Effect of cell shape on proteinase secretion by epithelial cells
Effect of cell shape on proteinase secretion by epithelial cells
H. L. HONG and D. M. BRUNETTE* Department of Oral Biology, Faculty of Dentistry, University of British Columbia, 2199 Westbrook Mall, Vancouver, British Columbia, Canada V6T 1Z7 * Author for correspondence
Summary Cell proliferation has been found to correlate with increased secretion of proteinases, such as plasminogen activator, in several different cell populations. In addition, the shape of the cell may also play a role in regulating proteinase secretion. However, the relationship between cell proliferation, cell shape and proteinase secretion has not been studied in diploid epithelial (E) cells cultured from porcine periodontal ligament (PL). We have modified PLE cell shape by physical means, such as growth on less-adhesive substrata and mechanical stretching, and by exposure to cholera toxin and 12-O-tetradecanoylphorbol-13acetate (TPA). Neutral proteinase and plasminogen activator secretion were found to correlate
with cell shape, the more round the cells, the greater the amount of proteinase secreted. PLE cells, stimulated to proliferate by cholera toxin or dibutyryl cyclic AMP, were more spread than control cells, but secreted less neutral proteinase and plasminogen activator. TPA stimulated cell proliferation slightly but, in contrast to cholera toxin, increased cell rounding and the secretion of neutral proteinase and plasminogen activator. Thus proteinase secretion was related more to cell shape than to cell proliferation.
Introduction
cells were used in this study since epithelial cultures differ in several ways from fibroblast cultures and it is from epithelium that the majority of human cancers arise (Weinsteinet al. 1978). Cell shape was altered by physical means, including growth on less-adhesive substrata (poly(HEMA)-coated dishes and grooved surfaces) and mechanical stretching. Cells were stimulated to proliferate by cholera toxin and dibutyryl cyclic AMP, which also caused cells to flatten, and by 12-O-tetradecanoylphorbol-13-acetate (TPA), which increased cell rounding. Secretion of PA and neutral proteinase (NP) by PLE cells (epithelial cells cultured from porcine periodontal ligament) was found to be related more to cell shape than to rate of proliferation.
For several transformed or malignant cell populations, tumour growth in vivo has been shown to be correlated with increased secretion of proteinases such as plasminogen activator (PA) in vitro (Mak et al. 1976; Butler et al. 1979; Mira-y-Lopezeia/. 1983; Pollacks al. 1975). This correlation between growth and pro : teinase secretion may be important in understanding tumour invasion because there is considerable evidence that tumour invasion is accompanied by proteolysis (Woolley, 1984; Quigley, 1979). It is not clear, however, whether the increased amount of proteinases secreted by transformed cell populations is associated specifically with the increased rate of proliferation, or whether other cellular properties, such as cell shape, are involved. Cell shape has been shown to regulate proteinase secretion in normal synovial fibroblasts (Aggeler et al. 1982), and has also been implicated in the regulation of proliferation of normal cells in vitro (Folkman & Moscona, 1978). This study was designed to investigate the relationship between cell proliferation, cell shape and proteinase secretion. Epithelial Journal of Cell Science 87, 259-267 (1987) Printed in Great Britain © The Company of Biologists Limited 1987
Key words: epithelial cells, cell shape, proteinases, cell proliferation.
Materials and methods Cell culture PLE cells were obtained from porcine periodontal ligament as described by Brunette et al. (1976) and separated from fibroblasts by their tendency to be more resistant than fibroblasts to removal by trypsin (Owens, 1974) or by a collagenase separation technique (Kanoza et al. 1978). The
259
cells were cultured in a minimal essential medium (aMEM) plus 15 % foetal calf serum (FCS) (Flow, MD) with penicillin G (Sigma, St Louis, MO), lOOfjgml"1, gentamycin (Sigma), 50 fig ml"', and amphoteracin B (Gibco, Grand Island, NY), 3/igml" 1 , at 37°C in a humidified atmosphere of air plus 5 % CGv For assaying proteinase secretion, PLE cells were plated on 60mm Falcon dishes at 4X 10s cells/dish and incubated overnight in a"-MEM + 15% FCS. The medium was removed and replaced with a modified low serum medium with lmM-K + and01mM-Ca 2 + , designated /3-MEM, plus 0-5% dialysed FCS (described by Brunette, 1984a), for 2-4 days to permit sufficient secretion of enzyme for the activity to be assayed. /3-MEM was used in these studies because it supports PLE cell growth at low concentrations of serum in the presence of growth promoters such as cholera toxin. It was desirable to keep the amount of serum low in these experiments because serum contains inhibitors to proteinases. At each time point of medium collection, two cultures were trypsinized and counted on an electronic cell counter (Coulter Electronics, Inc., Hialeah, Florida) to monitor cell growth. In some experiments, as noted in the text, /3-MEM was supplemented by the addition of 1 ngml" 1 cholera toxin (Sigma), or 0 5 mM-dibutyryl cyclic-3'5'-adenosine monophosphate (BtjcAMP) (Sigma), or 12-O-tetradecanoylphorbol-13-acetate (TPA) (Sigma) at either l n g m l " 1 , or lOngml" 1 . PLE cells plated at 4X10 5 cells/dish were processed for scanning electron microscopy (SEM) as described by Brunette et al. (1983).
Physical methods for altering cell shape Mechanical stretching. Mechanical stretching was applied using the method of Hasegawa et al. (1985). PLE cells were plated on Petriperm dishes (Tekmar, Cincinnati, OH). Stretching the flexible plastic membrane of the Petriperm dish, to which the cells are firmly attached by hemidesmosomes, stretches the cells and makes them more flattened than cells in unstretched dishes (Brunette, 19846). Conversely, growing the cells on a stretched membrane and then taking the Petriperm dish off the template allows the plastic to return to its initial state and the attached cells to become more rounded. PLE cells were plated at 8x 105 cells/dish in ff-MEM + 15 % FCS and grown for 3 days under the conditions stated below. For the detection of proteinase secretion the cells were washed twice with /3-MEM + 0-5 % dialysed FCS and incubated in the same medium for 2 days. To determine the effect of stretching two groups were used. (1) Unstretched control cultures: the cells were grown on Petriperm dishes over the experimental period. (2) Stretched cultures: the Petriperm dishes were stretched during the 2-day proteinase collection period. To determine the effect of relaxing cells grown on a stretched membrane another two groups were used. (1) Controls: the cells were plated and grown on a stretched membrane and the dish remained stretched during the 2-day proteinase collection period. (2) Relaxed: the cells were plated and grown on a stretched membrane for 3 days and then the dishes were taken off the templates during the 2-day proteinase collection period. Preparation ofgrooved substratum. With the exception of macrophages (Rich & Harris, 1981), cultured cells adopt a
260
H. L. Hong and D. M. Brunette
more spherical shape when cultured on grooved (Rovensky et al. 1971) or rough substrata. Silicon wafers with V-shaped grooves were produced as described by Brunette et al. (1983). The V-shaped grooves were 79 jxm wide at the top of the V and 60fim deep; the grooves separated by flat areas 13 [im in width. Impressions of the groove pattern on the silicon wafer were made with Exaflex (G-C Dental Industrial Corp., Japan). The groove pattern was then reproduced onto Epotek resin (Epoxy Technology Inc., Billerica, MA) in 60 mm Falcon dishes. Control dishes with a flat surface were also made with Epotek resin. Preparation of poly (HEMA) plates. Cell shape was also altered by changing the culture dish surface adhesiveness. Tissue culture dishes were coated with various dilutions of a 12% stock solution of poly-2-hydroxyethyl methacrylate (poly(HEMA)) (Polysciences, Inc., Warrington, PA) according to the method of Folkman & Moscona (1978). Following the convention of Folkman & Moscona (1978) the poly(HEMA) concentrations stated are the amounts of dilution of the 12% (v/v) stock solution in ethanol. For example, to obtain a concentration of 10~3, 0'lml of the stock solution was added to 99'9ml ethanol.
Assays Cyclic AMP. PLE cell cultures were processed as described by Oey et al. (1974) and intracellular cyclic AMP was determined using the Becton Dickinson cAMP radioimmunoassay kit (125I) (Becton Dickinson Immunodiagnostics, Orangeburg, NY). Neutral proteinase assay. The harvested media were concentrated 20 times using Millipore CX-10 filters (Mr cutoff 10000) (Millipore Corp., Bedford, MA) and dialysed against 50mM-Tris-HC1 buffer containing 200mM-sodium chloride and 5 mM-calcium chloride (pH7-5) at 4°C. The neutral proteinase activity in the media was measured using A20C0II (Calbiochem, San Diego, CA) as the substrate, as described by Pettigrewe* al. (1978). Activation of the latent enzyme was achieved by the addition of mersalyl acid (Sigma) to the assay at a final concentration of 2mM. All assays were done in duplicate and a mersalyl activation control was included. The degradation of Azocoll was linear with respect to enzyme concentration as well as incubation time up to 56h. As defined by Werb & Reynolds (1974), 1 unit of enzyme activity was equivalent to 1 mg Azocoll splubihzed per h at 37°C. Plasminogen activator assay. The two-stage assay described by Jackson et al. (1981) was used to determine PA activity. In the first stage, plasminogen is converted into plasmin by PA and in the second stage, the plasmin acts on the substrate S-2251. The reaction is initiated by adding 20/il of the sample to 40 jJ.\ of plasminogen (Kabi, Stockholm, Sweden) ( l u n i t m F 1 ) . After 30min, 80^1 of the substrate mixture (S-2251 (Kabi) ( 5 m g m r ' ) and 1-77MNaCl in 32mM-Tris- HC1 buffer mixed in a 2:3 ratio) was added. The second stage of the reaction was terminated after 30min by the addition of 0-8ml of 40% acetic acid. All assays were done in duplicate and included controls with buffer only and without plasminogen. The standard curve was prepared using increasing concentrations of streptokinase (Behring Institute, Marburg, West Germany). The assay was linear with respect to increasing amounts of the
sample up to an absorbance of 0'9 at 405 nm. Activities were expressed as IU 10~5 cells. The coefficient of variation for the NP and PA assays was 3 % and 6%, respectively.
Statistical analysis In experiments where replicate cultures were used, the data were analysed using the analysis of variance (ANOVA). In other experiments, the sign test (Zar, 1974) was used to test whether the differences observed under different conditions were due to chance alone.
Results Effect of cell population density As cell population density has been found to influence many properties of epithelial cells in culture including proliferation (Zetterberg & Auer, 1970), prostaglandin production (Bireket* al. 1980) and dome formation (Young et al. 1978), the effect of cell population density on proteinase secretion was investigated. PLE cells were plated at 2, 4, 8, 12 and 15 (X 10s) cells/dish in /S-MEM + 0-5% dialysed FCS. These conditions kept the cells in a non-proliferative state. The amount of activity present in the medium was assayed at days 7 and 10. As the cell population density increased, both NP (Fig. 1) and PA secretion (Fig. 2), measured in units 10~ s cells, decreased. Therefore, experiments were designed so that the cell densities were normally in the plateau regions of the curves where the effects of cell population density on proteinase secretion were small. Effect of cholera toxin and dibutyryl cyclic AMP PLE cells were induced to proliferate by the addition of cholera toxin at Ing ml" 1 to cell cultures as described by Brunette (1984a). By day 10, there was a significant ( P < 0 - 0 1 ) , increase in cell number in the cholera toxin-treated group (Fig. 3). Cholera toxin-
treated cells also became more flattened compared to control cultures as described by Brunette (1984a). Thus in this experiment, both an increase in cell proliferation and an altered cell shape were seen. Secretion of proteinases, however, was reduced significantly (P 0 5
J
0 10~ 5xlO" 10"' poly(HEMA) concentration Fig. 8. Effect of poly(HEMA) concentration on activity. Duplicate groups of E cells were plated in dishes precoated with various concentrations of poly(HEMA) solution. Medium collected on days 3 (open) and 7 (striped) were assayed for PA activity. Error bars indicate S.E.
It CO
4
Q.
3
7
Time (days) Fig. 6. Effect of grooved substrata. E cells were plated at 8x 105 cells/dish. Proteinase activity from cultures grown on flat substrata (open bars) and grooved substrata (stippled bars) as measured on days 3 and 7. The data shown are from one representative experiment that was repeated twice. The trends were the same in all experiments.
5x10-" 0 1CT3 poly(HEMA) concentration
10-
Fig. 7. Effect of poly(HEMA) concentration on NP activity. Duplicate groups of E cells were plated in dishes precoated with various concentrations of poly(HEMA) solution. The medium was collected on days 3 (open) and 7 (striped) and assayed for NP activity. Error bars indicate S.E.
Auer, 1970) decreases with population density. Similarly, a number of metabolic functions in mature rat hepatocytes have been shown to be regulated by cell population density (Jurin & McCune, 1985; Nakamura et al. 1983). The mechanisms underlying these
effects are not known, but the effect of cell population density must be considered when designing experiments that include treatments that influence cell multiplication. The addition of cholera toxin, Bt2cAMP or TPA to a medium containing a low concentration of serum stimulated cell proliferation and permitted direct comparisons of proliferating and quiescent cells. Because serum contains inhibitors to proteinases, the use of a low-serum medium enabled NP activity to be assayed in the proliferating or quiescent cultures. Although TPA increased both PLE cell number and PA and NP secretion, cholera toxin and Bt2cAMP caused an increase in PLE cell proliferation but decreased the amount of PA and NP secreted. Our observations, obtained with normal diploid PLE cells (Brunette et al. 1979), thus do not provide support for the hypothesis, based originally on observations on tumour cells (Mak et al. 1976; Butler et al. 1979; Mira-y-Lopez et al. 1983), that PA secretion is positively correlated with cell multiplication. However, it should be noted that the production of a high level of PA in culture is a distinctive property of many transformed cells, and that the PLE cells used in this study secrete only small amounts of this enzyme. It is possible that PA secretion is regulated differently in transformed cells than in the normal cells used in this study. In summary, six methods of changing cell shape including chemical (cholera toxin, Bt2cAMP, TPA) or physical (mechanical stretching, poly(HEMA) dishes, grooved substrata) stimuli, consistently showed that, irrespective of the method used, the more-flattened cells secreted less proteinase than the more-rounded cells. There was, however, no consistent relationship between the amount of proteinase secreted and the rate of proliferation. Thus for diploid PLE cells proteinase E-cell shape and proteinase secretion
265
secretion was related more to cell shape than to cell proliferation. References AGGELER, J., FRISCH, S. M. & WERB, Z. (1984). Changes
in cell shape correlate with collagenase gene expression in rabbit synovial fibroblasts. J. Cell Biol. 98, 1662-1671. AGGELER, J., KAPP, L. N., TSENG, S. C. G. & WERB, Z.
Molecular Action of Toxins and Viruses (ed. P. Cohen & S. van Heyningen), pp. 33-49. Amsterdam: Elsevier Biomedical Press. JURIN, R. R. & MCCUNE, S. A. (1985). Effect of cell density on metabolism in isolated rat hepatocytes. J. cell. Physiol. 123, 442-448. KANOZA, R. J. J., BRUNETTE, D. M., PURDON, A. D. &
SODEK, J. (1978). Isolation and identification of epithelial-like cells in culture by a collagenase separation technique. In Vitro 14, 746-753. MAK, T. W., RUTLEDGE, G. & SUTHERLAND, D. J. A.
(1976). Androgen-dependent fibrinolytic activity in a murine mammary carcinoma (Shionogi SC-115 cells) in vitro. Cell 7, 223-226.
(1982). Regulation of protein synthesis in Chinese hamster ovary cells by cell cycle position and cell density. Expl Cell Res. 139, 275-283. BEN-ZE'EV, A. (1985). Cell shape, the complex cellular networks, and gene expression. In Gene Expression in Muscle (ed. R. C. Strohman & S. Wolf), pp. 23-53. New York: Plenum.
MAKMAN, M. H., MORRIS, S. A. & AHN, H. S. (1972).
BIREK, C , BRUNETTE, D. M., HEERSCHE, J. N. M.,
MIRA-Y-LOPEZ, R., REICH, E. & OSSOWSKI, L. (1983).
WANG, H. M. & JOHNSTON, M. G. (1980). A reverse
hemolytic plaque assay for the detection of prostaglandin production by individual cells in vitro. Expl Cell Res. 129, 95-101. BRUNETTE, D. M. (1984a). Cholera toxin and dibutyryl cyclic-AMP stimulate the growth of epithelial cells derived from epithelial rests from porcine periodontal ligament. Archs oral Biol. 29, 303-309. BRUNETTE, D. M. (19846). Mechanical stretching increases the number of epithelial cells synthesizing DNA in culture. J.CellSci. 1, 35-45. BRUNETTE, D. M., HEERSCHE, J. N. M., PURDON, A. D., SODEK, J., MOE, H. K. & ASSURAS, J. N. (1979). In
vitro cultural parameters and protein and prostaglandin secretion of epithelial cells derived from porcine cell rests of Malassez. Archs oral Biol. 24, 199-203. BRUNETTE, D. M., KENNER, G. S. & GOULD, T. R. L.
(1983). Grooved titanium surfaces orient growth and migration of cells from human gingival explants. jf. dent. Res. 62, 1045-1048. BRUNETTE, D. M., MELCHER, A. H. & MOE, H. K.
(1976). Culture and origin of epithelial-like and fibroblast-like cells from porcine periodontal ligament explants and cell suspensions. Archs oral Biol. 21, 393-400. BUTLER, W. B., KIRKLAND, W. L. & JORGENSEN, T. L.
(1979). Induction of plasminogen activator by estrogen in a human breast cancer cell line (MCF-7). Biochem. biophys. Res. Commun. 90, 1328-1334. FoLKMAN, J. & MOSCONA, A. (1978). Role of cell shape in growth control. Nature, Land. 273, 345-349. HASEGAWA, S., SATO, S., SAITO, S., SUZUKI, Y. &
BRUNETTE, D. M. (1985). Mechanical stretching increases the number of cultured bone cells synthesizing DNA and alters their pattern of protein synthesis. Calcif. Tiss. Int. 37, 431-437. JACKSON, K. W., ESMON, N. & TANG, J. (1981).
Streptokinase and staphylokinase. In Methods in Enzymology, vol. 80 (ed. L. Lorand), pp. 387-388. New York: Academic Press. JOHNSON, G. L. (1982). Cholera toxin action and the regulation of hormone-sensitive adenylate cyclase. In
266
H. L. Hong and D. M. Brunette
Cyclic nucleotides. In Growth, Nutrition and Metabolism of Cells in Culture (ed. G. H. Rothblat & V. J. Cristofalo), pp. 295-336. New York: Academic Press. Modulation of plasominogen activator in rodent mammary tumors by hormones and other effectors. Cancer 43, 5467-5477. NAKAMURA, T. (1983). Density-dependent controls of growth and various cellular activities of mature rat hepatocytes in primary culture. In Isolation, Characterization and Use of Hepatocytes (ed. R. A. Harris & N. W. Cornell), pp. 193-198. New York: Elsevier Biomedical. OEY, J., VOGEL, A. & POLLACK, R. (1974). Intracellular
cyclic AMP concentration responds specifically to growth regulation by serum. Proc. natn. Acad. Sci. U.SA. 71, 694-698. OWENS, R. B. (1974). Glandular epithelial cells from mice: A method for selective cultivation. J. natn. Cancer Inst. 52, 1375-1378. PETTIGREW, D. W., H O , G. H., SODEK, J., BRUNETTE, D.
M. & WANG, H. M. (1978). Effect of oxygen tension and indomethacin on production of collagenase and neutral proteinase enzymes and their latent forms by porcine gingival explants in culture. Archs oral Biol. 23, 767-777. POLLACK, R., RISSER, R., CONLON, S., FREEDMAN, V.,
SHIN, S. & RIFKIN, D. B. (1975). Production of
plasminogen activator and colonial growth in semisolid medium are in vitro correlates of tumorigenicity in the immune-deficient nude mouse. In Proteases and Biological Control, vol. 2 (ed. E. Reich, D . B. Rifkin & E. Shaw), pp. 885-899. New York: Cold Spring Harbor Laboratory Press. QUIGLEY, J. P. (1979). Proteolytic enzymes of normal and malignant cells. In Surfaces of Normal and Malignant Cells (ed. R. O. Hynes), pp. 247-285. New York: Wiley. RICH, A. & HARRIS, A. K. (1981). Anomolous preferences of cultured macrophages for hydrophobic and roughened substrata. J. Cell Sci. 50, 1-7. ROVENSKY, Y. A., SLAVNAJA, I. L. & VASLIER, J. M.
(1971). Behaviour of fibroblast-like cells on grooved surfaces. Expl Cell Res. 65, 193-201. WEINSTEIN, I. B., WlGLER, M. & PlETROPAOLO, C. (1977). The action of tumor-promoting agents in cell culture. In
Origins of Human Cancer (ed. H. H. Hiatt, J. D. Watson & J. A. Winsten), pp. 751-772. New York: Cold Spring Harbor Laboratory Press. WEINSTEIN, I. B., WIGLER, M., YAMASAKI, H., L E E , L.,
FISHER, P. B. & MUFSON, A. (1978). Regulation of the
expression of certain biologic markers of neoplasia. In Biological Markers of Neoplasia: Basic and Applied Aspects (ed. R. W. Ruddon), pp. 451-471. North Holland, New York: Elsevier. WERB, Z., HEMBRY, R. M., MURPHY, G. & AGGELER, J.
(1986). Commitment to expression of the metalloendopeptidases, collagenase and stromelysin: relationship of inducing events to changes in cytoskeletal architecture. J. Cell Biol. 102, 697-702. WERB, Z. & REYNOLDS, J. J. (1974). Stimulation by
endocytosis of the secretion of collagenase and neutral proteinase from rabbit synovial fibroblasts. J. exp. Med. 140, 1482-1497.
WIGLER, M. & WEINSTEIN, I. B. (1976). Tumor promoter induces plasminogen activator. Nature, Land. 259, 232-233. WOOLLEY, D. E. (1984). Collagenolytic mechanisms in tumour cell invasion. In Cancer Metastasis Reviews, vol. 3, pp. 361-372. Boston, MA: Matinus Nijhoff. YOUNG, L. J. T., CARDIFF, R. D. & SEELEY, T . (1978).
Quantitative characterization of domes in primary mouse mammary epithelial tumor cell cultures. In Vitro 14, 895-902. ZAR, J. H. (1974). The binomial distribution. In Biostatistical Analysis, pp. 290-291. New Jersey: Prentice-Hall, Inc. ZETTERBERG, A. & AUER, G. (1970). Proliferative activity and cytochemical properties of nuclear chromatin related to local cell density of epithelial cells. Expl Cell Res. 62, 262-270. {Received 31 July 1986 -Accepted
E-cell shape and proteinase
14 October 1986)
secretion
267
H. L. HONG and D. M. BRUNETTE* Department of Oral Biology, Faculty of Dentistry, University of British Columbia, 2199 Westbrook Mall, Vancouver, British Columbia, Canada V6T 1Z7 * Author for correspondence
Summary Cell proliferation has been found to correlate with increased secretion of proteinases, such as plasminogen activator, in several different cell populations. In addition, the shape of the cell may also play a role in regulating proteinase secretion. However, the relationship between cell proliferation, cell shape and proteinase secretion has not been studied in diploid epithelial (E) cells cultured from porcine periodontal ligament (PL). We have modified PLE cell shape by physical means, such as growth on less-adhesive substrata and mechanical stretching, and by exposure to cholera toxin and 12-O-tetradecanoylphorbol-13acetate (TPA). Neutral proteinase and plasminogen activator secretion were found to correlate
with cell shape, the more round the cells, the greater the amount of proteinase secreted. PLE cells, stimulated to proliferate by cholera toxin or dibutyryl cyclic AMP, were more spread than control cells, but secreted less neutral proteinase and plasminogen activator. TPA stimulated cell proliferation slightly but, in contrast to cholera toxin, increased cell rounding and the secretion of neutral proteinase and plasminogen activator. Thus proteinase secretion was related more to cell shape than to cell proliferation.
Introduction
cells were used in this study since epithelial cultures differ in several ways from fibroblast cultures and it is from epithelium that the majority of human cancers arise (Weinsteinet al. 1978). Cell shape was altered by physical means, including growth on less-adhesive substrata (poly(HEMA)-coated dishes and grooved surfaces) and mechanical stretching. Cells were stimulated to proliferate by cholera toxin and dibutyryl cyclic AMP, which also caused cells to flatten, and by 12-O-tetradecanoylphorbol-13-acetate (TPA), which increased cell rounding. Secretion of PA and neutral proteinase (NP) by PLE cells (epithelial cells cultured from porcine periodontal ligament) was found to be related more to cell shape than to rate of proliferation.
For several transformed or malignant cell populations, tumour growth in vivo has been shown to be correlated with increased secretion of proteinases such as plasminogen activator (PA) in vitro (Mak et al. 1976; Butler et al. 1979; Mira-y-Lopezeia/. 1983; Pollacks al. 1975). This correlation between growth and pro : teinase secretion may be important in understanding tumour invasion because there is considerable evidence that tumour invasion is accompanied by proteolysis (Woolley, 1984; Quigley, 1979). It is not clear, however, whether the increased amount of proteinases secreted by transformed cell populations is associated specifically with the increased rate of proliferation, or whether other cellular properties, such as cell shape, are involved. Cell shape has been shown to regulate proteinase secretion in normal synovial fibroblasts (Aggeler et al. 1982), and has also been implicated in the regulation of proliferation of normal cells in vitro (Folkman & Moscona, 1978). This study was designed to investigate the relationship between cell proliferation, cell shape and proteinase secretion. Epithelial Journal of Cell Science 87, 259-267 (1987) Printed in Great Britain © The Company of Biologists Limited 1987
Key words: epithelial cells, cell shape, proteinases, cell proliferation.
Materials and methods Cell culture PLE cells were obtained from porcine periodontal ligament as described by Brunette et al. (1976) and separated from fibroblasts by their tendency to be more resistant than fibroblasts to removal by trypsin (Owens, 1974) or by a collagenase separation technique (Kanoza et al. 1978). The
259
cells were cultured in a minimal essential medium (aMEM) plus 15 % foetal calf serum (FCS) (Flow, MD) with penicillin G (Sigma, St Louis, MO), lOOfjgml"1, gentamycin (Sigma), 50 fig ml"', and amphoteracin B (Gibco, Grand Island, NY), 3/igml" 1 , at 37°C in a humidified atmosphere of air plus 5 % CGv For assaying proteinase secretion, PLE cells were plated on 60mm Falcon dishes at 4X 10s cells/dish and incubated overnight in a"-MEM + 15% FCS. The medium was removed and replaced with a modified low serum medium with lmM-K + and01mM-Ca 2 + , designated /3-MEM, plus 0-5% dialysed FCS (described by Brunette, 1984a), for 2-4 days to permit sufficient secretion of enzyme for the activity to be assayed. /3-MEM was used in these studies because it supports PLE cell growth at low concentrations of serum in the presence of growth promoters such as cholera toxin. It was desirable to keep the amount of serum low in these experiments because serum contains inhibitors to proteinases. At each time point of medium collection, two cultures were trypsinized and counted on an electronic cell counter (Coulter Electronics, Inc., Hialeah, Florida) to monitor cell growth. In some experiments, as noted in the text, /3-MEM was supplemented by the addition of 1 ngml" 1 cholera toxin (Sigma), or 0 5 mM-dibutyryl cyclic-3'5'-adenosine monophosphate (BtjcAMP) (Sigma), or 12-O-tetradecanoylphorbol-13-acetate (TPA) (Sigma) at either l n g m l " 1 , or lOngml" 1 . PLE cells plated at 4X10 5 cells/dish were processed for scanning electron microscopy (SEM) as described by Brunette et al. (1983).
Physical methods for altering cell shape Mechanical stretching. Mechanical stretching was applied using the method of Hasegawa et al. (1985). PLE cells were plated on Petriperm dishes (Tekmar, Cincinnati, OH). Stretching the flexible plastic membrane of the Petriperm dish, to which the cells are firmly attached by hemidesmosomes, stretches the cells and makes them more flattened than cells in unstretched dishes (Brunette, 19846). Conversely, growing the cells on a stretched membrane and then taking the Petriperm dish off the template allows the plastic to return to its initial state and the attached cells to become more rounded. PLE cells were plated at 8x 105 cells/dish in ff-MEM + 15 % FCS and grown for 3 days under the conditions stated below. For the detection of proteinase secretion the cells were washed twice with /3-MEM + 0-5 % dialysed FCS and incubated in the same medium for 2 days. To determine the effect of stretching two groups were used. (1) Unstretched control cultures: the cells were grown on Petriperm dishes over the experimental period. (2) Stretched cultures: the Petriperm dishes were stretched during the 2-day proteinase collection period. To determine the effect of relaxing cells grown on a stretched membrane another two groups were used. (1) Controls: the cells were plated and grown on a stretched membrane and the dish remained stretched during the 2-day proteinase collection period. (2) Relaxed: the cells were plated and grown on a stretched membrane for 3 days and then the dishes were taken off the templates during the 2-day proteinase collection period. Preparation ofgrooved substratum. With the exception of macrophages (Rich & Harris, 1981), cultured cells adopt a
260
H. L. Hong and D. M. Brunette
more spherical shape when cultured on grooved (Rovensky et al. 1971) or rough substrata. Silicon wafers with V-shaped grooves were produced as described by Brunette et al. (1983). The V-shaped grooves were 79 jxm wide at the top of the V and 60fim deep; the grooves separated by flat areas 13 [im in width. Impressions of the groove pattern on the silicon wafer were made with Exaflex (G-C Dental Industrial Corp., Japan). The groove pattern was then reproduced onto Epotek resin (Epoxy Technology Inc., Billerica, MA) in 60 mm Falcon dishes. Control dishes with a flat surface were also made with Epotek resin. Preparation of poly (HEMA) plates. Cell shape was also altered by changing the culture dish surface adhesiveness. Tissue culture dishes were coated with various dilutions of a 12% stock solution of poly-2-hydroxyethyl methacrylate (poly(HEMA)) (Polysciences, Inc., Warrington, PA) according to the method of Folkman & Moscona (1978). Following the convention of Folkman & Moscona (1978) the poly(HEMA) concentrations stated are the amounts of dilution of the 12% (v/v) stock solution in ethanol. For example, to obtain a concentration of 10~3, 0'lml of the stock solution was added to 99'9ml ethanol.
Assays Cyclic AMP. PLE cell cultures were processed as described by Oey et al. (1974) and intracellular cyclic AMP was determined using the Becton Dickinson cAMP radioimmunoassay kit (125I) (Becton Dickinson Immunodiagnostics, Orangeburg, NY). Neutral proteinase assay. The harvested media were concentrated 20 times using Millipore CX-10 filters (Mr cutoff 10000) (Millipore Corp., Bedford, MA) and dialysed against 50mM-Tris-HC1 buffer containing 200mM-sodium chloride and 5 mM-calcium chloride (pH7-5) at 4°C. The neutral proteinase activity in the media was measured using A20C0II (Calbiochem, San Diego, CA) as the substrate, as described by Pettigrewe* al. (1978). Activation of the latent enzyme was achieved by the addition of mersalyl acid (Sigma) to the assay at a final concentration of 2mM. All assays were done in duplicate and a mersalyl activation control was included. The degradation of Azocoll was linear with respect to enzyme concentration as well as incubation time up to 56h. As defined by Werb & Reynolds (1974), 1 unit of enzyme activity was equivalent to 1 mg Azocoll splubihzed per h at 37°C. Plasminogen activator assay. The two-stage assay described by Jackson et al. (1981) was used to determine PA activity. In the first stage, plasminogen is converted into plasmin by PA and in the second stage, the plasmin acts on the substrate S-2251. The reaction is initiated by adding 20/il of the sample to 40 jJ.\ of plasminogen (Kabi, Stockholm, Sweden) ( l u n i t m F 1 ) . After 30min, 80^1 of the substrate mixture (S-2251 (Kabi) ( 5 m g m r ' ) and 1-77MNaCl in 32mM-Tris- HC1 buffer mixed in a 2:3 ratio) was added. The second stage of the reaction was terminated after 30min by the addition of 0-8ml of 40% acetic acid. All assays were done in duplicate and included controls with buffer only and without plasminogen. The standard curve was prepared using increasing concentrations of streptokinase (Behring Institute, Marburg, West Germany). The assay was linear with respect to increasing amounts of the
sample up to an absorbance of 0'9 at 405 nm. Activities were expressed as IU 10~5 cells. The coefficient of variation for the NP and PA assays was 3 % and 6%, respectively.
Statistical analysis In experiments where replicate cultures were used, the data were analysed using the analysis of variance (ANOVA). In other experiments, the sign test (Zar, 1974) was used to test whether the differences observed under different conditions were due to chance alone.
Results Effect of cell population density As cell population density has been found to influence many properties of epithelial cells in culture including proliferation (Zetterberg & Auer, 1970), prostaglandin production (Bireket* al. 1980) and dome formation (Young et al. 1978), the effect of cell population density on proteinase secretion was investigated. PLE cells were plated at 2, 4, 8, 12 and 15 (X 10s) cells/dish in /S-MEM + 0-5% dialysed FCS. These conditions kept the cells in a non-proliferative state. The amount of activity present in the medium was assayed at days 7 and 10. As the cell population density increased, both NP (Fig. 1) and PA secretion (Fig. 2), measured in units 10~ s cells, decreased. Therefore, experiments were designed so that the cell densities were normally in the plateau regions of the curves where the effects of cell population density on proteinase secretion were small. Effect of cholera toxin and dibutyryl cyclic AMP PLE cells were induced to proliferate by the addition of cholera toxin at Ing ml" 1 to cell cultures as described by Brunette (1984a). By day 10, there was a significant ( P < 0 - 0 1 ) , increase in cell number in the cholera toxin-treated group (Fig. 3). Cholera toxin-
treated cells also became more flattened compared to control cultures as described by Brunette (1984a). Thus in this experiment, both an increase in cell proliferation and an altered cell shape were seen. Secretion of proteinases, however, was reduced significantly (P 0 5
J
0 10~ 5xlO" 10"' poly(HEMA) concentration Fig. 8. Effect of poly(HEMA) concentration on activity. Duplicate groups of E cells were plated in dishes precoated with various concentrations of poly(HEMA) solution. Medium collected on days 3 (open) and 7 (striped) were assayed for PA activity. Error bars indicate S.E.
It CO
4
Q.
3
7
Time (days) Fig. 6. Effect of grooved substrata. E cells were plated at 8x 105 cells/dish. Proteinase activity from cultures grown on flat substrata (open bars) and grooved substrata (stippled bars) as measured on days 3 and 7. The data shown are from one representative experiment that was repeated twice. The trends were the same in all experiments.
5x10-" 0 1CT3 poly(HEMA) concentration
10-
Fig. 7. Effect of poly(HEMA) concentration on NP activity. Duplicate groups of E cells were plated in dishes precoated with various concentrations of poly(HEMA) solution. The medium was collected on days 3 (open) and 7 (striped) and assayed for NP activity. Error bars indicate S.E.
Auer, 1970) decreases with population density. Similarly, a number of metabolic functions in mature rat hepatocytes have been shown to be regulated by cell population density (Jurin & McCune, 1985; Nakamura et al. 1983). The mechanisms underlying these
effects are not known, but the effect of cell population density must be considered when designing experiments that include treatments that influence cell multiplication. The addition of cholera toxin, Bt2cAMP or TPA to a medium containing a low concentration of serum stimulated cell proliferation and permitted direct comparisons of proliferating and quiescent cells. Because serum contains inhibitors to proteinases, the use of a low-serum medium enabled NP activity to be assayed in the proliferating or quiescent cultures. Although TPA increased both PLE cell number and PA and NP secretion, cholera toxin and Bt2cAMP caused an increase in PLE cell proliferation but decreased the amount of PA and NP secreted. Our observations, obtained with normal diploid PLE cells (Brunette et al. 1979), thus do not provide support for the hypothesis, based originally on observations on tumour cells (Mak et al. 1976; Butler et al. 1979; Mira-y-Lopez et al. 1983), that PA secretion is positively correlated with cell multiplication. However, it should be noted that the production of a high level of PA in culture is a distinctive property of many transformed cells, and that the PLE cells used in this study secrete only small amounts of this enzyme. It is possible that PA secretion is regulated differently in transformed cells than in the normal cells used in this study. In summary, six methods of changing cell shape including chemical (cholera toxin, Bt2cAMP, TPA) or physical (mechanical stretching, poly(HEMA) dishes, grooved substrata) stimuli, consistently showed that, irrespective of the method used, the more-flattened cells secreted less proteinase than the more-rounded cells. There was, however, no consistent relationship between the amount of proteinase secreted and the rate of proliferation. Thus for diploid PLE cells proteinase E-cell shape and proteinase secretion
265
secretion was related more to cell shape than to cell proliferation. References AGGELER, J., FRISCH, S. M. & WERB, Z. (1984). Changes
in cell shape correlate with collagenase gene expression in rabbit synovial fibroblasts. J. Cell Biol. 98, 1662-1671. AGGELER, J., KAPP, L. N., TSENG, S. C. G. & WERB, Z.
Molecular Action of Toxins and Viruses (ed. P. Cohen & S. van Heyningen), pp. 33-49. Amsterdam: Elsevier Biomedical Press. JURIN, R. R. & MCCUNE, S. A. (1985). Effect of cell density on metabolism in isolated rat hepatocytes. J. cell. Physiol. 123, 442-448. KANOZA, R. J. J., BRUNETTE, D. M., PURDON, A. D. &
SODEK, J. (1978). Isolation and identification of epithelial-like cells in culture by a collagenase separation technique. In Vitro 14, 746-753. MAK, T. W., RUTLEDGE, G. & SUTHERLAND, D. J. A.
(1976). Androgen-dependent fibrinolytic activity in a murine mammary carcinoma (Shionogi SC-115 cells) in vitro. Cell 7, 223-226.
(1982). Regulation of protein synthesis in Chinese hamster ovary cells by cell cycle position and cell density. Expl Cell Res. 139, 275-283. BEN-ZE'EV, A. (1985). Cell shape, the complex cellular networks, and gene expression. In Gene Expression in Muscle (ed. R. C. Strohman & S. Wolf), pp. 23-53. New York: Plenum.
MAKMAN, M. H., MORRIS, S. A. & AHN, H. S. (1972).
BIREK, C , BRUNETTE, D. M., HEERSCHE, J. N. M.,
MIRA-Y-LOPEZ, R., REICH, E. & OSSOWSKI, L. (1983).
WANG, H. M. & JOHNSTON, M. G. (1980). A reverse
hemolytic plaque assay for the detection of prostaglandin production by individual cells in vitro. Expl Cell Res. 129, 95-101. BRUNETTE, D. M. (1984a). Cholera toxin and dibutyryl cyclic-AMP stimulate the growth of epithelial cells derived from epithelial rests from porcine periodontal ligament. Archs oral Biol. 29, 303-309. BRUNETTE, D. M. (19846). Mechanical stretching increases the number of epithelial cells synthesizing DNA in culture. J.CellSci. 1, 35-45. BRUNETTE, D. M., HEERSCHE, J. N. M., PURDON, A. D., SODEK, J., MOE, H. K. & ASSURAS, J. N. (1979). In
vitro cultural parameters and protein and prostaglandin secretion of epithelial cells derived from porcine cell rests of Malassez. Archs oral Biol. 24, 199-203. BRUNETTE, D. M., KENNER, G. S. & GOULD, T. R. L.
(1983). Grooved titanium surfaces orient growth and migration of cells from human gingival explants. jf. dent. Res. 62, 1045-1048. BRUNETTE, D. M., MELCHER, A. H. & MOE, H. K.
(1976). Culture and origin of epithelial-like and fibroblast-like cells from porcine periodontal ligament explants and cell suspensions. Archs oral Biol. 21, 393-400. BUTLER, W. B., KIRKLAND, W. L. & JORGENSEN, T. L.
(1979). Induction of plasminogen activator by estrogen in a human breast cancer cell line (MCF-7). Biochem. biophys. Res. Commun. 90, 1328-1334. FoLKMAN, J. & MOSCONA, A. (1978). Role of cell shape in growth control. Nature, Land. 273, 345-349. HASEGAWA, S., SATO, S., SAITO, S., SUZUKI, Y. &
BRUNETTE, D. M. (1985). Mechanical stretching increases the number of cultured bone cells synthesizing DNA and alters their pattern of protein synthesis. Calcif. Tiss. Int. 37, 431-437. JACKSON, K. W., ESMON, N. & TANG, J. (1981).
Streptokinase and staphylokinase. In Methods in Enzymology, vol. 80 (ed. L. Lorand), pp. 387-388. New York: Academic Press. JOHNSON, G. L. (1982). Cholera toxin action and the regulation of hormone-sensitive adenylate cyclase. In
266
H. L. Hong and D. M. Brunette
Cyclic nucleotides. In Growth, Nutrition and Metabolism of Cells in Culture (ed. G. H. Rothblat & V. J. Cristofalo), pp. 295-336. New York: Academic Press. Modulation of plasominogen activator in rodent mammary tumors by hormones and other effectors. Cancer 43, 5467-5477. NAKAMURA, T. (1983). Density-dependent controls of growth and various cellular activities of mature rat hepatocytes in primary culture. In Isolation, Characterization and Use of Hepatocytes (ed. R. A. Harris & N. W. Cornell), pp. 193-198. New York: Elsevier Biomedical. OEY, J., VOGEL, A. & POLLACK, R. (1974). Intracellular
cyclic AMP concentration responds specifically to growth regulation by serum. Proc. natn. Acad. Sci. U.SA. 71, 694-698. OWENS, R. B. (1974). Glandular epithelial cells from mice: A method for selective cultivation. J. natn. Cancer Inst. 52, 1375-1378. PETTIGREW, D. W., H O , G. H., SODEK, J., BRUNETTE, D.
M. & WANG, H. M. (1978). Effect of oxygen tension and indomethacin on production of collagenase and neutral proteinase enzymes and their latent forms by porcine gingival explants in culture. Archs oral Biol. 23, 767-777. POLLACK, R., RISSER, R., CONLON, S., FREEDMAN, V.,
SHIN, S. & RIFKIN, D. B. (1975). Production of
plasminogen activator and colonial growth in semisolid medium are in vitro correlates of tumorigenicity in the immune-deficient nude mouse. In Proteases and Biological Control, vol. 2 (ed. E. Reich, D . B. Rifkin & E. Shaw), pp. 885-899. New York: Cold Spring Harbor Laboratory Press. QUIGLEY, J. P. (1979). Proteolytic enzymes of normal and malignant cells. In Surfaces of Normal and Malignant Cells (ed. R. O. Hynes), pp. 247-285. New York: Wiley. RICH, A. & HARRIS, A. K. (1981). Anomolous preferences of cultured macrophages for hydrophobic and roughened substrata. J. Cell Sci. 50, 1-7. ROVENSKY, Y. A., SLAVNAJA, I. L. & VASLIER, J. M.
(1971). Behaviour of fibroblast-like cells on grooved surfaces. Expl Cell Res. 65, 193-201. WEINSTEIN, I. B., WlGLER, M. & PlETROPAOLO, C. (1977). The action of tumor-promoting agents in cell culture. In
Origins of Human Cancer (ed. H. H. Hiatt, J. D. Watson & J. A. Winsten), pp. 751-772. New York: Cold Spring Harbor Laboratory Press. WEINSTEIN, I. B., WIGLER, M., YAMASAKI, H., L E E , L.,
FISHER, P. B. & MUFSON, A. (1978). Regulation of the
expression of certain biologic markers of neoplasia. In Biological Markers of Neoplasia: Basic and Applied Aspects (ed. R. W. Ruddon), pp. 451-471. North Holland, New York: Elsevier. WERB, Z., HEMBRY, R. M., MURPHY, G. & AGGELER, J.
(1986). Commitment to expression of the metalloendopeptidases, collagenase and stromelysin: relationship of inducing events to changes in cytoskeletal architecture. J. Cell Biol. 102, 697-702. WERB, Z. & REYNOLDS, J. J. (1974). Stimulation by
endocytosis of the secretion of collagenase and neutral proteinase from rabbit synovial fibroblasts. J. exp. Med. 140, 1482-1497.
WIGLER, M. & WEINSTEIN, I. B. (1976). Tumor promoter induces plasminogen activator. Nature, Land. 259, 232-233. WOOLLEY, D. E. (1984). Collagenolytic mechanisms in tumour cell invasion. In Cancer Metastasis Reviews, vol. 3, pp. 361-372. Boston, MA: Matinus Nijhoff. YOUNG, L. J. T., CARDIFF, R. D. & SEELEY, T . (1978).
Quantitative characterization of domes in primary mouse mammary epithelial tumor cell cultures. In Vitro 14, 895-902. ZAR, J. H. (1974). The binomial distribution. In Biostatistical Analysis, pp. 290-291. New Jersey: Prentice-Hall, Inc. ZETTERBERG, A. & AUER, G. (1970). Proliferative activity and cytochemical properties of nuclear chromatin related to local cell density of epithelial cells. Expl Cell Res. 62, 262-270. {Received 31 July 1986 -Accepted
E-cell shape and proteinase
14 October 1986)
secretion
267
