Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Chloroethylnitrosoureas (CENUs) such as 1,3-bis-(2-chloroethyl)-1-nitrosourea (Carmustine, BCNU) are clinically useful cancer chemotherapeutic agents. However, their antineoplastic activity is much lower against human tumors than against experimental animal tumors (Carter and Wasserman, 1976). The DNA repair protein O⁶-methyguanine DNA methyltransferase (MGMT), also known as alkylguanine DNA alkyltransferase, is expressed in more than 80% of human tumor cell lines and plays a major role in CENU resistance (Pegg, 1990; Erickson, 1991; Mitra and Kaina, 1993). The cytotoxic lesion induced by CENU agents is a DNA interstrand cross-link between guanine and cytosine. MGMT prevents the formation of the cross-link by stoichiometrically transferring the chloroethyl monoadduct from the O⁶ position of guanine to the active site of the enzyme, resulting in a covalent bond that inactivates the protein. The replacement of inactivated MGMT appears to require de novo protein synthesis (Kroes and Erickson, 1995).

Biochemical strategies have been developed to inactivate MGMT by pretreating tumor cells with methylating agents such as streptozotocin (Futscher et al., 1989), or alkylguanine analogs such as O⁶-methylguanine (Dolan et al., 1989) and O⁶-benzylguanine (Dolan et al., 1991). These chemical agents can deplete cellular MGMT and thus temporarily sensitize tumor cells resistant to CENU treatment (Dolan et al., 1989; Futscher et al., 1989). O⁶-Benzylguanine is currently in phase II clinical trials (Friedman et al., 1998; Spiro et al., 1999). One limitation of biochemical strategies to deplete MGMT is the inability to do so selectively in tumor cells and not in normal tissues. Moreover, the MGMT inhibition is only temporary because the de novo synthesized MGMT proteins replace the inactivated molecules, resulting in the gradual recovery of MGMT activity within 24 h and the ultimate restoration of CENU resistance (Kroes and Erickson, 1995; Kreklau et al., 1999).

An alternative strategy to deplete MGMT in tumor cells is to inhibit de novo MGMT synthesis using hammerhead ribozymes (Potter et al., 1993; Citti et al., 1999) designed to degrade the long-lived MGMT mRNA (Kroes and Erickson, 1995). Ribozymes are small RNA molecules that hybridize to a complementary sequence of RNA and catalyze site-specific cleavage of the substrate (Cech, 1987). Ribozymes have been used to down-regulate gene expression in a sequence-specific manner (Denman et al., 1994; Larsson et al., 1994; Phylactou et al., 1998). The hammerhead ribozyme, composed of only 30 to 40 nucleotides, is the smallest and simplest ribozyme. In addition, the potential cleavage sites of hammerhead ribozymes are abundant in most mRNAs. The consensus sequence of the cleavage domain within the target RNA contains any GUX (where X can be C, A, or U), UUC, and CUC recognition sites. The optimal cleavage activity occurs 3^' to the GUC target sites (Ruffner et al., 1990). The rates of ribozyme hybridization to the substrate, dissociation from the substrate, and dissociation of the cleavage products from the ribozyme depend upon the length and nucleotide composition of the hybridizing arms. A hybridizing arm of 7 to 10 base pairs has been found to be effective for a ribozyme to bind with the target mRNA GUC flanking sequences with antisense hybridization and to dissociate from the cleaved products (Herschlag, 1991; Tuschl and Eckstein, 1993; Bertrand et al., 1994; Birikh et al., 1997; Bramlage et al., 1998; Verma and Eckstein, 1998). We selected ribozyme cleavage sites by first identifying the GUC triplets within the nucleotide sequence of the target MGMT mRNA deduced from the cDNA sequence (Tano et al., 1990). The next round of selection was to find the GUC sites within the mRNA that were most accessible to ribozymes. We then designed eight hammerhead ribozymes that target specific sites of MGMT mRNA and used this strategy to modulate BCNU resistance in human tumor cells. We have used a pooled transfection strategy with all eight ribozymes in BCNU-resistant HeLa cells. Stably transfected clones were screened for BCNU sensitivity, and the BCNU-sensitive clones were then further characterized.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. BCNU was purchased from Sigma Chemical Co. (St. Louis, MO) and stored at 20°C. Immediately prior to use, BCNU was dissolved in 95% EtOH, maintained on ice, and diluted at least 100-fold in culture medium for experiments. Unless otherwise stated, other chemical reagents were obtained from Sigma Chemical Co., Fisher Scientific (Chicago, IL), or USB (Cleveland, OH).

Cell Culture. HeLa and HeLa MR cells were cultured in -minimum essential medium (Life Technologies, Rockville, MD) supplemented with 10% bovine calf serum (Hyclone Laboratories, Logan, UT), 1% glutamine, 1% HEPES, and 2% penicillin-streptomycin (all from Life Technologies) and maintained at 37°C in a humidified 5% CO₂ atmosphere.

Synthesis, Subcloning, and Purification of Ribozymes. Ribozymes (Rz) were designed against eight GUC sites in the 5' region of the MGMT mRNA, specifically cleavage sites at nucleotides 41, 57, 161, 178, 212, 234, 262, and 374, deduced from the MGMT cDNA sequence (Tano et al., 1990). The general structure of the hammerhead ribozymes and the variable sequences are shown in Fig. 1 and Table 1, respectively. Two complementary single-stranded oligodeoxynucleotides containing the sequence encoding each of the ribozymes were synthesized by Life Technologies. When annealed, these oligonucleotides possessed sticky ends to allow directional subcloning into the pcDNA3.1/Zeo() mammalian expression vector (Invitrogen, Carlsbad, CA) between the XhoI and HindIII restriction sites. Following ligation with the expression vector, competent Escherichia coli DH5 cells (Life Technologies) were transformed. Plasmid DNA from ampicillin-resistant colonies was purified by anion exchange chromatography (Qiagen Maxiprep kit; Qiagen, Valencia, CA).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Structure of hammerhead ribozymes targeting against MGMT mRNA. The generic hammerhead ribozyme structure is depicted (adapted from Von Ahsen and Schroeder, 1993). The core sequence of the hammerhead structure, consisting of 24 nucleotides, is conserved and essential for catalytic activity.

Stable Transfection of HeLa Cells. The eight ribozyme-containing plasmids were pooled and cotransfected into HeLa cells using LipofectAMINE-PLUS reagent according to the manufacturer's instructions (Life Technologies). Briefly, 2 × 10⁶ cells were seeded into 150-mm dishes in complete medium and incubated under standard conditions for 24 h. Plasmid DNA (1 µg/ribozyme-containing vector, 8 µg total) was precomplexed with the PLUS reagent and then incubated with serum-free medium containing the LipofectAMINE reagent. The DNA-LipofectAMINE-PLUS complex and fresh serum-free medium (total volume 10 ml) were added to the cells, and the cells were incubated under standard culture conditions for 3 h. The medium was then replaced with fresh complete medium. One day after transfection, Zeocin (Invitrogen) was added to the medium to a final concentration of 400 µg/ml. About 10 days later, Zeocin-resistant colonies were picked using cloning discs (Scienceware, Pequannock, NJ) and expanded in six-well plates in complete medium containing 100 µg/ml Zeocin.

Cell Survival Assays. To screen for the BCNU-sensitive HeLa/Rz clones, the 4-[3-(4-lodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]-1,3-benzene disulfonate (WST-1; Roche Molecular Biochemicals, Mannheim, Germany) cell proliferation assay was used. Briefly, 2 × 10³ cells were seeded per well in a 96-well plate and incubated under standard conditions for 24 h. BCNU to a final concentration of 100 µM or 95% EtOH (final concentration <1%) was then added to the medium for 1 h. The cells were washed and incubated with complete medium for 72 h. WST-1 reagent (10 µl/well) was added directly to the medium and cells were incubated under standard conditions for 1 h. Cell viability as a function of metabolic activity was quantified by measuring absorbance at a wavelength of 450 nm using a multiwell spectrophotometer (ELISA reader). For colony formation assays, 2.5 × 10³ cells were seeded in 20 ml of complete medium in 75-cm² flasks and incubated under standard conditions for 24 h. Various doses of BCNU or 95% EtOH (final concentration <1%) were separately added to each flask, and cells were treated for 1 h. The cells were then washed with phosphate-buffered saline (PBS; Life Technologies), trypsinized, counted with a Coulter counter (Coulter Corporation, Miami, FL), and seeded into triplicate 100-mm cell culture dishes (Falcon, Franklin Lakes, NJ) at densities of 300, 500, 1,000, 3,000, and 10,000 cells/dish in complete medium. After incubation under standard conditions for 15 days, colonies were fixed with methanol, stained with 10% methylene blue in PBS, and counted.

PCR Analysis. Genomic DNA was prepared from cells using Release-IT reagent (CPG, Lincoln Park, NJ) according to the manufacturer's instructions. Deoxynucleotides and Taq polymerase were purchased from Life Technologies and Roche Molecular Biochemicals, respectively. The cytomegalovirus forward (5'-CGCAAATGGGCGGTAGGCGTG-3') and bovine growth hormone reverse (5'-TAGAAGGC ACAGTCGAAG-3') primers, flanking the pcDNA3.1/Zeo() multiple cloning sites, were synthesized by Life Technologies. The products were amplified using an initial denaturation step at 94°C for 3 min; followed by 35 cycles 94°C for 1 min, 55°C for 1 min, 72°C for 1 min; and finally 1 cycle 72°C for 8 min in a thermal cycler (PerkinElmer, Norwalk, CT). The amplified products were then sequenced at the Indiana University Biotechnology Facility (Indianapolis, IN).

MGMT Activity Assay. MGMT activity was measured as previously described (Wu et al., 1987; Futscher et al., 1989; Kreklau et al., 2001). Briefly, 1 × 10⁶ PBS-washed cells were suspended in 0.5 ml of cold assay buffer (50 mM Tris-HCl, 0.5 mM EDTA, 1 mM dithiothreitol, 5% glycerol; pH 8.0), pulse-sonicated on ice for 5 × 5 s, and centrifuged at 14,000g at 4°C for 30 min. Protein concentration was determined by measuring absorbance at 595 nm using the Coomassie blue protein reagent (Pierce, Rockford, IL). The MGMT substrate consisted of a fluorometric (5'-HEX labeled), double-stranded 18-bp oligonucleotide containing a single O⁶-methylguanine residue nested within a PvuII restriction site (Genosys Biotechnologies, Inc., The Woodlands, TX), which yielded a 10-bp, labeled PvuII cleavage product. To measure the activity, 0.2 pmol of the oligonucleotide substrate (200 pmol/ml) was incubated with 50 µg of cellular protein in a total volume of 150 µl of assay buffer at 37°C for 2 h. Following two phenol/chloroform/isoamyl alcohol (25:24:1) extractions, the oligonucleotide was precipitated overnight in ethanol in the presence of sodium acetate and 10 µg of carrier tRNA. The purified oligonucleotide was then digested with PvuII restriction enzyme (Roche Molecular Biochemicals) in a total reaction volume of 20 µl at 37°C for 2 h. The reaction was terminated by adding formamide (10 µl) absent any dyes to each sample, and the samples were denatured by heating at 95°C for 5 min. The resulting HEX-labeled DNA fragments were analyzed by electrophoresis on a denaturing 20% polyacrylamide gel containing 8 M urea in 1× Tris-borate EDTA buffer run at 26 milliamps for 75 min at approximately 50°C. The HEX-labeled 18- and 10-bp fragments were detected using a Hitachi FMBIO II Fluorescence Imaging System (Hitachi Genetic Systems, South San Francisco, CA), and the fluorescence was quantitated using FMBIO Analysis software (Hitachi Genetic Systems). Cellular MGMT activity is directly proportional to the substrate cleavage and is thus represented as the percentage of the 10-bp fragment fluorescence relative to the total amount of substrate fluorescence (10 bp + 18 bp).

Western Blot Analysis. Following electrophoresis of 50 µg of cellular protein on a 12% SDS-polyacrylamide gel, the protein was transferred to a 0.22 µM nitrocellulose membrane (MSI/Osmonics Inc., Minnetonka, MN). The transfer was carried out in transfer buffer (20 mM Tris, 10 mM glycine, 19% ethanol, pH 9.12), run at 150 mA for 80 min at room temperature. The immunoblots were stained with 0.5% Ponceau S stain to assess evenness of loading and transfer. The immunoblots were immersed overnight at 4°C in blocking agent prepared with wash buffer (20 mM Tris-HCl, pH 7.4, 137 mM NaCl, and 0.05% Tween 20) containing 10% (w/v) skim milk. For immunodetection of MGMT, the blots were incubated with 15 ml of primary anti-MGMT antibody (Lab Vision, Fremont, CA) diluted 1:400 in blocking buffer for 1 h and rinsed with wash buffer. Fifteen milliliters of goat anti-mouse horseradish peroxidase-conjugated second antibody (Amersham Pharmacia Biotech, Buckinghamshire, UK) diluted 1:5000 in blocking buffer was added to the blots, which were then incubated for 1 h at room temperature, and rinsed in washing buffer. The immunoblot was overlaid with 4 ml of chemiluminescence substrate (Pierce) for 5 min and exposed to film for about 2 min.

Northern Blot Analysis. Total cellular RNA was isolated and purified using TRI reagent (Sigma Chemical Co.) according to the manufacturer's instructions. Following electrophoresis of total RNA in a 1% formaldehyde/agarose gel, RNA was transferred to a nylon membrane (GeneScreen Plus; PerkinElmer Life Science Products, Boston, MA) by capillary blotting. The membrane was prehybridized in ULTRAhyb hybridization solution (Ambion, Austin, TX) at 42°C for 30 min in a hybridization bottle in a mini hybridization oven (NLC, Woodbridge, NJ). MGMT and -actin cDNA oligonucleotides were ³²P-radiolabeled by random primer labeling (Prime-It II kit; Stratagene, Cedar Creek, TX). The membrane was sequentially hybridized with ³²P-labeled MGMT cDNA probe (2 × 10⁶ cpm/ml) overnight and ³²P-labeled -actin cDNA probe (0.5 × 10⁶ cpm/ml; cDNA purchased from Ambion) for 8 h at 42°C in ULTRAhyb hybridization solution. The membrane was washed with 2× standard saline citrate/0.1% SDS, 2 × 5 min, and then washed with 0.1× standard saline citrate/0.1% SDS, 2 × 15 min, at 42°C. The hybridized membrane was exposed to film at 80°C overnight. The Northern blot was quantitated using a Storm 860 PhosphorImager (Molecular Dynamics, Sunnyvale, CA) according to the manufacturer's instructions.

CpG Methylation Analysis. Southern blot analysis for DNA methylation of the SmaI site within the MGMT promoter (Harris et al., 1991; Von Wronski et al., 1992) was performed as described by Potter et al. (1993) with minor modifications. Briefly, genomic DNA was isolated and purified using DNAzol reagent (Molecular Research Center, Cincinnati, OH) according to the manufacturer's instructions. The DNA was sequentially digested with SacI (3 U/µg DNA) at 37°C and SmaI (3 U/µg DNA) at 25°C for 12 h each. After 1% agarose gel electrophoresis, the gel to be transferred was denatured in denaturation solution (0.5 M NaOH, 0.4 M NaCl) 2 × 20 min and neutralized in neutralization solution (0.5 M Tris-HCl, 1.5 M NaCl, pH 7.0) 2 × 20 min at room temperature. The DNA was transferred to Magnacharge nylon membrane (Osmonics, Minnetonka, MN). The membrane was prehybridized as described above for Northern analysis. A 291-bp fragment containing the SmaI restriction site (nucleotides 676-967 of the MGMT promoter, graciously provided by Dr. Bernard Futscher, AZ Cancer Center, Tucson, AZ) was ³²P-radiolabeled by random primer labeling as described above. The membrane was hybridized with the ³² P-labeled probe (1 × 10⁶ cpm/ml) in ULTRAhyb hybridization solution at 42°C overnight. The membrane was washed and exposed to film for 18 h at 80°C.

Statistical Analysis. Individual values were compared with respective controls using the Student's t test, and differences were considered to be significantly different when p < 0.05.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Selection of BCNU-Sensitive HeLa/Rz Clones. Eight hammerhead ribozyme constructs were designed targeting eight specific GUC sites of MGMT mRNA (Fig. 1; Table 1). After pooled transfection of the eight Rz into HeLa cells, stably transfected HeLa/Rz clones were picked, expanded, and screened for BCNU sensitivity using a replica plating strategy. Briefly, the clones were expanded in a master plate and then split into replica plates to measure cell viability following treatment with either vehicle alone or a sublethal dose of BCNU (100 µM) using the WST-1 assay. Clones in the replica plates that displayed decreased viability in response to both vehicle and BCNU were considered to be false positives. Clones that displayed sensitivity only to BCNU but not vehicle were considered to be prospective ribozyme-expressing BCNU-sensitive clones, and were expanded for further characterization from cells remaining in the master plate. Several HeLa clones stably transfected with the empty vector were also picked and expanded. In the WST-1 cell proliferation assay (Fig. 2), wild-type HeLa cells, which are Mer⁺ (methylation repair) with high MGMT activity, were resistant to a sublethal dose of 100 µM BCNU. In contrast, HeLa MR cells, which are Mer and MGMT-deficient with no detectable MGMT activity (Day et al., 1980; Yarosh et al., 1983), were highly sensitive to 100 µM BCNU. A HeLa/vec clone, HeLa cells transfected with the empty vector, exhibited resistance to BCNU similar to that of the wild-type HeLa cells. Among 50 HeLa/Rz clones that were screened using the WST-1 cell proliferation assay, four clones (clone 4, 14, 29, and 36) displayed levels of BCNU sensitivity that were similar to HeLa MR cells (Fig. 2).

View larger version (43K): [in this window] [in a new window]   Fig. 2.   Screening of BCNU-sensitive HeLa/Rz clones by WST-1 cell proliferation assay. BCNU-sensitivity was determined in wild-type HeLa, HeLa MR, and 50 stably transfected HeLa/Rz clones using a short-term WST-1 assay. Data shown represent the mean ± S.D. from three independent experiments (*p < 0.05).

PCR Analysis. To determine which ribozyme construct(s) had been transfected into the BCNU-sensitive HeLa/Rz clones, the ribozyme cDNA inserts were amplified by PCR from the genomic DNA of HeLa/Rz clones 4, 14, 29, and 36 (data not shown). Sequencing of the PCR products showed that the active ribozyme inserts were Rz161 in HeLa/Rz clone 4, Rz178 in clones 14 and 29, and both Rz178 and Rz212 in HeLa/Rz clone 36, whose target sites within the MGMT mRNA are nucleotide 161, 178, and 212, respectively (deduced from MGMT cDNA sequence, Tano et al., 1990). Henceforth, HeLa/Rz clones with active ribozyme insert(s) were termed according to the targeted nucleotide(s) of MGMT mRNA as HeLa/Rz161, HeLa/Rz178a, HeLa/Rz178b, and HeLa/Rz212/178 (Table 2).

Effects of Ribozymes on MGMT Expression and Activity. To determine whether MGMT was down-regulated in the HeLa/Rz clones, MGMT activity, protein, and mRNA levels were measured (Figs. 3-5, respectively). Wild-type HeLa cells exhibited a high level of MGMT activity, generating about 65% substrate cleavage (Fig. 3). Under the same conditions, MGMT activity was undetectable in the HeLa MR cells. Among the HeLa/Rz clones, HeLa/Rz178a and HeLa/Rz178b displayed intermediate levels of MGMT activity, showing about 50 and 32% reduction in activity compared with wild-type HeLa cells, respectively (*p < 0.05). In both HeLa/Rz161 and HeLa/Rz212/178, the MGMT activities were dramatically reduced compared with the wild-type HeLa cells (***p < 0.001) and nearly as low as that in HeLa MR cells.

View larger version (39K): [in this window] [in a new window]   Fig. 3.   MGMT activity in HeLa/Rz clones. Cellular MGMT activity was compared among wild-type HeLa, HeLa MR, and HeLa/vec, HeLa/Rz161, 178a, 178b, and 212/178 clones. Percentage (%) cleavage = fluorescence intensity units of 10 mer/(10 mer + 18 mer), which is directly proportional to MGMT activity. Data shown represent mean ± S.D. from three independent determinations (*p < 0.05; ***p < 0.001).

View larger version (31K): [in this window] [in a new window]   Fig. 4.   Western blot analysis of MGMT protein levels in HeLa/Rz clones. MGMT protein levels in wild-type HeLa, HeLa MR and HeLa/vec, HeLa/Rz161, 178a, 178b, and 212/178 clones was determined by Western blot analysis using 50 µg of total cellular protein. Data shown are of a representative experiment from at least three independent determinations.

View larger version (43K): [in this window] [in a new window]   Fig. 5.   Northern blot analysis of MGMT mRNA levels in HeLa/Rz clones. MGMT mRNA in wild-type HeLa, HeLa MR, HeLa/vec, and HeLa/Rz161, 178a, 178b, and 212/178 clones was determined by Northern blot analysis using 20 µg of total RNA. Data shown are of a representative experiment from at least three independent determinations.

CpG Methylation of the MGMT Promoter. It has been shown that cells lacking MGMT activity can be distinguished from MGMT-expressing cells by examination of the CpG methylation status in the MGMT promoter region (Harris et al., 1991; Pieper et al., 1991; Von Wronski et al., 1992; Watts et al., 1997). Genomic DNA digested sequentially with SacI and SmaI and then hybridized with a cDNA probe (nucleotide 676-967 in MGMT gene) flanking the SmaI restriction site in the MGMT promoter region shows a 2.3-kb band in MGMT-expressing cells and a 2.6-kb band in cells in which the MGMT gene has been silenced by promoter methylation. This analysis determines CpG methylation status because cytosine methylation within the SmaI (^mCC^mCGGG) restriction site prevents digestion by this enzyme (Harris et al., 1991; Von Wronski et al., 1992; Potter et al., 1993; Watts et al., 1997). As shown in Fig. 6, genomic DNA from all HeLa/Rz and HeLa/vec clones sequentially digested by SacI and SmaI displayed the 2.3-kb band, as did the parental HeLa cell line. This observation indicates that the lack of MGMT expression in HeLa/Rz clones is not likely to be a result of methylation silencing of the MGMT gene.

View larger version (35K): [in this window] [in a new window]   Fig. 6.   Southern blot analysis of CpG methylation of the MGMT promoter. Methylation status of a CpG site located within the MGMT promoter region was examined in genomic DNA from wild-type HeLa, HeLa/vec, and HeLa/Rz161, 178a, 178b, and 212/178 clones. The presence of a 2.3-kb band indicates a lack of methylation at the CpG site. Data shown are of a representative experiment from at least three independent determinations.

BCNU Sensitivity. To further investigate the extent to which the HeLa/Rz clones were sensitized to BCNU compared with wild-type HeLa cells, cell survival was measured using colony formation assays. As shown in Fig. 7, BCNU-induced cell kill was shown to be increased by 2 to 3 logs in HeLa/Rz161 and HeLa/Rz178/212 clones compared with wild-type HeLa cells at a BCNU dose of 100 µM. In contrast, the HeLa/Rz178a and HeLa/Rz178b clones were only modestly sensitized to BCNU compared with wild-type HeLa cells.

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Colony formation assays for BCNU sensitivity in HeLa/Rz clones. BCNU sensitivity was determined in wild-type HeLa (), HeLa MR (), HeLa/vec (), and HeLa/Rz161 (*), 178a (), 178b (), and 212/178 (+) clones by colony formation 10 to 14 days after treatment with 100 µM BCNU for 1 h. Survival curves depict the mean ± S.D. of three independent experiments.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Several approaches have been undertaken to reverse tumor resistance to CENU agents by depleting cellular MGMT activity, including the use of DNA methylation agents and biochemical MGMT substrates (Futscher et al., 1989; Dolan et al., 1991; Erickson, 1991; Pieper et al., 1991). These biochemical strategies temporarily deplete cellular MGMT activity. Due to the continuous translation of pre-existing, long-lived MGMT mRNA (Kroes and Erickson, 1995), the MGMT protein is rapidly regenerated, which is thought to be responsible for the recovery of MGMT activity. In the current study, we investigated a strategy to deplete MGMT activity in human tumor cells using hammerhead ribozymes targeted to MGMT mRNA.

One of the major problems in directing destruction of mRNA by ribozymes is the selection of a suitable target site that is free of secondary structure and thus accessible to ribozyme cleavage. Computational techniques for predicting the folding of the secondary structure of the ribozyme and the target mRNA continue to be explored (Denman, 1993; Thomson et al., 1993). However, the secondary structure of MGMT mRNA is often too complex to be predicted even with refined computer-derived structure. To circumvent the problem of trying to predict the higher order structure of the MGMT mRNA and determine the optimum ribozyme target sites precisely, a pooled ribozyme approach has been developed. In this study, HeLa cells resistant to CENU and possessing a high level of MGMT activity were transfected with a pool of ribozyme constructs that were designed against eight potential GUC triplet sites in the 5' region of MGMT mRNA. The stably transfected cell clones with active ribozyme insert(s) were identified by exposure to a sublethal dose of BCNU on replica plates, thereby testing their enhanced sensitivity. The revealed BCNU-sensitive clones were further characterized for MGMT activity, protein, and mRNA levels.

By MGMT activity, Western blot, and Northern blot analyses, the HeLa/Rz clones sensitized to BCNU showed dramatic or intermediate decreases in MGMT activity, protein, and mRNA levels. It has been reported that the high immunoreactivity, prevalent cytoplasmic localization (>90%), and long half-life of MGMT molecules may mask the decrease in the fraction of MGMT protein actually involved in nuclear DNA repair (Citti et al., 1998). However, we observed a substantial decrease in MGMT protein and a concomitant reduction in enzyme activity, paralleled by the significant reduction of MGMT mRNA in the HeLa/Rz161 and HeLa/Rz212/178 clones. No cleaved MGMT mRNA was detected in the Northern analysis, likely due to the rapid degradation of the truncated mRNA molecules in the cellular environment. In other ribozyme studies, it is not uncommon for the truncated mRNA to elude detection (Potter et al., 1993). We are currently attempting to detect these fragments with reverse transcription-PCR.

Hypermethylation of the MGMT promoter is oftentimes thought to be associated with a loss of MGMT expression (Harris et al., 1991; Von Wronski et al., 1992; Watts et al., 1997). To exclude the possibility of intrinsic MGMT expression deficiency in HeLa/Rz clones due to promoter methylation, we analyzed the methylation status of one of the CpG hotspots within the MGMT gene promoter. Southern analysis showed that in all of the HeLa/Rz clones we examined, this CpG site was unmethylated, suggesting that hypermethylation of the MGMT promoter was not responsible for the down-regulation of MGMT in these cells.

The BCNU sensitivity of HeLa/Rz clones exhibiting ribozyme(s) expression was further characterized by colony formation, a sensitive and reliable cell assay for measuring cell survival. It was shown that BCNU-induced cell killing was increased by 2 to 3 logs in HeLa/Rz161 and HeLa/Rz212/178 clones compared with wild-type HeLa cells, when measured using a BCNU dose that is sublethal in the wild-type cells. It should be noted that one of the most BCNU-sensitive clones, HeLa/Rz212/178, contains two ribozyme constructs against nucleotides 178 and 212 in the 5' region of MGMT mRNA. However, two HeLa/Rz clones containing Rz178 alone showed only modest reductions in MGMT levels in this study. It is not known whether the significantly greater level of MGMT reduction in HeLa/Rz212/178 was due to an additive effect of Rz212 and Rz178 or the activity of Rz212 alone. We are currently investigating the role of Rz212 in BCNU sensitization.

Previously, Potter et al. (1993) reported that nucleotide 161 in the hMGMT mRNA was a target site for hammerhead ribozyme-mediated cleavage. Although that ribozyme was shown to be catalytically active, it was not used in studies to modulate CENU resistance. In our study, a ribozyme directed against nucleotide 161 was also found to be effective in down-regulating endogenous MGMT. Citti et al. (1999) have recently reported that a transient transfection into cells of a chemically synthetic hammerhead ribozyme, targeted against a GUU triplet in the hMGMT mRNA (nucleotide 743), could potentiate the genotoxicity of the alkylation damage induced by mitozolomide (Citti et al., 1999). In the current study, our combined approach identified two additional target sites in the MGMT mRNA sensitive to ribozyme attack, nucleotide 178 and 212. The ribozyme target sites, nucleotides 161, 178, and 212, are proximal to one another within the 5' region of the MGMT mRNA. Because the secondary and tertiary structures of the MGMT mRNA are unknown, we can only speculate whether this portion of the mRNA molecule is preferably accessible to hybridization with the ribozyme constructs we designed. Alternatively, this region of the MGMT mRNA molecule may be particularly susceptible to ribozyme-mediated cleavage based on the kinetics of ribozyme catalysis or dissociation. Due to the small number of positive clones (4 of 50), however, it is possible that this region of the molecule is over-represented simply by chance; and that if more positive clones had been obtained, others of the eight Rz sites may also have been found to be active. Experiments to distinguish among these possibilities are currently ongoing in our laboratory.

It is known that by complementary binding to the specific sequence of the target mRNA, antisense RNA can inhibit translation, thus down-regulating the synthesis of a specific gene product (Scanlon et al., 1995). On the other hand, after binding with the mRNA substrate, the ribozyme can catalyze the cleavage reaction at specific GUC triplet sites. Thus, as a powerful inhibitor designed against specific gene expression, the mechanism(s) of the hammerhead ribozyme could be either the catalytic endoribonuclease reaction at unique GUC triplet sites of the target mRNA, or the antisense inhibition of mRNA function, or both. To characterize the molecular mechanism of these ribozymes, further studies with catalytically inactive ribozymes, whose active counterparts were identified above, is currently underway in our laboratory.

The use of hammerhead ribozymes as modulators of specific gene expression has been investigated widely. The most active area of investigation into the application of ribozymes has been the development of gene therapy for cancer and acquired immunodeficiency syndrome. The molecular strategy developed here to deplete MGMT might be useful in future gene therapy strategies in which nonspecific toxic effects can be avoided by coupling ribozyme delivery with tumor-specific immunoliposomes or by using tumor-specific retroviruses and tumor-specific promoters to drive ribozyme expression (Shi and Pardridge, 2000). It is anticipated that the development of tumor-specific delivery systems by the many academic and biotech laboratories currently studying this important area will facilitate the clinical application of antisense and ribozyme therapeutic strategies in the future.