Efficacy of auranofin as an inhibitor of desmoid progression

Anticancer drugs and molecular targeted therapies are used for refractory desmoid-type fibromatosis (DF), but occasionally cause severe side effects. The purpose of this study was to identify an effective drug with fewer side effects against DF by drug repositioning, and evaluate its efficacy. FDA-approved drugs that inhibit the proliferation of DF cells harboring S45F mutations of CTNNB1 were screened. An identified drug was subjected to the investigation of apoptotic effects on DF cells with analysis of Caspase 3/7 activity. Expression of β-catenin was evaluated with western blot analysis, and immunofluorescence staining. Effects of the identified drug on in vivo DF were analyzed using Apc1638N mice. Auranofin was identified as a drug that effectively inhibits the proliferation of DF cells. Auranofin did not affect Caspase 3/7 activity compared to control. The expression level of β-catenin protein was not changed regardless of auranofin concentration. Auranofin effectively inhibited the development of tumorous tissues by both oral and intraperitoneal administration, particularly in male mice. Auranofin, an anti-rheumatic drug, was identified to have repositioning effects on DF. Since auranofin has been used for many years as an FDA-approved drug, it could be a promising drug with fewer side effects for DF.

www.nature.com/scientificreports/ The purpose of this study is to identify an agent from among existing drugs that effectively suppresses DF cell proliferation using the drug repositioning method, clarify its underlying mechanism, and evaluate its efficacy in a mouse model spontaneously developing DF.

Results
Drug repositioning. The results of drug screening are shown in Fig. 1. Ninety-three drugs suppressed desmoid cell proliferation by less than 0.9-fold. Of the 93 drugs, highly toxic and expensive drugs were excluded. Auranofin was extracted as an effective and safe drug, which had a proliferation rate of 0.324, approved by the Pharmaceuticals and Medical Devices Agency (PMDA, Japan). Among the drugs that had an inhibitory effect on DF cells, felodipine was selected as a control drug for subsequent experiments because it was a PMDA-approved and safe drug having the second-most inhibitory effect (proliferation rate is 0.869) after auranofin. The top 10 drugs that showed high inhibitory effects on DF cell proliferation are shown in Table 1.

MTS assays of DF cells under drug administration.
The inhibitory effect of auranofin on the proliferation of DF cells was evaluated not only for cells with S45F mutation, which was used in drug screening, but also those with T41A mutation, and MRC-5 cells as a control fibroblast. Proliferation of S45F-mutated cells was significantly inhibited by 15.5% at 2 μM and 56.4% at 5 μM compared to 0 μM of auranofin (P = 0.004 and P = 0.002, respectively), and that of T41A was also significantly inhibited by 61.9% at 2 μM and 76.3% at 5 μM (P < 0.001 Figure 1. Screening for FDA-approved drugs. The S45F-mutated cells were cultured for 24 h in a 96 well culture plate of 5 × 10 3 cells/well in the presence of 10 μM 1200 FDA-approved compounds. The cell proliferation was measured using the MTS assay, and the absorbance (at 490 nm) was measured using a microplate reader. The number of drugs is plotted for each growth inhibition rate of desmoid cells. If the relative absorption is less than 1, the growth inhibition will be stronger. Auranofin was extracted as an effective and safe drug, which had a proliferation rate of 0.324. www.nature.com/scientificreports/ for each). Proliferation of MRC-5 cells was not inhibited up to 2 μM, but was by 27.4% at 5 μM (Fig. 2). MTS assays were also performed under felodipine treatment, however, felodipine did not inhibit proliferation of cells with T41A mutation compared to MRC-5 cells (Supplementary Figure S1 online). In addition, MTS assays were performed in T41A-mutated cells every 6 h up to 24 h after auranofin administration, with little change from 6 to 24 h (Supplementary Figure S2 online).
Apoptosis and reactive oxygen species assay. Auranofin, felodipine, and meloxicam did not show any significant increase in Caspase 3/7 activity as compared to the control. The results were similar for all DF cells of wild type (WT), T41A and S45F. (Fig. 3). The reactive oxygen species (ROS) transition of T41A-mutated cells treated with auranofin was observed, but there were no significant changes (Supplementary Figure S3 online).
Effects of auranofin on expression of β-catenin and downstream genes of Wnt/β-catenin signaling pathway. With T41A-mutated cells, the expression level of β-catenin was not affected by the administration of various concentrations of auranofin as compared with control in Simple Western assays (Fig. 4A) (see Supplementary Figure S4 online). We observed the distribution of β-catenin in T41A-mutated cells under 0-5 μM auranofin treatment using fluorescence immunostaining. The distribution of β-catenin in the nucleus and cytoplasm did not appear to change according to the auranofin concentration. However, the cell shape was slightly rounded by the administration of 5 μM auranofin (Fig. 4B).
The administration of auranofin, felodipine, and meloxicam resulted in various changes in mRNA expression of AXIN2 (AXIN-2), CCND1 (CYCLIN D1), MYC (c-MYC), and PTGS2 (COX-2). Notably, auranofin suppressed CCND1 in all of WT, T41A, and S45F cells. In the S45F-mutated cells, auranofin suppressed the expression of all four genes (Fig. 4C). Analysis of protein expression levels of these four genes in T41A-mutated cells with auranofin using Simple Western assay showed reduced expression of all proteins (Supplementary Table S1 online).  www.nature.com/scientificreports/ The mRNA expression levels of TCF1 / TCF1a / TCF3 / TCF4 on treatment with auranofin were measured. We found that TCF3 increased in a concentration-dependent manner with auranofin (Fig. 4D).
Effects on a spontaneous DF developing mouse model. Tumors had developed in the subcutaneous tissue, fascia, and muscle of Apc1638N mice by the time of sacrifice (6 months of age) (Fig. 5A). First, regarding the oral administration experiment, the average number of tumors in male mice treated with auranofin was 41 ± 4.6 in the control group (n = 7) and 29 ± 9.0 in auranofin group (n = 10), showing a statistically significant decrease in the number of tumors in auranofin group (P = 0.007) (Fig. 5B). On the other hand, the average number of tumors in female mice was 9 ± 2.9 in control group (n = 8) and 10 ± 2.5 in auranofin group (n = 13), with no significant difference (P = 0.86). There was no significant difference in the bodyweight of the mice between the control and auranofin groups (P = 0.36 for males and P = 0.41 for females).
In the experiments with intraperitoneal administration, the average number of tumors in male mice was 35 ± 5.6 in control group (n = 8) and 20 ± 4.0 in auranofin group (n = 8), indicating a statistically significant decrease in the number of tumors in auranofin group (P < 0.001). In female mice, the average number of tumors in control group was 16 ± 6.4 (n = 12) and 6 ± 2.5 in auranofin group (n = 9), showing a statistically significant decrease in the number of tumors in auranofin group (P < 0.001). There was no significant difference in the bodyweight of the mice between the control and auranofin groups (P = 0.44 for males and P = 1.0 for females).
The average size of the tumors that developed in all male mice was 1.9 mm and that in females was 1.6 mm. There was no significant difference in tumor size between control and auranofin groups for either oral or intraperitoneal administration (for oral administration in male mice P = 0.14, in female mice P = 0.38; for intraperitoneal administration in male mice P = 0.32, in female mice P = 0.24).
Histological evaluation showed less cellularity in the tumorous tissues in the auranofin-treated group as compared with control ( Fig. 5C). There seemed to be no difference in the nuclear staining of β-catenin between the auranofin-treated and non-treated groups (see Supplementary Fig. S5 online).

Discussion
When the "wait and see policy", which is the initial treatment modality for DF, fails, the use of effective medication is recommended 2 in locations other than the abdominal wall. This means Dox-based 6,7 , low-dose chemotherapy with MTX + VBL 8,9,11,13 , and molecular-targeted therapy with sorafenib 12 or pazopanib 9 . The therapeutic effects of these drugs have been proven, but severe side effects have also been reported. The development of a promising www.nature.com/scientificreports/ drug treatment for DF with fewer side effects is awaited. In the present study, by a drug repositioning method auranofin was identified as a possible agent. The pathogenesis of DF is mainly divided into two types: a germline mutation in the APC and a sporadic mutation in exon 3 of CTNNB1 gene. The treatment algorithm has been reported to be quite similar for both types 2 . However, several studies have shown that mutation types of CTNNB1 affect various treatment outcomes. Previous studies indicated that patients with S45F mutation have a high recurrence rate after surgical treatment 14,15 . Recently, some studies have analyzed whether mutation types affect the outcome of medical treatment. One study showed that DF patients with S45F mutation had a poor outcome with meloxicam, which is a selective COX-2 inhibitor NSAID 16 . Based on these previous reports, the S45F type is considered potentially to have an aggressive character in DF. For that reason, DF cells with the S45F mutation were used for drug screening in the present study.
Hotspot sites have been reported in the CTNNB1 mutation, which are thought to underlie approximately 85% of the cases of DF 17 . A review of the CTNNB1 mutation type in DF summarized eight analyzes, with T41A being the most common, followed by S45F, and WT 18 . Based on these reports, it seemed to make sense to perform the in vitro evaluation on these three types of cell in the present study.
Auranofin is an alkylphosphine gold coordination complex that has long been used clinically in the treatment of rheumatoid arthritis, and has shown stable safety in humans 19,20 . Several studies have been conducted to determine whether auranofin can be used as a therapeutic agent for malignant tumors, neurodegenerative diseases such as Alzheimer's disease and Parkinson's disease, parasitic infections, and bacterial infections 21 . One mechanism of auranofin activity has been thought to be mediated by the inhibition of reduction/oxidation (redox) enzymes such as thioredoxin reductase (TrxR) that are essential for maintaining intracellular levels of ROS 21 . By inhibiting these enzymes, auranofin increases ROS and induces apoptosis and further induces DNA damage by causing mitochondrial dysfunction [22][23][24] . Furthermore, referring to the relationship between ROS and the Wnt/β-catenin signaling, it has been reported that Ape1 and Nrf2 degrade or alter the localization of β-catenin via ROS 25,26 . In the present study, no obvious induction of apoptosis or increase in ROS was observed, suggesting that auranofin may inhibit cell proliferation without relating to these pathways. www.nature.com/scientificreports/ To our knowledge no reports have focused on the effects of auranofin on Wnt/β-catenin signaling. The results of the present study indicated that auranofin treatment partially suppressed the expression of genes downstream of the Wnt/β-catenin signal, suggesting that auranofin may act on this signaling pathway. In addition, when the intracellular localization of β-catenin was examined, changes in the distribution after drug administration were unclear with western blot and fluorescent immunostaining in vitro. These results suggest that auranofin acts after translocation of β-catenin into the nucleus, that is, in the process of binding to T-cell factor (TCF), and/or transcription of downstream genes. Therefore, we assayed the mRNA expression of TCF1 / TCF1a / TCF3 / TCF4 and found an increase in TCF3. Since TCF3 is generally known to suppress Wnt target gene expression 27,28 , the results of the present study may suggest that auranofin inhibits tumor growth through increasing TCF3. On the other hand, the possibility that some kind of gold transporter might play a role in the weak effect of auranofin in the S45F mutant cells cannot be ruled out. Further experiments are needed to elucidate the detailed mechanism of the action.
Two mouse models in which DF develops have been reported, a Ctnnb1 gene mutation model in which Ng2/ Cspg4-CreER mouse is crossed with a Ctnnb1 lox(exon3) mouse 29 , and an Apc1638N mouse model (Apc gene mutant model) 30 . To determine the more appropriate mouse model for use in the present experiment, a preliminary experiment was conducted to determine whether both models developed tumors. Since the number of tumors was low, and there was no reproducibility in the former model, the latter model, in which tumors sometimes occur, was used in the present study. APC forms a complex with glycogen synthase kinase 3 beta (GSK3B) and contributes to the stabilization of β-catenin. However, when a mutation occurs in APC, this complex inhibits phosphorylation of β-catenin, and β-catenin is no longer degraded and accumulates in cells. The translocation of accumulated β-catenin into the nucleus has been thought to result in the onset of desmoid [31][32][33] . Therefore, it seems to be appropriate to use Apc mutant mice to evaluate drug efficacy against DF although APC mutations have been reported to account for DF onset in 15% of cases. In the present study, auranofin was administered to these mice at a dose calculated to be equivalent to that used in humans as an anti-rheumatic drug suggesting clinical applicability.
The results of the preliminary experiments showed that the number of tumors of DF increased sharply after 6 weeks of age, which is in consistent with a report on Apc1638N mice by Smits et al. 30 . Regarding the number of DF tumors that develop, there is no major difference between the reports of Smits and the present study (Smits: 46 for male mice, 16 for females at the age of 6 months, present study: 41 for males, 9 for females). Similar results were also seen regarding the more frequent occurrence in male mice.
The results of the present in vivo study revealed that auranofin treatment suppressed the number of tumors but not reduce the tumor size. Due to the small size of the tumor that develops in this mouse model, it may be difficult to use this model for the purpose of assessing tumor size and growth. Although no size difference was observed between the auranofin-treated and control groups, the histological findings indicated that cellular density was lower in auranofin group than in control group, indicating the suppression of tumor aggressiveness by auranofin.
To apply therapeutic treatments for human, there is some advantages but it also remains unknown things for auranofin. First, in the mouse model used, the number of tumors that developed in males was high. On the other hand, in humans, DF occurs more frequently in women. Although the mechanism underlying the higher frequency of DF occurrence in male mice is unknown, intraperitoneal administration of auranofin significantly suppressed the onset of DF not only in male mice but also in female ones, suggesting that a tumor-suppressing effect can be expected in both genders. Second, since human desmoid shows mutations more frequently in CTNNB1 than APC, there is concern about the effectiveness of auranofin for patients with CTNNB1 mutation. There is a consensus statement that there is no need to change drug treatment options between desmoid that develops with an APC mutation in the germ line (Gardner's syndrome) and desmoid that has a CTNNB1 mutation in sporadic cases 2 based on the findings that both tumor types respond to drug treatment similarly 11 . Third, the time required for the effect of auranofin to appear has not been determined yet. For human use, it is necessary to clarify the time until the therapeutic effect of the drug appears.
Finally, the present study identified and confirmed auranofin as a potentially effective drug for DF by drug repositioning method. Auranofin is a drug that has been used for RA patients for many years and is inexpensive with relatively low toxicity. Although there are anticancer drugs and molecular-targeted drugs that have been shown to be effective against DF, auranofin is an FDA-approved drug and considered to be an easy-to-use drug for DF, which is an intermediate tumor. In order to obtain a proof of concept for clinical use, an investigatorinitiated clinical trial is now being planned. Methods DF cell culture. Informed consent was obtained from all patients. All research was performed in accordance with relevant guidelines/regulations, and the Declaration of Helsinki. Tumor tissues pathologically diagnosed as DF were obtained during incisional biopsy or excision. DF cells with three different mutation types on the gene, human CTNNB1 encoding β-catenin, were prepared: WT, T41A-and S45F. Tumorous tissues were cut into small fragments, and lysed by incubation at 37 °C for 3 h in Dulbecco's modified Eagle's medium (DMEM) supplemented with 0.2 mg/mL proteinase. Cells separated from lysates were mono-layer cultured in DMEM containing 10% fetal bovine serum (FBS) supplemented with 100 U/mL penicillin and 100 mg/mL streptomycin at 37 °C with 5% CO 2 . The culture medium was changed every 3-4 days, and cells were passaged by trypsinization before reaching confluence. Since beyond 20 passages, DF cells show senescence and are not suitable for experimentation, DF cells of passages 5-15 were used for the experiments. www.nature.com/scientificreports/ approved compounds (Prestwick Chemical Libraries, Illkirch-Graffenstaden, France). The cell proliferation was measured using the MTS assay (CellTiter 96 AQueous Assay, Promega, Madison, WI) according to the manufacturer's instructions, and the absorbance (at 490 nm) was measured using a microplate reader, Rainbow RC (Tecan, Männedorf, Switzerland).
Cell proliferation assay. The identified drug was auranofin, an anti-rheumatic drug. Auranofin was subsequently used for evaluation of various inhibitory effects on DF. The DF cells were seeded on a 96-well plate (5 × 10 3 cells/well) in DMEM containing 10% FBS for 12 h. Thereafter, the medium was replaced with a medium containing 10% FBS and 0.68% dimethyl sulfoxide (DMSO) or auranofin (dissolved in DMSO) at each concentration of 0.2 to 10 µM. The cell proliferation after 24 h was measured using the MTS assay kit, and the absorbance was measured using Rainbow RC. MRC-5 cells, fibroblasts originally developed from the lung tissue of male fetus (provided by the Riken BioResource Center, Tsukuba, through the National BioResource Project of the MEXT/AMED, Japan) were used as control cells. For felodipine, which was identified as a control drug, the MTS assay was performed in the same manner.
Apoptosis and ROS assay. The DF cells were seeded on a 96-well plate (5 × 10 3 cells/well) for 12 h in DMEM containing 10% FBS. Thereafter, the medium was replaced with a medium containing 10% FBS with DMSO or 5 µM of each drug (dissolved in DMSO), and the plate was incubated for 24 h. The drugs used in the analysis are: auranofin, felodipine that was the second most effective drug after auranofin, and meloxicam, which has been used for DF in our institution as an NSAID 4 . The activity of Caspase 3/7, which is an indicator of the apoptosis-inducing ability, was measured using a Caspase assay kit (Caspase-Glo 3/7 Assay, Promega) at the 24 h time point, and the luminous intensity was measured using a microplate reader, PowerScan4 (DS Pharma Biomedical, Osaka, Japan). Similarly, in the ROS assay, 2.5 × 10 4 T41A-mutated cells per well were seeded and incubated for 12 h, and then treated with a medium containing 10% FBS with DMSO or 1-10 µM of auranofin (dissolved in DMSO) for 24 h. The ROS Detection Assay Kit (BioVision, Waltham, MA) was used as described in its instructions, and fluorescence intensity was measured in the microplate reader.
Simple Western assays. The T41A-mutated cells were seeded on a 6-well plate (1 × 10 5 cells per well) in DMEM containing 10% FBS for 12 h. Thereafter, the medium was replaced with DMEM containing 10% FBS and DMSO with or without auranofin (dissolved in DMSO) at 0-10 µM concentration, and the plate was incubated for 24 h. These cells were collected and lysed on ice with radioimmunoprecipitation (RIPA) lysis buffer, and then centrifuged (1800 rpm, 10 min, 4 °C). Protein concentrations were determined using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA). Western blotting was performed using a Jess Simple Western system, and an automated capillary-based size sorting system (ProteinSimple, San Jose, CA). The expression level of β-catenin in these DF cells was evaluated according to the manufacturer's standard method for 12-230 kDa Jess separation module (SM-W004). Briefly, the extracted samples were mixed with 0.1X Sample buffer and Fluorescent 5X Master mix (ProteinSimple) to achieve a final concentration of 1 μg/μL in the presence of fluorescent molecular weight markers and 400 mM dithiothreitol (ProteinSimple). This preparation was denatured at 95 °C for 5 min. Ladder (12- Fluorescent immunostaining. The T41A-mutated cells in DMEM containing 10% FBS were seeded onto chamber slides (FALCON, Big Flats, NY) at 1.0 × 10 5 cells/ml and incubated for 12 h. The culture medium was replaced with 10% FBS-containing DMEM containing auranofin (dissolved in DMSO) at each 0-5 µM concentration and cultured for 24 h, followed by fixation with 4% paraformaldehyde for 2 h at room temperature. Cells were permeabilized in phosphate-buffered saline (PBS) containing 0.2% Triton X-100 by incubation at room temperature for 10 min. Preparations were incubated in blocking buffer containing 10% goat serum for 60 min at room temperature and incubated with mouse anti-β-catenin antibody (1:500 dilution, BD Biosciences, Franklin Lakes, NJ) at 4 °C overnight. Next, they were incubated with goat anti-mouse fluorescein isothiocyanate (FITC) secondary antibody (1:100 dilution, Proteintech) in a dark environment for 1 h at room temperature. Hoechst 33342 (5 μL/mL, Thermo Fisher Scientific) was added to them and incubated for 10 min at room temperature in a dark environment. mRNA expression analysis: reverse transcriptase-polymerase chain reaction (RT-PCR). To evaluate the influence of each drug on the Wnt/β-catenin signaling pathway, the mRNA expression levels of downstream genes AXIN2, CCND1, MYC, and PTGS2, 34 were measured. The DF cells (1 × 10 4 cells per well) with WT, T41A and S45F mutation were seeded in a 96-well plate in DMEM containing 10% FBS for 12 h. Subsequently the cells were cultured in a medium containing 5 μM of each drug for 6 h, and total cellular RNA was extracted using RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the instructions. The cDNA obtained by the RT was subjected to a real-time PCR reaction using LightCycler (Roche Diagnostics, Mannheim, Germany). In addition, for only T41A-mutated cells, mRNA expression of TCF1, TCF1a, TCF3, and TCF4 was www.nature.com/scientificreports/ evaluated in a similar manner. The expression level of mRNA of each gene was evaluated as a relative value, with reference to the expression level of GAPDH used as an internal control.  30 . In the same report, desmoid-like tumors were reported to develop frequently after approximately 6 weeks of age. Therefore, auranofin administration was started at 5 weeks of age in the present study. Mice were fed a diet with or without auranofin. As another experiment to change the route of administration, intraperitoneal administration and a control group were set up, and administration was started at 5 weeks of age. In terms of dose of the drug, auranofin normally used in humans is 6 mg/day, which is calculated to be 0.1 mg/kg/day assuming a bodyweight of 60 kg. Since the human equivalent dose for mice is 12.3 35 , oral dose for mice should be set at 1.23 mg/kg/day, but was set at 1.0 mg/kg/day in the present study. In terms of intraperitoneal dose, the bioavailability of auranofin in rats has been reported to be 20-30% with oral administration 36 compared with intraperitoneal administration. Oral dose (1.23 mg/kg X 7 days) appears to be the same as the intraperitoneal dose (3 mg/kg/week, bioavailability 20-30%). Therefore, the intraperitoneal dose was set as 1 mg/kg three times a week. Mice were sacrificed with the cervical dislocation procedure by a proficient person while sedated with 4% isoflurane. The number of tumors developed, tumor size, body weight, and histological features at 6 months after birth were compared between the auranofin and control groups, in male and female mice, respectively. A scheme of this experimental protocol is shown in Fig. 5D. In the oral experiment, 23 cases of auranofin group (10 male mice and 13 female mice) and 15 cases of control group (7 male mice and 8 female mice) were analyzed.
In the intraperitoneal experiment, 17 cases of auranofin group (8 male mice and 9 female mice) and 20 cases of control group (8 male mice and 12 female mice) were analyzed. Initially, we estimated at least 8 mice were required in each group, and assigned more mice to each. Some mice dying of unknown cause were present in each group, resulting in differences in the number of mice per group. All tumor samples detected were obtained by resection, fixed in 10% formaldehyde, and embedded in paraffin. Sections were cut to a thickness of 4 μm and subjected to hematoxylin-eosin staining and immunostaining using β-catenin antibody. According to a previous report 37 , the slides were treated overnight at 4 °C with anti-β-catenin antibody (ab2365, Abcam), counterstained with hematoxylin, dehydrated, and mounted.

Statistical analysis.
Mann-Whitney U test was used to compare the results of the cell proliferation assay, apoptosis assay, and RT-PCR. Differences in tumors developing in mice treated and untreated with auranofin were also analyzed by Mann-Whitney U test. Statistical analyses were conducted using SPSS 28 (IBM Corporation, Armonk, NY). In all analyses, P-values < 0.05 were considered statistically significant.
Institutional review board statement. This study was approved by Institutional Review Board of Nagoya University Hospital (approved number: 2014-0280).
Informed consent statement. Informed consent was obtained from the patients with DF involved in the study.

Data availability
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.