Dual Blockade of Cyclic AMP Response Element- (CRE) and AP-1-directed Transcription by CRE-transcription Factor Decoy Oligonucleotide

Alteration of gene transcription by inhibition of specific transcriptional regulatory proteins has important therapeutic potential. Synthetic double-stranded phosphorothioate oligonucleotides with high affinity for a target transcription factor can be introduced into cells as decoy cis-elements to bind the factors and alter gene expression. The CRE (cyclic AMP response element)-transcription factor complex is a pleiotropic activator that participates in the induction of a wide variety of cellular and viral genes. Because the CRE cis-element, TGACGTCA, is palindromic, a synthetic single-stranded oligonucleotide composed of the CRE sequence self-hybridizes to form a duplex/hairpin. Herein we report that the CRE-palindromic oligonucleotide can penetrate into cells, compete with CRE enhancers for binding transcription factors, and specifically interfere with CRE- and AP-1-directed transcriptionin vivo. These oligonucleotides restrained tumor cell proliferation, without affecting the growth of noncancerous cells. This decoy oligonucleotide approach offers great promise as a tool for defining cellular regulatory processes and treating cancer and other diseases.

Eukaryotic transcription is regulated by the interplay of various protein factors at promoters (1,2). It has been shown that prokaryotic repressors can function as negative regulators of eukaryotic promoters (3,4). This observation suggests that displacement of activating proteins might provide a general strategy for gene-specific repression in eukaryotes. Several approaches have been undertaken to control eukaryotic gene expression through such displacement.
In one approach, trans-dominant mutants are generated that interfere with the function of transactivators. Mutants are generated that retain the ability to bind to cis-regulatory DNA sequences but that have dysfunctional transcriptional activation domains. These mutant transcription factors compete with their functional, wild-type counterparts for binding to the enhancer sequences and prevent the activation or repression of the target gene. Although this strategy has been successful, in vitro (5)(6)(7)(8), the generation of such mutants is not always possible. The transcription factor must be well characterized such that the activation domain(s) is identified and can be mutated. Also, even with sufficient knowledge to generate such mutants, difficult gene therapy procedures would be required to express these proteins in vivo.
Promoter competition strategy has also been utilized whereby plasmids containing cis-acting elements in common with the targeted gene are introduced in high copy number into cells (9). At high copy number, a majority of the transcription factors can be competitively bound away from the native enhancer sequences with gene expression accordingly regulated. Because these plasmids must be stably maintained at high copy number in target cells, a requirement that is difficult to achieve in vivo, this approach has also been limiting.
Another alternative is to employ oligonucleotides to form triple helices with enhancer elements. Pyrimidine oligonucleotides were found to bind in a sequence-specific dependence to homopurine sites in duplex DNA by triple helix formation and had sufficient specificity and affinity to compete with sitespecific DNA binding proteins for occupancy of overlapping target sites (10). However, such oligonucleotide-directed triple helix formation has not been shown in cells in vitro or in vivo.
A more successful oligonucleotide-based approach has been the use of synthetic double-stranded phosphorothioate oligonucleotide containing a cis-transcription element that can penetrate cells, can bind sequence-specific DNA-binding proteins, and can interfere with eukaryotic transcription in vivo (11,12).
The CRE 1 (cyclic AMP response element)-transcription factor complex is a pleiotropic activator that participates in the induction of a wide variety of cellular and viral genes (13). Because the CRE cis-element, TGACGTCA (13), is palindromic, a synthetic single-stranded oligonucleotide composed of the CRE sequence self-hybridizes to form a duplex/hairpin. We sought to ascertain whether the CRE-palindromic oligonucleotide can penetrate cells, bind sequence-specific DNA-binding proteins, and interfere with the CRE-directed transcription in vivo.
Because there are many cAMP-regulated genes and because they are ubiquitous in all cell types, the use of CRE decoy could be harmful to cells and organisms. Surprisingly however, the CRE decoys were harmless for normal cells but were potent inhibitors for cancer cell growth in vitro and in vivo.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

EXPERIMENTAL PROCEDURES
Oligonucleotides-CRE-decoy and control oligonucleotides used in the present studies were phosphorothioate oligonucleotides. Their sequences are as follows: 24-mer CRE palindrome, 5Ј-TGACGTCA  TGACGTCA TGACGTCA-3Ј; 24-mer CRE mismatch control, 5Ј-TGTG-GTCA TGTGGTCA TGTGGTCA-3Ј; and 24-mer nonsense-sequence  palindrome, 5Ј-CTAGCTAG CTAGCTAG CTAGCTAG-3Ј. The synthesis of the 24-mer phosphorothioate oligonucleotides was carried out using ␤-cyanoethylphosphoramidite chemistry on automated DNA synthesizer (Amersham Pharmacia Biotech Oligo Pilot II). Deprotection and purification of the oligonucleotides followed the experimental procedures previously described (14). Analysis of the purified oligonucleotide was carried out using capillary gel electrophoresis and polyacrylamide gel electrophoresis. The purity of the oligonucleotide based on capillary gel electrophoresis was 95% full-length and 5% n Ϫ 1, n Ϫ 2 products.
Treatment of Cells in Culture with CRE Oligonucleotides-Cells (0.25-1 ϫ 10 5 cells/well) were plated in 6-well plates containing the growth medium at 37°C. To increase the delivery of oligonucleotide into the cell, cationic lipid (DOTAP) (Boehringer Mannheim) was used in the oligonucleotide treatment. The CRE-decoy and control oligonucleotides were added (1 day after seeding) to the wells at varying concentrations (50 -200 nM) in the presence of DOTAP. At 5 h of incubation, the medium was removed, and fresh medium without oligonucleotide and DOTAP was added. Cells were harvested at indicated times, and cell numbers were counted in duplicate by a Coulter Counter.
Production of Stable Transfectants-MCF-7 cells (3 ϫ 10 5 cells/ 60-mm dish) were transfected with 6 g of KCREB plasmid (kindly provided by Richard H. Goodman), a dominant negative mutant form of CREB using DOTAP. Stably transfected cells were selected by growing cells in the presence of Geneticin (400 g/ml) (G418, Life Technologies, Inc.). The G418-resistant colonies were isolated after 3 weeks of selection.
Transient Transcription Assay of Somatostatin-Chloramphenicol Acetyltransferase Fusion Gene-Cells (5 ϫ 10 5 cells/60-mm dish) were transfected with 3 g of somatostatin-chloramphenicol acetyltransferase (CAT) fusion gene plasmid and 4 g of CRE or control oligonucleotide using DOTAP. After 24 h, fresh medium was added, and the cells were harvested at 48 -72 h, then assayed for CAT activity. When indicated, cells were also treated with forskolin (10 M) for the final 24 h. Cell lysates were prepared as described by Gorman (17). Lysates (100 g of protein) were incubated with 0.4 Ci of [ 14 C]chloramphenicol, 0.53 mM acetyl-CoA, and 250 mM Tris-HCl, pH 7.8, for 90 min at 37°C. Under these conditions, CAT activity was linear with time. Reaction products were analyzed by thin layer chromatography (17), and the plate was autoradiographed.
Cellular Uptake of Oligonucleotide-Cells (2 ϫ 10 5 cells/well) were incubated with 1.5 ng of 32 P-labeled oligonucleotide (5.5 ϫ 10 6 cpm/ pmol) in a 12-well plate containing the growth medium at 37°C for 1-24 h; at indicated times, cells were washed four times with phosphatebuffered saline, and radioactivities in the cell pellets and culture media were determined. The cell uptake of oligonucleotides was calculated as percent radioactivity in cell pellets relative to the total radioactivity in cell pellet plus culture medium.
Stability Tests of Oligonucleotide-Cells (4 ϫ 10 5 cells/60-mm dish) were incubated in growth medium containing 6 ng of 32 P-labeled CRE or palindromic control oligonucleotide (5.5 ϫ 10 5 cpm/pmol) at 37°C for 1-2 days. Cell-incorporated CRE oligonucleotides were extracted with phenol:chloroform, precipitated with ethanol, and subjected to 20% nondenaturing polyacrylamide gel electrophoresis. UV thermal melting experiments were carried out in 200 mM NaCl, 10 mM sodium dihydro-gen phosphate, pH 7.4, buffer and at the oligonucleotide concentration of 1.5 M. Thermal denaturation profiles were recorded on a Perkin-Elmer Lambda 20 Spectrophotometer equipped with a 6-cell linear movement module and connected to a PTP-6 Peltier thermal controller. All the data were collected and processed on a personal computer attached to the spectrophotometer using the software supplied by Perkin-Elmer. The melting temperatures (T m ) were measured from the first derivative plots.
Cellular Localization of Fluorescein Isothiocyanate-labeled Oligonucleotide-Cells (5 ϫ 10 5 per dish) were incubated with 150 nM of FITClabeled CRE palindrome or palindromic control oligonucleotide (5Ј-end labeled, The Midland Certified Reagent Co.) in the presence of DOTAP in 60-mm dishes containing the growth medium at 37°C. Six h after incubation, the medium was removed, and cells were washed three times with phosphate-buffered saline and were cultured in fresh growth medium. At indicated times, the intracellular distribution of FITClabeled oligonucleotides was analyzed by fluorescence inverted microscope (Axiovert 35, Carl Zeiss).
Western Blot Analysis of CREB-Nuclear extracts were prepared (see "Experimental Procedures" for EMSA), nuclear proteins (20 g) were separated on 12% SDS-polyacrylamide gel, and separated proteins were transferred onto nitrocellulose membrane using semidry blotting. Anti-CREB antibody (1:200 dilution, Santa Cruz Biotech.) and anti-rabbit-IgG antibody (1:2500 dilution, Amersham Pharmacia Biotech) conjugated with horseradish peroxidase were used as primary and secondary antibodies, respectively. Immunodetection was performed using enhanced chemiluminescence method recommended by the manufacturer (Amersham Pharmacia Biotech).
RNA Preparation and Northern Blotting Analysis-Total cellular RNA was prepared by the use of TRIreagent (Molecular Res. Center, Inc.). Northern blot analysis and hybridization of RNA with 32 P-labeled DNA probes were as described earlier (18). DNA was labeled with [␣-32 P]dCTP according to a standard protocol for random prime labeling using Amersham Pharmacia Biotech multiprime DNA labeling kit. The specific radioactivity of labeled DNA equaled 3.7 ϫ 10 6 cpm/g DNA.
Induction of c-fos Gene Expression-Cells were treated with 150 nM CRE or control oligonucleotide for 2 days in the serum-containing medium. Cells were washed twice with phosphate-buffered saline. Cells were further incubated in the absence of oligonucleotides in the serumfree medium containing 100 ng/ml TPA (Sigma) at 37°C. At various times, cells were harvested and fos mRNA was measured by Northern blotting.

RESULTS AND DISCUSSION
We examined whether a 24-mer single-stranded phosphorothioate oligodeoxynucleotide comprising a CRE palindrome (i.e. triplet copies of TGACGTCA, a CRE consensus sequence) can compete for binding the sequence-specific DNA binding proteins. We used 32 P-labeled double-stranded oligonucleotide (unmodified) that contained the CRE element (Promega) and nuclear extracts from MCF-7 breast cancer cells in the electrophoretic mobility shift assay. As shown in Fig. 1A, protein binding to the 32 P-labeled CRE probe was inhibited by the unlabeled CRE-palindromic oligonucleotide (lanes 2 and 3) but not by two-base mismatched control oligonucleotide, nonsensesequence palindrome oligonucleotide containing no CRE, or Oct-1, AP-1, and Sp-1 sequences.
We next examined the ability of CRE-decoy oligonucleotide to penetrate cells and compete with the cellular CRE elements for binding transcription factors in vivo. The nuclear extracts from cells treated with the 24-mer CRE oligonucleotide (150 nM, for 2 days) demonstrated a marked decrease in formation of the CRE-protein complex in the mobility shift assay as compared with control (saline-treated) cells (Fig. 1B, lanes 4, 8, and  12). A CREB (19) (CRE-binding protein) antibody caused supershift, indicating the presence of CREB protein within the labeled protein-DNA complexes (Fig. 1B, lanes 2, 6, and 10). The two-base mismatched control oligonucleotide treatment did not affect the CRE-protein complex formation (Fig. 1B,  lanes 3, 7, and 11). A Western blot analysis for CREB protein in untreated and in CRE oligo-treated and control oligo-treated cells demonstrated no change in CREB protein level (Fig. 1C), indicating that the CRE-decoy treatment did not affect CREB levels in the cell. These results were demonstrated in MCF-7 (breast carcinoma), MCF-10A (normal human mammary epithelial cell) and LNCaP (prostate carcinoma) cells. The above results demonstrate that the CRE-decoy oligonucleotide successfully competed with the cellular CRE enhancer for binding of sequence-specific DNA-binding proteins. Standard deviation for each CAT assay was less than 10%, and results are representative of two to four independent transfections. Panel E, CRE-decoy inhibition of endogenous cAMP-responsive genes. Total cellular RNA (from 2 ϫ 10 6 cells) preparation and Northern blotting were performed as described under "Experimental Procedures." 32 P-labeled probes were: the 1.5-kb cDNA clone containing the entire coding region of human RI␣ (kindly provided by Tore Jahnsen); the 1.1-kb full-length sequence of the human C␣ (kindly provided by Steven K. Hanks); 0.9-kb cDNA clone containing the human PEPCK (ATCC); 29-mer oligonucleotide probe for human 28 S rRNA (CLONTECH). C, saline-treated control cells; CRE, CRE-palindromic oligonucleotide-treated cells (150 nM, 2days); CREC, control oligonucleotide (nonsense sequence palindrome) treated cells (150 nM, 2days). The data represent one of three to four independent experiments that gave similar results. Panel F, inhibition of AP-1 DNA-protein complex formation. Binding site specificity was tested by EMSA with the 32 P-labeled oligonucleotide (double-stranded AP-1, Sp-1, and Oct-1 (Promega)) in the absence and presence of unlabeled competitor as indicated. Nuclear extracts were prepared from cells treated with saline (C), CRE oligonucleotide (CRE), or control mismatched oligonucleotide (CREC) at 150 nM for 2 days, and the EMSA was performed as described under "Experimental Procedures." Data represent one of three independent experiments that gave similar results. Panel G, inhibition of c-fos induction. At indicated times, cells were harvested. RNA preparation and Northern blot analysis were performed as described under "Experimental Procedures." 32 P-labeled probes were: 1.7-kb v-fos (ATCC), 4.2-kb PKC-␣ (kindly provided by C.A. Stein), and human ␤-actin (Oncor p7000␤ actin). C, saline-treated control cells; CRE, cells treated with CRE oligonucleotide (150 nM, 2days). Data represent one of three independent experiments that gave similar results. We next examined whether CRE oligonucleotides can modulate the transcriptional activity of sequence-specific DNA binding proteins in vivo. As shown in Fig. 1D, transfection of MCF-7 cells with somatostatin ⌬-71-CAT (20) plus CRE-palindromic oligonucleotide (lane CRE) resulted in a greater than 90% inhibition of the CRE-directed transcription compared with cells transfected with ⌬-71-CAT alone (lane C). Addition of the nonsense-sequence palindromic oligonucleotide that contains no CRE sequence (lane CREC) or the two-base mismatched control oligonucleotide (data not shown), which does not self-hybridize to form a duplex, had no inhibitory effect on the CAT activity. None of the oligonucleotides inhibited the simian virus 40 (SV40) enhancer (data not shown), which contains no recognizable CRE enhancer element.
It was noted that the untreated MCF-7 cells exhibited relatively high levels of basal somatostatin-CAT activity, and forskolin treatment had only a small stimulatory effect (Fig. 1D). In contrast, LNCaP prostate cancer cells and noncancerous MCF-10A cells exhibited very low levels of basal CAT activity, and forskolin greatly stimulated CAT activity (5-10-fold) (data not shown). This forskolin-stimulated CAT activity was also almost completely abolished by the CRE oligonucleotide treatment (data not shown). The disparity in the basal CAT activity observed in different cells may reflect varying degrees of cAMPdependent protein kinase activation and CREB-phosphorylation in the cell. Significantly, the CRE oligonucleotide was capable of inhibiting both the basal and cAMP-stimulated CAT activities.
A group of cAMP-responsive genes, such as somatostatin and phosphoenolpyruvate carboxykinase (PEPCK) contain the CRE which lies within the first 150 base pairs of the 5Ј-flanking region of the gene (13). Therefore, these elements could be regarded as basal enhancer, in addition to functioning as inducible enhancer (1). A role for the CRE as a basal transcription element was suggested in deletion analysis of the PEPCK promoter-regulatory region (22). When the CRE was deleted from the promoter, the basal level of gene transcription was reduced, and the responsiveness of the promoter to cAMP in hepatoma cells was abolished. CRE binding protein binds to the CRE in a cAMP-independent manner (13,19). This binding, as it has been proposed (13), may stimulate basal transcription by interacting with proximal promoter element, such as the TATA box binding factor and/or RNA polymerase II. The cAMP-induced phosphorylation of CREB could lead to a higher order complex formation with the basic transcription factor. Thus, dual role for the CRE as both a basal and an inducible transcription element has been shown. Importantly, the CRE oligonucleotide interfered with both the basal and cAMP-induced transcription of an exogenously supplied CRE-containing gene (Fig. 1D).
We next examined the ability of CRE-palindromic oligonucleotide to inhibit expression of the endogenous cAMP-responsive genes. The CRE-palindromic oligonucleotide treatment brought about a marked reduction in the mRNA levels of the catalytic (C␣) and regulatory (RI␣) subunits of cAMP-dependent protein kinase (PKA) and PEPCK in MCF-7 breast cancer cells (Fig. 1E). In noncancerous MCF-10A cells, the reduction in the mRNA levels was small, although the expression of these genes, particularly of RI␣ and C␣ genes was very low to begin with. In contrast, the control oligonucleotide had no effect on the mRNA levels of the RI␣ and C␣ and PEPCK (Fig. 1E).
PEPCK and the PKA R and C subunit genes are cAMPinducible genes (13,23). The promoter region of the porcine RI␣ gene contains a CRE consensus sequence as well as the AP-2 recognition site (24). Transcription factor binding to both of these sequences is believed to be involved in cAMP regulation of gene transcription (13). The promoters of PKA R and C subunit genes are TATA-less and GC-rich, and multiple transcription initiation sites were identified within the GC-rich region (24,25). Such GC-rich regions have been associated with transcription initiation sites of many constitutively expressed housekeeping genes (26). Our observation that the CRE oligonucleotide inhibits the basal expression of RI␣ and C␣ genes (Fig. 1E) strongly indicates that the CRE oligonucleotide can indeed compete with the cis-CRE element in binding CREB. Because the CRE oligonucleotide can interfere with CREB binding to the cis-element, it is expected that the oligonucleotide could produce a more profound effect on the mRNA reduction under cAMP-induced conditions. Thus, the CRE oligonucleotide can interfere with both basal and cAMP-induced expression of the endogenous CRE-containing genes.
We therefore examined whether the CRE oligonucleotide treatment affects AP-1 binding. The nuclear extracts from cells treated with the CRE-palindromic oligonucleotide demonstrated a marked reduction in formation of the AP-1 DNAprotein complex in the mobility shift assay compared with control (saline-treated) cells (Fig. 1F, lane 4). Two-base mismatched control oligonucleotide treatment had no effect on the AP-1 DNA-protein complex formation (Fig. 1F, lane 3). By comparison, CRE-decoy oligonucleotide had no effect on Sp-1 or Oct-1 DNA-protein binding (Fig. 1F, lanes 8 and 12).
We then examined the effect of CRE-decoy oligonucleotide treatment on the expression of c-fos gene that is cAMP responsive (13). As shown in Fig. 1G, the CRE decoy brought about a marked decrease in the TPA-inducible mRNA level of c-fos. The control mismatched oligonucleotide had no effect on the c-fos expression (data not shown). A cAMP-unresponsive gene, such as PKC-␣ was not affected by the CRE oligonucleotide treatment (Fig. 1G). These results show that the CRE-decoy oligonucleotide treatment resulted in inhibition of transcription factor binding at two different cis-elements, the CRE and the AP-1.
The above results showed that the CRE-decoy oligonucleotides effectively competed with the cellular cis-element for binding the transcription factors and interfered with the function of transactivators in intact cells. To correlate such effects of decoy oligonucleotides with their cellular uptake, we incubated 32 P-labeled 24-mer CRE-palindromic oligonucleotide or mismatched or nonsense sequence palindromic control oligonucleotide with MCF7 and MCF-10A cells. Cell-associated radioactivity was quantified. Within 5 h, about 10% of the total input oligonucleotide accumulated in the cell and the incorporation continued to rise thereafter, reaching 20 -25% maximum levels at 24 h of oligonucleotide incubation (Fig. 2A). The amount and the rate of the incorporation of the oligonucleotides were similar between MCF-7 ( Fig. 2A) and MCF-10A (data not shown) cells and between the CRE-decoy oligonucleotide and control oligonucleotide ( Fig. 2A). Cell-associated DNA was isolated and analyzed by nondenaturing polyacrylamide gel electrophoresis. Up to 48 h of examination, the 24-mer CRE oligonucleotide accumulated in MCF-7 cells at a size consistent with the duplex/hairpin forms (Fig. 2B). Consistent with these data, the 24-mer CRE oligonucleotide exhibited a high melting temperature (Fig. 2C).
The intracellular distribution of the fluorescence signal at different times after the treatment of cells with FITC-labeled CRE-decoy and control oligonucleotides is illustrated in Fig. 3. In MCF-7 breast cancer cells, within 6 h of treatment with CRE-palindromic oligonucleotide, a strong fluorescence labeling was observed in both cytoplasm and nucleus (Fig. 3, A and  B). In addition, a large amount of labeling was also observed in the extracellular space. 12 h after the treatment, the nuclear fluorescence had become much more intense, and the extracellular fluorescence had largely disappeared (Fig. 3, C and D), and by 24 h, the intensity of fluorescence was reduced in both nucleus and cytoplasm (Fig. 3, E and F). This pattern of fluorescence was also observed in noncancerous MCF-10A cells except that the fluorescence had a more punctated appearance (Fig. 3, G, H, I, and J). The control oligonucleotide exhibited the same pattern of fluorescence as did the CRE-decoy oligonucleotide (Fig. 3, K and L).
Because there are many cAMP-regulated genes, and they are ubiquitously expressed, we examined whether the CRE oligonucleotide treatment could interfere with cell growth. Surprisingly, the CRE-decoy oligonucleotide produced selective growth inhibition of cancer cells without adversely affecting the normal cell growth (Fig. 4A). The 24-mer CRE-palindromic oligonucleotide produced potent growth inhibition in a variety of cancer cells including MCF-7 (breast cancer), A549 (lung carcinoma), LNCaP (prostate cancer), LS174T and SW480 (colon carcinomas) (data not shown), KB (epidermoid carcinoma) (data not shown), and multidrug-resistant (MDR) cancer lines of MCF7TH (MDR-breast cancer) and HCT-15 (MDR-colon carcinoma) (data not shown). In contrast, the CRE oligonucleotide had little (Ͻ20%) or no effect on the growth of noncancerous cells, MCF-10A (human mammary epithelial cell), L132 (human lung epithelial cell), Hs68 (human newborn foreskin fibroblast) (data not shown), and NIH/3T3 fibroblasts (data not shown). The growth inhibition of cancer cells was achieved at nanomolar concentrations of CRE oligonucleotide (IC 50 , 100 -200 nM) without obvious cytotoxicity and accompanied by changes in cell morphology and appearance of apoptotic nuclei (programmed cell death) (data not shown). The growth inhibition was CRE sequence-specific as the two-base mismatched control oligonucleotide (Fig. 4A) or the nonsense-sequence palindromic oligonucleotide that contains no CRE sequence (data not shown) had little (Ͻ30%) or no growth inhibitory effect.
CRE oligonucleotide treatment also inhibited in vivo tumor growth. Treatment of nude mice bearing HCT-15 human MDR colon carcinoma with 24-mer CRE oligonucleotide (0.1 mg/ mouse, intraperitoneal, daily, 5ϫ week for 4 weeks) resulted in greater than 85% inhibition of tumor growth as compared with the saline-treated control tumors without causing systemic toxicity (Fig. 4B). Two-base mismatched control oligonucleotide had no growth inhibitory effect (Fig. 4B).
The growth inhibition may have been because of actions other than blockade of CRE gene transcription, as nonspecific binding of oligonucleotide or its degradation products to biological targets has been shown (30). As discussed below, however, our data show that the binding of decoy oligonucleotide at the transcription factor DNA-binding domain is clearly related to the inhibition of cell growth. First, in undifferentiated F9 teratocarcinoma cells, a cell line that is unresponsive to cAMP, the CRE decoy produced no growth inhibition (Fig. 4C). This suggests that the decoy may act as growth inhibitor, at least in part, through binding to CREB because the CRE is nonfunctional in F9 cells although CREB is present (31). Second, KCREB, a CREB mutant that contains a mutation of a single amino acid in the DNA-binding domain, is known not to bind to native CRE sequences (32). Cancer cells overexpressing KCREB exhibited decreased cell growth as compared with parental cells and showed little or no response to the decoy oligonucleotide treatment (Fig. 4D).
We demonstrated here that the synthetic single-stranded CRE oligonucleotide of palindrome structure functioned as effective and stable transcription factor decoys to alter gene expression in vivo. Importantly, the CRE-decoy oligonucleotides achieved gene-specific regulation in vivo, leading to selective inhibition of cancer cell growth without adversely affecting the growth of normal cells.
The specificity of the growth inhibitory effect of the CREdecoy oligonucleotides on cancer cells is supported by several lines of evidence. (i) The CRE-decoy oligonucleotide produced growth inhibition of cancer cells but not normal cells, in vitro and in vivo, whereas mismatched control oligonucleotides or nonsense-sequence palindromic oligonucleotide that self-hybridizes but contains no CRE, did not inhibit growth. (ii) Administration of CRE-decoy oligonucleotides, but not mismatched oligonucleotides, markedly inhibited CRE DNA-protein complex formation, CRE-directed transcription activity, and endogenous cAMP-responsive gene expression in both cancer cells and normal cells. (iii) Cellular uptake of decoy oligonucleotides and control oligonucleotides was similar for cancer cells and normal cells. (iv) The specific growth inhibitory effect toward cancer cells correlated with induction of phenotypic change and apoptosis.
Importantly, the CRE-decoy oligonucleotides not only blocked the CRE-PKA pathway via repression of the PKA genes but also brought about the blockade of AP-1-PKC pathway by inhibiting c-fos gene expression that is CRE-responsive. This dual blockade of two important signal transduction pathways could be causally related, at least in part, to the CRE oligonucleotide-inhibition of cancer cell growth.
CREB, a critical regulator of immediate early gene transcription, has been shown to be activated by growth factors (33) and play an important role in the acquisition of the metastatic phenotype of human melanoma cells (21). Although the exact mechanism of action remains to be elucidated, our results suggest that the CRE-decoy oligonucleotides may provide a powerful means of combating cancers by regulating the expression of cAMP-sensitive genes.