There is growing evidence showing that prostate cancer cells have perturbed cholesterol homoeostasis, accumulating cholesterol to promote cell growth. Consequently, cholesterol-lowering drugs such as statins are being evaluated in prostate cancer treatment. Furthermore, natural products such as betulin (from birch tree bark) and tocotrienol (a minor form of vitamin E) have been shown to lower cholesterol levels. Using these drugs and oxysterols, we have determined which aspects of cholesterol homoeostasis should be targeted in prostate cancer, e.g. cellular cholesterol levels are increased by the transcription factor SREBP-2 (sterol-regulatory-element-binding protein isoform 2), whereas LXR (liver X receptor) promotes cholesterol efflux. Whereas betulin exerted non-specific effects on cell viability, tocotrienols produced a strong direct correlation between SREBP-2 activity and cell viability. Mechanistically, tocotrienols lowered SREBP-2 activity by degrading mature SREBP-2 independently of the proteasome. In contrast, no correlation was seen between LXR activity and cell viability, implying that SREBP-2 is a better target than LXR for prostate cancer treatment. Lastly, androgen-dependent and -independent LNCaP cells were both sensitive to tocotrienols. Overall, this suggests that tocotrienols and other drugs targeting the SREBP-2 pathway are a potential therapeutic option for prostate cancer.

INTRODUCTION

In developed countries, prostate cancer is the most common cancer in males and contributes significantly to cancer-related mortality [1]. Epidemiological evidence has associated prostate cancer risk with high cholesterol levels [2,3], and laboratory studies have observed that the aging prostate and prostate cancer cells have elevated intracellular cholesterol levels [4]. Such cholesterol accumulation could promote prostate cancer development by providing raw material for membrane synthesis, androgen production and other signalling pathways [5].

Consequently, there has been considerable interest in cholesterol-lowering drugs in prostate cancer treatment [2,3]. One example is the statins, a class of drugs that target HMGCR (3-hydroxy-3-methylglutaryl-CoA reductase), which catalyses a rate-limiting step in cholesterol synthesis. Statins have proven effective in animal models (see, e.g., [6]), and there is some evidence that statins reduce the risk of advanced prostate cancer (reviewed in [3]). However, statin treatment has limitations, including their side effects [7] and the potential inability to reach sufficient concentrations at the tumour site [3]. Thus there is a need to find better methods of manipulating cholesterol levels in prostate cancer.

Cellular cholesterol levels are largely regulated by two master transcription factors: SREBP-2 (sterol-regulatory-element-binding protein isoform 2) and LXR (liver X receptor). The regulation of SREBP-2 has been thoroughly elucidated by the laboratory of Brown and Goldstein [8]. Briefly, SREBP-2 initially exists in its precursor form, tethered to the endoplasmic reticulum. When cholesterol levels are low, SREBP-2 is escorted by SCAP (SREBP-2 cleavage activation protein) to the Golgi, where SREBP-2 is cleaved. This releases the active N-terminal portion of SREBP-2, which migrates to the nucleus to up-regulate genes involved in cholesterol synthesis (e.g. HMGCR) and uptake [e.g. LDLR (low-density lipoprotein receptor)]. This increases cholesterol levels, which promotes the binding of SCAP to Insig (insulin-induced gene), retaining SREBP-2 in the endoplasmic reticulum. Thus SREBP-2 increases cholesterol levels, which then feed back on its activity.

In contrast, LXR is activated by oxygenated cholesterol derivatives (oxysterols), and serves to lower cellular cholesterol levels by up-regulating genes involved in cholesterol efflux, such as ABC (ATP-binding cassette transporter) A1 and G1. Consequently, LXR agonists have been put forward as a therapeutic option for prostate cancer [9,10]. However, given that LXR also up-regulates fatty acid synthesis (e.g. via fatty acid synthase), this could cause side effects such as enhancing lipid levels in the liver (hepatic steatosis), or even in prostate cancer cells.

Recent studies have characterized novel lipid-lowering drugs isolated from natural products. For example, the active ingredient in birch tree bark, betulin, inhibits SREBP-2 activation by promoting SCAP–Insig binding [11]. In vivo, betulin improved the outcome of diet-induced metabolic diseases, such as atherosclerosis and obesity [11]. Another natural product of interest is vitamin E, which consists of two families of lipophilic compounds: tocotrienols and tocopherols. Tocotrienols are sourced from cereal grains and palm oil, whereas tocopherols are found in vegetable oil [12]. Both families share a chromanol ring and an isoprenoid-derived side chain, but the side chain is saturated in tocopherols and unsaturated in tocotrienols [12]. Well-known as antioxidants, they have also been investigated in cholesterol homoeostasis (see, e.g., [1315]). The tocotrienols, particularly the γ and δ isoforms, blocked SREBP-2 activation and degraded HMGCR by an Insig-dependent mechanism [15]. There have been intense efforts evaluating tocotrienols for cancer therapy (reviewed in [16]), but not in the context of manipulating prostate cancer cholesterol levels.

In the present study, we examined the potential of toggling cholesterol homoeostasis in prostate cancer, using betulin and tocotrienol. We correlate this with their mechanism of action, finding that SREBP-2 activity determines resistance to the tocotrienols. Furthermore, we propose that SREBP-2 is a better target than LXR for treating prostate cancer, supporting the potential for tocotrienols in prostate cancer therapy.

EXPERIMENTAL

Materials

FBS (fetal bovine serum) was obtained from Bovogen, NCS (newborn calf serum) and penicillin/streptomycin were from Life Technologies, and all other media components were from Sigma–Aldrich. NCLPDS (lipoprotein-deficient NCS) and FBLPDS (lipoprotein-deficient FBS) were prepared from NCS and FBS respectively, as described previously [17,18]. α-, γ- and δ-tocotrienol were obtained from Cayman Chemical. Betulin, compactin (mevastatin), casodex (bicalutamide) and 25-HC (25-hydroxycholesterol) were obtained from Sigma–Aldrich. Zeocin was obtained from Life Technologies. 24,25-EC [(24S,25)-epoxycholesterol] was obtained from Enzo Life Sciences. Testosterone was a gift from Dr David Handelsman (ANZAC Research Institute, Concord, NSW, Australia).

Cell culture

The prostate cancer cell lines LNCaP and PC-3 were a gift from Dr Pamela Russell (Australian Prostate Cancer Research Center, Buranda, QLD, Australia), and were maintained in medium A [RPMI 1640, supplemented with 10% (v/v) FBS, 100 units/ml penicillin and 100 μg/ml streptomycin]. Before plating LNCaP and PC-3 cells, plates and dishes were treated with polyethyleneimine (Sigma–Aldrich) to enhance cellular adhesion as described previously [19], with slight modifications: polyethyleneimine was prepared as a 25 μg/ml solution in 0.15 mM NaCl, with the pH neutralized using HCl. As specified in experiments, prostate cancer cells were treated in medium B [RPMI 1640, supplemented with 10% (v/v) FBLPDS, 100 units/ml penicillin and 100 μg/ml streptomycin], to remove the influence of exogenous cholesterol.

The CHO (Chinese-hamster ovary) cell lines CHO-7 and SRD-1 were a gift from Dr Michael Brown and Dr Joseph Goldstein (University of Texas Southwestern Medical Center, Dallas, TX, U.S.A.). SRD-15 cells were a gift from Dr Russell DeBose-Boyd (University of Texas Southwestern Medical Center). CHO-7 cells were maintained in medium C [DF-12, supplemented with 5% (v/v) NCLPDS, 100 units/ml penicillin and 100 μg/ml streptomycin], and SRD-1 and SRD-15 cells were maintained in medium C supplemented with 1 μg/ml 25-HC. As specified in experiments, these cells were treated in medium D (medium C, supplemented with 5 μM compactin and 50 μM mevalonate), to maximize SREBP-2 activity [20].

Plasmid constructs and transfection

The sources of plasmids are listed in Table 1. For simplicity, the plasmid names have been condensed. To generate the SCAP-Myc expression construct, PIPE (polymerase incomplete primer extension) cloning [21] was used to clone the SCAP gene into pcDNA4-MycHisC, retaining the Myc and His tags within the open reading frame. SCAP-Y234A-Myc was generated from SCAP-Myc using PCR-based SDM (site-directed mutagenesis) [22]. To generate the HA-SREBP-2(M) expression construct, the HA-SREBP-2(M) gene was cloned from pCMV5-HA-SREBP-2(M) into pcDNA4-MycHisC using PIPE cloning; in this construct, the stop codon is located upstream of the His and Myc tags, thus excluding these tags from the open reading frame. Point mutations were made in the DNA-binding domain (Y342R [23]) and CPD (Cdc4 phosphodegron) motif (S432A/S436A [24]) by PCR-based SDM [22]. The primers used in these cloning and mutagenesis protocols are available from A.J.B. upon request. Successful constructs were confirmed by sequencing and expression by Western blotting.

Table 1
Plasmid constructs used in the present study
Alias Original name Source 
SRE-luc 6×SRE-luc Generated by our laboratory group [28
LXRE-luc 3×LXRE-luc Generated by our laboratory group [26
TK-luc pGL3-TK Dr Malcolm Lyons, Western Australian Institute of Medical Research, Perth, WA, Australia 
Renilla-luc phRL-TK Promega 
pCMV5-SCAP pCMV5-SCAP Dr Ryuichiro Sato, University of Toyko, Tokyo, Japan 
SCAP-Myc pcDNA4-SCAP-MycHisC Generated in the present study 
SCAP-Y234A-Myc pcDNA4-SCAP-MycHisC-Y234A Generated in the present study 
pCMV5-HA-SREBP2(M) pCMV5-HA-SREBP2(N) Dr John Shyy, University of California Riverside, Riverside, CA, U.S.A. 
HA-SREBP-2(M) pcDNA4-HA-SREBP-2(M) Generated in the present study 
HA-SREBP-2(M)-DNABmut pcDNA4-HA-SREBP-2(M)-Y342R Generated in the present study 
HA-SREBP-2(M)-CPDmut pcDNA4-HA-SREBP-2(M)-S432A/S436A Generated in the present study 
Alias Original name Source 
SRE-luc 6×SRE-luc Generated by our laboratory group [28
LXRE-luc 3×LXRE-luc Generated by our laboratory group [26
TK-luc pGL3-TK Dr Malcolm Lyons, Western Australian Institute of Medical Research, Perth, WA, Australia 
Renilla-luc phRL-TK Promega 
pCMV5-SCAP pCMV5-SCAP Dr Ryuichiro Sato, University of Toyko, Tokyo, Japan 
SCAP-Myc pcDNA4-SCAP-MycHisC Generated in the present study 
SCAP-Y234A-Myc pcDNA4-SCAP-MycHisC-Y234A Generated in the present study 
pCMV5-HA-SREBP2(M) pCMV5-HA-SREBP2(N) Dr John Shyy, University of California Riverside, Riverside, CA, U.S.A. 
HA-SREBP-2(M) pcDNA4-HA-SREBP-2(M) Generated in the present study 
HA-SREBP-2(M)-DNABmut pcDNA4-HA-SREBP-2(M)-Y342R Generated in the present study 
HA-SREBP-2(M)-CPDmut pcDNA4-HA-SREBP-2(M)-S432A/S436A Generated in the present study 

For plasmid transfection, cells were plated and transfected in their maintenance medium, using TransIT-2020 reagent (MirusBio) according to the manufacturer's instructions. In addition, the medium was refreshed before adding the TransIT-2020–DNA complexes.

Generation of stable cell lines

To generate stable clones LNCaP-EV, LNCaP-SCAP-WT and LNCaP-SCAP-mut, LNCaP cells were each plated in a 100-mm-diameter dish and transfected with pcDNA4-MycHisC, SCAP-Myc or SCAP-Y234A-Myc constructs respectively. Transfected cells were selected in medium A using 250 μg/ml zeocin (Life Technologies), according to the manufacturer's instructions. Single clones were selected and expanded, and stable expression was assessed by Western blotting. Stable cells were maintained in medium A, supplemented with 250 μg/ml zeocin.

To generate androgen-independent LNCaP cells, LNCaP cells were initially grown for eight passages in medium A, supplemented with 10 nM testosterone, generating the LNCaP-305 cell line. These cells can tolerate physiological androgen concentrations (~10 nM) (see, e.g., [25]). To simulate androgen ablation, LNCaP-305 cells were grown for eight passages in medium A, supplemented with 10 μM casodex, generating the LNCaP-364 cell line. These cells grow independently of the medium androgen levels. These stable populations were maintained in their respective selection medium.

Luciferase assay

Cells were plated in 60-mm-diameter dishes for transfection. Cells were transfected with 6 μg of firefly luciferase plasmid and 0.12 μg of Renilla luciferase plasmid. Following transfection, cells were trypsinized and seeded into 24-well plates in medium B, and allowed to adhere overnight. As described previously [17], the treatment was delivered in a small quantity of plating medium, added to the existing medium in the wells. After treatment, the luciferase assay was conducted using the Dual-Luciferase Reporter Assay System (Promega), according to the manufacturer's instructions. The firefly luciferase values were divided by their corresponding Renilla luciferase values and made relative to the vehicle condition to obtain ‘Renilla/vehicle-normalized luciferase activity’.

Furthermore, since LXRE-luc and SRE-luc are driven by both their respective LXR or SREBP-2 response elements and the viral TK (thymidine kinase) promoter [26], additional cells were transfected with the TK-luc construct in the same experiment and treated in parallel. Renilla/vehicle-normalized luciferase activity values of LXRE-luc or SRE-luc were divided by those of TK-luc to determine LXRE-specific and SRE-specific promoter activity. For simplicity, this has been depicted as ‘relative luciferase activity’ in the Figures.

WST-1 (water-soluble tetrazolium salt 1) assay

Cells were seeded in 96-well plates in Phenol Red-free RPMI 1640 medium, supplemented with 0.1% fatty-acid-free BSA, 100 units/ml penicillin and 100 μg/ml streptomycin. Phenol Red was avoided since it was found to interfere with the WST-1 assay, and cells were seeded in serum-free conditions to provide an equal baseline when treatment began in different serum conditions. Owing to different growth rates between cell lines, PC-3 cells were seeded at 7500 cells/well, whereas LNCaP cells (and related sublines) were seeded at 10000 cells/well.

The next day, treatment was delivered in Phenol Red-free RPMI 1640 medium to the existing medium in the well, resulting in a final serum concentration of 10% (v/v) and drug concentrations as indicated in the Figure legends. The blank wells were cell-free, and treated similarly to the vehicle condition. As a positive control, antibiotic (750 μg/ml G418, obtained from Roche Diagnostics) was included, and caused significant cell death in each experiment (results not shown).

After 72 h, the WST-1 assay (Roche Diagnostics) was performed according to the manufacturer's instructions, with some modifications. Briefly, instead of 0.1 vol. of WST-1, 0.2 vol. of a 1:1 mixture of WST-1 reagent and Phenol Red-free RPMI 1640 was added to the wells; dispensing this larger volume reduced variation. Each plate was mixed and the A450 was measured every 1 h until the highest A450 (typically from the vehicle condition) began to plateau. The data obtained at the time point before the plateau was used for analysis; A450 values were adjusted to those of blank wells, and made relative to the vehicle condition to obtain ‘relative cell viability’.

Western blotting

Cells were plated, transfected and treated as described in the Figure legends. Cells were harvested for protein and lysates subjected to SDS/PAGE (7.5% gels), Western blotting and visualization as described previously [19], with a modification in the lysis buffer: 3% (w/v) SDS, 10 mM Tris/HCl (pH 7.6) and 100 mM NaCl, supplemented with 0.05 vol. of protease inhibitor cocktail (Sigma–Aldrich). The additional SDS was included to ensure the solubilization of membrane proteins such as SREBP-2 and SCAP. Primary antibodies included anti-Myc (mouse clone 9E10, from Santa Cruz Biotechnology), anti-SCAP (rabbit clone IgG-R139 [27], herein denoted ‘R139’, a gift from Dr Michael Brown and Dr Joseph Goldstein), anti-HA (haemagglutinin) (mouse clone 16B12, from Covance) and anti-α-tubulin (mouse clone B-5-1-2, from Sigma–Aldrich). Densitometry was performed using Adobe Photoshop (version 7.0).

qRT-PCR (quantitative real-time reverse transcription–PCR)

Cells were plated and treated as described in the Figure legends. RNA was harvested and reverse-transcribed, and mRNA levels were quantified by qRT-PCR as described previously [19]. All primers for human genes have been described previously [19], except that different ABCG1 primers were used, with the ABCG1 primer sequences provided by Dr Etienne Lefai (Faculté de Médecine Lyon Sud, Lyon, France). All primers for hamster genes have been described previously [26,28,29].

siRNA (small interfering RNA) knockdown

Cells were plated in medium A at 350000 cells/well in six-well plates. The next day, the medium was refreshed and cells were transfected with 25 nM siRNA, using Lipofectamine™ RNAi-MAX (Life Technologies) according to the manufacturer's instructions. As a negative control (siNC), scrambled siRNA (SIC001, Sigma–Aldrich) was used. siRNA oligonucleotides against SCAP (siSCAP) were designed by Sigma–Aldrich and their sequences (sense strand provided only) are as follows: siSCAP1, 5′-CACUCAAAGACUUGUUAUUTT-3′; siSCAP2, 5′-GCAUCAUCCUCAUCGGCUATT-3′; siSCAP3, 5′-CUCAAUGGCGGCGAGAUUUTT-3′; siSCAP4, 5′-GCAUUACCGCUCUGGUGUUTT-3′; siSCAP5, 5′-GGAAAUUGUCCUUCCGCCATT-3′; siSCAP6, 5′-GUGCUAUGCCCGCGCAACUTT-3′. Their efficiency was assessed by Western blotting (see Supplementary Figure S3 at http://www.BiochemJ.org/bj/446/bj4460191add.htm.

RESULTS

Cell viability correlates with changes in SREBP-2 activity

Previously, we have shown that PC-3 cells have higher basal SREBP-2 activity than LNCaP cells [17]. Thus we hypothesized that SREBP-2 activity would be less affected by betulin and tocotrienols in PC-3 cells. We tested this using a luciferase reporter assay: in this experiment, cells were co-treated with the HMGCR inhibitor compactin to eliminate HMGCR-specific effects (Figure 1A) and increase basal SREBP-2 activity. As expected, SREBP-2 activity in PC-3 cells was less affected than in LNCaP cells (Figure 1B), with γ- and δ-tocotrienol being more potent than the other drugs. In contrast, the drugs had little effect on LXR activity in either cell line, in contrast with the LXR agonist GW683965 (Figure 1C).

Betulin and tocotrienols down-regulate SREBP-2 activity, but do not affect LXR activity

Figure 1
Betulin and tocotrienols down-regulate SREBP-2 activity, but do not affect LXR activity

(A) Schematic diagram of the actions of betulin (Bet) and tocotrienols (T3) in the context of cholesterol homoeostasis. See the text for details. (B) Cells were transfected with SRE-luc, and treated for 24 h with betulin (Bet), α-tocotrienol (αT3), γ-tocotrienol (γT3) or δ-tocotrienol (δT3) at the concentrations indicated, in the presence of compactin (10 μM). Following treatment, cells were assayed for luciferase activity. (C) Cells were transfected with LXRE-luc, and treated for 24 h with vehicle (Veh) or betulin (Bet), α-tocotrienol (αT3), γ-tocotrienol (γT3) or δ-tocotrienol (δT3) at 10 μM or GW683965 (GW) at 1 μM. Following treatment, cells were assayed for luciferase activity. (B and C) Results are means+S.E.M. from at least three separate experiments, each performed with triplicate wells per condition.

Figure 1
Betulin and tocotrienols down-regulate SREBP-2 activity, but do not affect LXR activity

(A) Schematic diagram of the actions of betulin (Bet) and tocotrienols (T3) in the context of cholesterol homoeostasis. See the text for details. (B) Cells were transfected with SRE-luc, and treated for 24 h with betulin (Bet), α-tocotrienol (αT3), γ-tocotrienol (γT3) or δ-tocotrienol (δT3) at the concentrations indicated, in the presence of compactin (10 μM). Following treatment, cells were assayed for luciferase activity. (C) Cells were transfected with LXRE-luc, and treated for 24 h with vehicle (Veh) or betulin (Bet), α-tocotrienol (αT3), γ-tocotrienol (γT3) or δ-tocotrienol (δT3) at 10 μM or GW683965 (GW) at 1 μM. Following treatment, cells were assayed for luciferase activity. (B and C) Results are means+S.E.M. from at least three separate experiments, each performed with triplicate wells per condition.

Next, we considered the effect of these drugs on cell viability. This was examined in lipoprotein-deficient serum (FBLPDS) to prevent interference from exogenous sources of cholesterol (e.g. low-density lipoprotein). Whereas betulin had comparable effects between LNCaP and PC-3 cells, the tocotrienols revealed a trend similar to that seen with SREBP-2 activity: LNCaP cells were more affected than PC-3 cells, with γ- and δ-tocotrienol being more potent than α-tocotrienol (Figure 2A). In fact, there was a strong linear correlation between SREBP-2 activity and cell viability (Figure 2B). Furthermore, co-treatment with the HMGCR product mevalonate rescued cells from compactin, but not betulin or the tocotrienols (Figure 2C), suggesting that their mode of action does not require HMGCR inhibition (Figure 1A). In contrast, providing full serum rescued the cells from all drugs (Figure 2D). Together, these data suggest that cell viability is dependent on cholesterol supply and, at least for the tocotrienols, sensitive to changes in SREBP-2 activity.

Betulin and tocotrienols influence cell viability in a cholesterol-dependent manner

Figure 2
Betulin and tocotrienols influence cell viability in a cholesterol-dependent manner

(A) Cells were treated with compactin (CPN), betulin (Bet), α-tocotrienol (αT3), γ-tocotrienol (γT3) or δ-tocotrienol (δT3) at the concentrations indicated, in the presence of FBLPDS. After 3 days, cell viability was measured using the WST-1 assay. (B) The effects of α-tocotrienol (αT3), γ-tocotrienol (γT3) and δ-tocotrienol (δT3) on cell viability (from A) are plotted against their respective effects on SREBP-2 activity (from Figure 1B) for both cell lines (PC-3 and LNCaP). For the combined dataset, the correlation is y=0.85x (R2≈0.81, P<0.01, regression analysis). (C) Cells were treated and assayed as in (A), except using 10 μM drug (indicated) with or without mevalonate (5 mM). The change in viability with mevalonate treatment is presented as a percentage, relative to the vehicle (Veh) condition. (D) Cells were treated and assayed as in (A), except using 10 μM drug (indicated) with either cholesterol-deficient (FBLPDS) or full (FBS) serum. Viability is relative to the vehicle (Veh)/FBLPDS condition for each cell line. (AD) Results are means+S.E.M. from at least three separate experiments for each cell line, each performed with triplicate wells per condition.

Figure 2
Betulin and tocotrienols influence cell viability in a cholesterol-dependent manner

(A) Cells were treated with compactin (CPN), betulin (Bet), α-tocotrienol (αT3), γ-tocotrienol (γT3) or δ-tocotrienol (δT3) at the concentrations indicated, in the presence of FBLPDS. After 3 days, cell viability was measured using the WST-1 assay. (B) The effects of α-tocotrienol (αT3), γ-tocotrienol (γT3) and δ-tocotrienol (δT3) on cell viability (from A) are plotted against their respective effects on SREBP-2 activity (from Figure 1B) for both cell lines (PC-3 and LNCaP). For the combined dataset, the correlation is y=0.85x (R2≈0.81, P<0.01, regression analysis). (C) Cells were treated and assayed as in (A), except using 10 μM drug (indicated) with or without mevalonate (5 mM). The change in viability with mevalonate treatment is presented as a percentage, relative to the vehicle (Veh) condition. (D) Cells were treated and assayed as in (A), except using 10 μM drug (indicated) with either cholesterol-deficient (FBLPDS) or full (FBS) serum. Viability is relative to the vehicle (Veh)/FBLPDS condition for each cell line. (AD) Results are means+S.E.M. from at least three separate experiments for each cell line, each performed with triplicate wells per condition.

To explore this further, we utilized the oxysterols 24,25-EC and 25-HC. These oxysterols are potent regulators of cholesterol homoeostasis, lowering cholesterol levels by reducing SREBP-2 activity, activating LXR and degrading HMGCR [30]. Thus these oxysterols were used to determine the dependence of cell viability on SREBP-2 compared with LXR activity. In terms of SREBP-2 activity, 25-HC was more potent, and LNCaP cells were more affected (Figure 3A). The reverse trend could be seen for LXR activity, with 24,25-EC being more agonistic and PC-3 cells more responsive (Figure 3B). Furthermore, cell viability was more affected in LNCaP cells, with 25-HC being more potent (Figure 3C), correlating better with SREBP-2 rather than LXR activity (Figure 3D). In addition, 1 μM GW68395, which activated LXR more than 25-HC (Figures 1C and 3B), had little effect on LNCaP or PC-3 cell viability (results not shown). Overall, these data suggest that cell viability is more influenced by SREBP-2 than by LXR activity.

The influence of oxysterols on cell viability correlates better with SREBP-2 activity than with LXR activity

Figure 3
The influence of oxysterols on cell viability correlates better with SREBP-2 activity than with LXR activity

(A and B) Cells were transfected with SRE-luc (A) or LXRE-luc (B), and treated for 24 h with 24,25-EC or 25-HC at the concentrations indicated. In (A), cells were also co-treated with compactin (10 μM). Following treatment, cells were assayed for luciferase activity. (C) Cells were treated with 24,25-EC or 25-HC (3.3 μM) in the presence of FBS. After 3 days, cell viability was measured using the WST-1 assay, and made relative to the vehicle condition. (D) Summary of trends in (A)–(C). (AC) Results are means+S.E.M. from at least three separate experiments, each performed with triplicate wells per condition.

Figure 3
The influence of oxysterols on cell viability correlates better with SREBP-2 activity than with LXR activity

(A and B) Cells were transfected with SRE-luc (A) or LXRE-luc (B), and treated for 24 h with 24,25-EC or 25-HC at the concentrations indicated. In (A), cells were also co-treated with compactin (10 μM). Following treatment, cells were assayed for luciferase activity. (C) Cells were treated with 24,25-EC or 25-HC (3.3 μM) in the presence of FBS. After 3 days, cell viability was measured using the WST-1 assay, and made relative to the vehicle condition. (D) Summary of trends in (A)–(C). (AC) Results are means+S.E.M. from at least three separate experiments, each performed with triplicate wells per condition.

Androgen-independence does not completely determine tocotrienol sensitivity

We explored why PC-3 cells were more resistant to tocotrienols. Given that LNCaP cells are androgen-sensitive, whereas PC-3 cells are androgen-independent, could it be related to their androgen-independence? To test this, we generated LNCaP-364 cells, an LNCaP subline that grows independently of medium androgen status.

In response to betulin, LNCaP-364 cells resembled PC-3 cells (Figures 4A and 4B), particularly with SREBP-2 activity being more resistant than in LNCaP cells (Figure 4A, leftmost panel). For the tocotrienols, LNCaP-364 cells exhibited a similar SREBP-2 response to LNCaP cells (Figure 4A), with their SREBP-2 activities having a comparable correlation with cell viability (Figure 4C). Furthermore, at high tocotrienol concentrations, LNCaP-364 cells displayed an intermediate sensitivity between LNCaP and PC-3 cells (Figure 4B). Together, these data suggest that androgen-dependence only partially determines tocotrienol sensitivity.

Androgen-independence does not confer the tocotrienol resistance observed in PC-3 cells

Figure 4
Androgen-independence does not confer the tocotrienol resistance observed in PC-3 cells

(A) LNCaP-364 cells were transfected and treated as described in Figure 1(B). (B) LNCaP-364 cells were transfected and treated as described in Figure 2(A). (C) The effects of α-tocotrienol (αT3), γ-tocotrienol (γT3) and δ-tocotrienol (δT3) on viability and SREBP-2 activity were correlated as described in Figure 2(B). The correlation for the LNCaP-364 dataset is y=0.99x (R2≈0.89) and for the LNCaP dataset is y=1.21x−0.15 (R2≈0.92). There is no significant difference between the datasets (P≈0.16), but a significant correlation between viability and SREBP-2 activity (P<0.01), by two-way ANOVA assuming interaction between cell type and SREBP-2 activity. The LNCaP and PC-3 datasets included in this Figure were sourced from Figures 1 and 2; however, for each experiment with LNCaP-364 cells, representative conditions were repeated in parallel with PC-3 and LNCaP cells as controls. (A and B) Results are means+S.E.M. from at least three separate experiments, each performed with triplicate wells per condition.

Figure 4
Androgen-independence does not confer the tocotrienol resistance observed in PC-3 cells

(A) LNCaP-364 cells were transfected and treated as described in Figure 1(B). (B) LNCaP-364 cells were transfected and treated as described in Figure 2(A). (C) The effects of α-tocotrienol (αT3), γ-tocotrienol (γT3) and δ-tocotrienol (δT3) on viability and SREBP-2 activity were correlated as described in Figure 2(B). The correlation for the LNCaP-364 dataset is y=0.99x (R2≈0.89) and for the LNCaP dataset is y=1.21x−0.15 (R2≈0.92). There is no significant difference between the datasets (P≈0.16), but a significant correlation between viability and SREBP-2 activity (P<0.01), by two-way ANOVA assuming interaction between cell type and SREBP-2 activity. The LNCaP and PC-3 datasets included in this Figure were sourced from Figures 1 and 2; however, for each experiment with LNCaP-364 cells, representative conditions were repeated in parallel with PC-3 and LNCaP cells as controls. (A and B) Results are means+S.E.M. from at least three separate experiments, each performed with triplicate wells per condition.

Tocotrienols act by degrading mature SREBP-2

To provide insight into the determinants of tocotrienol resistance, we explored the mechanism by which tocotrienols inhibit SREBP-2 activity. We employed the CHO-7 cell line and its mutant derivatives, which are commonly used to study cholesterol homoeostasis (see, e.g., [20,31]).

Since tocotrienols degrade HMGCR by promoting HMGCR–Insig binding [15], we hypothesized that tocotrienols would reduce active SREBP-2 by inhibiting its processing in an Insig-dependent mechanism. To test this, we compared CHO-7 cells with Insig-deficient SRD-15 cells [32]. For instance, 25-HC ablates SREBP-2 processing in CHO-7 cells (Figure 5A, left-hand panel), demonstrated by a loss of the active mature form (M) and accumulation of the precursor form (P). This effect was severely blunted in SRD-15 cells (Figure 5A, right-hand panel), confirming that 25-HC acts through Insig to inhibit SREBP-2 processing [33]. In contrast, the tocotrienols had a similar effect in both cell lines (Figure 5A), suggesting that tocotrienols reduce SREBP-2(M) independently of Insig.

Tocotrienols lower SREBP-2 activity by degrading nuclear SREBP-2

Figure 5
Tocotrienols lower SREBP-2 activity by degrading nuclear SREBP-2

(A) CHO-7 and SRD-15 cells were plated in medium C, and starved for 16 h in medium D. The cells were then treated for 4 h with vehicle (Veh) or betulin (Bet), α-tocotrienol (αT3), γ-tocotrienol (γT3), δ-tocotrienol (δT3) or 25-HC, all at 10 μM in medium D. Protein was harvested and subjected to SDS/PAGE and Western blotting against SREBP-2 and α-tubulin. Densitometry was performed on the bands representing the precursor (P) and mature (M) forms of SREBP-2, and processing was calculated as M/(P+M). (B) SRD-1 cells were plated in medium C, and starved for 16 h in medium D. The cells were then treated with δ-tocotrienol (δT3) (10 μM) and/or 25-HC (10 μM) as indicated, in medium D. Following treatment, protein was harvested and subjected to SDS/PAGE and Western blotting as in (A). (C) CHO-7 cells were seeded in 60-mm-diameter dishes and transfected with HA-SREBP-2(M) constructs [wild-type (WT), -DNABmut or -CPDmut]. The next day, cells were then seeded into six-well plates in medium C, and starved for 16 h in medium D. The cells were either untreated or treated for 6 h with or without δ-tocotrienol (δT3) (10 μM) in the presence of cycloheximide (CHX) (10 μg/ml). Following treatment, protein was harvested and subjected to SDS/PAGE and Western blotting as in (A). For HA, the ‘low exposure’ was 5 s and ‘high exposure’ was 20 s. The asterisk (*) denotes phosphorylated SREBP-2(M). (D) CHO-7 cells were transfected with wild-type HA-SREBP-2(M), seeded and starved as in (C). The cells were then treated for 6 h with or without δ-tocotrienol (δT3) (10 μM) or MG132 (10 μM), in the presence of cycloheximide (CHX) (10 μg/ml). Following treatment, protein was harvested and subjected to SDS/PAGE and Western blotting as in (A). (AD) Blots are each representative of at least two separate experiments. Molecular masses are indicated in kDa.

Figure 5
Tocotrienols lower SREBP-2 activity by degrading nuclear SREBP-2

(A) CHO-7 and SRD-15 cells were plated in medium C, and starved for 16 h in medium D. The cells were then treated for 4 h with vehicle (Veh) or betulin (Bet), α-tocotrienol (αT3), γ-tocotrienol (γT3), δ-tocotrienol (δT3) or 25-HC, all at 10 μM in medium D. Protein was harvested and subjected to SDS/PAGE and Western blotting against SREBP-2 and α-tubulin. Densitometry was performed on the bands representing the precursor (P) and mature (M) forms of SREBP-2, and processing was calculated as M/(P+M). (B) SRD-1 cells were plated in medium C, and starved for 16 h in medium D. The cells were then treated with δ-tocotrienol (δT3) (10 μM) and/or 25-HC (10 μM) as indicated, in medium D. Following treatment, protein was harvested and subjected to SDS/PAGE and Western blotting as in (A). (C) CHO-7 cells were seeded in 60-mm-diameter dishes and transfected with HA-SREBP-2(M) constructs [wild-type (WT), -DNABmut or -CPDmut]. The next day, cells were then seeded into six-well plates in medium C, and starved for 16 h in medium D. The cells were either untreated or treated for 6 h with or without δ-tocotrienol (δT3) (10 μM) in the presence of cycloheximide (CHX) (10 μg/ml). Following treatment, protein was harvested and subjected to SDS/PAGE and Western blotting as in (A). For HA, the ‘low exposure’ was 5 s and ‘high exposure’ was 20 s. The asterisk (*) denotes phosphorylated SREBP-2(M). (D) CHO-7 cells were transfected with wild-type HA-SREBP-2(M), seeded and starved as in (C). The cells were then treated for 6 h with or without δ-tocotrienol (δT3) (10 μM) or MG132 (10 μM), in the presence of cycloheximide (CHX) (10 μg/ml). Following treatment, protein was harvested and subjected to SDS/PAGE and Western blotting as in (A). (AD) Blots are each representative of at least two separate experiments. Molecular masses are indicated in kDa.

Next, we tested whether tocotrienols reduce SREBP-2(M) by inhibiting SREBP-2 processing. We focused on δ-tocotrienol, since it had the most pronounced effect in CHO-7 cells (Figure 5A). We used the SRD-1 cell line, a mutant CHO-7 cell line that constitutively expresses SREBP-2(M) [34]. Its higher basal SREBP-2(M) levels required a longer treatment time before a clear fall in SREBP-2(M) levels was observed (8 h, compared with 4 h in CHO-7 cells) (Figure 5B). Interestingly, this δ-tocotrienol effect was enhanced by 25-HC. This oxysterol could promote SREBP-2(M) degradation by the Akt/GSK-3 (glycogen synthase kinase 3)/Fbw7 (F-box and WD repeat domain-containing 7) axis. This axis has been elucidated for SREBP-1 and generalized to SREBP-2 [5,35]: when mature SREBP binds DNA, it is phosphorylated by GSK-3 at the CPD motif. This primes SREBP for ubiquitination by the E3 ligase Fbw7 and subsequent proteasomal degradation [24]. This is prevented by Akt, which phosphorylates GSK-3. In turn, 25-HC degrades Akt [36], which would derepress GSK-3 and promote SREBP-2(M) degradation. Given that δ-tocotrienol has been shown to reduce Akt activity [37,38], and 25-HC and δ-tocotrienol act synergistically (Figure 5B), could δ-tocotrienol influence the Akt/GSK-3/Fbw7 axis of SREBP-2(M) degradation?

To test this, we used ectopic SREBP-2(M), with mutations in the DNA-binding domain (Y342R [23], denoted ‘DNABmut’) and CPD motif (S432A/S436A [24], denoted ‘CPDmut’). The phosphorylation doublet seen with wild-type SREBP-2(M) (WT) was lost with CPDmut (Figure 5C). Treatment with δ-tocotrienol was accompanied by co-treatment with a translation inhibitor, cycloheximide, to ensure only effects on degradation were being observed. δ-Tocotrienol promoted the degradation of both wild-type and mutant SREBP-2(M), suggesting that this occurs independently of the Akt/GSK3/Fbw7 axis. Furthermore, although the proteasomal inhibitor MG132 increased SREBP-2(M), it could not rescue the δ-tocotrienol effect (Figure 5D). Together, these data suggest that tocotrienols degrade SREBP-2(M) independently of the proteasome.

To confirm that this SREBP-2 degradation has downstream effects, we examined the tocotrienol effect on SREBP-2 target gene expression. Similarly to SREBP-2 processing (Figure 5A), δ-tocotrienol reduced the expression of SREBP-2 target genes (LDLR and HMGCR) in CHO-7 cells (Figure 6). Similar effects were observed at 4 h (n=3, results not shown). In SRD-1 cells, δ-tocotrienol reduced LDLR expression by 37±3% at 8 h under the same experimental conditions (n=3). These effects could be translated to a prostate setting, with SREBP-2 target gene expression being suppressed in LNCaP cells (Figure 6). In contrast, LXR target gene (ABCA1 and ABCG1) expression was unaffected by δ-tocotrienol (Figure 6), but increased with an LXR ligand (25-HC), together showing a similar trend to LXR-driven luciferase activity (Figure 1C). Overall, tocotrienols degrade SREBP-2(M), having downstream effects on target gene expression in both the CHO system and prostate cancer cells.

δ-Tocotrienol has downstream effects on SREBP-2 target gene expression

Figure 6
δ-Tocotrienol has downstream effects on SREBP-2 target gene expression

CHO-7 cells were plated in medium C, and starved for 16 h in medium D. LNCaP cells were plated in medium A, and then starved for 16 h in medium B, supplemented with 5 μM compactin and 50 μM mevalonate to be consistent with the CHO-7 cells. Cells were treated for 8 h with 10 μM δ-tocotrienol (δT3) or 10 μM 25-HC in their respective starvation medium. Following treatment, RNA was harvested and mRNA expression was determined by qRT-PCR, normalized to the vehicle (Veh) condition. Results are means+S.E.M. from three separate experiments per cell line, each performed with triplicate wells per condition.

Figure 6
δ-Tocotrienol has downstream effects on SREBP-2 target gene expression

CHO-7 cells were plated in medium C, and starved for 16 h in medium D. LNCaP cells were plated in medium A, and then starved for 16 h in medium B, supplemented with 5 μM compactin and 50 μM mevalonate to be consistent with the CHO-7 cells. Cells were treated for 8 h with 10 μM δ-tocotrienol (δT3) or 10 μM 25-HC in their respective starvation medium. Following treatment, RNA was harvested and mRNA expression was determined by qRT-PCR, normalized to the vehicle (Veh) condition. Results are means+S.E.M. from three separate experiments per cell line, each performed with triplicate wells per condition.

Modulating SCAP levels alters sensitivity to tocotrienols

Since tocotrienols degrade SREBP-2(M), could higher SREBP-2 processing provide resistance to tocotrienols? In support of this, we found that SCAP mRNA levels were higher in PC-3 than LNCaP cells (Figure 7A). Higher SCAP levels would promote SREBP-2 processing, generating more SREBP-2(M) to protect against tocotrienols.

Changing SCAP levels alters sensitivity to tocotrienols

Figure 7
Changing SCAP levels alters sensitivity to tocotrienols

(A and B) LNCaP and PC-3 cells were grown in medium A. (A) RNA was harvested and SCAP mRNA levels were determined by qRT-PCR, normalized to LNCaP cells. (B) Protein was harvested and subjected to SDS/PAGE and Western blotting against SCAP and α-tubulin. The asterisk (*) denotes a non-specific band explored further in Supplementary Figures S2 and S3 at http://www.BiochemJ.org/bj/446/bj4460191add.htm. Densitometry was performed on the SCAP band (150 kDa) and α-tubulin, and relative density was calculated as SCAP/α-tubulin, normalized to LNCaP cells. (C) PC-3 cells were transfected with negative-control scrambled siRNA (siNC) or SCAP-specific siRNA (siSCAP2), then seeded into 96-well plates. The next day, cells were treated with different concentrations of δ-tocotrienol (δT3) in FBLPDS for 3 days, after which the WST-1 assay was performed. (D) LNCaP cells stably transfected with empty vector (EV), wild-type SCAP (WT) or Y243A-mutant SCAP (mut) were grown in medium A, supplemented with 250 μg/ml zeocin. Protein was harvested and subjected to SDS/PAGE and Western blotting against Myc, SCAP and α-tubulin. For SCAP, the ‘low exposure’ was 5 s and ‘high exposure’ was 30 s. (E) Left-hand panel: LNCaP clones described in (D) were transfected with SRE-luc and treated with different concentrations of δT3, in the presence of compactin (5 μM). Following treatment, cells were assayed for luciferase activity. Middle panel: LNCaP clones were treated with different concentrations of δT3 for 3 days as in (C), after which the WST-1 assay was performed. Right-hand panel: the effects of δT3 on cell viability and SREBP-2 activity were correlated as described in Figure 2(B). The correlation for the combined dataset is y=1.21x−0.24 (R2≈0.90, P<0.01, regression analysis). (B and D) Blots are representative of at least two separate experiments. Molecular masses are indicated in kDa. (AC and E) Results are means+S.E.M. from at least three separate experiments, with each experiment performed with triplicate wells per condition.

Figure 7
Changing SCAP levels alters sensitivity to tocotrienols

(A and B) LNCaP and PC-3 cells were grown in medium A. (A) RNA was harvested and SCAP mRNA levels were determined by qRT-PCR, normalized to LNCaP cells. (B) Protein was harvested and subjected to SDS/PAGE and Western blotting against SCAP and α-tubulin. The asterisk (*) denotes a non-specific band explored further in Supplementary Figures S2 and S3 at http://www.BiochemJ.org/bj/446/bj4460191add.htm. Densitometry was performed on the SCAP band (150 kDa) and α-tubulin, and relative density was calculated as SCAP/α-tubulin, normalized to LNCaP cells. (C) PC-3 cells were transfected with negative-control scrambled siRNA (siNC) or SCAP-specific siRNA (siSCAP2), then seeded into 96-well plates. The next day, cells were treated with different concentrations of δ-tocotrienol (δT3) in FBLPDS for 3 days, after which the WST-1 assay was performed. (D) LNCaP cells stably transfected with empty vector (EV), wild-type SCAP (WT) or Y243A-mutant SCAP (mut) were grown in medium A, supplemented with 250 μg/ml zeocin. Protein was harvested and subjected to SDS/PAGE and Western blotting against Myc, SCAP and α-tubulin. For SCAP, the ‘low exposure’ was 5 s and ‘high exposure’ was 30 s. (E) Left-hand panel: LNCaP clones described in (D) were transfected with SRE-luc and treated with different concentrations of δT3, in the presence of compactin (5 μM). Following treatment, cells were assayed for luciferase activity. Middle panel: LNCaP clones were treated with different concentrations of δT3 for 3 days as in (C), after which the WST-1 assay was performed. Right-hand panel: the effects of δT3 on cell viability and SREBP-2 activity were correlated as described in Figure 2(B). The correlation for the combined dataset is y=1.21x−0.24 (R2≈0.90, P<0.01, regression analysis). (B and D) Blots are representative of at least two separate experiments. Molecular masses are indicated in kDa. (AC and E) Results are means+S.E.M. from at least three separate experiments, with each experiment performed with triplicate wells per condition.

To confirm that increased SCAP mRNA levels translate into more SCAP protein, we used the R139 antibody [27]. This antibody was raised against hamster SCAP; we have characterized it for use against human SCAP in the present study (Supplementary Figure S1 at http://www.BiochemJ.org/bj/446/bj4460191add.htm). We found that PC-3 cells have SCAP protein levels comparable with those in LNCaP cells (Figure 7B), despite differences at the mRNA level. In addition, we observed a 110 kDa band that was expressed prominently in PC-3 cells, but found it to be not specific to SCAP (Supplementary Figures S2 and S3 at http://www.BiochemJ.org/bj/446/bj4460191add.htm). Overall, it is unlikely that differences in SCAP expression explain the tocotrienol resistance in PC-3 cells.

Nevertheless, to confirm the role of SREBP-2 activity in tocotrienol resistance, we manipulated the levels of its regulator, SCAP. First, SCAP knockdown with siRNA lowered basal viability in PC-3 cells, which could not be reduced further with δ-tocotrienol (Figure 7C). Thus eliminating SCAP expression ablates the effect of δ-tocotrienol. Secondly, we generated LNCaP clones that stably express either ectopic human wild-type SCAP or mutant SCAP (Y234A), which cannot mediate SREBP-2 processing [39] (Figure 7D). Overexpressing wild-type SCAP increased resistance to δ-tocotrienol with regard to SREBP-2 activity and cell viability, whereas overexpressing mutant SCAP did not (Figure 7E). Again, a strong correlation exists between SREBP-2 activity and cell viability (Figure 7E, right-hand panel). Together, these results demonstrate that altering SCAP expression influences sensitivity to tocotrienols.

DISCUSSION

In the present study, we have used natural products to toggle cholesterol homoeostasis in prostate cancer cells, determining potential targets for prostate cancer therapy. We found that (i) tocotrienols degrade mature SREBP-2, (ii) altering SREBP-2 activity, such as by adjusting SCAP expression, influences sensitivity to tocotrienols, and (iii) targeting SREBP-2 is more lethal to prostate cancer cells than is LXR.

Initially, we focused our attention on both betulin and the tocotrienols. Betulin influenced SREBP-2 activity more in LNCaP cells than PC-3 cells (Figure 1B), but had a similar effect on cell viability (Figure 2A), suggesting that it exerts other effects that influence cell viability. Betulin has been shown to induce apoptosis in PC-3 cells [40], but we did not consider it further because, in contrast, tocotrienols demonstrated a strong correlation between SREBP-2 activity and cell viability (Figures 2B, 4C and 7E).

We found that γ- and δ-tocotrienols were more potent inhibitors of SREBP-2 than was α-tocotrienol. This trend was also observed with HMGCR degradation [15], and given that tocotrienols degrade HMGCR via Insig-1, we hypothesized that SREBP-2 activity was also reduced in an Insig-dependent manner. However, this was not the case, instead degrading SREBP-2(M) (Figure 5). It has been demonstrated that SREBP-2(M) is ubiquitinated [41] and degraded by the proteasome [42]. Furthermore, a well-characterized GSK-3-dependent degradation pathway has been elucidated for SREBP-1, with evidence suggesting that this also occurs for SREBP-2 [5,35]. However, δ-tocotrienol promoted SREBP-2(M) degradation independently of GSK-3 and the proteasome (Figure 5), with further work required to elucidate this mechanism. Nevertheless, this has downstream effects on SREBP-2 target gene expression (Figure 6), adding a new dimension to previous work that showed tocotrienols promote HMGCR degradation [15], and hence providing an additional means by which tocotrienols regulate cholesterol metabolism.

Furthermore, this implies that PC-3 cells could be more resistant to tocotrienols than LNCaP cells (Figure 1 and 2) because they have higher mature SREBP-2 levels than LNCaP cells [17]. We hypothesized the latter was due to high mRNA levels of the SREBP-2 activator, SCAP, in PC-3 cells (Figure 7A), but found this did not translate to the protein level (Figure 7B). Despite this, manipulating SCAP levels via siRNA knockdown or overexpression constructs had noticeable effects on tocotrienol sensitivity (Figure 7). Although we have not ruled out effects of tocotrienols on other transcription factors such as PPARs (peroxisome-proliferator-activated receptors) and RXR (retinoid X receptor), these SCAP modulation experiments (Figure 7) demonstrate that tocotrienol sensitivity is dependent on SREBP-2 activity. Moreover, mevalonate treatment rescued cell viability against HMGCR inhibition (compactin), but not against tocotrienols (Figure 2C), suggesting that the tocotrienols acted largely through SREBP-2, which regulates cholesterol levels more broadly than HMGCR. In addition to cholesterol synthesis, degrading SREBP-2 also reduces cholesterol uptake by down-regulating LDLR expression (Figure 6). Combined with the oxysterol data (Figure 3), this supports the notion that targeting SREBP-2 would be preferable over targeting LXR.

Several studies have explored LXR as a prostate cancer target, finding that LXR ligands arrested prostate cancer cell growth, via ABCA1, p27 and Akt inhibition [9,10]. However, these studies used the LXR ligand TO-901317, which has been shown to antagonize the androgen receptor [43] and activate nuclear receptors FXR (farnesoid X receptor) [44] and PXR (pregnane X receptor) [45]. Oxysterols, which additionally inhibit HMGCR and SREBP-2 activity [30], were also used. In contrast, we found that a more specific LXR ligand, GW683965, had little effect on cell viability and that SREBP-2 activity has a greater influence over cell viability than LXR activity. However, we have only shown this in a prostate cancer setting. Unique to prostate cells is the androgen receptor, a key regulator of prostate cell growth and differentiation, which also reduces LXR activity [19]. This is evident with GW683965 (Figure 1C), and oxysterols (Figure 3) being less potent LXR ligands in LNCaP cells compared with androgen-receptor-null PC-3 cells. The androgen receptor/LXR cross-talk would dampen the effect of LXR ligands on prostate cell viability, but targeting LXR may still be useful for other cancers.

In the context of tocotrienols, other studies have also found that γ- and δ-tocotrienol are more lethal to prostate cancer cells than α-tocotrienol [4648], with tocotrienols being more effective than tocopherols [38,46,48]. In contrast with the present study, one study found the viability of PC-3 cells to be more sensitive than LNCaP to tocotrienols [47], but this could be attributed to shorter treatment times (24–48 h, compared with 72 h in the present study) and serum conditions (FBS compared with FBLPDS). In fact, all other studies used full serum, achieving IC50 values above the low-micromolar range assessed in the present study. Despite this, similarly high concentrations had little effect on prostate epithelial cells [47,48], as seen with the tocotrienol-rich fraction of palm oil [49], suggesting a cancer-specific phenomenon. Nevertheless, attainable serum tocotrienol levels are in the high-nanomolar range (see, e.g., [50]), with postprandial levels peaking at ~2–3 μM after 4 h and returning to a submicromolar baseline after 8 h [50], which also suggests a low bioavailability. Thus physiologically relevant concentrations of tocotrienols only seem to influence cell viability under cholesterol-deficient conditions (compare Figures 2A and 2D).

Consequently, serum cholesterol levels may need to be lowered to enhance the effective tocotrienol dose. It should be noted that, although tocotrienols lower LDLR expression (Figure 6), this does not necessarily eliminate uptake of serum cholesterol by other means, such as via passive diffusion. Thus lowering serum cholesterol in vivo would confirm that the deficiency of cholesterol in FBLPDS is what causes tocotrienol sensitivity (compare Figures 2A and 2D), as opposed to other elements removed during FBLPDS preparation. As proof-of-principle, co-treatment of a statin with γ- or δ-tocotrienol worked synergistically to impair melanoma growth in vivo [51]; this could be tested in prostate cancer. Furthermore, although we found no changes in the expression of cholesterol-efflux genes (Figure 6), in vivo studies should consider other whole-body compensatory mechanisms, such as cholesterol reabsorption. In this respect, tocotrienols have been found to reduce serum-cholesterol levels (see, e.g., [52]), and could thus enhance their own observed effects against prostate cancer cells in vitro.

In addition, mechanistic insights from previous studies should be tested in this setting, such as increased apoptotic signalling [46,47], reduced metastasis by enhanced E-cadherin expression [47], increased PPARγ activity, and reduced TGFβ2 (transforming growth factor β2) signalling [48] by tocotrienols. Interestingly, one study observed tocotrienols that induced caspase-dependent apoptosis in LNCaP, but not PC-3 or DU-145, cells, from which the authors proposed that androgen-dependence influences sensitivity [46]. Using our progression model, we have shown that both androgen-dependent (LNCaP) and -independent (LNCaP-364) cells were sensitive to tocotrienols (Figure 5C) and oxysterols (results not shown), supporting the idea that cholesterol homoeostasis could be an alternative therapeutic target for androgen-independent prostate cancer [53].

There is very little in the way of treatment options to substantially extend survival in patients with advanced recurrent prostate cancer, with existing strategies including chemotherapeutic and immunotherapeutic approaches [54]. Could tocotrienols be useful here? For instance, γ-tocotrienol preferentially accumulated in prostate cancer xenografts, reduced xenograft growth, and sensitized xenografts to chemotherapy [55,56]. To our knowledge, no large-scale clinical trials have been conducted to evaluate tocotrienols for prostate cancer treatment. The potential for tocotrienols, and other drugs that manipulate cholesterol homoeostasis via SREBP-2, should be explored further in a prostate cancer setting. Given that SREBP-2 has also been considered for metabolic diseases such as diabetes, cardiovascular disease and hyperlipidaemia [11,57], this has implications for other disease settings.

Abbreviations

     
  • ABC

    ATP-binding cassette transporter

  •  
  • CHO

    Chinese-hamster ovary

  •  
  • CPD

    Cdc4 phosphodegron

  •  
  • 24,25-EC

    (24S,25)-epoxycholesterol

  •  
  • FBLPDS

    lipoprotein-deficient fetal bovine serum

  •  
  • FBS

    fetal bovine serum

  •  
  • Fbw7

    F-box and WD repeat domain-containing 7

  •  
  • GSK-3

    glycogen synthase kinase 3

  •  
  • HA

    haemagglutinin

  •  
  • 25-HC

    25-hydroxycholesterol

  •  
  • HMGCR

    3-hydroxy-3-methylglutaryl-CoA reductase

  •  
  • Insig

    insulin-induced gene

  •  
  • LDLR

    low-density lipoprotein receptor

  •  
  • LXR

    liver X receptor

  •  
  • NCLPDS

    lipoprotein-deficient newborn calf serum

  •  
  • NCS

    newborn calf serum

  •  
  • PIPE

    polymerase incomplete primer extension

  •  
  • PPAR

    peroxisome-proliferator-activated receptor

  •  
  • qRT-PCR

    quantitative real-time reverse transcription PCR

  •  
  • SDM

    site-directed mutagenesis

  •  
  • SREBP-2

    sterol-regulatory-element-binding protein isoform 2

  •  
  • SCAP

    SREBP-2 cleavage activation protein

  •  
  • siRNA

    small interfering RNA

  •  
  • siNC

    negative control siRNA

  •  
  • siSCAP

    siRNA against SCAP

  •  
  • TK

    thymidine kinase

  •  
  • WST-1

    water-soluble tetrazolium salt 1

AUTHOR CONTRIBUTION

The study was conceived and designed by James Krycer and Andrew Brown. Experimentation and data analysis was performed by James Krycer and Lisa Phan. The paper was written by James Krycer and Andrew Brown, and proofread by Lisa Phan before submission.

We thank the members of the Brown Research Laboratory for feedback throughout the investigation. We are grateful to numerous colleagues for providing the cell lines, reagents, plasmids and qRT-PCR primer designs used in the present study.

FUNDING

J.R.K. is the recipient of the Petre Foundation scholarship. A.J.B.'s research is supported by the Prostate Cancer Foundation of Australia [grant number PG2710].

References

References
1
Jemal
A.
Bray
F.
Center
M. M.
Ferlay
J.
Ward
E.
Forman
D.
Global cancer statistics
Ca Cancer J. Clin.
2011
, vol. 
61
 (pg. 
69
-
90
)
2
Brown
A. J.
Cholesterol, statins and cancer
Clin. Exp. Pharmacol. Physiol.
2007
, vol. 
34
 (pg. 
135
-
141
)
3
Solomon
K. R.
Freeman
M. R.
Do the cholesterol-lowering properties of statins affect cancer risk? Trends Endocrinol
Metab.
2008
, vol. 
19
 (pg. 
113
-
121
)
4
Schaffner
C. P.
Prostatic cholesterol metabolism: regulation and alteration
Prog. Clin. Biol. Res.
1981
, vol. 
75A
 (pg. 
279
-
324
)
5
Krycer
J. R.
Sharpe
L. J.
Luu
W.
Brown
A. J.
The Akt–SREBP nexus: cell signaling meets lipid metabolism
Trends Endocrinol. Metab.
2010
, vol. 
21
 (pg. 
268
-
276
)
6
Zhuang
L.
Kim
J.
Adam
R. M.
Solomon
K. R.
Freeman
M. R.
Cholesterol targeting alters lipid raft composition and cell survival in prostate cancer cells and xenografts
J. Clin. Invest.
2005
, vol. 
115
 (pg. 
959
-
968
)
7
Brown
A. J.
Does upsizing statins have a downside?
Drug Saf.
2010
, vol. 
33
 (pg. 
435
-
436
)
8
Brown
M. S.
Goldstein
J. L.
Cholesterol feedback: from Schoenheimer's bottle to Scap's MELADL
J. Lipid Res.
2009
, vol. 
50
 
(Suppl)
(pg. 
S15
-
S27
)
9
Chuu
C. P.
Kokontis
J. M.
Hiipakka
R. A.
Liao
S.
Modulation of liver X receptor signaling as novel therapy for prostate cancer
J. Biomed. Sci.
2007
, vol. 
14
 (pg. 
543
-
553
)
10
Pommier
A. J.
Alves
G.
Viennois
E.
Bernard
S.
Communal
Y.
Sion
B.
Marceau
G.
Damon
C.
Mouzat
K.
Caira
F.
, et al. 
Liver X receptor activation downregulates AKT survival signaling in lipid rafts and induces apoptosis of prostate cancer cells
Oncogene
2010
, vol. 
29
 (pg. 
2712
-
2723
)
11
Tang
J. J.
Li
J. G.
Qi
W.
Qiu
W. W.
Li
P. S.
Li
B. L.
Song
B. L.
Inhibition of SREBP by a small molecule, betulin, improves hyperlipidemia and insulin resistance and reduces atherosclerotic plaques
Cell Metab.
2011
, vol. 
13
 (pg. 
44
-
56
)
12
Sen
C. K.
Khanna
S.
Roy
S.
Tocotrienols: vitamin E beyond tocopherols
Life Sci.
2006
, vol. 
78
 (pg. 
2088
-
2098
)
13
Pearce
B. C.
Parker
R. A.
Deason
M. E.
Qureshi
A. A.
Wright
J. J.
Hypocholesterolemic activity of synthetic and natural tocotrienols
J. Med. Chem.
1992
, vol. 
35
 (pg. 
3595
-
3606
)
14
Parker
R. A.
Pearce
B. C.
Clark
R. W.
Gordon
D. A.
Wright
J. J.
Tocotrienols regulate cholesterol production in mammalian cells by post-transcriptional suppression of 3-hydroxy-3-methylglutaryl-coenzyme A reductase
J. Biol. Chem.
1993
, vol. 
268
 (pg. 
11230
-
11238
)
15
Song
B. L.
DeBose-Boyd
R. A.
Insig-dependent ubiquitination and degradation of 3-hydroxy-3-methylglutaryl coenzyme a reductase stimulated by δ- and γ-tocotrienols
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
25054
-
25061
)
16
Ling
M. T.
Luk
S. U.
Al-Ejeh
F.
Khanna
K. K.
Tocotrienol as a potential anticancer agent
Carcinogenesis
2012
, vol. 
33
 (pg. 
233
-
239
)
17
Krycer
J. R.
Kristiana
I.
Brown
A. J.
Cholesterol homeostasis in two commonly used human prostate cancer cell-lines, LNCaP and PC-3
PLoS ONE
2009
, vol. 
4
 pg. 
e8496
 
18
Goldstein
J. L.
Basu
S. K.
Brown
M. S.
Receptor-mediated endocytosis of low-density lipoprotein in cultured cells
Methods Enzymol.
1983
, vol. 
98
 (pg. 
241
-
260
)
19
Krycer
J. R.
Brown
A. J.
Cross-talk between the androgen receptor and the liver X receptor: implications for cholesterol homeostasis
J. Biol. Chem.
2011
, vol. 
286
 (pg. 
20637
-
20647
)
20
Gill
S.
Stevenson
J.
Kristiana
I.
Brown
A. J.
Cholesterol-dependent degradation of squalene monooxygenase, a control point in cholesterol synthesis beyond HMG-CoA reductase
Cell Metab.
2011
, vol. 
13
 (pg. 
260
-
273
)
21
Klock
H. E.
Lesley
S. A.
The Polymerase Incomplete Primer Extension (PIPE) method applied to high-throughput cloning and site-directed mutagenesis
Methods Mol. Biol.
2009
, vol. 
498
 (pg. 
91
-
103
)
22
Sanchis
J.
Fernandez
L.
Carballeira
J. D.
Drone
J.
Gumulya
Y.
Hobenreich
H.
Kahakeaw
D.
Kille
S.
Lohmer
R.
Peyralans
J. J.
, et al. 
Improved PCR method for the creation of saturation mutagenesis libraries in directed evolution: application to difficult-to-amplify templates
Appl. Microbiol. Biotechnol.
2008
, vol. 
81
 (pg. 
387
-
397
)
23
Amemiya-Kudo
M.
Shimano
H.
Hasty
A. H.
Yahagi
N.
Yoshikawa
T.
Matsuzaka
T.
Okazaki
H.
Tamura
Y.
Iizuka
Y.
Ohashi
K.
, et al. 
Transcriptional activities of nuclear SREBP-1a, -1c, and -2 to different target promoters of lipogenic and cholesterogenic genes
J. Lipid Res.
2002
, vol. 
43
 (pg. 
1220
-
1235
)
24
Sundqvist
A.
Bengoechea-Alonso
M. T.
Ye
X.
Lukiyanchuk
V.
Jin
J.
Harper
J. W.
Ericsson
J.
Control of lipid metabolism by phosphorylation-dependent degradation of the SREBP family of transcription factors by SCFFbw7
Cell Metab.
2005
, vol. 
1
 (pg. 
379
-
391
)
25
Sedelaar
J. P.
Isaacs
J. T.
Tissue culture media supplemented with 10% fetal calf serum contains a castrate level of testosterone
Prostate
2009
, vol. 
69
 (pg. 
1724
-
1729
)
26
Wong
J.
Quinn
C. M.
Brown
A. J.
Synthesis of the oxysterol, 24(S),25-epoxycholesterol, parallels cholesterol production and may protect against cellular accumulation of newly-synthesized cholesterol
Lipids Health Dis.
2007
, vol. 
6
 pg. 
10
 
27
Sakai
J.
Nohturfft
A.
Cheng
D.
Ho
Y. K.
Brown
M. S.
Goldstein
J. L.
Identification of complexes between the COOH-terminal domains of sterol regulatory element-binding proteins (SREBPs) and SREBP cleavage-activating protein
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
20213
-
20221
)
28
Wong
J.
Quinn
C. M.
Brown
A. J.
SREBP-2 positively regulates transcription of the cholesterol efflux gene, ABCA1, by generating oxysterol ligands for LXR
Biochem. J.
2006
, vol. 
400
 (pg. 
485
-
491
)
29
Gelissen
I. C.
Harris
M.
Rye
K. A.
Quinn
C.
Brown
A. J.
Kockx
M.
Cartland
S.
Packianathan
M.
Kritharides
L.
Jessup
W.
ABCA1 and ABCG1 synergize to mediate cholesterol export to apoA-I
Arterioscler., Thromb., Vasc. Biol.
2006
, vol. 
26
 (pg. 
534
-
540
)
30
Gill
S.
Chow
R.
Brown
A. J.
Sterol regulators of cholesterol homeostasis and beyond: the oxysterol hypothesis revisited and revised
Prog. Lipid Res.
2008
, vol. 
47
 (pg. 
391
-
404
)
31
Metherall
J. E.
Goldstein
J. L.
Luskey
K. L.
Brown
M. S.
Loss of transcriptional repression of three sterol-regulated genes in mutant hamster cells
J. Biol. Chem.
1989
, vol. 
264
 (pg. 
15634
-
15641
)
32
Lee
P. C.
Sever
N.
Debose-Boyd
R. A.
Isolation of sterol-resistant Chinese hamster ovary cells with genetic deficiencies in both Insig-1 and Insig-2
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
25242
-
25249
)
33
Radhakrishnan
A.
Ikeda
Y.
Kwon
H. J.
Brown
M. S.
Goldstein
J. L.
Sterol-regulated transport of SREBPs from endoplasmic reticulum to Golgi: oxysterols block transport by binding to Insig
Proc. Natl. Acad. Sci. U.S.A.
2007
, vol. 
104
 (pg. 
6511
-
6518
)
34
Yang
J.
Sato
R.
Goldstein
J. L.
Brown
M. S.
Sterol-resistant transcription in CHO cells caused by gene rearrangement that truncates SREBP-2
Genes Dev.
1994
, vol. 
8
 (pg. 
1910
-
1919
)
35
Bengoechea-Alonso
M. T.
Ericsson
J.
SREBP in signal transduction: cholesterol metabolism and beyond
Curr. Opin. Cell Biol.
2007
, vol. 
19
 (pg. 
215
-
222
)
36
Liu
J.
Netherland
C.
Pickle
T.
Sinensky
M. S.
Thewke
D. P.
Stimulation of Akt poly-ubiquitination and proteasomal degradation in P388D1 cells by 7-ketocholesterol and 25-hydroxycholesterol
Arch. Biochem. Biophys.
2009
, vol. 
487
 (pg. 
54
-
58
)
37
Samant
G. V.
Sylvester
P. W.
γ-Tocotrienol inhibits ErbB3-dependent PI3K/Akt mitogenic signalling in neoplastic mammary epithelial cells
Cell Proliferation
2006
, vol. 
39
 (pg. 
563
-
574
)
38
Jiang
Q.
Rao
X.
Kim
C. Y.
Freiser
H.
Zhang
Q.
Jiang
Z.
Li
G.
γ-Tocotrienol induces apoptosis and autophagy in prostate cancer cells by increasing intracellular dihydrosphingosine and dihydroceramide
Int. J. Cancer
2012
, vol. 
130
 (pg. 
685
-
693
)
39
Motamed
M.
Zhang
Y.
Wang
M. L.
Seemann
J.
Kwon
H. J.
Goldstein
J. L.
Brown
M. S.
Identification of luminal Loop 1 of Scap protein as the sterol sensor that maintains cholesterol homeostasis
J. Biol. Chem.
2011
, vol. 
286
 (pg. 
18002
-
18012
)
40
Li
Y.
He
K.
Huang
Y.
Zheng
D.
Gao
C.
Cui
L.
Jin
Y. H.
Betulin induces mitochondrial cytochrome c release associated apoptosis in human cancer cells
Mol. Carcinog.
2010
, vol. 
49
 (pg. 
630
-
640
)
41
Kim
W.
Bennett
E. J.
Huttlin
E. L.
Guo
A.
Li
J.
Possemato
A.
Sowa
M. E.
Rad
R.
Rush
J.
Comb
M. J.
, et al. 
Systematic and quantitative assessment of the ubiquitin-modified proteome
Mol. Cell
2011
, vol. 
44
 (pg. 
325
-
340
)
42
Hirano
Y.
Yoshida
M.
Shimizu
M.
Sato
R.
Direct demonstration of rapid degradation of nuclear sterol regulatory element-binding proteins by the ubiquitin–proteasome pathway
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
36431
-
36437
)
43
Chuu
C. P.
Chen
R. Y.
Hiipakka
R. A.
Kokontis
J. M.
Warner
K. V.
Xiang
J.
Liao
S.
The liver X receptor agonist T0901317 acts as androgen receptor antagonist in human prostate cancer cells
Biochem. Biophys. Res. Commun.
2007
, vol. 
357
 (pg. 
341
-
346
)
44
Houck
K. A.
Borchert
K. M.
Hepler
C. D.
Thomas
J. S.
Bramlett
K. S.
Michael
L. F.
Burris
T. P.
T0901317 is a dual LXR/FXR agonist
Mol. Genet. Metab.
2004
, vol. 
83
 (pg. 
184
-
187
)
45
Mitro
N.
Vargas
L.
Romeo
R.
Koder
A.
Saez
E.
T0901317 is a potent PXR ligand: implications for the biology ascribed to LXR
FEBS Lett.
2007
, vol. 
581
 (pg. 
1721
-
1726
)
46
Constantinou
C.
Hyatt
J. A.
Vraka
P. S.
Papas
A.
Papas
K. A.
Neophytou
C.
Hadjivassiliou
V.
Constantinou
A. I.
Induction of caspase-independent programmed cell death by vitamin E natural homologs and synthetic derivatives
Nutr. Cancer
2009
, vol. 
61
 (pg. 
864
-
874
)
47
Yap
W. N.
Chang
P. N.
Han
H. Y.
Lee
D. T.
Ling
M. T.
Wong
Y. C.
Yap
Y. L.
γ-Tocotrienol suppresses prostate cancer cell proliferation and invasion through multiple-signalling pathways
Br. J. Cancer
2008
, vol. 
99
 (pg. 
1832
-
1841
)
48
Campbell
S. E.
Rudder
B.
Phillips
R. B.
Whaley
S. G.
Stimmel
J. B.
Leesnitzer
L. M.
Lightner
J.
Dessus-Babus
S.
Duffourc
M.
Stone
W. L.
, et al. 
γ-Tocotrienol induces growth arrest through a novel pathway with TGFβ2 in prostate cancer
Free Radical Biol. Med.
2011
, vol. 
50
 (pg. 
1344
-
1354
)
49
Srivastava
J. K.
Gupta
S.
Tocotrienol-rich fraction of palm oil induces cell cycle arrest and apoptosis selectively in human prostate cancer cells
Biochem. Biophys. Res. Commun.
2006
, vol. 
346
 (pg. 
447
-
453
)
50
Khosla
P.
Patel
V.
Whinter
J. M.
Khanna
S.
Rakhkovskaya
M.
Roy
S.
Sen
C. K.
Postprandial levels of the natural vitamin E tocotrienol in human circulation
Antioxid. Redox Signaling
2006
, vol. 
8
 (pg. 
1059
-
1068
)
51
McAnally
J. A.
Gupta
J.
Sodhani
S.
Bravo
L.
Mo
H.
Tocotrienols potentiate lovastatin-mediated growth suppression in vitro and in vivo
Exp. Biol. Med.
2007
, vol. 
232
 (pg. 
523
-
531
)
52
Qureshi
A. A.
Sami
S. A.
Salser
W. A.
Khan
F. A.
Dose-dependent suppression of serum cholesterol by tocotrienol-rich fraction (TRF25) of rice bran in hypercholesterolemic humans
Atherosclerosis
2002
, vol. 
161
 (pg. 
199
-
207
)
53
Twiddy
A. L.
Leon
C. G.
Wasan
K. M.
Cholesterol as a potential target for castration-resistant prostate cancer
Pharm. Res.
2011
, vol. 
28
 (pg. 
423
-
437
)
54
Di Lorenzo
G.
Buonerba
C.
Autorino
R.
De Placido
S.
Sternberg
C. N.
Castration-resistant prostate cancer: current and emerging treatment strategies
Drugs
2010
, vol. 
70
 (pg. 
983
-
1000
)
55
Yap
W. N.
Zaiden
N.
Luk
S. Y.
Lee
D. T.
Ling
M. T.
Wong
Y. C.
Yap
Y. L.
In vivo evidence of γ-tocotrienol as a chemosensitizer in the treatment of hormone-refractory prostate cancer
Pharmacology
2010
, vol. 
85
 (pg. 
248
-
258
)
56
Luk
S. U.
Yap
W. N.
Chiu
Y. T.
Lee
D. T.
Ma
S.
Lee
T. K.
Vasireddy
R. S.
Wong
Y. C.
Ching
Y. P.
Nelson
C.
, et al. 
γ-Tocotrienol as an effective agent in targeting prostate cancer stem cell-like population
Int. J. Cancer
2011
, vol. 
128
 (pg. 
2182
-
2191
)
57
Krycer
J. R.
Brown
A. J.
Putative fat fighter hits the middle man
Chem. Biol.
2009
, vol. 
16
 (pg. 
798
-
800
)

Supplementary data