Time-dependent effect of REV-ERBα agonist SR9009 on nonalcoholic steatohepatitis and gut microbiota in mice

ABSTRACT The circadian clock is involved in the pathogenesis of nonalcoholic steatohepatitis (NASH), and the target pathways of many NASH candidate drugs are controlled by the circadian clock. However, the application of chronopharmacology in NASH is little considered currently. Here, the time-dependent effect of REV-ERBα agonist SR9009 on diet-induced NASH and microbiota was investigated. C57BL/6J mice were fed a high-cholesterol and high-fat diet (CL) for 12 weeks to induce NASH and then treated with SR9009 either at Zeitgeber time 0 (ZT0) or ZT12 for another 6 weeks. Pharmacological activation of REV-ERBα by SR9009 alleviated hepatic steatosis, insulin resistance, liver inflammation, and fibrosis in CL diet-induced NASH mice. These effects were accompanied by improved gut barrier function and altered microbial composition and function in NASH mice, and the effect tended to be stronger when SR9009 was injected at ZT0. Moreover, SR9009 treatment at different time points resulted in a marked difference in the composition of the microbiota, with a stronger effect on the enrichment of beneficial bacteria and the diminishment of harmful bacteria when SR9009 was administrated at ZT0. Therefore, the time-dependent effect of REV-ERBα agonist on NASH was partly associated with the microbiota, highlighting the potential role of microbiota in the chronopharmacology of NASH and the possibility of discovering new therapeutic strategies for NASH.


Introduction
As one of the most common chronic liver diseases, nonalcoholic fatty liver disease (NAFLD) is characterized by abnormal liver function caused by excessive fat deposition in the liver (Ni et al. 2020a). NAFLD has a series of pathological processes, from simple fatty liver to nonalcoholic steatohepatitis (NASH), liver fibrosis, and cirrhosis, greatly increasing the risk of liver cancer (Diehl et al. 2017). NASH is also known as metabolic steatohepatitis, which is closely related to metabolic disorders such as obesity, insulin resistance, type 2 diabetes, and hyperlipidemia (Vernon et al. 2011). Hepatocyte apoptosis induced by liver fibrosis accompanied by chronic inflammation is the main feature of NASH (Henderson et al. 2020). Over the past decades, researchers have devoted extensive resources to discovering and developing NASH drugs, and unfortunately, no approved NASH drug is available yet (Dufour et al. 2022). Therefore, new targets and more effective strategies for NASH are needed.
Increasing evidence has proved the involvement of the intestinal barrier function and the gut microbiota in the pathogenesis of NASH ). The intestinal barrier not only ensures the body obtains sufficient nutrients (Yinhua et al. 2022), but also effectively prevents pathogenic bacteria and toxic metabolites from entering the systemic circulation (Ni et al. 2020. The intestinal mucosal epithelial barrier comprises complete intestinal epithelial cells and adjacent epithelial cells, including tight junctions, adhesion junctions, and gap junctions (Assimakopoulos et al. 2018). We have previously found impaired tight junctions in the intestinal barrier during the progression of metabolic diseases, such as obesity and NASH (Ma et al. 2020;Ni et al. 2021), suggesting a potential strategy for the treatment of NASH by targeting the barrier function. On the other hand, hormones released from the gut regulate the lipid metabolism of the liver through the intestinal liver axis, and intestinal hormones are often affected by the diversity and abundance of intestinal microbiota (Ni et al. 2019). Inflammatory signals produced by intestinal microbial abnormalities cause liver inflammation, insulin resistance, and affect the susceptibility to develop obesity, liver steatosis, NASH, liver cirrhosis, and liver cancer (Sherif et al. 2016). In addition, the severity of NAFLD/NASH is directly related to intestinal microbial imbalance and metabolic dysfunction (Miele et al. 2009).
Liver homeostasis is strongly influenced by the circadian clock, and the disruption of the circadian clock has been heavily related to the progression of NASH. The metabolism activities in the host are controlled by the circadian clock, leading to temporal variability in multiple pathways involved in glucose uptake, gluconeogenesis, and lipogenesis (Saran et al. 2020). Genetically deletion of core circadian clock genes or desynchronization of the circadian clock could cause a range of adverse metabolic consequences, including insulin resistance, hepatic steatosis, and even heptocarcinogenesis (Kettner et al. 2016;Turek et al. 2005). More importantly, the target pathways of many current NASH candidate drugs are under the control of the circadian clock, such as fatty acid synthesis and signaling via the peroxisome proliferator-activated receptor α and γ, thyroid hormone receptor, farnesoid X receptor, fibroblast growth factor 19 and 21, and glucagon-like peptide 1 (Marjot et al. 2022). However, the application of chronopharmacology in NASH is little considered currently.
REV-ERBs are typical nuclear receptors acting as master regulators of metabolism, mitochondrial biogenesis, inflammatory response, and fibrosis (Raza et al. 2022). A number of synthetic REV-ERB agonists have been developed for the treatment of relevant diseases regarding their role in positively influencing dysregulated metabolism and inflammation (Griffett et al. 2022). For example, REV-ERBα agonist SR9009 treatment reduces fat content and significantly improves dyslipidemia and hyperglycemia in diet-induced obese mice (Griffett et al. 2020). A recent study also found that the beneficial effects of REV-ERBα activation by SR9009 are associated with an overall improvement of hepatic health by suppressing hepatic fibrosis and inflammatory response in mice (Griffett et al. 2020). However, a broad range of the effect of REV-ERB activation in NASH mice remains exclusive. We have previously found that REV-ERBα directly regulates the tight junction protein, and the activation of REV-ERBα protects against the impaired gut barrier function, which might be partly related to the amelioration of NASH in mice (Ni et al. 2021). However, the time-dependent effect of REV-ERB activation on microbiota and the involvement of gutliver interaction in SR9009-mediated alleviation of NASH remains unknown.
To that end, the REV-ERB agonists, SR9009, was administrated at ZT0 and ZT12 in the NASH mice, and the time of SR9009 administration was chosen based on the rhythm of Rev-erbs in the liver (Ni et al. 2021) and other studies. For example, REV-ERB agonist SR9009 is found to be injected at ZT0 when REV-ERB activity is normally low or ZT0 and ZT12 (Amador et al. 2016;Griffett et al. 2020;Hong et al. 2021;Roby et al. 2019).

Animals and experimental design
Eight-week-old male C57BL/6J mice were purchased from the China National Laboratory Animal Resource Center (Shanghai, China). All mice were kept in a temperature-controlled room (22°C ± 2°C) under a 12/12 h light/dark cycle. The on and off times of light were defined as Zeitgeber time 0 (ZT0) and ZT12, respectively. Water and food were available ad libitum. After one week of adaptation, the mice were fed with normal chow (NC, P1101F-25, Salcom Co., Ltd., Shanghai, China) or a high-cholesterol and high-fat diet (CL, 60% fat, 1.25% cholesterol, 0.5% sodium cholate, D06061403, Research Diets, Brunswick, NJ) for 12 weeks to induce NASH as described previously (Ni et al. 2021). To compare the chronopharmacological effect of REV-ERBα agonist SR9009 on NASH, mice were then divided into four groups and treated for another 6 weeks as follows: (1) NC, mice were continued to feed an NC diet and injected with saline at ZT0, n = 6; (2) CL, mice were continued to feed a CL diet, n = 8; (3) ZT0, mice were continued to feed a CL diet and injected with SR9009 (100 mg/kg/day, i.p.) at ZT0, n = 8; (4) ZT12, mice were continued to feed a CL diet and injected with SR9009 (100 mg/kg/day, i.p.) at ZT12, n = 8. The food intake was measured during the experiment, and all mice were sacrificed after overnight fasting, and blood and tissue samples were collected and stored at−80°C until use. All experiments were approved by the laboratory animal ethical committee of Zhejiang University of Technology and followed the NIH Guide for Laboratory Animals (NIH Publication No. 85-23, revised 1996) for the Care and Use of Animals.

Biochemical analysis
Plasma levels of triglycerides (TGs), total cholesterol (TC), nonesterified fatty acids (NEFAs), alanine aminotransferase (ALT), aspartate aminotransferase (AST), insulin, and hepatic TG, TC, and NEFA levels were measured as described previously . After 4 weeks of SR9009 treatment, a glucose tolerance test (GTT) was performed after 16 h of fasting by intraperitoneal injection of glucose (2 g/kg BW). Blood glucose levels were measured at the indicated time points. Homeostatic model assessment-insulin resistance (HOMA-IR) was calculated using the following equation: fasting glucose (mg/dL) × fasting insulin (μU/ mL)/405.

Determination of intestinal permeability
Mice were fasted for 4 h and gavaged with 0.5 mg/g FITCdextran (46944-500 MG-F, Sigma). Blood was taken from the tail vein after 4 h, and fluorescence in the plasma was monitored by a spectrophotofluorometer using excitation at 485 nm and emission at 528 nm. A standard curve was obtained by serial dilutions of the stock dextran-FITC solution.

Quantitative RT-PCR analysis
Total RNA from the liver and colon was extracted using Biozol reagent (Biomiga, San Diego, CA, USA) according to the manufacturer's instructions. cDNA was synthesized using the GoldenstarTM RT6 cDNA Synthesis Kit (Tsingke Co., Ltd, Beijing, China). Quantitative real-time PCR (qPCR) was performed on a CFX Connect Optics Module (Bio-rad, USA) using Fast qPCR Mix (SYBR Green, Tsingke Co., Ltd, Beijing, China). The primers used in the current study were same with our previous study (Ni et al. 2021). Gene expression was normalized to glyceraldehyde 3-phosphate dehydrogenase (Gapdh). The results are shown as the fold changes compared with the control group.

Histological examination
For histological evaluation of the liver and colon, tissues were fixed in 4% paraformaldehyde and embedded with paraffin. Hematoxylin & eosin (H&E) staining, Alcian blue-periodic acid Schiff (AB-PAS) staining, and Sirius red staining were performed as described previously (Ni et al. 2021). The goblet cell number was independently assessed in colonic tissue sections, and only crypts that were cut longitudinally from the crypt opening to the bottom of the crypt were considered. The positive staining of mucin proteins was morphometrically quantified by IpWin 60 software and presented as a percentage of the field examined. For immunohistochemistry, liver sections were stained with F4/80 (Cell Signaling, 70076T) and α-SMA (Proteintech, 14395-1-AP), and the quantification of the positive area was analyzed, as described previously (Ni et al. 2020).

16S rRNA sequencing and data analysis
According to the manufacturer's instructions, total genomic DNA from cecal contents was extracted using Ezup Column Soil DNA Purification Kit (Sangon, Shanghai, China). The V3-V4 region of 16S rRNA genes was amplified by a specific primer (338F-806 R). DNA libraries were validated by Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA) and quantified using PicoGreen dsDNA quantitation Reagent (Yeasen, Shanghai, China) by UV spectroscopy. The Illumina Hiseq sequencing was conducted at Novogene (Tianjin, China). Raw sequence data were demultiplexed using the demux plugin followed by primers cutting with cutadapt plugin. Sequences were then quality filtered, denoised, merged and chimera removed using the DADA2 plugin. Non-singleton amplicon sequence variants (ASVs) were aligned with mafft and used to construct a phylogeny with fasttree2. The composition of the microbiota was mainly determined by QIIME2 bioinformatic analysis and R packages (v3.2.0) ).

Statistical analysis
All data are given as the mean ± SEM. The student's t-test was performed comparisons between two different groups. Multiple-group comparisons were performed by analysis of variance (ANOVA). A p value < 0.05 was considered statistically significant. Pearson correlation analysis of differentially enriched species was performed using the R package (v3.2.0) as described previously ).

Time-dependent effect of SR9009 on hepatic steatosis in NASH mice
To determine whether the administration of REV-ERBα agonist at different times could affect its therapeutic effect on NASH, mice were intraperitoneally injected with SR9009 either at ZT0 or ZT12. SR9009 treatment significantly reduced the weight of NASH mice without affecting the food intake, while no significant difference was found between the two SR9009-treated groups (Figures 1a and S1b). The decrease in weight gain caused by SR9009 was mainly due to the significant reduction in fat mass and adipocyte size rather than muscle weight (Figure S1c-e). The liver weight of NASH mice increased slightly and decreased markedly by SR9009 treatment at ZT0 but not at ZT12 (Figure 1b). In addition, the elevated plasma levels of AST and ALT in CL-fed mice were decreased by SR9009 treatment at both time points (Figure 1c). H&E staining showed severe lipid accumulation in the liver of mice fed the CL diet, which was attenuated by SR9009 treatment at both time points (Figure 1d). Consistent with these results, plasma TC was significantly decreased by SR9009 at ZT0 and ZT12, and plasma NEFA level was only reduced by SR9009 at ZT0, though plasma TG level was not affected (Figure 1e). Moreover, hepatic lipid contents, including TG and NEFA levels, were all attenuated by SR9009 at both time points, and liver TC level was only reduced by SR9009 at ZT0 (Figure 1f). Next, the mRNA expression of lipid metabolism-related markers was assessed by qPCR. Firstly, the expression of Cd36, a key regulator of lipid uptake, was up-regulated in the liver of NASH mice, and it was significantly down-regulated by SR9009 treatment at ZT0 but not ZT12 (Figure 1g). Secondly, the expression of lipogenic genes, such as acetyl-coenzyme A carboxylase (Acc), sterol-regulatory element binding protein-1c (Srebp-1c), and fatty acid synthase (Fas), were all decreased by SR9009 at ZT0 (Figure 1g). No significant difference was found in the SR9009 treatment at ZT12 (Figure 1g). On the other hand, SR9009 treatment upregulated the expression of Rev-erbα, Rev-erbβ, and Bmal1 in the liver of NASH mice. Specifically, the effect of SR9009 on Rev-erbα expression was stronger at ZT0 than at ZT12 ( Figure S2).

Time-dependent effects of SR9009 on insulin resistance in NASH mice
The results of GTT showed that SR9009 significantly improved glucose intolerance induced by the CL diet, and the effect at ZT0 tended to be stronger than that of ZT12 (Figure 2a). In addition, the elevated fasting blood glucose level in NASH mice was significantly reduced by SR9009, with the effect at ZT0 more potent than ZT12 (Figure 2b). CL-induced hyperinsulinemia both at the fasting and fed state was also attenuated by SR9009 treatment at ZT0, but not ZT12 (Figure 2c,d). Consistent with these findings, the HOMA-IR index was increased significantly in NASH mice, and it was substantially decreased by SR9009 at both time points (Figure 2e). These results were associated with enhanced insulin-stimulated Akt phosphorylation in the liver and epididymal white adipose tissue (eWAT) of SR9009treated mice compared with CL-fed NASH mice (Figure 2f).

Time-dependent effects of SR9009 on hepatic inflammation and fibrosis in NASH mice
The effects of REV-ERBα activation on hepatic inflammation and fibrosis in NASH were next evaluated. F4/80 staining, as a marker of hepatic macrophage, indicated that SR9009 attenuated CL diet-induced liver macrophage activation at both time points (Figure 3a,b). qPCR analysis of pro-inflammatory cytokines expression, including interleukin 1β (Il1β), Il6, and tumor necrosis factor α (Tnfα), in the liver of NASH mice found that SR9009 treatment at ZT0 significantly downregulated the expression of all these markers, while no significant difference was found when treated at ZT12 (Figure 3c). On the other hand, Sirus red and α smooth muscle actin (α-SMA) staining revealed that clear liver fibrosis and the activation of hepatic stellate cells in the liver of NASH mice, which was ameliorated by SR9009 (Figure 3a,b). Similarly, the mRNA expression of the fibrogenic genes, such as transforming growth factor β (Tgf-β) and α-Sma, was also downregulated by SR9009 at both time points (Figure 3d).

Time-dependent effects of SR9009 on the intestinal barrier function
Consistent with our previous study, the colon length of NASH mice was shorter than that of NC-fed mice, and SR9009 significantly increased colon length at ZT0 but not at ZT12 (Figure 4a). In addition, the increased intestinal permeability induced by the CL diet was improved by SR9009, as determined by the level of FITC-labeled dextran and plasma LPS levels, and the effect of SR9009 on FITC dextran was more potent at ZT0 (Figure 4b,c). Histology analysis with H&E and AB-PAS staining further revealed that CL diet-induced increase in the inflammatory cells infiltration was attenuated by SR9009 at both time points, and the decrease of mucin secretion was improved by SR9009 at ZT0 (Figure 4d,e). Moreover, the mRNA expression of tight junction-related markers in the colon, such as claudin 1 (Cldn1), Cldn3, occludin (Ocln), zonula occludin-1 (Zo-1), Zo-2, and Zo-3, was downregulated in the colon of NASH mice, which was reversed by SR9009 treatment at both time points (Figure 4f). However, the expression of mucin-related marker, mucin-1 (Muc-1) and antimicrobial peptides-related markers, such as regenerating family member 3 gamma (Reg3g) and α-defensins (Defa), was only up-regulated by SR9009 at ZT0, but not at ZT12 (Figure 4f). In contrast, the mRNA expression of inflammatory cytokines, including Il6 and Tnfα, were decreased by SR9009 at both time points, and the expression of Il1β was only decreased by SR9009 at ZT0 (Figure 4f). Consistently, the protein levels of Claudin-1 and Occludin in the colon were increased by SR9009 (Figure 4g). Moreover, the autophagy-enhancing and apoptosis-inhibiting effect of SR9009 was also verified by the level of LC3B and cleaved Caspase 3 in the colon (Figure 4g).

Time-dependent effects of SR9009 on the composition and function of gut microbiota
Next, the effect of SR9009 on the gut microbiota was assessed by 16S rRNA sequencing of cecal content. The principal coordinate analysis (PCoA) revealed clear discrimination in the β-diversity among NC-fed control, NASH, and different time SR9009-treated mice, leading the SR9009-treated mice close to that of control mice (Figure 5a). Further analysis of the overall composition of microbiota at the phylum level indicated that the abundance of Firmicum increased by 10.75%, while the abundance of Bacteroides decreased by 37.68% in the NASH mice, and this dysregulated F/B ratio was reversed by SR9009 mainly at ZT0, but not at ZT12 (Figure 5b). In addition, the phylogenetic dendrogram obtained by LEfSe analysis showed significant differences in the taxonomic groups among all groups (Figure 5c). Consistent with these findings, heat map analysis of top 30    (Figure 5d). These results were associated with marked alterations in the microbial function of NASH and different time SR9009-treated mice (Figure 5e). Specifically, the functions related to fatty acid and lipid biosynthesis, carboxylate degradation, cofactor, prosthetic group, electron carrier, and vitamin biosynthesis, generation of precursor metabolite and energy, and amino acid biosynthesis were all significantly alter by CL diet feeding (Figure 5e). SR9009 treatment at ZT0 greatly reversed these microbial function alterations, while SR9009 treatment at ZT12 slightly affected these functions (Figure 5e).
Next, Pearson correlation analysis of the top 40 abundant bacterial genera and the NASH parameters was performed to identify the bacteria that might be responsible for the effects of SR9009 in NASH mice. The results found that several genera displayed strong positive or negative correlations with plasma parameters, including insulin-and lipid-related parameters and hepatic lipid levels (Figure a). Specifically, Oscillospiraceae uncultured, Incertae Sedis, Colidextribacter, Desulfovibrionaceae uncultured, and Anaerotruncus were positively correlated with these parameters, and Blautia, Lachnospiraceae NK4A136 group, Muribaculaceae, Faecalibaculum, and Dubosiella were negatively correlated with these parameters (Figure 6a). Further correlation analysis of these bacteria genera and microbial function revealed that those bacteria, which positively or negatively correlated with NASH parameters, also displayed distinguished functions related to lipid metabolism (Figure 6b). In addition, SR9009 treatment at different time points also showed different effects on the abundance of particular bacterial taxa. In detail, the increased abundance of Oscillospirales uncultured, Colidextribacter, Incertae Sedis and Desulfovibrionaceae uncultured were all decreased significantly by SR9009 treatment at ZT0, while only the first two genera were decreased by SR9009 treatment at ZT12 (Figure 7a-d). In contrast, the abundance of Eubacterium coprostanoligenes group, Lachnospiraceae NK4A136 group, Muribaculaceae, and Akkermansia was significantly increased by SR9009 at ZT0, but not ZT12 (Figure 7e-h).

Discussion
In the present study, we found that the time-dependent effect of REV-ERBα agonist SR9009 on NASH tended to be stronger when treated at ZT0, evidenced by alleviated hepatic steatosis, insulin resistance, inflammation, and fibrosis, and improved gut barrier function. Moreover, CL diet-induced NASH was accompanied by gut microbiota dysbiosis, and different time treatments with SR9009 resulted in marked differences in the composition and function of microbiota. Thereby, the timedependent effect of SR9009 might be associated with the different impacts of SR9009 on the microbiota. However, the relatively minor chronotherapeutic effects of REV-ERB activation on the NASH phenotypes might be caused by the terminal sample collection method. In addition, our results suggested that SR9009 activated the expression of clock genes, and these changes could lead to a more robust rhythm of the liver, which might contribute to the enhancement of liver functions. Though REV-ERBs function to repress Bmal1 expression, the feedback loop may then enhance the Bmal1 expression to preserve the robust oscillation of the liver clock. However, how SR9009 alters the circadian clock remains unclear. Further studies that performed multiple time points of sample collection would provide more details about the chronopharmacological effects of SR9009 and its impact on the circadian rhythm.
An increased number of researches have demonstrated that the gut barrier is involved in the pathogenesis of NASH, in addition to its defensive role against the invasion of harmful substances from the intestines (Cui et al. 2019). The intestinal permeability is increased in NAFLD patients and is associated with the degree of hepatic steatosis (De Munck et al. 2020). Therefore, the gut barrier has emerged as a novel target for preventing and treating NAFLD/NASH (Ni et al. 2020). In our previous study, we found that REV-ERBα directly bound to the promoters of tight junction genes to regulate the intestinal permeability, and chronopharmacological activation of REV-ERBα by SR9009 protected against lipopolysaccharide (LPS)-induced increased intestinal permeability (Ni et al. 2021). Altered gut barrier function is associated with changes in the composition and function of the microbiota, and the main purpose of the present study is to determine whether the effect of REV-ERB activation by SR9009 on NASH is related to microbiota. Therefore, the same time points were selected in the present study, and consistent with the acute effect of SR9009 in LPS-treated mice, we found that the chronic effect of SR9009 on gut barrier function in NASH mice also tended to be stronger at ZT0 than at ZT12. These results suggested that clock gene-targeted therapies should follow the rhythm of the circadian clock, despite the acute or chronic utilization.
REV-ERBα acts as a transcription inhibitor due to the lack of the coactivator's binding domain and activation function (Burke et al. 1996). Several studies have shown that REV-ERBα plays an important role in lipid metabolism (Le Martelot et al. 2009), such as regulating SREBP-1C, a key transcription regulator in fatty acid synthesis (Yang et al. 2021). Activation of REV-ERBα by SR9009 is found to downregulate the mRNA expression of Srebp-1c and Fas in the liver of NASH mice (Raspe et al. 2002). In contrast, the lack of REV-ERBα results in dyslipidemia and impaired muscle function, indicating the pharmacological potential for the treatment of related diseases (Ruan et al. 2021;Woldt et al. 2013). Consistent with these findings, our results also found that the activation of REV-ERBα ameliorated hepatic steatosis by downregulating the expression of lipogenic genes in the liver. One recent study suggested that REV-ERBα is implicated in the alteration of β-cell autophagy and survival in vitro, thereby may be involved in the pathogenesis of type 2 diabetes (Brown et al. 2022). Here we also found the amelioration of glucose intolerance and insulin resistance by SR9009 in NASH mice, while the detailed mechanism remained unknown due to the tissue-and disease-specific effect of REV-ERBα.
Hepatic inflammation and fibrosis are key features during the progression of NASH, and the circadian rhythm controls a series of physiological processes, including the immune response and inflammation (Xu et al. 2021). The inflammation status may affect the transcription of the core clock genes, such as Rev-erbα, and its circadian oscillation (Yang et al. 2014). Studies have found that REV-ERBα controls the homeostatic regulation of inflammation in the lung by interacting with the pro-inflammatory cytokines (Pariollaud et al. 2018). In addition, Rev-erbα modulates the inflammatory function of macrophages directly through a Rev-erbα-binding motif in the Ccl2 promoter region and suppresses CCL2-activated pathways (Sato et al. Figure 6. Correlation analysis of NASH parameters and gut microbiota. (a) correlation analysis of top 40 genera with NASH-related parameters between the control, CL, and SR9009-treated mice. Rows correspond to NASH-related parameters, and columns correspond to the specific genus. Red and blue colors denote positive and negative associations, respectively. The intensity of the colors represents the degree of association between the abundance of bacteria and host parameters assessed by Pearson correlation analysis. Stars mean p < 0.05. (b) Correlation analysis between the top 40 bacteria genera and microbial function related to lipid metabolism. Row corresponds to lipid metabolism-related functions, and columns correspond to the specific genus. Stars mean p < 0.05. 2014). We consistently found that the activation of REV-ERBα attenuated the expression of inflammatory cytokines, thereby alleviating hepatic inflammation in NASH mice.
Gut microbiota plays a key role in the progress of NASH, and the microbial composition of patients with NASH has changed significantly (Ma et al. 2020). For example, the abundance of Proteus and Scleroderma in NASH patients increased significantly, while Bacteroides decreased significantly (de Sant'Ana LP et al. 2019). The overrepresentation of Lachnospiraceae bacteria 609 and Barnesiella intestinihominis was found to have a potency to induce NAFLD (Le Roy et al. 2013). Thus, microbial community disorder and the destruction of metabolites derived from the microbiota are now found to be related to liver steatosis, inflammation, and fibrosis (Li et al. 2011;Perez and Briz 2009). On the other hand, diurnal variations in the microbiota also affect the circadian rhythms in the host to regulate many core processes, such as metabolism and immune response (Brooks et al. 2021;Frazier and Chang 2020;Kuang et al. 2019). In addition to the microbial changes in the NASH mice, we found that the activation of REV-ERBα itself caused a marked alteration in the composition and function of microbiota, and different time points treatment with SR9009 also resulted in discriminated microbial phenotypes. Importantly, the microbiota differences in ZT0 and ZT12 SR9009 treatment were associated with significant changes in microbial function, and the dysregulated functions related to fatty acids and lipid biosynthesis tended to be amended more effectively in SR9009 treatment at ZT0. These changes might also be partly responsible for the effect of SR9009 on NASH phenotypes.
correlate positively with oxidative stress and hyperlipidemia-related parameters in the serum (Duan et al. 2021;Wang et al. 2021). Desulfovibrionaceae is abundant in the NASH model and strongly correlated with obesity and metabolic syndrome in mice (Panasevich et al. 2018;Ussar et al. 2015). In addition, Eubacterium coprostanoligenes is a cholesterol-reducing anaerobe (Li et al. 1996), and Lachnospiraceae NK4A136 is related to short-chain fatty acids synthesis, as verified by our previous study (Ma et al. 2020). Muribaculaceae is enriched after dietary interventions with compound fibers , and Akkermansia is a well-known next-generation beneficial microorganism with various health-promoting effects (Cani et al. 2022). In our study, we found that SR9009 treatment was strongly correlated with the abundance of these bacteria, and the differences between different time of SR9009 administration in the abundance of these bacterial genera may partly be associated with the efficiency of SR9009 for NASH treatment. Our results highlighted that microbial changes might also affect the chronopharmacological effect of the potential NASH drugs, and further studies are needed to clarify the role and mechanism of circadian rhythm of microbiota and the pathogenesis of NASH.

Conclusion
In summary, pharmacological activation of REV-ERBα by SR9009 alleviated hepatic steatosis, insulin resistance, liver inflammation, and fibrosis in CL diet-induced NASH mice. These effects were accompanied by improved gut barrier function and altered microbial composition and function in NASH mice, and the effect tended to be more potent when SR9009 was injected at the valley time of the Rev-erbα rhythmicity. The chronobiological effect of SR9009 might be partly associated with enriching beneficial bacteria and diminishing harmful bacteria, in addition to its role in tight junction and autophagy. Therefore, our study highlighted the involvement of gut-liver interaction in SR9009mediated alleviation of NASH. While more research is warranted to further understand the role of REV-ERB in NASH and microbiota, and these studies might contribute to the discovery of new therapeutic strategies.

Disclosure statement
No potential conflict of interest was reported by the author(s).