Effect of seeding density on stability of the differentiated phenotype of pig articular chondrocytes in culture

Transcription

1 Effect of seeding density on stability of the differentiated phenotype of pig articular chondrocytes in culture FIONA M. WATT Imperial Cancer Research Fund, IK) Box 123, Lincoln's Inn Fields, London WC2A 3PX, UK Summary Articular chondrocytes are known to be phenotypically unstable in culture. One condition that has been reported to suppress dedifferentiation is cultivation at high density on tissue-culture plastic. The aim of the experiments described here was to study the effect of seeding density on chondrocyte proliferation and 35 SO 4 incorporation, and on the types of collagen and proteoglycan synthesized. I found that cells seeded at low or high density reached the same final density at confluence, and that 35 SO 4 incorporation, while initially higher (per cell) in high-density cultures, fell under both conditions, reaching the same low level after 3 weeks. The proportion of cells expressing keratan sulphate fell in low- but not high-density cultures and the decline was not prevented by inhibition of cell division. In all the cultures cells expressing keratan sulphate tended to have a rounded morphology. After 21 days in culture, chondrocytes grown at high density expressed predominantly large proteoglycans that aggregated with hyaluronic acid, whereas in lowdensity cultures a smaller, non-aggregating form was also present. By 21 days in culture cells at both high and low density were expressing type I collagen, although the high-density cells also had an extensive extracellular matrix of type II collagen. These observations support the conclusion that high seeding density stabilizes the chondrocyte phenotype to a greater extent than low seeding density. They also suggest that enhanced dedifferentiation at low density may be due to cell spreading, rather than to selective proliferation of a phenotypically unstable subpopulation of cells. Key words: chondrocytes, differentiation, cell shape, proteoglycans, collagen. Introduction Articular cartilage contains a single cell type, the chondrocyte, which secretes a matrix consisting predominantly of type II collagen and high molecular weight proteoglycans aggregated with hyaluronic acid. Chondrocytes can readily be isolated from cartilage; however, during prolonged cultivation they tend to stop synthesizing the extracellular matrix molecules characteristic of cartilage and begin to synthesize type I collagen and small, non-aggregating proteoglycans (reviewed by von der Mark, 1986). This loss of differentiated phenotype has been termed dedifferentiation, although the word is not intended to imply regression to an embryonic state (Holtzer et al. 1960). There is controversy as to whether dedifferentiation is Journal of Cell Science 89, (1988) Printed in Great Britain The Company of Biologists Limited 1988 an irreversible process or whether it reflects environmental modulation that can be reversed under certain conditions (Benya & Shaffer, 1982; von der Mark, 1986). Culture conditions that promote dedifferentiation include: low starting density, increasing time in culture and passaging. Dedifferentiation can be prevented or slowed by seeding chondrocytes at high density (Amadio et al. 1983; Kuettner et al. 1982o,6); by growth in suspension (Srivastava et al. 1974; Bcnya & Shaffer, 1982) or on a substratum that restricts spreading (Glowacki et al. 1983; F. M. Watt and J. Dudhia, unpublished results). The purpose of the experiments described here was to investigate the effect of seeding density in more detail, and, specifically, to discover whether enhanced dedifferentiation at low density is 373

2 due to cell spreading or to selective proliferation of a phenotypically unstable subpopulation of cells. Materials and methods Chondrocyte isolation and culture Pig's trotters were purchased from local butchers' shops and the cartilage dissected from the articular surfaces. Chondrocytes were released from the cartilage matrix by sequential digestions with O'Ol % hyaluronidase, 0-1% pronase and 0-25% collagenase, as described previously (Zanetti et al. 1985). The cells were filtered through 60^m aperture nylon cloth (Cadisch and Sons, London) to remove undigested matrix and grown on tissue-culture plastic in Dulbecco's modified Eagle's medium containing 10% foetal calf serum. Chondrocytes were seeded at (high density) or 10 5 (low density) per 35-mm diameter Petri dish in 2 ml medium. For determining the growth rate of chondrocytes in culture, duplicate dishes were harvested with 0-05% trypsin and 0'01 % EDTA and counted in a haemocytometer. Column chromatography The amount and type of proteoglycan synthesized by chondrocytes was analysed after labelling the cultures for 24 h with loficimr 1 35 SO 4 (specific activity Cimg" 1 ; Amersham International, Amersham, England). To separate unincorporated label from 35 SC>4 incorporated into proteoglycan, labelled medium and cells were mixed with an equal volume of 8M-guanidine hydrochloride in 005 M-sodium acetate, ph 6-8, and chromatographed on columns of Sephadex G50 (Mitchell & Hardingham, 1981); incorporated material was eluted in the void volume and counted for the time course shown in Fig. 1. The culture medium from each dish was combined with the trypsin/edta solution used to harvest the chondrocytes; thus the level of incorporation includes both counts in the medium and counts secreted but retained in the cell layer. Duplicate dishes were harvested for each time point. To determine the hydrodynamic size and ability to aggregate with hyaluronic acid of secreted proteoglycans, SO4- labelled medium was chromatographed on Sepharose 2B columns in 0-5 M-sodium acetate, ph 6'8, in the presence of unlabelled carrier proteoglycan and hyaluronic acid (associative conditions) or 4 M-guanidine hydrochloride and 005 M- sodium acetate ph6'8 (dissociative conditions) (Mitchell & Hardingham, 1981). VQ and V, were determined as the fractions in which aggregated cartilage proteoglycan and 3 H2O, respectively, eluted. The A.' av of each peak was calculated as described by Mitchell & Hardingham (1981). Immunofluorescence The primary antibodies used were MZ15 (a monoclonal antibody to keratan sulphate; Zanetti et al. 1985) and affinity-purified rabbit antibodies to chick type I and type II collagen (von der Mark et al. 1976). Fluorescein-conjugated second antibodies were purchased from Miles Scientific. For MZ15 staining, cells were fixed in 3-7 % formaldehyde in phosphate-buffered saline (PBS) at room temperature for min, then permeabilized with methanol for 5 min on ice. For anti-collagen staining, cells were fixed in 70% ethanol in distilled water for 5 min and in 50 % ethanol, 50 % ether for 5 min. Antibody incubations were for 45 min at room temperature followed by thorough washing in PBS. Cells were mounted in Gelvatol (Monsanto) and viewed with a Zeiss Photomicroscope III. Results Proliferation and Jj SO^ incorporation by chondrocytes seeded at high or low density Chondrocytes isolated from pig articular cartilage were seeded at 10 s (low density) or (high density) per 35- mm diameter Petri dish. Duplicate dishes were labelled for 24 h with 35 SO4 and harvested 24 h later, in order to determine the number of cells per dish and the level of 35 SO4 incorporation in the medium and cell layer. Fig. 1 illustrates the changes in cell number and 3S SO4 incorporation at the two cell densities. At both densities cell number levelled off at approximately 2x per dish, but during the first 2 weeks in culture cells at low density underwent more rounds of division than cells at high density. In both cultures 35 SO4 incorporation per cell fell slowly during the first week, then more rapidly, reaching a plateau in the third week. During the first 12 days, the level of incorporation was higher in the high-density than in the low-density cultures. 3x10* Fig. 1. Number of cells per dish (A) and disintsmin ' 3S SO 4 incorporated per 10 3 cells (B) at intervals after plating at high ( ) or low (A) density F. M. Watt

3 Proteoglycan sytithesis The type of proteoglycan synthesized by ehondrocytes in culture was assessed in two ways. Sepharose 2B columns of 35 SO4-labelled medium from 21-day cultures were run under dissociative and associative conditions to determine the hydrodynamic size of the proteoglycan monomers and their ability to aggregate with hyaluronic acid. The cell layers were also stained with MZ15, a monoclonal antibody that recognizes keratan sulphate, a glycosaminoglycan that is found attached to the cartilage proteoglycan core (Zanetti et al. 1985). The Sepharose 2B profiles (Fig. 2) showed that most of the 3s SO 4 -labelled proteoglycan from high- and lowdensity cultures eluted in the void volume under associative conditions, indicating that it was aggregated with hyaluronic acid. Under dissociative conditions the proteoglycan from high-density cultures ran as a single peak (K ev 0-3), in the same position as proteoglycan isolated directly from cartilage (not shown). In lowdensity cultures there was a major peak o(k BV 0-3 and a second, less abundant, peak of K 3V 0-7. This second peak co-elutes with 35 SC>4-labelled macromolecules secreted into the culture medium by pig dermal fibroblasts (F. M. Watt, unpublished observations) and may, therefore, be non-aggregating. These results suggest that in both types of culture the majority of the 3s SO4-labelled proteoglycan was of large hydrodynamic size and was capable of aggregating with hyaluronic acid; however, a second, smaller, species was also expressed at low density Associative Dissociative Expression of keratan sulphate and effect of inhibiting proliferation The high- and low-density cultures differed in the proportion of cells expressing keratan sulphate (Table 1). This was measured by scoring the proportion of cells with extracellular or intracellular MZ15 staining in paired fluorescence and phase-contrast photographs (Fig. 3). At all the times measured (Table 1), the proportion of MZ15-positive cells was greater than 60% in the high-density cultures. In contrast, the proportion fell to less than 50% by day 5 in low-density cultures and then stabilized at about 30% from day 9 onwards. In both high- and lowdensity cultures, positive staining was usually associated with rounded cells (Fig. 3), and the proportion of rounded cells was higher in the high-density cultures. Since cells at low density underwent more rounds of cell division before confluence than cells at high density (Fig. 1), it was possible that the decline in the proportion of MZ15-positive cells was due to selective proliferation of a subpopulation of MZ15-negative cells already present in cartilage. It is known, for example, that most cells in the upper third of pig articular cartilage lack surface keratan sulphate after isolation, whereas the deeper zone cells are MZ15-positive (Zanetti et al. 1985). To test this idea, ehondrocytes were seeded at low density for 2 days to allow full attachment, then were exposed to 4^gml~' mitomycin C for 2h. Mitomycin C treatment resulted in inhibition of proliferation, as judged by inhibition of [ 3 H]thymidine incorporation (results not shown). The cultures were fixed 3 days later and the proportion of cells stained with MZ15 was measured (Table 1). The proportion had fallen from 70 % to 47 %, the same level as in control cultures. This suggests that the decline in MZ15 staining was not due to selective proliferation of an MZ15-negative subpopulation of cells. Table 1. Proportion of cells positively stained with MZ15 at intervals after plating Days in culture Plating density % Positive No. cells counted Effluent mass (g) Fig. 2. Sepharose 2B profiles of 3s SO 4 -labelled medium from cultures grown at high ( ) or low (10 5 ) cell density for 21 days. Left, associative conditions; right, dissociative conditions s 10' lo'+mc* * Cells were treated with mitomycin C two days after plating. Seeding density and chondrocyte phenotype 375

4 Fig. 3. Phase-contrast and fluorescence micrographs of chondrocytes grown at low (A,B) or high (C,D) cell density for 6 days, and then stained with MZ15. Bar, 100 f.im. Expression of type I and type II collagen Chondrocytes grown at low or high density for 21 days were fixed and stained with affinity-purified rabbit antiserum to type I or type II collagen (Fig. 4). In both cultures, type 1 collagen was detected in the extracellular matrix. Type II collagen was abundant in the highdensity cultures, but barely detectable in the lowdensity cultures. Discussion The experiments that I have reported suggest that the differentiated phenotype of pig articular chondrocytes is more stable when the cells are seeded at high density than at low density. Immunofluorescence analysis showed that cultures seeded at 10 cells per 35-mm diameter Petri dish expressed more type II collagen and keratan sulphate than cultures seeded at 10s cells per dish. 35SC>4 incorporation fell in both cultures, but was initially greater (per cell) in the high-density cultures. In the low-density cultures a small proteoglycan species was detected which was not present in the high-density cultures or extracts of cartilage (not shown). Even though the high-density cultures remained differentiated in that they continued to express type II 376 F. M. Watt collagen and large, aggregating proteoglycans, there was clearly some abnormal gene expression, because of the appearance of type I collagen in the extracellular matrix. Whether this reflects an early stage of dedifferentiation or, rather, environmental modulation of matrix synthesis remains to be determined, von der Mark (1986) has distinguished between reversible changes (modulation) of phenotype and irreversible dedifferentiation. How does seeding density influence the stability of the differentiated phenotype? One possibility is that the number of rounds of division prior to confluence is important and that in low-density cultures there is selective overgrowth by phenotypically unstable chondrocytes. Two distinct subpopulations of chondrocytes can be distinguished by their distance from the articular surface and ability to bind MZ15 (Zanetti et al. 1985). To test this idea, proliferation in low-density cultures was inhibited by exposure to mitomycin C. The proportion of cells expressing keratan sulphate declined to the same extent as in control low-density cultures. Thus, the loss of keratan sulphate-positive cells was not due to overgrowth by cells lacking keratan sulphate. A number of reports have suggested a link between differentiated gene expression and rounded chondrocyte morphology (von der Mark et al. 1977; Pacifici et

5 Fig. 4. Fluorescence micrographs of chondrocytes grown at low (A,C) or high (B,D) density for 21 days, and stained with antibodies to type II (A,B) or type I (C,D) collagen. Bar, 25 Jim. al. 1977; West et al. 1979; Benya & Shaffer, 1982; Glowacki et al. 1983). Both the high- and low-density cultures contained a mixture of rounded and spread cells, but the proportion of rounded cells was higher in the high-density cultures (Fig. 3). Thus it is possible that cell shape could provide the basis for the effect of seeding density on gene expression. This is supported by the observation of Zanetti & Solursh (1984) that cytochalasin D-induced rounding induces chondrogenesis of mesenchymal cells; and by the finding that cytochalasin D treatment of mature chondrocytes stimulates proteoglycan synthesis, an effect which is reversed when the cells respread after removal of the drug (P. G. Newman and F. M. Watt, unpublished results). Indeed, cell shape may regulate differentiated gene expression in a number of cell types (Watt, 1986). In addition to its effect on cell morphology, seeding density could affect intercellular communication, either via cell cell contact or secreted factors. It could also influence the ability of chondrocytes to incorporate newly synthesized collagen and proteoglycan into the extracellular matrix. Clearly further experiments are necessary to test these ideas, and techniques are required that allow changes in one variable without affecting others. In this regard, new methods that allow precise control over cell shape, in the absence of cell-cell contact, are likely to be valuable (O'Neill et al. 1986; F. M. Watt and J. Dudhia, unpublished results). Much of the work reported in this paper was carried out at the Kennedy Institute of Rheumatology in Hammersmith with the financial support of the Arthritis and Rheumatism Council, which I gratefully acknowledge. I thank I. R. Kill for expert technical assistance. 1 am grateful to K. von der Mark for providing the affinity-purified anti-collagen antibodies. References AMADIO, P. C, EHRLICH, M. G. & MANKIN, H. J. (1983). Matrix synthesis in high-density cultures of bovine epiphyseal plate chondrocytes. Connect. Tiss. Res. 11, BENYA, P. D. & SHAFFER, J. D. (1982). Dedifferentiated chondrocytes reexpress the differentiated collagen phenotype when cultured in agarose gels. Cell 30, GLOWACKI, J., TREPMAN, E. & FOLKMAN, J. (1983). Cell shape and phenotypic expression in chondrocytes. five. Soc. exp. Biol. Med. Ill, Seeding density and chondrocyte phenotype 377

6 HOLTZER, H., ABBOT, J., LASH, J. & HOLTZER, S. (1960). The loss of phenotypic traits by differentiated cells in vitro. I. Dedifferentiation of cartilage cells. Proc. natn. Acad. Sci. U.SA. 46, KUETTNER, K. E., MEMOLI, V. A., PAULI, B. U., WROBEL, N. C, THONAR, E. J.-M. A. & DANIEL, J. C. (1982a). Synthesis of cartilage matrix by mammalian chondrocytes in vitro. II. Maintenance of collagen and proteoglycan phenotype.j. Cell Biol. 93, KUETTNER, K. E., PAULI, B. U., GALL, G., MEMOLI, V. A. & SCHENK, R. K. (1982*). Synthesis of cartilage matrix by mammalian chondrocytes in vitro. I. Isolation, culture characteristics, and morphology. J. Cell Biol. 93, MITCHELL, D. & HARDINGHAM, T. (1981). The effects of cycloheximide on the biosynthesis and secretion of proteoglycans by chondrocytes in culture. Biochem. jf. 196, O'NEILL, C, JORDAN, P. & IRELAND, G. (1986). Evidence for two distinct mechanisms of anchorage stimulation in freshly explanted and 3T3 Swiss mouse fibroblasts. Cell 44, PACIFICI, M., BOETTIGER, D., ROBY, K. & HOLTZER, H. (1977). Transformation of chondroblasts by Rous sarcoma virus and synthesis of the sulfated proteoglycan matrix. Cell 11, SRIVASTAVA, V. M. L., MALEMUD, C. J. & SOKOLOFF, L. (1974). Chondroid expression by lapine articular chondrocytes in spinner culture following monolayer growth. Connect. Tiss. Res. 2, VON DER MARK, H., VON DER MARK, K. & GAY, S. (1976). Study of differential collagen synthesis during development of the chick embryo by immunofluorescence. 1. Preparation of collagen type I and type II specific antibodies and their application to early stages of the chick embryo. Devi Biol. 48, VON DER MARK, K. (1986). Differentiation, modulation and dedifferentiation of chondrocytes. Rheumatology 10, VON DER MARK, K., GAUSS, V., VON DER MARK, H. & MULLER, P. (1977). Relationship between cell shape and type of collagen synthesised as chondrocytes lose their cartilage phenotype in culture. Nature, Land. 267, WATT, F. M. (1986). The extracellular matrix and cell shape. Trends Biochem. Sci. 11, WEST, C. M., LANZA, R., ROSENBLOOM, J., LOWE, M., HOLTZER, H. & AVDALOVIC, N. (1979). Fibronectin alters the phenotypic properties of cultured chick embryo fibroblasts. Cell 17, ZANETTI, M., RATCLIFFE, A. & WATT, F. M. (1985). Two subpopulations of differentiated chondrocytes identified with a monoclonal antibody to keratan sulfate. J. Cell Biol. 101, ZANETTI, N. C. & SOLURSH, M. (1984). Induction of chondrogenesis in limb mesenchymal cultures by disruption of the actin cytoskeleton. J. Cell Biol. 99, (Received 18 November Accepted 23 November 1987) 378 F. M. Watt