Rapid Apoptosis Induction by IGFBP-3 Involves an Insulin-like Growth Factor-independent Nucleomitochondrial Translocation of RXRα/Nur77*

Insulin-like growth factor-binding protein-3 (IGFBP-3) induces apoptosis by its ability to bind insulin-like growth factors (IGFs) as well as its IGF-independent effects involving binding to other molecules including the retinoid X receptor-α (RXRα). Here we describe that in response to IGFBP-3, the RXRα binding partner nuclear receptor Nur77 rapidly undergoes translocation from the nucleus to the mitochondria, initiating an apoptotic cascade resulting in caspase activation within 6 h. This translocation is a type 1 IGF receptor-signaling independent event as IGFBP-3 induces Nur77 translocation in R-cells. IGFBP-3 and Nur77 are additive in inducing apoptosis. GFP-Nur77 transfection into RXRα wild-type and knock-out mouse embryonic fibroblasts and subsequent treatment with IGFBP-3 show that RXRα is required for IGFBP-3-induced Nur77 translocation and apoptosis. Addition of IGFBP-3 to 22RV1 cell lysates enhanced the ability of GST-RXRα to “pull down” Nur77, and overexpression of IGFBP-3 enhanced the accumulation of mitochondrial RXRα. This unique nongenotropic nuclear pathway supports an emerging role for IGFBP-3 as a novel, multicompartmental signaling molecule involved in induction of apoptosis in malignant cells.

Over the past decade, multiple lines of investigation have validated insulin-like growth factor-binding protein-3 (IGFBP-3) 1 as an inducer of cellular apoptosis, effects that can be unrelated to its IGF binding (1). Most importantly, several groups have now reported successful in vivo treatment of cancer models with IGFBP-3, either as a single agent or in combination with chemotherapeutic agents (2)(3)(4). However, the molecular mechanisms by which IGFBP-3 induces apoptosis remain largely unknown at present.
Several novel IGFBP-3 binding partners have been recently identified that may participate in its IGF-independent proapoptotic effects (1). We and others demonstrated that retinoid X receptor-␣ (RXR␣) is a binding partner for IGFBP-3 (5,6) and that RXR␣ is required for IGFBP-3 apoptotic effects (5). Indeed, IGFBP-3 potentiates RXRE-mediated signaling while inhibiting signaling via other RXR␣ heterodimeric partners (5)(6)(7). Our discovery of IGFBP-3 binding to RXR␣ suggested that its apoptotic effects might involve an RXR␣-dependent transcriptional mechanism. Most published reports have evaluated IGFBP-3-induced apoptosis at 24 -72 h (8 -11), consistent with a transcriptional mechanism. However, we have recently described apoptosis activation by IGFBP-3 (as evidenced by caspase activation and histone associated DNA fragmentation ELISA) as early as 1-6 h after IGFBP-3 exposure, suggesting a mechanism that does not require de novo gene transcription (12,13). The orphan nuclear receptor Nur77 (also known as NGFI-B (14) and TR3 (15)) is a nuclear receptor transcription factor and is an important regulator of apoptosis in different cells (16). It is a member of the orphan steroid receptor family, which also includes Nor1 and Nurr1. This family is essential for apoptosis of self-reactive immature thymocytes following stimulation of the T-cell receptor (17,18). In response to synthetic apoptotic stimuli, Nur77 translocates from the nucleus to the mitochondria to induce cytochrome c release and apoptosis in leukemia (19), lung (20), ovary (21), stomach (22), colon (23), and prostate cancer cells (24). Subcellular localization of Nur77 is important for its biologic function. In the nucleus, it functions as a transcription factor to mediate cell proliferation events. Targeted to the mitochondria, it takes on a novel role as a mediator of apoptosis, not unlike the role played at the mitochondria by another transcription factor, p53 (25). Most importantly, Nur77 can also heterodimerize with RXR␣ (26) and participate in its transcriptional activities (26 -28). The mitogenic effect of Nur77 requires its DNA binding and transactivation functions in the nucleus, whereas both are dispensable for the apoptotic effects of Nur77 at the mitochondria (29).
Because the nuclear receptor RXR␣ is an intracellular binding partner for IGFBP-3, we hypothesized that IGFBP-3 would modify RXR␣/Nur77 heterodimeric DNA binding, shifting this heterodimer from a DNA binding state to one that targets mitochondria. Mitochondrial translocation of RXR␣/Nur77 would then result in the release of cytoplasmic cytochrome c, activation of intracellular caspases, and induction of apoptosis.
Here we report evidence that IGFBP-3 is a rapid biological signal molecule for RXR␣/Nur77 translocation. Our results reveal a new interaction between the nuclear receptor and IGFBP superfamilies and identify IGFBP-3 as a unique signal modulator of both traditional and novel nuclear receptor roles at the junction of cellular proliferation and apoptosis. (Mountain View, CA) provided recombinant human IGFBP-3. IGF-1 was a generous gift from Amersham Biosciences.  The commercial antibodies used are as follows: anti-human IGFBP-3  from DSL (Webster, TX); anti-Nur77 from Geneka Biotechnology (Montreal, Canada); anti-RXR␣ from Santa Cruz Biotechnology (Santa Cruz, CA); anti-cytochrome c from Pharmingen; and anti-␤-actin from Sigma. PMP70 Antibody and cathepsin S antibodies were from Zymed Laboratories Inc. (South San Francisco) and R&D Systems (Minneapolis, MN), respectively. For the Western immunoblot utilizing the R-MEFs, polyclonal rabbit anti-Nur77 antibody (Harlan Biosciences, Indianapolis, IN) was generated against two specific N-terminal peptides (Genemed Synthesis, San Francisco) derived from the human Nur77 peptide sequence. Sera were purified on a protein A/G column (Amersham Biosciences) and verified by Western blotting. Nur77 banding pattern was confirmed using CCRF-CEM nuclear extract (Active Motif, Carlsbad, CA). SDS-PAGE reagents, Tween, and fat-free milk were purchased from Bio-Rad. ECL reagents were from Amersham Biosciences. Full-length IGFBP-3 and NUR77 cDNAs were cloned into pLP-IRESneo mammalian expression vector via pDNR-mediated Creator TM technology (Clontech). The cloning of GFP-Nur77 has been described previously (24). Lipofectamine and PLUS Reagent were from Invitrogen. All other chemicals were from Sigma.
MEF Generation-Fibroblasts from an IGF-I receptor knock-out and corresponding wild-type mouse were generated from 18-day embryos as described previously (30) and were designated RϪ and WT MEFs, respectively. The R-cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and geneticin (G418). All cells were used before passage 6.
Apoptosis ELISA-Cells (2500 cells/cm 2 ) were seeded on 96-well plates. Following overnight attachment, cells were washed with PBS and serum-starved overnight, before incubation with the indicated conditions in a total volume of 100 l. Photometric Cell Death ELISA (Roche Applied Science) was performed according to the manufacturer's instructions to quantify histone-associated DNA fragments (mono-and oligonucleosomes) generated by apoptotic cells. This immunoassay is based on the sandwich-enzyme principle and used separate mouse monoclonal antibodies directed against DNA fragments and histones. Briefly, cell lysates were placed into a 96-well plate coated with streptavidin-linked, anti-histone antibody. Peroxidase-labeled mouse monoclonal DNA antibodies were used to localize and detect the bound, fragmented DNA by photometry of 2,2Ј-azino-bis-(3-ethylbenzathiazoline sulfonate) as the substrate. Calcium ionophore treatment served as the positive control, and serum-free medium served as the negative control. Each experimental condition was performed in triplicate. Reaction products in each 96-well plate were read using a Bio-Rad microplate reader. Mean absorbance data at 405 nm (ϮS.D.) were plotted. Studies to provide evidence that the absorbance values generated by the assay are linearly related to growth were performed and confirmed in a published paper from our lab (31).
Caspase Assays-The caspase assay was done using Apo-ONE TM homogenous caspase Ϫ3/Ϫ7 assay (Promega) and performed according to manufacturer's instructions. Recombinant human IGFBP-3 was used at a final concentration of 1 g/ml.
Immunofluorescence Confocal Microscopy-Ten thousand cells were plated on cover glasses in serum-containing media for 2 days. Cells were then incubated in serum-free media with or without IGFBP-3 before staining for immunofluorescence. After three washes in PBS, fixation and permeabilization of the cells were performed with 1% paraformaldehyde in PBS for 15 min at room temperature and 0.2% Triton X-100 in PBS for 15 min on ice, and cells were washed twice with PBS. Nur77 or IGFBP-3 protein localization was detected using human IGFBP-3/Nur77 polyclonal antibodies (diluted 1:1,000) followed by fluorescein/Texas Red antibody from Vector Laboratories. Specimens were incubated with primary antibodies in PBS for 1 h at room temperature, with secondary antibodies in PBS for 40 min at room temperature, and then incubated with Hoechst (Electron Microscopy Sciences, Ft. Washington, PA) for 2 min. Samples were analyzed by using an inverted confocal microscope (Leica, Inc., Germany) equipped with a digital camera (Himamatsu, Japan) and operated by QED image software.
Subcellular Fractionation Procedures-Nu-CLEAR protein extrac-tion kit TM was from Sigma. Subcellular fractions were isolated according to the manufacturer's protocol. The Apoalert TM cell fractionation kit (Clontech) was used to isolate a mitochondrial fraction from the cytoplasm of cells. The purity of the fraction was assessed by immunoblot for PMP70 (peroxisomal) and cathepsin S (lysosomal) contamination. Transient Transfections-Cells (2 ϫ 10 4 ) were seeded in 96-well culture plates. Reagents were appropriately scaled up to 6-well plates for transfections that were followed by mitochondrial isolation and subsequent Western immunoblotting. Transfections were done with Lipofectamine PLUS Reagent as directed by the manufacturer (Invitrogen). Typically, 50 ng of ␤-galactosidase expression vector (pSV-␤-gal, Promega, Madison, WI) and 50 ng of expression vector containing IGFBP-3 and/or Nur77 were mixed with carrier DNA to give 0.2 g of total DNA per well. After 24 -48 h of transfection, caspase activity was quantitated and normalized for transfection efficiency to measurements of aliquots of co-transfected ␤-galactosidase gene activity (␤-galactosidase enzyme assay system, Promega).
GST Pull-down-The GST-RXR␣ fusion vector encoding the fulllength RXR␣ molecule and was the generous gift of Dr. D. J. Mangelsdorf and has been described previously (32). GST-RXR␣ fusion protein was produced in GST-RXR␣-transformed Escherichia coli DH5 cells, which were lysed and loaded on glutathione-Sepharose 4B beads (Sigma). Ten g of purified GST-RXR␣ bound to beads was incubated with 500 g of cell lysate, with or without 200 ng of recombinant IGFBP-3 protein, and then separated by centrifugation. The bound proteins were analyzed by nonreducing SDS-PAGE followed by Western blotting using anti-Nur77 antibody. The experiments were repeated three times.
Densitometric and Statistical Analysis-Densitometric measurement of autoradiographs was performed by using computer-scanned densitometry. All experiments were repeated at least three times. Means Ϯ S.D. are shown. Statistical analyses were performed using analysis of variance utilizing InStat (GraphPad, San Diego). Differences were considered statistically significant when p Ͻ 0.005, denoted by ** as shown in the figures.

IGFBP-3 Induces Rapid Induction of Apoptosis-
We have observed previously that in A172 glioblastoma cells, IGFBP-3induced caspase activation reaches its maximum at 1-6 h, after which it decreases (12). Likewise, in human macrovascular umbilical vein endothelial cells, the vascular endothelial growth factor-induced survival of the human umbilical vein endothelial cells is inhibited by IGFBP-3, via the induction of apoptosis in a type 1 IGF receptor-independent manner utilizing the neutralizing antibody ␣IR3 (13). We therefore confirmed these rapid effects in a variety of cell lines. MEFs exhibited a 70% increase in apoptosis as evidenced by fluorometric assessment of caspase 3/7 activation (Fig. 1A) as early as 2 h post-treatment. This induction was maximal to nearly 2.5-fold over base line at 6 h. Similarly, in the human glioblastoma cell line A172, an almost 2-fold increase in apoptosis was detected as early as 1 h (Fig. 1B). In the 22RV1 prostate cancer cell line (Fig. 1C), a significant 32% increase in caspase activation was induced by the addition of IGFBP-3. This subsequently rose to a 40 and 51% increase over serum-free levels at 6 and 24 h, respectively. These results confirm that IGFBP-3 induction of apoptosis, assessed in multiple cell lines, is a rapid event. Experiments were repeated three times.
IGFBP-3 Induces Rapid Nucleomitochondrial Translocation of Nur77-To investigate whether Nur77 translocation could mediate the pro-apoptotic effects of IGFBP-3 in CaP cells, we first established that IGFBP-3 leads to nucleomitochondrial translocation of Nur77. We performed a time course of IGFBP-3 treatment and observed the subcellular localization of Nur77 by indirect immunofluorescence confocal microscopy in 22RV1 CaP cells. Within 15 min there was strong cytoplasmic appearance of red-staining Nur77 compared with minimal cytoplasmic labeling at time 0 ( Fig. 2A). The faint nuclear staining seen at time 0 may reflect inaccessibility of anti-Nur77 antibody to the protein secondary to RXR heterodimerization or Nur77 homodimerization. The cytoplasmic Nur77 presence was maintained throughout this 2-h time course, although some Nur77 reappeared in a nuclear location by the end of 2 h.
To confirm that IGFBP-3 is a biologic Nur77 translocation signal, we also assessed relative Nur77 concentrations in nuclear and cytoplasmic fractions of IGFBP-3-treated 22RV1 CaP cells by Western immunoblot at 0, 0.5, 1, and 3 h (Fig. 2B). Cells treated with IGFBP-3 showed the gradual disappearance of nuclear Nur77 over the course of 3 h associated with some increase in the amount of cytoplasmic Nur77. Furthermore, a dramatic increase was shown to be via mitochondrial targeting as subdivision of the cytoplasmic fraction into a mitochondrial enriched fraction revealed the appearance of a prominent Nur77 band as detected by Western immunoblot as early as 1 h (Fig. 2C). Membranes were probed with PMP70 and cathepsin S to show that the isolated mitochondrial fraction was free of peroxisomal/lysosomal contamination. Similar results were obtained for the LAPC-4 prostate cancer cell line and A172 glioblastoma cell line (data not shown). A representative of three separate experiments is shown.
Rapid Mitochondrial Translocation of Nur77by IGFBP-3 Occurs via a Type 1 Receptor Independent Mechanism-The possibility that IGFBP-3 acts to induce apoptosis independently of IGFs and IGF receptors was investigated by testing the ability of IGFBP-3 to induce apoptosis in the IGF receptor-negative (RϪ) embryonic fibroblast cells derived from an IGF-1R knockout mouse (30). This effect was mediated in part by a type 1 IGF receptor independent mechanism as IGFBP-3 was still able to induce apoptosis, with a 32% increase over base line at 2 h, that was maximal at 6 h in type 1 IGF receptor-disrupted MEFs in a fragmented DNA/histone ELISA (Fig. 3A). These cells have been shown previously to neither bind nor respond to IGFs. IGFBP-3-induced mitochondrial translocation of Nur77 in this unique system was further demonstrated by immunoblotting analysis, which showed accumulation of Nur77 in the mitochondrial fraction (Fig. 3B). To demonstrate the purity of the mitochondrial fraction, expression of mitochondrial-specific protein Hsp60 and nuclear-specific protein poly(ADP-ribosyl) polymerase (PARP) is shown as well as immunoblots to PMP70 and cathepsin S (peroxisome/lysosome markers, respectively). These data suggest that Nur77 mitochondrial translocation by IGFBP-3 occurs independent of signaling via the type 1 receptor. This experiment was repeated three times.
Additive Apoptotic Effects of Overexpression of IGFBP-3 and Nur77-To determine whether mitochondrial targeting of Nur77 by IGFBP-3 plays a role in regulating the release of cytochrome c from mitochondria into cytosol, the location of cytochrome c was examined during the course of IGFBP-3 treatment. Immunoblotting of a cytoplasmic fraction that was depleted of mitochondria showed that the addition of 1 g/ml of IGFBP-3 caused a greater than 3-fold increase in the appearance of cytoplasmic cytochrome c at 30 min, which increased to almost 6-fold at 60 min and was sustained at 180 min (Fig. 4A).
The release of cytoplasmic cytochrome c is directly upstream of caspase activation in the mitochondrial pathway of apoptosis (33). Overexpression of Nur77 in thymocytes induces massive apoptosis (34). We co-expressed by transient transfection both IGFBP-3 and Nur77 in mammalian expression vectors to assess the effects on caspase activation in 22RV1 cells. Expression of both IGFBP-3 and Nurr7 in combination resulted in additive effects on caspase activation (Fig. 4B), indicating additivity between IGFBP-3 and Nur77 in their apoptotic effects. Protein expression was confirmed with immunoblots of whole

FIG. 2. IGFBP-3 induces nucleomitochondrial translocation of Nur77.
A, indirect immunofluorescent confocal microscopy of 22RV1 cells after treatment with IGFBP-3. Nur77 is labeled in red. Nuclei are labeled in blue. Note rapid appearance of cytoplasmic Nur77. SFM, serum-free medium. B, Western immunoblot of subcellular fractions of 22RV1 prostate cancer cells after treatment with 1 g/ml IGFBP-3, probed for Nur77. PARP and Hsp60 are loading controls and show purity of the nuclear fraction. C, 22RV1 mitochondrial fraction isolated after treatment with 1 g /ml IGFBP-3. Membrane was probed with PARP, PMP70, and cathepsin S to show purity of the mitochondrial fraction. cell lysates from transfected cells (Fig. 4C).

IGFBP-3-induced Nur77 Translocation Is RXR␣-dependent and Involves Co-migration of RXR␣/Nur77 Heterodimers to
Mitochondria-We reported previously that the nuclear receptor RXR␣ is a nuclear binding partner for IGFBP-3 and is required for IGFBP-3-induced apoptosis (5). Also, in response to nerve growth factor, nuclear RXR␣-Nur77 heterodimeric complexes translocate to the cytoplasm in PC12 pheochromocytoma cells (35). A recent paper (36) also describes the carrier role of RXR␣ to assist Nur77 translocation in the 9-cis-retinoic acid-dependent apoptosis of gastric cancer cells. We hypothesized that pro-apoptotic IGFBP-3 would translocate RXR␣/ Nur77 heterodimers to the mitochondria. To investigate the role of RXR␣ in IGFBP-3-induced Nur77 translocation, F9 RXR␣ ϩ/ϩ (wild type (WT)) embryonal carcinoma cells and F9 RXR␣ Ϫ/Ϫ cells were treated with 1 g/ml IGFBP-3 overnight, and nuclear fractions were isolated, resolved via SDS-PAGE, and immunoblotted for the presence of Nur77. Bands were quantitated by densitometric analysis and normalized to nuclear PARP. As expected, IGFBP-3 induced a Ͼ50% reduction of nuclear Nur77 in the wild-type cells, consistent with an RXR␣/Nur77 translocation event (Fig. 5A). In contrast, treatment of the sister RXR␣ Ϫ/Ϫ line with IGFBP-3 resulted in an increase, rather than decrease, in nuclear Nur77. To confirm this observation, we transfected green fluorescent protein (GFP)-Nur77 in the same cell lines, treated with IGFBP-3, and visualized these cells utilizing confocal microscopy. Expression of GFP-Nur77 in the RXR␣ WT line showed a predominantly nuclear distribution consistent with its function as a transcription factor at basal conditions (Fig. 5B). Treatment with 1 g/ml of IGFBP-3 resulted in the rapid appearance of extranuclear GFP-Nur77 observed within 15 min. This effect was not seen in the sister RXR␣ Ϫ/Ϫ line, as IGFBP-3 treatment had no effect on the translocation of nuclear GFP-Nur77. In addition, transfection of the Nur77 overexpression vector into F9 RXR␣ ϩ/ϩ cells induced a marked increase in caspase activation, whereas transfection into the sister F9 RXR␣ Ϫ/Ϫ line failed to induce any significant increase in caspase activity (Fig. 5C). Indeed, addition of 200 ng of IGFBP-3 to 500 g of 22RV1 cell lysate demonstrated that IGFBP-3 enhanced the ability of GST-RXR␣ to "pull down" Nur77 (Fig. 5D), indicating that IGFBP-3 augments the ability of RXR␣/Nur77 to physically associate. In fact, isolation of mitochondria from cells transfected with IGFBP-3 revealed a 3-fold increase in mitochondrial RXR␣ (Fig. 5E) that was not seen in cells transfected with control expression vector, demonstrating that IGFBP-3 leads to the co-export of RXR␣ and Nur77. Together, our results demonstrate that the mechanism of IGFBP-3-induced Nur77 translocation, like IGFBP-3-induced apoptosis, requires RXR␣ and involves co-migration of RXR␣/Nur77 heterodimers to mitochondria. DISCUSSION Multiple lines of in vitro, in vivo, and clinical evidence point to IGFBP-3 as an anti-cancer molecule (37). In vitro, IGFBP-3 has been shown to induce apoptosis in a variety of cancer models via both IGF-dependent and -independent mechanisms (1). In vivo, recent publications report efficacy either as single agent or in chemotherapy combinations in non-small cell lung cancer (2) and colon cancer (3). IGFBP-3 gene expression is commonly lost in human cancer cell lines and xenografts as detected in DNA microarray analysis of cancerous cells compared with noncancerous cells (38). Decreased IGFBP-3 expression is associated with prostate cancer progression, demonstrating more frequent loss of expression in advanced disease, in both human (39) and mouse (40) models. In addition, recent evidence demonstrates that methylation of the IGFBP-3 promoter is one mechanism by which the silencing of IGFBP-3 expression in cancer cells is achieved (41).
Despite promising pre-clinical evidence using IGFBP-3 as cancer therapy (2)(3)(4), controversy remains as to the complex role of IGFBP-3 in various tumors. IGFBP-3 modulates cellular proliferation with dual actions that either enhance IGFs or inhibit their actions as well as actions that are independent of its binding to IGFs (1). Evidence for this duality has been reported in renal cell (42,43), lung (44,45), breast, and other cancers (46). Most interestingly, an outcome prediction model for prostate cancer was established utilizing HoxC6 and IGFBP-3 expression, as IGFBP-3 was positively associated with Gleason score (47). However, recent expression profiling of HoxC6 small interfering RNA transfections and HoxC6 overexpression identified IGFBP-3 as a potential proapoptotic repression target of HoxC6 in prostate cancer (48). Clearly, more work needs to be done to examine the role of IGFBP-3 in cellular proliferation and apoptosis.
We have demonstrated recently (49) in a prostate cancer model, the requirement for IGFBP-3 secretion and re-uptake by endocytic pathways (specifically caveolin-and transferrin receptor-mediated) for apoptosis induced by transforming growth factor-␤. After internalization, IGFBP-3 rapidly localizes to the nucleus where it interacts with RXR␣ and other factors (1). Nuclear import is a nuclear localization signal-dependent process and is mediated by importin-␤ factor (50).
Our observation that IGFBP-3 translocates Nur77 has several important implications. First, because IGFBP-3 is a biological signal versus the previously described chemical apoptosis inducers (i.e. calcium ionophores, etoposide, rexinoids) (24), this implies that the normal prostate epithelial cell has an endogenous signal (IGFBP-3), which can induce a programmed cell death cascade upon cancer surveillance. In fact, we have

FIG. 3. Rapid mitochondrial accumulation of Nur77 by IGFBP-3 occurs via a type 1 receptor independent mechanism.
A, type 1 IGF receptor-disrupted mouse embryonic fibroblasts were treated with IGFBP-3 (1 g/ml) for the indicated times, and apoptosis was measured in a fragmented DNA/histone ELISA. B, the mitochondrial fraction was analyzed for expression of Nur77 by Western blotting. To demonstrate the purity of the mitochondrial fraction, expression of mitochondrial specific protein Hsp60 and nuclear specific protein PARP, as well as peroxisomal PMP70 and lysosomal cathepsin S are shown. A representative of three separate experiments is shown. Values are represented as percent of serum-free. **, p Ͻ 0.005 relative to serum-free (SF) conditions. reported recently that EWS/FLI-1, an abnormal transcription factor resulting from oncogenic fusion in Ewing's tumor, binds the IGFBP-3 promoter in vitro and in vivo and represses its activity. Moreover, IGFBP-3 silencing can partially rescue the apoptotic phenotype caused by EWS/FLI-1 inactivation. IG-FBP-3-induced Ewing cell apoptosis relies on both IGF-1-dependent and -independent pathways. These findings therefore identify the repression of IGFBP-3 as a key event in the development of Ewing's sarcoma (51).
IGFBP-3 mediates the effects of multiple anti-proliferative and pro-apoptotic biological agents including transforming growth factor-␤ (52), tumor necrosis factor-␣ (53), retinoids (54), p53 (55), and 1,25-dihydroxyvitamin D 3 (56). In addition, IGFBP-3 gene expression is commonly lost in human prostate cancer cell lines and xenografts and was detected in DNA microarray analysis of normal compared with cancerous cells (38). Decreased IGFBP-3 expression is associated with prostate cancer progression, demonstrating more frequent loss of expression in advanced disease in both human and mouse models (37)(38)(39). Low IGFBP-3 levels in prostate cancer imply impairment of RXR␣/Nur77 translocation and subsequent apoptosis of cancerous cells.
We have shown that expression of IGFBP-3 and NUR77 together are additive in their pro-apoptotic effects. The importance of both these genes that are inactivated on the cellular path to immortalization is supported by the fact that IGFBP-3 binding and proteolysis are the targets of the E7 protein encoded by human papillomavirus type 15, one of the few viral genes that can immortalize primary human cells and thereby override cellular senescence (57), and that NUR77 is inactivated by the Epstein-Barr virus transactivator EBNA2, essential for the immortalization of B-cells (58). These small, lean, viral genomes would presumably selectively inactivate critical apoptosis-inducing host proteins that would hinder the viral program of self-propagation. Additionally, two recent papers (36, 59) have now described a nongenotropic carrier function of RXR␣ to transport Nur77 to the mitochondria to initiate a mitochondria-dependent apoptotic pathway. Values are normalized to nuclear PARP to adjust for loading. **, p Ͻ 0.005 relative to no treatment. B, confocal microscopy of F9 cells transfected with GFP-Nur77. Note extranuclear appearance of GFP-Nur77 after treatment with IGFBP-3. C, caspase activation posttransfection of Nur77 expression vector. **, p Ͻ 0.005 relative to control vector alone. D, GST-RXR␣ pull down of 22RV1 cell lysates treated with IGFBP-3. IGFBP-3 enhances the ability of RXR␣ and Nur77 to physically associate. CCRF-CEM nuclei was used as a positive control for Nur77 protein expression. E, overexpression of IGFBP-3 enhances mitochondrial RXR␣ accumulation. Mitochondrial fraction of 22RV1 cells is transiently transfected with IGFBP-3 expression vector. Fraction is immunoblotted with anti-RXR␣. Finally, we have described a novel interface between the nuclear receptor superfamily and the growth and survivalregulating IGF-IGFBP axis. Beyond its initial description as a serum carrier for the growth-promoting IGFs, IGFBP-3 has emerged as a multifunctional, intrinsic, and IGF-independent signaling protein that mediates important autocrine and paracrine regulation of growth and homeostasis in a variety of tissues (1). Although we have described previously IGFBP-3 binding to the nuclear receptor RXR␣ and supershifting RXR⅐RXRE complexes in electrophoretic mobility shift assays, modulating traditional nuclear receptor roles as transcription factors via modulation of signaling via the RXRE and presumably taking on a co-activator/co-repressor role in the nucleus (5-7), we currently describe IGFBP-3 as a modulator of novel nuclear receptor roles as extra-nuclear mediators of cellular processes. The fact that both IGFBP-3 and Nur77 are dramatically suppressed by androgens and are up-regulated during apoptosis induced by castration in the ventral rat prostate affords another unique in vivo model to study IGFBP-3-induced Nur77 translocation (60,61). This phenomenon also suggests that uncontrolled androgen receptor signaling implicated in androgen-independent prostate cancer involves the loss of the IGFBP-3/Nurr77 apoptotic pathway.
It is now well recognized that Nur77 mediates apoptosis (62) through both transactivation-dependent (63,64) and -independent (17,19,20,22,24) pathways. The movement of transcription factors (such as RXR␣ and Nur77), kinases, and DNA replication factors between the nucleus and cytoplasm is important in regulating their activity (65). Nur77 interacts with Bcl-2 at the mitochondria, inducing a conformational change that exposes its BH3 domain, resulting in Bcl-2 conversion from an anti-to pro-apoptotic molecule (66). Our results regarding RXR␣/Nur77 suggest a model that may explain how RXR␣/Nur77 activity is regulated, at least partially, by the presence of IGFBP-3. Abnormal Nur77 transcriptional activity may have oncogenic potential because a Nur77 fusion protein that is 270 times as active as the native receptor in activating gene expression is produced through chromosomal translocation in extraskeletal myxoid chondrosarcoma (67). Recent x-ray crystallographic analysis of the Nur77 Drosophila homologue revealed the absence of both a classic ligand binding pocket and co-activator-binding site (68). Nur77 is often overexpressed in cancer cells, because of the uncontrolled expression of growth factors that induce its synthesis and subsequent transactivation (28,61). Agents, such as IGFBP-3, that induce Nur77 translocation may inhibit growth and promote apoptosis of cancer cells.
In conclusion, the elucidation of the signaling pathways involved in the anti-proliferative, pro-apoptotic effects of IG-FBP-3 is important both for our basic understanding of the mechanism of action of IGFBP-3 on a cellular level and to devise new therapeutic approaches to treat prostate cancer constituting IGFBP-3 or related derivatives alone or in combination with synergistic agents. Our present findings indicating that IGFBP-3 induces a rapid RXR␣/Nur77 translocation event along with previous findings of modulation of slower DNA transcriptional events may herald an "amplification" loop in prostate cancer apoptosis signaling. Mutagenesis and identification of the specific peptide regions of each molecule involved in apoptosis signaling will prove to be a beneficial adjunct to the delivery of these molecules in a broad range of cancer therapeutics.