Minimal and inducible regulation of tissue factor pathway ... - Nature

Oncogene (2002) 21, 921 ± 928 ã 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00 www.nature.com/onc

Minimal and inducible regulation of tissue factor pathway inhibitor-2 in human gliomas Santhi D Konduri1, Francis Ali Osman5, Chilukuri N Rao6, Harish Srinivas4, Niranjan Yanamandra1, Anastasia Tasiou1, Dzung H Dinh2, William C Olivero2, Meena Gujrati3, Donald C Foster7, Walter Kisiel8, Gregory Kouraklis9 and Jasti S Rao*,1,2 1

Division of Cancer Biology, Department of Biomedical and Therapeutic Sciences, University of Illinois, Peoria, Illinois, USA; Department of Neurosurgery, University of Illinois, Peoria, Illinois, USA; 3Department of Neuro-pathology, University of Illinois, Peoria, Illinois, USA; 4Department of Thoracic/Head and Neck Medical Oncology, The University of Texas, M.D. Anderson Cancer Center, Houston, Texas, USA; 5Department of Neurosurgery, The University of Texas, M.D. Anderson Cancer Center, Houston, Texas, USA; 6Center for Prostate Disease Research, Rockville, Maryland, USA; 7ZymoGenetics, Inc., Seattle, Washington, USA; 8University of New Mexico Health Sciences Center, Albuquerque, New Mexico, USA; 9Department of Propedeutic Surgery, Athens University School of Medicine, Athens, Greece

2

Tissue factor pathway inhibitor-2 (TFPI-2), a serine protease inhibitor abundant in the extra cellular matrix, is highly expressed in non-invasive cells but undetectable levels in highly invasive human glioma cells. The mechanisms responsible for its transcriptional regulation are not well elucidated. In this study, we made several deletion constructs from a 3.6 kb genomic fragment from Hs683 cells containing the 5'-¯anking region of the TFPI-2 gene, transiently transfected with these constructs into non-invasive (Hs683) and highly invasive (SNB19) human glioma cells, and assessed their expression by using a luciferase reporter gene. Three constructs showed high promoter activity (pTF5, 7670 to +1; pTF6, 7312 to +1; pTF2, 71511 to +1). Another construct, pTF8 (781 to +1), showed no activity. PTF9, a variant of pTF5 in which a further 231 bp fragment (7312 to 781) was deleted, from the [7670 to +1] pTF5 region, also showed no promoter activity. Hence, (7312 to 781) this region is essential for the transcription of TFPI-2 in glioma cells. Sequencing of this promoter region revealed that it has a high G+C content, contains potential SP1 and AP1 binding motifs, and lacks canonical TATA and CAAT boxes immediately upstream of the major transcriptional initiation site, although CAAT boxes were found about 73000 bp upstream of the transcription start site. We also found a strong repressor in the region between 7927 to ± 1181, upstream of the major transcriptional initiation site, followed by positive elements or enhancers between ± 1511 to 71181. These positive elements masked the silencer eect. Finally TFPI-2 was induced in Hs683 cells transfected with the pTF6 construct (7312 to +1) and stimulated with phorbol-12-myristate13-acetate (PMA). We conclude that the 7312 to +1

*Correspondence: JS Rao, Chief, Division of Cancer Biology, DBTS and Neurosurgery, School of Medicine at Peoria, One Illini Drive, Peoria, Illinois, IL-61656, USA; E-mail: [email protected] Received 2 June 2001; revised 13 August 2001; accepted 20 August 2001

region is critical for the minimal and inducible regulation of TFPI-2 in non-invasive (Hs683) and highly invasive (SNB19) human glioma cell lines. Oncogene (2002) 21, 921 ± 928. DOI: 10.1038/sj/onc/ 1204983 Keywords: promoter; transcriptional regulation; TFPI2; gliomas Introduction Tissue factor pathway inhibitor-2 (TFPI-2), also known as matrix associated serine protease inhibitor (MSPI) and placental protein 5 (PP-5), consists of three tandemly-arranged Kunitz-type protease inhibitor domains homologous to tissue factor pathway inhibitor (TFPI), a major negative regulator of the extrinsic pathway of blood coagulation (Sprecher et al., 1994; Butzow et al., 1988). The gene for human TFPI-2 has been mapped to chromosome 7q22 by ¯uorescence in situ hybridization (Miyagi et al., 1996). Human TFPI-2 is synthesized and secreted by several cell types including endothelial cells (Rao et al., 1995a), smooth muscle cells, ®broblasts (Rao et al., 1995b), keratinocytes (Rao et al., 1995c) and syncytiotrophoblasts (Udagawa et al., 1998). Most (70-90%) of this protease inhibitor is associated with the extra cellular matrix (ECM) (Iino et al., 1998), a complex mixture of proteoglycans that participates in cell adhesion, migration, growth and dierentiation (Hascall et al., 1991). ECM also serves as a reservoir for several growth factors (Vlodavsky et al., 1987; Vukicevic et al., 1992), proteases (McGuire and Seeds, 1989; Korner et al., 1993; Nagase and Woessner, 1999) and protease inhibitors (Mimuro et al., 1987; Chen et al., 1992; Gomez et al., 1997) that are important for homeostasis. However, the functional signi®cance of the association of TFPI-2 with the ECM, as well as the identity of the matrix components that interact with TFPI-2, is largely unknown. Recently, we reported that TFPI-2 expres-

Transcriptional regulation of TFPI-2 in human gliomas SD Konduri et al

922

sion in gliomas declines with increasing malignancy (Rao et al., 2001; Konduri et al., 2001a) and that TFPI-2 has a role in tumor invasion in other cancers (Konduri et al., 2000; Lakka et al., 2000; Konduri et al., 2001b). Since, TFPI-2 is present in abundance in non-invasive cell lines but minimal or basal in highly invasive cell lines, we sought to clarify the transcriptional regulation of TFPI-2 in human glioma cell lines by attempting to identify minimal promoter region sucient for transcription of TFPI-2. Experimental procedures Oligonucleotides Oligonucleotides for genomic polymerase chain reaction (PCR) and electrophoretic mobility shift assay (EMSA) were synthesized by Sigma-Genosys (The Woodlands, TX, USA). Genomic PCR primers were designed on the basis of human BAC clone GS1-345D13 from 7q31-q32 (Acc. No. AC002076). The following oligonucleotides were used to amplify a 3.6-kb genomic fragment (pTF1) from Hs683 cells: 5'-TCAAGGTACCAGCTTCATACATCTTGGTTG-3' 5'-GTATCTCGAGAGTGCAGCCTCCGTCAGGAA-3'. Further subsequent fragments (pTF2-pTF8) were ampli®ed from 3.6-kb pTF1 genomic fragment by using several internal primers. All fragments were cloned into the KpnI/XhoI site of PGL-3 basic vector. To create the 429-bp pTF9 clone, a 231 bp fragment (from 7312 to 781) was deleted from the pTF5 region (7670 to +1) with the following oligonucleotides: 5'-CTGAGGTACCGATTACAGGCGCGAGCTACCG-3'; 5'-GATCGAGCTCAATGGGAAATCTGCAAGCTAG-3'; 5'-GATCGAGCTCTGGAGCAGAAAGCCGCGCAC-3'; 5'-GTATCTCGAGAGTGCAGCCTCCGTCAGGAA-3' and the resulting fragment was cloned into KpnI/XhoI site in the PGL-3 basic vector. The oligos used in the EMSAs were as follows: SP1 5'-TCAGGCTCCGCCCCGGCGGGGGTCGGCCGGA-3'; 5'-TCCGGCCGACCCCCGCCGGGGCGGAGCCTGA-3'. Position of AP-1 sites in the three dierent constructs: AP1 (pTF6a) 5'-TGACAGTCCCCGTGCATGAATCAGCCACCCCT-3'; 5'-AGGGGTGGCTGATTCATGCACGGGGACTGTCA-3'; (pTF6b) 5'AATTCCTCTCCCTCTTACACAGTTTGCAGCG-3'; 5'-CGCTGCAAACTGTGTAAGAGGGAGAGGAATT-3'; (pTF6) 5'-ATTTCCCATTACTGACACAAACGCTCCCTCA-3' 5'-TGAGGGAGCGTTTGTGTCAGTAATGGGAAAT-3'; Plasmid constructs Genomic DNA was isolated from Hs683 cells using DNA easy columns (Qiagen, USA). The 3.6-kb pTF1 genomic fragment was ampli®ed by using Pfu Turbo DNA polymerase (Stratagene, CA, USA) from the genomic DNA of Hs683 cells. Two reporter gene vectors were used in this study: the PGL3 basic vector (Promega Corporation, Madison, Wisconsin, USA), which is used to examine TFPI-2

Oncogene

promoter activity and PRL-TK (Promega), is used to monitor transfection eciency. AP1 constructs were made according to the presence of number of AP1 sites. PTF6 (7312 bp to +1) has three AP1 sites, whereas in pTF6b (7221 bp to +1) has two AP1 sites and pTF6a has single AP1 site (7167 bp to +1). All these constructs were cloned into the Kpn1/Xho1 site of PGL-3 basic vector. These constructs were used for luciferase assays and oligos were used for electrophoretic mobility shift assays. Each oligo represents the single AP1 site in three dierent construct regions (pTF6, pTF6a and pTF6b). DNA sequencing All constructs were con®rmed by automated sequencing (Applied Biosystems Inc., USA). For sequencing, 200 ng of the double-stranded templates and 10 ng of the primer were used in the dideoxy method with a Taq Dye Deoxy terminator cycle sequencing kit. Tissue culture The non-invasive and highly invasive human glioma cell lines (Hs683 and SNB19) were grown in Dulbecco's Modi®ed Eagle's Medium (DMEM). All cultures are supplemented with 1% glutamine, 100 mg/ml streptomycin, 100 U/ml penicillin, and 10% fetal bovine serum (pH 7.2 ± 7.4) and maintained in a humidi®ed atmosphere containing 5% CO2 at 378C. Subcultures were taken every 3 ± 5 days. The Hs683 cell line was obtained from the American Type Culture Collection (Manassas, VA, USA), and the SNB19 was gifted from Dr Richard Morrison. Nuclear extraction procedure Nuclear extracts were made as described elsewhere (Schreiber et al., 1989; Chaturvedi et al., 1994). Brie¯y, 26106 cells were treated with 300 ± 400 ml of lysis buer (50 mm KCl, 0.5% Nonidet P-40, 25 mM HEPES [N-2 hydroxyethylpiperazine-N'-2 ethanesulfonic acid] [pH 7.8], 25 mm phenylmethylsulfonyl ¯uoride, 10 mg/ml of leupeptin, 20 mg/ml of aprotinin per ml, 100 mM dithiothreitol) on ice for 4 min. After 1 min of centrifugation at 14 000 r.p.m., the supernatant was removed and saved as a cytoplasmic extract. The nuclei were washed once with the same volume of buer without Nonidet P-40 and then 50 ± 100 ml of extraction buer (500 mm KCl and 10% glycerol with the same concentrations of HEPES, phenylmethylsulfonyl ¯uoride, leupeptin, aprotinin, and dithiothreitol as in the lysis buer) was added depending on pellet size and pipetted several times. After centrifugation at 14 000 r.p.m. for 5 min, the supernatant was harvested as the nuclear protein extract and stored at 7708C. Protein concentration was quanti®ed with a protein assay reagent kit (Bio-Rad, CA, USA). Antibodies Antibodies against nuclear proteins for SP1 were purchased from Santa Cruz Biotechnology, Inc (Santa Cruz, CA, USA). Electrophoretic mobility shift assays (EMSAs) Nuclear extracts from 26106 cells were prepared as


described above and then either used immediately or stored at 7708C. EMSA was performed by incubating 4 mg of nuclear extract for 15 min at 378C with 8 fmol of 32P-end-labeled with double stranded oligonucleotide containing the SP1 and AP1 binding site. The DNA protein complex formed was separated from free oligonucleotide on 7% native polyacrylamide gels, and the gel was then dried. The ®lm was developed after overnight exposure at 7708C. For supershift assays, the nuclear extracts were incubated with SP1 and AP1 for 30 min at 378C before the complex was analysed by EMSA. Transfection assays The non-invasive cell line Hs683 and highly invasive cell line (SNB19) were grown in complete medium as described above, and then 16106 cells were transiently transfected with 1 ± 2 mg of the reporter gene vector containing the dierent deletion constructs and 6 ml of Fugene 6 transfection reagent (Roche Molecular Biochemicals, USA). The internal control was 300 ng of PRL-TK (Renilla luciferase). After transfection, the cells were washed twice with phosphate-buered saline solution, lysed with the lysis buer in the DLR Assay kit (Promega), and luciferase activity was measured as described below. In some experiments, the cells were stimulated after the washes with 200 nm phorbol-12-myristateacetate (PMA) for 24 h. Reporter gene analysis was performed according to the manufacturer's instructions. Luciferase assays Luciferase activity was measured with Promega's DLR luciferase assay kit. Following 24 ± 48 h incubation period, cells were washed twice with phosphate-buered saline and lysed with 100 ml of reporter lysis buer. The lysate was shaken at room temperature for 10 ± 15 min, after which 20 ml of each cell lysate was mixed with 100 ml of buer and measured for luciferase activity in a Turner luminometer (Turner Designs, Sunnyvale, CA, USA) over an integration period of 15 s. Values obtained were normalized to the levels of PRL-TK (Renilla luciferase). Results We generated eight luciferase constructs (pTF1 to pTF8) from the human TFPI-2 promoter region isolated from genomic DNA of non-invasive Hs683 cells (Figure 1). A 670-bp fragment (7670 to +1, pTF5) was found to have slightly higher promoter activity compared to pTF6 (7312 to +1) and pTF2 (71511 to +1) constructs. A strong repressor was found between 71181 to 7927 bp upstream of the major transcriptional initiation site, followed by positive elements between 71511 to 71181 bp. Luciferase activity in non-invasive and highly invasive glioma cell lines (Hs683 and SNB19) TFPI-2 is abundantly expressed in the non-invasive cell line (Hs683) and undetectable expression in the highly

invasive cell line SNB19 (Konduri et al., 2001a). To assess whether the 5'-¯anking region of the TFPI-2 gene can support its transcription in these cell lines, we transiently transfected with all these deletion constructs into non-invasive cells (Hs683) and highly invasive cells (SNB19) and found that transfection with the pTF5 construct led to the slightly higher luciferase activity, followed by the pTF6 and pTF2 clones in Hs683 cells (Figure 2a) and found substantially less luciferase activity in SNB19 cells under any transfection conditions (Figure 2b). Further deletion of a 231-bp (7312 to 781, pTF9) fragment from pTF5 region (7670 to +1) resulted in no reporter activity being expressed in Hs683 cells. The pTF8 fragment (781 to +1) also produced no reporter activity. Hence, the 231-bp fragment (from 7312 to 781) apparently was important for TFPI-2 transcription in Hs683 cells.

923

Luciferase activity in PMA stimulated cell lines Noninvasive cells (Hs683) and highly invasive cells (SNB19) were transfected with all these deletion constructs pTF1 to pTF9 and stimulated with or without PMA (200 nm) after 24 h. PRL ± TK was used as an internal control for transfection eciency. Hs683 cells were transfected with pTF6 construct (7312 to +1) showed the higher luciferase activity (Figure 3a); whereas SNB19 cells were showed substantially less activity regardless of transfection and it is slightly increased upon PMA stimulation (Figure 3b). PMA induction exhibits the presence of AP1 sites in the sequence. Hence, activation or induction of the region between 7312 to +1, pTF6, is important for inducible regulation of TFPI-2. Moreover, deletion of the 231bp fragment 7312 to 781 from 7670 to +1 sequence led to no reporter activity being found in Hs683 cells regardless of PMA treatment (Figure 3a). Nucleotide sequence of human TFPI-2 promoter We next analysed the major-transcriptional region of TFPI-2 obtained from the pTF6 construct (7312 to +1) for the presence of potential transcription factor binding sites by using the BIMAS computer algorithm (National Institutes of Health, Bethesda, Maryland, USA). The pTF6 construct was found to contain many putative transcriptional factors, including SP1, AP1, Lyf-1 and NF-kB. Transcriptional start site (Sprecher et al., 1994) is denoted in Figure 4 with an asterisk (*). Dierential regulation of AP1 in PMA-stimulated Hs683 cells To further narrow down the AP1 region, two more constructs were developed from the pTF6 region that diered with regard to number of AP1 sites: the pTF6a construct had a single AP1 site, the pTF6b construct had two AP1 sites, and the pTF6 had all three AP1 sites. Luciferase activity, with or without PMA treatment, was highest in the pTF6 transfected cells where 3 AP1 sites are present in the construct, followed in descending order by that of the pTF6b (two AP1 sites) and pTF6a (single AP1 site) transfected cells (Figure 5). However, deleting the 7312 to +1 region led to no detectable luciferase activity being Oncogene


924

Figure 1 Locations of TFPI-2 promoter sequences in Hs683 human glioma cells. The roman numerals stand for the ®ve exons of TFPI-2

Figure 2 Expression of luciferase in transfected human glioma non-invasive cell line Hs683 and highly invasive cell line SNB19. Percentage luciferase activity is shown over that of the PGL-3 basic control plasmid. The results are from three independent experiments performed in triplicate. PRL-TK was used to normalize the transfection. The pTF5, pTF6 and pTF2 constructs showed higher promoter activity in Hs683 cells (A) and substantially less promoter activity was observed under any transfection condition in SNB19 cells (B)

Oncogene

produced with or without PMA treatment. Hence, all three AP1 sites in non-invasive cells (Hs683) showed dierential inducible regulation.

in the invasive SNB19 cells. Use of an Sp1 antibody led to supershift of the DNA-protein complexes in the Hs683 cells (Figure 6).

Electrophoretic mobility shift assay of SP1 Electrophoretic mobility shift assays were performed to show the presence of Sp1 sites in the sequence. Nuclear extracts from Hs683 and SNB19 cells were combined with a TFPI-2 promoter-speci®c SP1 oligonucleotide to test binding capacity. More DNA-protein binding complexes formed in the non-invasive Hs683 cells than

Electrophoretic mobility shift assay of AP1 EMSA's were performed to con®rm that three dierent AP1 sites are present at three dierent places in the transcriptional region and these three AP1 sites are induced by PMA. Nuclear extracts from stimulated (200 nM PMA) and unstimulated Hs683 cells were incubated with oligonucleotides for the three AP1 sites


925

Figure 4 Nucleotide sequence of the minimal promoter region for TFPI-2 [7312 to +1]. Potential SP1 and AP1 binding sites are shown. (+1) indicates translational start site; (*) denotes the major transcription initiation site (Sprecher et al., 1994)

Figure 3 Expression of luciferase in transfected human glioma non-invasive cell line Hs683 and highly invasive cell line SNB19. Percentage luciferase activity is shown over that of the PGL-3 basic control plasmid. The results are from three independent experiments performed in triplicate. PRL-TK was used to normalize the transfection. The Hs683 cells transfected with the pTF6 construct expressed high levels of luciferase activity when treated with PMA (A); substantially less luciferase activity was expressed by SNB19 cells regardless of transfection or stimulation condition (B). (+) Indicates PMA stimulation and (7) indicates without PMA treatment

in the pTF6, pTF6a and pTF6b regions. The oligos were made from pTF6, pTF6a and pTF6b regions, has single AP1 site. More DNA-protein complexes formed at all three AP1 sites when cells were treated with PMA than when they were not (Figure 7). Discussion TFP1 and TFPI-2 are both Kunitz-type serine proteinase inhibitors (Sprecher et al., 1994). The human TFPI-1 gene spans about *100 kb and the human TFPI-2 gene does not contain additional exons in the 5'-¯anking region, or exons encoding a junction domain between each Kunitz-type domain (Girard et al., 1991; Enjyoji et al., 1993). TFPI is known as a negative regulator or the extrinsic pathway of blood coagulation, and synthesized primarily in endothelial cells (Bajaj et al., 1990; Werling et al., 1993). Although the domain structure and amino acid sequence of TFPI-2 are similar to those of TFPI, their inhibitory

Figure 5 Dierential regulation of AP1 in Hs638 cells stimulated with PMA. Percentage luciferase activity is shown over that of the PGL-3 basic control plasmid. The results are from three independent experiments performed in triplicate. PRL ± TK was used to normalize the transfection. The highest activity was associated with the pTF6 construct (in which three AP1 sites are present), followed by the pTF6b construct (in which two AP1 sites are present) and the pTF6a construct (in which a single AP1 site is present). (+) Indicates PMA stimulation and (7) indicates without PMA treatment

spectra are dierent (Sprecher et al., 1994; Peterson et al., 1996a,b). In endothelial cells, most TFPI-2 is associated with ECM (Iino et al., 1998). The TFPI-2 gene is ubiquitous among human tissues, expressed in liver, skeletal muscle, heart, kidney and pancreas; its expression is particularly prominent in placenta (Miyagi et al., 1994). Recent complete nucleotide sequencing of the murine TFPI-2 gene revealed that it consists of ®ve exons and four introns, with the 5'¯anking region containing a TATA box, a GC box, two CAAT boxes, and several transcription factorbinding sites (Kazama et al., 2000). No TATA or CAAT boxes were found immediately upstream of the transcription start site but CAAT boxes were present around *3000 bp upstream of that site (Kazama et al., Oncogene


926

NE

+

+

+

+ + +

+

PMA

–

+

–

+

– +

–

pTF6a

+

+

–

–

–

–

+

+

pTF6b

–

–

+

+

–

–

–

–

pTF6

–

–

–

–

+ +

–

cold probe

–

–

–

–

–

+

1

2

3

4

NE-SNB19

+

–

–

–

NE-Hs683

–

+

+

+

Probe

+

+

+

+

Cold Probe

–

–

–

SP1

–

–

+

–

Super Shift SP1

Figure 6 Binding of Hs683 and SNB19 nuclear extracts (NE) with TFPI-2 promoter-speci®c SP1 oligonucleotides. Doublestranded oligonucleotides (50 fmol) were incubated with nuclear extracts (4 mg) in the presence of SP1 antibody (2 mg) or a 100fold molar excess of unlabeled cold probes in Hs683 cells. Greater amounts of the DNA ± protein complex were evident in the noninvasive Hs683 cell line than the invasive SNB19 cell line

2000). TATA-less promoters are often activated by SP1 and depend on the presence of clusters of SP1 binding sites near the transcription start site (Blake et al., 1990; Lee et al., 1992; Lu et al., 1994; Baker et al., 1996; Huber et al., 1998). In this respect, TFPI-2 seems to be typical of a TATA-less promoter, as the 312-bp fragment of human TFPI-2 (7312 to +1); pTF6 with high TFPI-2 promoter activity contained a cluster of SP1 binding sites (Figure 4). The size and organization of the human TFPI-2 gene is similar to that of the murine TFPI-2 gene (Kamei et al., 2001). Several possible locations have been identi®ed as being the TFPI-2 transcription initiation site, i.e., 75 bp upstream from the translation start site (Kamei et al., 2001) 37 bp upstream from the 5'-terminal sequence (Sprecher et al., 1994) and 19 bp upstream of this Oncogene

Figure 7 Binding of Hs683 nuclear extracts (NE) with TFPI-2 promoter-speci®c oligonucleotides for three AP1 sites. Nuclear extracts were made from cells that had or had not been treated with PMA. Double-stranded oligonucleotides (50 fmol) were incubated with nuclear extracts (4 mg) in the presence of 100fold molar excess of unlabeled cold probes in Hs683 cells. Binding was higher for all three AP1 sites when the cells had been stimulated with PMA

sequence (Miyagi et al., 1994). The 535-bp 3'-¯anking region contains two probable polyadenylation sites (AATAAA), one at +4297 and the other at +4314 (Huber et al., 1998). These results suggested that no additional exons are present in the 5'- and 3'-¯anking regions of TFPI-2. The present study used the translation start site (Sprecher et al., 1994) and a minimal promoter activity was observed at 7312 to 781. In the Kamei et al., 2001 study, the transcriptional initiation site (TIS) was 75 bp 5' away from this site. Using this TIS, in our study promoter activity would be equivalent to 7246 to 75, which is


essentially identical to what was observed in the Kamei study looking at expression in transformed endothelial cells. The previous study showed minimal promoter between 7224 and 7139 (89 bp) (Kamei et al., 2001). We chose human glioma cell lines for our study. Hs683 cell line is non-invasive and does not form tumors in animal models and it has very high expression of TFPI-2 protein and mRNA. Hs683 is less invasive in matrigel and spheroid invasion assays and which is similar to low-grade glioma biopsies in spheroid assays. In contrary, SNB19 cell line is highly invasive in both matrigel and spheroid assays and form tumors in animal models and which has undetectable expression TFPI-2 protein and mRNA. Over expression or down regulation of TFPI-2 showed decreased/ increased invasion in these cell lines (Konduri et al., 2001a). The same cell lines were used for our transient transfections for luciferase assays. Initially, we ampli®ed a 3.6-kb (pTF1) and 1.2-kb (pTF3) genomic fragments from the genomic DNA of Hs683 cells and found luciferase activity in the 3.6-kb fragment; later, we used that 3.6-kb fragment as a template and developed several deletion constructs with various 5'ends between 73601 and 781, and a common 3'-end at +1 (where the translational start site begins). Each of these plasmids were transfected into non-invasive (Hs683) and highly invasive (SNB19) human glioma cell lines. We found that the pTF5 construct (7670 to +1) produced slightly higher luciferase expression in the non-invasive cell line and that the highly invasive cell line expressed very little luciferase activity regardless of the transfection condition (Figure 2a,b). Further study is needed to determine precisely how the TFPI-2 transcription is aected in invasive gliomas. In our study, the pTF6 (7312 to +1) and the pTF2 (71511 to +1) produced activity close to that produced by pTF5 (7670 to +1). The 1.2-kb pTF3 construct (71181 to 7927) apparently included a strong repressor, as no luciferase was expressed in this clone. Upstream of that region, however, positive elements or enhancers were found, as evidenced by luciferase expression in the pTF2 construct (71511 to +1). These positive elements masked the silencer eect. Kamei and colleagues (Kamei et al., 2001) recently reported that transfected human transformed bone marrow micro vascular endothelial cells and JEG-3 choriocarcinoma cells both expressed TFPI-2, which is consistent with the ability of these cells to synthesize and secrete TFPI-2 antigen (Udagawa et al., 1998; Iino et al., 1998). This group also found silencer elements between nucleotides 72372 to ± 982 (Kamei et al., 2001). In contrast, our results indicated the presence of a strong repressor at the 5'-end between nucleotides at

71181 to 7927, followed by a positive regulator between nucleotides at 71511 to 71181. pTF8 construct, (781 to +1) has no activity, which is like a negative control (empty vector). It has been reported that PMA induces the synthesis and secretion of TFPI proteins in several cell types (Rao et al., 1995a,b; 1998) and analysed mRNA by Northern blotting and protein by Western blotting. Recently it has been shown that PMA increased VEGF protein in human bladder transitional carcinoma cells (Chabannes et al., 2001) and PMA also induced the activation of PKC z during monocytic dierentiation of HL-60 cells (Kim et al., 2001). Finally, we found that expression of the pTF6 region (7312 to +1) could be induced or activated by PMA in Hs683 cells (Figure 3a) but not in SNB19 (Figure 3b). Hence, we conclude that the 231-bp region between 7312 and 781 contains the minimal and inducible promoter sequence for the expression of the human TFPI-2 gene. A previous study showed that the minimal promoter region was in the 89-bp region between 7224 and 7139 (Kamei et al., 2001). In our study inducible regulation was observed dierentially by PMA in Hs683 cells (Figure 5) and more SP1 promoter speci®c DNA protein complex was observed in non-invasive cell line (Hs683) compared to highly invasive cell line (SNB19) (Figure 6). When EMSA's were conducted for AP1, we observed more DNA protein complex at all three AP1 sites, when cells treated with PMA compared to untreated Hs683 cells (Figure 7). The potential transcription factor binding sites of the 5'-¯anking region that could regulate the gene expression of human TFPI-2 using BIMAS, NCBI, NIH revealed cluster of SP1 and three AP1 sites Lyf-1 and NF-kB were observed in the minimal and inducible promoter sequence (Figure 4). The importance of these factors needs to be addressed in future studies.

927

Abbreviations TFPI-2, tissue factor pathway inhibitor-2; PMA, phorbol myristate acetate; kb, kilobase pair(s); bp, base pair(s); PCR, polymerase chain reaction; SP1, stimulatory protein 1; AP1, activator protein-1; NE, nuclear extract; EMSA, electrophoretic mobility shift assay

Acknowledgments The nucleotide sequence reported in this paper has been submitted to GenBank with accession number AF266279. Supported by National Cancer Institute grants, CA-76350, CA-75557 and CA75692 (to JS Rao) and HL-64119 (to W Kisiel).

References Bajaj MS, Kuppuswamy MN, Saito H, Spitzer SG and Bajaj SP. (1990). Proc. Natl. Acad. Sci. USA, 87, 8869 ± 8873. Baker DL, Dave V, Reed T and Periasamy M. (1996). J. Biol. Chem., 271, 5921 ± 5928.

Blake MC, Jambou RC, Swick AG, Kahn JW and Azizkhan JC. (1990). Mol. Cell. Biol., 10, 6632 ± 6641. Butzow R, Huhtala ML, Bohn H, Virtanen I and Seppala M. (1988). Biochem. Biophys. Res. Commun., 150, 483 ± 490. Oncogene


928

Oncogene

Chabannes E, Fauconnet S, Bernardini S, Wallerand H, Adessi G and Bittard H. (2001). Cell. Signal., 13, 585 ± 591. Chaturvedi MM, LaPushin R and Aggarwal BB. (1994). J. Biol. Chem., 269, 14575 ± 14583. Chen L, Mao SJ and Larsen WJ. (1992). J. Biol. Chem., 267, 12380 ± 12386. Enjyoji K, Emi M, Mukai T, Imada M, Leppert ML, Lalouel JM and Kato H. (1993). Genomics, 17, 423 ± 428. Girard TJ, Eddy R, Wesselschmidt RL, MacPhail LA, Likert KM, Byers MG, Shows TB and Broze Jr GJ. (1991). J. Biol. Chem., 266, 5036 ± 5041. Gomez DE, Alonso DF, Yoshiji H and Thorgeirsson UP. (1997). Eur. J. Cell. Biol., 74, 111 ± 122. Hascall VC, Heinegard DK and Wight TN. (1991). Cell Biology of Extra Cellular Matrix: New York: Plenum Press. Huber R, Schlessinger D and Pilia G. (1998). Gene, 214, 35 ± 44. Iino M, Foster DC and Kisiel W. (1998). Arterioscler. Thromb. Vasc. Biol., 18, 40 ± 46. Kamei S, Kazama Y, Kuijer JL, Foster DC and Kisiel W. (2001). Biochimica. Et. Biphysica. Acta., 1517, 430 ± 435. Kazama Y, Kamei S, Kuijper JL, Foster DC and Kisiel W. (2000). Thromb. Haemost., 83, 141 ± 147. Kim M, Lim W, Cha J, An N, Yoo S, Park J, Kim H and Lee Y. (2001). Cancer. Lett., 171, 79 ± 85. Konduri SD, Rao CN, Nirmala C, Anastasia T, Sanjeeva M, Kin Y, Lakka SS, Dinh D, Olivero WC, Gujrati M, Foster DC, Kisiel W and Rao JS. (2001a). Oncogene, 20, 6938 ± 6945. Konduri SD, Tasiou A, Chandrasekar N and Rao JS. (2001b). Int. J. Oncology., 18, 127 ± 131. Konduri SD, Tasiou A, Chandrasekar N, Nicolson GL and Rao JS. (2000). Clin. Exp. Metastast., 18, 303 ± 308. Korner G, Bjornsson TD and Vlodavsky I. (1993). J. Cell. Physiol., 154, 456 ± 465. Lakka SS, Konduri SD, Mohanam S, Nicolson GL and Rao JS. (2000). Clin. Exp. Metastast., 18, 239 ± 244. Lee JK, Tam JW, Tsai MJ and Tsai SY. (1992). J. Biol. Chem., 267, 4638 ± 4645. Lu J, Lee W, Jiang C and Keller EB. (1994). J. Biol. Chem., 269, 5391 ± 5402. McGuire PG and Seeds NW. (1989). J. Cell. Biochem., 40, 215 ± 227.

Mimuro J, Schleef RR and Loskuto DJ. (1987). Blood, 70, 721 ± 728. Miyagi Y, Koshikawa N, Yasumitsu H, Miyagi E, Hirahara F, Aoki I, Misugi K, Umeda M and Miyazaki K. (1994). J. Biochem., 116, 939 ± 942. Miyagi Y, Yasumitsu H, Eki T, Miyata S, Koshikawa N, Hirahara F, Aoki I, Misugi K and Miyazaki K. (1996). Genomics, 35, 267 ± 268. Nagase H and Woessner Jr JF. (1999). J. Biol. Chem., 274, 21491 ± 21494. Petersen LC, Sprecher CA, Foster DC, Blumberg H, Hamamoto T and Kisiel W. (1996a). Biochemistry, 35, 266 ± 272. Petersen LC, Bjorn SE, Olsen OH, Nordfang O, Norris F and Norris K. (1996b). Eur. J. Biochem., 235, 310 ± 316. Rao CN, Cook B, Liu Y, Chilukuri K, Stack SM, Foster DC, Kisiel W and Woodley DT. (1998). Int. J. Cancer, 76, 749 ± 759. Rao CN, Gomez DE, Woodley DT and Thorgeirsson UP. (1995a). Arch. Biochem. Biophys., 319, 55 ± 62. Rao CN, Liu YY, Peavey CL and Woodley DT. (1995b). Arch. Biochem. Biophys., 317, 311 ± 314. Rao CN, Peavey CL, Liu YY, Lapiere JC and Woodley DT. (1995c). J. Invest. Dermatol., 104, 379 ± 383. Rao CN, Lakka SS, Kin Y, Konduri SD, Fuller G, Mohanam S and Rao JS. (2001). Clin. Cancer. Res., 7, 570 ± 576. Schreiber E, Matthias P, Muller MM and Schaner W. (1989). Nucleic. Acids. Res., 17, 6419. Sprecher CA, Kisiel W, Mathewes S and Foster DC. (1994). Proc. Natl. Acad. Sci. USA, 91, 3353 ± 3357. Udagawa K, Miyagi Y, Hirahara F, Miyagi E, Nagashima Y, Minaguchi H, Misugi K, Yasumitsu H and Miyazaki K. (1998). Placenta, 19, 217 ± 223. Vlodavsky I, Folkman J, Sullivan R, Fridman R, IshaiMichaeli R, Sasse J and Klagsbrun M. (1987). Proc. Natl. Acad. Sci. USA, 84, 2292 ± 2296. Vukicevic S, Kleinman HK, Luyten FP, Roberts AB, Roche NS and Reddi AH. (1992). Exp. Cell. Res., 202, 1 ± 8. Werling RW, Zacharski LR, Kisiel W, Bajaj SP, Memoli VA and Rousseau SM. (1993). Thromb. Haemost., 69, 366 ± 369.