Oestradiol regulation of the components of the plasminogen-plasmin system in MDA-MB-231 human breast cancer cells stably expressing the oestrogen receptor.

To understand the hormonal regulation of the components of the plasminogen-plasmin system in human breast cancer, we examined the oestradiol (E2) regulation of plasminogen activators (PAs), namely urokinase-type plasminogen activator (uPA) and tissue-type plasminogen activator (tPA), plasminogen activator inhibitor type 1 (PAI-1) and uPA receptor (uPAR), in our model system. We used stable transfectants of the MDA-MB-231 human breast cancer cells that express either the wild-type (S30 cells) or the mutant 351asp-->tyr oestrogen receptor (ER) (BC-2 cells). Northern blot analysis showed that there was a concentration-dependent down-regulation of uPA, tPA and PAI-1 mRNAs by E2. In contrast, uPAR mRNA was not modulated by E2. The pure anti-oestrogen ICI 182,780 was able to block E2 action, indicating that the regulation of these genes is ER mediated. The E2 also inhibited the expression and secretion of uPA, tPA and PAI-1 proteins as determined by enzyme-linked immunosorbent assay (ELISA) in cell extracts (CEs) and conditioned media (CM). Zymography of the CM confirmed the inhibitory effect of E2 on uPA activity. Thus, we now report the regulation of uPA, PAI-1 and tPA by E2 in both mRNA and protein levels in ER transfectants. The association between down-regulation of the uPA by E2 and known E2-mediated growth inhibition of these cells was also explored. Our findings indicate that down-regulation of uPA by E2 is an upstream event of inhibitory effects of E2 on growth of these cells as the addition of exogenous uPA did not block the growth inhibition by E2.

tumour growth. invasion and metastasis is well documented (Kwaan. 1992. Tumour cell invasion is accomplished by the concerted action of several extracellular proteolytic enzyme systems. one of which is the plasminogenplasmin system. The different components of this system. e.g. urokinase-type plasminogen activator (uPA). its receptor (uPAR). tissue-type plasminogen activator (tPA) and the plasminogen activator inhibitor type 1 (PAI-1), along with other proteolytic enzymes. are involved in the process of activation of plasminogen to plasmin, which, directly or indirectly through the activation of other matrix metalloproteinases, degrade most components of the extracellular matrix and basement membrane.
The uPA/uPAR system plays a key role in tumour cell invasion and dissemination (Dan0 et al. 1985: Kwaan. 1992). Binding of both double chain uPA (tcuPA) and single-chain uPA (scuPA) to its receptor (uPAR) with the concomitant cell-surface binding of plasminogen enhances plasmin generation . uPAmediated proteolysis is modulated by PAI-1. the inhibitor for both uPA and tPA. PAI-I binding to receptor-bound uPA results in internalization of the uPA/PAI-l complexes (Olson et al. 1992 Whether the uPAR is intemalized at the same time (Bastholm et al. 1994) or the receptor just plays an enhancing role in internalization (Olson et al. 1992) is not clear.
In carcinoma of the breast, the level of uPA antigen. but not tPA antigen. in tumour homogenates was found to be a strong unfavourable prognostic factor for relapse and overall survival (Janicke et al, 1989(Janicke et al, : 1990Foekens et al. 1992;Sumiyoshi et al, 1992). High PAI-I and uPAR levels have also been found to be unfavourable (Duffy et al. 1988;Janicke et al. 1990;Foekens et al. 1992;Grondahl-Hansen et al, 1993. Several studies have found no correlation between high uPA (Grondahl-Hansen et al, 1997), PAI-l or uPAR levels and the oestrogen receptor (ER) status of the tumour (Duggan et al. 1995Gohring et al, 1995), whereas others (Foekens et al, 1994;Grondahl-Hansen et al. 1993;Ferno et al. 1996) have demonstrated that high uPA.
PAI-1 and uPAR contents in tumours are negatively correlated with the ER and progesterone receptor (PgR). In contrast, Duffy and colleagues (1986) state that levels of tPA activity had the highest correlation with ER and PgR positivity in human breast carcinomas. Determination of uPA levels and. to a lesser extent, PAI-1 levels was found to be useful in predicting the rate of response to tamoxifen therapy for metastatic disease . Cell lines derived from human carcinoma of breast have also been studied for the expression of uPA, uPAR. PAI-I and tPA (Butler et al. 1979;Shyamala and Dickerman. 1982;Huff andLippman. 1984: Mangel et al, 1988;Madsen andBriand. 1990: Holst-Hansen et al. 1996) and their role in cancer cell invasion and metastasis (Madsen and Briand, 1990;Holst-Hansen et al. 1996: Long and Oestadiol regulation of uPA, tPA, PAI-1 in breast cancer cells 89 Rose, 1996). The plasminogen activators' (PAs) activity in ERcontaining human breast cancer cell lines is stimulated by oestradiol (E,) and suppressed by anti-oestrogens (Butler et al, 1979;Shyamala and Dickerman, 1982;Huff and Lippman, 1984;Katzenellenbogen et al. 1984). Most investigators, with only one exception (Yang et al, 1983), have found that tPA and not uPA is regulated by E, in MCF-7 cells (Ryan et al, 1984;Dickerman et al, 1989;Butler et al, 1979;Mizoguchi et al, 1990). On the other hand, Mangel et al (1988) have shown that uPA was stimulated by E, in the T47-D and ZR-75-1 cells containing lower levels of ER, whereas MDA-MB-231 cells, which do not contain ER, showed a high level of both PAs activity that was not modulated by E,.
Interactions between oestrogen, tamoxifen and retinoic acid are also reflected in the expression of PAs by breast cancer cells (Butler and Fontana, 1992). PAI-1 levels were not influenced by E, in MCF-7 culture media (Davis et al, 1995), whereas in endometrial Ishikawa cells both E, and progesins induced the expression of PAI-1 and its mRNA (Fujimoto et al, 1996). Thus, in hormonedependent breast cancers, dissolution of the extracellular matrx may be modulated by PAs and PAI-1 under hormonal control.
To understand the hormonal regulation of components of plasminogen-plasmin system, we examined their E, regulation in our model system, using ER-negative MDA-MB-231 breast cancer cells transfected with either the wild-type (S30 cells) or the codon 351 AsT mutant ER (BC-2 cells). The purpose of this study was, using S30 and BC-2 cell lines, to: (a) examine the oestogen responsiveness of the components of the plasminogen-plasmin system, namely uPA, tPA, PAI-1 and uPAR; (b) determine if the process is ER-mediated; and (c) correlate the oestrogen responsiveness of the uPA with known growth inhibitory effect of E, on these cells.

Northem blot analysis
Northern blot analysis was performed as described previously (Levenson et al, 1997). Briefly, total RNA was isolated from cells following 48 h of treatment with compounds using the Trizol Reagent (Gibco, BRL). Twenty micrograms of total RNA sample was fractionated in 1.2% agarose-formaldehyde gel and transferred to a nylon membrane (Hybond-N+; Amersham, Arlington Heights, IL, USA). The membranes were hybridized at 420C with the corresponding 32P-labelled probes. The membranes were then washed and autoradiographed by exposure to Hyperfilm (Amersham) at -8C with intensifying screens. The expected 2.5-kb (uPA), 2.6-kb (tPA), 3.2-and 2.3-kb (PAI-1), and 1.4-kb (uPAR) tanscripts were detected. Because the sizes of mRNAs were similar, probing was performed on separate blots for each gene. Subsequently, the blots were stripped and reprobed with a cDNA to 1-actin. The signals were quantitated using phosphorimage analysis (Molecular Dynamics phosphorimager, Image Quant software).
Enzymelinked immunosorbent assay (ELISA) Two hundred thousand cells were plated per well in six-well dishes in 3 ml of media as described above. The cells were then treated with 10-9 M oestradiol or ethanol (control) for 48 h. Cell extracts (CEs) were isolated using Camiolo buffer (potassium acetate 75 mM, sodium chloride 300 mM, L-arginine 100 mM, EDTA 10 mM and 0.25% Triton X-100). The conditioned media (CM) were centrifuged at 5000 g and aliquots were stored at -00C until use. The corresponding cells were counted so that the levels of proteins could be standardized to the cell number. For the ELISA, the CM were diluted 1:10, and CEs were diluted 1:50 in PBS EDTAJTween and samples were analysed for uPA, tPA, PAI-l antigen levels using the ELISA kits TintElize uPA, tPA or PAI-I (Biopool, Sweden) with the corresponding antibodies. The uPA ELISA recognizes the scuPA as well as active tcuPA, whether it is free, receptor bound or complexed with PAI-1. For the quantitation of uPAR as well as uPAR/uPA and uPAR/uPA/PAI-l complexes the uPAR ELISA kit (American Diagnostica, Greenwich, CT, USA) was used. All assays were performed as described by the supplier.

Chromogenic assays
The proteolytic activity of uPA was assayed using a commercial Chromolize kit (Biopool, Ventura, CA, USA). The CM were added to a microtest plate well coated with monoclonal antibody against uPA, enabling the adsorption of uPA to the plate wall. After washing the non-absorbed material, plasmin was added to the well for the conversion of all scuPA to the active tcuPA. The uPA activity was determined by the addition of plasminogen and a plasmin-sensitive chromogenic substrate (D-But-CHT-Lys-pNA). uPA converted the plasminogen to plasmin, which reacts with the chromogenic substrate. This reaction was assayed by measurement of the absorbance at 405 um.
Zymography uPA, tPA and PAI-1/PA complex activities were analysed using zymography. Samples were electrphoresed in the presence of sodium dodecyl sulphate (SDS) under denaturing but non-reducing conditions in 10% polyacrylamide gels (SDS-PAGE) using the buffer system of Laemmli (1970) by a modification of the methods of Granrelli-Piperno and Reich (1978) and DePetro et al (1984).
Brffiish Journal of Cancer (1998) 78(1) uPA Diagnostica) were applied in each gel. In addition, standards for uPA/PAI-l complex and for tPA/PAI-l complex were applied. They were produced by first activating the 3.5 gg of PAI-I by boiling for 30 s, cooling slightly, adding excess uPA and tPA, respectively, and incubating at 37°C for 30 min to permit complex formation.
Growth assays Growth assays were performed as described previously (Levenson et al, 1997). Briefly, cells maintained in the media described above were plated at a density of 4x104 cells per well in 24-well plates and cultured for the 6 days either in media containing different concentrations of E,, uPA or in combinations. The media were changed every other day. Cells were then harvested and sonicated for 20 s with an ultrasonic cell disrupter.    Figure IA shows that expression of uPA mRNA is down-regulated by E, and that this effect is concentration dependent. In the presence of physiological concentrations of E, (IO9 M), there was a 3.6-fold decrease in uPA mRNA levels. At E, concentations of 106 M, the decrease was 7.4-fold. The expression of PAT-I mRNA was also down-regulated by E, in a concentration-dependent manner ( Figure IB). In the presence of 109 M E, there was a 1.7-fold decrease in PAI-I mRNA levels compared with unteated cells. In the case of tPA, the E, effect was even more pronounced ( Figure IC). In the presence of 1X9 M E,, there was 6.2-fold decrease compared with untreate cells. In contrast, uPAR mRNA renmained unaffected by E, at any given concentration ( Figure ID). To examine further the E,-mediated repression of uPA, tPA and PAI-I gene expression we used the pure antioestrogen ICI 182,780 alone and in combination with E, (Figure  2). The compound ICI 182,780 alone did not have any effect on these mRNA expressions, and in combination with E, was able to block E,-induced down-regulation of uPA, PAT-I and tPA mRNAs (Figure 2A-C). These data suggest that E, regulation of uPA, tPA and PAI-I genes occurs through the ER-mediated pathway. Once again, uPAR mRNA was not affected either by E, or by ICI 182,780 ( Figure 2D). The effect of E2 on evels ot uPA, PAM-1, tPA and uPAR as determned by EUSA CE and CM of S30 cels as descrbed Mateials and methods.
The data repesent dilleres between the l ev of comIponent in untreated contrl (100%) and levels in cels teated with 10-9 m E2 (% reduction) + s.e.m. *P< 0.005; *P= 0.06 in uPA levels in all cases. The uPA level when expressed as a ratio of E, treatment vs baseline was 0.49, indicating that E, has an inhibitory effect (Figure 3). This inhibitory effect of E, on uPA production reached statistical significance in CEs (P < 0.005) and less so in CM because of the variability of CM samples. Assay of PAI-I showed that E, also decreases levels of PAI-I in both CEs and CM, but to a lesser extent than uPA. Examination of tPA levels also revealed a definite decrease in this protease in expressed (CEs) and secreted (CM) levels in hormone-trea cells compared with controls ( Figure 3). By contrast, the uPAR levels in these cells were unaffected by E, treatment. We have previously shown that S30 and BC-2 cells are growth inhibited by 10-l-104 M concentrations of E, (Jiang and Jordan. 1992; Levenson et al, 1997). The fact that E, simultaneously inhibits endogenous uPA synthesis and activity while having no effect on uPAR prompted us to examine whether exogenous addition of intact uPA could (a) exert any effect on proliferative behaviour of S30 and BC-2 cells; and (b) reverse the inhibitory effect of E2 on growth. We performed a series of expenrments using various concentrations of uPA in a range from 10 nm to 5000 nm. The proliferative effect of added uPA was not different from that of untreated control cells at days 2, 4 and 6 (data not shown). In combination experiments with various concentrations of E,, uPA in 106 M (1000 nM) was not able to reverse the E,-inhibitory effect on cell growth ( Figure 5). To dissect possible reasons for added uPA failure, we performed both zymography ( Figure 6A) and ELISA ( Figure 6B) using CM from S30 cells treated with E,, uPA and combinations of E, and uPA from day 2 (Fig 6). E, inhibited activity of uPA. tPA and uPAJ PAI-1. and tPA/PAI-1 complexes. Interestingly. the addition of exogenous (lane 5) uPA alone resulted in its localization as both free uPA and as uPA/PAI-l complex. When added to E, (lane 6), the activity of uPA is decreased by both E2 and formation of the uPAIPAI-1 complex. The uPAIPAI-l complex is also decreased because of the inhibitory effect of E, on PAI-l levels. Samples from this experiment were examined using ELISA to determine the total (endogenous + exogenous) amount of uPA ( Figure 6B).

Effect of
Results showed that the addition of exogenous uPA incrased the total amount of uPA; however, when added to uPA. E,. was able to decrease amount of total uPA below control levels. high levels of uPA, PAI-I and uPAR were found to be independent prognostic factors with respect to relapse-free and/or overall survival, particularly in post-menopausal women (Gr6ndahl-Hansen et al, 1992;Spyratos et al, 1992;Gohring et al, 1995;Femo et al, 1996;Grondahl-Hansen et al, 1992). When analysed with respect to the ER status of the tumour, no clear answer was found: some found no correlation with the ER (Duggan et al, 1995;Foekens et al, 1995;Grondahl-Hansen et al, 1997) whereas others (Grondahl-Hansen et al, 1992;Foekens et al, 1994;Ferno et al, 1996) demonstrated negative correlation between high levels of prognostic factors and ER/PgR status of the tumour. As to the hormonal regulation of the components of the plasminogen-plasmin system in breast cancer cells in culture and its ER-mediated nature, the findings are controversial, especially for uPA. tPA but not uPA, PAI-1 or uPAR has been found to be regulated by E, in breast cancer cells containing the ER (Butler et al, 1979;Dickerman et al, 1989;Davis et al, 1995).
In the surgical specimens studied the expression of uPA in breast carcinoma was not in the tumour cells but in the myofibroblasts and other stromal cells (Nielsen et al, 1996). However, this picure is far from being clear. For example, xenografts of MDA-MB-231 cells in nude mice produced tmours that, in in situ hybridization studies, showed 'mRNA for human uPA in virtually all the cancer cells' (Romer et al, 1994). In either case, the stromal cells or the MDA-MB-231 cells do not express the ER. Thus, we wish to stress that we are evaluating a model system of uPA regulation in the laboratory that does not directly replicate the clinical situation in breast cancer. The availability of cells derived from MDA-MB-231 human breast cancer cells, which stably express ER and contain high levels of PAs, prompted us to examine the relationship between ER content and hormonal regulation of expression of the components of the plasminogen-plasmin system. To control at least three steps of E, regulation of final PA activity, e.g. from gene transcription to the enzyme activity, we concentrated on studying the regulation of mRNA accumulation, intra-and extracellular protein concentrations, and enzyme activities.
We herein report an observation that not only tPA but also uPA Because the PA protein levels do not necessarily reflect the PA activity, we performed zymographic analyses to examine the effect of E2 on the activity of uPA and tPA in the media conditioned by cells in the absence and pesence of 10-9 M E2. We found that E2 inhibits activity of uPA, tPA and PA/PAI-1 complexes. Thus, the inhibitory effect of E2 on PA activities in these cells occurs at any level examined from mRNA to enzyme activity.
It has been reported that uPAR is present in breast cancer tissue but not in normal breast tissue (Needham et al, 1987). uPAR is expressed in a variety of cancer cell lines and its synthesis is regulated by growth factors, such as epidennal growth factor (EGF), transforming-growth factors (TGF-1 and TGF-f2) and by the umour prmoter phorbol myristate acetate (Lumd et al, 1991a(Lumd et al, , 1995. In its native form uPAR is a glycolipid-anchored integral membrane protein. Therefore, we were interested in examining uPAR levels in cell extracts that represent potentially functional receptors compared with water-soluble tion products of uPAR found in cytosols (Gr6ndahl-Hansen et al, 1995). We found that our cells express uPAR, and that its synthesis is not rgulated by E2.
A mitogenic effect of uPA has been demonstra in vitro in different cell lines, both normal and neoplastic (Rabbai et al, 1990;He et aL 1991;De Petro et al, 1994;luperello et al, 1996).
Conversely, the inhibition of endogenously produced uPA by human malignant melanoma cells impairs cell proliferation (Kirchheimer et aL 1989). It is believed that the mitogenic and growth factor activity of uPA occurs throgh the amnotermnal fragment (AIT) of uPA, containing the growth factor domain, which shares sucntures homologous to the EGF, the TGF-a and the Krinkle domain. The ATF is a binding site of the uPA molecule for uPAR. The activity of uPA depends on the presence of uPAR on the cells as its mitogenic effect is selectively blocked by the addition of antibodies specific for the receptor (De Petro et al, 1994 this effect should be reversible upon the addition of exogenous uPA. However, the addition of exogenous uPA alone to culture media did not have any growth stimulatory effects on S30 and BC-2 cells, indicating that stimulatory effects of uPA varies with different cell types. When added to E2, uPA was not able to block the inhibitory effect of E2 on the growth. Tlere are several possible explanations: (a) down-regulation of uPA by E, is an upstream event of inhibitory effect of E2 on growth. Thus, replacement of uPA would not affect El-mediated growth inhibition. Parallel to our findings, Long et al showed (1996) that although EGF and TGF-a exerted stimulatory effects on uPA expression in S30 cells, neither of these growth factors was able to reverse the suppressive action of E, on growth of these cells. (b) Exogenously added uPA is inactivated by forming a uPA/PAI-l complex ( Figure  6A). It is possible that recombinant tcuPA added to the culture is more predisposed to complex formation with PAI-1 than endogenous cellular uPA, which is secreted as a scuPA. This can explain the detection of a large uPA/PAI-l band ( Figure 6A, lane 5) and subsequent absence of proliferative effect of uPA on these cells. Tlne addition of E, along with exogenous uPA to the culture affects endogenous uPA levels as well as uPA/PAI-l complex formation ( Figure 6A, lane 6).
In sumnmay, we have shown that uPA, tPA and PAI-1 are regulated by E, via the ER-mediated pathway in breast cancer cells stably transfected with the ER inreasing levels of endogenous uPA by transfection experiments with expression plasmids containing the uPA gene might be useful in future investigations.