Control of protein and matrix-molecule synthesis in isolated ovine fetal growth-plate chondrocytes by the interactions of basic fibroblast growth factor, insulin-like growth factors-I and -II, insulin and transforming growth factor-\g=b\1 D. J. Hill, A. Logan, M. McGarry and D. De Sousa MRC Group in Fetal and Neonatal Health and Development, Lawson Research Institute, St Joseph's Health Centre, 268 Grosvenor Street, London, Ontario N6A 4V2, Canada and Departments of Physiology and Medicine, University of Western Ontario, London, Ontario n6a sci, Canada *Department of Clinical Chemistry, University of Birmingham, Queen Elizabeth Medical Centre,
Birmingham bi5 2th, U.K.
received
4 November 1991
ABSTRACT
Chondrogenesis is thought to be controlled by interactions between circulating anabolic hormones and locally produced peptide growth factors, and involves ordered changes in matrix composition which ultimately allow endochondral calcification. We have used a model of isolated ovine fetal growth-plate chondrocytes to examine the actions and interactions of basic fibroblast growth factor (basic FGF), insulin\x=req-\ like growth factors-I and -II (IGF-I and -II), insulin and transforming growth factor-\g=b\1 (TGF-\g=b\1) on total protein, collagen or non-collagenous protein and sulphated glycosaminoglycan synthesis. These parameters were determined by assessment of the incorporation by monolayer cultures of early passage chondrocytes of [3H]leucine, [14C]proline and [35S]sulphate respectively, followed by partial molecular
and insulin each caused significant increases in the synthesis of collagen and sulphated glycosaminoglycans. TGF-\g=b\1 had no effect on total protein synthesis by chondrocytes when present alone at concentrations of 200 pmol/l or less, but was inhibitory at 400 pmol/l. However, the use of this parameter masked a stimulatory action of 50 or 100 pmol TGF-\g=b\1on sulphated glycosaminoglycan synthesis and a relative shift in the ratio of collagen : non-collagenous protein synthesis in favour of the former. A synergistic interaction existed between TGF-\g=b\1 (20\p=n-\100pmol/l) and basic FGF which potentiated total protein and collagen synthesis, and their actions on sulphated glycosaminoglycan production were additive. The same concentrations of TGF-\g=b\1inhibited the ability of IGF-I or insulin to stimulate total protein or collagen synthesis, but were additive to their stimulatory effects on sulphated glycosaminoglycan synthesis. The results suggest that matrix-molecule composition and the anabolic status of the epiphyseal growth-plate may be modulated in utero by multiple interactions between peptide growth factors produced locally, such as basic FGF, IGF-II and TGF-\g=b\1,and circulating hormones such as insulin and IGF-I. Journal of Endocrinology (1992) 133, 363\p=n-\373
INTRODUCTION
is the synthesis of a defined cartilage matrix. This con¬ sists predominantly of type-II collagen and a variety of sulphated mucopolysaccharides, the most abun¬ dant of which are chondroitin and keratin sulphates, and hyaluronic acid (Campo & Tourtellotte, 1967). During terminal differentiation of chondrocytes to
characterization. Basic FGF enhanced total protein synthesis with a half-maximal effective concentration of 270 \m=+-\60 pmol/l (mean \m=+-\s.e.m., four animals) and was sixfold more active on a molar basis than IGF-I or insulin, and 28-fold more active that IGF-II which is the endogenously synthesized IGF. The actions of basic FGF were additive to those of IGF-I or insulin. More detailed analysis of extracellular-matrix component synthesis showed that basic FGF, IGF-I
The ordered proliferation and maturation of chondro¬ cytes within epiphyseal growth-plates is necessary for progressive longitudinal skeletal growth in the fetus and child. Central to the maturation of chondrocytes
hypertrophie postmitotic cells, changes in matrix composition occur, the synthesis of sulphated muco¬ polysaccharides being reduced while that of collagen is increased, including the site-specific type-X collagen (Wuthier, 1982; Hunziker, Schenk & Cruz-Orive, 1987; Leboy, Shapiro, Uschmann et al. 1988). These changes may be necessary for subsequent calcification
and new bone formation. A number of peptide growth factors have been shown to stimulate DNA synthesis and replication by isolated epiphyseal growth-plate chondrocytes, including basic fibroblast growth factor (basic FGF), insulin-like growth factors-I and -II (IGF-I and -II), insulin and transforming growth factor-ß (TGF-ß) (O'Keefe, Ruzas, Brand & Rosier, 1988; Makower, Wroblewski & Pawlowski, 1989; Trippel, Corvol, Dumontier et al. 1989; Hill & De Sousa, 1990; Kato & Iwamoto, 1990). However, peptide growth factors have also been shown to influence chondrocyte extracellular-matrix synthesis. One of the first biologi¬ cal properties of IGFs to be recognized was their
ability to promote sulphated-mucopolysaccharide synthesis by articular cartilage (Froesch, Zapf, Audhya et al. 1976). IGF-I not only promoted sulphation of glycosyl residues of proteoglycans, which represents the final stage in their biosynthesis, but enhanced the synthesis of the core proteins and caused an elongation of existing chondroitin sulphate chains (Kilgore, McNatt, Meador et al. 1979; Silbergeld, Mamet, Laron & Nevo, 1981; Kemp, Mutchnick & Hintz, 1984). Articular chondrocytes cultured in the
absence of basic FGF became spindle shaped and lost their ability to synthesize type-II collagen within a few passages. However, the presence of basic FGF in the medium delayed the loss of chondrocytic phenotype and maintained sulphated-mucopolysaccharide pro¬ duction (Kato & Gospodarowicz, 1985). Conversely, under certain conditions, basic FGF caused a sup¬ pression of type-II collagen synthesis, as shown by Horton, Higginbotham & Chandrasekhar (1989) using embryonic-chick sternal chondrocytes. This discrepancy in the actions of basic FGF may be explained in part by the presence or absence of endogenous TGF-ß, since a synergy has been described between TGF-ß and basic FGF in both chick embryo and immature rabbit growth-plate chondrocytes, resulting in an inhibition of collagen synthesis and a potentiation of sulphated proteoglycan production (O'Keefe et al. 1988; Inoue, Kato, Iwamoto et al.
1989).
recently showed, using a model of isolated epi¬ physeal growth-plate chondrocytes from the ovine fetus, synthesis, release and mitogenic activity of both IGF-II and basic FGF (Hill & De Sousa, 1990; Hill, Logan, Ong et al. 1992). A further growth factor likely to be available to the growth-plate of the ovine fetus is We
TGF-ß,, since mRNA encoding this peptide is expressed in bone cells of the human fetus, adjacent to the region of hypertrophie cartilage (Sandberg, Vuorio, Hirvonen et al. 1988). Isolated chondrocytes have also been reported to synthesize TGF-ß (Rosier, O'Keefe, Crabb & Puzas, 1989). It therefore seems probable that the defined changes in growth-plate cartilage matrix composition which accompany chondrocyte proliferation and maturation are heavily dependent on the presence and actions of several endogenous peptide growth factors. In this study we
have looked at the actions and interactions of basic FGF, IGF-I and -II, insulin and TGF-ß, upon total protein, collagen and sulphated glycosaminoglycan synthesis by isolated ovine fetal growth-plate chondrocytes. This has allowed an assessment of the relative functional importance of each factor in a mammalian system with respect to both biological potency and likely endogenous availability. MATERIALS AND METHODS
Human
TGF-ß,
was
purchased
from R and D
Systems Inc., Minneapolis, MN, U.S.A., recombinant human IGF-I from IMCERA Bioproducts Inc., Terre Haute, IN, U.S.A., recombinant human IGF-II from Bachern Inc., Torrance, CA, U.S.A. and porcine insu¬ lin from Sigma Chemical Co., St Louis, MO, U.S.A.
Recombinant human basic FGF was kindly provided Dr P. Barr, Chiron Corporation, Emeryville, CA, U.S.A. Plastic culture flasks (75 cm2 area, NUNC), six-well culture plates (9 cm2 area per well), Hank's buffered salts solution, Dulbecco's modified Eagle's medium (DMEM), fetal calf serum (FCS) and trypsin-EDTA were each purchased from Gibco Ltd, Burlington, Ontario, Canada. A glucose-free formulation of Dulbecco's phosphate-buffered saline
by
(PBS), chromatographically pure collagenase (typeIll), chondroitin sulphate and papain were from Sigma. Collagenase for tissue dissociation and Trasylol (aprotinin) were purchased from BoehringerIngelheim (Canada) Ltd, Dorval, Quebec, Canada, Hoechst fluorochrome 33258 from Aldrich Chemical
Co., Milwaukee, MI, U.S.A., Euthanyl from MCT Pharmaceuticals, Mississauga, Ontario, Canada
and [3H]leucine, [14C]proline and [35S]sulphate from Amersham International, Mississauga, Ontario, Canada. Centricon microconcentrators were obtained from Amicon, Danvers, MA, U.S.A. Tissue source and
chondrocyte isolation Fifteen pregnant sheep of mixed breed
with single known insemination dates were used in this study. Animals were killed with sodium pentabarbitone (Euthanyl) on days 70-85 of gestation (term 145
a laparotomy was performed, and the fetuses delivered. Epiphyseal growth-plate cartilage was dissected from the proximal tibia and chondrocytes were isolated by incubation with 0-2% (w/v) col¬ lagenase as described previously (Hill & De Sousa, 1990). Isolated chondrocytes were plated into 75 cm2 plastic culture flasks in DMEM supplemented with 10% (v/v) FCS, penicillin (100pU/ml), streptomycin (100 pg/ml) and fungizone (2-5 pg/ml), and allowed to grow to confluency over 2-3 days. Cultures were maintained at 37 °C in a humidified atmosphere of 5% C02-95% air. Chondrocytes were passaged with trypsin-EDTA three times before use in experiments between passages 3 and 6. We have previously shown a continued production of type-II collagen, indicating a chondrocytic phenotype, over this period (Hill & De
days), were
Sousa, 1990).
Estimation of total
protein synthesis plated into six-well culture dishes
Chondrocytes density of 20 000 cells per well in 1 ml DMEM + 10% (v/v) FCS and grown for 3-4 days until approxi¬ mately 80% confluent. Cells were then growth restricted by incubation in 1 ml DMEM + 0-1 % (v/v) FCS before the medium was replaced with 1 ml glucose-free DMEM supplemented with 2-7 mmol glucose/1 with or without 002-42 nmol insulin/1,01320 nmol IGF-I/1 or IGF-II/1, 20-100 pmol TGF-ß,/l or 6-1100 pmol basic FGF/1. After a 20 h culture, 5 pCi [3H]leucine was added in 20 pi medium to each were
at a
well and the incubation continued for a further 4 h. Cells were washed with PBS (pH 7-4) to remove unin¬ corporated isotope, and the protein and DNA were precipitated with 1 ml ice-cold trichloroacetic acid (10%, w/v) before solubilization overnight at 37 °C in 0-1 mol NaOH/1 (500 pi per well). Isotope incorpor¬ ation into newly synthesized protein was estimated by liquid-scintillation counting and expressed as d.p.m./ pg cell DNA. The DNA content of each culture well was measured by fluorimetry using Hoechst fluorochrome 33258 (Hill & De Sousa, 1990). A glucose con¬ centration of 2-7 mmol/1 was previously found to be optimal for insulin-stimulated DNA synthesis by ovine fetal growth-plate chondrocytes (Hill & De Sousa, 1990). Certain studies utilizing basic FGF and IGFs were also performed in media containing 8-7 mmol glucose/1.
Estimation of collagen and
synthesis
non-collagenous protein
Cells were growth restricted and medium was replaced with 1 ml glucose-free DMEM supplemented with 2-7 mmol glucose/1 and added growth factors as described above. After the 20-h incubation, the medium was removed and replaced with 500 pi
glucose-free DMEM supplemented with 2-7 mmol glucose/1 and 3 pCi [14C]proline/ml. After a further 4 h incubation, the plates were placed on ice and the cells scraped from the well surface into the culture medium. The isotope incorporated into new protein within and excreted by the cells was measured and separated into collagen and non-collagenous protein (NCP) as described previously by Peterkofsky & Diegelmann (1971). Briefly, proteins were precipitated by the addition of 1 ml 10% (w/v) trichloroacetic acid con¬ taining 0-5% (w/v) tannic acid per well for 30 min at 4°C. The precipitated proteins were collected by centrifugation at 2500 g for 15 min and the pellet was washed twice more with trichloroacetic acid/tannic
acid and once with ice-cold acetone. Proteins were then solubilized in 1 ml 005 mol NaOH/1 for 3 h in a water bath at 60 °C. An aliquot of solubilized protein was used to estimate the [l4C]proline incorporated into total new protein by liquid-scintillation counting, and another to estimate DNA content. The remainder was neutralized with 0-2 ml Tris-HCl buffer (0-2 mol/ 1; pH 7-6) containing 10 mmol CaCl2/l and 20 mmol N-ethylmaleimide/1 before the addition of chromatographically pure collagenase (20 pg/ml). The samples were incubated at 37 °C for 3 h, 1ml 10% (w/v) trichloroacetic acid was then added and the nondigested proteins were precipitated by incubation overnight at 4 °C before collection by centrifugation. After two further washes in trichloroacetic acid, the pellet was solubilized in 0-05 mol NaOH/1 and an aliquot used for liquid-scintillation counting. The iso¬ tope incorporated in this fraction was taken to rep¬ resent that contained in NCP. This amount was subtracted from the total protein incorporation to obtain an estimate of the [14C]proline which had been incorporated into collagen. Titration of the collage¬ nase batch used showed that concentrations in excess of 20 pg/ml resulted in no further destruction of cellderived collagen. Results are expressed both as d.p.m. [l4C]proline/pg cell DNA and as a ratio of d.p.m. collagen to d.p.m. associated with NCP. In the latter case the fraction associated with NCP was increased by a factor of 5-4 to correct for the relative amount of proline in collagen compared with other proteins
(Peterkofsky, 1972).
Estimation of sulphated-glycosaminoglycan
synthesis
growth restricted and the culture Chondrocytes medium was replaced with 1 ml glucose-free DMEM supplemented with 2-7 mmol glucose/1, added growth factors as described above, and 5 pCi [35S]sulphate/ml. After a 20-h incubation, the medium was removed, carrier chondroitin sulphate added (2mg/ml), and cells were stored at —70 °C. Cells were washed three times with PBS (pH 7-4) and 0-5 ml PBS containing were
aprotinin (Trasylol; 120k.i.u. per well), and carrier chondroitin sulphate (1 mg) was added to each well. Cells were scraped free from the culture well and the suspension was dispersed by sonication. Sulphated glycosaminoglycans labelled with [35S]sulphate were isolated and quantitated as described by Dorfman & Ho (1970) and modified by Levitt, Ho & Dorfman (1975). Briefly, dispersed cell solutions and conditioned culture media were each digested with papain (10 mg/ml) in 01 mol acetate buffer/1 (pH 5-5) containing 0005mol EDTA/1 and 0-005mol cysteine hydrochloride/1 for 36 h at 65 °C. Proteins were pre¬ cipitated with 10% (v/v final concentration) trichloro¬
acetic acid and the soluble material was desalted on Centricon microconcentrators (molecular mass cut-off 10 kDa) and washed with 003 mol NaCl/1. Sulphated glycosaminoglycans were precipitated by the addition of 5 volumes of cetylpyridinium chloride (10%, w/v) in 003 mol NaCl/1 and the precipitate was solubilized in 2 mol NaCl/1. Glycosaminoglycans were reprecipitated with 80% (v/v) ethanol followed by 100% (v/v) ethanol, air-dried, resuspended in 0-5 ml distilled water and the incorporated [35S]sulphate was measured by liquid-scintillation counting. Counts derived from the cell layer and culture medium from each well were combined and the results expressed as d.p.m. per pg cell DNA. The DNA content of each well was estimated from the material precipitated with trichloroacetic acid, which was solubilized with
Similar increases in protein synthesis above control values were observed after the incu¬ bation of cells with IGF-I, IGF-II or insulin, their ED50 values being 1-51+0-11 nmol/l (1-26-1-75), 7-43 ± 0-36 nmol/l (6-41 + 7-88) and 1 -64 + 0-09 nmol/l (1-34-1-85) (n 4) respectively. The ability of insulin to increase protein synthesis was biphasic, an initial response achieving a plateau in the presence of 4 nmol insulin/1 with a further rise in protein synthesis seen in the presence of 84 nmol insulin/1, a pharmacological concentration. Basic FGF was therefore sixfold more potent than IGF-I or insulin, and 28-fold more potent than IGF-II when inducing new protein synthesis by isolated chondrocytes. IGF-I was fivefold more potent than IGF-II. TGF-ß, had little effect on new protein formation at concentrations between 5 and 200 pmol/1, but a significant reduction of approxi¬ mately 25% compared to control values was seen in the presence of 400 pmol TGF-ß,/ (Fig. 1). Exper¬ iments were repeated with culture medium containing 8-7 mmol glucose/1. The ability of IGF-I, IGF-II and basic FGF to increase [3H]leucine incorporation was similar if the medium contained 2-7 mmol/1 or 8-7 mmol/1 glucose. The ability of insulin (5 nmol/l) to potentiate [ H]leucine incorporation was reduced by approximately 30% when medium contained 8-7 mmol glucose/1 compared with 2-7 mmol/1 (data not shown).
experiments).
=
OlmolNaOH/1. Statistical
analysis Experiments were repeated between three and six times. For analysis of the protein synthetic rate, six replicate incubations were included for each exper¬ imental variable. Results are expressed as mean values ± s.e.m. for replicate incubations within a single experiment. Differences between mean values were assessed by analysis of variance (ANOVA). RESULTS 0 0005 0-02
ability of basic FGF, IGF-I, IGF-II, insulin and to potentiate new protein synthesis by cultures of isolated growth-plate chondrocytes was examined by the incorporation of [3H]leucine. The concentration of glucose in the culture medium was 2-7 mmol/1. Increasing concentrations of basic FGF caused a dose-related increase in protein synthetic The
TGF-ß,,
rate, with a maximal effective concentration of 550 pmol/1 (Fig. 1). This amount of basic FGF caused a threefold rise in protein synthetic rate compared with that seen in control cultures, and the halfmaximal effective concentration (ED50) of basic FGF was 270 + 60 pmol/1 (range: 210-355; four separate
008
0-32
1-28
Concentration
512
20
(nmol/l)
1. Incorporation of [3H]leucine into the newly syn¬ thesized proteins of growth-plate chondrocytes incubated with increasing concentrations of basic fibroblast growth factor (basic FGF, ·), insulin-like growth factor-I (IGF-I, A ), IGF-II ( ), insulin ( O ) or transforming growth factorß, (TGF-ß,, ). Figures represent mean values ± s.e.m. for six replicate incubations. The minimum concentrations at which isotope incorporation was significantly elevated when compared with control incubations were: for basic FGF, 0-28 nmol/l; for IGF-I, 1 -3 nmol/l; for IGF-II, 6-6 nmol/l and for insulin, 1 -7 nmol/l. A significant decrease in isotope incorporation compared with controls was seen with 0-4 nmol TGF-ß,/ (P< 005 or greater) (ANOVA). figure
Although concentrations of TGF-ß, of 100 pmol/1 less had no effect on protein synthesis by chondro¬ cytes when present alone, a small but statistically sig¬ nificant synergy was seen in the presence of 60 pmol/1 or 300 pmol/1 basic FGF, resulting in an 18-35% increase in the protein synthetic rate compared with or
that seen with basic FGF alone (Fig. 2). At submaxi¬ mal concentrations, the effects of IGF-I and basic FGF on protein synthesis were additive (Fig. 3). Similar interactions were seen with basic FGF and insulin (data not shown). However, TGF-ß, at con¬ centrations of 20 pmol/1 or 100 pmol/1 substantially reduced the ability of a physiological concentration of insulin (1 -67 nmol/l, Fig. 4) or IGF-I (6-6 nmol/l, data not shown) to induce protein synthesis. No dosedependency of TGF-ß, was seen at the concentrations used.
I
.2
Birmingham bi5 2th, U.K.
received
4 November 1991
ABSTRACT
Chondrogenesis is thought to be controlled by interactions between circulating anabolic hormones and locally produced peptide growth factors, and involves ordered changes in matrix composition which ultimately allow endochondral calcification. We have used a model of isolated ovine fetal growth-plate chondrocytes to examine the actions and interactions of basic fibroblast growth factor (basic FGF), insulin\x=req-\ like growth factors-I and -II (IGF-I and -II), insulin and transforming growth factor-\g=b\1 (TGF-\g=b\1) on total protein, collagen or non-collagenous protein and sulphated glycosaminoglycan synthesis. These parameters were determined by assessment of the incorporation by monolayer cultures of early passage chondrocytes of [3H]leucine, [14C]proline and [35S]sulphate respectively, followed by partial molecular
and insulin each caused significant increases in the synthesis of collagen and sulphated glycosaminoglycans. TGF-\g=b\1 had no effect on total protein synthesis by chondrocytes when present alone at concentrations of 200 pmol/l or less, but was inhibitory at 400 pmol/l. However, the use of this parameter masked a stimulatory action of 50 or 100 pmol TGF-\g=b\1on sulphated glycosaminoglycan synthesis and a relative shift in the ratio of collagen : non-collagenous protein synthesis in favour of the former. A synergistic interaction existed between TGF-\g=b\1 (20\p=n-\100pmol/l) and basic FGF which potentiated total protein and collagen synthesis, and their actions on sulphated glycosaminoglycan production were additive. The same concentrations of TGF-\g=b\1inhibited the ability of IGF-I or insulin to stimulate total protein or collagen synthesis, but were additive to their stimulatory effects on sulphated glycosaminoglycan synthesis. The results suggest that matrix-molecule composition and the anabolic status of the epiphyseal growth-plate may be modulated in utero by multiple interactions between peptide growth factors produced locally, such as basic FGF, IGF-II and TGF-\g=b\1,and circulating hormones such as insulin and IGF-I. Journal of Endocrinology (1992) 133, 363\p=n-\373
INTRODUCTION
is the synthesis of a defined cartilage matrix. This con¬ sists predominantly of type-II collagen and a variety of sulphated mucopolysaccharides, the most abun¬ dant of which are chondroitin and keratin sulphates, and hyaluronic acid (Campo & Tourtellotte, 1967). During terminal differentiation of chondrocytes to
characterization. Basic FGF enhanced total protein synthesis with a half-maximal effective concentration of 270 \m=+-\60 pmol/l (mean \m=+-\s.e.m., four animals) and was sixfold more active on a molar basis than IGF-I or insulin, and 28-fold more active that IGF-II which is the endogenously synthesized IGF. The actions of basic FGF were additive to those of IGF-I or insulin. More detailed analysis of extracellular-matrix component synthesis showed that basic FGF, IGF-I
The ordered proliferation and maturation of chondro¬ cytes within epiphyseal growth-plates is necessary for progressive longitudinal skeletal growth in the fetus and child. Central to the maturation of chondrocytes
hypertrophie postmitotic cells, changes in matrix composition occur, the synthesis of sulphated muco¬ polysaccharides being reduced while that of collagen is increased, including the site-specific type-X collagen (Wuthier, 1982; Hunziker, Schenk & Cruz-Orive, 1987; Leboy, Shapiro, Uschmann et al. 1988). These changes may be necessary for subsequent calcification
and new bone formation. A number of peptide growth factors have been shown to stimulate DNA synthesis and replication by isolated epiphyseal growth-plate chondrocytes, including basic fibroblast growth factor (basic FGF), insulin-like growth factors-I and -II (IGF-I and -II), insulin and transforming growth factor-ß (TGF-ß) (O'Keefe, Ruzas, Brand & Rosier, 1988; Makower, Wroblewski & Pawlowski, 1989; Trippel, Corvol, Dumontier et al. 1989; Hill & De Sousa, 1990; Kato & Iwamoto, 1990). However, peptide growth factors have also been shown to influence chondrocyte extracellular-matrix synthesis. One of the first biologi¬ cal properties of IGFs to be recognized was their
ability to promote sulphated-mucopolysaccharide synthesis by articular cartilage (Froesch, Zapf, Audhya et al. 1976). IGF-I not only promoted sulphation of glycosyl residues of proteoglycans, which represents the final stage in their biosynthesis, but enhanced the synthesis of the core proteins and caused an elongation of existing chondroitin sulphate chains (Kilgore, McNatt, Meador et al. 1979; Silbergeld, Mamet, Laron & Nevo, 1981; Kemp, Mutchnick & Hintz, 1984). Articular chondrocytes cultured in the
absence of basic FGF became spindle shaped and lost their ability to synthesize type-II collagen within a few passages. However, the presence of basic FGF in the medium delayed the loss of chondrocytic phenotype and maintained sulphated-mucopolysaccharide pro¬ duction (Kato & Gospodarowicz, 1985). Conversely, under certain conditions, basic FGF caused a sup¬ pression of type-II collagen synthesis, as shown by Horton, Higginbotham & Chandrasekhar (1989) using embryonic-chick sternal chondrocytes. This discrepancy in the actions of basic FGF may be explained in part by the presence or absence of endogenous TGF-ß, since a synergy has been described between TGF-ß and basic FGF in both chick embryo and immature rabbit growth-plate chondrocytes, resulting in an inhibition of collagen synthesis and a potentiation of sulphated proteoglycan production (O'Keefe et al. 1988; Inoue, Kato, Iwamoto et al.
1989).
recently showed, using a model of isolated epi¬ physeal growth-plate chondrocytes from the ovine fetus, synthesis, release and mitogenic activity of both IGF-II and basic FGF (Hill & De Sousa, 1990; Hill, Logan, Ong et al. 1992). A further growth factor likely to be available to the growth-plate of the ovine fetus is We
TGF-ß,, since mRNA encoding this peptide is expressed in bone cells of the human fetus, adjacent to the region of hypertrophie cartilage (Sandberg, Vuorio, Hirvonen et al. 1988). Isolated chondrocytes have also been reported to synthesize TGF-ß (Rosier, O'Keefe, Crabb & Puzas, 1989). It therefore seems probable that the defined changes in growth-plate cartilage matrix composition which accompany chondrocyte proliferation and maturation are heavily dependent on the presence and actions of several endogenous peptide growth factors. In this study we
have looked at the actions and interactions of basic FGF, IGF-I and -II, insulin and TGF-ß, upon total protein, collagen and sulphated glycosaminoglycan synthesis by isolated ovine fetal growth-plate chondrocytes. This has allowed an assessment of the relative functional importance of each factor in a mammalian system with respect to both biological potency and likely endogenous availability. MATERIALS AND METHODS
Human
TGF-ß,
was
purchased
from R and D
Systems Inc., Minneapolis, MN, U.S.A., recombinant human IGF-I from IMCERA Bioproducts Inc., Terre Haute, IN, U.S.A., recombinant human IGF-II from Bachern Inc., Torrance, CA, U.S.A. and porcine insu¬ lin from Sigma Chemical Co., St Louis, MO, U.S.A.
Recombinant human basic FGF was kindly provided Dr P. Barr, Chiron Corporation, Emeryville, CA, U.S.A. Plastic culture flasks (75 cm2 area, NUNC), six-well culture plates (9 cm2 area per well), Hank's buffered salts solution, Dulbecco's modified Eagle's medium (DMEM), fetal calf serum (FCS) and trypsin-EDTA were each purchased from Gibco Ltd, Burlington, Ontario, Canada. A glucose-free formulation of Dulbecco's phosphate-buffered saline
by
(PBS), chromatographically pure collagenase (typeIll), chondroitin sulphate and papain were from Sigma. Collagenase for tissue dissociation and Trasylol (aprotinin) were purchased from BoehringerIngelheim (Canada) Ltd, Dorval, Quebec, Canada, Hoechst fluorochrome 33258 from Aldrich Chemical
Co., Milwaukee, MI, U.S.A., Euthanyl from MCT Pharmaceuticals, Mississauga, Ontario, Canada
and [3H]leucine, [14C]proline and [35S]sulphate from Amersham International, Mississauga, Ontario, Canada. Centricon microconcentrators were obtained from Amicon, Danvers, MA, U.S.A. Tissue source and
chondrocyte isolation Fifteen pregnant sheep of mixed breed
with single known insemination dates were used in this study. Animals were killed with sodium pentabarbitone (Euthanyl) on days 70-85 of gestation (term 145
a laparotomy was performed, and the fetuses delivered. Epiphyseal growth-plate cartilage was dissected from the proximal tibia and chondrocytes were isolated by incubation with 0-2% (w/v) col¬ lagenase as described previously (Hill & De Sousa, 1990). Isolated chondrocytes were plated into 75 cm2 plastic culture flasks in DMEM supplemented with 10% (v/v) FCS, penicillin (100pU/ml), streptomycin (100 pg/ml) and fungizone (2-5 pg/ml), and allowed to grow to confluency over 2-3 days. Cultures were maintained at 37 °C in a humidified atmosphere of 5% C02-95% air. Chondrocytes were passaged with trypsin-EDTA three times before use in experiments between passages 3 and 6. We have previously shown a continued production of type-II collagen, indicating a chondrocytic phenotype, over this period (Hill & De
days), were
Sousa, 1990).
Estimation of total
protein synthesis plated into six-well culture dishes
Chondrocytes density of 20 000 cells per well in 1 ml DMEM + 10% (v/v) FCS and grown for 3-4 days until approxi¬ mately 80% confluent. Cells were then growth restricted by incubation in 1 ml DMEM + 0-1 % (v/v) FCS before the medium was replaced with 1 ml glucose-free DMEM supplemented with 2-7 mmol glucose/1 with or without 002-42 nmol insulin/1,01320 nmol IGF-I/1 or IGF-II/1, 20-100 pmol TGF-ß,/l or 6-1100 pmol basic FGF/1. After a 20 h culture, 5 pCi [3H]leucine was added in 20 pi medium to each were
at a
well and the incubation continued for a further 4 h. Cells were washed with PBS (pH 7-4) to remove unin¬ corporated isotope, and the protein and DNA were precipitated with 1 ml ice-cold trichloroacetic acid (10%, w/v) before solubilization overnight at 37 °C in 0-1 mol NaOH/1 (500 pi per well). Isotope incorpor¬ ation into newly synthesized protein was estimated by liquid-scintillation counting and expressed as d.p.m./ pg cell DNA. The DNA content of each culture well was measured by fluorimetry using Hoechst fluorochrome 33258 (Hill & De Sousa, 1990). A glucose con¬ centration of 2-7 mmol/1 was previously found to be optimal for insulin-stimulated DNA synthesis by ovine fetal growth-plate chondrocytes (Hill & De Sousa, 1990). Certain studies utilizing basic FGF and IGFs were also performed in media containing 8-7 mmol glucose/1.
Estimation of collagen and
synthesis
non-collagenous protein
Cells were growth restricted and medium was replaced with 1 ml glucose-free DMEM supplemented with 2-7 mmol glucose/1 and added growth factors as described above. After the 20-h incubation, the medium was removed and replaced with 500 pi
glucose-free DMEM supplemented with 2-7 mmol glucose/1 and 3 pCi [14C]proline/ml. After a further 4 h incubation, the plates were placed on ice and the cells scraped from the well surface into the culture medium. The isotope incorporated into new protein within and excreted by the cells was measured and separated into collagen and non-collagenous protein (NCP) as described previously by Peterkofsky & Diegelmann (1971). Briefly, proteins were precipitated by the addition of 1 ml 10% (w/v) trichloroacetic acid con¬ taining 0-5% (w/v) tannic acid per well for 30 min at 4°C. The precipitated proteins were collected by centrifugation at 2500 g for 15 min and the pellet was washed twice more with trichloroacetic acid/tannic
acid and once with ice-cold acetone. Proteins were then solubilized in 1 ml 005 mol NaOH/1 for 3 h in a water bath at 60 °C. An aliquot of solubilized protein was used to estimate the [l4C]proline incorporated into total new protein by liquid-scintillation counting, and another to estimate DNA content. The remainder was neutralized with 0-2 ml Tris-HCl buffer (0-2 mol/ 1; pH 7-6) containing 10 mmol CaCl2/l and 20 mmol N-ethylmaleimide/1 before the addition of chromatographically pure collagenase (20 pg/ml). The samples were incubated at 37 °C for 3 h, 1ml 10% (w/v) trichloroacetic acid was then added and the nondigested proteins were precipitated by incubation overnight at 4 °C before collection by centrifugation. After two further washes in trichloroacetic acid, the pellet was solubilized in 0-05 mol NaOH/1 and an aliquot used for liquid-scintillation counting. The iso¬ tope incorporated in this fraction was taken to rep¬ resent that contained in NCP. This amount was subtracted from the total protein incorporation to obtain an estimate of the [14C]proline which had been incorporated into collagen. Titration of the collage¬ nase batch used showed that concentrations in excess of 20 pg/ml resulted in no further destruction of cellderived collagen. Results are expressed both as d.p.m. [l4C]proline/pg cell DNA and as a ratio of d.p.m. collagen to d.p.m. associated with NCP. In the latter case the fraction associated with NCP was increased by a factor of 5-4 to correct for the relative amount of proline in collagen compared with other proteins
(Peterkofsky, 1972).
Estimation of sulphated-glycosaminoglycan
synthesis
growth restricted and the culture Chondrocytes medium was replaced with 1 ml glucose-free DMEM supplemented with 2-7 mmol glucose/1, added growth factors as described above, and 5 pCi [35S]sulphate/ml. After a 20-h incubation, the medium was removed, carrier chondroitin sulphate added (2mg/ml), and cells were stored at —70 °C. Cells were washed three times with PBS (pH 7-4) and 0-5 ml PBS containing were
aprotinin (Trasylol; 120k.i.u. per well), and carrier chondroitin sulphate (1 mg) was added to each well. Cells were scraped free from the culture well and the suspension was dispersed by sonication. Sulphated glycosaminoglycans labelled with [35S]sulphate were isolated and quantitated as described by Dorfman & Ho (1970) and modified by Levitt, Ho & Dorfman (1975). Briefly, dispersed cell solutions and conditioned culture media were each digested with papain (10 mg/ml) in 01 mol acetate buffer/1 (pH 5-5) containing 0005mol EDTA/1 and 0-005mol cysteine hydrochloride/1 for 36 h at 65 °C. Proteins were pre¬ cipitated with 10% (v/v final concentration) trichloro¬
acetic acid and the soluble material was desalted on Centricon microconcentrators (molecular mass cut-off 10 kDa) and washed with 003 mol NaCl/1. Sulphated glycosaminoglycans were precipitated by the addition of 5 volumes of cetylpyridinium chloride (10%, w/v) in 003 mol NaCl/1 and the precipitate was solubilized in 2 mol NaCl/1. Glycosaminoglycans were reprecipitated with 80% (v/v) ethanol followed by 100% (v/v) ethanol, air-dried, resuspended in 0-5 ml distilled water and the incorporated [35S]sulphate was measured by liquid-scintillation counting. Counts derived from the cell layer and culture medium from each well were combined and the results expressed as d.p.m. per pg cell DNA. The DNA content of each well was estimated from the material precipitated with trichloroacetic acid, which was solubilized with
Similar increases in protein synthesis above control values were observed after the incu¬ bation of cells with IGF-I, IGF-II or insulin, their ED50 values being 1-51+0-11 nmol/l (1-26-1-75), 7-43 ± 0-36 nmol/l (6-41 + 7-88) and 1 -64 + 0-09 nmol/l (1-34-1-85) (n 4) respectively. The ability of insulin to increase protein synthesis was biphasic, an initial response achieving a plateau in the presence of 4 nmol insulin/1 with a further rise in protein synthesis seen in the presence of 84 nmol insulin/1, a pharmacological concentration. Basic FGF was therefore sixfold more potent than IGF-I or insulin, and 28-fold more potent than IGF-II when inducing new protein synthesis by isolated chondrocytes. IGF-I was fivefold more potent than IGF-II. TGF-ß, had little effect on new protein formation at concentrations between 5 and 200 pmol/1, but a significant reduction of approxi¬ mately 25% compared to control values was seen in the presence of 400 pmol TGF-ß,/ (Fig. 1). Exper¬ iments were repeated with culture medium containing 8-7 mmol glucose/1. The ability of IGF-I, IGF-II and basic FGF to increase [3H]leucine incorporation was similar if the medium contained 2-7 mmol/1 or 8-7 mmol/1 glucose. The ability of insulin (5 nmol/l) to potentiate [ H]leucine incorporation was reduced by approximately 30% when medium contained 8-7 mmol glucose/1 compared with 2-7 mmol/1 (data not shown).
experiments).
=
OlmolNaOH/1. Statistical
analysis Experiments were repeated between three and six times. For analysis of the protein synthetic rate, six replicate incubations were included for each exper¬ imental variable. Results are expressed as mean values ± s.e.m. for replicate incubations within a single experiment. Differences between mean values were assessed by analysis of variance (ANOVA). RESULTS 0 0005 0-02
ability of basic FGF, IGF-I, IGF-II, insulin and to potentiate new protein synthesis by cultures of isolated growth-plate chondrocytes was examined by the incorporation of [3H]leucine. The concentration of glucose in the culture medium was 2-7 mmol/1. Increasing concentrations of basic FGF caused a dose-related increase in protein synthetic The
TGF-ß,,
rate, with a maximal effective concentration of 550 pmol/1 (Fig. 1). This amount of basic FGF caused a threefold rise in protein synthetic rate compared with that seen in control cultures, and the halfmaximal effective concentration (ED50) of basic FGF was 270 + 60 pmol/1 (range: 210-355; four separate
008
0-32
1-28
Concentration
512
20
(nmol/l)
1. Incorporation of [3H]leucine into the newly syn¬ thesized proteins of growth-plate chondrocytes incubated with increasing concentrations of basic fibroblast growth factor (basic FGF, ·), insulin-like growth factor-I (IGF-I, A ), IGF-II ( ), insulin ( O ) or transforming growth factorß, (TGF-ß,, ). Figures represent mean values ± s.e.m. for six replicate incubations. The minimum concentrations at which isotope incorporation was significantly elevated when compared with control incubations were: for basic FGF, 0-28 nmol/l; for IGF-I, 1 -3 nmol/l; for IGF-II, 6-6 nmol/l and for insulin, 1 -7 nmol/l. A significant decrease in isotope incorporation compared with controls was seen with 0-4 nmol TGF-ß,/ (P< 005 or greater) (ANOVA). figure
Although concentrations of TGF-ß, of 100 pmol/1 less had no effect on protein synthesis by chondro¬ cytes when present alone, a small but statistically sig¬ nificant synergy was seen in the presence of 60 pmol/1 or 300 pmol/1 basic FGF, resulting in an 18-35% increase in the protein synthetic rate compared with or
that seen with basic FGF alone (Fig. 2). At submaxi¬ mal concentrations, the effects of IGF-I and basic FGF on protein synthesis were additive (Fig. 3). Similar interactions were seen with basic FGF and insulin (data not shown). However, TGF-ß, at con¬ centrations of 20 pmol/1 or 100 pmol/1 substantially reduced the ability of a physiological concentration of insulin (1 -67 nmol/l, Fig. 4) or IGF-I (6-6 nmol/l, data not shown) to induce protein synthesis. No dosedependency of TGF-ß, was seen at the concentrations used.
I
.2
