Farnesoid X Receptor Agonist GW4064 Protects Lipopolysaccharide-Induced Intestinal Epithelial Barrier Function and Colorectal Tumorigenesis Signaling through the αKlotho/βKlotho/FGFs Pathways in Mice

Compared with the WT mice, LPS-treated mice present with colon injuries that disrupt the epithelial barrier and cause infiltration of inflammatory cells into the mucosa and submucosa (Figure 1A). The colons of thirteen-month-old FXR-KO mice showed villi-form papillary folds and had moderately reduced colon crypt height and inflammatory cell infiltration (Figure 1A). However, LPS treatment in FXR-KO mice resulted in serious injuries such as loss of histological structures and increased inflammatory cell infiltration compared to LPS-treated WT mice. GW4064 restored the intestinal mucosal morphology and tight-junction disruption in LPS-treated WT mice but not in FXR-KO mice (Figure 1A). GW4064 treatment increased FXR protein levels in LPS-treated WT mice (Figure 1A–D). Histological score analysis, introduced in Figure 1B, further substantiates these observations. Treatment with LPS or GW4064 did not alter the FXR protein levels in KO mice (Figure 1A–D). The zonula occludens-1 (ZO-1) and claudin-1 protein levels appeared to be increased by GW4064 administration in WT mice after LPS treatment, whereas no such effect was observed in FXR-KO mice (Figure 1A,D,E). These data suggest that GW4064 rectifies LPS-induced intestinal barrier disruption and that this process requires functional FXR signaling.

GW4064 alters bile acid composition in LPS-treated mice. To determine whether GW4064 treatment alters BA profiles in the plasma, liver, and colon, BA species were assayed (Figure 2B–H). Compared with WT mice, LPS increased BA levels in the plasma and liver (~3-fold and ~2.25-fold, respectively). In contrast, the BA levels were decreased in the colon of WT mice (Figure 2D). FXR-KO mice displayed an increase in total BA pool, which was decreased in the liver and further enhanced the level of BAs in the colon of LPS-treated FXR-KO mice (Figure 2B–D). GW4064 reversed BA levels in the plasma, liver, and colon of WT and FXR-KO mice treated with LPS. Moreover, all FXR-KO mice showed considerably higher BA levels compared with WT mice, suggesting that FXR deficiency causes BA homeostasis disorder. Of particular interest, the levels of tauro-cholic acid (TCA), tauro-chenodeoxycholate (TCDCA), tauro-ursodesoxy cholic acid (TUDCA), and tauro-deoxycholic acid (TDCA) were increased in the plasma and liver of WT mice after LPS injection (Figure 2E–H), and GW4064 treatment was able to reverse most BAs and tauro-conjunction BA levels in the plasma and liver. In FXR-KO mice, the levels of TCA increased over 100-fold in the plasma and 25-fold in the liver compared with WT mice. Interestingly, TCA, TUDCA, and TDCA levels were decreased in plasma and liver as well as TCDCA in liver of LPS-treated FXR-KO mice. Although GW4064 administration almost reversed bile acid levels in the plasma and liver of WT and FXR-KO LPS-treated mice-, GW4064 markedly enhanced tauro-conjunction BA levels (Figure 2G,H), suggesting that these conditions lead to inflammatory disease susceptibility.

To test whether or not the alteration in the bile acid profile due to LPS treatment affects intestinal FXR signaling, the FXR signaling-related factors protein and mRNA expression was assessed. LPS enhanced MRP2, MRP3, and OATP1 and decreased OSTβ and ASBT protein levels in WT mice, and GW4064 almost completely abolished this condition (Figure 3A,B). FXR-KO mice had higher MRP2, MRP3, ASBT, and OATP1 and lower OSTβ compared with WT mice. LPS further enhanced MRP2, MRP3, and ASBT and reduced OSTβ protein levels in FXR-KO mice, and GW4064 administrated did not reverse this condition (Figure 3C,D). LPS reduced Fxr and pregnane X receptor (Pxr) and enhanced constitutive androstane receptor (Car), cytochrome P4503A11 (Cyp3a11), and sulfotransferase family 2A member 1 (Sult2a1) mRNA expression in WT mice, and GW4064 reversed this condition (Figure 3E). FXR-KO mice had lower Fxr and Pxr mRNA expression than WT mice, and their expression did not change after LPS and GW4064 administration in colon tissues (Figure 3E). Additionally, compared to WT, FXR-KO mice had higher Car and Cyp3a11, and lower Sult2a1 mRNA expression, which increased furthermore following LPS administration (Figure 3E). Interestingly, GW4064 further enhanced Car and Sult2a1 mRNA expression in LPS-treated FXR-KO mice (Figure 3E).

LPS also induced organic anion-transporting polypeptide 2B1 (Oatp2b1), Mrp2, Mrp3, multidrug resistance protein 1b (Mdr1b), Mdr2, and breast cancer resistance protein (Bcrp) and reduced apical sodium-dependent bile acid transporter (Asbt), ileal bile acid-binding protein (Ibabp), organic solute transporter α (Ostα), and Ostβ mRNA expression in WT mice, and GW4064 almost completely abolished this condition. FXR-KO mice had higher Asbt, Ibabp, Mrp2, Mrp3, Mdr1b, Mdr2, and Bcrp and lower Oatp2b1 and Ostβ as well as unperturbed Ostα mRNA expression (Figure 3F,G) compared with WT mice. GW4064 reduced Oatp2b1 and Mrp2 expression and increased Ostβ, Mdr2, and Bcrp mRNA expression furthermore in LPS-treated FXR-KO mice (Figure 3F,G). Hence, our results demonstrate that the imbalance in bile acid transporters due to LPS treatment could be significantly improved by GW4064 administration in WT mice, but this only partially reversed mRNA expression in FXR-KO mice (Figure 3).

To assess the underlying mechanistic basis whereby GW4064 could alter αKlotho /FGF23 and βKlotho/FGF19/FGF21 pathways, these pathways were subsequently analyzed via immunohistochemical and Western blot analyses. As shown in Figure 4A–C, GW4064 induced αKlotho, βKlotho, FGF19, FGF21, and FGF23 protein levels in LPS-treated mice. Compared with WT mice, we found significantly decreased αklotho, βKlotho, FGF19, FGF21, and FGF23 protein levels in FXR-KO mice (Figure 4A–D). Therefore, αKlotho, βKlotho, FGF19, FGF21, and FGF23 protein levels were reduced in LPS-treated FXR-KO mice; moreover, GW4064 only partial enhanced FGF19 and FGF23 in FXR-KO mice treated with LPS (Figure 4B,D).

In LPS-treated WT mice, GW4064 significantly decreased colonic mRNA expression of the proinflammatory genes, tumor necrosis factor α (Tnfα), interferon gamma (Ifnγ), and Il-1β (Figure 5A). GW4064 alleviated LPS-induced pro-caspase-1, adaptor protein apoptosis-associated speck-like protein containing a caspase recruitment domain (Asc), Nlrp3, and pannexin-1 mRNA expression in WT mice. Compared to WT mice, the FXR-KO mice had higher Tnfα, Il-1β, pro-caspase-1, Asc, Nlrp3, and pannexin-1 mRNA expression. GW4064 did not affect LPS-induced inflammatory mediator and Nlrp3 inflammasome in FXR-KO mice (Figure 5A). Furthermore, we investigated the effect of GW4064 on LPS-induced endoplasmic reticulum (ER) stress in the intestinal inflammation model. Compared to WT mice, the mRNA levels of activating transcription factor 4 (Atf4), Atf6, glucose-regulated protein 78 (Grp78), CCAAT-enhancer-binding protein homologous protein (Chop), and X-box binding protein 1 spliced (Xbp1s) increased remarkably upon LPS administration, which was markedly suppressed by GW4064 treatment (Figure 5B). FXR-KO mice had higher Atf4, Atf6, and Xbp1s mRNA levels than WT mice under un-induced conditions. Our results demonstrate that GW4064 had no effect on LPS-induced increase in ER stress in FXR-KO mice (Figure 5B). The LPS effects on colonic inflammatory response and tight-junction disruption and microbiota dysbiosis might be due to the upregulation of TLR4 signaling. Our results showed that GW4064 repressed LPS-induced TLR4, myeloid differentiation primary response 88 (MyD88), nuclear factor kappa-B (NF-κB), and caspase 3 protein levels in WT mice but not in FXR-KO mice (Figure 5C–G).

To investigate the mechanism contributing to the formation of angiogenesis and intestinal epithelial cell proliferation in mice treated with LPS, GW4064 significantly diminished β-catenin and c-myc protein levels in WT mice treated with LPS but not in FXR-KO mice (Figure 5H–J). Additionally, GW4064 significantly decreased LPS-induced transforming growth factor beta receptor II (TGFβRII), intercellular adhesion molecule (ICAM), vascular cell adhesion molecule (VCAM), vascular endothelial growth factor (VEGF), vascular endothelial growth factor receptor 1 (VEGFR1), and matrix metallopeptidase 9 (MMP9) protein levels in WT mice (Figure 5K). FXR-KO mice displayed higher levels of angiogenesis protein than WT mice. LPS stimulation in FXR-KO mice resulted in a dramatic increase in TGRβRII, VCAM, VEGF1, and MMP9 protein levels compared to untreated mice. GW4064 treatment had no effect on TGRβRII, VCAM, VEGF1, and MMP9 protein levels in FXR-KO mice under LPS treatment (Figure 5L). These data indicate that FXR activation decreases the expression of several inflammatory, Nlrp3 inflammasome, and ER stress genes by inhibiting TLR4 and β-catenin signaling, contributing to the amelioration of LPS-induced intestinal injury in WT mice.

We hypothesized that FXR activation may alter the growth of colon stem cells (CSCs), which is generally expressed in the intestine. Mechanistically, we found that the expression levels of intestinal stem cell markers and their proliferation, including CD44, LGR5, PCNA, CD34, BrdU, CD133, GCSF, and cyclin D1, were downregulated 30–40% upon GW4064 administration in WT mice treated with LPS (Figure 6A–E). In addition, GW4064 decreased the Lgr5, Olfm4, and cyclin D1 mRNA expression (Figure 6F) in WT mice treated with LPS but not in FXR-KO mice. Our results corroborated the attenuation of stem cell proliferation by FXR, as expression of cancer stem cells expansion and proliferation marker protein increased significantly in FXR-KO mice and was further enhanced with LPS treatment. These data suggest that FXR may play an important role in the early stage of colon cancer development.

To investigate the effects of GW4064 on gut microbiota in LPS-treated WT and FXR-KO mice, we examined the stool samples from WT and FXR-KO mice post treatment. Alpha diversity indices, including observed Chao1 and Shannon’s diversity indices, were lower in the LPS-treated WT mice than in the untreated mice; however, the diversity was restored to a normal level by GW4064 treatment (Figure 7A,B). The alpha diversity was reduced in FXR-KO mice compared with WT mice, and it further decreased with LPS treatment. Notably, GW4064 treatment did not alter the alpha diversity in these mice (Figure 7A,B). Principal component analyses (PCA) revealed that the gut microbiota in the LPS-treated group deviated from the untreated group, and GW4064 partially restored the level of gut microbiome in LPS-treated mice (Figure 7C). In addition, the gut microbiota profile of the FXR-KO group differed from that of the WT group.

Here, an analysis of 16S rRNA gene sequencing results revealed distinct clustering of colon microbiome communities isolated from WT and FXR-KO mice. Changes in the composition of the gut microbiota induced by GW4064 were noted, and it was observed that the colon community structure in LPS-injected mice was altered following GW4064 treatment in WT and FXR-KO mice (Figure 7D–Z). The phyla Proteobacteria, Verrucomicrobia, Bacteroidetes, class Betaproteobacteria, Gammaproteobacteria, Verrucomicrobiae, order Enterobacteriales, Verrucomicrobiales, genus Bacteroides, Escherichia, and species Bacteroides thetaiotaomicron, Escherichia coli, Bacteroides acidifaciens, Akkermansia muciniphila, and Clostridium saccharogumia were significantly increased, while phyla Firmicutes, the Firmicutes/Bacteroidetes (F/B) ratio, class Clostridia, and genus Clostridium were reduced in WT mice treated with LPS (Figure 7D–Z). However, compared with WT mice, specifically, the phyla Bacteroidetes, genus Bacteroides, and species Bacteroides thetaiotaomicron were increased, and the phyla Verrucomicrobia, the Firmicutes/Bacteroidetes (F/B) ratio, class Betaproteobacteria, Gammaproteobacteria, Verrucomicrobiae, order Verrucomicrobiales, Burkholderiales, genus Clostridium, Escherichia, species Escherichia coli, Akkermansia muciniphila, and Clostridium saccharogumia were significantly reduced in FXR-KO mice (Figure 7H–Z). The phyla Proteobacteria, Firmicutes, order Enterobacteriales, species Bacteroides acidifaciens, and Helicobacter hepaticus did not significantly change in FXR-KO mice (Figure 7E–Z). Interestingly, FXR-KO mice treated with LPS showed increased levels of Bacteroidetes, Betaproteobacteria, Clostridia, Bacteroides, Clostridium, and Helicobacter hepaticus. The abundance of specific bacterial species from Proteobacteria, Verrucomicrobia, Gammaproteobacteria, Clostridia, Verrucomicrobiae, Enterobacteriales, Verrucomicrobiales, Clostridium, Escherichia, Escherichia coli, and Akkermansia muciniphila exhibited oppositing changes in LPS treatment between WT and FXR-KO mice as compared to their respective control groups (Figure 7E–X). After GW4064 treatment, the abundance of Proteobacteria, Verrucomicrobia, Firmicutes, Betaproteobacteria, Gammaproteobacteria, Clostridia, Verrucomicrobiales, Enterobacteriales, Verrucomicrobiales, Clostridium, Escherichia, Bacteroides thetaiotaomicron, Escherichia coli, Bacteroides acidifaciens, Akkermansia muciniphila, and Clostridium saccharogumia were significantly reversed in LPS-treated WT mice, while only a few of these microbes were restored in LPS-treated FXR-KO mice (Figure 7E–Z). These results indicate that GW4064 has a potential role in the regulation of microbiota dysbiosis.

To understand the impact of GW4064 on intestinal injury, bile acid signaling, and changes in the gut microbiota, we further analyzed the variations in short-chain fatty acids (SCFA) and branched-chain fatty acids (BCFA). The presence of LPS has been observed to significantly decrease the concentrations of total SCFA and BCFA in the feces of normal mice (Figure 8A,B). The administration of GW4064 further reduces total SCFA, but interestingly, it significantly elevates the concentration of total BCFA. In the feces of FXR-KO mice, the concentrations of total SCFA and BCFA are significantly lower than those in normal mice. However, post LPS treatment, there is a significant increase in the concentrations of total SCFA and BCFA. When treated concurrently with GW4064, the concentration of total BCFA experiences a significant decrease (Figure 8A,B). Further analysis reveals that LPS significantly reduces the concentrations of acetic acid, butyric acid, valeric acid, isobutyric acid, isovaleric acid, 2-methylbutyric acid, and 4-methylvaleric acid in normal mouse feces. GW4064 administration significantly reverses these changes induced by LPS. However, GW4064 also reduces acetic acid concentration, which constitutes a large proportion of total SCFA. As a result, there is no significant improvement in the concentration of total SCFA. On the other hand, GW4064 significantly increases the changes in propionic acid, butyric acid, valeric acid, hexanoic acid, isobutyric acid, isovaleric acid, 2-methylbutyric acid, and 4-methylvaleric acid concentrations (except for formic acid and acetic acid) (Figure 8C–H). In FXR-deficient mice, apart from formic acid, the concentrations of other acids such as acetic acid, butyric acid, valeric acid, hexanoic acid, isobutyric acid, isovaleric acid, 2-methylbutyric acid, and 4-methylvaleric acid are significantly lower than those in normal mice. LPS treatment leads to a significant increase in the concentration of these acids. However, following the administration of GW4064, there is a significant decrease observed. The absence of FXR in FXR-KO mice primarily results in no significant changes in isovaleric acid concentration regardless of whether treatment is with LPS or GW4064 (Figure 8J). Therefore, FXR and GW4064 administration affects SCFA and BCFA concentrations in mice.

The FXR/αKlotho/βKlotho/FGFs pathway is crucial for maintaining the integrity of the intestinal barrier and preventing colon cancer. Upregulation of FXR helps protect against intestinal barrier dysfunction and colon carcinogenesis by improving tight-junction markers, reducing inflammation, and regulating β-catenin levels. GW4064, an FXR agonist, enhances FXR, αKlotho, βKlotho, FGF19, FGF21, and FGF23 protein levels in the colon, while FXR deficiency leads to a decrease in FXR/Klothos/FGFs protein levels and increased colon tumorigenesis markers. FXR plays a role in modulating gut microbiota and bile acid metabolism, highlighting the interplay between these factors (Figure 9).

Thirteen-month-old male C57BL/6J wild-type (WT) mice (Taiwan National Laboratory Animal Center, Taipei City, Taiwan) and FXR-knockout (FXR-KO) mice (strain B6.129X1(FVB)-Nr1h4tm1Gonz/J; stock number 007214, Jackson Laboratory, Bar Harbor, ME, USA) were maintained in the Animal Experimental Center of Chang Gung University under a 12-h light/dark cycle. Food and water were provided ad libitum. Lipopolysaccharide (LPS) is a vital component of the outer membrane in Gram-negative bacteria. LPS serves as the principal toxin derived from the gut microbiota, triggering inflammation in gut epithelial cells. Notably, the serum concentration of LPS is significantly elevated in patients with inflammatory bowel disease (IBD) [53], and LPS has been employed in experimental animal models to investigate intestinal injury [54]. In this study, WT and FXR-KO mice were intraperitoneally injected with a single dose of LPS (5 mg/kg) and after 12 h injected with GW4064 (20 mg/kg i.p. every 12 h) twice. After 6 h of the last GW4064 treatment, all animals were terminated by gradual fill isoflurane asphyxiation.

The colon tissues were fixed in 4% paraformaldehyde for 48 h and cut into 5 μm-thick serial sections, while hematoxylin and eosin (HE) staining was performed for histology. Then, the colon tissue damage was scored according to the histopatho-logical index (Supplementary Table S1) [55]. Slides were incubated with 7.5% non-immune goat serum for 30 min and then incubated with primary antibody for 2 h. Immunofluorescence (IF) and immunohistochemical (IHC) staining were performed as previously described [56]. The antibodies used for IF and IHC are listed in Supplementary Table S2.

The colon tissue protein extraction used NE-PER nuclear and cytoplasmic extraction reagents. The protein concentration was determined by the use of the Bio-Red Protein Assay Kit, in which protein extracts were added to the prepared SDS-PAGE for protein electrophoresis. Next, protein extracts were transferred onto PVDF membranes and blocked with 5% skim milk powder solution and then incubated with diluted primary antibodies overnight at 4 °C and incubated with HRP-conjugated secondary antibody for 1 h. The signals were detected using an enhanced chemiluminescence kit and quantified using ImageQuant 5.2 software.

RNA was isolated from the colon of mice using TRIzol reagent, and cDNA was generated from 1 μg total RNA using the High-Capacity cDNA Reverse transcription kit. qRT-PCR analysis was carried out using SYBR green PCR mastermix and analyzed on a LightCycler^® 1.5 Real-Time PCR System. Values were normalized to glyceraldehyde-3-phosphate dehydrogenase (Gapdh). The sequences of primers used for qRT-PCR are listed in Supplementary Table S3.

Fecal samples were collected and immediately transferred to a −80 °C freezer. Total bacterial DNA was extracted and purified using the QIAGEN mini stool kit (QIAGEN, Hilden, Germany). The total DNA was extracted with the reported methods [57]. The primers 16s_illumina_V3F (5′-CCTACGGGNGGCWGCAG-3′) and 16s_illumina_V4R (5′-GACTACHVGGGTATCTAATCC-3′) were used for bacterial 16S rRNA variable region V3-V4. The 16S rRNA sequences were compared for similarity with the reference species of bacteria using the NCBI nucleotide database and PhiX Control Library. To begin, sequences were clustered into operational taxonomic units (OTUs) using UPARSE [58]. To generate taxonomic assignments, Bowtie2 was used to align sequencing reads against the collection of a 16S rRNA sequences database [58]. A principal component analysis (PCA) plot was performed using ClustVis 2.0 software [59], which is a web tool for visualizing clustering of multivariate data using principal component analysis and heatmap from Oxford University.

Plasma (50 uL) was added to 150 μL of methanol with internal standard deoxycholic-2,2,4,4,11,11-d6 acid (DCA-d6; 809675, Merck, Rahway, NJ, USA) for protein precipitation and remained in −80 °C for 10 min [60]. Samples were centrifuged for 30 min at 12,000× g at 4 °C. The frozen liver and colon tissue were weighed and placed in homogenization tubes containing ceramic beads. The tissues were homogenized in a Precelly24 homogenizer two times over 15 s at 6000 rpm with 30 s pause intervals with extraction solvent (150 μL) containing internal standard. After homogenization, samples were centrifuged for 30 min at 12,000× g at 4 °C. The supernatant was analyzed with UPLC/MS/MS [61].

Frozen fecal samples were thawed at room temperature, then 50 mg of fecal material was combined with 1 mL of 10% isobutanol (J.T. baker 9044-01). The mixture was homogenized for 20 s at 6000 rpm with two cycles and a 30 s pause in between, using a Precelly24 homogenizer (Bertin Technologies, Montigmy le Bretonnexux, France). Subsequently, the homogenized mixture was centrifuged for 30 min at 12,000 rpm at 4 °C. Following centrifugation, 270 µL of the supernatant was transferred into a fresh glass vial. Sequentially, 50 µL of 20 mM NaOH and 160 µL of chloroform were added to the same glass vial for extraction. After vortexing for 5 min, the sample was centrifuged for 15 min at 1200× g. The upper aqueous layer (200 µL) was then carefully transferred to a new glass vial for further derivatization. For both the calibration standards and the aqueous layer (200 µL), the following steps were taken: 70 µL of the internal standard (100 µM butyrate-d8), 80 µL of isobutanol, 100 µL of pyridine, and 200 µL of ultrapure water were added to the sample. Subsequently, 50 µL of isobutyl chloroformate was carefully introduced into the sample or standard solution. To release the gases generated by the reaction, the cap was left open for 20 s, then closed and mixed for 1 min. The mixture was then sonicated for 5 min. Finally, 150 µL of hexane was added and thoroughly mixed. The vials were centrifuged at 1200× g for 10 min. The upper hexane phase was transferred into an autosampler sample vial and securely capped for GC/MS analysis.

The GC/MS analysis was conducted using a Bruker Scion 436 gas chromatograph/mass spectrometry system (Bruker Daltonics, Billerica, MA, USA) equipped with a VF-5 ms capillary column (30 m in length, 0.25 mm in diameter, and a 0.5 µm film thickness) from Agilent Technologies, Santa Clara, CA, USA. Analyte quantification was performed in the selected ion monitoring (SIM) mode. The temperature settings for various components were as follows: the injector was set at 260 °C, the ion source at 250 °C, and the transfer line at 280 °C. The helium carrier gas flowed at a rate of 1 mL/min. A 2 µL volume of derivatized sample was injected with a split ratio of 50:1. The temperature program for the column was as follows: it started at 40 °C and was maintained for 5 min, then ramped to 200 °C at a rate of 10 °C/min, and held for 1 min. Finally, the temperature was increased to 310 °C at a rate of 50 °C/min and maintained at this level for 3 min.

Results are presented as mean ± standard error of the mean (SEM). Statistical determined by the Student’s t-test and analysis of variance with Student’s Newman–Keuls multiple-range test as appropriate. All p-values values <0.05 were considered statistically significant.

Data is contained within the article and. supplementary material