Canagliflozin helps fat cells change their energy structures through the AMPK-Sirt1-Pgc-1α pathway

C57BL/6 J male mice (8 weeks old; n = 6) were treated with Cana (Meilunbio®, Dalian, China) or equal amounts of starch (as vehicle group) for 8 weeks (0.03% w/w Cana was mixed with powdered chow diet). On the basis of daily food intake and body weight, 0.03% (w/w) was approximately equivalent to 20 ~ 30 mg/kg [16]. This dosage is known to inhibit SGLTs efficiently [16]. All mice were maintained in ventilated cages with a 12 h light/dark cycle and free access to food and water. CO₂ inhalation euthanasia was used to euthanize the mice. Kidney and subcutaneous adipose tissue were dissected and divided into portions for mRNA or protein analysis. All animal procedures were performed in accordance with the Animal Care and Use Committee of Sichuan University.

Stromal-vascular fractions (SVFs) were isolated from subcutaneous adipose tissue as described [24]. In brief, subcutaneous adipose tissue was dissected from 4- to 6-week-old C57BL/6 J mice, then digested for 30 min at 37°C in isolation buffer (containing 1 mg/ml collagenase, 0.123 M NaCl, 5 mM KCl, 1.3 mM CaCl₂, 5 mM glucose, 100 mM HEPES, and 4% BSA), and centrifuged at 700 g for 10 min at room temperature. Cell pellets were washed twice and resuspended in DMEM/F12 medium containing 10% foetal bovine serum (FBS) (Gibco, NY, USA). For differentiation, cells at 100% confluence (Day 0) were exposed to induction medium containing 10% FBS, 0.5 mM isobutylmethylxanthine (Sigma-Aldrich, MO, USA), 1 μM dexamethasone (Sigma-Aldrich, MO, USA), 1 μM rosiglitazone (Meilunbio®, Dalian, China), 1 nM T3 (Sigma-Aldrich, MO, USA) and 10 μg/ml insulin (Sigma-Aldrich, MO, USA). Two days after induction (from Day 3), cells were switched to maintenance medium containing 10% FBS, 1 nM T3, 10 μg/ml insulin until they were ready for harvesting (generally days 6–7 post differentiation). Differentiated primary adipocytes were treated with 10 μM Cana (Meilunbio®, Dalian, China) or drug vehicle (DMSO, Sigma-Aldrich, MO, USA) for 48 h and collected for further analysis. All group data subjected to statistical analysis were repeated in at least three independent experiments, each in duplicate or triplicate.

Differentiated primary adipocytes were treated with 10 μM Cana or drug vehicle (DMSO) for 48 h and collected for further analysis. For mitochondrial labelling, differentiated adipocytes were treated with 10 μM Cana or vehicle for 48 h and labelled with a mitochondrial probe. Following protocols recommended by the manufacturer, Mito Tracker Red CMXRos (200 nM) (Thermo Fisher Scientific, MA, USA) was used to stain mitochondria for 30 min at 37°C, then cell nucleus was stained with DAPI (Meilunbio®, Dalian, China). The fluorescence images were taken by confocal laser scanning microscopy (CLSM) (Leica, Wetzlar, Germany). The red fluorescence intensity quantification was detected (excitation wavelength of 579 nm and emission wavelength at 599 nm) using a Bio-Tek fluorescent microplate reader (BioTek Instruments, Inc.). For immunofluorescence, TOMM20 primary antibody was used at a dilution of 1:200 and incubated overnight at 4°C, then adipocytes were incubated with secondary anti-rabbit Alexa Fluor 488 antibody for 1 h at room temperature, and incubated with DAPI dye for 2 min. The fluorescence images were taken by confocal laser scanning microscopy. All representative images were repeated in at least three independent experiments.

For adenoviral infection, differentiated adipocytes were infected with control or (Ad)-shPgc-1α adenovirus (10 μl of 2 × 10¹⁰ Pfu) for 48 h, and then treated with 10 μM Cana or vehicle for 48 h. All data obtained from cell samples were repeated in at least three independent experiments, each in triplicate.

Total RNA of tissue and adipocytes were extracted with Trizol (Invitrogen, CA, USA). Reverse transcription of total RNA (1 μg) was performed with a high-capacity cDNA reverse transcription kit (Takara, Shiga, Japan). Amplification reactions were conducted using a SYBR Green qRT-PCR Kit (Bio-Rad Laboratories, CA, USA) with the Bio-RAD CFX 96 PCR system [25]. The amplification of each gene was verified by melting points. All primer sequences are listed in Table S1.

Total DNA of adipocyte was extracted by conventional phenol-chloroform method as described [24]. DNA concentration was assessed with a Nanodrop 2000 (Thermo Fisher Scientific, MA, USA). Mitochondria DNA (mtDNA) was amplified by using primers specific for the mitochondrial cytochrome oxidase subunit 2 (Cox 2), a marker of mitochondrial copy number [26], then normalized to genomic DNA coded by 18s. Primer sequences for Cox 2 and 18s were as follows: COX 2: forward 5ʹ-ATAACCGAGTCGTTCTGCCAAT-3ʹ and reverse 5ʹ-TTTCAGAGCATTGGCCATAGAA-3ʹ; and 18s: forward 5ʹ- TGTGTTAGGGGACTGGT GGACA −3ʹ and reverse 5ʹ- CATCACCCACTTACCCCCAAAA-3ʹ. mtDNA copy number relative to genomic DNA content was quantitatively analysed with the Bio-RAD CFX 96 PCR system.

Primary fat SVF cells from subcutaneous adipose tissue were plated in an XFp-well microplate (Seahorse Bioscience, MA, USA), and differentiated for 8 d and treated with vehicle (DMSO) or 10 μM Cana for 48 h. Oxygen consumption rate (OCR) was measured at 37°C by using an XFp analyser (Seahorse Bioscience, MA, USA) in accordance with the manufacturer’s instructions. In total, 1 μM oligomycin (Seahorse Bioscience, MA, USA), 0.5 μM carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) (Seahorse Bioscience, MA, USA) and 0.5 μM rotenone/antimycin A (Rot/AA) (Seahorse Bioscience, MA, USA) were delivered to detect basal respiration, uncoupled respiration, maximal respiration and non-mitochondrial respiration, respectively. Relative OCR is calculated by subtracting OCR measured after antimycin addition from basal OCR, from OCR after oligomycin addition, or from OCR after FCCP addition.

Western blotting analyses of tissue and adipocytes were performed as described [25]. Total protein was extracted from cells using RIPA buffer (Sigma-Aldrich, MO, USA), protease inhibitor cocktail and phosphatase inhibitor cocktail (Thermo Fisher Scientific, MA, USA). Protein concentration was measured using a commercially available assay (Thermo Fisher Scientific, MA, USA) according to the manufacturer’s protocol. Extracted proteins were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to polyvinylidene fluoride membranes (Millipore, MA, USA). Primary antibodies for western blotting analysis are listed in Table S2. All immunoreactive bands were visualized and analysed by densitometric scanning and Image J software. All representative images were repeated in at least three independent experiments.

Student t-test was used for comparing two groups. One-way ANOVA followed by the Dunnett post hoc test was used for multiple comparisons versus the control group (GraphPad Software). p < 0.05 was considered statistically significant. Data are shown as mean ± SEM.

To determine the direct effect of Cana on adipose tissue, we isolated mesenchymal stem cells from subcutaneous adipose tissue of C57BL/6 J mice, differentiated into mature adipocyte, and then treated with Cana or DMSO for 48 h (Figure 1(a)). The concentration of Cana was justified according to the peak plasma concentrations achieved following therapeutic doses in humans [27]. Oil red O staining results showed that more than 90% of cells were differentiated in both groups (Figure 1(b)). No significant differences were observed in the levels of fatty acid binding protein 4 (Fabp4) and lipoprotein lipase (Lpl), two fat cell-specific markers, between Cana-treated and vehicle groups (Figure 1(c)). Given the previously reported role of SGLT2 inhibitors on kidney, we next examined the expression of Solute carrier family 5 member 1 and 2 (Slc5a1/2, gene encoding SGLT1 and SGLT2). RT-PCR analysis revealed that Slc5a1 and Slc5a2 were highly expressed in the kidney and barely detectable in adipose tissue (Figure 1(d)).

Then, we used a Mito-Tracker to label mitochondria in adipocytes, and the results indicated that the fluorescence intensity in Cana-treated group was higher than the control group (Figure 2(a-b)). We further analysed the mitochondria content of primary adipocytes via determination of copy number and immunofluorescent intensity. The results confirmed Cana can increase the mitochondrial biogenesis (Figure 2(c–d)). Mitochondrial transcription factor A (Tfam, encoded by the nuclear genome) is a master transcription factor to regulate the transcription of the mitochondrial genome and regulated by the transcription factors NRF (nuclear respiratory factor)-1/2 [8,28]. NRFs are under the control of coactivators of mitochondrial gene transcription including Pgc-1α and Pgc-1β [7]. RT-PCR analysis showed that Cana treatment significantly increased the expression of Pgc-1α and Pgc-1β, and their target genes Nrf-1 and Tfam in primary subcutaneous adipocytes (Figure 2(e)). Changes in the levels of Pgc-1α and Tfam were also confirmed by western blotting analysis (Figure 2(f–g)). Taken together, these results indicate that Cana can induce mitochondrial biogenesis in adipocytes.

Oxidative phosphorylation (OXPHOS) of metabolic substrates and fatty acids oxidation (FAO) are two critical biological processes of mitochondria in adipocytes [29]. Several studies have reported an association between insulin resistance and impaired mitochondrial oxidative function [8,30]. Considering the effects of Cana on mitochondrial biogenesis, we examined whether the increased number of mitochondria in adipocytes could translate to enhanced oxidative capacity. As shown in Figure 3, this compound significantly increased the mRNA and protein levels of mitochondrial enzymes involved in the respiratory chain, including cytochrome c oxidase subunit 4 (Cox4) and ubiquinol-cytochrome c reductase core protein 2 (Uqcrc2) (Figure 3(a–c)). In adipocytes, the enhanced activity of mitochondria can increase lipid utilization. Our results indicated that Cana significantly increased the mRNA levels of genes mediating fatty acid metabolism such as peroxisome proliferator activated receptor α (Pparα) and medium-chain acyl-CoA dehydrogenase (Mcad) (Figure 3(d)). In addition, the Cana-induced increase in mitochondrial function was supported by the accumulation of nicotinamide adenine dinucleotide (NAD+) (Figure 3(e)). These findings indicate that Cana can promote oxidative phosphorylation and fatty acid oxidation in mitochondria.

It has been reported that activation of brown fat and white fat browning enhances thermogenesis, which can improve some metabolic disorders such as obesity and diabetes [31], and these processes associated with increased mitochondrial function [32,33]. This prompted us to investigate the effects of Cana on thermogenesis. Results of RT-PCR and western blotting analyses revealed that the level of Ucp-1, a mitochondrial protein responsible for thermogenic respiration, was significantly up-regulated (Figure 4(a–c)). Other thermogenic genes like Cell death-inducing DFFA-like effector A (Cidea), iodothyronine deiodinase 2 (Dio2) and Cytochrome c oxidase subunit 8β (Cox8β), or beige-selected genes inducing transmembrane protein 26 (Tmem26) and homeobox A9 (Hoxa9) were also increased by Cana treatment (Figure 4(a)). High metabolic rate of adipocyte depends on multiple biological processes including OXPHOS, FAO and thermogenesis, accompanied by the increased oxygen consumption [8,34,35]. We then investigated whether Cana could alter the cellular oxygen consumption measured by using an Agilent Seahorse XFp Analyser. Indeed, Cana significantly increased the basal and maximal respiratory capacity (Figure 4(d–e)), thereby confirming the effect of Cana in the regulation of oxidative respiration and energy expenditure.

Based on the observation that Cana promoted adipocyte mitochondrial biogenesis, thermogenic activity and oxygen consumption in vitro, we next explore the specific mechanism of Cana in these biological processes. Given the essential role of Pgc-1α in white fat browning and mitochondrial function, we hypothesized that the effects of Cana on primary adipocytes might be mediated by Pgc-1α. A shPgc-1α adenovirus was used to ablated the expression of Pgc-1α in primary adipocytes. We found Cana could significantly increase the mRNA or protein levels of Pgc-1α and its target genes in the control group (Figure 5(a–c)). However, these effects were largely abolished when Pgc-1α knockdown (Figure 5(a–c)). These results confirmed that Pgc-1α was necessary for Cana’ metabolic action in adipocytes.

Db-cAMP, a cAMP analogue, has been reported to activate the PKA pathway and up-regulate Pgc-1α expression [36]. The levels of Pgc-1α and its downstream genes were increased by Db-cAMP treatment and further increased after Cana treatment (Figure 6(a)). Cana increased Pgc-1α expression by approximately 7-fold and Ucp-1 expression by approximately 10-fold (Figure 6(a)), consistent with the results of western blotting analysis (Figure 6(b–c)). These findings indicate that Pgc-1α is a regulator of mitochondrial biogenesis and thermogenesis.

AMPK and Sirt1 are important energy regulatory factors, and both can stimulate mitochondrial biogenesis and regulate glucose homoeostasis [5,37]. AMPK phosphorylation or Sirt1 deacetylation have been reported to directly activate Pgc-1α, thereby regulating mitochondrial function [28]. Indeed, Cana treatment increased the levels of phosphorylated AMPK, and the expression of Sirt1 and liver kinase B1 (LKB1), a kinase upstream of AMPK, in adipocytes (Figure 7(a–e)). Moreover, the expression of Sirt1-downstream nuclear transcriptional regulators such as peroxisome proliferator-activated receptor γ (Pparγ), CCAAT/enhancer-binding protein (C/ebp)-α and-β were also up-regulated (Figure 7(f)), which have been reported to regulate the activity of Pgc-1α [29]. Therefore, the effects of Cana on mitochondrial function might be, at least partly, mediated by an AMPK/Sirt1/Pgc-1α signalling pathway. To determine whether Cana can activate AMPK in vivo, mice were treated with Cana by oral gavage for 8 weeks. Cana treatment was found to up-regulate the levels of phosphorylated AMPK, and the expression of Sirt1 and LKB1 (Figure 8(a–d)), and promote mitochondrial remodelling (Figure 8(e)) in white adipose tissue.