Ultrasound triggered topical delivery of Bmp7 mRNA for white fat browning induction via engineered smart exosomes

Background Efficient and topical delivery of drugs is essential for maximized efficacy and minimized toxicity. In this study, we aimed to design an exosome-based drug delivery platform endowed with the ability of escaping from phagocytosis at non-target organs and controllably releasing drugs at targeted location. Results The swtichable stealth coat CP05-TK-mPEG was synthesized and anchored onto exosomes through the interaction between peptide CP05 and exosomal surface marker CD63. Chlorin e6 (Ce6) was loaded into exosomes by direct incubation. Controllable removal of PEG could be achieved by breaking thioketal (TK) through reactive oxygen species (ROS), which was produced by Ce6 under ultrasound irradiation. The whole platform was called SmartExo. The stealth effects were analyzed in RAW264.7 cells and C57BL/6 mice via tracing the exosomes. To confirm the efficacy of the engineered smart exosomes, Bone morphogenetic protein 7 (Bmp7) mRNA was encapsulated into exosomes by transfection of overexpressing plasmid, followed by stealth coating, with the exosomes designated as SmartExo@Bmp7. Therapeutic advantages of SmartExo@Bmp7 were proved by targeted delivering Bmp7 mRNA to omental adipose tissue (OAT) of obese C57BL/6 mice for browning induction. SmartExo platform was successfully constructed without changing the basic characteristics of exosomes. The engineered exosomes effectively escaped from the phagocytosis by RAW264.7 and non-target organs. In addition, the SmartExo could be uptaken locally on-demand by ultrasound mediated removal of the stealth coat. Compared with control exosomes, SmartExo@Bmp7 effectively delivered Bmp7 mRNA into OAT upon ultrasound irradiation, and induced OAT browning, as evidenced by the histology of OAT and increased expression of uncoupling protein 1 (Ucp1). Conclusions The proposed SmartExo-based delivery platform, which minimizes side effects and maximizing drug efficacy, offers a novel safe and efficient approach for targeted drug delivery. As a proof, the SmartExo@Bmp7 induced local white adipose tissue browning, and it would be a promising strategy for anti-obesity therapy. Graphical Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s12951-021-01145-3.


Background
Exosomes are a type of nanovesicles, 30-150 nm in diameter, that have the same topology as the parental cells and are enriched in proteins, lipids, nucleic acids and glycoconjugates [1]. In recent years, exosomes have emerged as a promising platform in drug delivering system. Despite the significant advantages in low immunogenicity, biocompatibility, ability of targeting different tissues with their surface ligands, etc., most of the exosomes have been found distributed into liver, spleen and other tissues with mononuclear phagocyte system (MPS), limiting their therapeutic efficacies [2][3][4][5][6].
Intensive efforts have focused on fusing targeting peptides and other moieties to exosomal surface to improve the delivering efficiency [7][8][9]. However, those strategies are challenged as some targeting ligands are refractory to proper expression and prone to degradation [10]. Alternatively, the targeting delivery efficiency is significantly rely on overexpression of targeting peptides, which can be difficult to control and may lead to massive side effects [7][8][9]. In addition, unlike the synthetic materials, decorated exosomes cannot thoroughly escape from MPS since its good biocompatibility [7][8][9]. Therefore, to realize the controllable and targeting delivery of therapeutic molecules using exosomes as drug carriers, a smart enough exosome capable of escaping of MPS is urgently required.
It is well known that the hydrophilic polymer polyethylene glycol (PEG) corona, i.e., PEGylation, is an efficient strategy to protect drug carriers from aggregation, opsonization, and phagocytosis, consequently prolonging their circulation time and decreasing collateral damage to healthy tissues [11,12]. It has been reported that reactive oxygen species (ROS) could readily cleave thioketal (TK) bonds hence expose the drug carrier in core by removing the TK linked corona [13]. Herein, we propose a smart exosomes-based delivery system constructed with a novel conjugate designated as CP05-TK-mPEG and a sonosensitizer chlorin e6 (Ce6). In this system, CP05 is the peptide, CRHSQMTVTSRL, with potent capacity bound to tetraspinin CD63 [14], allowing PEG anchored onto exosomes. Moreover, Ce6 was loaded in exosomes through direct incubation. The smart exosomes could escape from phagocytosis and locally deliver drugs ondemand by ultrasound.
As a proof-of-concept study, we confirmed that smart exosome-based system could deliver Bmp7 locally and induce the local omental adipose tissue (OAT) browning efficiently, serving as a promising strategy for anti-obesity therapy.

Identification and characterization of exosomes
CP05-TK-mPEG was anchored onto exosomes by direct incubation (Fig. 1A).To characterize CP05-TK-mPEG functionalized exosomes (Exo CP05−TK−mPEG ), morphology analysis through transmission electron microscope, particle size distribution analysis by dynamic light scattering and exosomal marker expression by western blot were applied. Transmission electron microscopy verified the shape and size of Exo CP05−TK−mPEG were nearly the same as Exo (Fig. 1B). Particle size analysis showed that the size of Exo CP05−TK−mPEG was slightly larger than the parental Exo, with whose size fell into the diameter range of 30-160 nm (Fig. 1C). As expected, CP05-TK-mPEG functionalization didn't change the expression profile of the inclusive markers (CD9, TSG101 and CD63) and exclusive marker GM130 (Fig. 1D).

Evidence of Escaping from Phagocytosis of Exo CP05−TK− mPEG in vitro and in vivo
In order to determine the ability of escaping from phagocytosis, Exo and Exo CP05−TK−mPEG were incubated with cells for different hours to observe the phagocytosis ( Fig. 2A). Compared to the control group, there were much less Exo CP05−TK−mPEG uptaken in different incubation period (Fig. 2B, C), suggesting that Exo CP05−TK−mPEG effectively escaped from endocytosing.
To visualize and compare the organ distribution of Exo CP05−TK−mPEG , Exo and Exo CP05−TK−mPEG were labeled with fluorescent dye under dark conditions and then injected into mice (Fig. 2D). As expected, Exo CP05−TK− mPEG were less concentrated in major organs and tissue, including heart, liver, spleen, lung, kidney, gut and abdominal adipose tissue, while more concentrated in circulating blood compared to natural Exo (Fig. 2E, F). In addition, confocal fluorescence imaging also evidenced the ability of escaping from phagocytosis of Exo CP05−TK− mPEG in above organs (Additional file 1: Fig. S1). Notably, Exo CP05−TK−mPEG has just mild ability of escaping from  (Fig. 2G). For further evidence, Exo and Exo CP05−TK−mPEG were electroporated with cel-miR-54 (Fig. 2D) followed by in vivo injection. qPCR analysis revealed that the expression level of cel-miR-54 is significantly lower in Exo CP05−TK− mPEG than that in Exo in most of the observed organs (Additional file 1: Fig. S2), which was consistent with fluorescence tracking results.

Construction of SmartExo and ultrasound triggered local removal of PEG corona
The above data suggest that Exo CP05−TK−mPEG prolongs the circulation time in blood while removal of the PEG is prerequisite for boost local delivery. To this end, Exo were incubated with Ce6 followed by incubation with CP05-TK-mPEG, with the exosome system designated as SmartExo. SmartExo was then DiI-labeled and incubated with RAW264.7 cells for 30 min. Ultrasound was applied to produce ROS by the sonodymamic effects of Ce6, which would theoretically cleave the TK bond and free exosomes from PEG corona (Fig. 3A). The confocal images showed that under ultrasound on (US on) condition, the effect of phagocytosis was much stronger than that in ultrasound off (US off) condition (Fig. 3B). Moreover, in vivo experiments showed that ultrasound irradiation significantly improved the delivery of SmartExo in adipose tissue ( Fig. 3C-E). Cel-miR-54 tracing by qPCR analysis further confirmed the findings (Additional file 1: Fig. S3).

Construction of SmartExo@Bmp7 for targeted delivery of Bmp7 mRNA
Currently, the whole world is struggling with a pandemic of obesity and seeking for efficient therapeutic strategies for its treatment is a constant need [15]. Researchers have found that the amount of brown adipose tissue (BAT) is positively in correlation with the energy expenditure and is significantly lower than in slim adults [16]. Among the strategies aimed at induction of BAT, the bone morphology protein 7 (Bmp7) is an important inducer of brown adipocyte differentiation [17,18]. Both in vitro and in vivo studies have evidenced that Bmp7 could induce fully differentiated brown adipocytes with high uncoupling protein 1 (Ucp-1) expression [19,20]. In the following experiments, we aimed to deliver Bmp7 mRNA with the above SmartExo system. Breifly, plasmid overexpressing Bmp7 mRNA was transfected in 293T cells, and after 48 h exosomes loaded with Bmp7 (Exo@Bmp7) were isolated by ultracentrifugation. Then Exo@Bmp7 was functionalized to SmartExo@Bmp7 using the method constructed above (Fig. 4A). For verification of delivery specificity in abdominal adipose tissue, Exo@Bmp7 and SmartExo@Bmp7 were injected into mice via tail followed by ultrasound induction on abdominal adipose tissue, and qPCR analysis of Bmp7 mRNA expression of various organs was applied (Fig. 4B). As expected, Bmp7 mRNA was efficiently loaded into exosomes while Smart-Exo engineering had no obvious effects on the abundance (Fig. 4C). In addition, relative expression of Bmp7 in adipose tissue was significantly increased, while that in other organs were slightly decreased or not significant (Fig. 4D). Accordingly, there was a significant increase of BMP7 protein expression in adipose tissue of mice treated with SmartExo@Bmp7 (Additional file 1: Fig. S4).

Smart exosome-based system applied in Bmp7 targeting delivery for OAT browning
In view of the above data, we aimed to explore whether SmartExo@Bmp7 induce adipose browning in obese mice. Mice were fed in high-fat-diet for induction of obesity. After 8 weeks of feeding, models of obesity were successfully built considering the weight was at least 20% more than the normal-chow-diet ones. The obesity mice were randomly allocated to PBS, Exo, Exo@ Bmp7 or SmartExo@Bmp7 treatment group. The exosome treatment was performed one every three day for 3 weeks. Ultrasound irradiation was conducted once every 1 h in the first 6 h after tail vein injection (Fig. 5A). It was evidenced that average body weight of the group treated with SmartExo@Bmp7 significantly dropped, while the Exo@Bmp7 treated group slightly dropped, in comparison with other groups (Fig. 5B). Consistently, the weight of OAT in group treated with SmartExo@Bmp7 The nuclei were counter-stained with Hoechst (blue). DiI-labeled exosomes were added to RAW264.7 cells and ultrasound irradiation was applied or not applied. The relative fluorescence intensity was shown at the right side. C Decorated exosomes were loaded with cel-miR-54 to evidence the phagocytosis at target organ with or without ultrasound irradiation, and ultrasound irradiation was applied once a hour in the first 6 h after injection. D Representative confocal images of DiI-labeled exosomes (red) in abdominal adipose tissue of mice with or without ultrasound irradiation. The average radiant efficiency is at right side. E Representative fluorescence showing the phagocytosis of DiR-labeled exosomes by abdominal adipose tissue with or without ultrasound irradiation. Relative fluorescence intensity of confocal images was shown at the right side was much lower than all other three groups (Additional file 1: Fig. S5). Histology examination of OAT revealed prominent decrease of adipocyte size after SmartExo@ Bmp7 treatment (Fig. 5C, D). Accordingly, the relative of expression of Ucp1 in OAT of SmartExo@Bmp7 treated mice was significantly higher than other groups (Fig. 5E). Immunohistochemical staining of adipose tissue for Ucp1 further confirmed that SmartExo@Bmp7 induced browning of OAT (Fig. 5F).

Discussion
We here developed a smart exosome-based drug delivery strategy by coating CP05-TK conjugated PEG to the surface of exosomes. In the meantime, sonosensitizer Ce6 was loaded into exosomes. The resultant exosomes (SmartExo) successfully prolonged the circulation time in blood shown as escaping phagocytosis from major organs. Moreover, since Ce6 was loaded in those exosomes, with the intervention of ultrasound at the location of target organ, the redox reaction between ROS, which was produced by Ce6, and TK freed exosomes from PEG corona so that it can be eaten off properly by the target organ. In vitro and in vivo experiments revealed that the system exerts the ability of escaping from phagocytosis at non-target sites and controllable releasing drugs at right organs. Last but not the least, using SmartExo system, with irradiation of ultrasound at abdominal region, Bmp7 was successfully delivered on site which was shown as OAT browning. The results showed that SmartExo system was a great prospect of targeting delivering strategy, and its efficiency was proved by targeting delivery of Bmp7 on OAT for obesity treatment. Natural exosomes are of poor cell-or tissue-specific targeting features [8]. Thus, the therapeutic application of exosomes as a delivery platform necessitates the design of controlled and targeted exosomes to promote therapeutic efficacy [7,21,22]. Previous researches showed that the targeting function of exosomes can be achieved by either parental cell-based engineering or direct exosome engineering, displaying specific proteins or RNAs on the surface or into the lumen of exosomes so that it can be identified by target organs or tissue [4,22,23]. Although the targeting peptides or proteins of exosomes could deliver molecules to specific cells, it is not avoidable that phagocytosis happened on non-target organ [24]. Therefore, a prospective exosome-based delivery platform would be capable of both evasion of phagocytosis on non-target sites and controllable release drugs in desirable places.
Researchers have found that PEG coatings on nanoparticles, i.e., PEGylation, shielded the surface against aggregation, opsonization and phagocytosis, and it was a common approach for improve the circulation time of nanoparticles in vivo [11,12]. Combining the requirement of targeting delivery, we thought a switchable PEGylation of exosomes would hold great promise for shielding them from interaction with cells and improving drug delivery efficiency. Recently, a particular peptide, CP05, was found of capable of direct painting of exosomes irrespective of the origin of exosomes through specifically binding to the second extracellular loop of CD63 [14]. In the light of that, we thought PEGylation of exosomes could be accomplished by construction of a conjugate in which PEG corona wrapped exosomes through CP05 anchored on CD63. In order to get rid of the PEG protection allowing phagocytosis of exosomes at target organs, ROS-sensitive covalent bonds can be introduced as a part of conjugate [25]. In our study, thioketal was used as a link between CP05 and PEG forming CP05-TK-mPEG. Under the assistance of photosensitizers or sonosensitizers, lights or ultrasound generates ROS, which readily cleave thioketal. In fact, thioketal could be replaced with phenylboronic ester, diselenide bonds, or any other ROSsensitive non-cytotoxic covalent bonds if it is possible [26].
Both light and ultrasound irradiation can trigger the respective sensitizers producing ROS. Although the sonodynamic therapeutic is developed from photodynamic therapeutic, ultrasound is more applicable than laser [27]. Unlike laser, ultrasound trigger sound-sensitizer by sound waves under certain wave length with deeper penetration [27,28]. According to previous researches, ultrasound is capable of penetrating into deep tissue, about tens of centimeters, allowing treatment with deep nidus or interaction with sonosensitizer in deep site [29]. In addition, ultrasound treatment has good repeatability since it is nonradiative and costs lower than laser application [27]. As a representative photo/sonosensitizer, Ce6 has been extensively used in photo/sonodynamic treatment due to its high singlet oxygen generation efficiency [30]. In our study, Ce6 was chosen as the sonosensitizer loaded in exosomes and triggered by ultrasound producing ROS in order to cleave TK.
As far as we have known, current drugs for weight loss approved by the FDA focus primarily on the reduction of energy intake, and few of them provides adequate long-term clinical efficacy [31]. BAT is essential for thermogenesis and energy balance in mammals, and in vivo experiments' results revealed its promotion for energy expenditure, reduction of adiposity, and protecting mice from diet-induced obesity [15,32,33]. According to previous studies, the mesodermal growth factor Bmp7 can stimulate BAT activation and promote the energy expenditure reducing weight [19,34]. However, given the fact that Bmp7 is also recognized as a cytokine with pleiotropic functions being offered as a treatment for cardiovascular, fibrotic, metabolic and neurodegenerative diseases, systematic administration of Bmp7 could bring unexpected side effects [35][36][37][38][39]. In our study, Bmp7 was loaded in SmartExo system and induced at topical OAT assisted by ultrasound. Since the exosomes derived from HEK293T cells were reported as having high biocompatible and low immunogenicity, the Bmp7 engineered exosomes induced no significant inflammatory response (Data not shown) [40]. This strategy guarantees the browning of local white adipose tissue with little leakage of Bmp7 to other organs. This newly smart exosome-based delivering strategy of Bmp7 could provide new directions in the development of therapeutics for obesity.

Conclusions
In summary, we here have established an ultrasoundassisted and smart exosome-based delivery system, namely SmartExo, which delivers the drugs to the local area. In contrast to most drug targeting delivery mechanism which concentrate on expressing of various peptide or moieties that can vary between tissues or organs, SmartExo provides a controllable delivery platform based on ultrasound, regardless of the type or targeting location, thus offering a robust and broadly applicable targeting strategy. With further optimization, the Smart-Exo-based targeting delivery strategy described here might be applied in the context wherever prevention of non-target phagocytosis and promotion of drug releasing at target organs are badly needed.

Plasmid construction
Bmp7 cDNA was synthesized in Genscript and subcloned into pWPI vector using PacI and BstB1, with the resultant correct clone designated as pWPI-Bmp7. All the clones were confirmed by sequencing and the right clones were stored at − 80 °C for following application.

Transfection
In order to load Bmp7 mRNA into exosomes, HEK293T cells were transfected with the corresponding plasmid with Lipofectamine 2000 (Invitrogen) following standard instruction. In brief, culture medium was replaced with antibiotic-free DMEM medium when cell density grew to 60-70% in 100-mm culture dish. Then, 20 µL Lipofectamine 2000 and 10 µg plasmid were incubated separately in two centrifuge tubes containing 500 µL of DMEM medium at room temperature for 5 min. After that, contents of two centrifuge tubes were mixed and incubated at room temperature for 20 min. The above mixture was added to HEK293T cell culture dish mentioned before.

Exosome isolation and purification
The HEK293T cells were used as the exosome donor cells in this study. Cell culture supernatants were replaced with serum-free medium at least 48 h before isolation process. The supernatants were collected and differentially centrifuged (250g for 10 min at 4 °C to eliminate dead cells and 10,000g for 15 min at 4 °C to remove residual cellular debris) followed by filtration (0.22 μm filter; Milipore-Sigma, Bilicera, MA). Next, the resulting supernatant was centrifuged at 100,000g and 4 ℃ for 2 h to obtain the exosomes. Finally, the extracted exosomes were re-suspended in PBS and stored at − 80 °C before using.

Transmission electron microscopy
The morphology of isolated exosomes was analyzed by transmission electron microscopy. Briefly, the exosome suspension was added onto the copper mesh, then the exosomes were stained with 2% uranyl acetate for 1 min and imaged by the electron microscope (JEM-2000EX TEM, JEOL Ltd, Tokyo, Japan).

Particle size analysis
For measure the size of distribution, the isolated exosomes were diluted to 1 µg/mL and the measurement was performed and analyzed using NanoPlus (Otsuka Electronics, Japan). Measurements were repeated at least three times for each sample.

Determination of concentration of exosomes by BCA method
The concentration of suspended exosomes was determined by total protein concentration. A small aliquot of exosomes was dissolved RIPA Lysis Buffer. Determination of exosomal protein concentration was performed by Pierce BCA Protein Assay Kit (Thermo, USA) under standard instruction [7][8][9].

Synthesis of CP05-TK-mPEG
About 100 mg of Maleimide-Thioketal-PEG2000 (Mal-TK-mPEG) were dissolved in 5 mL N,N-dimethylformamide (DMF), then, 1.1 eq of the peptide CRHSQMTVTSRL (CP05) was added to the solution and incubated at room temperature for 12 h. After that, the mixture was loaded into dialysis tubing with molecular cut off at 3500 Da, and the dialyse in purified water for 24 h. Finally, the dialysate was collected and the final product CP05-TK-mPEG was obtained using freeze-drying method.

Construction of SmartExo
Ce6 (diluted in DMSO, 30 mg/mL ) were encapsulated into exosomes (1 µg/µL ) through electroporation. Then, Ce6 loaded exosomes were isolated and resuspended with CP05-TK-mPEG solutions (1 mg CP05-TK-mPEG powder dissolved in 1 mL PBS) and incubated at 4 °C overnight. The decorated exosomes were isolated and washed twice using PBS to remove residuals, and the products were resuspended in PBS as SmartExo.

In vitro tracing of exosomes
Exosomes (about 1 µg/µL of protein concentration) were labeled with DiI o by incubation with the dye (1 mM) at the ratio of 500:1 for 30 min in 37 °C dark room, followed by exosome isolation as described above. RAW 264.7 cells were incubated with DiI-labeled exosomes for 4 h, 8 h, and 12 h. For those were incubated with DiI-labeled SmartExo, ultrasound was applied after 40 min of incubation. The ultrasound irradiation intensity was set as 0.1 W/cm 2 with a duty cycle of 20%, and the irradiation time was 30 s (Additional file 1: Table S1). The cells were then washed with PBS for three times and fixed with 4% paraformaldehyde for 10 min and again washed with PBS twice. Cell nuclei were counter-stained with 1 µg/ mL Hoechst 33,342 (Beyotime, China) diluted at 1:1000 for 10 min in 37 °C dark room then washed with sodium acetate solution to remove the nonspecific adhesion. The prepared samples were observed and captured by Confocal Microscope (Eclipse C2, Nikon, Japan).

Animal experiments
Healthy male C57/BL6J mice aged 8 weeks were purchased from Animal Center of the Fourth Military Medical University. The animal experimental and housing procedures were performed in accordance with the protocols of Animal Experimentation and Ethics Committee of Fourth Military Medical University.

In vivo fluorescence tracing of exosomes
Exosomes (about 1 µg/µL of protein concentration) were labeled with DiI or DiR by incubation with the dye (1 mM) at the ratio of 500:1 for 30 min in 37 °C dark room. Organ distribution analysis of exosomes was performed as previously described. Briefly, DiRlabeled exosomes were washed with PBS to remove free dyes and tail-vein injected into mice. The dosages of exosomes treatment vary from study to study, and 4 µg/g of exosomes per mouse for each time was chosen in our study, which is similar as the dose in published literatures [21]. For those were injected with SmartExo, ultrasound was applied on the abdominal area once every hour in the first 6 h after injection. For in vivo experiments, the irradiation intensity was set as 2 W/cm 2 , and the duration was 180 s for each irradiation (Additional file 1: Table S1). The distribution of exosomes in organs was detected 8 h after injection using an IVIS Lumina II in vivo imaging system (Perki-nElmer, Thermo Fisher, US).
For test of the sliced section, exosomes labeled with fluorescent dye DiI (Beyotime, China) were administered into mice intravenously. Eight hours after the injection or ultrasound, the fresh organs were harvested, then washed with PBS and fixed in 4% paraformaldehyde for 30 min before being embedded in optimal cutting temperature compound. Then, the embedded organs were sliced into 10 μm sections on a freezing microtome (Leica, Germany). Tissue sections were stained with 1 µg/mL Hoechst 33,342 (Beyotime, China) diluted at 1:1000 for 10 min then washed with PBS before sealing with neutral gum under dark conditions. The prepared samples were observed and captured by Confocal Microscope (Eclipse C2, Nikon, Japan) for exosomes tracking.

Anti-obesity therapy
All mice were fed with high-fat-diet and body weights of mice were monitored weekly. At the 8th week, mice fed a high-fat-diet developed obesity (weighing at least 20% more than mice fed with normal chow diet). All animals were randomly allocated to four groups (four mice for each group) to start treatments for obesity once three days for totally 3 weeks. The first group was treated with physiological saline solution (PBS), the second group was treated with natural exosomes (Exo, the third group was treated with Bmp7 loaded exosomes (Exo@Bmp7) and the last group was treated with Bmp7 loaded smart exosome-based drug delivery system (SmartExo@Bmp7). The dosage of exosomes in each group was determined as 4 µg/g per mouse as mentioned above. Ultrasound irradiation was applied on the abdominal area once every hour in the first 6 h after injection for each group in order to remove as much as PEG corona from SmartExo. The parameters were set same as described above (Additional file 1: Table S1). All animals were sacrificed after 3 weeks of treatments, and abdominal fat were obtained to evaluation the browning situation.

RNA isolation and qRT-PCR
Total RNA of the sample was extracted using TRIzol (Invitrogen, 15596018) followed the standard instructions. For mRNA, 2 µg of RNA was reverse transcribed into cDNA by SMART ® MMLV Reverse Transcriptase (Takara). While, for miRNA, 2 µg of RNA was reverse transcribed into cDNA using miRcute Plus miRNA qPCR detection kit (Tiangen). Experiment of qPCR was performed with FastStart Essential DNA Green Master Kit (06924204001, Roche). Relative expression of mRNA and miRNA was respectively normalized to GAPDH/Gapdh and U6 levels and calculated by 2 −ΔΔCt . The PCR primer sequences are supplied in Additional file 1: Table S2.

Histology
OAT were excised and fixed in 10% formalin. Then they were paraffin-embedded and sectioned prior to H&E staining or immunohistochemistry for Ucp1 (Abcam; 1:400) detection. Signal was detected using the Vector ABC Elite kit.

Statistical analysis
Data and results are expressed as mean ± SEM as indicated. Student's t-test was used for comparison between two groups (Graphpad Prism 7.0). P values of < 0.05 indicate statistical difference.