Prostate cancer chemoprevention by sulforaphane in a preclinical mouse model is associated with inhibition of fatty acid metabolism

Increased fatty acid synthesis is a rather unique mechanism of prostate cancer. This study demonstrates, for the first time, that prostate cancer prevention by SFN in TRAMP mice is associated with inhibition of fatty acid metabolism.

Abstract

Increased de novo synthesis of fatty acids is a rather unique and targetable mechanism of human prostate cancer. We have shown previously that oral administration of sulforaphane (SFN) significantly inhibits the incidence and/or burden of prostatic intraepithelial neoplasia and well-differentiated adenocarcinoma in TRansgenic Adenocarcinoma of Mouse Prostate (TRAMP) mice. The present study used cellular models of prostate cancer and archived plasma/adenocarcinoma tissues and sections from the TRAMP study to demonstrate inhibition of fatty acid synthesis by SFN treatment in vitro and in vivo. Treatment of androgen-responsive (LNCaP) and castration-resistant (22Rv1) human prostate cancer cells with SFN (5 and 10 μM) resulted in downregulation of protein and mRNA levels of acetyl-CoA carboxylase 1 (ACC1) and fatty acid synthase (FASN), but not ATP citrate lyase. Protein and mRNA levels of carnitine palmitoyltransferase 1A (CPT1A), which facilitates fatty acid uptake by mitochondria for β-oxidation, were also decreased following SFN treatment in both cell lines. Immunohistochemistry revealed a significant decrease in expression of FASN and ACC1 proteins in prostate adenocarcinoma sections of SFN-treated TRAMP mice when compared with controls. SFN administration to TRAMP mice resulted in a significant decrease in plasma and/or prostate adenocarcinoma levels of total free fatty acids, total phospholipids, acetyl-CoA and ATP. Consistent with these results, number of neutral lipid droplets was lower in the prostate adenocarcinoma sections of SFN-treated TRAMP mice than in control tumors. Collectively, these observations indicate that prostate cancer chemoprevention by SFN in TRAMP mice is associated with inhibition of fatty acid metabolism.

Introduction

Sulforaphane (SFN) is a phytochemical occurring naturally as glucoraphanin (thioglucoside conjugate) in edible cruciferous vegetables like broccoli, and generated upon plant damage from cutting or chewing in a reaction catalysed by myrosinase (1). SFN is effective for chemoprevention of prostate cancer in preclinical studies (2,3). We were the first to show chemoprevention of prostate cancer by oral administration of SFN (6 μmol/mouse, three times/week for 17–19 weeks) in TRansgenic Adenocarcinoma of Mouse Prostate (TRAMP) mice (2). With this regimen of oral SFN treatment, the incidence of the prostatic intraepithelial neoplasia (PIN) and well-differentiated adenocarcinoma in the dorsolateral prostate of TRAMP mice was decreased by ~23–28% when compared with controls (2). The burden (area) of well-differentiated adenocarcinoma in the dorsolateral prostate was also lower by ~44% in the SFN treatment group compared to control group (2). We showed further that SFN treatment provoked cytoprotective autophagy in cultured human prostate cancer cells leading to inhibition of cytosolic release of cytochrome c and apoptotic cell death (4). Accordingly, prostate cancer chemoprevention by SFN in TRAMP mice was augmented by pharmacological inhibition of autophagy (3). The TRAMP mice fed 240 mg of broccoli sprout/mouse/day for 16 weeks exhibited a significant decrease in prostate tumor growth (5).

Preclinical studies, including those from our laboratory, have led to clinical trials investigating the effect of SFN or broccoli sprout extract on androgen receptor (AR) function (6–8). In a phase II study, 20 patients with recurrent prostate cancer were treated with 200 μmol/day of SFN-rich broccoli sprout extract for a maximum period of up to 20 weeks (8). Even though only one subject experienced the primary end point of a ≥50% decline in serum prostate-specific antigen (PSA) level, a significant lengthening of PSA doubling time compared with the pre-treatment (6.1 months pre-treatment versus 9.6 months on-treatment; P = 0.044) was also reported (8). In another multisite French study with daily oral administration of 60 mg of a stabilized free SFN preparation for 6 months, the primary end point of 0.012 log (ng/ml)/month decrease in the log PSA slope was not achieved, but the median log PSA slopes were consistently lower in SFN-treated men (9). These results are encouraging but necessitate additional studies to identify novel mechanistic targets of SFN.

Prostate cancer is rather unique among epithelial solid tumors because of its heavy reliance on β-oxidation of fatty acids to meet cellular energy demand (10–12). Overexpression of mRNA or protein levels of key fatty acid synthesis enzymes, including ATP citrate lyase (ACLY), acetyl-CoA carboxylase (ACC) and fatty acid synthase (FASN) were observed in human prostate cancers (11,13,14). Pharmacologic or genetic suppression of ACLY, ACC1 and FASN inhibited prostate cancer cell growth in vitro and in vivo (10,11,15–19). Increased expression of FASN in human prostate cancer correlated with a more aggressive phenotype (20,21). Transgenic expression of FASN in mice resulted in PIN development (22). Furthermore, blood levels of saturated and monounsaturated fatty acids were suggested to be markers of de novo lipogenesis and the risk of prostate cancer (23). Therefore, inhibition of synthesis and/or β-oxidation of fatty acids represent a viable mechanistic strategy for chemoprevention of prostate cancer. The present study was designed to test whether fatty acid metabolism is a mechanistic target of prostate cancer chemoprevention by SFN.

Materials and methods

Ethics statement

Plasma and prostate adenocarcinoma tissues/sections from the previously published TRAMP studies (2,3), which were approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh, were used to determine the in vivo effect of oral SFN administration on expression of fatty acid metabolism proteins (ACC1, FASN and CPT1A) as well as levels of acetyl-CoA, total free fatty acids, total phospholipids and ATP.

Reagents

SFN (purity ≥98%) was purchased from LKT Laboratories (St. Paul, MN). Fetal bovine serum, phosphate-buffered saline (PBS), antibiotic mixture and other cell culture reagents were purchased from Life Technologies-Thermo Fisher Scientific (Waltham, MA). RPMI 1640 medium was purchased from Mediatech (a Corning subsidiary, Manassas, VA). Anti-ACC1 (cat. #4190), anti-acetyl-CoA carboxylase 2 (ACC2; cat. #8578) and anti-FASN (cat. #3180) antibodies were purchased from Cell Signaling Technology (Beverly, MA). The anti-ACLY (cat. #ab40793), anti-carnitine palmitoyltransferase 1A (CPT1A; cat. #ab128568) and anti-sterol regulatory element binding protein 1 (SREBP1; cat. #ab28481) antibodies were purchased from Abcam (Cambridge, MA). Anticarbohydrate response element binding protein (ChREBP; cat. #NB400-135) antibody was purchased from Novus Biologicals (Littleton, CO). The anti-AR (cat. #sc-816) and anti-α-smooth muscle actin (α-SMA; cat. #sc-32251) antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, Dallas, TX). A kit (cat. #ab118183) for determination of levels of fatty acid β-oxidation dehydrogenases, including very long-chain specific acyl-CoA dehydrogenase (ACADVL), acyl-coenzyme A dehydrogenase for medium-chain fatty acids (ACADM) and hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase, α subunit (HADHA) was also purchased from Abcam. Kits for determination of acetyl-CoA (cat. #K317-100; 24) and ATP (cat. #K354-100; 25) levels were purchased from BioVision (Milpitas, CA). Kits for determination of total free fatty acids (cat. #MAK044; 26) and total phospholipids (cat. #MAK122; 27), and anti-ACC1 antibody for immunohistochemistry (cat. #SAB4501396) were purchased from Sigma-Aldrich (St. Louis, MO).

Cell lines

Monolayer cultures of LNCaP and 22Rv1 cells, purchased from the American Type Culture Collection (Manassas, VA), were maintained as recommended by the supplier. These cells were authenticated by us in 2012 (22Rv1), 2015 (LNCaP) and 2017 (LNCaP and 22Rv1). Each cell line was of human origin and devoid of inter-species contamination. Frozen stocks of the authenticated cells were used in this study. Immortalized mouse embryonic fibroblasts (MEF) from wild-type (WT) and systemic nuclear factor (erythroid-derived 2)-like 2 (Nrf2)-knockout C57BL/6J mice were generously provided by Dr. Thomas W. Kensler and cultured as described by Palliyaguru et al. (28).

Confocal microscopy

LNCaP (2 × 10⁴) or 22Rv1 (3 × 10⁴) cells were plated on coverslips in 24-well plates. After overnight incubation, they were treated with dimethyl sulfoxide (DMSO) or SFN (10 μM) for 24 h. The cells were then fixed and permeabilized with 2% paraformaldehyde and 0.5% Triton X-100, respectively. After blocking with PBS containing 0.5% bovine serum albumin and 0.15% glycine, cells were incubated with anti-ACLY (1:3000 dilution), anti-ACC1 (1:1000 dilution) or anti-FASN (1:5000 dilution) antibody overnight at 4°C. Cells were treated with Alexa Fluor 488-conjugated rabbit antibody (1:2000 dilution) for 1 h at room temperature in the dark. Cells were counterstained with DRAQ5 (nuclear stain) for 5 min at room temperature in the dark prior to microscopic examination. In the case of CPT1A, cells were treated with DMSO or 10 μM SFN for 24 h, and then treated with 100 nM MitoTracker Red in complete medium at 37°C for 1 h to label mitochondria. After fixation with 4% formaldehyde in complete medium at 37⁰C for 15 min, cells were permeabilized with 0.5% Triton X-100 for 5 min and incubated for 1 h in blocking buffer containing 0.5% bovine serum albumin and 0.15% glycine in PBS. Cells were incubated with anti-CPT1A (1:3000) primary antibody (4°C; overnight) followed by treatment with Alexa Fluor 488-conjugated goat anti-mouse secondary antibody for 1 h at room temperature. Cells were examined under a Leica confocal microscope. Corrected total cell fluorescence (CTCF) was quantitated using ImageJ software.

Real-time reverse transcription polymerase chain reaction (RT-PCR)

LNCaP and 22Rv1 cells were plated in culture dishes, allowed to attach and then treated with DMSO or desired doses of SFN for specified time periods. Total RNA was isolated using RNeasy kit. 2 µg RNA was used for cDNA synthesis with the use of SuperScript III reverse transcriptase and oligo (dT)₂₀ primer. Real-time quantitative RT-PCR was performed using 2× SYBR Green master mix (Applied Biosystems–Life Technologies). Primers for ACC1, FASN, CPT1A and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were as follows. Forward (ACC1): 5ʹ-ACCACCAATGCCAAAGTAGC-3ʹ; Reverse (ACC1): 5ʹ-CTGCAGGTTCTCAATGCAAA-3ʹ; Forward (FASN): 5ʹ-CTGGCTCAGCACCTCTATCC-3ʹ; Reverse: (FASN) 5ʹ-CAGGTTGTCCCTGTGATCCT-3ʹ; Forward (CPT1A): 5ʹ-CAAGGACATGGGCAAGTTTT-3ʹ; Reverse: (CPT1A) 5ʹ-AAAGGCAGAAGAGGTGACGA-3ʹ; Forward (GAPDH): 5ʹ-GGACCTGACCTG CCGTCTAGAA-3ʹ; Reverse (GAPDH): 5ʹ-GGTGTCGCTGTTGAAGTCAGAG-3ʹ. The PCR conditions were as follows: 95°C for 10 min followed by 40 cycles of 95°C for 15 s, 60°C for 1 min and 72°C for 30 s. Relative gene expression levels were calculated using the method of Livak and Schmittgen (29).

Western blotting

LNCaP (3.5 × 10⁵) or 22Rv1 (5 × 10⁵) cells were plated in 6 cm dishes, allowed to attach by overnight incubation and then treated with DMSO or desired concentrations of SFN for 8, 16 or 24 h. Details for preparation of cell lysates and prostate tumor supernatants as well as western blotting have been described previously (30). The membranes were stripped and re-probed with anti-β-Actin or anti-GAPDH antibody for normalization of quantitative values. In some experiments, cells were treated with DMSO or 5 µM SFN for 8 h. After treatment, cells were treated with 10 µg/ml cycloheximide for specified time periods prior to preparation of cell lysates and immunoblotting. For the experiment involving proteasomal inhibitor MG132, cells were treated with DMSO or the indicated doses of SFN in the absence or presence of 10 µM MG132 for 8 h followed by western blotting for SREBP1. Protein expression changes were determined by densitometric analysis using UN-SCAN-IT5.1 (Silk Scientific Orem, UT) and corrected for the loading control.

Cell proliferation and apoptosis assay

LNCaP cells (750 cells/well) or 22Rv1 cells (1000 cells/well) were seeded in 96-well plates. After 18 h of incubation, cells were treated with DMSO (control) or the indicated doses of SFN or Etomoxir or Cerulenin for 48 h. Subsequently, 20 µl of manufacturer supplied color development reagent (MTS, Promega, Madison, WI) was added to each well and the plates were incubated at 37°C for 2 h. Absorbance was measured at 492 nm. For apoptosis assay, 22Rv1 (2 × 10⁵/well) or LNCaP (1 × 10⁵/well) cells were seeded in six-well plates and allowed to attach by overnight incubation. Cells were treated with DMSO or the indicated doses of SFN or Cerulenin for 48 h. Apoptosis was determined by flow cytometry as described previously (31).

Quantitation of acetyl-CoA, free fatty acids, total phospholipids and ATP levels

Levels of acetyl-CoA, total free fatty acids, total phospholipids and ATP in the plasma or prostate tumor supernatants of control- and SFN-treated TRAMP mice and in cells were determined using commercially available kits and by following the manufacturer’s protocol. For acetyl-CoA assay, prostate tumors from control- and SFN-treated TRAMP mice were sonicated in acetyl-CoA assay buffer and centrifuged at 10000g for 10 min. Tumor lysates and plasma were filtered using a 10 kDa membrane. Fluorescence was measured at excitation and emission wavelengths of 535 and 587 nm, respectively. For total free fatty acids (C8 and longer) assay, cells or tumor tissues were sonicated in 200 µl of 1% (w/v) Triton X-100 in chloroform. The homogenates were centrifuge at 14 000 rpm for 10 min, and organic phase was collected and dried at 50°C to remove chloroform. The dried lipids were dissolved in fatty acid assay buffer and used for the assay. Plasma was diluted in fatty acid assay buffer and used for the assay. For assaying total phospholipids, prostate tumor was sonicated in phospholipid assay buffer and centrifuged at 10 000 g for 10 min. The supernatant was used for the phospholipid assay. Plasma was directly used for this assay. A colorimetric assay kit was used to measure ATP levels in the tumor by following manufacturer’s protocol.

Flow cytometric analyses of levels of fatty acid oxidation enzyme proteins

Levels of ACADVL, ACADM and HADHA proteins were determined by the flow cytometry using a kit and by following supplier’s instructions.

Immunohistochemistry

Prostate tumor sections (4–5 μm thick) were de-paraffinized, hydrated, washed with PBS and incubated in citrate retrieval buffer solution (pH 6.0) for 20 min at 100°C followed by treatment with 0.3% hydrogen peroxide in 100% methanol for 20 min at room temperature. Sections were treated with blocking buffer for 1 h followed by incubation with primary antibody (anti-ACC1: 1:500 dilution; anti-FASN: 1:500 dilution; anti-CPT1A: 1:1000 dilution; anti-α-SMA: 1:100 dilution and anti-AR: 1:50 dilution) overnight in humidified chambers at room temperature. Sections were washed with PBS and incubated with Alexa Fluor 488- or horseradish peroxidase-conjugated secondary antibody (1:1000 dilution) for 1 h at room temperature. For the immunofluorescence staining sequential labeling was performed and sections were washed PBS after secondary antibody incubation, then overnight incubated with AR primary antibody at 4°C, washed in PBS, incubated in Alexa Fluor 568 conjugated secondary antibody. After washing with PBS, sections were mounted in media containing 4ʹ,6-diamidino-2-phenylindole. For α-SMA and AR, color was developed by incubation with 3,3ʹ-diaminobenzidine tetrahydrochoride and counterstained with hematoxylin. Stained sections were examined under Nikon A1 confocal microscope or Leica microscope equipped with DFC 450C digital camera. Immunohistochemical analysis of ACC1, FASN and CPT1A protein expression intensity was performed using ImageJ software at 16-bit gray scale in the AR co-stained area, and the results are presented as expression of proteins as % of AR positive area.

Visualization and quantitation of neutral lipids in TRAMP tumors

Neutral lipid droplets in prostate tumor sections were visualized using BODIPY^493/503. Tumor sections were de-paraffinized and hydrated. Sections were treated at room temperature with 25 μg/ml BODIPY^493/503 for 1 h, washed with PBS and mounted in medium containing 4ʹ,6-diamidino-2-phenylindole. The sections were examined under Nikon A1 confocal microscope. Number of lipid droplets/cell was quantified using ImageJ software from at least 50 cells per high-power field from five non-overlapping areas of each section.

Statistical analysis

Statistical tests were performed using GraphPad Prism software (version 7.02). One-way analysis of variance (ANOVA) followed by Dunnett’s or Bonferroni’s test was applied for dose–response or multiple comparisons, respectively. Student’s t-test was used for two group comparisons.

Results

SFN decreased ACC1, FASN and CPT1A expression in LNCaP and 22Rv1 cells

We used two different human prostate cancer cell lines (LNCaP and 22Rv1) to determine the effect of SFN treatment on expression of enzyme proteins responsible for fatty acid synthesis. LNCaP is an androgen-responsive cell line expressing mutant AR (T877A) and mutant phosphatase and tensin homolog, whereas 22Rv1 cell line expresses wild-type phosphatase and tensin homolog but mutant AR as well as its splice variants. Figure 1A shows representative microscopic images for ACLY, ACC1, FASN and CPT1A protein expression in control- and SFN-treated LNCaP cells. The protein level of ACLY, which is responsible for conversion of citrate to acetyl-CoA (10,11), was not affected by SFN treatment in either LNCaP (Figure 1A) or 22Rv1 cells (Figure 1B). To the contrary, protein level of ACC1, which is the rate-limiting enzyme in the de novo fatty acid synthesis and responsible for conversion of acetyl-CoA to malonyl-CoA (10,11), was decreased significantly upon SFN treatment in both LNCaP and 22Rv1 cell lines (Figure 1A and B). One molecule of acetyl-CoA and seven molecules of malonyl-CoA are utilized to produce fatty acids through a series of catalytic domains of FASN complex (10). Exposure of both prostate cancer cells to SFN resulted in suppression of FASN protein level (Figure 1A and B).

Figure 1.

SFN treatment decreased ACC1, FASN and CPT1A protein and mRNA expression in LNCaP and 22Rv1 cells. (A) Confocal microscopy images depicting expression of ACLY, ACC1, FASN and CPT1A proteins in LNCaP cells after 24 h treatment with DMSO (control) or 10 µM SFN. Similar results were observed in 22Rv1 cells (microscopy data not shown). Nuclei and mitochondria were visualized by staining with DRAQ5 and Mitotracker Red, respectively. (B) Quantitation of corrected total cell fluorescence (CTCF) for ACLY, ACC1, FASN and CPT1A expression using ImageJ software. Results shown are mean ± SD (n = 20). Statistically significant (*P < 0.05) compared with DMSO-treated control by Student’s t-test. (C) Real-time quantitative RT-PCR for ACC1, FASN and CPT1A mRNA expression in LNCaP and 22Rv1 cells after 8, 16 or 24 h treatment with DMSO or 5 and 10 µM SFN. Results shown are mean ± SD (n = 3). Significantly different (*P < 0.05) compared with corresponding DMSO-treated control by one-way ANOVA followed by Dunnett’s test. Results were reproducible in replicate experiments.

The CPT1A is associated with the outer mitochondrial membrane and plays an important role in fatty acid uptake by mitochondria for β-oxidation. Merging of the CPT1A-associated green fluorescence and MitoTracker Red signal resulting in yellow-orange fluorescence confirmed mitochondrial localization of CPT1A in DMSO-treated control cells, which was decreased significantly in SFN-treated LNCaP and 22Rv1 cells (Figure 1A and B). Collectively, these results indicated downregulation of ACC1, FASN and CPT1A protein expression after SFN treatment in human prostate cancer cell lines. SFN treatment also resulted in downregulation of ACC1, FASN and CPT1A mRNA expression in both cell lines (Figure 1C).

Next, western blotting was performed to determine the dose–response and time-course kinetics of SFN-mediated downregulation of the above proteins (Figure 2A). Protein level of ACLY was either marginally increased or not affected in SFN-treated cells in comparison with control (Figure 2A). A decrease in ACLY protein level visible in the immunoblot in Figure 2A (16 h after SFN treatment in the LNCaP cell line or 24 h after SFN treatment in the 22Rv1 cell line) was not consistent in replicate experiments. However, consistent with microscopic analyses, the levels of ACC1, FASN and CPT1A proteins were dose-dependently decreased after SFN treatment in both cell lines (Figure 2B). The level of ACC2 protein was also decreased upon SFN treatment in the LNCaP cell line but not in 22Rv1 (Figure 2B). Next, we determined the role of Nrf2 in SFN-mediated suppression of CPT1A protein level using immortalized MEF from WT and Nrf2-disrupted mice (28). Lack of Nrf2 expression in knockout MEF was confirmed by immunoblotting (Supplementary Figure 1, available at Carcinogenesis Online). Basal CPT1A expression was minimal in Nrf2-disrupted MEF implicating Nrf2 in regulation of CPT1A expression. SFN treatment caused induction of Nrf2 protein in WT MEF. At the same time, protein level of CPT1A was decreased by about 50 and 80% upon 24 h treatment of WT MEF with 5 and 10 μM SFN, respectively (Supplementary Figure 1, available at Carcinogenesis Online). These results indicated that SFN-mediated downregulation of CPT1A was not related to Nrf2.

Figure 2.

Dose-response and time course kinetic effect of SFN treatment on expression of fatty acid metabolism proteins in LNCaP and 22Rv1 cells. (A) Immunoblotting for ACLY, ACC1, ACC2, FASN and CPT1A proteins using lysates from LNCaP and 22Rv1 cells. The blots were stripped and re-probed with anti-β-Actin or anti-GAPDH antibody as a loading control. The arrow indicates correct size for ACC2 protein. (B) Densitometric quantitation of ACC1, ACC2, FASN and CPT1A protein expression from three independent experiments. Significantly different (*P < 0.05) compared with corresponding DMSO-treated control by one-way ANOVA followed by Dunnett’s test.

Comparison of SFN with other fatty acid metabolism inhibitors

Cerulenin (an antifungal agent that also inhibits fatty acid and steroid biosynthesis) and Etomoxir (an irreversible inhibitor of CPT1) were compared with SFN for their ability to inhibit cell proliferation (Figure 3A), induce apoptosis (Figure 3B) or decrease cellular total free fatty acid level (Figure 3C). Cell proliferation was not affected by Etomoxir in either LNCaP or 22Rv1 cells except at 10 μM dose in 22Rv1 cells (Figure 3A). SFN and Cerulenin exhibited more or less comparable antiproliferative effect. Interestingly, SFN was a relatively more potent apoptosis inducer in comparison with Cerulenin in both LNCaP and 22Rv1 cells (Figure 3B). In LNCaP cells, but not in 22Rv1, decrease in levels of total free fatty acids was relatively more pronounced with SFN when compared with Cerulenin (Figure 3C).

Figure 3.

SFN treatment decreased levels of total free fatty acids and expression of fatty acid β-oxidation dehydrogenases in LNCaP and 22Rv1 cells. (A) Effects of SFN, Cerulenin and Etomoxir treatment (48 h) on proliferation of LNCaP and 22Rv1 cells. Experiment was repeated twice in triplicate and representative data from one such experiment are shown as mean ± SD (n = 3). Significantly different (*P < 0.05) compared with DMSO-treated control, and #between SFN treatment group and Cerulenin or Etomoxir group by one-way ANOVA followed by Bonferroni’s test. (B) Total apoptosis induced by SFN and Cerulenin in LNCaP and 22Rv1 cells after 48 h of treatment. Results shown are mean ± SD. Experiment was repeated twice in triplicate and representative data from one such experiment are shown as mean ± SD (n = 3). Significantly different (*P < 0.05) compared with DMSO-treated control, and #between SFN and Cerulenin groups by one-way ANOVA followed by Bonferroni’s test. (C) Total free fatty acid levels in LNCaP or 22Rv1 cells after 16 h treatment with DMSO or the indicated doses of SFN or Cerulenin. Results shown are mean ± SD (n = 3). Significantly different (*P < 0.05) compared with DMSO-treated control, and #between SFN and Cerulenin groups by one-way ANOVA followed by Bonferroni’s test. (D) Bar graphs depicting the flow cytometric analysis for the expression of fatty acid β-oxidation dehydrogenases after 16 h treatment with DMSO or SFN in LNCaP and 22Rv1 cells. Results shown are mean ± SD (n = 3). *Significantly different (P < 0.05) compared with corresponding DMSO-treated control by one-way ANOVA followed by Dunnett’s test.

SFN treatment decreased expression of β-oxidation dehydrogenases

The levels of fatty acid β-oxidation enzymes ACADVL (catalyzes formation of a C2-C3 trans-double bond in the very long-chain fatty acids typically C16-acyl-CoA and longer), ACADM (plays an important role in the breakdown of C-4 to C-12 straight chain fatty acids) and HADHA (catalyzes the last three steps of mitochondrial β-oxidation of fatty acids) were decreased after SFN treatment in both LNCaP and 22Rv1 cells (Figure 3D). These results suggested that SFN may inhibit β-oxidation of fatty acids.

SFN administration significantly decreased ACC1 and FASN protein levels in vivo

Next, we compared the levels of ACC1, FASN and CPT1A proteins in prostate adenocarcinoma sections of control- and SFN-treated TRAMP mice by immunohistochemistry after co-staining with AR to determine the in vivo relevance of the cellular observations. Prostate sections from TRAMP mice with PIN and adenocarcinoma, which was confirmed by histopathological analysis (2,3), were first stained for a stromal (α-SMA) and a tumor epithelial (AR) marker (Figure 4A). The α-SMA staining was observed in stroma surrounding PIN but not in adenocarcinoma. On the other hand, AR expression was observed in both PIN and adenocarcinoma. Representative immunohistochemical images for ACC1, FASN and CPT1A protein expression with co-staining for AR in prostate adenocarcinoma sections of control- and SFN-treated TRAMP mice are shown in Figure 4B. Expression of ACC1, FASN and CPT1A was primarily cytoplasmic but AR localization was observed in both cytoplasmic and nuclear compartments. The ACC1, FASN and CPT1A expression intensity was computed as % of AR positive area. Immunohistochemistry revealed a statistically significant decrease in ACC1 and FASN protein expression in prostate adenocarcinoma of SFN-treated TRAMP mice when compared with controls (Figure 4C). SFN administration also resulted in downregulation of CPT1A protein (Figure 4B) but the difference was not significant at P < 0.05 (Figure 4C). These results provided evidence for in vivo downregulation of at least ACC1 and FASN proteins in prostate tumors upon SFN administration to TRAMP mice.

Figure 4.

Effect of SFN treatment on the expression of fatty acid metabolism proteins in the prostate adenocarcinoma of TRAMP mice. (A) Representative immunohistochemical images for α-SMA and AR expression in PIN and adenocarcinoma lesions of TRAMP mice (20× objective magnification; scale bar = 50 μm). (B) Representative images depicting immunohistochemical determinations of ACC1, FASN and CPT1A protein expression (green fluorescence) co-stained with AR (red fluorescence) in the prostate adenocarcinoma of control- and SFN-treated TRAMP mice (40× oil optical magnification, scale bar = 50 μm). (C) Bar graphs showing quantitation of ACC1, FASN and CPT1A protein expression intensity as % of AR positive area (n = 5 for both groups). Expression intensity of ACC1, FASN and CPT1A protein was quantified using ImageJ software at 16-bit gray scale in the AR co-stained area, and the results are presented as expression of proteins in % of AR positive area. Statistically significant (*P < 0.05) compared with control by Student’s t-test. The term ‘ns’ signifies not significantly different.

SFN administration caused in vivo suppression of acetyl-CoA, total free fatty acids and total phospholipids levels in the plasma and/or prostate adenocarcinoma of TRAMP mice

Prostate cancer prevention by SFN in TRAMP mice was also associated with a significant decrease in plasma and/or prostate adenocarcinoma levels of acetyl-CoA (Figure 5A), total free fatty acids (Figure 5B), total phospholipids (Figure 5C) and ATP (Figure 5D). Figure 5E depicts BODIPY staining for neutral lipids in a representative prostate tumor section of a control- and a SFN-treated TRAMP mouse. The number of neutral lipid droplets/cell was significantly lower in the prostate adenocarcinoma sections of SFN-treated mice when compared with controls (Figure 5F).

Figure 5.

SFN administration decreased fatty acid metabolism intermediates in the plasma and/or prostate adenocarcinoma of TRAMP mice. Levels of acetyl-CoA (A), total free fatty acids (B), total phospholipids (C) and ATP (D) in the plasma and prostate adenocarcinoma of control- and SFN-treated TRAMP mice. Results shown are mean ± SD (n = 6 for both groups). Plasma and prostate tumor samples from different mice of each group were used for determination of metabolite levels. Statistically significant (*P < 0.05) by Student’s t-test. (E) Representative confocal microscopy images for BODIPY staining depicting neutral lipid droplets in representative prostate adenocarcinoma of a control mouse and a mouse of the SFN treatment group. (F) Quantitation of number of lipid droplets/cell in the prostate adenocarcinoma of control- and SFN-treated TRAMP mice. Results shown are mean ± SD (n = 5). Significantly different (*P < 0.05) compared with control by Student’s t-test.

SFN treatment decreased SREBP1 protein level in prostate cancer cells

Because SREBP1 plays an important role in regulation of fatty acid synthesis (32), we determined the effect of SFN on its protein levels using LNCaP and 22Rv1 cells. As can be seen in Figure 6A and B, SREBP1 protein level was decreased significantly upon 24 h treatment with 10 μM SFN in both cell lines. The basal expression of SREBP1 protein seemed lower at 16 h and 24 time points relative to that at 8 h time period (Figure 6A). We therefore raised the question of whether expression of SREBP1 protein was affected by cell confluency. To address this question, LNCaP or 22Rv1 cells were plated at 17.5, 35 and 70% confluency and allowed to attach by incubation for 16 h. Subsequently, the cells were harvested at 8, 16 and 24 h time points and processed for immunoblotting. The expression of SREBP1 protein was variable at 16 and 24 h time points but the difference was not statistically significant when compared with 8 h time point in both cell lines (data not shown; n = 4 independent experiments). A role for ChREBP in regulation of lipid synthesis has also been suggested (33), and its protein level was decreased significantly after SFN treatment in the LNCaP cell line but not in 22Rv1 (Figure 6A and B). Western blotting also revealed that the half-life of SREBP1 protein was not affected by SFN treatment in either 22Rv1 (Figure 6C and D) or LNCaP cell line (data not shown). On the other hand, SFN-mediated downregulation of SREBP1 protein was partially attenuated by treatment with a proteasomal inhibitor MG132 (Figure 6E).

Figure 6.

SFN treatment decreased SREBP1 protein expression in LNCaP and 22Rv1 cells. (A) Immunoblotting for SREBP1 and ChREBP proteins using lysates from LNCaP or 22Rv1 cells. The blots were stripped and re-probed with anti-β-Actin antibody as a loading control. Representative blots from replicate experiments for each protein are shown. (B) Densitometric quantitation of SREBP1 and ChREBP protein expression. Results shown are mean ± SD (n = 3 independent experiments). Significantly different (*P < 0.05) compared with corresponding DMSO-treated control by one-way ANOVA followed by Dunnett’s test. (C) Immunoblotting for SREBP1 and β-actin proteins using lysates from 22Rv1 cells after treatment with cycloheximide for indicated time points. (D) Quantitation of the effect of SFN treatment on SREBP1 protein stability (half-life) in 22Rv1 cells (cycloheximide treatment). Combined results from three independent experiments are presented as mean ± SD (n = 3). (E) Immunoblotting for SREBP1 and β-actin proteins using lysates from 22Rv1 cells after 8 h treatment with DMSO or SFN in the absence or presence of 10 µM MG132. Experiment was done two times and the results were consistent. (F) A cartoon summarizing key steps in fatty acid metabolism and their inhibition by SFN.

Discussion

The de novo fatty acid synthesis is very low in most normal adult human tissues with the exception of some organs including liver, adipose tissue and brain (10,11). Fatty acid demand in normal tissues with low de novo synthetic capacity is mainly met from dietary sources (10,11). Fatty acid synthesis is elevated in neoplastic cells, a phenomenon recognized in early fifties (34). In fact, fatty acid oxidation is a dominant bioenergetic pathway in prostate cancer (12). Functions of fatty acids in cancerous cells include membrane biogenesis, protection from oxidative stress and activation of oncogenes through post-translational modifications (e.g. palmitoylation) (35,36). A fatty acid synthesis inhibitor (C75) was shown to prevent breast cancer development in a transgenic mouse model (37). As summarized by a simplified cartoon in Figure 6F, the present study is the first to provide in vitro and in vivo evidence for SFN-mediated inhibition of de novo fatty acid synthesis in prostate cancer. Inhibition of de novo fatty acid synthesis by SFN is reflected by a decrease in plasma and/or prostate adenocarcinoma levels of fatty acid synthesis precursor acetyl-CoA, total free fatty acids and total phospholipids. The protein levels of dehydrogenases implicated in β-oxidation of fatty acids are also decreased upon SFN treatment at least in cultured cancer cells.

The present study reveals that prostate cancer chemoprevention by SFN is associated with suppression of FASN at least in the TRAMP model. FASN is considered an oncogene for prostate cancer. Expression profiling of 20 invasive human prostate cancers and 10 high-grade PIN by cDNA microarray representing 23 040 genes revealed upregulation of FASN along with other genes in transition from PIN to invasive cancer (38). Shah et al. (39) also observed overexpression of FASN in prostate adenocarcinoma. Overpression of FASN in the prostate of transgenic mice under the control of prostate-specific ARR2 probasin promoter caused PIN that did not progress to invasive cancer (22). Genetic as well as pharmacological suppression of FASN inhibits growth of LNCaP cells in vitro and in vivo (19). In addition, stable knockdown of FASN in prostate cancer cells resulted not only in cell proliferation inhibition but also suppression of cell adhesion, migration and invasion suggesting a role for FASN in tumor metastasis (40). In this context, we have shown previously that spontaneous pulmonary metastasis in TRAMP mice is also inhibited by SFN administration (2). It is reasonable to postulate that antimetastatic effect of SFN may, at least in part, be linked to its inhibitory effect on FASN.

ACC1 is the rate limiting enzyme in de novo synthesis of fatty acids and its expression is decreased by SFN treatment in both cultured prostate cancer cells in vitro and in TRAMP adenocarcinoma in vivo. Inhibition of ACC1 with 5-tetradecyl-oxy-2-furoic acid resulted in caspase-dependent apoptotic cell death in prostate cancer cells that was not affected by p53 (41). Apoptosis induction in prostate cancer cells after SFN treatment has been documented (42,43). Apoptosis induction by SFN in prostate cancer cells is mediated by reactive oxygen species that are generated due to inhibition of mitochondrial electron transport chain (42,43). It is possible that ACC1 downregulation contributes to proapoptotic effect of SFN.

Prior efforts in clinical development of FASN inhibitors have been hampered due to a variety of reasons, including stability (Cerulenin), poor bioavailability (Orlistat) and side effects such as appetite suppression and weight loss through direct activation of CPT1 (44). On the other hand, SFN has an excellent oral bioavailability and is well-tolerated in rodent and human studies (2,3,7–9). We were intrigued by our findings of CPT1A suppression by SFN treatment in both LNCaP and 22Rv1 cell lines (Figures 1 and 2). Alternate mechanisms have been implicated in regulation of CPT1A expression (45) that may explain its downregulation by SFN treatment in LNCaP and 22Rv1 cells. Data in WT and Nrf2-disrupted MEF further support this possibility. However, further work is necessary to test this possibility.

A few studies have indicated negative regulation of ACLY, ACC1 and FASN by the transcription factor nuclear factor (erythroid-derived 2)-like 2 (Nrf2) in the liver of mice (46–48). However, a role for Nrf2 in regulation of fatty acid metabolism in normal or cancerous prostate is yet to be established. Because SFN is a known inducer of Nrf2, it is possible that Nrf2 plays a role in downregulation of fatty acid synthesis enzymes by SFN treatment. The present study reveals that the expression of ACLY protein or mRNA is not affected by treatment with SFN (Figures 1 and 2) suggesting an Nrf2-independent mechanism. On the other hand, protein levels of ACC1 and FASN are decreased significantly upon SFN exposure even at a time point when cell viability is minimally affected. We have shown previously that roughly 90% of LNCaP cells are viable upon 8 h treatment with 10 μM SFN (49). The protein levels of ACC1 and FASN are decreased by 40–60% by a similar SFN treatment in LNCaP cells (Figure 2A and B). Collectively, these results indicate that inhibition of fatty acid synthesis by SFN is likely not due to its toxic effect on the cells.

Tissue bioavailability is an important consideration for a chemopreventive intervention. In this context, SFN and its metabolites were detectable in mouse prostate (about 0.05–0.10 nmol/mg tissue) at 2 and 6 h after gavage with 5 μmol (we used 6 μmol SFN in our TRAMP study) and 20 μmol SFN, respectively (50). In addition, accumulation of SFN and its metabolites in the mouse prostate was highest after small intestine in comparison with other organs including kidney, lung, colon, liver and brain (50). Residual levels of SFN and its metabolites were detectable in the mouse prostate even after 24 h post-treatment (50). These results are encouraging but the bioavailability of SFN or its metabolites in human prostate/tumor after oral administration is yet to be determined.

In conclusion, the present study demonstrates that prostate cancer prevention by SFN in a preclinical mouse model is associated with suppression of fatty acid synthesis and possibly their β-oxidation. Our data also suggest that ACC1 and FASN are novel mechanistic targets of prostate cancer prevention by SFN at least in the TRAMP model.

Supplementary material

Supplementary materials can be found at Carcinogenesis online.

Funding

This study was supported by the NCI grants CA115498 and CA101753. This research used the Animal Facility, Flow Cytometry Facility and the Tissue and Research Pathology Facility supported in part by a grant from the National Cancer Institute at the National Institutes of Health (P30 CA047904; Dr. Robert L. Ferris, Principal Investigator).

Conflict of Interest Statement: None declared.

Abbreviations

ACADM

acyl-CoA dehydrogenase for medium-chain fatty acids

ACADVL

very long-chain specific acyl-CoA dehydrogenase

ACC1

acetyl-CoA carboxylase 1

ACC2

acetyl-CoA carboxylase 2

ACLY

ATP citrate lyase

ANOVA

analysis of variance

AR

androgen receptor

ChREBP

carbohydrate response element binding protein

CPT1A

carnitine palmitoyltransferase 1A

CTCF

corrected total cell fluorescence

DMSO

dimethyl sulfoxide

FASN

fatty acid synthase

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

HADHA

hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase, α subunit

MEF

mouse embryonic fibroblasts

Nrf2

nuclear factor (erythroid-derived 2)-like 2

PBS

phosphate-buffered saline

PIN

prostatic intraepithelial neoplasia

PSA

prostate-specific antigen

RT-PCR

reverse transcriptase polymerase chain reaction

α-SMA

α-smooth muscle actin

SREBP1

sterol regulatory element-binding protein 1

SFN

D,L-sulforaphane

TRAMP

TRansgenic Adenocarcinoma of Mouse Prostate

WT

wild type