Basic Fibroblast Growth Factor Induces Proteolysis of Secreted and ...

0013-7227/99/$03.00/0 Endocrinology Copyright © 1999 by The Endocrine Society

Vol. 140, No. 7 Printed in U.S.A.

Basic Fibroblast Growth Factor Induces Proteolysis of Secreted and Cell Membrane-Associated Insulin-Like Growth Factor Binding Protein-2 in Human Neuroblastoma Cells* VINCENZO C. RUSSO, GEORGIA REKARIS, NAOMI L. BAKER, LEON A. BACH, AND GEORGE A. WERTHER Centre for Hormone Research (V.C.R., G.R., N.L.B., G.A.W.), Royal Children’s Hospital Research Institute, Parkville 3052, Victoria, Australia; University of Melbourne, Department of Paediatrics (V.C.R., G.R., N.L.B., G.A.W.), Royal Children’s Hospital, Melbourne 3052, Victoria, Australia; and Department of Medicine (L.A.B.), Austin and Repatriation Medical Centre, Heidelberg 3084, Victoria, Australia ABSTRACT Insulin-like growth factor (IGF) action in the brain is modulated by IGF-binding proteins (IGFBPs) whose abundance can be altered by other locally expressed growth factors. However, the mechanisms involved are unclear. We here employed the neuroblastoma cell line SK-N-MC as a model to define the mechanisms involved in modulation of IGFBPs in neuronal cells. Western ligand blotting analysis and immunoprecipitation of conditioned media (CM) from SK-N-MC cells showed that in these cells, as in the brain, the most abundantly expressed IGFBP was IGFBP-2. However, IGFBP-2 was barely detectable in CM from cells treated with basic fibroblast growth factor (bFGF) without a change in IGFBP-2 messenger RNA (mRNA) abundance. These CM contained specific IGFBP-2 proteolytic activity, resulting in two IGFBP-2 fragments of 14 and 22 kDa. The activity

T

HE insulin-like growth factors (IGFs) are peptides that regulate cell growth, differentiation, and survival (1– 3). They are synthesized and active in most tissues, including the developing nervous system, and may act in an endocrine, autocrine, or paracrine manner (1, 2). The key role of the IGFs in the nervous system has been demonstrated in several in vivo and in vitro models (3). Studies of IGF transgenic and knock-out mouse models have shown that IGF-I is required for normal development and viability in both the fetal and newborn mouse (3). IGF-I knock-out mice have small brains, whereas overexpression of the IGF-I gene generates mice with larger brains. The latter is the result of suppressed apoptotic neuronal cell death, which occurs during normal brain development, when all of the components of the IGF system are expressed in a precise spatial and temporal manner (3). In vivo expression of IGF-I and IGF-binding proteins (IGFBPs) is dramatically altered in damaged neuronal tissue Received October 19, 1998. Address all correspondence and requests for reprints to: Associate Professor George A. Werther, Centre for Hormone Research, Royal Children’s Hospital, Flemington Road, Parkville, Victoria 3052, Australia, E-mail: [email protected]. * This project was supported by a grant from the National Health and Medical Research Council of Australia. (Grant No. 960265).

was inhibited by EDTA/phenylmethylsulfonyl fluoride or aprotinin. Competitive binding studies indicated that IGFBP-2 fragments had reduced binding affinity for IGF-I. bFGF induced IGFBP-3 mRNA and protein. Affinity cross-linking of [125I]IGF-I to neuroblastoma cell membranes followed by immunoprecipitation revealed a ;38 kDa [125I]IGF-I/IGFBP-2 complex. Cell surface-associated IGFBP-2 was also susceptible to bFGF-induced proteolysis, with the appearance of a single cross-linked 21-kDa complex with low affinity for IGF-I. These findings indicate that intact IGFBP-2 and the 14-kDa, but not the 22-kDa fragment, bind to the cell surface. Our data suggest that induction of IGFBP-2 proteolysis on neuronal cell surface is a novel mechanism whereby IGF availability is modulated by the local growth factor bFGF. (Endocrinology 140: 3082–3090, 1999)

(4), and administration of IGF-I rescues and prevents neuronal loss in such areas (4). IGF action is controlled at different levels including tissue and developmentally specific transcriptional and posttranscriptional regulation of the IGFs (5, 6). IGF cellular responses are further modulated by the six IGFBPs (2) via IGF/IGFBP interactions, which occur in the pericellular and/or extracellular space, affecting IGF diffusion/stability and regulating IGF-I targeting to its receptors (2). Furthermore, specific IGFBP proteases, including members of the kallikrein family (7), neutral and acid-activated cathepsins (8), or the matrix metalloproteinase family (9), generate fragments that have reduced binding affinity for IGFs (10), thereby facilitating the release of free IGF peptide. IGFBP-2 has been shown to associate with cell surface in several systems (11–13), thus creating additional binding sites for the IGFs. In the rat brain these nonreceptor IGF-I binding sites (13) are colocalized with areas rich in IGF-I receptors (14) and with high expression of IGF-I (15). The further observation that the affinity of IGFBP-2 for IGF-I is reduced when bound to PGs (13) suggests that membranebound IGFBP-2 could play a key role in targeting IGF-I to its receptors. The IGF system and IGF-I signaling can be modulated by other local growth factors (16, 17). Basic fibroblast growth

3082

IGFBP-2 PROTEOLYSIS IN SK-N-MC CELLS

factor (bFGF) acts synergistically with IGF-I in controlling growth and proliferation of chromaffin cells (18) and central nervous system precursors (19). Growth and survival of rat brain explants are sustained by a combination of bFGF and IGF-I (17, 20). In the rat brain olfactory bulb, bFGF differentially regulates IGFBP expression (21). In addition, bFGF up-regulates expression of the IGF-I receptor in neuroblastoma cells (22), with consequent increase in IGF tumorigenic activity. In the present study we employed the neuroblastoma cell line SK-N-MC, which synthesizes IGF-I receptors (23), IGF-I (24), and IGFBPs (25), as a model system for neuronal cells, to examine novel mechanisms involved in the regulation of IGFBP-2 abundance and, consequently, IGF-I bioavailability, induced by other local growth factors and hormones active in the brain. Materials and Methods

3083

course and dose-response experiments (data not shown), or obtained from previously published studies on this cell line (34) or related neuroblastoma cell lines (35). Medium was changed every 24 h, and conditioned medium (CM) was collected and stored at 220 C until analyzed. Cell numbers were determined by dye binding assay using napthalene blue black as previously described (36)(data not shown).

Western ligand blot (WLB) analysis To identify the IGFBPs secreted by the SK-N-MC cells CM (250 ml) samples were concentrated by ethanol precipitation (21) and analyzed by WLB analysis (37) using [125I]IGF-I (1.5 3 106 cpm/50 ml). Filters were finally exposed to a phospho-screen for 16 h and then to Kodak X-Omat AR films with intensifying screens for 5–15 days at 270 C.

Immunoprecipitation CM (500 ml) from cells cultured in the absence or presence of 25 ng/ml of bFGF, or [125I]IGF-I-cross-linked SK-N-MC membranes were immunoprecipitated with antisera to IGFBP-2, -3, -4, and -5 or normal rabbit serum (1:100), as previously described (21), and and analyzed by WLB or direct autoradiography.

Reagents Recombinant human-IGF-I was a gift from Dr. A. Skottner (KabiPharmacia, Peptide Hormones, Stockholm, Sweden). Bovine bFGF, human epidermal growth factor (EGF), human platelet-derived growth factor (PDGF), protease inhibitors, phenylmethylsulfonyl fluoride (PMSF), E-64, AEBSF, and pepstatin, and the random primed DNA labeling kit were from Boehringer Mannheim (North Ryde, New South Wales, Australia). [125I]IGF-I (;2000 Ci/mmol) and Hybond-N nylon membrane were purchased from Amersham (North Ryde, NSW, Australia). Anti-IGFBP-2 and anti-IGFBP-5 antisera were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). a-Hec1 rabbit antiserum to human IGFBP-3 was a gift from Dr. R. G. Rosenfeld (Oregon Health Science University, Portland, OR). Rabbit anti-IGFBP-4 was kindly supplied by Dr. S. D. Chernausek (University of Cincinnati, OH). Rat IGFBP-2 was purified as previously described (26). Recombinant human nonglycosylated Escherichia coli IGFBP-3 was supplied courtesy of Dr. C. Maack and A. Sommer (Celtrix Pharmaceuticals, Inc., Santa Clara, CA). Human pregnant serum, 31 weeks of gestation, was kindly supplied by Ms. L. De Rosa (Royal Women’s Hospital, Melbourne, Australia). Total cellular RNA was extracted from cells using RNAzol (Biotecx Laboratories, Inc., Houston, TX). Prepubertal rat kidney total RNA was supplied courtesy of Dr. S. Ymer (Centre for Hormone Research, Royal Children’s Hospital, Melbourne, Australia). Complementary DNAs for human IGFBP-2, -3, and -4 were obtained from Dr. S. Shimasaki (Whittier Institute, La Jolla, CA). Protein-A-Sepharose CL-4B, RIA grade BSA, 5-a-dehydrotestosterone (DHT), and aprotinin were purchased from Sigma Chemical Co. (St. Louis, MO). The neuroblastoma cell line SKN-MC was kindly supplied by Professor V. R. Sara (Queensland University of Technology, Brisbane, Australia). Disuccinimidyl suberate (DSS) was obtained from Pierce Chemical Co. (Rockford, IL). Chemical reagents (Analar grade) were purchased from BDH-Merck Pty Ltd (Kilsyth, Victoria, Australia). Nitrocellulose membranes (0.45 mm) were obtained from Schleicher & Schuell, Inc. (Dassel, Germany). PhosphorImager screens were from Molecular Dynamics, Inc. (Sunnyvale, CA) and X-Omat AR films from Eastman Kodak Co. (Rochester, NY).

Cell culture and growth factor stimulation To define growth factor regulation of IGFBPs synthesized by the neuroblastoma cells, cells were cultured in T 75 flasks or 24-well plates to about 60% confluency in Iscove’s medium (19) containing 10% FCS. Cells were then washed twice in PBS, incubated for 48 h in Iscove’s medium-2.5% FCS (low serum) followed by a 24-h incubation in serumfree Iscove’s medium (starvation medium). Cells were further incubated for 72 h in the presence or absence of 25 ng/ml of either IGF-I, bFGF, EGF, PDGF, or DHT at 1028 m. Growth factors and hormones were selected for their ability to elicit mitogenic, differentiative, and metabolic effects in neuroblastoma cells as shown in previous studies (27–33). Growth factor concentrations, time of exposure, and pretreatment in low serum and starvation serum were determined in preliminary time

Northern analysis Total cellular RNA was extracted from cells using RNAzol as per manufacturer’s instruction. Northern blot analysis was performed on 25 mg of total RNA after electrophoresis on 0.8% agarose formaldehydedenaturing gels, transfer to nylon filters, and cross-linking by alkali fixation. DNA probes (IGFBP-2, -3, -4) were labeled with [a-32P]deoxycytosine triphosphate (3000 Ci/mmol, 10 mCi/ml) at specific activity of more than 108 dpm/mg DNA. Filters were probed for 16 h at 65 C, washed in 2– 0.13 SSC/0.1% SDS at 65 C, and exposed to x-ray film at 270 C. Consistency of RNA loading was confirmed by reprobing stripped filters for glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

Protease assays and proteolytic fragmentation of IGFBP-2 To investigate the presence of proteolytic activity for IGFBP-2 in CM from cells cultured in the presence of bFGF (F-CM), we performed a CM mixing assay. CM containing IGFBP-2 (serum free, SF-CM) was mixed 1:1 with F-CM (250 ml SF-CM 1 250 ml F-CM) and incubated for 16 h at 37 C in the presence (1) or absence (2) of protease inhibitors (PI) (3 mm PMSF, 10 mm EDTA). As controls, each CM (SF-CM or F-CM) was mixed individually with fresh Iscove’s medium at 1:1 (250 ml CM 1 250 ml Iscove’s medium) and incubated as above. Samples were analyzed by WLB (37). Dried filters were exposed to x-ray film for 15 days. In addition, 20 ng of rat IGFBP-2 (26) or [125I]IGFBP-3 (20,000 cpm; specific activity, ;16 mCi/mg) (38) in 250 ml Iscove’s medium, were mixed with 250 ml of F-CM 6 PI and incubated for 16 h at 37 C as above. F-CM 1 PI was run as control. Proteolysis of [125I]IGFBP-3 (20,000 cpm) was induced by incubation with 10 ml of human pregnant serum (31 weeks gestation) 6 PI as above and run as control (not shown).

Immunoblotting Filters were incubated for 16 h at 4 C in 10 mm Tris/HCl, pH 7.4/ NaCl, 150 mm/1% BSA/0.1% Tween 20 (TBS-BT) with the anti-IGFBP-2 antiserum (1:3000). IGFBP-2 immunoreactivity was detected with the Vectastain Elite ABC kit following the manufacturer’s instructions (Vector Laboratories, Inc., Burlingame, CA).

Protease inhibitor assay To define the family specificity of the IGFBP-2 protease, 20 ng of purified rat IGFBP-2 were incubated with CM from bFGF-treated cells at 37 C for 16 h in the presence or absence of the following proteases inhibitors: 10 mm EDTA, 5 mm PMSF, 5 mm AEBSF, 2 mm pepstatin, 5 mm aprotinin, 20 mg/ml E-64. Incubation was terminated by placing the tubes on ice before WLB analysis and immunoblotting as described above.

3084


Affinity cross-linking of [125I]IGF-I 125

To determine and quantify variations in binding affinity of [ I]IGF-I for IGFBP-2 and its fragments, CM (85 ml) from cells treated for 48 h with 25 ng/ml of bFGF (with some intact IGFBP-2 present) was incubated in a final volume of 100 ml with [125I]IGF-I (; 60,000 cpm/tube) in the presence of unlabeled IGF-I (10 –1000 ng/ml) for 16 h at 4 C. Crosslinking with DSS and analysis of reduced samples were performed as we previously described (13). Dried gels were exposed overnight to a PhosphorImaging screen or to x-ray film for 5–15 days. Bands were then quantified by densitometry as described below. Experiments were performed twice and samples were in triplicate.

Densitometric analysis Cross-linking gels were analyzed by the STORM PhosphorImager system and quantified by Image QuaNT software (Molecular Dynamics, Inc.) or exposed to x-ray film and analyzed by Bio-Rad GS-670 Imaging Densitometer and quantified by Molecular Analyst TM/PC Image Analysis software (Bio-Rad Laboratories, Inc., Hercules, CA). Sigma Plot software (Jandel Corp., San Rafel, CA) was used to plot the data.

Membrane preparation, [125I]IGF-I binding, and cross-linking SK-N-MC membranes were obtained by modification of a previously described method (13). SK-N-MC cells, grown to confluency in a T-75 flask or cultured in the presence of bFGF as described above, were scraped in ice-cold homogenizing buffer (10 mm Tris/HCl, 2 mm PMSF, 1 trypsin inhibitor unit/ml of Aprotinin) and transferred to 15-ml tubes. Cells were disaggregated through 19-gauge and 23-gauge needle syringes, and the cell membrane suspension was then aliquoted into 1.5-ml tubes and centrifuged at 800 rpm for 5 min at 4 C. The supernatant was centrifuged at 14,000 rpm for 1 h at 4 C, and the pellet (membrane fraction, MF) was resuspended in ice-cold 10 mm Tris/HCl, pH 7.4. MF protein content was then adjusted to a total concentration of 100 mg/80 ml. BSA (0.1%) was then added for storage at 220 C until used. [125I]IGF-I binding and cross-linking to cell membranes was performed as we previously described (13). As a control, rat IGFBP-2 (26) was incubated with [125I]IGF-I followed by cross-linking.

Proteolysis of membrane-associated IGFBP-2 and [125I]IGFI cross-linking To investigate whether membrane-associated IGFBP-2 is proteolysed, SK-N-MC membranes were incubated with CM from bFGF-treated cells

FIG. 1. IGFBP-2 is the major IGFBP synthesized by SK-N-MC cells. A, CM from SK-N-MC cells cultured without growth factors (SF) or with bFGF (F), IGF-I (I), EGF (E), PDGF (P), or DHT (D) was analyzed by WLB as described in Materials and Methods. B, SF-CM 72 h and F-CM 72 h were immunoprecipitated with anti-IGFBP-2, -2/3, and -5 antisera and analyzed by WLB as described. Film exposures were for 5 days (panel A) or 15 days (panel B).

Endo • 1999 Vol 140 • No 7

in the presence or absence of protease inhibitors as follows. Membranes were pelleted and resuspended in 100 ml of F-CM 6 5 mm PMSF/5 mm EDTA. Pure rat IGFBP-2 (20 ng) was also incubated as above as control. All samples were incubated for 6 h at 37 C with periodic vortexing. Samples were harvested on ice, and those containing SK-N-MC membrane suspension were spun at 14,000 rpm for 20 min at 4 C. The supernatant was transferred to fresh tubes, and the pellet was resuspended in fresh Iscove’s medium. All samples, as well as SK-N-MC membranes or IGFBP-2 not incubated with F-CM, and F-CM used as control, were then incubated with [125I]IGF-I, cross-linked, and analyzed by 12% SDS-PAGE (13). Samples were analyzed in triplicate in each of two experiments.

Results IGFBP-2 is the major IGFBP synthesized by SK-N-MC cells

We investigated the presence of IGFBPs in CM from SKN-MC cells and studied their regulation by growth factors (bFGF, IGF-I, EGF, PDGF) and hormone (DHT) (Fig. 1). WLB revealed a major band of about 30 –32 kDa in the 24 h CM after each of the treatments (Fig. 1A). This band was identified as IGFBP-2 by immunoprecipitation (Fig. 1B, lanes b and d). Two additional faint autoradiographic bands were seen at about 40 kDa and 24 kDa, consistent with IGFBP-3 and IGFBP-4, respectively. The 40-kDa band was identified as IGFBP-3 by immunoprecipitation (Fig. 1B, lane h). Although the 24-kDa band is consistent with IGFBP-4, it was expressed at extremely low levels that made it undetectable by WLB after immunoprecipitation with an anti-IGFBP-4 antiserum (data not shown). bFGF reduces IGFBP-2 abundance in CM via posttranslational events

In the presence of bFGF, the 30- to 32-kDa band, identified as IGFBP-2 (Fig. 1B), was diminished at 48 h (Fig. 1A, lane F-48) and barely detectable in CM after 72 h (lane F-72). This reduced abundance of IGFBP-2 in F-CM was not related to variation in cell number (data not shown). Treatment with all the other growth factors and hormones, as also found in


preliminary dose-response and time course experiments (data not shown), did not affect IGFBP-2 protein abundance. Northern analysis of total RNA from 72 h bFGF-treated cells (Fig. 2A, lane c) showed no change in IGFBP-2 mRNA levels compared with untreated cells (Fig. 2A, lane a). IGFBP-2 mRNA was similarly expressed after 72 h of incubation with each of the growth factors (Fig. 2A, lanes a–f), as confirmed by the ratio of IGFBP-2 to GAPDH mRNA bands determined by PhosphoImager (data not shown). IGFBP-3 mRNA expression was only detectable in cells treated with bFGF (Fig. 2B, lane c) and correlated with the IGFBP-3 band detected by immunoprecipitation and WLB analysis (Fig. 1B, lanes e and h). In support of the low level of IGFBP-4 protein expression, IGFBP-4 mRNA was detected in all samples (Fig. 2A, lanes a–f). bFGF induces a protease for IGFBP-2

Because our data showed that decreased levels of IGFBP-2 after exposure to bFGF was not explained by decreased levels of IGFBP-2 mRNA, we investigated the presence of an IGFBP-2 protease. When CM from cells cultured in serumfree medium (Fig. 3, lanes a and b) was mixed with CM from cells cultured in the presence of bFGF (lanes i and j), without protease inhibitors (lanes e and f), the IGFBP-2 band was decreased. This decrease was blocked by addition of protease inhibitors (EDTA, PMSF, aprotinin) (lanes g and h). IGFBP-3 abundance (Fig. 3, lanes e– h) or [125I]IGFBP-3 (data not shown) was unaffected during the incubation, suggesting the presence of a specific IGFBP-2 protease. A cation-dependent serine-protease proteolyses IGFBP-2 to 14-kDa and 22-kDa fragments. Immunoblotting with an IGFBP-2 antiserum identified two IGFBP-2 fragments of 14 kDa and 22 kDa in CM from cells treated with bFGF (Fig. 4, lane a). CM from cells treated with bFGF was also able to

FIG. 2. bFGF does not alter IGFBP-2 mRNA while increasing IGFBP-3 mRNA. Total RNA extracted form SK-N-MC cells cultured for 72 h without growth factors SF (a) or with EGF (b), bFGF (c), IGF-I (d), DHT (e), or PDGF (f), and prepubertal rat kidney total RNA (g) (positive control for IGFBP-3 mRNA) were subjected to Northern analysis for IGFBP-2, -4, and GAPDH (panel A) and IGFBP-3 (panel B) mRNA. Loading controls are indicated by GAPDH hybridization (panel A) and ethidium bromide staining of 18S (panel A) or 28S RNA (panel B). Film exposures were for up to 15 days.

3085

FIG. 3. bFGF induces proteolysis of IGFBP-2. CM from cells cultured without growth fators (SF-CM) (lanes a– d) was mixed with CM from cells cultured in the presence of bFGF (F-CM) (lanes i–l) 6 protease inhibitors (EDTA/PMSF) and analyzed by WLB, as described in Materials and Methods. The mixed CM samples were run in lanes e– h. Filters were exposed to x-ray films for 15 days.

proteolyse pure rat IGFBP-2 to give fragments of comparable sizes (16 kDa, 22 kDa) (Fig. 4, right panel, lane i). The IGFBP-2 protease activity was completely inhibited by aprotinin and combined EDTA/PMSF and, to a lesser extent, by pepstatin, AEBSF and, E-64 (Table 1), indicating that a cation-dependent serine-protease or a cascade of proteases is probably involved in IGFBP-2 proteolysis. Addition of IGF-I (150 ng/ ml) to our samples did not prevent IGFBP-2 fragmentation (Fig. 4, lane d). Addition of bFGF (25 ng/ml) to CM from serum-free cultured cells did not induce fragmentation of IGFBP-2 in a cell-free system (Fig. 4, lane e), indicating that bFGF was acting indirectly to stimulate proteolytic activity. The 14-kDa and 22-kDa IGFBP-2 fragments bind IGF-I

IGFBP-2 fragments were detected by immunoblotting (Fig. 4) but not by WLB (Fig. 1), suggesting that these fragments have reduced or no binding affinity for IGF-I after SDSPAGE and transfer to nitrocellulose membranes. However, affinity cross-linking of [125I]IGF-I to CM from bFGF-treated cells revealed bands sized ;22 kDa, ;28 kDa, and ;38 kDa (Fig. 5A) most likely representing IGF-I cross-linked to 14kDa and 22-kDa fragments and intact IGFBP-2, respectively. Unlabeled IGF-I decreased [125I]IGF-I binding to intact IGFBP-2 (Fig. 5A, 38-kDa band) more readily than the IGFBP-2 fragments (Fig. 5A, 28- and 22-kDa bands). Thus, in the presence of 100 ng/ml of unlabeled IGF-I, the 38-kDa band was barely detectable (;10%) [Fig. 5A, 38-kDa band; panel B, 38-kDa (F) at 100 ng/ml] while the two fragments retained about 50% of binding (Fig. 5, panel A, 28- to 22-kDa bands; panel B, 28 kDa (Œ) and 22 kDa (n) at 100 ng/ml]. IGFBP-2 fragments therefore bind IGF-I in solution, although the affinity of this binding is reduced at least 10-fold. The 14-kDa fragment appears to bind IGF-I with lower affinity than the 21-kDa fragment. IGFBP-2 and its 14-kDa fragment associate with the cell membrane: two nonreceptor-IGF-I binding sites on the surface of SK-N-MC cells

To investigate whether IGFBP-2 was similarly bound to SK-N-MC cells, we performed [125I]IGF-I binding and crosslinking to SK-N-MC membranes. An affinity-labeled 38-kDa band, whose molecular size is consistent with an IGFBP/ [125I]IGF complex, was identified on SK-N-MC membranes (Fig. 6, lane d). Immunoprecipitation with the anti-IGFBP-2

3086

Endo • 1999 Vol 140 • No 7


FIG. 4. Proteolysis of human and rat IGFBP-2. Conditioned medium (CM) from cells cultured without growth fators (SF-CM) was mixed with CM from cells cultured in the presence of bFGF (F-CM) (lanes a, c, and d), and protease inhibitors (EDTA/PMSF) (lane a), or IGF-I (150 ng/ml) (lane d); SF-CM (lanes b and e) was incubated with bFGF (lane e); F-CM (lane f) was mixed with purified rat IGFBP-2 6 protease inhibitors (lanes h and i). Samples were analyzed by immunoblotting with an IGFBP-2 antiserum, as described in Materials and Methods.

TABLE 1. Inhibition of IGFBP-2 proteolytic activity Inhibitor

EDTA (10 mM) PMSF (5 mM) AEBSF (5 mM) Pepstatin (2 mM) E-64 (20 mg/ml) Aprotinin (5 mM) EDTA/PMSF

Specificity

Metalloprotease Serine (e.g. chymotrypsin, trypsine, and trombine) Serine protease Aspartic (e.g. Cathepsin D) Cysteine protease Plasmin, kallikrein, trypsin, chymotrypsin Metallo/serine protease

Degree of inhibition

111 11 1 111 11 1111 1111

IGFBP-2 protease blockade: A panel of protease inhibitors was used to prevent IGFBP-2 proteolytic activity in CM from bFGFtreated cells as described in Materials and Methods. Protease inhibitors were assigned activities of 1111, 111, 11, or 1 according to their ability to inhibit the proteolysis of IGFBP-2 by .90%, 60 –90%, 30 – 60%, or ,20%, respectively. Samples were analyzed by WLB, and band intensity was quantified by densitometry.

antiserum showed that the 38-kDa complex was indeed IGFBP-2/[125I]IGF-I (Fig. 6, lanes b and c). We further investigated whether the IGFBP-2 protease present in CM from cells treated with bFGF (F-CM) could affect membrane-associated IGFBP-2. Cell membranes from bFGF-treated cells show decreasing levels of membranebound IGFBP-2 after 48 h and more clearly at 72 h (Fig. 7, lanes 72 h). This finding correlates well with the observed decreased level of IGFBP-2 in F-CM shown in Fig. 1. Furthermore, coincubation of F-CM with SK-N-MC membranes induced disappearance of the cross-linked complex at 38 kDa (Fig. 8, lane k) and the appearance of a cross-linked complex at 22 kDa (Fig. 8, lanes e and f), suggesting proteolysis of cell surface-associated IGFBP-2 (38 kDa, lane k) to IGFBP-2 fragment (22 kDa, lanes e and f). There was no evidence that fragments in solution (contained in the F-CM, Fig. 8, lanes h and i) were passively binding to membranes (Fig. 8, lanes j and k). This further suggests that the appearance of IGFBP-2 fragments on membrane after F-CM-induced proteolysis is specifically occurring at the membrane, rather than in the

medium after IGFBP-2 dissociation from cell membrane. Therefore, these findings indicate that IGF-I potentially can bind to both intact IGFBP-2 and its 14-kDa fragment on the cell surface. In contrast, there was no evidence of the 22-kDa fragment on the cell membrane, suggesting that this fragment is unable to bind to cell surface. Discussion

Neuroblastomas are neural crest-derived tumors that rely on IGF-I and -II autocrine stimulation (39 – 42) and have been widely used as models to study the role of IGFs in neuronal cells (23–25, 42). The data reported here illustrate that, similar to the study of Kiess et al. (25), the neuroblastoma cell line SK-N-MC predominantly expresses and synthesizes IGFBP-2. In addition to its presence in CM, we found that IGFBP-2 also is associated with SK-N-MC membranes. In these cells we further found that IGFBP-2 is proteolysed by a bFGF-inducible protease that is active on both soluble and membrane-bound IGFBP-2. The resulting IGFBP-2 fragments have reduced binding affinity for IGF-I, and the smaller of these retains some binding to cell membrane. Proteolysis of IGFBP-2 has been demonstrated in vivo, under physiological conditions (10, 43), in pathophysiological conditions (44), and in vitro (10, 12, 45). Proteolysis of IGFBP-2 generates fragments of variable size, depending on cell line and species (40, 43, 46, 47), including a small fragment of about 14 –18 kD and a larger one of about 21–25 kDa. These fragments have been reported either not to bind or to have markedly reduced binding affinity for the IGFs (43, 46). In addition, IGFBP-2 proteases can be activated or induced by hormones, cytokines, and drugs (12, 40, 46, 48). These findings suggest that proteolysis may represent a further level of regulation of IGFBP-2 and possibly affects IGFBP2/IGF interactions and IGF action. Previous studies (12) however, have not identified such processes occurring at the cell surface. In the present study, proteolysis of soluble and membrane-


3087

FIG. 5. The 14-kDa and 22-kDa IGFBP-2 fragments bind IGF-I with reduced affinity. [125I]IGF-I was incubated with CM from cells cultured with bFGF for 48 h [48 h F-CM, containing both intact IGFBP-2, 38-kDa (F), and IGFBP-2 fragments, 28-kDa (Œ) 22-kDa (n) 6 unlabeled IGF-I], and cross-linking was performed with DSS. Samples were analyzed as described in Materials and Methods. A, Autoradiography; B, densitometric analysis; values represent mean 6 SD with n 5 3. Film exposure was for 15 days.

FIG. 6. IGFBP-2 associates with cell membranes. SK-N-MC cell membranes were incubated with [125I]IGF-I, cross-linked (CXL) with DSS (lanes a– d), and immunoprecipitated with an anti-IGFBP-2 antiserum (lanes b and c) or control serum (lane a). Samples were analyzed as described in Materials and Methods. Film exposure was for 10 days.

FIG. 7. bFGF alters levels of membrane-bound IGFBP-2. SK-N-MC cells were cultured with bFGF for up to 72 h, and their membranes were incubated with [125I]IGF-I binding and cross-linking in the presence or absence (1/2) of unlabeled IGF-I as described in Materials and Methods. Film exposure was for 7 days.

bound IGFBP-2 was detectable 48 h after the addition of bFGF. bFGF is a potent angiogenic factor (49) that regulates expression of matrix metalloproteinases (50, 51) and plasminogen activators (52, 53). Limited proteolysis of IGFBP-2 to fragments of similar sizes to those described in the present study has been recently described in a neuroblastoma cell line by Menouny et al. (40). In that study it was shown that the addition of plasminogen

FIG. 8. IGFBP-2 and its 14-kDa fragment associate with the cell membrane, two nonreceptor-IGF-I-binding sites on the surface of the SK-N-MC cells. SK-N-MC membranes (M) were incubated with CM from cells cultured with bFGF (F-CM) (IGFBP-2 proteolytic activity present) 1/2 protease inhibitors followed by [125I]IGF-I binding and cross-linking in the presence or absence of unlabeled IGF-I as described in Materials and Methods. S, Supernatant. Film exposure was for 10 days.

induced proliferation of neuroblastoma cells and IGFBP-2 proteolysis. This effect was blunted by TGFb, but enhanced by retinoic acid. Thus, bFGF could potentially induce proteolysis of IGFBP-2 by a protease of the plasmin family as described by Menouny et al. (40). Our finding that IGFBP-2 mRNA is unchanged during IGFBP-2 proteolysis, whereas in the study by Menouny et al. (40) IGFBP-2 mRNA is downregulated, suggests that differing mechanisms may be involved in this process in our cell line. Similar to our study in the rat brain (21) and described elsewhere (54 –57), we here show for the first time in neuroblastoma cells, that bFGF induces IGFBP-3 expression. The role of IGFBP-3 in the CNS is unclear. IGFBP-3 mRNA, which

3088

Endo • 1999 Vol 140 • No 7


is in low abundance during normal brain development (21, 58, 59), is induced in the rat brain after hypoxic/ischemic injury (58, 59). IGFBP-3 expression correlates with the processes of tissue repair, remodeling, maturation, and cellular differentiation of specific brain regions (21, 58, 59). However, IGFBP-3 inhibits cellular proliferation (60, 61), including mitogenic stimuli induced by bFGF and IGF-I (62). Some IGFBPs associate with the extracellular matrix or cell surface via glycoproteins, collagens, integrins, and glycosaminoglycans (2, 11, 13, 63). We have recently demonstrated that IGFBP-2 binds to cell membrane chondroitin-sulfate proteoglycans in the rat brain (13). Similar to our finding in the rat brain, we demonstrated IGFBP-2/IGF-I complexes on membranes from SK-N-MC cells, suggesting that cell surface association of IGFBP-2 is a common mechanism for differential localization of IGFBP-2 in the nervous system. In addition, we have shown, for the first time, processing of membrane-bound IGFBP-2 by a specific IGFBP-2 protease, generating a 14-kDa fragment, which is capable of binding IGF-I while simultaneously being bound to the cell surface. This fragment was not detectable on membranes of cells treated with bFGF and only seen on cell membranes after incubation with medium from bFGF-treated cells. The presence of this fragment in F-CM during incubation did not lead to its binding to membranes. These findings suggest that such proteolysis is specifically occurring on the cell surface rather than in the medium after IGFBP-2 secretion. Cell surface association of IGFBP-2 and its 14-kDa fragment raise intriguing questions about a potential role in regulating IGF-I access to its receptor. We have recently demonstrated that the binding affinity of IGFBP-2 for IGF-I is modestly reduced when it associates with glycosaminoglycans (13). In the present study, we have shown that the 14-kDa fragment of IGFBP-2 has markedly reduced affinity for IGF-I in solution. It is thus possible that proteolysis of membrane-bound IGFBP-2 provides a mechanism for creating perireceptor low-affinity IGF-I-binding sites. The detection of a reduced amount of intact IGFBP-2, but not IGFBP-2, fragments on cells treated with bFGF is likely to be the result of IGFBP-2 proteolysis and generation of such perireceptor low-affinity IGF-I-binding sites. These sites, as shown for the IGFBP-2 fragments in solution, are not readily displaceable and are thus inefficiently labeled by IGF-I. However, the precise localization of IGFBP-2 and its fragments relative to the IGF-I receptors is not known. Cell surface association of IGFBP-2 and its 14-kDa but not 21-kDa fragment suggests that the smaller fragment contains a cell surface-binding domain. We suggested that the heparin-binding domain (PKKLRP), present in rat IGFBP-2 at residues 160 –166 and in human IGFBP-2 at residues 179 –185, may be involved in these interactions (13). Two recent studies by Ho and Baxter (43) (human breast milk) and Ishikawa et al. (47) (rat meningeal cells) reported the isolation and aminoterminal sequence of IGFBP-2 fragments of 14 kDa and 16 kDa, respectively. The amino-terminal sequence of the human 14-kDa fragment (GGKHHLGLEEPKKLRPPPAR) (43) was identical to residues 169 –189 of human IGFBP-2. Similarly, the amino-terminal sequence of rat 16-kDa fragment (MGKGAKHL) (47) matched residues 148 –155 of rat IGFBP-2. Thus, both fragments contain the putative heparin-

binding domains. Although further studies are required to verify whether the 14-kDa fragment identified in SKNMC cells contains the heparin-binding domain, it is very likely that, similar to the 14-kDa fragment described by Ho and Baxter (43), this domain is present and possibly involved in the association of the 14-kDa fragment and intact IGFBP-2 wih the SK-N-MC cell surface. In conclusion, in our hands bFGF modulates the IGF system at several levels in neuroblastoma cells by inducing proteolysis of soluble and cell-associated IGFBP-2 to generate fragments with reduced binding affinity for IGF-I, and by enhancement of IGFBP-3 synthesis. It has also been shown in other neuroblastoma cell lines to increase expression of the type I IGF receptor (22), and enhance IGF-I expression (18, 19). While proteolysis of secreted IGFBP-2 might increase levels of free IGF-I, release of IGFBP-3 and proteolysis of membrane-associated IGFBP-2, a likely membrane reservoir of IGF-I, might further optimize IGF-I availability for receptors. The IGFBP-2 fragments thus may play a key role in mediating bFGF modulation of IGF-I action in these cells. Our findings further suggest a role for IGFBP-2 in modulating IGF action, with relevance for IGFBP-2 biology in degenerative brain diseases (64), brain injury (4), and in a wide variety of common malignancies (65– 69). Acknowledgments We would like to thank David Casley for iodination of IGFBP-3 and Dr. Amanda J. Fosang for helpful discussion of the manuscript.

References 1. Bach LA, Rechler MM 1995 Insulin like growth factor binding proteins. Diabetes Rev 3:38 – 61 2. Jones JI, Clemmons DR 1995 Insulin-like growth factors and their binding proteins: biological actions. Endocr Rev 16:3–34 3. D’Ercole AJ, Ye P, Calikoglu AS, Gutierrez-Ospina G 1996 The role of the insulin-like growth factors in the central nervous system. Mol Neurobiol 13:227–255 4. Guan J, Williams C, Gunning M, Mallard C, Gluckman P 1993 The effects of IGF-1 treatment after hypoxic-ischemic brain injury in adult rats. J Cereb Blood Flow Metab 13:609 – 616 5. Pao CI, Zhu JL, Robertson DG, Lin KW, Farmer PK, Begovic S, Wu GJ, Phillips LS 1995 Transcriptional regulation of the rat insulin-like growth factor-I gene involves metabolism-dependent binding of nuclear proteins to a downstream region. J Biol Chem 270:24917–24923 6. Sussenbach JS, Rodenburg RJ, Scheper W, Holthuizen P 1993 Transcriptional and post-transcriptional regulation of the human IGF-II gene expression. Adv Exp Med Biol 343:63–71 7. Cohen P, Peehl DM, Graves HC, Rosenfeld RG 1994 Biological effects of prostate specific antigen as an insulin-like growth factor binding protein-3 protease. J Endocrinol 142:407– 415 8. Conover CA, Perry JE, Tindall DJ 1995 Endogenous cathepsin D-mediated hydrolysis of insulin-like growth factor-binding proteins in cultured human prostatic carcinoma cells. J Clin Endocrinol Metab 80:987–993 9. Fowlkes JL, Suzuki K, Nagase H, Thrailkill KM 1994 Proteolysis of insulinlike growth factor binding protein-3 during rat pregnancy: a role for matrix metalloproteinases. Endocrinology 135:2810 –2813 10. Rajaram S, Baylink DJ, Mohan S 1997 Insulin-like growth factor-binding proteins in serum and other biological fluids: regulation and functions. Endocr Rev 18:801– 831 11. Reeve JG, Morgan J, Schwander J, Bleehen NM 1993 Role for membrane and secreted insulin-like growth factor-binding protein-2 in the regulation of insulin-like growth factor action in lung tumors. Cancer Res 53:4680 – 4685 12. Cohen P, Noveral JP, Bhala A, Nunn SE, Herrick DJ, Grunstein MM 1995 Leukotriene D4 facilitates airway smooth muscle cell proliferation via modulation of the IGF axis. Am J Physiol 269:l151–1157 13. Russo VC, Bach LA, Fosang AJ, Baker NL, Werther GA 1997 Insulin-like growth factor binding protein-2 binds to cell surface proteoglycans in the rat brain olfactory bulb. Endocrinology 138:4858 – 4867

IGFBP-2 PROTEOLYSIS IN SK-N-MC CELLS 14. Werther GA, Hogg A, McKinley M, Oldfield B, Figdor R, Mendelsohn FAO 1989 Localization and characterization of insulin-like growth factor-I receptor in rat brain and pituitary, using in vitro autoradiography and computerised densitometry. J Neuroendocrinol 370 –377 15. Werther GA, Abate M, Hogg A, Cheesman H, Oldfield B, Hards D, Hudson P, Power B, Freed K, Herington AC 1990 Localization of insulin-like growth factor-I mRNA in rat brain by in situ hybridization—relationship to IGF-I receptors. Mol Endocrinol 4:773–778 16. Recio-Pinto E, Ishii DN 1984 Effects of insulin, insulin-like growth factor-II and nerve growth factor on neurite outgrowth in cultured human neuroblastoma cells. Brain Res Mol Brain Res 302:323–334 17. Werther GA, Cheesman H, Russo V 1993 Olfactory bulb organ culture is supported by combined insulin-like growth factor-I and basic fibroblast growth factor. Brain Res Mol Brain Res 617:339 –342 18. Frodin M, Gammeltoft S 1994 Insulin-like growth factors act synergistically with basic fibroblast growth factor and nerve growth factor to promote chromaffin cell proliferation. Proc Natl Acad Sci USA 91:1771–1775 19. Drago J, Murphy M, Carroll SM, Harvey RP, Bartlett PF 1991 Fibroblast growth factor-mediated proliferation of central nervous system precursors depends on endogenous production of insulin-like growth factor I. Proc Natl Acad Sci USA 88:2199 –2203 20. Russo VC, Cheesman H, Werther GA 1994 Organ culture: an in vitro system for the sustained growth of neonatal olfactory bulb. Neurosci Protocols 3:1–11 21. Russo VC, Edmondson SR, Mercuri FA, Buchanan CR, Werther GA 1994 Identification, localization, and regulation of insulin-like growth factor binding proteins and their messenger ribonucleic acids in the newborn rat olfactory bulb. Endocrinology 135:1437–1446 22. Hernandez-Sanchez C, Werner H, Roberts Jr CT, Woo EJ, Hum DW, Rosenthal SM, LeRoith D 1997 Differential regulation of insulin-like growth factor-I (IGF-I) receptor gene expression by IGF-I and basic fibroblastic growth factor. J Biol Chem 272:4663– 4670 23. Ota A, Wilson GL, Spilberg O, Pruss R, LeRoith D 1988 Functional insulinlike growth factor I receptors are expressed by neural-derived continuous cell lines. Endocrinology 122:145–152 24. Hall LJ, Kajimoto Y, Bichell D, Kim SW, James PL, Counts D, Nixon LJ, Tobin G, Rotwein P 1992 Functional analysis of the rat insulin-like growth factor I gene and identification of an IGF-I gene promoter. DNA Cell Biol 11:301–313 25. Kiess W, Koepf G, Christiansen H, Blum WF 1997 Human neuroblastoma cells use either insulin-like growth factor- I or insulin-like growth factor-II in an autocrine pathway via the IGF-I receptor: variability of IGF, IGF binding protein (IGFBP) and IGF receptor gene expression and IGF and IGFBP secretion in human neuroblastoma cells in relation to cellular proliferation. Regul Pept 72:19 –29 26. Bach LA, Hsieh S, Sakano K, Fujiwara H, Perdue JF, Rechler MM 1993 Binding of mutants of human insulin-like growth factor II to insulin-like growth factor binding proteins 1– 6. J Biol Chem 268:9246 –9254 27. Anchan RM, Reh TA, Angello J, Balliet A, Walker M 1991 EGF and TGF-alpha stimulate retinal neuroepithelial cell proliferation. Neuron 6:923–936 28. Leventhal PS, Randolph AE, Vesbit TE, Schenone A, Windebank A, Feldman EL 1995 Insulin-like growth factor-II as a paracrine growth factor in human neuroblastoma cells. Exp Cell Res 221:179 –186 29. Ochoa A, Domenzain C, Clapp C, Martinez de la Escalera G 1997 Differential effects of basic fibroblast growth factor, epidermal growth factor, transforming growth factor-a and insulin-like growth factor-I on hypothalamic gonadotropin-releasing hormone neuronal cell line. J Neurosci Res 49:739 –749 30. Smits A, Funa K 1998 Platelet-derived growth factor (PDGF) in primary brain tumours of neuroglial origin. Histol Histopathol 13:511–520 31. Li S, Baserga R 1996 Epidermal growth factor and platelet-derived growth factor regulate the activity of insulin-like growth factor I gene promoter. Exp Gerontol 31:195–206 32. Melcangi RC, Maggi R, Martini L 1993 Testosterone and progesterone metabolism in the human neuroblastoma cell line SH-SY5Y. J Steroid Biochem Mol Biol 46:811– 818 33. Shughrue PJ, Dorsa DM 1994 Estrogen and androge differentially modulate the growth-associated protein GAP-43 (neuromodulin) messenger ribonucleic acid in postnatal rat. Endocrinology 134:1321–1328 34. Janet T, Ludecke G, Otten U, Unsicker K 1995 Heterogeneity of human neuroblastoma cell lines in their proliferative responses to basic FGF, NGF, and EGF: correlation with expression of growth factors and growth factor receptors. J Neurosci Res 40:707–715 35. Meghani MA, Martin DM, Singleton JR, Feldman EL 1993 Effects of serum and insulin-like growth factors on human neuroblastoma cell growth. Regul Pept 48:217–224 36. Hansen MB, Nielsen SE, Berg K 1989 Re-examination and further development of a precise and rapid dye method for measuring cell growth/cell kill. J Immunol Methods 119:203–210 37. Hossenlopp P, Seurin D, Segovia-Quinson B, Hardouin S, Binoux M 1986 Analysis of serum insulin-like growth factor binding proteins using Western

38. 39. 40.

41.

42. 43. 44.

45. 46.

47. 48. 49. 50.

51. 52.

53.

54. 55. 56. 57.

58. 59.

60. 61.

3089

blotting: use of the method for titration of the binding proteins and competitive binding studies. Anal Biochem 154:138 –143 Bach LA, Leeding KS, Leng SL 1997 Regulation of IGF-binding protein-6 by dexamethasone and IGFs in PC12 rat phaeochromocytoma cells. J Endocrinol 155:225–232 Martin DM, Feldman EL 1993 Regulation of insulin-like growth factor-IL expression and its role in autocrine growth of human neuroblastoma cells. J Cell Physiol 155:290 –300 Menouny M, Binoux M, Babajko S 1997 Role of insulin-like growth factor binding protein-2 and its limited proteolysis in neuroblastoma cell proliferation: modulation by transforming growth factor-b and retinoic acid. Endocrinology 138:683– 690 Cianfarani S, Germani D, Rossi P, Spagnoli A, Mercanti D 1996 Do insulinlike growth factor binding proteins (IGFBPs) modulate the IGF-I growth promoting and differentiating effects in human neuroblastoma cells? Eur J Endocrinol 135:716 –723 Singleton JR, Randolph AE, Feldman EL 1996 Insulin-like growth factor I receptor prevents apoptosis and enhances neuroblastoma tumorigenesis. Cancer Res 56:4522– 4529 Ho PJ, Baxter RC 1997 Characterization of truncated insulin-like growth factor-binding protein-2 in human milk. Endocrinology 138:3811–3818 Matsumoto T, Gargosky SE, Iwasaki K, Rosenfeld RG 1996 Identification and characterization of insulin-like growth factors (IGFs), IGF-binding proteins (IGFBPs), and IGFBP proteases in human synovial fluid. J Clin Endocrinol Metab 81:150 –155 Rajah R, Bhala A, Nunn SE, Peehl DM, Cohen P 1996 7S nerve growth factor is an insulin-like growth factor-binding protein protease. Endocrinology 137:2676 –2682 Cohick WS, Gockerman A, Clemmons DR 1995 Regulation of insulin-like growth factor (IGF) binding protein-2 synthesis and degradation by plateletderived growth factor and the IGFs is enhanced by serum deprivation in vascular smooth muscle cells. J Cell Physiol 164:187–196 Ishikawa K, Ohe Y, Tatemoto K 1995 Synthesis and secretion of insulin-like growth factor (IGF)-II and IGF binding protein-2 by cultivated brain meningeal cells. Brain Res Mol Brain Res 697:122–129 Rajah R, Nunn SE, Herrick DJ, Grunstein MM, Cohen P 1996 Leukotriene D4 induces MMP-1, which functions as an IGFBP protease in human airway smooth muscle cells. Am J Physiol 271:L1014 –L1022 BikfalviA 1995 Significance of angiogenesis in tumour progression and metastasis. Eur J Cancer 31A:1101–1104 Bigg HF, Cawston TE 1995 All-trans-retinoic acid interacts synergistically with basic fibroblast growth factor and epidermal growth factor to stimulate the production of tissue inhibitor of metalloproteinases from fibroblasts. Arch Biochem Biophys 319:74 – 83 Allen LE, Maher PA 1993 Expression of basic fibroblast growth factor and its receptor in an invasive bladder carcinoma cell line. J Cell Physiol 155:368 –375 Blei F, Wilson EL, Mignatti P, Rifkin DB 1993 Mechanism of action of angiostatic steroids: suppression of plasminogen activator activity via stimulation of plasminogen activator inhibitor synthesis. J Cell Physiol 155:568 –578 LalouC, Silve C, Rosato R, Segovia B, Binoux M 1994 Interactions between insulin-like growth factor-I (IGF-I) and the system of plasminogen activators and their inhibitors in the control of IGF-binding protein-3 production and proteolysis in human osteosarcoma cells. Endocrinology 135:2318 –2326 Bradshaw SL, Han VK 1993 Growth factors modulate the biosynthesis of rat astroglial insulin-like growth factor binding proteins. Ann NY Acad Sci 692:249 –252 Villaudy J, Delbe J, Blat C, Desauty G, Golde A, Harel L 1991 An IGF binding protein is an inhibitor of FGF stimulation. J Cell Physiol 149:492– 496 Loret C, Janet T, Labourdette G, Schneid H, Binoux M 1991 FGFs stimulate IGF binding protein synthesis without affecting IGF synthesis in rat astroblasts in primary culture. Glia 4:378 –383 Kummer JL, Zawada WM, Freed CR, Chernausek SD, Heidenreich KA 1996 Insulin-like growth factor binding proteins in fetal rat mesencephalic cultures: regulation by fibroblast growth factor and insulin-like growth factor I. Endocrinology 137:3551–3556 Gluckman PD, Guan J, Beilharz EJ, Klempt ND, Klempt M, Miller O, Sirimanne E, Dragunow M, Williams CE 1993 The role of the insulin-like growth factor system in neuronal rescue. Ann NY Acad Sci 692:138 –148 Beilharz EJ, Russo VC, Butler GE, Baker NL, Connor B, Sirimanne ES, Dragunow M, Werther GA, Gluckman PD, Williams CE, Sheepens A 1998 Coordinate and cellular specific induction of the components of the IGF/ IGFBP axis in rat brain following hypoxic-ischemic injury. Mol Brain Res 59:119 –134 Cohen P, Lamson G, Okajima T, Rosenfeld RG 1993 Transfection of the human IGFBP-3 gene into Balb/c fibroblasts: a model for the cellular functions of IGFBPs. Growth Regul 3:23–26 Liu L, Delbe J, Blat C, Zapf J, Harel L 1992 Insulin-like growth factor binding protein (IGFBP-3), an inhibitor of serum growth factors other than IGF-I and -II. J Cell Physiol 153:15–21

3090


62. Mohseni-Zadeh S, Binoux M 1997 Insulin-like growth factor (IGF) binding protein-3 interacts with the type 1 IGF receptor, reducing the affinity of the receptor for its ligand: an alternative mechanism in the regulation of IGF action. Endocrinology 138:5645–5648 63. Conover CA, Clarkson JT, Bale LK 1996 Factors regulating insulin-like growth factor-binding protein-3 binding, processing, and potentiation of insulin-like growth factor action. Endocrinology 137:2286 –2292 64. Tham A, Nordberg A, Grissom FE, Carlsson-Skwirut C,Viitanen M, Sara VR 1993 Insulin-like growth factors and insulin-like growth factor binding proteins in cerebrospinal fluid and serum of patients with dementia of the Alzheimer type. J Neural Transm Park Dis Dement Sect 5:165–176 65. Korc-Grodzicki B, Ren N, Hilf R 1996 Effects of estradiol on the expression and production of IGFBP-2 by R3230AC mammary tumor cells. Oncol Res 8:473– 483 66. Ho PJ, Baxter RC 1997 Insulin-like growth factor-binding protein-2 in patients

Endo • 1999 Vol 140 • No 7

with prostate carcinoma and benign prostatic hyperplasia. Clin Endocrinol (Oxf) 46:333–342 67. Nordqvist AC, Peyrard M, Pettersson H, Mathiesen T, Collins VP, Dumanski JP, Schalling M 1997 A high ratio of insulin-like growth factor II/insulin-like growth factor binding protein 2 messenger RNA as a marker for anaplasia in meningiomas. Cancer Res 57:2611–2614 68. Muller HL, Oh Y, Lehrnbecher T, Blum WF, Rosenfeld RG 1994 Insulin-like growth factor-binding protein-2 concentrations in cerebrospinal fluid and serum of children with malignant solid tumors or acute leukemia. J Clin Endocrinol Metab 79:428 – 434 69. Zumkeller W, Schwander J, Mitchell CD, Morrell DJ, Schofield PN, Preece MA 1993 Insulin-like growth factor (IGF)-I, -II and IGF binding protein-2 (IGFBP-2) in the plasma of children with Wilms’ tumour. Eur J Cancer 29A: 1973–1977