INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Transforming growth factor- (TGF-)¹ belongs to a superfamily that includes activins, inhibins, bone morphogenetic proteins, and müllerian-inhibiting substances (1, 2). TGF- plays an important role in cellular proliferation, differentiation, and synthesis of extracellular matrix proteins (1, 2). Three major TGF--binding proteins have been identified. They are referred to as type I (RI), type II (RII), and type III (RIII). RI and RII are glycoproteins of 53 and 75 kDa, respectively, whereas RIII is a 280-330-kDa proteoglycan (3). RI and RII are serine/threonine kinases and form a hetero-oligomeric complex that is required for the TGF--mediated signaling cascade (4-7). RIII lacks a signaling motif, and its role appears to be limited to presenting TGF- to the signaling receptors (8).

One of the important effects of TGF- is the inhibition of growth of epithelial cells as well as some cancer cells. Because RI and RII are both required for TGF--mediated growth suppression, loss of either receptor may contribute to TGF- resistance and subsequent malignant progression. TGF- resistance caused by defects in RII expression has been reported in various cell lines (9-11). Previous work has indicated an association between defective RII expression and malignant progression of several cell types (9, 10, 12-14) including breast carcinoma cells (15). RII replacement in breast and colon carcinoma cells restored TGF- response and reduced malignant behavior (12, 15). It has also been demonstrated that exogenous RI expression in an RI-defective colon carcinoma cell line reversed malignancy (16). These studies underline the importance of both RI and RII as tumor suppressors. Loss of RII expression was observed in gastric cancer cells as well as a subset of colon cancer cells in association with deletions or gene mutations (10, 17). RII repression caused by decreased binding of nuclear proteins to the positive regulatory elements of the RII promoter has been shown to cause TGF- resistance of adenovirus E1A-transformed mouse keratinocytes (18).

Breast cancer cell lines that express estrogen receptor (ER⁺) are refractory to TGF- effects, whereas estrogen receptor-negative (ER) cells are often TGF--sensitive (19). Loss or undetectable expression of RII has been reported to contribute to TGF- resistance in ER⁺ breast cancer cells (11, 15). Several different ER⁺ MCF-7 strains have been reported in the literature. Comparison of ER⁺ MCF-7 early (MCF-7E) and MCF-7 late (MCF-7L) passage cells from our laboratory has shown that MCF-7E cells express RII and are TGF--responsive, but MCF-7L cells lack RII and are TGF--resistant, suggesting possible defects at the transcriptional or post-transcriptional level (20). A transiently transfected RII promoter element exhibited markedly decreased activity in the MCF-7L cells compared with MCF-7E cells, pointing toward a possible defect in transcription.

Gene inactivation caused by methylation of CpG sites in the vicinity of promoter regions has long been associated with tissue-specific and developmentally regulated genes (21). However, recent studies have cited gene methylation as a mode of inactivation of several genes including some that are involved in cell cycle control (14). DNA methyltransferase inhibitors 5-azacytidine and 5-aza-2'-deoxycytidine (5-aza-2'-dC) are the agents used most frequently to reverse methylation and reconstitute the expression of these genes (22).

To delineate the mechanism of RII repression in the MCF-7L and other ER⁺ breast cancer cells, we have carried out studies using 5-aza-2'-dC and now provide evidence that RII expression is low or undetectable because of suboptimal activity of Sp1 transcripton factor. Treatment with 5-aza-2'-dC leads to increased Sp1 steady-state levels as a result of increased protein stability and, consequently, concomitant induction of RII expression. RII expression resulted in restoration of TGF- sensitivity. These results shed light on a novel mechanism by which epithelial cells escape negative regulatory effects of TGF- leading to uncontrolled growth and hence tumor formation and progression.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Cell Culture-- All of the breast cancer cell lines used were obtained from American Type Culture Collection (ATCC). The BT20 strain in our laboratory has a constitutively active mutated estrogen receptor, hence we refer to it as ER⁺. Cells were grown in McCoy's 5A medium supplemented with 10% fetal bovine serum (Sigma), amino acids, antibiotics, pyruvate, and vitamins (Life Technologies, Inc.). Cultures were maintained at 37 °C in a humidified atmosphere of 5% CO₂. For experiments in which 5-aza-2'-dC was used, cells were seeded at a density of 4 × 10⁵ cells/10-cm culture dish (day 0). 5-Aza-2'-dC (Sigma) was added to the growth medium (1 or 2 µg/ml as indicated) in two 24-h pulses on days 2 and 5. Cells were used on day 7 for RNA determinations, transfections, or isolation of nuclear extracts for electrophoretic mobility shift assays (EMSAs) and Sp1 Western immunoblots.

RNA Analysis-- Total RNA from the breast cancer cells was extracted by guanidine thiocyanate homogenization and ultracentrifugation through a cesium gradient as described previously (23). RI, RII, and actin riboprobes were described previously (15). RNase protection assays were performed as described previously (24). Briefly, radioactive riboprobes were allowed to hybridize overnight with the RI and RII mRNA in 40 µg of total RNA. After RNase A and T₁ treatment, the protected double-stranded RNA fragments were analyzed by urea-PAGE and visualized by autoradiography. Actin was used to normalize sample loading.

Receptor Cross-linking-- Simian TGF-1 was purified as described (25) and iodinated by the chloramine-T method (26). Cells were seeded at a density of 6 × 10⁴/well in a six-well plate (day 0). Wherever indicated, 1 or 2 µg of 5-aza-2'-dC was added in two 24-h pulses on days 2 and 5, and receptor binding studies were carried out on day 7 using 200 pM ¹²⁵I-TGF- as described previously (27). Labeled cells were solubilized in 200 µl of 1% Triton X-100 with 1 mM phenylmethylsulfonyl fluoride. Equal amounts of cell lysate protein were separated by 4-10% gradient SDS-PAGE under reducing conditions and exposed for autoradiography.

Luciferase Assay-- The TGF--responsive plasminogen activator inhibitor promoter-luciferase reporter construct (p3TP-Lux) was used to determine TGF- sensitivity as described previously (5). The p21 promoter in which the TGF--responsive element was deleted (p21/WAF1/CIP1/smaI) was used as a control (28). 5-Aza-2'-dC-treated and untreated MCF-7L cells were transiently transfected with 30 µg of p3TP-Lux and 5 µg of -galactosidase plasmid by electroporation with a Bio-Rad gene pulser at 250 mV and 960 microfarads. The electroporated cells were plated into six-well tissue culture plates. Cells were grown for 24 h and then treated with 5 or 10 ng/ml TGF-1 for 24 h. Cells were harvested in 200 µl of lysis buffer (luciferase assay system, Promega) 48 h after transfection. In the first 10 s after the addition of substrate, luciferase activity was measured using a luminometer (Berthold lumat LB 9501) and was expressed as relative units after normalization to -galactosidase activity.

Chloramphenicol Acetyltransferase (CAT) Assay-- RII and insulin-like growth factor II (IGF-II) promoter-CAT constructs were described previously (29, 30). The RII construct (274/+50) containing the core promoter with two Sp1 sites and two enhancer regions (PRE1 and PRE2), a 47 RII-CAT (wild type Sp1 site), a 47 Spm RII-CAT (mutated Sp1 site), and IGF-II constructs (58/+124) containing a distinct TATA box in combination with either two wild type Sp1 (58 IGF-II-CAT) or two mutated Sp1 sites (58 Spm IGF-II-CAT) sites were used in this study. MCF-7L cells that were untreated or treated with 1 µg/ml 5-aza-2'-dC, respectively, were transiently transfected with 30 µg of 274/+50 RII promoter-CAT construct or 58/+124 IGF-II promoter-CAT construct by electroporation with a Bio-Rad gene pulser at 250 mV and 960 microfarads. For normalization of transfection efficiency, 5 µg of Rous sarcoma virus -galactosidase was cotransfected into the cells. Cells were plated into 10-cm Petri dishes, and 48 h later cells were harvested to carry out the standard -galactosidase (31) and CAT assays (32). -Galactosidase was analyzed using a molecular dynamics microtiter plate reader and the Softmax software package. Results from CAT assays were analyzed by TLC, and the TLC plate was quantitated directly using an Ambis system as well as by autoradiography.

EMSA-- Double-stranded oligonucleotides representing the two Sp1 sites (25 bp and 143 bp relative to start site), the two positive regulatory regions (PRE1 and PRE2), and a mutant Sp1 oligonucleotide (18, 29) were custom designed and obtained from Genosys. The oligonucleotides were end labeled using [-³²P]ATP (50 µCi at 3,000 Ci/mmol) and T₄ polynucleotide kinase. The labeled oligonucleotides were purified using probe Quant^TM G-50 microcolumns (Amersham Pharmacia Biotech). Binding reactions were performed using 3 µg of nuclear extracts, buffer (10 mM Tris, pH 7.5, 50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, and 5% glycerol), 2 µg of poly(dI-dC), and 10,000 cpm of ³²P-labeled oligonucleotide in a volume of 20 µl. Reactions were incubated at room temperature for 15 min. Competition reactions were performed by adding an unlabeled double-stranded oligonucleotide to the reaction mixture. Samples were analyzed by nondenaturing 4% polyacrylamide gel at 150 V for 1.5 h in 100 mM Tris borate-EDTA buffer. Gels were vacuum dried and analyzed by autoradiography. For the supershift assay, the ³²P-labeled oligonucleotide plus nuclear extract was incubated further at room temperature for 15 min with 2 µg of Sp1 antibody (anti-rabbit, Santa Cruz) before electrophoresis and autoradiography.

Northern Analysis-- Total RNA from the control and 5-aza-2'-dC-treated MCF-7L cells was extracted by guanidine thiocyanate homogenization and ultracentrifugation through a cesium gradient as described previously (23). Total RNA (10 µg) was fractionated on 1.2% agarose gel containing formaldehyde and transferred to nitrocellulose membrane. Prehybridization and hybridization were performed at 65 °C using RIPA buffer (Amersham Pharmacia Biotech). The cDNA probe for Sp1 was labeled with [³²P]dCTP (>3,000 Ci/mmol; Amersham Pharmacia Biotech) using a random primed DNA labeling kit (Boehringer Mannheim).

Western Immunoblot Analysis of Sp1-- Nuclear extracts (4 µg) from 5-aza-2'-dC-treated and untreated MCF-7L cells were resolved using 7.5% SDS-PAGE and transferred to nitrocellulose membranes by wet electrophoretic transfer (Bio-Rad). Nonspecific binding was blocked with 5% non-fat milk in TTBS (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Tween 20) overnight at 4 °C. The membrane was probed with a rabbit anti-human Sp1 polyclonal antibody (Santa Cruz) in the same buffer for 1 h at room temperature and washed three times with TTBS for 10 min each. Bound antibodies were detected with an anti-rabbit peroxidase-conjugated IgG and an enhanced chemiluminescence detection system (Amersham Pharmacia Biotech). The membrane was reprobed with rabbit anti-human c-Jun polyclonal antibody (Santa Cruz). The MCF-7L cells were treated with 10 µg/ml cyclohexamide and harvested at the indicated time points for the Sp1 protein stability studies (33).

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Expression of TGF- Receptors-- Previous studies indicate that ER⁺ breast cancer cells express low RII levels and hence are TGF--resistant (11, 15). However, several different strains of various ER⁺ breast cancer cell lines have been reported which exhibited differential sensitivities to TGF-. Hence we have screened for RI and RII transcripts in TGF--resistant ER⁺ breast cancer cell lines (MCF-7L, BT20, ZR75, T47D) using an RNase protection assay. The TGF--sensitive ER cell line MDA MB 231 was used as a positive control. All of the ER⁺ cell lines expressed RI mRNA, but RII mRNA was undetectable (Fig. 1). The ER MDA MB 231 cell line has both RI and RII transcripts.

View larger version (61K): [in this window] [in a new window]   Fig. 1.   Expression of TGF- RI and RII transcripts in the ER+ breast cancer cells. Riboprobes for RI and RII were incubated with total RNA (40 µg) from confluent cells. The resulting hybridized RNA was analyzed by RNase protection assay as described under "Experimental Procedures." The ER MDA MB 231 cell line was used as a positive control. Actin controls are shown for normalization of sample loading. First four lanes from left, ER+ cells; far right lane, MDA MB 231 cells.

View larger version (73K): [in this window] [in a new window]   Fig. 2.   Cell surface expression of TGF- receptors in ER+ breast cancer cells. Receptor cross-linking assays were performed to detect the cell surface expression of RI and RII. Confluent cells were incubated with 200 pM 125I-TGF-1, cross-linked with 0.3 mM disuccinimidyl suberate, and separated by 4-10% gradient SDS-PAGE under reducing conditions. Third through sixth lanes from left, ER+ cell lines; far right lane, ER cell line; far left lane, mink lung cells. Binding specificity was demonstrated by competing with a 50-fold excess of cold TGF-1 (second lane from left).

Effect of 5-Aza-2'-dC on the Induction of RII-- RNase protection assays were performed to determine whether treatment with the DNA methyltransferase inhibitor 5-aza-2'-dC leads to expression of RII transcript. Accumulation of RII transcript was observed in all of the ER⁺ cell lines (Fig. 3). To examine whether RII expression permitted TGF- binding to RI, cell surface receptor binding studies with ¹²⁵I-TGF-1 were carried out. The data show that expression of RII resulted in the cell surface binding of TGF- to both RII and RI (Fig. 4). Increased binding for RIII was also noted. This phenomenon has also been observed after RII transfection in previous studies (12, 15). Binding specificity was demonstrated by competing with a 50-fold excess of TGF-1.

View larger version (86K): [in this window] [in a new window]   Fig. 3.   Expression of RII mRNA after 5-aza-2'-dC treatment. Total RNA was isolated from control and 5-aza-2'-dC-treated breast cancer cells as described under "Experimental Procedures." RII riboprobe was incubated with 40 µg of total RNA, and the hybridized RNA was analyzed by RNase protection assay as described under "Experimental Procedures." Actin RNA levels were used for normalization.

View larger version (46K): [in this window] [in a new window]   Fig. 4.   Detection of cell surface RI and RII in 5-aza-2'-dC-treated breast cancer cells. Breast cancer cells were treated as described under "Experimental Procedures." Cells were incubated with 200 pM 125I-TGF-1, cross-linked with 0.3 mM disuccinimidyl suberate, and separated by 4-10% SDS-PAGE under reducing conditions. Binding specificity was demonstrated by a 50-fold excess of cold TGF-1 (lane 4).

Responsiveness to TGF-1-- To evaluate whether 5-aza-2'-dC-mediated RII expression in MCF-7L cells restored TGF- sensitivity, TGF--dependent promoter activity was analyzed using the p3TP-Lux plasminogen activator inhibitor promoter-TGF--responsive element in tandem with a luciferase reporter. The TGF--responsive promoter-luciferase construct or a p21/WAF1/CIP1 control promoter lacking the TGF--responsive element (p21/WAF1/CIP1/smaI) was transiently transfected into 5-aza-2'-dC-treated or untreated MCF-7L cells. 24 h after transfection, cells were subjected to treatment with 5 or 10 ng/ml TGF-1 for an additional 24 h, and luciferase activity was then determined (5). TGF- had no effect on the control cells, but 5-aza-2'-dC-treated cells showed 2- and 5-fold increases in luciferase activity at 5 or 10 ng/ml TGF-1, respectively (Fig. 5).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Restoration of TGF- response after 5-aza-2'-dC treatment. 5-Aza-2'-dC-treated and untreated MCF-7L cells were transiently transfected with a TGF--responsive promoter construct (p3TP-Lux) and a TGF--nonresponsive p21 promoter-reporter construct, p21/smaI (43). At 24 h post-transfection cells were treated for an additional 24 h with 5 or 10 ng/ml TGF-1. Cells were then harvested, and luciferase activity was measured and presented as relative units after normalizing for the -galactosidase activity. a and b are MCF-7L cells transfected with TGF--responsive promoter construct, and c, those transfected with TGF--nonresponsive promoter construct.

RII Promoter Activity-- Southern analysis of the RII promoter region did not reveal any methylated sites (data not shown). We have compared the RII promoter activity in control and 5-aza-2'-dC-treated MCF-7L cells using an RII promoter-CAT construct (Fig. 6A) to examine whether the induction of RII after 5-aza-2'-dC treatment is caused by the activation of a transactivator. This promoter element contains two Sp1 binding sites (at 25 bp and 143 bp relative to the start site) and two positive regulatory elements (PRE1, 219 bp/172 bp; PRE2, +1 bp/+50 bp relative to the start site). 5-Aza-2'-dC-treated MCF-7L cells showed higher promoter activity than control cells, thus suggesting the activation of nuclear proteins that bind to the RII promoter and enhance its activity (Fig. 6B).

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Panel A, schematic of the 274 RII promoter-CAT construct. The construct contains two Sp1 binding sites and two positive regulatory elements (PRE1 and PRE2). Panel B, RII promoter-CAT activity in the 5-aza-2'-dC-treated MCF-7L cells. The RII promoter-CAT construct was transiently transfected into the control and 5-aza-2'-dC-treated MCF-7L cells as described under "Experimental Procedures." 48 h after transfection, cells were harvested, normalized for -galactosidase activity, and CAT assays were performed (see "Experimental Procedures").

EMSAs-- To narrow the identity of the proteins that are enhancing the RII promoter activity in the 5-aza-2'-dC-treated MCF-7L cells, EMSAs were performed using nuclear extracts and the ³²P-labeled oligonucleotides corresponding to the two Sp1 binding sites as well as the two positive regulatory elements (PRE1 and PRE2), which have been recognized previously (18). Oligonucleotides corresponding to the wild type RII promoter PRE1, PRE2, and the two Sp1 sites as well as a mutated Sp1 oligonucleotide were analyzed for binding to nuclear proteins (Figs. 7A and 8A). Gel shift analysis of nuclear extracts from 5-aza-2'-dC-treated and control MCF-7L cells do not reveal any differences in the proteins that bind to PRE1 and PRE2 (Fig. 7B). Cold oligonucleotides competed with ³²P-labeled PRE oligonucleotides for binding to the protein-DNA complexes (Fig. 7B). However, there was enhanced binding of nuclear proteins from the 5-aza-2'-dC-treated MCF-7L cells to both of the ³²P-labeled Sp1 oligonucleotides compared with untreated control cells (Fig. 8B). The low mobility complexes that were present in the 5-aza-2'-dC-treated extracts were absent in the control nuclear extracts. The mobilities of these complexes were similar to that of recombinant human Sp1 (Promega)-bound ³²P-labeled Sp1 oligonucleotide complex, which was used as a positive control. Wild type Sp1 oligonucleotides competed with ³²P-labeled Sp1 oligonucleotides for binding to the protein complexes, whereas the mutant Sp1 oligonucleotide could not, thus indicating the specificity of the shifts. To confirm further that the enhanced protein-DNA complex contains Sp1, supershift assays were carried out by incubating the protein-DNA complexes with 2 µg of Sp1 antibody (described under "Experimental Procedures"). Sp1 antibody recognized the Sp1 in the protein-DNA complexes, resulting in a clear shift of the mobility of the protein bound to the ³²P-labeled Sp1 oligonucleotides (Fig. 9).

View larger version (38K): [in this window] [in a new window]   Fig. 7.   Panel A, sequences of the sense strands of double-stranded PRE1 and PRE2 oligonucleotides used for the gel shift analysis. Panel B, binding of nuclear proteins to positive regulatory elements, PRE1 and PRE2. EMSA was performed using 32P-labeled double-stranded PRE1 and PRE2 oligonucleotides by incubating with 3 µg of nuclear extracts from control and 5-aza-2'-dC-treated MCF-7L cells. Cold PRE1 and PRE2 oligonucleotides were used as competitors.

View larger version (58K): [in this window] [in a new window]   Fig. 8.   Panel A, sequences of the sense strands of the double-stranded Sp1 and mutant (Spm) oligonucleotides used for the gel shift analysis. Panel B, binding of the nuclear proteins to Sp1 oligonucleotides. EMSA was performed using 32P-labeled Sp1 oligonucleotides by incubating with 3 µg of nuclear extracts from the control and 5-aza-2'-dC-treated MCF-7L cells. Wild type and mutant Sp1 oligonucleotides were used as competitors.

View larger version (70K): [in this window] [in a new window]   Fig. 9.   Detection of protein-DNA complexes using Sp1 antibody. To confirm that protein-DNA complexes contained Sp1, supershift assays were carried out by incubating the nuclear extract plus 32P-labeled Sp1 oligonucleotide complexes for 15 min at room temperature with 2 µg of Sp1 antibody. A lower mobility complex resulting from the binding of Sp1 antibody to protein-DNA complexes was observed.

Effect of Sp1 Mutation on RII Promoter Activity-- To confirm the role of Sp1 in the enhanced RII promoter activity, the 47 RII promoter-CAT construct containing either the wild type or mutant Sp1 site was transiently transfected into control and 5-aza-2'-dC-treated MCF-7 cells (Fig. 10A). 5-Aza-2'-dC-treated MCF-7L cells expressed higher wild type RII promoter activity compared with control untreated MCF-7L cells. The Sp1 mutant RII promoter-CAT construct was not expressed (Fig. 10B).

View larger version (17K): [in this window] [in a new window]   Fig. 10.   Panel A, schematic of wild type/mutated 47 RII promoter-CAT construct. The construct contains one Sp1 binding site. Wild type and mutated sequences are indicated. Panel B, wild type and mutated RII promoter-CAT activity in 5-aza-2'-dC-treated MCF-7L cells. The 47 RII promoter-CAT construct with wild type or mutated Sp1 sites was transiently transfected into control or 5-aza-2'-dC-treated MCF-7L cells as described under "Experimental Procedures." 48 h after transfection, cells were harvested, normalized for -galactosidase activity, and CAT assays were performed (see "Experimental Procedures").

IGF-II Promoter Activity-- To determine if 5-aza-2'-dC-mediated Sp1 activity leads to enhanced expression of other Sp1-dependent promoters, we transiently expressed either the wild type or Sp1 site-mutated IGF-II promoter-CAT constructs (Fig. 11A) in 5-aza-2'-dC-treated or untreated MCF-7L cells. The IGF-II promoter contains a distinct TATA box and two Sp1 sites (30). 5-Aza-2'-dC-treated MCF-7L cells exhibited a significantly higher IGF-II promoter activity compared with untreated control cells. The 5-aza-2'-dC-induced IGF-II promoter activity disappeared but retained the basal promoter activity when the Sp1 sites were mutated (58 Spm IGF-II-CAT). These data further confirm the activation of Sp1 by 5-aza-2'-dC in the MCF-7L cells (Fig. 11B).

View larger version (28K): [in this window] [in a new window]   Fig. 11.   Panel A, schematic of IGF-II promoter-CAT constructs. The promoter element contains a TATA box and two Sp1 recognition sequences (shown in bold). Panel B, IGF-II promoter-CAT activity in the 5-aza-2'-dC-treated MCF-7L cells. IGF-II promoter-CAT constructs were transiently transfected into the control and 5-aza-2'-dC-treated MCF-7L cells as described under "Experimental Procedures." 48 h after transfection, cells were harvested, normalized for -galactosidase activity, and CAT assay was performed (see "Experimental Procedures").

Effect of 5-Aza-2'-dC on Sp1 Transcription-- To determine if 5-aza-2'-dC treatment induces Sp1 transcript, Northern analysis using Sp1 cDNA as a probe was performed. Control and 5-aza-2'-dC-treated MCF-7L cells both expressed similar levels of Sp1 mRNA (Fig. 12), suggesting that demethylation was not affecting Sp1 transcription.

View larger version (21K): [in this window] [in a new window]   Fig. 12.   Expression of Sp1 transcript in control and 5-aza-2'-dC-treated MCF-7L cells. Total RNA was isolated from control and 5-aza-2'-dC-treated MCF-7L cells as described under "Experimental Procedures," and 10 µg RNA from each sample was resolved on a formaldehyde/agarose gel. RNA was transferred to nitrocellulose membranes and probed with [32P]dCTP-labeled Sp1 cDNA probe.

Expression of Sp1 Protein-- To determine whether 5-aza-2'-dC treatment stimulates the Sp1 protein expression, Western immunoblot was performed using 4 µg each of nuclear extracts from 5-aza-2'-dC-treated and untreated MCF-7L cells. Western analysis showed two protein species of 95 and 105 kDa. The two species are the result of differential post-translational modification of the Sp1 polypeptide (34, 35). Significant increases of both the species were observed in the nuclear extracts of 5-aza-2'-dC-treated MCF-7L cells, whereas there was no change in the c-Jun levels (Fig. 13). Sp1 protein stability studies after treatment with 10 µg/ml cyclohexamide were performed as described previously for this protein (33) to determine if 5-aza-2'-dC stabilizes Sp1 protein indirectly. Sp1 protein from 5-aza-2'-dC-treated MCF-7L cells showed enhanced stability compared with untreated MCF-7L control cells (Fig. 14), thus indicating that the increased steady-state levels of Sp1 protein and oligonucleotide binding activity were the result of indirect actions of the demethylating agent on Sp1.

View larger version (12K): [in this window] [in a new window]   Fig. 13.   Western immunoblot analysis of Sp1. Nuclear extracts (4 µg) from the control and 5-aza-2'-dC-treated MCF-7L cells were resolved on a 7.5% SDS-PAGE, transferred to nitrocellulose membranes, and probed with rabbit anti-human Sp1 and c-Jun polyclonal antibodies. Sp1 antibody recognizes two protein species of 95 and 105 kDa, which are the result of differential post-translational modifications (40, 41), but c-Jun antibody recognizes a 39-kDa species.

View larger version (15K): [in this window] [in a new window]   Fig. 14.   Sp1 stability in control and 5-aza-2'-dC-treated MCF-7L cells. Control (panel A) and 5-aza-2'-dC-treated MCF-7L cells (panel B) were treated with 10 µg/ml cyclohexamide, and cells were harvested at the indicated time points to isolate nuclear extracts. Nuclear extracts (10 µg) were resolved on a 7.5% SDS-PAGE, transferred to nitrocellulose membrane, and probed with rabbit anti-human Sp1 polyclonal antibody.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

Breast cancer cells that express estrogen receptor (ER⁺) escape negative growth regulation by TGF-, leading to malignant behavior. Previous work from our laboratory (15) and Kalkhoven et al. (11) indicated that ER⁺ breast cancer cells acquire resistance to TGF- because of a lack of or inadequate expression of RII. Replacement of RII in a TGF- resistant ER⁺ MCF-7L cell line restored TGF- response and reduced tumorigenicity in athymic nude mice (15). Reversal of malignancy of a human colon carcinoma cell line was also reported after RII expression (12). Hence, targeting re-expression of RII may offer potential novel approaches for treatment or chemoprevention of breast cancer. In this study, we have examined the ability of the DNA methyltransferase inhibitor 5-aza-2'-dC to restore endogenous RII expression in the ER⁺ breast cancer cells.

Treatment with 5-aza-2'-dC led to induction of RII expression in all three of the ER⁺ breast cancer cell lines examined (Figs. 3 and 4). Significantly, 5-aza-2'-dC-mediated RII induction resulted in restoration of TGF- response in the MCF-7L cells (Fig. 5). However, the induction of RII expression after 5-aza-2'-dC treatment was not found to be a result of the direct demethylation of the RII gene. This raised the possibility of the involvement of increased activation of a transactivator as a cause for enhanced RII transcription.

RII repression resulting from decreased binding of nuclear proteins to the enhancer regions (PRE1 and PRE2) of the RII promoter has been reported in adenovirus E1A-transformed mouse keratinocytes (18). However, 5-aza-2'-dC-treated and untreated MCF-7L cells showed no differences in the nuclear proteins that bind to these enhancer regions (PRE1 and PRE2). The RII promoter lacks a distinct TATA box, and Sp1 has been reported to play an important role in the initiation of transcription from promoters lacking distinct TATA boxes (36). The human RII promoter contains two Sp1 sites at 25 and 143 bp relative to the start site (29). Enhanced RII promoter activity (Fig. 6) as well as the increased binding of nuclear proteins to the ³²P-labeled Sp1 oligonucleotides (Figs. 8B and 9) in the 5-aza-2'-dC-treated MCF-7L cells indicate that increased Sp1 activity was induced by 5-aza-2'-dC. An RII promoter-CAT construct with a mutated Sp1 site was not expressed in the 5-aza-2'-dC-treated MCF-7L cells, further demonstrating the specificity of enhanced Sp1 activity resulting from 5-aza-2'-dC treatment (Fig. 10B). Enhancement of Sp1 levels also led to the increased expression of the Sp1-dependent IGF-II promoter in 5-aza-2'-dC-treated MCF-7L cells (Fig. 11B). Consequently, the results presented in this study indicate that suboptimal activity of Sp1 results in transcriptional repression of RII in the ER⁺ breast cancer cells. This appears to have a role in uncontrolled growth and subsequent malignant progression as evidenced by studies showing that RII replacement reverses malignancy in MCF-7L cells (15).

However, 5-aza-2'-dC studies on MCF-7L cells have raised some interesting questions. A similar drug (5-azacytidine) has been reported to increase Sp1 activity without altering protein expression leading to TGF- transcription in melanoma cells (37). In our study, 5-aza-2'-dC-treated MCF-7L cells did not show any increase in the Sp1 transcript levels (Fig. 12). Thus, Sp1 induction as a result of the demethylation at the Sp1 gene locus was eliminated. However, 5-aza-2'-dC-treated MCF-7L cells exhibited enhanced Sp1 activity as well as increased Sp1 nuclear protein levels (Fig. 13). This may result from the effects of 5-aza-2'-dC at a different gene locus, whose product may be required for Sp1 expression and activity. Sp1 undergoes post-translational modifications such as phosphorylation and glycosylation (34, 35). Glycosylation may aid in nuclear localization and DNA binding, and phosphorylation may assist in stabilizing the Sp1·DNA complex. It has been shown that glycosylation stabilizes Sp1, and hypoglycosylated Sp1 is susceptible to proteasome degradation (38). However, recent modeling studies of Sp1 indicated that the modification adversely affected protein-protein interactions involving the transcription factor (39). Protein stability studies after cyclohexamide treatment suggested a significant increase in the Sp1 protein stability from 5-aza-2'-dC-treated MCF-7L nuclear extracts compared with control untreated MCF-7L nuclear extracts (Fig. 14). The mechanism for increased stability is not clear at this time. The available data indicate that the affected gene resulting in Sp1 protein stabilization is probably not affecting glycosylation. A transiently transfected RII promoter-CAT element exhibited enhanced promoter activity in MCF-7L cells when cotransfected with Sp1 cDNA, demonstrating further that low expression levels of Sp1 contribute to repression of RII expression in the MCF-7L cells.²

In summary, the results of our present study demonstrated that the tumor suppressor gene RII is repressed in the ER⁺ breast cancer cells because of suboptimal activity of Sp1. 5-Aza-2'-dC treatment indirectly stabilizes and activates Sp1, thus leading to enhanced RII transcription and subsequent restoration of TGF- response. These findings suggest a novel mechanism by which epithelial cells escape the negative growth regulatory effects of TGF- leading to malignant behavior.