The role of insulin-like growth factor binding protein-3 in the growth inhibitory actions of androgens in LNCaP human prostate cancer cells

Abstract

Insulin-like growth factor binding protein-3 (IGFBP-3), an antiproliferative and proapoptotic protein, has been shown to be upregulated by growth inhibitory concentrations of androgens in LNCaP human prostate cancer (PCa) cells, but the mechanism of regulation and the role of IGFBP-3 in the modulation of PCa cell proliferation are unknown. In this study, we have examined the effects of a range of concentrations of the synthetic androgen R1881 on IGFBP-3 expression and cell growth in LNCaP cells. We have also investigated the role of androgen-stimulated IGFBP-3 in androgen-induced growth inhibition. We show that low doses of R1881 stimulate LNCaP cell proliferation, but do not induce IGFBP-3 expression, whereas high doses of R1881 that inhibit cell growth, significantly increase expression of IGFBP-3. Importantly, we demonstrate that the combination of calcitriol and androgens not only synergistically upregulates IGFBP-3 expression but also inhibits cell growth better than either hormone alone. siRNA knockdown of IGFBP-3 expression partially reverses the growth inhibition by calcitriol and by androgens. Furthermore, we find that the growth inhibitory dose of R1881 leads to increases in the cyclin dependent kinase inhibitors (CDKIs), p21 and p27 as well as to G1 arrest. These changes can be blocked or partially reversed by IGFBP-3 siRNA, indicating that the induction of CDKIs is downstream of IGFBP-3. Our data suggest, for the first time, that IGFBP-3 is involved in the antiproliferative action of high doses of androgens partly through p21 and p27 pathways and that IGFBP-3 may contribute significantly to androgen-induced changes in LNCaP cell growth. © 2007 Wiley-Liss, Inc.

Insulin-like growth factor binding protein-3 (IGFBP-3) is the major binding protein for circulating insulin-like growth factor-I (IGF-I), a potent mitogen.1, 2, 3 The IGF-IGFBP-3 axis plays an important role in cellular growth, differentiation and proliferation. IGFBP-3 not only modulates IGF action by sequestering IGF but also regulates cell growth and apoptosis in an IGF-independent manner.1, 2, 3 Epidemiological4, 5, 6 and other studies7, 8 suggest that reduced IGFBP-3 is a potential marker of increased risk of certain cancers including prostate cancer (PCa), although this has recently been refuted for PCa.9 Both animal studies and preclinical evidence suggest that IGFBP-3 has the potential to be a cancer therapy either by activating apoptosis and initiating cell growth arrest, or by suppressing angiogenesis.1, 7, 8, 10, 11, 12, 13, 14

IGFBP-3 is regulated by a number of growth stimulators and hormones including calcitriol15, 16, 17 and androgens.18, 19, 20, 21, 22, 23 We have previously demonstrated that calcitriol or androgens directly upregulate IGFBP-3 expression, an action mediated through interaction of each receptor with its corresponding response element in the IGFBP-3 promoter.17, 23 Our previous studies also demonstrate that induction of IGFBP-3 by calcitriol in LNCaP cells is essential for the growth inhibitory actions of calcitriol.16 Increasing evidence has shown that upregulation of IGFBP-3 in PCa cells plays an important role in the in vivo growth inhibitory effects of calcitriol and its analogs.24, 25 However, the mechanism of the interaction of calcitriol and androgens to regulate IGFBP-3 expression and the role of IGFBP-3 in the modulation of PCa cell proliferation remain unclear.18, 19, 22, 25, 26, 27 Thus, the purpose of this study was to clarify the interaction between androgens and calcitriol in regulating IGFBP-3 expression and to investigate the potential role of IGFBP-3 in androgen-induced cell growth inhibition using the androgen-responsive LNCaP cell line as a model.

LNCaP cells, derived from a lymph node metastasis from a patient with PCa, exhibit a biphasic growth response to androgens.26, 27, 28, 29 Low doses of androgens stimulate cell growth whereas high concentrations result in inhibition of cell growth. In contrast, the prostate-specific antigen (PSA) is upregulated by androgens in a dose-dependent manner.26, 27, 28, 29 Only limited information is available on the mechanisms underlying this differential phenomenon. Lee et al.29 demonstrated that the abundance of androgen receptors (ARs) may be one of the determinants affecting the growth response to the hormone, as high levels of AR are found in the cells in the proliferative phase while low AR levels are found in the cells in the nonproliferative phase. Several other groups30, 31, 32 found that the biphasic dose-response of LNCaP proliferation to androgens is closely reflected by the cell cycle regulators including retinoblastoma protein (Rb), E2F and cyclin-dependent kinase inhibitor (CDKI) p27. In fact, low concentrations of androgens, known to promote LNCaP cell proliferation, induce Rb phosphorylation accompanied by increased E2F activity and decreased p27; whereas high concentrations of androgens, that inhibit cell proliferation, greatly reduce Rb phosphorylation along with a decrease in E2F activity and an increase in p27 protein levels.30, 31, 32 Elevated activity of transcription factor activator protein-1 (AP-1) by 1 nM R1881 was also reported to contribute to androgen-induced growth arrest in LNCaP cells.33 Furthermore, Schatten et al.34 found that 1 nM R1881 can alter mitochondria and centrosome proteins, thus leading to decreased cell proliferation in LNCaP cells.

In the present study, we treated LNCaP cells with various concentrations of the synthetic androgen R1881 alone or in combination with calcitriol and studied their effects on IGFBP-3 expression and cell proliferation. For the first time we show that IGFBP-3 is involved in mediating the antiproliferative effects of high concentrations of R1881 at least in part through stimulation of the p21/p27 pathway. Our findings provide new insights into the mechanisms by which androgens cause a biphasic growth response in LNCaP cells.

Abbreviations:

AP-1, activation protein-1; AR, androgen receptor; CDKI, cyclin dependent kinase inhibitor; DHT, dihydrotestosterone; FACS, fluorescence activated cell sorting; FBS, fetal bovine serum; GAPDH, glyceraldehydes-3-phosphate dehydrogenase; IGFBP-3, insulin-like growth factor binding protein-3; PCa, prostate cancer; PCR, polymerase chain reaction; PSA, prostate specific antigen; TBP, TATA-box-binding protein; VDR, vitamin D receptor.

Material and methods

Cell culture

The LNCaP human prostate carcinoma cell line (ATCC, Manassas, VA) was maintained in RPMI-1640 medium (Invitrogen, Carlsbad, CA) containing 5% fetal bovine serum (FBS; HyClone, Logan, UT) at 37°C in a humidified atmosphere with 5% CO₂. During all the treatments with androgen R1881 (PerkinElmer Life Sciences, Boston, MA) or calcitriol (a generous gift from Dr. Milan Uskokovic, BioXell, Nutley, NJ), the serum concentration was reduced to 2% FBS.

Isolation of RNA and quantitative real-time reversetranscription-PCR

Total RNA was prepared using the RNeasy Mini Kit (Qiagen, Valencia, CA). cDNA was synthesized from 2.5 μg of total RNA with random hexamer primers and SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA). Real-time PCR analysis was performed using the DNA Engine Opticon fluorescence detection system (MJ Research, Foster City, CA). The PCR reaction was set up as described previously.23 Comparative C_T method (2^−ΔΔCT)35 was used to semiquantitate the target signal change in a treatment group versus a control group. The TATA-box-binding protein (TBP) was used as an internal control to normalize the amount of cDNA added to the PCR reaction. The following primers were used: PSA, 5′-GCAG CATTGAACCAGAGGAG-3′ (sense), 5′-CACCATTACAGACAAGTGG GC-3′ (antisense); IGFBP-3, 5′-AAGTTCCACCCCCTC CATTC-3′ (sense), 5′-TCTTCCATTTCTCTACGGCAG-3′ (antisense); TBP, 5′-TGCTGAGAAGAGTGTGCTGGAG-3′ (sense), 5′-TCTGAATAG GCTGTGGGGTC-3′ (antisense).

IGFBP-3 and PSA enzyme-linked immunosorbent assay

IGFBP-3 and PSA secreted into the growth medium by LNCaP cells were measured using enzyme-linked immunosorbent assay (ELISA) kits (DSL, Webster, TX) according to the manufacturer's instructions.

Cell growth assays

LNCaP cells were plated at low density (5 × 10⁴) in 6-well plates and grown in 5% FBS-RPMI medium. After 24 hr incubation, the cells were treated with vehicle or the indicated compounds or recombinant human IGFBP-3 (from ProSpec-Tany TechnoGene, Rehovot, Israel) in 2% FBS-containing medium for 6 days. Fresh media with compounds or IGFBP-3 protein were replenished 72 hr after initiation in all the experiments. At Day 6, the conditioned medium was collected for analyses of IGFBP-3 or PSA, and the cells were processed for DNA quantitation using the Burton reagent.36

[³H]-thymidine incorporation assay

DNA synthesis was assessed by incorporation of [³H]-thymidine as described with modifications.23, 25 Briefly, 2 hr before cell collection, 0.5 μCi/ml of [³H]-thymidine (35 Ci/mmol, ICN Biomedicals, Irvine, CA) was added to each well. The cells were fixed in 1 ml/well methanol-acetic acid (3:1) for 5 min, rinsed with methanol for 5 min, then with 5% trichloroacetic acid for 5 min and finally washed twice with methanol. The DNA was solubilized with 0.2 N NaOH and the amount of [³H]-thymidine incorporated per sample was determined using a Beckman scintillation counter.

Fluorescence activated cell sorting

The cell cycle distribution was determined by fluorescence activated cell sorting (FACS) following the protocol described by Husbeck et al.37 Briefly, the cells were collected and washed in PBS buffer containing 5 mM EDTA and then fixed in ice-cold 70% ethanol for at least 1 hr at 4°C. The fixed cells were treated with 100 μg/ml RNase A for 30 min and stained with 50 μg/ml propidium iodide (Sigma, St. Louis, MO) for 10 min at ambient temperature. FACS data were analyzed using the FlowJo software (Tree Star, Ashland, OR).

siRNA transfection

LNCaP cells were seeded at 1.5–2 × 10⁵ cells/well and grown in 6-well plates for 24 hr before siRNA transfection. IGFBP-3 siRNA oligonucleotides (Ambion, Austin, TX), directed at exon 4, were transfected using siPORT NeoFX reverse transfection reagent (Ambion) following the manufacturer's instructions. A negative control siRNA (Ambion), comprised of a 19-bp scrambled sequence with 3′ dT overhangs, was used to demonstrate that the transfection does not induce nonspecific effects on gene expression. A second set of experiments used a different IGFBP-3 siRNA oligonucleotide directed at exon 5. After 24 hr, the cells were treated with vehicle, androgens or calcitriol (at 10 nM) or the combination of both hormones for 48 hr. The conditioned media were collected for IGFBP-3 ELISA to document IGFBP-3 knockdown and the cells were subjected to either [³H]-thymidine incorporation assay or FACS or Western blot analysis.

Western blot analysis

Extracts of the siRNA-transfected cells were prepared in RIPA buffer (50 mM Tris-HCl (pH 7.4), 1% NP-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EGTA) containing a protease inhibitor tablet (Roche Molecular Biochemicals, Indianapolis, IN). Supernatants were subjected to SDS-PAGE and Western blot analysis using p21, p27 and β-actin mouse monoclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, CA).23 Western blot for AR (Santa Cruz Biotechnology) was performed as described previously.38

Statistical analysis

Statistical comparisons were performed by ANOVA tests. *, p < 0.05; **, p < 0.01; ***, p < 0.001, when compared with the vehicle-treated group.

Results

Synergistic upregulation of IGFBP-3 gene expression by the synthetic androgen R1881 and calcitriol in LNCaP cells

We have previously shown that androgens and calcitriol individually increase the expression of IGFBP-3 by acting through their respective hormone response elements located in the IGFBP-3 promoter.17, 23 To determine whether there is an interaction between androgens and calcitriol in regulating IGFBP-3 expression, LNCaP cells were treated with 10 nM R1881 in the absence or presence of 10 nM calcitriol for different time periods. Treatment with the individual hormones resulted in significant increases in IGFBP-3 mRNA levels after 6 hr (Fig. 1a) and IGFBP-3 protein levels after 12 hr (Fig. 1b). Most interestingly, the combination of R1881 and calcitriol produced a highly significant synergistic upregulation of both IGFBP-3 mRNA and protein. IGFBP-3 mRNA levels were increased ∼17-fold at the end of 6 hr by the combination treatment that further increased up to 46-fold after 48 hr (Fig. 1a). Consistent with the increase in IGFBP-3 mRNA levels, IGFBP-3 protein levels rose proportionately with the time of incubation. An ∼50-fold increase in IGFBP-3 protein was observed when compared with the control after 48 hr of the combination treatment (Fig. 1b).

Figure 1. Synergistic upregulation of IGFBP-3 expression by R1881 and calcitriol. LNCaP cells were treated with vehicle or 10 nM R1881, a synthetic androgen, 10 nM calcitriol (1,25D) or 10 nM R1881 plus 10 nM calcitriol (R+D) for different time periods. (a) At each time point, cells were harvested and RNA prepared for quantitative real-time reverse transcription-PCR (qRT-PCR) analyses. (b) The conditioned medium was collected for determining IGFBP-3 levels by ELISA. Values are expressed as fold-change over control. The data is represents the mean ± SE of 3 separate experiments.

Mechanism of synergistic regulation of IGFBP-3 by R1881 and calcitriol

We have previously shown that calcitriol upregulates AR expression in LNCaP cells.38 To determine whether the calcitriol-mediated increase in AR contributes to the synergistic effect on IGFBP-3 transactivation by the 2 hormones, LNCaP cells were pretreated with ethanol vehicle or 10 nM calcitriol for 24 hr. Following the preincubation, R1881 or calcitriol or the combination was applied to the cells a second time and the incubation continued for an additional 24 hr. As expected, calcitriol pretreatment increased AR levels of mRNA (Fig. 2a) and protein (Fig. 2b, lane 4). No further significant increase in AR levels was observed with the addition of the second dose of calcitriol to calcitriol-pretreated cells (Figs. 2a and 2b, lane 5). R1881 did not affect AR mRNA expression but increased the AR protein level presumably because of the increased stability of the hormone-bound AR (Fig. 2b, lane 3).29 This androgen effect also caused a further increase in AR protein in the calcitriol-pretreated sample (Fig. 2b, lane 6). As shown in Figure 2c, IGFBP-3 levels (mRNA and protein) pretreated with calcitriol were further increased ∼5.5- to 7.5-fold by the androgen treatment when compared with the vehicle-pretreated control cells. The data suggest that the synergistic effect of the 2 hormones on IGFBP-3 may in part be due to the upregulation of AR, thus enhancing androgen induction of IGFBP-3 by the combination of both hormones.

Figure 2. Pretreatment with calcitriol increases IGFBP-3 induction by R1881. LNCaP cells were preincubated for 24 hr with vehicle or 10 nM calcitriol (1,25D). The media were replaced with fresh media containing vehicle, 10 nM R1881, or 10 nM 1,25D or 10 nM R1881 plus 10 nM calcitriol and incubated for an additional 24 hr. The conditioned medium was collected and the cells were harvested for RNA preparation or protein extraction. (a) AR mRNA levels determined by qRT-PCR. (b) Western blot analysis with AR specific antibody. Relative density represents the level of AR protein corrected by β-actin. (c) IGFBP-3 mRNA levels were quantitated by qRT-PCR and IGFBP-3 protein levels in the media were determined by ELISA. Each value in panel A and C is displayed as fold-change over the vehicle-pretreated control and represents 3 experiments (mean ± SE).

A growth stimulatory concentration of androgens has no effect on IGFBP-3 expression

As shown in Figure 3a, 0.1 nM R1881, which caused ∼4-fold increase in PSA mRNA, did not produce a significant impact on IGFBP-3 mRNA level. Surprisingly, 0.5 nM of R1881 resulted in an ∼3-fold increase in IGFBP-3 mRNA. Further increases in R1881 concentrations led to greater increases in IGFBP-3 mRNA in a dose-responsive manner similar to the increases in PSA mRNA across the entire dose range. We also observed this phenomenon in the promoter transactivation assays. As shown in Figure 3b, no induction of the IGFBP-3 promoter was observed at 0.1 nM R1881, in contrast to the PSA promoter that showed an over 2-fold induction at this concentration of R1881.

Figure 3. Induction of IGFBP-3 requires high concentrations of androgens. (a) LNCaP cells were incubated with increasing concentrations of R1881 (0–100 nM) for 18 hr. IGFBP-3 or PSA mRNA levels are expressed as fold-change relative to the vehicle-treated control. Each value represents 3 independent experiments (mean ± SE). (b) The luciferase reporter constructs driven by the 3.6 kb of IGFBP-3 promoter sequence17 or by the PSA promoter were transfected into LNCaP cells. Eighteen hours posttransfection, the cells were incubated with R1881 (0–100 nM) for an additional 18 hr. The cells were then subjected to luciferase activity assay. The data represents at least 3 independent experiments, each performed in triplicate (mean ± SE).

Androgen induction of IGFBP-3 correlates with the biphasic growth pattern in LNCaP cells

LNCaP cells display a biphasic growth pattern in response to androgens with low concentrations (DHT, <1 nM) stimulating cell proliferation and higher concentrations inducing cell growth arrest.26, 27, 28, 29 To determine whether androgen induction of IGFBP-3 correlates with this biphasic growth phenomenon, we treated LNCaP cells with increasing concentrations of R1881 (0–100 nM) and examined cell growth and induction of IGFBP-3 protein expression. As shown in Figure 4a, 0.1 nM R1881 stimulated cell growth but did not change IGFBP-3 protein level relative to the control. In contrast, concentrations equal to or higher than 0.5 nM R1881 increased IGFBP-3 protein ∼3-fold, while reducing cell proliferation ∼48% compared with the control. Higher concentrations of R1881 further increased IGFBP-3 and reduced growth. Furthermore, treatment with 10 nM of R1881 and calcitriol together produced a synergistic increase of IGFBP-3 expression (Fig. 4b) and also resulted in a greater cell growth inhibition. The data suggest that induction of IGFBP-3 is correlated with androgen-induced growth inhibition in LNCaP cells.

Figure 4. R1881 increases IGFBP-3 expression and enhances calcitriol-induced cell growth inhibition. LNCaP cells were treated with increasing concentrations of R1881 (0–100 nM) (in panel a) or treated with ethanol vehicle (control) or 10 nM R1881 or 10 nM calcitriol (1,25D) or 10 nM R1881 plus 10 nM calcitriol (R+D) (panel b). Media were changed every 3 days and fresh hormones added. On Day 6, the conditioned media were collected for IGFBP-3 protein assay and cells were harvested for DNA assay. The data for the DNA content of each treatment are expressed as percentage of control, whereas the IGFBP-3 protein levels are expressed as fold-change over the vehicle treatment. Each value is the mean ± SE of 3 experiments.

IGFBP-3 contributes to androgen-induced growth inhibitory action

To further determine the role of IGFBP-3 in androgen-induced growth inhibition in LNCaP cells, we used siRNA oligonucleotides to block IGFBP-3 production. As shown in Figure 5a, IGFBP-3 siRNA targeted to exon 4 prevented most of the rise in IGFBP-3 protein levels. Importantly, IGFBP-3 siRNA abolished the growth inhibition induced by androgens or by calcitriol used individually (Fig. 5b). In the case of the combination, however, IGFBP-3 siRNA caused a substantial reversal of the growth inhibition seen with the combination treatment, leaving a small residual inhibitory effect. As shown in Figure 5a, the hormone combination still induced a small amount of IGFBP-3 protein in the IGFBP-3 siRNA-transfected cells, which may contribute to the incomplete blockade of growth inhibition triggered by the combination treatment (Fig. 5b). The control (scrambled) siRNA had no effect on growth inhibition induced by calcitriol or by R1881 (Fig. 5b). In other experiments, a second siRNA oligonucleotide targeted to exon 5 of IGFBP-3 was tested. Similar results were obtained that confirm the target specificity of the siRNA (data not shown). These findings using siRNA knockdown of IGFBP-3 indicate that the induction of IGFBP-3 significantly contributes to the growth inhibitory action of androgens in LNCaP cells.

Figure 5. Growth inhibitory actions of R1881 in LNCaP cells can be blocked by IGFBP-3 siRNA. LNCaP cells were transfected with scrambled control siRNA or IGFBP-3 siRNA as described in Material and methods. Transfected cells were treated with ethanol vehicle (control) or 10 nM R1881, or 10 nM calcitriol (1,25D) or 10 nM R1881 plus 10 nM calcitriol (R + D) for 48 hr. Then, the conditioned media were collected for IGFBP-3 ELISA (panel a) and the cells were subjected to the [³H]-thymidine incorporation assay to assess the cell growth in each well (panel b). (c, d) LNCaP cells were treated with different doses of calcitriol (nM) or recombinant human IGFBP-3 (ng/ml) in 2% FBS-containing medium for 6 days. The cells were processed for assay of DNA content as a measure of cell growth (panel c) and conditioned media were collected for IGFBP-3 ELISA (panel d). The data for the DNA content of each treatment are expressed as percentage of control. The IGFBP-3 levels in the medium (endogenously secreted by calcitriol induction or exogenously added) are the total amount of immunoreactive IGFBP-3 per ml per well at the end of the experiment (mean ± SE of triplicate samples) from one representative experiment.

Exogenous addition of recombinant human IGFBP-3 protein

To substantiate the role of IGFBP-3, we tested the effect of exogenous addition of recombinant IGFBP-3 protein on cell growth. As demonstrated in Figure 5c, addition of 500 ng/ml of recombinant IGFPB-3 caused ∼35% growth inhibition, which was close to the effect induced by 1 nM of calcitriol. Calcitriol at 1 nM resulted in ∼40 ng/ml of extracellular IGFBP-3 (Fig. 5d), while addition of 500 ng/ml of IGFBP-3 protein led to an extracellular concentration of ∼245 ng/ml in the medium measured by IGFBP-3 ELISA. Thus, it takes a higher concentration of exogenous IGFBP-3 added to the medium to achieve the degree of growth inhibition seen with endogenously stimulated IGFBP-3. Nevertheless, the data confirm that IGFBP-3 can function as a growth inhibitor as reported previously1, 2, 7, 8, 10, 11, 12 and that IGFBP-3 plays an important role in androgen- or calcitriol-mediated growth inhibition in LNCaP cells.

The CDK Inhibitors p21 and p27 are involved in the growth inhibition mediated by IGFBP-3

Next, we examined the role of the CDKIs p21 and p27 in androgen-induced growth inhibition in LNCaP cells. Cell extracts were isolated from siRNA-transfected LNCaP cells following the various treatments. The levels of p21 and p27 were determined by Western blot analysis. As shown in Figure 6, in the control siRNA-transfected cells, 10 nM R1881 or calcitriol treatment resulted in an increase in p21 (Fig. 6a, lanes 2–3) and p27 (Fig. 6b, lanes 2–3). Transfection of IGFBP-3 siRNA, on the other hand, blocked the increase in p21 and p27 induced by androgens or calcitriol (Figs. 6a and 6b, lanes 5–6). The changes in the levels of p21 and p27 were reflected in alterations in cell cycle distribution measured by FACS (Fig. 6c). In the control siRNA-transfected cells, 48 hr of incubation with R1881 or calcitriol caused significant increases in the accumulation of the cell population in the G1 phase accompanied by a decrease in the proportion of cells in S/G2/M. The combination of R1881 and calcitriol produced a greater effect on the cell cycle distribution. Treatment with IGFBP-3 siRNA significantly affected this distribution with a reduction in the increase in G1 cells by R1881 or calcitriol and a smaller decrease in S/G2/M fraction compared with the ethanol-treated IGFBP-3 siRNA-transfected cells. Even the IGFBP-3 siRNA itself caused a slight increase in cells in G1 and some decrease in cells in S/G2/M. The incomplete reversal of the induced changes in the cell cycle distribution by IGFBP-3 siRNA may partly be due to incomplete knockdown of IGFBP-3 production; however, other growth factors or pathways are also likely to be involved.39

Figure 6. Induction of p21 and p27 by R1881 in LNCaP cells can be blocked by IGFBP-3 siRNA. siRNA-transfected LNCaP cells were treated with ethanol (control) or compounds, as described in the legend of Figure 5, for 24 hr (panel a) or for 48 hr (panel b and c). Then, the conditioned media were collected for IGFBP-3 ELISA to monitor the effect of siRNA transfection, and the cells were subjected to Western blot analysis of p21 (panel a) and p27 (panel b) or flow cytometry (panel c). The protein expression levels of p21 (a) and p27 (b) are represented by relative intensity from densitometry quantitation of the Western blot after internal control (β-actin) correction. The FACS results (c) show the percentage of the cells in each phase (G1 or G2/M/S). The data represent one of two independent experiments.

Discussion

In the current study, we show that androgens and calcitriol act synergistically to induce IGFBP-3 expression that results in a greater inhibition of cell growth than either hormone alone in the androgen-dependent LNCaP cell line. Low concentrations of R1881, that stimulate cell growth, fail to induce IGFBP-3, whereas higher concentrations of R1881, that inhibit cell growth, upregulate IGFBP-3 expression with a dose-dependent increase that is associated with growth inhibition. Transfection of IGFBP-3 siRNA abrogates the growth inhibition by calcitriol and by R1881 and partially abolishes the growth inhibition seen with the combination. Furthermore, we demonstrate that the expression of p21 and p27 are stimulated by R1881 and calcitriol, which can be blocked by IGFBP-3 siRNA. The changes in the levels of p21 and p27 are reflected in a shift of cell cycle distribution with an increase in G1 arrest accompanied by a decline in S/G2/M fractions due to the hormone treatment.

Consistent with previous reports,16, 32, 40, 41, 42, 43 we have also observed an increase in protein levels of p21 and p27 in the LNCaP cells treated with 10 nM R1881 or calcitriol. However, the combination of R1881 and calcitriol does not further enhance p21/p27 induction more than the individual hormones (data not shown). Since the combination produces more IGFBP-3 and more growth inhibition, other antiproliferative pathways besides the p21/p27 CDKIs must be involved. Knockdown of IGFBP-3 expression using siRNA attenuates the growth inhibition as well as the induction of p21/p27 by androgens or by calcitriol when used individually suggesting that p21/p27 at least in part, mediate the antiproliferative actions of IGFBP-3 induced by androgens and calcitriol. Previous studies have also suggested that the antiproliferative action of calcitriol is partially mediated by p21.16, 42 Boyle et al.16 further demonstrated that IGFBP-3 was the mediator of the growth inhibitory effect of calcitriol and that p21 was a downstream effect, since immunoneutralization of IGFBP-3 or IGFBP-3 antisense oligonucleotides abolished p21 induction and relieved the LNCaP growth inhibition by calcitriol. That study further showed that addition of recombinant IGFBP-3 could induce p21 and inhibit cell growth, thus demonstrating that IGFBP-3 was mediating these activities. The study by Martin and Pattison19 also showed that the cell proliferative effects of androgens and calcitriol in LNCaP cells were partly mediated by IGFBP-3. A study in human breast cancer cells has demonstrated that recombinant IGFBP-3 protein significantly inhibited cell growth, while IGFBP-3 antisense oligonucleotides reduced retinoid-induced cell growth inhibition, suggesting that growth inhibition by retinoids is mediated by IGFBP-3 as well.44 Thus, our current study not only confirms previous findings but also presents the first evidence that only growth inhibitory doses of androgens induce expression of IGFBP-3, which increases p21/p27 expression leading to cell growth arrest in LNCaP cells.

We have previously shown that IGFBP-3 is directly upregulated by both calcitriol and androgens in LNCaP cells.16, 17, 23 In this study, we further demonstrate that androgens synergistically enhance calcitriol-induced IGFBP-3 production. This synergistic effect by the two-hormone combination can be the result of several pathways. Pretreatment of LNCaP cells with calcitriol potentiates the androgen actions on IGFBP-3 expression, partially as a result of the increase in AR induced by calcitriol during the pretreatment as previously shown.38 Androgen binding to AR has been reported to slow the degradation of AR leading to prolonged androgen action.29 A recent study by Lou et al.45 demonstrated that androgens can strongly inhibit calcitriol-induced 24-hydroxylase gene expression and retard the catabolism of calcitriol thus enhancing calcitriol actions in LNCaP cells. Furthermore, the coexistence of functional hormone response elements within the IGFBP-3 promoter17, 23 may act as a hormone response unit with shared accessory elements and accessory factors causing the synergistic transcriptional responses.46 Although the magnitude of growth inhibition induced by the hormone combination is much less than the degree of IGFBP-3 induction, a small increase in endogenous IGFBP-3 expression may initiate a growth inhibitory action. Further increases in IGFBP-3 expression are not directly proportional to increased growth inhibition, suggesting that there likely are other growth inhibition pathways triggered by calcitriol and/or androgens.

It is interesting that it takes a higher concentration of exogenously added IGFBP-3 to cause growth arrest when compared with endogenously stimulated IGFBP-3. The reason for the endogenously stimulated IGFBP-3 concentration advantage is unclear. Perhaps the exogenous IGFBP-3 is less fully bioactive, has reduced cellular uptake, is less able to interact with IGFBP-3 receptors or is more susceptible to enzymatic degradation in the medium. Nevertheless, increasing evidence supports IGFBP-3 as a potential PCa therapeutic molecule.1, 7, 8, 10, 11, 12, 13, 14

There exist some differences between our data16, 23 (this study) and that of others18, 25 regarding the relationship between the antiproliferative actions of calcitriol or androgens and IGFBP-3 induction in PCa cells. Goossens et al.18 failed to observe any effect of recombinant IGFBP-3 protein on LNCaP proliferation either in the presence or in the absence of androgens. Similar to our finding, Stewart and Weigel25 also observed a significant growth inhibition in the presence of high concentrations of recombinant IGFBP-3 protein, but they did not find a blockade of calcitriol-induced growth inhibition by IGFBP-3 siRNA in LNCaP cells in serum-containing media. The reason for the conflicting results reported by different research groups remains unclear. Since the current study was performed in serum-containing medium, serum does not appear to completely explain the discrepancy. The differences in cell lines and cell culture conditions adopted by different laboratories, which can cause differential sensitivity to growth factors and hormones even within the same cell line47, 48 may be one of the causes for the variation in findings.

Whereas the action of calcitriol to stimulate an antiproliferative pathway is expected,24 androgens are well known to promote PCa growth in patients and androgen deprivation therapy causes regression of PCa. Since our current and previous data23 show that androgens stimulate the expression of IGFBP-3, an antigrowth proapoptotic protein, these results raise an interesting question about the role of androgen-stimulated IGFBP-3 in PCa. Recently Hendriksen et al.49 using microarrays found that in the early stage of PCa, the androgen-AR pathway stimulates genes involved in growth promotion or inhibition as well as cell differentiation with a lower growth rate produced. However, the growth inhibitory AR target genes in high-grade PCa are selectively downregulated. These findings provide potential explanations for our somewhat unexpected conclusion that androgens induce a growth-inhibiting gene, IGFBP-3 in LNCaP cells.

In conclusion, our results reveal a novel mechanism by which high concentrations of androgens exert growth inhibitory effects at least partially through the IGFBP-3-p21/p27 pathway in the well-characterized LNCaP human PCa cell line.

Acknowledgements

The authors thank Dr. Donald J. Tindall (Mayo Clinic, Rochester, MN) for the PSA-LUC reporter construct. We greatly appreciate Ms. Gu Zhang (Genentech, South San Francisco) for her help on FACS data analysis. The authors also thank Dr. Srilatha Swami (Stanford University Medical Center, SUMC) for her assistance in the preparation of the figures and Dr. Aruna Krishnan (SUMC) for her helpful suggestions.