Nemo-like kinase as a negative regulator of nuclear receptor Nurr1 gene transcription in prostate cancer

Human PCa specimens were derived from patients undergoing radical prostatectomy, and the benign prostate tissue samples were derived from benign prostatic hyperplasia (BPH) patients. Specimens were formalin-fixed and paraffin-embedded for histopathologic diagnosis and immunohistochemical analysis. Eight paired fresh specimens were frozen in liquid nitrogen immediately following surgical removal and maintained at −80 °C until use (Western blots). Patient clinical features, including age, Gleason score and tumor node metastasis (TNM) staging, are shown in Table 1. This study was approved by the ethics committee of Jinzhou People’s Hospital. Written informed consent for anonymized use of tissue specimens was obtained from all patients.

Human prostate tissues were processed for immunostaining with rabbit anti-Nurr1 polyclonal antibody (2 μg/ml, Santa Cruz, CA, USA) and rabbit anti-NLK polyclonal antibody (4 μg/ml, Abcam, Cambridge, MA, USA) as described previously [16]. Stained tissue sections were scored independently by two pathologists blinded to the clinical parameters. The intensity of staining was evaluated subjectively on a scale of 0–3, where 0 = no staining, 1 = weak equivocal staining, 2 = unequivocal moderate staining and 3 = strong staining. The extent of staining, defined as the percentage of positive staining cells in relation to the whole field, was scored on a scale of 0 to 4 as 0 (≤5 %), 1 (6 to 25 %), 2 (26 to 50 %), 3 (51 to 75 %) or 4 (>75 %). Values for staining intensity and positive area were multiplied to give the final score. For statistical analysis, final staining scores of < 4 and ≥ 4 were considered to be low and high expression levels, respectively.

Three prostate cell lines (LNCaP, PC-3 and BPH-1) were obtained from American Type Culture Collection (Rockville, MD, USA). Cells were cultured in RPMI-1640 media (Invitrogen, Carlsbad, CA, USA) supplemented with 100 IU/mL penicillin G sodium, 100 mg/mL streptomycin sulfate and 10 % (v/v) fetal bovine serum (Hyclone, Logan, UT, USA). Cultures were maintainedat 37 °C in a 5 % CO₂ environment.

NLK expression in PCa cells was knocked down using NLK-specific shRNA plasmid vectors. The shRNA-NLK (target sequence: 5′-GAATATCCGCTAAGGATGC-3′), shRNA-CBP (target sequence: 5′-AACAGTGGGAACCTTGTTCCA-3′) and shRNA-control plasmids were obtained from GenePharma (Shanghai, China). Cells were grown to 70 % confluency in 6-well plates and transfected with 2 μg specific or nonspecific shRNA plasmid using Lipofectamine LTX Plus (Invitrogen, Carlsbad, USA) according to the manufacturer′s instructions. Cells were harvested 48 h after transfection and used in further experiments.

The human NLK cDNA clone was obtained from Sino Biological Inc (Beijing, China). The NLK coding sequence was amplified by PCR using the following primers: 5′-GAGCGGATAACAATTTCACACAGG-3′ (forward) and 5′-CGCCAGGGTTTTCCCAGTCACGAC-3′ (reverse). The resultant PCR fragment was cloned into the pcDNA3.1 expression vector using the HindIII and XbaI restriction sites. Proper insertion of the clone was confirmed by DNA sequencing, and the plasmid was prepared for transfection using the Genopure Plasmid Maxi Kit (Roche, Basel, Switzerland). Transfection was performed using the Lipofectamine LTX Plus (Invitrogen, Carlsbad, USA) according to the manufacturer’s protocol. Cells were grown to 70 % confluency in 6-well plates and transfected with pcDNA3.1-NLK vector (1 μg) or pcDNA3.1 vector (1 μg). Cells were harvested and used in further experiments 48 h following transfection.

Western blots were performed in triplicate as previously described [16]. The following antibodies were used for Western blot: rabbit anti-Nurr1 polyclonal antibody (1:500, Santa Cruz, CA, USA), rabbit anti-NLK polyclonal antibody (1:250, Abcam, Cambridge, MA, USA), rabbit anti-NF-κB p65 polyclonal antibody (1:200, Santa Cruz, CA, USA), mouse anti-IκBα monoclonal antibody (1:1000, Abcam, Cambridge, MA, USA), mouse anti-CREB monoclonal antibody (1:500, Santa Cruz, CA, USA), rabbit anti-phospho-CREB polyclonal antibody (Ser133; 1:100, Santa Cruz, CA, USA), mouse anti-β-tubulin monoclonal antibody (1:500, Santa Cruz, CA, USA), rabbit anti-Histone H3 polyclonal antibody (1:250, Santa Cruz, CA, USA), mouse anti-β-actin monoclonal antibody (1:1000, Santa Cruz, CA, USA) and mouse anti-CBP monoclonal antibody (1:500, Abcam, Cambridge, MA, USA). Band density was measured with Image J software and normalized against internal control levels.

Total RNA was isolated from cultured cells 48 h post-transfection using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Reverse transcription was carried out using the iScript^TM cDNA synthesis kit (Bio-Rad Laboratories Inc., Hercules, CA, USA). Fluorescent real-time PCR was carried out in the Bio-Rad iCycler System (Bio-Rad) with a final volume of 10 μl for each reaction. The following primers were used: 5′-TCCAACGAGGGGCTGTGCG-3′ (forward) and 5′-CACTGTGCGCTTAAAGAAGC-3′ (reverse) for human Nurr1; and 5′-GAAGGTGAAGGTCGGAGTC-3′ (forward) and 5′-GAAGATGGTGATGGGATTTC-3′ (reverse) for human GAPDH (internal control). The PCR program was as follows: initial cycle of 95 °C for 3 min, followed by 40 cycles of 95 °C for 15 s, 59.5 °C for 60 s. PCR products were detected using the fluorescent dye SYBR Green I (BioTeke, Beijing, China), and threshold cycle (C_t) values were generated after every cycle in the run. Fluorescent readings from real-time PCR were quantitatively analyzed by finding the difference between C_t values (delta C_t) of the target and its internal control, with target gene expression determined using the formula 2^{-delta Ct} [25]. This experiment was repeated three times.

A fragment of the human Nurr1 promoter region was amplified by PCR using specific primers (forward 5′-AGTTGGTGGGCACAGAGGAGTATC-3′ and reverse 5′-TCACGGAGGGAGGGAGCAG-3′), then cloned into the pGL3-Basic luciferase reporter vector (Promega, Madison, USA) and sequenced to confirm generation of the pGL3-Nurr1-promoter plasmid. A site-directed mutagenesis kit (Stratagene, La Jolla, CA) was used to mutate the core GA sequence of the NF-κB-binding site or the core AC sequence of the CREB-binding site in luciferase constructs. Cells were transfected with wild-type (wt) or mutant (mt) Nurr1 promoter luciferase constructs (1 μg) and with the effector plasmids pcDNA3.1 (1 μg), pcDNA3.1-NLK (1 μg), shRNA-control (2 μg) or shRNA-NLK (2 μg). The TK renilla luciferase plasmid (2 ng) was transfected under the same conditions to enable normalization of transfection efficiencies. Twelve hours after transfecting with WT-Nurr1 promoter constructs, cells were treated with DMSO (control, 1 % DMSO in PBS), BAY 11–7082 (NF-κB inhibitor, 10 μM), KG-501 (CREB inhibitor, 25 μM) or ICG-001 (CBP inhibitor, 25 μM). Firefly and renilla luciferase activities were measured 48 h after transfection in cell extracts using the Dual-Luciferase Reporter Assay System (Promega, Madison, USA). This experiment was repeated three times.

LNCaP cells grown on coverslips were washed 3 times with PBS, fixed in 4 % paraformaldehyde for 30 min at room temperature and permeabilized with 0.2 % Triton X-100 solution in PBS for 10 min at room temperature. Next, the samples were incubated in 5 % BSA solution in PBS for 1 h at room temperature to saturate non-specific binding sites. Incubation with the primary antibodies, diluted in 5 % BSA solution in PBS, was performed overnight at 4 °C. Primary antibodies were rabbit anti-NLK polyclonal antibody (1:100, Abcam, Cambridge, MA, USA) and mouse anti-CBP monoclonal antibody (1:200, Abcam, Cambridge, MA, USA). The next day, cells were incubated for 1 h at room temperature in secondary FITC-conjugated anti-rabbit and Cy3-conjugated anti-mouse antibodies (1:300, Sigma-Aldrich, St. Louis, MO, USA). Confocal analysis was performed on a Leica TCS SP5 microscope, with excitation wavelengths set to 488 and 543 nm for dual channel imaging.

Five hundred microgramme of protein lysates from LNCaP cells were incubated with 2 μg of rabbit anti-NLK antibody or mouse anti-CBP antibody overnight at 4 °C on an oscillation shaker. Next, 50 μl of suspensions were generated by mixing samplesina 1:1 ratio with protein A-Sepharose beads and then incubated for 2 h at 4 °C with gentle rotation. To ensure specificity of the detected associations, control IPs with rabbit IgG or mouse IgG in place of the primary antibodies were performed under the same conditions. After extensive washing, precipitates were subjected to Western blotting analyses for detection of potential interacting proteins.

Chromatin immunoprecipitations were carried out 48 h after transfection with shRNA-NLK using the MAGnify™ Chromatin Immunoprecipitation System (Invitrogen, Carlsbad, USA) according to the manufacturer’s guidelines. Briefly, LNCaP cells (6 × 10⁶) were fixed with formaldehyde and sonicated. Immunoprecipitation was carried out with 5 μg of mouse anti-CBP antibody (Abcam, Cambridge, MA, USA) or mouse IgG at 4 °C overnight with rotation. The immunoprecipitates were washed and eluted, and purified DNA was obtained. Purified immunoprecipitated DNA and input DNA were used as templates for subsequent real-time PCR. ChIP primer sets were checked for linear amplification and designed to amplify two separate regions of the human Nurr1 promoter: NF-κB element forward (5′-TGGGACAGGAAAAGGGAGTA-3′) and reverse (5′-ACAGATGTCCCATGACACGA-3′), spanning nucleotides −633 to −536; and CREB element forward (5′- AACCACCCAAGCTGGCTAC-3′) and reverse (5′-TGGTATATTTCCGACCTGACG-3′), spanning nucleotides −252 to −163.

Results are expressed as mean ± SD and are representative of at least three separate experiments. The student’s t-test was used to determine statistical differences in the means of two columns. The Pearson χ ² test and Fisher’s exact test were used to analyze the relationship between the expression of NLK and the patient’s various clinicopathological parameters. The nonparametric Spearman’s rank correlation coefficient was used to determine the statistical dependence between the expression of NLK and Nurr1. A p value < 0.05 was regarded as statistically significant. All computations were performed using the SPSS 13.0 software.

To explore the clinical significance of NLK in the occurrence and progression of PCa and further characterize the relationship between NLK and Nurr1, we examined the levels of NLK and Nurr1 using immunohistochemical staining in 118 PCa and 50 benign prostate tissue samples. Representative examples of staining are shown in Fig. 1a (I-IX), which show that epithelial cells from benign prostate gland samples have strong nuclear NLK staining (Fig. 1a IV) and weak Nurr1 staining (Fig. 1a VII), and also that low NLK levels (Fig. 1a VI) correlate with high Nurr1 levels (Fig. 1a IX) in the same PCa specimens (high-grade PCa). Correlation analysis demonstrated a significant negative correlation between NLK and Nurr1 expression levels in PCa tissue specimens (Fig. 2). Furthermore, we investigated the abundance of NLK and Nurr1 in eight tumors relative to the adjacent normal tissues (Fig. 1b) by Western blot. The results indicate that compared with the non-tumorous adjacent tissue, NLK expression was dramatically lower and Nurr1 expression much higher in the tumor tissues. To further characterize the relationship between NLK and Nurr1, we investigated their abundance in a normal human prostate epithelial cell line (BPH-1) and two human prostate cancer cell lines (PC-3 and LNCaP) by Western blot analysis. Different expression levels of NLK and Nurr1 were observed in all of the cells (Fig. 1c). As expected, relative abundances of NLK and Nurr1 appeared to be inversely correlated in BPH-1, PC-3 and LNCaP cells. PC-3 cells displayed the lowest abundance of NLK and the highest expression of Nurr1 among the three cell lines.

The frequencies of NLK immunostaining in relation to the different clinicopathological parameters and Nurr1 expression levels are shown in Table 1. For statistical analysis of NLK and Nurr1 expression, the specimens were divided into two groups (high expressers and low expressers), according to staining intensity and the percentages of positive cells. The level of NLK protein was significantly less in the 118 PCa tissue samples compared to the 50 benign prostate tissue samples (p < 0.001). In our PCa tissue specimens, the frequency of NLK low expression increased with progressive T classification (p = 0.007), N classification (p = 0.015) and Gleason score (p < 0.001), though no correlation was observed with other prognostic factors such as age and M classification. As expected, NLK expression was negatively correlated with Nurr1 expression in all of the analyzed PCa tissue samples (p < 0.001).

To elucidate the biological importance of NLK in the regulation of Nurr1 gene expression, we used RNA interference to knock down NLK expression in LNCaP cells with a higher level of NLK expression. In contrast, we used transient transfection to upregulate NLK expression in PC-3 cells with lower levels of NLK expression. Levels of Nurr1 protein and mRNA were determined by Western blot analysis and real-time PCR, respectively. As shown in Fig. 3a, b, transfection of the shRNA-NLK vector into LNCaP cells strongly reduced NLK protein levels, whereas shRNA-control vector transfection had no effect on NLK expression. In addition, levels of endogenous Nurr1 mRNA and protein were significantly upregulated by shRNA-NLK transfection. In contrast, NLK protein levels were dramatically increased after transfection of the pcDNA3.1-NLK vector into PC-3 cells, and this upregulation of NLK resulted in a significant downregulation of Nurr1 mRNA and protein expression. These cells were then transfected with the Nurr1 promoter luciferase constructs and either pcDNA3.1-NLK or shRNA-NLK. The results confirm that Nurr1 promoter luciferase activities vary similarly to Nurr1 mRNA and protein levels (Fig. 3c), and strongly suggest that NLK can inhibit Nurr1 promoter activity, thus leading to downregulation of Nurr1 gene expression in human PCa cells.

Analysis of the promoter region of the human Nurr1 gene shows several putative cis elements, including the upstream NF-κB-site and the downstream CREB-site [26]. To determine the role of these binding sites in NLK-mediated inhibition of Nurr1 transcription, we created two mutant Nurr1 promoter luciferase constructs (NF-κB-site mutated or CREB-site mutated), as indicated in Fig. 4. Compared to the wild-type Nurr1 promoter construct, mutation of either NF-κB- or CREB-binding sites significantly impaired the upregulation of Nurr1 promoter activity induced by shRNA-NLK transfection into LNCaP cells (Fig. 4). We also used three inhibitors, including BAY 11–7082 (NF-κB inhibitor), KG-501 (CREB inhibitor) and ICG-001 (CREB binding protein, CBP, inhibitor). Figure 4 shows the upregulation of Nurr1 promoter luciferase activity induced by shRNA-NLK under control conditions and following treatment with BAY 11–7082, KG-501 or ICG-001. All three inhibitors markedly inhibited Nurr1 promoter activation induced by shRNA-NLK, indicating that NF-κB- and CREB-mediated transcriptional activation are required for upregulation of Nurr1 gene expression by shRNA-NLK.

Having demonstrated that NLK inhibits NF-κB- and CREB-mediated transcriptional activation, we sought to determine changes in the protein levels of p65, IκBα, CREB and phospho-CREB in the presence of shRNA-NLK or shRNA-control. As shown in Fig. 5a, protein levels of p65, CREB and phospho-CREB did not change, whereas IκBα levels decreased. We next performed Western blotting to determine the contents of p65 included in cytoplasmic and nuclear extracts collected from shRNA-NLK- and control-transfected LNCaP cells, which showed that knockdown of NLK increased the level of p65 in nuclear extracts (Fig. 5a). Since ICG-001 (CBP inhibitor) can inhibit Nurr1 promoter activation induced by shRNA-NLK (Fig. 4) and CBP is a nuclear transcriptional co-activator of NF-κB and CREB, we suspected that CBP was a target of NLK in LNCaP cells. To test this hypothesis, dual-color fluorescence and co-immunoprecipitation experiments were used to investigate whether interaction occurs between the two proteins. Immunofluorescent microscopy (Fig. 5b) demonstrated co-localization (yellow) of NLK and CBP in LNCaP cells. Their interaction was further confirmed through immunoprecipitation analysis, which detected the CBP protein in anti-NLK immunoprecipitates and the NLK protein in anti-CBP immunoprecipitates from whole cell lysates (Fig. 5c). To determine if CBP is critical for NF-κB- and CREB-mediated transcriptional activation of Nurr1, we used ChIP to examine the endogenous occupancy of CBP at the Nurr1 promoter in the presence and absence of shRNA-NLK. As shown in Fig. 5d, compared with controls, knockdown of NLK markedly increases the recruitment of CBP on both the NF-κB- and CREB-binding regions. Finally, we employed siRNA approach to knock down CBP expression and examined if CBP was involved in regulation of NLK on Nurr1 expression. As shown in Fig. 5e, f, the regulation of NLK on Nurr1 expression was abrogated by knockdown of CBP in LNCaP cells. Taken together, these results support the notion that NLK directly interacts with CBP to regulate transcriptional activation of Nurr1 in LNCaP cells.