Rel transcription factors contribute to elevated urokinase expression in human ovarian carcinoma cells

Eur. J. Biochem. 259, 143±148 (1999) q FEBS 1999

Rel transcription factors contribute to elevated urokinase expression in human ovarian carcinoma cells Ute Reuning1, Luisa Guerrini2, Tomizo Nishiguchi3, Sharon Page4, Hildegard Seibold1, Viktor Magdolen1, Henner Graeff1 and Manfred Schmitt1 1

Frauenklinik der Technischen UniversitaÈt MuÈnchen, Germany; 2University of Milan, Department of Genetics and Microbiology, Italy; Hamamatsu University, School of Medicine, Japan; 4Institut fuÈr Klinische Chemie der Technischen UniversitaÈt MuÈnchen, Germany

3

Elevated levels of the urokinase-type plasminogen activator (uPA) in tumor cells are conductive to tumor cell spread and metastasis. In a previous study we observed that suppression of RelA dramatically reduced endogenous uPA synthesis in the human ovarian cancer cell line OV-MZ-6. Because the uPA promoter contains three potential Rellike protein binding motifs (RRBE, 5 0 -NF-kB, and 3 0 -NF-kB) we conducted the first thorough systematic uPA promoter analysis to examine the direct impact of Rel proteins on uPA gene transcription. Disruption of RRBE resulted in a ,40% decrease in uPA promoter activity, mutation of the 5 0 -NF-kB motif led to an additional 20% decrease. The 3 0 -NF-kB motif was not active. Overexpression of RelA significantly enhanced uPA promoter activity, whereas IkB-a overexpression reduced uPA promoter activity by 40%. These data were supported by the finding that endogenous uPA was also increased sixfold by overexpression of RelA and decreased by 30% upon overexpression of IkB-a. Transfection of OV-MZ-6 cells with antisense deoxynucleotides directed to RelA expression reduced uPA promoter activity by at least 40%. Our data clearly suggest that by binding to uPA promoter elements, Rel transcripton factors contribute directly to elevated uPA gene expression in human ovarian cancer cells, thereby promoting the multiple functions of uPA during tumor growth and metastasis. Keywords: ovarian cancer; Rel transcription factors; tumor invasion; tumor metastasis; uPA promoter; urokinasetype plasminogen activator (uPA). Tumor invasion and metastasis depend on the capacity of tumor cells to remodel extracellular matrix in order to cross tissue boundaries, extravasate into blood and lymph vessels, distribute into various compartments and form metastases. This complex scenario requires the concerted and regulated expression of pericellular proteolytic enzyme systems [1, 2]. Many tumor cells synthesize and secrete increased levels of the serine proteinase, urokinase-type plasminogen activator (uPA) crucial for tumor invasion and metastasis [2,3]. Several in vitro and in vivo experimental approaches have been undertaken to directly target uPA in tumor cells and thus reduce its unwanted effects [4±8]. In a previous study [9] we reasoned that an early downregulation of uPA synthesis at the transcriptional level might be even more effective than blocking the already produced uPA protein. Studies to elucidate regulatory events leading to uPA overexpression in tumor cells have led to a detailed characterization of the human uPA promoter [10±14]. This promoter contains two binding sites for Rel transcription factors around nucleotide positions 21865 and 21835, respectively, referred to here as the 5 0 -NF-kB motif and 3 0 -NF-kB motif [15], and an additional Rel motif between nucleotide position 21596 and 21588, the Correspondence to U. Reuning, Frauenklinik der Technischen UniversitaÈt MuÈnchen, Klinikum rechts der Isar, Ismaninger Str. 22, D-81675 MuÈnchen, Germany. Fax: +49-89-4140-7410, Tel: +49-89-4140-2493, E-mail: [email protected] Abbreviations: ASODN, antisense deoxynucleotide; EMSA, electromobility shift assay; FITC, fluorescein isothiocyanate; ONPG, ortho-nitrophenyl b-d-galactopyranosid; RRBE, Rel-related binding element; uPA, urokinasetype plasminogen activator. (Received 11 August 1998, revised 6 October 1998, accepted 9 October 1998)

Rel-related binding element (RRBE) [16]. Rel proteins are thought to be involved in the development and progression of cancer. Members of the Rel family, including RelA (p65), NFKB1 (p50), NFKB2 (p52), c-Rel and RelB, rapidly transmit signals from the cytoplasm to the nucleus. As dimers, Rel proteins associate with the cytoplasmic inhibitors (IkB) thus blocking their nuclear translocation and DNA binding activity. Upon phosphorylation and degradation of IkB, Rel proteins are liberated from IkB and translocated to the nucleus where they bind to a decameric kB-DNA consensus sequence, resulting in immediate changes in gene transcription [17]. In our earlier study the successful suppression of uPA mRNA and protein by RelA-antisense deoxynucleotide (ASODN) might have been an indirect effect due to changes of uPA regulatory proteins, therefore we systematically analyzed the still unclear impact of Rel transcription factors on the activity of the fulllength human uPA promoter in human ovarian cancer cells.

EX PERIM E NTAL PR OC EDUR ES Materials Lipofectin was obtained from Gibco, BRL (Eggenstein, Germany). g[32P]-ATP (3000 Ci´mmol21) was from Amersham (Braunschweig, Germany). Poly (dI±dC) and ortho-nitrophenyl b-d-galactopyranosid (ONPG) were purchased from Sigma (Deisenhofen, Germany). Polyclonal rabbit antibodies directed against Rel proteins were bought from Santa Cruz Inc. (Santa Cruz, CA, USA). The in vitro mutagenesis kit was purchased from BioRad (MuÈnchen, Germany). The plasmids pGL2Control, pGL2-Basic and pSV-b-Gal were obtained from

144 U. Reuning et al. (Eur. J. Biochem. 259)

q FEBS 1999 Fig. 1. Scheme of the in vitro mutagenesis performed on the Rel-like binding motifs of the human uPA promoter, 5 0 -NF-kB, 3 0 -NF-kB and RRBE. uPA promoter sequences, altered at the indicated nucleotide positions, are underlined. Restriction sites created upon mutagenesis are indicated (Asp 700, SpeI). The PCR fragments spanning the uPA wild-type promoter sequence stretch from ± 2062 to + 27, as well as the promoter sequence containing the mutagenized Rel-like binding sites were inserted into the vector pGL2-Basic for transient transfections of OV-MZ-6 cells and luciferase reporter gene assays.

Promega (Madison, WI, USA). ODN were synthesized and purified by HPLC at MWG Biotech (Ebersberg, Germany).

c-Rel, and IkB-a were as described previously [9]. Cell lysates were prepared as described [14].

Cell culture

Luciferase and b-galactosidase assay

The origin and cultivation of the human ovarian cancer cell line OV-MZ-6 has been described previously [9,18]. 4b-Phorbol 12myristate 13-acetate stimulation of OV-MZ-6 cells was performed at a concentration of 20 nm for 6 h following 24-h of serum depletion.

Luciferase activity was determined as described previously [20]. Transfection of cells with the plasmid pGL2-Control served as positive controls. The activity of b-galactosidase was measured using ONPG as the substrate [14]. The ratio of luciferase activity to b-galactosidase activity served as a measure for normalized relative luciferase units. Randomly initiated transcriptional events occurred upon transfection with the promoterless plasmid pGL2-Basic, normalized values for this were subtracted from all values obtained.

In vitro mutagenesis and transient transfection In vitro mutagenesis was conducted in Bluescript II KS (+) as described previously [19]. The ODN used were as follows: 5 0 NF-kB-Mut: 5 0 -ATC AGG AAA TGA TTC GGG GGC GCG3 0 ; 3 0 -NF-kB-Mut: 5 0 -CCC CGA AAC TAG TAG GTT AGT T3 0 ; RRBE-Mut: 5 0 -CTG CCT GCT GAC TAG TGT ACA AGT TAG-3 0 and 3 0 -NF-kB-con: 5 0 -GAC TTA CCC GGA AAT TCC CAG GTT-3 0 (Fig. 1). The full-length human uPApromoter (± 2062 to + 27) was cloned into the luciferase reporter gene plasmid pGL2-Basic. OV-MZ-6 cells were transfected via liposomes with the luciferase reporter gene plasmids (10 mg), the pSV-b-Gal plasmid (5 mg) and the CMVdriven expression plasmids (5 mg). ASODN for RelA, NFKB1,

ELISA uPA antigen concentrations in cell lysates and cell culture supernatants were determined using the ELISA kit IMUBIND uPA # 894 (American Diagnostica, Greenwich, CT, USA). Electromobility shift assay Nuclear and cytosolic extracts were prepared as described previously [21]. Electromobility shift assays (EMSA) were performed as published elsewhere [22]. The sequences of the

Fig. 2. Analysis of the impact of Relprotein binding sites on human uPA promoter activity in OV-MZ-6 cells. Relprotein binding sites within the human uPA promoter were mutagenized individually and/or in all possible combinations by sitedirected in vitro mutagensis (Fig. 1). Disruption of the Rel-binding sites are indicated by crosses. The 3 0 -NF-kB-motif is depicted by an open box which is changed into a solid box upon alteration of the 3 0 -NF-kB-motif into a perfect Rel binding site. uPA promoter/luciferase reporter gene constructs were transfected and luciferase and bgalactosidase assays performed as described. Luciferase data were normalized and corrected as outlined.

q FEBS 1999

Elevated urokinase expression in human ovarian carcinoma cells (Eur. J. Biochem. 259) 145

decrease (Fig. 2c) in transcriptional activity was noticed (Fig. 2d). The 3 0 -NF-kB motif did not contribute to uPA promoter activity (Fig. 2e). Mutagenesis of the 3 0 -NF-kB motif into a perfect Rel-protein binding site did not enhance uPA promoter activity (Fig. 2f). The reduced activity of the uPA promoter mutant lacking the 5 0 -NF-kB motif (Fig. 2c) was reversed by the insertion of a perfect 3 0 -NF-kB-motif (Fig. 2g). Upon OV-MZ-6 cell stimulation by PMA, uPA wild-type promoter activity was increased sixfold, whereas the activity of the uPA promoter mutant in which all potential Rel-protein binding sites had been eliminated was increased only twofold (data not shown). Effect of Rel-ASODN on uPA promoter activity in OV-MZ-6 cells

Fig. 3. Effect of ASODN directed against Rel-protein expression on uPA promoter activity in OV-MZ-6 cells. ASODN targeting the expression of RelA, NFKB1 and c-Rel, respectively, were cotransfected with the uPA wild-type promoter/luciferase constructs and luciferase activity was determined as described. Cotransfection of OV-MZ-6 cells with ODN spanning the nucleotide content of the RelA-ASODN in scrambled order (Mis) served as controls. Transfections with the promoterless plasmid pGL2-Basic served for the evaluation of randomly started transcriptional events. Data are evaluated as described; a representative experiment is shown.

double-stranded ODN used were as follows: uPA-RRBE: 5 0 GCT GCC TGC TGG GGA AAG TAC-3 0 , uPA-5 0 -NF-kB: 5 0 TAA CCT GGG AGT TTC GGG GTA-3 0 and uPA-3 0 -NF-kB: 5 0 -CCC CCT GGG AAT TTC CTG ATA AC-3 0 [15, 16]. Oligodeoxynucleotide (ODN) were 5 0 -end labeled using g[32P]ATP and polynucleotide kinase. Immunocytochemistry OV-MZ-6 cells were fixed in 4% (wt/vol.) paraformaldehyde, 0.1% (wt/vol.) saponin for 30 min at room temperature and subsequently blocked in 1% (wt/vol.) bovine serum albumin (BSA) for 30 min at room temperature. A polyclonal rabbit IgG directed against RelA was incubated at a concentration of 1 mg´mL21 in blocking buffer overnight at 4 8C and detected with a fluorescein isothiocyanate (FITC)-labeled goat antibody directed against rabbit IgG. Staining procedures in the presence of control antibodies and FITC-labeled secondary antibodies alone, served as controls. R E S U LT S Rel-protein binding sites within the human uPA promoter significantly contribute to its activity in human ovarian cancer cells In the present analysis we investigated the direct contribution of Rel transcription factors to uPA promoter activity. The three potential Rel-like sites 5 0 -NF-kB, 3 0 -NF-kB and RRBE contained within the full-length human uPA promoter were mutagenized individually and/or in all possible combinations (Fig. 1). Compared with the basal uPA promoter activity (Fig. 2a), the activity of the uPA promoter mutant lacking the RRBE was reduced by at least 40% (Fig. 2b). When, in addition, the 5 0 -NF-kB motif was disrupted, a further and significant 20%

ASODN suppressing the expression of RelA [9], cotransfected with the uPA promoter/luciferase constructs, reduced uPA wildtype promoter activity by at least 40% (Fig. 3). An ODN exhibiting the nucleotide content of the RelA-ASODN in scrambled order (missense) served as a control and did not elicit any effect on uPA promoter activity. The extent of the reduced uPA promoter activity by RelA-ASODN was similar to that noted after the disruption of all possible Rel-protein binding sites (Fig. 2h). As expected, treatment of OV-MZ-6 cells with RelA-ASODN did not affect the basal activity of the uPA promoter mutant totally deficient in Rel-protein binding sites. ASODN targeting the expression of NFKB1 and c-Rel were not effective in either cases (data not shown). Detection and identification of Rel proteins in OV-MZ-6 nuclei and cytosols Nuclear and cytoplasmic cell extracts were subjected to EMSA using double-stranded ODN encompassing the Rel-protein binding motifs contained within the human uPA promoter. Nuclear OV-MZ-6 proteins bound to the RRBE-spanning ODN in a concentration-dependent manner (Fig. 4A, lanes 1±4) as well as to the 5 0 -NF-kB-motif (data not shown). The specificity of the gel-shifted bands was demonstrated by the inclusion of an excess of unlabeled ODN (Fig. 4A, lanes 9±11). Rel proteins were also detected in cytoplasmic OV-MZ-6 cell extracts after dissociation of the Rel-protein dimers from IkB upon detergent treatment (Fig. 4A, lanes 5±8). In contrast, nuclear OVMZ-6 proteins did not specifically associate with either the ODN encompassing the 3 0 -NF-kB motif, or with the doublestranded ODN spanning the mutagenized sequences (data not shown). In EMSA supershift experiments a shift was provoked by antibodies directed against RelA and NFKB1, respectively (Fig. 4B). RelA-ASODN were tested for their efficiency to suppress RelA expression and thus binding of nuclear OVMZ-6 proteins. OV-MZ-6 cells treated with RelA-ASODN exhibited a markedly reduced nuclear binding activity to ODN encompassing the RRBE (Fig. 4C). Scrambled control ODN did not significantly decrease Rel-protein binding to the RRBE. These data were cross-supported by the immunocytochemical detection of RelA (data not shown). Under non-induced culture conditions, a strong signal for RelA was detected in OV-MZ-6 cytosols compared with a significantly weaker nuclear staining for RelA, thus proving the existence of constitutively active Rel-protein dimers in cultured OV-MZ-6 cells. After OV-MZ-6 cell stimulation by PMA, the ratio of nuclear to cytoplasmic RelA staining increased significantly compared with nonstimulated OV-MZ-6 cells because of the enhanced nuclear translocation of Rel-protein dimers.

146 U. Reuning et al. (Eur. J. Biochem. 259)

q FEBS 1999

Fig. 4. Detection of nuclear and cytosolic Rel proteins in OV-MZ-6 cells by EMSA. Nuclear and cytosolic extracts were prepared as described [21]. EMSA was performed as published with ODN spanning the RRBE [22]. (A) Concentration-dependent binding of OV-MZ-6 nuclear extracts (lanes 1±4) and detergent (CHAPS)-activated cytosolic fractions (lanes 5±8) to ODN spanning the RRBE. A 50-fold molar excess of homologous unlabeled RRBE/ODN was included into the binding reaction in order to prove specific binding of OV-MZ-6 nuclear proteins to RRBE (lanes 9±11). The arrows mark specific and unspecific association of proteins with the probe. (B) Identification of Rel transcription factors in DNA binding protein complexes. EMSA supershift experiments were conducted with OV-MZ-6 nuclear extracts using rabbit polyclonal antibodies directed against RelA (lane 2), NFKB2 (lane 3), NFKB1 (lane 4), c-Rel (lane 5) and RelB (lane 6). DNA binding in the absence of antibodies served as control (lane 1). Arrows indicate supershifted bands with antibodies directed against RelA and NFKB1. (C) Effect of RelA-ASODN transfection of OV-MZ-6 cells on the DNA binding activity of nuclear OV-MZ-6 proteins in EMSA. OV-MZ-6 cells were transfected with RelA-ASODN for 8 h. Extracts from OV-MZ-6 cells treated with lipofectin alone or transfections with ODN spanning the nucleotide content of the RelAASODN in scrambled order (Missense) served as controls.

Effect of RelA and IkB-a overexpression in OV-MZ-6 cells on uPA promoter activity We transiently increased expression of Rel proteins by cotransfection experiments using CMV-promoter-driven plasmids encoding RelA, NFKB1, c-Rel and IkB-a respectively. Overexpression of RelA led to an approximately twofold increase in uPA wild-type promoter activity (Fig. 5). Increased expression of NFKB1 and c-Rel was not effective (data not shown). In contrast, overexpression of IkB-a reduced uPA wildtype promoter activity by < 40%. However, neither RelA nor IkB-a overexpression was effective in producing changes in the activity of the uPA promoter mutant deficient in Rel-protein binding (Fig. 5). In line with this, overexpression of RelA in OV-MZ-6 cells resulted in an approximately sixfold increase in endogenous uPA protein, whereas no effect was seen upon overexpression of NFKB1 and c-Rel (data not shown). Overexpression of IkB-a led to a reduction of up to 30% in uPA protein content in OV-MZ-6 cell supernatants. DISCUSSION In the search for new therapeutic targets to suppress unwanted uPA-mediated tumor biological effects, many attempts have been made to characterize the regulatory events underlying the

elevated expression of uPA in malignant cells. Rel proteins are implicated in tumor development and the progression of cancer [23]. The genes for c-Rel, NFKB1, NFKB2 and Bcl-3 are located at sites of recurrent translocations and genomic rearrangements in cancer [24±26]. Tax-induced fibrosarcomas in transgenic mice were thus suppressed by ASODN targeting of NF-kB expression [27]. In a previous study [9] we observed a prominent decrease in uPA protein and mRNA levels in OV-MZ-6 cells by ASODN directed against RelA expression, accompanied by impaired fibrin degradation, OV-MZ-6 cell invasiveness and adhesion. On the contrary, ASODN directed against the expression of the Rel inhibitor IkB-a significantly increased uPA expression [9]. From the antisense approach, however, we could not judge whether the pleiotropic Rel proteins were affecting uPA promoter activity directly or whether RelA-ASODN provoked downregulation of other uPA regulatory proteins translated into suppressed uPA synthesis. In direct uPA promoter studies we thus demonstrated that RRBE is the most potent Rel-binding element contained within the human uPA promoter acting in an additive manner with the 5 0 -NF-kB motif. The finding that the 3 0 -NF-kB-motif did not contribute to uPA promoter activity is in line with observations by Novak et al. [15]. They demonstrated using a uPA minimal promoter construct that the

q FEBS 1999

Elevated urokinase expression in human ovarian carcinoma cells (Eur. J. Biochem. 259) 147

more pronounced upon PMA stimulation of OV-MZ-6 cells, since PMA provokes the phosphorylation of IkB, finally leading to the dissociation of IkB from the Rel-protein dimers and their nuclear translocation. PMA also significantly increased uPA wild-type promoter activity. The uPA promoter mutant deficient in Rel-protein binding was also induced by PMA but to a lower extent, most probably due to phorbol ester-mediated induction of other uPA promoter elements including AP-1, AP-2, PEA-3 and CREB. In summary, we have shown that Rel transcription factors significantly and directly contribute to elevated uPA expression in human ovarian cancer cells aside from their effects on the expression of proteins such as cytokines and growth factors implicated in uPA regulation. A C K N O W L ED G M EN T

Fig. 5. Effect of RelA overexpression on the activity of the uPA wildtype promoter as well as the uPA promoter mutant deficient in Relprotein binding. OV-MZ-6 cells were cotransfected with reporter gene constructs containing the uPA wild-type promoter (open bars) or the uPA promoter mutant deficient in Rel-protein binding sites (closed bars), and CMV-driven expression plasmids for RelA and IkB-a. Relative luciferase activity was determined and evaluated as described.

5 0 -NF-kB motif conferred inducibility by PMA to the uPA promoter, whereas the 3 0 -NF-kB motif did not, possibly because of its two deviations from the Rel consensus sequence. As expected from our data of the reporter gene assays, OV-MZ-6 nuclear proteins specifically and exclusively associated with ODN spanning the RRBE and the 5 0 -NF-kB motif, respectively. The direct impact of Rel proteins on the transactivation of the uPA promoter reporter plasmids was shown by two approaches: positively, via overexpression of Rel proteins such as RelA; and negatively, by suppressing Rel proteins using ASODN [9]. The fact that ASODN targeting RelA expression significantly reduced uPA promoter activity was perfectly in line with the observed decrease in endogenous uPA protein and mRNA levels upon treatment of OV-MZ-6 cells with RelA-ASODN [9]. The extent of the decrease in uPA protein and mRNA levels was similar to the reduced uPA promoter activity noted upon disruption of all Rel-protein binding sites within the human fulllength uPA promoter. On the contrary, transient RelA overexpression significantly enhanced uPA wild-type promoter activity. Both reduction of RelA expression by RelA-ASODN and overexpression of RelA had no effect on the activity of the uPA promoter mutant lacking functional Rel-protein binding sites. The importance of Rel proteins for uPA transcription was further underlined by the prominent decline in uPA wild-type promoter activity as well as uPA protein levels by the overexpression of IkB-a. These findings were in good agreement with the increase in uPA protein levels after treatment of OVMZ-6 cells with ASODN directed against IkB-a expression. The constitutive nuclear occurrence of activated, DNAbinding Rel-protein dimers under nonstimulated culture conditions was demonstrated in OV-MZ-6 cells by immunofluorescence and EMSA, respectively. Rel-protein dimers are present in OV-MZ-6 cytosols in a latent non-DNA-binding form that can be released from IkB-a and activated for DNA binding by detergent treatment. As expected, RelA nuclear staining was

This work was supported by the Deutsche Forschungsgemeinschaft (Klinische Forschergruppe GR 280/4-5), the BIOMED-1 program of the European Union (Project CBMH1-CT93-1346) and the Deutsche Krebshilfe e. V. (Dr Mildred Scheel-Stiftung). The substantial support of Dr Richard Hart, American Diagnostica, Greenwich, CT, USA, is highly acknowledged.

R E F E RE N C E S 1. Blasi, F. (1993) Urokinase and urokinase receptor: a paracrine/autocrine system regulating cell migration and invasiveness. Bioessays 15, 105± 111. 2. Schmitt, M., Harbeck, N., Thomssen, C., Wilhelm, O., Magdolen, V., Reuning, U., Ulm, K., HoÈfler, H., JaÈnicke, F. & Graeff, H. (1997) Clinical impact of the plasminogen activation system in tumor invasion and metastasis: prognostic relevance and target for therapy. Thromb. Hemost. 78, 285±296. 3. Besser, D., Verde, P., Nagamine, Y. & Blasi, F. (1996) Signal transduction and the u-PA/u-PAR system. Fibrinolysis 10, 215±237. 4. Ossowski, L. & Reich, E. (1983) Antibodies to plasminogen activator inhibit human tumor metastasis. Cell 35, 611±619. 5. Crowley, C.W., Cohen, R.L., Lucas, B.K., Lui, G., Shuman, M. & Levinson, A.D. (1993) Prevention of metastasis by inhibition of the urokinase receptor. Proc. Natl Acad. Sci. USA 90, 5021±5025. 6. Schlechte, W., Murano, G. & Boyd, D. (1989) Examination of the role of the urokinase receptor in human colon cancer mediated laminin degradation. Cancer Res. 9, 6064±6069. 7. Wilhelm, O., Weidle, U., HoÈhl, S., Rettenberger, P., Schmitt, M. & Graeff, H. (1994) Recombinant soluble urokinase receptor as a scavenger for urokinase-type plasminogen activator (uPA). Inhibition of proliferation and invasion of human ovarian cancer cells. FEBS Lett. 337, 131±134. 8. Wilhelm, O., Schmitt, M., HoÈhl, S., Senekowitsch, R. & Graeff, H. (1995) Antisense inhibition of urokinase reduces spread of human ovarian cancer in mice. Clin. Exp. Metast. 13, 296±302. 9. Reuning, U., Wilhelm, O., Nishiguchi, T., Blasi, F., Graeff, H. & Schmitt, M. (1995) Inhibition of NF-kB-RelA expression by antisense oligodeoxynucleotides suppresses synthesis of urokinase-type plasminogen activator (uPA) but not its inhibitor PAI-1. Nucleic Acids Res. 23, 3889±3893. 10. Riccio, A., Grimaldi, G., Verde, P., Sebastio, G., Boast, S. & Blasi, F. (1985) The human urokinase-plasminogen activator gene and its promoter. Nucleic Acids Res. 13, 2759±2771. 11. Verde, P., Boast, S., Franze, A., Robbiati, F. & Blasi, F. (1988) An upstream enhancer and a negative element in the 5 0 flanking region of the human urokinase plasminogen activator gene. Nucleic Acids Res. 16, 10699±10716. 12. Nerlov, C., Rorth, P., Blasi, F. & Johnsen, M. (1991) Essential AP-1 and PEA-3 binding elements in the human urokinase enhancer display cell type-specific activity. Oncogene 6, 1583±1592. 13. Nerlov, C., De Cesare, D., Pergola, F., Caracciolo, A., Blasi, F., Johnsen, M. & Verde, P. (1992) A regulatory element that mediates cooperation between a PEA-3-AP-1 element and an AP-1 site is

148 U. Reuning et al. (Eur. J. Biochem. 259)

14.

15. 16.

17. 18.

19. 20.

required for phorbol ester induction of urokinase enhancer activity in HepG2 hepatoma cells. EMBO J. 11, 4573±4582. Lengyel, E., Gum, R., Stepp, E., Juarez, J., Wang, H. & Boyd, D. (1996) Regulation of urokinase-type plasminogen activator expression by an ERK-1-dependent signaling pathway in a squamous cell carcinoma cell line. J. Cell. Biochem. 61, 430±443. Novak, U., Cocks, B.G. & Hamilton, J.A. (1991) A labile repressor acts through the NF-kB-like binding site of the human urokinase gene. Nucleic Acids Res. 19, 3389±3393. Hansen, S.K., Nerlov, C., Zabel, U., Verde, M., Johnson, M., Baeuerle, P.A. & Blasi, F. (1992) A novel complex between the p65 subunit of NF-kB and c-Rel binds to a DNA element involved in the phorbol ester induction of the human urokinase gene. EMBO J. 11, 205±231. Grilli, M., Chiu, J.J. & Lenardo, M.J. (1993) NF-kB and Rel: participants in a multiform transcriptional regulatory system. Int. Rev. Cytol. 143, 1±62. MoÈbus, V., Gerharz, C.D., Press, U., Moll, R., Beck, T., Mellin, W., Pollow, K., Knapstein, P.G. & Kreienberg, R. (1992) Morphological, immunohistochemical and biochemical characterization of 6 newly established human ovarian carcinoma cell lines. Int. J. Cancer 52, 76± 84. Kunkel, T.A., Roberts, J.D. & Zakour, R.A. (1987) Rapid and efficient site-specific mutagenesis without phenotypic selection. Methods Enzymol. 154, 367±382. Bierhaus, A., Zhang, Y., Deng, Y., Mackman, N., Quehenberger, P., Haase, M., Luther, T., MuÈller, M., BoÈhrer, H., Greten, J., Martin, E., Baeuerle, P.A., Waldherr, R., Kisiel, W., Ziegler, R., Stern, D.M. &

q FEBS 1999

21. 22. 23. 24.

25.

26.

27.

Nawroth, P.P. (1995) Mechanism of the tumor necrosis factor amediated induction of endothelial tissue factor. J. Biol. Chem. 270, 26419±26432. Dignam, J.D., Lebovitz, R.M. & Roder, R.G. (1983) Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 11, 1475±1489. Hansen, S.K., Guerrini, L. & Blasi, F. (1994) Differential DNA sequence specificity and regulation of HIV-1 enhancer activity by cRel-Rel A transcription factor. J. Biol. Chem. 269, 22230±22237. Moore, B.E. & Bose, H.R. (1988) Expression of the v-rel oncogene in reticuloendotheliosis virus-transformed fibroblasts. Virology 162, 377±387. Neri, A., Chang, C.C., Lombardi, L., Salina, M., Corradini, P., Maiolo, A.T., Chaganti, R.S.K. & Dalla-Favera, R.B. (1991) B cell lymphomaassociated chromosomal translocation involves candidate oncogene lyt-10, homologues to NF-kB p50. Cell 67, 1075±1087. Liptay, S., Schmid, R.M., Perkins, N.D., Meltzer, P., Altherr, M.R., McPherson, J.D., Wasmuth, J.J. & Nabel, G.J. (1992) Related subunits of NF-kappa B map to two distinct loci associated with translocation in leukemia, NFKB1 and NFKB2. Genomics 13, 287±292. Mathews, S., Murty, V.V.V.S., Dalla-Favera, R. & Chaganti, R.S.K. (1993) Chromosomal localization of genes encoding the transcription factors c-rel, NF-kBp50, NF-kBp65 and lyt-10 by fluorescence in situ hybridization. Oncogene 8, 191±193. Kitajima, I., Shinohara, T., Bilakovics, J.D., Xu, X. & Nerenberg, M. (1992) Ablation of transplanted HTLV-1 tax-transformed tumors in mice by antisense inhibition of NF-kB. Science 258, 1792±1795.