JOURNAL OF CELLULAR PHYSIOLOGY 143:534-545 (19901

Differential Effects of Transforming Growth Factor$ and Epidermal Growth Factor on the Cell Cycle of Cultured Rabbit Articular Chondrocytes DENIS VIVIEN,*

PHILIPPE GALERA, EMMANUEL LEBRUN, GERARD LOYAU, JEAN-PIERRE PUJOL

AND

Laboratoire de Biochirnie du Tissu Conjonctif, C H U CBte de Nacre, 14033 Caen Cedex (D.V., P.C., C.I., ].-P.PJ, and Centre de Transfusion Sanguine, Cacn (€.I.) France We examined the effect of transforming growth factor (TGF-P) on the proliferative rate and cell cycle of cultured rabbit articular chondrocytes using cell counting, cytofluorometry, and [-%-thymidine incorporation. In the presence of 2% or 10% FCS (fetal calf serum), TGF-p at O . O l , O . l , 1, and 10 ng/ml had an inhibitory effect on cell proliferation after 24 h exposure with a dose dependence only for 2% FCS. Flow cytometric analysis of cell DNA content at that time showed that a high proportion of cells were arrested in late S-phase (SQ or G2Q) in either 2% or 10% FCS-containing medium. In both cases, a disappearance of the cell blockage occurred between 24 and 48 h after TGF-P addition. However, whereas a stimulation of cell proliferation rate was observed at that time in cultures containing 10% FCS, a dose-dependence inhibition of cell growth was detected, in contrast, for 2% FCS-treated cells. Presence of TGF-P during the last 24 h was not necessary to release the arrested cells. Furthermore, platelet-poor plasma at 10% produced the same effects as FCS, suggesting that platelet-derived factors, such as platelet-derived growth factor (PDGF), could not be responsible for the release of blocked cells in this case. We conipared the effect of TGF-P to that of epidermal growth factor (EGF), used at an optimal concentration (1 0 ngiml). In both slowly growing (2% FCS) and proliferating chondrocytes (10% FCS), EGF caused a significant increase of cell proliferation as early as 24 h. No arrest in late S-phase but an augmentation of the percentage of cells in S- and G2M-phases were observed. When combined, TGF-(3 and EGF did not induce synergistic effect on the chondrocyte proliferation, as estimated by cell counting. [ 'HI-thymidine labeling showed that the factors induced identical maxima of incorporation but the peak occurred earlier for TGF-P than for EGF (approximately 6 h versus 12 h, respectively). Although both factors induce similar cell-number increases at 48 h in 10% FCS-containing medium, these proliferative effects were due to different actions on the cell cycle. The present study indicates that TGF-P induces first a recruitment of chondrocytes in noncycling SQ- or G2Q-blocked cells. Then, the release of these cells may produce either apparent stimulation of cell proliferation if sufficient levels of an unknown serum factor are present (10% FCS) or an inhibition of growth rate when only reduced amounts of this factor are available (2% FCS). This mechanism is quite different from that of EGF, which make more cells enter S-phase, whatever the serum concentration in medium, and might be related to the transforming capacity of TGF-p.

Transforming growth factor+ (TGF-P) is a 22-25 kDa peptide derived from both normal tissues and transformed cells (Roberts e t al., 1980, 1981), which is composed of two subunits held together by interchain disulfide bonds. TGF-P has been described for its ability to confer anchorage-independent proliferation on rat normal rat kidney (NRK) fibroblasts, in the presence of epidermal growth factor (EGF) (Roberts et al., 1981) and that of mouse AKR-2B cells without the necessity of EGF (Tucker et al., 1983). It has been found 'Q 1990 WILEY-LISS, INC

to b e an important regulator of cellular growth and metabolism (Sporn et al., 1986). The effects of this factor are multifunctional. TGF-P acts as an inhibitor of growth in several cell types, including epithelial cells (Tucker et al., 1984; Masui et al., 19861, T and B lymphocytes (Kherl et al., 19861, fibroblasts (Roberts et a]., Received May 3, 1989; accepted January 26, 1990.

*To whom reprint requestsicorrespondence should be addressed.

TGF-8 MODULATION OF CHONDROCYTE PROLIFERATION

1985; Anzano et al., 19861, and hepatocytes (Hayashi and Carr, 1985).However, TGF-P is a bifunctional regulator of cell growth in mesenchymal cells and may inhibit or stimulate cell proliferation depending on a number of factors such as the culture procedure (monolayers or soft agar, for example), the growth factors acting together with TGF-P, or the developmental stage of the cells (review in Roberts et al., 1988). In addition, TGF-P has been shown to stimulate production of matrix molecules such as fibronectin, collagen (Ignotz and Massague, 1986; Fine and Goldstein, 1987; Redini et al., 19881, and proteoglycans (Chen et al., 1987; Redini et al., 1988). TGF-(3 is homologous with cartilage-inducing factor A (CIF-A), a growth factor isolated from bone which induces embryonic rat mesenchymal cells in culture to differentiate into chondrocytes, with production of cartilage specific proteoglycan and type I1 collagen (Seyedin et al., 1986). Chondrocytes have been shown to possess TGF-P mRNA (Young et al., 1986), and this factor has also been found in articular cartilage (Ellingsworth et al., 1986). Recently, we demonstrated that the TGF-p stimulates the synthesis of both collagen and proteoglycan in monolayer cultures of rabbit chondrocytes (Redini et al., 1988), a datum which is slightly different from the results found by others in three-dimensional gel cultures, where a n inhibition of collagen production was observed (Skantze et al., 1985). Finally, TGF-P appears a s a n important regulator of the matrix synthesis in cartilage and, therefore, a role for this factor in osteoarticular diseases may be postulated. The stimulative effect of TGF-P could be related to repair process of cartilage since several studies on pathological joints have shown that articular cartilage can form tissue resembling fibrocartilage or fibrous tissue in response to injury (Meachum and Roberts, 1971; Nakata and Bullough, 1983). In this respect, it was of interest to determine whether TGF-P can also influence the proliferation of articular chondrocytes. So far, the factors that regulate this proliferative activity have not been defined, although several local autocrine or paracrine growth factors have been discovered in cartilage (Beckoff and Klagsbrun, 1982; Kato et al., 1983). To test the hypothesis that TGF-P can regulate chondrocyte growth in vitro, we studied its effects on the proliferation rate of cultured rabbit articular chondrocytes and compared them to the action of epidermal growth factor (EGF) a s a nontransforming and typical growth factor.

MATERIAL AND METHODS Culture of articular chondrocytes Articular cartilage slices were taken from the shoulders and the knees of 3-week-old rabbits. Chrondrocytes were obtained by enzymatic dissociation (Benya et al., 1977) and cultured in DMEM (Dulbecco’s modified Eagle’s medium) supplemented with glutamine (2 mM), penicillin (100 Uiml), streptomycin (100 pgiml), fungizone (0.25 pgiml), and 10% heat-inactivated fetal calf serum (FCS). The cells were grown a t 37°C in a 5% CO, environment with medium change a t 2-3 day intervals. For the experiments, they yere generally plated in Petri dishes (9.6 cm2) at 2.10” cellsiwell, except for preliminary studies designed to determine the

535

effect of cell plating density on the growth curves (1,2, and 3.105 cellsiwell). Primary cultures were used in order t o avoid dedifferentiation of the chondrocytes. They were maintained in medium supplemented with 10% FCS for 3 days after seeding. Then, the medium was replaced with same medium containing different serum concentrations depending on the experiment.

Growth factors Human derived transforming growth factor (TGF-PI was supplied by R and D Systems (Minneapolis, MN) and epidermal growth factor (EGF) was obtained from Sigma (St Louis, MO). Cell counting The cells were harvested by trypsinization (0.025% trypsin for 10 min at 37°C) and counted in a Coulter counter (Hycel) in triplicate. Flow-cytometric analysis Flow-cytometric analysis of DNA content distribution was performed with a modification of the technique of Crissman and Steinkamp (1982), using the same dishes that were prepared for cell counting. After trypsinization, the cells were collected in 1 ml of DMEM + 10% FCS and a n aliquot of 250 p1 was used for counting. The remaining cell suspensions from the triplicate dishes were then pooled for subsequent DNA analysis except when the countings differed by more than 5%.The cells were centrifuged, washed with PBS (phosphate buffered saline), and finally resuspended in 1 ml of PBS containing 50 pgiml RNase (previously treated for 30 min at 75°C to destroy DNase traces) and 50 pgiml propidium iodide. The reaction was stopped a t 4°C and the cells, still in the staining solution, were analysed in a Cytofluorograf FASCAN (Becton Dickinson, Mountain View, CA), equipped with a n argon laser operating a t a wavelength of 488 nm. The relative percentages of cells in GOil-, S-, and G2+M- phases were estimated according to the SFIT mathematical method (Krishan and Frei, 1976). When this analytical method was not suitable because of a too small S-phase for that model, Fried’s method of histograms (Fried, 1976) was used. Both statistical methods were performed on 7-10.103 events and only if the coefficient of variation was less than 8%. [’HI-Thymidine incorporation

I6-3H]-thymidine (25 Ciimmol, CEA France) was added to the culture medium a t a final concentration of 1or 2 pCiiml depending on the experiment. At the end of incubation, the medium was removed and the monolayers were rinsed once with 1 ml PBS. This rinse medium was aspirated and the culture dishes were placed on a n ice block while 1 ml ice-cold 5% trichloroacetic acid (TCA) was added slowly. The culture dishes were placed in a cold room a t 4°C for 30 min, whereupon the TCA was removed and the fixed monolayer was carefully rinsed three additional times with ice-cold 5% TCA. Then, 0.5 ml of 0.1 M NaOH was added to each well for 1 h a t 50°C to solubilize the precipitated material. Aliquots of 0.4 ml NaOH extract were removed from each well for liquid scintillation counting.

536

VIVIEN ET AL

Preparation of calf platelet-rich and platelet-poor plasma Young calf blood was collected in plastic syringes, using 25 mM citrate phosphate buffer as anticoagulant, and divided into two fractions. The first fraction was centrifuged at 250g for 15 min to remove red cells. The second fraction was spun twice a t 4,OOOg for 15 min in order to separate red cells and platelets. The resultant platelet-rich and platelet-poor plasma (PRP and PPP) were used in similar fashion to FCS.

Protein a s s a y The cell layers were washed with PBS and solubilized in 1 ml of 0.2 M NaOH. Aliquots of the solution were used to estimate the protein amount by a modification of the Lowry’s method (Hartree, 1972). Statistics Statistical analysis for cell counting and thymidine incorporation was performed using Student’s t-test. RESULTS

serum concentration present. However, for the lowest serum levels (0.4% and 2%), t h e chondrocytes proliferate slowly all along the experiment. The profiles in these cases displayed a small slope. They significantly differed from the patterns obtained with 5% and 10% FCS that showed a logarithmic phase followed by a plateau. In this study, we determined that 10% FCS caused maximal level of proliferation since 8% FCS already gave a n identical growth curve (not shown). As seen in Figure 2 (inserts), the cytofluorometric analysis of DNA content clearly shows that most of the cells exposed t o 0.4% or 2% FCS-containing medium were slowly growing, with 70% to 80% GO/l-phases. In contrast, the cell population from cultures incubated with G2M5% and 10% FCS contained about 50% S phases in the period day 3-day 5, reflecting a high proliferative rate. From the preceding data t h a t define the experimental system, we chose to investigate the effect of TGF-P between day 3 and day 5, in both proliferative (10% FCS) and slowly growing cells ( 2 8 FCS). These conditions allowed us to study the effect of a factor on the proliferation activity independently of the cell contact inhibition, since cells are either actively growing or potentially capable of proliferating.

+

Cell growth kinetics of rabbit articular chondrocytes in monolayer culture Effect of TGF-p on c h o n d r o c y t e proliferation As a prerequisite t o the study of TGF-P action on and cell cycle proliferation, we determined first the effect of cell platWhen articular chondrocytes in slowly growing state ing density on the growth curve. When rabbit articular chondrocytes were seeded at three different densities (2% FCS) were incubated for 24 h and 48 h with 0.01, (lo5, %lo5,and 3.105 cells/dish) in 10% FCS-containing 0.1, 1, and 10 ngiml TGF-p, we observed a dose-depenmedium, typical growth curves were obtained by cell dent inhibition of proliferation rate, as estimated by counting (Fig. 1). They showed three successive peri- cell counting (Fig. 3A). In cell proliferative conditions ods: a lag phase (day l 4 a y 31, a logarithmic period (10% FCS), TGF-p induced a non-dependent reduction (day 3-day 6), and a plateau beginning by day 5 or 6. of cell number during the first 24 h after addition, The final stationary phase was associated with a n while i t increased the cell number by approximately abundant deposition of extracellular matrix that could 20-30% at 48 h, compared to controls (Fig. 3B). The be visualized by standard microscopy staining (not stimulatory effect observed a t 48 h was induced by all shown). By day 10 or 11, the cultures became highly the concentrations of TGF-p used, but dose dependence confluent and some cells detached from the bottom of was only seen between 0.01 and 0.1 ng/ml. The data the dish. Therefore the cultures were not studied be- suggest t h a t TGF-(3 could exert opposite effects on the proliferation process. TGF-P induces either a decrease yond that time. More precise information about the cellular state or a n increase of the cell number after 48 h exposure, was obtained from plotting the cell protein level ex- depending on the serum concentration present in the pressed on a cell basis (Fig. 1,insert). The profile shows medium. The results of DNA content analysis (Table 1) india n ascending phase (days 2-4), reflecting that cell protein synthesis had a greater rate than proliferation, cate that the cultures treated by TGF-P (0.01 to 10 followed by a decrease of the index corresponding to a ng/ml) for 24 h in presence of 2 and 10% FCS contained high proportion of dividing cells in that period (days a high proportion of chondrocytes arrested a t the end of 4-6). Then the ratio raised again from day 6 to day 8, S-phase. The cytofluorometric profiles showed that the but, in that case, this was mainly due to a n increase of peak corresponding to the G2M-phase was shifted to protein synthesis since the cell number remained es- the left, in the zone usually attributed to S-phase in sentially stable over that time. Thus, the two ascend- controls (Fig. 4). This phenomenon impeded analysis ing parts of the profile, despite their identical slopes, with the SFIT model because of a too small S-phase. correspond to different states of the cell population Using Fried‘s method of histograms a s a n alternative, which could not have been detected by cell counting nor we tentatively estimated the percentage of cells arprotein assay alone. It is also clear from this experi- rested in late S-phase, i.e., SQ or G2Q. This phase is ment that the different cell plating densities lead to known to contain noncycling cells with a nuclear DNA similar growth curves with same maximal levels, prob- content equal to t h a t of cells in S- or G2-phase (4 c) ably a s a combined result of three critical parameters: which are able t o rapidly reenter the cell cycle (Gelsurface, density, and serum concentration. Therefore, fant, 1977). Though the experimental values for this we studied in the same experimental conditions the phase in Table 1 are only indicative, we found up to effect of changing the serum concentration on day 3, 41% chondrocytes arrested in t h a t portion of the cell i.e., a t the end of the lag phase where cells were defin- cycle a t 24 h after TGF-p addition. At 48 h of exposure itively settled. Figure 2 indicates that the onset of cell to TGF-p, SQ- or G2Q-phase has completely disapproliferation occurred at the same time, whatever the peared in both 2% and 10% FCS-treated cultures, in-

537

TGF-p MODULATION OF CHONDROCYTE PROLIFERATION

1

50

4

2

0

8

6

10

Time (days) Fig. 1. Effect of cell plating density on the growth curve of rabbit articular chondrocytes. The cells were seeded in 35-mm plates at densities of lo5 cellsidish (dl:o), 2.10" cellsidish (d2:W, and 3,105 cells/ dish (d3:0 1 in 3 ml of DMEM plus 109 PCS. At the indicated times, the cell number was determined by Coulter counting and the cell

layer protein was estimated on parallel dishes as described in Methods. Each point corresponds to 3 plates and error bars have been included only when they exceed the size of the symbol. Insert shows the ratio of protein to cell number for the density d2.

30

r

a"lG$]

4

3

5

7

TI

20

L

10

2 IFCS >G 0

1

3

5

7

'

0

0

2

8

6

4

10

1 2

Time (days) Fig. 2. Effect of serum concentration on cell proliferation and cell distribution in the phases of cell cycle. Cells were seeded in 35-mm plates at a density of %lo5 cellsidish in 3 ml of DMEM plus 10% FCS. On day 3 the media were changed for DMEM containing different concentrations of FCS. Then, the media were replaced every 3 days. At indicated times, the cell number was determined and cytofluoro-

metric analysis of DNA content was performed on the pool of 3 samples as described in Methods (inserts). The data were obtained on 7.103 events and statistical treatment was performed only when the coefficient of variation was less than 8%. Each point corresponds to 3 dishes and error bars have been included only when they exceed the size of the symbol.

dicating that the previously arrested cells have entered the cycle by the time. However, this was accompanied by a relative increase of cell number only in cultures

containing 10% FCS, whereas reduced cell counts were observed a t the same time for 2% FCS-containing samples (Fig. 3). The mechanism of such a dual effect of

538

VIVIEN ET AL.

Fig. 3. Effect of TGF-8 on cell proliferation.Cells were seeded as in Figure 2 in 3 ml of DMEM plus lo%, FCS. On day 3, the media were changed for 1.5 ml of DMEM containing 2% (A) or 10% FCS (B) with or without TGF-p. At the end of incubation (24 and 48 h) the cells were harvested and counted. Each point represents the mean S.E.M. of triplicate dishes. **Significant at P

Differential effects of transforming growth factor-beta and epidermal growth factor on the cell cycle of cultured rabbit articular chondrocytes.

We examined the effect of transforming growth factor (TGF-beta) on the proliferative rate and cell cycle of cultured rabbit articular chondrocytes usi...
1MB Sizes 0 Downloads 0 Views