Paradoxical actions of endogenous and exogenous insulin-like growth factor-binding protein-5 revealed by RNA interference analysis.

Insulin-like growth factor-binding protein-5 (IGFBP-5) is abundantly expressed in bone cells. To determine the physiological role(s) of endogenous IGFBP-5 in regulating bone cell growth, differentiation, and survival, we used short double-stranded RNA (siRNA) to trigger RNA interference of IGFBP-5 in human osteosarcoma cells. The IGFBP-5 siRNA, targeting against a sequence unique to the IGFBP-5 middle domain, efficiently reduced IGFBP-5 mRNA and protein levels. The IGFBP-5 siRNA did not change the levels of IGFBP-4, a structurally related protein, or glyceraldehyde-3-phosphate dehydrogenase, a housekeeping gene. Knock-down of IGFBP-5 resulted in a significant increase in the number of transferase-mediated dUTP nick end labeling-positive cells and a decrease in a bone differentiation parameter (alkaline phosphatase activity) but had little effect on basal or insulin-like growth factor I-induced proliferation. Overexpression of a siRNA-resistant IGFBP-5 mutant in the IGFBP-5 knock-down cells restored the levels of survival to the control level; overexpression of IGFBP-4 or wild type IGFBP-5 had no such effect. Paradoxically, the addition of exogenous IGFBP-5 not only failed to rescue IGFBP-5 knock-down-induced apoptosis, it caused a further increase in apoptosis. Furthermore, the addition of exogenous IGFBP-5 alone increased apoptosis. This pro-apoptotic action of exogenous IGFBP-5 was abolished when IGF-I was added in excess, suggesting that exogenous IGFBP-5 increases apoptosis by binding to and inhibiting the activities of insulin-like growth factors. These results indicate that endogenous and exogenous IGFBP-5 exhibits opposing biological actions on cell survival and underscore the necessity and utility of studying IGFBP functions through loss-of-function approaches.

Insulin-like growth factor-binding protein (IGFBP)-5 1 is a secreted protein belonging to the IGFBP gene family. IGFBP-5 was first identified and purified from human bone extracts and conditioned medium obtained from cultured human osteoblastlike cells (1,2). Mature human IGFBP-5 consists of 252 amino acids that can be divided into a highly conserved cysteine-rich N-terminal domain, a conserved C-terminal domain, and a variable middle L domain (3,4). In adult rats, IGFBP-5 is expressed in many tissues, with high levels in the kidney, muscle, bone, lung, heart, brain, ovary, and testes (3,5,6).
The role of IGFBP-5 in bone cell growth and differentiation has been studied extensively (7)(8)(9). Although it is generally agreed that IGFBP-5 is important, there is a host of inconsistent, and even contradictory, findings in the literature regarding its precise role(s) in regulating bone cell proliferation and differentiation. For example, when added in combination with IGF-I to cultured human osteoblastic cells, IGFBP-5 was found to potentiate IGF-I-induced DNA synthesis and differentiation (1, 2, 10 -12). The potentiating effects of IGFBP-5 were attributed to its ability to bind to bone extracellular matrix because IGFBP-5 has a high affinity for hydroxyapatite (1,11). In contrast to these studies, others have suggested an inhibitory role of IGFBP-5 in IGF-I-induced DNA synthesis in this cell type (13,14). Likewise, stable overexpression of IGFBP-5 was found to inhibit mouse osteosarcoma cell proliferation (15). Recent studies indicate that IGFBP-5 itself can act as a bone growth factor and exert cellular effects that are independent of IGFs (16). There are also inconsistent in vivo findings. Although systemic administration of IGFBP-5 to intact or ovariectomized mice has been shown to stimulate bone performance parameters, including alkaline phosphatase activity, and/or to increase the osteoblast cell number (17,18), transgenic mice overexpressing IGFBP-5 in bone tissues exhibited transient decreases in bone performance parameters (19).
All the previous studies relied on gain-of-function approaches, either by adding large amounts of purified IGFBP-5 to or by overexpressing IGFBP-5 in cultured cells or tissues. Although these approaches are useful in demonstrating the biological capabilities of IGFBP-5, they do not necessarily provide insight into the physiological function(s) of the endogenous protein. Because IGFBP-5 is abundantly expressed in bone cells, interpretation of data from these studies using the addition/overexpression of IGFBP-5 approach is not always straightforward. Another complication is the presence of one or more IGFBP-5 protease(s) secreted by these cells and the fact that some IGFBP-5 fragments can exert IGF-independent actions in bone cells (18,20).
In this study, we investigated the role(s) of endogenous IGFBP-5 in bone cells using a loss-of-function approach. Using short double-stranded RNA (siRNA) to trigger RNA interference, we knocked down the endogenously synthesized IGFBP-5 in U2 osteosarcoma (OS) cells. Our results indicate that endogenous IGFBP-5 is involved in regulating U2-OS cell survival and differentiation, but it plays limited role in cell proliferation.

EXPERIMENTAL PROCEDURES
Materials-All chemicals and reagents were purchased from Fisher unless noted otherwise. Human IGF-I and IGFBP-5 were purchased from GroPep (Adelaide, Australia). Fetal bovine serum, McCoy's 5␣ medium, OPTI-minimum essential medium, penicillin-streptomycin, and trypsin were purchased from Invitrogen.
Plasmid Construction-Vector pSUPER, which generates biologically active siRNA from the polymerase III H1-RNA gene promoter (21), was kindly provided by Dr. Reuven Agami (Netherlands Cancer Institute). To generate the pSUPER-BP5 plasmid, a synthetic doublestranded oligonucleotide (5Ј-GAAGCTGACCCAGTCCAAGttcaagaga-CTTGGACTGGGTCAGCTTC-3Ј) was introduced into pSUPER. This oligonucleotide contains sequences corresponding to a 19-nucleotide sequence from human IGFBP-5 (nucleotides 474 -492, numbering with A of the ATG (start) codon as 1), a 9-nucleotide linker (lowercase letters), and the reverse complement of the same 19-nucleotide sequence. For the construction of various IGFBP expression plasmids, DNA fragments corresponding to the entire coding region and various domain(s) of human IGFBP-5 and IGFBP-4 were amplified by PCR as described previously (22). The cDNA fragments were subcloned into the HindIII/KpnI, XhoI/KpnI, or NotI/XbaI sites of pEGFP-N1 or pRcCMV2 vector (Clontech Laboratories, Palo Alto, CA). All of the plasmids were sequenced.
Cell Culture and Transfection-Human U2-OS cells were purchased from American Type Culture Collection (Manassas, VA) and cultured in McCoy's 5␣ medium supplemented with 10% fetal bovine serum in a humidified air atmosphere containing 5% CO 2 . For transfection, 1.0 ϫ 10 5 cells were seeded in 6-well plates (Falcon, Corning, NY). 1 g of pSUPER-BP5 or pSUPER DNA was transfected into cells following a previously reported method (23). In co-transfection experiments, a mixture containing 1 g of pCMV-IGFBP plasmid and 1 g of pSUPER-BP5, empty pCMV, and/or pSUPER was added. Two days after transfection, the cells were washed and incubated with fresh serum-free medium for 48 h. At the end of the incubation, the conditioned media and total RNA were prepared for further analysis.
Western Blot and Northern Blot Analysis-Equal amounts of protein were separated by 12.5% SDS-PAGE and transferred to Immobilon P membranes (0.45-m pore size; Millipore Corp., Bedford, MA). Ligand blot analysis was performed using digoxigenin-labeled IGF-I following published procedure (24). Immunoblot was performed following a published method using a GFP antibody (25). The protein concentration was determined using a BCA protein assay kit (Pierce).
For Northern blot analysis, total RNA was extracted using TriReagent (Molecular Research Center, Inc., Cincinnati, OH) following the manufacturer's instructions and was quantified by measuring UV absorption at 260 nm. RNA samples were size-fractionated in 1.2% agarose gels, blotted, and fixed onto Hybond N membranes (Amersham Biosciences). Hybridization was performed using [ 32 P]dCTP-labeled (ICN, Irvine, CA) human IGFBP-5 or IGFBP-4 cDNA probes as reported previously (26). Labeled GAPDH cDNA was used as a control. The band densities were quantified by scanning the autoradiographs and using Quantity One quantitation software (Bio-Rad). IGFBP-5 mRNA levels were standardized by the levels of GAPDH mRNA and expressed as a percentage of the wild type cell control group.
Biological Assays-The effect of IGFBP-5 knock-down on cell proliferation was determined by bromo-2-deoxyuridine (BrdU) labeling assay as described previously except that BrdU was added in the final 4 h (27). The result was expressed as the BrdU labeling index (percentage of BrdU-labeled cells in total cells counted).
Alkaline phosphatase activity was measured following a published method (28) using a kit from Sigma. Briefly, the cells were rinsed twice with 1ϫ phosphate-buffered saline, lysed in 250 l of lysis buffer (0.2% Nonidet P-40 in 1 mM MgCl 2 ), and sonicated for 10 s. The reaction mixture (1 ml) containing the phosphatase substrate was added to 200 l of cell lysate. After 30 min of incubation at 37°C, the reaction was stopped by adding 12 l of 1 N NaOH, and absorbance was read at 405 nm. Alkaline phosphatase activities were calculated using p-nitrophenol as a standard. The results, normalized by the total protein level, are expressed as percentages of the wild type cell control group.
Cell death was analyzed by terminal deoxynucleotidyl transferasemediated dUTP nick end labeling (TUNEL) assay using the In Situ Cell Death Detection kit (Roche Applied Science) following the manufacture's instructions. Briefly, the cells were fixed with 4% paraformaldehyde in 1ϫ phosphate-buffered saline for 1 h at room temperature. The cells were then permeabilized in 0.1% Triton X-100 in 0.1% sodium citrate for 2 min on ice and incubated in a wet chamber at 37°C for 1 h with the TUNEL reaction mixture. The reaction was stopped by rinsing slides with 1ϫ phosphate-buffered saline, and the cells were counterstained with 0.5 g/ml DAPI. Omission of the enzyme in the TUNEL reaction was used as a negative control, and cells treated with DNase I were used as a positive control.
Statistical Analysis-The values are presented as means Ϯ S.E. Differences among groups were analyzed by the Student's t test or one-way analysis of variance followed by Fisher's protected least significance difference test, using Statview (Abacus Concepts, Inc., Berkeley, CA). Significance was accepted at p Ͻ 0.05.

RESULTS
siRNA Interference of IGFBP-5-The effect of the IGFBP-5 siRNA in silencing the IGFBP-5 gene in human U2-OS cells was examined by directly measuring changes in IGFBP-5 mRNA as well as changes in secreted IGFBP-5 protein levels. Transfection of cells with the pSUPER-BP5 plasmid resulted in a significant decrease in the level of IGFBP-5 mRNA, whereas mock transfection with the empty pSUPER plasmid had no such effect (Fig. 1A). This effect was specific because the IGFBP-5 siRNA did not change the levels of IGFBP-4 mRNA, a structurally related protein, or GAPDH mRNA, a housekeeping gene (Fig. 1A). Results from three independent experiments indicated a 75% reduction (p Ͻ 0.05) in IGFBP-5 mRNA abundance by siRNA interference (Fig. 1B). Likewise, ligand blot analysis of the conditioned media revealed a 71% decrease (p Ͻ 0.05) in IGFBP-5 protein levels ( Fig. 1, C and D). This is likely an underestimation because the autoradiography was overexposed to appreciate any IGFBP-5 signal in the pSUPER-BP5transfected group. Therefore, IGFBP-5 siRNA can specifically and efficiently knock-down IGFBP-5 expression in human U2-OS cells.
Effects of IGFBP-5 siRNA Interference on Cell Proliferation, Differentiation, and Apoptosis-We next investigated the impact of the IGFBP-5 knock-down on U2-OS cell proliferation. As shown in Fig. 2A, RNA interference of IGFBP-5 had no significant effect on BrdU labeling index. To prove that these cells were capable of responding to a mitogen, pSUPER and pSUPER-BP5-transfected cells were treated with 100 ng/ml IGF-I. IGF-I treatment increased the BrdU index from 31 ϩ 1.3 to 43.4 Ϯ 2.3% (p Ͻ 0.05) in the pSUPER-transfected cells. In the pSUPER-BP5-transfected cells, IGF-I increased the BrdU index from 33.1 Ϯ 1.9 to 42.6 Ϯ 1.5% (p Ͻ 0.05) ( Fig. 2A). In contrast to the lack of effect on cell proliferation, the pSUPER-BP5-transfected cells had significantly lower alkaline phosphatase activity than that of the pSUPER group (Fig. 2B), suggesting that the loss of IGFBP-5 impairs U2-OS cell differentiation.
During these experiments, we noticed a consistent reduction in cell number in the pSUPER-BP5-transfected group. To determine whether IGFBP-5 knock-down leads to elevated apoptosis, TUNEL assays were performed. Knock-down of IGFBP-5 resulted in a marked increase in the number of TUNEL-positive cells (Fig. 3A). These TUNEL-positive cells exhibited morphological characteristics of apoptotic cells, including nuclear shrinkage, chromatin condensation, and nuclear fragmentation that presumably represented cells in the late stage of the apoptosis (Fig. 3B). Quantitative analysis indicated that the percentage of TUNEL-positive cells in pSU-PER-BP5-transfected cells (2.93 Ϯ 0.55%) was significantly higher than that of the pSUPER-transfected group, which is 0.30 ϩ 0.01% (p Ͻ 0.05; Fig. 3C). These data suggest that endogenous IGFBP-5 plays a role in maintaining U2-OS cell survival and differentiation but has a limited effect on proliferation.
Rescuing IGFBP-5 Knock-down-induced Apoptosis by Overexpressing a siRNA-resistant IGFBP-5 Mutant-To prove that the elevated apoptosis in the IGFBP-5 knock-down group was indeed due to the loss of IGFBP-5, a chimeric IGFBP (termed as IGFBP-545), in which the siRNA-targeted IGFBP-5 middle domain was replaced with that of IGFBP-4, was generated and subcloned into the pCMV mammalian expression vector. U2-OS cells were co-transfected with pSUPER-BP5 and pCMV-IGFBP-545. Northern blotting analysis revealed the presence of two transcripts in the transfected cells: one is ϳ6 kb, corresponding to the size of endogenous IGFBP-5 mRNA, and the other is about 1.1 kb, corresponding to the size of the transcript derived from the IGFBP-545 transgene (Fig. 4A). Results from three independent experiments indicated that IGFBP-5 siRNA caused 74 and 75% reduction (p Ͻ 0.05) in the endogenous IGFBP-5 mRNA abundance in the absence and presence of IGFBP-545 (Fig. 4B), respectively. Although the pSUPER-BP5 efficiently reduced the 6-kb IGFBP-5 mRNA levels, it did not change the levels of the IGFBP-545 mRNA (Fig. 4C). These data indicate that IGFBP-545 is resistant to IGFBP-5 siRNA, and expression of IGFBP-545 does not affect the knock-down of IGFBP-5 by pSUPER-BP5.
To show that IGFBP-545 retains the IGF binding capability and to compare its IGF binding to that of native IGFBP-5, IGFBP-545 and IGFBP-5 were C-terminally fused with EGFP and introduced into U2-OS cells. Cells transfected with the empty pCMV-EGFP vector were used as a control. Immunoblot analysis of the media conditioned by these cells indicated the presence of a 55-kDa EGFP fusion protein in the pCMV-IGFBP-5-EGFP-transfected group and a 53-kDa EGFP fusion protein in the pCMV-IGFBP-545-EGFP-transfected group (Fig.  5A). A 24-kDa EGFP protein was detected in the empty vectortransfected group (Fig. 5A). Ligand blot analysis revealed that IGFBP-5-EGFP and IGFBP-545-EGFP, but not EGFP, were capable of IGF binding (Fig. 5B). Densitometry analysis of the ligand blot and immunoblot results indicates that IGFBP-545-EGFP and IGFBP-5-EGFP bind to IGF-I to a similar degree (Fig. 5C).
Next, the rescue effect of reintroducing IGFBP-545 on apoptosis was examined. As shown in Fig. 6A, compared with the pSUPER group, there were significantly more TUNEL-positive cells in the pSUPER-BP5-transfected group (460.1 Ϯ 83.5% of the pSUPER control, p Ͻ 0.05). The basal level (100%) in the pSUPER group was 1.03% TUNEL-positive cells/total cells. Co-transfection with pCMV-IGFBP-545 and pSUPER reduced apoptosis to the basal levels (133.3 Ϯ 31.5% of the pSUPER control). Expression of IGFBP-545 alone did not significantly increase cell death (171.8 Ϯ 60.5% of the pSUPER control group). Transfection of cells with the pCMV-IGFBP-5 (wild type) construct had no rescuing effect (Fig. 6B). To show that the rescue effect of IGFBP-545 was not due to any role of the IGFBP-4 L domain, we transfected the cells with pCMV-IGFBP-4. Overexpression of IGFBP-4 had no effect on the IGFBP-5 knock-down-induced apoptosis (Fig. 6C). Overexpression of IGFBP-4 in the mock transfected cells resulted in a significant increase in apoptosis, probably by binding to endogenously produced IGFs and inhibiting their actions.
Exogenous IGFBP-5 Increases Apoptosis by Inhibiting IGF Actions-We also attempted to "rescue" the IGFBP-5 knockdown-induced apoptosis by adding pure, recombinant IGFBP-5 protein to U2-OS cells transfected with pSUPER or pSUPER-BP5. As shown in Fig. 7A, knock-down of IGFBP-5 resulted in a significant increase in the number of TUNEL-positive cells (313 Ϯ 21.1% of the pSUPER control, p Ͻ 0.05). Addition of purified IGFBP-5 (400 ng/ml), however, not only failed to attenuate the IGFBP-5 knock-down-induced apoptosis, it caused a higher rate of apoptosis (472.5 Ϯ 23.1% of the control group; p Ͻ 0.05). Also, the addition of IGFBP-5 to the mock transfected cells resulted in a significant increase in TUNEL-positive cells (320 Ϯ 38.5% of the control group; p Ͻ 0.05). These results indicate that exogenously added IGFBP-5 increases apoptosis in U2-OS cells. Because U2-OS cells are known to synthesize and secrete IGF-I and IGF-II (29), we postulated that exogenously added IGFBP-5 may increase apoptosis by sequestering the action of endogenously produced IGFs, which are well known survival factors (8). To test this idea, exogenous IGF-I and IGFBP-5 were added to the cells at a 4:1 molar ratio. As shown in Fig. 7B, the addition of IGF-I in excess abolished the IGFBP-5-induced elevation in apoptosis. These data sup- port the idea that exogenous IGFBP-5 increases apoptosis by sequestering IGF actions. DISCUSSION In this study, we have determined the roles of endogenous IGFBP-5 in human osteoblastic cells by employing the siRNA gene silencing approach. Our results indicate that endogenous IGFBP-5 is important for maintaining bone cell survival and differentiation. This conclusion is supported by the following observations. First, knock-down of IGFBP-5 leads to a significant increase in the number of apoptotic cells. Loss of IGFBP-5 also resulted in a significant decrease in alkaline phosphate activity but had no effect in either basal or IGF-I-stimulated DNA synthesis in U2-OS cells. Furthermore, expression of a siRNA interference-resistant IGFBP-5 mutant, but not native IGFBP-5 or IGFBP-4, rescued IGFBP-5 knock-down-induced apoptosis. Although there is no previous report on the effect of IGFBP-5 in regulating bone cell survival/apoptosis, the regulatory role of endogenous IGFBP-5 in bone cell differentiation agrees with previous studies using the addition/overexpression of IGFBP-5 approach (11, 17-19).
Our loss-of-function study suggests that endogenous IGFBP-5 plays a limited role, if any, in DNA synthesis in these cells. Previous studies by several groups indicated that exogenously administered or overexpressed IGFBP-5 either potentiates or inhibits DNA synthesis and/or cell proliferation in human or rodent osteoblastic cells (1, 2, 10 -14, 20). The different findings on the mitogenic activity of IGFBP-5 could be due to the different experimental approaches. In the present study, we explored the physiological role of endogenous IGFBP-5 through a loss-of-function approach, whereas all previous studies examined the biological effects of exogenous or overexpressed IGFBP-5. In addition, we measured the mitogenic  (11,16,18,20), interpretation of data from assays using long incubation times may not be as straightforward as previously believed. Because the siRNA interference depleted only 75% of the endogenous IGFBP-5, it is also conceivable that the remaining IGFBP-5 may be sufficient for maintaining cell growth.
A novel and intriguing finding made in this study is that although endogenous IGFBP-5 is required for optimal cell survival, exogenously added IGFBP-5 has a pro-apoptotic effect in human osteosarcoma cells. Both pro-apoptotic and anti-apoptotic activities have been reported for IGFBP-5 in nonskeletal cell types. A number of in vitro and in vivo studies have shown that IGFBP-5 increases/induces apoptosis when added to mammary gland/cells and prostate cells (30 -37). Butt et al. (38) have reported that expression of IGFBP-5 in human breast cancer cells caused growth arrest and induced apoptosis. They further showed that the IGFBP-5-induced increase in apoptosis was accompanied by an increase in the pro-apoptotic protein bax and a decrease in the anti-apoptotic protein bcl-2. In contrary to the pro-apoptotic effects of IGFBP-5 in mammary/ breast cancer cells, overexpression of IGFBP-5 in mouse C2 myoblasts through stable transfection enhanced cell survival under low serum conditions (39). The IGFBP-5 overexpressing cells are more resistant to tumor necrosis factor-␣-induced apoptosis (40). A recent paper reported that overexpression of wild type and a non-IGF-binding mutant form of IGFBP-5 increased survival and decreased apoptosis in C2 myoblasts, suggesting that IGFBP-5 may promote myoblast cell survival through a ligand-independent mechanism (41). Likewise, two studies in Hs578T human breast cancer cells indicated that IGFBP-5 inhibits ceramide-induced apoptosis (42,43). Although the possibility that IGFBP-5 may play opposing roles in cell death in different cell types cannot be excluded, our data raise the possibility that these contradictory findings may be in part attributed to the different effects of exogenously added or endogenously expressed IGFBP-5.
The opposite biological actions observed with endogenous and exogenous IGFBP-5 are intriguing but also puzzling. Butt et al. (38) have reported that although overexpression of IGFBP-5 in human breast cancer cells increased apoptosis, exogenously added IGFBP-5 had no such effect. It was unclear whether the exogenously added IGFBP-5 was degraded in those systems. We speculated that the pro-apoptotic action of exogenously added IGFBP-5 is due to its ability to bind extracellular IGFs and modulate the interaction between IGFs and their cell surface receptors. IGFs are well known survival factors for a wide variety of cell types (8), and U2-OS cells have been shown to synthesize and secrete both IGF-I and IGF-II (29). Because the IGF-IR mediates the biological actions of the IGFs, exogenously added IGFBP-5 may bind to IGFs and block their actions. This notion is well supported by our observation that the addition of exogenous IGF-I in excess abolishes the pro-apoptotic action of exogenous IGFBP-5. It is also in consistent with the finding that overexpression of IGFBP-4 increases apoptosis in these cells.
Although the pro-apoptotic action of exogenously added IGFBP-5 can be rationally explained by its interaction with IGFs, the molecular mechanism(s) underlying the anti-apoptotic activity of endogenous IGFBP-5 is not clear at present. The opposing biological actions of exogenous and endogenous IGFBP-5 are reminiscent to those of parathyroid hormonerelated protein (PTHrP). PTHrP, a secreted protein, has been shown to enter the nucleus and influence cellular events in an intracrine fashion (44,45). When targeted to the nucleus, PTHrP stimulates vascular smooth muscle cell proliferation; the same PTHrP inhibits cell proliferation when it is present in the extracellular environment and interacts with its cell surface receptor (44). IGFBP-5 contains a functional nuclear localization sequence in its C-terminal domain and is found in the nuclear compartment of cultured beast cancer cells, vascular smooth muscle cells, and MG63 osteogenic sarcoma cells (46 -49). Recently, we have shown that the IGFBP-5 N-terminal domain possesses intrinsic transactivation activity and that this activity is ligand-independent (49). Preliminary studies using IGFBP-EGFP fusion proteins have shown that overexpressed IGFBP-545-EGFP, but not IGFBP-4-EGFP, is targeted to the nucleus in U2-OS cells (data not shown). It is possible that the nuclear targeting of IGFBP-5 may play a role in the anti-apoptotic effect observed with the endogenous IGFBP-5. Future studies are needed to determine whether the anti-FIG. 7. Exogenous IGFBP-5 induces apoptosis by blocking IGF actions. A, addition of exogenous IGFBP-5 increases apoptosis in pSUPER-and pSUPER-BP5-transfected cells. Two days after transfection, the cells were treated with or without IGFBP-5 (400 ng/ml) for 48 h. The cells were then fixed and stained with TUNEL as described under "Experimental Procedures." The results are expressed relative value of the pSUPER group. The values are the means Ϯ S.E. of three independent experiments. *, p Ͻ 0.05 compared with the pSUPER group. #, p Ͻ 0.05 compared with pSUPER-BP5 group. B, addition of IGF-I abolishes exogenous IGFBP-5-induced apoptosis. U2-OS cells were treated with serum-free medium, IGF-I (400 ng/ml), pure IGFBP-5 (400 ng/ml), or IGF-I plus pure IGFBP-5, respectively. Fortyeight hours later, apoptosis was determined by TUNEL staining. The results are expressed relative to the value of the pSUPER group. The values are means Ϯ S.E. of three independent experiments. *, p Ͻ 0.05 compared with the IGFBP-5 group. apoptotic activity of endogenous IGFBP-5 is due to its nuclear targeting.
In conclusion, this study has elucidated a novel role of endogenous IGFBP-5 in regulating human osteoblast cell survival and differentiation. The results of this study suggest that endogenous and exogenous IGFBP-5 exhibits opposing biological actions in these cells. Our current understanding of the biological actions of IGFBP-5 and other IGFBPs is in large part derived from studies using exogenous proteins. The finding on the opposing biological actions of exogenous and endogenous IGFBP-5 underscores the necessity and utility of studying IGFBP functions through loss-of-function approaches. This study has also provided necessary information and a useful model cell system to further elucidate the molecular mechanisms of IGFBP-5 actions in regulating cell survival.