Empagliflozin mitigates metabolic dysfunction‐associated steatotic liver disease by reducing de novo lipogenesis in a mouse model of lipoatrophic diabetes

Initially developed as an antidiabetic drug that acts by increasing urinary glucose excretion, sodium-glucose transport protein 2 inhibitors (SGLT2i) quickly showed a marked benefit in terms of cardiovascular and renal protection.1, 2 Numerous studies also indicated an improvement in hepatic steatosis with a decrease in plasma liver enzymes in subjects with metabolic dysfunction-associated steatotic liver disease (MASLD).3 In patients with well-controlled type 2 diabetes (T2D), dapagliflozin reduced hepatic steatosis and improved insulin sensitivity in the liver and white adipose tissue.4 Furthermore, a meta-analysis of 12 randomized trials confirms that SGLT2is decrease liver enzymes and liver fat content measured by magnetic resonance imaging in 90% of patients with MASLD and T2D.5 As SGLT2 is not expressed in the liver,6 these hepatoprotective effects have been suspected to be of extrahepatic origin. Seipin knock-out (SKO) mice are characterized by an almost complete absence of adipose tissue, leading to adverse metabolic consequences, and constitute an interesting model of T2D and MASLD.7 To get new insight on the effect of SGLT2is on MASLD and to better discriminate the role of adipose tissue in SGLT2i action, we treated SKO mice with empagliflozin and evaluated its effects on hepatic steatosis. SKO mice were generated as previously described8 and were housed in enriched cages in SPF conditions. Mice were fed a chow diet with or without (CONT) empagliflozin (100 mg/kg; Boehringer Ingelheim) for 8 weeks. The length of treatment and mouse number are indicated in the figure legend. Animals are killed at 08:00 h in a random fed state. The ethics committee of the French national veterinary agency approved all animal protocols used in this study. mRNA levels were analysed as previously described.8 All data are reported as mean ± standard error of mean. Data sets were analysed for statistical significance using the non-parametric Mann-Whitney t-test and the Kruskal-Wallis test. As a first preventive approach, we sought to assess whether empagliflozin could alter the development of hepatic steatosis in SKO mice. Eight-week-old SKO mice and their littermates wild-type (WT) mice were fed either an empagliflozin-supplemented diet (100 mg/kg) or a control diet for 8 weeks. In SKO mice, empagliflozin had no impact on body weight (BW) but fully corrected hyperglycaemia (Figure S1A,B). The liver mass was increased by four-fold in 16-week-old SKO mice as compared with WT mice, and empagliflozin treatment reduced liver mass (35%), triglyceride (TG) content (50%) and Oil Red O staining (Figure 1A-C). Consistently, the mRNA levels of the lipid droplet-associated protein Adrp, a marker of liver steatosis, were increased by three-fold in SKO mice compared with WT mice but were reduced by empagliflozin (Figure 1D). In addition, empagliflozin completely prevented the increased expression of the inflammation marker Mcp1 and the extracellular matrix component Col1a1 observed in SKO mice (Figure 1E,F). To understand further how empagliflozin limited liver steatosis development in SKO mice, we measured the mRNA levels of three main actors of de novo lipogenesis: Acly, Acc and Fas. They were found to be increased in SKO mice compared with WT littermates but nearly normalized by empagliflozin (Figure 1G-I). Finally, we tested whether empagliflozin can promote fatty acid oxidation (FAO). While mRNA levels of peroxisome proliferator-activated receptor alpha target genes involved in fatty acid utilization, such as Cpt1, Acot2 and Pdk4, were increased in SKO mice, empagliflozin had no additional effect (Figure 1J-L). Pdk4 levels were even decreased in SKO mice treated with empagliflozin. Of note, empagliflozin did not alter ketogenesis in SKO mice, as plasma levels of ketone bodies remained unchanged (Figure S1C). Altogether, these data showed that empagliflozin prevented the development of MASLD in SKO mice and that this was potentially because of a reduction in de novo lipogenesis. To test whether empagliflozin could improve MASLD in SKO animals that already display liver injury, 16-week-old mice were treated with empagliflozin for 7 weeks. The empagliflozin did not affect either BW or adipose tissue mass in SKO mice (Figure S2A,B). Empagliflozin reduced glycaemia (Figure S2C) without altering insulinaemia or insulin sensitivity (Figure S2D,E). Empagliflozin treatment reduced liver mass by 19% and liver TG by 26% (Figure 2A-C). It is important to note that while SKO mice livers displayed severe steatosis, they did not develop significant ballooning or inflammation (data not shown). The empagliflozin significantly reduced mRNA levels of Adrp but not Mcp1 or Col1a1, suggesting a primary action on liver steatosis rather than on inflammation or fibrosis. Empagliflozin treatment reduced the mRNA levels of Acly and Acc, reinforcing the hypothesis of its central action on de novo lipogenesis to improve hepatic steatosis. As observed in the preventive approach, empagliflozin did not alter gene expression levels of Cpt1 and Acot2 but reduced Pdk4 mRNA levels (Figure 2J-L). Overall, these data showed that empagliflozin was able to improve established liver steatosis in older SKO mice and that this was still associated with a decrease in lipogenic gene expression. SKO mice are an interesting model of lipoatrophic T2D, as they quickly develop severe insulin resistance and massive liver steatosis. In this work, we showed that empagliflozin blunted the development of MASLD in a preventive approach and reduced the extent of established liver steatosis in a curative approach. Using this preclinical model, we intended to decipher how empagliflozin could exert its beneficial effect on MASLD. First, in this study and in our previous one,9 SGLT2i normalized glycaemia without changing BW, plasma insulin levels or insulin sensitivity in SKO mice. This suggests that, in SKO mice, SGLT2i reduces liver steatosis by a mechanism independent of BW loss or improvement of insulin resistance. Importantly, empagliflozin neither increased nor decreased fat pad weights, indicating an effect independent of an action on the adipose tissue. This suggests that SGLT2i could be an interesting therapeutic option for the management of MASLD associated with lipodystrophies. Several reports have proposed that SGLT2is promote fatty acid utilization by increasing the glucagon/insulin ratio, leading to increased ketogenesis and FAO.10 This has been proposed as a central mechanism in the beneficial effect of SGLT2i on MASLD.11 For instance, Radlinger et al.12 suggested that empagliflozin prevents the development of liver steatosis in Western diet-fed mice by increasing the expression of FAO genes. In our current experiment, the plasma concentration of ketone bodies was unchanged, and the mRNA levels of peroxisome proliferator-activated receptor alpha target genes involved in FAO were not increased by empagliflozin. This suggests that in our model, FAO activation could not explain the anti-steatotic effect of SGLT2i. In another mouse model of MASLD (i.e. low-streptozotocin-treated mice under a high-fat diet), empagliflozin and liraglutide led to similar results with no effect on beta-oxidation-related gene expression but a decrease in lipogenic genes expression.13 In MASLD, the adipose tissue is an important source of free fatty acid (FFA) when lipolysis is elevated in insulin-resistant adipocytes. However, we have shown that TG and FFA levels are lower in SKO mice than in control littermates, indicating that FFA derived from adipose tissue is not the main contributor to the liver steatosis observed in SKO mice.7 On the other hand, empagliflozin decreased the expression of the lipogenic genes, suggesting that reduced de novo lipogenesis could be one of the main mediators of the beneficial effect of SGLT2i on MASLD. Preliminary clinical studies have shown that SGLT2is can reduce the accumulation of intrahepatic fat through mechanisms that are not yet fully understood.6 Here, it is tempting to propose that empagliflozin can reduce de novo lipogenesis by inhibiting hepatic glucose input. Beyond our specific mouse model, given that in obese patients with MASLD, de novo lipogenesis contributes to nearly 30% of liver lipid accumulation, limiting this biosynthetic pathway should indeed be beneficial.14 Altogether, our work shows that despite its lack of effect on insulin sensitivity, BW, adipose tissue function and the pro-catabolic pathways, empagliflozin mitigates liver steatosis in a mouse model of lipoatrophic diabetes through a repression of lipogenic gene expression. SS, TS, PD, SD, MJ and LA: research data and wrote the manuscript. SH: contributed to scientific discussion. BC, CM and XP: designed the experiments and wrote the manuscript. We acknowledge the MicroPICell facility, SFR-Santé, INSERM, CNRS, UNIV Nantes and CHU Nantes, Nantes, France, member of the national infrastructure France-BioImaging supported by the French National Research Agency (ANR-10-INBS-04). We thank the animal facility UTE-IRS-UN for the daily care of the animals. This work was supported by grants from the “Institut National de la Santé et de la Recherche Médicale” (INSERM) and the Fondation pour la recherche médicale (FRM: EQU201903007846, and FRM physiopathologie cardiovasculaire DPC20171138974). The authors have no conflicts of interest to declare. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.15634. Data available on request from the authors. Figure S1. 8-week-old Seipin KO (KO) and WT littermate mice were fed a chow diet with (EMPA) or without (CONT) empagliflozin (100 mg/kg) for 8 weeks. (A) Body weight, (B) random fed glycaemia and (C) random fed ketone bodies levels were measured at the end of the experiment. For all panels, bars: mean ± SEM. ns indicates p > .05, *p < .05, **p < .01, ***p < .005, as compared with WT-CONT, and #p < .05, ##p < .01, ###p < .005 as compared with KO-CONT. Figure S2. 16-week-old Seipin KO (KO) and WT littermate mice were fed a chow diet with (EMPA) or without (CONT) empagliflozin (100 mg/kg) for 7 weeks. (A) Body weight, (B) inguinal fat pad weight, (C) random fed glycaemia and (D) insulinaemia were measured at the end of the experiment. (E) At the 6th week of the treatment, an insulin tolerance test (0.75 UI/kg) was performed after a 4-h fasting period. For all panels, bars: mean ± SEM. ns indicates p > .05, *p < .05, **p < .01, ***p < .005, as compared with WT-CONT, and #p < .05, ##p < .01, ###p < .005 as compared with KO-CONT. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.