Cell, Vol. 7, 407-412.
March
1976,
Copyrightc
1976
by MIT
Effects of Protease Treatment on Growth, Morphology, Adhesion, and Cell Surface Proteins of Secondary Chick Embryo Fibroblasts Bruce Ft. Zetter, Lan Bo Chen, and John M. Buchanan Department of Biology Massachusetts Institute of Technology Cambridge, Massachusetts 02139
Summary Several proteolytic enzymes have been studied with regard to their ability to induce DNA synthesis and cell proliferation in resting chick embryo fibroblasts. Of the enzymes examined, thrombin, bromelin, and trypsln exhibit potent mitogenic activity, elastase has significant but less marked activity, whereas thermolysin, papain, and a-protease are inactive. The enzymes were also tested for their ability to induce morphological change or to remove two iodinatable proteins of 250,000 and 205,000 daltons. Although the larger proteln is removed by some but not all of the proteases examined, every protease tested removed the smaller cell surface protein. The ability of proteases to stimulate cell growth could not be correlated directly with removal of either of these cell surface proteins; however, loss of the smaller protein does correlate with the reduction of both cytoplasmic spreading and cell-cell interactions observed after protease treatment. A secondary, later event of migration of cells into clumps is observed in those instances when protease treatment did not result in a loss of the 250K protein. A role for each of these proteins in the processes of cellular adhesion is discussed. Introduction It is now well established that the densitydependent inhibition of growth observed in chick embryo fibroblast cultures can be released by mild treatment with certain proteolytic enzymes (Sefton and Rubin, 1970; Vaheri, Ruoslahti, and Hovi, 1974; Chen and Buchanan, 1975; Hale and Weber, 1975). Protease treatment induces not only DNA synthesis and cell proliferation in chick fibroblasts, but also can lead to changes in cellular morphology (Martin, 1970), stimulation of glucose uptake (Sefton and Rubin, 1970; Hale and Weber, 1975; Blumberg and Robbins, 1975a), and removal of proteins from cell surfaces (Hynes, 1974; Blumberg and Robbins, 1975a; Teng and Chen, 1975). It is tempting to view this complex response to proteolysis as a sequence or pathway of events leading to cell division. In this regard, certain specific responses to protease treatment, such as release of cell surface proteins and changes in cell morphology, have been suggested to bear a causative rela-
tionship to the induction of cell growth (Blumberg and Robbins, 1975a; Henneberry, Fishman, and Freese, 1975) as well as to transformation (Wickus, Branton, and Robbins, 1974; Hynes, 1974). It is of interest, however, that not every protease is capable of eliciting the same array of cellular responses (Sefton, 1972; Teng and Chen, 1975). The specificity of each protease for substrates at the cell surface probably has a role in determining whether a given enzyme can induce stimulation of nutrient transport, mitogenesis, or morphological changes. Of the several cell surface proteins sensitive to the action of proteolytic enzymes, none has yet been proved to be responsible for the induction of a specific cellular response. To study these phenomena, we have examined several proteases for their abilities to induce changes in the morphology, growth state, or cell surface composition of resting chick embryo fibroblasts. We find that whereas protease-induced cell growth cannot be correlated with the removal of any specific cell surface protein, a relationship may exist between the removal of a recently described 205,000 dalton protein (Teng and Chen, 1976) and the cellular retraction and elongation, which regularly follow limited protease treatment. In addition, we have provided evidence for the role of the protein of 250,000 daltons in cell adhesion of chick embryo fibroblasts, as has been postulated by Yamada, Yamada, and Pastan (1975). Results Protease-Induced Morphological Changes After 4 days in a medium containing 0.5% calf serum, quiescent secondary chick embryo fibroblasts appear as shown in Figure 1A. The cells are flat, firmly attached to the plate, and form a confluent monolayer in which the cells appear to have grown together in a random fashion. Figure 1B shows the same plate after a 12 hr incubation in fresh medium containing 2 pg/ml of purified thrombin and no serum. The cells have become more elongated and have undergone a change in their pattern of cell-cell interactions. Since the number of cells attached to the plate has not decreased during this 12 hr period, the appearance of bare areas on the plate is clearly a result of changes in cell morphology, adhesion, and migration, rather than loss of attached fibroblasts by proteolysis. This morphology is in marked contrast to the striking parallel arrays of cells observed after 12 hr incubation in a medium containing 5% (v/v) calf serum (Figure 1C). As in the resting state, there are no spaces between cells, but here the individual cells are more spindle-shaped and the intercellular contacts highly organized.
Cell 408
Protease 409
Effects
on Chick
Embryo
Fibroblasts
Cultures treated for 12 hr with bromelin (Figure 1 D) and oc-protease (1E) form aggregates similar to those observed following thrombin treatment. Four other proteases-trypsin (1 F), elastase (1 G), papain (1 H) and thermolysin (1 I)-cause the cells to retract and reduce cytoplasmic spreading without inducing the formation of large aggregates of cells. Although it is clear that there are two patterns of distribution of protease-treated cells shown in Figure 1, the shapes of the individual cells are basically the same. The initial morphological response to all proteases is a retraction of cells with a decrease in cytoplasmic spreading and cell-cell contact. This process gives rise to a pattern similar to that shown in Figure 1 F and correlates with the loss of a particular cell surface protein of 205,000 daltons. In the case of those protease-treated cultures in which aggregates are observed, we have shown by the aid of cinemicrophotography (unpublished results) and the photographs of cells (Figures lB,
Table
1. Protease
Stimulation
of DNA
Synthesis
(fwml)
None Serum Thrombin
DNA Synthesis and Cell Proliferation As shown in Table 1, several of the enzymes tested are capable of inducing increased DNA synthesis and cell division. When added to cells in a serumfree medium and incubated for 12 hr, thrombin, elastase, bromelin, and trypsin all cause significant incorporation of thymidine during a 1 hr pulse. Thermolysin, a-protease, and papain, on the other hand, do not significantly increase DNA synthesis. That this is not a result of toxicity of the proteases can be demonstrated by the ability of serum or thrombin to induce DNA synthesis when added to cells 5 hr after the addition of thermolysin, a-protease, or papain, and incubated an additional 12 hr (data not shown).
and Cell Proliferation
Amount of 13rlLabeled Protein (% of Control)
Concentration Treatment
lD, and 1 E) that a separate later event occurs which causes cells to migrate into clumps, leaving large open spaces in between. We show that this aggregation occurs only when a protease fails to remove a cell surface protein of 250,000 daltons.
250K
205K 100
1,700
100
100
0
33,000 32,000 40,000
19.4 18.8 23.5
10.6 9.2 10.1
5 IO 15
100
0
2,000 2,200 2,800 4,200
1.2 1.3 1.6 2.5
4.4 4.5 4.8 5.1
0 0
0 0
1,800 3,800 5,200
1.1 2.2 3.0
4.8 4.9 5.1
9,900 17,900 24,000
5.8 10.5 14.1
5.2 9.0 9.5
28,000 41,000
16.4 24.1
9.6 10.3
1,700 2,700 3,200
1 .o 1.8 1.9
4.5 4.7 4.7
14,000 20,000 20,000
8.2 11.8 11.8
5.8 7.5 7.6
Trypsin
0.1 0.2
Papain
0.5 1 5
100
0
0
0
0
0
0 0
0 0
Elastase 5 10
Figure
1, Morphology
1 .o
Cell Number0 at 36 Hr
100
1 2.5 5.0
incubated
Radioactivity: Experimental/ Control
100
5 10
were
Fibroblasts
(5%) 1 2.5
Thermolysin
aCells
Embryo
Incorporation ofa XH-Thymidine into DNA at 12 hr (cpm per Plate)
n-Protease
Bromelin
in Chick
continuously
of Chick
Cells
with
serum
Resting
or protease
and after
for the duration
Treatment
with
Serum
x IO-5
4.5
of the experiment.
or Proteases
(A) Resting chick embryo cells prepared according to Chen and Buchanan (1975). Resting cells treated for 12 hr with (B) thrombin (2.5 pg/ml); (C) calf serum (5%); (D) bromelin (1 pglml); (E) a-protease (5 pg/ml); (F) trypsin (0.1 pg/ml); (G) elastase (5 pg/ml); (H) papain (1 @g/ml); and (I) thermolysin (1 pg/ml). Photographs were taken using a Bausch-Lomb Dynazoom phase-contrast microscope. The varied cell density seen in photographs resulted from the use of different batches of resting cells for the individual experiments.
Cell 410
To determine whether the treated cells could traverse the entire cell cycle and divide, cell numbers were determined before and 36 hr after protease treatment. Thrombin, bromelin, and trypsin-treated cultures doubled in cell number during this time. Elastase-treated cultures increased in number but to a lesser degree, whereas thermolysin, a-protease, and papain were incapable of promoting cell growth. Removal of Cell Surface Proteins Figure 2 shows the gel migration patterns of cell surface proteins labeled by lactoperoxidasecatalyzed iodination with lx’-iodine. All iodinations were performed after 3 hr of incubation with the proteases. The heavily labeled protein at the top of the gels is the 250K or LETS protein. As previously reported by Teng and Chen (1975), the presence or absence of this protein on the cell surface cannot be correlated with the growth state of proteasetreated cells. Although this protein is removed by mitogens such as trypsin and elastase, it is also removed by the nonmitogenic proteases, thermolysin and papain (Figure 2). The high dose of bromelin necessary to remove this protein (Teng and Chen, 1975) causes the removal of large numbers of cells
AECDE
from the plate if bromelin is kept in the culture for 36 hr, whereas a lower dose (5 pg/ml) causes mitogenie stimulation without removal of this protein. Thrombin, as noted previously (Teng and Chen, 1975), causes no removal of the 250K protein at doses that are highly mitogenic. The removal by thrombin of a cell surface protein, in this case of 205,000 daltons, has recently been demonstrated (Teng and Chen, 1976). A protein of this size is removed by every protease tested in our experiments. Since this protein is sensitive to both mitogenic and nonmitogenic enzymes, its removal is not considered to be sufficient to bring about cell division. It may, however, be involved in the initial morphological alterations brought about by treatment with any of these enzymes. Incubation of cells with 5% (v/v) calf serum brings about no alteration in the observed pattern of iodinatable cell surface proteins, even though serum at this concentration is highly mitogenic. Discussion Protease treatment of resting blasts can induce a pleiotropic
chick embryo fibroresponse consisting
FGH
250 205
Figure
2. Effect
of Protease
Treatment
on Cell Surface
Proteins
Chick fibroblast cultures were labeled by lactoperoxidase catalyzed iodination with 1311after 3 hr incubation with the following enzymes: (A) untreated control; (B) thermolysin (1 cg/ml); (C) thermolysin (5 pg/ml); (D) elastase (5 pg/ml); (E) papain, (5 ag/ml); (F) a-protease (5 pg/ml); (G) bromelin (5 pg/ml); (H) trypsin (0.5 pg/ml); (I) thrombin (1 ag/ml). Vertical lines separate gels run on different occasions.
Protease 411
Effects
on Chick
Embryo
Fibroblasts
of morphological, physiological, and metabolic changes. The experiments described here were designed to determine whether these different responses are independent and dissociable, or whether they are part of a complex universal response to mitogenic protease treatment. It appears from our data that whereas no correlation can be made between the removal of cell surface proteins and the growth state of fibroblasts, a relationship may exist between these proteins and protease-induced morphological change. Proteases and Mitogenicity The mechanisms by which proteolytic enzymes stimulate cell growth are still undefined. It has been previously shown that brief treatments of resting chick cells with proteases (Sefton and Rubin, 1970; Teng and Chen, 1975) can induce only a low level of cell growth. Apparently, the longer the presence of proteases in the culture, the greater the growth stimulation that can be elicited (Sefton and Rubin, 1970). In our investigations, we have allowed the proteases to be incubated with the culture constantly until measurement of the growth parameter to be examined, that is, 12 hr after the addition of proteases for the determination of DNA synthesis, and 36 hr for determination of cell number. Of the enzymes tested, the most potent mitogens are thrombin, bromelin, and trypsin, each of which is capable of causing a doubling in cell number within the first 36 hr of protease treatment. Significant but less marked stimulation of growth is caused by treatment with elastase. All of these enzymes give rise to at least a 10 fold increase in the rate of DNA synthesis after 12 hr of continuous incubation. It is important to note that all enzyme treatments are performed in serum-free media, and that the observed increases in DNA synthesis and cell number take place under these conditions. It is clear, however, that not every type of proteolysis is capable of inducing cell growth, even when the protease is constantly present in culture. In our hands, papain, n-protease, and thermolysin alone cause no significant increase in DNA synthesis in resting cells and are unable to bring about a significant increase in cell number after a 36 hr incubation. An increasing amount of attention is currently given to labile proteins on cell surfaces, especially those which can be iodinated in the presence of lactoperoxidase. Viral transformation and protease treatment have both been shown to result in changes in the expression of certain cell surface proteins (Ruoslahti et al., 1973; Hynes, 1974; Robbins et al., 1974; Hogg, 1974; Stone, Smith, and Joklik, 1974; Gahmberg and Hakomori, 1974), but to date there has been no convincing evidence that
enzymatic removal of any specific protein can be directly correlated with the induction of DNA synthesis or cell division. Our data clearly demonstrate that removal of either of the two major high molecular weight protease-sensitive proteins (Hynes, 1974; Robbins et al., 1974; Teng and Chen, 1976) is not responsible for the mitogenic stimulation of resting chick embryo fibroblasts. The largest of these, the 250K or LETS protein, has previously been shown to be sensitive to trypsin, subtilisin, chymotrypsin, pronase, and high doses of bromelin, but not to thrombin. We have extended these findings to show removal of this protein by one more mitogen, elastase, and two nonmitogenic proteases, thermolysin and papain. Lower but still mitogenic doses of bromelin have been found not to remove the protein. In these experiments, all iodinations were performed after 3 hr of continuous protease treatment. Proteins removed from the surface by this time are still absent at 12 hr (during the peak of the S phase). In a recent paper, Blumberg and Robbins (1975b) reported that among the proteases that remove LETS protein, there is still a correlation between the amount of LETS removed and cell growth, even though thrombin does not remove LETS. They concluded that although the removal of LETS is not a necessary condition for growth, it may be a sufficient one. In this regard, this paper presents a new finding-that two proteases, papain and thermolysin, remove LETS protein, yet induce neither DNA synthesis nor cell growth. In a previous report, Teng and Chen (1975) showed that the removal of LETS is not a necessary condition; the new information contained here shows that the removal of LETS is not a sufficient condition for cell growth either. Proteases and Cell-Cell Interactions A second protease-sensitive cell surface protein has recently been reported by Teng and Chen (1976). This protein of 205,000 daltons does not appear to correlate with cell growth, since it is removed by growth-promoting enzymes such as thrombin, elastase, and trypsin, as well as bythermolysin and papain, which are unable tostimulate cell division. It is evident from our results that protease treatments of chick embryo fibroblasts consistently bring about two responses: removal of a 205,000 dalton protein from the cell surface, and an initial morphological change characterized by reduction of bothcytoplasmicspreadingandcell-cellinteractions. Since the protease-treated cells initially reduce contact with neighboring cells as well as with the plate surface, it is probable that some of the cell surface proteins affected by proteolysis are those involved in cell adhesion.
Cell 412
We propose a role of the 205K protein in this connection. Loss of this cell surface protein might result in the change of the cells from the morphology seen in Figure 1A to that observed in Figure 1F. These cells have reduced intercellular contact as well as a reduction in the amount of cytoplasm that is in contact with the substratum. This morphology is the one seen initially after treatment with every protease we have tested. In some cases, this is followed by a later independent event involving increased migration and aggregation, which we propose results from the retention and unique exposure of the 250K protein on the cell surface after the removal of the 205K protein. That the 250K protein may also have a role in cellular adhesion has been suggested by Yamada et al. (1975), who showed that 250K protein purified from chick embryo fibroblasts agglutinates sheep erythrocytes. Our data provide new evidence that this protein may be involved in cell-cell adhesion in the chick cells themselves. Of all the proteases tested, only thrombin, bromelin, and a-protease (Figures lB, ID, and 1E) induce the formation of large clumps of cells. This result is striking, because at the concentration used, only these three enzymes fail to remove the 250K protein. It is possible that removal of 205K protein increases the exposure of the 250K protein on the cell surface, thus enhancing intercellular adhesion in these three cases. Further work is in progress to elucidate the role of both these proteins in the reported alterations of cell-cell adhesion of tumor cells in vivo (McCutcheon, Coman, and Moore, 1948).
Highly purified thrombin, was provided by Dr. David of Technology.
Experimental
Robbins, P. W., Wickus, G. G.. Branton, P. E., Gaffney, B. J.. Hirschberg. C. B., Fuchs, P., and Blumberg. P. M. (1974). Cold Spring Harbor Symp. Quant. Biol. 39, 1173.
Procedures
Secondary cultures of chick embryo fibroblasts, prepared by the method of Rein and Rubin (1968) and provided by Dr. P. W. Robbins, were assayed for DNA synthesis and cell division by previously described methods (Chen and Buchanan, 1975). In all cases, the cultures were treated continuously with the proteases for 12 hr prior to labeling with ‘H-thymidine or for 36 hr prior to counting cell number. All protease treatments were carried out in serum-free medium at 37°C. Lactoperoxidase-catalyzed iodination of cell surface proteins with rsr-iodine was performed according to the method described by Teng and Chen (1976). The samples were heated to 100°C for 10 min in buffer containing 3% sodium-dodecylsulphate, 1% p-mercaptoethanol. and 2 mM phenylmethanesulfonylfluoride, and then subjected to electrophoresis. The polyacrylamide gel system was similar to that described by Huang and Lehman (1972) except that the running gel consisted of a gradient of 5-15% acrylamide. Autoradiograms of the dried slab gels were made with Kodak NSZT X-ray film. from New High concentration carrier-free Na 1311 was obtained England Nuclear. Enzyme preparations were from the following sources: bromelin (1.2 U/mg), Schwartz-Mann: elastase (500 U/mg), Sigma Chemical Co; thermolysin (93,904 U/mg), Calbiothem; a-protease, Gallard Schlesinger; TPCK-treated trypsin (250 U/mg), Worthington Chemical Co; twice-crystallized papain (12.5 U/mg), Sigma Chemical Co.
approximately 2000 NIH units F. Waugh of the Massachusetts
per mg, Institute
Acknowledgments This work was supported by grants-in-aid from the American Cancer Society and the National Cancer Institute. B.R.Z. was a postdoctoral fellow of the Damon Runyon-Walter Winchell Cancer Fund, and L.B.C. a predoctoral fellow of Johnson and Johnson Co. We wish to thank Mr. Norman Hochella for technical assistance. Received
August
28, 1975;
revised
November
11, 1975
References Blumberg, Biological p. 945.
P. M.. and Robbins, Control (New York:
Blumberg,
P. M., and Robbins,
Chen, L. B., and USA 72, 131. Gahmberg, Commun.
P. W. (1975b).
Buchanan,
J. M. (1975).
C. G.. and Hakomori, 59, 283.
Hale, A. H., and Weber. Henneberry, Hogg,
P. W. (1975a). In Proteases and Cold Spring Harbor Laboratory),
N. M. (1974).
Proc.
Huang,
W. M., and Lehman,
Hynes,
R. 0. (1974).
Martin, G. S. (1970). Lepetit Colloquium, land), p. 320. McCutcheon, 1, 460.
Nat.
Biochem.
Nat. Acad.
E. (1975).
Sci. USA
I. R. (1972).
Cell 5, 1,
77, 489.
J. Biol. Chem.
D. R., and Moore,
Rein,
H. (1988).
247, 7663.
P. P. (1974). Exp.
F. B. (1948).
Biochim.
P., and Linder,
Sefton, 8. M. (1972). ley, California.
University
Ph.D.
B. M., and Rubin,
Thesis, H. (1970).
R. E.. and
Teng, N. N. H., and Chen, 72, 413. N. N. H.. and Chen,
Cancer
Biophys.
Acta
Cell Res. 49, 666.
Ruoslahti, E.. Vaheri, A., Kuusela, chim. Biophys. Acta 322, 352.
Teng,
Res.
In The Biology of Oncogenic Viruses, Second L. G. Silveston, ed. (Amsterdam: North Hol-
Agin,
Stone, K. R.. Smith, 58. 86.
Biophys.
Sci.
Cell 5, 245.
P. H., and Freese,
Phillips, D. R., and 352, 218.
Sefton,
Acad.
Cell 1, 147.
M., Coman,
A., and Rubin,
S. (1974).
N. J. (1975).
R. C., Fishman.
Cell 6, 137.
Proc.
of California,
Nature Joklik.
L. B. (1975). L. B. (1976).
E. (1973).
BioBerke-
227, 843. W. K. (1974).
Proc.
Nat. Acad.
Nature,
Virology Sci. USA
in press.
Vaheri, A., Ruoslahti. E.. and Hovi, T. (1974). In Control of Proliferation in Animal Cells, B. Clarkson and R. Baserga, eds. (New York: Cold Spring Harbor Laboratory), p. 305. Wickus. G. G., Branton. P. E., and Robbins, of Proliferation in Animal Cells, B. Clarkson (New York: Cold Spring Harbor Laboratory), Yamada, K. M., Yamada, S. S., and Pastan, Acad. Sci. USA 72, 3158.
P. S. (1974). In Control and R. Baserga, eds. p. 541. I. (1975).
Proc.
Nat.
March
1976,
Copyrightc
1976
by MIT
Effects of Protease Treatment on Growth, Morphology, Adhesion, and Cell Surface Proteins of Secondary Chick Embryo Fibroblasts Bruce Ft. Zetter, Lan Bo Chen, and John M. Buchanan Department of Biology Massachusetts Institute of Technology Cambridge, Massachusetts 02139
Summary Several proteolytic enzymes have been studied with regard to their ability to induce DNA synthesis and cell proliferation in resting chick embryo fibroblasts. Of the enzymes examined, thrombin, bromelin, and trypsln exhibit potent mitogenic activity, elastase has significant but less marked activity, whereas thermolysin, papain, and a-protease are inactive. The enzymes were also tested for their ability to induce morphological change or to remove two iodinatable proteins of 250,000 and 205,000 daltons. Although the larger proteln is removed by some but not all of the proteases examined, every protease tested removed the smaller cell surface protein. The ability of proteases to stimulate cell growth could not be correlated directly with removal of either of these cell surface proteins; however, loss of the smaller protein does correlate with the reduction of both cytoplasmic spreading and cell-cell interactions observed after protease treatment. A secondary, later event of migration of cells into clumps is observed in those instances when protease treatment did not result in a loss of the 250K protein. A role for each of these proteins in the processes of cellular adhesion is discussed. Introduction It is now well established that the densitydependent inhibition of growth observed in chick embryo fibroblast cultures can be released by mild treatment with certain proteolytic enzymes (Sefton and Rubin, 1970; Vaheri, Ruoslahti, and Hovi, 1974; Chen and Buchanan, 1975; Hale and Weber, 1975). Protease treatment induces not only DNA synthesis and cell proliferation in chick fibroblasts, but also can lead to changes in cellular morphology (Martin, 1970), stimulation of glucose uptake (Sefton and Rubin, 1970; Hale and Weber, 1975; Blumberg and Robbins, 1975a), and removal of proteins from cell surfaces (Hynes, 1974; Blumberg and Robbins, 1975a; Teng and Chen, 1975). It is tempting to view this complex response to proteolysis as a sequence or pathway of events leading to cell division. In this regard, certain specific responses to protease treatment, such as release of cell surface proteins and changes in cell morphology, have been suggested to bear a causative rela-
tionship to the induction of cell growth (Blumberg and Robbins, 1975a; Henneberry, Fishman, and Freese, 1975) as well as to transformation (Wickus, Branton, and Robbins, 1974; Hynes, 1974). It is of interest, however, that not every protease is capable of eliciting the same array of cellular responses (Sefton, 1972; Teng and Chen, 1975). The specificity of each protease for substrates at the cell surface probably has a role in determining whether a given enzyme can induce stimulation of nutrient transport, mitogenesis, or morphological changes. Of the several cell surface proteins sensitive to the action of proteolytic enzymes, none has yet been proved to be responsible for the induction of a specific cellular response. To study these phenomena, we have examined several proteases for their abilities to induce changes in the morphology, growth state, or cell surface composition of resting chick embryo fibroblasts. We find that whereas protease-induced cell growth cannot be correlated with the removal of any specific cell surface protein, a relationship may exist between the removal of a recently described 205,000 dalton protein (Teng and Chen, 1976) and the cellular retraction and elongation, which regularly follow limited protease treatment. In addition, we have provided evidence for the role of the protein of 250,000 daltons in cell adhesion of chick embryo fibroblasts, as has been postulated by Yamada, Yamada, and Pastan (1975). Results Protease-Induced Morphological Changes After 4 days in a medium containing 0.5% calf serum, quiescent secondary chick embryo fibroblasts appear as shown in Figure 1A. The cells are flat, firmly attached to the plate, and form a confluent monolayer in which the cells appear to have grown together in a random fashion. Figure 1B shows the same plate after a 12 hr incubation in fresh medium containing 2 pg/ml of purified thrombin and no serum. The cells have become more elongated and have undergone a change in their pattern of cell-cell interactions. Since the number of cells attached to the plate has not decreased during this 12 hr period, the appearance of bare areas on the plate is clearly a result of changes in cell morphology, adhesion, and migration, rather than loss of attached fibroblasts by proteolysis. This morphology is in marked contrast to the striking parallel arrays of cells observed after 12 hr incubation in a medium containing 5% (v/v) calf serum (Figure 1C). As in the resting state, there are no spaces between cells, but here the individual cells are more spindle-shaped and the intercellular contacts highly organized.
Cell 408
Protease 409
Effects
on Chick
Embryo
Fibroblasts
Cultures treated for 12 hr with bromelin (Figure 1 D) and oc-protease (1E) form aggregates similar to those observed following thrombin treatment. Four other proteases-trypsin (1 F), elastase (1 G), papain (1 H) and thermolysin (1 I)-cause the cells to retract and reduce cytoplasmic spreading without inducing the formation of large aggregates of cells. Although it is clear that there are two patterns of distribution of protease-treated cells shown in Figure 1, the shapes of the individual cells are basically the same. The initial morphological response to all proteases is a retraction of cells with a decrease in cytoplasmic spreading and cell-cell contact. This process gives rise to a pattern similar to that shown in Figure 1 F and correlates with the loss of a particular cell surface protein of 205,000 daltons. In the case of those protease-treated cultures in which aggregates are observed, we have shown by the aid of cinemicrophotography (unpublished results) and the photographs of cells (Figures lB,
Table
1. Protease
Stimulation
of DNA
Synthesis
(fwml)
None Serum Thrombin
DNA Synthesis and Cell Proliferation As shown in Table 1, several of the enzymes tested are capable of inducing increased DNA synthesis and cell division. When added to cells in a serumfree medium and incubated for 12 hr, thrombin, elastase, bromelin, and trypsin all cause significant incorporation of thymidine during a 1 hr pulse. Thermolysin, a-protease, and papain, on the other hand, do not significantly increase DNA synthesis. That this is not a result of toxicity of the proteases can be demonstrated by the ability of serum or thrombin to induce DNA synthesis when added to cells 5 hr after the addition of thermolysin, a-protease, or papain, and incubated an additional 12 hr (data not shown).
and Cell Proliferation
Amount of 13rlLabeled Protein (% of Control)
Concentration Treatment
lD, and 1 E) that a separate later event occurs which causes cells to migrate into clumps, leaving large open spaces in between. We show that this aggregation occurs only when a protease fails to remove a cell surface protein of 250,000 daltons.
250K
205K 100
1,700
100
100
0
33,000 32,000 40,000
19.4 18.8 23.5
10.6 9.2 10.1
5 IO 15
100
0
2,000 2,200 2,800 4,200
1.2 1.3 1.6 2.5
4.4 4.5 4.8 5.1
0 0
0 0
1,800 3,800 5,200
1.1 2.2 3.0
4.8 4.9 5.1
9,900 17,900 24,000
5.8 10.5 14.1
5.2 9.0 9.5
28,000 41,000
16.4 24.1
9.6 10.3
1,700 2,700 3,200
1 .o 1.8 1.9
4.5 4.7 4.7
14,000 20,000 20,000
8.2 11.8 11.8
5.8 7.5 7.6
Trypsin
0.1 0.2
Papain
0.5 1 5
100
0
0
0
0
0
0 0
0 0
Elastase 5 10
Figure
1, Morphology
1 .o
Cell Number0 at 36 Hr
100
1 2.5 5.0
incubated
Radioactivity: Experimental/ Control
100
5 10
were
Fibroblasts
(5%) 1 2.5
Thermolysin
aCells
Embryo
Incorporation ofa XH-Thymidine into DNA at 12 hr (cpm per Plate)
n-Protease
Bromelin
in Chick
continuously
of Chick
Cells
with
serum
Resting
or protease
and after
for the duration
Treatment
with
Serum
x IO-5
4.5
of the experiment.
or Proteases
(A) Resting chick embryo cells prepared according to Chen and Buchanan (1975). Resting cells treated for 12 hr with (B) thrombin (2.5 pg/ml); (C) calf serum (5%); (D) bromelin (1 pglml); (E) a-protease (5 pg/ml); (F) trypsin (0.1 pg/ml); (G) elastase (5 pg/ml); (H) papain (1 @g/ml); and (I) thermolysin (1 pg/ml). Photographs were taken using a Bausch-Lomb Dynazoom phase-contrast microscope. The varied cell density seen in photographs resulted from the use of different batches of resting cells for the individual experiments.
Cell 410
To determine whether the treated cells could traverse the entire cell cycle and divide, cell numbers were determined before and 36 hr after protease treatment. Thrombin, bromelin, and trypsin-treated cultures doubled in cell number during this time. Elastase-treated cultures increased in number but to a lesser degree, whereas thermolysin, a-protease, and papain were incapable of promoting cell growth. Removal of Cell Surface Proteins Figure 2 shows the gel migration patterns of cell surface proteins labeled by lactoperoxidasecatalyzed iodination with lx’-iodine. All iodinations were performed after 3 hr of incubation with the proteases. The heavily labeled protein at the top of the gels is the 250K or LETS protein. As previously reported by Teng and Chen (1975), the presence or absence of this protein on the cell surface cannot be correlated with the growth state of proteasetreated cells. Although this protein is removed by mitogens such as trypsin and elastase, it is also removed by the nonmitogenic proteases, thermolysin and papain (Figure 2). The high dose of bromelin necessary to remove this protein (Teng and Chen, 1975) causes the removal of large numbers of cells
AECDE
from the plate if bromelin is kept in the culture for 36 hr, whereas a lower dose (5 pg/ml) causes mitogenie stimulation without removal of this protein. Thrombin, as noted previously (Teng and Chen, 1975), causes no removal of the 250K protein at doses that are highly mitogenic. The removal by thrombin of a cell surface protein, in this case of 205,000 daltons, has recently been demonstrated (Teng and Chen, 1976). A protein of this size is removed by every protease tested in our experiments. Since this protein is sensitive to both mitogenic and nonmitogenic enzymes, its removal is not considered to be sufficient to bring about cell division. It may, however, be involved in the initial morphological alterations brought about by treatment with any of these enzymes. Incubation of cells with 5% (v/v) calf serum brings about no alteration in the observed pattern of iodinatable cell surface proteins, even though serum at this concentration is highly mitogenic. Discussion Protease treatment of resting blasts can induce a pleiotropic
chick embryo fibroresponse consisting
FGH
250 205
Figure
2. Effect
of Protease
Treatment
on Cell Surface
Proteins
Chick fibroblast cultures were labeled by lactoperoxidase catalyzed iodination with 1311after 3 hr incubation with the following enzymes: (A) untreated control; (B) thermolysin (1 cg/ml); (C) thermolysin (5 pg/ml); (D) elastase (5 pg/ml); (E) papain, (5 ag/ml); (F) a-protease (5 pg/ml); (G) bromelin (5 pg/ml); (H) trypsin (0.5 pg/ml); (I) thrombin (1 ag/ml). Vertical lines separate gels run on different occasions.
Protease 411
Effects
on Chick
Embryo
Fibroblasts
of morphological, physiological, and metabolic changes. The experiments described here were designed to determine whether these different responses are independent and dissociable, or whether they are part of a complex universal response to mitogenic protease treatment. It appears from our data that whereas no correlation can be made between the removal of cell surface proteins and the growth state of fibroblasts, a relationship may exist between these proteins and protease-induced morphological change. Proteases and Mitogenicity The mechanisms by which proteolytic enzymes stimulate cell growth are still undefined. It has been previously shown that brief treatments of resting chick cells with proteases (Sefton and Rubin, 1970; Teng and Chen, 1975) can induce only a low level of cell growth. Apparently, the longer the presence of proteases in the culture, the greater the growth stimulation that can be elicited (Sefton and Rubin, 1970). In our investigations, we have allowed the proteases to be incubated with the culture constantly until measurement of the growth parameter to be examined, that is, 12 hr after the addition of proteases for the determination of DNA synthesis, and 36 hr for determination of cell number. Of the enzymes tested, the most potent mitogens are thrombin, bromelin, and trypsin, each of which is capable of causing a doubling in cell number within the first 36 hr of protease treatment. Significant but less marked stimulation of growth is caused by treatment with elastase. All of these enzymes give rise to at least a 10 fold increase in the rate of DNA synthesis after 12 hr of continuous incubation. It is important to note that all enzyme treatments are performed in serum-free media, and that the observed increases in DNA synthesis and cell number take place under these conditions. It is clear, however, that not every type of proteolysis is capable of inducing cell growth, even when the protease is constantly present in culture. In our hands, papain, n-protease, and thermolysin alone cause no significant increase in DNA synthesis in resting cells and are unable to bring about a significant increase in cell number after a 36 hr incubation. An increasing amount of attention is currently given to labile proteins on cell surfaces, especially those which can be iodinated in the presence of lactoperoxidase. Viral transformation and protease treatment have both been shown to result in changes in the expression of certain cell surface proteins (Ruoslahti et al., 1973; Hynes, 1974; Robbins et al., 1974; Hogg, 1974; Stone, Smith, and Joklik, 1974; Gahmberg and Hakomori, 1974), but to date there has been no convincing evidence that
enzymatic removal of any specific protein can be directly correlated with the induction of DNA synthesis or cell division. Our data clearly demonstrate that removal of either of the two major high molecular weight protease-sensitive proteins (Hynes, 1974; Robbins et al., 1974; Teng and Chen, 1976) is not responsible for the mitogenic stimulation of resting chick embryo fibroblasts. The largest of these, the 250K or LETS protein, has previously been shown to be sensitive to trypsin, subtilisin, chymotrypsin, pronase, and high doses of bromelin, but not to thrombin. We have extended these findings to show removal of this protein by one more mitogen, elastase, and two nonmitogenic proteases, thermolysin and papain. Lower but still mitogenic doses of bromelin have been found not to remove the protein. In these experiments, all iodinations were performed after 3 hr of continuous protease treatment. Proteins removed from the surface by this time are still absent at 12 hr (during the peak of the S phase). In a recent paper, Blumberg and Robbins (1975b) reported that among the proteases that remove LETS protein, there is still a correlation between the amount of LETS removed and cell growth, even though thrombin does not remove LETS. They concluded that although the removal of LETS is not a necessary condition for growth, it may be a sufficient one. In this regard, this paper presents a new finding-that two proteases, papain and thermolysin, remove LETS protein, yet induce neither DNA synthesis nor cell growth. In a previous report, Teng and Chen (1975) showed that the removal of LETS is not a necessary condition; the new information contained here shows that the removal of LETS is not a sufficient condition for cell growth either. Proteases and Cell-Cell Interactions A second protease-sensitive cell surface protein has recently been reported by Teng and Chen (1976). This protein of 205,000 daltons does not appear to correlate with cell growth, since it is removed by growth-promoting enzymes such as thrombin, elastase, and trypsin, as well as bythermolysin and papain, which are unable tostimulate cell division. It is evident from our results that protease treatments of chick embryo fibroblasts consistently bring about two responses: removal of a 205,000 dalton protein from the cell surface, and an initial morphological change characterized by reduction of bothcytoplasmicspreadingandcell-cellinteractions. Since the protease-treated cells initially reduce contact with neighboring cells as well as with the plate surface, it is probable that some of the cell surface proteins affected by proteolysis are those involved in cell adhesion.
Cell 412
We propose a role of the 205K protein in this connection. Loss of this cell surface protein might result in the change of the cells from the morphology seen in Figure 1A to that observed in Figure 1F. These cells have reduced intercellular contact as well as a reduction in the amount of cytoplasm that is in contact with the substratum. This morphology is the one seen initially after treatment with every protease we have tested. In some cases, this is followed by a later independent event involving increased migration and aggregation, which we propose results from the retention and unique exposure of the 250K protein on the cell surface after the removal of the 205K protein. That the 250K protein may also have a role in cellular adhesion has been suggested by Yamada et al. (1975), who showed that 250K protein purified from chick embryo fibroblasts agglutinates sheep erythrocytes. Our data provide new evidence that this protein may be involved in cell-cell adhesion in the chick cells themselves. Of all the proteases tested, only thrombin, bromelin, and a-protease (Figures lB, ID, and 1E) induce the formation of large clumps of cells. This result is striking, because at the concentration used, only these three enzymes fail to remove the 250K protein. It is possible that removal of 205K protein increases the exposure of the 250K protein on the cell surface, thus enhancing intercellular adhesion in these three cases. Further work is in progress to elucidate the role of both these proteins in the reported alterations of cell-cell adhesion of tumor cells in vivo (McCutcheon, Coman, and Moore, 1948).
Highly purified thrombin, was provided by Dr. David of Technology.
Experimental
Robbins, P. W., Wickus, G. G.. Branton, P. E., Gaffney, B. J.. Hirschberg. C. B., Fuchs, P., and Blumberg. P. M. (1974). Cold Spring Harbor Symp. Quant. Biol. 39, 1173.
Procedures
Secondary cultures of chick embryo fibroblasts, prepared by the method of Rein and Rubin (1968) and provided by Dr. P. W. Robbins, were assayed for DNA synthesis and cell division by previously described methods (Chen and Buchanan, 1975). In all cases, the cultures were treated continuously with the proteases for 12 hr prior to labeling with ‘H-thymidine or for 36 hr prior to counting cell number. All protease treatments were carried out in serum-free medium at 37°C. Lactoperoxidase-catalyzed iodination of cell surface proteins with rsr-iodine was performed according to the method described by Teng and Chen (1976). The samples were heated to 100°C for 10 min in buffer containing 3% sodium-dodecylsulphate, 1% p-mercaptoethanol. and 2 mM phenylmethanesulfonylfluoride, and then subjected to electrophoresis. The polyacrylamide gel system was similar to that described by Huang and Lehman (1972) except that the running gel consisted of a gradient of 5-15% acrylamide. Autoradiograms of the dried slab gels were made with Kodak NSZT X-ray film. from New High concentration carrier-free Na 1311 was obtained England Nuclear. Enzyme preparations were from the following sources: bromelin (1.2 U/mg), Schwartz-Mann: elastase (500 U/mg), Sigma Chemical Co; thermolysin (93,904 U/mg), Calbiothem; a-protease, Gallard Schlesinger; TPCK-treated trypsin (250 U/mg), Worthington Chemical Co; twice-crystallized papain (12.5 U/mg), Sigma Chemical Co.
approximately 2000 NIH units F. Waugh of the Massachusetts
per mg, Institute
Acknowledgments This work was supported by grants-in-aid from the American Cancer Society and the National Cancer Institute. B.R.Z. was a postdoctoral fellow of the Damon Runyon-Walter Winchell Cancer Fund, and L.B.C. a predoctoral fellow of Johnson and Johnson Co. We wish to thank Mr. Norman Hochella for technical assistance. Received
August
28, 1975;
revised
November
11, 1975
References Blumberg, Biological p. 945.
P. M.. and Robbins, Control (New York:
Blumberg,
P. M., and Robbins,
Chen, L. B., and USA 72, 131. Gahmberg, Commun.
P. W. (1975b).
Buchanan,
J. M. (1975).
C. G.. and Hakomori, 59, 283.
Hale, A. H., and Weber. Henneberry, Hogg,
P. W. (1975a). In Proteases and Cold Spring Harbor Laboratory),
N. M. (1974).
Proc.
Huang,
W. M., and Lehman,
Hynes,
R. 0. (1974).
Martin, G. S. (1970). Lepetit Colloquium, land), p. 320. McCutcheon, 1, 460.
Nat.
Biochem.
Nat. Acad.
E. (1975).
Sci. USA
I. R. (1972).
Cell 5, 1,
77, 489.
J. Biol. Chem.
D. R., and Moore,
Rein,
H. (1988).
247, 7663.
P. P. (1974). Exp.
F. B. (1948).
Biochim.
P., and Linder,
Sefton, 8. M. (1972). ley, California.
University
Ph.D.
B. M., and Rubin,
Thesis, H. (1970).
R. E.. and
Teng, N. N. H., and Chen, 72, 413. N. N. H.. and Chen,
Cancer
Biophys.
Acta
Cell Res. 49, 666.
Ruoslahti, E.. Vaheri, A., Kuusela, chim. Biophys. Acta 322, 352.
Teng,
Res.
In The Biology of Oncogenic Viruses, Second L. G. Silveston, ed. (Amsterdam: North Hol-
Agin,
Stone, K. R.. Smith, 58. 86.
Biophys.
Sci.
Cell 5, 245.
P. H., and Freese,
Phillips, D. R., and 352, 218.
Sefton,
Acad.
Cell 1, 147.
M., Coman,
A., and Rubin,
S. (1974).
N. J. (1975).
R. C., Fishman.
Cell 6, 137.
Proc.
of California,
Nature Joklik.
L. B. (1975). L. B. (1976).
E. (1973).
BioBerke-
227, 843. W. K. (1974).
Proc.
Nat. Acad.
Nature,
Virology Sci. USA
in press.
Vaheri, A., Ruoslahti. E.. and Hovi, T. (1974). In Control of Proliferation in Animal Cells, B. Clarkson and R. Baserga, eds. (New York: Cold Spring Harbor Laboratory), p. 305. Wickus. G. G., Branton. P. E., and Robbins, of Proliferation in Animal Cells, B. Clarkson (New York: Cold Spring Harbor Laboratory), Yamada, K. M., Yamada, S. S., and Pastan, Acad. Sci. USA 72, 3158.
P. S. (1974). In Control and R. Baserga, eds. p. 541. I. (1975).
Proc.
Nat.
