Requirement and Epigenetics Reprogramming of Nrf2 in Suppression of Tumor Promoter TPA-Induced Mouse Skin Cell Transformation by Sulforaphane

An increase in oxidative stress may result in genomic instability, genetic mutation, and neoplastic transformation, increasing the incidence of carcinogenesis (1). Endogenous reactive oxygen species (ROS) can be generated during normal cellular metabolism, immune reactions, and several pathologic conditions, including inflammation, whereas exogenous ROS may originate from air pollution, UV irradiation, microorganisms, viruses, and xenobiotics. Therefore, skin is a major organ that is subjected to exogenous ROS and endogenous oxidative stress, which may trigger skin carcinogenesis (2). Nuclear factor erythroid-2-related factor-2 (Nrf2 or NFE2L2) is a basic helix-loop-helix leucine zipper transcription factor that plays a key role in reducing cellular oxidative stress by regulating the defense system (3, 4). Nrf2 itself is regulated by Kelch-like ECH-associating protein 1 (Keap1), which serves as an adaptor protein that bridges Nrf2 and Cul3, resulting in the ubiquitination of Nrf2 (5). By binding to an antioxidant response element (ARE) in the promoter region of a target gene, nuclear Nrf2 induces the expression of antioxidative stress/detoxifying enzymes, such as heme oxygenase-1 (HO-1), NAD(P)H:quinone oxidoreductase-1 (NQO-1), UDP-glucuronosyltransferase (UGT), and glutathione-S-transferase (GST; 6, 7). It has been reported that the susceptibility to tumorigenesis induced by a carcinogen is higher in Nrf2-deficient mice (8, 9), and cancer chemoprevention is partially correlated with the induction of phase II enzymes (10, 11). Our recent study demonstrated that the expression of Nrf2 can be regulated by epigenetic alterations in both the prostate tissue of the transgenic adenocarcinoma of the mouse prostate (TRAMP) model and tumorigenic TRAMP C1 cells (12).

In addition to genetic changes, accumulating evidence suggests that carcinogenesis is regulated by epigenetic alterations in tumor suppressor genes and oncogenes, defined as gene expression that can be regulated without the alteration of DNA sequences (13, 14). DNA methylation and histone modification, managed by DNA methyltransferases (DNMT) and histone deacetylases (HDAC), respectively, can maintain cellular DNA stability and integrity and are 2 major epigenetic mechanisms for the transcriptional control of gene promoters. Drugs targeting the enzymes responsible for epigenetic silencing are continually being researched and developed as cancer chemopreventive/therapeutic agents, including 5-azadeoxycytidine (5-aza, a DNMT inhibitor) and trichostatin A (TSA, an HDAC inhibitor; refs. 15 and 16). However, owing to the side effects of such drugs, many dietary phytochemicals have recently been investigated for their cancer preventive potential and have been shown to induce epigenetic alterations (17, 18). For example, epigallocatechin-3-gallate (EGCG) from green tea has been reported to be an inhibitor of DNMTs and histone acetyltransferases (HAT; refs. 19 and 20). Our previous studies also demonstrate that curcumin from Curcuma longa and Z-ligustilide from Radix Angelica sinensis might induce the expression of Nrf2 in the TRAMP C1 cell line via an inhibitory effect on the activity of DNMTs (21, 22).

Sulforaphane (SFN) is a natural isothiocyanate found in cruciferous vegetables, including broccoli, cabbage, cauliflower, Chinese cabbage, and watercress. The cancer chemopreventive potential of SFN has been investigated in numerous animal models of different cancers, and one of the major mechanisms of SFN is to induce several phase II antioxidant and detoxification enzymes to block electrophilic chemicals that cause DNA damage, mutagenesis, and carcinogenesis (23). It has been reported that the topical application of an SFN-containing broccoli sprout extract can inhibit UVB-induced skin carcinogenesis in SKH-1 mice (24). SFN can also suppress skin tumorigenesis caused by 7,12-dimethylbenz(a)anthracene (DMBA)/12-O-tetradecanoylphorbol-13-acetate (TPA) in C57BL/6 mice, an effect that is potentially mediated through the regulation of Nrf2 (25). SFN also protects mouse skin from mustard gas analog 2-(chloroethyl) ethyl sulfide–induced mutagenesis through the induction of phase II detoxification enzymes (26). Moreover, these results suggest that epigenetic modification, including the inhibition of HDAC activity, may contribute to the multitargeted chemopreventive activity of SFN (27). Our recent study also found that SFN reactivates Nrf2 in TRAMP C1 cells, a result that involves a decrease in DNMT and HDAC expression and gene promoter demethylation (28). However, the molecular mechanisms by which SFN inhibits the skin carcinogenesis promoted by tumor promoters such as TPA, which may be mediated by epigenetic alteration, have not yet been clearly described. In this study, we investigated whether SFN can inhibit neoplastic transformation induced by TPA, which promotes skin tumor development, in cultured mouse skin epidermal JB6 P+ cells. Furthermore, the possible molecular mechanisms, including Nrf2-mediated pathways and epigenetic modification, were also investigated to aid in the understanding of the chemopreventive effect of SFN against skin carcinogenesis.

SFN was purchased from LKT Laboratories, Inc. 5-Aza, bacteriological agar, Eagle's basal medium (BME), puromycin, 12-O-tetradecanoylphorbol-13-acetate (TPA), and TSA were obtained from Sigma-Aldrich. FBS, minimum essential medium (MEM), and trypsin-EDTA solution were purchased from Gibco Laboratories.

JB6 P+ (JB6 Cl 41-5a, ATCC CRL-2010) cells from American Type Culture Collection were maintained in MEM containing 5% FBS at 37°C in a humidified 5% CO₂ atmosphere. JB6 P+ cells stably transfected with shMock- and shNrf2-knockdown were maintained in MEM supplemented with 5% FBS and 2 μg/mL puromycin. After incubation for 24 hours, the cells were treated with various concentrations of SFN or 5-aza (250 nmol/L) in MEM containing 1% FBS for 5 days. The medium was changed every 2 days. For the combination treatment of 5-aza and TSA, TSA (50 nmol/L) was added to the medium on day 4. The treated cells were harvested on day 5 for further assays.

Protein supernatant fractions from whole-cell lysates were prepared from the treated cells using radioimmunoprecipitation assay buffer (Cell Signaling) supplemented with a protease inhibitor cocktail (Sigma). For preparation of the nuclear protein fraction, the NEPER Nuclear and Cytoplasmic Protein Extraction Kit (Thermo Scientific) was used according to the manufacturer's instructions. The protein concentrations were determined using the Bicinchoninic Acid (BCA) Kit (Pierce).

Proteins separated by 4% to 15% SDS-PAGE (Bio-Rad) were transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore). After blocking with 5% BSA in Tris-buffered saline-0.1% Tween 20 buffer, the membrane was sequentially incubated with specific primary antibodies and horseradish peroxidase–conjugated secondary antibodies. The SuperSignal enhanced chemiluminescence detection system and Gel Documentation 2000 system (Bio-Rad) were used to detect and record the antibody-bound proteins on the membrane, and the densitometry of the bands was analyzed using the ImageJ program (Version 1.40g; NIH). The primary anti-Nrf2, anti-HO-1, anti-NQO-1, anti-UGT1A1, anti-β-actin, and anti-lamin A antibodies were obtained from Santa Cruz Biotechnology. The primary anti-DNMT (DNMT1, DNMT3a, and DNMT3b) antibodies were obtained from IMGENEX and the primary anti-HDAC (HDAC1, HDAC2, HDAC3, and HDAC4) antibodies were obtained from Cell Signaling.

Total RNA was extracted from the treated cells using TRizol reagent (Invitrogen) and the RNeasy Mini Kit (Qiagen). SuperScript III First-Strand cDNA Synthesis System (Invitrogen) was further used to synthesize first-strand cDNA from total RNA according to the manufacturer's instructions. The mRNA expression of specific genes was then determined by quantitative real-time PCR (qPCR) using the total first-strand cDNA as the template and Power SYBR Green PCR Master Mix (Applied Biosystems). The primer pairs used were described previously (22), and the β-actin mRNA expression level was utilized as an internal loading control.

The pretreated cells (8 × 10³/well) were transferred to 1 mL of BME containing 0.33% agar over 3 mL of BME containing 0.5% agar with 10% FBS in 6-well plates. The cells were maintained with TPA (20 ng/mL) alone or a combination of TPA (20 ng/mL) and various concentrations of SFN in a 5% CO₂ incubator at 37°C for 14 days. The cell colonies in soft agar were photographed using a computerized microscope system with the Nikon ACT-1 program (Version 2.20; LEAD Technologies) and counted using the ImageJ program (Version 1.40g; NIH).

Genomic DNA was isolated from the treated cells using the QIAamp DNA Mini Kit (Qiagen). Bisulfite conversion of genomic DNA was performed using the EZ DNA Methylation Gold Kit (Zymo Research Corp.) according to the manufacturer's instructions, as described previously (12). The DNA fragment containing the first 15 CpGs, which are located between −863 and −1226 in the murine Nrf2 gene with the translation start site defined as position +1, was amplified from the converted DNA by PCR using Platinum Taq DNA polymerase (Invitrogen). The primer sequences were as follows: sense, 5′-AGT TAT GAA GTA GTA GTA AAA A-3′; anti-sense, 5′-ACC CCA AAA AAA TAA ATA AAT C-3′. The PCR products were further cloned into a pCR4 TOPO vector using the TOPO TA Cloning Kit (Invitrogen). At least 10 colonies from each treatment group were randomly selected; the plasmids were prepared using a QIAprep Spin Miniprep Kit (Qiagen) and analyzed by sequencing (GeneWiz).

An HDAC activity assay was performed using Epigenase HDAC Activity/Inhibition Direct Assay Kit (Colorimetric; Epigentek) according to the manufacturer's instructions. The nuclear protein fraction from the treated cells was used, and the relative HDAC activity was calculated based on the ratio of the HDAC activity of the SFN treatment group to that of the control group.

The data are represented as the mean ± SD. The statistical analyses were performed using an ANOVA followed by Student t test. Statistically significant differences among the means were set at P < 0.05 and P < 0.01.

After 5 days of treatment, SFN at 2.5 and 5.0 μmol/L significantly increased the nuclear translocation of the Nrf2 protein in JB6 P+ cells compared with the control group (Fig. 1A). Nrf2 plays an important role in regulating the expression of phase II detoxifying and antioxidative stress enzymes (29). The expression of enzymes regulated by Nrf2 in the JB6 P+ cells treated with SFN for 5 days was measured using qPCR, and the results showed that SFN treatment increased the mRNA expression of HO-1, NQO1, and UGT1A1 in a concentration-dependent manner (Fig. 1B). The protein levels of Nrf2, HO-1, NQO1, and UGT1A1 in the JB6 P+ cells treated with SFN were further evaluated by Western blotting. SFN also increased the protein expression of Nrf2, HO-1, and NQO1 in a concentration-dependent manner; however, UGT1A1 expression was only slightly increased (Fig. 1C). These results suggest that SFN has the potential to increase the Nrf2-mediated mRNA and protein expression of antioxidant and detoxifying enzymes, which might be correlated with the increased cellular expression and nuclear translocation of Nrf2 in JB6 P+ cells.

The effect of SFN treatment on the TPA-induced anchorage-independent growth of JB6 P+ cells was examined in soft agar (Fig. 2). The JB6 P+ cells were pretreated with SFN for 5 days and subsequently incubated with TPA with or without SFN for an additional 14 days for the induction of transformation. The results showed that pretreatment with 2.5 or 5.0 μmol/L SFN significantly suppressed the TPA-induced colony formation of the JB6 P+ cells by approximately 55% (P < 0.05) and 66% (P < 0.01), respectively. When incubated without SFN in soft agar containing TPA for an additional 14 days, SFN (5.0 μmol/L) pretreatment also significantly decreased the number of JB6 P+ colonies compared with the TPA-treated control (P < 0.01). These results indicate that SFN has novel chemopreventive potential against TPA-induced carcinogenesis in JB6 P+ cells.

The efficiency of short hairpin RNA (shRNA) knockdown was examined, as shown in Fig. 3, and the protein expression of Nrf2 was significantly decreased after transfection of JB6 cells with shNrf2 in the absence of SFN treatment (P < 0.01). However, application of 5.0 μmol/L SFN significantly induced Nrf2 and HO-1 protein expression compared with the control JB6-shMock cells (P < 0.05); in contrast, treatment with 1.0 to 5.0 μmol/L SFN only caused a slight increase in Nrf2 and HO-1 expression in the JB6-shNrf2 cells. The induced levels of NQO1 protein in the JB6-shMock cells were higher than those in the JB6-shNrf2 cells after treatment with SFN (1.0–5.0 μmol/L). These results indicate that SFN might inhibit cell transformation in JB6 P+ cells through the upregulation of cellular Nrf2, resulting in an increase in the protein levels of Nrf2 downstream genes, including HO-1 and NQO1.

A stable Nrf2-knockdown JB6 P+ cell line was established using a shRNA to evaluate whether the resistance of the cells to TPA-induced cell transformation is directly correlated with the status of Nrf2 in the cells. Knocking down Nrf2 significantly increased the growth of the JB6-shNrf2 cells in soft agar compared with the JB6-shMock cells (P < 0.01; Fig. 4). Moreover, pretreatment with SFN (1.0–5.0 μmol/L) significantly inhibited the TPA-induced anchorage-independent growth of the JB6-shMock cells by approximately 28% to 56%. In contrast, this SFN-mediated inhibition was reduced in the JB6-shNrf2 cells, indicating that the protective effect of SFN in JB6-shNrf2 cells was decreased. These results suggest that Nrf2 plays an important role in the preventive effect of SFN on TPA-induced JB6 P+ cell transformation.

The expression of Nrf2 and its downstream genes has previously been shown to be decreased by SFN in the tumorigenic TRAMP C1 cell line via the epigenetic activation of the Nrf2 gene (28). In this study, JB6 P+ cells were pretreated with SFN for 5 days and subsequently incubated with TPA for an additional 6 hours. The mRNA expression level of Nrf2 was then measured. The results indicated that pretreatments with both 2.5 and 5.0 μmol/L SFN increased the ratio of relative Nrf2 mRNA expression (Fig. 5A). To examine whether SFN pretreatment caused demethylation of the promoter region resulting in increased Nrf2 transcription, bisulfite sequencing was performed to investigate the methylation status of the first 15 CpGs of Nrf2, similar to our previous study in TRAMP C1 cells (12). Figure 5B shows that the 15 CpGs were hypermethylated in the control group (methylation ratio, 88.0%). The ratio of methylated CpGs was decreased to 68.7% by SFN (2.5 μmol/L), whereas 5-aza (250 nmol/L)/TSA (50 nmol/L) reduced the methylation level to 50.0%. These data suggest that SFN can demethylate the Nrf2 gene promoter in JB6 P+ cells.

The effects of SFN on epigenetic modifying enzymes, including DNMTs and HDACs, were further examined to explore the epigenetic mechanism of SFN in promoter demethylation and the induction of Nrf2 gene transcription. We found that SFN (1.0–5.0 μmol/L) decreased the protein expression of DNMT1, DNMT3a, and DNMT3b in a concentration-dependent manner in JB6 P+ cells after 5 days of treatment (Fig. 6). In addition, treatment with either 2.5 or 5.0 μmol/L SFN significantly inhibited the relative HDAC activity by 50% (P < 0.01; Fig. 7A), which might be because of the reduced level of HDAC1, HDAC2, HDAC3, and HDAC4 protein expression caused by SFN treatment (Fig. 7B).

Carcinogenesis is caused by a cumulative, multistage process that mainly consists of initiation, promotion, and progression. ROS, which are produced as a result of the metabolism of molecular oxygen via biochemical reactions in cells, play a key role in tumor promotion. Some tumor promoters accelerate/induce the conversion of initiated cells (carcinogen-mediated mutation or potential “stem cells”) into tumorigenic cells possibly via the production of oxidative/inflammatory responses (30, 31). TPA (12-O-tetradecanoylphorbol-13-acetate), a phorbol ester, is a tumor promoter that induces the neoplastic/tumorigenic transformation of preneoplastic JB6 cells through the overproduction of ROS (32). In this study, we investigated the inhibitory effect of SFN on TPA-stimulated neoplastic transformation in the mouse epidermal JB6 P+ cell line to assess whether SFN is able to block tumor promoter-induced tumorigenesis in skin cells. Our results show that SFN was effective when it was given together with the tumor promoter TPA as well as when it was given before TPA that would have produced some biologic responses resulting in subsequent inhibition of transformation of JB6 P+ cells by TPA (Fig. 2). Pretreatment with SFN would increase the protein and mRNA expression of HO-1, NQO1, and UGT1A1 (Fig. 1B and C), among others, that could potentially explain the inhibitory effects on transformation. The multiple mechanisms of SFN in the induction of Nrf2 were subsequently been demonstrated, including increased protein expression and nuclear translocation of Nrf2 and enhanced the transcription level of Nrf2 by epigenetic regulation.

Nrf2 has been identified as an important transcription factor to regulate the expression of phase II antioxidant and detoxification enzymes, thus maintaining the balance of cellular ROS (6, 33). The transcriptional activation of antioxidant and detoxification genes is increased by Nrf2 through its interaction with AREs (34), and accumulating evidence indicates that dietary phytochemicals with chemopreventive potential could enhance the activities of antioxidant and detoxification enzymes via the Nrf2 pathway (35). This study shows that SFN increased HO-1, NQO1, and UGT1A1 mRNA and protein levels as well as the nuclear translocation and expression of total endogenous Nrf2 protein in JB6 P+ cells (Fig. 1). Keap1, a negative regulator of Nrf2, is a primary redox sensor with novel sulfhydryl residues that are sensitive to electrophiles and ROS, and Nrf2 itself may regulate its subcellular localization through its MES_TA motif (33, 36). Interestingly, it has been reported that SFN can modify multiple Keap1 domains and form thionoacyl adducts (37). The SFN-induced thiocarbamylation at cysteine 151 of Keap1 may disrupt the interaction between Keap1 and Nrf2, leading to the cytoplasmic and nuclear accumulation of Nrf2 (38–40). Therefore, the interaction between SFN and Keap1 is considered to be a potential mechanism for the induction of Nrf2 in JB6 P+ cells.

The chemopreventive efficacy of SFN against tumorigenesis via a critical Nrf2-mediated pathway has been demonstrated in several animal models, such as DMBA/TPA-induced skin carcinogenesis in Nrf2-deficient C57BL/6 mice (25) and benzo[a]pyrene-induced fore-stomach tumors in Nrf2-deficient ICR mice (41). In our previous study using TRAMP mice, broccoli sprouts, which contain abundant SFN, were found to suppress prostate tumorigenesis via the activation of the Nrf2 pathway (42). We have also used oligonucleotide microarray analysis to compare gene expression profiles between Nrf2-knockout mice and wild-type mice administered SFN via oral gavage; through these experiments, we identified Nrf2-dependent SFN-inducible antioxidant and detoxification genes (43). In this study, the critical role of Nrf2 induction via SFN pretreatment for the prevention of TPA-induced neoplastic transformation of JB6 P+ cells was demonstrated using a stable Nrf2-knockdown cell line (Fig. 4). In this line, reduced expression of the Nrf2 downstream enzymes HO-1 and NQO1 was induced by SFN (Fig. 3).

Epigenetic regulation, including DNA methylation and histone modification, is an important process in cancer development, and the epigenetic regulatory activity of certain dietary phytochemicals with cancer chemopreventive potential has been demonstrated, including curcumin (21), EGCG (20), and genistein (44). We have previously shown that the expression of Nrf2 can be attenuated by DNA CpG methylation in both the prostate tumor tissue of TRAMP mice and tumorigenic TRAMP C1 cells (12). Subsequently some effective cancer chemopreventive compounds were found to demethylate the CpG of Nrf2 promoter with the resulting increased of Nrf2 expression in in vivo (TRAMP mice) or in vitro (TRAMP C1 cells) studies (21, 22, 28, 45, 46). These studies suggest that the potent chemopreventive compounds that epigenetically restore the silenced Nrf2 gene via DNA demethylation could be used as epigenetic modifying cancer chemopreventive and therapeutic agents. Our present study provides further evidence about the demethylation activity of SFN on the first 15 CpGs of the Nrf2 promoter in JB6 P+ cells, and we show that the transcription of a hypomethylated Nrf2 gene might be more easily induced upon exposure to TPA (Fig. 5). These results also suggest that DNA demethylation is associated with Nrf2 activation in SFN-treated JB6 cells corroborating with previous published findings that DNA CpG methylation in the promoter of Nrf2 gene could serve as a potential key factor for Nrf2 gene silencing.

Epigenetic modification enzymes, such as DNMTs and HDACs that silence gene expression, are the major targets of cancer prevention and therapeutic strategies. Previous reports indicate that SFN can suppress DNMT1 and DNMT3a in TRAMP C1 cells (28) and human breast cancer cells (47), which may result in hypomethylation of the Nrf2 promoter and the first exon of the hTERT gene, respectively. We also found that the expression of the DNMT1, DNMT3a, and DNMT3b proteins was decreased in JB6 P+ cells after treatment with SFN (Fig. 6). In addition to DNMTs, the inhibitory effect of SFN on HDAC has been demonstrated in prostate epithelial BPH-1, LnCaP, and PC-3 cells (48). In addition to SFN, 2 major metabolites, SFN-cysteine and SFN-N-acetylcysteine, can suppress HDAC activity in vitro (49). In an in vivo study, the SFN-induced decrease in HDAC activity suppressed the growth of PC-3 xenografts, and SFN-rich broccoli sprouts inhibited HDAC activity in peripheral blood mononuclear cells in human subjects (50). SFN also causes histone modification in the colonic mucosa and polyps of Apc-minus mice through inhibition of HDAC activity, which might retard intestinal carcinogenesis (51). Histone modification was also observed in MCF-7 and MDA-MB-231 cells after SFN treatment, including an increase in active chromatin markers (acetyl-H3, acetyl-H3K9, and acetyl-H4) and a decrease in inactive chromatin markers (trimethyl-H3K9 and trimethyl-H3K27; ref. 47). Our present results show that the total activity of HDACs and the protein levels of HDAC1, HDAC2, HDAC3, and HDAC4 were significantly decreased by SFN in a dose-dependent manner (Fig. 7), suggesting that transcriptional activity of the Nrf2 gene may also be regulated by SFN through histone modification.

In conclusion, our present study shows that SFN induces the nuclear translocation of Nrf2, resulting in enhanced Nrf2-mediated expression of the cellular defense enzymes HO-1 and NQO1. Furthermore, our epigenetic analysis shows that SFN inhibits DNMT and HDAC protein expression, attenuates total HDAC activity, and decreases methylation of the first 15 CpGs of the Nrf2 promoter, resulting in enhanced Nrf2 mRNA and protein expression. Collectively, these results suggest that the SFN-mediated antineoplastic transformation of JB6 P+ cells induced by the tumor promoter TPA may involve multiple molecular mechanisms, including epigenetic reprogramming, Nrf2 reactivation, and enhanced expression of Nrf2 and Nrf2 target genes.

No potential conflicts of interest were disclosed.

Conception and design: Z.-Y. Su, C. Zhang, T.O. Khor, A.H. Conney, A.-N.T. Kong

Development of methodology: Z.-Y. Su, C. Zhang, L. Shu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z.-Y. Su, C. Zhang, J.H. Lee, T.-Y. Wu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z.-Y. Su, C. Zhang, T.-Y. Wu

Writing, review, and/or revision of the manuscript: Z.-Y. Su, A.H. Conney, Y.-P. Lu, A.-N.T. Kong

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.H. Lee, A.H. Conney, A.-N.T. Kong

Study supervision: T.O. Khor, A.H. Conney, A.-N.T. Kong

The authors thank all the members in Dr. A.-N.T. Kong's lab for their helpful discussion and preparation of this manuscript.

This work was supported by institutional funds of Rutgers, The State University of New Jersey.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.