Restoration of Insulin-Like Growth Factor Binding Protein-Related Protein 1 Has a Tumor-Suppressive Activity through Induction of Apoptosis in Human Prostate Cancer

The IGF⁴ system plays an important role in the development and/or progression of prostate cancer (1, 2, 3, 4, 5). Its ligands, IGFBPs, act as either positive or negative regulators of the IGF signaling pathway (6). On the other hand, IGFBP-rPs are cysteine-rich protein families that exhibit structural similarity with the IGFBP super family (6) and serve as mediator of various biological functions, including growth regulation. IGFBP-rPs such as mac25, connective tissue growth factor, nephroblastoma overexpressed, and CYR61 are also referred to as IGFBP-rP1, IGFBP-rP2, IGFBP-rP3, and IGFBP-rP4, respectively. IGFBP-rP1 (mac25), originally found in meningial and mammary cell lines (7), is also referred to as tumor cell adhesion factor (8), prostacyclin stimulating factor (9), angiomodulin (10), and IGFBP-7 (11).

IGFBP-rP1 resides on human chromosome 4q12 and encodes a precursor protein of 282 amino acids (11). This protein is processed to a mature M_r 27,000 of 256 amino acids after a M_r 32,000 protein is modified by N-glycosylation. The characteristics of IGFBP-rP1 appear to have low homology of amino acid sequence (∼30%) with the other IGFBPs (12, 13) and have low affinity to the known IGFBP ligands, including IGF-I and IGF-II (14).

Recent publications have shown that several functions of IGFBPs are not actually related to the modulation of IGF signaling pathway (15, 16, 17). In human prostate, lower expression of IGFBP-rP1 mRNA was observed in metastatic prostatic carcinoma than in normal prostate (18). In an in vivo study using nude mice, reexpression of IGFBP-rP1 in metastatic human epithelial cell inversely correlated with tumor formation and was related to induction of apoptosis (19). In the human mammary gland, loss of heterozygosity in the IGFBP-rP1 gene is a frequent genetic event in cancerous tissue (20).

As mentioned above, the functional role and regulatory mechanism of IGFBP-rP1 are not fully investigated. The aims of this study are to elucidate what mechanisms are involved in the regulation of IGFBP-rP1 expression and how IGFBP-rP1 exerts its growth suppressive effect in prostate cancer cells.

Total of 60 prostate samples (48 prostate cancer and 12 BPH), taken from either radical prostatectomy or radical cystectomy, were used for this study. Formalin-fixed, paraffin-embedded materials of these samples were used for H&E staining to determine pathological diagnosis.

The human prostate cancer cell lines DU145, DuPro, LNCaP, and PC-3 were obtained from the American Type of Culture Collection (Manassas, VA). Human prostate cancer cell line ND-1 and benign hyperplastic epithelial cell line BPH-1 were developed in our laboratory. All cell lines except PC-3 were maintained in RPMI 1640 with l-glutamine and sodium pyruvate. The PC-3 cell line was incubated in OPTI-MEM with l-glutamine, and 10% FCS was added to all media. The cells were maintained in a humidified atmosphere of 5% CO₂/95% air at 37°C.

Tissues were fixed in freshly prepared 4% paraformaldehyde/PBS solution (pH 7.4) at 4°C overnight and dehydrated in graded alcohol and then embedded in paraffin wax. IGFBP-rP1-specific antisense oligonucleotide probe was designed based on the cDNA sequence, and its specificity was analyzed by a GenEBML database search using the Genetics Computer Group sequence analysis program (Genetics Computer Group, Madison, WI). The probe was commercially synthesized with a 3′-biotinylated tail. The sequence used was 25 bp (5′-TCGCAGGGGCCGCAGGTGTCCGAAG-Brigati Tail-3′) with the Brigati tail consisting of 5′-(TAG)5-biotin-biotin-biotin-(TAG)2-biotin-biotin-biotin-3′ (21). A deoxythymidine-20 oligonucleotide was used to verify the integrity and lack of degradation of mRNA in each sample.

In situ hybridization was performed on 5-μm sections using a microprobe manual staining system (Fisher Scientific, Pittsburgh, PA; Ref. 22), as previously described by Yano et al. (23). Briefly, 5-μm sections were deparaffinized and rehydrated in Tris buffer. The oligoprobes were hybridized at 105°C for 5 min and at 4°C for 45 min with Micro Probe System (Fisher Scientific) according to the manufacture’s instructions. After washing, the signal was detected by alkaline phosphatase-linked streptavidin at 50°C for 10 min and counterstained with Auto Hematoxylin (Research Genetics) before mounting using Crystal/Mount (Biomedia, Foster City, CA).

PC-3 cells were used as parent cells and cultured in OPTI-MEM supplemented with FBS and 5 ng/ml kanamycin at 37°C under 5% CO₂. The IGFBP-rP1 cDNA was obtained from a cDNA library made from the murine liver cancer. A 1.1-kb IGFBP-rP1 cDNA fragment containing the full-length coding sequence was ligated and cloned into IRES 2 neo expression vector (Clontech Laboratories, Inc., Palo Alto, CA).

The PC-3 cells were transfected by an electroporation with IRES 2 neo expression vector bearing the full length of wild-type IGFBP-rP1 cDNA (PC-3-rP1). The PC-3 cells transfected with the vector not bearing IGFBP-rP1 cDNA served as control cells (control empty vector-transfected PC-3 cells). Both IGFBP-rP1 cDNA-transfected and empty vector-transfected PC-3 cells were selected after 36 h by G418 treatment (200 μg/ml; Promega Corp., Madison, WI) for 5–7 days, until all of the nonvector-transfected PC-3 cells were dead. Surviving transfected cells were maintained with G418 (200 μg/ml), and the formation of individual colonies was monitored. Visible colonies were cloned by the penicillin-cup method, and each colony was transferred to a new well in a 12-well tissue culture plate. These cells were cultured in OPTI-MEM supplemented with FBS and 5 ng/ml kanamycin at 37°C under 5% CO₂.

Total RNA was isolated by an acid guanidinium thiocyanate/phenol/chloroform extraction method (24). Approximately 10 μg of each RNA were electrophoresed on 1% agarose gel containing formaldehyde and then transferred onto Hybond N nylon membrane. After prehybridization, the membrane was hybridized for 16 h at 42°C with ³²P labeled probes prepared by the Megaprime DNA labeling system (Amersham). The hybridized membranes were washed twice with sodium citrate buffer containing 0.1% SDS at room temperature for 15 min. Autoradiography was done with XAR-5 film (Kodak, Rochester, NY) at −80°C overnight. Expression level of IGFBP-rP1 mRNA was semiquantified by a BAS 2000 Bioimaging analyzer (Fujix, Tokyo, Japan). Levels of β-actin mRNA as the internal control were quantified in the same manner. Expression level of IGFBP-rP1 mRNA in each sample was adjusted by that of β-actin mRNA and standardized according to the level of external standard marker/membrane.

The cDNA was constructed by reverse transcription (Promega Corp.) using the extracted total RNA as a template. Briefly, 1 μg of total RNA was added to 0.5 μg of oligodeoxythymidilic acid primer and prepared to a final volume of 20 μl. The samples were placed at 55°C for 5 min and then cooled on ice. The primer and RNA mixture was then combined with 0.25 units of Avian myeloblastosis virus reverse transcriptase and 0.5 units of RNase inhibitor. Finally, the mixture was added to 1 mm deoxynucleotide triphosphate, 50 mm Tris-HCl, 75 mm KCl, and 5 mm MgCl₂. The reverse transcriptase reaction was then carried out at 45°C for 45 min. The cDNA was then incubated at 95°C for 5 min to inactivate the reverse transcriptase. Samples were stored at −20°C until used.

The cDNA samples (2 μl) were diluted into 20 μl of solution containing 50 mm deoxynucleotide triphosphate, 0.5 units of RedTaq polymerase, and PCR reaction buffer provided by the manufacturer (Sigma, St. Louis, MO). For differential RT-PCR between IGFBP-rP1 and G3PDH, each primer concentration was 300 and 150 nm, respectively, and for differential RT-PCR between caspase-3 and G3PDH, primer concentrations were 200 nm in each. Annealing temperature was 55°C with 32 cycles for both differential RT-PCR reactions. The primers used for IGFBP-rP1, caspase-3 and G3PDH are follows: IGFBP-rP1 forward, 5′-CAAGAGGCGGAAGGGTAAAG-3′; IGFBP-rP1 reverse, 5′-CTGTCCTTGGGAATTGGATG-3′; caspase-3 forward, 5′-CAAACTTTTTCAGAGGGGATCG-3′; caspase-3 reverse, 5′-GCATACTGTTTCAGCATGGCA-3′; G3PDH forward, 5′-TCCCATCACCATCTTCCA-3′; and G3PDH reverse, 5′-CATCACGCCACAGTTTCC-3′. The PCR products were electrophoresed on 1.5% agarose gel and the expression level of these genes was evaluated by Image J software (http://rsb.info.nih.gov/ij), and the areas under the curves were calculated and analyzed. The expression level of IGFBP-rP1 and caspase-3 was quantified relative to G3PDH expression level and expressed as arbitrary units.

To investigate whether CpG methylation causes inactivation of IGFBP-rP1 gene in DU145, DUPro, LNCaP, ND-1, PC-3, TSU, and BPH-1 cell lines, demethylating agent (5-Aza-dC) was added to the culture medium at the concentrations of 5 μm in duplicate. The cultured cells were harvested after 4 days of 5-Aza-dC treatment. Using cDNA obtained from these cell lines, the difference in expression level of IGFBP-rP1 and caspase-3 mRNA transcripts before and after 5-Aza-dC treatments was analyzed by differential RT-PCR.

The newly established transfected cells (PC3-rP1) were seeded in dishes and cultured in RITC serum-free medium at 37°C under 5% CO₂. After 72 h, the condition-medium was collected and dialyzed for 12 h three times using Ultrafilter model (3.5K cutoff membrane) at 4°C and sterilized using Millipore filter (0.22 μm).

The growth/proliferation of transfected cells was measured by MTT kit (Promega Corp.). Transfected cells were grown in 96-wells plate with 100 μl of RITC serum-free medium (Kyokuto Pharmacoceutical, Tokyo, Japan) containing 10 μg/ml heparin and maintained at 37°C under 5% CO₂. On days 1, 3, and 5, MTT solution was added and incubated at 37°C for 4 h. To determine the effect of condition medium from PC-3-rP1 cells on cell growth rate, starved cells of PC-3-rP1 and control empty vector-transfected PC-3 cells were seeded in 96-well plates (3000 cells/well). The medium was replaced with 100 μl of prepared condition medium and maintained at 37°C under 5% CO₂. On days 1, 3, and 5, MTT solution was added and incubated at 37°C for 4 h. To determine the effect of IGF-I on cell growth rate, starved cells of PC-3-rP1 and control empty vector-transfected PC-3 cells were seeded in 96-well plates (3000 cells/well). The medium was replaced with 100 μl of RITC serum-free medium each containing 0, 1, 10, and 100 ng/ml IGF-I and maintained at 37°C under 5% CO₂. On days 1, 3, and 5, MTT solution was added and incubated at 37°C for 4 h. After incubation with MTT solution, solubilization/stop solution of 100 μl was added and incubated at 37°C to completely dissolve the formazan product. Plates were analyzed using an ELISA plate reader (Bio-Rad, Hercules, CA) at 570 nm with the reference wavelength of 630 nm.

To analyze anchorage-independent growth of PC-3-rP1 and control empty vector-transfected PC-3 cells, 60-mm dish was first layered with 1% agar/2X OPTI-MEM with 20% FBS. A top layer containing 15 × 10³ cells/dish suspended in 2× OPTI-MEM supplemented with 20% FBS and 0.5% agar was added. These dishes were maintained at 37°C under 5% CO₂ for 11 days. Those colonies > 50 μm in diameter were counted.

Starved cells of PC-3-rP1 and control empty vector-transfected PC-3 cells were seeded in 96-well plates (3000 cells/well). The medium was replaced with 100 μl each of RITC serum-free medium. After 48 h, cultured medium was replaced with fresh RITC medium containing APOPercentage Dye Label (Biocolor Ltd., United Kingdom) and incubated for 1 h at room temperature. Purple-red stained cells were identified as apoptotic cells. The number of purple-red stained cells/100 cells was counted.

The difference in the values between two groups was analyzed using the Student’s t test. P of <0.05 was considered statistically significant.

IGFBP-rP1 mRNA was expressed in all 12 BPH tissues. As shown in Fig. 1, expression of IGFBP-rP1 mRNA was localized in stromal cells, as well as hyperplastic epithelial cells. The epithelial expression of IGFBP-rP1 mRNA was almost lost in all prostate cancer tissues; however, stromal expression was weak, but detectable. Similarly, as shown in Fig. 2, differential RT-PCR revealed that the mRNA transcript expression level of IGFBP-rP1 was significantly lower in prostate cancer than in control prostate.

Before 5-aza-dC treatment, mRNA transcript of IGFBP-rP1 gene was not detected in DU145, LNCaP, ND-1, and PC-3 cell lines. On the other hand, in BPH-1 and DUPro cell lines, IGFBP-rP1 mRNA was strongly expressed. After 5-aza-dC treatment, expression of IGFBP-rP1 mRNA was restored in DU145, LNCaP, ND-1, and PC-3 cell lines. In BPH-1 and DUPro cell lines, there was no significant difference observed in the mRNA transcript expression of IGFBP-rP1 gene before and after 5-aza-dC treatment (Fig. 3).

After transfection, mRNA transcript levels of IGFBP-rP1 were analyzed by Northern blotting in the PC-3 (parental, control empty vector-transfected, and IGFBP-rP1 cDNA-transfected PC-3 cells). As shown in Fig. 4, among several clones that expressed IGFBP-rP1 mRNA, we have randomly selected three types of clones with each different expression of IGFBP-rP1 mRNA, namely PC-3-rP1-H (higher expression), PC-3-rP1-M (moderate expression), and PC-3-rP1-L (low expression). These 3 IGFBP-rP1 cDNA-transfected PC-3 cell lines were used for the following experiment.

MTT assay was performed to determine the relationship of IGFBP-rP1 with its growth suppressive effect on PC-3 cells. As shown in Fig. 5, the growth rate of each PC-3-rP1 (PC-3-rP1-H, PC-3-rP1-M, and PC-3-rP1-L) cells was reduced as compared with control empty vector-transfected PC-3 cells. On day 5, the difference in growth rate between each PC-3-rP1 (PC-3-rP1-H, PC-3-rP1-M, and PC-3-rP1-L) cell and control empty vector-transfected PC-3 cells reached statistical significance (P < 0.001).

The effect of condition medium taken from PC-3-rP1 cells on the growth rate of control PC-3 cells was analyzed by MTT assay. As shown in Fig. 6, the growth rate of control empty vector-transfected PC-3 cells cultured with condition medium obtained from PC-3-rP1 was significantly reduced after 5 days as compared with control empty vector-transfected PC-3 cells cultured with conventional RITC serum-free medium (P < 0.001). In addition, the growth of PC-3-rP1 cells in condition medium from PC-3-rP1 was also significantly reduced after 5 days, as compared with PC-3-rP1 cells cultured with conventional RITC serum-free medium (P < 0.001).

MTT assay was performed to determine the effect of IGF-I on the growth rate of PC-3-rP1 cells and control empty vector-transfected PC-3 cells. There were no significant differences in the growth rate of PC-3-rP1 cells among the conditions of RITC serum-free medium containing 0, 1, 10, and 100 ng/ml IGF-I (Fig. 7). Similarly, there were no significant differences observed in the growth rates of control empty vector-transfected PC-3 cells among the RITC serum-free medium containing 0, 1, 10, and 100 ng/ml IGF-I (Fig. 7).

After 11 days of incubation in soft agar, the numbers of colonies formed in each cell was analyzed. As shown in Fig. 8, colony-forming activity in the IGFBP-rP1 cDNA-transfected PC-3 cells was significantly lower (mean 4/dish) than in the control empty vector-transfected PC-3 cells (mean 46/dish; P < 0.001).

APOPercentage assay, where purple-stained cells are considered apoptotic cells, was used to determine the relationship between IGFBP-rP1 expression and apoptosis in the IGFBP-rP1 cDNA-transfected and control empty vector-transfected PC-3 cells. As shown in Fig. 9, the number of purple-red stained cells in the PC-3-rP1 cells was significantly higher (40 of 100 cells) than in the control empty vector-transfected PC-3 cells (2.3 of 100 cells; P < 0.001).

In the randomly selected two clones [indicated as (+) in Fig. 10], which were transfected with full length of wild-type IGFBP-rP1 cDNA (PC-3-rP1), strongly expressed mRNA transcript level of caspase-3 gene as compared with the clones without full-length IGFBP-rP1 transfection [indicated as (−) in Fig. 10].

IGFBP-rP1 is widely distributed in normal tissues, but its expression is reduced or lost in a variety of tumor cell lines (11). In human prostate, expression of IGFBP-rP1 mRNA has been reported to be significantly lower in metastatic tissues than in benign prostate (18). In the present study, using in situ hybridization, expression of IGFBP-rP1 mRNA was detected in the stroma and epithelium of nonmalignant prostate but only weakly detected in the cancerous tissues. In addition, RT-PCR in radical prostatectomy samples clearly showed that IGFBP-rP1 mRNA expression was significantly lower in prostate cancer than in benign prostatic tissues. These finding are in agreement with a previous study (18) that loss or down-regulation of IGFBP-rP1 expression is related to abnormal cell growth of prostate and/or pathogenesis of prostate cancer. Similarly in other tissues such as breast cancer, IGFBP-rP1 gene was strongly expressed in normal breast tissue but down-regulated or lost in hyperplastic and ductal carcinoma in situ and in invasive carcinoma (20). The loss of heterozygosity of the IGFBP-rP1 gene is a frequent genetic event in breast cancer (20). These findings suggest that IGFBP-rP1 is a tumor suppressor. Although the mechanism of inactivation of IGFBP-rP1 gene in cancer tissue is not fully investigated, hypermethylation of CpG island of IGFBP-rP1 gene has been reported to be associated with its down-regulation in hepatic carcinoma (25). In the present study, the expression of mRNA transcript of IGFBP-rP1 gene was restored after the treatment of demethylating agent (5-aza-dC) in four prostate cancer cell lines (Du145, LNCaP, ND-1, and PC-3), whereas in the benign BPH-1 cell line, there was no difference in IGFBP-rP1 mRNA expression before and after treatment. On the basis of these experiments, IGFBP-rP1 gene promoter function is inactivated by CpG methylation in prostate cancer.

To investigate the function of IGFBP-rP1, we established a PC-3 transfected cell line and determine whether IGFBP-rP1 exerts a tumor suppressive effect on human prostatic carcinoma cells. In this study, selection of PC-3 was based on two major pieces of evidence: (a) PC-3 cell preserves the property of androgen independence, which in turn could be a possible clinical model of hormone-refractory prostate cancer; and (b) PC-3 cell shows lack of endogenous IGFBP-rP1 expression and markedly restores the mRNA transcript of IGFBP-rP1 gene after 5-aza-dC treatment as compared with other prostate cancer cell lines. In the present study, transfected cells showed significant decrease in cell growth and included apoptosis through activation of caspase-3. In this regard, earlier studies have shown that a M12 cell, a metastatic human prostate cancer cell line, when transfected with IGFBP-rP1, resulted in increased doubling time and decreased colony formation (19). In the present study, transfected PC-3-rP1 cell showed (a) a significantly lower growth rate compared with control PC-3 cell, (b) the growth rate of both transfected PC-3-rP1 cells and control PC-3 cells was significantly decreased by the condition medium obtained from PC-3-rP1, and (c) the soft agar colony-forming activity of the PC-3-rP1 cells was significantly decreased compared with control PC-3 cells. These findings strongly suggest that IGFBP-rP1 acts as a tumor suppressor factor in prostate cancer cells.

The IGFBP-rP1 is a family of IGFBPs, which shows specific binding activity to IGF-I, IGF-II, and insulin (14). It has been reported that the interaction between IGFs and IGFBPs can either enhance or suppress mitogenic IGF signaling. Human IGFBPs 1–6 bind IGFs with high affinity, whereas IGFBP-rP1 binds IGFs with lower affinity (11). In the present study, the growth rate of both transfected PC-3-rP1 cells and control PC-3 cells did not change by treatment with IGF-I in the culture medium, suggesting that interaction between IGFBP-rP1 and IGF-I does not control growth of prostatic cancer cells. Therefore, it is possible that IGFBP-rP1-related signaling is through IGF-independent pathway in suppressing prostate cancer growth.

As discussed above, IGFBP-rP1-related signaling appears to be a negative growth regulator for tumor formation in an IGF-independent manner. In this regard, Kato et al. (26, 27) suggested that the murine mac25/IGFBP-rP1 functions as activated follistatin-like gene and also acts as growth suppressor by modulating transforming growth factor β signaling. Sprenger et al. (28) reported that overexpression of IGFBP-rP1/mac25 in the M12 prostate cancer cell line alters cell cycle kinetics, which in turn increased apoptosis in the M12 prostate cancer cell line. In the present study, the number of apoptotic cells were significantly increased in the IGFBP-rP1 cDNA-transfected PC-3-rP1 clones compared with controls. Moreover, IGFBP-rP1 overexpressed PC-3 clones also exhibited a higher level of mRNA transcript for caspase-3. Considering that activation of caspase-3 is the final and common pathway of apoptosis signal transduction (29), the positive correlation of IGFBP-rP1 overexpression with higher caspase-3 expression strongly suggested that increased expression of IGFBP-rP1 is associated with the induction of apoptosis in transfected PC-3-rP1 clones.

In conclusion, IGFBP-rP1 expression is down-regulated in prostate cancer cells. Treatment with demethylating agent (5-aza-dC) restored IGFBP-rP1 expression in prostate cancer cell lines, suggesting the role of methylation in inactivation of IGFBP-rP1 gene in prostate cancer. This is the first study that suggests inactivation of IGFBP-rP1 gene is through CpG hypermethylation, and tumor-suppressive activity of IGFBP-rP1 is through induction of apoptosis in an IGF-I-independent manner in prostate cancer. The results of these studies may provide us with better strategies for the management of prostate cancer.