Cryptotanshinone down-regulates androgen receptor signaling by modulating lysine-specific demethylase 1 function

Authors

  • Ching-Yuan Wu,

    1. Department of Chinese Medicine, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan
    2. Graduate Institute of Clinical Medical Sciences, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan
    Search for more papers by this author
  • Chang-Yi Hsieh,

    1. Graduate Institute of Clinical Medical Sciences, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan
    2. Center for Menopause and Reproductive Research, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan
    Search for more papers by this author
  • Ko-En Huang,

    1. Graduate Institute of Clinical Medical Sciences, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan
    2. Center for Menopause and Reproductive Research, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan
    Search for more papers by this author
  • Chawnshang Chang,

    1. Center for Menopause and Reproductive Research, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan
    2. George Whipple Laboratory for Cancer Research, University of Rochester, Rochester, New York
    Search for more papers by this author
  • Hong-Yo Kang

    Corresponding author
    1. Graduate Institute of Clinical Medical Sciences, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan
    2. Center for Menopause and Reproductive Research, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan
    • Graduate Institute of Clinical Medical Sciences, Center for Menopause and Reproductive Research, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, No.123, Ta-Pei Rd., Niaosong District, Kaohsiung, Taiwan 833, Republic of China
    Search for more papers by this author
    • Tel.: +886-7731-7123, ext: 8898, Fax: +886-7733-6970


Abstract

Development and progression of prostate cancer are intimately associated with androgen receptor (AR) signaling. The emergence of hormone-refractory prostate cancer and consequent failure of conventional androgen deprivation therapies make it necessary to bypass hormonal resistance by targeting the same signaling pathway at new intervention points. In our study, we showed that cryptotanshinone inhibited the growth of AR-positive prostate cancer cells, suggesting that cryptotanshinone affected AR function. Cryptotanshinone also profoundly inhibited the transcriptional activity of AR and suppressed the expression of several AR-target genes at the mRNA and the protein levels. At the molecular level, cryptotanshinone disrupted the interaction between AR and lysine-specific demethylase 1 (LSD1), and inhibited the complex of AR and LSD1 to the promoter of AR target genes without affecting the protein degradation and translocation of AR. Cryptotanshinone increased the mono-methyl and di-methylation of Histone H3 lysine 9 (H3K9), a repressive histone marker which is demethylated and activated by LSD1. These data suggest that cryptotanshinone functions via inhibition of LSD1, a protein that promotes AR-dependent transcriptional activity via derepression of H3K9. In summary, we describe a novel mechanism whereby cryptotanshinone down-regulates AR signaling via functional inhibition of LSD1-mediated demethylation of H3K9 and represses the transcriptional activity of AR. Our data suggest that cryptotanshinone can be developed as a potential therapeutic agent for prostate cancer.

Prostate cancer is the most common malignant disease and the second leading cause of death among male cancer patients in US. The American Cancer Society estimated that there were a total of 192,280 newly diagnosed cases of prostate cancer in 2009, while metastatic prostate cancer accounted for ∼ 273,600 deaths in US.1 Shortterm responsiveness towards the currently used hormone deprivation therapies is mainly caused by the development of hormonal resistance, leading to the emergence of disease known as the “hormone-refractory prostate cancer (HRPC).”2, 3 Androgen receptor (AR) mutation has been suggested to drive the progression of prostate cancer in HRPC. New drug development currently focuses on more effective inhibition of AR signaling which may lead to improved clinical benefits.4–6

AR, a ligand binding activated nuclear transcription factor, is a member of the steroid nuclear receptor superfamily. Androgens-AR pathway plays an important role in the stimulation of prostate cancer cell growth and the regulation of a number of target genes including the prostate-specific antigen gene (PSA).7, 8 Structurally, AR consists of four functional regions: an amino terminal domain (ARD), a DNA binding domain (DBD), a hinge region containing a nuclear localization signal and a carboxy-terminal ligand-binding domain. Inactive ARs, which are not bound to ligand, have been reported to bind heat shock proteins (HSPs),9, 10 and reside in the cytoplasm. Binding of AR ligands, such as dihydrotestosterone (DHT) results in dissociation of HSPs, AR dimerization, tyrosine kinase phosphorylation and nuclear translocation of the activated ARs.11, 12 Although androgens are dispensable for the survival of HRPC cells, the AR signaling pathway is necessary for the initiation and progression of prostate cancer7 and AR has been reported to be essential for the growth of HRPC cells. The dependence of HPRC on AR may be due to several factors such as point mutation in the AR gene, amplification of AR and changes in expression of AR coregulatory proteins.13, 14 These changes drive the transformation of native AR to develop into “super AR,” that is, stimulated by minimal concentrations of androgens to exert adverse effects and to become responsive to an increased range of ligands. Inhibition of AR function is therefore an important goal during the therapeutic development of novel antiprostate cancer growth drugs.

Traditional Chinese medicine has used dried roots of Salvia miltiorrhiza Bunge to treat a number of cardiovascular and endocrine diseases, including coronary artery disease, angina pectoris, hepatitis and menstrual disorders. Cryptotanshinone (Fig. 1a) is one among the more than 25 compounds of tanshinones and an active component of S. miltiorrhiza Bunge. While cryptotanshinone was previously shown to possess the most powerful antibacterial activity among the tanshinones and inhibit the growth of the androgen-independent prostate cancer cell line in vitro and in mice,15–17 the effects of cryptotanshinone and its mechanism of action on androgen-dependent prostate cancer cells remain unclear. As cryptotanshinone has been shown to play a role in lowering androgen synthesis in prenatally androgenized male rats,18 the goals in the present study were to investigate cryptotanshinone function in androgen-dependent prostate cancer cells and to explore the underlying mechanisms by investigating cryptotanshinone-AR interaction.

Figure 1.

LNCaP cell proliferation is inhibited by cryptotanshinone. (a) Structure of cryptotanshinone. (b,c) The growth rates of LNCaP cell (about 70th passage) line (b) and PC3 cell line (c) as measured by XTT assay after 1–3 days of culturing in the presence of 10 μM of cryptotanshinone (CR) with or without 10 nM DHT treatment. (d) The growth rates of LNCaP cell line and PC3 cell line as measured by BrdU assay after 3 days of culturing in the presence of 10 μM of cryptotanshinone (CR) with or without 10 nM DHT treatment. (e) Effect of CR over a range of concentrations on the cell viability of LNCaP cells after a 24 hrs incubation. There were three replicates for each experiment and the results represent the mean ± S.D. of five experiments. Asterisks (*,**,***) mark samples significantly different from cryptotanshinone treatment with p <0.05.

Material and Methods

Cell culture and treatment

The human prostate cancer cell lines (LNCaP, 22Rv1 and PC3) were obtained from the American Type Culture Collection. Human prostate cancer PC-3 cells were maintained in Dulbecco's Minimum Essential Medium (DMEM) (Invitrogen, Carlsbad, CA) containing penicillin (25 U/ml), streptomycin (25 μg/ml) and 10% fetal bovine serum (FBS). The LNCaP and 22Rv1 cells were cultured in RPMI 1640 medium (Invitrogen, Carlsbad, CA), supplemented with 10% FBS at 37°C and 5% CO2. Stable clones of PC3-AR cell line were obtained after puromycin selection as previous described.19 Cryptotanshinone was obtained from Fleton Natural Products, China (the purity of cryptotanshinone is >98% and its solubility in DMSO is >5 mg/ml). DHT was obtained from Sigma. LNCaP cells were cultured to 60–70% confluence before treatment. Medium was then replaced with fresh medium containing cryptotanshinone in DMSO (dimethyl sulfoxide) at the indicated concentrations. Cells treated with DMSO alone were used as untreated controls.

Cell viability and XTT assay

Human prostate cancer cell lines (LNCaP (about 70th passage) and PC3 cells) were plated at a density of 103 or 5 × 103 per well, respectively, in 96-well plates, in RPMI 1640 medium or DMEM containing 10% FBS. Once attached, the medium was replaced with RPMI 1640 or DMEM containing 10% charcoal-dextran–treated FBS. The cells were then treated with DHT or cryptotanshinone for 24, 48 or 72 hrs, and absorbance were measured using the XTT assay kit (Roche, Cat. No. 11465015001) according to the manufacturer's instructions as described previously.20 The XTT formazan complex was quantitatively measured at 492 nm using an ELISA reader (Bio-Rad Laboratories). For cell viability assay, LNCaP cells were treated with a range of concentrations of cryptotanshinone with or without DHT for 24 hrs. The cell numbers and viability were assessed by trypan blue exclusion under an inverse light microscope (Carl Zeiss).

BrdU cell proliferation assay

BrdU cell proliferation assay was performed as described previously.21 For this assay, cell proliferation was measured by colorimetric immunoassay based on bromodeoxyuridine (BrdU) incorporation by BrdU kit (Roche, Cat. No. 11647229001) depending to the manufacturer's instructions. In brief, the cells (4–5,000 cells/well) were seeded in 96-well plates with RPMI 1640 or DMEM containing 10% charcoal-dextran–treated FBS. After 24 hrs, cells were incubated with RPMI 1640 or DMEM containing 10% charcoal-dextran–treated FBS with or without DHT or cryptotanshinone for 3 days. Next, BrdU-labeling solution was added to each well and were reincubated for 4 hrs. After removal of the BrdU-labeling solution, cells were fixed and denatured with the kit's Fix Denat solution. Then cells were incubated for 90 min with peroxidase-conjugated anti-BrdU antibody. After washing, the color reaction was developed with the substrate solution and stopped by adding sulfuric acid. The optical densities of the samples were determined using an ELISA reader (Bio-Rad Laboratories) at 450 nm (reference value 690 nm).

Luciferase assay

LNCaP and PC3-AR cells were transiently transfected with the PSA-luciferase construct or the mouse mammary tumor virus (MMTV)–luciferases construct using the Superfect reagent (Qiagen, Cat. No. 301307) according to the manufacturer's instructions. Cells were cotransfected with the reni-luciferase (SV40) plasmid as the internal control. The cells were treated with the indicated drugs or left untreated for 48 hrs posttransfection. Cells were harvested and luciferase activity was measured using the Dual Luciferase Assay Kit (Promega, Cat. No. E1980) according to the manufacturer's instructions.22

Quantitative real time PCR

Total RNA was extracted from LNCaP and 22Rv1 cells using the TRIzol reagent (Invitrogen, Cat. No. 15596-026) according to the manufacturer's instructions. Reverse transcription was performed using the superscript first strand synthesis kit (Invitrogen, Number: 11904018). Quantitative real-time PCR analyses using the comparative CT method were performed on an ABI PRISM 7700 sequence detector system using the SYBR Green PCR Master Mix kit (Perkin–Elmer, Applied Biosystems, Wellesley, MA) according to the manufacturer's instructions. After initial incubation at 50°C for 2 min and 10 min at 95°C, amplification was performed for 40 cycles at 95°C for 20 sec, 65°C for 20 sec and 72°C for 30 sec. Specific primer pairs were determined with the PrimerExpress program (Applied Biosystems). Primer sequences are available on request.

Western blot analysis and fractionation assay

Western blot analyses were performed as described previously.23, 24 NE-PER nuclear and cytoplasmic Extraction Reagents kits (Pierce, Number: 89863) were used for the fractionation assay. For fractionation assay, cellular extracts of LNCaP cells treated with or without DHT and CR for 1 hr were prepared according to the manufacturer's instructions. The equal amounts of protein were fractionated on a 7.5 or 10% SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes. The membranes were then blocked with 5% nonfat dried milk for 30 minutes and incubated in primary antibody for 3 hrs in room temperature. The primary antibodies used were: AR (Santa Cruz, N-20, sc-816, ratio: 1:1,000), PSA (Santa Cruz, C-19, sc-7638, ratio: 1:1,000), mono-methyl H3K9 (Abcam, ab1771, ratio: 1:2,000), di-methyl H3K9 (Abcam, ab1220, ratio: 1:2,000), acetylated H3K9 (Abcam, ab10812, ratio: 1:2,000) and LSD1 (Abcam, ab17721, ratio: 1:1,000). The primary antibody and secondary antibody were diluted with 1% nonfat dried milk in 1× TBST (Tris-Buffered Saline Tween-20). Blots were washed by 1× TBST and incubated in horseradish peroxidase-conjugated secondary anti-mouse or anti-rabbit antibodies (Santa Cruz, ratio: 1:5,000) for 1 hr in room temperature. After washing by 1× TBST again, protein signal was detected by chemiluminescence, using the Super Signal substrate (Pierce, Number: 34087).

Chromatin immunoprecipitation assay (ChIP) and rechromatin immunoprecipitation assay (re-ChIP)

ChIP assays were performed using a modification of a previously described method.6 LNCaP cells were treated with the specified drug for 45 minutes, and then cells were harvested for ChIP. Untreated cells were used as control. Cell lysates were precleared with normal rabbit IgG (sc-2027, Santa Cruz Biotechnology) and protein A-agarose. Specific antibodies (anti-AR, anti-LSD1 and control IgG) were added to the lysates (1 μg primary antibody/1 mg protein extract) and immunoprecipitated overnight at 4°C. The following primer pairs, which span the region −3,061 to −3,262 of the AR promoter, were used for the amplification of PCR products: forward primer, 5′-ATGCTTTCCTGTTTACAAGTTTATTCTATACAC-3′; reverse primer, 5′-AGTTACTCTGAATAAAAAGCAGTCTGACAT-3′.

Immunoprecipitated complexes were sequentially washed with TSE I, TSE II, buffer II and TE for re-ChIP assays. Complexes were eluted by incubation with 10 mM DTT, followed by a second immunoprecipitation with the indicated antibodies. PCR primers for ARE I (PSA −459 to 121), ARE II+III (PSA −4,288 to −3,922) have been described previously.6

Immunoprecipitation (IP) and immunofluorescence assay (IFA)

IP was performed using a modification of a previously described method.12, 24 After indicated treatment for 1 hr, LNCaP or 22Rv1 cells were lysed with E1A lysis buffer [250 mM NaCl, 50 mM HEPES (pH 7.5), 0.1% NP-40, 5 mM EDTA, protease inhibitor cocktail (Roche)]. Cell lysates were precleared with normal rabbit IgG (sc-2027, Santa Cruz Biotechnology) and protein A-agarose. Specific antibodies (anti-AR, anti-LSD1 and control IgG) were added to the lysates (1 μg primary antibody/1 mg protein extract) and immunoprecipitated overnight at 4°C. After washing beads for four times with E1A lysis buffer, immunoprecipitated complexes were analyzed by immunoblotting. For IFA, cells were grown on chamber slides. After treated with or without indicated drugs, cells were washing with 1× PBS. Cells were then fixation with 4% paraformaldehyde and permeabilized in 0.2% Triton X-100 in PBS at room temperature. Cells were immunolabeled using specific primary and secondary antibodies with DAPI (1:1,000) and observed on microscopy.

Small interfering RNA (siRNA) transfection

siRNA was performed using a modification of a previously described method.24 LNCaP cells were transfected with siRNA against control (Dharmacon Research) or LSD1 (sequence: 5′-UGAAUUAGCUGAAACACAAUU-3′)25 using lipofectamine 2000 (invitrogen) for 48 hrs and treated with or without drugs. After 24 hrs, LNCaP cells were harvested for Western's blotting and q-PCR.

Quantification of protein level

For quantification of protein level, the images of bands from Western's blotting were analyzed by AlphaEaseFC® Software according to the manufacturer's instructions. After selecting the band of each group and the background, the densities of bands were automatically calibrated by subtracting the background. The density of the group without treatment was used as the standard to calculate the ratio value of the other groups.

Statistical analyses

All values were the means ± standard deviations (SD) of replicate samples (n = 3 to 6, depending on the experiment) and experiments were repeated a minimum of three times. Differences between two groups were assessed using the unpaired two-tailed Student's t-test or by ANOVA if more than two groups were analyzed. The Tukey test was used as a post hoc test in ANOVA for testing the significance of pairwise group comparisons. The p values <0.05 were considered statistically significant in all comparisons. SPSS version 13.0 for windows (LEAD technologies) was used for all calculations.

Results

The effect of cryptotanshinone on the growth of prostate cancer cell lines

To study whether cryptotanshinone have effects on androgen/AR-dependent prostate cancer cell growth, we used two different prostate cancer cell lines (AR-positive LNCaP cells and AR-negative PC3 cells) as a model to study cryptotanshinone function. As 10 μM or higher concentration of cryptotanshinone has been shown to have better suppression effect on prostate cancer cell growth,17 we first choose the concentration of 10 μM for further study. Our results showed that 10 μM of cryptotanshinone significantly inhibited the growth and proliferation of the LNCaP cells in a time-dependent manner (Figs. 1b and 1d). More importantly, the DHT-enhanced growth and proliferation of LNCaP cells both were blocked by cryptotanshinone (Figs. 1b and 1d). Exposure of LNCaP cells to cryptotanshinone reduced the cell viability in a concentration-dependent manner and the medium lethal dose (LD50) of cryptotanshinone was 20 μM (Fig. 1e). In contrast, the growth of PC3 cells, which are known to be androgen-independent, was partial affected by the same concentration of cryptotanshinone (Figs. 1c and 1d), suggesting a role for cryptotanshinone in growth-inhibition of androgen-dependent prostate cancer cells.

Cryptotanshinone inhibits AR function

As the transcriptional activity of activated AR plays an important role in maintaining the growth of AR dependent prostate cancer cells, we used the dual luciferase reporter gene assay to investigate the effect of cryptotanshinone on the transcriptional activity of AR. LNCaP and PC3-AR cells were transiently transfected with PSA-luciferase or the mouse mammary tumor virus (MMTV)–luciferase construct along with Reni- luciferase (SV40) as an internal control. Cells were treated with varying concentrations of cryptotanshinone and 10 nM DHT for 48 hrs. We showed a significant decrease in luciferase activity at the end of the treatment period in LNCaP cells (Figs. 2a and 2b) and PC3-AR cells (Figs. 2c and 2d) by 1 μM of cryptotanshinone, suggesting that low concentrations of cryptotanshinone maybe sufficient to inhibit the DHT-induced transcriptional activity of activated AR.

Figure 2.

Transcriptional activity of AR is inhibited by cryptotanshinone. LNCaP cells (a,b) or PC3-AR cells (c,d) was were transiently transfected with the PSA-luciferase construct or the mouse mammary tumor virus (MMTV)–Luc. Reni-luciferase (SV40) was used to control transfection efficiency in both cell types. Varying concentrations of cryptotanshinone were added in the presence or absence of 10 nM DHT and transcriptional function of AR was then quantified by the dual-luciferase assay after 48 hrs. Each experiment was performed in triplicate and the results represent the mean ± S.D. Asterisks (*) mark samples significantly different from cryptotanshinone treatment with p <0.05. DHT:10 nM of dihydrotestosterone. CR: 0.1, 1 or 10 μM cryptotanshinone.

Cryptotanshinone inhibited AR-mediated transcriptional activity but not AR mRNA and protein levels

As cryptotanshinone inhibited AR transcriptional activity in LNCaP cells, human prostate tumor cells with AR, as well as in PC3-AR cells, we used real-time PCR and Western blots to explore the effect of cryptotanshinone on mRNA as well as protein levels of AR, as well as a critical downstream protein, PSA. We showed that there was no significant difference in AR mRNA or protein levels between untreated and treated cells after one day of treatment with different concentrations of cryptotanshinone in LNCaP cells (Figs. 3a and 3d). However, cryptotanshinone inhibited PSA mRNA and protein levels at the concentration of 10 μM in LNCaP cells (Figs. 3a and 3e). There was also a significant decrease in the mRNA levels of other AR-target genes, such as NKX3.1 and TMPRSS2 in LNCaP cells (Figs. 3f and 3g) in the same concentration. 22Rv1 cell line is an androgen-responsive human prostate carcinoma cell line that expresses mutant (H874Y) ARs and secretes low level of PSA.26–28 As shown in Figure 3b, the DHT-induced PSA protein level of 22Rv1 cells was weaker than LNCaP cells (Fig. 3a). To confirm this result, we measured PSA mRNA of 22Rv1 cells by qPCR. As shown in Figure 3c, the expression level of PSA mRNA was increased to 2-folds in 22Rv1 cells after 10 nM of DHT treatment, that is, consistent with previous report.29 After treat with cryptotanshinone, we found that expression of PSA mRNA was inhibited in 22Rv1 cells. Together, our data showed that cryptotanshinone could effectively suppress the DHT-induced AR target gene expression.

Figure 3.

Cryptotanshinone treatment decreases the mRNA levels of AR target genes but not the protein level of AR. (a,b) Total cell extracts of LNCaP cells (a) or 22Rv1 cells (b) were harvested from untreated cells and cells treated with DHT and cryptotanshinone for 24 hrs. The protein was immunoblotted with polyclonal antibodies specific for AR and PSA. Each lane contained 100 μg of the total proteins. α-tubulin was used as an internal loading control. (c–g) Total mRNA was extracted from LNCaP cells or 22Rv1 cells after treat with drugs for 24 hrs. The coding regions of AR, PSA, NKX3.1 or TMPRSS2 were used as probes for real time polymerase chain reaction analysis. DHT: 10 nM of dihydrotestosterone. CR: 1 or 10 μM cryptotanshinone. All the results are representative of at least three independent experiments. Asterisks (*) mark samples significantly different from cryptotanshinone treatment with p <0.05.

Nuclear translocation of AR is not inhibited by cryptotanshinone

One of the critical steps in AR activation is the translocation of AR from cytoplasm into nucleoplasm in response to androgen treatment, suggesting that this step may be a target of cryptotanshinone modulation. We evaluated AR protein localization by immunofluorescence assay and fractionation assays followed by Western blot analysis. The immunofluorescence staining revealed that the majority of cells have a diffuse cytosolic and nucleus distribution of AR in unstimulated cells. While the activated AR translocated into the nucleus in the presence of DHT, the localization of AR did not altered significantly even with cryptotanshinone treatment (Fig. 4a). Results obtained by fluorescence microscopy were also confirmed by subcellular fractionation experiments to show that the localization of both AR proteins was marginally altered by cryptotanshinone both in the cytosolic and nuclear fraction of cell extracts (Fig. 4b), suggesting that cryptotanshinone may inhibit the transcriptional function of activated AR via other pathways than inhibition of AR nuclear translocation.

Figure 4.

Nuclear translocation of activated AR was marginally affected by cryptotanshinone. (a) The localization of AR in LNCaP cells after treatment with or without cryptotanshinone and DHT was detected by immunofluorescence staining with AR antibodies. Cells were stained with DAPI (top panel), anti-AR antibody (middle panel) and the bottom panel shows merged images. (b) Nuclear and cytoplasmic extracts of LNCaP cells treated with or without cryptotanshinone and DHT for 1 hr were prepared and immunoblotted with AR antibodies. PARP was detected exclusively in the nuclear fraction and served as a control for nuclear proteins. α-tubulin was detected exclusively in the cytoplasmic fraction and served as a control for cytoplasmic proteins. DHT: 10 nM of dihydrotestosterone. CR: 10 μM cryptotanshinone.

Cryptotanshinone induces dissociation of LSD1 from the promoter of AR target genes to inhibit binding of AR to DNA

We found activated, nuclear AR in DHT-treated cells in the presence of cryptotanshinone, suggesting that cryptotanshinone did not function solely by inhibition of nuclear translocation of AR (Figs. 4a and 4b). In recent reports,30–33 mTOR signal and cyclin D1 played important roles in regulation of AR functions. So we also test the effect of CR on these two signals. However, there is no difference in these signals between with or without CR, (Fig. 5a). We previously showed that AR cofactors were required to activate androgen-mediated gene transcription and that recruitment of cofactors and histone modifications were essential for AR-mediated transcriptional regulation of target genes.5, 7, 34 We therefore used immunoprecipitation (IP) assays to screen the interaction of AR with a number of cofactors in the presence of cryptotanshinone. To this end, we found that Lysine-specific demethylase 1 (LSD1) dissociated from AR by immunoprecipitation (IP) assay with AR antibody in cryptotanshinone-treated LNCaP cells and 22Rv1 cells as shown in Figures 5b and 5d. To confirm the effect of cryptotanshinone on interfering AR and LSD1 interaction, we performed the reverse IP with LSD1 antibody and also repeated the experiments in LNCaP cells to confirm androgen-mediated LSD1-AR complex were disrupted by cryptotanshinone treatment (Fig. 5c). LSD1 is detected in the epithelium of normal prostate and in prostate tumor cells and higher Gleason grade of prostate tumor tissue have higher LSD1 expression level. Importantly, these cells also express AR, showing that LSD1 and AR colocalize. The nuclear colocalization of LSD1 and AR was verified further in LNCaP cells.6 As LSD1 has been reported to possess a histone demethylating function, specifically demethylating H3K9 (lysine 9 of histone H3) in AR-targeted genes and demethylation of mono- and dimethyl residues of H3K9 by LSD1 is an important marker of AR-activated gene expression.6 To examine the effects of cryptotanshinone on the LSD1-mediated demethylation of AR-targeted genes, we checked the methylation status of H3K9 and found that both di-methyl and mono-methylation of H3K9 cannot be demethylated by androgens in the presence of cryptotanshinone treatment (Fig. 5e), suggesting that cryptotanshinone suppressed the AR-mediated demethylation activity of LSD1. As both LSD1 and AR are known to bind the promoter area of AR response element (ARE), we performed chromatin immunoprecipitation (ChIP) to examine the binding of LSD1 and AR complex to AR target gene promoters. Our ChIP assays revealed that while both LSD1 and AR bound in the ARE I of PSA gene promoters in the presence of DHT as expected, cryptotanshinone treatment caused a dissociation of LSD1 from ARE I in the promoter of AR target genes, PSA (Fig. 5f). Re-ChIP assay confirmed that cryptotanshinone significantly inhibited the ability of AR and LSD1 to bind together to the PSA promoter, including ARE I, II and III (Fig. 5g), indicating that cryptotanshinone may suppress the binding of LSD1 to activated AR on the ARE of PSA promoter by inhibiting the AR-LSD1 interaction. To study whether LSD1 downstream target genes are also influenced by cryptotanshinone, we examined the mRNA level of LSD1 target genes including S100A8, which encodes a calcium-binding protein, and DEK, a proto-oncogene.35 As shown in Figures 5h and 5i, we found that there is no statistically significance between different treatment groups. To confirm this result, we silenced LSD1 in LNCaP cells, as shown in Figure 5j. After treatment with drugs, we found there was no difference in the mRNA levels of the two genes after treatment of cryptotanshinone (Figs. 5k and 5l). Together, these results suggest while cryptotanshinone can interrupt the interaction between AR and LSD1 in AR-targeted genes such as PSA, the LSD1 downstream target genes such as S100A8 and DEK may not be affected by cryptotanshinone in the prostate cancer cells.

Figure 5.

LSD1-AR interaction is inhibited by cryptotanshinone. (a–e) Total cell extract were prepared from LNCaP cells or 22Rv1 cells after 1 hr of various treatment as indicated for immunoblotting and coimmunoprecipitation assays. (e) Total cell extract were prepared from LNCaP cells after 18 hr of various treatment as indicated for immunoblotting. (f, g) ChIP or Re-ChIP was carried out to study the effect of cryptotanshinone on LSD1 and AR interaction on PSA promoter by using the indicated antibodies after 45 minutes of incubation with or without treatment in LNCaP cells. Primers in the promoter region of the PSA gene (ARE I and ARE II & III) or actin genes were used to PCR-amplify the precipitated chromatin. The “input” samples represent equal fractions of extract collected before precipitation from which DNA was extracted and used as a control. (h,i) Total mRNA was extracted from LNCaP cells after treat with drugs for 24 hrs. The coding regions of S100A8 and DEK were examined as LSD1 targeted genes for real time polymerase chain reaction analysis. (j) LNCaP cells were transfected with siRNA against GFP or LSD1 for 48 hrs and harvested for WCE, followed by IB analysis. (k, l) Total mRNA was extracted from control- and LSD1-knockdown LNCaP cells after treat with drugs for 24 hrs. The coding regions of S100A8 and DEK were examined as LSD1 targeted genes for real time polymerase chain reaction analysis. All the results are representative of at least three independent experiments. DHT: 10 nM of dihydrotestosterone. CR: 10 μM cryptotanshinone.

Discussion

AR-mediated signaling pathways are very important for the growth and survival of primary prostate cancer cells. Although the antitumor activity of cryptotanshinone has been well documented in a number of cancers including prostate cancer,15, 17, 36, 37 its exact mode of action in AR-positive prostate cancer cells is not completely understood. Our current study demonstrated that cryptotanshinone inhibited the proliferation of LNCaP cells much more effectively when compared to PC3 cells. One of the important differences between these two cell types is that proliferation of LNCaP cells is AR-positive and androgen-responsive cells whereas PC3 cells grow independently of AR and androgens. The inhibitory activity of cryptotanshinone on LNCaP cells may therefore be related to AR function. As cryptotanshinone treatment inhibited the transcriptional activity of activated AR without altering AR mRNA or protein levels, these data suggest that cryptotanshinone regulates the function of activated AR without affecting AR production and degradation. Furthermore, we found that cryptotanshinone treatment resulted in a decrease in the mRNA levels of downstream-target proteins of the AR-activated pathway, such as NKX3.1, TMPRSS2 and PSA but not AR. Together, these results further strengthen the notion that cryptotanshinone-mediated inhibition of the DHT-activated AR pathway was brought about by inhibition of DHT-mediated AR function, rather than by down-regulating AR expression levels.

For AR, it needs the other regulatory proteins, like mTOR,30–32 cyclin D1,33 to control its function. However, in our data, CR affected the AR function without through these pathways. Gene regulation in eukaryotes requires the coordinate interaction of chromatin-modulating proteins with specific transcription factors such as AR. Gene activation and repression are specifically regulated by histone methylation of distinct lysine residues. LSD1 is a newly identified histone demethylase that promotes AR-targeted genes by interacting with AR and derepressing histone marks via demethylating histone H3 at lysine 9 (H3-K9).6 High levels of LSD1 along with a high Gleason score and grade, which indicated poor prognosis in prostate cancer, have been shown to correlate significantly with relapse during follow-up, suggesting that LSD1 may be a novel biomarker predictive for AR-mediated prostate cancer growth.38 We used immunoprecipitation assays to show that the AR-LSD1 interaction is a major target for cryptotanshinone-mediated inhibition. Cryptotanshinone treatment also inhibited LSD1-induced demethylation of di- and mono-methyl H3K9, which in turn, suppressed AR transcriptional activity. Based on our results, we propose a working model by which the traditional Chinese herb-derived compound, cryptotanshinone, regulates AR-mediated gene activation (Fig. 6). Cryptotanshinone-mediated inhibition of AR target gene activation occurs via inhibition of LSD1-AR demethylation function, resulting in inhibition of prostate cancer growth. Jumonji C (JMJC) domain-containing protein, JMJD2C, has been shown to colocalize with AR and LSD1 in normal prostate and in prostate carcinomas. Demethylases JMJD2C and LSD1 interact and cooperatively stimulate AR-dependent gene transcription. In addition, AR, JMJD2C and LSD1 assemble on chromatin to remove methyl groups from mono, di and trimethylated H3K9.39 It will be of interest to further investigate the effect of cryptotanshinone on the assembly of AR, JMJD2C and LSD1 complex on chromatin.

Figure 6.

Working model describing the role of cryptotanshinone in blocking the activated AR pathway. After stimulation with androgens activated AR binds to the promoter of AR target genes. Interaction of activated AR with LSD1 induces demethylation of H3K9 and activation of AR target genes transcription in the presence of androgens. Cryptotanshinone treatment disrupts the interaction between activated AR and LSD1 on chromatin. H3K9 demethylation is suppressed, resulting in the inhibition of AR target gene transcription.

S. miltiorrhiza Bunge was found to be effective in reducing both intimal hyperplasia and restenosis.40 Cryptotanshinone has also been reported to possess several desirable biological properties including anti-inflammatory, antibacterial, antitumor, antioxidative, antimutagenic and antiplatelet aggregation activities through blocking STAT or Bcl-2 and MAPK regulation,37, 41–45 which are unrelated to cryptotanshinone's effects on androgen-dependent prostate cancer. Our data demonstrating the inhibitory role of cryptotanshinone on AR-positive LNCaP cells, suggest that it can be used to treat prostate cancer even in patients with early androgen-dependent staged disease, before the cancer cells become refractory to hormone deprivation therapies. Although we showed that cryptotanshinone inhibited activated AR via LSD1, the mechanisms by which cryptotanshinone inhibits the association of LSD1 with AR remain unclear. Neo-tanshinlactone, a related constituent of tanshinones with a lactone rather than o-quinone ring-C, showed selective cytotoxicity against two estrogen receptor-positive (ER+) breast cancer cell lines, MCF-7 and ZR-75-1 and was 10-fold more potent than tamoxifen.46, 47 In initial Structure-Activity Relationship (SAR) analog studies, the aromatic rings A and D were found to be important for anticancer activity. In addition, certain ring C-opened analogs retained activity and had increased selectivity towards specific cancer subtypes. It is highly possible that the ring C structure of cryptotanshinone may possess the selective ability towards inhibiting the association of AR-LSD1 complex. In addition, the growth of prostate cancer cells and the function of histone H3 lysine 9 are influenced by several nuclear proteins other than LSD1, it is possible that the growth inhibition of prostate cancer cells in our study by cryptotanshinone may be dictated by factors other than the AR-LSD1 pathway and these will warrant attention in future studies.

DEK is a nuclear DNA-binding protein and the expression of DEK is correlates with hepatocellular tumor cell proliferation.48 S100A8 is the members of the S100 protein family and belongs to the calcium-binding EF-hand motif superfamily. They are involved in several cellular processes including transcription, proliferation and differentiation.49 In the recent study, the mRNA level of S100A8 could be the prognostic marker of bladder cancer.50 In our result, the mRNA levels of two genes were down-regulated when LSD1 was silenced which was consistent with the previous report.35 However, the mRNA levels of the two genes were not markedly affected by the treatment of DHT in both wild type and LSD1-knockdown LNCaP cells. It suggests activated AR cannot influence the transcription of S100A8 and DEK in the prostate cancer cells. As shown in Figures 5hl, we found that there was no statistically significance between different treatment groups. Together, these results supported while cryptotanshinone can selectively inhibit AR target genes such as PSA, the LSD1 downstream target genes, such as S100A8 and DEK, may not be influenced by cryptotanshinone in the prostate cancer cells.

Cryptotanshinone was recently found to inhibit the proliferation of androgen-independent prostate cancer cells, DU145, in 7 μM cryptotanshinone treatment by blocking STAT3 activation.17 In parallel with these results, our data also demonstrated that cryptotanshinone inhibited proliferation of androgen-dependent prostate cancer cells. Moreover, we are the first to demonstrate that low concentrations (0.1–1 μM) of cryptotanshinone inhibited AR transcriptional activity. These data point to the potential of cryptotanshinone as a novel antiandrogen drug. We are aware that the growth inhibitory function of cryptotanshinone can be achieved via other target molecules or pathways, such as STAT3 and the concentration of cryptotanshinone at 7 μM had no inhibitory effect on the growth of LNCaP and PC3 cells in the absence of androgens.17 However, we found that 10 μM of cryptotanshinone is sufficient to decrease the proliferation of LNCaP prostate cancer cells. Although it is important to investigate this discrepancy further, it is notable that the presence of androgens in our experiments is an indispensable variable when determining the effectiveness of cryptotanshinone on LNCaP cells, as LNCaP cells are AR positive and are androgen-dependent cells. Together the earlier report,17 10 μM cryptotanshinone could inhibit the growth of AR-positive and AR-negative prostate cancer cells. The cryptotanshinone interrupted AR-LSD1 interaction is one of the mechanisms responsible for cryptotanshinone-inhibited prostate cancer cell growth.

In summary, our results describe a novel mechanism of action of cryptotanshinone, whereby it functions via a novel-signaling pathway. Our data expand our understanding of cryptotanshinone-mediated signaling pathways and the mechanism underlying its antitumor effects. More importantly, the results of our study open the possibility for the use of cryptotanshinone as a therapeutic agent in androgen-sensitive as well as androgen refractory prostate cancers.

Acknowledgements

The authors thank Ms. Pei-Chun Lin for the technical assistance and Ms. Hsi-Ping Chi for critical reading of the article. This work was supported by grant CMRPD 8A0441 from Chang Gung Memorial Hospital, and NMRPD 1A0591 from the National Science Council to Dr. Hong-Yo Kang.

Ancillary