Twenty-eight tumor specimens from patients were included in the present study, including 8 specimens of castration-resistant prostate cancers (CRPC). These paraffin-embedded human primary PCa specimens were acquired from patients who provided informed consent at Chonnam National University Hospital between 1997 and 2005. The tumors used in the present study were acquired by transurethral resection of the prostate. All cases had clinical follow-up of at least 10 years (Table I). Immunohistochemical studies were performed on both paraffin-embedded tissue sections of human hormone dependent prostate cancers (HDPC, 20 specimens) and CRPC (8 specimens) as previously described (7). The antibodies used were anti-p21 monoclonal antibody (DCS60; Cell Signaling Technology, Danvers, MA, USA) and anti-HOXB13 antibodies. The stained slides were evaluated by two different investigators, including a pathologist, who were blinded to the patient's clinical features. The intensity of the staining was classified into one of the two grades (+, positive; −, negative).

The synthetic testosterone R1881, was purchased from NEN Life Science (Boston, MA, USA) and used at a final concentration of 10 nM (nmol/l). Both fetal bovine serum (FBS) and charcoal dextran-treated (CDT) FBS were purchased from Invitrogen (Carlsbad, CA, USA). The pFLAG-p21 and pCDNA-FLAG-p21 were generated by PCR cloning. The pFLAG-HOXB13 was previously described (11). Antibodies to JNK, phosphorylated JNK (Thr183/185), ERK, AKT, STAT3 and p53 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-FLAG M5, c-Fos, c-Jun, phosphorylated c-Jun (Ser73) and β-actin antibodies were purchased from Sigma-Aldrich (St. Louis, MO, USA). Antibodies against p21^WAF1/CIP1 were purchased from Upstate Biotechnology (Lake Placid, NY, USA). The dominant negative c-Jun mutant was kindly provided by Dr Kyung Keun Kim (Chonnam National University, Gwangju, Korea). An inhibitor of c-Jun N-terminal kinase (SP600125) was obtained from Sigma-Aldrich. Reporter vectors, including AP1-luc and c-Jun-luc, were purchased from Stratagene (Santa Clara, CA, USA).

Human prostate cell lines, including LNCaP and PC3, were routinely cultured in RPMI media (Invitrogen, Carlsbad, CA, USA) supplemented with 5% FBS at 37°C in an atmosphere containing 5% CO₂. HOXB13-suppressed LNCaP and HOXB13-overexpressed PC3 cells have been previously described (7). Wild-type HCT116-p21(+/+) and HCT116-p21(−/−) were kindly provided by Dr Bert Vogelstein. All cultures were fed with fresh medium every 3–4 days.

To determine the hormone effect, the cells were grown under 5% CDT-FBS for two days. The next day, cells were treated with either R1881 or ethanol and were grown for up to 7 days. After incubation for 24 h, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide in vitro proliferation assay was performed as previously described (11). Briefly, the cells were stained with 100 µl of 5 mg/ml MTT (Sigma-Aldrich) solution and incubated for 4 h at 37°C. The reaction was stopped and then the absorbance at 570 nm was measured using a microplate reader with SoftMax PRO software (Molecular Devices, Sunnyvale, CA, USA). Densitometric values were analyzed with the Student's t-test, using GraphPad Prism software (San Diego, CA, USA).

The cells were grown to 80% confluence and then lysed in protein extraction RIPA buffer (25 mM Tris-HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate and 0.1% SDS) containing a cocktail of protease inhibitors and phosphatase inhibitors. Twenty micrograms of total cell lysates were loaded onto a 10% Bis-Tris gel and separated using the electroporation system (Bio-Rad Laboratories, Hercules, CA, USA). After the proteins were transferred to a PVDF membrane, primary antibodies were applied, followed by incubation with horse peroxidase-conjugated secondary antibodies. The blots were developed by the ECL detection system (Thermo Fisher Scientific Inc., Rockford, IL, USA).

Approximately 1×10⁵ cells were plated in a 24-well plate 16 h before transfection. Each transfection was carried out using Lipofectamine 2000 (Invitrogen) with 0.1 µg of reporter and 2 ng of Renilla luciferase reporter as described in the manufacturer's protocol. Six hours later, the cells were washed and fed with medium containing 5% FBS. Addition of SP600125 was performed 18 h post-transfection. Subsequently, the cells were washed with PBS, harvested for 24 h, lysed with 100 µl of passive lysis buffer. Then luciferase activities were assayed as relative light units (RLU) using the Dual-Luciferase reporter assay system (Promega, Madison, WI, USA) as per the instructions. The transfection experiments were performed in triplicate and the results are reported as the mean ± SD. For each transfection, Renilla luciferase reporter was used to normalize for differences in transfection efficiency. The RLU from the experimental group were normalized to the control group and the values were represented as fold change.

Chromatin immuno-precipitation was performed using antibodies against c-Jun, HOXB13 or IgG as previously described (12). The immuno-precipitated DNA was amplified by using specific primers as follows: p21 (forward, ctcacatcctccttcttcag and reverse, cacacacagaatctgactccc).

The data were expressed as mean ± standard deviation and processed using the SPSS software 17.0 (SPSS, Inc., Chicago, IL, USA). The Pearson's correlation coefficient test was used to estimate the correlation between tumors and HOXB13 and p21. Statistical significance was set at P-value <0.05.

Our previous studies demonstrated that HOXB13 inhibits p21 expression without clearly indicating the underlying mechanism (7,8). Since HOXB13 is predominantly overexpressed in CRPC, we screened the expression of both HOXB13 and p21 in selected prostate cancers including CRPC in order to demonstrate correlated expression of these two proteins. Immunohistochemistry was performed on formalin-fixed and paraffin-embedded tumor specimens. Fig. 1 shows negatively correlated expression of HOXB13 and p21 in some selected tumors. Intensity of HOXB13 staining was classified into one of the four grades (0, absent; 1, weak; 2, intermediate; and 3, strong) to statistically correlate the expression of HOXB13 and p21. We further classified the grades as follows: 0–1 as negative and 2–3 as positive. Due to the heterogeneous nature of prostate cancer and small sample size, there were no statistically significant correlations between HOXB13 and p21 expression (Table II). However, HOXB13-deficient tumors had higher odds for expressing p21 than HOXB13-positive tumors with negative p21 expression. Although the sample size was small resulting in lack of statistical significance, CRPC showed a more negative correlation between HOXB13 and p21 than HDPC.

To demonstrate the growth-regulating role of p21 in the absence of androgen, androgen-responsive LNCaP PCa cells were transfected with either pCDNA or pCDNA-p21. Cells were grown up to 7 days in the presence of androgen (final concentration of 1 nM R1881) followed by MTT in vitro proliferation assays. As shown in Fig. 2, addition of R1881 promoted proliferation of LNCaP cells, while p21 generally suppressed cell proliferation. Noticeably, androgen did not stimulate cell proliferation in the presence of p21, suggesting that p21 suppression is an important step in survival of PCa cells under androgen-deprived conditions.

To explore the detailed mechanism by which HOXB13 suppresses p21 expression, PC3 cells stably transfected with HOXB13 were used for western blot analysis. HOXB13 suppressed not only p21 expression but also c-Jun N-terminal kinase (JNK)-related proteins (Fig. 3A). There were no significant alterations in extracellular signal-regulated kinase (ERK), AKT, STAT3 and p53 proteins. While HOXB13 inhibited expression of total forms of JNK and c-Jun and phosphorylated forms of JNK and c-Jun, there was no significant difference in c-fos proteins. JNK is known to bind and phosphorylate the DNA binding protein c-Jun, which is a central component of the AP-1 family of transcription factors. AP-1 complex consists of homodimers or heterodimers of c-Jun and other transcription factors such as c-fos. Consequently, the reporter transcription assay demonstrated that HOXB13 suppressed c-Jun and AP-1 mediated transcription (Fig. 3B). Suppression of HOXB13 in LNCaP cells also indicated stimulation of AP-1 activity (Fig. 3C).

To confirm the effect of HOXB13 on c-Jun and AP-1 activity, TAM67 (dominant-negative c-Jun) and SP600125 (pharmacological inhibitor of JNK) were used in the reporter transcription assay. HOXB13 further downregulated AP-1 activity suppressed by TAM67 (Fig. 4A). Treatment with SP600125 blocked not only the activation of c-Jun but also the activation of p21 (Fig. 4B). Chromatin immunoprecipitation assay demonstrated that HOXB13 accommodated less amount of c-Jun on the p21 promoter region while HOXB13 association did not change (Fig. 4C). These results suggest that HOXB13 affects both AP-1 activity and p21 activity and there might be a mutual communication between these two proteins.

c-Jun negatively regulates transcription of p53 by binding to multiple AP-1 sites in the p53 promoter and represses its target p21 expression (13). As shown in Fig. 3A, however, HOXB13 suppressed the expression of both p21 and c-Jun, implying that there is no negative correlation between c-Jun and p21. To assess if there is c-Jun mediated regulation of p21 activity, p21-deficient HCT116 cells were tested for c-Jun activity. Compared to wild-type cells, p21(−/−) cells showed inhibited activity of c-Jun (Fig. 5A) and AP-1 (Fig. 5B). A further study demonstrated that exogenous p21 stimulated c-Jun activity in a dose-responsive manner in HOXB13-deficient PC3 cells, while the effect of p21 on c-Jun activity was minimal in HOXB13-overexpressed PC3 cells (Fig. 5C). These results suggest that the cell cycle inhibitor p21 may have a novel role in the regulation of c-Jun signaling in the presence of HOXB13.