Farnesoid X receptor promotes renal ischaemia‐reperfusion injury by inducing tubular epithelial cell apoptosis

Abstract Purpose We investigated the role of farnesoid X receptor (FXR), a ligand‐dependent transcription factor, in renal ischaemia‐reperfusion (I/R) injury. Materials and Methods We performed unilateral renal I/R model in FXR knockout (Fxr−/−) and wild‐type (WT) mice in vivo and a hypoxia‐reoxygenation (H/R) model in vitro. The pathways by which FXR induces apoptosis were detected using a proteome profiler array. The effects of FXR on apoptosis were evaluated using immunoblotting, TUNEL assays and flow cytometry. Results Compared with WT mice, Fxr−/− mice showed improved renal function and reduced tubular injury scores and apoptosis. Consistent with the in vivo results, the silencing of FXR decreased the number of apoptotic HK‐2 cells after H/R, while FXR overexpression aggravated apoptosis. Notably, bone marrow transplantation (BMT) and immunohistochemistry experiments revealed the involvement of FXR in the tubular epithelium rather than in inflammatory cells. Furthermore, in vivo and in vitro studies demonstrated that FXR deficiency increased phosphorylated Bcl‐2 agonist of cell death (p‐Bad) expression levels and the ratio of Bcl‐2/Bcl‐xL to Bax expression in the kidney. Treatment with wortmannin, which reduced p‐Bad expression, inhibited the effects of FXR deficiency and eliminated the tolerance of Fxr−/− mouse kidneys to I/R injury. Conclusions These results established the pivotal importance of FXR inactivation in tubular epithelial cells after I/R injury. FXR may promote the apoptosis of renal tubular epithelial cells by inhibiting PI3k/Akt‐mediated Bad phosphorylation to cause renal I/R damage.

Ischaemia-reperfusion (I/R) injury is one of the leading causes of AKI. I/R generally induces significant tubular damage, which is characterized by cell flattening, tubular epithelium shedding, exposure of the underlying tubular basement membrane and luminal cast formation. 6,7 Therefore, it is of great clinical significance to understand the pathogenesis of renal I/R injury and to find effective preventative measures. Experimental animal models of renal I/R injury have been widely used to study the pathogenesis of ischaemic AKI. 1,8 In this study, we developed a unilateral I/R model with a consistent background to explore the pathophysiology of AKI.
Farnesoid X receptor (FXR), which was first found in the rat liver in 1995 by Forman, is a member of the nuclear receptor superfamily of ligand-dependent transcription factors. 9 It was named for its enhanced transcriptional activity in response to farnesol at concentrations above physiologic levels. 9 FXR has been shown to be essential in various physiological processes, including bile acid metabolism, glucose metabolism and lipid metabolism. 10,11 Recent research has revealed that in addition to the liver, small intestine and adrenal glands, FXR is also expressed in adipose tissue, heart, spleen and kidney. [12][13][14] Moreover, FXR plays an important role in the pathophysiological changes in blood vessels and participates in I/R injury of the small intestine, heart and other tissues. 12,15 The kidney expresses high levels of FXR. It has been reported that FXR has direct or indirect effects on the processes of kidney inflammation, oxidative stress, fibrosis and lipid metabolism. 16 However, the potential function of FXR in the kidney remains largely unknown. In this study, we investigated FXR expression in kidney tissues. We report that FXR is expressed in the renal epithelium, is a novel apoptosis mediator and contributes to renal I/R injury.

| Renal I/R model and drug treatment
A warm renal I/R model was established as described. 17,18 The details of the operation and the treatment of pharmacological agents are described in the Supplementary Information.

| Cell culture and treatment
The human proximal tubular cell line (HK-2) was acquired from the American Type Culture Collection (ATCC; Manassas, VA, USA). The details of the cell culture and the treatment of pharmacological agents are described in the Supplementary Information.

| Renal function, survival and histomorphological analyses
Plasma creatinine (Cr) and urea nitrogen (BUN) levels were measured with a standard spectrophotometric assay (Roche Diagnostic GmbH, Germany). The Kaplan-Meier survival analytical method was used to estimate the survival rate and to generate a survival curve for the mice.

| RNA sequencing (RNA-seq) and the identification of differentially expressed transcripts
Kidney tissues were sent to the Genminix Biological Company

| Mouse apoptosis proteome profiler array
To investigate the pathways by which FXR induces apoptosis, we examined apoptosis-related proteins using a proteome profiler array.
Details are provided in the Supplementary Information.

| Bone marrow transplantation (BMT)
BMT was performed as previously described ( Figure S1). 17,19 The details of BMT are described in the Supplementary Information. Renal I/R procedures were conducted 30 days after BMT.

| Transcriptional analysis and Western blot (WB) analysis
Kidney tissues or HK-2 cells were subjected to transcriptional or WB analyses. Experimental procedures, primer sequences and antibody information are described in the Supplementary Information.

| Small interfering RNA (siRNA)
siRNA duplexes targeting FXR, as well as non-targeted scrambled siRNA duplexes, were provided by Invitrogen (Life Technologies Corporation, NY, USA). The details of RNA interference and siRNA sequences are described in the Supplementary Information.

| Fluorescence-activated cell sorting (FACS) analysis
Flow cytometry was used to analyse apoptosis after H/R. Details are provided in the Supplementary Information.

| Polymerase chain reaction (PCR) genotyping
Routine PCR genotyping was performed to confirm the knockout allele in Fxr −/− mice. DNA was extracted from the tails of mice. Primer sequences were as follows: wild-type forward: TCTCTTTAAGTGATGACGGGAATCT; mutant forward: GCTCTAAGGAGAGTCACTTGTGCA; and common: GCATGCTCTGTTCATAAACGCCAT. These primers produced fragments of 291 bp in Fxr -/tissues. DNA from the tail of a wild-type mouse was 249 bp.

| Plasmid transfection
A plasmid for overexpression of FXR under the CMV was constructed by Genomeditech (Shanghai, China). The HK-2 cells were transfected with plasmid (3 μg) for 12 hours using Lipofectamine™ 3000 Transfection Reagent (Thermo Fisher Scientific, L3000150) according to the manufacturer's instructions.

| Statistics
All values are expressed as the mean ± the standard deviation of the mean (SD). Differences between two parameters were analysed by an unpaired Student's t test. Statistical significance was set at P < 0.05.

| Renal expression of FXR protein is upregulated after renal I/R
RT-PCR was used to detect FXR mRNA concentrations in the heart, liver, spleen, lungs, kidney and intestine. The highest FXR levels were in the liver. Low levels of FXR mRNA were detected in the heart, spleen and lungs. FXR mRNA was also highly expressed in the kidney, which was consistent with previous findings ( Figure 1A). 14 Previous studies have reported that renal I/R induces acute kidney disease. 6,20 Next, we established a renal I/R model. At different time points after the initiation of reperfusion following 20 minutes of ischaemia, the expression of FXR protein was analysed. FXR protein expression was increased after renal ischaemia by reperfusion at every point after ischaemia, demonstrating the highest expression at 6 hours after reperfusion ( Figure 1B). Time courses of renal FXR mRNA expression after I/R are shown in the Supplementary Information ( Figure S2).

| FXR deficiency alleviates kidney injury after I/R
To determine whether FXR was involved in renal I/R injury, we We further recorded the survival of Fxr −/− and WT mice for 7 days after reperfusion following 25 minutes of ischaemia ( Figure 1E).
Between 48 and 96 hours after renal I/R injury, all WT mice died, while most Fxr −/− mice survived (90%, P < 0.01). These data suggested that knockout of FXR protected mice from renal I/R injury.

| FXR deficiency attenuates I/Rinduced apoptosis
We assessed the role of FXR in renal I/R-triggered apoptosis by a TUNEL assay. As shown in Figure 2, TUNEL-positive tubular cells in I/R after excluding genes shared with controls. As shown in Figure 3B, we found that apoptosis, the HIF-1 signalling pathway, the ErbB signalling pathway, the PI3K-Akt signalling pathway and the MAPK signalling pathway were the five most enriched KEGG pathways. Strikingly, these pathways contributed to apoptosis.
Heat maps of the DEGs involved in these pathways are depicted in Figure 3C. In conclusion, RNA-seq revealed that FXR-mediated apoptosis contributed to the onset of I/R injury.

| FXR deficiency promotes post-I/R phosphorylation of Bad
To further determine the mechanism of apoptosis induced by FXR in I/R, we used a Mouse Apoptosis Signaling Pathway Array Kit.
After exposure to 24 hours of reperfusion, the kidneys were harvested, and apoptotic markers were examined using the array. In

| FXR deficiency in renal tubular epithelial cells alleviates renal I/R injury
Although the final result of renal I/R injury was renal parenchymal cell death, the severity of the injury depended in part on the inflammatory response. We performed BMT to investigate whether FXR induction in inflammatory cells was involved in renal I/R injury ( Figure S1). WT mice were lethally irradiated, and the bone marrow was reconsti-

| Silencing of FXR activates the phosphorylation of Bad and decreases the apoptosis of renal tubular epithelial cells
To determine the role of FXR in renal tubular epithelial cells, we knocked down FXR expression with specific siRNAs before these HK-2 cells were subjected to H/R. Compared with control siRNA-transfected cells, the cells transfected with FXR siRNA showed increased p-Bad expression after H/R ( Figure 6A,B). Consistent with the changes in vivo, after H/R, the level of p-Akt was significantly higher in FXR siRNAtransfected cells than in control siRNA-transfected cells ( Figure 6C,D).

| Overexpression of FXR contributes to apoptosis of renal tubular epithelial cells
To further elucidate the role of FXR in renal I/R injury, the effect of FXR overexpression on H/R injury was determined. We used a plasmid to overexpress FXR under control of the cytomegalovirus (CMV) promoter in HK-2 cells. After transfection, the cells were subjected to H/R. FXRoverexpressing cells were successfully generated, which was verified by WB analysis (Figure 7A

| Inactivation of Bad by wortmannin reduces the protection conferred by FXR knockout against I/R injury
Twenty-four hours after reperfusion, the levels of Cr and BUN were significantly higher in wortmannin-treated than vehicle-treated Fxr −/− mice ( Figure 8A,B), showing that wortmannin treatment in-

| D ISCUSS I ON
In this study, we demonstrated that (i) FXR is expressed in the kidney and that FXR deficiency protects the kidney from I/R-induced renal  AKI resulting from I/R involves multiple mechanisms, among which immune/inflammatory cells play a key role. [28][29][30] Our previous study demonstrated that renal macrophages and natural killer T (NKT) cells participate in the mediation of renal I/R injury and recovery. 18,31 FXR was found to be an important regulator of the inflammatory response. 32,33 FXR was confirmed to directly interact with NF-κB. NF-κB binds to the promoter of FXR and FXR target genes in models of inflammation. [34][35][36] To determine whether FXR in inflammatory cells was involved in I/R-mediated injury, we con- Previous studies have suggested that in renal I/R injury, necrotic cell death is rare and is limited to the outer medullary area, which is highly sensitive to ischaemia and hypoxia, whereas apoptosis is more common in other areas, especially in proximal and distal renal tubular cells. 7 Consistent with these reports, in our study, after I/R injury, there were more TUNEL-positive cells in In conclusion, the present study indicates that FXR insufficiency promotes the Bad phosphorylation-mediated anti-apoptotic effect, which has a protective role against renal I/R injury. These findings strongly suggest that FXR could be a therapeutic target for renal I/R injury. F I G U R E 6 Inhibition of Bad phosphorylation by wortmannin, an inhibitor of PI3K, aggravates the apoptosis of renal tubular epithelial cells. HK-2 cells were transfected with FXR siRNA (siFxr) or control siRNA (siCtrl). Cells were treated with wortmannin (Wort) or not and then exposed to hypoxia (1% O2) for 24 h followed by 6 h of reoxygenation. (A) Representative WB images and (B) expression ratio of p-Bad to t-Bad. (C) Representative WB images and (D) expression ratio of p-Akt to t-Akt. (E-G) Representative WB images of Bcl-2, Bcl-xL and Bax; left: expression ratio of Bax to Bcl-2; right: expression ratio of Bax to Bcl-xL. (H) and (I) H/R-induced apoptotic cells were counted by Flow cytometry. The densities of bands were quantified with ImageJ analytical software. Each column represents the mean ± SD. *P < 0.05 vs siCtrl after H/R, ## P < 0.01 vs siFxr after H/R F I G U R E 7 Overexpression of FXR in HK-2 cells leads to increased cell apoptosis. HK-2 cells were transfected with FXR plasmid (FXR OE ) or empty vector (Ctrl OE ). Cells were treated with wortmannin (Wort) or not and then exposed to hypoxia (1% O 2 ) for 6 h followed by 3 h of reoxygenation. (E) Representative images of TUNEL assays (original magnification, ×400; scale bar, 100 μm). Each column represents the mean ± SD. *P < 0.05 and **P < 0.01 vs wild-type mice F I G U R E 9 Schematic representation of FXR-induced renal I/R injury. The insufficiency of FXR facilitated Bad phosphorylation, isolating Bad from the Bcl-2/Bcl-xL complex and leading to less apoptosis after I/R

ACK N OWLED G EM ENTS
The Authors must give full details about the funding of any research relevant to their study, including sponsor names and explanations of the roles of these sources in the preparation of data or the manu-

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.