Inhibition of leydig tumor growth by farnesoid X receptor activation: The in vitro and in vivo basis for a novel therapeutic strategy



Leydig cell tumors (LCTs) are the most common tumors of the gonadal stroma and represent about 3% of all testicular neoplasms. In most cases, LCTs are benign; however, if the tumor is malignant, no effective treatments are currently available. We have recently reported that farnesoid X receptor (FXR) is expressed in R2C Leydig tumor cells, and it reduces the estrogen-dependent cell proliferation by negatively regulating aromatase expression. Here, we demonstrated that treatment with GW4064, a specific FXR agonist, markedly reduced Leydig tumor growth in vivo by inhibiting proliferation and inducing apoptosis. Indeed, the tumors from GW4064-treated mice exhibited a decrease in the expression of the proliferation marker Ki-67 and aromatase along with an increase in the apoptotic nuclei. FXR activation induced an enhanced poly(ADP-ribose) polymerase cleavage, a marked DNA fragmentation and a strong increase in TUNEL-positive R2C cells also in vitro. Moreover, in both in vivo and in vitro models, FXR ligands upregulated mRNA and protein levels of p53 and of its downstream effector p21WAF1/Cip1. Functional experiments showed that FXR ligands upregulated p53 promoter activity and this occurred through an increased binding of FXR/nuclear factor-kB (NF-kB) complex to the NF-kB site located within p53 promoter region as revealed by electrophoretic mobility shift assay and chromatin immunoprecipitation analysis. Taken together, results from our study show, for the first time, that treatment with FXR ligands induces Leydig tumor regression in vivo, suggesting that activation of FXR may represent a promising therapeutic strategy for LCTs.


The farnesoid X receptor (FXR) is an adopted member of the metabolic nuclear receptor superfamily, mainly expressed in the liver and the gastrointestinal tract, where it acts as a bile acid sensor.1, 2 FXR was first isolated from a rat-liver cDNA library and named after its weak activation by supraphysiological concentrations of farnesol, an intermediate in the mevalonate biosynthetic pathway.1 Shortly after its discovery, specific bile acids that both bind to the ligand-binding domain of the receptor and activate the transcription of FXR target genes were identified.3 Subsequent studies have led to the identification of potent synthetic FXR agonists, including GW4064, 6-ethyl chenodeoxycholic acids and fexaramine.4

Although FXR was originally identified in hepatocyte homeostasis, it has become increasingly clear that this nuclear receptor is important in a number of different cell types and mediates diverse functions including cell growth control and carcinogenesis.5, 6 Indeed, separate studies have established both positive and negative correlations between FXR expression and cancer. An early study showed that expression of FXR is inversely related to the progression of human colorectal cancers and the degree of malignancy of colon cancer cell lines.7 Activation of FXR markedly increases apoptosis, reduces proliferation and significantly blocks tumor growth in a xenograft mouse model of colon cancer cell lines by increasing the mRNA levels of several proapoptotic genes.8 Recently, it has been demonstrated that administration of chenodeoxycholic acid (CDCA) and GW4064 induces a significant inhibition of cholangiocarcinoma tumor growth.9 Additionally, FXR deficiency promotes intestinal10 and hepatic carcinogenesis.11, 12 In breast cancer cell lines, FXR agonists inhibit aromatase expression and induce apoptosis,13 and we have also shown that activated FXR reduces the membrane tyrosine kinase receptor HER2 expression and signaling, resulting in an inhibition of tamoxifen-resistant breast cancer cell growth.14 In contrast, the inhibition of FXR promotes apoptosis in Barrett's esophageal-derived cells15 and inhibits proliferation, migration and invasion in pancreatic cancer cells.16

In our previous study, we have demonstrated that FXR is expressed in tissues of normal and tumor Fisher rat testis and in Leydig normal and tumor cell lines.17 Particularly, in R2C rat Leydig tumor cells the FXR activators CDCA and GW4064 downregulate aromatase expression at both mRNA and protein levels, together with the inhibition of its enzymatic activity. FXR is able to compete with the steroidogenic factor 1 in binding to a common nuclear response element within the PII aromatase promoter, interfering negatively with its activity. Thus, FXR ligands exert antiproliferative effects on Leydig tumor cells at least in part through an inhibition of estrogen-dependent cell growth.

Leydig cell tumors (LCTs) are the most common tumors of the gonadal stroma and represent about 3% of all testicular neoplasms.18 Several studies on both rodents and humans indicate that excessive estrogen production plays a significant role in sustaining Leydig cell tumorogenesis. Transgenic mice overexpressing aromatase and exhibiting enhancement of 17β-estradiol (E2) circulating levels show Leydig cell hyperplasia and tumors.19 We have also reported that estrogens can elicit proliferative effects in rat tumor Leydig cells through an autocrine mechanism.20 Besides, elevated aromatase expression with subsequent high plasma E2 levels has been detected in human testicular LCTs.21, 22 While usually benign, about 10% of LCTs in adult patients exhibit a malignant phenotype, metastasizing to retroperitoneal lymphnodes, liver, lungs and bone. A malignant neoplasm is characterized by large size, marked nuclear pleomorphism, high mitotic rate, extensive necrosis and vascular invasion.23 Unfortunately, malignant LCTs respond poorly to chemotherapy or radiation and no effective treatments are currently available.

Here, we investigated the ability of FXR ligands to influence Leydig tumor cell growth in vivo using xenografts models. Our results provide further insights into the molecular mechanisms through which FXR may affect LTC malignancy.

Materials and Methods

Reagents and antibodies

Nutrient Mixture F-12 Ham, penicillin, streptomycin, fetal bovine serum, horse serum, CDCA were from Sigma (Milan, Italy). GW4064 was from Tocris, and Parthenolide was purchased from Alexias (San Diego, CA). TRIzol and Lipofectamine 2000 were provided by Invitrogen (Carlsbad, CA). FuGENE 6 was provided by Roche Applied Science (Indianapolis, IN). TaqDNA polymerase, RETROscript kit, Dual Luciferase kit, TNT master mix and nuclear factor-kB (NF-kB) protein were from Promega (Madison, WI). Antibodies against FXR, NF-kB, β-Actin, p53, p21, PARP, RNA polymerase II, Ki67 were by Santa Cruz Biotechnology (Santa Cruz, CA), antibody against Aromatase was by Serotec (Raleigh, NC). ECL system and Sephadex G-50 spin columns were from Amersham Biosciences (Buckinghamshire, UK). [γ32P]ATP was from PerkinElmer (Wellesley, MA).


The human p53 promoter-luciferase reporter constructs kindly provided by Dr. Stephen H. Safe24 were as follows: p53-1 (containing the −1,800 to + 12 region), p53-6 (containing the −106 to + 12 region), p53-13 (containing the −106 to −40 region) and p53-14 (containing the −106 to −49 region).

Cell culture

Rat Leydig tumor cells (R2C) were acquired in 2011 from American Type Culture Collection (LGC Standards, Teddington, Middlesex, UK) where they were authenticated, stored according to supplier's instructions and used within 4 months after frozen aliquots resuscitations.

In vivo experiments

The in vivo experiments were done in 45-day-old male nude mice (nu/nu Swiss; Charles River, Milan, Italy). At day 0, mice were inoculated with R2C cells (1.0 × 105/mouse) into the intrascapular region. GW4064 treatment was started at day 12 later and delivered daily to the animals by i.p. injection. Tumor growth was monitored as described.25 At the time of killing, 25 days, tumors were dissected out from the neighboring connective tissue, frozen in nitrogen and stored at −80°C. All the procedures involving animals and their care have been conducted in conformity with the institutional guidelines at the Laboratory of Molecular Oncogenesis, Regina Elena Cancer Institute, Rome.

Histopathological analysis

Tumors, livers, lungs, spleens and kidneys were fixed in 4% formalin, sectioned at 5 μm and stained with hematoxylin and eosin Y, as suggested by the manufacturer (Bio-Optica, Milan, Italy).

Immunohistochemical analysis

Paraffin-embedded sections, 5 μm thick, were mounted on slides precoated with poly-lysine, and then they were deparaffinized and dehydrated (seven to eight serial sections). Immunohistochemical experiments were performed as described,26 using rabbit polyclonal Ki67 primary antibody at 4°C overnight. Then, a biotinylated goat-anti-rabbit IgG was applied for 1 hr at room temperature, followed by the avidin biotin-horseradish peroxidase complex (Vector Laboratories, CA). Immunoreactivity was visualized by using the diaminobenzidine chromogen (Sigma-Aldrich). Counterstaining was carried out with methylene-blue (Sigma-Aldrich). The primary antibody was replaced by normal rabbit serum in negative control sections.


Tumor sections were incubated with anti-aromatase antibody, followed by rhodamine-conjugated secondary antibody. IgG primary antibody was used as negative control. 4′,6-Diamidino-2-phenylindole (DAPI, Sigma) was used for nuclei detection. Fluorescence was photographed using OLYMPUS BX51 microscope using a 20× objective.

Immunoprecipitation and immunoblot analysis

Cells were treated as indicated before lysis for total protein extraction.27 Nuclear extracts were prepared as described.28 For coimmunoprecipitation experiments, we used 1 mg of nuclear protein extract and FXR antibody (Santa Cruz Biotechnology) followed by protein A/G precipitation. Protein extracts from tumor tissues were homogenized in lysis buffer supplemented with 10% glycerol. Equal amounts of extracts and coimmunoprecipitated protein were subjected to SDS-polyacrylamide gel electrophoresis (PAGE), as described.27

TUNEL assay

Apoptosis was determined by Dead End TM Fluorometric terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) System (Promega) according to the manufacturer's instructions. Apoptotic nuclei were photographed by using a fluorescent microscope (20× objective).

DNA fragmentation

DNA fragmentation was determined as previously described.29

Total RNA extraction, reverse transcription polymerase PCR and real-time RT-PCR assay

Total RNA was extracted from R2C cells and from xenografts tissues using TRIzol reagent, and the evaluation of gene expression was performed by the reverse transcription-polymerase chain reaction (PCR) method using a RETROscript kit. The cDNAs obtained were amplified by PCR using the following primers: forward 5′-TTTCTACCCGCAACAACCGGAA-3′ and reverse 5′-GTGACAAAGAAGCCGCGAATGG-3′ (FXR); forward 5′-CAAGTCTGTTATGTGCACGTACTCA-3′ and reverse 5′-AACTGCACAGGGCATGTCTTC-3′ (p53); forward 5′-AGCAAAGTATGCCGTCGTCT-3′ and reverse 5′-ACACGCTCCCAGACGTAGTT-3′ (p21WAF1/Cip1); forward 5′-GAAATCGCCAATGCCAACTC-3′ reverse 5′-ACCTTCAGGTACAGGCTGTG-3′ (L19).

The PCR was performed for 35 cycles for FXR (94°C 1 min, 65°C 1 min and 72°C 1 min); 25 cycles for p53 (94°C for 1 min, 61°C for 1 min and 72°C for 1 min); 22 cycles for p21WAF1/Cip1 (94°C for 1 min, 57°C for 1 min and 72°C for 1 min), 18 cycles for L19 (94°C 1 min, 60°C 1 min and 72°C 2 min); in the presence of 1 μl of first strand cDNA, 1 μM each of the primers, 0.5 mM dNTP, Taq DNA polymerase (2 Units/tube) and 2.2 mM MgCl in a final volume of 25 μl.

Scavanger receptor class B type I (SR-BI) gene expression was evaluated by real-time RT-PCR. Five microliters of diluted (1:3) cDNA was analyzed in triplicates by real-time PCR in an iCycler iQ Detection System (Bio-Rad, Hercules, CA) using SYBR Green Universal PCR Master Mix, following the manufacturer's recommendations. Each sample was normalized on its 18S rRNA content. Primers used for the amplification were: forward 5′-CAAGAAGCCAAGCTGTAGGG-3′ and reverse 5′-CCCAACAGGCTCTACTCAGC-3′ (SR-BI). The relative gene expression levels were calculated as described.30

RNA interference

R2C cells were transfected with RNA duplex of stealth RNA interference (RNAi) targeted for the rat FXR mRNA sequence, 5-UCUGCAAGAUCUACCAGCCCGAGAA-3 (Invitrogen), or with a stealth RNAi control to a final concentration of 50 nM using Lipofectamine 2000 as recommended by the manufacturer. After 5 hr, the transfection medium was changed with serum-free medium and then the cells were exposed to treatments.

Transient transfection assay

R2C cells were transiently transfected using the FuGENE 6 reagent with different p53 promoter-luciferase reporter gene constructs (p53-1, p53-6, p53-13, p53-14). After transfection, R2C cells were treated with GW4064 3 μM for 24 hr. Thymidine kinase (TK) Renilla luciferase plasmid was used to normalize the efficiency of the transfection. Firefly and Renilla luciferase activities were measured by Dual Luciferase kit. The firefly luciferase data for each sample were normalized based on the transfection efficiency measured by Renilla luciferase activity.

Electrophoretic mobility shift assay

The probe was generated by annealing single-stranded oligonucleotides, labeled with [γ32P] ATP using T4 polynucleotide kinase and purified using Sephadex G50 spin columns. The DNA sequences used as probe or as cold competitors are the following (nucleotide motifs of interest are underlined and mutations are shown as lowercase letters): NF-kB, 5′-CTGGGATTGGGACTTTCCCCTCCCAC-3′; mutated NF-kB, 5′-CTGGGATTaatcCTTTCCCCTCCCAC-3′. In vitro transcribed and translated FXR proteins were synthesized using the T7 polymerase the rabbit reticulocyte lysate system. Electrophoretic mobility shift assay (EMSA) experiments were performed as described.31 For experiments involving anti-NF-kB and anti-FXR antibodies, the reaction mixture was incubated with these antibodies at 4°C for 12 hr before addition of labeled probe.

Chromatin immunoprecipitation assay

Chromatin immunoprecipitation (ChIP) assays were performed as described32 using anti-FXR, anti-NF-kB or anti polymerase II antibodies. A 3-μl volume of each sample and input were used for PCR using the primers flanking NF-kB sequence in the rat p53 promoter region 5′-AGCTTTGTGCCAGGAGTCTC-3′ and 5′-ACTGGAGCTTCAGAACTTTAG-3′ and the primers flanking FXR-RE sequence in SR-BI gene 5′-AAATCACCAGTCTGTCCTGATAGCCG-3′ and 5′-ACCTTTGTCTTCCCAGAGGGTCAT-3′. The amplification products were analyzed in a 2% agarose gel and visualized by ethidium bromide staining.

Statistical analyses

Each datum point represents the mean ± SD of three different experiments. Data were analyzed by Student's t test using the GraphPad Prism 4 software program. p < 0.05 was considered as statistically significant.


Activation of FXR inhibits R2C tumor xenograft growth

Based on our previous results demonstrating that FXR activation exerts antiproliferative effects on tumor Leydig cells in vitro through an inhibition of aromatase expression,17 we used the R2C tumor xenograft model to examine the effects of GW4064, a synthetic FXR agonist, on tumor growth in vivo. To this aim, we injected R2C cells into the intrascapular region of male nude mice and followed tumor growth after administration of GW4064 at 30 and 60 mg/kg/day. This administration was well tolerated because no change in body weight or in food and water consumption was observed along with no evidence of reduced motor function. In addition, no significant differences in the mean weights or histological features of the major organs (liver, lung, spleen and kidney) after sacrifice were observed between vehicle-treated mice and those that received treatment, indicating a lack of toxic effects at the dose given. As shown in Figure 1a, GW4064 (30 and 60 mg/kg/day) induced a regression in tumor growth. This effect was evident as early as day 6 of treatment, and tumor volumes continued to reduce over control for the duration of experiment. At the time of killing (25 days), tumor size was markedly smaller in animals treated with GW4064 when compared with vehicle-treated mice (Fig. 1b). Corresponding to their growth characteristics, histological examination of R2C xenografts by hematoxylin and eosin staining revealed necrotic central regions and regions of hemorrhage in GW4064-treated tumors (Fig. 1c). Expression mRNA levels of the known FXR target gene SR-BI33 have been evaluated as control (Fig. 1d).

Figure 1.

GW4064 treatment inhibits Leydig tumor growth in vivo. (a) R2C cells were injected in male nude mice. After 12 days, mice were assigned randomly in groups of six to receive 30 mg/kg/day GW4064, 60 mg/kg/day GW4064 or vehicle as control (C), and tumor growth was monitored over time. *p < 0.05, GW4064-treated groups versus control group. (b) Images of a representative individual tumor from each treatment group. (c) Hematoxylin and eosin stained histologic images of R2C xenograft tumors. (d) mRNA levels of the FXR target gene SR-BI, evaluated by real-time PCR, from R2C xenograft treated as indicated. *p < 0.05, GW4064-treated versus vehicle-treated tumors.

GW4064-treated tumors exhibit reduced cell proliferation and increased apoptosis

To determine whether the reduction in Leydig tumor growth induced by GW4064 was associated with any changes in the mitotic or apoptotic index, we evaluated in tumors the expression of Ki-67 as a marker of proliferation and the extent of DNA fragmentation as a marker of apoptosis. Sections of tumors from GW4064-treated mice exhibited a reduced expression of Ki-67 compared with that in tumors from control mice (Fig. 2a). In agreement with our previous in vitro findings,17 we observed in R2C xenografts a significant decrease of aromatase expression, evaluated by both immunoblot and immunofluorescence analysis (Figs. 2b and 2c). These latter results suggest that the reduced in vivo Leydig tumor growth is due at least in part to an inhibition of estrogen-dependent cell proliferation. Moreover, in GW4064-treated tumors we showed a marked increase in the apoptotic nuclei as evidenced by TUNEL assay (Fig. 2d), indicating that the induction of apoptotic pathways may represent an additional mechanism through which FXR ligands inhibit Leydig tumor cell growth.

Figure 2.

GW4064 treatment reduces proliferation and increases apoptosis in R2C tumor xenografts. (a) Representative pictures of Ki-67 immunohistochemical staining of R2C xenograft tumors. NC, negative control. (b) Protein extracts from xenografts excised from vehicle (C) and GW4064-treated mice were tested for aromatase expression by immunoblot analysis. β-Actin was used as a loading control. Numbers on the bottom of the blots represent the average fold change versus control of R2C xenografts normalized for β-actin. (c) Aromatase expression was determined by immunofluorescence analysis. NC, negative control. (d) TUNEL assay in R2C tumor sections from mice treated as indicated. DAPI staining was used to visualize the cell nucleus.

FXR ligands induce apoptosis in Leydig cancer cells “in vitro”

Next, the impact of FXR ligands on apoptosis was assessed in vitro by using different approaches. First, we evaluated the proteolysis of poly(ADP-ribose) polymerase (PARP), a known substrate of effector caspases, by immunoblot analysis. We found an increase in the levels of the proteolytic form of PARP in R2C cells upon both CDCA a natural ligand of FXR and GW4064 treatments (Fig. 3a). The second approach we used was to evaluate, in the same experimental condition, the internucleosomal fragmentation profile of genomic DNA, a diagnostic hallmark of cells undergoing apoptosis. FXR activation induced a marked DNA fragmentation in R2C cells (Fig. 3b). Finally, in TUNEL assay, we observed a strong induction of the apoptotic cells after CDCA and GW4064 exposure (Fig. 3c). A direct involvement of FXR in modulating apoptotic effects was provided by the evaluation of PARP cleavage after knocking down FXR in R2C cells with a specific short interfiring RNA (siRNA). As shown in Figure 3d, silencing the FXR gene reversed the increase in the levels of the proteolytic form of PARP induced by GW4064, whereas no changes were observed after transfection of cells with a control siRNA.

Figure 3.

Induction of apoptosis by FXR ligands in vitro. (a) Immunoblots of PARP from extracts of R2C cells treated with vehicle (C), CDCA (50 and 75 μM) or GW4064 (3 μM) for 24 hr. β-Actin was used as loading control. (b) DNA laddering and (c) TUNEL assay in R2C cells treated as indicated for 24 hr. DAPI staining was used to visualize the cell nucleus. (d) R2C cells were transfected with control siRNA or FXR siRNA for 24 hr and then treated with vehicle (C) or 3 μM GW4064 for 24 hr. Total proteins were extracted, and immunoblot analysis for PARP cleavage was performed. β-Actin was used as loading control. Numbers below the blots represent the average fold change versus control of R2C cells normalized for β-actin.

Activated FXR upregulates p53 and p21WAF1/Cip1 expression “in vitro” and “in vivo”

Considering that the tumor suppressor gene p53 plays an important role in apoptotic events,34, 35 we aimed to examine the ability of CDCA and GW4064 to modulate the expression of p53 along with its natural target gene p21WAF1/Cip1. Treatment with FXR ligands induced a significant increase in both p53 and p21WAF1/Cip1protein content and mRNA levels (Figs. 4a and 4b). These results were confirmed in vivo as we observed a significant upregulation in the expression of both p53 and p21WAF1/Cip1 genes in R2C tumor xenografts (Figs. 4c and 4d). These effects were completely reversed in the presence of the specific FXR siRNA, further supporting the direct involvement of this nuclear receptor (Fig. 4e).

Figure 4.

Upregulation of p53 and p21WAF1/Cip1 expression induced by FXR ligands. Immunoblot analysis for p53 and p21WAF1/Cip1 from extracts of R2C cells treated for 24 hr with vehicle (C), CDCA (50 and 75 μM) or GW4064 (3μM) (a), or from R2C xenografts treated with vehicle (C) or GW (30 mg/kg/day) (c) β-Actin was used as loading control. Total RNA from R2C cells (b) or from R2C xenografts (d) treated as indicated was reverse transcribed. cDNA was subjected to PCR using primers specific for p53, p21WAF1/Cip1 and L19 (internal standard). (e) R2C cells were transfected with control siRNA or FXR siRNA for 24 hr and then treated with vehicle (C) or 3 μM GW4064 for 24 hr. Total proteins were extracted, and immmunoblot analysis for p53 and p21WAF1/Cip1 was performed. β-Actin was used as loading control. Numbers below the images represent the average fold change versus control normalized for β-actin or L19 as indicated.

The FXR ligand GW4064 transactivates p53 gene promoter via NF-kB

The aforementioned observations prompted us to investigate whether FXR is able to transactivate p53 promoter gene. Thus, we transiently transfected R2C cells with a luciferase reporter construct (named p53-1) containing the upstream region of the p53 gene spanning from −1,800 to +12 (Fig. 5a). Treatment with GW4064 at 3 μM significantly induced luciferase activity (Fig. 5b). To identify the region within the p53 promoter responsible for its transactivation, we used constructs with deletions of different binding sites such as CTF-1 (CCAAT-binding transcription factor-1), NF-Y (nuclear factor-Y), NF-kB and GC sites (Fig. 5a). In transfection experiments performed using the mutants p53-6 and p53-13 encoding the regions from −106 to + 12 and from −106 to −40, respectively, the responsiveness to GW4064 was still observed (Fig. 5b). In contrast, in cells transfected with a construct encoding the sequence from −106 to −49 containing a deletion of the NF-kB domain (p53-14), the transactivation of p53 by FXR ligand was no longer noticeable (Fig. 5b). Concomitantly, the upregulation induced by the FXR ligand on p53 transcriptional activity was abolished in the presence of parthenolide, a specific inhibitor of NF-kB activation (Fig. 5c).

Figure 5.

Effects of GW4064 on p53 promoter activity. (a) Schematic map of the p53 promoter fragments used in this study. CTF-1, CCAAT-binding transcription factor-1; NF-Y, nuclear factor-Y, NF-kB, nuclear factor kB, GC, GC-rich motif. (b) p53 transcriptional activity in R2C cells transiently transfected with p53 gene promoter-luciferase reporter constructs and treated for 24 hr with GW4064 3 μM. (c) R2C cells were transiently transfected with p53 gene promoter-luciferase reporter construct p53-1 and treated for 24 hr with vehicle (-) or GW4064 3 μM alone or in combination with Parthenolide 20 μM. The luciferase activities were normalized to the Renilla luciferase as internal transfection control, and data were reported as fold change. Columns are mean ± SD of three independent experiments performed in triplicate. *p < 0.05; n.s.: nonsignificant.

Taken together, these data indicate that the integrity of the NF-kB-binding site is necessary for FXR modulation of p53 promoter activity in Leydig tumor cells.

The specific role of the NF-kB motif in the transcriptional regulation of p53 by FXR ligands was further investigated using EMSA. We observed the formation of a protein complex in nuclear extracts from R2C cells using synthetic oligodeoxyribonucleotides corresponding to the NF-kB motif (Fig. 6a, lane 1), which was abrogated by incubation with 100-fold molar excess of unlabeled probe (Fig. 6a, lane 2), demonstrating the specificity of the DNA-binding complex. This inhibition was no longer observed when mutated oligodeoxyribonucleotide was used as competitor (Fig. 6a, lane 3). Interestingly, treatment with GW4064 strongly increased the DNA-binding protein complex compared with control samples (Fig. 6a, lane 4). The inclusion of anti-FXR and anti-NF-kB antibodies in the reactions attenuated the specific bands, suggesting the presence of both proteins in the complex (Fig. 6a, lanes 5 and 6). Recombinant NF-kB protein revealed a complex migrating at the same level as that of nuclear extracts from cells (Fig. 6a, lane 7).

Figure 6.

FXR/NF-kB complex binds to NF-kB sequence within the p53 promoter region. (a) Nuclear extracts from R2C cells (lane 1) were incubated with a double-stranded NF-kB consensus sequence probe labeled with [32P] and subjected to electrophoresis in a 6% polyacrylamide gel. Competition experiments were done, adding as competitor a 100-fold molar excess of unlabeled probe (lane 2) or a 100-fold molar excess of unlabeled oligonucleotide containing a mutated NF-kB RE (Mut) (lane 3). Nuclear extracts from R2C were treated with GW4064 3μM (lane 4). Anti-FXR (lane 5), anti-NF-kB (lane 6) antibodies were incubated. Lane 7, NF-kB protein, lane 8, probe alone. (b) R2C cells were treated with vehicle (C) or GW4064 3 μM for 1 hr before lysis. FXR protein was immunoprecipitated using an anti-FXR polyclonal antibody (IP:FXR) and resolved in SDS-PAGE. Immunoblotting was performed using an anti-NF-kB and anti-FXR antibodies. Numbers below the blot represent the average fold change versus control. (c) R2C cells were treated with vehicle (C) or GW4064 3 μM for 1 hr, then crosslinked with formaldehyde and lysed. The precleared chromatin was immunoprecipitated with anti-FXR, anti-NF-kB, anti-RNA polymerase II or normal mouse IgG (NC) antibodies. The p53 promoter sequence containing NF-kB site and SR-BI regulatory region including FXR-RE were detected by PCR with specific primers. Input DNA was amplified from 30 μl of initial preparations of soluble chromatin (before immunoprecipitation). Numbers below the images represent the average fold change versus control normalized for input DNA.

To assess whether FXR and NF-kB may physically interact, we performed coimmunoprecipitation studies using nuclear protein fractions from R2C cells treated with GW4064. As shown in Figure 6b, the formation of an FXR and NF-kB complex was detected in untreated cells, and this association was enhanced upon FXR ligand treatment.

Moreover, to confirm the involvement of NF-kB in GW4064-mediated p53 upregulation at the promoter level, ChIP assays were performed. Using specific antibodies against FXR, NF-kB and RNA-polymerase II, protein–chromatin complexes were immunoprecipitated from cells cultured with or without GW4064 for 1 hr. The resulting precipitated DNA was then quantified using PCR with primers spanning the NF-kB-binding element in the p53 promoter region. The results indicated that FXR was constitutively bound to the p53 promoter in untreated cells, and this recruitment was increased upon GW4064 treatment. Interestingly, we observed upon GW4064 exposure a significant increase in NF-kB recruitment to the p53 promoter. The upregulatory role of FXR on p53 promoter is further evidenced by the enhanced RNA Pol II recruitment onto the p53 promoter (Fig. 6c). Next, the anti-FXR antibody immunoprecipitated on a region containing the FXR-RE site located within the SR-BI gene sequence was used as positive control (Fig. 6c).


The knowledge of the different functions of FXR has expanded rapidly from initial roles in controlling metabolism to regulating cell growth and malignancy. Indeed, recent studies have highlighted an oncosuppressive role for FXR in a variety of cancer cell types.8–14, 17, 36 However, direct evidence is missing for the in vivo effects of FXR activation in affecting Leydig carcinogenesis.

In our study, we have provided the first evidence that GW4064, a synthetic ligand of FXR, induces Leydig tumor regression by a mechanism that involves both inhibition of cell proliferation and induction of apoptosis. In line with these findings, FXR activation has been shown to increase apoptosis, reduce proliferation and block tumor growth in colon cancer cells.8 Our results evidenced in tumor sections from GW4064-treated mice a marked decrease in the expression of the nuclear proliferation antigen Ki67 as well as of the estrogen-producing enzyme aromatase. These observations well correlated with a recent published report from our research group showing that FXR activation by specific ligands, by modulating aromatase expression, is able to antagonize estrogen signaling and thus inhibit testicular tumor growth in vitro.17 Moreover, we observed that FXR activation induces apoptosis in GW4064-treated tumors as well as in R2C cells in vitro. Our data revealed upon FXR ligands exposure: (i) an increase in the levels of the proteolytic form of PARP, a crucial target that signals the presence of DNA damage and facilitates DNA repair; (ii) a marked DNA fragmentation, an hallmark of apoptosis and (iii) a strong increase of TUNEL-positive cells. These effects were abrogated in the presence of FXR siRNA confirming the crucial role of FXR in mediating apoptosis.

A large body of evidence has suggested the straightforward role of p53 signaling in the growth inhibition and in the apoptotic cascades.37–39 p53 acts as a tumor suppressor depending on its physical and functional interaction with diverse cellular proteins, like some nuclear receptors that, in turn, exert an inhibitory activity on p53 biological outcomes. Activation of p53 by UV damage or other agents/signals results in p53-mediated transcription or upregulation of genes such as the cyclin-dependent kinase inhibitor p21WAF1/Cip1 to induce apoptotic process, inhibiting the growth of cells with damaged DNA or cancer cells.39–41

We showed that FXR activation enhances both p53 mRNA and protein levels and induces the expression of the p53 target protein p21WAF1/Cip1 in R2C cells as well as in xenograft models. These effects were abrogated after knocking-down FXR addressing an FXR dependence of these events. Moreover, functional studies evidenced the ability of activated FXR to upregulate p53 promoter gene activity.

As a transcription factor, FXR binds to DNA either as a monomer or heterodimer with the retinoid X receptor to regulate the expression of various genes.33, 42–45 The analysis of p53 promoter sequence did not display any FXR response elements, thus it is reasonable to hypothesize that FXR-induced upregulation of p53 promoter activity may occur through its interaction with other transcriptional factors. Using different deletion mutants of p53 promoter gene, we demonstrated that the integrity of the NF-kB sequence is a prerequisite for FXR-mediated p53 transactivation.

NF-kB is an ubiquitously expressed transcription factor that regulates both proapoptotic and antiapoptotic signaling pathways depending on cell type, the extent of NF-kB activation and the nature of the apoptotic stimuli.46, 47 NF-kB has been shown to be required for the onset of apoptosis in several experimental systems,48 and inhibition or loss of NF-kB activity abrogated p53-induced apoptosis, indicating that NF-kB is essential in p53-mediated cell death.49 Moreover, it has been previously reported that NF-kB physically interacts with FXR,14 and the FXR agonist farnesol activates NF-kB signaling pathway and a number of NF-kB target genes.50 EMSA revealed a marked increase in a specific DNA-binding complex in nuclear extracts from R2C cells treated with GW4064, that was immunodepleted by anti-FXR or anti-NF-kB antibodies, suggesting the presence of these two proteins in the complex. Indeed, we observed a strong association of FXR with NF-kB in the nuclear fraction of untreated R2C cells that was further potentiated by GW4064 treatment. Furthermore, ChIP analysis showed that GW4064 increases the FXR/NF-kB occupancy of the NF-kB-containing promoter region concomitantly with an enhanced RNA Pol II recruitment. Finally, the FXR-mediated transactivation of p53 promoter was abolished in the presence of parthenolide, a specific NF-kB inhibitor.

Taken together, these data demonstrate that FXR enhances the recruitment of NF-kB to the NF-kB responsive element located within the p53 promoter region, resulting in an increase of p53 expression and thereby in Leydig tumor growth arrest.

In conclusion, the present and our previous findings17 show that the regression of Leydig tumors exerted by FXR ligands might rely on two mechanisms: a reduction of aromatase expression resulting in a decrease of estrogen-dependent growth and an increase of p53 expression inducing apoptosis. These observations suggest the possibility that FXR ligands could represent a promising pharmacological strategy for testicular cancer treatment.


This work was supported by Reintegration AIRC/Marie Curie International Fellowship in Cancer Research to I.B., AIRC MFAG 6180 to L.M., Lilli Funaro Foundation to S.P.