Omaveloxolone and TX63682 are hepatoprotective in the STAM mouse model of nonalcoholic steatohepatitis

Abstract Omaveloxolone is a potent activator of Nrf2, a master transcriptional regulator of a multitude of cytoprotective functions, including antioxidative, anti‐inflammatory, and mitochondrial bioenergetic effects. Some of the most potent known effects of Nrf2 involve hepatoprotective functions. The purpose of this study was to evaluate the effects of omaveloxolone and TX63682, a closely related structural analog with similar oral bioavailability, in the STAM mouse model of nonalcoholic steatohepatitis (NASH). C57Bl/6 mice received a single subcutaneous injection of streptozotocin two days after birth and were fed a high‐fat diet from 4 to 9 weeks of age. Omaveloxolone and TX63682 were orally administered at doses of 1, 3, and 10 mg/kg/d from 6 to 9 weeks of age. Consistent with the beneficial effects of Nrf2 on hepatoprotection and improved lipid handling, both omaveloxolone and TX63682 decreased hepatic fat deposition, hepatocellular ballooning, inflammatory cell infiltration, and collagen deposition. Omaveloxolone and TX63682 also improved blood glucose control, as evidenced by reductions in nonfasting blood glucose and glycated hemoglobin A1C concentrations. Reductions in liver and serum triglycerides with omaveloxolone and TX63682 treatment were also observed. Both omaveloxolone and TX63682 decreased leptin and increased adiponectin in serum, which is consistent with the anti‐inflammatory and antifibrotic effects observed in the liver. These results were associated with significant induction of Nrf2 target gene expression in the liver, including NAD(P)H:quinone oxidoreductase 1, sulfiredoxin 1, and ferritin heavy chain 1. Overall, these data suggest that omaveloxolone and related Nrf2 activators may be useful for the treatment of NASH.

cytoplasm into the mitochondria, where ROS homeostasis is critical to maintaining proper mitochondrial function and ATP production.
Nrf2 can also be physically bound to mitochondria, which aids in its ability to rapidly respond to alterations in mitochondrial function. [3] In addition to responding to ROS, Nrf2 can increase the efficient use of fatty acids and glucose by affecting fundamental pathways of metabolism. [4] For example, Nrf2 enhances the expression of genes encoding enzymes involved in the pentose phosphate pathway, which uses glucose to produce the reduced form of nicotinamide adenine dinucleotide phosphate, an important reducing cosubstrate for many antioxidative reactions, including the reductive recycling of oxidized glutathione. [5] Nrf2 also decreases the levels of acetyl-CoA carboxylase, the rate-limiting enzyme in fatty acid synthesis, which decreases the levels of malonyl-CoA, and results in increased fatty acid oxidation by relieving the suppression of carnitine palmitoyltransferase 1 (Cpt1). [6] Cpt1 catalyzes the rate-limiting step in the beta-oxidation of long-chain fatty acids, and Nrf2 is essential for Cpt1 expression, particularly in settings of high-fat diet (HFD) and oxidative stress. [7] Paramount to the effects of Nrf2 on mitochondrial function, a pivotal clinical trial of omaveloxolone in patients with Friedreich's ataxia was recently completed. [8,9] Friedreich's ataxia is a disease caused by a deficiency in frataxin, a protein involved in the assembly of iron-sulfur clusters, which leads to increased free radicals, oxidative stress, and disruption of mitochondrial homeostasis primarily in the cerebellum, liver, and heart. Administration of omaveloxolone once daily for 48 weeks to patients with Friedreich's ataxia significantly increased the modified Friedreich's ataxia Rating Scale score when compared with a placebo control group. [8,9] Further mechanistic support of the effects of omaveloxolone in the setting of mitochondrial dysfunction comes from studies in Friedreich's ataxia mouse and patient cell models. Omaveloxolone improved mitochondrial function and ATP generation and protected against oxidative stress in neurons from KIKO and YG8R Friedreich's ataxia mice and also in fibroblasts from patients with Friedreich's ataxia. [10] Consistent with the noteworthy improvements observed in Friedreich's ataxia patients and the known cytoprotective effects of Nrf2 in the central nervous system, [11] omaveloxolone has also demonstrated efficacy in rodent models of status epilepticus [12] and multiple sclerosis. [13] Furthermore, in line with the broad applicability of its effects on oxidative stress, inflammation, and mitochondrial dysfunction, omaveloxolone has also demonstrated efficacy in animal models of diseases outside of the central nervous system. For example, omaveloxolone demonstrates significant efficacy in rodent models of kidney ischemia-reperfusion injury, multiple sclerosis, diabetic wound healing, and radiationinduced dermatitis. [13][14][15][16] Nonalcoholic fatty liver disease (NAFLD) is the most common liver disease, affecting 80-100 million people in the United States. [17] Approximately, 25% of cases of NAFLD progress to nonalcoholic steatohepatitis (NASH), which is defined by steatosis, inflammation, cellular ballooning, and fibrosis. [17] NASH has the potential to progress to cirrhosis and hepatocellular carcinoma and is associated with many other comorbidities, such as colon cancer and chronic kidney disease. [18] Mitochondria control hepatic lipid metabolism and regulate oxidative stress in the liver, and, as such, NAFLD and NASH are associated with the disruption of mitochondrial homeostasis, often as a consequence of type 2 diabetes and/or obesity. [19] Because mitochondrial dysfunction leads to increased production of ROS and inflammation and decreased fatty acid oxidation, the role of mitochondria appears to be central to the mechanism contributing to the pathogenesis of NAFLD and NASH. [20] Therefore, because Nrf2 is critical for mitochondrial homeostasis, lipid metabolism, and overall hepatoprotection, [21] the effects of omaveloxolone and a closely related analog, TX63682, were investigated in the STAM mouse model of NASH. [22] Diabetes and obesity are key features of the STAM model, which has been described to recapitulate the known progression of the human disease. [23] Structures of omaveloxolone and TX63682 are presented in Figure 1A. In this model, NASH is induced by a single injection of the pancreatic β-cell toxin streptozotocin (STZ) at 2 days after birth with HFD initiated at 4 weeks of age. This NASH model is characterized by insulin resistance, hyperglycemia, dyslipidemia, and accumulation of hepatic lipids, which facilitates oxidative stress, inflammation, and liver fibrosis. [24] 2 | METHODS

| Animals
For the plasma drug concentration and liver content study, C57Bl/6 mice were obtained from Charles River Laboratories (Wilmington, MA). The study was approved by the Institutional Animal Care and Use Committee at the laboratory where the inlife study was conducted (Biomodels Inc, Watertown, MA). Mice were administered omaveloxolone or TX63682 at 10 mg/kg once daily for 14 days. Sesame oil was used as the vehicle with a dosing volume of 10 mL/kg. Mice were terminated 4 hours after the final dose, and plasma (supplemented with sodium sulfite, 0.25% wt/vol) and livers were collected for determination of drug concentrations.
The in-life portion of the NASH study was conducted at Stelic Institute & Co, (Tokyo, Japan). Thirty C57Bl/6 mice (15-day-pregnant females) were obtained from Charles River Laboratories Japan (Kanagawa, Japan). All animals were housed and cared for in accordance with the Japanese Pharmacological Society Guidelines for Animal Use.
NASH was induced in 68 male mice by a single subcutaneous injection of 20 μL STZ (Catalog #S0130, Lot #031M1287V; Sigma-Aldrich) solution (10 mg/mL) 2 days after birth, followed by feeding with HFD (57 kcal% fat, Catalog #HFD32; CLEA Japan Inc, Japan) ad libitum from 4 to 9 weeks of age. The mice were randomized into a vehicle group (n = 6) and drug treatment groups (n = 8/group) at 6 weeks of age.
Mice were terminated at 9 weeks of age, and whole blood, serum, liver, and pancreas were collected for the various biochemical and molecular biology assays listed in the following sections.

| Quantification of drug concentrations in plasma and liver
Omaveloxolone and TX63682 were extracted from mouse plasma and, then, quantified using liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods similar to those described previously. [25][26][27] Plasma calibration curves were prepared using blank C57Bl/6 mouse plasma obtained from a commercial supplier (BioIVT, Westbury, NY). All plasma samples and controls were supplemented with 0.25% sodium sulfite to prevent oxidative degradation of omaveloxolone or TX63682. Omaveloxolone and TX63682 were extracted from mouse livers and, then, quantified using liquid chromatography methods similar to those previously described. [25]

| Blood biochemistry
Nonfasting blood glucose levels were measured in whole blood samples

| Measurement of liver triglycerides
Liver total lipid extracts were obtained from right lobes by Folch's method. [28] The liver lobes were homogenized in chloroform-methanol (2:1, vol/vol) and incubated overnight at room temperature. After washing with chloroform-methanol-water (8:4:3, vol/vol/vol), the extracts were evaporated to dryness and dissolved in isopropanol. Liver triglyceride content was measured by the triglyceride E-test (Catalog #432-40201, Lot #TJ482; Wako Pure Chemical Industries, Japan).
When the triglyceride content exceeded the detection limit, extracts were diluted twofold in isopropanol and were then reanalyzed. F I G U R E 1 Chemical structures and plasma and liver concentrations of omaveloxolone and analog TX63682. A, Chemical structures of omaveloxolone and TX63682 are presented. B, Blood and liver were collected 4 hours following the final dose after 14 daily oral doses of omaveloxolone (n = 8) or TX63682 (n = 5 for plasma and n = 10 for liver) at a dose level of 10 mg/kg to male C57Bl/6 mice. Plasma and liver concentrations were determined using liquid chromatography-tandem mass spectrometry methodology. Plasma concentrations are presented as mean ng/mL ± SD, and liver concentrations are presented as mean ng/g ± SD. Omav, omaveloxolone REISMAN ET AL.  Enzyme activity data are presented as fold change from normal control ± SEM. Asterisks indicate a statistically significant difference from the vehicle control group (*P < .05, **P < .01, ***P < .001). DCPIP, 2,6-dichlorophenol-indophenol; mRNA, messenger RNA; Omav, omaveloxolone was determined by quantifying the reduction of 2,6-dichlorophenolindophenol in the presence and absence of dicumarol, as previously described [31] and modified. [32] Data were normalized to protein concentration and presented as fold mean normal control.

| Statistics
Statistical analyses were performed using a one-way analysis of  Figure 1B). Overall, pharmacologically relevant concentrations were readily achieved, and exposures were similar between compounds.

| Omaveloxolone and TX63682 do not affect body or liver weights in the STAM mouse model
Mean body weights for all groups of STAM mice were significantly lower than those of normal mice at the start and completion of the study (Table 1). There were no significant differences in mean body weight between the vehicle-treated STAM mice and any of the drug-treated STAM mice during the dosing period. Liver weights were also significantly higher in the vehicle-treated STAM mice compared with the normal nondiseased mice (Table 1). Treatment with omaveloxolone or TX63682 also did not affect liver weight.
Likewise, liver-to-body weight ratios were significantly higher in the vehicle-treated STAM mice and were not affected by treatment with omaveloxolone or TX63682 ( Table 1). The decrease in body weight and increase in liver weight between normal mice and vehicle-treated STAM mice are consistent with previous findings. [24] 3.3 | Omaveloxolone and TX63682 improve liver architecture and decrease collagen and fat deposits As expected in the STAM model of NASH, [22,24] livers from the vehicle group exhibited severe microvesicular and macrovesicular fat deposition, hepatocellular ballooning, and inflammatory cell infiltration, which were reflected by a marked increase in NAFLD activity scores compared to normal mice (Figure 2A,B). Treatment with either omaveloxolone or TX63682 ameliorated these histopathological features ( Figure 2A) and resulted in dose-dependent decreases in NAFLD activity scores with statistical significance achieved at 10 mg/kg for both compounds ( Figure 2B).

| Omaveloxolone and TX63682 induce Nrf2 target genes in liver
Expression of the Nrf2 target genes Nqo1, Srxn1, and Fth1 was quantified in liver ( Figure 3A). Nqo1 enzyme activity, another functional measurement of Nrf2 activity, was also evaluated in livers ( Figure 3B).

| Omaveloxolone and TX63682 improve glucose control and alter lipid profiles
Nonfasting blood glucose and serum glycated HbA 1c levels were significantly elevated in vehicle-treated STAM mice compared to normal nondiseased mice ( Figure 4A,B). Treatment with omaveloxolone or TX63682 at 10 mg/kg tended to decrease, or significantly decreased, blood glucose levels. Consistent with the decreases in fasting blood glucose, serum HbA 1c was also lower with both omaveloxolone and TX63682 at 10 mg/kg, with statistical significance observed with omaveloxolone.
Serum total cholesterol levels, HDL cholesterol, LDL cholesterol, serum triglycerides, and liver triglycerides were all higher in vehicletreated STAM mice compared to normal nondiseased mice ( Figure 5A-E).
Omaveloxolone and TX63682 mildly and dose-dependently increased serum total cholesterol with statistically significant increases observed at 3 and 10 mg/kg TX63682. Omaveloxolone and TX63682 only modestly increased HDL cholesterol ( Figure 5B); however, both compounds significantly and dose-dependently increased LDL cholesterol ( Figure 5C).
Reductions in both serum and liver triglycerides were observed for both omaveloxolone and TX63682 with a statistically significant decrease in liver triglycerides observed by TX63682 at 10 mg/kg ( Figure 5D,E).
Both leptin and adiponectin were significantly decreased in vehicle-treated STAM mice compared to normal nondiseased mice ( Figure 6A,B). Both omaveloxolone and TX63682 at 10 mg/kg decreased leptin further with statistical significance obtained by TX63682. In contrast, both omaveloxolone and TX63682 slightly increased adiponectin compared to vehicle with statistical significance obtained by TX63682 at 3 and 10 mg/kg.
Pancreatic tissue from the vehicle-treated STAM mice showed islet atrophy, a reduction in the number and sizes of islets, and fewer F I G U R E 5 Effect of omaveloxolone and analog TX63682 on lipids. Serum total cholesterol (A), HDL cholesterol (B), LDL cholesterol (C), and triglycerides (D) were analyzed by HPLC. E, Liver triglycerides were quantified with the commercially available Triglyceride E-test Kit (Wako Chemicals). Data are presented as mean ± SEM. Asterisks indicate a statistically significant difference from the vehicle control group (*P < .05, **P < .01, ***P < .001). HDL, high-density lipoprotein; HPLC, high-performance liquid chromatography; LDL, low-density lipoprotein; Omav, omaveloxolone insulin-positive β cells compared to pancreatic tissue from normal nondiseased mice (data not shown). However, no differences in pancreatic insulin staining, or pancreas histology, were observed between STAM mice treated with vehicle and STAM mice treated with any dose of omaveloxolone or TX63682 (data not shown).
Serum insulin was also significantly lower in the vehicle-treated STAM group compared to the normal group; however, no differences were observed between STAM mice treated with vehicle and STAM mice treated with any dose of omaveloxolone or TX63682 (Table 1).
These data indicate that the effects of omaveloxolone and TX63682 on glucose and lipid levels were not related to islet status.

| Omaveloxolone and TX63682 increase serum transaminases
Compared to normal nondiseased mice, vehicle-treated STAM mice had mild increases in both serum ALT and AST ( Figure 7A,B).
Treatment with omaveloxolone or TX63682 resulted in mild and dose-dependent increases in ALT with statistical significance obtained by TX63682 ( Figure 7A). Mild increases in serum AST were also observed with both omaveloxolone and TX63682 treatment ( Figure 7B). No meaningful differences among any of the NASH treatment groups were observed for serum GGT, though there were mild but significant elevations between the vehicle-treated NASH and normal mice ( Figure 7C). There was a minimal yet statistically significant increase in total bilirubin at 3 mg/kg omaveloxolone but no changes at 10 mg/kg, and mean total bilirubin trended lower in the 3 and 10 mg/kg dose groups for TX63682 ( Figure 7D). Based on the evidence, the minimal change observed at 3 mg/kg omaveloxolone was considered a statistical anomaly and not biologically meaningful.
Overall, the mild increases in serum ALT and AST occurred in the absence of meaningful changes in bilirubin and were concurrent with marked improvements in hepatic histology, lipid accumulation, and collagen deposition (Figure 2).

| DISCUSSION
Omaveloxolone and its analog TX63682 demonstrated significant hepatoprotection in the STAM mouse model of NASH, as evidenced by improvements in multiple parameters of hepatic physiology. At doses that significantly and dose-dependently increased Nrf2 target gene expression in the liver, both compounds decreased hepatocellular ballooning and inflammation, as well as lipid accumulation and collagen deposition. Enhanced blood glucose control and reduced serum and liver triglycerides were also observed. Consistent with its higher potency in vitro (not shown for TX63682, see previous data for omaveloxolone [33] ), TX63682 performed slightly better than omaveloxolone in this model, despite achieving similar plasma and liver exposures after oral dosing. The higher activity of TX63682 was associated with higher Nrf2 target gene induction in liver.
Interestingly, despite the improvements in hepatic lipid profiles with omaveloxolone and TX63682, mild, and in some cases significant, increases in serum total, HDL, and LDL cholesterol were observed. Another analog of omaveloxolone, CDDO-Im, also mildly increased serum cholesterol in an HFD mouse model of obesity, while attenuating weight gain and hepatic fat accumulation and increasing energy expenditure and oxygen consumption. [34] Strikingly, CDDO-Im did not affect serum cholesterol or produce the profound improvements on obesity in Nrf2-null mice. Consistent with these observations, untreated Nrf2-null mice have lower serum cholesterol levels than wild-type mice when fed an HFD. [35] Overall, these data suggest that Nrf2 plays an important role in the regulation of cholesterol homeostasis. The mechanism underlying this phenomenon is currently unknown, but we speculate that it may be related to the known improvements in lipid handling, fat mobilization, fatty acid oxidation, and mitochondrial function that have been shown to occur with omaveloxolone treatment and Nrf2 activation. [6] Additionally, the omaveloxolone analog CDDO-dhTFEA increases glutathionedependent bile flow and the biliary excretion of cholesterol, [36] and Nrf2 regulates the expression of the cholesterol efflux transporters ATP-binding cassette sub-family G member 5 (Abcg5) and Abcg8. [37] F I G U R E 6 Effect of omaveloxolone and analog TX63682 on leptin and adiponectin. Serum leptin (A) and adiponectin (B) concentrations were quantified by ELISA. Data are presented as mean ± SEM. Asterisks indicate a statistically significant difference from the vehicle control group (*P < .05, **P < .01, ***P < .001). ELISA, enzyme-linked immunosorbent assay; Omav, omaveloxolone which, in hepatocytes, facilitates pyruvate production that is then used to synthesize glucose. ALT also contributes glutamate to make glutathione, [39] an important cellular antioxidant that is regulated by Nrf2. AST, on the contrary, is an important enzyme in the malateaspartate shuttle and the generation of NADH, a necessary cosubstrate for ATP generation by mitochondria. [38] Induction of NASH in vehicle-treated STAM mice led to significant decreases in both serum leptin and adiponectin compared to normal nondiseased mice. Treatment with omaveloxolone or TX63682 further decreased leptin values, while increasing adiponectin values back toward normal levels. Leptin, an adipokine, is known to contribute to hepatic disrepair and fibrogenesis often observed in response to liver injury. [40] Leptin has also been shown to suppress Nrf2 activity, and the decrease in leptin observed with omaveloxolone or TX63682 treatment in this study is considered to be hepatoprotective and potentially related, at least in part, to the reduction in collagen deposition that was observed. [41] Adiponectin is a potent anti-inflammatory adipocytokine that has suppressive effects on tumor necrosis factor-α-mediated nuclear factor κB activity and has been shown to prevent liver fibrosis. [42] As such, lower F I G U R E 7 Effect of omaveloxolone and analog TX63682 on serum transaminases. Serum ALT (A), AST (B), GGT (C), and total bilirubin (D) were quantified using a FUJI DRI-CHEM7000 chemistry analyzer. Data are presented as mean ± SEM. Asterisks indicate a statistically significant difference from the vehicle control group (*P < .05, **P < .01). ALT, alanine transaminase; AST, aspartate transaminase; GGT, gamma-glutamyl transpeptidase; Omav, omaveloxolone adiponectin levels are also considered to be an independent risk factor for NASH. [43] The hepatoprotection observed with omaveloxolone and TX63682 in the STAM model of NASH presented herein is consistent with data from other models of liver injury that evaluated various analogs of omaveloxolone. For example, an analog suppressed hepatic steatosis and liver fibrosis in mice fed a high-fat/high-fructose diet. [44] However, the STAM model of NASH is considered to be more clinically relevant because the progression of disease closely mirrors the known progression in humans but within a much shorter time frame. [23] Significant hepatoprotection with omaveloxolone analogs has also been demonstrated in acetaminophen-induced hepatotoxicity, [45] chronic liver injury from carbon tetrachloride, [46] and ischemia-reperfusion. [47] Thus, the overall hepatoprotective pharmacology of omaveloxolone and its analogs are consistent with the profound hepatoprotective properties of Nrf2, which has been studied and reviewed extensively. [21,[48][49][50][51][52] In conclusion, omaveloxolone and TX63682 demonstrated significant hepatoprotection in a rodent model of NASH. Such hepatoprotection was observed in conjunction with improvements in glucose control and lipid handling, which is consistent with the overall phenotype of improved mitochondrial function that is associated with Nrf2 activation. Overall, these data suggest that omaveloxolone and its analogs may be useful for the treatment of liver diseases.