Slug, a Unique Androgen-Regulated Transcription Factor, Coordinates Androgen Receptor to Facilitate Castration Resistance in Prostate Cancer

Abstract

Prostate cancer (PCa) becomes lethal when cancer cells develop into castration-resistant PCa (CRPC). Androgen receptor (AR) gene mutation, altered AR regulation, or overexpression of AR often found in CRPC is believed to become one of the key factors to the lethal phenotype. Here we identify Slug, a member of the Snail family of zinc-finger transcription factors associated with cancer metastasis, as a unique androgen-responsive gene in PCa cells. In addition, the presence of constitutively active AR can induce Slug expression in a ligand-independent manner. Slug overexpression will increase AR protein expression and form a complex with AR. In addition, Slug appears to be a novel coactivator to enhance AR transcriptional activities and AR-mediated cell growth with or without androgen. In vivo, elevated Slug expression provides a growth advantage for PCa cells in androgen-deprived conditions. Most importantly, these observations were validated by several data sets from tissue microarrays. Overall, there is a reciprocal regulation between Slug and AR not only in transcriptional regulation but also in protein bioactivity, and Slug-AR complex plays an important role in accelerating the androgen-independent outgrowth of CRPC.

Prostate cancer (PCa) is a leading cancer incidence and second leading cause of cancer death among men in the United States (1). In general, primary PCa is an androgen-dependent disease and is highly responsive to androgen ablation therapy. However, PCa eventually will recur and progress to castration-resistant prostate cancer (CRPC) after hormone therapy. CRPC is the lethal form of PCa. At present, there is no therapy to cure CPRC (2) because mechanism(s) leading to the relapse of CRPC are poorly understood.

Androgens are essential hormone for the proliferation and survival of prostate cells, and the androgen receptor (AR) is the principal receptor responsible for mediating the physiological effects of androgens. In addition to androgens, AR can also be activated by nonandrogen ligands, such as IL or growth factors (3). Upon ligand binding, AR undergoes a conformational change, translocates into the nucleus, and binds to specific androgen response elements (ARE), in which it recruits the cofactors to modulate the gene transcription (4, 5). AR is a critical factor for not only the growth and survival of primary cancer but also the development of CRPC (6, 7). In CRPC cells, AR signaling becomes hyperactive through various AR gene alterations such as amplification, mutations, or splicing and altered AR coactivator interaction (7–14). All these data implicate multiple mechanisms by which PCa cells acquire resistance to hormone therapy at the late stage; AR with its signaling axis is still the key role in this event.

Recently Slug (SNAI2), a member of the Snail family of zinc-finger transcription factors (15), was identified as a potential oncogene in various cancer types (16–19). In addition, Slug is capable of repressing E-cadherin expression and triggering epithelial-mesenchymal transition (EMT), a complex process that allows cancer cells to escape from the primary tumor and metastasize (20–22). In general, Slug has been implicated in various physiological and pathological processes. For instance, Slug is involved in the protection of hematopoietic progenitors from radiation-induced apoptosis from Slug-knockout animals (23). Also, Slug-knockout mice exhibit unique phenotypes such as white forehead blaze, patchy depigmentation of the ventral body, tail and feet, and macrocytic anemia, which infer an important role for Slug in melanocyte and hematopoietic stem cells (24). Interestingly, these mice also develop infertility with defective functions of both spermatogonia and Leydig cells (24), suggesting that Slug may contribute to urogenital organ development. On the other hand, data from cDNA arrays have indicated that androgen can regulate Slug expression in the PCa cell line (i.e. LNCaP) and normal prostate epithelial cell line (i.e. HPr-1) (25, 26). However, the functional role of Slug in PCa is largely unknown.

In this study, we define Slug as an androgen-regulated gene in PCa cell lines expressing AR, whereas the presence of Slug in PCa cells will enhance AR protein expression. The increased Slug and AR proteins can form a complex, which may result in enhancing both AR transcriptional activities and AR-mediated cell growth induced by androgen. Noticeably, the elevated Slug expression in PCa cells can activate AR target gene expression and potentiate both in vitro and in vivo growth in the absence of androgen. In addition, data from clinical specimens also support this conclusion that there is a positive correlation of Slug expression with nuclear AR localization in PCa specimens, particularly from CRPC patients. Taken together, there is a reciprocal regulation and interaction between AR and Slug in which the Slug-AR complex appears to play an important role in accelerating the outgrowth of CRPC by increasing AR protein expression, enhancing AR activities, and potentiating castration resistance.

Results

Slug as an androgen-regulated gene in PCa cells

To determine the effect of androgen on the Slug gene expression in PCa cells, several AR-positive PCa cells were treated with androgen for 36 h. Dihydrotestosterone (DHT) treatment resulted in a dramatic induction of Slug mRNA and protein from LNCaP, C4-2, and CWR22RV1 cells in a dose-dependent manner (Fig. 1A), and similar results were observed in three cells after R1881 treatment. This induction could be observed as early as 2 h after androgen treatment, which is much earlier than other androgen-regulated genes such as prostate-specific antigen (PSA) and transmembrane protease serine 2 (TMPRSS2) (Supplemental Fig. 1A, published on The Endocrine Society's Journals Online web site at http://mend.endojournals.org). In contrast, androgen could not increase Slug mRNA levels in AR-negative DU145 and PC-3 cell lines (Supplemental Fig. 1B). In addition, the antiandrogen bicalutamide (Casodex) blocked Slug induction in LNCaP, C4-2, and CWR22RV1 cells treated with DHT (Fig. 1B) and knocking down endogenous AR expression using small interfering RNA (siRNA) could significantly suppressed androgen-induction of Slug protein expression (Fig. 1C).

Fig. 1.

Effect of androgen on the expression of the Slug gene in PCa cells. A, The levels of Slug mRNA (top panel) or protein (low panel) expression in LNCaP, C4-2, and CWR22RV1 cells after DHT treatment for 36 h were determined using qRT-PCR or Western blot. After normalizing with 18S RNA, the relative levels of Slug mRNA from each sample (fold of change) were calculated using nonandrogen treated cells as the control (value of 1). B, The levels of Slug mRNA (top panel) or protein (low panel) expression in LNCaP, C4-2, and CWR22RV1 cells pretreated with 10 μm bicalutamide (Casodex) 30 min before adding 10 nm DHT. After normalizing with 18S RNA, the relative levels of Slug mRNA from each sample (fold of change) were calculated using non-androgen-treated cells as the control (value of 1). C, LNCaP or C4-2 cells were transfected with 20 nm AR siRNA or control siRNA for 24 h and then stimulated with 10 nm DHT for 48 h. The protein levels of Slug or AR were determined using Western blot.

The presence of Slug leading to increased AR protein expression

To investigate the potential effect of Slug on AR signaling, we observed that ectopic expression of Slug in LNCaP cells increased the steady-state levels of AR protein in the absence or presence of androgen (Fig. 2A). On the other hand, knocking down the Slug expression could lead to decreased steady-state levels of AR protein in LNCaP and C4-2 cells (Fig. 2B and Supplemental Fig. 2, C and D). Similar results were observed in other PCa cell lines, such as C4-2 and CWR22RV1 (Fig. 2, C and D). Interestingly, this AR protein accumulation was not due to the transcriptional regulation of AR gene based on quantitative RT-PCR (qRT-PCR) (Supplemental Fig. 2, A and B). These results indicate that Slug can increase AR protein accumulation, which may be due to posttranscriptional regulation.

Fig. 2.

Effect of Slug on AR protein expression in PCa cells. A, After transfecting with different concentrations of pCI-neo-hSlug, LNCaP cells were treated with 10 nm DHT for 36 h and then subjected to Western blot. After normalizing with actin in each sample, the relative AR protein levels were calculated using control (value of 1). B, After transfecting with Slug siRNA or control siRNA, LNCaP cells were treated with 10 nm DHT for 36 h and then subjected to Western blot. After normalizing with actin in each sample, the relative AR protein levels were calculated using control (value of 1). C and D, After transfecting with different concentrations of pCI-neo-hSlug, C4-2 or CWR22RV1 cells were treated with 10 nm DHT for 36 h and then subjected to Western blot. After normalizing with actin in each sample, the relative AR protein levels were calculated using control (value of 1).

Interaction of Slug with AR at the DNA-binding domain (DBD) region

Most of Slug protein induced by DHT was detected in the nucleus of the cells (Fig. 3A). Thus, it is possible that Slug may interact with AR and then enhance AR activities in PCa cells. In Fig. 3B, the data clearly showed that the complex formation between endogenous Slug and AR can be detected in C4-2 cells, and an increased Slug formed a complex with AR in the C4-2 cells as well (Fig. 3C). To further determine the domain of AR that mediates the interaction with Slug, we performed a coimmunoprecipitation assay using flag-tagged AR domain constructs and Slug protein. The results from Fig. 3D showed that the DBD region (amino acids 559–624) of the AR is a key interactive domain with the Slug protein.

Fig. 3.

Determination of the interactive domain of AR with Slug. A, C4-2 cells were treated with 10 nm DHT for 36 h then subjected to immunostaining. DAPI, 4′,6-Diamidino-2-phenylindole. B, Total lysates of C4-2 cells cultured in normal condition were immunoprecipitated with anti-AR antibody, anti-Slug antibody, or IgG and then detected with anti-Slug or anti-AR antibodies. C, Left panel, C4-2 cells were treated with 10 nm DHT for 36 h and then subjected to nuclear protein extraction for IP with the indicated antibodies. Right panel, C4-2 cells were transfected with pPGS-hSlug.fl.flag or vector control (VC) for 48 h in the absence of androgen and then subjected to IP with the indicated antibodies. D, Human embryonic kidney 293 cells were cotransfected with pCI-neo-hSlug and flag-tagged AR vectors containing different domains (AR-F, AR-L, AR-D, AR-N-D) for 48 h and then subjected to IP with Slug antibody and Western blot analysis. AR-F, Full-length; AR-L, ligand-binding domain; AR-D, DBD domain; AR-N-D, NTD-DBD domain; NTD, N-terminal domain.

The effect of Slug on AR transcriptional activity

To examine the possible effect of Slug on AR transcriptional activity, we performed reporter gene assay using the ARE reporter gene construct in several PCa cell lines with or without androgen administration. As shown in Fig. 4A, in three different PCa cells (i.e. LNCaP, C4-2, and CWR22v1), ectopic expression of Slug increased not only ARE reporter gene activities without androgen but also androgen-elicited ARE reporter gene activities in a dose-dependent manner. In addition, in LNCaP cells without androgen administration, Slug could elicit ARE promoter gene activities in a dose-dependent manner (Supplemental Fig. 3). Moreover, knocking down the endogenous Slug expression with siRNA resulted in the reduction of DHT-elicited ARE promoter gene activities in LNCaP cells (Fig. 4B). Results from qRT-PCR analysis indicated a specific synergistic effect of Slug on androgen-regulated genes, such as PSA and TMPRSS2 (Fig. 4C) because Slug showed no effect on the transcription of the CDH1 gene, a typical Slug downstream target gene in EMT in LNCaP models at the same condition (Supplemental Fig. 3B). Furthermore, knocking down AR could decrease PSA or TMPRSS2 mRNA expression in LNCaP-Slug cells (Supplemental Fig. 3, C and D), indicating that the effect of Slug on PSA and TMPRSS2 gene transcription is AR dependent. On the other hand, knocking down endogenous Slug could result in decreased PSA and TMPRSS2 mRNA expression (Fig. 4D). Together, these data indicate that Slug is a potent AR coactivator to enhance its transcriptional activity.

Fig. 4.

Slug increased the AR transcription activity. A and B, Cells were cotransfected with pGL2-(ARE)_3-tk-luc vector, different concentrations of pCI-neo-hSlug (A), or Slug siRNA (B), and β-gal vector for 24 h. Twenty-four hours after incubating with 10 nm DHT, cells were subjected to luciferase and β-gal assay. After normalizing with the β-gal activity, the relative ARE-luc activity was calculated. Each result was carried out in triplicate. C and D, LNCaP cells were transfected with pCI-neo-hSlug (C) or Slug siRNA (D) for 6 h and then treated with 10 nm DHT for 48 h. Total RNA was extracted to analyze PSA and TMPRSS2 mRNA levels. The relative mRNA level from each gene was determined by normalizing 18S RNA.

The reciprocal interaction between Slug and AR splice variants

Recent studies indicate that the presence of constitutively active AR splice variants could facilitate the development of CRPC (10–14). Thus, we examined the impact of these variants on the induction of Slug. Indeed, different AR variants (ARv7, AR3, and ARv567es) could induce Slug expression in LNCaP and C4-2 cells (Fig. 5, A and B). In addition, we also noticed the androgen-induced Slug expression in C4-2 cells, a castration-resistant cell line developed from LNCaP, is much more prominent than that in LNCaP under castrated level of androgen (Supplemental Fig. 4), suggesting that AR hypersensitivity could lead to higher Slug expression in CRPC. Functionally, the elevated Slug was able to enhance different AR splice variant-mediated ARE reporter gene activities in the absence of androgen (Fig. 5C). Together, these data suggest that a cooperative effect between Slug and AR variants could be one of the underlying mechanisms of hyperactive AR in CRPC.

Fig. 5.

Reciprocal interaction between Slug gene expression and AR splice variants in PCa cells. A and B, The induction of Slug expression by AR splice variants in PCa cells. The expression of Slug in PCa cells transfected with different AR variant plasmids (ARv7, AR3, and ARv567es) for 48 h was determined using qRT-PCR (A) or Western blot (B). *, P < 0.05 vs. vector control (VC). C, The effect of Slug on the transcription activities of AR splice variants. Human embryonic kidney 293 cells were transfected with different AR variant plasmids (ARv7, AR3, and ARv567es) and increased amount of pCI neo-hSlug, pGL2-(ARE)_3-tk-luc and β-gal vector for 48 h, and then reporter gene activities were determined using luciferase assay. After normalizing with the β-gal activity, the relative ARE-luc activity was calculated. *, P < 0.05 vs. vector control.

The effect of Slug on the growth and survival of PCa cells in vitro and in vivo

To examine whether Slug can promote androgen-independent growth of PCa, stable transfection of Slug cDNA into an androgen-responsive LNCaP cell line was generated. Elevated Slug expression facilitated the in vitro growth of LNCaP cells under regular culture condition and in vivo growth in LNCaP tumor as well as tumor take rate (75%, Fig. 6A). LNCaP-Slug tumors were observed within 21 d after the injection, whereas no tumor was found from control cells (tumor take rate 33.3%) 32 d after the injection. The overall growth of LNCaP-Slug was significant higher than control cells (Fig. 6B). In addition, Slug was able to promote both in vitro growth and colony formation of PCa cells under hormone-free conditions (Fig. 6C). Consistent with this observation, PCa tumor-expressing Slug became hormone resistant (Supplemental Fig. 5, A and B). In addition, LNCaP-Slug cells could form tumors and grow in the precastrated mice (tumor take rate in seven of 11), whereas LNCaP did not form the detectable tumors in the same condition (Fig. 6D).

Fig. 6.

Slug promoted PCa cell growth under different hormonal condition. A, Cell growth of LNCaP stable [LNCaP vector control (VC) and LNCaP Slug], and parental cells were monitored by MTT assay in vitro. B, Subcutaneous xenograft assay was applied to monitor tumor growth of LNCaP VC and LNCaP Slug cells in vivo. The tumor volume of xenografts was measured at d 32 after the injection. C, Cell growth and colony formation of LNCaP stable (LNCaP VC and LNCaP Slug) and parental cells under hormone-free condition in vitro were monitored by MTT assay (left panel) and colony formation assay (right panel). D, The tumor volume of sc xenografts was measured for 8 wk after injection into the precastrated mice. *, P < 0.05 vs. LNCaP VC cells.

Elevated Slug in CRPC tissues and its correlation with nuclear staining of AR

Slug expression has not been examined in human PCa. Using a Slug-specific antibody for staining as previously reported, we performed immunohistochemical (IHC) analysis on tissue microarrays (TMA) containing 400 human PCa specimens.

However, the Slug staining pattern in PCa tissues is not consistent with previous study using the same antibody in other tumor types (27). In general, Slug staining in normal or benign prostate epithelial cells is very low. Hormone-naive PCa (Fig. 7A, left panel) showed a weak Slug staining, whereas CRPC (Fig. 7A, right panel) showed a strong Slug staining in the cytoplasm and nucleus. Overall, the majority of prostatic intraepithelial neoplasia and PCa tissues exhibited a significantly higher Slug expression compared with benign tissue (Fig. 7B, P < 0.05). Notably, both Slug expression and AR staining increased after neoadjuvant hormone therapy (NHT) and CRPC tumors (Fig. 7, C and D, P < 0.05). Similar results were also found from Gene Expression Omnibus data sets (7, 28). For example, Slug mRNA levels in metastatic or CRPC were higher than localized or hormone-naive PCa (Supplemental Fig. 6, A and B). Also, from the xenograft model (7), both AR and Slug elevated once PCa became castration resistant (Supplemental Fig. 6C). We further analyzed the correlation of Slug and AR expression in PCa, and data clearly indicated a positive correlation between Slug and AR in all three TMA arrays (overall P = 6.4e-08; Table 1; Gleason, correlation = 0.34, P = 6.6e-07; Table 2; CRPC, correlation = 0.30, P = 0.041; Table 3: NHT, correlation = 0.21, P = 0.0089). Collectively, these data support the notion that Slug is an androgen-regulated gene and has a potential role in increasing AR protein expression and promoting the onset of CRPC.

Fig. 7.

Correlation of Slug and AR expression in human PCa specimens. A, Representative IHC staining of AR and Slug in clinical specimens. Case 1, Low expression of AR and Slug in primary PCa with low Gleason score; case 2, high expression of AR and Slug in CRPC. The scale bar represents 50 μm. B, Histogram of Slug protein expression in PCa (n = 120) compared with the benign tissue (n = 32) (PCa vs. benign tissue, P = 0.0409). C, Histogram of Slug protein expression in PCa after NHT (n = 76) or recurrence (n = 21) (naive vs. NHT, P = 0.0284; naive vs. CPRC, P = 0.0296). The differences between two groups were compared by the Student's t test. D, Histogram of AR protein expression in PCa after NHT or recurrence (naive vs. NHT, P = 0.0012; naive vs. CRPC, P < 0.0001; NHT vs. CRPC, P = 0.0143). The differences between two groups were compared by the Student's t test.

Table 1.

The correlation between Slug and AR staining from the TMA-Gleason TMA set

Variable		Slug
		0	1	2	3
AR	0	3	9	11	1
	1	4	26	24	9
	2	0	13	30	24
	3	0	1	3	5

The number of patient samples with different staining score in each TMA array is listed. Kendall's tau rank correlation and P value were applied to analyze the correlation between Slug and AR staining. The overall P = 6.4e-08. Correlation = 0.34, P = 6.6e-07.

Table 2.

The correlation between Slug and AR staining from the TMA-CRPC TMA set

Variable		Slug
		0	1	2	3
AR	0	1	2	0	0
	1	0	5	1	0
	2	0	4	10	2
	3	1	4	6	2

The number of patient samples with different staining score in each TMA array is listed. Kendall's tau rank correlation and P value were applied to analyze the correlation between Slug and AR staining. The overall P = 6.4e-08. Correlation = 0.30, P = 0.041

Table 3.

The correlation between Slug and AR staining from the TMA-NHT TMA set

Variable		Slug
		0	1	2	3
AR	0	0	0	1	0
	1	0	18	22	3
	2	0	14	32	11

The number of patient samples with different staining score in each TMA array is listed. Kendall's tau rank correlation and P value were applied to analyze the correlation between Slug and AR staining. The overall P = 6.4e-08. Correlation = 0.21, P = 0.0089.

Discussion

The transition from androgen-dependent to castration-resistant state after androgen ablation therapy signifies an end stage of PCa, which represents a major obstacle to control this disease because the underlying mechanisms are still not fully understood (29). Nevertheless, most of studies have revealed that the AR gene, still expressed in most CRPC (30), undergoes mutation or expresses splice variants and becomes hyperactive. Thus, this hyperactive AR signaling axis and its interactive proteins are the major focus for the development of effective therapeutic regimen. This study unveiled Slug as a unique androgen-regulated gene and coactivator with a novel functional role in regulating the AR protein expression and enhancing the AR activities.

The potential roles of Slug in the development of cancer metastasis through transcriptional repression of E-cadherin and induction of EMT have been documented (15, 20–22, 31). Liu et al. (32) reported that Slug protein was highly expressed in PCa from transgenic adenocarcinoma of mouse prostate mice but not in normal prostate tissue from wild-type mice. Nevertheless, its expression and biological significance in human PCa development remains to be determined. Using the TMA of primary PCa, a very recent study has reported that Slug was increased at pathologically graded stage III and stage IV PCa in which invasive tumor cells had spread beyond the prostate (31). In contrast, Urbanucci et al. (33) recently showed that Slug mRNA was decreased in both PCa and CRPC by RT-PCR assays. For the first time, our data (Fig. 7) demonstrated that overexpression of Slug protein was detected in PCa specimens, particularly in CPRC, which also correlated with AR nuclear staining because of a reciprocal interaction between Slug and AR. Also, Slug overexpression not only increased both in vitro and in vivo growth of PCa but also provided a survival advantage for PCa during the androgen-deprived condition (Fig. 6). Obviously, Slug plays a critical role in the growth of CRPC.

Although Slug is involved in EMT, cell self-renewal, and survival of different cancer types (24), the regulation of Slug is not fully characterized. In one study, Slug can be transcriptionally regulated by p53 upon irradiation and then protects the damaged cell from apoptosis by directly repressing p53-mediated transcription of puma (23). In addition, Slug also can be regulated through murine double minute-2-mediated ubiquitylation and protein degradation, and it is not transcriptionally regulated by the wild-type p53 (27). Here we defined Slug as an androgen-regulated gene. Elevated Slug expression is detected in CRPC cells treated with castrated level of androgen (Supplemental Fig. 4) and also in the presence of constitutively active AR without ligand-binding domain (Fig. 5A), often detected in CRPC, can increase Slug expression under androgen-deprived condition. Our data provide a new insight of the gene regulation of Slug, particularly in PCa cells.

Like many other steroid hormone nuclear receptors, AR translocates into the nucleus upon androgen stimulation and binds to ARE to modulate gene transcription (34). Similar to the AR, we found that most of the Slug protein was located in the nucleus upon androgen administration, and Slug can form a complex with AR and sensitize AR transcriptional activity as a typical AR coactivator (Figs. 3 and 4). Interestingly, Slug could increase AR protein expression in PCa cells (Fig. 2). This reciprocal regulation between Slug and AR and subsequent complex formation of Slug and AR synergize AR transcriptional activity, which consequently provides a growth or survival advantages for PCa during hormone therapy. This event appears to be highly dependent on AR because earlier studies reported that Slug were not required for cell proliferation of AR-negative PCa cells, such as PC-3 and DU-145 (32).

Taken together, Slug cooperates with AR and plays a critical role in the onset of CRPC by enhancing AR protein expression and AR activity under castration condition. Therefore, Slug could be a potential prognostic marker and drug target in CRPC.

Materials and Methods

Cell culture and clinical specimens

All human prostate cancer cell lines were obtained from American Type Culture Collection (Manassas, VA). LNCaP, CWR22RV1, and DU145 cells were maintained in RPMI1640 medium (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (FBS; Invitrogen). C4-2 and PC-3 cells were maintained in T medium (Invitrogen) containing 5% FBS. Human embryonic kidney 293 cells were maintained in DMEM (Invitrogen) containing 10% FBS. For androgen administration, cells were switched to phenol-red free RPMI 1640 medium containing 5% charcoal-stripped FBS (Hyclone, Logan, UT) for 24 h before adding androgen (i.e. DHT).

This study was done on the total of 194 prostate cancer specimens obtained from the Vancouver Prostate Centre Tissue Bank (University of British Columbia). Seventy-six of those cases were subjected to NHT. The hematoxylin and eosin slides were reviewed and the desired areas were marked on them and their correspondent paraffin blocks. Three TMA were manually constructed (Beecher Instruments, Silver Spring, MD) by punching duplicate cores of 1 mm for each sample. All the specimen were from radical prostectomy except 12 CRPC samples. Tissue samples were classified arrayed according to Gleason score, primary or CRPC status, and with or without NHT, respectively (35, 36). The Institutional Review Board of the University of Texas Southwestern approved the tissue procurement protocol for this study, and appropriate informed consent was obtained from all patients.

Plasmid constructs and antibodies

The Slug expression vector pCI-neo-hSlug and Flag-tagged pPGS-hSlug.fl.flag expression vector were obtained from Drs. Pan-Chyr Yang (Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan) (16) and Eric R. Fearon (The University of Michigan, Ann Arbor, MI) (21), respectively. The AR splice variants (ARv7, AR3, and ARv567es) were obtained from Drs. Jun Luo (Johns Hopkins Medical School, Baltimore, MD), Yun Qiu (University of Maryland School of Medicine, Baltimore, MD), and Stephen R. Plymate (University of Washington, Seattle, WA), respectively. The Flag-tagged AR truncated expression vectors (AR-F, AR-L, AR-D, AR-DN) were obtained from Dr. Chawnshang Chang (University of Rochester, Rochester, NY). The pCMV-β-galactosidase vector (β-gal) was provided by Dr. Ching-Hai Kao (Indiana University, Indianapolis, IN). Cells (5 × 10⁵ cells/well) were seeded in a six-well plate (Costar, Corning, NY) with 70–80% confluence before transfection. Cell transfection was carried out using Lipofectamine LTX with PLUS (Invitrogen) according to the manufacturer's instructions.

Primary antibodies used in this study were as follows: Slug (G-18) (Santa Cruz Biotechnology, Santa Cruz, CA); AR, Flag (M2), and actin (Sigma-Aldrich, St. Louis, MO); and Slug (Abgent, San Diego, CA) and AR (N-20) (Santa Cruz Biotechnology) for IHC.

siRNA oligonucleotides and transfection

Three siRNA oligonucleotides for human Slug [5′-CCAUUCUGAUGUAAAGAAATT-3′ and 5′-UUUCUUUACAUCAGAAUGGGT-3′ (1); 5′-CAUUAGUGAUGAAGAGGAATT-3′ and 5′-UUCCUCUUCAUCACUAAUGGG-3′ (2); 5′-AGUGCAAUUUAUGCAAUAATT-3′ and 5′-UUAUUGCAUAAAUUGCACUGA-3′ (3)] and control siRNA were purchased from Invitrogen. AR siRNA and its control siRNA used in this study were described previously (37). Transfection was carried out using Lipofectamine RNAiMax (Invitrogen) according to the manufacturer's protocol.

Luciferase reporter gene assay

Cells were seeded in 24-well plates and then transfected with plasmid constructs for 6 h. After transfection, cells were subjected to androgen administration for 24 h. Luciferase assays were carried out as previously described (38). Each result was normalized with β-gal activity and presented as mean relative light units (RLU) ± sd in triplicates.

RNA isolation and quantitative RT-PCR analysis

Total cellular RNA was extracted with an RNeasy minikit (QIAGEN, Gaithersburg, MD), and 1 μg RNA was subjected to a cDNA synthesis kit (Bio-Rad Laboratories, Hercules, CA). One tenth of the cDNA was subjected to a 25-μl PCR carried out in an iCycler thermal cycler (Bio-Rad Laboratories) using iQ SYBR Green Supermix (Bio-Rad Laboratories). The primer sequences used were as follows: Slug, forward, 5′-GGGGAGAAGCCTTTTTCTTG-3′; reverse, 5′-TCCTCATGTTTGTGCAGGAG-3′; AR, forward, 5′-AGGAACTCGATCCTATCATTGC-3′; reverse, 5′-CTGCCATCATTTCCGGAA-3′; PSA, forward, 5′-GACCAAGTTCATGCTGTGTG-3′; reverse, 5′-ACTAGGGAGCCATGGAGGAC-3′; TMPRSS2, forward, 5′-GTGATGGTATTCACGGACTGG-3′, reverse, 5′-CAGCCCCATTGTTTTCTTGTA-3′; CDH1 (E-cadherin), forward, 5′-TGCCCAGAAAATGAAAAAGG-3′; reverse, 5′-GTGTATGTGGCAATGCGTTC-3′; 18S, forward, 5′-GGAATTGACGGAAGGGCACCACC-3′; reverse, 5′-GTGCAGCCCCGGACATCTAAGG-3′. qRT-PCR was performed as described previously (38).

IHC staining

The EnVision system (DAKO Corp., Carpinteria, CA) was used for IHC staining according to the protocol recommended by the manufacturer. The tissue sections were stained with specific antibodies against Slug (1:75) and AR (N-20, 1:50). Hematoxylin was used for counterstaining. The histology was assessed independently by two pathologists, and a consensus of grading was reached using a two-score system based on intensity score and proportion score as described (39).

Immunofluorescence staining

Cells were fixed in 4% paraformaldehyde for 15 min at room temperature and washed three times with PBS and then permeabilized with ice-cold 100% methanol for 10 min. The slides were blocked in PBS containing 0.3% Triton X-100 and 5% normal donkey serum for 1 h at room temperature. Primary antibodies were incubated for 1 h at room temperature. After washing with PBS, cells were incubated with secondary antibodies for 45 min at room temperature. Finally, cells were counterstained with 4′,6-diamidino-2-phenylindole before mounting.

Immunoprecipitation (IP) and Western blot analysis

For IP, cells were washed twice with cold PBS and lysed in 0.5 ml radioimmunoprecipitation assay lysis buffer (25 mm Tris-HCl, pH 7.6; 150 mm NaCl; 1% Nonidet P-40; 1% sodium deoxycholate; 0.1% sodium dodecyl sulfate) for 30 min on ice or subjected to nuclear and cytoplasmic extraction (NE-PER; Pierce Biotechnology, Rockford, IL). The cell lysates were incubated with a primary antibody. The immunocomplexes were precipitated with Dynabeads Protein G (Invitrogen) and then subjected to a Western blot analysis.

In vitro growth assay and colony formation assay

For in vitro cell growth assay, cells were seeded at the density of 5 × 10³ cells/well in 96-well plates overnight then treated with DHT (10 nm) or chemotherapeutic drugs for the indicated time. Growth was monitored using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Roche, Indianapolis, IN).

A total of 1000 cells per well were seeded in six-well plates. The medium was switched to phenol-red free RPMI 1640 medium containing 5% charcoal-stripped FBS after 24 h for 2 wk; fresh medium was changed every 3–4 d. The plates were then washed with ice-cold PBS, fixed with 4% paraformaldehyde, stained in crystal violet solution for 15 min at room temperature, and washed with distilled water to remove the excess dye. The number of colonies was counted for each sample.

Xenograft animal model

For the sc tumor model, 2 × 10⁶ cells mixed with Matrigel [1:1 (vol/vol)] were injected into both flanks of severe combined immunodeficient (SCID) mice. Tumor volume (cubic millimeters) was measured weekly by caliper and calculated by using the ellipsoid formula (π/6 × length × width × depth) for 8 wk. Animals bearing tumors were castrated when they reached the same primary tumor size (∼1000 mm³) after cell injection. In addition, the SCID mice were precastrated for 3 d before sc injection of 5 × 10⁶ two cells, and the tumor uptake and growth in precastrated SCID mice were monitored for 8 wk. All experimental procedures have been approved by the Institutional Animal Care and Use Committee.

Statistical analysis

Data are presented as the mean ± sem from at least three independent experiments, and the differences between two groups were compared by the Student's t test. All data analyses were done by software of SPSS13.0 for Windows (SPSS Inc., Chicago, IL). P < 0.05 was regarded as the threshold value for statistical significance.

NURSA Molecule Pages^†:

• Nuclear Receptors: AR;

• Ligands: Dihydrotestosterone.