Cholestyramine Administration Induces Gut Dysbiosis and Severe Liver Injury in ApoE-/- Mice


 Background & AimBile acids, as hydrophilicity/ hydrophobic molecules and natural ligands of bile acids receptors, are involved in the pathogenesis of liver injury and metabolic disorders. Cholestyramine acts as a bile acid chelating agent and is used to reduce cholesterol in clinic, but its hepatotoxicity and effects on metabolism are still unclear.MethodsApoE-/- mice were continuously administered with low fat diet with or without 2% cholestyramine 12 weeks to determine the hepatotoxicity and metabolic influences of cholestyramine on these mice. Serum transaminases, liver histological features and hepatic inflammatory markers expression were used to assess cholestyramine hepatotoxicity. Body weight, liver weight, serum lipids profile and liver lipid metabolism related genes expression were used to evaluate effects of cholestyramine on metabolism. Differences of endogenous bile acids composition, bile acids metabolism related genes and intestinal flora structure between groups were also analyzed by an ultraperformance liquid chromatography coupled to tandem mass spectrometry system, quantitative real time PCR and 16S rDNA sequencing.ResultsAfter 12 weeks of dietary intervention, although cholestyramine decreased ApoE-/- mice serum cholesterol level, it significantly increased the body and liver weight of these mice, and caused severe liver injury: increased serum transaminases levels, promoted liver vesicular steatosis and inflammatory cells infiltration, and enhanced the expression of lipid synthesis gene: FASN and inflammatory markers genes: TNF-α 、MCP-1、F4/80 and CD68. Moreover, cholestyramine administration caused hydrophobic transformation of bile acids pool and severe gut dysbiosis, which were characterized by decreased hydrophilic bile acids: muricholic acid (MCA) and ursodeoxycholic acid (UDCA), and increased abundance of potential pathogenic bacteria: s_Bacteroide_vulgatus.ConclusionThis study revealed that cholestyramine had the ability to induce severe liver injury and metabolic disorders. These effects might be attributed to inducing hydrophobic transformation of bile acids pool and gut dysbiosis by cholestyramine. Furthermore, our study provided a warning of hepatotoxicity and metabolic disorders for clinical application of cholestyramine, and suggested that cholestyramine was not a good scavenging tool to study the role of bile acids alternations in diseases progression.


Introduction
Bile acids are amphoteric metabolic molecules derived from cholesterol. In the liver, at least 17 enzymes are involved in the de novo synthesis of primary bile acids by hepatocytes using cholesterol as the substrate [1]. Among them, cholesterol 7α-hydroxylase (CYP7A1) is considered to be the rate limiting enzyme for bile acids synthesis, which determines the size of the bile acids pool [2]. Primary bile acids were synthesized in hepatocytes, which include cholic acid (CA) (human and mice), chenodeoxycholic acid (CDCA) (human and mice) and α/β muricholic acid (MCA) (mice). These bile acids continue to be conjugated with taurine (T) and glycine (G) to form conjugated primary bile acids: T (G) CA, T (G) CDCA, T α/β MCA in the liver [3]. Then, conjugated primary bile acids enter into the intestinal tract through the bile duct system and further form the secondary bile acids: deoxycholic acid (DCA), lithocholic acid (LCA) and ursodeoxycholic acid (UDCA) through bacterial biotransformation. About 95% of intestinal bile acids metabolites are reabsorbed by the portal vein circulation system at the jejunum end and ileum, and most of them are reuptake by hepatocytes to repeat the "enterohepatic circulation".
The number, position and direction of hydroxyl groups in the amphoteric structure of bile acids determine their hydrophilicity/ hydrophobicity. It is reported that DCA, LCA, CDCA and CA are hydrophobic bile acids, while MCA and UDCA are hydrophilic. Hydrophobic bile acids have been proved to be hepatotoxic, resulting in oxidative stress and in ammatory death of hepatocytes [4]. Previous studies have shown that hydrophobic transformation of bile acids pool usually occurred in liver diseases such as nonalcoholic steatohepatitis (NASH), cirrhosis and cholestatic hepatitis [5][6][7]. In addition, bile acids are also natural ligands of some receptors in vivo, such as the nuclear receptor: farnesoid X receptor (FXR), and G protein coupled receptors: takeda G-protein-coupled receptor 5 (TGR5) and sphingosine-1-phosphate receptor 2 (S1PR2) [8]. The abnormality of these signals has recently been considered to be involved in the pathogenesis of metabolic diseases due to their abilities to regulating glucose and lipid metabolism and in ammatory response. In endogenous bile acids pool, CDCA is the most potent ligand for FXR activation, followed by CA, DCA and LCA, while TαMCA and TβMCA have been identi ed as natural FXR antagonists; and TCA is a weak agonist of S1PR2; and TGR5 is activated mainly by the secondary bile acids: LCA and DCA [8]. Although bile acids signaling biofunction has been con rmed by many studies, the causal relationship between endogenous bile acids alternations and diseases progression has not been established.
Cholestyramine is a bile acids sequestrant, which is used to treat hypercholesterolemia in clinic [9]. Besides, in basic experiments, it is also used as a method to remove endogenous bile acids in laboratory animals to investigate the role of bile acids in regulating diseases progression [10][11]. However, as a strong CYP7A1 inducer, bile acids may not be removed by cholestyramine since overexpression of CYP7A1 can lead to an enlarged and hydrophobic bile acids pool [12]. Based on the above, we speculate that cholestyramine may cause hepatotoxicity and metabolic disorders by inducing hydrophobic transformation of endogenous bile acids pool and altering bile acids signaling. Therefore, we conducted this study to evaluate the relationship between the use of cholestyramine and liver injury and/ or metabolic disorders in mice.

Methods
Animals 8-week-old male ApoE knockout (ApoE-/-) mice on a C57BL/6 background purchased from Beijing Vital River Laboratory Animal Technology Co. Ltd were divided into two groups according to their diets: the low fat diet (LFD) control group and the cholestyramine supplement (LFD + Chy) group. Mice in groups of LFD and LFD + Chy were fed with a low fat diet (LF10B, Dyets, USA) for 12 weeks and a low fat diet (LF10B, Dyets, USA) supplemented with 2% cholestyramine (C4650, Sigma, USA) for 12 weeks. After 12 weeks feeding, mice were sacri ce after fasted overnight and then serum, liver and intestinal tissues were collected.

Biochemical and histopathologic analyses
Fasting serum samples were obtained from orbital blood and. Mice serum biochemical parameters including total cholesterol (TC), total triglyceride (TG), low density lipoprotein (LDL), high density lipoprotein (HDL), alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were measured by Hitachi 7180 automatic biochemical analyzer. Formalin-xed liver para n samples were cut into successive 4um sections and stained with hematoxylineosin (HE). The characteristics of liver tissues were observed by the light microscope.
Quantitative real-time PCR (Q-PCR) RNA was extracted from frozen liver and intestine tissues using TRIzol reagent (190906, Ambion, USA). Complementary DNA was synthesized from 1 mg of total RNA using Transcriptor cDNA Synth. kit (32460620, Roche, GER) following the instruction. Q-PCR was performed using SYBR green PCR master mix (04913914001, Roche, GER) in an ABI Prism 7900HT sequence detection system (Applied Biosystems, Foster City, CA). Each mRNA level was normalized to those of GAPDH and expressed as fold change relative to those of control group. Primer sequences used for Q-PCR were listed in Table 1. Bile acids were isolated from serum, liver and intestine tissues using acetonitrile/methanol (8:2) mix, and an ultraperformance liquid chromatography coupled to tandem mass spectrometry system (ACQUITY UPLC-Xevo TQ-S; Waters Corp., Milford, Massachusetts) was used to measure bile acids absolute concentration in serum (nM), liver (fmol/mg) and intestine (fmol/mg) as described previously [13]. Each bile acid concentration was then normalized to the ratio of its absolute concentration to total bile acids.
Fecal samples were collected from mouse appendix and frozen at -80℃. Bacterial genomic DNA of appendiceal feces was extracted using a QIAamp Fast DNA Stool Mini Kit (Qiagen, USA), and then bacterial 16 s rDNA was ampli ed by employing BIO-RAD thermal cycler (BIO-RAD Inc., USA). 16 s recombinant DNA sequencing data was generated using Illumina MiSeq (Illumina, San Diego, CA) sequencing. The sequencing data was processed by the ucluster method of Quantitative Insights Into Microbial Ecology (QIIME) (version 1.9.0; http://qiime.org/index.html). High similar (≥ 97%) sequences were assigned into the same operational taxonomic units (OUT) and the most abundant sequence was elected as the OTU representative sequence.

Statistical analysis
Student t test was performed to examine signi cance of measurement data between groups using GraphPad Prism, version 7.00 (GraphPad Software, CA, USA). Similarity of bacterial community structure was conducted by principal co-ordinates analysis (PCoA). LEfSe analysis was performed to identify the potential biomarker intestinal ora. Results were considered to be statistical signi cant with a P-value < 0.05.

Results
1. Cholestyramine administration induces obesity and severe liver injury in ApoE-/-mice fed with LFD After 12 weeks of continuous feeding, the effect of cholestyramine on reducing levels of serum TC (p< 0.001) and LDL (p< 0.01) was con rmed in this study. While, serum TG and HDL levels were not affected by cholestyramine administration (Fig.1 C-F). Interestingly, we found that cholestyramine administration signi cantly increased body (p< 0.01) and liver (p< 0.001) weight of these mice ( Fig.1 A-B).
Further serological tests showed that levels of liver enzymes: ALT (p< 0.001) and AST (p< 0.01) levels were much higher in LFD+Chy group than those in LFD group. (Fig.2 A). HE staining of liver tissues showed mice in LFD+Chy group had more severe vesicular steatosis and in ammation compared with these in LFD group (Fig.2 B). Furthermore, we investigated whether cholestyramine administration altered hepatic mRNA expression of lipid metabolism related genes and in ammatory factors genes in these mice. Fig.2 C-D showed that cholestyramine administration signi cantly promoted hepatic expression of FASN (p< 0.01), but did not altered the expression of SREBP1c SCD1 and PPAR-α (p> 0.05) in liver tissues of these mice. Moreover, cholestyramine administration increased the expression of hepatic TNF-α (p< 0.05) MCP-1 (p< 0.01) F4/80 (p< 0.05) and CD68 (p< 0.05) in these mice.
2. Cholestyramine does not remove endogenous bile acids, but induces a hydrophobic bile acid pool in ApoE-/-mice fed with LFD As shown in Fig.3 A-C, cholestyramine administration did not completely eliminate serum, hepatic and intestinal bile acids. Only hepatic total bile acids level was reduced by cholestyramine administration and there was no differences in serum and intestinal total bile acids between LFD and LFD+Chy groups.
Alternations In addition, we also analyzed the expression of molecules involved in bile acids enterohepatic circulation. The results indicated that cholestyramine administration signi cantly decreased the expression of the bile acid signaling molecule: SHP (p< 0.05) in the liver ( Fig.5 A), increased the expression of the ratelimiting enzyme of bile acids synthesis: CYP7A1 (p< 0.05) (Fig.5 B) and decreased the expression of molecules responsible for hepatic bile acids reuptake: NTCP (p< 0.001) and OATP (p< 0.05) (Fig.5 C). However, in the intestinal tissue, we did not nd that cholestyramine administration had any effects on the expression of molecules responsible for intestinal bile acids transport: ASBT and OSTα/β.
3. Cholestyramine administration causes the increase of Bacteroides abundance that closely associated with the decreased hydrophilic bile acids in ApoE-/-mice fed with LFD Considering that bile acids pro le and signaling alternations were closely related to gut microbiota, we further evaluated the effect of cholestyramine on intestinal ora in mice. The unweighted and weighted PCoA revealed signi cant differences in the structure of intestinal ora between cholestyramine intervention and non-intervention groups (Fig.6 A). The cladogram showed that cholestyramine intervention caused signi cant changes of Firmicutes and Bacteroides abundances in phylum, class, order, family, genus and species, especially increased p_Bacteroidetes, c_Bacteroidia, o_Bacteroidales, f_Bacteroidaceae, g_Bacteroides and s_Bacteroide_vulgatus abundances. (Fig.6 B-C). Further analysis showed the correlation between changes of hydrophilic bile acids and Bacteroides. As shown in Figure 6, the decrease of hydrophilic bile acids: MCA and UDCA in vivo showed a negative correlation with the rise of Bacteroides abundances, especially in the family and genus (Fig.6 D).

Discussion
Previous studies have shown that elimination of endogenous bile acids by cholestyramine could improve liver injury and metabolic abnormalities [11,14]. Conversely, our study for the rst time has demonstrated that cholestyramine was able to induced obesity and promoted severe liver injury in ApoE-/-mice. More importantly, cholestyramine could not remove endogenous bile acids, but induce hydrophobic transformation of endogenous bile acids pool and imbalance of bacterial ora, which were characterized by decreased hydrophilic bile acids: MCA and UDCA percentages in serum, liver and intestine, and increased abundance of pathogenic bacteria: s_Bacteroide_vulgatus. These results suggested that hydrophobic toxic bile acid pool formation and the increase of pathogenic bacteria might be the main cause of hepatotoxicity and metabolic disorders caused by cholestyramine in ApoE-/-mice.
Cholestyramine reduced cholesterol mainly by chelating intestinal bile acids. However, this effect would feedback activate liver CYP7A1 to promote bile acids synthesis [15]. In our study, cholestyramine not only upregulated the expression of CYP7A1, but also reduced the expression of NTCP and OATP to inhibit bile acids reuptake by hepatocytes. And, we only found total bile acids level was decreased in the liver rather than the serum and intestine. This might be attributed to insu cient uptake of bile acids by hepatocytes and unchanged intestinal bile acids output.  [4]. We also observed some hydrophobic bile acids: CA, CDCA, DCA and LCA were increased in circulation, liver or intestine. These hydrophobic bile acids play an important role in inducing oxidative stress, necrosis, apoptosis and plasma membrane damage of hepatocytes [4]. Therefore, we speculated that the formation of hydrophobic bile acids pool might be responsible for the severe liver injury induced by cholestyramine. Small heterodimer partner (SHP) is a unique orphan nuclear receptor and acts as an important factor for regulating metabolism and in ammation [20][21]. In our study, SHP expression was decreased by cholestyramine administration. This might be a reason for cholestyramine inducing obesity and liver in ammation. Gut dysbiosis was closely related with liver injury and metabolic disorders. We found a signi cant increase in the abundance of s_Bacteroide_vulgatus in mice fed with cholestyramine.
Based on the adverse effects of s_Bacteroide_vulgatus on in ammation and metabolic abnormalities reported in previous studies [18,22], we speculated that elevated abundance of s_Bacteroide_vulgatus might also participate in liver injury and obesity caused by cholestyramine.
Our research had some limitations: we have not investigated the role of intestinal ora in cholestyramine induced liver injury and obesity phenotype through eliminating intestinal ora; and the function of single bacteria and/ or bile acid in cholestyramine induced phenotype has not been studied in this study.

Conclusion
In summary, our study for the rst time con rmed that cholestyramine induced severe liver injury and gut dysbiosis in ApoE-/-mice fed with low fat diet, which not only provided a warning of hepatotoxicity and metabolic disorders for clinical application of cholestyramine, but also suggested that cholestyramine was not a good bile acids scavenger in basic research, and better tools were needed to study the role of bile acids in diseases progression.

Availability of data and materials
All data and materials used in this study are available when requested.

Ethics approval and consent to participate
The animal study was conducted in accordance with the guidelines of the Laboratory Animal Ethical and Welfare Committee at Hangzhou Normal University.

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests.           The effect of cholestyramine administration on total bile acids levels. Scatter and bar plots drawn with mean ± SD showed the differences of total bile acids levels in serum (A), liver (B) and intestine (C). *p < 0.05, **p < 0.01, ***p < 0.001. LFD: low fat diet; Chy: cholestyramine. The effect of cholestyramine administration on total bile acids levels. Scatter and bar plots drawn with mean ± SD showed the differences of total bile acids levels in serum (A), liver (B) and intestine (C). *p < 0.05, **p < 0.01, ***p < 0.001. LFD: low fat diet; Chy: cholestyramine.

Figure 6
Cholestyramine administration induces gut dysbiosis in ApoE-/-mice fed with LFD. PCoA scatter plots re ected differences of intestinal ora pro le between groups (A); Cladogram and LDA bar plots showed intestinal ora biomarkers between groups (B-C); The correlation's matrix drawn with correlation coe cient showed the correlation between differential Bacteroides and hydrophilic bile acids (D). LFD: low fat diet; Chy: cholestyramine; PCoA: principal co-ordinates analysis; LDA: linear discriminant analysis.

Figure 6
Cholestyramine administration induces gut dysbiosis in ApoE-/-mice fed with LFD. PCoA scatter plots re ected differences of intestinal ora pro le between groups (A); Cladogram and LDA bar plots showed intestinal ora biomarkers between groups (B-C); The correlation's matrix drawn with correlation coe cient showed the correlation between differential Bacteroides and hydrophilic bile acids (D). LFD: low fat diet; Chy: cholestyramine; PCoA: principal co-ordinates analysis; LDA: linear discriminant analysis.

Figure 6
Cholestyramine administration induces gut dysbiosis in ApoE-/-mice fed with LFD. PCoA scatter plots re ected differences of intestinal ora pro le between groups (A); Cladogram and LDA bar plots showed intestinal ora biomarkers between groups (B-C); The correlation's matrix drawn with correlation coe cient showed the correlation between differential Bacteroides and hydrophilic bile acids (D). LFD: low fat diet; Chy: cholestyramine; PCoA: principal co-ordinates analysis; LDA: linear discriminant analysis.

Figure 6
Cholestyramine administration induces gut dysbiosis in ApoE-/-mice fed with LFD. PCoA scatter plots re ected differences of intestinal ora pro le between groups (A); Cladogram and LDA bar plots showed intestinal ora biomarkers between groups (B-C); The correlation's matrix drawn with correlation coe cient showed the correlation between differential Bacteroides and hydrophilic bile acids (D). LFD: low fat diet; Chy: cholestyramine; PCoA: principal co-ordinates analysis; LDA: linear discriminant analysis.

Figure 6
Cholestyramine administration induces gut dysbiosis in ApoE-/-mice fed with LFD. PCoA scatter plots re ected differences of intestinal ora pro le between groups (A); Cladogram and LDA bar plots showed intestinal ora biomarkers between groups (B-C); The correlation's matrix drawn with correlation coe cient showed the correlation between differential Bacteroides and hydrophilic bile acids (D). LFD: low fat diet; Chy: cholestyramine; PCoA: principal co-ordinates analysis; LDA: linear discriminant analysis.