SEMINARS IN THROMBOSIS AND HEMOSTASIS—VOLUME 17, NO. 3, 1991
Epidermal Growth Factor and 12-Tetradecanoyl Phorbol 13-Acetate Induction of Urokinase in A431 Cells
The accumulated evidence suggests that plasminogen activators (PAs) do play a role in invasion and metastasis by cancer cells. 1 , 2 PAs, and in particular urokinase (u-PA), can be induced by a large variety of agents. Two agents, an epidermal growth factor (EGF) and a tumor promoter, 12-tetradecanoyl phorbol 13acetate (TPA), interact with specific receptors at the cell membrane, the EGF receptor and protein kinase C, respectively. In the study of Degen et al 3 TPA transiently increased the u-PA mRNA levels, following which the cells became unresponsive to TPA-inducible u-PA mRNA synthesis. This desensitization, however, by TPA did not impair the stimulation of u-PA mRNA synthesis by calcitonin, vasopressin, or 8-Br-cyclic adenosine monophosphate (cAMP), suggesting at least two separate induction pathways for u-PA. At the other end of the induction pathway, several regulatory elements in the 5'-flanking region of the u-PA gene have been identified. The regulatory sites so far identified are two cAMP-responsive elements (CRE), one reported by Nagamine and Reich 4 and the other by Nakagawa et al. 5 An AP-2 site, responsive to both cAMP and phorbol esters, was shown by Nagamine et al 6 and an AP-1 site by Nerlov et al. 7 The AP-1 site is known to interact with the fosljun heterodimer. Expression of fas and jun on EGF 8 and TPA stimulation was shown by Angel et al 9
and by Lee et al. 1 0 Thus, the initial (receptors) and terminal (regulatory elements) stages of the induction process have been investigated; however, little is known about the steps of signal transduction between these points for any of the u-PA inducers. The aim of the present study was to characterize, as far as possible, the basic mechanisms for signal transduction. Since information is already available on EGF- and TPA-inducing effects on u-PA, we chose to study the induction pathways utilized by these agents.
MATERIALS AND METHODS Cell Culture Human A431 epidermoid carcinoma cells were obtained from American Type Culture Collection and maintained in RPMI-1640 with 10% (vol/vol) fetal bovine serum (J.R. Scientific, Woodland, CA or Gibco Laboratories Grand Island, NY) in 5% carbon dioxide and 95% air at 37°C. Single-cell suspensions of A431 cells were counted with a hemocytometer after 10 minutes' incubation at 37°C with 0.05% trypsin (Gibco Laboratories). Cells were grown to more than 90% confluency.
RNA Isolation From the Department of Experimental Biology, Roswell Park Cancer Institute, New York State Department of Health, Buffalo, New York. This work was supported in part by the Mark Diamond Research Fund from the Graduate Student Association of SUNYAB (T.L.K.), and by grant 1 P01 CA28853 from the National Cancer Institute (G.M.). Reprint requests: Dr. Markus, Dept. of Exp. Biology, Roswell Park Cancer Institute, New York State Department of Health, Buffalo, NY 14263.
Total cellular RNA was isolated by the method described by Chirgwin et al;11 Northern blot analysis was described by Maniatis et al. 12
Sample Preparation for Chromogenic Assay For each experimental condition, cultures were analyzed in duplicate or quadruplicate, as indicated. The
Copyright © 1991 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
217
Downloaded by: University of British Columbia. Copyrighted material.
TIMOTHY L. KESSLER, Ph.D., and GABOR MARKUS, M.D., Ph.D.
SEMINARS IN THROMBOSIS AND HEMOSTASIS—VOLUME 17, NO. 3, 1991
cultures were washed with serum-free medium to wash out the fetal bovine serum. After treatment with inducers or other agents, the cells were washed with phosphatebuffered saline (PBS) twice, then scraped. The cell suspension was centrifuged at 120 × g for 4 minutes. The PBS was aspirated and to the cell pellet were added 250 µl of 0.01M Tris-hydrochloric acid (HC1) pH 7.9 and 0.5% Triton X-100 and mixed thoroughly. The cell lysates were then centrifuged at 1500 × g for 5 minutes in an IEC Centra-7R centrifuge to bring down fractionated membranes. The supernatant was then analyzed for PA activity using the chromogenic assay.
bromododecane and centrifuged. Up to 150 µg of protein was loaded onto a 5% polyacrylamide tube gel and focused (pH 3 to 10) for 10,000 Vhr. The second dimension was by electrophoresis on an 8% sodium dodecyl sulfate-polyacrylamide gel for 600 Vhr. The gels were fixed, dried, and exposed to X-omat film for 12 to 72 hours, and developed.
Spectrozyme PL Assay
In the human epidermoid carcinoma cell line A431, u-PA is induced on the addition of either EGF16 or TPA.17 Figure 1 shows that addition of both EGF and TPA increase the u-PA content of cell lysates in a time-dependent manner. The figure also shows the
One to 25 µl of cell lysate incubated at 37°C for 1 hour with 0.05M Tris-HCl pH 7.9, 0.01% Triton X-100, 5 µg of plasminogen, and 0.1 µMO1 of Spectrozyme-PL (H-D-norleucyl-hexahydrotyrosyl-lysine-p-nitroanilide diacetate salt; American Diagnostica, Greenwich, CT) in a final volume of 1 ml. The reaction was terminated by the addition of 50 µ1 of 50% acetic acid and read at 405 nm in a Varian DMS-90 spectrophotometer and the absorbances were converted to international units by comparison with standard curves made with purified u-PA (Winkinase).
RESULTS Induction of u-PA by EGF and TPA
Protein Determination One to 20 µl of cell lysate was incubated at 37°C for 30 minutes with BCA reagent (Pierce, Rockford, IL), and the color read at 562 nm. Bovine serum albumin was used as standard.
Two-Dimensional Gel Electrophoresis Cytosolic proteins were resolved by the 2-dimensional get electrophoresis system described by O'Farrell,13 with minor modifications. Cells were grown to confluency and incubated in serum-free medium for 24 hours. After washing with Krebs-Ringer bicarbonate buffer (KRBB), 250 µl of KRBB containing 100 µCi of 32 P i , 14 and a 30-minute incubation, the inducers were added for an additional hour. The cells were lysed in a hypertonic fractionation medium15 containing 10 mM Tris-HCl pH 7.4, 10 mM ethylene diaminetetraacetic acid, EDTA, 5 mM ethylene glycol-bis(P-aminoethyl ether)-N,N'-tetraacetic acid 4 mg/ml digitonin, 50 mM sodium fluoride, 4 mM sodium vanadate, 1 mM leupeptin, 1 mM sodium potassium tartrate, 1 mM benzamidine, 1 mM iodoacetamide, and 200 mM sucrose. The sample was passed through a hydrocarbon cushion of
FIG. 1. Time-response curves of EGF and TPA in A431 cells. Cells were exposed to 50 nM EGF (●), or 100 nM TPA (o) for the indicated times. Cells were then lysed and u-PA activity determined on triplicate cultures. For the time course of u-PA mRNA levels, cells were exposed to TPA for the indicated times and RNA was isolated and subjected to Northern blot analysis. Levels of steady-state mRNA were expressed as multiples of the uninduced level, after normalizing to triosephosphate isomerase mRNA levels.
Downloaded by: University of British Columbia. Copyrighted material.
218
INDUCTION OF UROKINASE IN CANCER CELLS—KESSLER, MARKUS
movements. u-PA induction by TPA was not expected to be Ca+ +-dependent, since the promoter was shown not to have an effect on Ca + + mobilization.21 We find, however, that in the absence of extracellular Ca + + , there is a significant decrease in the extent of u-PA induction by TPA. This may be due to a decrease in the Ca + + level needed for the activation of PKC. We carried out intracellular Ca + + measurements using Indo-1 as the Ca+ + indicator and confirmed the findings of Moolenaar et al21 and Hesketh et al22 that EGF induced a signal only in the presence of extracellular Ca + + , and that TPA did not induce Ca+ + flux.
Tyrosine Phosphorylations in the Induction Pathways In order to evaluate the role of tyrosine phosphorylations in the induction of u-PA by these two inducers, we made use of the plant isoflavonoid, genistein, a specific inhibitor of kinases carrying out this function.23 Figure 2 shows that genistein did, as expected, block
FIG. 2. The effect of genistein (GNSTN) on u-PA induction by EGF and TPA in A431 cells. Cells were incubated with 50 nM EGF, or 100 nM TPA in the presence or absence of 40 fig/ml genistein for 4 hours, lysed, and u-PA determination performed on duplicate cultures.
Downloaded by: University of British Columbia. Copyrighted material.
steady-state levels of u-PA mRNA in the cells following induction by TPA. Transcription starts immediately, peaks at 4 hours, and declines rapidly thereafter. The functional u-PA follows a similar course, with an approximately 4-hour lag. In these studies EGF was used at 50 nM, and TPA at 100 nM, giving maximal induction at 4 hours. The effect of the inducers in the system studied here is expressed mainly by modulation of transcription. EGF and TPA gave a 7- and a 19-fold induction, respectively; the two together, a 27-fold induction of u-PA. Thus, the effects of the two inducers seem to be additive, suggesting independent induction pathways. The isoquinoline derivative, H7, a specific inhibitor of protein kinase C (PKC)18 causes an up to 77% inhibition of TPA-inducible u-PA expression, without affecting EGF inducibility. Furthermore, after a 72-hour exposure of the cells to TPA, a procedure known to down-regulate this enzyme,3 u-PA production becomes completely unresponsive to reintroduction of TPA, although full reactivity toward EGF is retained. Down-regulation was verified by dot blots of cell lysates derived from control and TPA-treated cells, developed with a monoclonal anti-PKC antibody. The fact that down-regulation of PKC does not impair EGF induction of u-PA is a significant result, since it was shown by Wahl et al19 in A431 cells that EGF stimulation leads to a very substantial activation of phospholipase C, via tyrosine phosphorylation, with an attendant production of diacylglycerol, the endogenous activator of PKC. Thus, our results establish that induction of u-PA by EGF does not involve the EGF-induced tyrosine phosphorylation of phospholipase C. Contrary to experience with other cell lines,3,20 8-Br-cAMP showed no u-PA-inducing effect in A431 cells, and neither did the phosphodiesterase inhibitor isobutyl methylxanthine (IBMX). Both of these compounds inhibit the EGF induction of u-PA. These inhibitions indicate that phosphorylations due to cAMPdependent protein kinases antagonize the EGF induction of u-PA. Interestingly, 8-Br-cAMP doubled the effect of TPA on induction. These results suggest that cAMPdependent protein kinase is not directly involved in u-PA induction in these cells, but is capable of modulating both the EGF and the TPA induction pathways. EGF was shown in A431 cells21 to cause a transient increase in intracellular Ca + + . This rise depended entirely on the presence of extracellular Ca + + . It was of interest to find out whether Ca + + is involved in the induction of u-PA by EGF. The effects of EGF were examined in the regular, 0.4 mM Ca + +-containing medium, and in a Ca + + -free medium. The results were essentially the same whether extracellular Ca + + was present or not. This allows the conclusion that u-PA induction by EGF is independent of EGF-induced Ca + +
219
SEMINARS IN THROMBOSIS AND HEMOSTASIS—VOLUME 17, NO. 3, 1991
u-PA induction by EGF, but it also blocked induction by TPA, suggesting that the pathway initiated by the phorbol ester also contained essential tyrosine phosphorylations. Indeed, Gilmore and Martin24 have shown that in A431 cells treatment with TPA resulted in the tyrosine phosphorylation of a 42 kd protein. This protein may be part of the signal transduction pathway for u-PA. The question arose whether the tyrosine phosphorylations in the two pathways involve the same or different proteins. In order to decide this question it was necessary first to establish the time it takes for EGF and TPA to induce the phosphorylations necessary for full u-PA induction. This was done by adding genistein to the cultures at different times following induction. Figure 3 shows that genistein, after 2 hours of induction by TPA, can no longer block this process. For EGF induction, the critical time was 3 hours. With this information, the experiment shown in Figure 4 was carried out. The first part of the figure shows that after allowing 3 hours for all
FIG. 3. The effect of genistein (GNSTN), added at varying times after EGF and TPA induction, on u-PA. Genistein was added at 0, 1, 2, 3, 3.9 hours following the addition of either 50 nM EGF (●) or 100 nM TPA (o). u-PA activity was deter mined on duplicate cultures and expressed as percent of controls for each inducer activity following 4 hours of incu bation without genistein. Total induction time was 4 hours, regardless of the time genistein was added.
essential phosphorylations to take place following the addition of EGF, TPA was still able to achieve full u-PA induction. This suggests that the TPA pathway either utilizes a different set of proteins for tyrosine phosphorylation, or else it makes use of proteins already phosphorylated in the EGF pathway. In the second part of the experiment shown in Figure 4, genistein was added together with TPA, following the 3 hour induction initiated with EGF. In this case TPA failed to elicit u-PA formation, indicating that TPA induction requires the tyrosine phosphorylation of one or more proteins that are different from those that had been phosphorylated by the preceding addition of EGF. In an attempt to visualize possible differences in the phosphorylation pattern associated with the two inducers, we carried out 2-dimensional electrophoresis on cell lysates derived from cultures that had been incubated with 32 P i . The label was added 30 minutes before the inducer, and induction was allowed to proceed for 1 hour before the cells were lysed. It should be pointed out that these autoradiographs show all the phosphorylated proteins, not only the tyrosine-phosphorylated ones. Since it is a known fact that tyrosine phosphorylations represent a rather small fraction of the total, the differences expected will be subtle. Figure 5 shows the peptide patterns from cells induced with EGF and TPA, with or without the addition of genistein. A comparison of panels A and C shows some inducer-specific differences (see also Table 1). The numbered spots show differences of
FIG. 4. The effect of genistein (GNSTN) on TPA induction after full EGF induction had been completed. A431 cells were incubated with 50 nM EGF for 3 hours, after which buffer, or 40 fig/ml genistein alone, or 100 nM TPA alone, or TPA and genistein together were added for an additional 2 hours. Cells were lysed and u-PA activity was determined on duplicate samples.
Downloaded by: University of British Columbia. Copyrighted material.
220
INDUCTION OF UROKINASE IN CANCER CELLS—KESSLER, MARKUS
221
TABLE 2. Genistein-Specific Changes in Phosphorylated Proteins in A431 Cells* Spot No.
Molecular Weight
PI
Genistein (G)
effect†
EGF 23
21,000
8.03
+G
EGF TPA EGF TPA TPA
* Cells were incubated with 3 2 P i for 30 minutes, then for an additional hour with 50 nM EGF or 100 nM TPA in the presence or absence of 40 µg/ml genistein. Cells were lysed and subjected to 2-dimensional gel electrophoresis as described in "Materials and Methods." Numbers indicate spots that have suffered a marked decrease in intensity due to genistein. Vertical comparisons (A and C) give inducer- specific differences; horizontal comparisons (A with B and C with D) give the location of tyrosine-phosphorylated peptides. †The entries indicate relative radioactivity in the spots.
likely to represent tyrosine phosphorylations. Not all of these spots, however, necessarily participate in the signal transduction of u-PA. The most important element in the TPA induction pathway is the activation of PKC. In order to identify phosphorylations attendant on PKC activation, we used the plant flavonoid quercetin, a compound known to inhibit PKC activity25 in addition to inhibiting tyrosine phosphorylations. Figure 6 shows identical areas of a series of autoradiographs of lysates of cells that had been treated with the inducers in the presence and absence of quercetin. The details in both panels A and B of Figure 6 show that many spots that appear when TPA is used as the inducer are weakened or eliminated in the presence of quercetin. These are the spots that are likely to represent phosphorylations attendant on the activation of PKC. Table 3 shows the molecular weights and isoelectric points of these proteins. As expected, the effects of quercetin on EGF-induced phosphorylations are considerably less pronounced.
DISCUSSION In this work we have presented evidence to indicate that EGF and TPA utilize different pathways for the induction of u-PA. We have also identified a number of peptides that are phosphorylated on induction by both of these agents. By the use of genistein, we were able to distinguish some of the peptides that were phosphorylated on tyrosine residues, and by the use of quercetin,
Downloaded by: University of British Columbia. Copyrighted material.
11
SEMINARS IN THROMBOSIS AND HEMOSTASIS—VOLUME 17, NO. 3, 1991
Downloaded by: University of British Columbia. Copyrighted material.
222
FIG. 6. Details of autoradiographs of 2-dimensional peptide patterns of A431 cell lysates. Cells were incubated with 32P, for 30 minutes, then with 50 nM EGFor 100 nM TPA, in the presence or absence of 40 µg/ml quercetin for 1 hour. Cells were lysed and subjected to 2-dimensional gel electrophoresis as described in "Materials and Methods." Numbers indicate spots that suffered a marked decrease in intensity due to quercetin.
we identified peptides phosphorylated via PKC. As mentioned, we do not know which of these peptides are in the pathways of u-PA induction. If obtainable in sufficient quantities, however, inhibitory antibodies could be generated against such peptides, and their effects on u-PA induction evaluated by microinjection into cells. One of the early EGF responses is the transcription of the fos gene, 26,27 the product of which complexes with the jun gene product and combines with the AP-1 site.28 The inductions of c-fos and of u-PA show important
similarities: (1) Both are induced by TPA and by EGF; (2) the EGF induction of both of these is independent of the EGF-induced influx of extracellular Ca + + ; 29 and (3) down-regulation of PKC removes TPA induction of c-fos and of u-PA, but does not affect the EGF induction of either of these.29 It seems possible therefore that the EGF induction of u-PA may depend on c-fos expression and, by extension, would involve the AP-1 target site. Activation of PKC by TPA may result in interaction with both the AP-1 and the AP-2 sites. Recent data indicate that the EGF receptor exists in
INDUCTION OF UROKINASE IN CANCER CELLS—KESSLER, MARKUS
Spot No.
Molecular Weight
PI
Quercetin (Q)
Effect†
EGF
effect by itself; however, it inhibits the EGF effect and doubles the TPA induction of u-PA; (4) after EGFinduced tyrosine phosphorylations are completed, addition of TPA can still induce u-PA, an effect that can be blocked by the addition of the tyrosine phosphorylation inhibitor, genistein, indicating that the two agents induce different tyrosine phosphorylation; (5) in 2-dimensional electrophoresis of 32P-labeled cell lysates, inducer-specific differences in the intensity of spots can be observed. Some of the spots weakened by genistein (tyrosine phosphorylations) are different, depending on the inducer used.
12
21,000
8.72
+Q
Epidermal Growth Factor and 12-Tetradecanoyl Phorbol 13-Acetate Induction of Urokinase in A431 Cells
The accumulated evidence suggests that plasminogen activators (PAs) do play a role in invasion and metastasis by cancer cells. 1 , 2 PAs, and in particular urokinase (u-PA), can be induced by a large variety of agents. Two agents, an epidermal growth factor (EGF) and a tumor promoter, 12-tetradecanoyl phorbol 13acetate (TPA), interact with specific receptors at the cell membrane, the EGF receptor and protein kinase C, respectively. In the study of Degen et al 3 TPA transiently increased the u-PA mRNA levels, following which the cells became unresponsive to TPA-inducible u-PA mRNA synthesis. This desensitization, however, by TPA did not impair the stimulation of u-PA mRNA synthesis by calcitonin, vasopressin, or 8-Br-cyclic adenosine monophosphate (cAMP), suggesting at least two separate induction pathways for u-PA. At the other end of the induction pathway, several regulatory elements in the 5'-flanking region of the u-PA gene have been identified. The regulatory sites so far identified are two cAMP-responsive elements (CRE), one reported by Nagamine and Reich 4 and the other by Nakagawa et al. 5 An AP-2 site, responsive to both cAMP and phorbol esters, was shown by Nagamine et al 6 and an AP-1 site by Nerlov et al. 7 The AP-1 site is known to interact with the fosljun heterodimer. Expression of fas and jun on EGF 8 and TPA stimulation was shown by Angel et al 9
and by Lee et al. 1 0 Thus, the initial (receptors) and terminal (regulatory elements) stages of the induction process have been investigated; however, little is known about the steps of signal transduction between these points for any of the u-PA inducers. The aim of the present study was to characterize, as far as possible, the basic mechanisms for signal transduction. Since information is already available on EGF- and TPA-inducing effects on u-PA, we chose to study the induction pathways utilized by these agents.
MATERIALS AND METHODS Cell Culture Human A431 epidermoid carcinoma cells were obtained from American Type Culture Collection and maintained in RPMI-1640 with 10% (vol/vol) fetal bovine serum (J.R. Scientific, Woodland, CA or Gibco Laboratories Grand Island, NY) in 5% carbon dioxide and 95% air at 37°C. Single-cell suspensions of A431 cells were counted with a hemocytometer after 10 minutes' incubation at 37°C with 0.05% trypsin (Gibco Laboratories). Cells were grown to more than 90% confluency.
RNA Isolation From the Department of Experimental Biology, Roswell Park Cancer Institute, New York State Department of Health, Buffalo, New York. This work was supported in part by the Mark Diamond Research Fund from the Graduate Student Association of SUNYAB (T.L.K.), and by grant 1 P01 CA28853 from the National Cancer Institute (G.M.). Reprint requests: Dr. Markus, Dept. of Exp. Biology, Roswell Park Cancer Institute, New York State Department of Health, Buffalo, NY 14263.
Total cellular RNA was isolated by the method described by Chirgwin et al;11 Northern blot analysis was described by Maniatis et al. 12
Sample Preparation for Chromogenic Assay For each experimental condition, cultures were analyzed in duplicate or quadruplicate, as indicated. The
Copyright © 1991 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
217
Downloaded by: University of British Columbia. Copyrighted material.
TIMOTHY L. KESSLER, Ph.D., and GABOR MARKUS, M.D., Ph.D.
SEMINARS IN THROMBOSIS AND HEMOSTASIS—VOLUME 17, NO. 3, 1991
cultures were washed with serum-free medium to wash out the fetal bovine serum. After treatment with inducers or other agents, the cells were washed with phosphatebuffered saline (PBS) twice, then scraped. The cell suspension was centrifuged at 120 × g for 4 minutes. The PBS was aspirated and to the cell pellet were added 250 µl of 0.01M Tris-hydrochloric acid (HC1) pH 7.9 and 0.5% Triton X-100 and mixed thoroughly. The cell lysates were then centrifuged at 1500 × g for 5 minutes in an IEC Centra-7R centrifuge to bring down fractionated membranes. The supernatant was then analyzed for PA activity using the chromogenic assay.
bromododecane and centrifuged. Up to 150 µg of protein was loaded onto a 5% polyacrylamide tube gel and focused (pH 3 to 10) for 10,000 Vhr. The second dimension was by electrophoresis on an 8% sodium dodecyl sulfate-polyacrylamide gel for 600 Vhr. The gels were fixed, dried, and exposed to X-omat film for 12 to 72 hours, and developed.
Spectrozyme PL Assay
In the human epidermoid carcinoma cell line A431, u-PA is induced on the addition of either EGF16 or TPA.17 Figure 1 shows that addition of both EGF and TPA increase the u-PA content of cell lysates in a time-dependent manner. The figure also shows the
One to 25 µl of cell lysate incubated at 37°C for 1 hour with 0.05M Tris-HCl pH 7.9, 0.01% Triton X-100, 5 µg of plasminogen, and 0.1 µMO1 of Spectrozyme-PL (H-D-norleucyl-hexahydrotyrosyl-lysine-p-nitroanilide diacetate salt; American Diagnostica, Greenwich, CT) in a final volume of 1 ml. The reaction was terminated by the addition of 50 µ1 of 50% acetic acid and read at 405 nm in a Varian DMS-90 spectrophotometer and the absorbances were converted to international units by comparison with standard curves made with purified u-PA (Winkinase).
RESULTS Induction of u-PA by EGF and TPA
Protein Determination One to 20 µl of cell lysate was incubated at 37°C for 30 minutes with BCA reagent (Pierce, Rockford, IL), and the color read at 562 nm. Bovine serum albumin was used as standard.
Two-Dimensional Gel Electrophoresis Cytosolic proteins were resolved by the 2-dimensional get electrophoresis system described by O'Farrell,13 with minor modifications. Cells were grown to confluency and incubated in serum-free medium for 24 hours. After washing with Krebs-Ringer bicarbonate buffer (KRBB), 250 µl of KRBB containing 100 µCi of 32 P i , 14 and a 30-minute incubation, the inducers were added for an additional hour. The cells were lysed in a hypertonic fractionation medium15 containing 10 mM Tris-HCl pH 7.4, 10 mM ethylene diaminetetraacetic acid, EDTA, 5 mM ethylene glycol-bis(P-aminoethyl ether)-N,N'-tetraacetic acid 4 mg/ml digitonin, 50 mM sodium fluoride, 4 mM sodium vanadate, 1 mM leupeptin, 1 mM sodium potassium tartrate, 1 mM benzamidine, 1 mM iodoacetamide, and 200 mM sucrose. The sample was passed through a hydrocarbon cushion of
FIG. 1. Time-response curves of EGF and TPA in A431 cells. Cells were exposed to 50 nM EGF (●), or 100 nM TPA (o) for the indicated times. Cells were then lysed and u-PA activity determined on triplicate cultures. For the time course of u-PA mRNA levels, cells were exposed to TPA for the indicated times and RNA was isolated and subjected to Northern blot analysis. Levels of steady-state mRNA were expressed as multiples of the uninduced level, after normalizing to triosephosphate isomerase mRNA levels.
Downloaded by: University of British Columbia. Copyrighted material.
218
INDUCTION OF UROKINASE IN CANCER CELLS—KESSLER, MARKUS
movements. u-PA induction by TPA was not expected to be Ca+ +-dependent, since the promoter was shown not to have an effect on Ca + + mobilization.21 We find, however, that in the absence of extracellular Ca + + , there is a significant decrease in the extent of u-PA induction by TPA. This may be due to a decrease in the Ca + + level needed for the activation of PKC. We carried out intracellular Ca + + measurements using Indo-1 as the Ca+ + indicator and confirmed the findings of Moolenaar et al21 and Hesketh et al22 that EGF induced a signal only in the presence of extracellular Ca + + , and that TPA did not induce Ca+ + flux.
Tyrosine Phosphorylations in the Induction Pathways In order to evaluate the role of tyrosine phosphorylations in the induction of u-PA by these two inducers, we made use of the plant isoflavonoid, genistein, a specific inhibitor of kinases carrying out this function.23 Figure 2 shows that genistein did, as expected, block
FIG. 2. The effect of genistein (GNSTN) on u-PA induction by EGF and TPA in A431 cells. Cells were incubated with 50 nM EGF, or 100 nM TPA in the presence or absence of 40 fig/ml genistein for 4 hours, lysed, and u-PA determination performed on duplicate cultures.
Downloaded by: University of British Columbia. Copyrighted material.
steady-state levels of u-PA mRNA in the cells following induction by TPA. Transcription starts immediately, peaks at 4 hours, and declines rapidly thereafter. The functional u-PA follows a similar course, with an approximately 4-hour lag. In these studies EGF was used at 50 nM, and TPA at 100 nM, giving maximal induction at 4 hours. The effect of the inducers in the system studied here is expressed mainly by modulation of transcription. EGF and TPA gave a 7- and a 19-fold induction, respectively; the two together, a 27-fold induction of u-PA. Thus, the effects of the two inducers seem to be additive, suggesting independent induction pathways. The isoquinoline derivative, H7, a specific inhibitor of protein kinase C (PKC)18 causes an up to 77% inhibition of TPA-inducible u-PA expression, without affecting EGF inducibility. Furthermore, after a 72-hour exposure of the cells to TPA, a procedure known to down-regulate this enzyme,3 u-PA production becomes completely unresponsive to reintroduction of TPA, although full reactivity toward EGF is retained. Down-regulation was verified by dot blots of cell lysates derived from control and TPA-treated cells, developed with a monoclonal anti-PKC antibody. The fact that down-regulation of PKC does not impair EGF induction of u-PA is a significant result, since it was shown by Wahl et al19 in A431 cells that EGF stimulation leads to a very substantial activation of phospholipase C, via tyrosine phosphorylation, with an attendant production of diacylglycerol, the endogenous activator of PKC. Thus, our results establish that induction of u-PA by EGF does not involve the EGF-induced tyrosine phosphorylation of phospholipase C. Contrary to experience with other cell lines,3,20 8-Br-cAMP showed no u-PA-inducing effect in A431 cells, and neither did the phosphodiesterase inhibitor isobutyl methylxanthine (IBMX). Both of these compounds inhibit the EGF induction of u-PA. These inhibitions indicate that phosphorylations due to cAMPdependent protein kinases antagonize the EGF induction of u-PA. Interestingly, 8-Br-cAMP doubled the effect of TPA on induction. These results suggest that cAMPdependent protein kinase is not directly involved in u-PA induction in these cells, but is capable of modulating both the EGF and the TPA induction pathways. EGF was shown in A431 cells21 to cause a transient increase in intracellular Ca + + . This rise depended entirely on the presence of extracellular Ca + + . It was of interest to find out whether Ca + + is involved in the induction of u-PA by EGF. The effects of EGF were examined in the regular, 0.4 mM Ca + +-containing medium, and in a Ca + + -free medium. The results were essentially the same whether extracellular Ca + + was present or not. This allows the conclusion that u-PA induction by EGF is independent of EGF-induced Ca + +
219
SEMINARS IN THROMBOSIS AND HEMOSTASIS—VOLUME 17, NO. 3, 1991
u-PA induction by EGF, but it also blocked induction by TPA, suggesting that the pathway initiated by the phorbol ester also contained essential tyrosine phosphorylations. Indeed, Gilmore and Martin24 have shown that in A431 cells treatment with TPA resulted in the tyrosine phosphorylation of a 42 kd protein. This protein may be part of the signal transduction pathway for u-PA. The question arose whether the tyrosine phosphorylations in the two pathways involve the same or different proteins. In order to decide this question it was necessary first to establish the time it takes for EGF and TPA to induce the phosphorylations necessary for full u-PA induction. This was done by adding genistein to the cultures at different times following induction. Figure 3 shows that genistein, after 2 hours of induction by TPA, can no longer block this process. For EGF induction, the critical time was 3 hours. With this information, the experiment shown in Figure 4 was carried out. The first part of the figure shows that after allowing 3 hours for all
FIG. 3. The effect of genistein (GNSTN), added at varying times after EGF and TPA induction, on u-PA. Genistein was added at 0, 1, 2, 3, 3.9 hours following the addition of either 50 nM EGF (●) or 100 nM TPA (o). u-PA activity was deter mined on duplicate cultures and expressed as percent of controls for each inducer activity following 4 hours of incu bation without genistein. Total induction time was 4 hours, regardless of the time genistein was added.
essential phosphorylations to take place following the addition of EGF, TPA was still able to achieve full u-PA induction. This suggests that the TPA pathway either utilizes a different set of proteins for tyrosine phosphorylation, or else it makes use of proteins already phosphorylated in the EGF pathway. In the second part of the experiment shown in Figure 4, genistein was added together with TPA, following the 3 hour induction initiated with EGF. In this case TPA failed to elicit u-PA formation, indicating that TPA induction requires the tyrosine phosphorylation of one or more proteins that are different from those that had been phosphorylated by the preceding addition of EGF. In an attempt to visualize possible differences in the phosphorylation pattern associated with the two inducers, we carried out 2-dimensional electrophoresis on cell lysates derived from cultures that had been incubated with 32 P i . The label was added 30 minutes before the inducer, and induction was allowed to proceed for 1 hour before the cells were lysed. It should be pointed out that these autoradiographs show all the phosphorylated proteins, not only the tyrosine-phosphorylated ones. Since it is a known fact that tyrosine phosphorylations represent a rather small fraction of the total, the differences expected will be subtle. Figure 5 shows the peptide patterns from cells induced with EGF and TPA, with or without the addition of genistein. A comparison of panels A and C shows some inducer-specific differences (see also Table 1). The numbered spots show differences of
FIG. 4. The effect of genistein (GNSTN) on TPA induction after full EGF induction had been completed. A431 cells were incubated with 50 nM EGF for 3 hours, after which buffer, or 40 fig/ml genistein alone, or 100 nM TPA alone, or TPA and genistein together were added for an additional 2 hours. Cells were lysed and u-PA activity was determined on duplicate samples.
Downloaded by: University of British Columbia. Copyrighted material.
220
INDUCTION OF UROKINASE IN CANCER CELLS—KESSLER, MARKUS
221
TABLE 2. Genistein-Specific Changes in Phosphorylated Proteins in A431 Cells* Spot No.
Molecular Weight
PI
Genistein (G)
effect†
EGF 23
21,000
8.03
+G
EGF TPA EGF TPA TPA
* Cells were incubated with 3 2 P i for 30 minutes, then for an additional hour with 50 nM EGF or 100 nM TPA in the presence or absence of 40 µg/ml genistein. Cells were lysed and subjected to 2-dimensional gel electrophoresis as described in "Materials and Methods." Numbers indicate spots that have suffered a marked decrease in intensity due to genistein. Vertical comparisons (A and C) give inducer- specific differences; horizontal comparisons (A with B and C with D) give the location of tyrosine-phosphorylated peptides. †The entries indicate relative radioactivity in the spots.
likely to represent tyrosine phosphorylations. Not all of these spots, however, necessarily participate in the signal transduction of u-PA. The most important element in the TPA induction pathway is the activation of PKC. In order to identify phosphorylations attendant on PKC activation, we used the plant flavonoid quercetin, a compound known to inhibit PKC activity25 in addition to inhibiting tyrosine phosphorylations. Figure 6 shows identical areas of a series of autoradiographs of lysates of cells that had been treated with the inducers in the presence and absence of quercetin. The details in both panels A and B of Figure 6 show that many spots that appear when TPA is used as the inducer are weakened or eliminated in the presence of quercetin. These are the spots that are likely to represent phosphorylations attendant on the activation of PKC. Table 3 shows the molecular weights and isoelectric points of these proteins. As expected, the effects of quercetin on EGF-induced phosphorylations are considerably less pronounced.
DISCUSSION In this work we have presented evidence to indicate that EGF and TPA utilize different pathways for the induction of u-PA. We have also identified a number of peptides that are phosphorylated on induction by both of these agents. By the use of genistein, we were able to distinguish some of the peptides that were phosphorylated on tyrosine residues, and by the use of quercetin,
Downloaded by: University of British Columbia. Copyrighted material.
11
SEMINARS IN THROMBOSIS AND HEMOSTASIS—VOLUME 17, NO. 3, 1991
Downloaded by: University of British Columbia. Copyrighted material.
222
FIG. 6. Details of autoradiographs of 2-dimensional peptide patterns of A431 cell lysates. Cells were incubated with 32P, for 30 minutes, then with 50 nM EGFor 100 nM TPA, in the presence or absence of 40 µg/ml quercetin for 1 hour. Cells were lysed and subjected to 2-dimensional gel electrophoresis as described in "Materials and Methods." Numbers indicate spots that suffered a marked decrease in intensity due to quercetin.
we identified peptides phosphorylated via PKC. As mentioned, we do not know which of these peptides are in the pathways of u-PA induction. If obtainable in sufficient quantities, however, inhibitory antibodies could be generated against such peptides, and their effects on u-PA induction evaluated by microinjection into cells. One of the early EGF responses is the transcription of the fos gene, 26,27 the product of which complexes with the jun gene product and combines with the AP-1 site.28 The inductions of c-fos and of u-PA show important
similarities: (1) Both are induced by TPA and by EGF; (2) the EGF induction of both of these is independent of the EGF-induced influx of extracellular Ca + + ; 29 and (3) down-regulation of PKC removes TPA induction of c-fos and of u-PA, but does not affect the EGF induction of either of these.29 It seems possible therefore that the EGF induction of u-PA may depend on c-fos expression and, by extension, would involve the AP-1 target site. Activation of PKC by TPA may result in interaction with both the AP-1 and the AP-2 sites. Recent data indicate that the EGF receptor exists in
INDUCTION OF UROKINASE IN CANCER CELLS—KESSLER, MARKUS
Spot No.
Molecular Weight
PI
Quercetin (Q)
Effect†
EGF
effect by itself; however, it inhibits the EGF effect and doubles the TPA induction of u-PA; (4) after EGFinduced tyrosine phosphorylations are completed, addition of TPA can still induce u-PA, an effect that can be blocked by the addition of the tyrosine phosphorylation inhibitor, genistein, indicating that the two agents induce different tyrosine phosphorylation; (5) in 2-dimensional electrophoresis of 32P-labeled cell lysates, inducer-specific differences in the intensity of spots can be observed. Some of the spots weakened by genistein (tyrosine phosphorylations) are different, depending on the inducer used.
12
21,000
8.72
+Q
