Cholestane-3β, 5α, 6β-triol Suppresses Proliferation, Migration, and Invasion of Human Prostate Cancer Cells

Oxysterols are oxidation products of cholesterol. Cholestane-3β, 5α, 6β-triol (abbreviated as triol) is one of the most abundant and active oxysterols. Here, we report that triol exhibits anti-cancer activity against human prostate cancer cells. Treatment of cells with triol dose-dependently suppressed proliferation of LNCaP CDXR-3, DU-145, and PC-3 human prostate cancer cells and reduced colony formation in soft agar. Oral administration of triol at 20 mg/kg daily for three weeks significantly retarded the growth of PC-3 xenografts in nude mice. Flow cytometric analysis revealed that triol treatment at 10–40 µM caused G1 cell cycle arrest while the TUNEL assay indicated that triol treatment at 20–40 µM induced apoptosis in all three cell lines. Micro-Western Arrays and traditional Western blotting methods indicated that triol treatment resulted in reduced expression of Akt1, phospho-Akt Ser473, phospho-Akt Thr308, PDK1, c-Myc, and Skp2 protein levels as well as accumulation of the cell cycle inhibitor p27Kip. Triol treatment also resulted in reduced Akt1 protein expression in PC-3 xenografts. Overexpression of Skp2 in PC-3 cells partially rescued the growth inhibition caused by triol. Triol treatment suppressed migration and invasion of DU-145, PC-3, and CDXR-3 cells. The expression levels of proteins associated with epithelial-mesenchymal transition as well as focal adhesion kinase were affected by triol treatment in these cells. Triol treatment caused increased expression of E-cadherin protein levels but decreased expression of N-cadherin, vimentin, Slug, FAK, phospho-FAK Ser722, and phospho-FAK Tyr861 protein levels. Confocal laser microscopy revealed redistribution of β-actin and α-tubulin at the periphery of the CDXR-3 and DU-145 cells. Our observations suggest that triol may represent a promising therapeutic agent for advanced metastatic prostate cancer.


Introduction
Prostate cancer is the second most frequently diagnosed cancer of men and the fifth most common cancer overall in the world. In 2008, more than 899,000 new cases were diagnosed (GLOBO-CAN 2008 database, version 1.2). In Western countries, prostate cancer is the most common non-cutaneous carcinoma of men. According to the statistics of Surveillance Epidemiology and End Results (SEER) of the National Cancer Institute, more than 240,000 men were diagnosed with and more than 28,000 men died of prostate cancer in 2012 in the United States. Although surgery is often successful for organ-confined prostate cancer, androgen ablation therapy is the primary treatment for metastatic prostate cancer. Unfortunately, most prostate cancer patients receiving androgen ablation therapy will ultimately develop recurrent, castration-resistant tumors within 1-3 years after treatment. The median overall survival time is 1-2 years after cancer relapse [1,2]. No effective standard therapy exists for patients that relapse with advanced prostate cancer. Chemotherapy is often used to treat metastatic hormone-refractory prostate cancer [2,3]. However, chemotherapies generally show little effect on prolonging survival. Therefore, new treatments for advanced prostate cancers are needed.
Cholestane-3b, 5a, 6b-triol (abbreviated as triol) is one of the most abundant oxysterols. Triol is derived from cholesterol by oxidation via formation of 5a, 6a-epoxycholesterol and 5b, 6bepoxycholesterol [15,16] as intermediates. Previously, 5a, 6aepoxycholesterol was reported to exhibit anti-cancer activity [13]. In this study, we examined the ability of triol to suppress the proliferation of advanced human prostate cancer cell lines both in vitro and in vivo. We applied Micro-Western Arrays, a recently developed antibody-based, high-throughput Western blotting assay [17][18][19][20], to study the signaling proteins affected by triol in advanced prostate cancer cells. Our observations suggested that triol may represent a promising therapeutic agent for advanced metastatic prostate cancer.

Cell Proliferation Assay
Cells were seeded at a density of 3610 3 cells/well in 100 ml complete medium in 96-well plates. Proliferation assays were performed under maintenance conditions (DMEM with 10% FBS for PC-3 and DU-145; DMEM with 10% CS-FBS for LNCaP CDXR-3). Relative cell number was analyzed by measuring the DNA content of cell lysates with the fluorescent dye Hoechst 33258 (Sigma, St. Louis, MO, USA) as described previously [8,10,11,21,23,24,27,28]. All readouts were normalized to the average of the control condition (no triol treatment) in each individual experiment. The experiment was repeated three times. Ten wells were used for each condition. The mean and standard deviation represented the average and standard deviation respectively of the results from all 30 wells in the three experiments.

Cell Viability Assay
Cells were seeded at a density of 3610 3 cells per well in a 96well plate (BD Bioscience). After 24 hrs, the cells were treated with increasing concentrations of triol for 48 hrs or 96 hrs. Cell viability was assessed by an MTT (3,4,5-dimethylthiazol-2-yl)-2-5-diphenyltetrazolium bromide) assay [29]. The amount of formazan was determined by measuring the absorbance at 560 nm using an Tecan GENios TM plate reader (Tecan group Ltd, Mä nnedorf, Switzerland) [29]. All results were normalized to the average of the control condition (no triol treatment) in each individual experiment. The experiment was repeated three times. Each time ten wells were utilized for each condition. The mean and standard deviation represented the results from all 30 wells in the three experiments.

Flow Cytometric Analysis
Cells were seeded at a density of 5610 5 cells in 10-cm dishes in 10 mL of media and cultured for 24 hrs before addition of triol. After 48 hrs of culture in the presence of various concentrations of triol, cells were removed with trypsin and fixed in 70% ethanol in phosphate buffered saline (PBS) overnight at 220uC. Fixed cells were washed with PBS, treated with 0.1 mg/mL RNase A in PBS for 30 min, and then suspended in PBS containing 50 mg/mL propidium iodide. Cell cycle profiles and distributions were determined by flow cytometric analysis of cells using a BD Facscan flow cytometer (BD Biosciences). The cell cycle distribution was analyzed using ModFit LT software (Verity Software House, Topsham, ME, USA) as described [10,23,24,26,27].

TUNEL Assay
Cells were cultured on cover slips in 24 wells and were treated with 0, 20, and 40 mM triol for 48 or 96 hrs. Cells were rinsed twice with PBS and subjected to the TUNEL assay using ApoAlert DNA Fragmentation Assay Kit (catalog no. 630108 from Clontech, Mountain View, CA, USA) according to the manufacture's instruction. The TUNEL-stained cells were observed with Olympus confocal microscope at 200 6(FV300, Olympus, Tokyo, Japan).

Ethics Statement
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (Taiwan). The protocol was approved by the National Health Research Institutes Institutional Animal Care and Use Committee (NHRI- Figure 3. Effect of triol on the growth of PC-3 xenografts in nude mice. Male Balb/c nu/nu mice, 6-8 weeks of age, were injected subcutaneously in both flanks with 5610 5 PC-3 cells suspended in 0.5 ml of Matrigel. Tumors were measured daily using calipers and volume was calculated using the formula volume = length6width6height60.52. Monitoring of tumor growth was started one week after tumor inoculation. Mice were separated into control and treatment groups. The control group had 5 mice carrying 10 tumors and the treatment group had 5 mice carrying 8 tumors. Three weeks after the initial inoculation, the treatment group was orally administrated triol daily at a dose of 20 mg/kg, 5 days/week. Mice in the control group were gavaged with vehicle only. Treatment started on day 22 and ended on day 42 after tumor inoculation. Tumor volumes were shown as mean 6 standard error (A) while body weight of mice was shown as mean 6 Standard Deviation (B). doi:10.1371/journal.pone.0065734.g003  IACUC-101046-A). Mice were kept under controlled environmental conditions (22uC, 12 hrs alternate light-dark cycles, 50% humidity, food and water ad libitum). Anesthesia (Isoflurane 1%) was given to reduce pain in mice during inoculation of prostate cancer cells. Animal care was conducted in accordance with the standard ethical guidelines (National Institutes of Health's ''Guide for the care and use of Laboratory animals'' and Clinical laboratory animal medicine (K. Hrapkiewicz, L. Medina, and D.D. Holmes, 1998, Iowa State University Press). The experiment was designed to minimize the number of nude mice used.

Tumor Xenografts in Athymic Mice
Experiments involving mice were approved by National Health Research Institutes Institutional Animal Care and Use Committee (NHRI-IACUC-101046-A). Male Balb/c nu/nu mice (NCI Frederick, MD), 6-8 weeks of age, were injected subcutaneously in both flanks with 5610 5 PC-3 cells suspended in 0.5 ml of Matrigel (BD Bioscience, Franklin Lakes, NJ, USA). Monitoring of tumor growth was started one week after tumor inoculation. Mice were separated into control and treatment groups. 5 mice carrying 10 tumors comprised the control group and 5 mice carrying 8 tumors comprised the treatment group. Three weeks after inoculation of PC-3 cells, mice in the treatment group were treated orally 5 days/week with 20 mg/kg triol. Triol was administered in a vehicle containing 1% Tween 20, 25% dimethylsulfoxide in PBS. Mice in the control group were gavaged with vehicle only. Triol treatment was started 21 days after inoculation of cells and continued for 21 days. Tumors were measured weekly using calipers and volume was calculated using the formula volume = length6width6height60.52 [7,[21][22][23].  [18] and MWAs were performed to measure the changes in abundance and modification of indicated proteins. The six samples printed in each well from left to right were CDXR-3 (control), CDXR-3 (20 mM triol), DU-145 (control), DU-145 (20 mM triol), PC-3 (control), PC-3 (20 mM triol) ( Figure S2). The right half blot (well A7-H12) was the technical duplicate of the left half blot (well A1-H6). Antibody list is shown in Table S2. Red and green were artificial color for anti-rabbit and anti-mouse 2 nd antibody, respectively. The blot is the same size as a standard 96-well microtiter plate. Relative protein abundance of each signaling protein in each cell line is listed in Table S3. doi:10.1371/journal.pone.0065734.g006 Luciferase-reporter Assay PC-3 cells were seeded at 2610 5 cells/well in a 12-well plate in DMEM containing 10% FBS. 24 hrs after plating, PC-3 cells were transfected with pRL-TK-Renilla luciferase plasmid (normalization vector; 5 ng/well), pSG5RXRa and pSG5LXRa (400 ng/ well ), 4xDR4D56cfos pGL3 (reporter gene vector; 500 ng/well) using the PolyJet TM in vitro DNA transfection reagent (SigmaGen Laboratories, Rockville, MD, USA). 24 hrs after transfection, cells were treated with increasing concentrations of triol or T0901317. After an additional 24 hrs, cells were lysed in 100 ml passive lysis buffer (Promega, Madison, WI, USA) and luciferase activity was measured using a Dual-Luciferase kit (Promega) in a 20/20 n luminometer Turner Biosystems.

Skp2 Overexpression in PC-3 Cells
Ectopic expression of Skp2 was achieved by infecting PC-3 cells with a retrovirus generated from the LPCX plasmid containing wild-type human Skp2 cDNA as described [10]. Puromycinresistant colonies were expanded and screened for increased Skp2  protein expression by Western blot analysis. PC-3 cells infected with an empty LPCX retrovirus were used as controls [10]. Cells were lysed in Laemmli buffer without bromophenol blue dye.

Real-Time Quantitative Polymerase Chain Reaction
RNA was extracted from the PC-3 and DU-145 cells treated with 0, 10, and 20 mM triol for 48 hrs using the RNeasy Minikit (cat. No. 74104, Qiagen, Venlo, The Netherlands) following the manufacturer's instructions. The genomic DNA was removed by DNase-on-column treatment supplied with the mini kit. RNA concentration was determined spectrophotometrically at 260 nm. Equal amounts of RNA were used in the cDNA synthesis reactions using the Reverse Transciption SystemRevertAid TM H Minus First Strand cDNA Synthesis Kit (Fermentas/Thermo Scientific, Waltham, MA, USA). Real-time quantitative PCR was performed as previously described [7,8,[21][22][23]28] using SYBR Green system/ 'reagents in an optical 96-well plate and cycling conditions consisting of 2 min at 50uC, 10 min st 95uC, 40 cycles of 15 sec at 95uC, and 60 sec at 60uC on an ABI Prism system (Applied Biosystems, Foster City, CA, USA). The sequences of primers for ATP-binding cassette transporter A1 (ABCA1) were TGTCCAGTCCAGTAATGGTTC (forward) and AAGCGA-GATATGGTCCGGATT (reverse). The transcript level of ABCA1 was determined in PC-3 and DU-145 cells following treatment with 0, 10, and 20 mM triol for 48 hrs and was normalized to GAPDH levels in each sample.

Transwell Migration Assay
Migration assays with PC-3 cells were performed with a transwell kit from BD Bioscience (catalog number 353097). PC-3, DU-145, and CDXR-3 cells were treated with 0, 10, and 20 mM triol for 48 hrs. Cells were then removed from tissue culture plates with trypsin, and washed twice with PBS. Trioltreated cells (1610 4 ) in 250 ul of DMEM without serum were placed in the upper invasion chamber and the lower compartment was loaded with DMEM containing 10% FBS. The cell migration chamber was inserted into the lower compartment and incubated for either 6 (PC-3, DU-145) or 24 (CDXR-3) hrs at 37uC. Cells on the topside of the filter were removed with a cotton swab. Cells attached to the filter were then fixed with methanol for 10 min. Cells attached to the filter were then stained with Giemsa stain (5%) for 1 hour. Filters were de-stained by washing with water and the number of cells attached to the filter was then quantified by enumerating cells in photographs of the stained filters.

Transwell Invasion Assay
An invasion assay with PC-3 cells was performed with Growth Factor Reduced BD BioCoat Matrigel invasion chambers (BD Biosciences) according to the manufacturer's instructions. PC-3 cells were treated with different concentrations of triol for 48 hrs. Cells were then trypsinized and washed twice with PBS. Cells from each condition were seeded at 4610 4 cells per well in serum-free DMEM in the upper compartment, and DMEM medium with 10% FBS was placed in the lower compartment of the chamber as a chemo-attractant. After 24 hrs of incubation, the non-invading cells on the upper side of the chamber were removed, and the membranes were fixed with methanol, washed, and stained with Giemsa's solution. Invasiveness was evaluated by counting the invading cells under a light microscope. All experiments were conducted in triplicate.

Wound Healing Assay
CDXR-3 cells were pre-treated with 0, 10, and 20 mM triol for 24 hrs. Cells were then removed from tissue culture plates with trypsin, and washed twice with PBS. Cells were seeded at a concentration of 3.5610 4 cells/well in 12-well plates. After 24 hrs, a wound healing assay was performed by scraping the cells with a 200-ml pipette tip. Cells were observed at 0, 8, and 24 hrs after scraping and photographed with a live imaging microscope (Leica AF 6000 LX, Leica, Wetzlar, Germany). The migration distance was automatically measured by the program inside live imaging microscope.

Confocal Microscopy
CDXR-3 and DU-145 cells were treated with 0 or 10 mM triol for 24 hrs. Cells were then washed 3 times with PBS buffer for 5 min per wash at room temperature (RT), fixed with 4% paraformaldehyde for 15 min at RT, and rinsed with PBS 3 times for 5 min per wash and permeabilized for 10 minutes at RT with 0.1% Triton X-100 in PBS. Cells were blocked for non-specific protein binding with 2% bovine serum albumin in PBS for 30 min, rinsed with PBS three times for 5 min each. Cells were incubated with b-actin or a-tubulin antibody for one hour. After washing, cells were incubated with secondary FITC-conjugated antibody for one hour. After washing, cells were mounted under glass slides and sealed with Permount. Images of cells were observed with an Olympus confocal microscope at 4006 (ocular lens 106; objective lens 406) (FV300, Olympus, Tokyo, Japan).

Data Analysis
Data are presented as the mean +/2 SD of at least three experiments or are representative of experiments repeated at least three times. Student's t test (two-tailed, unpaired) was used to evaluate the statistical significance of results from the proliferation assay experiments. A Microsoft Excel add-in program ED50V10 was used for calculating half maximal effective concentration (EC 50 ).

Triol Suppressed the Proliferation of Human Prostate Cancer Cells
We first sought to determine the effect of triol treatment on the viability and proliferation of three commonly used human prostate cancer cell lines (Fig. 1). LNCaP CDXR-3 cells are androgen receptor (AR)-positive, relapsed, androgen-independent cells derived from parental AR-positive androgen-dependent LNCaP 104-S cells [27]. DU-145 and PC-3 both are AR-negative androgen-insensitive cells established from brain- [30] and bone- [31] derived metastasis, respectively. An MTT assay and a Hoechst dye-based proliferation assays indicated that triol suppressed both cell viability and cellular proliferation (Fig. 1). The inhibitory effect on proliferation was dose-dependent and increased over time (Fig. 1, Table S1). The EC 50 for triol in the MTT assay was similar to the EC 50 measured by Hoechst dyebased proliferation assay (Table S1), suggesting that inhibition of cell proliferation was responsible for the reduction of viable cells caused by triol treatment in all three human advanced prostate cancer cell lines.

Triol Treatment Inhibited Colony Formation of PC-3 and CDXR-3 Cells in Soft Agar
Treatment of PC-3 and LNCaP CDXR-3 cells with 25 mM and 50 mM triol markedly inhibited the formation of PC-3 and LNCaP CDXR-3 colonies in soft agar, confirming the anti-cancer activity of triol. DU-145 cells grew too slowly in soft agar to detect sufficient colonies for further analysis (Fig. 2).

Triol Treatment Retarded Growth of Androgeninsensitive Prostate Cancer Xenografts in Nude Mice
To determine if triol could suppress tumor growth in vivo, we performed a pilot study using DU-145 xenografts in nude mice. Intraperitoneal (i.p.) injection of 1 mg (50 mg/kg) of triol daily for 14 days caused a 36% reduction in the average volume of DU-145 xenografts growing in nude mice (Fig. S1). Triol treatment for 14 days did not affect the body weight of the mice (data not shown).
To determine if lower oral doses of triol could inhibit the growth of PC-3 prostate xenografts. we orally administered triol at 20 mg/kg five times a week. Oral administration of 20 mg/kg triol (5 times per week) from the 3 rd to 6 th week caused a 65% reduction in average tumor volume (p = 0.0002) in the treatment group compared to tumor volume in the vehicle control group (Fig. 3A). Body weight of both vehicle and triol-treated groups gradually decreased (Fig. 3B). This may be due to cachetic effects of PC-3 xenografts. These observations confirmed that oral administration of triol could retard the growth of prostate tumors in vivo.

Triol Treatment Caused G1 Cell Cycle Arrest in Prostate Cancer Cells
We next performed flow cytometric analysis to determine if cell cycle progression of prostate cancer cells was affected by triol. Treatment with 10 mM triol for 48 hrs caused an increase in the G1 phase cell population and a decrease in S and G2/M phase populations of LNCaP CDXR-3 and DU-145 prostate cancer cells ( Fig. 4A, 4B). Triol at 10 mM did not affect cell cycle progression of PC-3 cells. However, 20 mM triol caused a significant increase in G1 and decrease in the S and G2/M populations of PC-3 cells (Fig. 4C). Therefore, treatment with 10-20 mM triol induced G1 cell cycle arrest in LNCaP, DU-145, and PC-3 cells. We did not observe any increase in sub-G1 population in these three prostate cancer cell lines. To determine if higher concentrations of triol will augment the population of cells in sub-G1, we treated DU-145 cells with 0, 20, and 40 mM triol (Fig. 4D). Reduction of S phase population and induction of G1 phase population of DU-145 cells was even more significant after treatment with 40 mM triol. However, there was no significant increase of cells in the sub-G1 population.

Treatment with High Concentration of Triol Induced Apoptosis in Prostate Cancer Cells
As propidium iodide staining flow cytometric analysis is not a reliable method to detect apoptosis, we used the TUNEL assay to determine if triol treatment at higher concentrations induced apoptosis in prostate cancer cells. We treated CDXR-3, DU-145, and PC-3 cells with 0, 20, and 40 mM triol for 48 or 96 hrs. Consistent with the flow cytometry analysis data, treatment with triol for 48 hrs only produced a small population of apoptotic cells in all these cell lines (data not shown). Treatment with triol for 96 hrs dose-dependently resulted in an increased the population of apoptotic cells in all three prostate cancer cell lines (Fig. 5). Although treatment with 20 mM triol for 96 hrs only slightly increased the apoptotic population, treatment with 40 mM triol resulted in a significant increase in apoptosis in all three prostate cancer cell lines.

Micro-Western Array Revealed Signaling Proteins being Affected by Triol Treatment
We hypothesized that triol may reduce cell viability through its effect on protein signaling networks. In order to determine what signaling proteins might be affected by triol treatment, we used Micro-Western Arrays (MWAs), a high-throughput Western blotting assay [17][18][19][20]24], to screen for signaling proteins affected by treatment with 10 mM triol in CDXR-3 and DU-145 cells. PTEN is frequently deleted in prostate cancer, resulting in activation of PI3K/Akt signaling [32]. PI3K/Akt signaling plays an important role in survival and progression of prostate cancer cells [32,33]. Up-regulation of PI3K/Akt activity is associated with poor clinical outcome of prostate cancer [34]. We used 48 antibodies that are capable of detecting proteins regulating cell cycle progression, cell survival and apoptosis, Akt-related signaling pathways, and several epithelial-mesenchymal transition (EMT) markers ( Figure S2, Table S2) for the screening portion of our MWA study (Fig. 6). Differences in the protein expression profile in the absence and presence of 20 mM triol is shown in Fig. 7.  (Table S3).

Triol Decreased the Level or Phosphorylation Status of Signaling Proteins Involved in Cell Cycle and Akt Signaling
The MWA result suggested that triol treatment altered the abundance and phosphorylation status of signaling proteins involved in cell cycle and Akt signaling in prostate cancer cells. We performed conventional Western blotting analysis to confirm these results. Triol treatment caused a decrease of Akt1 (DU-145, CDXR-3), phospho-Akt Thr308 (DU-145), phospho-Akt Ser473

Overexpression of Skp2 Partially Rescued the Inhibitory Effect of Triol Treatment in PC-3 Cells
Since triol treatment caused a significant decrease in Skp2 expression and an increase in p27 Kip expression in all three prostate cancer cell lines (Fig. 8), we next wished to determine if overexpression of Skp2 would block the growth inhibition caused by triol in PC-3 cells. Indeed, overexpression of Skp2 partially rescued the cell proliferation rate of PC-3 cells under triol treatment (Fig. 9). Because overexpression of Skp2 did not  completely reverse the growth inhibitory effects, we hypothesized that other additional pathways or mechanisms may be influenced by triol to cause cell growth inhibition of prostate cancer cells.

Triol Treatment Suppressed Akt1 in PC-3 Xenograft
We examined the protein expression of Akt1, c-Myc, FASN, Ki-67, and a-tubulin in 5 tumors xenografts from the vehicle-treated group and 8 tumor xenografts from the triol-treated group from Fig. 3. Triol treatment resulted in significantly reduced Akt1 expression in PC-3 xenografts (p,0.001, Fig. 10). Tumors in the treatment group tended to express lower c-Myc and FASN levels, although the p value was larger than 0.05. We also determined the protein levels of E-cadherin, N-cadherin, MMP-9, vimentin, slug, snail, p27 Kip , and phospho-Akt Ser473 in these xenografts. The variation between individual tumors within both control and treatment groups of these proteins was large and no significant difference in specific protein abundances between control and treatment groups (data not shown).

Triol is as a Weak Liver X Receptor (LXR) Agonist
LXRs are ligand-activated transcriptional factors that belong to the nuclear receptor super family. LXRs are important regulators of cholesterol, fatty acid, and glucose homeostasis [9]. Two LXR isoforms exist: LXRa expression is most abundant in liver, kidney, intestine, fat tissue, macrophages, lung, and spleen; LXRb is expressed ubiquitously [9]. A specific group of oxysterols, including 22(R)-hydroxycholesterol, 20(S)-hydroxycholesterol, 24(S)-hydroxycholesterol, and cholestenoic acid are natural ligands for LXRs [9,10]. We previously reported that LXR agonists, such as T0901317, 22(R)-hydroxycholesterol, and 24(S)-hydroxycholesterol suppress the proliferation of prostate and breast cancer cells via induction of G1 cell cycle arrest [6][7][8][9][10]. Since triol is an oxysterol, it might produce some effects as an agonist for LXR. To determine if triol is an LXR agonist, we determined if triol could activate LXR and induce expression of a luciferase reporter gene containing a LXR response element. Triol only slightly increased the transcriptional activity of LXRa (Fig. 11A). However, compared to the well known LXR agonist T0901317, triol was a very weak LXRa agonist (Fig. 11A). Treatment of DU-145 cells with 20 mM triol resulted in a slight increase in expression of ABCA1 mRNA expression, a target gene of LXRa activity (Fig. 11B), confirming the agonistic activity of triol for LXRa. The agonistic activity of triol for LXRa may partially contribute to the growth inhibition and G1 cell cycle arrest of prostate cancer cells.

Triol Treatment Altered Actin and Tubulin Distribution in Prostate Cancer Cells
Triol was previously reported to cause disruption of actin microfilaments [35], redistribution of vimentin [36], and loss of vinculin [35] in 73/73 endothelial cells. We next sought to determine if the cytoskeleton was affected by triol treatment of prostate cancer cells. Confocal laser microscopy revealed a modification of the b-actin and a-tubulin microfilaments, especially redistribution at the periphery of the cells following 48 hrs of treatment with 10 mM triol in CDXR-3 and DU-145 cells (Fig. 12). We did not observe significant changes in b-actin and a-tubulin microfilaments in PC-3 cells treated with 10 mM triol (data not shown).

Triol Treatment Suppressed Migration and Invasion of Prostate Cancer Cells
Since triol resulted in changes in the structure and morphology of the cytoskeleton, we suspected that triol treatment might affect the migration and invasion of prostate cancer cells. Pre-treatment of prostate cancer cells with 10 mM or 20 mM triol for 48 hrs significantly inhibited the migration and invasiveness of DU-145 and PC-3 cells (Fig. 13A, 13B). LNCaP cells have poor migration and invasion ability. Extending the time for transwell migration assay revealed that migration of LNCaP CDXR-3 was suppressed dose-dependently by triol treatment (Fig. 13C). A wound healing assay confirmed that treatment with 20 mM triol significantly reduced the migration ability of LNCaP CDXR-3 cells (Fig. 14).

Triol Treatment Affected Epithelial-mesenchymal Transition Proteins in Prostate Cancer Cells
Epithelial-mesenchymal transition (EMT) marker proteins regulate the ability of cancer cells to migrate and invade. We determined if triol treatment altered the expression level of these EMT marker proteins. Triol treatment (20 mM) (Fig. 15).

Non-cancerous Human Prostate Epithelial Cells were More Resistant to Triol Treatment
We sought to determine the effect of triol on the proliferation of non-cancerous human cells. In place of normal prostate epithelial cells, we used PZ-HPV-7 cells. PZ-HPV-7 cells were derived from normal human prostate epithelial cells transformed by transfection with HPV18 DNA. Compared to the three prostate cancer cell lines in this study, PZ-HPV-7 cells were more resistant to triol treatment. After 48 hrs treatment with triol, the EC 5 0 for triol suppression of proliferation of PZ-HPV-7 cells was 110.5 mM (Fig. 16). This suggested the possibility that non-cancerous human prostate cells may be more resistant to triol treatment compared to prostate cancer cells. However, the higher resistance of PZ-HPV-7 cells to triol treatment might be due to the lower proliferation rate of these non-cancerous cells compared to that of cancer cells. Further studies are needed to determine if triol will adversely affect normal prostate cells.
We observed that triol treatment inhibited tumor growth, reduced cell proliferation, induced G1 cell cycle arrest, and altered cytoskeletal morphology of human prostate cancer cells. Oral administration of triol at 20 mg/kg for three weeks reduced the average tumor volume of AR-negative androgen-insensitive PC-3 prostate cancer cells by 65% (Fig. 3). PC-3, one of the most commonly used prostate cancer cell lines, was established from bone-derived metastases [31]. Prostate cancer has a predilection to metastasize to bone and more than 80% of patients who died from prostate cancer developed bone metastases [38][39][40]. Our animal study suggested that triol treatment might be a potential treatment for prostate cancer metastasis. Indeed, we demonstrated that triol suppressed migration and invasion of DU-145, PC-3, and LNCaP CDXR-3 cells. The inhibition of migration and invasion of prostate cancer cells was not due to reduction of cell proliferation because the triol was given as a pre-treatment only and the same number of prostate cancer cells was seeded in both control and pre-treatment groups for the transwell assay (Fig. 9). Cancer cells in certain circumstance recapitulate normal developmental processes to enhance cell migration and invasion, a procedure called epithelial-to-mesenchymal transition (EMT). During the EMT, polarity is lost, motility is increased, mesenchymal markers are up-regulated, and differentiation markers are down-regulated [41,42]. This process was linked to cancer progression, in which epithelial cells acquire increased capacity for migration, stromal invasion, and metastasis [41,42]. During EMT of prostate cancer cells, the expression of transcription factors such as Snail, Slug, Twist, E47, ZEB1, and ZEB2 is increased, which leads to increased expression of N-cadherin and vimentin and concomitantly decreased expression of E-cadherin and cytokeratins [42][43][44][45]. Our study suggested that triol treatment resulted in decreased abundance of the proteins Slug, vimentin, and N-cadherin proteins while resulting in increased expression of the protein Ecadherin protein (Fig. 10). Focal adhesion kinase (FAK) is a nonreceptor tyrosine kinase localized in focal contacts that becomes tyrosine phosphorylated. FAK signaling pathway is involved in the regulation of prostate cancer cell migration [46][47][48][49]. We observed that triol treatment resulted in reduced abundance of FAK in PC-3 cells as well as phospho-FAK Ser722 in DU-145 cells and phospho-FAK Tyr861 in PC-3 cells. This new finding indicated that triol treatment targeted both EMT markers and FAK signaling proteins and can thus be a potential therapeutic agent for treatment of prostate cancer metastasis. We observed that noncancerous human prostate epithelial PZ-HPV-7 cells were more resistant to triol treatment compared to LNCaP, DU-145, and PC-3 prostate cancer cells (Fig. 11). This result suggested that triol has selectivity against prostate cancer cells versus normal cells.
Micro-Western Arrays (MWAs) represent a new method for proteomic analyses [19]. The MWA is a modification of the antibody-based modified reverse phase protein array. A noncontact microarray (such as the GeSim Nanoplotter), a flatbed electrophoresis apparatus (such as the GE multiphor), and a near infrared scanner (such as the Licor Odyssey) comprise the MWA analysis system [18]. The MWA system allows detection of the abundance or modification status of up to 384 different proteins with as many as 15 samples simultaneously. The method has been shown to be effective in characterizing the EGFR pathway under multiple growth factor concentrations and in network modeling of complex receptor tyrosine kinase (RTK) cross-talk after stimulation [18]. The MWA has a technical advantage over other current technologies for assaying protein modifications such as Reverse Phase Lysate Arrays, quantitative Mass Spectroscopy, and Luminex through its ability to efficiently and accurately measure hundreds of chosen phosphosites simultaneously, expanding the number of pathways surveyed in a single experiment. We applied this technology to screen signaling proteins being affected by triol in prostate cancer cells. In our MWA screening using 48 antibodies against proteins involved in Akt signaling and cell cycle regulation, we observed that triol caused reduction of Akt1, Akt2, cyclin E2, cyclin B1, phospho-c-Myc Thr58/Ser62, c-Myc, and phospho-Akt Ser473 in all three prostate cancer cell lines. Treatment with 20 mM triol caused at least a 20% increase in protein abundance of the following proteins: phospho-p38 MAPK Thr180/Tyr182, CDK4, cyclin E1, cyclin D1, and p53; while triol decreased in protein expression of cyclin D1, phospho-c-Myc Thr58/Ser62, phospho-Rb Ser780, c-Myc, phospho-Akt Thr308, phospho-Akt Ser473, cyclin A, cyclin B1, phospho-GSK3a Ser21, NF-kB p65, and FASN (Table S3). These findings were confirmed by conventional Western blotting. Western blotting experiments also showed that the protein levels of Akt1, phospho-Akt Ser473, phospho-Akt Thr308, PDK1, c-Myc, Skp2 in prostate cancer cells was decreased by triol treatment (Fig. 8). PI3K/Akt signaling plays an important role in survival and progression of prostate cancer cells [32,33,50]. Akt is a serine/threonine protein kinase encoded in mammals by three unique genes: Akt1, Akt2, and Akt3 [51,52]. Akt1 is involved in cellular survival pathways and inhibits apoptotic processes. Akt contains at least two important phosphorylation sites: threonine 308 and serine 473, which regulate its activity. Phosphorylation of threonine 308 on Akt is activated by PDK1 [53]. Phosphorylation of serine 473 is activated by mTOR kinase, its associated protein rictor, and SIN1/MIP1 [54,55]. Upregulation of the PI3K/Akt activity is associated with poor clinical outcome of prostate cancer [33,34,50]. c-Myc protein, a wellknown proto-oncoprotein, is an important transcriptional regulator of the androgenic response and cell proliferation in prostate and other cancers [25,56]. Expression of both c-Myc mRNA and protein increase during the progression of prostate cancer [57,58]. Decrease in c-Myc protein abundance results in decreased prostate cancer cell proliferation and tumor growth [23,[25][26][27]59]. Forced expression of FASN promoted proliferation while siRNA knockdown of FASN induced apoptosis in prostate cancer cells [60]. Because triol treatment suppressed Akt1, phospho-Akt Ser473, phospho-Akt Thr308, PDK1, c-Myc, and FASN, reduction of Akt signaling, c-Myc, and FASN may all contribute to growth inhibition caused by triol. The only difference in results that we observed between MWAs and conventional Western blots was in the detection of phospho-p44/42 MAPK Thr202/Tyr204, possibly due to the low abundance of the Erk1/2 proteins in these prostate cancer cells and the relatively weak signal of the reporter antibody for this protein. Western blotting also indicated the induction of p27 Kip1 under triol treatment. The p27 Kip1 is a cell cycle inhibitory proteins and accumulation of p27 Kip1 can result in cell cycle arrest and reduced cell proliferation of prostate cancer cells [23,26,61]. Skp2 is a member of the F-box protein family that is responsible for ubiquitination and down-regulation of p27 Kip1 and other proteins [62,63]. Our observation that triol treatment resulted in decreased Skp2, increased p27 Kip1 , and G1 cell cycle arrest was consistent with the known functions of Skp2 and p27 Kip1 (Fig. 4, Fig.7, Fig. 8). Overexpression of Skp2 partially rescued the growth inhibition induced by triol (Fig. 9), indicating that other pathways or mechanisms, such as LXR signaling (Fig. 11) or cytoskeletal signaling (Fig. 12), were probably involved in the growth inhibition caused by triol in prostate cancer cells.
Our study suggested that triol treatment causes both G1 cell cycle arrest and apoptosis (Fig. 4, Fig. 5) in CDXR-3, DU-145, and PC-3 in vitro. Administration via gavage of 20 mg/kg/day triol five times a week retarded the growth of PC-3 xenografts (Fig. 3). The actual concentration of triol in serum will require further examination. However, the fact that Ki-67 was not significantly reduced in PC-3 tumors suggested the possibility that bioavailability of triol might be poor via oral administration. Local injection of triol might be a more efficient way to administer triol in future studies.
Triol was found to exert oxidative stress in rat hepatocytes as well as to increase levels of superoxide dismutase, catalase, and glutathione peroxidase activity [64]. Components of the cytoskeletal network, such as the actin microfilament meshwork or intermediate filaments, are known targets for oxidative and thioldepleting agents that induce cell injury. Treatments with 20 mM triol, 7-ketocholesterol, or 25-hydroxycholesterol for 6 hrs caused progressive disruption of actin microfilaments, loss of vinculin, redistribution of vimentin, induction of cell detachment, and stimulation of apoptosis in 73/73 endothelial cells [35,36]. In our study, confocal laser microscopy revealed a modification of the bactin and a-tubulin microfilaments, especially redistribution at the periphery of the cell following 48 hrs of treatment with 10 mM triol in CDXR-3 and DU-145 cells (Fig. 8). Alterations of actin and tubulin distribution have been reported to cause cell cycle arrest in human cancer cells [65,66]. Abnormal distribution of cytoskeletal components caused by triol may contribute to the induction of cell cycle arrest in human prostate cancer cells.
Different signaling proteins in different prostate cancer cells responded differently to triol treatment ( Fig. 6-8, 15). Several proteins in DU-145 responded oppositely to 20 mM triol treatment compared to PC-3 and CDXR-3 cells (Fig. 7). This observation might be due to the different characteristics of PC-3, DU-145, and CDXR-3 cells. Although triol treatment suppressed cell proliferation, migration, invasion, and tumor growth in all three of these different prostate cancer cells, it may target different signaling proteins involved in the same pathways (such as Akt signaling) or similar functions (such as regulation of cell cycle, EMT, and cell structure) to produce similar results in different cell lines.

Conclusions
Our study provided systems-level insight into the molecular mechanism of triol action that resulted in suppression of proliferation, migration, and invasion of prostate cancer cells. Our results suggested that triol might be a potential therapeutic agent for treatment of advanced metastatic prostate cancer. In addition, our study highlights the power of the MWA approach for rapidly examining the systems level impact of small molecules on protein signaling networks and relating these changes to cell phenotypes. Figure S1 Effect of triol on the growth of DU-145 xenografts in nude mice. Male Balb/c nu/nu mice, 6-8 weeks of age, were injected subcutaneously in both flanks with 1610 6 DU-145 cells suspended in 0.5 ml of Matrigel (BD Bioscience, Franklin Lakes, NJ, USA). Tumors were measured daily using calipers and volume was calculated using the formula volume = length6width6height60.52. Tumor volume was allowed to increase to larger than 100 mm 3 before treatment started. Both control and treatment groups had 3 mice carrying 5 tumors. Mice were given intraperitoneal (i.p.) injection of vehicle (5% DMSO and 1% Tween 20 in water) or 1 mg triol in vehicle daily for 14 days. Tumor volumes were shown as mean  S1 EC 50 of triol to suppress viability and proliferation of prostate cancer cells. LNCaP CDXR-3, DU-145, and PC-3 cells treated with triol for 48 hrs or 96 hrs were assayed with MTT assay or Hoechst dye-based proliferation assay to determine the EC 50 of triol to cause growth inhibition. (TIF) Table S2 List of antibodies used in the MWA study. Information of antibodies and location of antibodies on the antibody plate for MWA blot incubation is shown. The a-tubulin, b-actin, and GAPDH proteins were detected as loading controls. (TIF) Table S3 Relative expression level of signaling proteins in prostate cancer cells assayed with MWA. Relative abundance of signaling proteins in CDXR-3, DU-145, and PC-3 cells treated with 0 or 20 mM triol for 48 hrs was determined by MWA and shown as list. N/A represents signaling being too weak to be detected. Protein abundance was normalized to the average of a-tubulin, GAPDH, and b-actin. Proteins in different cell lines were normalized to the control condition (no triol treatment) of that cell line. (TIF)