Bile acids increase intestinal markers via FXR/SNAI2/miR-1 axis in the stomach

Intestinal (IM) risk However, factors governing the transformation from normal gastric epithelial cells to IM remain unclear. Previously, miR-1 turned out to play an essential role in the initiation of bile acids (BA)-induced IM. Here, we investigate the mechanism underlying miR-1 inhibition by BA in gastric cells. We conducted IPA analysis to determine the potential molecular interacting BA with miR-1. The changes of FXR and SNAI2 after BA treatment were detected by western blot and qRT-PCR. IHC was performed to assess the expression level of FXR and SNAI2 in normal and IM tissue microarrays. The transcriptional regulation of SNAI2 or miR-1 was veried by bioinfamatics, luciferase reporter assay and chromatin immunoprecipitation PCR.


Abstract Background
Intestinal metaplasia (IM) is a precancerous lesion that increases risk of subsequent gastric cancer (GC). However, factors governing the transformation from normal gastric epithelial cells to IM remain unclear.
Previously, miR-1 turned out to play an essential role in the initiation of bile acids (BA)-induced IM. Here, we investigate the mechanism underlying miR-1 inhibition by BA in gastric cells.

Methods
We conducted IPA analysis to determine the potential molecular interacting BA with miR-1. The changes of FXR and SNAI2 after BA treatment were detected by western blot and qRT-PCR. IHC was performed to assess the expression level of FXR and SNAI2 in normal and IM tissue microarrays. The transcriptional regulation of SNAI2 or miR-1 was veri ed by bioinfamatics, luciferase reporter assay and chromatin immunoprecipitation PCR.

Results
BA treatment caused aberrant expression of FXR and IM markers in gastric cells. Augmented FXR led to the transcriptional activation of SNAI2 which further stimulates the expression of downstream IM markers. Bioinformatics analysis indicated that SNAI2 had miR-1 promoter binding region and we identi ed that SNAI2 negatively regulated miR-1 transcription. Both FXR and SNAI2 were increased in patients with IM.

Conclusions
This study demonstrated that FXR might be responsible for a series molecular changes in gastric cells after BA, and FXR/SNAI2/miR-1 axis exert a crucial role in BA-induced IM progression. Blocking the activation of the FXR-oriented axis may provide a promising approach for IM or even GC treatment.

Background
Gastric cancer is the fth most common malignant tumor but the third leading cause of cancer-related death worldwide [1]. We gradually recognized that gastric IM is one of the most common precancerous lesions of intestinal-type GC which follows the development model from chronic atrophic gastritis, IM, atypical hyperplasia and then to GC [2,3]. Preventing the development of IM may block the arising of intestinal-type GC. However, the mechanism underlying the occurrence of IM in gastric mucosa is far from clear.
It is generally believed that chronic environmental stimulation leads to gastric IM, such as Helicobacter pylori (Hp) infection and bile acids (BA) re ux [4,5]. However, whether Hp eradication could prevent the occurrence or further development of gastric cancer is still controversial [6,7], indicating other factors may contribute to the development of IM. BA is one kind of the important factors inducing or exacerbating IM [8]. Interestingly, a large number of studies have shown that BA stimulation could cause IM in the esophagus and the mechanism of its occurrence have been largely revealed [9][10][11][12]. However, there are few studies on the mechanism underlying the prometalpastic role of BA in gastric cells.
According to clinical statistical results, the risk of gastric IM in patients with bile re ux is 11 times higher than that in patients without bile re ux [13]. However, the mechanism of BA promoting gastric IM remains to be further studied.
MicroRNAs (miRNAs) causes the degradation of mRNA or the inhibition of translation by binding the 3'noncoding region (3'-UTR) of the target gene messenger RNA (mRNA) and play important roles in regulating biologic processes of cells [14,15]. At present, many studies have provided evidence that miRNAs participate in the development, differentiation and formation of the digestive tract, so it also exert great in uence in the occurrence of IM, such as miR-30, miR-194 and miR-490 [16,17]. Moreover, it should be noted that our team has successfully constructed a cell model of BA-induced gastric IM, and found that miR-92 promotes the expression of IM markers Caudal-related homeobox 2 (CDX2), Krüppellike factor 4 (KLF4) and VILLIN 1 (VIL1) by targeting FOXD1 [18]. We also discovered that miR-1 was signi cantly reduced in tissues of patients with IM and cell models of BA-induced IM, as well as the reduction of which promoted the signi cant increase of CDX2, MUC2, KLF4 via targeting both HDAC6 and HNF4α (manuscript submitted for publication). But, the mechanism of miR-1 depression triggered by BA remains to be elucidated. In the current study, we performed IPA analysis between miR-1 and BA potential receptor Farnesoid X Receptor (FXR, also known as NR1H4), Takeda G-coupled Protein Receptor 5 (TGR5), Small Heterodimer Partner (SHP), Constitutive and Rostane Receptor (CAR) and Vitamin D Receptor (VDR), and the results suggested that FXR might be the interacting protein.
FXR, BA activated receptor, is expressed in intestine and liver predominantly to maintain BA homeostasis [19,20]. FXR can also serve as a transcription factor to regulate the expression of downstream genes by directly binding to their promoter. With the increasing recognition of the important role of BA in the induction of IM of gastric mucosa, researchers began to study the role of FXR in this process. Yu JH et al.
found that upregulation of CDX2 and MUC2 in BA-induced gastric IM cells resulted from activated FXR/NF-κB pathway [21]. Similarly, one of our previous studies found that FXR is responsible for the enhancement of CDX2 by a SHP-dependent way in gastric cells [22]. However, the study on the mechanisms of FXR promotes BA-induced gastric IM is still far from enough.
Herein, we identi ed that FXR/SNAI2/miR-1 axis was dramatically activated in gastric cells after BA treatment and triggered an intestinal-like phenotype. Based on our results and recent ndings suggesting the role of FXR in IM and tumor, we hypothesized that FXR may be responsible for the molecular changes result from BA stimulation and facilitate succeeding progression of IM and GC.

Cell lines
The BA-induced IM cell model was described previously [18]. GES-1 and AGS (purchased from ATCC) were cultured in PRIM-1640 medium (Gibco, USA) with 10% fetal bovine serum (Gibco, USA), 100 mg/ml streptomycin and 100 U/ml penicillin. The primary cells were incubated in ICell Primary Epithelial Cell Culture System (icellbioscience, China) with 2% fetal bovine serum (Biocytosci, USA). Deoxycholic acid (DCA) was chosen for the following experiments as it is a major hydrophobic bile acid with potent cytotoxicity (BiocytoSci, USA).

Tissue microarray and human IM samples
A normal tissue microarray containing 24 cases (BN01011b) and a gastric IM tissue microarray (ST8017a) containing 80 cases of gastric IM cases were purchased from Alenabio Biotech (China). Ten paired human gastric IM specimens were obtained from patients who underwent gastroscopy. The pathological data of these specimens was collected from the Department of Pathology. Our study was approved by the ethics committee of Xijing Hospital. All patients signed informed consent before the specimens were obtained. To exclude the effect of Hp infection, all of the patients we selected were Hp negative.

Immuno uorescence
Cells were plated in 4-well chamber slides (Millipore, USA). Then, the cells were washed by twice with cold PBS, xed with 500μl of 4% paraformaldehyde for 35 minutes, permeabilized with 200μl 1% Triton X-100, for 10 minutes and blocked with 500μl of goat serum, for 30 minutes. Then the cells were incubated with primary antibodies at 4°C overnight. Then the cells were incubated with the FITC secondary antibody at room temperature for 2 hours. Finally, the nuclei were stained with DAPI (Solarbio) for 10 minutes. IF staining for FXR was performed in GES-1 cells. The primary antibodies were a rabbit anti-human FXR antibody (1:100, abcam, #187735).

RNA extraction and real-time PCR
TRIzol® reagent (Invitrogen, USA) was used to extract the total RNA from cell lines and human tissue samples according to a standard protocol. Then RNA was reverse transcribed into cDNA using the PrimeScript® RT Reagent kit (Takara Biotechnology, Japan) at 37˚C for 15 minutes and 85˚C for 5 seconds, followed by temperature maintenance at 4˚C. And qPCR was conducted using SYBR Premix Ex Taq II (Takara Biotechnology, Japan) on a CFX96™ Real-Time PCR Detection system (Bio-Rad Laboratories, USA). β-actin and U6 were used as the internal control and the 2 ∆∆Cq method was utilized to quantify the relative mRNA expression of each gene. Sequences of gene-speci c primers are shown in Table 3.

Protein extraction and western blot analysis
Total protein was extracted by mixing cells with radioimmunoprecipitation assay (RIPA) buffer (Beyotime, China), which was supplemented with protease and phosphatase inhibitors. Western blot analysis was carried out following standard procedures. The following antibodies were used for this experiment: anti-

Transfection
Synthetic miR-1 agomir, the corresponding negative control oligonucleotides and plasmid small interfering RNA (siRNA) were purchased from Genepharma (China), and the sequence of them was shown in Table 4. Attractene transfection reagent was added to Opti-MEM to generate the transfection medium. Cells were cultured in the transfection medium and it was replaced with fresh medium after 6 hours. Then the cells were grown in 5% CO 2 at 37°C for an additional 48 hours. The transfection regent was purchased from Thermo Fisher Scienti c (USA) and was used following the manufacturer's protocol. SNAI2 and FXR overexpression lentiviral vector were designed and provided by Genechem Co. Ltd.
(China). GES-1 cells were infected with the lentivirus for 48 hours and then cultured with 2 μg/ml puromycin for 7 days.
Chromatin immunoprecipitation GES-1 cells were transiently transfected with a miR-1 promoter vector. Then, ChIP analysis was performed according to the standard method of the Magna ChIP G Assay kit (Millipore, USA). Chromatin was immunoprecipitated with anti-FXR, anti-SNAI2 or the negative control IgG. Finally, immunoprecipitated DNA-protein complexes were isolated and a real-time PCR assay was carried out to examine the quantity of speci c proteins. The primers for the SNAI2 and miR-1 promoters are listed in Table 3.

Statistical analysis
The SPSS software (V. 19.0, SPSS, USA) statistical package was used to conduct all statistical analyses.
All continuous data are expressed as the mean±SD. The χ 2 test was used to compare the frequencies of categorical variables. Mutual associations among clinical results were assessed by using Spearman's rank correlation. Statistical comparisons between two groups were analyzed with unpaired Student's t test. P values less than 0.05 were considered statistically signi cant.

FXR expression is increased in intestinal metaplastic cells
In our previous study, decreased miR-1 was found to promote the transformation from normal gastric cells to IM cells after BA through targeting both HDAC6 and HNF4α. To uncover the mechanism by which BA inhibiting miR-1, we conducted IPA data analysis which showed that FXR might correlate BA with miR-1 (Fig. 1a). Then, we examined the protein levels of FXR in GES-1 cells treated with DCA (100μM) for 24 hours. As expected, DCA caused the increase of FXR along with the high expression of CDX2, KLF4 and MUC2 (Fig. 1b). Next, because of the de ciency in the mouse model with bile re ux, we attempted to utilize Sprague-Dawley (SD) rat primary gastric epithelial cells suffered from DCA to mimic the condition. Similarly, the result showed that FXR and intestinal markers were enhanced in primary cultured cells after DCA (Fig. 1c). In addition, IF staining for FXR was performed in GES-1 cells and the results recon rmed that DCA caused the enhancement of FXR in gastric cells (Fig. S1a). These results indicated that FXR might participate in the downstream activation induced by BA.
Further, to determine the signi cance of FXR in IM tissues of patients, we detected and analyzed its expression level in IM and normal tissues. Besides, high expression of FXR was observed more frequently in IM tissues and mainly located in the nucleus (Fig. 1d, 1e, Table 1). Moreover, the correlation analysis showed that the expression of FXR was negatively related to miR-1 (r = -0.2390, p = 0.0328) and positively correlated with target molecules HDAC6 (r = 0.4959, p = 0.0001) and HNF4α (r = 0.7665, p = 0.0001) of miR-1 (Fig. S1b, 1f, 1g) in IM tissues. Additionally, mRNA level of FXR was found to be higher in IM tissue cells Compared with normal tissues (Fig. 1h). Collectively, these results suggested that FXR might involve in the occurrence and development of IM.

FXR regulates downstream IM markers and miR-1
We used loss-of-function and gain-of-function experiments to study FXR function in regulating intestinal markers. And the results showed that the overexpression of FXR in GES-1 cells resulted in the increase of HDAC6 and HNF4α, as well as the overexpression of CDX2 (Fig. 2a). At the same time, it can be seen that the enhancement of FXR led to the downregulation of miR-1 (Fig. 2b). On the contrary, HDAC6, HNF4α and CDX2 decreased signi cantly after knockdown of FXR in AGS cells (Fig. 2c). The decrease of it also caused the augment of miR-1 (Fig. 2d). Together, these results indicated that FXR could not only mediate the downstream intestinal markers and miR-1 target molecules positively, but also suppress the expression of miR-1.

SNAI2 is enhanced in BA-induced IM cells
IPA data analysis revealed that FXR might in uence the expression of miR-1 through SNAI2, NCL, AKT and MET. Through bioinformatics analysis, we found that as a transcription factor, SNAI2 has the binding sites on the promoter region of miR-1. Then we detected the protein and mRNA level of SNAI2 in DCAtreated cells and found that it increased substantially, in contrast to control cells (Fig. 3a, 3b). Besides that, SNAI2 was also upregulated obviously on both protein and RNA levels in primary cells in response to DCA treatment (Fig. 3c, 3d). Moreover, it could be observed that SNAI2 was enhanced as well as positively correlated with FXR in IM tissues (Fig. 3e, 3f, 3g, Table 2). Its mRNA level was also signi cantly increased in IM tissues of patients in contrast to adjacent normal tissues (Fig. 3h). Collectively, these results suggested that SNAI2 was overexpressed in BA-induced IM cells.

SNAI2 is transcriptionally activated by FXR in gastric cells
To further clarify the role of FXR and SNAI2 in the process of BA-induced IM, we explored the regulatory relation between FXR and SNAI2. As showed in gure 4a, CDX2 enhancement induced by DCA could be reversed by FXR speci c siRNA. After that, GES-1 cells were infected with FXR expression vector and transfected with siSNAI2 simultaneously. It can be seen that the upregulation of SNAI2 and CDX2 induced by FXR could be inversed by siSNAI2 (Fig. 4b). On the contrary, we could see that the decrease of FXR resulted in correspondingly diminishing of SNAI2 and CDX2 in AGS cells, whereas no inhibition was observed after overexpression of SNAI2 (Fig. 4c). According to the prediction results of JASPAR database (http://jaspar.binf.ku.dk/), we constructed a series of reporter gene vectors containing different sequences upstream of SNAI2 transcription starting point. The results showed that the activities of SNAI2-p1 and SNAI2-p2 were higher than other sequences without any treatment, and the transcription activity of them was signi cantly higher in GES-1 cells after transfected with FXR expression plasmid (Fig. 4d). The results of ChIP assays further indicated that ChIP-1(GAGGTAATTAT) sequence might be the putative binding site of FXR on SNAI2 promoter (Fig. 4e). The above results indicated that FXR could promote the transcription of SNAI2 and induce its elevation in BA-induced IM cells.

SNAI2 promotes IM by inhibiting miR-1 transcriptionally
To further uncover the mechanism of pro-metaplastic function of SNAI2, we examined its functional regulation on downstream intestinal markers and miR-1. At rst, we could see that high expression of SNAI2 resulted in an evident decrease of miR-1 (Fig. 5a). Conversely, miR-1 was augmented after downregulating SNAI2 (Fig. 5b). When SNAI2 and miR-1 were overexpressed simultaneously, it could be found that miR-1 reversed the elevation of HDAC6 and HNF4α induced by SNAI2 (Fig. 5c, 5d). Furthermore, to ascertain whether SNAI2 could regulate the activity of the promoter of miR-1, we co-transfected SNAI2 expression vector with miR-1 full-length promoter reporter gene in GES-1 cells. The results showed that SNAI2 could evidently reduce miR-1 promoter activity (Fig. 5e). Then we constructed different promoter sequences: ChIP 1 (-1660bp~-1389bp); ChIP 2 (-1205bp~-982bp); ChIP 3 (-649bp~-420bp), ChIP NC (-5930bp ~-5672bp). ChIP assays were utilized to detect the binding sites of SNAI2 to these sequences. Finally, the results showed that ChIP 1 and ChIP 2 could bind to SNAI2 promoter directly, and the combination of them was reinforced by DCA (Fig. 5f). In summary, our ndings suggested that SNAI2 overexpression in BA-induced IM cells contributed to the reduction of miR-1 promoter activity and elevation of intestinal markers.

Discussion
In this study, for the rst time, we identi ed the aberrant expression of FXR and SNAI2 in IM cells and their key roles in BA-stimulated metaplastic progression. Interestingly, SNAI2 regulated miR-1 transcription negatively which further targeted downstream intestinal markers. Therefore, FXR/SNAI2/miR-1 axis mediates the expression of intestinal markers in IM cells. In addition, the axis triggered by DCA may promote the transformation of normal gastric cells to IM cells and mediate the further development of IM to GC. [22][23][24][25][26][27]. Recently, what attracts people's attention is the regulatory mechanism of CDX2 in intestinal differentiation. In a previous study, our group con rmed that increased miR-92 is involved in the upregulation of CDX2 and the downstream IM markers in gastric cells treated by BA [18]. Besides, our in vivo and in vitro experiments provided evidence that downregulation of miR-1 facilitated the formation of a positive feedback loop involving HDAC6/HNF4α and then stimulated CDX2 expression in BA-induced IM cells. Herein, we sought to elucidate the regulatory mechanism of DCA on miR-1 and to establish a molecular network of the alterations during BA-induced gastric IM progression. GES-1, a kind of immortalized gastric cell, is regarded as a non-malignant cell line which was used to imitate gastric epithelial cells. In addition, we detected the changes of target molecules in gastric mucosa cells of SD rats treated with BA.

Now, the mechanism of the transformation from normal gastric mucosal cells to intestinal cells is not clear. It is widely accepted that IM is the result of the interaction of a series of intestine related transcription factors and CDX2 plays an indispensable role in stimulating intestinal differentiation
A retrospective prospective follow-up study conducted in Japan found that the incidence of GC in patients with high level of BA in gastric juice (≥ 1000 mmol / L) signi cantly higher than patients with low gastric BA level (< 1000 mmol / L) [13]. Another multicenter study also showed that the risk of GC was 2.4 times higher in the high concentration BA group than in the other groups [8]. FXR regulates the biosynthesis, secretion and transport of BA as well as participates in various metabolic diseases [28]. And it is not only the principal nuclear receptor of BA, but also a transcription factor to regulate gene transcription [29]. So we speculated that it might involve in the stimulation of BA on gastric cells. In this study, we observed that the expression level of FXR increased in IM tissues as well as in GES-1 cells and primary cultured cells after BA treatment. We also provide evidence that FXR could stimulate the expression of downstream intestinal markers and the FXR speci c siRNA could abolish the enhancement of these markers induced by DCA. Consistently, there are some studies showed evidence that FXR was augmented in intestinal-type GC with IM as well as contributed to the increase of MUC2 and CDX2 in gastric cells [30,31]. Hence, we guess that there might be possible that FXR mediates the prometaplastic effect of DCA on gastric cells.
Based on IPA and bioinformatics analysis, SNAI2 was considered to connect FXR with miR-1. SNAI2 (also known as Slug), a member of Snail superfamily, is one of the transcription factors of epithelial mesenchymal transition (EMT) [32]. In recent years, it has been found that SNAI2 is abnormally expressed in various kinds of malignant tumors and plays an important role in tumor progress and prognosis [33][34][35]. To investigate the role of SNAI2 in IM cells, we examined its expression level in IM tissues and BA-induced IM cells. The results con rmed that it is positively correlated with FXR in IM tissues and could be activated by DCA in GES-1 cells as well as primary cultured cells. Further, we analyzed the sequence of SANI2 promoter and found that FXR might have binding sites on its promoter.
Next, data of ChIP and luciferase reporter gene assays con rmed that FXR could directly bind to SNAI2 promoter and activate its transcription. And increased IM markers via FXR overexpression or DCA could be blocked by SNAI2 knock-down which means that it could mediate the prometaplastic function of DCA. We also provided evidence that SNAI2 could stimulate the expression of intestinal markers in gastric cells and primary cultured cells. These results are partly consistent with some studies showing that SNAI2 promotes the malignant transformation of gastric cells through an EMT dependent manner [36][37][38][39]. Nevertheless, in this study, we rst identi ed that SNAI2 was augmented in IM tissues and could promote the expression of intestinal markers in gastric cells.
In terms of the downstream markers regulation, miR-1 was bioinfamatically identi ed as the functional targets of SNAI2. The dysregulation of miR-1 was proved to contribute to the aggressive progression and poor prognosis of human gastric cancer [40]. More interestingly, miR-1 was previously found to posttranscriptionally regulate SNAI2 to inhibit EMT in different kind of cancer [41,42]. Conversely, in this study, we proved that SNAI2 directly binds to the two sites (-1205bp~ -982 bp and − 649bp~ -420 bp) on the promoter of miR-1 to inhibit its transcription. In our previous study has shown that miR-1 could target 3'UTR of both HDAC6 and HNF4α to suppress their expression. Moreover, DCA inhibited miR-1 in gastric cells to induce high expression of HDAC6 and HNF4α which stimulated intestinal markers. Together with the results here, we identi ed a complex signal network through witch DCA induced the arising and continuous progression of IM in the stomach. Collectively, this study, for the rst time, showed evidence that BA-induced gastric IM is realized, in part, by FXR/SNAI2/miR-1 axis. However, it is noteworthy that the model systems of bile re ux, such as mouse model or primary cells from human gastric mucosa, will be needed in our future studies.

Conclusions
In summary, the ndings of this study revealed a schematic model of gastric IM development (Fig. 6). This gure describes that BA induces the upregulation of FXR in gastric cells to further stimulate the activation of SNAI2, leading to the transcriptional inhibition of miR-1, which nally promotes the transcription of downstream intestinal markers. Hence, FXR might serve as a critical regulator of BAinduced intestinal phenotype in gastric epithelial cells. This new FXR/SNIA2/miR-1 axis may provide a novel insight into the mechanism underlying the initiation and development of gastric IM. Suppression of FXR to abrogate the pathway may be a potential approach for preventing gastric IM or even GC in patients with bile regurgitation.  Overexpression of SNAI2 in IM cells. a GES-1 cells were treated as gure 1B. The protein level of SNAI2 was detected by immunoblots (Left) and the quanti cation of western blot analysis results was normalized as to β-actin (Right). b The mRNA level of SNAI2 was examined in GES-1 cells after DCA by qRT-PCR. c, d Primary cultured cells were treated by DCA as in GES-1 cells, and SNAI2 was detected at protein and mRNA levels respectively. e Representative pictures of IHC staining for SNAI2 in normal and gastric IM tissues. Scale bars: 100 µm; 500 µm (insets). f Quanti cation of IHC staining for SNAI2. g The correlation between SNAI2 and FXR in IM specimens. h SNAI2 mRNA level in 10 pairs of matched human IM specimens. *p<0.05; **p<0.01. N.S., not signi cant.