Biochimica et Biophysica Acta, 1093 (1991) 229-233 © 1991 Elsevier Science Publishers B.V. 0167-4889/91/$03.50 ADONIS 016748899100205B
229
BBAMCR 12955
Differential effect of transforming growth factor fl on proteoglycan synthesis in human embryonic lung fibroblasts Manuel Romarls, Ascensi6n Heredia, Anna Molist and Anna Bassols Departament de Bmqulmica i Biologia Molecular, Facultat de Veterin~ria, UniversitatAutSnoma de Barcelona, Bellaterra (Spain) (Received 23 November 1990)
Key words: Transforming growth factor; Proteoglycan synthesis; (Human embryonic lung fibroblast)
The effect of transforming growth factor ~ (TGF-/]) on the biosynthesis of individual proteoglycans (PGs) by human embryonic lung fibroblasts has been investigated using specific antibodies and cDNA probes. Human lung fibroblasts secrete the two small chondroitin/dermatan sulfate PGs, PG-I or biglycan (300 kDa) and, in a larger proportion, PG-I! or decorin (130 kDa). Metabolic labeling experiments reveal that TGF-/3 induces selectively the expression of PG-I, whereas the level of PG-II remains unaltered. The effect of TGF-/3 on PG-I and PG-II has been studied by immunoprecipitation and Northern blot analysis. Either at the core protein or mRNA level, a specific 5-fold increase in PG-I can be observed. TGF-/] acts probably at the transcriptional level, as actinomycin D blocks completely the TGF-/3 induced proteoglycan synthesis. A low saturation density and a slower growth rate is also observed for TGF-/$ treated cells. The possible role of PG-I and PG.II as mediators of the growth inhibition caused by TGF-/] is discussed.
Introduction Transforming growth factor/3 (TGF-fl) is a polypeptide growth factor which can be found in a large variety of cells and tissues. It is one of the most powerful growth inhibitory factors and acts as a differentiation agent in several cell types [1]. It can inhibit the differentiation in vitro of 3T3-L1 preadipocytes [2], myoblasts [3] and hematopoietic cells [3], but induces differentiation of chondroblasts [5] and some epithelial derived cells [6]. TGF-/3 has also been implicated in vitro in different processes such as morphogenesis, angiogenesis and wound healing [7,8]. One of the most important phenotypic effects of TGF-/3 is the ability of this factor to modulate the extracellular matrix [1]. TGF-fl is a potent inducer of several extracellular matrix components: fibronectin, several types of collagen, integrins and chondroitin/ dermatan sulfate proteoglycans. On the other hand,
Abbreviations: TGF-/3, transforming growth factor/3; PG, proteoglycan Correspondence: A. Bassols, Departament de Bioquimica i Biologia Molecular, Facultat de Veterinhria, Universitat Aut6noma de Barcelona, 08193 Bellaterra, Spain.
TGF-fl is able to induce several extracellular matrixdegrading proteinase inhibitors. The effects on the extracellular matrix are supposed to mediate, at least partially, the effects on proliferation and differentiation. Proteoglycans (PGs) are one of the main components of the extracellular matrix [9,10]. Fibroblasts produce different types of PG including chondroitin and heparan sulfate proteoglycans [11]. The most abundant PGs secreted into the medium by fibroblasts are the so-called small PGs, PG-I or biglycan and PG-II or decorin, that can also be found in the extracellular matrix from numerous tissues and cells. PG-I and PG-II have been cloned from human bone [12] and purified from bovine skin and cartilage [13,14]. The derived protein sequence of both PGs shows 55% amino acid identity, suggesting that they probably arose through a gene duplication. PG-II contains one attached glycosaminoglycan (GAG) chain, while PG-I contains two chains. Both PGs have protein cores of approx. 37 kDa that contain a series of 12 (PG-I) and 10 (PG-II) tandem repeats of leucine-rich regions that make up the major part of the molecule (80%). These regions are thought to mediate the binding of PG-I and PG-II to other extracdlular matrix components, such as collagens.
230 The effect of TGF-/3 on PG metabolism has already been described [15], but the action on individual PGs has not been completely assessed. Using specific antibodies and cDNA probes against human PG-I and PG-II, we show in the present work that TGF-/3 affects more markedly the expression of PG-I (biglycan) in human lung fibroblasts and that TGF-/3 acts at a transcriptional level.
Materials and Methods TGF-/31 from bovine bone was kindly donated by Dr. J. Massagu6 (Sloan Kettering Cancer Center, New York, U.S.A.). Chondroitinase ABC and AC were from Seikagaku Kogyo. Heparitinase was from Sigma. Products for cell culture were from Gibco. All other reagents were of analytical grade from Merck or Sigma.
Cell culture and metaboJic labeling Human embryonic lung fibroblasts (HFL1, ATCC CCL 153) were grown in RPMI 1640 medium with 10% fetal calf serum, 100 I U / m l penicillin and 100 ~tg/ml streptomycin at 37 °C in a humidified 5% CO2 atmosphere. Subconfluent cultures were placed in serum free medium for 16 h before the experiment. At the beginning of the experiment, TGF-/3 was added at the indicated times and concentrations. 8 h before the end of the incubation, the medium was replaced with low sulfate (0.2 mM) minimal essential medium containing 100 ~Ci/ml of [35S]sulfate (New England Nuclear) or with methionine-free minimal essential medium containing 25 ~tCi/ml [~SS]methionine (New England Nuclear). After labeling, the medium was collected and made up to I mM PMSF. Aliquots of the medium were analyzed by SDS-PAGE and fluorography. Enzymatic digestions with chondroitinase ABC, chondrotinase AC and heparitinase were performed as previously described [15,16].
Immunoprecipitation Polyclonal antibodies directed against the aminoterminus of PG-I (LF-51) and PG-II (LF-30) were kindly provided by Dr. L.W. Fisher, HAM., U.S.A. The antibodies did not crossreact with each other's target protein [12]. lmmunoprecipitations were performed by adding 100 ~! of antiserum to 750 Izl of [35S]methionine labeled conditioned medium treated with chondroitinase ABC. Preimmune serum was used in parallel control experiments. The samples were incubated overnight at 4 °C with mixing. To precipitate the antigen-antibody complexes, 50 /~! of protein A-Sepharose (Sigma) was added to the samples [17]. The immunocomplexes were dissolved in sample buffer and analyzed by electrophoresis.
Northern blot analysis Total RNA from control and TGF-10 treated cells was isolated as described by Chomczynski and Sacchi [18]. The RNA was fractionated by 1% agarose gel electrophoresis and transferred to Genescreen Plus filters (Dupont). The membranes were hybridized using eDNA probes to human PG-I and PG-II, under the conditions described by Fisher et al. [12]. The probes were kindly donated by Dr. L.W. Fisher (N.I.H. U.S.A.).
Growth proliferation assays Cells were plated in triplicate in 24-well plates in the presence or absence of 100 pM TGF-/3 in serumcontaining medium. Each day, half of the medium was removed and new medium with or without TGF-fl was added. In this way it was possible to maintain the higher levels of PGs in the medium of TGF-/3 treated cells. At the stated times, cells were detached with trypsin and counted.
Results
Effect of TGF.~ on lung fibroblast proliferation TGF-/3 acts as a growth inhibitor on lung fibroblasts as it does in other cell types [1]. As shown in Fig. 1, TGF-/] added to the cultures for several days markedly decreases cell growth rate as well as saturation density.
Effect of TGF-~ on PG production by lung fibroblasts Subconfluent cells were incubated with [35S]suifate to label the PGs and the conditioned medium was analyzed by electrophoresis in order to characterize the profile of secreted PGs in this cell line. In control cells, three sulfated bands were observed (Fig. 2): a high molecular weight band that does not enter into the gel, which probably corresponds to versican [19], and two other bands with average molecular masses of 300 kDa and 130 kDa. These mobilities correspond to the small chondroitin/dermatan sulfate PGs, identified in other cell types as PG-I and PG-II. In TGF-fl treated cells, a significant increase in the secretion of PG-I was observed, as shown in Fig. 2. The increase was 4.5 + 0.5-fold (mean + S.E. from five different experiments) as quantified by densitometry. In contrast, the intensity of the band corresponding to PG-II did not change. A moderate increase in the molecular weight of PG-I and PG-II was also observed, as described in other cell types [15]. Dose-response and time-course experiments indicate that the effect of TGF-fl on PG production was maximal at 50 pM and at 12 h, decreasing slowly after 24 h (results not shown). Therefore, the conditions chosen for the subsequent experiments were 100 pM TGF-/] and 16 h exposure to the factor. To characterize the GAG chains on the PGs in-
231 120
CELL3/10000
O0
pig
0 100-
kDa I PG-I
80"
205
-
O0"
PG-II 40.
116-
97-
20'
0
0
,
,
,
,
2
4
0
8
6610
12
DAYS Fig. |. Effect of TGF-/3 on human embryonic lung fibroblast proliferation. Cells were plated in triplicate in the presence or absence of
100 pM TGF-/3. Each day, half of the medium was removed and new medium with or without TGF-B was added. At the stated times, cells were detached with trypsin and counted.
duced by TGF:/3, aliquots of the [aSS]sulfate labeled conditioned medium from control and TGF-/3 treated cells were incubated with the enzymes chondroitinase AC, chondroitinase ABC and heparitinase. Chondroitinase AC causes a partial degradation of both PG-I and PG-II, whereas they were completely degraded by chondroitinase ABC, heparitinase does not have any effect on these two PGs. These results indicate that the TGF-/3-inducible sulfated-PGs belong to the chondroitin/dermatan sulfate type (not shown).
Identification of PG-I and PG-H To confirm that the high and low molecular weight sulfated bands were PG-I and PG.II, respectively, immunoprecipitation experiments using specific antibodies raised against the protein core of the two PGs were performed. Conditioned medium from cells labeled with [35S]methionine in the presence or absence of TGF-/3 was treated with chondroitinase ABC and immunoprecipitated as described in Materials and Methods. As shown in Fig. 3, the anti-PG-I antibody immunoprecipitated the 45 kDa band corresponding to the protein core of PG-I. In TGF-/~ treated cells the intensity of this band was increased about 5-fold, similar to the results obtained in the metabolic labeling experiments. When the anti-PG-II antibody was used, two bands of 44 and 47 kDa corresponding to the protein core of PG-II were observed and there was no increase in its intensity after TGF-/3 treatment.
45Fig. 2. TGF-/] effects on secreted proteoglycans in human embryonic lung fibroblasts. Subconfluent cells were labeled with [35S]sulfate and treated with TGF-/3 for 16 h. Conditioned medium was removed and analyzed by SDS-PAGE in 6% polyacrylamide gels. Molecular weight markers are shown on the left.
Effect of TGF-~ at the mRNA level T o assess w h e t h e r TGF-/3 was a c t i n g at the m R N A level, total R N A was extracted f r o m TGF-/~ t r e a t e d
and control cells and hybridized with specific eDNA probes for PG-I and PG-II.
PG-I --
PG-li +
--
m
m
4"
TGF-/3
............. ~
~' -
~:
-
205 116
-45 kDa
Fig. 3. Immunoprecipitation of PG-I and PG-II from control and TGF-/3 treated cells. Subconfluent cultures were labeled with 25 /zCi/ml of [3SS]methionine and treated with 100 pM TGF-/3 for 16 h. The conditioned medium was treated with chondroitinase ABC and immunoprecipitated with specific antibodies against PG-I and PG-II as described in Materials and Methods. PG-I and PG-II protein cores are marked with arrows.
232 PG-I 0
6
PG-I 12
0
-
28S-
-
18S-
6
I 12
H
Fig, 4. Northern blot analysisof TGF-/3 effect on PG-i and PG-II. Total RNA was extracted fromcellstreated with 100 pM TGF-/3for different times. Blots were hybridizedwith cDNA probes for PG-I (left) and PG-II(right).
As can be seen in Fig. 4, when blots were probed with the cDNA cc~:esponding to PG-I, a band of 2.6 kb was observed which correlates with the message size observed in other tissues [12]. In cells treated with TGF-/3, and increase of 5-fold in the intensity of the band was observed with a maximum at 12 h of treatment, similar to the effects observed at the protein level (Figs. 2 and 3). When blots were hybridized with the cDNA corresponding to PG-11, two bands of 1.6 and 1.9 kb were shown (Fig. 4). Two messages of similar size have been already described in IMRg0 fibroblasts [20]. There was no increase in any of the bands after treatment with
TGF-0, Actinomycin D blocked the induction of PG-I by TGF-~, showing that the factor acts at the transcriptional level. Addition of cycloheximide to cells 1 h before the labeling period, inhibited the appearance of PC~ indicating that the effect of TGF-/3 required de novo protein synthesis, presumably synthesis of the core protein (results not shown). Discussion
The extracellular matrix is a complex structure that surrounds the cells and mediates the adhesion of cells to the substrate or to other cells. It has been suggested that the extracellular matrix has a role in morphogenesis and differentiation [21]. In this paper we have studied the effect of TGF-/3, one of the most important growth inhibitors and a m~or regulator of extracellular matrix components, on the synthesis of individual proteoglycans. PGs are involved in cell adhesion, cell
recognition and growth regulation, due to the capacity to bind several extracellular matrix components and growth factors [10]. In human lung fibroblasts, we have shown by metabolic labeling experiments and confirmed by immunoprecipitation using specific antibodies that TGF-/3 affects differently the secretion of small PGs, PG-I and PG-II. Therefore, these two PGs can be independently regulated. After treatment with TGF-fl, PG-II levels are not significantly changed, whereas PG-I is induced up to 5-fold. The results obtained at the protein level have been confirmed by mRNA experiments. The 2.6 kb mRNA corresponding to PG-I increases 5-fold following TGF-fl treatment, while the two mRNA band for PG-II are unaffected. When actinomycin D is used, the production of TGF-/~ induced PGs was blocked suggesting that TGF-/3 acts at the transcriptional level, as occurs with other TGF-fl-regulated matrix components [1]. TGF-/3 increases the molecular weight of PG-I and PG-II, indicating that the factor affects also PG metabolism at a posttranslational level. This effect has been described in other cell types and is due to an increase in the length of the glycosaminoglycans (GAGs), rather than to a higher number of GAG chains [16]. Similar findings to those observed in lung fibroblasts were obtained in human skin fibroblasts (results not shown). It is interesting to observe that, in contrast to these types of human ceils, there is a marked increase in a secreted sulfated PG with a molecular mass similar to that of PG-II when rat and murine cells are treated with TGF-/3 (Ref. 15 and results not shown). Moreover, during the preparation of this manuscript a report by Breuer et al. was published [22] showing that TGF-/3 selectively increases PG-I synthesis in human osteosarcoma ceils and human fetal skin fibroblasts. Taken together, these data indicate that the response to TGF-fl depends largely on the cell type. In conclusion, TGF-/3 induces specifically the synthesis and secretion of PG-I in human fibroblasts, whereas the lev~! of PG-II remains largely unaffected. It is interesting to speculate about the meaning of this difference. Whereas PG-II is ubiquitous, PG-I has a more restricted distribution, being found mostly in cartilage and bone [12]. Thus, the effect of TGF-/~ would be to confer a certain degree of differentiation to fibroblasts. Little is known about the physiological role of PG-I. The expression of PG-II in CHO cells has been described as inhibitory to growth [23,24]. The results presented here indicate that the effect of TGF-/3 on PGs is different in cells that are equally growth inhibited by it. This suggest that both PGs could have the same efect, or at least participate in the same controlling processes. In any case, the modification of the
233 extracellular space by TGF-/3, including PGs, is likely to mediate, at least partially, the effects on proliferation and differentiation.
Acknowledgements We thank Dr. J. Massagu6 for providing the TGF-~ and Dr. L.W. Fisher for the antibodies and cDNA probes. We also thank Dr. J. Arifio for his help with the Northern blot experiments and his helpful comments, and Dr. S. Cheifetz for reading the manuscript. This work was supported by grants from the Comisi6n para la Investigaci6n Cientlfica y T6cnica (PB87/0760) and the Fondo de Investigaci6n Sanitaria de ia Seguridad Social (FIS 89/0488). M.R. is recipient of a B.I.L Fellowship from the FIS and A.H. of a postdoctoral fellowship from the Spanish Ministerio de Educaci6n y Ciencia.
References 1 Massagu6, J. (1987) Cell 49, 437-438. 2 Ignotz, R.A. and Massagu6, J. (1985) Proc. Natl. Acad. Sci. USA 82, 8530-8534. 3 Massagu6, J., Cheifetz, S., Endo, T. and Nadal-Ginard, B. (1986) Proc. Natl. Acad. Sci. USA 83, 8206-8210. 4 0 h t a , M, Anklesaria, P., Greenberger, J.S., Bassols, A. and Massagu6, J. (1987) Nature 329, 539-541. 5 Seyedin, S.M., Sega~ini, P.R., Rosen, D.M., Thompson, A.Y., Bentz, H. and Graycar, J. (1987) J. Biol. Chem. 262, 1946-1949. 6 Kurokowa, M., Lynch, K. and Podolsky, D.K. (1987) Biochem. Biophys. Res. Commun. 142, 775-782.
7 Roberts, A.B., Sporn, M.B., Assoian, R.K., Smith, J.M., Roche, N.S., Wakefield, L.M., Heine, U.I., Liotta, L.A., Falanga, V., Kehri, J.H. and Fauci, A.S. (1986) Proc. Natl. Acad. Sci. USA 83, 4!67-4171. 8 Sporn, M.B., Roberts, A.B, Shull, J.H., Smith, J.M., Ward, J.M. a,~d Sodek, J. (1983) Science 219, 1329-1331. 9 Ruoslahti, E. (1988) Annu. Rev. Cell Biol. 4, 229-255. 10 Ruoslahti, E. (1989) J. Biol. Chem. 264, 13369-13372. 11 Schmidtchen, A., Carlstedt, I., Malmstr/Jm, A. and Fransson, L.-A. (1990) Biochem. J. 265, 289-300. 12 Fisher, L.W., Termine, J.D. and Young, M.F. (1989) J. Biol. Chem. 264, 4571-4576. 13 Rosenberg, L.C., Choi, H.U., Tang, L.H., Johnson, T.L., Pal, S., Webber, C., Reiner, A. and Poole, A.R. (1985) J. Biol. Chem. 260, 6304-6313. 14 Choi, H.U., Johnson, T.L., Pal, S., Tang, L.H., Rosenberg, L. and Neame, P.J. (1989) J. Biol. Chem. 264, 2876-2884. 15 Bassols, A. and Massagu6, J. (1988) J. Biol. Chem. 263, 3039-3045. 16 Lories, V., De Boeck, H., David, G., Cassiman, JJ. and Van Den Berghe, H. (1987) J. Biol. Chem. 262, 854-859. 17 Border, W.A., Okuda, S., Languino, L.R. and Ruoslahti, E. (1990) Kidney Int. 37, 689-695. 18 Chomczynski, P. and Sacchi, N. (1987) Anal. Biochem. 162, 156-159. 19 Zimmermann, D.R. and Ruoslahti, E. (1989) EMBO J. 8, 29752981. 20 Krusius, T. and Ruoslahti, E. (1986) Proc. Natl. Acad. Sci. USA 83, 7683-7686. 21 Reddi, A.H. (1984) in Extracellular Matrix Biochemistry (P'ez, K.A., ed.), pp. 375-404, Elsevier Science Publishing Co., New York. 22 Breuer, B., Schmidt, G. and Kresse, H. (1990) Biochem. J. 269, 551-554. 23 Yamag, chi, Y. and Ruoslahti, E. (1988) Nature 336, 244-246. 24 Yamaguchi, Y., Mann, D.M. and Ruoslahti, E. (1990) Natt~re 346, 281-284.
229
BBAMCR 12955
Differential effect of transforming growth factor fl on proteoglycan synthesis in human embryonic lung fibroblasts Manuel Romarls, Ascensi6n Heredia, Anna Molist and Anna Bassols Departament de Bmqulmica i Biologia Molecular, Facultat de Veterin~ria, UniversitatAutSnoma de Barcelona, Bellaterra (Spain) (Received 23 November 1990)
Key words: Transforming growth factor; Proteoglycan synthesis; (Human embryonic lung fibroblast)
The effect of transforming growth factor ~ (TGF-/]) on the biosynthesis of individual proteoglycans (PGs) by human embryonic lung fibroblasts has been investigated using specific antibodies and cDNA probes. Human lung fibroblasts secrete the two small chondroitin/dermatan sulfate PGs, PG-I or biglycan (300 kDa) and, in a larger proportion, PG-I! or decorin (130 kDa). Metabolic labeling experiments reveal that TGF-/3 induces selectively the expression of PG-I, whereas the level of PG-II remains unaltered. The effect of TGF-/3 on PG-I and PG-II has been studied by immunoprecipitation and Northern blot analysis. Either at the core protein or mRNA level, a specific 5-fold increase in PG-I can be observed. TGF-/] acts probably at the transcriptional level, as actinomycin D blocks completely the TGF-/3 induced proteoglycan synthesis. A low saturation density and a slower growth rate is also observed for TGF-/$ treated cells. The possible role of PG-I and PG.II as mediators of the growth inhibition caused by TGF-/] is discussed.
Introduction Transforming growth factor/3 (TGF-fl) is a polypeptide growth factor which can be found in a large variety of cells and tissues. It is one of the most powerful growth inhibitory factors and acts as a differentiation agent in several cell types [1]. It can inhibit the differentiation in vitro of 3T3-L1 preadipocytes [2], myoblasts [3] and hematopoietic cells [3], but induces differentiation of chondroblasts [5] and some epithelial derived cells [6]. TGF-/3 has also been implicated in vitro in different processes such as morphogenesis, angiogenesis and wound healing [7,8]. One of the most important phenotypic effects of TGF-/3 is the ability of this factor to modulate the extracellular matrix [1]. TGF-fl is a potent inducer of several extracellular matrix components: fibronectin, several types of collagen, integrins and chondroitin/ dermatan sulfate proteoglycans. On the other hand,
Abbreviations: TGF-/3, transforming growth factor/3; PG, proteoglycan Correspondence: A. Bassols, Departament de Bioquimica i Biologia Molecular, Facultat de Veterinhria, Universitat Aut6noma de Barcelona, 08193 Bellaterra, Spain.
TGF-fl is able to induce several extracellular matrixdegrading proteinase inhibitors. The effects on the extracellular matrix are supposed to mediate, at least partially, the effects on proliferation and differentiation. Proteoglycans (PGs) are one of the main components of the extracellular matrix [9,10]. Fibroblasts produce different types of PG including chondroitin and heparan sulfate proteoglycans [11]. The most abundant PGs secreted into the medium by fibroblasts are the so-called small PGs, PG-I or biglycan and PG-II or decorin, that can also be found in the extracellular matrix from numerous tissues and cells. PG-I and PG-II have been cloned from human bone [12] and purified from bovine skin and cartilage [13,14]. The derived protein sequence of both PGs shows 55% amino acid identity, suggesting that they probably arose through a gene duplication. PG-II contains one attached glycosaminoglycan (GAG) chain, while PG-I contains two chains. Both PGs have protein cores of approx. 37 kDa that contain a series of 12 (PG-I) and 10 (PG-II) tandem repeats of leucine-rich regions that make up the major part of the molecule (80%). These regions are thought to mediate the binding of PG-I and PG-II to other extracdlular matrix components, such as collagens.
230 The effect of TGF-/3 on PG metabolism has already been described [15], but the action on individual PGs has not been completely assessed. Using specific antibodies and cDNA probes against human PG-I and PG-II, we show in the present work that TGF-/3 affects more markedly the expression of PG-I (biglycan) in human lung fibroblasts and that TGF-/3 acts at a transcriptional level.
Materials and Methods TGF-/31 from bovine bone was kindly donated by Dr. J. Massagu6 (Sloan Kettering Cancer Center, New York, U.S.A.). Chondroitinase ABC and AC were from Seikagaku Kogyo. Heparitinase was from Sigma. Products for cell culture were from Gibco. All other reagents were of analytical grade from Merck or Sigma.
Cell culture and metaboJic labeling Human embryonic lung fibroblasts (HFL1, ATCC CCL 153) were grown in RPMI 1640 medium with 10% fetal calf serum, 100 I U / m l penicillin and 100 ~tg/ml streptomycin at 37 °C in a humidified 5% CO2 atmosphere. Subconfluent cultures were placed in serum free medium for 16 h before the experiment. At the beginning of the experiment, TGF-/3 was added at the indicated times and concentrations. 8 h before the end of the incubation, the medium was replaced with low sulfate (0.2 mM) minimal essential medium containing 100 ~Ci/ml of [35S]sulfate (New England Nuclear) or with methionine-free minimal essential medium containing 25 ~tCi/ml [~SS]methionine (New England Nuclear). After labeling, the medium was collected and made up to I mM PMSF. Aliquots of the medium were analyzed by SDS-PAGE and fluorography. Enzymatic digestions with chondroitinase ABC, chondrotinase AC and heparitinase were performed as previously described [15,16].
Immunoprecipitation Polyclonal antibodies directed against the aminoterminus of PG-I (LF-51) and PG-II (LF-30) were kindly provided by Dr. L.W. Fisher, HAM., U.S.A. The antibodies did not crossreact with each other's target protein [12]. lmmunoprecipitations were performed by adding 100 ~! of antiserum to 750 Izl of [35S]methionine labeled conditioned medium treated with chondroitinase ABC. Preimmune serum was used in parallel control experiments. The samples were incubated overnight at 4 °C with mixing. To precipitate the antigen-antibody complexes, 50 /~! of protein A-Sepharose (Sigma) was added to the samples [17]. The immunocomplexes were dissolved in sample buffer and analyzed by electrophoresis.
Northern blot analysis Total RNA from control and TGF-10 treated cells was isolated as described by Chomczynski and Sacchi [18]. The RNA was fractionated by 1% agarose gel electrophoresis and transferred to Genescreen Plus filters (Dupont). The membranes were hybridized using eDNA probes to human PG-I and PG-II, under the conditions described by Fisher et al. [12]. The probes were kindly donated by Dr. L.W. Fisher (N.I.H. U.S.A.).
Growth proliferation assays Cells were plated in triplicate in 24-well plates in the presence or absence of 100 pM TGF-/3 in serumcontaining medium. Each day, half of the medium was removed and new medium with or without TGF-fl was added. In this way it was possible to maintain the higher levels of PGs in the medium of TGF-/3 treated cells. At the stated times, cells were detached with trypsin and counted.
Results
Effect of TGF.~ on lung fibroblast proliferation TGF-/3 acts as a growth inhibitor on lung fibroblasts as it does in other cell types [1]. As shown in Fig. 1, TGF-/] added to the cultures for several days markedly decreases cell growth rate as well as saturation density.
Effect of TGF-~ on PG production by lung fibroblasts Subconfluent cells were incubated with [35S]suifate to label the PGs and the conditioned medium was analyzed by electrophoresis in order to characterize the profile of secreted PGs in this cell line. In control cells, three sulfated bands were observed (Fig. 2): a high molecular weight band that does not enter into the gel, which probably corresponds to versican [19], and two other bands with average molecular masses of 300 kDa and 130 kDa. These mobilities correspond to the small chondroitin/dermatan sulfate PGs, identified in other cell types as PG-I and PG-II. In TGF-fl treated cells, a significant increase in the secretion of PG-I was observed, as shown in Fig. 2. The increase was 4.5 + 0.5-fold (mean + S.E. from five different experiments) as quantified by densitometry. In contrast, the intensity of the band corresponding to PG-II did not change. A moderate increase in the molecular weight of PG-I and PG-II was also observed, as described in other cell types [15]. Dose-response and time-course experiments indicate that the effect of TGF-fl on PG production was maximal at 50 pM and at 12 h, decreasing slowly after 24 h (results not shown). Therefore, the conditions chosen for the subsequent experiments were 100 pM TGF-/] and 16 h exposure to the factor. To characterize the GAG chains on the PGs in-
231 120
CELL3/10000
O0
pig
0 100-
kDa I PG-I
80"
205
-
O0"
PG-II 40.
116-
97-
20'
0
0
,
,
,
,
2
4
0
8
6610
12
DAYS Fig. |. Effect of TGF-/3 on human embryonic lung fibroblast proliferation. Cells were plated in triplicate in the presence or absence of
100 pM TGF-/3. Each day, half of the medium was removed and new medium with or without TGF-B was added. At the stated times, cells were detached with trypsin and counted.
duced by TGF:/3, aliquots of the [aSS]sulfate labeled conditioned medium from control and TGF-/3 treated cells were incubated with the enzymes chondroitinase AC, chondroitinase ABC and heparitinase. Chondroitinase AC causes a partial degradation of both PG-I and PG-II, whereas they were completely degraded by chondroitinase ABC, heparitinase does not have any effect on these two PGs. These results indicate that the TGF-/3-inducible sulfated-PGs belong to the chondroitin/dermatan sulfate type (not shown).
Identification of PG-I and PG-H To confirm that the high and low molecular weight sulfated bands were PG-I and PG.II, respectively, immunoprecipitation experiments using specific antibodies raised against the protein core of the two PGs were performed. Conditioned medium from cells labeled with [35S]methionine in the presence or absence of TGF-/3 was treated with chondroitinase ABC and immunoprecipitated as described in Materials and Methods. As shown in Fig. 3, the anti-PG-I antibody immunoprecipitated the 45 kDa band corresponding to the protein core of PG-I. In TGF-/~ treated cells the intensity of this band was increased about 5-fold, similar to the results obtained in the metabolic labeling experiments. When the anti-PG-II antibody was used, two bands of 44 and 47 kDa corresponding to the protein core of PG-II were observed and there was no increase in its intensity after TGF-/3 treatment.
45Fig. 2. TGF-/] effects on secreted proteoglycans in human embryonic lung fibroblasts. Subconfluent cells were labeled with [35S]sulfate and treated with TGF-/3 for 16 h. Conditioned medium was removed and analyzed by SDS-PAGE in 6% polyacrylamide gels. Molecular weight markers are shown on the left.
Effect of TGF-~ at the mRNA level T o assess w h e t h e r TGF-/3 was a c t i n g at the m R N A level, total R N A was extracted f r o m TGF-/~ t r e a t e d
and control cells and hybridized with specific eDNA probes for PG-I and PG-II.
PG-I --
PG-li +
--
m
m
4"
TGF-/3
............. ~
~' -
~:
-
205 116
-45 kDa
Fig. 3. Immunoprecipitation of PG-I and PG-II from control and TGF-/3 treated cells. Subconfluent cultures were labeled with 25 /zCi/ml of [3SS]methionine and treated with 100 pM TGF-/3 for 16 h. The conditioned medium was treated with chondroitinase ABC and immunoprecipitated with specific antibodies against PG-I and PG-II as described in Materials and Methods. PG-I and PG-II protein cores are marked with arrows.
232 PG-I 0
6
PG-I 12
0
-
28S-
-
18S-
6
I 12
H
Fig, 4. Northern blot analysisof TGF-/3 effect on PG-i and PG-II. Total RNA was extracted fromcellstreated with 100 pM TGF-/3for different times. Blots were hybridizedwith cDNA probes for PG-I (left) and PG-II(right).
As can be seen in Fig. 4, when blots were probed with the cDNA cc~:esponding to PG-I, a band of 2.6 kb was observed which correlates with the message size observed in other tissues [12]. In cells treated with TGF-/3, and increase of 5-fold in the intensity of the band was observed with a maximum at 12 h of treatment, similar to the effects observed at the protein level (Figs. 2 and 3). When blots were hybridized with the cDNA corresponding to PG-11, two bands of 1.6 and 1.9 kb were shown (Fig. 4). Two messages of similar size have been already described in IMRg0 fibroblasts [20]. There was no increase in any of the bands after treatment with
TGF-0, Actinomycin D blocked the induction of PG-I by TGF-~, showing that the factor acts at the transcriptional level. Addition of cycloheximide to cells 1 h before the labeling period, inhibited the appearance of PC~ indicating that the effect of TGF-/3 required de novo protein synthesis, presumably synthesis of the core protein (results not shown). Discussion
The extracellular matrix is a complex structure that surrounds the cells and mediates the adhesion of cells to the substrate or to other cells. It has been suggested that the extracellular matrix has a role in morphogenesis and differentiation [21]. In this paper we have studied the effect of TGF-/3, one of the most important growth inhibitors and a m~or regulator of extracellular matrix components, on the synthesis of individual proteoglycans. PGs are involved in cell adhesion, cell
recognition and growth regulation, due to the capacity to bind several extracellular matrix components and growth factors [10]. In human lung fibroblasts, we have shown by metabolic labeling experiments and confirmed by immunoprecipitation using specific antibodies that TGF-/3 affects differently the secretion of small PGs, PG-I and PG-II. Therefore, these two PGs can be independently regulated. After treatment with TGF-fl, PG-II levels are not significantly changed, whereas PG-I is induced up to 5-fold. The results obtained at the protein level have been confirmed by mRNA experiments. The 2.6 kb mRNA corresponding to PG-I increases 5-fold following TGF-fl treatment, while the two mRNA band for PG-II are unaffected. When actinomycin D is used, the production of TGF-/~ induced PGs was blocked suggesting that TGF-/3 acts at the transcriptional level, as occurs with other TGF-fl-regulated matrix components [1]. TGF-/3 increases the molecular weight of PG-I and PG-II, indicating that the factor affects also PG metabolism at a posttranslational level. This effect has been described in other cell types and is due to an increase in the length of the glycosaminoglycans (GAGs), rather than to a higher number of GAG chains [16]. Similar findings to those observed in lung fibroblasts were obtained in human skin fibroblasts (results not shown). It is interesting to observe that, in contrast to these types of human ceils, there is a marked increase in a secreted sulfated PG with a molecular mass similar to that of PG-II when rat and murine cells are treated with TGF-/3 (Ref. 15 and results not shown). Moreover, during the preparation of this manuscript a report by Breuer et al. was published [22] showing that TGF-/3 selectively increases PG-I synthesis in human osteosarcoma ceils and human fetal skin fibroblasts. Taken together, these data indicate that the response to TGF-fl depends largely on the cell type. In conclusion, TGF-/3 induces specifically the synthesis and secretion of PG-I in human fibroblasts, whereas the lev~! of PG-II remains largely unaffected. It is interesting to speculate about the meaning of this difference. Whereas PG-II is ubiquitous, PG-I has a more restricted distribution, being found mostly in cartilage and bone [12]. Thus, the effect of TGF-/~ would be to confer a certain degree of differentiation to fibroblasts. Little is known about the physiological role of PG-I. The expression of PG-II in CHO cells has been described as inhibitory to growth [23,24]. The results presented here indicate that the effect of TGF-/3 on PGs is different in cells that are equally growth inhibited by it. This suggest that both PGs could have the same efect, or at least participate in the same controlling processes. In any case, the modification of the
233 extracellular space by TGF-/3, including PGs, is likely to mediate, at least partially, the effects on proliferation and differentiation.
Acknowledgements We thank Dr. J. Massagu6 for providing the TGF-~ and Dr. L.W. Fisher for the antibodies and cDNA probes. We also thank Dr. J. Arifio for his help with the Northern blot experiments and his helpful comments, and Dr. S. Cheifetz for reading the manuscript. This work was supported by grants from the Comisi6n para la Investigaci6n Cientlfica y T6cnica (PB87/0760) and the Fondo de Investigaci6n Sanitaria de ia Seguridad Social (FIS 89/0488). M.R. is recipient of a B.I.L Fellowship from the FIS and A.H. of a postdoctoral fellowship from the Spanish Ministerio de Educaci6n y Ciencia.
References 1 Massagu6, J. (1987) Cell 49, 437-438. 2 Ignotz, R.A. and Massagu6, J. (1985) Proc. Natl. Acad. Sci. USA 82, 8530-8534. 3 Massagu6, J., Cheifetz, S., Endo, T. and Nadal-Ginard, B. (1986) Proc. Natl. Acad. Sci. USA 83, 8206-8210. 4 0 h t a , M, Anklesaria, P., Greenberger, J.S., Bassols, A. and Massagu6, J. (1987) Nature 329, 539-541. 5 Seyedin, S.M., Sega~ini, P.R., Rosen, D.M., Thompson, A.Y., Bentz, H. and Graycar, J. (1987) J. Biol. Chem. 262, 1946-1949. 6 Kurokowa, M., Lynch, K. and Podolsky, D.K. (1987) Biochem. Biophys. Res. Commun. 142, 775-782.
7 Roberts, A.B., Sporn, M.B., Assoian, R.K., Smith, J.M., Roche, N.S., Wakefield, L.M., Heine, U.I., Liotta, L.A., Falanga, V., Kehri, J.H. and Fauci, A.S. (1986) Proc. Natl. Acad. Sci. USA 83, 4!67-4171. 8 Sporn, M.B., Roberts, A.B, Shull, J.H., Smith, J.M., Ward, J.M. a,~d Sodek, J. (1983) Science 219, 1329-1331. 9 Ruoslahti, E. (1988) Annu. Rev. Cell Biol. 4, 229-255. 10 Ruoslahti, E. (1989) J. Biol. Chem. 264, 13369-13372. 11 Schmidtchen, A., Carlstedt, I., Malmstr/Jm, A. and Fransson, L.-A. (1990) Biochem. J. 265, 289-300. 12 Fisher, L.W., Termine, J.D. and Young, M.F. (1989) J. Biol. Chem. 264, 4571-4576. 13 Rosenberg, L.C., Choi, H.U., Tang, L.H., Johnson, T.L., Pal, S., Webber, C., Reiner, A. and Poole, A.R. (1985) J. Biol. Chem. 260, 6304-6313. 14 Choi, H.U., Johnson, T.L., Pal, S., Tang, L.H., Rosenberg, L. and Neame, P.J. (1989) J. Biol. Chem. 264, 2876-2884. 15 Bassols, A. and Massagu6, J. (1988) J. Biol. Chem. 263, 3039-3045. 16 Lories, V., De Boeck, H., David, G., Cassiman, JJ. and Van Den Berghe, H. (1987) J. Biol. Chem. 262, 854-859. 17 Border, W.A., Okuda, S., Languino, L.R. and Ruoslahti, E. (1990) Kidney Int. 37, 689-695. 18 Chomczynski, P. and Sacchi, N. (1987) Anal. Biochem. 162, 156-159. 19 Zimmermann, D.R. and Ruoslahti, E. (1989) EMBO J. 8, 29752981. 20 Krusius, T. and Ruoslahti, E. (1986) Proc. Natl. Acad. Sci. USA 83, 7683-7686. 21 Reddi, A.H. (1984) in Extracellular Matrix Biochemistry (P'ez, K.A., ed.), pp. 375-404, Elsevier Science Publishing Co., New York. 22 Breuer, B., Schmidt, G. and Kresse, H. (1990) Biochem. J. 269, 551-554. 23 Yamag, chi, Y. and Ruoslahti, E. (1988) Nature 336, 244-246. 24 Yamaguchi, Y., Mann, D.M. and Ruoslahti, E. (1990) Natt~re 346, 281-284.
