Sodium–glucose cotransporter 2 (SGLT2) inhibitors are widely used in the treatment of type 2 diabetes, heart failure and chronic kidney disease. Their role in the prevention of diet-induced metabolic deteriorations, such as obesity, insulin resistance and fatty liver disease, has not been defined yet. In this study we set out to test whether empagliflozin prevents weight gain and metabolic dysfunction in a mouse model of diet-induced obesity and insulin resistance.
C57Bl/6 mice were fed a western-type diet supplemented with empagliflozin (WDE) or without empagliflozin (WD) for 10 weeks. A standard control diet (CD) without or with empagliflozin (CDE) was used to control for diet-specific effects. Metabolic phenotyping included assessment of body weight, food and water intake, body composition, hepatic energy metabolism, skeletal muscle mitochondria and measurement of insulin sensitivity using hyperinsulinaemic–euglycaemic clamps.
Mice fed the WD were overweight, hyperglycaemic, hyperinsulinaemic and insulin resistant after 10 weeks. Supplementation of the WD with empagliflozin prevented these metabolic alterations. While water intake was significantly increased by empagliflozin supplementation, food intake was similar in WDE- and WD-fed mice. Adipose tissue depots measured by MRI were significantly smaller in WDE-fed mice than in WD-fed mice. Additionally, empagliflozin supplementation prevented significant steatosis found in WD-fed mice. Accordingly, hepatic insulin signalling was deteriorated in WD-fed mice but not in WDE-fed mice. Empagliflozin supplementation positively affected size and morphology of mitochondria in skeletal muscle in both CD- and WD-fed mice.
Empagliflozin protects mice from diet-induced weight gain, insulin resistance and hepatic steatosis in a preventative setting and improves muscle mitochondrial morphology independent of the type of diet.
Obesity prevalence is dramatically increasing worldwide. Recently, overweight and obesity were estimated to account for 4 million deaths globally every year . According to the WHO a vast 1.9 billion adults were overweight in 2016 and over 650 million were obese . Chronic overfeeding, wrong choice of diet and an increasingly sedentary lifestyle are accepted as central contributing factors to the obesity pandemic . There is a long and well-documented epidemiological and pathophysiological relationship between obesity and risk of type 2 diabetes. Despite extensive knowledge of the positive effects of a Mediterranean diet and physical activity on energy and glucose metabolism, sedentary lifestyle and intake of high-energy, fat- and sucrose-rich diets are still predominant in the western world .
Sodium–glucose cotransporter 2 (SGLT2) inhibitors are now widely used for their specific beneficial effects in individuals with type 2 diabetes [5, 6]. Beside their glucose-lowering properties these effects include cardio- and nephroprotection, and a modest reduction in BP and body weight [7, 8]. Interestingly, a weight loss of roughly 2–4 kg is usually only seen within the first 3 months of treatment despite ongoing glucosuria, probably explained by a compensatory increase in food intake . Besides direct effects on hepatic inflammation, reactive oxygen species generation and mitochondrial function SGLT2 inhibitor-induced weight reduction might also contribute to beneficial effects on non-alcoholic fatty liver disease (NAFLD) [10, 11].
While SGLT2 inhibitors are well established in the treatment of type 2 diabetes, heart failure and chronic kidney disease, their role in the primary prevention of metabolic dysfunction is unclear. Prospective studies utilising SGLT2 inhibitors in the primary prevention of obesity are scarce and to our knowledge have not been conducted in humans. In preclinical studies, empagliflozin, a specific SGLT2 inhibitor, showed increased energy expenditure, browning of adipose tissue and decreased adipose tissue inflammation in mice fed a high-fat diet .
In this study we set out to test whether empagliflozin prevents weight gain and metabolic dysfunction in a mouse model of diet-induced obesity and insulin resistance.
A total of 140 male 6-week-old C57Bl/6 mice (Charles River Laboratories, Germany) were used in this study. Mice were kept under standard conditions of 12 h night–day cycle, 23±2–3°C and controlled humidity. After 1 week of acclimatisation, mice were fed four different diets ad libitum (SSNIFF Spezialdiäten, Germany): standard control diet (CD) (SSNIFF art. no. 1534-00); CD with added empagliflozin (CDE); western-type diet (WD) (21.2% fat and 33.3% sugar content, corresponding to 42% energy from fat and 43% energy from carbohydrates; SSNIFF art. no. E15721-34); and WD with added empagliflozin (WDE). Empagliflozin was provided by Boehringer Ingelheim and added directly to diets by SSNIFF to aim at an empagliflozin dose of 30 mg/kg body weight (260 mg empagliflozin/1000 g CD, 192 mg empagliflozin/1000 g WD). Dosing was based on previously published studies [13,14,15]. Group allocation was chosen randomly.
Diets were fed to the mice for 10 weeks, with weekly measurements of capillary blood glucose via tail vein puncture and a handheld glucometer (Accu-Chek Performa Nano; Roche, Switzerland) and weekly measurements of body weight and water and food consumption. Before weekly measurements of blood glucose, mice were fasted for 6 h. At the end of study mice were anaesthetised after a 6 h fast. Mice were killed via a combination of central blood collection with cardiac puncture and cervical dislocation. Tissue samples were collected immediately afterwards, snap-frozen in liquid nitrogen and stored until further processing at −80°C. All animal procedures were performed in accordance with the guidelines of the Austrian Animal Testing Act of 1988. Approval for this animal study was granted by the Austrian Federal Ministry for Education, Science and Research (application no. BMWF-66.011/0066/ − V/3b/2018).
After 10 weeks of diet and/or treatment with empagliflozin, a silicone catheter was surgically inserted into the right external jugular vein of the mice to provide i.v. access as described before . The catheter was connected to an i.v. access point placed behind the neck of the mouse (VABM1BSM/25; Instech Laboratories, USA). During surgery, mice were anaesthetised using a combination of breathable isoflurane (5% [vol./vol.] for induction of anaesthesia and 2–3% [vol./vol.] as anaesthesia maintenance; Zoetis, USA) and given an s.c. injection of piritramide (0.1 mg/kg; Piramal, India). Mice were kept at 37°C via a heating pad during surgery. At 3–5 days after surgery, mice had regained their pre-surgery weight (±10%) and hyperinsulinaemic–euglycaemic clamping was performed as described . In short, after a 6 h fast, i.v. access was established via connecting the i.v. access point. Capillary blood glucose levels were measured using the cut tail method and a handheld glucometer during the experiment. Readings were taken with mice in the fasted state, during the set-up phase, and every 5–10 min during the clamp until stable euglycaemia was reached with a glucose target of 5.55–6.66 mmol/l (Fig. 3e and f). Insulin (insulin aspart; Novo Nordisk, Denmark) was infused steadily at 8 mU kg−1 min−1. A variable infusion of 20% (wt/vol.) glucose (Merck, USA) was given to reach and maintain the set glucose target.
At the end of the clamp, mice were anaesthetised and killed via cervical dislocation. Tissue and blood samples from these mice were not used for further analysis due to likely interference of the clamp experiment with subsequent analysis (especially analysis of the insulin signalling pathway).
MRI studies were performed after the clamp studies in a subset of mice (n=6) using a 3T MRI scanner (Magnetom Skyra; Siemens, Germany). Transverse relaxation time (T2)-weighted Dixon Turbo Spin Echo sequences were acquired via coronal plane sections covering the whole mouse volume. T2-weighted fat- and water-separated images were generated. Total body volume and total fat volume was calculated using Fiji/ImageJ  and employing threshold-based segmentation with manual corrections (electronic supplementary material [ESM] Fig. 1). Fat depots were separated according to their anatomical location into three different compartments: subcutaneous fat; visceral fat; and retroperitoneal fat. Treatment groups were blinded in the data for quantitative analysis.
Liver tissue samples were fixed using a standard formaldehyde fixation/paraffin embedding protocol. Samples were cut in sections and routine H&E staining was done. Grading of steatosis was done on five fields of view per sample in a blinded manner. A standard steatosis grading system was applied according to the percentage of lipid droplets (score: 0, <5%; 1, 5–33%; 2, 33–66%, 3, >66%) .
Triacylglycerol and glycogen measurement
Hepatic triacylglycerol and glycogen levels were measured using an automated analyser (Hitachi Cobas C 311; Roche, Switzerland) with commercially available kits. For analysis of hepatic glycogen, tissue samples were weighed and minced, and glycogen was hydrolysed using 2 mol/l HCl at 100°C for 1 h. After pH neutralisation and centrifugation (10,000 g for 10 min), glucose was measured in the supernatant fraction. NaOH treatment was used as negative controls.
Quantitative PCR and immunoblotting
Homogenisation of tissue samples, RNA isolation, quantification and reverse transcription was done as described previously in detail . Quantitative PCRs were run on a 7900 HT Fast Real-Time cycler (Applied Biosystems, USA). Hypoxanthine guanine phosphoribosyl transferase (encoded by Hprt) mRNA expression was used for normalisation of data. All quantitative PCRs were performed with two technical replicates. Western blotting was performed as described previously .
Citrate synthase assay
Citrate synthase activity in tissue samples was determined using a commercially available kit (Sigma-Aldrich, USA). Protein lysate was freshly prepared before the experiment. Oxaloacetate and acetyl-CoA were added to the samples. Formation of free -SH groups was photometrically measured using a shift in absorption via 5,5′-dithiobis-(2-nitrobenzoic acid)/2-nitro-5-thiobenzoic acid (DTNB/TNB) at 512 nm over the course of 2 min. Baseline activity was subtracted for subsequent analysis. Analysis was performed with two technical replicates. Protein concentration of samples were used for normalisation of data.
Transmission electron microscopy
Preparation of specimens for transmission electron microscopy (TEM) was done according to a previously published protocol . For fixation, Karnovsky fixative (2% [wt/vol.] paraformaldehyde, 2.5% [wt/vol.] glutaraldehyde) was used. TEM sampling was done in three mice per group. Image magnification was 3200×; one image covered an area of 8.87×8.87 μm. Skeletal muscle fibre types were differentiated into type I and type II fibres using both Z-plane thickness and intramyofibrillary mitochondrial content and appearance based on previously published work [19, 20]. Only longitudinal sections were used for further analysis. Using this approach, 220 images, with a total of 9453 mitochondria, were collected for subsequent analysis. Fiji/ImageJ  was used for quantitative analysis of mitochondrial morphology. Group allocation was blinded before subsequent analysis.
Plasma levels of insulin, glucagon and adiponectin were measured using commercially available ELISAs according to the manufacturer’s instructions (respectively, Ultra Sensitive Mouse Insulin ELISA Kit [Crystal Chem, USA], Glucagon ELISA [10 μl kit; Mercodia, Sweden] and Mouse Adiponectin ELISA Kit [R&D Systems, USA]). β-Hydroxybutyrate levels were UV photometrically measured using an available kit (Ketonebody Assay Kit; Sigma-Aldrich, USA).
GraphPad Prism 8.4.1 was used for statistical analysis (GraphPad Software, USA). ANOVA analysis was performed according to assumptions and normality of residuals was checked using the Kolmogorov–Smirnov test. For skewed residuals, the non-parametric Kruskal–Wallis test was used. Post hoc group comparisons were performed with pre-specified comparisons (CD vs WD, CD vs CDE, WD vs WDE) and either Sidak’s or Dunn’s method were used to correct for multiple testing. Systemic data from the feeding period (body weight, blood glucose, water and food intake) were analysed via a two-way ANOVA model with Geisser–Greenhouse correction. To address potential confounding in the analysis of TEM data, multiple linear regression was performed using R Version 4.0.2 and RStudio Version 1.4.615 . Statistical significance was conferred with p values <0.05. Unless otherwise stated, data are expressed as means ± SD. For highly skewed data (mitochondrial data from TEM analysis), boxplots were used to present data.
Empagliflozin treatment prevents diet-induced obesity, hyperglycaemia and hyperinsulinaemia
After 10 weeks on the diets, WD-fed mice had a significantly higher body weight than CD-fed mice. Treatment with empagliflozin prevented diet-induced weight gain during the study, resulting in comparable body weights for WDE-fed and CD-fed mice (Fig. 1a). Accordingly, WDE-fed mice displayed a significantly lower total body fat content and smaller subcutaneous, visceral and retroperitoneal adipose tissue depots when compared with WD-fed mice (Fig. 2). WD-fed mice had elevated fasting blood glucose levels during the study period, with the difference being significant at week 10 when compared with CD-fed mice (Fig. 1b). Empagliflozin prevented hyperglycaemia as well as the accompanying hyperinsulinaemia in WDE-fed mice (Fig. 1b, c). Glucagon, ketone body (β-hydroxybutyrate) and adiponectin levels were comparable between all groups (Fig. 1d–f). Empagliflozin treatment did not significantly affect daily food intake in CD-fed or WD-fed mice (Fig. 1g, h). Water intake was significantly increased upon addition of empagliflozin to the diet, irrespective of whether mice were fed CD or WD (Fig. 1i, j). Glucosuria was seen in all mice upon empagliflozin treatment, irrespective of the type of diet (ESM Fig. 2).
Empagliflozin treatment prevents diet-induced insulin resistance
After 10 weeks of being fed the diets, the whole-body insulin sensitivity of the mice was measured using hyperinsulinaemic–euglycaemic clamps. The glucose infusion rate (GIR) required to maintain stable euglycaemia under hyperinsulinaemic conditions was found to be significantly higher in WDE-fed mice than in WD-fed mice, indicating increased insulin sensitivity (Fig. 3a). Remarkably, the mean GIR (and thus insulin sensitivity) was comparable in WDE- and CD-fed mice. When differences in plasma insulin concentration under clamp conditions and differences in lean body mass of mice were taken into consideration, there was an even greater difference seen when comparing insulin sensitivity in WD and WDE mice (Fig. 3b, c). The GIR was strongly correlated with body weight, suggesting that lack of excess weight gain significantly contributed to the improved insulin sensitivity in WDE mice (Fig. 3d).
Empagliflozin provides protection against hepatic steatosis and improves hepatic insulin signalling
The grade of steatosis was significantly lower in WDE-fed mice than in WD-fed mice and measurement of intrahepatic lipid content corroborated histological findings (Fig. 4a–c). Accordingly, hepatic insulin signalling as assessed by phosphorylated Akt/total Akt (p-Akt/tAkt) was significantly higher in WDE-fed mice than in WD-fed mice (Fig. 4d, e). Levels of insulin receptor expression were comparable between all groups (Fig. 4d, f).
Empagliflozin supplementation affects hepatic energy metabolism
mRNA expression of Pdk4, encoding pyruvate dehydrogenase kinase 4 (PDK4, an isoform known to regulate hepatic insulin signalling and energy metabolism), was significantly lower in the liver of WDE-fed mice compared with WD-fed mice (Fig. 5a). Increased expression of Adipor1 (encoding adiponectin receptor isoform I) and Cpt1a (encoding carnitine palmitoyltransferase I) mRNA suggested increased β-oxidation in livers of WDE-fed mice compared with WD-fed mice (Fig. 5b, c). On the other hand, liver expression of Pparγ (also known as Pparg, encoding peroxisome proliferator-activated receptor γ) and Cd36 mRNA was lower in WDE-fed mice than in WD-fed mice, suggesting lower hepatic fatty acid uptake (Fig. 5d, e). Glycogen levels (Fig. 5f) were not affected by empagliflozin, irrespective of the type of diet, while Pepck (encoding PEPCK) and G6pc (encoding glucose-6 phosphatase) mRNA levels were increased in WDE-fed mice when compared with WD-fed mice, indicative of higher gluconeogenesis (Fig. 5g, h). Gck (encoding glucokinase) mRNA expression was significantly increased in WD-fed mice compared with CD-fed mice (Fig. 5i); addition of empagliflozin to the WD led to a normalisation of Gck mRNA expression.
Empagliflozin treatment affects mitochondrogenesis in skeletal muscle
Citrate synthase activity, as a biomarker of cumulative mitochondrial mass, was increased by addition of empagliflozin to the WD (Fig. 6a). In line, expression levels of Pgc1α (also known as Ppargc1a, encoding peroxisome proliferator-activated receptor γ coactivator 1-α [PGC1α]) mRNA as well as Nrf1 (encoding nuclear respiratory factor 1) and Tfam (encoding mitochondrial transcription factor A) mRNA were increased in WDE-fed mice when compared with WD-fed mice (Fig. 6b–d). Irrespective of the type of diet, empagliflozin supplementation was associated with alterations of subsarcolemmal and intermyofibrillar mitochondrial size and morphology in both types of skeletal muscle fibres as assessed by TEM (Fig. 6e–h). Dietary empagliflozin supplementation led to larger mitochondria and significant changes in the aspect ratio and circularity of mitochondria (ESM Table 1). Skeletal muscle fibre type and myocellular location of mitochondria are known to affect size and shape of mitochondria. To account for this confounding effect, we performed multiple linear regression analysis after adjusting for fibre type and subcellular localisation, and confirmed a drug-specific effect (ESM Table 1). Compared with liver tissue, changes in the p-Akt/tAkt ratio were modest in skeletal muscle (ESM Fig. 3a, b). Skeletal muscle triacylglycerol content was higher in WD-fed mice than in WDE-fed mice, without the difference reaching statistical significance (ESM Fig. 3c). Skeletal muscle glycogen content was unchanged with addition of empagliflozin (ESM Fig. 3d)
SGLT2 inhibitors have an indispensable role in the treatment of type 2 diabetes, chronic kidney disease and heart failure. Despite clinical evidence on cardio- and renoprotection the underlying beneficial mechanisms are still under discussion and include moderate weight loss, haemodynamic changes and cellular effects. While the glucose-lowering, cardioprotective and renoprotective properties of SGLT2 inhibitors are well established and these drugs are widely used in clinical practice, it is not clear whether SGLT2 inhibitors are capable of preventing metabolic disease in a high-risk scenario.
The aim of this study was to investigate whether SGLT2 inhibition provides protection against diet-induced weight gain, disturbances in glucose metabolism and fatty liver disease. The study was performed in male C57BL/6 mice on a WD as this setting is known to induce a distinct phenotype of metabolic disease including obesity, insulin resistance and hepatic steatosis .
Various mechanisms, especially in relation to possible benefits of SGLT2 inhibition regarding fatty liver disease, have been proposed and reviewed before [10, 23]. Mechanisms include attenuation of inflammation and macrophage polarisation [12, 24], autophagy , endoplasmic reticulum stress [25,26,27], and attenuation of steatosis and fibrosis  to name a few examples. However, whether SGLT2 inhibition is capable of preventing the development of obesity and insulin resistance is currently poorly studied. Therefore we chose to study empagliflozin in a diet-induced mouse model and to carry out extensive metabolic phenotyping to assess the metabolic health of the mice.
Adipose tissue plays a key role in the pathophysiology of insulin resistance, type 2 diabetes and fatty liver disease mainly via the release of NEFA, proinflammatory cytokines and adipocytokines [29, 30]. In our study, we show that empagliflozin supplementation prevents excess weight gain in WD-fed mice while it had no relevant effect on adipose tissue depots in CD-fed mice. MRI studies revealed that empagliflozin supplementation reduced diet-induced expansion of subcutaneous, visceral and retroperitoneal adipose tissue depots. In a previous study by Vallon et al , epididymal fat adipocyte size was found to be reduced upon empagliflozin treatment in C57BL/6-background mice, while increased adipocyte size was found in insulin-deficient Akita mice.
In individuals with type 2 diabetes, SGLT2 inhibitor treatment is associated with a modest weight reduction, which is typically seen during the initial phase of treatment. As expected, in our study empagliflozin-treated mice displayed marked glycosuria irrespective of type of diet. The weight difference between empagliflozin-treated and control mice might be explained by urinary energy loss and increased energy expenditure upon SGLT2 inhibitor treatment, as reported previously . Glucose excretion into the urine was measured in fasted mice only in this study. However, a contribution by urinary glucose excretion in empagliflozin-treated mice to overall GIR in mice from the hyperinsulinaemic–euglycaemic clamp part of the study cannot be ruled out by the study design.
Interestingly, in humans, the initial weight loss is attenuated by a compensatory increase in energy intake . In previous murine studies, both unchanged and increased food intake upon empagliflozin treatment was reported [31,32,33]. In our study, empagliflozin supplementation was associated with increased food intake in CD- and WD-fed mice, without the difference reaching statistical significance, as estimated by twice weekly determination of the food remaining in each cage. However, the sensitivity of this method might be too low to detect small differences in food intake.
Importantly, in our model, empagliflozin protected the mice against WD-induced insulin resistance as shown by hyperinsulinaemic–euglycaemic clamp studies. When adjusting pure GIRs for lean body mass and increased insulin levels at clamp conditions the results indicate an even more drastic increase in insulin sensitivity upon addition of empagliflozin to WD. A full dataset for lean body mass, clamp data and insulin levels was only available in a small sample size, so results should be seen as hypothesis-generating despite being statistically significant.
We found increased hepatic Akt phosphorylation, indicating increased hepatic insulin sensitivity with dietary empagliflozin supplementation. Akt phosphorylation in WD-fed compared with CD-fed mice was similar, although fasting insulin plasma levels were not. Mice on the WD were hyperinsulinaemic compared with CD-fed mice and this might explain similar Akt phosphorylation despite the presence of insulin resistance.
Body weight is a major contributing factor to whole-body insulin sensitivity, with some authors arguing it is the most important predicting factor especially after weight loss [34, 35]. Here, we corroborate this association by correlating body weight with GIR (Fig. 3d). A limitation of the study is the relative per-body-weight dose adjustment of insulin for the clamp procedure instead of adjustment for lean body mass.
Diet-induced obesity and insulin resistance are strongly associated with NAFLD. NAFLD is commonly associated with hepatic insulin resistance and results in further deterioration of systemic glucose metabolism [36, 37]. In our study, empagliflozin protected the WD-fed mice against diet-induced hepatic steatosis. In obese people with moderately controlled type 2 diabetes, empagliflozin treatment for 20 weeks resulted in improved blood glucose control, moderate weight reduction and significant improvement in liver triacylglycerol content . In previous studies, significant weight loss of greater than 10% of initial body weight led to the resolution of steatosis in individuals with obesity [36, 37]. A study by Kahl et al showed that empagliflozin reduces liver fat content in individuals with well-controlled type 2 diabetes . Interestingly, in that study no change in insulin sensitivity was observed, in contrast to findings of improved insulin sensitivity with empagliflozin in our study. In obese, insulin-resistant C57BL/6J mice Xu and colleagues reported that empagliflozin diminished weight gain and reduced deteriorations of insulin sensitivity and hepatic steatosis of ongoing high-fat-diet feeding [12, 24]. Mechanistically, increased energy expenditure and browning of adipose tissue as well as reduced inflammation in adipose tissue and the liver were found upon SGLT2 inhibition.
Here we show that empagliflozin not only exerts beneficial effects on overt NAFLD but also protects against diet-induced hepatic steatosis. Interestingly, we found reduced Pdk4 mRNA expression in WDE-fed mice. Hepatic PDK4 expression has been linked to impaired insulin sensitivity and fatty liver disease [39,40,41] through stimulation of fatty acid uptake and synthesis in the liver. Accordingly, in our study reduced Pdk4 expression was accompanied by decreased fatty acid uptake, estimated by expression of key regulators of cellular fatty acid metabolism. Barres et al  reported altered promotor methylation of PDK4 in obese individuals, with the alteration being restored by significant weight loss, suggesting that empagliflozin protects against obesity-induced alterations in PDK4 expression upon WD feeding.
The p-Akt/tAkt ratio was increased in WDE-fed mice compared with WD-fed mice, suggesting enhanced hepatic insulin sensitivity. While empagliflozin supplementation did not affect hepatic glycogen content, expression data suggest increased gluconeogenesis and decreased glycolysis in WDE-fed mice when compared with WD-fed mice. While increased gluconeogenesis and enhanced insulin signalling seems controversial, these data might be explained by the significantly reduced PDK4 expression in WDE-fed mice. PDK4 inhibits the pyruvate dehydrogenase complex, which links fatty acid and glucose metabolism by catalysing the oxidative decarboxylation of pyruvate [40, 43]. PDK4 deficiency was shown to prevent hepatic steatosis upon high-fat-diet feeding, probably due to increased PGC1⍺ activity. This decreased activity is associated with increased levels of PEPCK and reduced capacity for de novo fatty acid synthesis . Importantly, PDK4 expression is increased and its methylation decreased in type 2 diabetes . We hypothesise that decreased Pdk4 mRNA expression might underlie the increased hepatic gluconeogenesis observed in WDE-fed mice when compared with WD-fed mice. It might be speculated that the reduced Pdk4 mRNA expression in livers of WDE-fed mice might be associated with reduced oxidation of carbohydrates leading to a shift towards fatty acid oxidation and stimulation of gluconeogenesis. In contrast, reduced glycolysis, as indicated by decreased glucokinase mRNA expression levels, in WDE-fed mice in comparison with WD-fed mice might reflect improved hepatic insulin sensitivity.
Unexpectedly, expression markers of key enzymes of gluconeogenesis were decreased in WD-fed mice when compared with CD-fed mice, suggesting preserved insulin action in insulin resistance partially compensated by hyperinsulinaemia. Further dynamic tests will be necessary to better understand the effects of SGLT2 inhibitors on hepatic glycogen metabolism.
Mechanistically, high influx of adipose-tissue-derived fatty acids, diminished suppression of fatty acid synthesis in the liver, reduced hepatic insulin signalling, and impaired adiponectin signalling are major drivers of triacylglycerol accumulation in the liver. Our data suggest that empagliflozin supplementation reduces fatty acid uptake and synthesis (probably by affecting PDK4 expression), enhances adiponectin activity (due to increased adiponectin receptor expression) and improves hepatic insulin signalling (as shown by increasing Akt phosphorylation).
Besides liver and adipose tissue, glucose disposal of skeletal muscle is another major determinant of systemic insulin sensitivity. Here, we did not see relevant changes in p-Akt/tAkt ratio (ESM Fig. 3). However, assessment of Akt activation was performed in fasted mice, whereas data from clamp studies were acquired under hyperinsulinaemic conditions. In skeletal muscle, abnormal mitochondrial function is thought to play a key role in pathophysiology of cellular insulin resistance [46, 47].
Remarkably, our data suggest diet-independent effects on mitochondrial morphology and mass, as shown by increased citrate synthase activity and elevated mRNA expression levels of Pgc1α and downstream targets Nrf1 and Tfam upon empagliflozin supplementation. Finally, ultrastructural analysis revealed skeletal muscle mitochondria to be larger and have a more rounded shape after empagliflozin supplementation compared with no supplementation. Altered skeletal muscle mitochondrial morphology is a hallmark of type 2 diabetes and obesity . Our group and others have shown weight-independent beneficial effects of empagliflozin on mitochondrial morphology  and improved mitochondrial respiration in cardiac muscle . Additionally, empagliflozin treatment improved skeletal muscle mitochondrial function in a murine model of heart failure .
Although previous studies have shown the beneficial effects of empagliflozin treatment in obese, insulin-resistant or diabetic humans or rodents, this is to our knowledge the first study that demonstrates preventative effects of empagliflozin in healthy mice. Our study suggests that empagliflozin treatment not only reverses obesity (or diabetes-specific alterations in metabolically important tissues such as adipose tissue, liver and skeletal muscle) but also protects against systemic and tissue-specific metabolic defects seen upon high-fat diet intake.
In conclusion, empagliflozin treatment protects mice against obesity, insulin resistance and hepatic steatosis in a setting of high-energy, high-sucrose and high-fat intake. Although many metabolic effects of empagliflozin might predominantly result from prevention of excess weight gain, our data suggest that additional diet-independent benefits on mitochondrial structure and mass in skeletal muscle might also contribute to preservation of systemic insulin sensitivity.
Original data are available upon reasonable request from the corresponding author.
Standard control diet
CD + empagliflozin
Glucose infusion rate
Non-alcoholic fatty liver disease
Pyruvate dehydrogenase kinase 4
Peroxisome proliferator-activated receptor γ coactivator 1-α
Sodium–glucose cotransporter 2
Transverse relaxation time
Transmission electron microscopy
WD + empagliflozin
GBD 2015 Obesity Collaborators, Afshin A, Forouzanfar MH et al (2017) Health effects of overweight and obesity in 195 countries over 25 years. N Engl J Med 377(1):13–27. https://doi.org/10.1056/NEJMoa1614362
World Health Organization (2021) Obesity and overweight. https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight. Accessed 5 Oct 2021
Heymsfield SB, Wadden TA (2017) Mechanisms, pathophysiology, and management of obesity. N Engl J Med 376(3):254–266. https://doi.org/10.1056/NEJMra1514009
Lingvay I, Sumithran P, Cohen RV, le Roux CW (2021) Obesity management as a primary treatment goal for type 2 diabetes: time to reframe the conversation. Lancet S0140-6736(21)01919-X. https://doi.org/10.1016/S0140-6736(21)01919-X
Tentolouris A, Vlachakis P, Tzeravini E, Eleftheriadou I, Tentolouris N (2019) SGLT2 inhibitors: a review of their antidiabetic and cardioprotective effects. Int J Environ Res Public Health 16(16):2965. https://doi.org/10.3390/ijerph16162965
Cowie MR, Fisher M (2020) SGLT2 inhibitors: mechanisms of cardiovascular benefit beyond glycaemic control. Nat Rev Cardiol 17(12):761–772. https://doi.org/10.1038/s41569-020-0406-8
Zinman B, Wanner C, Lachin JM et al (2015) Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med 373(22):2117–2128. https://doi.org/10.1056/NEJMoa1504720
Packer M, Anker SD, Butler J et al (2020) Cardiovascular and renal outcomes with empagliflozin in heart failure. N Engl J Med 383(15):1413–1424. https://doi.org/10.1056/NEJMoa2022190
Ferrannini G, Hach T, Crowe S, Sanghvi A, Hall KD, Ferrannini E (2015) Energy balance after sodium-glucose cotransporter 2 inhibition. Diabetes Care 38(9):1730–1735. https://doi.org/10.2337/dc15-0355
Scheen AJ (2019) Beneficial effects of SGLT2 inhibitors on fatty liver in type 2 diabetes: a common comorbidity associated with severe complications. Diabetes Metab 45(3):213–223. https://doi.org/10.1016/j.diabet.2019.01.008
Kahl S, Gancheva S, Straßburger K et al (2020) Empagliflozin effectively lowers liver fat content in well-controlled type 2 diabetes: a randomized, double-blind, phase 4, placebo-controlled trial. Diabetes Care 43(2):298–305. https://doi.org/10.2337/dc19-0641
Xu L, Nagata N, Nagashimada M et al (2017) SGLT2 inhibition by empagliflozin promotes fat utilization and browning and attenuates inflammation and insulin resistance by polarizing M2 macrophages in diet-induced obese mice. EBioMedicine 20:137–149. https://doi.org/10.1016/j.ebiom.2017.05.028
Radlinger B, Hornsteiner F, Folie S et al (2020) Cardioprotective effects of short-term empagliflozin treatment in db/db mice. Sci Rep 10:19686. https://doi.org/10.1038/s41598-020-76698-8
Takahashi H, Nomiyama T, Terawaki Y et al (2019) Combined treatment with DPP-4 inhibitor linagliptin and SGLT2 inhibitor empagliflozin attenuates neointima formation after vascular injury in diabetic mice. Biochem Biophys Rep 18:100640. https://doi.org/10.1016/j.bbrep.2019.100640
Shao Q, Meng L, Lee S et al (2019) Empagliflozin, a sodium glucose co-transporter-2 inhibitor, alleviates atrial remodeling and improves mitochondrial function in high-fat diet/streptozotocin-induced diabetic rats. Cardiovasc Diabetol 18(1):165. https://doi.org/10.1186/s12933-019-0964-4
Jelenik T, Séquaris G, Kaul K et al (2014) Tissue-specific differences in the development of insulin resistance in a mouse model for type 1 diabetes. Diabetes 63(11):3856–3867. https://doi.org/10.2337/db13-1794
Schindelin J, Arganda-Carreras I, Frise E et al (2012) Fiji: an open-source platform for biological-image analysis. Nat Methods 9(7):676–682. https://doi.org/10.1038/nmeth.2019
Kleiner DE, Brunt EM, Van Natta M et al (2005) Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatol Baltim Md 41(6):1313–1321. https://doi.org/10.1002/hep.20701
Nielsen J, Christensen AE, Nellemann B, Christensen B (2017) Lipid droplet size and location in human skeletal muscle fibers are associated with insulin sensitivity. Am J Physiol Endocrinol Metab 313(6):E721–E730. https://doi.org/10.1152/ajpendo.00062.2017
Sjöström M, Angquist KA, Bylund AC, Fridén J, Gustavsson L, Scherstén T (1982) Morphometric analyses of human muscle fiber types. Muscle Nerve 5(7):538–553. https://doi.org/10.1002/mus.880050708
R Core Team (2021) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria URL: https://www.R-project.org
Dobner J, Ress C, Rufinatscha K et al (2017) Fat-enriched rather than high-fructose diets promote whitening of adipose tissue in a sex-dependent manner. J Nutr Biochem 49:22–29. https://doi.org/10.1016/j.jnutbio.2017.07.009
Androutsakos T, Nasiri-Ansari N, Bakasis A-D et al (2022) SGLT-2 inhibitors in NAFLD: expanding their role beyond diabetes and cardioprotection. Int J Mol Sci 23(6):3107. https://doi.org/10.3390/ijms23063107
Xu L, Nagata N, Chen G et al (2019) Empagliflozin reverses obesity and insulin resistance through fat browning and alternative macrophage activation in mice fed a high-fat diet. BMJ Open Diabetes Res Care 7(1):e000783. https://doi.org/10.1136/bmjdrc-2019-000783
Nasiri-Ansari N, Nikolopoulou C, Papoutsi K et al (2021) Empagliflozin attenuates non-alcoholic fatty liver disease (NAFLD) in high fat diet Fed ApoE(-/-) mice by activating autophagy and reducing ER stress and apoptosis. Int J Mol Sci 22(2):818. https://doi.org/10.3390/ijms22020818
Swe MT, Thongnak L, Jaikumkao K, Pongchaidecha A, Chatsudthipong V, Lungkaphin A (2019) Dapagliflozin not only improves hepatic injury and pancreatic endoplasmic reticulum stress, but also induces hepatic gluconeogenic enzymes expression in obese rats. Clin Sci Lond Engl 133(23):2415–2430. https://doi.org/10.1042/CS20190863
Shibusawa R, Yamada E, Okada S et al (2019) Dapagliflozin rescues endoplasmic reticulum stress-mediated cell death. Sci Rep 9(1):9887. https://doi.org/10.1038/s41598-019-46402-6
Dwinata M, Putera DD, Hasan I, Raharjo M (2020) SGLT2 inhibitors for improving hepatic fibrosis and steatosis in non-alcoholic fatty liver disease complicated with type 2 diabetes mellitus: a systematic review. Clin Exp Hepatol 6(4):339–346. https://doi.org/10.5114/ceh.2020.102173
Shulman GI (2014) Ectopic fat in insulin resistance, dyslipidemia, and cardiometabolic disease. N Engl J Med 371(12):1131–1141. https://doi.org/10.1056/NEJMra1011035
Roden M, Shulman GI (2019) The integrative biology of type 2 diabetes. Nature 576(7785):51–60. https://doi.org/10.1038/s41586-019-1797-8
Vallon V, Gerasimova M, Rose MA et al (2014) SGLT2 inhibitor empagliflozin reduces renal growth and albuminuria in proportion to hyperglycemia and prevents glomerular hyperfiltration in diabetic Akita mice. Am J Physiol Renal Physiol 306(2):F194–F204. https://doi.org/10.1152/ajprenal.00520.2013
Nambu H, Takada S, Fukushima A et al (2020) Empagliflozin restores lowered exercise endurance capacity via the activation of skeletal muscle fatty acid oxidation in a murine model of heart failure. Eur J Pharmacol 866:172810. https://doi.org/10.1016/j.ejphar.2019.172810
Rieg T, Masuda T, Gerasimova M et al (2014) Increase in SGLT1-mediated transport explains renal glucose reabsorption during genetic and pharmacological SGLT2 inhibition in euglycemia. Am J Physiol Renal Physiol 306(2):F188–F193. https://doi.org/10.1152/ajprenal.00518.2013
Clamp LD, Hume DJ, Lambert EV, Kroff J (2017) Enhanced insulin sensitivity in successful, long-term weight loss maintainers compared with matched controls with no weight loss history. Nutr Diabetes 7(6):e282–e282. https://doi.org/10.1038/nutd.2017.31
Lillioja S, Bogardus C (1988) Obesity and insulin resistance: lessons learned from the Pima Indians. Diabetes Metab Rev 4(5):517–540. https://doi.org/10.1002/dmr.5610040508
Hannah WN, Harrison SA (2016) Lifestyle and dietary interventions in the management of nonalcoholic fatty liver disease. Dig Dis Sci 61(5):1365–1374. https://doi.org/10.1007/s10620-016-4153-y
Rachakonda V, Wills R, DeLany JP, Kershaw EE, Behari J (2017) Differential impact of weight loss on nonalcoholic fatty liver resolution in a North American cohort with obesity. Obes Silver Spring Md 25(8):1360–1368. https://doi.org/10.1002/oby.21890
Kuchay MS, Krishan S, Mishra SK et al (2018) Effect of empagliflozin on liver fat in patients with type 2 diabetes and nonalcoholic fatty liver disease: a randomized controlled trial (E-LIFT Trial). Diabetes Care 41(8):1801–1808. https://doi.org/10.2337/dc18-0165
Zhao Y, Tran M, Wang L, Shin D-J, Wu J (2020) PDK4-deficiency reprograms intrahepatic glucose and lipid metabolism to facilitate liver regeneration in mice. Hepatol Commun 4(4):504–517. https://doi.org/10.1002/hep4.1484
Zhang S, Hulver MW, McMillan RP, Cline MA, Gilbert ER (2014) The pivotal role of pyruvate dehydrogenase kinases in metabolic flexibility. Nutr Metab 11(1):10. https://doi.org/10.1186/1743-7075-11-10
Zhang M, Zhao Y, Li Z, Wang C (2018) Pyruvate dehydrogenase kinase 4 mediates lipogenesis and contributes to the pathogenesis of nonalcoholic steatohepatitis. Biochem Biophys Res Commun 495(1):582–586. https://doi.org/10.1016/j.bbrc.2017.11.054
Barres R, Kirchner H, Rasmussen M et al (2013) Weight loss after gastric bypass surgery in human obesity remodels promoter methylation. Cell Rep 3(4):1020–1027. https://doi.org/10.1016/j.celrep.2013.03.018
Gudi R, Bowker-Kinley MM, Kedishvili NY, Zhao Y, Popov KM (1995) Diversity of the pyruvate dehydrogenase kinase gene family in humans. J Biol Chem 270(48):28989–28994. https://doi.org/10.1074/jbc.270.48.28989
Hwang NR, Yim S-H, Kim YM et al (2009) Oxidative modifications of glyceraldehyde-3-phosphate dehydrogenase play a key role in its multiple cellular functions. Biochem J 423(2):253–264. https://doi.org/10.1042/BJ20090854
Kulkarni SS, Salehzadeh F, Fritz T, Zierath JR, Krook A, Osler ME (2012) Mitochondrial regulators of fatty acid metabolism reflect metabolic dysfunction in type 2 diabetes mellitus. Metabolism 61(2):175–185. https://doi.org/10.1016/j.metabol.2011.06.014
Koliaki C, Roden M (2016) Alterations of mitochondrial function and insulin sensitivity in human obesity and diabetes mellitus. Annu Rev Nutr 36:337–367. https://doi.org/10.1146/annurev-nutr-071715-050656
Szendroedi J, Phielix E, Roden M (2011) The role of mitochondria in insulin resistance and type 2 diabetes mellitus. Nat Rev Endocrinol 8(2):92–103. https://doi.org/10.1038/nrendo.2011.138
Kelley DE, He J, Menshikova EV, Ritov VB (2002) Dysfunction of mitochondria in human skeletal muscle in type 2 diabetes. Diabetes 51(10):2944–2950. https://doi.org/10.2337/diabetes.51.10.2944
Parts of this study were presented as a short oral discussion at the 58th annual meeting of the EASD 2022.
Authors’ relationships and activities
The authors declare that there are no relationships or activities that might bias, or be perceived to bias, their work.
BR, CR, SF, AL, BW, GS, KS, MG, JaH and JoH performed all experiments. WS and BR performed TEM sample preparation, image acquisition and data analysis. TJ contributed technical expertise for performing clamps in mice. CK and BR designed and performed MRI scans and subsequent image analysis. PM prepared H&E staining from tissue samples. BR, CR, MR and SK drafted an early version of the manuscript. All authors contributed to analysis and discussion of the data. All authors revised and approved the manuscript in its final form. SK is the guarantor of this work.
Open access funding provided by University of Innsbruck and Medical University of Innsbruck. The financial support by the Federal Ministry for Digital and Economic Affairs and the National Foundation for Research, Technology and Development is gratefully acknowledged. We would like to thank Boehringer Ingelheim for providing financial support for these studies (to SK).
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Radlinger, B., Ress, C., Folie, S. et al. Empagliflozin protects mice against diet-induced obesity, insulin resistance and hepatic steatosis. Diabetologia 66, 754–767 (2023). https://doi.org/10.1007/s00125-022-05851-x