0021-972X/91/7302-0401 $03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1991 by The Endocrine Society
Vol. 73, No. 2 Printed in U.S.A.
Insulin-Like Growth Factors (IGFs), IGF Receptors, and IGF-Binding Proteins in Primary Cultures of Prostate Epithelial Cells* PINCHAS COHENt, DONNA M. PEEHL, GEORGE LAMSON, AND RON G. ROSENFELD Departments of Pediatrics and Urology (DM.P.), Stanford University Medical Center, Stanford, California 94305
3 documented the expression of mRNA for hIGFBP-2 only. Modifications of the serum-free conditions of PEC did not significantly alter the IGFBP profile of PEC CM. The ability of IGF-I, IGF-II, and insulin to stimulate clonal growth of PEC was examined. IGF-I stimulated PEC growth with an ED50 of 0.1 ng/mL. IGF-II and insulin, respectively, were 1 and 3 orders of magnitude less effective than IGF-I in stimulating the growth of PEC. Radioimmunoassayable IGF-I and IGF-II levels in PEC CM were below the assay detection levels. In conclusion, we suggest that IGFs are important growth stimulators of PEC in culture, that their actions are mediated through the type 1 IGF receptor, and that PEC produce hIGFBP-2 and a 24-kDa IGFBP which may modulate IGF action in these cells. (J Clin Endocrinol Metab 73: 401-407,1991)
ABSTRACT. Insulin-like growth factors (IGFs) are potent mitogens that bind with high affinity and specificity to IGF receptors and IGF-binding proteins (IGFBPs). We studied the roles of these three groups of proteins in prostate epithelial cells (PEC) in primary culture grown under serum-free conditions. Affinity cross-linking of IGF-I and IGF-II to crude membranes prepared from PEC revealed an abundance of type 1 IGF receptors and no evidence of type 2 IGF receptors. Western ligand blots of conditioned media (CM) from PEC demonstrated the presence of two specific IGFBP bands similar to those previously demonstrated in seminal plasma, with approximate mol wt of 31 and 24 kDa. The 31-kDa band was immunoprecipitable with an antibody to IGFBP-2, and neither band could be deglycosylated with endoglycosidase-F. Northern blot analysis of poly(A)+ RNA prepared from PEC with cDNAs for hIGFBP-1, -2, and -
I
NSULIN -like growth factor-I (IGF-I) and IGF-II belong to a family of peptide hormones that includes relaxin and insulin and share a high degree of structural similarity with proinsulin (1). These peptides interact with specific receptors, designated type 1 and type 2 IGF receptors, as well as with the insulin receptor (2). IGFs have direct effects on the proliferation of many tissues and cell types, both in vivo and in vitro, and are thought to be significant autocrine-paracrine factors involved in normal and malignant cellular proliferation (3). A recently recognized class of proteins with high affinity for the IGFs, the IGF-binding proteins (IGFBPs), has been suggested to have a role in the modulation of the proliferative and mitogenic effects of IGFs on cells (4, 5). The molecular mechanisms involved in the interaction of the IGFBPs with the IGFs and their receptors
remain unclear. The human (h) IGFBP family consists of at least five proteins, three of which have been cloned and sequenced: IGFBP-1, IGFBP-2, and IGFBP-3 (6). Two more BPs have been purified and partially analyzed: a 32-kDa protein from cerebrospinal fluid (CSF) (7) and a 24-kDa protein from bone extract (8). This low mol wt IGFBP has also been described in serum (4,5,9), seminal plasma (10), and conditioned media from many cell lines (4, 5,9). In certain systems, IGFBPs are tightly regulated by various endocrine factors (4) and are uniquely expressed ontogenetically (11). Purified or recombinant IGFBPs have been shown to have seemingly conflicting effects when added to various cellular systems. Inhibition of IGF action by excess IGFBPs (12) as well as enhancement of the mitogenic actions of IGF (13) have been reported by different investigators. The presence and function of IGFs and IGFBPs in the female reproductive system have been extensively studied (4, 14). In the male reproductive system, IGFs and their receptors have been shown to be present in the murine testis (15), and we have documented the presence of IGF, as well as hIGFBP-2 and a low mol wt (24-kDa) IGFBP in seminal plasma (10). The source of seminal
Received September 17,1990. Address all correspondence and requests for reprints to: Pinchas Cohen, M.D., Division of Pediatric Endocrinology, Department of Pediatrics, Room S-322, Stanford University Medical Center, Stanford, California 94305. * This work was supported in part by Grant DK-28229 from the NIH (to R.G.R.). t Juvenile Diabetes Foundation Fellow.
401
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2014. at 21:42 For personal use only. No other uses without permission. . All rights reserved.
COHEN ET AL.
402
plasma IGFs and IGFBPs has yet to be established. The prostate is composed of mesenchymal and epithelial cells. The latter are prone to proliferative diseases and malignancy (16). The study of the role of IGFs in the prostate has been limited so far to animal tissues and human malignancies. A serum-free primary culture technique for human prostate tissue obtained during surgery has recently been developed (17), thus allowing us to employ normal prostatic epithelial cells (PEC) in culture to study the role of the IGFs, their receptors, and their BPs in mediating prostatic growth.
Materials and Methods Materials Recombinant [Thr59] hIGF-I was purchased from Amgen Biologicals (Thousand Oaks, CA). Synthetic hIGF-II for biological experiments was purchased from Bachem (Torrance, CA). Synthetic hIGF-II for iodination was generously provided by M. Smith (Eli Lilly Co., Indianapolis, IN). Peptides were iodinated by a modification of the chloramine-T method to a specific activity of 150-300 ^Ci/Mg- Rabbit polyclonal aHECl antibody, raised against multiplication-stimulating activity-affinity-purified medium from a human endometrial adenocarcinoma cell line, HEC1A, recognizes hIGFBP-2 and hIGFBP-3, as previously described (18). Endoglycosidase-F (Endo-F) was purchased from Calbiochem (La Jolla, CA). Bovine pituitary extract was obtained from Hammond/Cell Tech (Alameda, CA). Cell cultures and clonal growth assays Wedges of prostate tissue with normal anatomical and histological appearance were dissected from the central and peripheral zones during radical prostatectomies and processed, as previously described, to select against the growth of fibroblasts (17). The epithelial nature of the selected cells was verified by immunocytochemical staining for keratin, prostatic acid phosphatase, and prostate-specific antigen. Primary cultures of epithelial cells were maintained in basal medium PFMR-4A (prepared in our laboratory from materials purchased from Sigma, St. Louis, MO), supplemented with phosphoethanolamine (0.1 mM), cholera toxin (10 ng/mL), selenium (3 X 10~8 M), epidermal growth factor (10 ng/mL), hydrocortisone (1 ng/ mL), bovine pituitary extract (10 /ig/mL), gentamicin (100 ng/ mL), a-tocopherol (2.3 X 10~6 M), retinoic acid (3 X 10~n M), and, unless otherwise specified, insulin (4 ng/mL). Cells were grown on dishes coated with type I collagen (Collagen Corp., Palo Alto, CA). Procedures previously described for clonal growth assay of PEC were followed (19). Briefly, trypsin-EDTA was used to prepare single cell suspensions from secondary cultures after eight population doubling. Cells were then plated (200 cells/60-mL dish) in the experimental medium. After an incubation period of 10 days, cells were fixed with crystal violet. Quantification of cell number was performed with an Artek image analyzer (Chantilly, VA). Previous work (19) has shown correlation between cell count and Artek quantification. Six
JCE & M • 1991 Vol73-No2
strains of prostate epithelial cells were studied (three each of peripheral and central zone origin). Data were compared with unpaired t tests, and statistical significance was confirmed with analysis of variance. Affinity cross-linking Cross-linking of crude membrane preparations was performed as previously described (20). Cells were harvested, resuspended in homogenization buffer, and sonicated for 90 s. The lysate was centrifuged at 4,000 x g for 30 min at 4 C, and the resulting supernatant was centrifuged at 40,000 x g for 1 h at 4 C. The pellet was resuspended in HEPES buffer with 1 mM phenylmethylsulfonylfluoride and frozen at -70 C. The crude membranes (100 fig membrane protein) were incubated with 150,000 cpm radioligand, with or without excess unlabeled peptide, and then cross-linked with 8 mg/mL disuccinimidyl suberate (Pierce, Rockford, IL). Membranes were heated to 95 C for 5 min before electrophoresis on a 6% sodium dodecyl sulfate (SDS)-polyacrylamide gel in the presence of 100 mM dithiothreitol at 60 V overnight. Gels were then stained, destained, dried, and autoradiographed. Western ligand blots (WLB) Serum samples (2 nD or conditioned medium (50 (iL) were electrophoresed on 10% SDS-polyacrylamide gels, electroblotted onto nitrocellulose, incubated with [125I]IGF-II, and exposed to film for 5-10 days, as previously described (21). IGFBPs were immunoprecipitated from serum or conditioned medium samples with protein-A-purified antibodies or normal rabbit or mouse serum, solubilized, and then electrophoresed, as previously described (18). For deglycosylation studies, samples were preincubated with 0.25 U Endo-F for 3.5 h at 37 C, as previously described (18). RNA analysis Poly(A)+ RNA was isolated by proteinase-K (Merck, Darmstadt, Germany) treatment of freshly lysed cells, followed by affinity chromatography on oligo(dT)-cellulose (Pharmacia, Piscataway, NJ), as previously described (22). Five-microgram RNA samples were size-fractionated on a 1.2% agarose formaldehyde gel, blotted onto nitrocellulose, and hybridized, as previously described (18). The cDNA probes for hIGFBP-1, -2, and -3 were the kind gifts of David Powell (Houston, TX) (23), Jurg Schwander (Basel, Switzerland) (24), and William Wood (Genentech, South San Francisco, CA) (25), respectively. A chicken /?-actin probe was used to assess relative amounts of RNA in the Northern blots. cDNA probes were labeled with 32 P by random oligo priming (Pharmacia, Uppsala, Sweden). Peptide assays Concentrations of IGF-I and IGF-II in the conditioned medium of cells after 48-h incubation in serum-free medium were measured by RIA. Samples were chromatographed on a Sephadex G-50 column with 0.25 M formic acid to separate the peptides from binding proteins. The RIA for IGF-I uses a rabbit antiserum (derived by Drs. Underwood and Van Wyk, Chapel
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2014. at 21:42 For personal use only. No other uses without permission. . All rights reserved.
403
IGF, RECEPTORS, AND BP IN PROSTATE Hill, NC) distributed by the National Hormone and Pituitary Program. The RIA for IGF-II uses a monoclonal antibody to rat IGF-II (Amano, Troy, VA). Possible contamination of the chromatographed IGF peptide fraction with IGFBPs was assessed by WLB of the pooled peptide fraction from the G-50 column.
Results IGF receptors
IGF receptors in PEC were assessed by affinity crosslinking under reducing conditions. Figure 1 shows an autoradiograph of crude PEC membrane preparations (100 /xg membrane protein) crossed-linked to [125I]IGF-I or [125I]IGF-II (150,000 cpm) in the presence or absence of 500 ng/mL unlabeled peptide. Lane 1 demonstrates binding of [125I]IGF-I to prostate peripheral zone cell membranes. Note intense bands at approximately 135 kDa, representing the a-subunit of the type 1 IGF receptor, and at 270 kDa, representing an a-a dimer of the same receptor, an additional higher mol wt band was also present and probably represents aggregates of the type 1 receptor. Binding was completely inhibited by excess unlabeled IGF-I (lane 2). There was no evidence of Bo NSB Bo NSB
Mr x10' 3 211 -
significant [125I] IGF-II binding to crude membranes from these same cells, except for very faint bands mirroring the [125I] IGF-I binding and probably representing IGFII binding to the type 1 receptor. This faint band, however, could also represent a reduced type 2 receptor at low concentrations (lane 3). IGF-II did bind to the type 2 IGF receptor in crude membranes prepared from rat placenta run on a distal lane in the same gel (data not shown). While no IGF-II binding to the type 2 receptor was demonstrable in this experiment, the possibility, however, of low concentrations of type 2 IGF receptors cannot be completely excluded by such studies. Identification and characterization of IGFBPs in prostatic cell-conditioned medium Secretion of IGFBPs into medium conditioned by PEC for 48 h was assessed by the WLB technique. Figure 2A shows a WLB with [125I] IGF-II of conditioned medium from primary cultures of PEC from the peripheral zone (lane 1) and the central zone (lane 2). Bands are observed at apparent mol wt of 31 and 24 kDa. An identical pattern was observed in four other strains with similar histologies (data not shown). For comparison, Fig. 2B demonstrates IGFBPs from other sources. Human serum (lane 1) has two major bands at 40 and 44 kDa (representing two glycosylation forms of hIGFBP-3) and minor bands at 31, 27, and 24 kDa. HEP-G2-conditioned medium (lane 2) contains an intense band at 27 kDa, which has been previously demonstrated to be IGFBP-1 (4). Normal human seminal plasma (lanes 3-5) contains a band at 31 kDa, which has been previously described as being A
B
1 2
1
2
3
4
5
1
2
3
4 »
Mr x10"J
Mrx1(T 3
r
•
44-
-27
27-
10718
18-
%
1 2 125 I-IGF-I
3 4 125 MGF-II
FIG. 1. IGF receptors in prostate cells. Autoradiograph of a 6% SDSpolyacrylamide gel of crude PEC membrane preparations affinity crosslinked to [125I]IGF-I (lanes 1 and 2) or [125I]IGF-II (lanes 3 and 4) in the absence (Bo) or presence (nonspecific binding; NSB) of 500 ng/mL of the corresponding unlabeled peptide and run under reducing conditions. Mr, Mol wt.
0
FIG. 2. IGFBPs in prostate cell-conditioned medium. Autoradiograph of a WLB with [125I]IGF-H in conditioned medium from primary cultures of PEC from the peripheral zone (PZ; lane Al) and the central zone (CZ; lane A2) compared with human serum (lane Bl), HEP-G2conditioned medium (lane B2), and seminal plasma (SP; lanes B3-5). Shown is WLB of conditioned medium from prostate peripheral zone cells (lane Cl), after Endo-F deglycosylation (lane C2), and after immunoprecipitation with aHECl antibody (lane C3) or nonimmune (NI) rabbit serum (lane C4). Mr, Mol wt.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2014. at 21:42 For personal use only. No other uses without permission. . All rights reserved.
COHEN ET AL.
404
hIGFBP-2 (10), and a band at 24 kDa; a minor band at 27 kDa is also seen in lane 3. Of note is that the IGFBP pattern in seminal plasma and that in prostate cellconditioned medium are essentially identical. Further characterization of the prostate IGFBPs was performed by deglycosylation and immunoprecipitation. Figure 2C shows a WLB with [125I]IGF-II of conditioned medium from prostate peripheral zone epithelial cells (lane 1). The same medium shows an identical pattern of [125I] IGF-II binding after deglycosylation with Endo-F, with bands observed at 31 and 24 kDa (lane 2). The prominent band at 31 kDa (but not the band at 24 kDa) was immunoprecipitated by aHECl antibody (lane 3), indicating that this nonglycosylated 31-kDa band in prostate-conditioned medium is hIGFBP-2. This band was not precipitated by normal rabbit serum (lane 4). Of note is that the hIGFBP-2 band sometimes appears as a doublet, possibly representing a slight shift in the electrophoretic mobility of hIGFBP-2 due to unknown causes. Both bands forming the doublet are immunoprecipitable with the aHECl antibody. Expression of IGFBP mRNA Figure 3 shows a Northern blot of poly(A)+ RNA isolated from primary cultures of PEC (5 /ig/lane) hybridized with the cDNA for hIGFBP-2. Lanes 1 and 2 are from two different strains of cells originating from
JCE&M-1991 Vol 73 • No 2
the central zone of the prostate, and lanes 3 and 4 are from two strains from the peripheral zone. While this gel demonstrates the expression of the 1.8-kilobase mRNA for hIGFBP-2, probing with cDNAs for hIGFBP-1 and 3 did not demonstrate expression of mRNA for these proteins (data not shown). Regulation of IGFBPs The effects of modifications of the hormonal milieu of PEC in primary culture on IGFBP production was examined by WLB with [125I]IGF-II. Prostate cell culture serum-free medium was conditioned under the specified modified conditions for 48 h at near confluency. Figure 4A shows the effects of eliminating growth supplements required for serum-free growth from the complete medium. Conditioned basal PFMR-4A medium lacking all supplements (lane 2) and complete serum-free conditioned medium containing all of the supplements required for serum-free growth listed in Materials and Methods (lane 3) have an identical IGFBP pattern, although hIGFBP-2 is slightly enhanced in the complete medium sample (lane 3). Eliminating cholera toxin, hydrocortisone, epidermal growth factor, and bovine pituitary extract (lanes 4-7) from the complete medium resulted in no change in the IGFBP pattern. For comparison, pooled human serum is shown in lane 1. No significant changes in the binding of [125I]IGF-II to conditioned medium samples were observed in spite of pre-
1 7.5-
2
3
4
5
6
7
1
2
3
4
5
MrxKT3
6
7
Mr »1
Vol. 73, No. 2 Printed in U.S.A.
Insulin-Like Growth Factors (IGFs), IGF Receptors, and IGF-Binding Proteins in Primary Cultures of Prostate Epithelial Cells* PINCHAS COHENt, DONNA M. PEEHL, GEORGE LAMSON, AND RON G. ROSENFELD Departments of Pediatrics and Urology (DM.P.), Stanford University Medical Center, Stanford, California 94305
3 documented the expression of mRNA for hIGFBP-2 only. Modifications of the serum-free conditions of PEC did not significantly alter the IGFBP profile of PEC CM. The ability of IGF-I, IGF-II, and insulin to stimulate clonal growth of PEC was examined. IGF-I stimulated PEC growth with an ED50 of 0.1 ng/mL. IGF-II and insulin, respectively, were 1 and 3 orders of magnitude less effective than IGF-I in stimulating the growth of PEC. Radioimmunoassayable IGF-I and IGF-II levels in PEC CM were below the assay detection levels. In conclusion, we suggest that IGFs are important growth stimulators of PEC in culture, that their actions are mediated through the type 1 IGF receptor, and that PEC produce hIGFBP-2 and a 24-kDa IGFBP which may modulate IGF action in these cells. (J Clin Endocrinol Metab 73: 401-407,1991)
ABSTRACT. Insulin-like growth factors (IGFs) are potent mitogens that bind with high affinity and specificity to IGF receptors and IGF-binding proteins (IGFBPs). We studied the roles of these three groups of proteins in prostate epithelial cells (PEC) in primary culture grown under serum-free conditions. Affinity cross-linking of IGF-I and IGF-II to crude membranes prepared from PEC revealed an abundance of type 1 IGF receptors and no evidence of type 2 IGF receptors. Western ligand blots of conditioned media (CM) from PEC demonstrated the presence of two specific IGFBP bands similar to those previously demonstrated in seminal plasma, with approximate mol wt of 31 and 24 kDa. The 31-kDa band was immunoprecipitable with an antibody to IGFBP-2, and neither band could be deglycosylated with endoglycosidase-F. Northern blot analysis of poly(A)+ RNA prepared from PEC with cDNAs for hIGFBP-1, -2, and -
I
NSULIN -like growth factor-I (IGF-I) and IGF-II belong to a family of peptide hormones that includes relaxin and insulin and share a high degree of structural similarity with proinsulin (1). These peptides interact with specific receptors, designated type 1 and type 2 IGF receptors, as well as with the insulin receptor (2). IGFs have direct effects on the proliferation of many tissues and cell types, both in vivo and in vitro, and are thought to be significant autocrine-paracrine factors involved in normal and malignant cellular proliferation (3). A recently recognized class of proteins with high affinity for the IGFs, the IGF-binding proteins (IGFBPs), has been suggested to have a role in the modulation of the proliferative and mitogenic effects of IGFs on cells (4, 5). The molecular mechanisms involved in the interaction of the IGFBPs with the IGFs and their receptors
remain unclear. The human (h) IGFBP family consists of at least five proteins, three of which have been cloned and sequenced: IGFBP-1, IGFBP-2, and IGFBP-3 (6). Two more BPs have been purified and partially analyzed: a 32-kDa protein from cerebrospinal fluid (CSF) (7) and a 24-kDa protein from bone extract (8). This low mol wt IGFBP has also been described in serum (4,5,9), seminal plasma (10), and conditioned media from many cell lines (4, 5,9). In certain systems, IGFBPs are tightly regulated by various endocrine factors (4) and are uniquely expressed ontogenetically (11). Purified or recombinant IGFBPs have been shown to have seemingly conflicting effects when added to various cellular systems. Inhibition of IGF action by excess IGFBPs (12) as well as enhancement of the mitogenic actions of IGF (13) have been reported by different investigators. The presence and function of IGFs and IGFBPs in the female reproductive system have been extensively studied (4, 14). In the male reproductive system, IGFs and their receptors have been shown to be present in the murine testis (15), and we have documented the presence of IGF, as well as hIGFBP-2 and a low mol wt (24-kDa) IGFBP in seminal plasma (10). The source of seminal
Received September 17,1990. Address all correspondence and requests for reprints to: Pinchas Cohen, M.D., Division of Pediatric Endocrinology, Department of Pediatrics, Room S-322, Stanford University Medical Center, Stanford, California 94305. * This work was supported in part by Grant DK-28229 from the NIH (to R.G.R.). t Juvenile Diabetes Foundation Fellow.
401
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2014. at 21:42 For personal use only. No other uses without permission. . All rights reserved.
COHEN ET AL.
402
plasma IGFs and IGFBPs has yet to be established. The prostate is composed of mesenchymal and epithelial cells. The latter are prone to proliferative diseases and malignancy (16). The study of the role of IGFs in the prostate has been limited so far to animal tissues and human malignancies. A serum-free primary culture technique for human prostate tissue obtained during surgery has recently been developed (17), thus allowing us to employ normal prostatic epithelial cells (PEC) in culture to study the role of the IGFs, their receptors, and their BPs in mediating prostatic growth.
Materials and Methods Materials Recombinant [Thr59] hIGF-I was purchased from Amgen Biologicals (Thousand Oaks, CA). Synthetic hIGF-II for biological experiments was purchased from Bachem (Torrance, CA). Synthetic hIGF-II for iodination was generously provided by M. Smith (Eli Lilly Co., Indianapolis, IN). Peptides were iodinated by a modification of the chloramine-T method to a specific activity of 150-300 ^Ci/Mg- Rabbit polyclonal aHECl antibody, raised against multiplication-stimulating activity-affinity-purified medium from a human endometrial adenocarcinoma cell line, HEC1A, recognizes hIGFBP-2 and hIGFBP-3, as previously described (18). Endoglycosidase-F (Endo-F) was purchased from Calbiochem (La Jolla, CA). Bovine pituitary extract was obtained from Hammond/Cell Tech (Alameda, CA). Cell cultures and clonal growth assays Wedges of prostate tissue with normal anatomical and histological appearance were dissected from the central and peripheral zones during radical prostatectomies and processed, as previously described, to select against the growth of fibroblasts (17). The epithelial nature of the selected cells was verified by immunocytochemical staining for keratin, prostatic acid phosphatase, and prostate-specific antigen. Primary cultures of epithelial cells were maintained in basal medium PFMR-4A (prepared in our laboratory from materials purchased from Sigma, St. Louis, MO), supplemented with phosphoethanolamine (0.1 mM), cholera toxin (10 ng/mL), selenium (3 X 10~8 M), epidermal growth factor (10 ng/mL), hydrocortisone (1 ng/ mL), bovine pituitary extract (10 /ig/mL), gentamicin (100 ng/ mL), a-tocopherol (2.3 X 10~6 M), retinoic acid (3 X 10~n M), and, unless otherwise specified, insulin (4 ng/mL). Cells were grown on dishes coated with type I collagen (Collagen Corp., Palo Alto, CA). Procedures previously described for clonal growth assay of PEC were followed (19). Briefly, trypsin-EDTA was used to prepare single cell suspensions from secondary cultures after eight population doubling. Cells were then plated (200 cells/60-mL dish) in the experimental medium. After an incubation period of 10 days, cells were fixed with crystal violet. Quantification of cell number was performed with an Artek image analyzer (Chantilly, VA). Previous work (19) has shown correlation between cell count and Artek quantification. Six
JCE & M • 1991 Vol73-No2
strains of prostate epithelial cells were studied (three each of peripheral and central zone origin). Data were compared with unpaired t tests, and statistical significance was confirmed with analysis of variance. Affinity cross-linking Cross-linking of crude membrane preparations was performed as previously described (20). Cells were harvested, resuspended in homogenization buffer, and sonicated for 90 s. The lysate was centrifuged at 4,000 x g for 30 min at 4 C, and the resulting supernatant was centrifuged at 40,000 x g for 1 h at 4 C. The pellet was resuspended in HEPES buffer with 1 mM phenylmethylsulfonylfluoride and frozen at -70 C. The crude membranes (100 fig membrane protein) were incubated with 150,000 cpm radioligand, with or without excess unlabeled peptide, and then cross-linked with 8 mg/mL disuccinimidyl suberate (Pierce, Rockford, IL). Membranes were heated to 95 C for 5 min before electrophoresis on a 6% sodium dodecyl sulfate (SDS)-polyacrylamide gel in the presence of 100 mM dithiothreitol at 60 V overnight. Gels were then stained, destained, dried, and autoradiographed. Western ligand blots (WLB) Serum samples (2 nD or conditioned medium (50 (iL) were electrophoresed on 10% SDS-polyacrylamide gels, electroblotted onto nitrocellulose, incubated with [125I]IGF-II, and exposed to film for 5-10 days, as previously described (21). IGFBPs were immunoprecipitated from serum or conditioned medium samples with protein-A-purified antibodies or normal rabbit or mouse serum, solubilized, and then electrophoresed, as previously described (18). For deglycosylation studies, samples were preincubated with 0.25 U Endo-F for 3.5 h at 37 C, as previously described (18). RNA analysis Poly(A)+ RNA was isolated by proteinase-K (Merck, Darmstadt, Germany) treatment of freshly lysed cells, followed by affinity chromatography on oligo(dT)-cellulose (Pharmacia, Piscataway, NJ), as previously described (22). Five-microgram RNA samples were size-fractionated on a 1.2% agarose formaldehyde gel, blotted onto nitrocellulose, and hybridized, as previously described (18). The cDNA probes for hIGFBP-1, -2, and -3 were the kind gifts of David Powell (Houston, TX) (23), Jurg Schwander (Basel, Switzerland) (24), and William Wood (Genentech, South San Francisco, CA) (25), respectively. A chicken /?-actin probe was used to assess relative amounts of RNA in the Northern blots. cDNA probes were labeled with 32 P by random oligo priming (Pharmacia, Uppsala, Sweden). Peptide assays Concentrations of IGF-I and IGF-II in the conditioned medium of cells after 48-h incubation in serum-free medium were measured by RIA. Samples were chromatographed on a Sephadex G-50 column with 0.25 M formic acid to separate the peptides from binding proteins. The RIA for IGF-I uses a rabbit antiserum (derived by Drs. Underwood and Van Wyk, Chapel
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2014. at 21:42 For personal use only. No other uses without permission. . All rights reserved.
403
IGF, RECEPTORS, AND BP IN PROSTATE Hill, NC) distributed by the National Hormone and Pituitary Program. The RIA for IGF-II uses a monoclonal antibody to rat IGF-II (Amano, Troy, VA). Possible contamination of the chromatographed IGF peptide fraction with IGFBPs was assessed by WLB of the pooled peptide fraction from the G-50 column.
Results IGF receptors
IGF receptors in PEC were assessed by affinity crosslinking under reducing conditions. Figure 1 shows an autoradiograph of crude PEC membrane preparations (100 /xg membrane protein) crossed-linked to [125I]IGF-I or [125I]IGF-II (150,000 cpm) in the presence or absence of 500 ng/mL unlabeled peptide. Lane 1 demonstrates binding of [125I]IGF-I to prostate peripheral zone cell membranes. Note intense bands at approximately 135 kDa, representing the a-subunit of the type 1 IGF receptor, and at 270 kDa, representing an a-a dimer of the same receptor, an additional higher mol wt band was also present and probably represents aggregates of the type 1 receptor. Binding was completely inhibited by excess unlabeled IGF-I (lane 2). There was no evidence of Bo NSB Bo NSB
Mr x10' 3 211 -
significant [125I] IGF-II binding to crude membranes from these same cells, except for very faint bands mirroring the [125I] IGF-I binding and probably representing IGFII binding to the type 1 receptor. This faint band, however, could also represent a reduced type 2 receptor at low concentrations (lane 3). IGF-II did bind to the type 2 IGF receptor in crude membranes prepared from rat placenta run on a distal lane in the same gel (data not shown). While no IGF-II binding to the type 2 receptor was demonstrable in this experiment, the possibility, however, of low concentrations of type 2 IGF receptors cannot be completely excluded by such studies. Identification and characterization of IGFBPs in prostatic cell-conditioned medium Secretion of IGFBPs into medium conditioned by PEC for 48 h was assessed by the WLB technique. Figure 2A shows a WLB with [125I] IGF-II of conditioned medium from primary cultures of PEC from the peripheral zone (lane 1) and the central zone (lane 2). Bands are observed at apparent mol wt of 31 and 24 kDa. An identical pattern was observed in four other strains with similar histologies (data not shown). For comparison, Fig. 2B demonstrates IGFBPs from other sources. Human serum (lane 1) has two major bands at 40 and 44 kDa (representing two glycosylation forms of hIGFBP-3) and minor bands at 31, 27, and 24 kDa. HEP-G2-conditioned medium (lane 2) contains an intense band at 27 kDa, which has been previously demonstrated to be IGFBP-1 (4). Normal human seminal plasma (lanes 3-5) contains a band at 31 kDa, which has been previously described as being A
B
1 2
1
2
3
4
5
1
2
3
4 »
Mr x10"J
Mrx1(T 3
r
•
44-
-27
27-
10718
18-
%
1 2 125 I-IGF-I
3 4 125 MGF-II
FIG. 1. IGF receptors in prostate cells. Autoradiograph of a 6% SDSpolyacrylamide gel of crude PEC membrane preparations affinity crosslinked to [125I]IGF-I (lanes 1 and 2) or [125I]IGF-II (lanes 3 and 4) in the absence (Bo) or presence (nonspecific binding; NSB) of 500 ng/mL of the corresponding unlabeled peptide and run under reducing conditions. Mr, Mol wt.
0
FIG. 2. IGFBPs in prostate cell-conditioned medium. Autoradiograph of a WLB with [125I]IGF-H in conditioned medium from primary cultures of PEC from the peripheral zone (PZ; lane Al) and the central zone (CZ; lane A2) compared with human serum (lane Bl), HEP-G2conditioned medium (lane B2), and seminal plasma (SP; lanes B3-5). Shown is WLB of conditioned medium from prostate peripheral zone cells (lane Cl), after Endo-F deglycosylation (lane C2), and after immunoprecipitation with aHECl antibody (lane C3) or nonimmune (NI) rabbit serum (lane C4). Mr, Mol wt.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2014. at 21:42 For personal use only. No other uses without permission. . All rights reserved.
COHEN ET AL.
404
hIGFBP-2 (10), and a band at 24 kDa; a minor band at 27 kDa is also seen in lane 3. Of note is that the IGFBP pattern in seminal plasma and that in prostate cellconditioned medium are essentially identical. Further characterization of the prostate IGFBPs was performed by deglycosylation and immunoprecipitation. Figure 2C shows a WLB with [125I]IGF-II of conditioned medium from prostate peripheral zone epithelial cells (lane 1). The same medium shows an identical pattern of [125I] IGF-II binding after deglycosylation with Endo-F, with bands observed at 31 and 24 kDa (lane 2). The prominent band at 31 kDa (but not the band at 24 kDa) was immunoprecipitated by aHECl antibody (lane 3), indicating that this nonglycosylated 31-kDa band in prostate-conditioned medium is hIGFBP-2. This band was not precipitated by normal rabbit serum (lane 4). Of note is that the hIGFBP-2 band sometimes appears as a doublet, possibly representing a slight shift in the electrophoretic mobility of hIGFBP-2 due to unknown causes. Both bands forming the doublet are immunoprecipitable with the aHECl antibody. Expression of IGFBP mRNA Figure 3 shows a Northern blot of poly(A)+ RNA isolated from primary cultures of PEC (5 /ig/lane) hybridized with the cDNA for hIGFBP-2. Lanes 1 and 2 are from two different strains of cells originating from
JCE&M-1991 Vol 73 • No 2
the central zone of the prostate, and lanes 3 and 4 are from two strains from the peripheral zone. While this gel demonstrates the expression of the 1.8-kilobase mRNA for hIGFBP-2, probing with cDNAs for hIGFBP-1 and 3 did not demonstrate expression of mRNA for these proteins (data not shown). Regulation of IGFBPs The effects of modifications of the hormonal milieu of PEC in primary culture on IGFBP production was examined by WLB with [125I]IGF-II. Prostate cell culture serum-free medium was conditioned under the specified modified conditions for 48 h at near confluency. Figure 4A shows the effects of eliminating growth supplements required for serum-free growth from the complete medium. Conditioned basal PFMR-4A medium lacking all supplements (lane 2) and complete serum-free conditioned medium containing all of the supplements required for serum-free growth listed in Materials and Methods (lane 3) have an identical IGFBP pattern, although hIGFBP-2 is slightly enhanced in the complete medium sample (lane 3). Eliminating cholera toxin, hydrocortisone, epidermal growth factor, and bovine pituitary extract (lanes 4-7) from the complete medium resulted in no change in the IGFBP pattern. For comparison, pooled human serum is shown in lane 1. No significant changes in the binding of [125I]IGF-II to conditioned medium samples were observed in spite of pre-
1 7.5-
2
3
4
5
6
7
1
2
3
4
5
MrxKT3
6
7
Mr »1
