The hypolipidemic effect of MI-883, the combined CAR agonist/ PXR antagonist, in diet-induced hypercholesterolemia model

Abstract

Constitutive androstane receptor (CAR) and pregnane X receptor (PXR) are closely related nuclear receptors with overlapping regulatory functions in xenobiotic clearance but distinct roles in endobiotic metabolism. Car activation has been demonstrated to ameliorate hypercholesterolemia by regulating cholesterol metabolism and bile acid elimination, whereas PXR activation is associated with hypercholesterolemia and liver steatosis. Here we show a human CAR agonist/PXR antagonist, MI-883, which effectively regulates genes related to xenobiotic metabolism and cholesterol/bile acid homeostasis by leveraging CAR and PXR interactions in gene regulation. Through comprehensive analyses utilizing lipidomics, bile acid metabolomics, and transcriptomics in humanized PXR-CAR-CYP3A4/3A7 mice fed high-fat and high-cholesterol diets, we demonstrate that MI-883 significantly reduces plasma cholesterol levels and enhances fecal bile acid excretion. This work paves the way for the development of ligands targeting multiple xenobiotic nuclear receptors. Such ligands hold the potential for precise modulation of liver metabolism, offering new therapeutic strategies for metabolic disorders.

Fig. 1. MI-883 is a CAR agonist and PXR antagonist/inverse agonist.

a Formula of compound MI-883. b MI-883 interacts with the recombinant human CAR LBD in the TR-FRET CAR coactivation assay (n = 3). c MI-883 activates human CAR in the CAR LBD assembly assay (n = 3). d MI-883 activates the CAR3 variant in the luciferase gene reporter assay with an expression vector for human CAR variant 3 (XM_005245697.4) in HepG2 cells treated with CITCO or MI-883 for 24 h (n = 3). e MI-883 stimulates EGFP-CAR+A fusion protein cytoplasm-nucleus translocation in COS-1 cells (magnification 40×)(n = 3). f MI-883 significantly induces CYP2B6 mRNA in HepaRG and g primary human hepatocytes, but not in HepaRG KO CAR cells as determined using RT-qPCR expression analysis. h MI-883 inhibits the binding of a fluorescent probe to the PXR LBD in TR-FRET PXR LBD Competitive assay. i MI-883 inhibits activation of CYP3A4 gene promoter-based luciferase construct (CYP3A4-luc) in the absence and presence of rifampicin in luciferase reporter gene assay performed in HepG2 cells treated with rifampicin (Rif) and MI-883 for 24 h (n = 3). j MI-883 does not activate PXR LBP with mutations (S208W/S247W/C284W, mutant PXR LBD) in luciferase reporter gene assay with a CYP3A4 gene promoter-based luciferase construct in HepG2 cells treated with rifampicin (RIF) and MI-883 for 24 h (n = 5). ***P value < 0.001; statistically significant effects vs. wtPXR LBD control (ctrl) sample. n.s. not statistically significant. k MI-883 stimulates PXR LBD assembly with a fusion protein of GAL4 and helix 1 (residues 132–188) and a fusion protein of VP16 and helix 2-12 part (residues 189-434 of PXR-LBD) in HepG2 cells (n = 7). l MI-883 suppresses rifampicin- and SR12813-mediated CYP3A4 mRNA induction and basal CYP3A4 mRNA expression in LS174T cells. m The PXR-responsive CYP3A4(pER6)-luc luciferase reporter construct is inhibited by MI-883 in the presence or absence of PXR ligands rifampicin (Rif) and SR12813 in transiently transfected HepG2 cells treated for 24 h. *^ƒP value < 0^.05; ** P < 0.01; ***^ƒƒƒ P < 0.001; statistically significant effects vs. control (ctrl, *) or vs. PXR ligand-mediated effect (^ƒ) (ANOVA with multiple comparisons). All experiments are biological triplicates (n ≥ 3) each done in technical triplicates. Data represent means ± S.D.

Fig. 2. Molecular dynamics (MD) simulations analysis of wtCAR-LBD complexed with CITCO and MI-883, and PXR-LBD complexed with SR12813 and MI-883.

a, b Proposed binding mode of CITCO and MI-883 in the LBD of wtCAR represented by the most populated cluster conformation from the MD simulations, performed using hierarchical clustering of the backbone atoms from the aligned trajectory. Residues performing H-bond are labeled in bold. c Protein–ligand hydrophobic interactions include: π−π contacts and hydrophobic interactions. d Protein–ligand hydrogen bond (H-bond) and water-mediated interactions. Each simulation replica runs for 1 microsecond (μs) (5 independent replicas were conducted). e Distance between Helix3 (H3, center of mass of residues 157-178) and Helix12 (H12, center of mass of residues 341-348) in CAR-CITCO and CAR-MI 883. Boxplots span from 25% to 75% quartiles (IQR: 25–75%); the black line is the median. Minimum and maximum are the smallest and the largest data points, respectively. Whiskers are the data within 1.5*IQ. Outliers are indicated with diamonds. Distances were monitored each ns, i.e., there are: ∼5000 individual datapoints for each ligand. Full data is available in Zenodo repository. f Distance between K195 (side chain) and S348 (oxygen atoms). g, h A representative snapshot of PXR-LBP with SR12813 and MI-883. i Protein–ligand hydrophobic interactions include: π−π contacts and hydrophobic interactions. j Protein–ligand H-bond and water-mediated interaction. k Box plot represents the distribution of the distance between the center of mass of Helix3 (H3, residues 240–260) and Helix 12 (H12, residues 423–430). l Cα–Cα distance between C207 (of H2′) and A312 (of β4-H6 loop). m Heatmap of MARCoNI array data indicating interactions of human CAR with coactivators and corepressors in the presence of CITCO, MI-883, and clotrimazole, an inverse agonist (all at 100 μM). Modulation index (MI) is the log₁₀-transformed relative binding value, which is calculated by the compound’s binding value, relative to the vehicle control. red color -stimulated interaction, blue color -release of interaction. n Mammalian two-hybrid assays for PXR/coactivator NCOA1, NCOA2, and MED1 recruitment and PXR/corepressor NCOR2 release. Transfected cells were treated with vehicle or 10 µM rifampicin (RIF) and/or the indicated concentrations of MI-883 for 24 h (n = 5). *P value < 0.05; ** ††P value < 0.01; ***P value < 0.001; statistically significant effects vs. control (*) or vs. rifampicin-mediated effect (†) (ANOVA). Data represent means ± S.D.

Fig. 3. Demonstration of CAR agonistic/PXR antagonistic activities of MI-883 in HepaRG cells, 2D and 3D primary human hepatocytes, and humanized PXR-CAR-CYP3A4/3A7 mice.

a CAR and PXR superimposition demonstrate the similarity of the related receptors of the NR1I group and their ligand binding domains (LBDs). Induction of CYP3A4 mRNA and CYP2B6 mRNA b expression in HepaRG, HepaRG KO PXR, HepaRG KO CAR cells or c 2D primary human hepatocytes (PHH, Biopredic, two donors) after treatment with MI-883 (range from 1 to 10 or 20 µM) for 24 h. Each analysis in HepaRG cells was done using independent triplicates; data in PHH represent a mean of technical triplicates from each donor. d Induction of CYP3A4 and CYP2B6 mRNA in 3D human hepatocyte spheroids (3D PHH) from three donors after treatment with MI-883 (5 μM) over 120 h in culture. Data are means±S.D. from 3D hepatocyte cultures from tree donors. e Hepatic induction of CYP3A4 and Cyp2b10 mRNA in humanized PXR-CAR-CYP3A4/3A7 mice after i.p. treatment of randomly assigned animal groups (n = 3-4) with the vehicle, rifampicin (Rif, 3×5 mg/kg, every 24 h for 3 days), MI-883 (10 mg/kg every 24 h for 3 days) or their combination. CITCO was applied at 10 mg/kg in two doses after 24 h. Data in bar charts represent means±S.D. f, g mRNA expression of key genes involved in cholesterol synthesis (SQLE), cholesterol metabolism regulation (SREBF1, SREBF2, INSIG1, LDLR), and in bile acid synthesis (CYP7A1) in HepaRG, HepaRG KO CAR or HepaRG KO PXR cells, and h 2D PHHs (OFA donor) after treatment with CITCO (1 μM), SPA70 (10 μM) and MI-883 (5 μM) in the presence or absence of rifampicin 5 μM (rif) for 48 h. All experiments with HepaRG cell line have been done at least in three independent triplicates (n = 3). Data represent means ± S.D. *^ƒP value < 0.05; ***P value < 0.001; statistically significant effects vs. control (*) or vs. rifampicin-mediated (^ƒ) effect in HepaRG cells or animals (ANOVA with multiple comparisons). ns - nonsignificant.

Fig. 4. Metabolic effects of MI-883 in the model of diet-induced hypercholesterolemia with high-fat diet (HFD) or high-cholesterol diet (HCD).

a Both male and female humanized PXR-CAR-CYP3A4/3A7 mice were fed for 8 weeks with HFD and treated with MI-883 (5 mg/kg 3× per week) via p.o. gavage for one month (n ≥ 15). In parallel, control group were treated with vehicle via p.o. gavage. b Body weight (BW), liver weight (LW), LW/BW ratio, and c food intake were monitored. d Representative microscopic (H&E staining, bar 100 μm) and macroscopic pictures of the livers in the vehicle- or MI-883-treated mice fed with HFDs. e Plasma total cholesterol, LDL cholesterol, HDL cholesterol, and triglyceride levels were assessed at the end of the studies after 3 h of fasting. f Both male and female humanized PXR-CAR-CYP3A4/3A7 mice were fed for two weeks with a high-cholesterol diet HCD (n ≥ 7). MI-883 (5 mg/kg 3× per week) was applied via p.o. gavage for 10 days. (5 mg/kg every other day). g Body weight (BW), liver weight (LW), LW/BW ratio were monitored. h Plasma total cholesterol, LDL cholesterol, HDL cholesterol, triglyceride, and glucose levels were assessed at the end of the studies after 3 h of fasting. In parallel, control groups were treated with vehicle via p.o. gavage. Animals were randomly and independently assigned to control or treatment groups before interventions. *P value < 0.05; **P value < 0.01; ***P value < 0.001, the significant effect of MI-883 to vehicle-treated control mice (Mann–Whitney U test, two-sided). Data are presented in box-and-whisker representing the 25th to 75th percentiles (box) and median (line). Whiskers represent minimum to maximum values. The dotted lines represent average plasma concentrations in humanized PXR-CAR-CYP3A4/3A7 mice on a chow diet. The figure in panel a was created using the BioRender tool.

Fig. 5. Lipidomic and bile acid analyses of cholesterol, bile acids, and lipids in plasma, stool, and livers of humanized PXR-CAR-CYP3A4/3A7 mice treated with MI-883 in the study with the high-fat diet.

a Total liver cholesterol (Ch) content, free cholesterol, cholesteryl esters (CE), cholesterol precursor lathosterol, and b bile acid precursor C4 contents were analyzed using LC-MS in the male liver samples (n ≥ 8). c Plasma bile acid concentration in males (n = 7), d total liver bile acid (BA) content, and e total BA content in stool per 24 h of males hPXR/hCAR/hCYP3A male mice treated with MI-883 (both n = 6). f Bile acid analysis of individual acids in stool (n = 6). g Enzymatic analysis of triglyceride (TG) content in the livers (n ≥ 7). TG and glycerol levels in samples were measured using enzymatic assays. Data are presented as the difference between TG and glycerol content after spectrophotometric detection in arbitrary units (arb. units). Lipidomic analysis in the livers of MI-883-treated animals. Lipidomics - h Principal component analysis (PCA) analysis, i triglycerides (TG), sphingomyelins (SM), ether/plasmalogen phosphatidylcholines (PC-O/PC-P), phosphatidylcholines (PC), monoglycerides (MG), lysophosphatidylcholines (LPC), diglycerides (DG), ceramides (Cer), and CE contents in the male liver homogenates. Data are presented as log2 fold changes (log₂FC) of lipid compounds between the MI-883-treated and control groups, showing individual values (grey) and their means (blue) for each lipid class. j Analysis of fatty acid saturation in TG lipid class (total chain double bonds) in relation to the total chain length. k Fatty acid saturation (total chain double bonds) in relation to the total chain length in CE and l LPC lipid classes. Animals were randomly and independently assigned to control or treatment groups before interventions (five animals per groups, n = 5). Box-and-whisker plots represent the 25th to 75th percentiles (box) with median (line). Whiskers represent minimum to maximum values. *P value < 0.05; statistically significant increase or decrease in lipid or individual bile acid content to control (vehicle-treated) animals (Mann–Whitney U test, two-sided).

Fig. 6. Next-generation sequencing transcriptomic, RT-qPCR, and Western blotting data of liver samples in humanized PXR-CAR-CYP3A4/3A7 male mice treated with MI-883 in the study with the high-fat diet.

NGS RNA-Seq transcriptomic data. a Principal component analysis (PCA) analysis, b volcano plot and c GO pathway analyses of top eight up-regulated and down-regulated pathways, and d heatmap of genes significantly regulated including the key genes of cholesterol homeostasis. e Detailed RNA-seq data of selected genes involved in cholesterol homeostasis significantly impacted by MI-883. f Expression of Mrp4 transporter, conjugation enzymes, and g, h enzymes involved in bile acid synthesis analyzed using RT-qPCR or Western blotting in the livers of MI-883-treated and control mice. i Expression of Srebf1/Srepb1 and Srepf2/Srepb2 mRNA and proteins in the livers. Western blotting of Srebp1 and Srepb2 proteins in the livers using the antibody detecting both precursor and mature nuclear forms (mSrebp−1c and mSrebp2). j Expression of Insig1 mRNA and protein employing RT-qPCR and Western blotting in the livers of MI-883-treated and control mice. k RT-qPCR and Wester blotting expression analysis of CAR target genes and bile acid transporters in the ileum of humanized PXR-CAR-CYP3A4/3A7 male mice in the model of HFD diet-induced hypercholesterolemia. Animals were randomly and independently assigned to control or treatment groups before interventions. For RNA-seq transcriptomic analysis, five independent liver samples from five animals per group (n = 5) were used. For RT-qPCR, at least ten animals (n ≥ 10) were analyzed in each group. Western blotting analyses were performed with at least four independent samples from four animals (n = 4). Violin plots represent the frequency distribution curve with the median (central line) and two quartile lines. Data in bar charts represent means ± S.D. *P < 0.05; **P < 0.01; ***P < 0.001; statistically significant expression when compared to control (vehicle-treated) animals (Mann–Whitney U test, two-sided).

Fig. 7. MI-883 does not up-regulate biomarkers of human hepatocyte proliferation in humanized PXR-CAR-CYP3A4/3A7 male mice, 3D spheroids of primary human hepatocytes, and PXB humanized liver mice.

RT-qPCR analyses of mRNA expression for pro-proliferative, antiapoptotic, and proapoptotic biomarkers in a humanized PXR-CAR-CYP3A4/3A7 mice (n = 6), b, c 3D spheroids of human hepatocytes from three donors (JEL, AKB, HLK) after 72 h treatment, and d PXB humanized liver mice (n = 5) with 70% population of liver parenchyma with human hepatocytes treated for seven days with MI-883. c Effects of MI-883 and CITCO (at 5 μM) on ATP production in 3D PHH spheroids from three donors after 72 h treatment. d Humanized PXR-CAR-CYP3A4/3A7 mice and PXB humanized liver were treated with MI-883 for 1 month (3 × 5 mg/kg per week p.o.) or 2.5 mg/kg every day for 7 consecutive days p.o. (n = 5 in each group). RT-qPCR analyses have been performed in technical triplicates of frozen liver or hepatocyte samples. Expression has been related to vehicle-treated samples (set to 1). P value, Mann-Whitney U test, two-sided. # gene with very low mRNA expression; n.d. - expression not determined - C_t expression has been detected after 35 cycles of PCR. Box-and-whisker plots show the median (line) with 25th to 75th percentiles. Whiskers represent minimum to maximum values.