Phorbol Ester Activation of a Proteolytic Cascade Capable of Activating Latent Transforming Growth Factor-β

12-O-Tetradecanoylphorbol-13-acetate (TPA) suppresses the proliferation of the human breast epithelial cell line MCF10A-Neo by initiating proteolytic processes that activate latent transforming growth factor (TGF)-β in the serum used to supplement culture medium. Within 1 h of treatment, cultures accumulated an extracellular activity capable of cleaving a substrate for urokinase-type plasminogen activator (uPA) and tissue plasminogen activator (tPA). This activity was inhibited by plasminogen activator inhibitor-1 or antibodies to uPA but not tPA. Pro-uPA activation was preceded by dramatic changes in lysosome trafficking and the extracellular appearance of cathepsin B and β-hexosaminidase but not cathepsins D or L. Co-treatment of cultures with the cathepsin B inhibitors CA-074 or Z-FA-FMK suppressed the cytostatic effects of TPA and activation of pro-uPA. In the absence of TPA, exogenously added cathepsin B activated pro-uPA and suppressed MCF10A-Neo proliferation. The cytostatic effects of both TPA and cathepsin B were suppressed in cells cultured in medium depleted of plasminogen/plasmin or supplemented with neutralizing TGF-β antibody. Pretreatment with cycloheximide did not suppress the exocytosis of cathepsin B or the activation of pro-uPA. Hence, TPA activates signaling processes that trigger the exocytosis of a subpopulation of lysosomes/endosomes containing cathepsin B. Subsequently, extracellular cathepsin B initiates a proteolytic cascade involving uPA, plasminogen, and plasmin that activates serum-derived latent TGF-β.

TGF-␤s mediate their biological activities via interaction with high affinity, cell surface receptors (16). Several cell types synthesize and secrete TGF-␤s in an inactive latent form. Latency is a consequence of intracellular processing. Specifically, after translation TGF-␤1 polypeptides dimerize and subsequently undergo cleavage to yield amino-terminal latency-associated peptides and carboxyl-terminal peptides (11,12,17). Latency-associated peptides remain associated with the dimerized carboxyl-terminal peptides via electrostatic interactions, thus forming latent TGF-␤. Latency-associated peptides must be released from the latent complex before TGF-␤ can activate its receptor (11,12).
A variety of agents and treatments activate latent TGF-␤1. Heat, acidic pH, chaotropic agents, plasmin, substilisin-like endopeptidases, and the extracellular matrix protein thrombospondin promote the release of latency-associated peptides from latent TGF-␤1 in vitro (11,12,(17)(18)(19)(20). In vivo studies with thrombospondin-deficient mice support a role for thrombospondin in the physiological activation of latent TGF-␤1 (21). It has been inferred from analyses of plasminogen-deficient mice that plasminogen/plasmin do not play a major role in the in vivo processing of latent TGF-␤1 (22,23). Nevertheless, studies by Grainger et al. (24) suggest that these proteases may contribute to the in vivo activation of latent TGF-␤1 in some situations, and data continue to be published documenting plasmin-mediated TGF-␤ activation in cell lines (25,26).
The phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) is cytostatic to cells of various lineages (Ref. 27 and references therein). We recently reported that the spontaneously immortalized human breast epithelial cell line MCF10A-Neo arrests in G 1 following exposure to 10 nM TPA (27). This arrest developed quickly and was paralleled by the production of a cytostatic medium. A variety of approaches suggested that ϳ50% of the cytostatic activity of TPA was mediated by a TGF-␤ family member. Surprisingly, the serum used to supple-ment the culture medium, and not the MCF10A-Neo cells, was the source of the latent cytokine (27).
The current study was initiated to identify the processes activated by TPA in MCF10A-Neo cultures responsible for the activation of latent TGF-␤. Preliminary studies (27) demonstrated that the cytostatic effects of TPA could be inhibited by co-treatment with plasminogen activator inhibitor 1 (PAI-1). Consequently, we hypothesized that a proteolytic cascade involving urokinase type plasminogen activator (uPA) or tissue plasminogen activator (tPA) and plasmin was involved in this activation. This hypothesis proved to be correct. However, it involved an unexpected mechanism for the activation of pro-uPA. Specifically, TPA treatment of MCF10A-Neo cultures triggered the exocytosis of a subpopulation of lysosomes/endosomes, and released lysosomal/endosomal cathepsin B catalyzed the activation of pro-uPA.
Cell Lines and Culturing Conditions-The MCF10A-Neo cell line was obtained from the Cell Lines Resource (Karmanos Cancer Institute, Detroit, MI). It was derived by transfection of the spontaneously immortalized, nontumorigenic human breast epithelial MCF10A cell line with plasmid Homer 6 (pHo6) and subsequent selection for G418 resistance. The MCF10A-Neo line consists of the pooled survivors. Detailed characterizations and descriptions of the derivations of the MCF10A and MCF10A-Neo cell lines have been published (28,29). The morphological and growth properties of the two cell lines are very similar (29). MCF10A cultures, like MCF10A-Neo cultures, exocytose cathepsin B following exposure to TPA. 2 MCF10A-Neo cells were cultured in supplemented Dulbecco's modified Eagle's medium/Ham's F-12 medium as previously described (30). The supplements consisted of human insulin (10 g/ml), epidermal growth factor (10 ng/ml), cholera toxin (100 ng/ml), hydrocortisone (0.5 g/ml), 100 units/ml penicillin G, 100 g/ml streptomycin sulfate, and 5% horse serum.
Culture Treatment-Proliferation studies were performed with cultures plated at a density that ensured exponential growth for a minimum of 4 days. The cultures were routinely treated 40 -48 h after plating. The details of treatment are provided in the text. TPA, cycloheximide, E-64, CA-074, Z-GGR-AMC, Z-RR-AMC, and Z-FA-FMK were all dissolved and diluted in Me 2 SO. Organic solvent never exceeded 0.2% of total culture/assay volume. Aprotinin and PAI-1 were diluted in sterile water. Purified human cathepsin B was diluted in a solution containing 1 mM EDTA and 20 mM sodium acetate, pH 5.0. The various antibodies used in this study were reconstituted/diluted according to the manufacturer's instructions. For estimation of cell numbers, the cultures were harvested by exposure to a solution of 0.25% trypsin, 0.1 mM EDTA. Viability was assessed by determining the ability to exclude trypan blue.
Removal of Latent TGF-␤ and Plasmin/Plasminogen from Culture Medium-Culture medium supplemented with 5% horse serum was incubated with 12 g/ml of a rabbit neutralizing polyclonal TGF-␤ pan antibody or 12 g/ml of control rabbit IgG for 2 h at 37°C. The medium was then chromatographed on a column of protein G-agarose that had been washed and equilibrated in 10 mM sodium phosphate, pH 7.0, 0.15 M NaCl. The eluant was used as a source of TGF-␤-depleted medium. Medium supplemented with horse serum was also chromatographed on a column of L-lysine-agarose that had been washed with 0.1 M NaCl, 1 mM EDTA, and 50 mM Tris-HCl, pH 7.5. The eluant was used as a source of plasminogen/plasmin-depleted medium.
[ 3 H]Leucine Incorporation Studies-Cells grown in 35-mm culture dishes were treated with varied concentrations of cycloheximide for 30 min. The cultures were then washed twice with PBS prior to being pulsed with L- [3,4, H(N)]leucine (1 Ci/ml of culture medium) in the presence of fresh cycloheximide. The cultures were harvested at varied times thereafter for analyses of [ 3 H]leucine incorporation into protein.
The procedure used for the processing of labeled cells has been described in detail (31). Radioactivity was detected by scintillation counting and expressed as dpm/10 6 cells.
Fluorometric Assays for tPA and uPA-The cells were plated in either 35-mm culture dishes or 96-well culture plates. Culture medium was also added to a parallel set of wells/dishes that lacked cells. After 1.5-2 days in culture, the cells were treated with various antibodies or protease inhibitors 1 h prior to the addition of either Me 2 SO or TPA (10 nM). Culture medium (200 l) was periodically removed thereafter and transferred to 96-well plates appropriate for fluorescent assay measurements. Analyses of uPA/tPA were initiated by the addition of 100 M Z-GGR-AMC. The release of AMC was monitored for 12-15 min on a SPECTRAmax Gemini dual scanning microplate spectrofluorometer using an excitation wavelength of 380 nm and an emission wavelength of 440 nm. Parallel analyses were performed on culture medium taken from dishes/wells lacking cells. Changes in fluorescence over a set time were determined and converted to fmol of AMC by comparison to a standard curve. Cell-derived activity was calculated as the difference in activities of media derived from dishes/wells having and lacking cells. For the calculation of specific activities, cells were released from dishes/ wells by treatment with 0.25% trypsin, 0.1 mM EDTA, and counted. Z-GGR-AMC cleavage activity is reported as fmol of AMC produced per min per 10 3 cells.
Fluorometric Assays for Cathepsins B, L, and D-The protocol described by Linebaugh et al. (32) was used for the assay of extracellular and intracellular cathepsin B. The cultures were treated as described above and washed twice with PBS at varied times after Me 2 SO or TPA addition before being covered with pericellular assay buffer (PAB), which consisted of Hanks' balanced salt solution lacking sodium bicarbonate but containing 0.6 mM CaCl 2 , 0.6 mM MgCl 2 , 2 mM L-cysteine, and 25 mM PIPES, pH 7.0. After a 10-min incubation at 37°C, 200 l of the PAB solution was transferred to a 96-well plate. The assay was initiated by the addition of 100 M Z-RR-AMC to each well in the absence or presence of 5-20 M CA-074 and followed for 15 min on a fluorescence plate reader using an excitation wavelength of 380 nM and an emission wavelength of 440 nm. To measure intracellular cathepsin B activity, the cultures were washed twice with PBS and then lysed with PAB supplemented with 0.1% Triton X-100. The lysates were assayed as described above. Changes in fluorescence over time were determined and converted to fmol of AMC by comparison with a standard curve. Activity inhibited by CA-074 was considered to represent cathepsin B. With Z-RR-AMC as a substrate, this represented ϳ90 -95% of the total activity measured in Triton X-100 lysates. Duplicate cultures not used for the assay of cathepsin B were treated with trypsin/ EDTA and counted. Cathepsin B specific activities are reported as fmol of AMC produced per min per 10 3 cells.
Extracts prepared for cathepsin B analyses were also used for the assay of cathepsins L and D. The assay for cathepsin L was similar to that used for cathepsin B except for the substitution of Z-FR-AMC (100 M) as substrate. Cleavage activity not inhibited by 10 M CA-074 was considered to be cathepsin L-like activity. Cathepsin L specific activities are reported as fmol of AMC produced per min per 10 3 cells.
Cathepsin D was assayed by a published procedure (33) that monitors the cleavage of Ac-GE(Edans)-Z-EVNLDAEF-Z-K(Dabcyl)-G-NH 2 , a highly selective substrate for cathepsin D (33,34). This peptide corresponds to a mutant of the amyloid B protein precursor having an Asn-Leu substitution at amino acids 670 -671 and contains Dabcyl and Edans groups, which act as internal quencher and fluorophore, respectively. Release of the Edans fluorophore was followed on a Shimadzu RF-540 spectrofluorophotometer using an excitation wavelength of 340 nM and an emission wavelength of 490 nm. The assays contained 40 mM sodium formate, pH 3.5, and 100 M peptide substrate (dissolved in Me 2 SO). The reactions were initiated by the addition of extract. Activity inhibited by the addition of 10 M pepstatin A represented cathepsin D. Changes in fluorescence over time were determined and converted to 2 M. Guo and J. J. Reiners, Jr., unpublished observations. fmol of Edans by comparison with a standard curve made with Ac-RE(Edans)-A-NH 2 . Cathepsin D specific activities are reported as fmol of Edans fluorophore released per min per 10 3 cells.
N-Acetyl-␤-D-Glucosaminidase (␤-Hexosaminidase) Assay-The assay described by Storrie and Madden (35) was modified slightly and used to assay ␤-hexosaminidase. The cells were plated in 60-mm culture dishes. After 2 days the cultures were washed, refed with 1.5 ml of growth medium, and treated with Me 2 SO or TPA (10 nM). Samples of 200 l were periodically removed and incubated with 200 l of assay mixture (0.1 mM sodium acetate, pH 4.0, 1 mM 4-methyl-umbelliferyl-N-acetyl-␤-glucosaminide). The reaction was terminated by the addition of 0.5 ml of 0.5 M glycine, 0.5 M Na 2 CO 3 . Samples (200 l) of the assay mixture were transferred to a 96-well plate and analyzed with a dual scanning microplate spectrofluorometer using an excitation wavelength of 364 nm and an emission wavelength of 448 nm. The above assay represents conditions used to monitor the extracellular accumulation of ␤-hexosaminidase over time following the addition of TPA. To determine the rate of ␤-hexosaminidase secretion, 2-day-old cultures were treated with TPA or Me 2 SO and at varied times thereafter washed and refed with 1 ml of fresh medium. After 15 min the culture medium was removed for assay as described above. Total cellular ␤-hexosaminidase activity was assayed as above using lysates prepared by exposing cultures to 1 ml of 1% Nonidet P-40 in PBS for 2 min. Additional plates were exposed to a solution of 0.25% trypsin, 0.1 mM EDTA and counted. ␤-Hexosaminidase activity is reported as either relative fluorescent units/10 3 cells or relative fluorescent units/10 3 cells/min. Untreated/ unconditioned medium (with or without Nonidet P-40 depending on the protocol) was used as a control and subtracted from each sample.
Cathepsin B Localization by Confocal Microscopy-MCF10A-Neo cells were plated on poly-L-lysine-coated coverslips. Two days later they were treated with either Me 2 SO or 10 nM TPA. At various times after treatment coverslips were washed, fixed, and processed for indirect immunolocalization of cathepsin B by confocal microscopy. The procedure used has been described in detail (36). The primary antibody was a rabbit polyclonal antibody that recognizes human cathepsin B and procathepsin B (37). The secondary antibody was goat anti-rabbit Alexa 488 (A 11008, Molecular Probes). The images were captured using a Zeiss 310 confocal microscope and a 63ϫ/1.4 oil immersion lens, PH3 phase condenser, and a 488-nm laser light source with a pinhole setting of 17. All of the images were captured at the same contrast and brightness settings.
Statistical Analyses-All enzymatic assays involved analyses of a minimum of four culture wells or three culture dishes/treatment/experiment. Cell counts were performed on three or four culture wells or culture dishes. The data were analyzed by the Tukey HSD test. The Statistica 5.0 software package (StaSoft Inc., Tulsa, OK) was used to perform these calculations.

RESULTS
TPA Activation of Z-GGR-AMC Cleavage Activity-The tripeptide Z-GGR-AMC is a substrate for a limited number of serine proteases including uPA and tPA (38 -40). MCF10A-Neo cultures constitutively expressed a non-cell-associated, extracellular activity capable of cleaving Z-GGR-AMC (Fig. 1A). This activity was not affected by exposure to Me 2 SO but was elevated after exposure to TPA (Fig. 1A). Maximal increases were noted 1-2 h after TPA treatment. Thereafter, cleavage activity declined and returned to control levels within 8 h of TPA treatment.
The assay employed in Fig. 1A was not designed to measure extracellular, cell-associated Z-GGR-AMC cleavage activity. To measure such activity, cleavage assays were performed directly in the culture wells (cell-associated ϩ non-cell-associated activities) and compared with the activities in medium assayed after removal from the culture wells (non-cell-associated activity). No additional activity was detected when the assays were performed in the presence of cells (Fig. 1B). Hence, the TPAinduced extracellular Z-GGR-AMC cleavage activity was not cell-associated.
Exposure of cultures to the serine protease inhibitor aprotinin prior to Me 2 SO treatment strongly inhibited extracellular Z-GGR-AMC cleavage activity ( Fig. 2A). Aprotinin co-treatment also suppressed TPA-induced extracellular cleavage activity. Indeed, the extracellular Z-GGR-AMC cleavage activity of cultures co-treated with TPA and aprotinin was similar to the activity measured in cultures co-treated with Me 2 SO and aprotinin ( Fig. 2A).
Treatment of cultures with the tPA/uPA inhibitor PAI-1 had no effect on basal extracellular Z-GGR-AMC cleavage activity (Fig. 2B). However, co-treatment of TPA cultures with PAI-1 completely suppressed the TPA-dependent induction of extracellular Z-GGR-AMC cleavage activity (Fig. 2B).
Neutralizing antibodies to tPA and uPA were used to determine whether either protease was responsible for extracellular Z-GGR-AMC hydrolysis (Fig. 3). Antibodies to tPA affected neither basal nor TPA-induced Z-GGR-AMC cleavage activity. Neutralizing uPA antibody had no effect on basal Z-GGR-AMC cleavage activity. However, the uPA antibody inhibited completely the extracellular Z-GGR-AMC cleavage activity induced by TPA (Fig. 3). Hence, the extracellular Z-GGR-AMC cleavage activity induced by TPA appeared to be uPA.
Co-treatment of cultures with uPA antibody suppressed the cytostatic effects of TPA by ϳ50% (Table I). In contrast, uPA antibody had no effect on the growth of solvent-treated cultures. Control IgG had no effect on the growth of either solventor TPA-treated cultures ( Table I). The magnitude of protection (ϳ50%) afforded by co-treatment with uPA antibody is significant because TGF-␤ was shown previously to mediate ϳ50% of the cytostatic effects of TPA on MCF10A-Neo proliferation (27).
TPA-stimulated Exocytosis of Lysosomal/Endosomal Cathepsins B, L, and D and ␤-Hexosaminidase-Cathepsin B is a cysteine protease and can activate pro-uPA in vitro (41,42). Z-RR-AMC is commonly used as a substrate for cathepsin B (32,43). Two-day old MCF10A-Neo cultures constitutively produced an extracellular Z-RR-AMC cleavage activity (Fig. 4A). This activity was neither enhanced by exposure to Me 2 SO nor inhibited by CA-074, a highly selective irreversible inhibitor of cathepsin B (44). In vitro analyses indicated that Ն5 M CA-074 inhibited completely the activity of purified cathepsin B in our assay system, whereas concentrations as high as 20 M did not affect the activities of the cysteine proteases cathepsins C and L or the aspartate protease cathepsin D. 2 Exposure of cultures to TPA enhanced the rate of secretion and accumulation of an extracellular Z-RR-AMC cleavage activity, which could be inhibited by CA-074 (Fig. 4A). This TPAderived cathepsin B activity was detectable within 1 h of treatment and peaked 1 h later. Within 4 h of TPA treatment, pericellular Z-RR-AMC cleavage activity had returned to basal levels (Fig. 4A).
The appearance of extracellular cathepsin B was paralleled by the loss of intracellular cathepsin B activity (Fig. 4B). Decreases in intracellular cathepsin B activity were observed within 1 h of treatment. Maximum decreases occurred 2-4 h post-treatment. Intracellular cathepsin B activities recovered to pretreatment levels within 24 h of exposure to TPA (Fig. 4B).
Phorbol ester-induced changes in intracellular and extracellular cathepsin B activities were paralleled by alterations in the intracellular location and trafficking of cathepsin B. In solvent-treated cultures cathepsin B exhibited punctuate, perinuclear staining (Fig. 5, A-C). Cathepsin B staining remained perinuclear after 0.5 h of TPA treatment but had become more polarized (Fig. 5D). Within 1 h of treatment cathepsin B appeared to stream toward the periphery of some cells (Fig. 5E). Staining patterns similar to those depicted in Fig. 5E were also seen after 2 h of TPA treatment. 2 By 4 h the streaming pattern had disappeared (Fig. 5F). However, cathepsin B staining remained polarized (Fig. 5F).
Co-localization studies suggest that the lysosomes/endosomes of a single cell may contain different types of cathepsins (45)(46)(47). Solvent-treated MCF10A-Neo cultures contained de-   tectable intracellular cathepsin L and D activities but no measurable extracellular activities (Fig. 4, C and D). Exposure to TPA did not promote the extracellular release of either cathepsin L or D (Fig. 4, C and D).
␤-Hexosaminidase has been used as a lysosomal marker in studies of calcium-triggered lysosome exocytosis in fibroblasts and epithelial cells (48,49). Exposure of MCF10A-Neo cultures to Me 2 SO had no effect on extracellular ␤-hexosaminidase activity (Fig. 6A). However, dramatic accumulations of extracellular ␤-hexosaminidase occurred within 1 h of TPA treatment (Fig. 6A). Activity peaked 1-2 h post-treatment and declined thereafter. The accumulation of extracellular ␤-hexosaminidase activity was paralleled by both the loss of intracellular ␤-hexosaminidase activity (Fig. 6B) and an increased rate of ␤-hexosaminidase release into the medium (Fig. 6C).
Pro-uPA Activation by Extracellular Cathepsin B-Co-treatment of MCF10A-Neo cultures with the broad spectrum cysteine protease inhibitor E-64 (50) or the cathepsin B selective inhibitor CA-074 (44) completely inhibited the activation of pro-uPA by TPA (Fig. 7). Co-treatment of MCF10A-Neo cultures with CA-074 also suppressed partially (ϳ53%) the cytostatic activity of TPA (Table I).
To confirm the role of extracellular cathepsin B in the activation of pro-uPA, we pretreated cultures with Z-FA-FMK, a cell-permeable inhibitor of lysosomal cathepsins B and L (51,52). Intracellular cathepsin B activity was inhibited Ͼ95% by 1 M Z-FA-FMK (Fig. 8A). This concentration of Z-FA-FMK was neither cytostatic nor cytotoxic to MCF10A-Neo cultures. 2 However, it completely inhibited the appearance of extracellular cathepsin B activity, 2 and the activation of pro-uPA follow-ing TPA treatment (Fig. 8B). It should be noted that preactivated uPA was not inhibited by Z-FA-FMK. 2 If released lysosomal/endosomal cathepsin B was responsible for the activation of extracellular pro-uPA in TPA-treated cultures, one would anticipate that direct addition of cathepsin B  to culture medium should mimic the effects of TPA. Fig. 9 shows that this was the case. Exogenously added cathepsin B was as effective as TPA in activating pro-uPA. Furthermore, treatment of cultures with a combination of TPA and cathepsin B yielded uPA activities comparable with those measured following treatment with only TPA or cathepsin B (Fig. 9).
MCF10A-Neo proliferation was partially suppressed by the addition of cathepsin B to culture medium (Fig. 10). The magnitude of this suppression equalled ϳ50% of the suppressive effect of TPA. Approximately 50% of the cytostatic activity of TPA and all of the cytostatic activity of exogenously added cathepsin B could be suppressed by co-treatment with a neutralizing TGF-␤ pan antibody (Fig. 10). In contrast, co-treatment with control IgG had no effect on the cytostatic activities of either TPA or cathepsin B (Fig. 10). Hence, the cytostatic effects of cathepsin B in our model appear to be mediated exclusively by TGF-␤.
Effects of Cycloheximide on Pro-uPA Activation-TPA-mediated exocytosis of lysosomal/endosomal cathepsin B and the activation of pro-uPA occurred within 1 h of phorbol ester exposure. Although this process occurred quickly, it was conceivable that the observed effects of TPA required de novo protein synthesis. Indeed, TPA increases uPA expression in a variety of models (53,54). To address this issue we determined concentrations of cycloheximide sufficient to inhibit protein synthesis in our system. A dose of 5 g/ml of cycloheximide completely inhibited protein synthesis in MCF10A-Neo cultures. 2 Co-treatment of cultures with TPA and 5 g/ml of cycloheximide inhibited neither the extracellular release of cathepsin B nor the activation of pro-uPA. 2 Role of Plasminogen/Plasmin in Mediating the Cytostatic Activity of TPA-uPA converts plasminogen to plasmin, and plasmin can activate latent TGF-␤ (17). Because our studies demonstrated that the cytostatic effects of TPA were in part mediated by TGF-␤ and required pro-uPA activation, we reasoned that plasminogen was involved in the proteolytic cascade leading to TGF-␤ activation. To test this possibility we passed medium over L-lysine-agarose columns to remove plasminogen/ plasmin (55,56). The cytostatic effects of TPA were reduced ϳ50% in cultures maintained in plasminogen/plasmin-deficient medium (Fig. 11A). It should be noted that chromatography over L-lysine-agarose columns did not remove pro-uPA. uPA activities in chromatographed and nonchromatographed medium were identical following in vitro activation with cathepsin B (Fig. 11B). DISCUSSION We recently reported that low nanomolar concentrations of TPA suppress the proliferation of MCF10A-Neo cells via a mechanism involving TGF-␤ (27). Surprisingly, the latent form of the cytokine was supplied by the serum used to supplement the culture medium and not the cells (27). Unknown was the mechanism responsible for the activation of the latent cytokine. The current investigation demonstrates that TPA triggers the exocytosis of lysosomes/endosomes and the release of active cathepsin B into the culture medium. Released cathepsin B subsequently initiates the pro-uPA/uPA/plasminogen/plasmin proteolytic cascade, which ultimately activates latent TGF-␤. Although the activation of latent TGF-␤ by plasmin is well documented (17,18), this is the first report to demonstrate FIG. 9. Addition of cathepsin B to culture medium activates pro-uPA. Two-day-old MCF10A-Neo cultures in exponential growth were exposed to Me 2 SO or 10 nM TPA for 1. cathepsin B-mediated activations of the uPA/plasminogen/ plasmin pathway and latent TGF-␤, in a cell culture system.
In addition to the current investigation, two other studies implicate a role for cathepsin B in the activation of latent TGF-␤. Specifically, cultured normal human osteoblast-like cells constitutively synthesize and secrete latent TGF-␤ (57). Treatment of human osteoblast-like cells with dexamethasone stimulates the production of a conditioned medium containing both active TGF-␤ and cathepsin B. Inhibitor studies suggest that cathepsins, in conjunction with additional unidentified extracellular proteases, were responsible for the activation of latent TGF-␤. In contrast to this latter study and our own, in vitro studies suggest that cathepsin B can directly activate recombinant and human serum platelet-derived latent TGF-␤1 (58). Under no circumstance was extracellular cathepsin B in the absence of pro-uPA/plasminogen able to activate serumsupplied latent TGF-␤ in our culture model. Although speculative, the differential effects of cathepsin B in the various studies may reflect differences in the relative amount of cathepsin B available for proteolysis of latent TGF-␤.
Cathepsin D, in addition to cathepsin B, has been implicated in the activation of TGF-␤. Specifically, addition of cathepsin D to a conditioned medium produced by NRK-49F fibroblasts causes the conversion of latent TGF-␤ to the 25-kDa processed peptide (59). It is not known whether this cleavage is mediated solely by cathepsin D or by cathepsin D in combination with another protease. Nevertheless, although MCF10A-Neo cells contain significant amounts of cathepsin D, we detected no extracellular cathepsin D activity following exposure to TPA. Furthermore, pro-uPA activation by TPA was not inhibited by co-incubation with pepstatin A, an inhibitor of cathepsin D. 2 Hence, the pro-uPA and latent TGF-␤ activations observed in the current study were not mediated by cathepsin D.
In vitro studies have shown that cathepsin B can convert both soluble and uPA receptor-bound pro-uPA to uPA (41,42). However, this activity is rarely mentioned in reviews of cathepsin B activities or processes involved in the activation of uPA. Such omissions may reflect the relative absence of data implicating an in vivo role for cathepsin B in uPA activation. Other than the current investigation, we know of one other study showing cathepsin B activation of pro-uPA in vivo. Specifically, ovarian cancer HOC-1 cells constitutively express pro-uPA/ uPA and cathepsin B on their cell surfaces (60). Inhibition of cathepsin B by inclusion of the protease inhibitor E-64 suppressed the conversion of surface-associated pro-uPA to surface-associated two-chain active uPA (60).
A variety of agents and treatments trigger the exocytosis of lysosomes in a cell type-dependent manner. Lysosome exocytosis occurs in alveolar macrophages following exposure to zymosan particles, agents that inhibit vacuolar type (H ϩ )-ATPase or agents that elevate lysosomal pH (61)(62)(63). The arachidonate metabolite 12-S-hydroxy-eicosatetraenoic acid (12-S-HETE) induces lysosome exocytosis in a variety of cultured cell types (43,64). Finally, conditions that rapidly elevate cytosolic Ca 2ϩ concentrations also induce lysosome exocytosis in cells of various lineages (48). Recent studies have also identified agents that will potentiate but not initiate lysosome exocytosis. For example, cAMP and agents that increase cAMP levels potentiate Ca 2ϩ -triggered lysosome exocytosis (49). Similarly, TPA has been shown to potentiate but not trigger lysosome exocytosis in macrophages (62). In the current study TPA by itself induced the exocytosis of a lysosomal/endosomal subpopulation containing cathepsin B and ␤-hexosaminidase but not cathespins L and D. Although TPA can induce lysosome exocytosis in platelets (65,66) in the absence of additional stimuli, this is the first report we know of to document TPA-mediated lysosome exocytosis in an epithelial cell line.
The mechanism by which TPA triggers lysosome exocytosis in MCF10A-Neo cultures is not known. TPA is a well characterized activator of "conventional" and "novel" protein kinase C (PKC) (67,68). However, it also binds to several non-PKC proteins that may have signaling functions (67). We have two pieces of data suggesting a role for PKCs in lysosome exocytosis. First, pretreatment with TPA to down-regulate PKC contents is commonly used to implicate a role for PKCs in biological phenomena (63,69,70). We have found that a 24-h pretreatment of MCF10A-Neo cultures with TPA selectively reduces the contents of some PKC isoforms (71) and suppresses the triggering of cathepsin B release by subsequent TPA treatment. 2 These effects correlate with the cultures becoming refractory to the cytostatic effects of a second TPA application. 2 Second, we have found that co-treatment with low nanomolar concentrations of the PKC inhibitor Ro-32-0432 will suppress completely the TPA-triggered release of cathepsin B. 2 We are currently attempting to identify the PKC species responsible for triggering lysosome exocytosis in our model.
Co-localization analyses of cells stained with antibodies to cathepsins B, L, and D suggest lysosome heterogeneity (45)(46)(47)72). Such observations raise the question of whether all lysosomes or only subpopulations of lysosomes are earmarked for exocytosis following receipt of an exocytotic signal. Germane to this issue is a report demonstrating that 12-S-HETE induces the exocytosis of lysosomes containing FIG. 11. Role of plasminogen/plasmin in mediating the cytostatic effects of TPA. The conditioned media of 2-day-old MCF10A-Neo cultures were depleted of plasminogen/plasmin by chromatography on lysine agarose columns. A, two-day-old MCF10A-Neo cultures were washed with PBS and refed with either conditioned medium or depleted conditioned medium immediately before the addition of either Me 2 SO or 10 nM TPA. The cultures were harvested 27 h later for determination of cell numbers. The data were normalized to the number of cells in cultures treated with Me 2 SO and grown in conditioned medium and represent the means Ϯ S.D. for a minimum of three plates/treatment group. B, conditioned medium and conditioned depleted medium in the presence or absence of 5 nM cathepsin B (CB) were analyzed for Z-GGR-AMC cleavage activity. The data represent the means Ϯ S.D. of triplicate analyses performed on media used in studies depicted in A. *, significantly less than all other treatments (p Ͻ 0.001). **, significantly greater than corresponding controls (p Ͻ 0.001). cathepsin B, but not L, from alveolar macrophages, Wi-38 fibroblasts, and a series of human lung tumor cell lines (64). Conditions capable of triggering selective exocytosis of cathepsin L were not identified. Very recent studies have colocalized cathepsin L to a lysosome population that appears to undergo exocytosis following changes in cytosolic Ca 2ϩ concentration (73). It is not known whether this latter lysosome population contains exclusively cathepsin L or contains cathepsin L and some B. The current investigation demonstrates that TPA induces the exocytosis of a lysosome population that contains cathepsin B but not cathepsin D or L. The similarity in the actions of TPA and 12-S-HETE are more than just coincidental. The effects of 12-S-HETE in cells of the MCF10A lineage have been attributed to its activation of PKCs (43,74,75). Collectively, these findings emphasize the heterogeneous nature of lysosomes/endosomes and demonstrate that subpopulations of these organelles differ in their susceptibility to signals that trigger exocytosis. These studies also suggest that there may be multiple sensors and mechanisms involved in lysosome exocytosis.
Cathepsin L, like cathepsin B, can activate pro-uPA (40,42). Two lines of research suggest that cathepsin L was not responsible for the pro-uPA activation noted in the current study. First, unlike cathepsin B, we were unable to detect any extracellular cathepsin L-like activity in TPA-treated cultures. Second, co-treatment of cultures with CA-074 completely suppressed the activation of pro-uPA by TPA. CA-074 inhibits cathepsin B but not cathepsin L (76). Nevertheless, conditions have been defined that lead to the exocytosis of cathepsin L in some cell types (73), and it is anticipated that activation of pro-uPA would occur in such systems, providing it was present.
A variety of tumor types constitutively express elevated levels of extracellular cell associated/non-cell-associated cathepsin B, uPA, and matrix metalloprotease (MMP) activities (77)(78)(79)(80). Individually and/or collectively, these proteases contribute to processes involved in invasion, metastasis, inflammation, tissue and vascular remodeling, fibrogenesis, etc. Recent studies have demonstrated or implicated plasmin in the activation of pro-MMP-1 (collagenase I), pro-MMP-3 (stromelysin), pro-MMP-9/gelatinase B/collagenase IV, and pro-membrane-type 1 MMP (80 -84). These latter findings, coupled with the present study, provide a plausible explanation for how MMPs may be activated in cells expressing elevated activities of extracellular cathepsin B and uPA.