Activation of alternative growth factor pathways after androgen withdrawal is one mechanism mediating androgen-independent (AI) progression in advanced prostate cancer. Insulin-like growth factor (IGF) I activation is modulated by a family of IGF binding proteins (IGFBPs). Although IGFBP-2 is one of the most commonly overexpressed genes in hormone refractory prostate cancer, the functional significance of changes in IGF-I signaling during AI progression remains poorly defined. In this article, we characterize changes in IGFBP-2 in the LNCaP tumor model after androgen withdrawal and evaluate its functional significance in AI progression using gain-of-function and loss-of-function analyses. IGFBP-2 mRNA and protein levels increase 2–3-fold after androgen withdrawal in LNCaP cells in vitro in LNCaP tumors during AI progression in vivo. Increased IGFBP-2 levels after castration were also identified using a human prostate tissue microarray of untreated and posthormone therapy-treated prostatectomy specimens. LNCaP cell transfectants that stably overexpressed IGFBP-2 progressed more rapidly after castration than control tumors. Antisense oligonucleotides (ASOs) targeting the translation initiation site of IGFBP-2 reduced IGFBP-2 mRNA and protein expression by >70% in a dose-dependent and sequence-specific manner. ASO-induced decreases in IGFBP-2-reduced LNCaP cell growth rates and increased apoptosis 3-fold. LNCaP tumor growth and serum prostate-specific antigen levels in mice treated with castration plus adjuvant IGFBP-2 ASOs were significantly reduced compared with mismatch control oligonucleotides. Increased IGFBP-2 levels after androgen ablation may represent an adaptive response that helps potentiate IGF-I-mediated survival and mitogenesis and promote androgen-independent tumor growth. Inhibiting IGFBP-2 expression using ASO technology may offer a treatment strategy to delay AI progression.
Prostate cancer is the most common cancer that affects men and the second leading cause of cancer deaths in men in the Western world. Because prostate cancer is an androgen-sensitive tumor, androgen withdrawal, for example, via castration, is used in some therapeutic regimens for patients with advanced prostate cancer. Androgen withdrawal is the only effective form of systemic therapy for men with advanced disease, producing symptomatic and/or objective response in 80% of patients. Unfortunately, AI3 progression and death occurs within a few years in the majority of these cases (1). HRPC is highly chemoresistant with objective response rates of 10% and no demonstrated survival benefit (2). More recently, Phase II studies using taxane-based combination regimens are reporting objective responses in 20–30% and PSA responses in >50% of cases (3, 4, 5). HRPC remains the main obstacle to improving the survival and quality of life in patients with advanced disease. Novel therapeutic strategies that target the molecular basis of androgen and chemoresistance may provide an opportunity to delay progression and prolong survival.
One mechanism mediating AI progression in advanced prostate cancer involves activation of alternative growth factor pathways after androgen withdrawal. The IGF system, which consists of two ligands, IGF-I and IGF-II, plays an important role in cell growth regulation, transformation, and tumorigenesis (6, 7). The mitogenic and antiapoptotic activity of IGF is modulated by six IGFBPs (7, 8, 9, 10). These IGFBPs are synthesized locally in most tissues, including cancer cells, to inhibit or enhance IGF actions and may even possess ligand-independent activity (11, 12, 13). During tissue development, IGFBPs are expressed in a tissue- and stage-specific manner (14, 15).
Among the IGFBPs, IGFBP-2 has been shown to function as an enhancer of IGF-I function in several cell lines (16, 17, 18). Although IGFBP-2 is normally expressed in fetal cells and low or absent in many adult tissues, high levels of IGFBP-2 are associated with carcinogenesis or tumor progression in several tumor tissues. Elevated serum levels of IGFBP-2 have been reported in ovarian (19), colon (20), central nervous system (21), and prostate (22, 23, 24) cancers. IGFBP-2 immunohistochemical staining levels are higher in high Gleason grade cancers compared with low-grade cancers or benign epithelial cells (25) and in recurrent AI tumors compared with untreated hormone naïve tumors (26). Experimentally, IGFBP-2 expression levels also increase after castration in the rat ventral prostate (27) and during AI progression in AD Shionogi (28) and CRW22 (29) prostate cancer models and in human prostate tumors (25, 30, 31). Additionally, long-term overexpression of IGFBP-2 in adenocortical tumor cells enhanced cell proliferation and increased cloning efficiency (18). Transfection of IGFBP-2 to human epidermoid carcinoma increases tumorigenesis (32). Collectively, these observations suggest that changes in IGF/IGFBP-2 signaling may represent activation of an alternative growth factor pathway involved in AI progression.
Elucidation of the pathogenic role of candidate genes implicated in tumor progression is a rapidly progressing field of cancer research and has provided a steadily growing list of candidates. Known nucleotide sequences of cancer-relevant genes offer the possibility to design tailored anticancer agents that lack many of the toxic side effects displayed by conventional therapeutics. The objective of this study was to define changes in IGFBP-2 expression after castration and during AI progression in the LNCaP tumor model and to evaluate the functional relevance of these changes in AI progression using gain-of-function and loss-of-function analyses. As part of our ongoing investigations to identify important pathways mediating AI progression, we have used ASOs to evaluate the functional role of castration-induced changes in Bcl-2 (33), Bcl-xL (34), clusterin (35), and IGFBP-5 (36) in prostate xenograft models, some of which are now being evaluated in clinical trials (37, 38). Phosphorothioate ASO are chemically modified stretches of single-stranded DNA complementary to mRNA regions of a target gene, thereby inhibiting gene expression and providing a useful tool for in vitro and in vivo functional genomics (39, 40, 41, 42). For this study, overexpressing IGFBP-2 LNCaP cell transfectants were created to determine whether this changed tumor growth rates and time to AI, whereas IGFBP-2 ASOs were used to determine their effects on IGFBP-2 levels and cell proliferation in LNCaP cells in vitro and AI LNCaP tumor growth and PSA levels after castration.
MATERIALS AND METHODS
LNCaP Tumor Growth.
The human prostatic carcinoma cell line LNCaP was kindly provided by Dr. Leland W. K. Chung (Emery University, Atlanta, GA) and used in this study. Cells were cultured and maintained in RPMI medium (Life Technologies, Inc.) supplemented with 5% heat-inactivated FCS as described previously (43, 44). For in vivo study, ∼1 × 106 LNCaP cells were inoculated s.c. with 0.1 ml of Matrigel (Becton Dickinson Labware, Franklin Lakes, NJ) in the flank region of 6–8 week-old male athymic nude mice (Harlan Sprague Dawley, Inc., Indianapolis, IN) via a 27-gauge needle under methoxyfluorane anesthesia.
Human Prostate Tissue Microarray Preparation.
Archival formalin-fixed, paraffin-embedded human prostate tumor specimens were used to construct a human prostate cancer tissue array of hormone naïve and NHT-treated samples. A total of 190 specimens was obtained from hormone naïve benign (n = 4), hormone naïve cancer (n = 21), AI (n = 24), and after treatment with NHT for 1 month (n = 56), 3 months (n = 24), and 6–8 months (n = 56). The benign samples were obtained from transition zone biopsies of radical prostatectomy specimens, whereas the AI specimens were obtained from transurethral resections of patients with HRPC. Two samples/tumor specimen were arrayed for a total of 380 cores/array. Core tissue biopsy specimens (diameter, 0.6 mm) were taken from preselected regions of individual paraffin-embedded donor blocks and precisely arrayed into a new recipient paraffin block with a tissue arrayer (Beecher Instrument, Silver Spring, MD). After the block construction was completed, 5-μm sections were cut with a microtome by use of an adhesive-coated tape sectioning system (Instrumedics, Hackensack, NJ) to support the adhesion of the array elements.
Mounted tissues on the slides were rehydrated, and endogenous peroxidase activity was blocked with methanol containing H2O2. Antigen retrieval was enhanced using Dako Target Retrieval Solution (Dako Corporation, Carpinteria, CA) with the autoclave method. BSA (Promega, Madison, WI) was applied for 1 h at room temperature to block the nonspecific binding sites on the slides, which were then incubated in a humidified chamber overnight at 4°C with a 1:50 dilution (200 μg/ml) of a rabbit antihuman IGFBP-2 polyclonal IgG antibody (Upstate Biotechnology). After primary incubation, tissue was washed three times with PBS and incubated with a horseradish peroxidase-conjugated goat antirabbit IgG secondary antibody (Upstate Biotechnology) used at a 1:400 dilution for 30 min at 25°C. The antigen was visualized by subsequent 5-min incubation with diaminobenzidine tetrahydrochloride before counterstaining with hematoxylin. Tissues were covered with mounting media (Permount; Fisher Scientific, Fair Lawn, NJ) and a coverslip. Negative control slide was processed in an identical fashion to that above, with the substitution of normal goat nonimmune serum for the primary antiserum. No color reactions were observed in negative control slide. Photomicrographs were taken through a microscope coupled to a digital camera, Photometrics CoolSNAP (Roper Scientific, Inc., Glenwood, IL), and corresponding imaging software.
Scoring of IGFBP-2 Staining.
The staining intensity of cytoplasmic IGFBP-2 was evaluated and scored by a pathologist (M. A.) and the Image Plus software (Media Cybernetics). Specimens were graded from 0 to +4 intensity, representing the range from no staining to heavy staining. All comparisons of staining intensity and percentages were made at ×400 magnification. Image Pro Plus software quantifies the percentage of stained area (combination of intensity and surface area).
Lentiviral Infection of IGFBP-2 into LNCaP.
The full-length cDNA for human IGFBP-2 was a generous gift of Dr. Anand Swaroop (University of Michigan). The 1.4-kb IGFBP-2 cDNA was subcloned into the lentiviral vector pHR′-CMV-EGFP at the BamHI and XhoI sites. Two vectors were created for study: pHR′-CMV-IGFBP-2 and pHR′-CMV (empty vector). Clone identity was verified using restriction digest analysis and plasmid DNA sequencing. Infectious lentivirus was generated by cotransfection of 1.5 × 106 293T cells with target plasmids with pCMVΔR8.2 (carries sequence necessary for viral assembly of lentivirus) and pMD.G, which expresses the vesicular stomatitis virus envelop glycoprotein G(VSV-G) pseudotype as described previously (45). The 293T cells were transfected with either IGFBP-2-expressing, EGFP-expressing, or empty lentiviral vector (10 μg of each) using the calcium phosphate precipitation method (Promega Profection Mammalian Transfection Systems). The 293T cells were transfected for 12–15 h, after which fresh media were added for 24 h. After this, the virus containing media was collected and filtered through 0.45-μm filter. Early-passage LNCaP cells (passage 30) were plated in 10-cm plates, and competent retrovirus was added to 30–40 multiplicity of infection. The media were changed after incubation for 16 h. The cells were passaged and harvested for UV microscopy to verify green fluorescent protein expression. Total RNA, conditioned media, and whole cell lysate were collected to ensure expression of IGFBP-2 by Western blotting and real-time PCR.
Antisense IGFBP-2 Oligonucleotides.
Phosphorothioate oligonucleotides used in this study were generated by Operon (Alameda, CA). Three different sequences of IGFBP-2 ASOs were designed. One was designed to hybridize with the translation initiation site (5′-GCAGCCCACTCTCGGCAGCAT-3′), with two others designed randomly (5′-CGCCCAGTAGCAGCAGCAGCA-3′ and 5′-TCCCGGAACACGGCCAGCTCC-3′). Two-base IGFBP-2 mismatch oligonucleotides (5′-GCAGCCCACTGTCCGCAGCAT-3′, 5′-CGCGCACTA-GCAGCAGCAGCA-3′, and 5′-TCCCGGAACTGCCCCAGCTCC-3′, respectively) were also synthesized and used as controls. Oligonucleotides were stored at −20°C in 10 mm Tris and 1 mm EDTA.
Treatment of Cells with Oligonucleotides.
Lipofectin, a cationic lipid (Life Technologies, Inc., Gaithersburg, MD), was used to increase uptake of oligonucleotides into cells. LNCaP cells were treated with various concentrations of oligonucleotides by adding 4 mg/ml lipofectin after preincubation for 20 min in serum-free RPMI. Four h after the beginning of the incubation, steroid hormone-depleted, charcoal-stripped media with or without 1 nm DHT (Sigma, St. Louis, MO) were added to the medium, as described previously (36). A second cycle treatment was applied 24 h later.
Northern Blotting Analysis.
Total RNA was extracted from cultured LNCaP cells or LNCaP tumors using the acid-guanidinium thiocyanate-phenol chloroform method. A total of 20 μg of RNA from each sample was subjected to electrophoresis on 1.2% agarose-formaldehyde gels and transferred to nylon membranes overnight according to standard procedures (35). Human IGFBP-2 and G3PDH probes were generated by reverse-transcription PCR from RNA of LNCaP cells with using primers 5′-ACAATGGCGATGACCACTCAGA-3′ (sense) and 5′-ACAGCACCATGAACATGTTTG-3′ (antisense) for IGFBP-2 and 5′-TGCTTTTAACTCTGGTAAAGT-3′ (sense) and 5′-ATATTTGGCAGGTTTTTCTAGA-3′ (antisense) for G3PDH. The RNA blots were hybridized with human IGFBP-2 probe labeled with [32P]dCTP by random primer labeling. Washing and densitometric analysis were carried out as reported previously (36, 37, 38). After detecting the IGFBP-2 expression, the membranes were reprobed and rehybridized with human G3PDH probe to verify the integrity of the RNA.
Western Blotting Analysis.
Samples containing equal amounts of protein (20 μg) from lysates of the cultured cells were subjected to SDS-PAGE and transferred to nitrocellulose filters. The filters were blocked in PBS containing 7% nonfat skim milk powder at 4°C overnight and then incubated for 1 h with 1:400-diluted anti-IGFPB-2 rabbit polyclonal antiserum (Upstate Biotechnology, Lake Placid, NY) and 1:400-diluted β-actin (Sigma). The filters were then incubated for 30 min with 1:10,000-diluted horseradish peroxidase-conjugated antirabbit or antimouse IgG antibody (Amersham Life Science, Arlington Heights, IL), and specific proteins were detected using an enhanced chemiluminescence Western blotting analysis system (Amersham Life Science).
In Vitro Mitogenic Assay.
The in vitro growth of LNCaP was assessed by the in vitro mitogenic assay, as described previously (43). When LNCaP cells in 10-cm dishes reached 70% confluence, medium was changed to serum-free media for serum starvation. After 24 h of serum starvation, cells were treated with IGFBP-2 ASO or mismatch control with or without DHT, as described above. After treatment, LNCaP cells were fixed with 1% glutaraldehyde (Sigma), stained with 0.5% crystal violet (Sigma), followed by incubation for 4 h at 37°C, and eluted with 10 ml of Sorensen’s solution (9 mg of trisodium citrate in 305 ml of distilled water, 195 ml of 0.1 n HCl, and 500 ml of 95% ethanol). The absorbance was determined with a microculture plate reader (Becton Dickinson Labware) at 550 nm. Absorbance values were normalized to the values obtained for the vehicle-treated cells to determine percent survival. Each assay was performed in triplicate.
Flow Cytometric Analysis.
Flow cytometric analysis of propidium iodide-stained nuclei was performed as described previously (36). Briefly, the LNCaP cell sublines were plated in 10-cm dishes and, when 70% confluent, were treated as described above. The cells were trypsinized, washed twice with PBS, and fixed in 70% ethanol for 5 h at 4°C. The fixed cells were washed twice with PBS, incubated with 1 μg/ml RNase A (Sigma) for 1 h at 37°C, and stained with 5 μg/ml propidium iodide (Sigma) for 1 h at room temperature. The stained cells were analyzed for relative DNA content on a dual laser flow cytometer (Beckman Coulter epics elite; Beckman, Inc., Miami, FL).
Treatment of Mice with Oligonucleotides in Vivo.
Mice bearing tumors between 300 and 500 mm3 in volume were castrated via a scrotal approach and randomly assigned to treatment with 10 mg/kg of either antisense or mismatch IGFBP-2 oligonucleotides i.p. once daily. Treatment commenced beginning 1 day after castration. Tumor volume and serum PSA measurements were performed once weekly. Tumor volumes were calculated by the formula L × W × H × 0.5236 (43). Blood samples were obtained from tail vein incisions of mice, and serum PSA levels were determined by an enzymatic immunoassay kit with a lower limit of sensitivity of 0.2 μg/liter (Abbott IMX, Montreal, Quebec, Canada), according to the manufacturer’s protocol. Selected tumors were excised for Northern analysis of IGFBP-2 mRNA levels. Data points were expressed as average tumor volumes and serum PSA levels ± SE. All animal procedures were performed according to the guidelines of the Canadian Council on Animal Care and with appropriate institutional certification.
All data were analyzed by student’s t test. Levels of statistical significance were set at P < 0.05 (two-sided), and all statistical calculations were done by use of the Statview 4.5 (Abacus Concepts, Inc., Berkeley, CA).
Changes in IGFBP-2 Expression after Androgen Withdrawal.
We first examined the protein expression pattern of IGFBP-2 in the LNCaP cells cultured in charcoal-stripped serum with or without added androgens (R1881). We found that, over time, there was a relative increase in the protein expression of IGFBP-2 in the cells cultured in conditions that lacked androgens (Fig. 1). Increases in IGFBP-2 levels were also seen under conditions using serum-free media (data not shown).
Northern blotting was used to characterize changes in IGFBP-2 mRNA expression in AD intact tumors before castration and at various time points after castration. As shown in Fig. 2, IGFBP-2 expression increased gradually beginning 14 days after castration and, by 28 days after castration, was >2-fold compared with levels before castration (two-sided P < 0.05, student’s t test).
Increased IGFBP-2 levels after castration was also identified using a human prostate tissue microarray (Fig. 3). IGFBP-2 staining was absent in benign prostate epithelium (Fig. 3,A), whereas mean intensity increased from 0.73 in hormone naïve cancers (Fig. 3,B) to 1.1 after 1 month NHT (Fig. 3,C), 1.7 after 3 months (Fig. 3,D), 1.9 after 8 months (Fig. 3,E), and 2.1 in AI tumors (Fig. 3 F). Differences in staining between benign and cancer and between hormone naïve and AI tumors were statistically significant (two-sided P < 0.05, student’s t test). Temporal changes in immunostaining in the human prostate tissue microarray corresponded with results from Northern blotting of LNCaP xenografts.
IGFBP-2 Overexpression Promotes AI Tumor Growth in Vivo.
To characterize the role of IGFBP-2 in AI progression, we generated a cell line that overexpressed IGFBP-2. Using the lentiviral transfection system, we generated a population of LNCaP cells that stably expressed either human IGFBP-2 cDNA (labeled LNBP-2) or an empty lentiviral construct (labeled LNCaP). We verified elevated expression of IGFBP-2 protein in both whole cell lysates and in conditioned media (Fig. 4). Quantitative expression of IGFBP-2 transcripts in the resultant cell lines was determined by real-time PCR and revealed a 100-fold greater expression of IGFBP-2 in the transfected cells (data not shown). We compared the rate of progression to AI in the LNCaP tumor model after castration (defined by a rising PSA after castration) in LNBP-2- and control-transfected LNCaP cells and followed tumor volumes and PSA levels before castration. At a PSA of 75 ng/liter, the mice were castrated, and tumor volumes and PSA measurements were done on a weekly basis. We found that by 2 weeks after castration, the LNBP-2 tumors demonstrated a rapid recovery in tumor growth and rise in PSA consistent with more rapid AI tumor growth compared with mock transfected controls (statistically significant using student t test P < 0.05, Fig. 5).
Suppression of IGFBP-2 mRNA and Protein Levels by IGFBP-2 ASO in LNCaP Cells in Vitro.
The data presented above suggests that increased IGFBP-2 after castration may help accelerate time to AI progression. We next set out to test whether inhibiting IGFBP-2 could inhibit IGF-I signaling and delay time to AI progression. To identify the most effective IGFBP-2 ASO, three different IGFBP-2 ASOs were synthesized and tested for their ability to decrease IGFBP-2 mRNA levels in LNCaP cells. IGFBP-2 mRNA expression was suppressed most effectively by 1000 nm of the ASO targeting the translation initiation site of IGFBP-2 (data not shown). This ASO sequence (5′-GCAGCCCACTCTCGGCAGCAT-3′), along with its 2-base mismatch control (5′-GCAGCCCACTGTCCGCAGCAT-3′) was chosen for subsequent in vitro and in vivo experiments.
Northern blotting was used to measure changes in IGFBP-2 mRNA expression in LNCaP cells after treatment with IGFBP-2 ASO and mismatch control oligonucleotide. Four days after treatment, IGFBP-2 ASO suppressed mRNA by >70% in a dose-dependent and sequence-specific manner (Fig. 6,A). These results were confirmed using real-time PCR (data not shown). Similarly, Western blotting analysis was used to show that the observed decrease in IGFBP-2 mRNA levels was paralleled by dose-dependent and sequence-specific decreases in IGFBP-2 protein levels (Fig. 6 B).
Effects of IGFBP-2 ASO on LNCaP Cell Growth in Vitro.
To determine whether reduced IGFBP-2 levels alter LNCaP cell growth rates, a mitogenic assay was performed after treatment with IGFBP-2 ASO or mismatch control oligonucleotide in cells cultured in serum-containing media. As shown in Fig. 7, after treatment with 500 nm IGFBP-2 ASO, LNCaP cell numbers decreased in a dose-dependent manner by up to 80% compared with the untreated or mismatch control-treated groups.
To investigate whether androgens altered IGFBP-2 ASO-induced LNCaP cell growth inhibition, LNCaP cells were treated with 500 nm IGFBP-2 ASO with or without 1 nm DHT for 2 days under serum-free conditions (Fig. 8). IGFBP-2 ASO reduced LNCaP cell number by 56% with DHT (P < 0.05) and 40% without DHT (P < 0.05), compared with the no treated or mismatch control oligonucleotide-treated groups, respectively. Addition of IGF-I with IGFBP-2 ASO blunted the inhibitory effects of IGFBP-2 ASO only when DHT was present, whereas recombinant IGF-I had no effect on LNCaP cell growth in the absence of DHT regardless of IGFBP-2 ASO treatment.
Changes in Cell Cycle Distribution after Treatment with IGFBP-2 ASO.
To additionally investigate the effects of IGFBP-2 ASO on LNCaP cell growth, flow cytometric assay was performed. LNCaP cells were treated with IGFBP-2 ASO with or without 1 nm DHT for 2 days. As shown in Fig. 9, IGFBP-2 ASO significantly increased the sub-G1-G0 proportion 3-fold (P < 0.05) and decreased the S + G2-M proportion by 50% compared with untreated or mismatch control oligonucleotide-treated cells. The addition of DHT did not significantly alter the effects of IGFBP-2 ASO on cell cycle distribution of sub-G1-G0 proportion.
Effects of IGFBP-2 ASO on AI LNCaP Tumor Growth in Vivo.
To evaluate the efficacy of IGFBP-2 ASO in vivo, nude mice bearing LNCaP tumors were castrated and treated with adjuvant IGFBP-2 ASO or mismatch control oligonucleotide (n = 7 in each group). Phosphorothioate oligonucleotides were administered once daily by i.p. injection for 8 weeks. Tumor volume and serum PSA levels were similar before treatment both groups. After castration, mean tumor volume increased more rapidly in the mismatch control oligonucleotide-treated group compared with IGFBP-2 ASO-treated group. By 8 weeks after castration, mean tumor volume in mismatch control treated mice was 1.6-fold higher than the IGFBP-2 ASO-treated group (P < 0.05; Fig. 10,A). After castration, serum PSA levels decreased 60% in both groups but increased more rapidly in the mismatch control oligonucleotide-treated group. By 8 weeks after castration, mean serum PSA level in IGFBP-2 ASO-treated mice was significantly lower than in control mice (211.7 versus 388.5 ng/ml, P < 0.005; Fig. 10,B). No significant toxicity or side effects were observed in either group. When LNCaP tumors were harvested for Northern blotting for IGFBP-2, IGFBP-2 mRNA expression in IGFBP-2 ASO-treated groups was significantly lower than mismatch control oligonucleotide-treated groups (after normalization with G3PDH, P < 0.05; Fig. 10 C).
The sensitivity of tumor cells to undergo apoptosis in response to cytotoxic stress and their ability to adaptively alter gene expression determines their susceptibility to therapies and survive under adverse physiological conditions. Androgen ablation causes apoptotic prostate cancer cell death, but despite high initial response rates, remissions are temporary because surviving tumor cells eventually recur. Cellular and molecular events that mediate AI progression consist of a complex process involving both clonal selection and adaptive mechanisms occurring in heterogeneous tumors composed of subpopulations of cells that respond differently to androgen withdrawal. Changes in expression of various apoptosis-associated genes, including bcl-2 (33) and clusterin (34), or recruitment of alternative autocrine, paracrine, or endocrine stimulatory pathways that replace androgen requirement for growth and survival (36, 46) are key elements of AI progression.
IGF-I is a potent mitogenic and antiapoptotic factor for various types of normal and malignant tissues, including prostate. Although circulating levels of IGF-I principally arise from the liver, the biological response of cells to IGFs is regulated by several factors in the microenvironment, including the IGFBPs (7, 8, 9, 10). To date, at least six IGFBPs have been identified that modulate the biological action of the IGF through high affinity binding interactions that influence the ability of IGF-I to function as ligands for the type I IGF receptor. IGFBP physiology is complex, with both stimulatory and inhibitory effects of cell proliferation and with certain actions occurring independent of IGF-I (11, 12, 13, 47). The positive or negative modulatory effects of IGFBPs on IGF action are additionally altered by secreted proteases, including PSA and cathepsins (48, 49, 50).
Reports of positive correlation between elevated serum IGF-I levels and increased prostate cancer risk (51), of dysregulated IGF-I expression promoting development of prostatic neoplasia in transgenic mice (52, 53), and of up-regulated IGFBP-5 expression potentiating AI development in tumor models (36) have raised considerable interest in the role of the IGF axis in prostate carcinogenesis and progression. Of the various components of the IGF axis, the most significant castration-induced changes occur in levels of various IGFBPs. Expression patterns of certain IGFBPs in normal prostate epithelial cells change after castration or treatment with antiandrogens (27, 30, 31). Similarly, changes in expression of various IGFBPs have also been reported in prostate cancer, with increases in IGFBP-2 and IGFBP-5 and decreases in IGFBP-3 from the benign to malignant state (28, 29, 30, 31, 54). IGFBP-3 and IGFBP-4 have apoptosis-inducing effects on prostate cancer cells, presumably by binding and inhibiting IGF-I activation of IGF-IR signaling (12, 48). Because IGFBP-3 antagonizes IGF-I signaling and IGFBP-2 and IGFBP-5 are thought to promote IGF responsiveness, castration-induced changes in IGFBP levels are consistent with a loss of IGF responsiveness in prostate cells undergoing androgen withdrawal-induced apoptosis and suggest that the ability to maintain IGF responsiveness through adaptive changes in IGFBP levels may be a key mechanism for survival of prostatic epithelial cells. We previously reported increasing IGFBP-5 levels in AD Shionogi tumors after castration and that IGFBP-5 ASOs inhibited IGF-I responsiveness and delayed AI progression postcastration (55). Forced overexpression of IGFBP-5 in LNCaP cells potentiated the antiapoptotic effects of IGF-I both in vitro and in vivo (36).
Increases in IGFBP-2 are associated with prostate cancer progression. IGFBP-2 immunohistochemical staining is higher in malignant compared with benign prostatic epithelial cells (30, 25). Elevated serum levels of IGFBP-2 have been positively correlated with the tumor stage of the patients and are assumed to originate from tumor cells (24). Bubendorf et al. (26) reported high IGFBP-2 expression in 100% of AI tumors, 36% of primary tumors, and 0% of benign prostate tissue in their human prostate cancer analysis by microarrays, suggesting a link between IGFBP-2 and AI progression.
In this study, we set out to characterize changes of IGFBP-2 expression after androgen withdrawal and during AI progression and to assess the functional relevance of these changes using overexpression and antisense inhibition approaches in the LNCaP tumor model. The LNCaP tumor model is an androgen-sensitive, PSA-secreting human prostate cancer cell line that can be induced to form tumors in athymic mice under a variety of conditions. As in human prostate cancer, serum PSA levels in this model are regulated by androgen and are directly proportional to tumor volume. After castration, serum and tumor cell PSA levels decrease up to 80% and remain suppressed for 3–4 weeks. Beginning 4 weeks after castration, however, PSA production gradually increases above precastration levels in the absence of testicular androgens, heralding the onset of AI progression (43, 44). IGFBP-2 levels increased after androgen withdrawal in vitro and in AI tumors in vivo compared with AD tumors before castration, consistent with the human prostate cancer microarray analysis by Bubendorf et al. (26). Increases in IGFBP2 from benign to cancer and after androgen withdrawal of prostate cancers using tissue microarrays in this study support earlier reports and illustrate that increases in IGFBP-2 begin within months of androgen withdrawal. Increased IGFBP-2 levels after androgen withdrawal appear to play a functional role in progression because mean serum PSA and tumor volume increased more rapidly after castration in LNBP-2 transfectants and were delayed when IGFBP-2 levels were reduced by ASO treatment. Collectively, these findings link rising IGFBP-2 levels with the hormone refractory phenotype and indicate that inhibition of IGFBP-2 up-regulation after androgen withdrawal can delay AI growth.
These results confirm that targeting cell survival or mitogenic genes up-regulated after androgen withdrawal offers one strategy to delay tumor progression. ASO inhibit expression of target genes and can assess the functional relevance of candidate targets in vivo. The main advantage of ASO therapy is the capacity for systemic delivery and endocytic uptake into target cells, without the need for creation of special and complex delivery vectors. This is especially relevant when studying processes leading to AI progression, which cannot be reproduced in monolayer tissue cultures and must be studied in a limited number of animal tumor model systems. Phosphorothioate oligonucleotides are chemically stabilized to resist nuclease digestion and hybridize with target mRNA with high affinity. The specificity and efficacy of ASOs rely on precise targeting afforded by strand hybridization, which activates RNaseH and prevents translation (56). Use of a 2-base mismatch oligonucleotide helps control for nonspecific immunomodulatory effects with phosphorothioate backbones.
In this study, an ASO targeted against IGFBP-2 translation initiation site was identified that potently suppressed IGFBP-2 mRNA and protein expression in vitro and in vivo in a dose-dependent and sequence-specific manner. In the in vivo experiments, mean serum PSA levels and tumor volume in mice bearing LNCaP tumors were significantly lower in mice treated with IGFBP-2 ASO compared with the mismatch control-treated group. These findings illustrate that in vivo systemic administration of an IGFBP-2 ASO can result in sequence specific down-regulation of the target gene in tumor tissues with delayed time to AI progression.
In vitro, ASO-induced decreases in IGFBP-2 inhibited serum-mediated LNCaP cell growth, both in the presence and absence of DHT. Flow cytometry analysis indicated that reduced cell numbers after IGFBP-2 ASO treatment resulted mainly from increased apoptosis either in the presence or absence of DHT. Although IGFBP-2 ASO decreased cell number to a greater degree in the presence of DHT, baseline absolute reduction in cell number was similar regardless of presence or absence of DHT. Fig. 8 indicates that IGF-I increases LNCaP cell proliferation and can partially overcome IGFBP-2 ASO growth inhibition only in the presence of DHT; indeed, IGF-I is not mitogenic and does not compensate for IGFBP-2 ASO inhibition in the absence of DHT. In the presence of DHT, IGF-I incompletely compensates for IGFBP-2 ASO inhibition of cell growth. These observations suggest that IGFBP-2 expression is necessary for efficient IGF-I-mediated cell growth by positively influencing the balance between apoptosis and proliferation.
LNCaP cells express trace levels of IGF-I mRNA, low levels of IGF-IR and no IRS-1 (45, 46) or trace amounts of IGFBP-5 (55). We (55) and others (57) have previously shown that IGF-I induces proliferation of LNCaP cells only in the presence of androgens. Furthermore, androgen ablation causes decreased IGF-IR mRNA levels in human LAPC-9 prostate cancer xenografts (54) and rat prostates (58). Taken together, these observations support a model in which androgen ablation results in reduced IGF-I responsiveness through reduced DHT and IGF receptor levels. The ability of the IGFBP-2 ASO treatment to significantly reduce LNCaP cell growth in both the presence and absence of IGF-I and under conditions in which LNCaP cells are both responsive (+DHT) and nonresponsive (−DHT) to IGF-I suggests that there may also be an IGF-I-independent pathway for IGFBP-2 to facilitate cell proliferation.
In summary, the results of this study support the hypothesis that increased IGFBP-2 expression after androgen ablation is an adaptive response that potentiates IGF-I-mediated cell survival and mitogenesis, thereby promoting AI tumor growth. Inhibiting IGFBP-2 expression using ASO technology may offer a treatment strategy to delay progression of prostate cancer after androgen withdrawal.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Grant 009002 from the National Cancer Institute of Canada and the Hudson Fund at the Vancouver General Hospital.
The abbreviations used are: AI, androgen independent; AD, androgen dependent; ASO, antisense oligonucleotide; HRPC, hormone refractory prostate cancer; IGF, insulin-like growth factor; IGFBP, IGF binding protein; IGF-IR, IGF-I receptor; PSA, prostate-specific antigen; NHT, neoadjuvant hormone therapy; DHT, dihydrotestosterone; G3PDH, glyceraldehyde-3-phosphate dehydrogenase.
We thank Eliana Beraldi, Mary Bowden, and Howard Tearle for their excellent technical assistance, as well as Dr. Lawrence Meyer and Visia Dragowska (Department of Molecular Sciences, British Columbia Cancer Agency, Vancouver, British Columbia, Canada) for their support and technical assistance in flow cytometry.