Ipragliflozin attenuates non-alcoholic steatohepatitis development in an animal model

Non-alcoholic steatohepatitis (NASH) is a common chronic liver disease with no decisive treatment. The sodium glucose cotransporter 2 (SGLT2) inhibitor ipragliflozin was developed as a new oral hypoglycemic drug, which can improve NASH via an insulin-independent glucose-lowering effect by inhibiting glucose reabsorption in the renal proximal tubules. However, ipragliflozin appears to modulate steatosis or inflammation via different pathways. To elucidate the new mechanism of ipragliflozin for the treatment of NASH, we evaluated its effects in a NASH mouse model (STAM mice) with beta cell depletion, and compared the expression of microRNAs (miRNAs) in STAM mice treated with or without ipragliflozin (16.7 μg/day for 5 weeks). Ipragliflozin reduced aspartate transaminase and alanine aminotransferase levels, along with reduced hepatic steatosis, hepatocyte ballooning, lobular inflammation, and liver fibrosis. In addition, ipragliflozin upregulated mitochondrial transport-related and antioxidant defensive system-related genes in the liver. Among 2555 mouse miRNA probes, miR-19b-3p was commonly differentially expressed with ipragliflozin treatment for 5 weeks in both the liver and serum but in different directions, with a decrease in the liver and increase in the serum. Therefore, ipragliflozin can improve NASH development likely through the antioxidative stress pathway and by regulating miR-19b-3p.


Introduction
Non-alcoholic steatohepatitis (NASH) is the advanced stage of non-alcoholic fatty liver disease (NAFLD) and is one of the most common chronic liver diseases worldwide with a global prevalence of approximately one-fourth of the adult population [1]. NAFLD has been recognized as the hepatic manifestation of the metabolic syndrome, whose major components are comorbidities such as obesity and type 2 diabetes mellitus (T2DM) [2][3][4]. In fact, liver diseases, including liver cirrhosis and hepatocellular carcinoma (HCC), represent the third leading cause of death among T2DM patients in Japan [5]. NAFLD comprises various pathological conditions ranging from simple steatosis to NASH, cirrhosis, and HCC [6]. Various drugs such as vitamin E [7,8], peroxisome proliferator-activated receptor gamma agonists [9,10], ursodeoxycholic acid [11,12], antioxidants [11], insulin-sensitizing agents, glucose-lowering drugs [13][14][15][16][17][18], and lipid-lowering drugs [19][20][21] have been used for the treatment of NASH. Despite numerous clinical trials and tremendous efforts, there is still no definitive therapy for NASH. Lifestyle alterations, such as exercise and diet control, are currently one of the most effective therapies for NASH [22]. Although such changes bring about better outcomes, they are hard to achieve and maintain for the majority of NASH patients, resulting in poor condition requiring pharmacological therapy. Sodium glucose cotransporter 2 (SGLT2) inhibitors have been developed as a new class of oral hypoglycemic drugs that inhibit glucose reabsorption at the proximal tubules for the treatment of T2DM [23,24]. The oral selective SGLT2 inhibitor ipragliflozin has been shown to improve hyperglycemia, dyslipidemia, liver steatosis, oxidative stress, inflammation, and liver injury in T2DM animal models [25][26][27]. The mechanism of ipragliflozin in the treatment of NASH is mainly considered to be an insulin-independent glucose-lowering effect with subsequent caloric loss via the inhibition of glucose reabsorption in the renal proximal tubules [25][26][27]. Komiya et al. [18] demonstrated that ipragliflozin also promotes epididymal fat accumulation without worsening adipose inflammation and inhibits ectopic fat accumulation in the liver. This suggested that ipragliflozin is involved in other pathways to modulate steatosis or inflammation, including energy homeostasis and balance between adipose and non-adipose tissues.
MicroRNAs (miRNAs) are 18-22-nucleotide-long interfering, endogenous noncoding RNAs, which are abundant in the human genome, including over 1000 reported miRNA sequences [28]. MiRNAs have broad effects in regulating the expression of various genes, in which one miRNA can target and modulate more than 200 genes [29]. Several reports have recently demonstrated that the miRNA expression profile in the liver is associated with the progression of liver diseases, such as liver fibrosis, cirrhosis, and HCC [30][31][32][33][34]. However, what the effect of ipragliflozin is associated with miRNA expression in the NASH liver remains unexplored.
Therefore, the aim of the present study was to elucidate the effect and mechanism of ipragliflozin for the treatment of NASH using a miRNA array in a mouse model to identify differentially expressed miRNAs in NASH with and without ipragliflozin treatment.

Biochemical and histological analyses
Serum AST and ALT levels were significantly reduced in the ipragliflozin group as compared to those of the control group (Fig 1A and 1B). In addition, serum ALP levels were slightly lower in the ipragliflozin group, although the difference was not statistically significant ( Fig  1C). Fasting blood glucose levels were very high in both groups bcause pancreatic beta cells were destroyed by the streptozotocin treatment in establishment of this model (Fig 1D). Body weights between control and ipragliflozin groups were not changed ( Fig 1E).
As shown in Fig 2, the control group developed hepatocyte steatosis, ballooning, and scattered inflammatory cell infiltration including F4/80 positive cells at 10 weeks, which were all clearly reduced in the ipragliflozin group (Fig 2A-2E). The NAFLD activity score was also significantly lower in the ipragliflozin group (Fig 2A-2C).
In addition, liver fibrosis was also significantly diminished in the liver tissues of the ipragliflozin group as compared with those of the control group (Fig 2F-2H). We also confirmed the effect of ipragliflozin for the reduction of liver fibrosis using NAFLD fibrosis score (S1 Fig).

Ipragliflozin upregulated antioxidative genes and inflammatory markers
We further examined the effect of ipragliflozin on oxidative stress in the liver tissues. Ipragliflozin upregulated the expression of mitochondrial transport-related and antioxidant defensive  system-related genes in the liver. Specifically, the mRNA levels of Sod2 and Cat were significantly increased in the ipragliflozin group relative to those of the control group (Fig 3A and  3B). In addition, the mRNA levels of Tnf and Il1b were significantly decreased in the ipragliflozin group relative to those of the control group (Fig 3C and 3D). These results suggest that ipragliflozin improves NASH development through the antioxidative stress and anti-inflammation pathway.

miRNA expression in the liver tissue and serum samples of STAM mice
Finally, to uncover the miRNA profiles during the development of NASH, we analyzed the expression levels of 2555 mouse miRNA probes using liver tissue and serum samples from the control and ipragliflozin groups. As shown in Table 1, three miRNAs were significantly upregulated and three miRNAs were downregulated in liver tissue samples of the ipragliflozin group compared with those of the control group. Unsupervised hierarchical clustering analysis using Pearson's correlation coefficient showed that the miRNA profile of the ipragliflozin group clustered separately from that of the control group in the liver tissue ( Fig 4A). In addition, 59 miRNAs were significantly upregulated and 84 miRNAs were downregulated in serum samples of the ipragliflozin group (Table 2). Unsupervised hierarchical clustering analysis using Pearson's correlation coefficient showed that the serum miRNA profile of the ipragliflozin group clustered separately from that of the control group ( Fig 4B). Among these differentially expressed miRNAs after ipragliflozin treatment, only miR-19b-3p was commonly altered in both the liver tissue and serum samples ( Fig 4C). In addition, RT-qPCR analysis confirmed that miR-19b-3p expression in the serum was significantly upregulated in the ipragliflozin group (Fig 4D).

Discussion
The prevalence of NAFLD and NASH has increased worldwide [35][36][37]. T2DM is an important risk factor for NAFLD; Matteoni et al. [38] demonstrated that 33-50% of T2DM patients have NAFLD. Although the antidiabetic drug pioglitazone, which is an insulin-sensitizing    agent, has a beneficial effect for NASH, long-term outcomes about its efficacy and safety, including the side effects of cardiovascular disease, congestive heart failure, bladder cancer, and bone loss, have not yet been elucidated [39]. To date, no effective drug for NASH has been discovered; thus, it is critical to investigate an effective therapy for patients with NASH. Therefore, we explored the effects of ipragliflozin on NASH and the underlying mechanism using STAM mice. Our results demonstrated that ipragliflozin had a therapeutic effect on NASH in STAM mice. Ipragliflozin has been reported to improve hyperglycemia by inducing the urinary excretion of glucose [40,41]. Ipragliflozin has been shown to dose-dependently increase urinary glucose excretion in various types of mice, including normal and streptozotocin-nicotinamide-induced diabetic mice [25][26][27]42]. In the present study, ipragliflozin improved the lobular inflammation score, although our STAM mouse model is an insulin-deficient NASH model. Hence, we hypothesize that ipragliflozin improved hyperglycemia, liver inflammation, and liver fibrosis through not only urinary glucose excretion but also other mechanisms. Reactive oxygen species (ROS) are recognized as an origin of oxidative stress and produced through the metabolism of free fatty acids in mitochondria, microsomes, and peroxisomes. Migration of the electrons supplied to the mitochondrial respiratory chain eventually reach the cytochrome C oxidase, combining with oxygen and protons to form water [43]. However, some of these electrons leak out and form the superoxide anion radical. SOD2 dismutes this radical into hydrogen peroxide, which is detoxified into water by GPXS and CAT [44]. Low levels of ROS are associated with important cellular processes. Therefore, proper control of oxidative stress and a balance between oxidative and antioxidative responses are pivotal [45]. Excess superoxide can be produced in damaged mitochondria by electron leakage and this excess is converted to hydrogen peroxide by SOD2. Although GPXS or CAT can metabolize hydrogen peroxide to nontoxic H 2 O, iron-mediated Fenton and/or Haber-Weiss reactions produce highly reactive and toxic hydroxyl radicals. The levels of iron which is an inducer of oxidative stress are increased in NASH. Reduced iron levels brings about the improvement of liver functions in patients with chronic liver diseases [46]. In our present study, SOD2 and CAT expression was upregulated in the liver after ipragliflozin administration (Fig 3A and 3B). Thus, most mitochondrial ROS might be detoxified by ipragliflozin. These data suggest that ipragliflozin might inhibit NASH development in the liver, reinforcing the activity of the key ROS-scavenging antioxidant enzymes SOD2 and CAT directly or indirectly.
Furthermore, ipragliflozin is said to inhibit SGLT1 and SGLT2. Since the intestinal and hepatic axes are known to play an important role in NAFLD/NASH, we also examined the expression of SGLT1 and SGLT2 in the intestinal cells of STAM mice. As a result, the expression of SGLT1 was upregulated, but the expression of SGLT2 was not expressed in small intestinal cells (S4 Fig). Surprisingly, ipragliflozin treatment resulted in the expression of p21 in intestinal cells in vitro and in vivo (S5 and S6 Figs). These results suggest that ipragliflozin may be involved in the proliferation of intestinal cells, resulting in changes in enterohepatic immunity. Surprisingly, SGLT2 was not expressed in normal liver, but was expressed in various HCC cell lines (S7 Fig). Therefore, our results suggest that ipragliflozin does not act directly on the normal liver, but inhibits the development of NASH by suppressing SGLT2 in the kidney and SGLT1 in the small intestine.
Among the upregulated miRNAs identified in the serum, miR-19b-3p was found to be significantly downregulated in the liver following ipragliflozin treatment in STAM mice; however, miR-34a-5p, miR-149-5p, and miR-122-5p were not changed in the serum (S2 Fig) and  liver tissue (S3 Fig). There are several potential reasons for the increase of miR-19b-39 in the serum and its decrease in the liver, including (i) leaked miR-19b-3p increased due to the destruction of hepatocytes and (ii) miR-19b-3p was secreted from hepatocytes as a circulating miRNA. Recently, serum miR-19b-3p up-regulation was shown to alleviate the lipopolysaccharide (LPS)-induced inflammatory response of human umbilical vein endothelial cells, thereby inhibiting the LPS-induced expression of the pro-inflammatory cytokines interleukin (IL)-6 and tumor necrosis factor (TNF)-α [47]. In addition, Duan et al. [48] demonstrated that miR-19b-3p attenuated IL-1β-induced extracellular matrix degradation and inflammatory injury. These reports suggest that ipragliflozin might have attenuated NASH development through not only the improvement of hyperglycemia by inducing the urinary excretion of glucose but also by modulation of miR-19b-3p to inhibit IL-6 and TNF-α expression. Furthermore, circulating miR-19b-3p has been reported to be involved in liver fibrosis via inhibition of C-C motif chemokine receptor (CCR2), indicating that up-regulation of circulating miR-19b-3p suppresses liver fibrogenesis via the regulation of inflammation [49].
In conclusion, we confirmed that ipragliflozin improves NASH development and identified a new mitochondrial pathway and a key miRNA that were modulated by ipragliflozin in a NASH mouse model. Our findings might provide a simple and novel therapeutic strategy for NASH.

Animals and experimental design
STAM mice, a NASH-cirrhosis-hepatocarcinogenic model, were purchased from Stelic Institute & Co., Inc. (Tokyo, Japan). This mouse model progresses from NAFLD to NASH at 8 weeks of age and develops HCC at 16 weeks of age [50]. In brief, for model establishment, C57BL/6J male mice were injected with 20 μL solution (10 mg/mL) of streptozotocin (Sigma-Aldrich Japan, Tokyo, Japan) 2 days after birth to destroy pancreatic beta cells, and were subsequently fed a high-fat diet (HFD-32; CLEA-Japan, Tokyo, Japan). For NASH experiments, 5-week-old male STAM mice were divided into two experimental groups: the control group was fed the HFD and the ipragliflozin group was fed the HFD mixed with 16.7 μg/day of ipragliflozin (Astellas Pharma Inc., Tokyo, Japan) for 5 weeks. The feed was replaced 3 times per week, so that it was kept dry, as HFDs tend to retain moisture. After 10 weeks, the mice were sacrificed using carbon dioxide. Blood samples were obtained from the right atrium by cardiac puncture and the livers were excised. The livers were cut into pieces and fixed in 10% formalin for histological analysis or fresh-frozen in liquid nitrogen and stored at -80˚C in a freezer until use. The animals had free access to water and food and were maintained under specific pathogen-free conditions in a temperature-controlled animal facility with a 12-h light-dark cycle. All protocols and procedures conformed to the guidelines of the Kagawa University Committee for Care and Use of Laboratory Animals and were approved by the Animal Experiments Ethics Committee of Kagawa University (approved number 15145). This study was carried out in compliance with the ARRIVE guidelines.

Measurement of fasting blood glucose, plasma aminotransferases, and triglycerides
The levels of fasting blood glucose were measured after 12 h of fasting using a blood glucose meter (GT1820; Arkray Inc., Kyoto, Japan). Aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALP) levels were determined by standard methods at Shikoku Chuken Central Biolaboratory (Takamatsu, Japan).

Histological analysis of the liver
Sections of formalin-fixed livers were used for hematoxylin-eosin and Masson's trichrome staining for pathological analysis. NAFLD activity was assessed by the NAFLD activity score, as described by Kleiner et al., with separate scores for steatosis (0-3), hepatocellular ballooning (0-2), and lobular inflammation (0-3) [51]. All liver specimens were assessed by two hepatologists (AM and TM) blinded to the identities of the study groups.

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
The whole RNA extracted from murine liver using RNeasy kit (Qiagen) was used for RT-qPCR with manufacturer's directions. Each extracted RNA was reverse-transcribed with a TaqMan Reverse Transcription kit (Life Technologies Japan) and quantified using TaqMan Gene Expression Assay (Applied Biosystems, Foster City, USA) with the manufacturer's directions. We analyzed the relative quantification of targeting mRNAs by using the comparative cycle threshold (CT) method (2 −ΔΔCT ). Actb was quantified as a house keeping gene.
Circulating miRNA purified from murine serum using the QIAGEN miRNeasy serumplasma kit (Qiagen K.K., Tokyo) and tissue miRNA extracted from murine liver using the miRNeasy Mini Kit (Qiagen, Hilden, Germany) were used for RT-qPCR with the manufacturer's directions. Each extracted miRNA was reverse-transcribed with TaqMan MicroRNA Reverse Transcription kit (Life Technologies Japan, Tokyo). For the relative quantification of serum miRNA, Caenorhabditis elegans miR-39 (cel-miR-39) was spiked in each sample as an external control. The serum miRNA was quantified using the TaqMan MicroRNA assay with the manufacturer's directions. We analyzed the relative quantification of targeting miRNAs by using the comparative cycle threshold (CT) method (2 −ΔΔCT ) with spiked cel-miR-39 serving as an external control. For the quantification of liver tissue miRNA, Rnu6b was used as an internal control for relative quantification.

miRNA array
Whole miRNA extracted from the serum or liver tissues using a miRNeasy Mini Kit (Qiagen, Hilden, Germany) with the manufacturer's directions was determined by an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). The miRNA samples were measured using RNA 6000 Nano Kit (Agilent Technologies) and labeled by miRCURYHy3/Hy5 Power Labeling Kit. Subsequently these miRNAs were hybridized to a mouse miRNA Oligo chip (v. 21.0; Toray Industries, Tokyo, Japan). Hybridized miRNAs were detected using 3D-Gene Scanner 3000 (Toray Industries) and analyzed by 3D-Gene extraction version 1.2 software (Toray Industries). Comprehensive microarray analysis of miRNA expression between the control and ipragliflozin groups were performed using GeneSpring GX 10.0 software (Agilent Technologies). We used quantile normalization for making the background identical. Differentially expressed miRNAs were verified using the Mann-Whitney U test and the furthest-neighbor method with the absolute uncentered Pearson's correlation coefficient as a metric was used for hierarchical clustering. The base-2 logarithm of the intensity was median-centered for each row in a heat map which was created using the relative expression intensity for each miRNA.

Statistical analysis
All analyses were conducted using computer-assisted GraphPad Prism 8.4.2 (GraphPad Software, San Diego, CA, USA). Paired analysis between the groups was conducted using Student's t test. P < 0.05 was considered to indicate a significant difference between the groups.

Conclusions
Our present study demonstrated that ipragliflozin improved NASH development and identified a new mitochondrial pathway and a key miRNA that were regulated by ipragliflozin in a NASH mouse model. Ipragliflozin also affected to the small intestine for preventing NASH development. Therefore, our findings might provide a novel therapeutic strategy for NASH.