The coding region of the human EphA3 cDNA was cloned into the EcoRI and BamHI sites of pcDNA3.1(−)/Myc-His B vector (Invitrogen). The following double-stranded oligonucleotides were cloned into the pGP-U6-Hygro siRNA expression vector (Shanghai GenePharma Co., Ltd., Shanghai, China): EphA3, 5′-CACCGC GGTCAGCATCACAACTAATTTCAAGAGAATTAGTTGT GATGCTGACCGCTTTTTTG-3′. The negative control oligo, 5′-CACCGTTCTCCGAACGTGTCACGTTTCAAGAGAAC GTGACACGTTCGGAGAACTTTTTTG-3′, had no significant homology to any human coding cDNA. All oligos were inserted into the pGP-U6-Hygro vector, and then transiently transfected into 293T cells to examine the efficiency of RNAi by western blot analysis.

LNCaP and C4-2B cell lines were a generous gift from Leland Chung (Emory University, Atlanta, GA, USA). The LNCaP-PC-1 cell line was maintained in this laboratory, cultured in RPMI-1640 (Invitrogen) with 10% fetal bovine serum (FBS; HyClone). All cells were cultured at 37°C, with 5% CO₂. The EphA3 cDNA vector was transfected into LNCaP cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol. An empty vector was used as a control. Transfected LNCaP cells were selected in the presence of 1,000 μg/ml G418 (Sigma). Stable clones were selected after 5 weeks, and clones overexpressing EphA3 were examined by western blot analysis.

Total RNA from cultured LNCaP-PC-1 or control LNCaP-Neo cells was isolated individually using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and quantified spectrophotometrically at 260/280 nm. The integrity of all RNA samples was evaluated on a 1% agarose gel, and all RNA samples were found to be pure with no degradation caused by the isolation procedure. Affymetrix GeneChips (Human Genome U133 Plus 2.0 arrays; Affymetrix, Santa Clara, CA, USA) were used for hybridization and data collection. The protocol was performed by the Microarray Facility at the Shanghai Biochip Co., Ltd., Shanghai, China.

To obtain growth curves, 2–3×10³ cells/well were seeded in 96-well plates. Cell growth was measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays according to the manufacturer’s instructions (Amresco). Before testing, 20 μl of MTT reagent (2.5 mg/ml MTT in PBS; Amresco, Inc., Solon, OH, USA) was added, and the cells were incubated for a further 4 h at 37°C. Then 150 μl of dissolving reagent DMSO (Amresco) was added to dissolve the formazan crystals. The absorbance was detected at a wavelength of 490 nm on a microplate reader.

A total of 2–6×10³ cells were resuspended in 3 ml of 0.35% low melting point agarose (BD Biosciences) in RPMI-1640/10% FBS and plated on top of a 2-ml underlayer of 0.5% agarose in the same medium in 6-well culture plates. After 3 weeks, the colonies containing >50 cells were counted.

Male nude athymic BALB/c mice, 6–8 weeks old (Vital River Experimental Animals Technique Ltd., Beijing, China) were injected s.c. on the side of the abdomen with a total of 8×10⁶ cells in an exponential growth phase suspended in 200 μl serum-free RPMI-1640 containing Matrigel (1:1, v/v; BD Biosciences). Tumor volume measurements were recorded once a week and calculated as follows: L × W × H × 0.5236.

Two human prostate cancer tissue microarrays (Shanxi Chaoying Biotechnology Co., Ltd., Xi’an, China) were used to detected the expression of EphA3. A total of 110 formalin-fixed, paraffin-embedded benign prostate hyperplasia (BPH) prostate cancer tissues and corresponding normal epithelia were included in the tissue microarray. Another microarray contained 207 dots of 69 clinical cases including normal prostate, BPH, prostate cancer and adjacent normal prostate tissues in triplicate. Briefly, immunohistochemical (IHC) analysis was conducted as follows. Following deparaffinization, the samples were treated for antigen retrieval by the microwave method. A monoclonal antibody to EphA3 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) (1:100) was used. Staining intensity and the percentage of immunoreactive cells were quantitated using image analysis (n=8 images), and then the slides were sealed with Crystal/Mount (Thermo Fisher Scientific, Pittsburgh, PA, USA) and analyzed separately by two pathologists.

Protein extracts were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA, USA). Blots were blocked in the nitrocellulose membranes (Amersham Biosciences), then probed with antibodies against EphA3 (1:1000;), β-actin (1:2,000; Santa Cruz Biotechnology), phospho-Akt (pAkt; 1:1,000; Cell Signaling Technology), total Akt (1:1,000; Cell Signaling Technology) and tublin (1:1,000; Cell Signaling Technology). After a series of washes, blots were incubated with goat anti-mouse or anti-rabbit IgG antibody conjugated to horseradish peroxidase (Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd., Beijing, China) then detected using an enhanced chemiluminescence kit (Pierce).

Analyses were conducted using the statistical software SPSS13.0. All tests of significance were set at P<0.05. Data analyses over time were undertaken by repeated measures analysis. For the staining of tissue microarrays, the average optical density, integral optical density, and area density data were subjected to ANOVA analysis, with the post hoc Scheffe test used for multiple comparisons between groups.

Our previous results indicated the potential linkage of PC-1 to the progression of PCa in vitro and in vivo. In order to gain more insight into the molecular mechanisms of PC-1-induced promotion of PCa, we analyzed RNA from LNCaP-PC-1 and control LNCaP-Neo cells. Three cultures for each cell line were used for the RNA extraction. RNA was labeled and hybridized to Affymetrix U133 Plus 2.0 GeneChips. For each gene, the difference in expression level and the statistical significance of this difference were calculated. One hundrend and ninety-one genes were evaluated for 1.4-fold relative increases and decreases in expression levels. These included 30 upregulated and 161 downregulated genes. Representative tumor-associated genes are shown in Table I. Expression levels of several oncogenes, such as RAP2B, MAF and NDRG (downstream of N-Myc), were increased by PC-1, while expression levels of several tumor suppressors including ST7 were decreased. Expression of apoptosis-associated genes, such as FAF1, CASP9, TNFRSF21 (DR6) was downregulated in the PC-1-overexpressing LNCaP cells. Expression levels of differentiation-associated genes, such as Nkx3.1, JAG1 and NOTCH3, were decreased by PC-1 as well. The data of the microarray were validated by RT-PCR and a portion of the results is shown in Fig. 1A. Notably, one of the migration and metastasis-associated genes, EphA3, was increased by 42-fold in the LNCaP-PC-1 cells.

Real-time PCR and western blot analysis were performed to confirm the PC-1-induced expression of EphA3. The results showed that EphA3 was induced by PC-1 in LNCaP cells at the mRNA (Fig. 1C) and protein (Fig. 1D) levels. We also detected the expression of PC-1 and EphA3 in six different prostate cell lines. As shown in Fig. 1B, no PC-1 or EphA3 expression was noted in benign prostatic hyperplasia (BPH1) cells. The LNCaP cell line, established from a metastatic deposit in a lymph node, is androgen-dependent, non-metastatic, and weakly tumorigenic (17). The LNCaP lineage-derived sublines C4-2 and C4-2B, obtained from tumor-stoma interaction, possess the capabilities of androgen-independent growth and distant organ metastasis (18,19). The mRNA expression of PC-1 increased progressively in the LNCaP, C4-2 and C4-2B cells. PC-3 and DU145 are two prostate cancer cell lines derived from lumbar and central nervous system metastasis, respectively, and are both androgen-independent. There was minimal expression of EphA3 and undetectable PC-1 expression in these two cell lines. Consistent with the mRNA level, the EphA3 protein was increased in the PC-1-overexpressing LNCaP cells compared with the control. PC-1-targeted siRNA was used to suppress the PC-1 expression in C4-2B cells. The amount of EphA3 protein was decreased in the si-PC-1-transfected C4-2B cells compared with the negative RNAi control and parental C4-2B cells. In another part of our study, a 2011-bp length of the EphA3 promoter sequence was cloned into a luciferase report gene vector, and PC-1 was found to enhance the activity of the EphA3 promoter in luciferase assays, suggesting that PC-1 may induce the EphA3 expression through increasing the promoter activity (20). Taken together, these results above indicate that EphA3 is induced by PC-1 in prostate cancer cells.

In view of the fact that PC-1 promoted prostate cancer progression and PC-1-induced EphA3 expression, we were interested in investigating the role of EphA3 in prostate cancer. For this purpose, EphA3 stably expressing and control LNCaP cell lines were constructed by transfection of an EphA3-expressing vector or a mock vector. Two LNCaP-EphA3 clones (#23 and #25) were determined to have higher levels of EphA3 expression than the control cells (LNCaP-Neo) (Fig. 2A). MTT assays were performed to evaluate the growth of the transfected cells. The growth rate of the LNCaP-EphA3 cells increased by 1.6-, 2.0- and 2.2-fold compared with the control at 2, 4 and 6 days, respectively (Fig. 2B). In the plate colony-formation assay, it was found that EphA3 markedly enhanced the colony-formation ability of the LNCaP cells. LNCaP-EphA3 cells formed more and larger colonies than the control after 10 days in culture (Fig. 2C and D). Based on these results, the effect of EphA3 expression on the anchorage-independent growth of LNCaP cells was evaluated in colony formation assays. As shown in Fig. 2E, LNCaP-EphA3 cells exhibited enhanced clonogenicity in semi-solid agar compared with the mock-transfected control cells. These results suggest that overexpression of EphA3 enhances the proliferation and survival of LNCaP cells.

Due to its high expression of EphA3 protein in C4-2B cells, they were chosen to construct the EphA3 knockdown cell model. EphA3 targeted siRNA was transfected into C4-2B cells, and the expression of EphA3 was markedly suppressed compared with the si-control and parental cells (Fig. 3A). In the MTT assay, the growth rate of the si-EphA3-transfected cells was decreased compared with this rate in the si-control cells (Fig. 3B). In the plate colony-formation assay, EphA3 suppression led to a markedly reduction in colony-formation ability of the C4-2B cells (Fig. 3C and D). Similarly, EphA3 suppression decreased the clonogenicity of the C4-2B cells compared with that noted in the control cells (Fig. 3E). These results suggest that knockdown of EphA3 suppresses the proliferation and survival of C4-2B cells, coinciding with the results in the LNCaP cells.

EphA3 enhances the proliferation and survival of prostate cancer cell in vitro. Furthermore, we determined the in vivo tumor growth and tumor take of EphA3-overexpressing LNCaP cells in nude mice. Similar to PC-1, data from the s.c. model indicated that ectopic EphA3 expression facilitated the tumor take rate and accelerated tumor growth in the LNCaP-EphA3 cells (Fig. 4A). Detectable tumors were observed within 5 weeks after injection (Fig. 4B). Visible tumors were noted in 4/7 mice by the end of 9 weeks. In contrast, the control LNCaP cells did not form tumors. In addition, tumor-bearing mice also exhibited a more emaciated phenotype than mice without tumors. These data indicate that EphA3 enhances the tumorigenicity of prostate cancer cells in vivo.

A tissue array containing 110 clinical specimens of BPH, PCa and adjacent normal prostate tissues, was analyzed by immunohistochemistry. Excluding 4 cases of non-adenocarcinoma, medium to strong expression of EphA3 was observed in 59/73 (80.8%) cancer tissues, while weak expression of EphA3 was detected in 8/27 (29.6%) normal prostate tissues (Table II). Immunohistochemisty staining was quantitated using image analysis software, and statistical analyses identified a positive correlation between levels of EphA3 and the Gleason grade of PCa (Fig. 5A). Cellular localization of EphA3 was also found to change as the grade of PCa increased. EphA3 was expressed preferentially in the cytoplasm in Gleason 1 samples, EphA3 was distributed equally between the cytoplasm and nucleus in Gleason 2 and 3 samples, predominantly nuclear staining was observed in Gleason 4 samples, and a nodular expression pattern was observed in a cluster of cell nuclei in Gleason 5 samples (Fig. 5B). To confirm these results, a second tissue array containing 207 dots of 69 clinical specimens (triplicate for each), including normal prostate, BPH, PCa, and adjacent normal prostate tissues was assembled and analyzed. Results consistent with the first array were observed (data not shown). EphA3 was overexpressed in PCa specimens and its expression increased and its cellular localization changed as the Gleason grade was elevated.

The Akt signaling pathway has been well-characterized for its role in promoting cell growth, invasion and metastasis during tumorigenesis (21–23). We examined the impact of EphA3 on the Akt pathway. Overexpression of EphA3 was found to increase the phosphorylation of Akt at Ser473 and Thr308 in LNCaP cells. The phosphorylation of Akt was deduced with inhibition of EphA3 in the C4-2B cells (Fig. 6A). The quantitation of the expression of the genes showed that the expression levels of EphA3, pAkt (Ser473), pAkt (Thr308) changed significantly (Fig. 6B and C). These results indicate that EphA3 activates the Akt signaling pathway.