Mice Lacking FXR Are Susceptible to Liver Ischemia-Reperfusion Injury

Activation of bile acid (BA) receptor, farnesoid X receptor (FXR) has been shown to inhibit inflammatory responses and improve tissue ischemia-reperfusion injury (IRI). This study investigated the effect of FXR deficiency on liver IRI, using a liver warm IRI mouse model. We demonstrate that liver IRI resulted in decreased FXR expression in the liver of WT mice. FXR-/-mice displayed greater liver damage and inflammatory responses than WT mice, characterized by significant increases in liver weight, serum AST and ALT, hepatocyte apoptosis and liver inflammatory cytokines. Liver IRI increased expression of X box binding protein 1 (XBP1) and FGF21 in WT liver, but not in FXR-/- liver, which conversely increased CHOP expression, suggesting a loss of ER stress protection in the absence of FXR. FXR deficiency increased circulating total BAs and altered BA composition with reduced TUDCA and hepatic BA synthesis markers. FXR deficiency also reshaped gut microbiota composition with increased Bacteroidetes and Proteobacteria and decreased Firmicutes. Curiously, Bacteroidetes were positively and Firmicutes were negatively correlated with serum ALT levels. Administration of FXR agonist CDCA inhibited NF-κB activity and TNFα expression in vitro and improved liver IRI in vivo. Our findings demonstrate that FXR signaling plays an important role in the modulation of liver IRI.


Introduction
Recent improvements in surgical techniques, liver preservation and immunosuppression continue to improve liver operations and increase survival of liver grafts in transplantation (1,2). However, liver ischemia and reperfusion injury (IRI) remains an important problem in the clinical scenario of liver surgery (3). In major liver resection, a primary side effect is liver warm IRI, i.e. ischemia of remnant liver following temporary vascular occlusion, and additional reperfusion injury added to the damage sustained during ischemia (4). In liver transplantation, liver warm and cold IRI may occur in situ during recipient surgery or donor liver harvest, which may cause cellular injury, organ dysfunction, or even complete graft failure (5).
Tissue ischemia and reperfusion initiates a defensive process known as the unfolded protein response (UPR) for adaptation and safeguard of cell survival. Activation of the inositolrequiring enzyme (IRE)/X-box binding protein 1 (XBP1) pathway in response to the endoplasmic reticulum (ER) stress protects hepatocytes from apoptosis. However, sustained activation of ER triggers proapoptotic signals via C/EBP homologous protein (CHOP), which is responsible for cellular dysfunction (6,7). Evidence for the role of BAs and their receptors in the regulation of inflammation and tissue IRI already exists in experimental and animal settings (8,9). Activation of the BA membrane receptor, G protein-coupled BA receptor 1 (TGR5), by TGR5 agonists inhibits inflammatory responses and attenuates liver IRI by suppressing the Toll like receptor 4 (TLR4)-NF-B mediated pathway (10). Recent experimental data suggest that activation of BA nuclear receptor, farnesoid X receptor (FXR), stimulates the IRE/XBP1 pathway and may be protective during liver injury (11). Similarly, administration of the FXR-agonist obeticholic acid (OCA) improves survival in a rodent model of intestinal IRI, through gut barrier preservation and mitigated inflammation (12). However, in the setting of liver IRI, the role and fundamental mechanism of FXR signaling in the modulation of ER stress and inflammatory responses remain unexplored.
There is growing evidence of bidirectional interactions between BAs and the gut microbiota, i.e. BAs in the distal intestine influence the composition of the gut microbiota that in turn modify primary BAs to generate secondary BAs (13). FXR deficiency significantly reduces the abundance of Firmicutes and increased Bacteroidetes (14). Dysbiosis directly impairs intestinal barrier function and increases gut permeability. Bacteria and their components, such as lipopolysaccharide (LPS), pass through the intestinal barrier to the liver, causing inflammatory responses. Modulation of the gut microbiota may inhibit inflammatory responses and represent a novel therapeutic approach for the prevention of tissue IRI (15).
In this study, we examined FXR deficiency in the aggravation of liver IRI using a mouse model of liver warm IRI. We hypothesized that a lack of FXR exaggerates hepatocyte ER stress and apoptosis during liver IRI. FXR deficiency may alter BA and gut microbiota homeostasis, which may also contribute to the pathogenesis of liver IRI. Our results demonstrate that mice lacking FXR are more susceptible to liver IRI than wild type (WT) mice.

Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) analysis
Blood samples were collected at 6 hours (WT vs. FXR -/-) or 24 hours (WT with and without FXR agonist CDCA) after ischemia and were centrifuged to obtain serum. ALT and AST were measured to assess the extent of hepatocyte damage using an automated chemical analyzer (Olympus Automated Chemistry Analyzer AU5400, Tokyo, Japan).

Bile acid analysis
Total BAs and BA composition were analyzed as described in our previous reports (17,18

Cecal content microbiota analysis
Gut microbiota in the cecal content was analyzed as described in our previous report (17,18).
Briefly, primers specific for 16S rRNA V4-V5 region (Forward: 515F: 5'-GTGYCAGCMGCCGCGGTAA -3' and Reverse: 806R: 5'-GGACTACHVGGGTWTCTAAT-3') that contained Illumina 3' adapter sequences, as well as a 12-bp barcode were used. Sequences were generated by an Illumina MiSeq DNA platform at Argonne National Laboratory and analyzed by the program Quantitative Insights Into Microbial Ecology (QIIME) (19). Operational Taxonomic Units (OTUs) were picked at 97% sequence identity using open reference OTU picking against the Greengenes database (20). OTUs generated in QIIME were then analyzed using linear discriminant analysis (LDA) effect size (LEfSe) where non-parametric factorial Kruskal-Wallis sum-rank testing (p < 0.05) identified significantly abundant taxa followed by unpaired Wilcoxon rank-sum test to determine LDA scores > 2 (21). BMDMs were maintained in DMEM (Invitrogen, Carlsbad, CA) supplemented with 10% FCS at 37°C in a 5% CO 2 incubator for 4 hours. Luciferase activity in BMDMs was determined by the tissue luciferase assay using the Bradford Luciferase Reporter Assay kit (Promega, Madison, WI, USA) following the manufacturer's instructions as described in our previous report (25). TNFα mRNA was analyzed by qRT-PCR as described above.

Histological examination
Liver tissue was collected 6 hours following ischemia and subjected to H&E staining. Hepatocyte apoptosis was analyzed by TUNEL assay. TUNEL labeling was performed using a TUNEL kit (Abcam, Cambridge, UK) according to the manufacturer's instructions. Quantitation of apoptotic cells was accomplished by calculating of the labeling index, which was defined as the ratio between the number of labeled cells and the total cells counted in triplicate field by two blinded investigators.

Statistical analysis
Data from the current studies were analyzed by ANOVA tests (StatView 4.5, Abacus Concepts, Berkeley, CA) with Tukey post-hoc multiple comparisons, or by 2-tailed Student's t test when appropriate, where p-value < 0.05 was considered significant. The results are presented as mean ± SEM.

Mice lacking FXR were more susceptible to liver IRI
A lack of FXR increased TGR5 expression in naïve FXR -/mice, indicating a compensation of TGR5 for the deficiency of FXR. Liver IRI reduced hepatic expression of FXR in WT mice, and conversely, the same procedure increased expression of TGR5 in WT and FXR -/mice (Fig. 1, A and B). FXR -/mice also revealed an increase in liver weight (liver/body weight) in untreated FXR - A lack of FXR resulted in increased TNFα, IL-1β and IL-6 transcript in FXR -/control mice Fig. 2, A to C). As expected, liver IRI also increased TNFα, IL-1β and IL-6 transcripts in both WT and FXR -/mice (p < 0.01), and TNFα was significantly increased in FXR -/-IRI mice, compared to WT IRI mice ( Fig. 2A). Superoxide dismutase 2 (SOD2) binds to the superoxide to improve tissue injury. Reduced SOD2 expression was observed in WT mice following IRI, but more dramatically in FXR -/mice (p = 0.05, Fig. 2D).

FXR deficiency triggered hepatocyte apoptosis via activation of ER stress
ER stress activates a number of proteins that straddle ER membranes. Activated IRE1α functions as an endoribonuclease splicing a 26 base pair intron from XBP1 mRNA. Spliced XBP1 (sXBP1) mRNA is translated into a stable and active UPR transcription factor. Therefore, measuring XBP-1 splicing represents a reliable indirect method of determining IRE1α activation (27). Liver IRI promoted expression of total XBP1 (TXBP1, Fig. 3A), sXBP1 (Fig. 3B) and unconventional splicing XBP1 (usXBP1, Fig. 3C) in WT mice. However, mice deficient in FXR showed no change in these genes (Fig. 3, A to C). Liver IRI decreased expression of ATF4 and GFP94 in the absence of FXR (WT IRI vs. FXR -/-IR, p < 0.05, Fig. 3, D and E). Furthermore, FXR deficiency resulted in decreased ERAD-enhancing α-mannosidase-like protein (EDEM, Fig. 3F), ultimately resulting in hepatocyte susceptibility to apoptosis. Consistently, we observed increased CHOP with FXR deficiency (Fig. 3G). TUNEL tests showed increased apoptotic cells in untreated FXR -/control mice, but without statistical difference. Liver IRI increased apoptotic cells in all groups; however, apoptotic cells were significantly increased in FXR -/mice compared with WT IRI mice (Fig. 3H).
Both the lack of FXR and liver IRI altered BA homeostasis FXR represses transcription of the gene encoding CYP7A1 that is the rate-limiting enzyme in BA synthesis, and thus, the lack of FXR thus resulted in increased circulating total BAs (Fig. 4A). BA composition analysis demonstrated that the lack of FXR specifically increased circulating TbMCA (Fig. 4B), TCA (Fig. 4D) and DCA (Fig. 4F) when compared to WT control mice. Liver IRI increased circulating total BAs (Fig.4A), TbMCA (Fig. 4B), bMCA (Fig. 4C), TCA (Fig. 4D) and TUDCA (Fig. 4G) in WT mice. Although CA and DCA were increased in WT liver IRI mice, there were no statistical differences (p > 0.05) (Fig. 4, E and F). There were no changes to circulating LCA in any groups (Fig. 4H). Given that the mutation of FXR affects BA synthesis and transport (28), and IRI may also impacts liver BA homeostasis, we analyzed total BAs in the liver. As expected, the absence of FXR causes BA accumulation of total BAs in the liver; however, IRI results in decreases of liver total BAs in both WT and FXR -/mice (Fig. 4I).
Enzyme CYP7A1 initiates the classic pathway of BA synthesis, followed by CYP8B1 that produces the majority of the BA pool.  5A). Liver IRI and the lack of FXR reduced FGFR4 gene, but no statistical difference (Fig. 5B).
Fibroblast growth factor 21 (FGF21) is identified as one of tissue protective proteins (30).
To test whether FGF21 is altered under liver IRI, liver FGF21 expression was analyzed. Results showed that liver IRI increased FGF21 transcript in the WT liver. However, IRI failed to enhance liver FGF21 expression in the absence of FXR (Fig. 5G), suggesting FGF21 expression may be, at least in part, regulated by FXR (31).

FXR deficiency altered gut microbiota composition
Taxonomic analysis of gut microbiota composition showed changes at the phylum level, where Bacteroidetes was increased and Firmicutes decreased -with increased Bacteroidetes/Firmicutes ratios -in FXR -/-IRI mice with and without liver IRI (Fig. 6, A and B) consistent with a previous report (14). Our data showed that Proteobacteria was increased in untreated FXR -/and FXR -/-IRI mice (Fig. 6C). More interestingly, altered Bacteroidetes and Firmicutes were correlated with markers of liver function; Bacteroidetes was positively and Firmicutes was negatively correlated with serum ALT levels (Fig. 6, D and E) following liver IRI. Increasing data identify Proteobacteria as lipopolysaccharide (LPS) producers that act as possible microbial signature of host disease (32). Within the phylum Bacteroidetes, the relative abundance of Bacteroides was increased in naïve and FXR -/-IRI mice (Fig. 6, F and G). Within the phylum Firmicutes, the relative abundances of Turicibacter and Clostridiales were decreased in FXR -/mice with and without liver IRI (Fig. 6, H and I). Finally, within the phylum Proteobacteria, the relative abundance of Desulfovibrio was increased in FXR -/mice with and without liver IRI (Fig. 6J).

Activation of FXR inhibited inflammatory reaction in vitro and improved liver IRI in vivo
Macrophage-induced inflammatory responses are mediated through inflammatory signaling pathways, such as NF-B, causing TNFα expression. In the bone marrow-derived macrophage (BMDM) culture, LPS increased luciferase activity, indicating increased NF-B activity, which was inhibited by administration of chenodeoxycholic acid (CDCA, Supplemental Fig. 1SA). TNFα mRNA was increased by the stimulation of LPS but also inhibited by CDCA (Supplemental Fig.   1SB).
To test whether FXR activation improved liver IRI, FXR agonist CDCA was administered to WT mice undergone 90 minutes of ischemia and 24 hours of reperfusion. The results showed that serum ALT levels were significantly decreased in CDCA-treated WT mice compared to Vehicle-treated WT mice (Supplemental Fig. 1SC). Although CDCA also decreased AST levels in the same animals, differences between CDCA-treated and control mice did not reach significance (p > 0.05, Supplemental Fig. 1SD).
A recent report suggests that FXR signaling modulates ER stress via activation of the IRE/XBP1 pathway (11). Our findings show that mice lacking FXR resulted in reduced expression of GRP94, ATF4, EDEM, XBP1 (including total, spliced and unconventional spliced XBP1) in the liver when suffering from IRI, suggesting that the absence of FXR forfeits protection of hepatocytes from apoptosis, contributing to the susceptibility to liver IRI.

ROS generation inflicts tissue damage and initiates a cascade of deleterious cellular
responses leading to inflammation, cell death and ultimately liver failure. SOD2 transforms toxic superoxide and clears mitochondrial ROS and therefore confers protection against hepatocyte apoptosis (36). In the absence of FXR, liver IRI fails to promote production of SOD2, and accumulation of ROS may initiate the inflammatory reaction and liver damage.
FGF21 is a member of the fibroblast growth factor family that regulates cell growth, differentiation, and glucose and lipid metabolism. FXR activation promotes FGF21 expression (31), which alleviates hepatic ER stress under the physiological condition (37). Increased FGF21 protects against acetaminophen-induced hepatotoxicity by enhancing antioxidant capacity (38).
Clinical findings show that serum FGF21 increases following liver or cardiac ischemia and is associated with protective responses (39). We observed that liver IRI promoted expression of liver FGF21 in WT mice, but deficiency of FXR failed to increase FGF21 expression, thereby losing liver protection.
The lack of FXR stimulates BA synthesis, resulting in increased circulating total BAs, TbMCA, MCA, TCA and DCA. Due to the damage of BA transports in the absence of FXR (28), BAs are accumulated in the liver as evidenced by increased total BAs in the liver of FXR -/mice.
Liver IRI causes immediate reduction of synthesis as confirmed by decreased BA synthetic enzymes and reduced total BAs in the liver in both WT and FXR -/mice. However, liver IRI induces rapid increases of circulating total BAs and alterations of BA composition in both WT and FXR -/mice. Among the notable changes in circulating BAs was TUDCA, a hydrophilic BA that regulates ER stress and mediates cytoprotective responses during liver IRI (40); however, FXR -/mice failed to demonstrate elevated TUDCA. Nevertheless, whether the alteration of BA composition mediates liver IRI needs further study.
A biochemical link between the gut microbiota and FXR signaling has been demonstrated (41). Taxonomic tree analysis displaying differential taxa in hierarchical layers supports the concept that FXR deficiency is associated with distinct microbiota and liver IRI may generate selection pressure for certain gut bacterial taxa. Recent work shows that FXR -/mice exhibit reduced levels of the phylum Firmicutes and destabilizes the gut microbiota when compared with WT animals (42). Our findings demonstrate that liver IRI decreased the relative composition of the bacterial phylum Bacteroidetes in WT animals, consistent with previous work (43). FXR deficiency leads to increased Bacteroidetes and reduced Firmicutes. Increased Bacteroidetes and reduced Firmicutes were correlated with higher levels of circulating ALT after liver IRI across all animals. This analysis also showed increased cecal Proteobacteria in the absence of FXR.   (D) Increased phylum Bacteroidetes was positively related with increased serum ALT levels in