Journal of Surgical Oncology 1 1 :27 5-282 (1 979)
Production and Secretion of Proteolytic Enzymes by Normal and Neoplastic Cells .......................................................................................... .......................................................................................... J . BARRY BOYD, MD, RODERICK M. FARB, PhD, FRED J. YOST, Jr, PhD, NICHOLAS GEORGIADE, DDS, MD,and GERALD S. LAZARUS, MD A possible mechanism for tumor cell invasion of normal tissue might be secretion of proteolytic enzymes. This study compares and contrasts production and secretion of proteinases by cell cultures of normal and chemically transformed mouse epithelial cells. Lysates of normal and neoplastic cells contain similar amounts of neutral proteinase, cathepsin D and plasminogen activator. Neither collagenase nor elastase could be identified in lysates of, or serum-free culture medium bathing, normal or neoplastic cells. Neoplastic cells secrete ten times more plasminogen activator than normal cells. Our data support the hypothesis that plasminogen activator produced by neoplastic cells could function to activate latent proteolytic enzymes secreted by connective tissue cells which might result in spread of neoplastic cells into normal tissue.
.......................................................................................... .......................................................................................... Key words: neoplastic cells, collagenase, plasminogen activator
INTRODUCTION Neoplastic cells possess the capacity to invade normal tissue. The mechanism of tumor spread is unknown but a reasonable hypothesis might be that cancer cells produce proteolytic enzymes which degree connective tissue directly. Another possibility is that tumor cells could produce proteinases which activate precursors of connective tissue proteinases produced by other cells. Either mechanism would result in connective tissue degradation and spread of neoplastic cells. Collagenase activity has been reported in explants of human squamous cell carcinomas, basal cell epitheliomas and melanomas [ 1 ] . Tumor
From the Division of Plastic, Maxillofacial, and Reconstructive Surgery, Department of Surgery, and the Division of Dermatology, Department of Medicine, Duke University Medical Center, Durham, North Carolina. Resented at the Surgical Forum of the American College of Surgeons, October 17, 1977, Dallas. Address reprint request to,Gerald Lazarus, MD, Division of Dermatology, Department of Medicine, Duke Medical Center, Durham, NC 27710. 0022-4790/79/1103-0275$01.70 @ 1979 Alan R. Liss. Inc.
276
Boyd et a1
cells have also been reported to secrete increased amounts of plasminogen activator [2]. The purpose of this study is to compare the production and secretion of a number of tissue proteinases by cell cultures of chemically transformed mouse epithelial cells and their normal epithelial progenitor cells.
MATERIALS AND METHODS Reparation of Normal Epithelial Cell Cultures Skin from 24-48-hr-old newborn Balb C mice was removed and trypsinized (0.25% trypsin in Hank’s Balanced Salts) dermal side down for 18 hours at 4°C using the technique of Yuspa [ 3 ] . The epidermis was then mechanically separated from the dermis and minced. The preparation was agitated in media 199 plus 15% fetal calf serum (FCS) for 1 hour at 37°C. The cell suspension was centrifuged in a discontinuous Ficoll (Sigma, St Louis, Missouri) density gradient (24% w/v Ficoll diluted with media 199 to 16, 14, and 12% v/v final concentration. The gradient consisted of 5 ml of 16% Ficoll successively overlayed with 10 ml aliquots of 14% Ficoll and 12% Ficoll. The cells were centrifuged through the gradient at 1,200s for 30 minutes. The pellet at the bottom of the tube, consisting of purified epidermal cells, was resuspended in media 199 containing 15% fetal calf serum and counted in a Neubauer Ultraplane counting chamber. Assay for cell viability by Eosin Y exclusion demonstrated that 95% of the cells were alive. The cells were then plated at 9.6 X l o 6 cells/l00 mm’ surface area and grown to confluence in media 199 with 10% FCS prior to analysis. Reparation of Transformed Epithelial Cultures Mouse epithelial cells chemically transformed by methyl nitronitrosoguanidine were obtained from Drs. S. Yuspa and N. H. Coburn, National Cancer Institute, Bethesda, Maryland. The cells were plated in Dulbecco’s Modified Eagle’s Media (DMEM) plus 10% FCS and grown t o confluence prior to the study [4]. A description of the transformation of these cells has been previously reported [4]. Collection and Reparation of Culture Media Confluent plates of normal and transformed epithelial cells were cultured in media 199 or DMEM containing either 1% lactalbumin hydrolysate (LAH) or 5% (heat inactivated 56”C, 30 minutes; acid treated pH 2, 1 hour) fetal calf serum. Media was collected daily for 5 days. The media was centrifuged at 2,OOOg and the supernatant was concentrated 1O-fold by dialysis against Aquacide I11 (polyethylene glycol). Aliquots of concentrated and unconcentrated media were dialyzed against 3 M NaSCN (10 mM Tris, 250 mM NaCI, 1 mM CaCl’, pH 7.4,24 hours, 4°C) followed by dialysis against 1 0 mM Tris pH 7.4 (250 mM NaCl, 1 mM CaC12, 24 hours, 4°C) to denature a2-macroglobulin (a2-M), a proteinase inhbitor found in serum. Media was also treated with trypsin (10 p g , 30 minutes, 25°C) followed by SBTI (80 pg, 15 minutes, 25°C) prior to assay to facilitate activation of latent procollagenase or elastase [5]. Reparation of Cell Extracts Intracellular proteinases were measured in cell extracts from cultures of normal and transformed epithelial cells grown under identical conditions as previously described. In addition, extracts were prepared from cells remaining after the five-day media collection.
Proteolytic Enzymes and Neoplastic Cells
277
Cells were washed five times with phosphate buffered saline (PBS) and lysed with 0.1% Triton X-100 in M KCl 50 mM KP04, 0.01% sodium azide, pH 7.5. The lysates were frozen and thawed 5 times and centrifuged at 40,OOOg for 15 minutes. The supernatants were used for all asays. Neutral Proteinase Assay Neutral proteinase activity was measured at pH 7.5 using 3H-casein as a substrate [ 6 ] . This assay has been used routinely in our laboratory t o measure neutral proteinase activity [7,81* Cathepsin D Assay Cathepsin D activity was assayed by our standard method at pH 3.4 using [HI hemoglobin as a substrate [ 6 ] .
Elastase Assay Elastolytic activity was determined by a modification of the method of Takahashi et a1 [9] . 3Helastase was used as a substrate. Eastin (500 mg) was suspended in 25 ml of H 2 0 , the pH adjusted t o 8.9 with Na2C03, and 3.4 mg of tritiated NaBH4was mixed with the elastin suspension which was incubated at room temperature for one hour. The reaction was terminated by adding acetic acid and decreasing the pH to 3.0. Labelled elastin was centrifuged and washed until the supernatant was free of radioactivity. The elastin was then resuspended in H20 and lyophilized. Ten micrograms of H-elastin contained 3.2 X lo4 cpm. One-hundred microliter aliquots of substrate (1 mg 3H-elastin), sample, and buffer (500 mM Tris pH 8.0) were agitated for one hour at 37°C. The incubations were then centrifuged at 48,OOOg (4°C) for 15 minutes. Supernatant solution (100 pl) was counted in 10 ml of Aquasol I1 in a liquid scintillation counter. Collagenase Assay The reconstituted l4 C-collagen fibril assay was used to measure collagenase activity [ 101 . Collagen was solubilized in 0.2 M acetic acid (4 mg/ml, 18 hours, 4OC) prior t o dialysis against large volumes of 10 mM Tris (250 mM NaCl, 1 mM CaC12, pH 7.4, 18 hours, 4°C). The dialyzed collagen was centrifuged at 48,OOOg (4OC) for 15 minutes and lOO-/.d aliquots of the supernatant were pipette into microfuge tubes. The tubes were incubated at 37°C for 18 hours to form reconstituted collagen fibrils prior to assay. One hundred microliters of sample and 100 pl of buffer (1 0 mM Tris, 1 mM CaC12,pH 7.4) were added to the fibrils and incubated at 37°C for 1 8 hours. At the conclusion of the assay the tubes were centrifuged at 48,OOOg (4°C) for 15 minutes. One hundred microliters of supernatant were counted in 10 ml of Aquasol I1 in a liquid scintillation spectrometer. Polyacrylamide Gel Electrophoresis ''C-Glycine-labeled rat skin collagen was solubilized in 0.2 M acetic acid (4 mg/ml, 18 hours, 4°C) prior to dialysis against large volumes of 10 mM Tris (1 M NaCl, 10 mM CaC12, pH 7.4, 18 hours, 4°C). The dialyzed collagen was centrifuged at 48,OOOg (4°C) for 15 minutes. One-hundred microliter aliquots of the supernatant were incubated with 100 pl of buffer (10 mM Tris, 1 mM CaClz, pH 7.4) for 1 8 hours a t 25°C. The reaction was terminated by addition of solid urea t o a final concentration of 7.5 M, and the mixture was heated at 50°C for 30 minutes. The reaction mixture was dialyzed against 8 M urea (pH 3.5, 18 hours, 4"C), and 100 pl of the preparations were analyzed by polyacrylamide disc gel
278
Boyd et al
electrophoresis to determine products of collagen digestion [l I ] . The gels were run at 3 mA/tube for 1 hour, stained with 1.4% Coomassie Blue and 2.7% Amido Black for 30 minutes and then destained with 7% acetic acid and 5% methanol for 18 hours.
Autoradiography Four-hundred-microliter aliquots of l4 C-collagen was incubated with 400-pl samples of media and 400 pl of buffer for 18 hours at 25°C. The preparations were dialyzed against large volumes of H20 (1 8 hours, 4"C), lyophilized and resuspended in 150 pl of 8 M urea (pH 3.5) and heated at 50°C for 30 minutes. Aliquots of this preparation were then analyzed by slab-gel electrophoresis [12]. Gels were prepared for autoradiography by a slight modification of the method of Bonner and Laskey [12]. Following electrophoresis the gels were soaked twice in dimethyl sulfoxide (DMSO) followed by 2,s diphenyloxazole (20 gm in 100 ml DMSO, three hours, room temperature), H 2 0 for one hour and 3% glycerol for 20 minutes. The gels were dried in a gel dryer for one hour and then exposed to Gaf-1000 photographic film in a Kodak exposure holder for three weeks at -70°C. After warming, the film was processed in a Kodak M-6 developer. Plasminogen Activator Assay Plasminogen activator activity was determined by measuring plasminogen-dependent lysis of '21 I-fibrin as described by Berger [ 131.
SDS Electrophoresis Ten-milliliter aliquots of media and cell lysate were incubated with 50 1.11 of Hdiisopropyl fluorophosphate (DFP, 30 minutes, room temperature) prior to dialysis against 0.05 M phosphate buffer (18 hours, 4OC). The preparations were lyophilized and redissolved in 200 pl of H 2 0 . One hundred microliters was analyzed by sodium dodecylsulfate (SDS) polyacrylamide disc-gel electrophoresis to identify serine class neutral proteinase [ 6 ] .
Protein and DNA Protein was determined by the microbiuret assay of Leggett-Bailey [14] . Cells were hydrolyzed in 1 .O N perchloric acid and DNA determined colorimetrically as described by Burton [15].
RESULTS Extracts of normal and transformed epithelial cells contained similar amounts of neutral proteinase activity. After growth for five days in 1% LAH media, normal epithelial cells increased the specific activity of neutral proteinase sixfold. By contrast, the specific activity of neutral proteinase in neoplastic cells did not change. The increase in neutral proteinase activity in normal cells may reflect autolytic processes since normal epithelial cells tolerate serum-free medium less well than tumor cells. Neither cell type secreted neutral proteinase into the medium. Lysates from normal and transformed epithelial cells contain similar amounts of cathepsin D activity. No cathepsin D activity could be measured in the media from either normal or tumor cell cultures, indicating that cathepsin D is not secreted by these cells
Proteolytic Enzymes and Neoplastic Cells
279
and that under the conditions of culture the cells were not dying and subsequently releasing lysosomal enzymes into the media. Elastase activity could not be found in lysates of either normal or neoplastic cells. Likewise, no activity could be measured in media from normal or transformed cell cultures. Even lysates of media activated with thiocyanate or with trypsin contained no measurable elastase activity. Collagenase activity could not be detected by radioactive fibril collagenase assay in cell lysates or media from either normal or tumor cell cultures. Lysates or media activated with thiocyanate or with trypsin contained no measurable collagenase activity. Collagenase activity could not be detected in media or cell lysates utilizing the more sensitive polyacrylamide disc-gel electrophoresis techniques. Incubation of medium containing heat- and aciddenatured serum from normal and neoplastic cells with radioactive collagen did not produce collagen degradation peptides when mixtures were analyzed by acrylamide gel electrophoresis autoradiographic techniques. Normal and transformed cells were found to contain a similar amount of plasminogen activator as shown in Figure 1. There was a threefold increase in plasminogen activator activity at the end of the culture period in both cell lines. Neoplastic cells in culture secrete tenfold more plasminogen activator than normal epithelial cells (Fig. 2). Plasminogen activator activity is totally inhibited by DFP. SDS polyacrylamide gel electrophoresis of media treated with tritium-labeled diisopropyl fluorophosphate (DFP) demonstrated a single peak of radioactivity (Fig. 3). Since no diisopropyl fluorophosphate inhibitable caseinolytic activity is secreted into the medium this peak probably represents plasminogen activator. The apparent molecular weight of this single peak corresponds with the expected molecular weight of 55,000 found in plasminogen activator [ 161. 4
8
30r
3 I0
Normal epithelio cell extrocts
;I
cell extrocts
$ 20 0
E
15
5 10 0 0,
E E
1
s 5
a
c
0
v)
c .2 c
I
5
Fig. 1. Production of plasminogen activator. The ordinate is specific activity of plasminogen activator per cell number. Unshaded bars are normal epithelial cell extracts, and shaded bars are transformed cell extracts at day 1 (prior to collection) and day 5 (after collection).
280
Boydet al 25
sa
Transformed epithelial cell media
-g g 20 G z c alx c
g
6
5 5 € a
I5
L -g ism 10 ==:h c x
=
5 2
I
3
5
4
h Y Fig. 2. Secretion of plasminogen activator. The ordinate is specific activity of plasminogen activator per cell number. Unshaded bars are normal epithelial cell media, and shaded bars are transformed cell media on days 1 through 5 .
400 1
350 300 -
2 200 250
150
-
100
-
50 L
I
,
I
DISCUSSION In previous studies, whole tissue explants of human carcinomas demonstrated collagenase activity [ 1 1 .These mixed cell systems contain normal epithelial cells, carcinoma cells, inflammatory cells, and connective tissue cells [ 1] . The complexity of these organ culture systems makes it difficult to analyze the contribution of the neoplastic cells per se. Our study demonstrates that normal epithelial and neoplastic cells in culture contain similar intracellular levels of neutral proteinase, cathepsin D, and plasminogen activator. We were not able to demonstrate the presence of elastase or collagenase in cells or in the culture medium, under our experimental conditions. Neoplastic epithelial cells secreted more plasminogen activator into the medium than normal cells.
Proteolytic Enzymes and Neoplastic Cells
281
Plasminogen activator activity has been demonstrated in normal processes such as rupture of ovarian follicles [17, 183 and in embryogenesis [19]. Some lines of neoplastic cells synthesize and secrete increased amounts of plasminogen activator in a continuous and uncontrolled manner [20]. Plasminogen activator may play a direct role in the loss of contact inhibition in neoplasia [20]. Plasmin is the product of plasminogen activator activity. Werb et al [21] have recently demonstrated that latent collagenase secreted by synovial cells in culture is activated by plasminogen activator from the same cells. This observation suggests that plasminogen activator could induce activation of connective tissue proteinases. These published experiments coupled with our observations suggest the following hypothesis. Neoplastic transformation of mouse epithelium results in induction of plasminogen activator activity. This proteinase then activates collagenase proenzyme which is produced in fibroblasts [22]. Activation of connective tissue proteinase could then facilitate spread of neoplastic cells into normal tissue. It is possible that other epithelial neoplastic cells produce connective tissue proteinases directly, but our hypothesis does not require such a mechanism.
ACKNOWLEDGMENTS Supported by the National Institute of Arthritis, Metabolic, and Digestive Diseases (7 R01 AM 17370) and the Fred Hutchison Cancer Foundation. We deeply appreciate the transformed mouse epidermal cell line and helpful discussions provided by Dr. N. H. Coburn and Dr.S. Yuspa, National Cancer Institute, Bethesda, Maryland. We think Dr.H. Berger, Wellcome Research Laboratories, Research Triangle Park, North Carolina, for performing the plasminogen activator assays. Ms Deborah A. Cleary provided expert technical assistance in the preparation of the manuscript. This is Publication No. 41 of the Dermatological Research Laboratories of the Duke University Medical Center.
REFERENCES 1. Hashimoto K, Yaminishi Y, Maeyens E, Dabbous MK, Kanzaki T: Collagenolytic activity of squamous cell carcinoma of the skin. Cancer Res 33:2790-2801,1973. 2. Coldberg AR, Wolf BA, Lefebore PA: Plasminogen activators of transformed and normal cells. In Reich E, Rifkin DB, Shaw E (eds): “Proteases and Biological Control.” New York: Cold Spring Harbor Laboratory, 1975, pp 857-868. 3. Elgjo K, Hennings H, Michael D,Yuspa SH: Natural synchrony of newborn mouse epidermal cells in vitro. J Invest Dermatol 66:292-296, 1976. 4. Coburn N, Vorder Bruegge WS, Bates JR, Gray RH, Rossen JD, Kelsey WH, Shimata T: Correlation of anchorage-independent growth with tumorigenecity of chemically transformed mouse epidermal cells. Cancer Res (in. press). 5 . Vaes G: The release of collagenase as an inactive proenzyme by bone explants in culture. Biochem J 126~275-289,1972. 6. Hatcher VB, Lazarus GS, Levine N, Burk P, Yost FJ: Characterization of a chemotactic and cytotoxic proteinase from human skin. Biochim Biophys Acta 483: 160-171, 1977. 7. Lazarus GS. Yost FJ, Thomas CA: Polymorphonuclear leukocytes: Possible mechanism of accumulation in psoriasis. Science 198:1162-1163, 1977. 8. Thomas CA, Yost FJ, Snyderman R, Hatcher VB, Lazarus CS: Cellular serine proteinase induces chemotaxis by complement activation. Nature 269521, 1977. 9. Takahashi S, Seifter S, Yang GC: A new radioactive assay for enzymes with elastolytic activity using reduced tritiated elastin. The effect of sodium dodecyl sulfate on elastolysis. Biochim Biophysl Acta 327:138-145, 1973. 10. Lazarus GS, Daniels JR, Brown RS, Bladen HA, Fullmer HM: Degradation of collagen by a human granulocyte collagenolytic system. J Clin Invest 47:2622-2629, 1968.
282
Boyd et a1
1 1 . Sakai T, Gross J: Some properties of the products of reaction of tadpole collagenase with collagen. Biochem. 6:518-523,1967. 12. Bonner WM,Laskey RA: A film detection method for tritium-labelled proteins and nucleic acids in polyacrylamide gels. Eur. J Biochem 46:83-88, 1974. 13. Berger H: Secretion of plasminogen activator by rheumatoid and nonrheumatoid synovial cells in culture. Arthritis Rheum 20: 1198- 1205, 1977. 14. Leggett-Bailey J: In “Techniques in Protein Chemistry.” New York: American Elsevier, 1962, pp 294-296. 15. Burton K : A study of the conditionsand mechanism of the diphenylamine reaction from the colorimetric estimation of deoxynucleic acid. Biochem J 62: 315-323, 1966. 16. Christman JK. Silverstein SC, Acs G: Plasminogen activators. Barrett AJ (ed): “Proteinases in Mammalian Cells and Tissues.:’ Amsterdam: North-Holland, 1977, pp 91 - 150. 17. Espey LL: Ovarian proteolytic enzymes and ovulation. Biol Reprod 10:216-235, 1974. 18. Koutsk YJ, Rybakov M, Rybakov AB: Plasminogen activator and other trypsin-like proteases in the uterus wall and their participation in tissue bleeding. Arch Cynaekol 220:8-11, 1975. 19. Strickland S , Reich E, Sherman MI: Plasminogen activator in early embryogenesis: Enzyme production by trophoblast and parietal endoderm. Cell 9:231-240, 1976. 20. Christman JK, Acs G, Silagi S, Silverstein SC: Plasminogen activator: Biochemical characterization and correlation with tumorigenecity. In Reich E, Rifkin DB, Shaw E (eds): “Proteases and Biological Control.” New York: Cold Spring Harbor Laboratory, 1975, pp 827-839. 21. Werb 2, Mainardi MD, Vater BA, Harris ED Jr: Endogenous activation of latent collagenase by rheumatoid synovial cells: Evidence for a role of plasminogen activator. N Engl J Med 296: 10171023,1977. 22. Bauer EA, Stricklin CS, Jeffrey J J , Eisen AZ: Collagenase production by human skin fibroblasts. Biophys Res Commun 64:232,1975.
Production and Secretion of Proteolytic Enzymes by Normal and Neoplastic Cells .......................................................................................... .......................................................................................... J . BARRY BOYD, MD, RODERICK M. FARB, PhD, FRED J. YOST, Jr, PhD, NICHOLAS GEORGIADE, DDS, MD,and GERALD S. LAZARUS, MD A possible mechanism for tumor cell invasion of normal tissue might be secretion of proteolytic enzymes. This study compares and contrasts production and secretion of proteinases by cell cultures of normal and chemically transformed mouse epithelial cells. Lysates of normal and neoplastic cells contain similar amounts of neutral proteinase, cathepsin D and plasminogen activator. Neither collagenase nor elastase could be identified in lysates of, or serum-free culture medium bathing, normal or neoplastic cells. Neoplastic cells secrete ten times more plasminogen activator than normal cells. Our data support the hypothesis that plasminogen activator produced by neoplastic cells could function to activate latent proteolytic enzymes secreted by connective tissue cells which might result in spread of neoplastic cells into normal tissue.
.......................................................................................... .......................................................................................... Key words: neoplastic cells, collagenase, plasminogen activator
INTRODUCTION Neoplastic cells possess the capacity to invade normal tissue. The mechanism of tumor spread is unknown but a reasonable hypothesis might be that cancer cells produce proteolytic enzymes which degree connective tissue directly. Another possibility is that tumor cells could produce proteinases which activate precursors of connective tissue proteinases produced by other cells. Either mechanism would result in connective tissue degradation and spread of neoplastic cells. Collagenase activity has been reported in explants of human squamous cell carcinomas, basal cell epitheliomas and melanomas [ 1 ] . Tumor
From the Division of Plastic, Maxillofacial, and Reconstructive Surgery, Department of Surgery, and the Division of Dermatology, Department of Medicine, Duke University Medical Center, Durham, North Carolina. Resented at the Surgical Forum of the American College of Surgeons, October 17, 1977, Dallas. Address reprint request to,Gerald Lazarus, MD, Division of Dermatology, Department of Medicine, Duke Medical Center, Durham, NC 27710. 0022-4790/79/1103-0275$01.70 @ 1979 Alan R. Liss. Inc.
276
Boyd et a1
cells have also been reported to secrete increased amounts of plasminogen activator [2]. The purpose of this study is to compare the production and secretion of a number of tissue proteinases by cell cultures of chemically transformed mouse epithelial cells and their normal epithelial progenitor cells.
MATERIALS AND METHODS Reparation of Normal Epithelial Cell Cultures Skin from 24-48-hr-old newborn Balb C mice was removed and trypsinized (0.25% trypsin in Hank’s Balanced Salts) dermal side down for 18 hours at 4°C using the technique of Yuspa [ 3 ] . The epidermis was then mechanically separated from the dermis and minced. The preparation was agitated in media 199 plus 15% fetal calf serum (FCS) for 1 hour at 37°C. The cell suspension was centrifuged in a discontinuous Ficoll (Sigma, St Louis, Missouri) density gradient (24% w/v Ficoll diluted with media 199 to 16, 14, and 12% v/v final concentration. The gradient consisted of 5 ml of 16% Ficoll successively overlayed with 10 ml aliquots of 14% Ficoll and 12% Ficoll. The cells were centrifuged through the gradient at 1,200s for 30 minutes. The pellet at the bottom of the tube, consisting of purified epidermal cells, was resuspended in media 199 containing 15% fetal calf serum and counted in a Neubauer Ultraplane counting chamber. Assay for cell viability by Eosin Y exclusion demonstrated that 95% of the cells were alive. The cells were then plated at 9.6 X l o 6 cells/l00 mm’ surface area and grown to confluence in media 199 with 10% FCS prior to analysis. Reparation of Transformed Epithelial Cultures Mouse epithelial cells chemically transformed by methyl nitronitrosoguanidine were obtained from Drs. S. Yuspa and N. H. Coburn, National Cancer Institute, Bethesda, Maryland. The cells were plated in Dulbecco’s Modified Eagle’s Media (DMEM) plus 10% FCS and grown t o confluence prior to the study [4]. A description of the transformation of these cells has been previously reported [4]. Collection and Reparation of Culture Media Confluent plates of normal and transformed epithelial cells were cultured in media 199 or DMEM containing either 1% lactalbumin hydrolysate (LAH) or 5% (heat inactivated 56”C, 30 minutes; acid treated pH 2, 1 hour) fetal calf serum. Media was collected daily for 5 days. The media was centrifuged at 2,OOOg and the supernatant was concentrated 1O-fold by dialysis against Aquacide I11 (polyethylene glycol). Aliquots of concentrated and unconcentrated media were dialyzed against 3 M NaSCN (10 mM Tris, 250 mM NaCI, 1 mM CaCl’, pH 7.4,24 hours, 4°C) followed by dialysis against 1 0 mM Tris pH 7.4 (250 mM NaCl, 1 mM CaC12, 24 hours, 4°C) to denature a2-macroglobulin (a2-M), a proteinase inhbitor found in serum. Media was also treated with trypsin (10 p g , 30 minutes, 25°C) followed by SBTI (80 pg, 15 minutes, 25°C) prior to assay to facilitate activation of latent procollagenase or elastase [5]. Reparation of Cell Extracts Intracellular proteinases were measured in cell extracts from cultures of normal and transformed epithelial cells grown under identical conditions as previously described. In addition, extracts were prepared from cells remaining after the five-day media collection.
Proteolytic Enzymes and Neoplastic Cells
277
Cells were washed five times with phosphate buffered saline (PBS) and lysed with 0.1% Triton X-100 in M KCl 50 mM KP04, 0.01% sodium azide, pH 7.5. The lysates were frozen and thawed 5 times and centrifuged at 40,OOOg for 15 minutes. The supernatants were used for all asays. Neutral Proteinase Assay Neutral proteinase activity was measured at pH 7.5 using 3H-casein as a substrate [ 6 ] . This assay has been used routinely in our laboratory t o measure neutral proteinase activity [7,81* Cathepsin D Assay Cathepsin D activity was assayed by our standard method at pH 3.4 using [HI hemoglobin as a substrate [ 6 ] .
Elastase Assay Elastolytic activity was determined by a modification of the method of Takahashi et a1 [9] . 3Helastase was used as a substrate. Eastin (500 mg) was suspended in 25 ml of H 2 0 , the pH adjusted t o 8.9 with Na2C03, and 3.4 mg of tritiated NaBH4was mixed with the elastin suspension which was incubated at room temperature for one hour. The reaction was terminated by adding acetic acid and decreasing the pH to 3.0. Labelled elastin was centrifuged and washed until the supernatant was free of radioactivity. The elastin was then resuspended in H20 and lyophilized. Ten micrograms of H-elastin contained 3.2 X lo4 cpm. One-hundred microliter aliquots of substrate (1 mg 3H-elastin), sample, and buffer (500 mM Tris pH 8.0) were agitated for one hour at 37°C. The incubations were then centrifuged at 48,OOOg (4°C) for 15 minutes. Supernatant solution (100 pl) was counted in 10 ml of Aquasol I1 in a liquid scintillation counter. Collagenase Assay The reconstituted l4 C-collagen fibril assay was used to measure collagenase activity [ 101 . Collagen was solubilized in 0.2 M acetic acid (4 mg/ml, 18 hours, 4OC) prior t o dialysis against large volumes of 10 mM Tris (250 mM NaCl, 1 mM CaC12, pH 7.4, 18 hours, 4°C). The dialyzed collagen was centrifuged at 48,OOOg (4OC) for 15 minutes and lOO-/.d aliquots of the supernatant were pipette into microfuge tubes. The tubes were incubated at 37°C for 18 hours to form reconstituted collagen fibrils prior to assay. One hundred microliters of sample and 100 pl of buffer (1 0 mM Tris, 1 mM CaC12,pH 7.4) were added to the fibrils and incubated at 37°C for 1 8 hours. At the conclusion of the assay the tubes were centrifuged at 48,OOOg (4°C) for 15 minutes. One hundred microliters of supernatant were counted in 10 ml of Aquasol I1 in a liquid scintillation spectrometer. Polyacrylamide Gel Electrophoresis ''C-Glycine-labeled rat skin collagen was solubilized in 0.2 M acetic acid (4 mg/ml, 18 hours, 4°C) prior to dialysis against large volumes of 10 mM Tris (1 M NaCl, 10 mM CaC12, pH 7.4, 18 hours, 4°C). The dialyzed collagen was centrifuged at 48,OOOg (4°C) for 15 minutes. One-hundred microliter aliquots of the supernatant were incubated with 100 pl of buffer (10 mM Tris, 1 mM CaClz, pH 7.4) for 1 8 hours a t 25°C. The reaction was terminated by addition of solid urea t o a final concentration of 7.5 M, and the mixture was heated at 50°C for 30 minutes. The reaction mixture was dialyzed against 8 M urea (pH 3.5, 18 hours, 4"C), and 100 pl of the preparations were analyzed by polyacrylamide disc gel
278
Boyd et al
electrophoresis to determine products of collagen digestion [l I ] . The gels were run at 3 mA/tube for 1 hour, stained with 1.4% Coomassie Blue and 2.7% Amido Black for 30 minutes and then destained with 7% acetic acid and 5% methanol for 18 hours.
Autoradiography Four-hundred-microliter aliquots of l4 C-collagen was incubated with 400-pl samples of media and 400 pl of buffer for 18 hours at 25°C. The preparations were dialyzed against large volumes of H20 (1 8 hours, 4"C), lyophilized and resuspended in 150 pl of 8 M urea (pH 3.5) and heated at 50°C for 30 minutes. Aliquots of this preparation were then analyzed by slab-gel electrophoresis [12]. Gels were prepared for autoradiography by a slight modification of the method of Bonner and Laskey [12]. Following electrophoresis the gels were soaked twice in dimethyl sulfoxide (DMSO) followed by 2,s diphenyloxazole (20 gm in 100 ml DMSO, three hours, room temperature), H 2 0 for one hour and 3% glycerol for 20 minutes. The gels were dried in a gel dryer for one hour and then exposed to Gaf-1000 photographic film in a Kodak exposure holder for three weeks at -70°C. After warming, the film was processed in a Kodak M-6 developer. Plasminogen Activator Assay Plasminogen activator activity was determined by measuring plasminogen-dependent lysis of '21 I-fibrin as described by Berger [ 131.
SDS Electrophoresis Ten-milliliter aliquots of media and cell lysate were incubated with 50 1.11 of Hdiisopropyl fluorophosphate (DFP, 30 minutes, room temperature) prior to dialysis against 0.05 M phosphate buffer (18 hours, 4OC). The preparations were lyophilized and redissolved in 200 pl of H 2 0 . One hundred microliters was analyzed by sodium dodecylsulfate (SDS) polyacrylamide disc-gel electrophoresis to identify serine class neutral proteinase [ 6 ] .
Protein and DNA Protein was determined by the microbiuret assay of Leggett-Bailey [14] . Cells were hydrolyzed in 1 .O N perchloric acid and DNA determined colorimetrically as described by Burton [15].
RESULTS Extracts of normal and transformed epithelial cells contained similar amounts of neutral proteinase activity. After growth for five days in 1% LAH media, normal epithelial cells increased the specific activity of neutral proteinase sixfold. By contrast, the specific activity of neutral proteinase in neoplastic cells did not change. The increase in neutral proteinase activity in normal cells may reflect autolytic processes since normal epithelial cells tolerate serum-free medium less well than tumor cells. Neither cell type secreted neutral proteinase into the medium. Lysates from normal and transformed epithelial cells contain similar amounts of cathepsin D activity. No cathepsin D activity could be measured in the media from either normal or tumor cell cultures, indicating that cathepsin D is not secreted by these cells
Proteolytic Enzymes and Neoplastic Cells
279
and that under the conditions of culture the cells were not dying and subsequently releasing lysosomal enzymes into the media. Elastase activity could not be found in lysates of either normal or neoplastic cells. Likewise, no activity could be measured in media from normal or transformed cell cultures. Even lysates of media activated with thiocyanate or with trypsin contained no measurable elastase activity. Collagenase activity could not be detected by radioactive fibril collagenase assay in cell lysates or media from either normal or tumor cell cultures. Lysates or media activated with thiocyanate or with trypsin contained no measurable collagenase activity. Collagenase activity could not be detected in media or cell lysates utilizing the more sensitive polyacrylamide disc-gel electrophoresis techniques. Incubation of medium containing heat- and aciddenatured serum from normal and neoplastic cells with radioactive collagen did not produce collagen degradation peptides when mixtures were analyzed by acrylamide gel electrophoresis autoradiographic techniques. Normal and transformed cells were found to contain a similar amount of plasminogen activator as shown in Figure 1. There was a threefold increase in plasminogen activator activity at the end of the culture period in both cell lines. Neoplastic cells in culture secrete tenfold more plasminogen activator than normal epithelial cells (Fig. 2). Plasminogen activator activity is totally inhibited by DFP. SDS polyacrylamide gel electrophoresis of media treated with tritium-labeled diisopropyl fluorophosphate (DFP) demonstrated a single peak of radioactivity (Fig. 3). Since no diisopropyl fluorophosphate inhibitable caseinolytic activity is secreted into the medium this peak probably represents plasminogen activator. The apparent molecular weight of this single peak corresponds with the expected molecular weight of 55,000 found in plasminogen activator [ 161. 4
8
30r
3 I0
Normal epithelio cell extrocts
;I
cell extrocts
$ 20 0
E
15
5 10 0 0,
E E
1
s 5
a
c
0
v)
c .2 c
I
5
Fig. 1. Production of plasminogen activator. The ordinate is specific activity of plasminogen activator per cell number. Unshaded bars are normal epithelial cell extracts, and shaded bars are transformed cell extracts at day 1 (prior to collection) and day 5 (after collection).
280
Boydet al 25
sa
Transformed epithelial cell media
-g g 20 G z c alx c
g
6
5 5 € a
I5
L -g ism 10 ==:h c x
=
5 2
I
3
5
4
h Y Fig. 2. Secretion of plasminogen activator. The ordinate is specific activity of plasminogen activator per cell number. Unshaded bars are normal epithelial cell media, and shaded bars are transformed cell media on days 1 through 5 .
400 1
350 300 -
2 200 250
150
-
100
-
50 L
I
,
I
DISCUSSION In previous studies, whole tissue explants of human carcinomas demonstrated collagenase activity [ 1 1 .These mixed cell systems contain normal epithelial cells, carcinoma cells, inflammatory cells, and connective tissue cells [ 1] . The complexity of these organ culture systems makes it difficult to analyze the contribution of the neoplastic cells per se. Our study demonstrates that normal epithelial and neoplastic cells in culture contain similar intracellular levels of neutral proteinase, cathepsin D, and plasminogen activator. We were not able to demonstrate the presence of elastase or collagenase in cells or in the culture medium, under our experimental conditions. Neoplastic epithelial cells secreted more plasminogen activator into the medium than normal cells.
Proteolytic Enzymes and Neoplastic Cells
281
Plasminogen activator activity has been demonstrated in normal processes such as rupture of ovarian follicles [17, 183 and in embryogenesis [19]. Some lines of neoplastic cells synthesize and secrete increased amounts of plasminogen activator in a continuous and uncontrolled manner [20]. Plasminogen activator may play a direct role in the loss of contact inhibition in neoplasia [20]. Plasmin is the product of plasminogen activator activity. Werb et al [21] have recently demonstrated that latent collagenase secreted by synovial cells in culture is activated by plasminogen activator from the same cells. This observation suggests that plasminogen activator could induce activation of connective tissue proteinases. These published experiments coupled with our observations suggest the following hypothesis. Neoplastic transformation of mouse epithelium results in induction of plasminogen activator activity. This proteinase then activates collagenase proenzyme which is produced in fibroblasts [22]. Activation of connective tissue proteinase could then facilitate spread of neoplastic cells into normal tissue. It is possible that other epithelial neoplastic cells produce connective tissue proteinases directly, but our hypothesis does not require such a mechanism.
ACKNOWLEDGMENTS Supported by the National Institute of Arthritis, Metabolic, and Digestive Diseases (7 R01 AM 17370) and the Fred Hutchison Cancer Foundation. We deeply appreciate the transformed mouse epidermal cell line and helpful discussions provided by Dr. N. H. Coburn and Dr.S. Yuspa, National Cancer Institute, Bethesda, Maryland. We think Dr.H. Berger, Wellcome Research Laboratories, Research Triangle Park, North Carolina, for performing the plasminogen activator assays. Ms Deborah A. Cleary provided expert technical assistance in the preparation of the manuscript. This is Publication No. 41 of the Dermatological Research Laboratories of the Duke University Medical Center.
REFERENCES 1. Hashimoto K, Yaminishi Y, Maeyens E, Dabbous MK, Kanzaki T: Collagenolytic activity of squamous cell carcinoma of the skin. Cancer Res 33:2790-2801,1973. 2. Coldberg AR, Wolf BA, Lefebore PA: Plasminogen activators of transformed and normal cells. In Reich E, Rifkin DB, Shaw E (eds): “Proteases and Biological Control.” New York: Cold Spring Harbor Laboratory, 1975, pp 857-868. 3. Elgjo K, Hennings H, Michael D,Yuspa SH: Natural synchrony of newborn mouse epidermal cells in vitro. J Invest Dermatol 66:292-296, 1976. 4. Coburn N, Vorder Bruegge WS, Bates JR, Gray RH, Rossen JD, Kelsey WH, Shimata T: Correlation of anchorage-independent growth with tumorigenecity of chemically transformed mouse epidermal cells. Cancer Res (in. press). 5 . Vaes G: The release of collagenase as an inactive proenzyme by bone explants in culture. Biochem J 126~275-289,1972. 6. Hatcher VB, Lazarus GS, Levine N, Burk P, Yost FJ: Characterization of a chemotactic and cytotoxic proteinase from human skin. Biochim Biophys Acta 483: 160-171, 1977. 7. Lazarus GS. Yost FJ, Thomas CA: Polymorphonuclear leukocytes: Possible mechanism of accumulation in psoriasis. Science 198:1162-1163, 1977. 8. Thomas CA, Yost FJ, Snyderman R, Hatcher VB, Lazarus CS: Cellular serine proteinase induces chemotaxis by complement activation. Nature 269521, 1977. 9. Takahashi S, Seifter S, Yang GC: A new radioactive assay for enzymes with elastolytic activity using reduced tritiated elastin. The effect of sodium dodecyl sulfate on elastolysis. Biochim Biophysl Acta 327:138-145, 1973. 10. Lazarus GS, Daniels JR, Brown RS, Bladen HA, Fullmer HM: Degradation of collagen by a human granulocyte collagenolytic system. J Clin Invest 47:2622-2629, 1968.
282
Boyd et a1
1 1 . Sakai T, Gross J: Some properties of the products of reaction of tadpole collagenase with collagen. Biochem. 6:518-523,1967. 12. Bonner WM,Laskey RA: A film detection method for tritium-labelled proteins and nucleic acids in polyacrylamide gels. Eur. J Biochem 46:83-88, 1974. 13. Berger H: Secretion of plasminogen activator by rheumatoid and nonrheumatoid synovial cells in culture. Arthritis Rheum 20: 1198- 1205, 1977. 14. Leggett-Bailey J: In “Techniques in Protein Chemistry.” New York: American Elsevier, 1962, pp 294-296. 15. Burton K : A study of the conditionsand mechanism of the diphenylamine reaction from the colorimetric estimation of deoxynucleic acid. Biochem J 62: 315-323, 1966. 16. Christman JK. Silverstein SC, Acs G: Plasminogen activators. Barrett AJ (ed): “Proteinases in Mammalian Cells and Tissues.:’ Amsterdam: North-Holland, 1977, pp 91 - 150. 17. Espey LL: Ovarian proteolytic enzymes and ovulation. Biol Reprod 10:216-235, 1974. 18. Koutsk YJ, Rybakov M, Rybakov AB: Plasminogen activator and other trypsin-like proteases in the uterus wall and their participation in tissue bleeding. Arch Cynaekol 220:8-11, 1975. 19. Strickland S , Reich E, Sherman MI: Plasminogen activator in early embryogenesis: Enzyme production by trophoblast and parietal endoderm. Cell 9:231-240, 1976. 20. Christman JK, Acs G, Silagi S, Silverstein SC: Plasminogen activator: Biochemical characterization and correlation with tumorigenecity. In Reich E, Rifkin DB, Shaw E (eds): “Proteases and Biological Control.” New York: Cold Spring Harbor Laboratory, 1975, pp 827-839. 21. Werb 2, Mainardi MD, Vater BA, Harris ED Jr: Endogenous activation of latent collagenase by rheumatoid synovial cells: Evidence for a role of plasminogen activator. N Engl J Med 296: 10171023,1977. 22. Bauer EA, Stricklin CS, Jeffrey J J , Eisen AZ: Collagenase production by human skin fibroblasts. Biophys Res Commun 64:232,1975.
