Journal o j Orfhopaedic Research 1063M46 Raven Press. Ltd., New York 0 1992 Orthopaedic Rcwarch Society

Interaction of Basic Fibroblast Growth Factor with Bovine Growth Plate Chondrocytes S. €3. Trippel, M. C. Whelan, *M. Klagsbrun, and *S. R. Doctrow Department of Orthopaedics, Massachusetts General Hospital and Hurvurd Medical School; and *Depurtment of Surgery, The Children's Hospital and Harvurd Medicul School, Boston, Massachusetts, U.S.A.

Summary: The basic fibroblast growth factor (bFGF) family of peptides influences a wide range of cellular actions. To better understand the possible role of bFGF in the growth plate, we have characterized the interaction of this growth factor with isolated bovine growth plate chondrocytes. Basic FGF interacts with two classes of binding sites on these cells. One is consistent with high-affinity bFGF receptors and the other with low-affinity heparin-like binding sites on the chondrocyte surface. Radiolabeled bFGF binding studies revealed approximately 4 x lo6 binding sites per cell, with a K , of approximately 42 nM. Graded concentrations of heparin or NaCl competed with ['2'I]-labeled bFGF in a dose-dependent Fashion, reducing ['2SI]-labeledbFGF binding by 75 and 97%, respectively. The data suggest the presence of a high-capacity, lowaffinity class of binding sites with the properties of a heparin-like moiety. Affinity cross-linking of ['2SI]-labeled bFGF to chondrocytes labeled two principal species with apparent molecular masses of 135 and 160 kDa. Labeled bFGF was specifically displaced from both species by subnanomolar concentrations of unlabeled bFGF. These high-affinity, low-capacity binding sites are characteristic of classical bFGF receptors. Binding of ['2'I]-labeled bFGF to these sites was also influenced by heparin, consistent with coregulation of binding to the two classes of binding sites. The data suggest that bFGF participates in the regulation of skeletal growth at the growth plate and that this regulation may involve bFGF interaction with at least two distinct classes of binding sites. Key Words: Basic fibroblast growth factor-Chondrocytes.

other polypeptide growth factors, cellular responses to bFGF appear to be initiated by interaction with target cell surface receptors. High-affinity bFGF receptors have been identified on cells of epidermal, neuroectodermal, and mesenchymal origin (18,27,32,35,38,39,50,58). These have been shown to be composed of 125-165 kDa glycosylated proteins (9,18,28,32,35,36,38), which appear to be similar but distinct gene products (6). On some cells, bFGF appears to also bind to low-affinity binding sites characterized as heparin-like (32) or heparan sulfate (3,45) moieties. These low-affinity sites appear to be necessary for bFGF to bind to its highaffinity receptors (59).

The basic fibroblast growth factor (bFGF) family of polypeptides regulates cell functions as diverse as mitogenesis (1 1,12), differentiation (23), protease induction (30,34), receptor modulation (5,41), and cell maintenance (1,31). Members of this family are mitogenic for cartilage (14,22,40,42,44,47,53) and present in cartilage tissue (53). As is the case for

Received August 13, 1991; accepted March 14. 1992. Address correspondence and reprint requests to Dr. S. B. Trippel at Massachusetts General Hospital and Harvard Medical School, Boston, MA 02115, U.S.A. Dr. S. R. Doctrow is currently at Alkermes. Inc., Cambridge, Massachusetts, U.S.A.

638

CHONDROCYTE BASIC FIBROBLAST GROWTH FACTOR RECEPTORS

Because the division and maturation of cartilage cells in the growth plate are responsible for skeletal growth and development, there is considerable interest in the mechanisms by which these chondrocyteq are regulated. Although receptors for several factors have been identified on these cells, including growth hormone, insulin-like growth factors (IGF) I and 11, parathyroid hormone, and bFGF (7,18,37,54,55), bFGF studies are limited and interaction of this peptide with growth plate chondrocytes remains to be characterked. Our studies were undertaken to elucidate this interaction. MATERIALS AND METHODS bFGF Human recombinant bFGF (19) was a generous gift of the Takeda Chemical Company (Osaka, Japan). Specific activity of the bFGF was 1-5 unitdng in the Balb/c/3T3 synthesis assay (57). Confluent monolayers of 3T3 cells were prepared in 96 well plates and after addition of bFGF the incorporation of [3H]thymidine into 3T3 DNA in a 36-48-h period was measured. One unit of activity was defined as the amount of growth factor needed to stimulate half-maximal synthesis of 3T3 DNA. Iodination of bFGF lodination of bFGF was by the method of Bolton and Hunter (4), modified a5 previouqly described (25). ['2s1]-labeled bFGF was purified by gel filtration on a Sephadex G-25 column, equilibrated with 50 mM Tris HC1, 0.3 M NaCI, 0.1% gelatin, and 1 mM dithiothreitol, pH 7.5. The biological activity of ['"I]-labeled bFGF was determined by its ability to bind to heparin-Sepharose and to stimulate DNA synthesis in Balbici3T3 cells and bovine capillary endothelial cells. Aliquots of ['251]-labeled bFGF were stored in the gel filtration buffer described earlier at - 20°C. The specific radioactivity of [1251]labeled bFGF was 71-197 nCi/ng at the time of use. Isolation of Growth Plate Chondrocytes Chondrocytes were isolated by a modification of previously described techniques (55). Long bones were obtained from newborn (- 1-week-old) calves at a nearby abattoire and the epiphyseal plates were dissected from metaphyseal and epiphyseal bone, taking care to excise perichondrium. The harvested

639

-

growth plate cartilage was diced into 2-mm cubes and incubated in Dulbecco's Modified Eagle Medium (DMEM) containing 0.1% collagenase (No. 0130, Sigma, St. Louis, MO, U.S.A.), 5% calf serum, 100 p/ml of penicillin G, and 100 kg/ml of streptomycin. Digestion was carried out in spinner bottles at 37°C overnight (12-16 h). Undigested matrix was removed by filtration over a 125-km pore size nylon mesh followed by centrifugation at 200 g for 6 min on a Ficoll-Hypaque (lO/lI%) gradient. The cells were recovered from the mediumgradient interface, washed with phosphate-buffered saline solution (PBS) (pH 7.4), and resuspended in calcium and magnesium-free DMEM with 2% calf serum, 100 p/ml of penicillin G, 100 pg/ml of streptomycin, and 0.1% Pluronic F68 (BASF Wyandotte, Parsippany, NJ, U.S.A.), a nontoxic polyol used to diminish cell agglutination (52). Cells were incubated in spinner bottles at 37°C for 30 h to permit recovery of receptors destroyed by enzymatic isolation (55). Binding of ['251]-Labeled bFGF to Growth Plate Chondrocytes Binding studies were performed as previously described (54). Cells were resuspended in binding medium consisting of DMEM with 1% bovine serum albumin and 20 mM HEPES (pH 7.4), and cooled to 4°C. Duplicate aliquots of cells were incubated with ["'I]-labeled bFGF and designated additives in a total volume of 0.4 ml. After the designated time periods, the contents were transferred to microfuge tubes. Bound ligand was separated from free ligand by centrifugation (2 min at 9,500 g) followed by a wash with ice-cold binding medium and repeat centrifugation (1.5 min at 9,500 g). The tip of the microfuge tube containing the cell pellet was cut off and counted in a gamma spectrometer. Nonspecific binding was defined as binding occurring in the presence of an excess (1 X lo-' - 1 x lo-' IM)of unlabeled ligand. Specific binding was defined as total binding minus nonspecific binding. Cross-Linking of ['251]-LabeledbFGF to Growth Plate Chondrocyte To estimate the relative molecular mass of ligandbinding site complexes, ['251]-labeledbFGF was covalently linked to its binding sites using the bifunctional cross-linking agent disuccinimidyl suberate (DSS) (Pierce Chemical C o . , Rockford, IL,

J Orthop Res, Vol. 10, No. 5 , 1992

S . B . TRIPPEL ET AL.

640

U.S.A.). Chondrocytes isolated as described earlier were resuspended in PBS (pH 7.4) and cooled to 4°C. Aliquots of 7 x lo6 cells were incubated in polystyrene tubes containing 0.5 or 0.75 X lo6 cpm of ['251]-labeledbFGF in the presence of designated additives in a total volume of 0.5 ml. After 24 h at 4"C, freshly solubilized DSS in dimethyl sulfoxide (DMSO) was added to the reaction mixture in a final concentration of 0.5 m M . After 30 min at 22"C, the reaction was quenched with an excess of NH,Br. Following the addition of 1 ml of cold PBS, the cells were centrifuged at 9,500 g for 15 min at 4"C, washed with I ml of cold PBS, and recentrifuged. Cell pellets were solubilized in SO p,l of 10 mM Tris HCl, 0.5% Nonidet P-40,0. I mM EDTA, and I mM phenylmethylsulfonyl fluoride, pH 7.0. Insoluble debris was removed by centrifugation at 9,500 g for 10 min. The supernatant was either immediately combined with concentrated sample buffer and subjected to SDS-polyacrylamide gel electrophoresis (PAGE), or stored at -80°C until analysis. SDS-PAGE and Autoradiography The SDS-PAGE (26) was performed on 10% acrylamide-resolving gels with 0.1% SDS. Gels were stained with Coomassie blue, destained, and dried. Autoradiograms were made on Kodak X-omat AR film in the presence, when needed, of an image-intensifying screen (Cronex "Lightning Plus," Dupont, Wilmington, DE, U.S.A.) for 2-14 days at - 70°C. Molecular weights were assigned by comparison to the log plot of standard marker proteins (8). Densitometric analysis of autoradiograms was with an Ultroscan densitometer (LKB, Piscataway, N J , U.S.A.).

1

0

' I

2

A 6

; h

7

;:,w?

a

HOURS

FIG. 1. Time course of specific ['251]-labeled basic fibroblast growth factor (bFGF) binding to isolated growth plate chondrocytes. Cells were incubated in the presence of [1251]labeled bFGF for the designated time period at 4°C (X line), 15°C (open circle line), and 37°C (closed circle line).

5.0-9.0. Binding was maximal from pH 6-7, with 94% maximal binding at pH 5 and 55% maximal binding at pH 9. Based on these data, subsequent binding studies were performed at 15°C for 12-1 4 h at pH 7.0. Binding of bFGF to growth plate chondrocytes was concentration dependent. Unlabeled bFGF competed with ['251]-labeledbFGF for binding with M. half-maximal displacement at 0.8-1.0 x Saturalion was approached at 3-4 x mol of bFGF bound per 7.6 x lo5 cells (Fig. 2 ) . By Scatchard analysis (481, the affinity constant (K;3 of bFGF binding in duplicate experiments was 2.4 X lo7 Wmol (Kd = 42 nM) and the number of binding sites per cell (R,) was 4 x 10' (Fig. 2). The Scatchard plots are consistent with a single class of 4-

RESULTS Interaction of bFGF with Growth Plate Chondrocytes Binding of [ '2S11-labeledbFGF to isolated bovine growth plate chondrocytes occurred rapidly at 4, 15, and 37°C. Equilibrium was reached by 30 min at 4 and 15"C, and little loss of cell-associated radioactivity was observed at these temperatures to 19 h, the maximum duration tested. At 37°C. maximal binding occurred at 15-30 min. followed by a gradual decrease in cell-associated radioactivity consistent with internalization and degradation (Fig. 1). Speciiic binding was measured over a pH range of

J Orthop Res, Vof. 10, No. 5 , 1992

bf GF ROUND imoles I 10'7

a&

~

L-

8

FREE bFGF

-1--

10

I

I

I2

14

46

18

fprnol/OlrnI/

FIG. 2. Concentration dependence of ['251]-labeled basic fibroblast growth factor (bFGF) binding to isolated growth plate chondrocytes. Cells were incubated in the presence of a constant quantity of ['251]-labeled bFGF and increasing concentrations of unlabeled bFGF. Specifically bound [lz5I]labeled bFGF is plotted against the concentration of free bFGF. Inset: Representative Scatchard plot of [1z51]-labeled bFGF binding data. The line represents a linear regression with a correlation coefficient of -0.84.

CHONDROCYTE BASIC FIBROBLAST GROWTH FACTOR RECEPTORS

010

.A4

28

.56

12

NOC/IM/ FIG. 3. Effect of NaCl on ['251]-labeled basic fibroblast growth factor (bFGF) binding to growth plate chondrocytes. Cells were incubated in the presence of ['251]-labeled bFGF and graded concentrations of NaCI. Binding is expressed as the ratio of bound ['251]-labeled bFGF (B) to ['2"1]-labeled bFGF in the presence of 0.10 M NaCl (Bo). Error bars represent 2 standard deviation.

low-affinity binding sites and do not provide evidence for a high-affinity bFGF receptor. Basic FGF specifically binds to heparin (10) and heparin-like molecules (3) and is displaced by high salt levels (53). It has previously been shown that specific bFGF binding to heparan sulfate (3) but not to its high-affinity receptor is sensitive to salt concentration (32,58). To determine whether bFGF binding to growth plate chondrocytes is salt concentration dependent, ['251]-labeledbFGF binding was performed in the presence of graded concentrations of NaCl. The NaCl displaced ["'I]-labeled bFGF in a concentration-dependent fashion. Ninety percent of ['2511-labeledbFGF bound at 0.1 M NaCl was displaced by 1.2 M NaCl (Fig. 3). At 2.0 and 3.0 M, NaCl reduced binding by 97 and 99%, respectively. To characterize bFGF binding to NaCl-independent sites, cells were exposed to [12sI]-labeled bFGF and graded concentrations of unlabeled bFGF in the presence of 1.7 M NaCl. Under these conditions, specific binding of ['251]-labeled bFGF was below the detection limits of the assay. Less than 1% of the added radiolabeled ligand was specifically bound, suggesting the presence of relatively few NaCI-insensitive binding sites on these cells. To further characterize this class of low-affinity binding sites, ['251]-labeled bFGF was bound to growth plate chondrocytes in the presence of graded concentrations of heparin. When coincubated with ['251]-labeled bFGF, heparin dccreased

641

growth factor binding in a dose-dependent fashion. Half-maximal inhibition by heparin occurred at 0.5 pglml. At 100 pg/ml, the maximal concentration tested, heparin reduced binding of ['"I]-labeled bFGF by 94% (data not shown). Although the 12-14-h duration of binding is presumed to be sufficient to establish equilibrium between bFGF, heparin, and cellular binding sites, the observed inhibition by heparin may reflect sequestration of bFGF by heparin, diminishing the availability to cellular binding sites of free bFGF. To determine whether heparin competes for bFGF already bound to growth plate chondrocytes, cells were exposed to ['251]-labeledbFGF for 14 h at 15°C and then to graded concentrations of heparin for 4 h at 15°C. Heparin displaced labeled bFGF in a dosedependent fashion with half-maximal displacement at 0.4 kgirnl and maximal displacement of 75% at 100 pglml, the highest dose tested (Fig. 4). To determine whether, after heparin displacement of bound ['2SI]-labeJed bFGF, there remains cell-associated salt-displaceable or bFGF-displaceable ligand, the cells were incubated with [lZ5Illabeled bFGF followed by heparin and then exposed to either 3 x lop7 M bFGF or 1.25 M NaCl (Fig. 4). The 1.25 M NaCl displaced approximately 35% of ['251]-labeledbFGF that remained cell associated in the presence of 100 pgiml of heparin. The 3 X M bFGF failed to produce further displacement (Fig. 4). These data wggest that heparin

.40

20

B

HEPARIN @g/mil

FIG. 4. Competition for ['251]-labeledbasic fibroblast growth factor (bFGF) binding to growth plate chondrocytes by heparin. The effect of graded concentrations of heparin on ['"I]labeled bFGF binding is expressed as a percentage of [1251]labeled bFGF bound in the absence of heparin (B,,). Binding of ['251]-labeledbFGF was unchanged by the addition of 3 x l o - ' M unlabeled bFGF (X) to 100 pg/ml of heparin, but was further decreased by the addition of 1.25 M NaCl to 100 pg/ml of heparin (open triangles). There was dissociation of bound [1251]-labeledbFGF in ['251]-labeledbFGF-free buffer at 15°C for 4 h (black triangles). Error bars represent 5 standard deviation.

J Orthop Res, Vol. 10, No. 5 , 1992

642

S.B . TRIPPEL ET AL.

facilitates dissociation of bFGF bound to cells and that an ['251]-labeledbFGF binding to putative heparin-insensitive high-affinity receptors is too low to detect by this binding method. These data indicate the presence on growth plate chondrocytes of a high-capacity , low-affinity system of binding sites consistent with a cellassociated heparin-like moiety. 'The data further suggest that the quantity of bFGF bound to any specific surface membrane receptors on these cells is sufficiently small, in comparison to that bound to this low-affinity class of binding sites, to be poorly accessible to accurate kinetic analysis by these binding methods.

kDa

200 -

92 69 -

CROSS-LINKING STUDIES

Affinity cross-linking techniques have previously been used to characterize insulin-like growth factor receptors on growth plate chondrocytes (54). In particular, it provides a method for detecting binding of bFGF to high-affinity receptor sites without interference by much more abundant low-affinity sites. To identify a receptor for bFGF on growth plate chondrocytes, ['251]-labeled bFGF was crosslinked to its cellular binding sites with DSS and analyzed by SDS-PACE and autoradiography. The predominant radiolabeled species appeared as a relatively broad band with an apparent M, of -135 kDa and an associated fainter band of -160 kDa (Fig. 5). Binding of ['"I]-labeled bFGF in the presence of graded concentrations of unlabeled bFGF generated progressive displacement of label from the 135 and 160 kDa bands but had little effect on other bands. On some separations, two separate bands of -MI 128 and 145 kDa replaced the 135kDa species without alteration in the 160-kDaband. These were similarly displaced by unlabeled bFCiF. Prominent labeling was also observed of a high molecular mass species (Mr > 300 kDa) from which the label was displaced by heparin (Fig. 6) and NaCl (data not shown), but not by unlabeled bFGF (Fig. 5). Graded concentrations of heparin progressively displaced label proportionately from all bands, including several less distinct bands with an apparent M, < 100 kDa (Fig. 6). Graded concentrations of unlabeled bFGF generated a distinctly different pattern, displacing label almost exclusively from the 135-160 kDa moieties (Fig. 5 ) . Although affinity cross-linking methods provide only approximate measures of competition potency, the observation that 10p'o M unlabeled

4

2

3

4

5

FIG. 5. Affinity labeling of ['251]-labeled fibroblast growth factor (bFGF) t o growth plate chondrocytes: effect of unlabeled bFGF. Labeling is in the presence of no unlabeled ligand (lane 1) or unlabeled bFGF 1 x lo-'' M (lane 2), 1 x 10 M (lane 3), 1 x lo-' M (lane 4),or 1 x M (lane 5).

bFGF displaced greater than 60% of [ '"11-labeled bFCF from the 135 and 160 kDa moieties (Fig. 7) suggests that these represent one or more highaffinity binding sites. Densitometric scanning of autoradiograms showing the effect of unlabeled bFGF (Fig. 7A) or heparin (Fig. 7B) on ['2SI]-labeled bFGF binding reveals a progressive decrease in intensity of both the predominant labeled complex at M, 135 kDa and the fainter complex at 160 kDa. Both complexes show an approximately parallel decrease in intensity with increasing concentrations of either bFGF or heparin. The ratio between the peak intensity of the 135 and 160 kDa moieties rose slightly with increasing concentrations of bFGF (1.9-2.7) and heparin (1.92.4). Although these data suggest that ['251]-labeled bFGP may be displaced somewhat more efficiently from the 160 than the 135 kDa complex, any such difference appears to be small. DISCUSSION

These data suggest that growth plate chondrocytes possess at least two classes of binding sites for bFGF, including specific high-affinity surface membrane receptors and heparin-like low-affinity binding sites. Competition binding studies demonstrate a class of low-affinity, high-capacity binding

643

CHONDROCYTE BASIC FIBROBLAST GROWTH FACTOR RECEPTORS

between heparin and bFGF may reflect heparin displacement of hFGF from the cells or heparin association with free bFGF in solution. These properties suggest that growth plate chondrocytes possess cell surface heparin-like binding sites similar to those reported on endothelial cells

kDa

200 -

A

i35,OOO

t

+

160,000

92 69

bFGF (M)

-

1

2

3

4

FIG. 6. Affinity labeling of ['"I]-labeled basic growth factor t o growth plate chondrocytes: effect of heparin. Labeling is in the presence of no heparin (lane 1) or 0.1(lane Z ) , 1 .O (lane 3),or 10 kg/ml (lane 4).

sites with a K , of -4 x 10 * M and an R , of approximately 4 x lo6 sitesicell. This compares to previously reported studies of bFGF binding to other cell lines with respective values of 2 x lo-' M and 6 x lo5 for endothelial cells (32) and 2 x lo-' and 1 X lo6 M for multiple other cell types (35). Similarly, the K , for endothelial cell-derived extracellular matrix has been reported as 6.1 x lo-' (3). The variability of R, but constancy of K , among chondrocyte preparations is similar to that described for subendothelial matrix (3). Our data may overapproximate the number of low-affinity sites and their K , since displacement of ['"I]-labcled bFGF by unlabeled bFGF was incomplete, and nonspecific binding remained high even at micromolar concentrations of unlabeled ligand. For these reasons and because such binding studies are subject to various intrinsic sources of error (15,21,43), these values represent estimates. Both heparin and NaCl displaced bFGF from these binding sites. Heparin displaced bFGF that was already bound to chondrocytes and also competed with labeled bFGF binding when coincubated with ['251]-labeledbFGF. The displacement by heparin of bound bFGF i q consistent with competition for the same cellular binding sites or with heparin binding to dissociated bFGF in equilibrium with bound bFGF. Similarly, the observcd competition

B

+

135,000

460,000

1

FIG. 7. Densitometric analysis of autoradiograms of labeled basic growth factor (bFGF) affinity labeled growth plate chondrocytes. Displacement of ['251]-labeled bFGF by unlabeled bFGF (A) or by heparin (B). Densitometer tracings are of the region between apparent M, 120,000and 180,000 kDa, with arrows designating M, 135,000and 160,000kDa. The concentration of competitor is indicated on the right. The area under each curve is expressed as a percentage of the area under the respective competition-free curve (B,,). A: bFGF 0, 100%; bFGF 1 x lo-'' M, 67%; bFGF 1 x lo-' M, 32%; bFGF 1 x 10 * M, 28%; bFGF 1 x M , 17%. B: Heparin 0,100%: heparin 0.1 pg/ml, 86%; heparin 1 .O pg/mI, 58%; heparin 10 pgirnl, 36%.

. I Orthop Kes. Vol. 10, N o . 5 , 1992

644

S.B . TRIPPEL ET AL.

(3,32) and in subendothelial extracellular matrix (3). Heparan sulfate, a predominantly cell-surface associated glycosaminoglycan (17), appears to be a major site of low-affinity bFGF binding in some cells (3,45). Heparan sulfate is also synthesized by chondrocytes (56) and, interestingly, its exogenous administration has been associated with chondrogenesis (46). The high M, ( > N O kDa) cross-linking species observed in our studies has the displacement pattern of a high-capacity low-affinity heparin and salt-dependent binding site. While this would be consistent with a heparan sulfate proteoglycan (27), its identity is not known. The low-affinity bFGF binding sites on growth plate chondrocytes, while not fully defined, share several characteristics with surface-associated hepardn sulfate. The time course of bFGF binding to growth plate chondrocytes is comparable to that of bFGF binding to bFGF-responsive cells and to extracellular matrix. Basic FGF binding to neuronal sites ha5 been detected within as little as 2 min (58) and to ECM in a similar time period (3). The observed gradual decrease in cell-associated ['251]-labeled bFGF with time at 37°C may represent internalization and degradation followed by release, or dissociation and degradation. A similar gradual loss of bound bFGF from matrix has been noted in capillary endothelial cells (33). Affinity cross-linking revealed bFGF binding to sites with an apparent M, of 135-160 kDa, complexed with bFGF. Displacement of labeled bFGF from these sites by subnanomolar concentrations of unlabeled bFGF suggest that this is a high-affinity class of binding sites, consistent with one or more classical bFGF receptors. Studies of certain other cell types reveal a receptor that is similar in its apparent M, and occurrence on SDS-PAGE as a doublet of 135/160 kDa (25,35,58). The doublet probably represents two distinct high-affinity receptor forms (6,20). The occurrence on some gels of three cross-linked species in the 130-160 kDa range may reflect partial degradation or receptor variants (16). A recently described bFGF receptor on rabbit chondrocytes appeared in the 150-160 kDa range, with binding at 100-120 kDa observed at high concentrations of [ 12511-labeledbFGF. No low-affinity binding sites were found on the rabbit chondrocytes, possibly due to differences in experimental technique (18). The apparent higher affinity of bFGF for the bFGF receptor than for heparin may explain the observed ability of bFGF to bind to this receptor in

J Orthop Res, Voi. 10, No. 5 , 1992

the presence of heparin concentrations as high as 10 pg/ml (approximately 4 x lop7 M>. Basic FGF possesses multiple binding domains (2,49) that are distinct for binding to its high-affinity receptor and for binding to heparin (25). Nevertheless, heparin substantially reduced receptor-bound bFGF, suggesting that it is capable of modulating bFGF availability to the receptor. Consistent with this interpretation is the recent demonstration that low-affinity heparin-like binding sites are required for bFGF binding to its high-affinity receptor (59). The data may alco reflect the ability of soluble heparin to decrease the local concentration of bFGF available to the receptor, as suggested by the ability of graded concentrations of heparin to inhibit endothelial cell DNA synthesis stimulated by low (0.1 ngiml) but not high (10 ng/ml) bFGF concentrations (S Doctrow, unpublished data). The failure to detect these high-affinity sites in kinetic studies most likely reflects their scarcity in cornparkon to low-affinity sites. The number of bFGF receptors per cell reported for other cells ranges from 700 to 82,000 (18,32,39). Although affinity cross-linking competition studies do not permit an accurate estimation of receptor concentration, if bovine growth plate chondrocytes possess a similar number of receptors, then the number of low-affinity surface binding sites per cell would exceed this by two to three orderc of magnitude. This difference would be sufficient to obscure detection of receptors by our competition binding methods. The functional significance of bFGF binding to heparin-like molecules on these cells is uncertain. These sites may serve as a storage depot for bFGF awaiting translocation to receptor sites, or conversely as a scavenger system, sequestering bFGF from extracellular or intracellular compartments. Binding of bFCF to low-affinity sites on bovine capillary endothelial cells does not appear to trigger its cellular effects (32), but the observations that heparin-like molecules may influence chondrocyte function (46) and may be required for high-affinity binding (59) suggest that bFGF and heparin-like molecules interact in cell regulation. Association of bFGF with heparin protects the polypeptide from inactivation (13,44) and it is plausible that the observed cell-associated low-affinity heparin-like sites confer a similar protective effect. The biological role of bFGF in the growth plate is not yet known. Unlike other polypeptides with receptors on growth plate chondrocytes, including growth hormone and IGF-I, bFGF has no well-

CHONDROCYTE BASIC FIBROBLAST GRO W7'H FACTOR RECEPTORS established effect on skeletal growth. Fibroblast growth factor is, however, a potent mitogen for chondrocytes ( 14,22,40,42,44,47,53), and recent studies have ascribed to it a role in embryonic development (24,29,51) and in expression of the chondrocyte phenotype (18,32). Data from our studies support the hypothesis that bFGF participates in the regulation of skeletal growth. They further suggest that this regulation involves at least two distinct classes of bFGF binding sites. The mechanisms governing the distribution of bFGF among these sites and their respective roles in modulating bFGF action remain to be determined. Acknowledgment: We wish to thank H e n r y J . Mankin, M.D., an d J u d ah F o lk m a n , M.D., w h o helped m a k e this work possible, an d Mrs. B r e n d a White f o r e x p e rt assistance in preparing th e manuscript. This work was supported in p a r t by USPHS Research

Grant AR3 1068 and an Orthopedic Research and Education Foundation Zirnmer Career Development Award (S.B.T.), an d NCI Research Grant CA-37392 ( M . K . ) .

REFERENCES 1. Anderson KJ, Dam D, Lee S , Cotrnan CW: Basic fibroblast

2. 3.

4.

5.

6.

7.

8. 9. 10. 11.

growth factor prevents death of lesioned cholinergic neurons in vivo. Nutiire 332:36G361, 1988 Baird A, Schubert D. Ling N, Chillemin R: Receptor and heparin binding domain.; of basic fibrohlast growth factor. Proc Nut1 Acud Sci U S A 85:232&2328, 1988 Bashkin P, Doctrow S , Klagsbrun M, Svahn CM, Folkman J , Vlodavsky I: Basic fibroblast growth factor binds to suhendothelial extracellular matrix and is released by heparatinase and heparin-like molecules. Biochernisfry 28: 17371743. 1989 Bolton AE, Hunter WM: The labelling of proteins to high specific radioactivities by conjugation to a '"I-containing acetylating agent. Application to the radioimmunoassay. Biochern J 133599-538, 1973 Chandrasekhar S , Harvey AK: lnduction of interleukin-I receptors on chondrocytes by fibroblast growth factor: a possible mechanism for modulation of interleukin-I activity. J Cell Physiol 138:236246, 1989 Dionne CA. Crumley G , Bellot F. et al.: Cloning and expression of two distinct high affinity receptors cross-reacting with acidic and basic fibroblast growth factors. EMBO J 9:2685-2692, 1990 Enomoto M, Kinoshita A, Pan H - 0 , Suzuki F. Yamamoto I , Takigawa M: Demonstration of receptors for parathyroid hormone on cultured rabbit costal chondrocytes. Biochem Biophys Kes Cornmun 162: 1222-1229, 1989 Fairbanks G. Steck TL. Wallach DFH: Electrophoretic analysis of the ma,ior polypeptides of the human erythrocyte membrane. Biochemistry 10:2606-26 17, 1971 Feige J-J, Raird A: Glycosylation of the basic fibroblast growth factor receptor: the contribiilion of carbohydrate to receptor function. J Biol C'heni 263: 14023-14029, 1988 Folkman J. Klagshrun M: Angiogenic factors. Science 235: 442-447. 1987 Gospodarowicz D: L.ocalization of a fibroblast growth factor

645

and its effect alone and with hydrocortisone on 3T3 cell growth. Nutiire 249: 123-127, 1974 12. Gospodarowicz D: Structural characterization and biological functions of fibroblast growth factor. Endocr Res 8 9 - 1 14, 1987 3 . Gospodarowicr D, Cheng J: Heparin protects basic and acidic FGF from inactivation. J Ce/l Physiol 128:475484, 1986 4. Gospodarowicz D, Mescher AL: A comparison of the responscs of cultured myoblasts and chondrocytes to fibroblast and epidermal growth factors. J Cell Plzysiol 93: 117128, 1977 5. Hollenberg MD, Cuatrecasas P: Methods for the biochemical identification of insulin receptors. In: Methods in M e m brnne Research, ed by M Blecher. New York, Marcel Dekker, 1976, pp 429477 16. Hou J , Kan M. McKeehan K, McBride G , A d a m P, MCKeehan WL: Fibroblast growth factor receptors from liver vary in three structural domains. Scierice 2.51 :665668, 1991 17. lozzo RV: Biology of disease: proteoglycan structure, function and role in neoplasia. T,ab Invest 53:373-396, 1985 18. lwamoto M, Shimazu A, Nakashima K , Suzuki F, Kato Y: Reduction in basic fibroblast growth factor receptor is coupled with terminal differentiation of chondrocytes. J B i d Chetn 266:461467, 1991 19. Iwane M, Kurokawa T , Sasada K. Seno M. Nakagawa S , Tgarashi K: Expression of cDNA encoding basic fibroblast growth factor in E. coli. Biochetn Biophys Res Commun 146:470-471, 1987 20. Johnson DE, Lee PL, Lu J, Williams LT: Diverse forms of a receptor for acidic and basic fibroblast growth factors. Mol Cell BicJ/ 10:4728-4736, 1990 21. Kahn CR: Membrane receptors for polypeptide hormones. In: Me2hod.s in Mernhrnne Biology, Vol. 3, ed by ED Korn, New York, Plenum Press, 1975, 81-146 22. Kasper S , Friesen HG: Human pituitary tissue secretes a potent growth factor for chondrocyte proliferation. J Clin Endocrind 62:7&76. 1986 23. Kato Y, Iwamoto M: Fibroblast growth factor is an inhibitor of chondrocyte terminal differentiation. J Biol Chem 265: 5903-5909. 1990 24. Kimelman D , Abraham JA , Haaparanta T , Palisi l M , Kirschner MW: The presence of fibroblast growth factor in the frog egg: its role as a natural mesoderm inducer. Science 242: 1053-1056, 1988 25. Kurokawa M, Doctrow SR, Klagsbrun M: Neutralizing antibodies inhibit the binding of basic fibroblast growth factor to its receptor but no1 to heparin. J Biol Chem 264:76867691, 1989 26. Laemmli OK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:68& 685, 1970 27. Ledbetter SR. Tyree B, Hassell JR. Horigan E: Identificalion of the precursor protein to basement membrane heparan sulfate proteoglycans. J Biol Chern 260:81064113, 1985 28. Lee PL. Johnson DE, Cousens LS, Fried VA, Williams LT: Purification and complementary DNA cloning of a receptor for basic fibroblast growth factor. Science 24557-60, 1909 29. Liu L. Nicoll CS: Evidence for a role of basic fibroblast growth factor in rat embryonic growth and differentiation. Endocrino/ogy 123:2027-203 1, 1988 30. Mignatti I', Tsuboi R. Robbins E, Rifkin DB: In vitro angiogenesis on the human amniotic membrane: requirement for basic fibroblast growth factor-induced proteinases. J Cell Biol 108:671482, 1989 31. Morrison RS. Sharma A, De Vellis J, Bradshaw RA: Basic fibroblast growth factor supports the survival of cerebral cortical neurons in primary culture. Proc Nurl Acud Sci USA 83:7537-7541, 1986

J Orfhop R C S , Vol. 10, N o . 5 , 1992

646

S. B . TRIPPEL ET AL.

32. Moscatelli D: High and low affinity binding sites for basic tibroblast growth faclor on cultured cells. Absence of a role for low affinity binding in the stimulation of plasminogen activator production by bovine capillary endothelial cells. J Cell Physiol 13 I :123-1 30, 1987 33. Moscatelli D: Metabolism of receptor bound and matrix bound basic fibroblast growth factor by bovine capillary endothelial cells. J Cell B i d 102753-760. 1988 34. Moscatelli D, Presta M, Rifkin DB: Purification of a factor from human placenta that stimulates capillary endot helial cell protease production, DNA synthesis and migration. Proc Natl Acnd Sri IJSA 83:2091-2095, 1986 35. Neufeld G , Gospodarowicz D: Basic and acidic fibrobla.;t growth factors interact with the same cell surface receptors. J Bid Chem 261:5631-5637, 1986 36. Neufeld G , Gospodarowict D: Identification of the fibroblast growth factor receptor in human vascular endothelial cells. J Cell Physiol 136:537-542, 1988 37. Nilsson A, Lindahl A, Edin S, lsaksson OGP: Demonstration of growth hormone rcccptors in culture of rat epiphyseal chondrocytes by specific binding of growth hormone and immunohistochemistry. J Endocrind 122:69-77, 1989 38. Olwin BB, Hauschka SD: Identification of the fibroblast growth factor receptor of Swiss 3T3 cells and mouse skeletal muscle myoblasts. Biochemistry 25:3487-3492, 1986 39. Olwin BB, Hauschka SD: Cell type and distribution of the fibroblast growth factor receptor. J Cell Biochem 39:443454, 1989 40. Osborn KD. Trippel SR. Mankin HJ: Growth factor stimulation of adult articular cartilage. J Orfhop Res 7 5 5 4 2 , 1989 41. Oury F, Darbon J-M: Fibroblast growth factor regulates the expression of luteiniring hormone receptors in cultures rat granulosa cells. Biochem Biophys Res Cornrnun 156:634643. 1988 42. Prins APA. Lipman JM, Sokoloff L,: Effect of purified growth factors on rabbit articular chondrocytes in monolayer culture. I. DNA synthesis. Arthritis Rheum 25: 12171227, 1982 43. Kodbard D: Mathematics of hormone-receptor interaction. Adv Exp Med Biol36:289-326, 1973 44. Sachs BL, Goldberg VM, Moskowitr RW, Malemud CJ: Response of articular chondrocytes to pituitary fibroblast growth factor (FGF). J Cell Plrysiol 112:51-59, 1982 45. Saksela 0, Moscatelli D, Sommer A , Rifkin DR: Endothelial cell derived heparan sulfate binds basic fibroblast growth factor and protects it from proteolytic degradation. J Cell Biol 107:743-751, 1988

J Orthop Res, Vol. 10, N o . 5 , 1992

46. San Antonio JD, Winstron BM. Tuan RS: Regulation of chondrogenesis by heparan sulfate and structurally related glycosaminoglycans. Dev Biol 123:17-24, 1987 47. Sasse J, Sullivan R, Klagsburn M: Purification of cartilagederived growth factors. Methods f k : ) ~ ? f O / 146:32&325, 1987 48. Scatchard G: The attractions of proteins from small molecules and ions. Ann IVY Acad Sci 51:660-672, 1949 49. Schubert D, Ling N, Baird A: Multiple influences of a heparin binding growth factor on ncuronal development. .I Cell Biol 104:635443, 1987 SO. Schweigerer L. Neufeld G , Mergia A, Abraham JA, Fiddes JC. Cospodarowiu D: Basic fibroblast growth factor in human rhabdomyosarcoma cells: implications for the proliferation and neovascularization of myoblast derived tumors. Proc N u f l Acad Sci USA 84:842-846, 1987 51. Slack JMW, Darlington BG, Heath JK. Godsave SP: Mesoderm induction in early xenopus embryos by heparinbinding growth factors. Nufirre 326: 197-200, 1987 52. Srivistava VML, Malemud CJ. Sokoloff L: Chondroid expression by lapine articular chondrocytes in spinner culture following monolayer growth. Connect Tis.Yw Res 2: 127-136, I974 53. Sullivan R, Klagsbrun M: Purification o f cartilage-derived growth factor by heparin affinity chromatography. J Biol Chem 260:2399-2403, I985 54. Trippel S B , Chernausek SD, Van Wyk JJ, Moses AC, Mankin HJ: Demonstration of type I and type 11 somatomedin receptors on bovine growth plate chondrocytes. J Orf h o p Rc~s6:817-826, 1988 55. Trippel SB, Van Wyk JJ, Foster MB. Svoboda ME: Characterization of a specific aomatoniedin-C receptor on isolated bovine growth plate chondrocytes. Endocrinahgy 112: 2128-2136, 1983 56. Venn G. Mason RM: Absence ofkeratan sulphate from skeletal tissues of mouse and rat. Biochern J 228:443450, 1985 57. Vlodavsky I , Folkman J, Sullivan R. Fridman R, IshaiMichaeli K. Sasse J , Klagsbrun M: Endothclial cell-derivcd basic fibroblast growth Factor: synthe. subendothelial extracellular matrix. Pruc, Null Acad Sci USA 84:2292-2296, 1987 58. Walickc PA, Feige J-J. Baird A: Characterization of the neuronal receptor for basic fibroblast growth factor and comparison to rcccptors on mesenchymal cells. J Biol C‘hem 264:412WI 26, 1989 59. Yayon A, Klagsbrun M. Esko J , Leder 1’. Ornitz D: Cell surface heparin-like molecules area required for binding of bFGF to its high affinity receptor. Cell 64:841-848, 1YYl