Farnesoid X receptor signaling activates the hepatic X-box binding protein 1 pathway in vitro and in mice

MATERIAL

Bile acids and cholestyramine were purchased from Sigma-Aldrich (St. Louis, MO). GW4064, (z)-guggulsterone and 6α-ethyl-chenodeoxycholic acid (6-ECDCA) (obeticholic acid) were purchased from Cayman Chemical (Ann Arbor, MI). FXR, SHP small interfering RNA (siRNA), and scramble siRNA were purchased from GE Dharmacon (Lafayette, CO). Lipofectamine RNAiMax and LTX transfection reagents were purchased from ThermoFisher (Waltham, MA). Antibodies against XBP1 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were from Proteintech (Chicago, IL), β-actin, FLAG, and green fluorescent protein (GFP) antibodies were from Sigma-Aldrich (St. Louis, MO), IRE1α antibody was from Cell Signaling Technology (Danvers, MA), and phosphorylated IRE1α (p-IRE1α) antibody was from Novus Biologicals (Littleton, CO). SHP antibody was from Santa Cruz Biotechnology (Dallas, TX).

ANIMAL EXPERIMENTATION

FVB/NJ, C57BL/6J and FXR (–/–) mice (8-10 weeks) were purchased from Jackson Laboratory (Bar Harbor, ME). SHP (–/–) mice colonies were maintained as reported.17 Mice were maintained under 12-hour light/dark cycles with unlimited access to regular chow and water until the first day of the study. For feeding experiments, male FVB/NJ mice received chow, chow supplemented with 0.3% (w/w) DCA, or chow supplemented with 2% (w/w) cholestyramine for up to 7 days. Male C57BL/6J mice were gavaged with GW4064 (150 mg/kg) as described.18 At the conclusion of each experiment, mice were fasted for 4 hours and livers were excised, flash-frozen in liquid nitrogen, and stored at −70°C. Whole blood was obtained from the right atrium by cardiac puncture. For mice used in bile acid analysis, liver, gallbladder, and small intestine were collected and minced in 100% ethanol. Bile duct ligation (BDL) in female C57BL/6J mice was performed by the Northwestern University Microsurgery Core. All protocols and procedures were approved by the Northwestern University Institutional Animal Care and Use Committee guidelines.

SERUM ALANINE AMINOTRANSFERASE ASSAY

Serum alanine aminotransferase (ALT) was measured using a spectrophotometric assay according to the manufacturer's protocol (Teco Diagnostics, Anaheim, CA).

ANALYSIS OF BILE ACID POOLS

Total bile acid pool size and composition were measured by high-performance liquid chromatography as described.19 Samples were spiked with glycocholic acid as an internal standard to control for extraction efficiency. Individual bile salt species were identified by their characteristic retention times and by using bile acid standards.

CELL CULTURE AND siRNA TRANSFECTION

HepG2 cells (ATCC, Manassas, VA) were cultured in Eagle's minimal essential medium with 10% fetal bovine serum (FBS) and maintained at 37°C in 5% CO₂. Lenti-X HEK293T cells (Takara Bio., Mountain View, CA) were cultured in Dulbecco's modified Eagle's medium with 10% FBS and 4 mM of glutamine. Huh7 cells stably expressing rat sodium-taurocholate polypeptide (Huh7-Ntcp) were kindly provided by Dr. Simon Hohenester (University of Munich, Munich, Germany) and were grown in minimal essential medium containing 10% FBS, 1% nonessential amino acids, 2 mM of glutamine, 1 mM of sodium pyruvate, 100 units/mL of penicillin, 100 μg/mL of streptomycin, and 0.25 μg/mL of amphotericin B.20 Cells were grown to 80% confluence and serum-deprived for 24 hours before treatment with taurocholic acid (TCA), taurodeoxycholic acid (TDCA), taurochenodeoxycholic acid (TCDCA), 6-ECDCA, GW4064, or vehicle. In guggulsterone studies, cells were pretreated with 50 μM of guggulsterone for 1 hour before receiving either dimethyl sulfoxide (DMSO) or 6-ECDCA. For siRNA transfection studies, HepG2 cells were transfected with 7.5 pmol of FXR siRNA, 9.5 pmol of SHP, or scramble siRNA using Lipofectamine RNAiMax for 48 hours before GW4064 treatment, according to the manufacturer's protocol.

LUCIFERASE REPORTER ASSAY

The XBP1-luc reporter vector was kindly provided by Dr. Albert Koong (Stanford University, Palo Alto, CA) to assay the atypical splicing of the XBP1 mRNA.21 Huh7-Ntcp or HepG2 cells were transfected with 0.45 μg of XBP1-luc reporter gene and 0.05 μg of Pol III-Renilla luciferase vector as a control for transfection efficiency using Lipofectamine LTX transfection reagent per the manufacturer's instruction. After a 48-hour transfection, cells were treated with 6-ECDCA, GW4064, or DMSO for 8 hours. For SHP overexpression study, Huh7-Ntcp or HepG2 cells were cotransfected with a total of 2.0 μg of vector using pcDNA3 or human GFP-SHP (0.0-2.0 μg) expression plasmid22 with or without 0.45 μg of XBP1-luc reporter gene vector and 0.05 μg of Pol III-Renilla luciferase vector for 48 hours. At the end of the experiment, cells were harvested in passive reporter lysis buffer. Firefly and Renilla luciferase activities were measured with a dual-luciferase reporter assay kit (Promega, Madison, WI). All experiments were performed in triplicate. The luciferase activity was calculated as the ratio of Firefly/Renilla luminescence.

RNA EXTRACTION AND REAL-TIME QUANTITATIVE PCR

Total RNA was extracted from frozen liver or cell culture using TRIzol reagent according to the manufacturer's protocol (ThermoFisher, Waltham, MA). Real-time quantitative PCR (qPCR) was performed as described.7 All primers were synthesized by Integrated DNA Technology (Coralville, CA).

Co-immunoprecipitation (Co-IP)

HEK293T cells were transfected with pcDNA3, human SHP-GFP, or human FLAG-IRE1α23 plasmids using Calfectin reagent (SignaGen Laboratories, Rockville, MD) according to the manufacturer's protocol. Forty-eight hours posttransfection, cells were rinsed with cold phosphate-buffered saline, scraped into tubes, centrifuged, and pellets flash frozen in liquid nitrogen before –70°C storage. For immunoprecipitation (IP) experiments, pellets were lysed in 150 mM of NaCl, 20 mM of Tris, and 1 mM of ethylenediaminetetraacetic acid (NTEN) buffer containing protease and phosphatase inhibitors, cleared by centrifugation, and protein concentration estimated. Lysates (750 μg) were precleared with Protein A/G beads then immunoprecipitated overnight, 4°C, with 2 μg of SHP or 1 μg of FLAG antibodies. Immunoprecipitates were captured with Protein A/G beads and washed extensively with NTEN buffer, before elution by heating in 2× loading buffer. NTEN buffer alone was used as a negative control. Eluates and reserved lysate (input) were electrophoresed and transferred to polyvinylidene fluoride membranes and probed with indicated antibodies.

WESTERN IMMUNOBLOTTING

Protein homogenates from frozen livers or cultured cells were isolated and used with western immunoblotting as described.7 Nuclei preparations were performed using a nuclear extraction kit (Caymen Chemical, Ann Arbor, MI) according to the instructions of the manufacturer.

STATISTICAL ANALYSIS

Each experiment was repeated a minimum of three times. Data are shown as means ± SEM. Comparison between two groups was performed by Student t test. Statistical significance was defined as P values of less than 0.05.

DCA FEEDING INCREASES BILE ACID POOL AND HEPATIC XBP1 PATHWAY EXPRESSION

It has pbeen reported that feeding cholic acid to mice increases hepatic Xbp1s expression, presumably attributed to the presence of ER stress.24 We hypothesized that Xbp1s induction may be mediated, at least in part, by FXR signaling activation of p-IRE1α with a resultant increase in hepatic XBP1s expression. We fed mice with a diet containing 0.3% DCA for 0, 1, 3, and 7 days to enrich the bile acid FXR ligands. Nuclear expression of XBP1s protein was increased after 1 and 3 days of DCA feeding. Hepatic nuclear XBP1s was not detectable at baseline or at 7 days (Fig. 1A). p-IRE1α was not detected at any time point because of low level of expression in wild-type (WT) mice. Figure 1B demonstrates that hepatic gene expression of Xbp1s and its downstream target, ERdj4, increased 6-fold and 1.5-fold, respectively, in DCA-fed mice compared to chow-fed controls (P < 0.05). Figure 1C demonstrates that DCA feeding for 1 day resulted in a 2-fold increase in the total bile acid pool (P < 0.001) and significant increases in the FXR bile acid ligands, TCDCA, TCA, and TDCA. TCDCA and TDCA contents were 1.2 ± 0.3 and 1.8 ± 0.6 μmol/100g mouse, respectively, in DCA-fed mice, whereas both bile acids were nondetectable in chow-fed mice (P < 0.05). TCA increased 7-fold in DCA-fed mice, being 52.0 ± 3.6 versus 7.3 ± 1.6 μmol/100g mouse (P < 0.001). Serum ALT levels were similar in the mice fed chow or DCA for 1 day and increased at days 3 and 7 (Supporting Fig. S1); all liver histology was normal. Hepatic gene expression of the UPR genes, Atf4, Atf6, BIP/GRP78, binding immunoglobulin protein/glucose-regulated protein, 78 kDa (Bip), and C/EBP-homologous protein (Chop), did not increase at any time points, and, in fact, gene expression of Atf4 and Atf6 decreased at 7 days (Fig. 1B). This suggests that bile acid feeding may activate the liver XBP1s pathway.

BDL INCREASES HEPATIC XBP1s EXPRESSION

In order to determine whether the IRE1α/XBP1 pathway is activated in another model of cholestasis, BDL or sham laparotomy was performed on C57BL/6J mice for 48 hours. Serum ALT and alkaline phosphatase (ALP) were higher in the BDL group (ALT, 734 ± 412 vs. 28 ± 3 U/L; ALP, 251 ± 66 vs. 83 ± 6 U/L; P < 0.05 in BDL vs. sham, respectively). Figure 1D demonstrates that 48 hours after BDL, hepatic Xbp1s mRNA increased 4-fold (P < 0.001), and Fig. 1E shows that XBP1s nuclear protein increased significantly.

REDUCING BILE ACID POOL WITH CHOLESTYRAMINE DECREASES HEPATIC XBP1 PATHWAY EXPRESSION

Because expansion of the bile acid pool with DCA feeding may potentially cause ER stress, we fed mice a diet containing 2% (w/w) cholestyramine for 7 days to reduce the bile acid pool and measured the effect on liver IRE1α/XBP1 signaling. Figure 2 demonstrates that feeding cholestyramine for 7 days resulted in significant reductions of basal gene expression of Xbp1s, Xbp1u, and Ire1α by 58%, 46%, and 35%, respectively (P < 0.05), whereas basal ERdj4 expression did not change. Baseline p-IRE1α and XBP1s protein levels were not detected. Hepatic gene expression of Atf4, Atf6, Bip, and Chop was not diminished by cholestyramine feeding, and, in fact, Atf6 expression increased slightly at 7 days.

BILE ACIDS ACTIVATE IRE1α/XBP1 PATHWAY IN VITRO

Because the hepatic XBP1 pathway was activated by expansion of the bile acid pool and in response to BDL, and basal gene expression of the XBP1 pathway was reduced by contraction of the bile acid pool, we next investigated the in vitro effects of bile acids on the XBP1 pathway. Initial experiments demonstrated that treating Huh7-Ntcp cells for 6 hours with 0- to 100-μM concentrations of TDCA, TCDCA, and TCA increased XBP1s protein expression in a dose-dependent manner, with XBP1s protein expression increases evident at 10-μM bile acid concentration (Supporting Fig. S2). Therefore, we treated Huh7-Ntcp cells for 6 hours with 10 μM of TDCA, TCDCA, TCA, or vehicle and determined the effect on the XBP1 pathway. Gene expression of XBP1s, ERdj4, and IRE1α all increased in response to treatment with each bile acid, and SHP expression increased 5- to 8-fold (Fig. 3A). XBP1s protein expression similarly increased, with greater effects occurring in response to the FXR ligands, TCDCA and TCA (Fig. 3B). Therefore, the bile acid species that were enriched in vivo by DCA feeding also activate the IRE1α/XBP1 pathway in vitro.

FXR AGONISTS ACTIVATE IRE1α/XBP1 PATHWAY IN VITRO

Because bile acids are endogenous ligands of the nuclear receptor, FXR, we then treated Huh7-Ntcp cells and HepG2 cells with high-affinity pharmacological FXR agonists to determine the effects of FXR signaling on the XBP1 pathway. Figure 4A,B demonstrates that treating Huh7-Ntcp cells with the high-affinity FXR agonist, 6-ECDCA (0-10 μM), for 4 hours increased gene expression of XBP1s, ERdj4, and IRE1α, as well as XBP1s protein expression, in a dose-dependent manner. Figure 4C,D shows that treating HepG2 cells with the FXR ligand, GW4064 (0.0-2.5 μM), for 6 hours similarly increased both gene expression of XBP1s, ERdj4, and IRE1α and protein expression of XBP1s in a dose-dependent manner. No changes of XBP1s expression occurred within 2 hours of treatment in either cell type (Supporting Fig. S3A,B). As expected, both FXR agonists increased gene expression of SHP (Fig. 4A,C and Supporting Fig. S3C,D).

FXR INHIBITION AND KNOCKDOWN DECREASE FXR AGONISTS-ACTIVATED XBP1s PROTEIN EXPRESSION

We subsequently investigated the effects of the FXR inhibitor, guggulsterone, and FXR knockdown on both basal and FXR agonist-induced XBP1s protein expression. Huh7-Ntcp cells were treated with 10 μM of 6-ECDCA or vehicle, in the presence or absence of 50 μM of guggulsterone. Figure 5A demonstrates that both basal and 6-ECDCA-induced XBP1s protein expression were blocked by pretreatment with guggulsterone. In fact, protein expression of XBP1s in 6-ECDCA and guggulsterone cotreated cells remained below the basal XBP1s expression levels of untreated controls. Given that inhibitors may lack specificity, and to further confirm the inhibitory effect of guggulsterone, we used HepG2 cells transfected with FXR siRNA (or scramble siRNA) to demonstrate that FXR knockdown similarly decreased both basal and GW4064-induced increases in XBP1s protein expression (Fig. 5B). Supporting Fig. S4A confirms that FXR siRNA treatment reduced expression of FXR mRNA by over 70%.

FXR AGONISTS INDUCE XBP1 SPLICING AND IRE1α PHOSPHORYLATION

The atypical splicing of XBP1 into its active spliced form by IRE1α is a major regulatory mechanism for the IRE1α/XBP1 pathway.6, 25 We utilized an XBP1 splice activity luciferase-reporter gene construct (XBP1-luc) to measure the effect of FXR agonists on XBP1 splicing. Figure 6A demonstrates that treatment of Huh7-Ntcp and HepG2 cells with either 10 μM of 6-ECDCA or 2.5 μM of GW4064 for 8 hours increased XBP1 splicing compared to DMSO vehicle-treated cells (P < 0.01). DMSO-treated cells had identical luciferase activity to untreated cells.

The phosphorylated form of IRE1α (p-IRE1α) is the active form that splices Xbp1u into Xbp1s. Thus, we measured the expression of p-IRE1α in Huh7-Ntcp and HepG2 cells treated with 6-ECDCA and GW4064, respectively. Figure 6B shows that the expression of p-IRE1α in Huh7-Ntcp cells increased as early as 2 hours after the administration of 6-ECDCA, and Fig. 6C shows that p-IRE1α expression in HepG2 cells increased by 4 hours after treatment with GW4064. There were no increases in p-IRE1α expression at earlier time points in either cell types (Supporting Fig. S3A,B). These data indicate that FXR agonist-induced XBP1s expression was associated with increases in both p-IRE1α expression and XBP1 splice activity. Of note, the gene expression of polo-like kinase 3 (PLK3) was not changed at 1 or 2 hours after 6-ECDCA or GW4064 treatment, indicating that increased p-IRE1α expression was not associated with PLK3 gene induction.

SHP IS ASSOCIATED WITH FXR AGONIST-INDUCED ACTIVATION OF IRE1α/XBP1 PATHWAY

Because many FXR signaling effects are mediated by SHP, and SHP mRNA expression was up-regulated as early as 30 minutes or 1 hour in GW4064-treated HepG2 cells or 6-ECDCA-treated Huh7-rNTCP cells, respectively (Supporting Fig. S3C,D), we next overexpressed human SHP in HepG2 cells to determine the effect on the activation of IRE1α/XBP1 pathway. Figure 7A demonstrates a dose-response curve of transfections (0-2 μg) overexpressing SHP using a GFP-tagged SHP plasmid. SHP overexpression resulted in a dose-dependent increase of the XBP1-luc reporter luciferase activity, indicative of increased splicing activity. Figure 7B further shows that this was associated with increased XBP1s and p-IRE1α protein expression.

To further confirm the specific role of SHP in IRE1α/XBP1 pathway activation, we performed SHP knockdown experiments using SHP siRNA. Figure 7C demonstrates that SHP knockdown decreased GW4064-induced XBP1s and p-IRE1α protein expression in HepG2 cells. Supporting Fig. S4B confirms that SHP siRNA treatment reduced expression of SHP mRNA by 79%. These data suggest that FXR agonist-induced IRE1α/XBP1 pathway activation is, at least in part, mediated by SHP.

Although SHP acts as a transcriptional repressor localized primarily in the nucleus, SHP also localizes and interacts with proteins in other cellular compartments.26 SHP is highly expressed in microsome-enriched subcellular fraction.27 Because IRE1α is an ER resident protein, we investigated a potential interaction between SHP and IRE1α which may lead to the activation of the IRE1α/XBP1 pathway. SHP and IRE1α were overexpressed in HEK293T cells with GFP-SHP and FLAG-IRE1α plasmids. Co-IP experiments demonstrate a physical interaction between SHP and IRE1α using IP assays with a FLAG (Fig. 7D) or SHP antibody (Fig. 7E). These data suggest that SHP could possibly activate IRE1α/XBP1s through a nontranscriptional mechanism.

FXR/SHP SIGNALING REGULATES HEPATIC XBP1 EXPRESSION IN VIVO

In order to determine the in vivo effects of FXR/SHP signaling on hepatic XBP1s expression, we administered GW4064 (150 mg/kg, 16 hours) to WT mice. Hepatic XBP1s protein was minimally detected at baseline, but increased in response to GW4064 administration (Fig. 8A). Hepatic Xbp1s gene expression also increased in response to GW4064 (Fig. 8B). Hepatic Xbp1u gene expression did not alter after GW4064 treatment (Supporting Fig. S5). We also examined baseline expression of hepatic Xbp1s gene expression in FXR (–/–) and SHP (–/–) mice. Hepatic gene expression of Xbp1s was reduced by 52% in FXR (–/–) mice and 45% in SHP (–/–) mice compared to WT mice (Fig. 8C,D). Because baseline hepatic XBP1s protein expression was not detected, reductions could not be observed.