Phosphorylation by DNA-Dependent Protein Kinase Is Critical for Apoptosis Induction by Insulin-Like Growth Factor Binding Protein-3

The activity of insulin-like growth factor (IGF)-I and IGF-II is regulated by a family of six high-affinity binding proteins. IGF binding protein (IGFBP)-3 is the most abundant of the IGFBPs in serum, where it forms a ternary complex with acid labile subunit and IGF (1). In addition to its role in regulating IGF action, IGFBP-3 exerts many IGF-independent effects to inhibit cell proliferation and enhance apoptosis in many cell types, including prostate (2) and breast (3–5) cancers.

IGFBP-3 has been reported in the nucleus of many cell types and contains a nuclear localization sequence (NLS) that facilitates nuclear uptake (6–8). Extracellular IGFBP-3 is rapidly internalized via transferrin receptor and caveolin and is transported into the nucleus by importin-β (9, 10). Once localized to the nucleus, IGFBP-3 interacts with the nuclear receptor retinoid X receptor (RXR) α to promote apoptosis by a mechanism that involves the nucleo-mitochondrial shuttling of RXRα/Nur77 (11, 12). However, IGFBP-3 may function in different ways to induce apoptosis because IGFBP-3 lacking a functional NLS is reported to promote apoptosis in breast cancer cells (13). However, little is understood about the cellular mechanisms regulating IGFBP-3 action.

IGFBP-3 is subject to post-translational modifications, such as glycosylation and proteolysis, and also contains consensus phosphorylation sites for a variety of protein kinases. In particular, Ser¹¹¹ and Ser¹¹³ have been described as phosphoacceptor residues possibly for CK2 (14, 15). Phosphorylation of these sites may affect the ability of IGFBP-3 to become glycosylated because the S111A/S113A double mutant showed a strongly reduced glycosylation pattern (14). Phosphorylation of IGFBP-3 at the cell membrane of T-47D cells was reported to enhance IGF binding (16). IGFBP-3 can also be phosphorylated by DNA-dependent protein kinase (DNA-PK) and cyclic AMP–dependent protein kinase A (PKA) after incubation with recombinant enzyme and [γ-³²P]ATP (17). DNA-PK is a predominantly nuclear serine/threonine protein kinase, which is activated in response to DNA damage. It plays a role in numerous cellular processes, including DNA double-strand break repair, V(D)J recombination, telomere maintenance, and gene transcription (18). DNA-PK phosphorylates many transcription factors in vitro, including p53, a tumor suppressor that also functions to regulate the transcription of IGFBP-3 (19). Exogenously added IGFBP-3 that had been phosphorylated by DNA-PK displayed enhanced nuclear accumulation in Chinese hamster ovary (CHO) cells and decreased IGF binding compared with the nonphosphorylated form (17).

We investigated the significance of phosphorylation by DNA-PK for the cellular actions of IGFBP-3 in prostate cancer. We identify phosphorylation to be a critical step in the growth-inhibitory and apoptosis-promoting actions of IGFBP-3. DNA-PK-mediated phosphorylation enhances the nuclear accumulation of IGFBP-3 and is critical for interactions with its nuclear binding partner RXRα. Moreover, we reveal that Ser¹⁵⁶ is the phosphoacceptor residue for DNA-PK and that this phosphorylation event is crucial for IGFBP-3 to exert these effects.

Reagents. Recombinant nonglycosylated IGFBP-3 was provided by Insmed (Glen Allen, VA). Goat anti-human IGFBP-3 antibody was purchased from Diagnostic Systems Laboratories (Webster, TX); rabbit anti-DYKDDDDK (FLAG tag) and rabbit anti-caspase-3 antibodies were from Cell Signaling Technology (Danvers, MA). The mouse anti-β-actin and mouse anti-Hsp60 antibodies, pCMV-FLAG expression vector, and the CelLytic NuCLEAR cell fractionation kit were purchased from Sigma (St. Louis, MO). Mouse anti-DNA-PK catalytic subunit antibody was from Kamiya (Seattle, WA). I-Block was purchased from Applied Biosystems (Foster City, CA). The rabbit anti-RXRα antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-TBP (nuclear loading control) antibody was purchased from Abcam (Cambridge, MA). Pfx DNA polymerase, T4 DNA ligase, LipofectAMINE 2000 transfection reagent, and all cell culture reagents were purchased from Invitrogen (Carlsbad, CA). NU7026, horseradish peroxidase–conjugated rabbit anti-goat, goat anti-rabbit, and goat anti-mouse secondary antibodies were from Calbiochem (San Diego, CA). The rabbit anti-phosphorylated serine/threonine antibody was from Chemicon (Temecula, CA). SDS-PAGE precast gels and blotting equipment were purchased from Bio-Rad (Hercules, CA). Restriction enzymes were from Fermentas (Hanover, MD). Recombinant DNA-PK and its substrate, ATP, CellTiter 96 AQueous One Solution Cell Proliferation Assay, and Apo-ONE Homogeneous Caspase-3/7 Assay were purchased from Promega (Madison, WI).

Cloning and mutagenesis. The putative DNA-PK phosphorylation sites of IGFBP-3 (Ser¹⁵⁶, Ser¹⁶⁵, and Thr¹⁷⁰) have been previously described (17) and were confirmed using NetPhos program in the CBS prediction servers (20). The three putative phosphorylation sites were individually mutated to alanine to prevent their phosphorylation. IGFBP-3 in PBS was mutated using PCR-based mutagenesis (sense primers: S156A, 5′-AAGAAAGGGCATGCTAAAGACGCCCAGCGCTACAAAGTTGACTACGAGTCTCA-3′; S165A, 5′-AGCCAGCGCTACAAAGTTGACTACGAGGCTCAGAGCACAGATACCCAGAACTT-3′; T170A, 5′-AAAGTTGACTACGAGTCTCAGAGCACAGATGCCCAGAACTTCTCCTCCGAGTCCAA-3′ and their reverse complement antisense copies) using Pfx DNA polymerase. Template DNA was digested using DpnI (Fermentas), and all constructs (termed BP3, 156A, 165A, and 170A) were cloned into pCMV-FLAG (Sigma).

In vitro phosphorylation assay. IGFBP-3 was phosphorylated in vitro by DNA-PK in the presence of ATP following the manufacturer's instructions. Phosphorylated IGFBP-3 was then analyzed by SDS-PAGE followed by phospho-specific immunoblotting.

Cell culture. The LAPC4 prostate cancer cell line was a generous gift from Charles Sawyers (University of California at Los Angeles, Los Angeles, CA). LAPC4 cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, and 10 nmol/L R1881 (Perkin-Elmer Life Sciences, Wellesley, MA). 22RV1 prostate carcinoma cell line [American Type Culture Collection (ATCC), Manassas, VA] was maintained in RPMI 1840 supplemented with 10% FBS and 1% penicillin/streptomycin. M059K and M059J glioblastoma cell lines (ATCC) were cultured in F-12/DMEM supplemented with 10% FBS, 1% penicillin/streptomycin, and 1% nonessential amino acids. For individual experiments, cells were seeded at a final density of 1 × 10⁵/cm² in 96-well, six-well, or 10-cm plates and grown to 80% confluence in a humidified atmosphere of 5% CO₂ at 37°C before treatment. All treatments were carried out as indicated in serum-free medium.

Transient transfection. Cells growing on six-well plates were transfected using LipofectAMINE 2000 following the manufacturer's instructions. Briefly, 4 μg DNA was diluted in serum-free medium and combined with LipofectAMINE transfection reagent. Complexes were applied to cells in culture and incubated for 24 to 48 hours before analysis.

Immunoblotting. Cell lysates containing 20 μg protein were resolved by SDS-PAGE and transferred to polyvinylidene difluoride membrane (Bio-Rad). Membranes were blocked in 0.2% I-Block in PBS containing 0.1% Tween 20 for 3 hours at room temperature and then probed with the appropriate primary and secondary antibodies. Antibody-antigen complexes were visualized by Western Lightning Chemiluminescence reagents (Perkin-Elmer Life Sciences) and autoradiography.

Cell fractionation. Cells on 10-cm plastic dishes were treated as indicated. Nuclear and cytoplasmic fractions were harvested using CelLytic NuCLEAR cell fractionation kit following the manufacturer's instructions. Separated fractions were quantified and analyzed by SDS-PAGE. Validity of separation was determined by immunoblotting for TBP and Hsp60.

Immunoprecipitation. Cell lysate (50 μg) or conditioned medium (10 mL) was incubated with 5 μL goat anti-human IGFBP-3 antibody overnight at 4°C. Protein A-Sepharose (50 μL, 25%) was added and samples were incubated at 4°C for 1 hour. Bound protein was eluted in Laemmli sample buffer [60 mmol/L Tris (pH 6.8), 2% SDS, 10% glycerol, 0.1% bromphenol blue], and the phosphorylation status of IGFBP-3 was assessed by immunoblotting with phosphorylated-specific antibodies. For coimmunoprecipitation experiments, samples were immunoprecipitated as above and analyzed by SDS-PAGE followed by immunoblotting.

Analysis of apoptosis. Apoptosis was assessed in cells growing in 96-well plates using Apo-ONE Homogeneous Caspase-3/7 Assay following the manufacturer's instructions.

3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium cell proliferation assay. To assess cell viability/proliferation, cells growing in 96-well plates were treated as appropriate and analyzed by CellTiter 96 AQueous One Solution Cell Proliferation Assay following the manufacturer's instructions.

Statistical analysis. Statistical analyses were analyzed using Student's t test and are presented as mean ± SE. Differences were considered statistically significant when P < 0.05.

IGFBP-3 is phosphorylated by DNA-PK in vitro. It has been reported that adenoviral-derived IGFBP-3 can be phosphorylated by DNA-PK in the presence of [γ-³²P]ATP or using HeLa cell extract in the presence of a DNA-PK-specific substrate (17). To confirm our ability to detect in vitro DNA-PK-phosphorylated IGFBP-3, we incubated recombinant DNA-PK with recombinant IGFBP-3 in the presence of ATP. The resulting proteins were separated by SDS-PAGE and analyzed by phosphorylated-specific serine/threonine and IGFBP-3 antibodies. Phosphorylated-specific (serine/threonine) antibodies recognize IGFBP-3 only after incubation with DNA-PK, confirming that DNA-PK phosphorylates IGFBP-3 and that generic phosphorylated antibodies can be used to detect phosphorylated IGFBP-3 (Fig. 1A). We set out to confirm the phosphorylation of IGFBP-3 by DNA-PK in vitro using two systems. First, 22RV1 prostate cancer cells were incubated with 2 μg/mL IGFBP-3 in the presence and absence of 10 μmol/L NU7026, a specific ATP-competitive inhibitor for DNA-PK cells. IGFBP-3 was immunoprecipitated from cell lysates, and its phosphorylation status was analyzed by phosphorylated-specific immunoblotting. In the presence of NU7026, serine/threonine phosphorylation of both exogenously added (nonglycosylated, 29 kDa) and endogenous IGFBP-3 (glycosylated, 44 kDa) was reduced >3-fold (Fig. 1B). To confirm these observations, we used a paired cell system of glioblastoma cell lines that either lack (M059J) or express (M059K) DNA-PK (21). Endogenous IGFBP-3 was immunoprecipitated from M059K and M059J cells after 24 hours of incubation in serum-free medium in the presence or absence of NU7026, and phosphorylation was assessed using phosphorylated serine/threonine immunoblotting. Three-fold reduced phosphorylation of IGFBP-3 was observed in M059J cell lysates compared with M059K (Fig. 1C). In addition, phosphorylation of IGFBP-3 in M059K, but not M059J, cells was partially inhibited by coincubation with NU7026, confirming that IGFBP-3 is phosphorylated by DNA-PK in vitro. To determine whether secreted IGFBP-3 has been phosphorylated by DNA-PK, we incubated 22RV1 cells in serum-free medium in the presence or absence of NU7026 for 24 hours. IGFBP-3 was immunoprecipitated from conditioned medium, and its phosphorylation status was assessed by phosphorylated-specific immunoblotting. Similar amounts of both total and phosphorylated IGFBP-3 were detected in conditioned medium regardless of the presence of NU7026, suggesting that DNA-PK phosphorylation of IGFBP-3 does not occur during its secretion and that secreted IGFBP-3 does not get phosphorylated by DNA-PK in 22RV1 prostate cancer cells (Fig. 1D).

Phosphorylation of IGFBP-3 by DNA-PK is necessary for its growth-inhibitory and apoptosis-inducing actions. IGFBP-3 directly inhibits proliferation (22, 23) and induces cell death in many tumor cell types, including prostate, lung, colon, and breast cancers (2, 3, 24, 25). Many post-translational modifications of IGFBP-3 have been reported, including phosphorylation by kinases, such as DNA-PK (15, 17). However, little physiologic relevance for such modifications is understood. Potential effects of DNA-PK activity on the ability of IGFBP-3 to inhibit cell growth were investigated by incubating LAPC4 cells with increasing concentrations of IGFBP-3 (0, 1, 2, and 4 μg/mL) in the presence or absence of NU7026 for 72 hours in serum-free medium. Treatment of LAPC4 cells with IGFBP-3 significantly inhibited cell growth in a dose-dependent manner (Fig. 2). However, coincubation with NU7026 completely prevented the growth-inhibitory actions of IGFBP-3. This suggests that phosphorylation of IGFBP-3 by DNA-PK is essential for its antiproliferative actions in prostate cancer. To determine whether phosphorylation of IGFBP-3 by DNA-PK plays a role in enhancing or inhibiting its apoptotic actions, we incubated 22RV1 and LAPC4 prostate cancer cells with 2 μg/mL human recombinant nonglycosylated IGFBP-3 ± 10 μmol/L NU7026 and assessed apoptosis induction using a fluorogenic caspase-3/caspase-7 substrate. In both cell types, treatment with exogenous IGFBP-3 caused a 40% increase in cleavage of caspase substrate compared with serum-free control (P < 0.05 in LAPC4; P < 0.01 in 22RV1; Fig. 3A). However, when cells were incubated with IGFBP-3 in the presence of the DNA-PK inhibitor NU7026, apoptosis induction by IGFBP-3 was completely abrogated and levels of caspase substrate cleavage were comparable with control cells. Incubation of either LAPC4 or 22RV1 cells with NU7026 alone caused no increase in caspase activity compared with serum-free controls, suggesting that NU7026 inhibits apoptosis induction by IGFBP-3 directly by inhibiting its phosphorylation as opposed to an indirect cellular effect.

To confirm these findings, we assessed caspase-3/caspase-7 activity in M059K and M059J cells incubated in the presence and absence of 2 μg/mL IGFBP-3 for 24 hours. The addition of IGFBP-3 to M059K cells led to a 30% increase in caspase activation (P < 0.05; Fig. 3B). In contrast, M059J cells, which completely lack the catalytic subunit of DNA-PK, have no significant response to treatment with IGFBP-3 (Fig. 3B). Taken together, these data suggest that DNA-PK activity is essential for the growth-inhibitory and apoptosis-inducing actions of IGFBP-3.

Phosphorylation of IGFBP-3 by DNA-PK enhances nuclear accumulation and is essential for interactions with RXRα. We have previously shown that the apoptosis-inducing actions of IGFBP-3 require its internalization, nuclear localization, and interaction with the nuclear receptor RXRα (10, 11). Because phosphorylation of IGFBP-3 by DNA-PK is also essential for its apoptosis-inducing actions, we hypothesized that phosphorylation of IGFBP-3 is necessary for its interaction with RXRα. 22RV1 prostate cancer cells were incubated for 24 hours in serum-free medium in the presence or absence of NU7026, and cytoplasmic and nuclear fractions were isolated. IGFBP-3 immunoblotting showed impaired nuclear localization of IGFBP-3 after treatment with NU7026 (Fig. 4A). To determine whether impaired nuclear localization resulted in reduced RXRα binding, we incubated 22RV1 cells with 2 μg/mL IGFBP-3 in the presence and absence of 10 μmol/L NU7026 for 24 hours in serum-free medium and harvested whole-cell extracts. IGFBP-3 was immunoprecipitated from lysates and analyzed by SDS-PAGE. Coimmunoprecipitation of RXRα was detected by immunoblotting in cells incubated with IGFBP-3 alone (Fig. 4B). However, RXRα was no longer detected in the IGFBP-3 immunoprecipitation complex when cells were coincubated with IGFBP-3 and NU7026, correlated with reduced serine/threonine phosphorylation, suggesting that inhibiting the phosphorylation of IGFBP-3 by DNA-PK prevents its interaction with RXRα. These data were confirmed by assessing the ability of IGFBP-3 to interact with RXRα in M059K and M059J cell lines (Fig. 4C). When incubated with exogenous IGFBP-3, RXRα could be coimmunoprecipitated from M059K but not M059J cells, correlated with the phosphorylation status of IGFBP-3. When M059K cells were preincubated with NU7026, IGFBP-3 and RXRα no longer coimmunoprecipitated, confirming that DNA-PK activity is necessary for this interaction to occur. This provides a potential mechanism for the lack of apoptosis induction by IGFBP-3 observed in the absence of active DNA-PK.

Phosphorylation of Ser¹⁵⁶ is critical for apoptosis induction by IGFBP-3. A cluster of three potential DNA-PK phosphorylation sites (Q/E/D-S/T-Q) have been identified in the central nonconserved domain of IGFBP-3, Ser¹⁵⁶, Ser¹⁶⁵, and Thr¹⁷⁰, which are highly conserved among human, mouse, rat, bovine, and porcine IGFBP-3 (17, 26). To determine the contribution of each residue to the functional regulation of the apoptotic actions of IGFBP-3, we mutated each residue individually to alanine by PCR-based site-directed mutagenesis. We then assessed the ability of pCMV-IGFBP-3-FLAG, pCMV-IGFBP-3/S156A-FLAG, pCMV-IGFBP-3/S165A-FLAG, and pCMV-IGFBP-3/T170A-FLAG expression to induce apoptosis in LAPC4 and 22RV1 prostate cancer cells. The expression of transfected constructs was verified by IGFBP-3 immunoblotting (Fig. 5A). Equivalent levels of endogenous IGFBP-3 (lower band) were detected in control cell lysates and in cells transfected with IGFBP-3. Slightly higher molecular weight IGFBP-3 was detected in cell lysates transfected with all forms of IGFBP-3 but was absent in pCMV-FLAG control transfected cells, corresponding to FLAG-tagged transfected constructs. Equivalent expression of all forms of transfected IGFBP-3 was observed. We assessed the ability of IGFBP-3/S156A, IGFBP-3/S165A, and IGFBP-3/T170A to induce apoptosis compared with wild-type IGFBP-3 by evaluating caspase-3/caspase-7 activity in transfected LAPC4 cells. Transfection of wild-type IGFBP-3 caused a 60% increase in apoptosis compared with control transfected cells (P < 0.01) that was completely abrogated by incubation with 10 μmol/L NU7026 (Fig. 5B). Similarly, LAPC4 cells transfected with either pCMV-IGFBP-3/S165A or pCMV-IGFBP-3/T170A caused a 60% increase in caspase-3/caspase-7 activity that was also inhibited by coincubation with NU7026 (P < 0.05). In contrast, pCMV-IGFBP-3/S156A was unable to promote caspase activation either in the presence or in the absence of NU7026. To confirm these observations, we also analyzed caspase-3/caspase-7 activity in transfected 22RV1 cells (Fig. 5C). Cells transfected with wild-type IGFBP-3, IGFBP-3/S165A, or IGFBP-3/T170A displayed ∼40% increased levels of caspase-3/caspase-7 activity compared with control transfected cells (P < 0.01 and P < 0.05, respectively). The increased caspase activation was inhibited by coincubation with NU7026. In contrast, 22RV1 cells overexpressing IGFBP-3/S156A displayed comparable levels of caspase activation with control cells and were unaffected by the addition of NU7026. Together, these studies suggest that phosphorylation of Ser¹⁵⁶ by DNA-PK is essential for the apoptosis-inducing actions of IGFBP-3 in prostate cancer cells. Incubation of 22RV1 cells with NU7026 causes reduced nuclear localization of IGFBP-3 (Fig. 4A). To determine if reduced nuclear expression of IGFBP-3/S156A occurred, we transiently transfected 22RV1 cells with pCMV-FLAG, pCMV-IGFBP-3, pCMV-IGFBP-3/S156A, pCMV-IGFBP-3/S165A, and pCMV-IGFBP-3/T170A and isolated nuclear and cytoplasmic fractions. Cellular localization of transfected IGFBP-3 was assessed by SDS-PAGE followed by FLAG immunoblotting (Fig. 5D). Consistent with data obtained using NU7026 (Fig. 4A), we observed 3-fold reduced nuclear accumulation of IGFBP-3/S156A compared with wild-type IGFBP-3, IGFBP-3/S165A, and IGFBP-3/T170A. Validity of fractionation was confirmed by immunoblotting for Hsp60 (cytoplasmic fraction) and TBP (nuclear fraction). As expected, no FLAG immunoreactivity was detected in control transfected cell lysates. These data identify phosphorylation of Ser¹⁵⁶ by DNA-PK as critical for the apoptotic actions of IGFBP-3 in prostate cancer cell lines.

In addition to its role as the principal serum carrier of IGFs, IGFBP-3 also functions to potentiate and inhibit IGF action by regulating the bioavailability of IGFs to interact with the IGF type I receptor (1). In this way, IGFBP-3 can both ameliorate and abrogate IGF-stimulated cell proliferation and survival. Beyond its role of modulating IGF action, IGF-independent actions of IGFBP-3 have been described. For example IGFBP-3 is known to promote apoptosis in an IGF-independent manner in many cancer models (2, 3, 24, 25). In addition, the absence of apoptosis in senescent fibroblasts has been associated with the absence of nuclear IGFBP-3 (27). Interestingly, human papillomavirus type 16 E7 oncoprotein, which can override senescence to immortalize human primary cells, can directly bind to and target IGFBP-3 for degradation (28). Several other cellular binding partners for IGFBP-3 have also been identified, including RXRα and humanin (11, 29, 30). Although factors, including p53, vitamin D, and transforming growth factor-β, are known to regulate IGFBP-3 expression (19, 31, 32), the mechanism of action for many IGF-independent roles of IGFBP-3 and how such functions are regulated are poorly understood.

Protein phosphorylation and dephosphorylation are common mechanisms for regulating the activity of numerous proteins and transcription factors in response to changing stimuli and environmental conditions (33). Of the six IGFBPs, phosphorylation has been reported for IGFBP-1 and IGFBP-3 (14–17, 34, 35). Previous reports of the effects of phosphorylation on IGFBP-3 action have suggested that post-translational modification of IGFBP-3 in this way may play a role in the regulation of IGF binding and nuclear localization (16, 17). Interestingly, phosphorylation has been reported to both enhance (16) and inhibit (17) IGF binding by IGFBP-3. Although these seem to be conflicting reports, it is possible that these are cell-specific effects or that different kinases play distinct roles in enhancing or preventing IGF-IGFBP binding. We have now identified phosphorylation of Ser¹⁵⁶ of IGFBP-3 to be a critical step in the induction of apoptosis by IGFBP-3 in prostate cancer cells.

We have described the relevance of phosphorylation by DNA-PK for the roles of IGFBP-3 in prostate cancer. However, it is unclear what role phosphorylation by other kinases may play in regulating IGFBP-3 action. In vitro phosphorylation has been described by PKA and at residues consistent with consensus CK2 phosphorylation sites (14, 15, 17). Coverley et al. (15) showed an 80% decrease in [³²P]phosphate incorporation in CHO cells transfected with IGFBP-3 in which Ser¹¹¹ and Ser¹¹³ (potential CK2 phosphoacceptor sites) had been mutated to alanine. In addition to suggesting that Ser¹¹¹ and Ser¹¹³ can be phosphorylated, these data also suggested that other residues in IGFBP-3 are also phosphorylated. Similarly, comparing the phosphorylation status of IGFBP-3 in the presence and absence of active DNA-PK reveals partial but not complete reduction of phosphorylation without active DNA-PK, again suggesting that phosphorylation of IGFBP-3 by multiple kinases may occur. Although Ser¹¹¹/Ser¹¹³ phosphorylation may influence IGF binding by IGFBP-3, what significance phosphorylation by CK2, PKA, DNA-PK, and other unidentified kinases may have on other actions of IGFBP-3 is yet to be determined.

IGFBP-3 interacts with its nuclear partner RXRα to induce apoptosis in prostate cancer cells (11) in a nuclear localization–dependent manner. However, recent reports have revealed that IGFBP-3 is also able to induce apoptosis independent of nuclear localization. For example, a form of IGFBP-3 with a mutated NLS, which was unable to interact with the cell membrane and had impaired internalization, was still able to promote apoptosis in breast cancer cells (13). This suggests that IGFBP-3 is also able to promote apoptosis without being internalized [e.g., by interacting with a specific extracellular receptor (36)]. It is therefore possible that IGFBP-3 can function in different ways to promote cell death in cancer cells possibly in a cell type–specific manner.

Because DNA-PK is also predominantly a nuclear protein, it is likely that the phosphorylation of IGFBP-3 by DNA-PK occurs in the nucleus, promoting the association of IGFBP-3 with RXR and resulting in the induction of apoptosis. As Ser¹⁵⁶ is in a region of IGFBP-3 distinct from the RXR-binding domain, it is possible that phosphorylation causes a conformational change in IGFBP-3 to facilitate interaction with RXRα. Such a mechanism would support data describing that nuclear localization is necessary for apoptosis induction by IGFBP-3 (11, 12). However, although DNA-PK acts predominantly as a nuclear kinase, low levels have been reported in cytoplasmic extracts derived from HTC rat hepatoma and HeLa cells (17, 37, 38), suggesting that phosphorylation of IGFBP-3 by DNA-PK may indeed occur in the cytoplasm or at the cell membrane. Indeed, phosphorylation of Akt by DNA-PK has been reported to occur at the cell membrane (39), suggesting that IGFBP-3 could potentially be phosphorylated during secretion or cellular uptake. However, our data suggest it is unlikely that phosphorylation by DNA-PK occurs during secretion in prostate cancer cells because there was little difference in phosphorylation status detected in IGFBP-3 from the conditioned medium of incubated with or without NU7026. However, phosphorylation by other kinases may occur either at the cell membrane or during secretion because phosphorylated IGFBP-3 is detectable in conditioned medium.

DNA-PK belongs to a family of large phosphatidylinositol 3-kinase-like proteins, which also includes ataxia-telangiectasia mutated and FRAP (40). Intracellular targets of DNA-PK kinase activity include p53, Mdm2, RNA polymerase II large subunit, and chromatin components (18, 41, 42). Functional DNA-PK consists of a catalytic subunit (DNA-PKcs) and a DNA-targeting heterodimer, Ku (43). Ku is tightly associated with DNA and functions by stimulating DNA-PKcs kinase activity toward DNA-bound targets and functions most effectively when the target protein is bound to the same DNA strand as DNA-PK itself (43, 44). Because DNA-PK requires DNA for its kinase activity, the demonstration of phosphorylation of IGFBP-3 by DNA-PK by ourselves and others would therefore seem to support growing evidence for either a direct or indirect role of IGFBP-3 in DNA binding and the modulation of gene transcription.

In summary, phosphorylation of IGFBP-3 by DNA-PK at Ser¹⁵⁶ is a critical step in the cellular functions of IGFBP-3 in modulating apoptosis and growth inhibition. The generation of a novel nonphosphorylated mutant of IGFBP-3 will provide a crucial tool for future study of the biological actions of IGFBP-3.

Grant support: Public Health Service grants NCI-P50CA92131, NIA-R01AG20954, NCI-R01CA100938, and DOD-PC050485 (P. Cohen) and DOD-PC050754 (L.J. Cobb).

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.