JOURNAL OF BONE AND MINERAL RESEARCH Volume 5, Number 11, 1990 Mary Ann Liebert, Inc., Publishers

Synergistic Effect of Transforming Growth Factor and Fibroblast Growth Factor on DNA Synthesis in Chick Growth Plate Chondrocytes IAN D. CRABB, REGIS J. O'KEEFE, J. EDWARD PUZAS, and RANDY N. ROSIER

ABSTRACT Transforming growth factor 0 and fibroblast growth factor are mitogens for chick growth plate chondrocytes. TGF-0 stimulated a 3.5-fold increase, and FGF a 13.5-fold increase in the rate of thymidine incorporation after a 24 h exposure, TGF-/3 and FGF were synergistic in chondrocytes, causing a 73-fold stimulation in thymidine incorporation compared with control. This synergistic response was not dependent upon the simultaneous presence of both mitogens. Sequential exposure of chondrocytes to TGF-0 and FGF in either order reproduced in large part the synergistic interaction observed when both growth factors were present simultaneously. The time required for induction of the subsequent synergistic response was brief and, in the case of TGF-6, corresponded to the time required for ["51]TGF-fi receptor binding. EGF and PDGF were not mitogenic for chondrocytes, and neither of these factors enhanced the response of the cells to either TGF-/3 or FGF. Finally, TGF-/3 and FGF did not, either alone or in combination, elevate intracellular CAMP levels. These results emphasize the importance of examining growth factor effects in the context of other growth regulators. Furthermore, this specific and dramatic synergistic stimulation of thymidine incorporation may provide a useful tool in elucidating the mitogenic mechanism of the individual growth factors.

INTRODUCTION

6 (TGF-0) and fibroblast growth factor are protein regulators of cellular function that have been isolated from many varied tissues, including Since they have been shown to regulate the growth and differentiation of these cells in cult u ~ e , ' * - they ~ ) are termed autocrine and/or paracrine regulators. Two forms of fibroblast growth factor have been identified: they are referred to as acid fibroblast growth factor and basic fibroblast growth factor. The acid form appears to be localized in brain and retina, but the basic form is u b i q u i t o ~ s . ~ 'Basic - ~ ~ fibroblast growth factor (FGF) is homologous with cartilage-derived growth factor, a protein purified from cartilage that stimulates both DNA and proteoglycan synthesis in chondrocytes."' Most of the cell

T

RANSFORMING GROWTH FACTOR

types that produce FGF have specific high-affinity FGF receptors on their surface, suggesting an autocrine role for this protein.lJ l o ) For both osteoblasts and articular chondrocytes in culture, FGF is mitogenic."'' This effect of FGF appears to be the result of a shortened G I phase of the cell cycle.(111 In addition to its effect on proliferation, FGF stabilizes the differentiated phenotype of articular chondrocytes. Articular chondrocytes cultured in the absence of FGF become spindle shaped and cease production of type I I collagen and proteoglycans. However, when FGF is present in the medium the chondrocytes retain their phenotype and produce an apparently normal extracellular matrix."21 Transforming growth factor fil is a 25,000 dalton polypeptide initially isolated from transformed cells1111 and defined by its ability to stimulate anchorage-independent growth of rat fibrobIasts.'l4l As with FGF, many cell types

Department of Orthopaedics, The University of Rochester, Rochester, N Y 14642. 1105

1106

CRABB ET AL.

synthesize TGF-(3 and have specific high-affinity receptors was added to the cells for an overnight 16 h digestion in a for the TGF-(3 is homologous with cartilage shaking water bath at 37°C. The cells were then filtered inducing factor A (CIF-A), a growth factor isolated from through a 40 pm nylon screen mesh, centrifuged for 4 minbovine bone that causes mesenchymal cells to differentiate utes at 600 x g, and washed twice with modified F-12 meinto chondrocytes.''.'g) TGF-(3 inhibits the proliferation of dium. The cells were subsequently rinsed twice in a solucells of epithelial ~ r i g i n . " ~ - ~The ' ) response of mesenchy- tion containing citrate-buffered saline (125 mM NaCI, 19 ma1 cells to TGF-(3, however, depends on the full comple- mM citric acid, and 10 mM K'HPO,, p H 6.0) at 4°C to ment of growth factors to which the cells are exposed. dissolve and remove any mineral debris. After a final TGF-(3 stimulates anchorage-independent growth of myc- washing in modified F-12 medium, the cells were counted transfected fibroblasts in the presence of platelet-derived with a hemocytometer. growth factor (PDGF) but inhibits their growth in the presence of epidermal growth factor (EGF).123)For most Chondrocyte cell culture cell populations, including articular chondrocytes, TGF-(3 Chondrocytes were plated in 2 cm' multiwell culture is an antiproliferative agent and inhibits the effects of other growth factor^.''^) However, TGF-(3 is mitogenic for plates (Corning) at a density of 200,000 cells per cm' in Dulbecco's modified Eagle's medium (DMEM) containing growth plate chondrocytes.(z41 Efforts in growth factor research have by necessity fo- 5% fetal bovine serum (FBS, GIBCO, lot 28N 3863). After cused on isolating, purifying, and testing individual pep- 24 h in culture, the plating medium was removed and retides on cells in culture. In vivo, however, osteoblasts and placed with fresh identical medium to which varying conchondrocytes are exposed to multiple local growth factors centrations of growth factors had been added. In sequenand systemic hormones simultaneously, and the effects of tial experiments culture wells were rinsed twice with media one of these multifunctional peptides are likely modulated plus 5% FBS between treatments. Terminal assays were by the presence of others.(2s.z6)Factors present in serum performed on the cell cultures at intervals as reported in have been shown to modify the response of chondrocytes Results. to TGF-(3.f'"1 Recent work by Centrella et al. investigated the ability of combinations of TGF-(3 and FGF to stimuThymidine incorporation late DNA synthesis in confluent osteoblasts.(z7)Although Thymidine incorporation was determined through the TGF-(3 and FGF were both found to stimulate proliferation individually, TGF-(3 failed to enhance the maximal labeling of cell cultures with 1 or 5 pM [3H]thymidine(New England Nuclear, Boston, MA) with a specific activity of 8 effect of FGF. We have used cultured chick growth plate chondrocytes or 2.4 mCi/pmol, respectively. Following a 4 h exposure, to investigate the role of growth factors in the regulation the radioactive medium was removed and the cells were of cellular proliferation. The present study demonstrates a rinsed with 1 ml of 0.15 M NaCI. The cells were removed dramatic synergistic effect of TGF-0 and FGF on DNA from the culture wells by scraping after the addition of synthesis in isolated chick epiphyseal chondrocytes. This 0.25 ml of 0.25 M NaOH and placed in 5 ml tubes. The effect can be reproduced by sequential exposure of the culture wells were rinsed with an additional 0.25 ml of 0.25 cells to the individual growth factors, and the time course M NaOH to recover any remaining radiolabeled DNA. of the induction of the effect appears to be related to re- The labeled solution was neutralized by the addition of 0.5 ml of 0.25 M HCI. Culture medium (1 ml) containing 5% ceptor binding. serum was added to the test tubes and the DNA precipitated by the addition of 0.5 ml of 10 N perchloric acid. MATERIALS AND METHODS The samples were chilled to 4°C and centrifuged at 14,000 x g for 30 minutes. The supernatant was removed, and Chondrocyte isolation the pellet was dissolved in 0.75 ml of 0.25 M NaOH. The Chondrocytes were isolated from 3- to 4-week-old solution was counted in 10 ml Liquiscint (National Diagchicks. After sacrifice in a CO, cannister, the long bones nostics, Mannville, NJ) in a liquid scintillation counter. of the leg were sterilely dissected free of soft tissue. The Standards of radiolabeled thymidine were prepared and cartilaginous tissue of the growth plate was then removed counted so that an accurate determination of the femtoand the shavings placed in modified F-12 medium (magne- moles of thymidine incorporated into DNA could be obsium free, 0.5 mM CaCI,, Sigma Chemical, St. Louis, tained. MO). After rinsing and weighing, the cartilage was subjected to an initial 30 minute digestion with trypsin (0.1070, /' 's]TGF-pbinding Sigma type 111) in modified F-12 medium at 37°C. For each 20 mg cartilage, 1 ml enzyme solution was used. The After 24 h in culture, the medium was removed and retrypsin solution was subsequently removed and the tissue placed with binding buffer (DMEM, 2 mg/ml of bovine rinsed twice with modified F-12 medium and then sub- serum albumin, BSA, 25 mM HEPES, and 50 pg/ml of asjected to a 1 h 0.1% hyaluronidase digestion in an equal corbate). The cultures were allowed to equilibrate for 1 h, volume of modified F-12 medium at 37°C. After removal after which the medium was aspirated and 40 pM ['"I]of the hyaluronidase solution, an equal volume of 0.1% TGF-(3 (50-100 pCi/pg, R & D, Minneapolis, MN) in fresh collagenase (Sigma, type I1 A) in modified F-12 medium binding buffer was added. Binding was allowed to proceed

1107

TGF-/3 AND FGF ON CHONDROCYTES at 37°C for varied lengths of time and terminated by four rinses with ice-cold binding buffer. Bound ['z51]TGF-/3was solubilized by the addition of 0.25 ml of l"/b Triton X-100 in isotonic saline pH 7.4 for 1 h, followed by a second rinse with the same solution. The bound fraction was combined with 10 ml Liquiscint and counted in a liquid scintillation counter. Nonspecific binding was delermined in the presence of a 250-fold excess of cold TGF-0 and subtracted from total binding to determine the specific receptor binding. Standards of 40 pM ['zsI]TGF-O in binding buffer were prepared so that counts per minute could be converted into femtomoles bound TGF-B.

CAMP assay After 24 h in culture the medium was removed and replaced with 1.0 ml serum-free DMEM per well. The cultures were allowed to equilibrate for l h, after which 10 pl isobutylmethylxanthine (10 mM in 95% ethanol) was added and the cultures were incubated for 20 minutes at 37°C. Growth factors were then added ar the indicated concentrations, and the cultures were returned to the incubator. After 20 minutes the reaction was stopped by aspirating the medium, adding 0.5 ml of 95% ethanol per well, and allowing cultures to air dry.("' The dried cell samples were reconstituted in 300 PI reaction buffer (50 mM Tris and 4.0 mM EDTA, pH 7.5). Aliquots of 50 pl were removed and assayed for ["HICAMPa:; previously described.'2VJRadiolabeled CAMP (41 Ci/mM, Amersham, Arlington Heights, IL) and binding protein (protein kinase, Sigma) were prepared as previously reported.'2gJ

ulated cell division. For this reason all subsequent experiments were performed after a 24 h exposure to the growth factors. TGF-P produced a dose-dependent stimulation of thymidine incorporation into the DNA of chick epiphyseal chondrocytes. The effect was biphasic, with a 3.5-fold maximal stimulation at 1.0 n g h l (Fig. 2). Exposure of cultured chick epiphyseal chondrocytes to FGF also caused a dose-dependent increase in the rate of thymidine incorporation into DNA (Fig. 3). The lowest effective concentration of FGF was 0.1 ng/ml, which resulted in a 2-fold increase in the rate of thymidine incorporation. At 10 ng/ml the FGF effect was essentially maximal, yielding a 13.5-fold stimulation. Unlike TGF-P the dose response to FGF was not biphasic within the concentration range tested. To examine whether the stimulatory effects of TGF-0 and FGF on DNA synthesis are interdependent, experiments were performed using both growth factors in combination. Figure 4 displays the effects of combining three different FGF concentrations with the dose response to TGF-0. The maximum stimulation of thymidine uptake was observed with a TGF-fl concentration of 0.3 ng/ml and a FGF concentration of 100 ng/ml. This combination generated a 73-fold stimulation over control (12.8 + 0.5 versus 0.17 0.02 pmol thymidine per 10" cells per h). For all combinations of TGF-/3 greater than 0.03 ng/ml and FGF greater than 1 ng/ml there was a synergistic stimulation of thymidine incorporation. The synergy resulted in 2.3- to 4.4-fold greater stimulation than would be predicted by adding the individual responses. The TGF-/3

*

Growth factors TGF-0, FGF, and PDGF were purchased from R & D Systems (Minneapolis, MN). EGF was purchased from Bachem. All growth factors were individually reconstituted in solutions containing 4 mM HCI and 1 mg/ml of BSA and stored at 4°C.

1.2 I

1

Sfafisficalanalysis Student's f-test was used for statistical comparisons. 1 standard Each data point is presented as the mean error of the mean (SEM) with n = 4. Treai.ed values were considered significantly different from controls when the p value was less than 0.05.

*

RESULTS Time course experiments with TGF-P (0.3 ng/ml) and FGF (10 ng/ml) demonstrated maximal stimulation of thymidine incorporation at 24 h o f exposure (Fig. 1). For both factors the rate of thymidine incorporation increased exponentially to 24 h and then decreased exponentially to 48 h, the longest time measured. We have interpreted these results as reflecting the time course for a single cycle of stim-

U

0

10

20

30 Hours

40

50

FIG. 1. Time course of the effect of FGF (W), TGF-0 (A ), or medium alone (0)on thymidine incorporation in chick growth plate chondrocytes. Freshly isolated chondrocytes were plated in 5% FBS for 24 h as described in Materials and Methods. After this time the medium was removed and replaced with DMEM and 5% FBS alone or in combination with TGF-(3 (0.3 ng/ml) or FGF (10 ng/ml). Cultures were exposed to 1 pM [jHIthymidine for 4 h intervals starting at 4, 8, 12, 24, 36, and 48 h. Data are mean f SEM of n = 4 replicate cultures per condition. Asterisks denote significant differences (p < 0.02) from control.

CRABB ET AL.

1108

15

0.6

8 "0

-

8

(D

0.4

10

0 7

h

= L

0.3 u)

a, -

g 0.2

Q

0.1

I

I

.1

1

0 10

TG F- 0 (nglmI) FIG. 2. Dose-response curve for TGF-@. Chondrocytes were plated in monolayer cultures in DMEM plus 5% FBS and after 24 h were exposed to varying concentrations of TGF-@ in identical fresh medium for an additional 24 h. Cultures were then labeled for 4 h with 1 pM [3H]thymidine. Data are mean SEM of n = 4 replicate cultures per condition. Asterisks denote significant differences (p < 0.05).

*

C

.o 1

.1

1

TGF-R (ng/ml) FIG. 4. Chondrocytes were plated in monolayer culture in DMEM plus 5% FBS and after 24 h were exposed to varying concentrations of TGF-@ alone (O), with 0.1 ng/ ml of FGF (x), with 1.O ng/ml of FGF (A),with 10 ng/ml of FGF (.'>), or with 100 ng/ml of FGF (B). All combinations were in the presence of DMEM plus 5% FBS. After 24 h cultures were labeled for 4 h with 1 gM ['Hlthymidine. Data are mean + SEM of n = 4 replicate cultures per condition. All values are significantly different from control (no added agents), with p < 0.05. Error bars for TGF-/3 alone and with 0.1 ng/ml of FGF were smaller than their respective symbols and omitted for clarity.

EGF and PDGF to determine whether they had interactive effects with TGF-@ or FGF. By itself, 0.3-10 ng/ml of EGF had no effect on thymidine incorporation. The combination of EGF and TGF-@ or E G F and FGF yielded no stimulation above that observed with only TGF-@or FGF, respectively. Similarly, 0.15-15 ng/ml of PDGF was without effect on DNA synthesis either alone or in combination with TGF-@ or FGF (data not shown). The synergistic simulation of thymidine incorporation by TGF-@and FGF suggests an amplification in the regulation of DNA synthesis. To determine whether this amplifiC .1 1 10 100 cation was dependent upon concomitant exposure of the cells to both groth factors, a series of sequential treatment FGF Dose (nglml) experiments were performed. These experiments used FIG. 3. Dose-response curve for FGF. Chondrocytes TGF-@and FGF at concentrations of 0.3 and 10 ng/ml, rewere plated in monolayer cultures in DMEM plus 5 % FBS spectively. Figure 5 demonstrates the results of pretreatand after 24 h were exposed to varying concentrations of FGF in identical fresh medium for an additional 24 h. Cul- ment with TGF-@on FGF-stimulated thymidine incorporatures were then labeled for 4 h with l pM [3H]thymidine. tion. TGF-@pretreatment of the cells generated 78-94% of Data are mean f SEM of n = 4 replicate cultures per con- the synergistic response observed when both factors are dition. Asterisks denote significant differences (p < 0.01). present simultaneously. The effect of pretreatment was maximal by 1 h, as longer TGF-@pretreatment intervals (4 and 16 h) did not increase the FGF-stimulated thymidine dose-response curve maintained a biphasic form in the incorporation. The data for FGF pretreatment are presented in Fig. 6. presence of FGF. Moreover, the time course of maximal thymidine incorporation with the combination of TGF-@ FGF pretreatment also caused a synergistic response, but and FGF paralleled the time course observed with the indi- the effect was only 4556% of that seen with the simulvidual growth factors, with a peak response after 24 h of taneous addition of both growth factors. As with TGF-@, exposure. Additional experiments were carried out with the effect of FGF pretreatment was maximal at 1 h.

1109

TGF-@AND FGF ON CHONDROCYTES

Incubation Times (hours) FIG. 5. Effects of TGF-0 pretreatment on FGF-stimulated thymidine incorporation. Chondrocytes in monolayer culture were pretreated for varying lengths of time ( 1 , 4, and 16 h) with TGF-@ (M)or medium (a),following which both cultures were rinsed and exposed to FGF for 24 h. The thymidine incorporation of these cultures is compared to both untreated cultures (Ill) and the combination of TGF-/3 and FGF (El). At 25, 28, and 40 h, respectively, cultures were labeled for 4 h with 1 pM ['Hlthymidine. Data are mean + SEM of n = 4 replicate cultures per condition. All treated values were significantly different (p < 0.02) from control (no added agents).

Longer times (4 and 16 h) demonstrated no further stimulation of thymidine incorporation. Thus sequential exposure of these chondrocytes to TGF-@ and FGF in either order can in large part reproduce the synergistic interaction observed when both growth factors are present simultaneously. To further elucidate the time course of this cellular induction, chondrocytes were again exposed to growth factors in a sequential manner, but with shorter pretreatment intervals. As demonstrated in Fig. 7, pretreatment for only 20 minutes with either growth factor was sufficient to yield maximal induction of the synergistic effect. TGF-@is known to mediate cellular effects through specific membrane receptors. To determine whether the rapid onset of cellular induction by TGF-@corresponded to specific binding of the molecule to cells, the binding time course with [IZSI]TGF-Pwas examined (Fig. 8). The results indicate that the specific binding of TGF-/3 paralleled the subsequently observed stimulation of thymidine incorporation. In fact, when the TGF-/3 binding data and thymidine incorporation data were compared at corresponding time points, they demonstrated a significant linear correlation, with a correlation coefficient of 0.90 and a p value less than 0.02. The mechanisms of action of TGF-/3 and FGF remain largely unknown. A class of mitogens, of which forskolin is the prototype, are known to generate CAMP as an inte-

2 4

4

c,

c

L

0

5 (D =3

0

G 3

8

Q)

(D

0

--'2 3.

607

a,

11.111.1.1.1.11

5 1

EQ

0

0 1,24

4,24

16,24

20

40 60 80 100 120 M i nut es

FIG. 7. The effect of brief pretreatment with 10 ng/ml of FGF (0)and 0.3 ng/ml of TGF-P (m),presented as the FIG. 6. Effects of FGF pretreatment on TGF-@-stirnu- percentage stimulation of thymidine incorporation versus lated thymidine incorporation. Chondrocyi es in monolayer control. Pretreatment times were 5, 10, 20, 35, 60,and 120 culture were pretreated for varying lengths of time (1, 4, minutes, following which the cultures were rinsed and the and 16 h) with FGF (M)or medium (El), following which cells were exposed to the other growth factor for 24 h. both cultures were rinsed and exposed to W F - @ for 24 h. Controls were treated in an identical manner except using The thymidine incorporation of these cultures is compared medium alone for pretreatment. Cultures were labeled at to both untreated cultures (Ill) and the combination of 24 h for 4 h with 5 pM [3H]thymidine. Results are exTGF-P and FGF (El). At 25, 28, and 40 h, respectively, pressed as the ratio of thymidine incorporation of TGF-0cultures were labeled for 4 h with 1 p M of [3H]thymidine. pretreated cultures to that of control pretreated cultures. Data are mean * SEM of n = 4 replicate cultures per con- Data are mean * SEM of n = 4 replicate cultures per condition. A11 treated values were significantly different (p < dition. All values are significant @ value < 0.01) com0.01) from control (no added agents). pared to the briefest time point.

Incubation Times (hours)

CRABB ET AL.

1110

f f 20

LL

.01

0

60 80 Minutes of TGF-D 40

100

120

FIG. 8. TGF-8 binding time course. Chondrocytes were exposed to ['z51]TGF-/3 (40 pM) in binding buffer, for the intervals in Fig. 7 at 37°C. Nonspecific binding (approximately 20%) was subtracted from total binding 10 deter-

mine the specific receptor binding. Data are mean * the SEM of n = 3 replicate cultures per condition. All time points are significantly different @ < 0.02) from control (1 minute incubation with "'II-TGF-/3).

gral part of their stimulation of DNA synthesis. Assays of cAMP levels in response to growth factor addition were undertaken to determine whether TGF-(3, FGF, or the combination exerted mitogenic effects through a CAMPmediated mechanism. As shown in Fig. 9, the generation, of cAMP by either the individual growth factors or the combination was not significantly different from control.

DISCUSSION This report demonstrates that TGF-(3 and FGF synergistically stimulate DNA synthesis in chick growth plate cells cultured in monolayer. The synergistic effect of TGF-/3 and FGF is not dependent upon the simultaneous presence of these growth factors, as sequential treatment of the cells with TGF-@and FGF leads to a synergistic stimulation of thymidine incorporation. In the case of TGF-@ pretreatment, induction of the subsequently observed synergistic effect on thymidine incorporation takes place in the same time span required for TGF-(3 binding. The stimulatory effects are not mediated through CAMP. We utilized short-term cultures to observe the cells in a representative phenotype. These short-term cultures maintained growth plate chondrocyte specific characteristics, including type I1 and type X collagen synthesis, high levels of alkaline phosphatase, and maintenance of proteoglycan synthesis. Growth plate chondrocytes are slow growing in monolayer cultures and thus provide a sensitive system in which to observe stimulatory growth factor effects on proliferation. ( * 4 . 3 0 ) Interactions of TGF-(3 with other growth factors have been reported in other cell ~ y s t e m s . ' * ~In. ~myc-trans~.~~~

TGF-0 FIG. 9. Effect of growth factors on cAMP levels. Chondrocytes were incubated with growth factors for 20 minutes after being previously exposed to the phosphodiesterase inhibitor IBMx. The amount of cAMP produced was determined as described in Methods. Data are mean f SEM of n = 4 replicate cultures per condition. Only forskolin treatment was significantly different from control @ < 0.01).

fected fibroblasts, PDGF had no effect on supporting anchorage-independent growth and TGF-P only a small effect. However, the combination yielded a large, synergistic stimulation of anchorage-independent colony formation. Surprisingly, under the same conditions TGF-(3 inhibited the stimulation of anchorage-independent colony growth normally seen with EGF.(z31The demonstration that in the same cell line TGF-P can cause a synergistic stimulation or a strong inhibition, depending on the presence of other growth factors, suggests that the biologic response is determined by the net effect of all local regulators. In bone and cartilage TGF-8 and FGF are both present in significant a r n o u n t ~ ( ~and ' . ~ ~thus ) might be expected to have interactive effects on chondrocytes and osteoblasts. In osteoblasts, however, no synergistic stimulation of thymidine incorporation by TGF-(3 and FGF was observed.'z71F G F actually decreased the stimulation of DNA synthesis seen with TGF-(3 alone. The specific and dramatic response of growth plate chondrocytes to the combination of TGF-(3 and FGF presented here suggests a regulatory role for these proteins in chondrocyte proliferation. Although growth plate chondrocytes replicate slowly in culture, they undergo rapid proliferation and subsequent differentiation in the process of endochondral ossification. The stimulus for chondrocyte proliferation in endochondral ossification. The stimulus for chondrocyte proliferation in enchondral ossification remains unknown. Somatomedins generate a 2-fold increase in chondrocyte DNA synthesis, and it has been suggested that these molecules are involved in regulating chondrocyte proliferation. Is' The data presented here suggest that TGF-/3 and FGF, with a 73-fold stimulation of DNA synthesis, are possibly important regulators of this process.

1111

TGF-L?AND FGF ON CHONDROCYTES The finding of the synergistic effects reported in this study underscores the inherent difficulty in interpreting biologic significance from effects of individual growth factors on cells. The effect of one growth factor may be amplified or diminished by the presence of another. Furthermore, identifying and purifying stimulatory substances may be complicated by growth factor interactions, as discovery o f an important regulator may be masked by the absence of an essential synergy-producing “cofactor .” The synergistic effects of TGF-0 and FGF on DNA synthesis have not been reported in other cell culture systems and thus may be specific for growth plate chondrocytes. Furthermore, other growth factor combinations did not produce a stimulatory effect. EGF or PDGF alone had no stimulatory effect on chondrocyte proliferation and no enhancing effect in combination with TGF or FGF. The lack of response to EGF was surprising as it had been previously shown to be a mitogen for costal chondrocytes.”’ This discrepancy may be explained by the different cell types and/or culture conditions. The synergistic interaction of TGF-0 and FGF implies that these growth factors mediate their effects via different, but complementary, pathways. I f FGF initiated the same sequence of subcellular events as elicited by TGF-0, then the effect of the combination should tle equivalent to increasing the concentration of TGF-(3. As high concentrations of TGF-0 are inhibitory, the addition of FGF should also be inhibitory. When TGF-0 and FGF were added in combination, however, the observed rate of thymidine incorporation by any concentration of TGF.0 was only increased by adding FGF. Since the effect of adding FGF was not the same as increasing the concentration of TGF0,the utilization of separate mechanisms by each growth factor can be inferred. Although these molecules have distinct mechanisms of action, each can amplify the mitogenic signal of the other. It is interesting that in the presence of FGF the biphasic shape of TGF-P’s dose-response curve was maintained. The synergistic response elicited by sequential exposure of the cells to growth factors suggests that a sensitization or induction has occurred. One explanation for this induction on the molecular level might include the activation by a growth factor of a cellular proto-oncogene. The product of this gene might complement the action of the other growth factor, leading to a synergistic response. Among the many known proto-oncogenes, c-myc appears to be important in mediating the actions of TGF-0. Cells that are unresponsive to TGF-P can be rendered responsive by transfection with the myc gene.L34)Since the TGF-0 response depends on an activated c-myc gene and FGF has been shown to elevate c-myc mRNA levels in a stable and prolonged fashion,(3s36) c-myc may underlie the synergy observed between these two growth factors. Although the actual mechanism of cellular induction is undoubtedly complex, proto-oncogenes, such as c-myc and c-fos, may well mediate important steps in preparing a cell for DNA synthesis. In summary, we report here a synergistic interaction of TGF-0 and FGF on DNA synthesis in chick growth plate chondrocytes. The response of these cells appears to be

unique and as such may be important in regulating proliferation in such processes as enchondral ossification. The presence of a cellular induction was demonstrated and may provide a useful tool in elucidating the mitogenic mechanism. Most importantly, we have demonstrated that the effect of a growth factor depends to a large degree on the presence of other local regulators. This has important implications for the investigation of isolated and purified growth factors.

ACKNOWLEDGMENTS The authors greatly appreciate the technical assistance

of Laura Hanson. The work was supported in part from Public Health Service Grants AR 28420 (Puzas) and T32AM07092-12 (O’Keefe) and Orthopedic Research and Educational Foundation Grants 87-462 (Crabb) and 84-1-CA (Rosier).

REFERENCES I . Ellingsworth LR, Brennan JE, Fok K. Rosen DM, Bentz H, Piez KA, Seyedin SM 1986 Antibodies to the N-terminal portion of cartilage-inducing factor beta. J Biol Chem 261: 12362-12367. 2. Sullivan R, Klagsbrun M 1985 Purification of cartilage-derived growth factor by heparin affinity chromatography. J Biol Chem 260:2399-2403. 3. Gospodarowicz D, Ferrara N, Schweigerer 1.. Neufeld G 1987 Structural characterization and biological functions of fibroblast growth factor. Endoc Rev 8:95-114. 4. Kato Y, Nomura Y, Tsuji M, Ohmae H, Kinoshita M, Hamamoto S, Suzuki F 1981 Cartilage-derived factor (CDF). Exp Cell Res 132:339-347. 5. Kato Y, Hiraki Y, lnoue H , Kinoshita M, Yutani Y, Suzuki F 1983 Differential and synergistic actions of 3omatomedinlike growth factors, fibroblast growth factor and epidermal growth factor in rabbit costal chondrocytes. Eur J Biochem 129~685-690. 6. Kato Y, Nomura Y , Tsuji M, Ohmae H, Nakazawa T, Suzuki F 1981 Multiplication-stimulating activity (MSA) and cartilage-derived factor (CDF): Biological actions in cultured chondrocytes. J Biochem 90:1377-1386. 7. Gospodarowicz D, Neufeld G , Schweigerer L 1986 Fibroblast growth factor. Mol Cell Endocrinol 46:187-204. 8. Baird A, Esch F, Mormede P, Ueno N, Ling N, Bohlen P, Ying S, Wehrenberg W, Guillemin R 1986 Molecular characterization of fibroblast growth factor: Distribution and biological activities in various tissues. Recent Prog Horm Res 42:143-205. 9. Lobb R , Sasse J, Sullivan RI, Shing Y, D’Amore P, Jacobs J , Klagsbrun M 1986 Purification and characterization of heparin binding endothelial cell growth factor. J Biol Chem 261: 1924- 1928. 10. Neufeld G , Gospodarowicz D 1985 The identification and partial characterization of the fibroblast growth factor receptor of baby hamster kidney cells. J Biol Chem 260:1386013868. 11. Gospodarowicz D, Vlodavsky I , Savion N 1981 The role of fibroblast growth factor and the extracelllar matrix in the control of proliferation and differentiation of corneal endothelial cells. Vision Res 21:87-92.

CRABB ET AL.

1112 12. Kato Y, Gospodarowicz D 1985 Sulfated proteoglcan synthesis by rabbit costal chondrocytes grown in the presence and absence of fibroblast growth factor. J Cell Biol 100:477-485. 13. Roberts A, Lamb L, Newton D, Sporn M, DeLarco J , Todaro G 1980 Transforming growth factors: Isolation of polypeptides from virally and chemically transformed cells by acid/ethanol extraction. Proc Natl Acad Sci USA 77: 3594-3498. 14. Sporn M, Roberts A, Wakefield L, Assoian R 1986 Transforming growth factor$: Biological function and chemical structure. Science 233:532-534. 15. Massague J , Like B 1985 Cellular receptors for type 0 transforming growth factor. J Biol Chem 260:2636-2645. 16. Cheifetz S, Like B, Massague J 1986 Cellular distribution of type 1 and type 11 receptors for transforming growth factor8. J Biol Chem 261:9972-9978. 17. Cheifetz S, Weatherbee J , Tsang M, Anderson J, Mole J , Lucas R, Massague J 1987 The tramforming growth factor-B system, a complex pattern of cross-reactive ligands and receptors. Cell 48:409-415. 18. Seyedin S, Thompson A, Bentz H, Rosen D, McPherson J , Conti A, Siege1 N, Galluppi G , Piez K 1986 Cartilage inducing factor-A. Apparent identity to transforming growth factor-beta. J Biol Chem 261:5693-5695. 19. Knabbe C , Lippman M, Wakefield L, Flanders K, Kasid A, Derynck R, Dickson RB 1987 Evidence that transforming growth factor$ is a hormonally regulated negative growth factor in human breast cancer cells. Cell 48:417-428. 20. Takehara K , LeRoy E, Grotendorst G 1987 TGF-beta inhibition of endothelial cell proliferation: Alteration of EGF binding and EGF-induced growth-regulatory gene expression. Cell 49:415-422. 21. Holley R, Bohlen P, Fava R, Baldwin J , Kleeman G, Armour R 1980 Purification of kidney epithelial cell growth inhibitors. Proc Natl Acad Sci USA 77:5989-5992. 22. Tucker R , Shipley G , Moses H 1984 Growth inhibitor from BSC-I cells closely related to platelet type beta transforming growth factor. Science 226:705-707. 23. Roberts A, Anzano M, Wakefield L, Roche N, Stern D, Sporn M 1986 Type fi transforming growth factor: A bifunctional regulator of cellular growth. Proc Natl Acad Sci USA 82:119-123. 24. O’Keefe R, Puzas E, Brand J , Rosier R 1988 Effects of transforming growth factor-beta on the prolif eration of growth plate chondrocytes. Calcif Tissue Int 43:352-358. 25. Rizzino A, Ruff E, Rizzino H 1986 Induction and modulation of anchorage-independent growth by platelet-derived growth factor, fibroblast growth factor, and transforming growth factor-beta. Cancer Res 46:2816-2820. 26. Massague J , Kelly B, Mottola C 1985 Stimulation by insulinlike growth factors is required for cellular transformation by

27.

28.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

type beta transforming growth factor. J Biol Chem 260: 4551-4554. Brinckerhoff C 1983 Morphologic and mitogenic responses of rabbit synovial fibroblasts to transforming growth factor 9 , require transforming growth factor alpha o r epidermal growth factor. Arthritis Rheum 261370-1379. Pfeilschifter J , Mundy G 1987 Modulation of type p transforming growth factor activity in bone cultures by osteotropic hormones. Proc Natl Acad Sci USA 84:2024-2028. Brinckerhoff C 1983 Morphologic and mitogenic responses of rabbit synovial fibroblasts to transforming growth factor 0 require transforming growth factor alpha or epidermal growth factor. Arthritis Rheum 26:1370-1379. Pfeilschifter J, Mundy G 1987 Modulation of type 0 transforming growth factor activity in bone cultures by osteotropic hormones. Proc Natl Acad Sci USA 84:2024-2038. Northington F, Hamill R, Banergee S 1985 Dopamine-stimulated adenylate cyclase and tyrosine hydroxylase in diabetic rat retina. Brain Res 337:151-159. O’Keefe R, Puzas E, Brand J , Rosier R 1988 Effects of transforming growth factor-beta on matrix synthesis by chick growth plate chondrocytes. Endocrinology 13:2953-2961. Roberts A, Anzano M, Lamb L, Smith J, Sporn M 1981 New class of transforming growth factors potentiated by epidermal growth factor: Isolation from non-neoplastic tissues. Proc Natl Acad Sci USA 78:5339-5343. Hauschka P , Mavrakas A, lafrati M, Doleman S, Klagsbrun N 1986 Growth factors in bone matrix. Isolation of multiple types by affinity chromatograpy on heparin-sepharose. J Biol Chem 261:12665-12674. Young MF, Robey PG, Reddi AH, Roberts AB, Sporn MB, Termine J D 1986 TGF-,13 expression in fetal bovine bone forming cells. J Bone Min Res 1:155. Leof E, Proper J , Moses H 1987 Modulation of transforming growth factor type 0 action by activated rus and c-myc. Mol Cell Biol 7:2649-2652. Muller R, Bravo R, Burckhardt J 1984 Induction of c-fos gene production by growth factors precedes activation of c-myc. Nature 312:716-720. Greenberg M, Ziff E 1984 Stimulation of 3T3 cells induces transcription of the c-fos proto-oncogene. Nature 311:433438.

Address reprint requests to: Ian D. Crabb Department of Orthopaedics, Box 665 University of Rochester Rochester, N Y 14642 Received for publication March 5 , 1990; in revised form June 7, 1990; acdepted June 8, 1990.