Absence of Bsep/Abcb11 attenuates MCD diet‐induced hepatic steatosis but aggravates inflammation in mice

Abstract Background Bile acids (BAs) regulate hepatic lipid metabolism and inflammation. Bile salt export pump (BSEP) KO mice are metabolically preconditioned with a hydrophilic BA composition protecting them from cholestasis. We hypothesize that changes in hepatic BA profile and subsequent changes in BA signalling may critically determine the susceptibility to steatohepatitis. Methods Wild‐type (WT) and BSEP KO mice were challenged with methionine choline‐deficient (MCD) diet to induce steatohepatitis. Serum biochemistry, lipid profiling as well as intestinal lipid absorption were assessed. Markers of inflammation, fibrosis, lipid and BA metabolism were analysed. Hepatic and faecal BA profile as well as serum levels of the BA synthesis intermediate 7‐hydroxy‐4‐cholesten‐3‐one (C4) were also investigated. Results Bile salt export pump KO MCD‐fed mice developed less steatosis but more inflammation than WT mice. Intestinal neutral lipid levels were reduced in BSEP KO mice at baseline and under MCD conditions. Faecal non‐esterified fatty acid concentrations at baseline and under MCD diet were markedly elevated in BSEP KO compared to WT mice. Serum liver enzymes and hepatic expression of inflammatory markers were increased in MCD‐fed BSEP KO animals. PPARα protein levels were reduced in BSEP KO mice. Accordingly, PPARα downstream targets Fabp1 and Fatp5 were repressed, while NFκB subunits were increased in MCD‐fed BSEP KO mice. Farnesoid X receptor (FXR) protein levels were reduced in MCD‐fed BSEP KO vs WT mice. Hepatic BA profile revealed elevated levels of TβMCA, exerting FXR antagonistic action, while concentrations of TCA (FXR agonistic function) were reduced. Conclusion Presence of hydroxylated BAs result in increased faecal FA excretion and reduced hepatic lipid accumulation. This aggravates development of MCD diet‐induced hepatitis potentially by decreasing FXR and PPARα signalling.


| INTRODUC TI ON
Non-alcoholic fatty liver disease (NAFLD) comprises a wide disease spectrum ranging from simple steatosis to steatohepatitis (NASH), fibrosis, cirrhosis and cancer. [1][2][3] The mechanisms underlying the progression from benign steatosis to NASH and more advanced disease stages are still poorly understood. Free fatty acids (FAs), especially saturated fatty acids, were proposed to act as lipotoxic triggers, 4 driving disease progression from steatosis to NASH. However, polyunsaturated FAs also serve as ligands for PPARα, 5 which correlate negatively with the severity of NASH in humans. 6 Accordingly, a beneficial role of PPARα and PPARδ agonists has been demonstrated in several (pre)clinical NAFLD/NASH studies. [7][8][9][10][11] Moreover, bile acids (BAs), via signalling through their dedicated nuclear receptor farnesoid X receptor (FXR; NR1H4) as key regulator of glucose and lipid metabolism, as well as inflammation [12][13][14][15][16][17] may play an important role in the pathogenesis and treatment of NAFLD/NASH. FXR KO mice exert decreased insulin sensitivity and have a pro-atherogenic lipoprotein profile with substantially elevated serum and hepatic cholesterol and TG levels. 17 The severity of NAFLD/NASH in humans has also been linked to reduction of FXR signalling 18,19 and changes in BA levels and composition. 20 Conversely, pharmacological activation of FXR is beneficial in patients with NAFLD and NASH. 21 Bile acids are excreted from hepatocytes into bile by the bile salt export pump (BSEP, ABCB11). 22,23 Notably, associations between BSEP variants and increased serum triglycerides (TG) 24 as well as cholesterol 24,25 and obesity 26 have been reported in humans. In line, mice overexpressing BSEP display reduced hepatic steatosis when fed a lithogenic diet 27 or a methionine choline-deficient (MCD) diet. 28 In the present study, we aimed to explore how increased BA hydroxylation and subsequent changes in BA signalling, only seen in BSEP KO mice conferring protection against cholestatic liver injury, 29 may impact on development of fatty liver and progression to steatohepatitis. Therefore, wild-type (WT) and BSEP KO mice were subjected to MCD diet, as a model of hepatic steatosis associated with profound inflammation.

| Animal experiments
Male BSEP KO and WT FVB/N mice were kindly provided by the British Columbia Cancer Research Center. 30

| Serum biochemistry
Blood was collected at harvesting and centrifuged for 20 minutes at 1500 g. Serum was stored at −80°C until analysis. Levels of transaminases (aspartate aminotransferase, AST; alanine aminotransferase, ALT), alkaline phosphatase (AP), total cholesterol, TG (Roche Diagnostics, Mannheim, Germany), FAs (Wako Chemicals GmbH, Neuss, Germany) and BAs (DiaSys Diagnostic Systems GmbH, Holzheim, Germany) were measured using enzymatic methods according to the manufacturer's instructions. High-, low-and very low-density lipoprotein (HDL, LDL and VLDL) cholesterol were assessed by quantitative agarose gel electrophoresis (Helena Biosciences, Gateshead, UK).

| Liver histology
For conventional light microscopy, livers were fixed in 4% neutralbuffered formaldehyde solution for 24 hours and embedded in paraffin. Sections were cut 4 µm thick and stained with haematoxylin and eosin and Sirius Red.
Endogenous peroxidase was blocked with 1% H 2 O 2 in methanol.

| Messenger RNA analysis and real-time quantitative polymerase chain reaction
RNA isolation, complementary DNA synthesis and real-time quantitative PCR were performed according to the manufacturer's instructions. All data were normalized to 36b4 and shown as mean ± SD. All oligonucleotide sequences are listed in Table S1.

| Western blotting
Protein isolation (nuclear extracts for FXR, PPARα and tubulin; total protein for αSMA and β-actin) and western blotting were performed. 32

| Hepatic lipid content
Lipids were extracted by a methyl tert-butyl ether (MTBE) protocol as previously described. 33 Briefly, liver pieces were extracted in methanol/MTBE/water and the organic phase was dried, re- 100% B were held for 10 minutes and the column was re-equilibrated with 50% B for 8 minutes before the next injection. The flow rate was 150 μL/min, the samples were kept at 8°C and the injection volume was 2 μL. The mass spectrometer was operated in Data-Dependent Acquisition mode using a HESI II ion source. In brief, samples and lipid standards were added to a microtiter plate and incubated at 55°C for 20-30 minutes. Isopropanol was added to the wells followed by a 30-minute incubation step at 37°C.
Fluorometric reagent was added and fluorescence was measured as Ex/Em = 490 nm/585 nm. Assay details: Fluorescent lipid assay kit for neutral lipids from Abcam (ab242307). Amount of neutral lipids was normalized to tissue weight.

| Faecal non-esterified fatty acid concentration
About 30-40 mg faeces (of non-fasted mice, collected at the end of the experiment from the individual mice) were homogenized in methanol/chloroform (2/1 v/v). Phase separation was achieved by addition of water. In brief, samples and fatty acid standards were added to a microtiter plate and incubated at 37°C for 30 minutes.
Acyl-CoA synthetase reagent was added to the wells followed by a 30-minute incubation step at 37°C. Reaction reagent was added to the wells followed by a 30-minute incubation step at 37°C (protected from light). Optical density was measured at 570 nm. Assay details: Fatty Acid Quantification Kit from abcam (ab65341). Nonesterified fatty acid (NEFA) concentration was normalized to the amount of stool.

| Bile acid profiling by ultra-performance liquid chromatography-tandem mass spectrometry
Profiles of murine primary and secondary unconjugated and conjugated C24-BAs in liver and faeces were analysed as published previously 38 on an Applied Biosystems AB SCIEX QTRAP 5500 platform.
Unconjugated and taurine-conjugated tetra-and pentahydroxylated BAs were identified from their molecular anions at m/z 463, 479, 530 and 546, respectively, and quantified in relation to D 4 -CA and D 4 -taurocholic acid (TCA).

| Statistical analysis
Results were evaluated using SPSS V.23.0. Statistical analysis was performed using multifactorial ANOVA. Data are reported as means of seven to nine (WT Ctrl n = 9; BSEP KO Ctrl n = 7; WT MCD n = 9; BSEP KO MCD n = 8) animals per group ± SD. A P value ≤.05 was considered as statistically significant.

| Alterations of BA metabolism and signalling in MCD-fed BSEP KO mice
Since MCD feeding was shown to interfere with BA metabolism, 39 recently found to be involved in progression of NASH, 20 expression of key determinants of BA homeostasis, such as Cyp27a1, Cyp2c70 (alternative BA synthesis pathway) and Cyp7a1, Cyp8b1 (classic BA synthesis pathway), serum levels of 7-hydroxy-4-cholesten-3-one (C4) as marker of BA synthesis as well as expression of FXR (main BA sensor) and its downstream target Shp were investigated in our mouse model. Gene expression of Cyp27a1 and Cyp2c70 was reduced because of MCD feeding independent of the genotype ( Figure 1A).
While mRNA expression of Cyp7a1 remained unchanged by MCD feeding (in line with unchanged C4 levels, Figure 1D), Cyp8b1 levels were significantly reduced in BSEP KO mice at baseline as well as under MCD feeding ( Figure 1B). In line, levels of TCA (endogenous FXR agonist) represented only 6%-7% of total hepatic BA concentration in BSEP KO mice independent of MCD dietary feeding (Table 1), arguing for reduced FXR signalling in BSEP KO mice.
Moreover, protein levels of FXR were reduced in MCD-fed BSEP KO mice ( Figure 1E). Accordingly, mRNA expression levels of Shp were reduced in BSEP KO MCD-fed mice but remained unchanged in WT MCD-fed mice ( Figure 1F). Cyp3a11 (enzyme involved in BA detoxification), known to be increased in BSEP KO mice at baseline,29 was significantly repressed during MCD feeding ( Figure 1C). Accordingly, hepatic levels of (poly) hydroxylated BAs dropped from 61% in BSEP KO Ctrl group to 35% in BSEP KO MCD-fed mice (Table 1).

| Lipid metabolism is changed in MCD-fed BSEP KO mice
After 5 weeks of MCD feeding, liver histology as well as hepatic TG and diacylglycerol (DG) quantification revealed that BSEP KO mice, despite same food as well as caloric intake ( Figure S1), accumulated less hepatic lipids than WT mice (Figure 2A,B). However, in MCD-fed BSEP KO mice, serum levels of transaminases (ALT, AST) as well as AP were higher than in MCD-fed WT mice ( Figure 2C). Serum TGs were reduced by MCD feeding independent of the genotype ( Figure 2D). Total and high-density lipoprotein (HDL) cholesterol levels were reduced in MCD challenged WT and BSEP KO mice with a more pronounced reduction in BSEP KO mice ( Figure 2D), while non-HDL cholesterol fraction was reduced equally in WT and BSEP KO mice fed a MCD diet ( Figure 2D). Nonesterified fatty acids ( Figure 2E) were also reduced in WT and BSEP KO mice fed with MCD diet, with a more evident difference in BSEP KO mice ( Figure 2E).

| Absence of BSEP results in impaired intestinal lipid absorption
To investigate whether differences in hepatic lipid metabolism in MCD-fed mice may result from impaired intestinal lipid absorption in BSEP KO mice, intestinal lipid metabolism was investigated.
Gene expression profiling revealed reduced expression of intestinal CD36 (FA uptake) in BSEP KO mice already at baseline. MCD feeding reduces the expression levels even more. Accordingly, expression of Mgat and Dgat1 and 2 (enzymes involved in TG formation) was reduced due to MCD feeding. In case of Dgat1 and

| Hepatic FA metabolism is altered in MCD-fed BSEP KO mice
To further explore whether differences in lipid accumulation after MCD feeding might be also influenced by differences in de novo lipogenesis, FA β-oxidation and/or FA uptake and transport, we measured the mRNA expression of hepatic de novo lipogenesis markers Srebp1c, Scd1 ( Figure 4A), β-oxidation master regulator

TA B L E 1 Hepatic BA profile
PPARα, (Figure 4B,C) as well as Fatp5 (FA uptake) and Fabp1 (intracellular FA transport) ( Figure 4D). Genes involved in de novo lipogenesis were repressed by MCD feeding independent of the BSEP genotype ( Figure 4A). Notably, PPARα expression was already significantly reduced in BSEP KO mice at baseline at both mRNA and protein levels. This reduction compared to WT mice was maintained under MCD challenge ( Figure 4B,C). PPARα downstream targets, Fabp1 and Fatp5, were significantly lower in MCD-fed BSEP KO mice compared to their control group and MCD-fed WT mice. Of note, Fabp1 expression was also significantly reduced in BSEP KO mice at baseline (following the PPARα expression levels). These findings suggest that in addition to impaired lipid absorption also reduced FA transport and uptake could account for differences in hepatic lipid content.

| D ISCUSS I ON
In this study, we examined the impact of increased levels of hydrophilic BAs in the liver on the development of fatty liver and its progression to NASH ( Figure 6). To this purpose, BSEP KO mice were challenged with MCD diet to induce steatohepatitis, a model which has been used previously by others to induce profound steatosis and inflammation and some degree of fibrosis in rodents, [41][42][43][44][45] although several short comings such as weight loss, lack of insulin resistance and obesity in this model need to be acknowledged. 46 Notably, a dissociation between hepatic steatosis and inflammation was also seen in MCD-fed WT mice where hepatic TG synthesis was inhibited via DGAT2 antisense oligonucleotide. 50 As a result of elevated hepatic free FA content with subsequent lipotoxicity, oxidative damage, liver inflammation and fibrosis were aggravated while hepatic steatosis was reduced. 50   In summary, our data show that altered BA composition and sig-