgc-1α and Nrf1 (6).

Certainly, PPARγ is the central regulator of both white fat and brown fat cell development. C/EBP family members function cooperatively with PPARγ and stimulate a transcriptional cascade to maintain a stable differentiated state of adipocytes. One report suggests that PPARγ is necessary for brown-fat development but not sufficient to drive the full brown-fat program (which requires additional factors) (7). Other studies offer evidence that transcriptional regulators including PPARγ, PGC-1α, PRDM16, and C/EBPβ can stimulate the development of brown fat (7, 8). Recently, research revealed that phytochemicals such as rutin activate brown fat, and cryptotanshinone promotes brown-adipocyte commitment among mesenchymal stem cells (9, 10).

In rodents, external stimuli lead to activation of β-adrenergic receptor 3 (β-3-AR) by sympathetic stimulation and increased cAMP production, which activates PKA and downstream genes of PKA (11). The activation of PKA promotes the release of fuel (free fatty acids) for thermogenesis. In mice, PKA modulates lipolysis activation involving β3-AR signaling which increases adenylyl cyclase activity and raises intercellular cAMP concentration (12). The elevated cAMP level triggers PKA, which phosphorylates HSL, ATGL, and perilipin, and thus subsequently initiates lipolysis (13). One report suggests that PKA upregulates UCP1 and plays role in both brown-adipocyte differentiation and mitochondrial biogenesis (11).

Sinapic acid (SA) is a natural alkaloidal amine found in black mustard seeds, wine, and vinegar. SA has been reported to have several biological functions, including antioxidant, anti-inflammatory, anticancer, antimicrobial, antimutagenic, and antianxiety activities (14). On the other hand, to date, the effects of SA on thermogenesis remain unexplored. In this study, we used BAT cells to find out the effect of SA on thermogenesis. We propose that SA holds promise as a potential secondary metabolite with a thermogenesis-promoting ability.

RESULTS

SA increases the expression of thermogenic markers and promotes mitochondrial biogenesis in brown adipocytes

During screening of 854 phytochemicals as described in our previous study (15) to find potential browning-thermogenic compounds, we found that SA can promote the activation of brown adipocytes. To decipher its function, at first, we tested the toxicity of SA toward brown adipocytes at indicated concentrations (Fig. 1A). It turned out that SA is not cytotoxic up to 72 h. Here, we compared the effects of SA with those of a well-known thermogenic inducer, Rosi, and BMP7 on the expression of thermogenic signature genes. SA treatment significantly enhanced Ucp1 mRNA expression, by 1.8-fold (Fig. 1B). Additionally, SA treatment increased the mRNA expression levels of Prdm16 (1.3-fold) and Pgc-1α (5.4-fold; Fig. 1B). As shown in Fig. 1C, SA treatment raised the protein expression levels of PRDM16, PGC-1α, PPARγ, and UCP1. The quantification data revealed that SA treatment increased UCP1 protein expression 1.7-fold, PGC-1α 1.9-fold, PRDM16 1.5-fold, and PPARγ protein expression 1.8-fold (Fig. 1C). It has been reported that PPARγ and PGC1a activation can promote mitochondrial biogenesis. So, we tested some of the mitochondrial-biogenesis-related genes: Cox7a, Cox8b and Nrf1. As shown in Fig. 2A, SA increased mitochondrial biogenesis related markers. Then, we performed immunostaining with MitoTracker and quantify the stain. As shown in Fig. 2B, SA treated cells showed higher mitochondrial mass than MDI. Next, we measured mtDNA expression and found that SA significantly increased mtDNA (Fig. 2C). After that, we needed to test whether mitochondrial biogenesis leads to increased oxygen consumption. As shown in Fig. 2D, SA treatment increased oxygen consumption. These data enabled us to propose that SA can induce the expression of thermogenic and mitochondrial biogenesis markers.

Fig. 1.

SA induces expression of Ucp1 along with other thermogenic genes. (A) Viability of brown adipocytes was assessed by the MTT assay at 24, 48, and 72 h. Data are expressed as mean ± SD of three independent experiments. qRT-PCR analysis was performed as described in the Materials and Methods section. (B) mRNA expression of brown-fat–specific markers: Ucp1, Pgc-1α, and Prdm16. Data are presented as mean ± SEM of three individual experiments. (C) Protein expression levels of brown-fat–specific markers UCP1, PRDM16, PPARγ, and PGC-1α. β-actin served as a loading control. Quantification of the protein expression levels of the brown-fat–specific markers. *A significant difference from group MDI (*P < 0.05, **P < 0.01, ***P < 0.001).

Fig. 2.

SA increases mitochondrial biogenesis. (A) mRNA expression of mitochondrial-biogenesis-related markers: Cox7a, Cox8b, and Nrf1. Data are presented as mean ± SEM of three individual experiments. (B) MitoTracker immunostaining images of BAT cells. Data are representative of three independent experiments. Quantification of the immunostaining images. Data are expressed as mean ± SD of three independent experiments. (C) Relative mitochondrial Cox1 DNA content (ratio of nDNA p0 and mtDNA Cox1). (D) A Luxcel MitoXpress fluorescence assay was conducted to determine the oxygen consumption rate. Gox = glucose oxidase (positive control), AA = antimycin A (negative control). Data are shown as % of control (group MDI) ± SD from three independent experiments. *A significant difference from group MDI (*P < 0.05, **P < 0.01, ***P < 0.001).

SA activates the PKA pathway in brown adipocytes

Of note, we found that SA and H89 share the ATP-binding pockets on the PKA surface via both hydrogen and hydrophobic interactions (Fig. 3A). SA and H89 seem to share amino acid residues Phe³²⁷, Glu⁹¹, Ala⁷⁰, Lys⁷², Thr¹⁸³, Leu¹⁷³, Val⁵⁷, and Asp¹⁸⁴ in the active site of PKA (Fig. 3A). Structural analysis revealed that the binding-affinity values of SA and H89 are −6.2 and −8.3 kcal/mol, respectively, implying that H89 binds more strongly than SA, but SA has the ability to slightly inhibit the activity of H89, as confirmed by western blotting (Fig. 3B). We increased the SA concentration and treated PKA with H89; we found that SA treatment inhibited the H89 activity (Fig. 3B). The molecular docking analysis suggests that SA may be a potent agonist for induction of PKA phosphorylation.

Fig. 3.

SA activates PKA pathways in BAT cells. (A) Structural analysis of SA and the pan-PKA inhibitor H89 in complex with PKA. They share the same binding pocket on the PKA surface, and the active-site amino acid residues Phe³²⁷, Glu⁹¹, Ala⁷⁰, Lys⁷², Thr¹⁸³, Leu¹⁷³, Val⁵⁷, and Asp¹⁸⁴ were found to be common for the interactions of SA and H89 with PKA. (B) Competition between the pan-PKA inhibitor H89 and SA. (C) WB images of phosphorylation levels of PKA substrates during SA (50 μM) treatment at the indicated time points. (D) Phosphorylation of CREB (Ser¹³³), p38, and ERK during SA (50 μM) treatment at the indicated time points. (E) WB images of phosphorylation levels of HSL (Ser⁵⁶³ and Ser⁶⁶⁰) under the influence of SA treatment. Rosi: Rosiglitazone, BMP7: bone morphogenic protein 7. β-actin was used as a loading control.

On the basis of this finding, we investigated PKA signaling during SA treatment. We found that SA treatment increases PKA substrate phosphorylation at 60 min as compared to the MDI medium (Fig. 3C). We also tested the phosphorylation of other downstream kinases of PKA signaling, where we found that SA stimulates phosphorylation of CREB, p38, and ERK as well (Fig. 3D). We also observed that SA treatment promoted phosphorylation of HSL (Ser⁵⁶³ and Ser⁶⁶⁰) on day 6 during brown-adipocyte differentiation (Fig. 3E). In addition, we assessed triglyceride (TG) accumulation by ORO staining (Supplemental Fig. 1A and B).

To understand the function of SA in thermogenic gene induction, we confirmed the expression of PKA downstream genes, Ucp1, Pgc1α and Prdm16 in SA treated condition along with an activator of cAMP-dependent protein kinase, 8-bromoadenosine 3’.5’-cyclicmomophosphate (8-Br-cAMP) (Fig. 4A and 4B). Data suggested SA can effectively induce thermogenic genes. To confirm PKA signaling in SA-treated brown adipocytes, we tried to assess the effect of PKA inhibition on SA-mediated PKA substrate phosphorylation (Fig. 4C and 4D). The expression of downstream thermogenic target genes of PKA, e.g., Ucp1, Pgc1α, Pparγ, and Cebpα was diminished by H89 treatment, but this effect was reversed by cotreatment with SA (Fig. 4C). As depicted in Fig. 4D, HSL (Ser⁵⁶³ and Ser⁶⁶⁰) and CREB (Ser¹³³) phosphorylation was dramatically inhibited during H89 treatment. By contrast, during cotreatment with SA, the phosphorylation was significantly recovered.

Fig. 4.

The function of PKA in SA-mediated differentiation of brown adipocytes. (A, B) Comparison of PKA downstream targets after SA and PKA activator, 8-bromoadenosine 3’.5’-cyclicmomophosphate (8-Br-cAMP) treatment. (C, D) The expression of PKA downstream targets after SA (50 μM) and/or H89 (10 μM) treatment. (E, F) mRNA expression levels of PKA-cat-α1 and downstream target genes of PKA, e.g., Ucp1, Pgc-1α, Pparγ, and Cebpα in BAT cells (6 days of differentiation) after the knockdown of the PKA-cat-α1 gene. Data are expressed as mean ± SEM of three independent experiments. *A significant difference from group MDI (*P < 0.05, **P < 0.01, ***P < 0.001). (G) Protein expression of PKA’s downstream genes (Ucp1, Pparγ, Pgc-1α, and Hsl) and phosphorylation levels of HSL (Ser⁵⁶³ and Ser⁶⁶⁰) in BAT cells (6 days of differentiation) after the knockdown of the PKA-cat-α1 gene. (H) A schematic model of SA-mediated lipolysis and differentiation of brown adipocytes via the PKA pathway.

We next studied the activities of the PKA signaling pathway during PKA-cat-α1 gene silencing and SA treatment (Fig. 4E). As illustrated in Fig. 4E, PKA-cat-αl siRNA inhibited the mRNA expression of PKA-cat-αl effectively. As expected, the expression of the downstream genes in the PKA signaling pathway was reduced by PKA-cat-αl gene silencing, but this effect was attenuated by SA treatment (Fig. 4F and 4G). Overall, we can propose that SA contributes to brown-adipocyte activation and lipolysis via activation of PKA signaling (Fig. 4H).

DISCUSSION

Stimulation of brown adipocytes and UCP1 mediated heat production require activation of adipocytes and muscle cell surface receptors, such as Trpv1, β3-AR, Ptch1, A2aR and TrkB. These receptors require cellular signaling cascades such as PKA, PKG, Sirt1, AMPK, and p38 MAPK, cytokines (IL-4 and IL-13), and transcriptional regulators, such as Prdm family, Pgc-1α, Ppar family, and Zfp516 for UCP1 expression (16).

Several plant-extracted compounds (berberine, butein, capsaicin, and fucoxanthin), artificially synthesized compounds (e.g., a Pparγ agonist, β3-AR agonist, and salsalate), and endogenous small-molecule factors (e.g., serotonin, lactate, and adenosine) have been identified as potential BAT activators that turn on thermogenic transcriptional factors by involving their cell surface receptors or by cellular signaling cascades modulation (16). Our data revealed that SA acts as a potent stimulator of brown-adipocyte activation. The protein factors that control brown-fat differentiation were upregulated by SA treatment, including PPARγ, PGC-1α, CEBPβ, and PRDM16 along with increased expression of UCP1. mRNA and protein expression levels of a key thermogenic marker, UCP1, were increased by SA treatment here, indicating activation of mature BAT adipocytes. HSL activity appears to be regulated by site-specific phosphorylation on at least five serine residues (Ser⁵⁶³, Ser⁵⁶⁵, Ser⁶⁰⁰, Ser⁶⁵⁹, and Ser⁶⁶⁰), and PKA phosphorylates HSL at Ser⁵⁶³, Ser⁶⁵⁹, and Ser⁶⁶⁰ (17, 18). SA increases the phosphorylation of HSL (Ser⁵⁶³ and Ser⁶⁶⁰); during browning and during thermogenic activation of BAT, similar changes with increased lipolysis have been observed (19).

The differentiation and physiological functions of brown adipocytes are closely related to enhance mitochondrial biogenesis. Researcher has demonstrated that PPARγ and transcriptional coactivator PGC-1α can contribute to mitochondrial function and biogenesis (20) which are similar to our findings although further investigation is needed for molecular mechanism.

Computational identification of PKA agonist has been described elsewhere (21) where, lead compound share same binding site as SA, but described as an agonist of PKA. H89 inhibits the activation of CREB and the CREB-mediated MKP-1 induction by lipopolysaccharide (LPS) resulting from the inhibition of PKA and MSK by H89 (22). Protein kinases catalyzes the transfer of γ-phosphate of ATP to the hydroxyl group present in a tyrosine, serine, or threonine residue to phosphorylate their substrates. The direct effects of H89 on active site of the enzyme causes its inhibitory actions. It also looked likely that H89 obstructs protein kinase activity via interacting with the free enzyme, not with the enzyme–ATP complex (23). PKA mediates forskolin-induced CREB phosphorylation through p38 and MSK1 in NIH 3T3 cells (24). In our study, H89 treatment inhibited PKA phosphorylation resulting lower effect of SA on HSL phosphorylation at PKA target sites, Ser⁵⁶³ and Ser⁶⁶⁰; our data are in agreement with the findings of other researchers (25). SA and H89 both can bind competitively to PKA by sharing the ATP-binding site on a PKA catalytic subunit, as we showed by our molecular docking analysis and confirmed by western blotting. Our data indicate that the expression of target genes of PKA, e.g., Ucp1, Pgc1α, Pparγ, and Cebpβ, was reduced by H89 treatment, but this suppression was reversed by SA treatment. Besides, SA promoted phosphorylation of CREB, p38, and ERK; CREB phosphorylation was inhibited by H89 treatment, and this change was attenuated by SA treatment. In this context, we propose that SA-mediated PKA activation can induce both phosphorylation of its downstream targets and expression of browning-related genes although a fine-tuned study is needed to confirm this crosstalk.

In conclusion, this study revealed that SA not only activates the thermogenic program but also induces lipolysis. SA-mediated thermogenic-signature upregulation is probably involved in the activation of UCP1 via PKA/CREB signaling. Our findings may open up a new avenue of research on the downstream thermogenic program in in vivo models.

MATERIALS AND METHODS

Chemicals, reagents, and antibodies

SA (purity ≥ 98.0%), insulin, dexamethasone, 3-isobutyl-1-methylxanthine (IBMX), rosiglitazone (Rosi), 8-bromoadenosine 3’.5’-cyclicmomophosphate (8-Br-cAMP), Oil Red O dye (ORO), and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) were purchased from Sigma-Aldrich (St. Louis, MO, USA), and human recombinant BMP7 from R & D Systems (Minneapolis, MN, USA). Fetal bovine serum (FBS) and High-glucose Dulbecco’s modified Eagle’s medium (DMEM) were bought from Atlas Biologicals (Fort Collins, CO, USA), and a penicillin-streptomycin solution from Hyclone Laboratories, Inc. (South Logan, NY, USA). Antibodies against UCP1, PGC-1α, PRDM16, and β-actin were purchased from Abcam (Cambridge, MA, USA), whereas antibodies against phospho-HSL (Ser⁵⁶³ and Ser⁶⁶⁰) and HSL from Cell Signaling Technology (Danvers, MA, USA).

Cell maintenance and differentiation

BAT cells were cultured and maintained in the DMEM GlutaMax medium supplemented with 10% of FBS and 1% of the penicillin-streptomycin solution and were kept at 37℃ in a 5% CO₂ incubator. BAT cells are an immortalized cell line created by Lee BS group (26). The cells were differentiated as described elsewhere (27). Cells after 6 days of differentiation were used in all the experiments unless stated otherwise.

Cell treatment procedures

Fully confluent BAT cells were treated with the MDI differentiation induction medium consisting of 0.5 μM 3-isobutyl-1-methylxanthine, 1 μM dexamethasone, and 10 μg/ml insulin in DMEM supplemented with 10% of FBS, followed by a maturation medium composed of DMEM, 10% of FBS, and 10 mg/ml insulin. The media were refreshed every other day. The following treatment groups were set up: MDI, Rosi with MDI (positive control), BMP7 with MDI (positive control), and SA with MDI. Treatments exceeding 2 days were continued until day 4 by mixing drugs with the maturation medium. After day 4, only the maturation medium was used until cell harvesting. Preadipocytes were maintained only in the culture medium (DMEM and 10% of FBS).

Cell viability assay

In 96-well plates, BAT cells were seeded at 80% to 90% confluence. Cell viability was evaluated by MTT assay as described elsewhere (15).

Analysis of mitochondrial DNA content

Mitochondrial biogenesis was quantified by real-time qPCR as described elsewhere (28) assuming that the ratio of mitochondrial DNA (mtDNA) to nuclear DNA (nDNA) increases but nDNA remains constant.

Oxygen consumption assay

BAT cells were seeded in a 96-well plate at a density of approximately 40,000-60,000 cells per well. Oxygen consumption assay was performed as kit protocol (Cayman, Ann Arbor, MI, USA).

Quantitative reverse-transcription PCR (qRT-PCR) analyses

Total-RNA extraction from (and qRT-PCR analyses of) BAT cells were performed as described previously (29). The primer sequences employed in this study are listed in Supplemental Table 1. Expression of target genes was normalized to that of TATA box–binding protein (Tbp).

Preparation of whole-cell extracts for western blot (WB) analyses

Whole-cell extracts from BAT cells were prepared as described elsewhere (30) with a minor modification. Briefly, cell extracts were collected after 6 days of differentiation. Next, 1% BSA was used for blocking the membranes instead of skim milk (because of analysis of phosphoproteins), and a phosphatase inhibitor cocktail (Sigma-Aldrich) was added into RIPA lysis buffer (Santa Cruz Biotechnology, Inc.).

Gene silencing experiments

Gene silencing assay was conducted as described previously (31).

In silico analysis

Western blot analysis revealed that SA upregulates phosphorylation of PKA in the absence of inhibitor H89 and attenuates H89-mediated dephosphorylation. To understand the mechanism better, molecular-docking–based structural analysis was performed. First, crystal structure of cAMP-dependent protein kinase (PDB ID: 1BX6) with a potent inhibitor, balanol (a natural product), bound to its catalytic subunit was obtained from Protein Data Bank (www.rcsb.org). After that, the ligand was removed, and a binding grid of (40 × 40 × 40, 1 Å) size was generated containing all possible active sites on PKA surface. Before docking, all the structures were prepared with AutoDock Tools (32), polar hydrogen was added, and finally molecular docking was performed in the AutoDock Vina software (33).

Statistics

Data are representative of three or more experiments and are shown as mean ± standard error of the mean (SEM). Student’s t test was conducted to identify significant changes between a control group and various treatment groups. Data with a P value of < 0.05 were considered statistically significant.