HER-2/neu Is Rate-limiting for Ovarian Cancer Growth

Amplification and overexpression of theHER-2/neu proto-oncogene frequently coincide with an aggressive clinical course of certain human adenocarcinomas. To assess whether HER-2/neu plays a rate-limiting role in ovarian cancer, we used human SK-OV-3 ovarian cancer cells as a model. We applied a conditional mRNA depletion strategy of HER-2/neu with anti-HER-2/neu-targeted hammerhead ribozymes expressed under the control of a tetracycline-regulated promoter system. In these ovarian cancer cells, we reduced HER-2/neu mRNA, protein expression, and tumor growth in nude mice by transfection with HER-2/neu-targeted ribozymes and generated cell lines expressing different levels of HER-2/neu. Expression of the most effective ribozyme (Rz3) quenched HER-2/neu mRNA levels by >90%. Concomitantly, fluorescence-activated cell sorting analysis revealed that expression of the HER-2/neu-encoded surface glycoprotein was almost completely abrogated. In nude mice, tumor growth was dramatically inhibited in the HER-2/neu-depleted Rz3-expressing SK-OV-3 cells. Furthermore, already established tumors started to regress when Rz3 expression was activated midstream by withdrawal of the tetracycline treatment. This study supports the thesis that HER-2/neu can be rate-limiting for the malignant phenotype of ovarian cancer in a gene dose-dependent manner.

The HER-2/neu proto-oncogene belongs to the epidermal growth factor receptor family and has been implicated in malignant transformation (reviewed in Ref. 1). HER-2/neu can be activated by at least three different genetic mechanisms including point mutation (2), gene amplification (3), and overexpression (4). These observations are relevant to human cancer because amplification and/or overexpression has been observed in 20 -30% of adenocarcinomas of the breast, ovary, lung, and stomach (reviewed in Ref. 5). Moreover, overexpression has been linked to an unfavorable prognosis in patients with breast (3) and ovarian (6) cancer.
Beyond this coincidence, it has been difficult to demonstrate directly that HER-2/neu is rate-limiting for tumor progression. One reason for this lack of understanding is that no ligand for HER-2/neu has been found, and HER-2/neu is now viewed merely as a signal-transducing subunit of epidermal growth factor and new differentiation factor/heregulin receptors (7). Depending on the cellular context, HER-2/neu-targeted antibodies can thus cause activating and inhibitory effects, which make it difficult to dissect the precise role of HER-2/neu. Recently, with a novel approach using single-chain antibodies that suppress cell-surface expression of HER-2/neu by retention in the endoplasmic reticulum, it was shown that reduction of HER-2/neu reversed the transformed phenotype of HER-2/ neu-transfected NIH/3T3 cells (8) and impaired growth factor signaling in T47D breast cancer cells (7,9). Here we utilize an independent approach by cleaving the HER-2/neu mRNA with specific ribozymes and thus deplete cells of the endogenous gene product. With this approach, the effects of a functional knockout can be studied in model cell lines, and thus, the contribution of a particular gene product delineated (reviewed in Ref. 10; see also . A major obstacle in achieving a constitutive ribozyme-mediated HER-2/neu depletion in stably transfected cells is that HER-2/neu expression may provide a growth advantage, thus making the selection of low expressing cell populations difficult. To circumvent this potential problem, we expressed hammerhead ribozymes under the control of a tetracycline-regulated promoter (15) to evaluate the effect of a conditional HER-2/neu depletion on in vitro and in vivo ovarian cancer cell growth. As a model, we used human SK-OV-3 ovarian cancer cells, which overexpress HER-2/neu due to a gene amplification event (16). We generated stably mass-transfected derivative cell lines that express anti-HER-2/neu-targeted ribozymes. Ribozyme expression almost completely abrogated HER-2/neu mRNA and protein expression, which resulted in a dramatic inhibition of tumor growth in nude mice. Furthermore, tumors that had been established in the absence of ribozyme expression started to regress when ribozyme expression was activated in vivo.

EXPERIMENTAL PROCEDURES
Plasmids and Generation of Constructs-Plasmids expressing the tetracycline transactivator (tTA) 1 /VP16 fusion protein (pUHG15-1 (15)) and the heptamerized tetracycline operator sequence (tet-O; pUHC13-3 (15)) were obtained from Dr. Bujard (Zentrum fü r Molekulare Biologie der Universitä t Heidelberg, Heidelberg, Germany). The ribozyme expression plasmid (pTET) was derived from pUHC13-3 and modified as described (17). The HER-2/neu-targeted hammerhead ribozymes Rz3 and Rz4 were designed and cloned as described (18). In brief, the following ribozyme coding sense and antisense oligonucleotides were annealed together and ligated into the HindIII restriction site of pTET: Rz3, 5Ј-agcttCCTGAAAGCTGATGAGTCCGTTAGGACG*AAAAAGT-CCTa-3Ј (sense) and 5Ј-agcttAGGACTTTTTC*GTCCTAACGGACTC-ATCAGCTTTCAGGA-3Ј (antisense); and Rz4, 5Ј-agcttCAAGACCAC-CTGATGAGTCCGTTAGGACGAAACCAGCAGa-3Ј (sense) and 5Ј- agcttCTGCTGGTTTCGTCCTAACGGACTCATCAGGTGGTCTTGa-3Ј (antisense) (with lower-case letters indicating HindIII restriction site overhangs, boldface upper-case letters showing HER-2/neu-specific antisense regions, and underlined upper-case letters indicating the hammerhead ribozyme core sequence). The resulting ribozyme expression plasmid (pTET/Rz3) contains HER-2/neu-specific antisense flanking regions of 8 nucleotides on both ends of the 22-nucleotide catalytic hammerhead ribozyme core sequence that target it to a central region in the HER-2/neu RNA just downstream of the transmembrane region of the protein. In addition, a control ribozyme was constructed (pTET/ Rz3mu) that lacks catalytic activity due to a single base pair mutation in the ribozyme core region (G to T exchange; * indicates the position of the mutated G). The ribozyme pTET/Rz4 contains HER-2/neu-specific antisense flanking regions of 9 and 8 nucleotides on the 5Ј-and 3Ј-ends, respectively, of the 22-nucleotide catalytic ribozyme core that target Rz4 to its cleavage site 1991 nucleotides downstream of the translation initiation site in the HER-2/neu RNA. Correct sequences of the ribozymes were verified by DNA sequencing, and specific catalytic ribozyme activity was demonstrated in in vitro cleavage assays as described (18).
Cell Lines and Transfections-Human ovarian cancer cells (SK-OV-3) were obtained from American Type Culture Collection and were maintained in continuous culture at 37°C in 5% CO 2 using Iscove's modified Eagle's medium (Life Technologies, Inc.) supplemented with glutamine, 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. SK-OV-3 cells were transfected using LipofectAMINE (Life Technologies, Inc.). Briefly, cells at 50 -70% confluence were incubated for 5 h with plasmid DNA mixed with Lipo-fectAMINE (7 l of LipofectAMINE/1 g of plasmid DNA) in serum-free medium (Opti-MEM, Life Technologies, Inc.) at 37°C in 5% CO 2 . The transfection medium was then replaced with normal growth medium and 36 h later supplemented with the respective drugs for selection of stable cell lines. SK-OV-3 cells stably expressing tetracycline-regulated HER-2/neu-targeted ribozymes were generated in a two-step transfection protocol. In the first step, SK-OV-3 cells were transfected with 10 g of pUHG15-1 plasmid DNA and 1 g of pcDNA3 plasmid DNA (Invitrogen, San Diego, CA) to provide G418 resistance. After selection for stable cell lines in the presence of G418 at 0.7 mg/ml, individual tTA-expressing clones were isolated with cloning rings. To test the clones for tTA expression and tetracycline regulation, the cells were transiently transfected in the absence and presence of 1 g/ml tetracycline (Sigma) with pUHC13-3 plasmid DNA containing a luciferase cDNA under the control of the tet-O-binding site (15). Cell lysates were prepared 36 h after the transfections, and luciferase activities were measured in a luminometer as described (17). Two clones (SK-OV-3/ tTA-7 and SK-OV-3/tTA-2) that demonstrated the best tetracycline regulation of luciferase activity were used for further transfections with the ribozyme expression plasmids. SK-OV-3/tTA-7 and SK-OV-3/tTA-2 cells were then transfected with 10 g of pTET/Rz3, pTET/Rz3mu, or pTET/Rz4 mixed with 1 g of pZero (Invitrogen) to provide Zeocin resistance. Mass-transfected derivative cell lines (SK-OV-3/Rz3, SK-OV-3/Rz3mu, and SK-OV-3/Rz4) were obtained after selection with 0.4 mg/ml Zeocin and 1 g/ml tetracycline.
Northern Analysis-Total cellular RNA was isolated with the RNA STAT-60 method (Tel-Test, Friendswood, TX), and 30 g were separated and blotted as described (19). A HER-2/neu cDNA probe (1.5kilobase pair EcoRI fragment) was hybridized, washed, and exposed to film for 16 h (19). To correct for variability in loading, blots were stripped, reprobed with a glyceraldehyde-3-phosphate dehydrogenase cDNA probe (CLONTECH), and exposed to film for 1 h. Relative band intensities were measured by densitometry.
Fluorescence-activated Cell Sorting (FACS)-To quantitate HER-2/ neu protein levels by FACS analysis, cells were trypsinized, washed once with growth medium containing serum and twice with phosphatebuffered saline (Sigma), and resuspended in phosphate-buffered saline at 5 ϫ 10 5 cells/100 l. The cells were incubated for 30 min at 4°C with a 1:100 dilution of a primary anti-human HER-2/neu mouse monoclonal antibody (clone 9G6.10; Neomarkers, Fremont, CA). Cells were washed twice with phosphate-buffered saline and incubated for 30 min at 4°C in the dark with a 1:200 dilution of a fluorescein isothiocyanate-labeled goat anti-mouse secondary antibody (Boehringer Mannheim). After two final washes with phosphate-buffered saline, the mean value of fluorescence intensity of 10,000 cells was determined by FACS (FACStar Plus, Becton Dickinson). Unlabeled cells and cells labeled with secondary antibody alone served as negative controls.
Tumor Growth in Animals-Female athymic nude mice (NCr nu/nu; NCI, Frederick, MD) were injected subcutaneously with 1 ϫ 10 6 cells in 100 l of Iscove's modified Eagle's medium (three mice/group and two injection sites/mouse). Mice in group 1 were injected with SK-OV-3/ tTA-7 control cells, and those in groups 2 and 3 with SK-OV-3/Rz3 ribozyme-expressing cells. In the mice in group 3, slow-release tetracycline pellets (Innovative Research of America), which release 0.7 mg of tetracycline/day, were implanted subcutaneously at the day of tumor cell injection. Tumor growth was monitored for up to 2 months, and tumor sizes were estimated from the product of the perpendicular diameters of the tumors. In one mouse in group 3, the tetracycline pellet was removed after 6 weeks to activate ribozyme expression in established tumors, and tumor growth was monitored for an additional 2 weeks. In a separate study, nude mice were injected subcutaneously with 2 ϫ 10 6 SK-OV-3/Rz4 cells in 100 l of Iscove's modified Eagle's medium (five mice/group and two injection sites/mouse). Tumor growth was monitored for 4 weeks in the presence (group 1) or absence (group 2) of tetracycline.
Data Analysis-Means Ϯ S.E. are depicted unless indicated otherwise. Student's t test or analysis of variance for repeated measures (Statview 4.02, Abacus Concepts, Inc.) was used for comparisons between data sets, and p Ͻ 0.05 was considered significant.

RESULTS
Generation of tTA-expressing SK-OV-3 Cells-To avoid promoter interference and to generate cells in which tet-O-controlled transgene expression can be tightly regulated by tetracycline, a two-step transfection protocol was used as originally described (15). In the first step, SK-OV-3 cells were transfected with pUHG15-1 plasmid DNA, and individual clones were screened for tetracycline regulation of tTA-driven tet-O/luciferase expression. Individual tTA-expressing clones were transiently transfected with pUHC13-3 plasmid DNA, coding for luciferase under the control of the tet-O-binding site, and luciferase activity was measured in the absence and presence of tetracycline. In the absence of tetracycline, luciferase activity was high in all clones. The best tetracycline regulation was observed in clone 7 (SK-OV-3/tTA-7), where tTA-driven luciferase activity was repressed by Ͼ90% in the presence of tetracycline (Fig. 1). Clones SK-OV-3/tTA-7 and SK-OV-3/tTA-2 were then used for further transfections with pTET/ribozyme expression plasmids. It is important to note that luciferase activity was not completely repressed to background levels, thus demonstrating some residual promoter activity in these cells. HER-2/neu expression as measured by FACS analysis did not vary between the clones, indicating that clonal selection and/or tTA expression had no significant effects on HER-2/neu expression in SK-OV-3 cells (data not shown).
Efficacy of HER-2/neu-targeted Ribozymes in Cells in Vitro-SK-OV-3/tTA-7 cells were transfected with pTET/Rz3 or pTET/Rz3mu plasmids, and mass-transfected cell pools were selected in the presence of tetracycline to repress ribozyme expression during the selection period. After 6 weeks, stably transfected cells were grown in the absence or presence of tetracycline, and HER-2/neu mRNA levels were quantified by Northern analysis (Fig. 2). HER-2/neu-targeted pTET/Rz3 ribozyme expression (Fig. 2, lane 2 (Ϫtet ϭ ribozyme on)) depleted endogenous HER-2/neu mRNA levels by 90% compared with control cells. Inactivation of ribozyme expression (Fig. 2,  lane 3) was followed by a 4.5-fold increase in HER-2/neu mRNA levels, demonstrating the specificity of the ribozyme effect. Expression of the pTET/Rz4 ribozyme construct was less efficacious and reduced HER-2/neu mRNA levels in a tetracyclinereversible manner by ϳ50% (data not shown).
In the next experiment, ribozyme-mediated depletion of HER-2/neu protein expression was assessed by FACS analysis. Cell-surface expression of the HER-2/neu protein was suppressed by 90% when cells were grown in the absence of tetracycline (Fig. 3A, Rz3-on). HER-2/neu levels almost completely reverted to control values when tetracycline was added to the medium. In addition, a catalytically inactive mutant ribozyme (SK-OV-3/Rz3mu) had no significant effects on HER-2/neu mRNA or protein expression (data not shown), which clearly indicates that cleavage of HER-2/neu RNA and not antisense inhibition is the main mode of ribozyme action. Expression of pTET/Rz4 reduced HER-2/neu protein levels by 50% (Fig. 3B), correlating very well with the results from the Northern analysis.
In parallel experiments with other cell lines, we verified the efficacy of the anti-HER-2/neu-targeted ribozymes. Ribozyme targeting caused a reduction of HER-2/neu expression in MKN7 gastric cancer, MDA-MB-361 breast cancer, and Colo357 pancreatic cancer cells (data not shown). This strongly suggests that this targeting approach will be useful in a variety of human adenocarcinomas.
The biological significance of a HER-2/neu depletion on in vitro growth of SK-OV-3 cells was assessed by anchorage-dependent as well as anchorage-independent growth assays. HER-2/neu-targeted ribozyme expression did not alter cell morphology or anchorage-dependent proliferation. However, anchorage-independent growth in soft agar was inhibited by Ͼ90% in SK-OV-3/Rz3 cells and was restored to control levels by the addition of tetracycline (data not shown). This demonstrates that HER-2/neu is a rate-limiting factor for anchorageindependent growth of SK-OV-3 cells.

Down-regulation of HER-2/neu Expression Inhibits Tumor
Growth in Vivo in Nude Mice-When injected subcutaneously at 1 ϫ 10 6 cells/injection site into nude mice, SK-OV-3/tTA-7 control cells grew to a mean tumor size of 100 Ϯ 20 mm 2 within 44 days (Fig. 4A, q). In contrast, tumor growth of ribozymeexpressing pTET/Rz3 cells was significantly inhibited (p Ͻ 0.01, Student's t test), and tumors did not grow beyond a very small inoculum size of 6 Ϯ 2.6 mm 2 (Fig. 4A, É). In animals that were treated with slow-release tetracycline pellets to turn the ribozyme off in vivo, tumor growth started to increase after 32 days, and mean tumor sizes were 27 Ϯ 7.8 mm 2 (Fig. 4A, å; significantly different from ribozyme-expressing tumors, p Ͻ 0.05, Student's t test). Tumor growth of SK-OV-3 cells transfected with the catalytically inactive ribozyme (pTET/Rz3mu) was not significantly different compared with control cells (data not shown).
In a subset of this study, we addressed the question of whether ribozyme-mediated abrogation of HER-2/neu expression can cause regression of established tumors. Ribozyme expression was activated in the tumor cells by removal of the tetracycline pellet after 44 days, and tumor growth was monitored for another 2 weeks. While the tumors in the tetracyclinetreated animals (ribozme off) grew continuously and reached a size of 76 Ϯ 20 mm 2 (Fig. 4B, solid lines), the tumors in which the ribozyme was activated by removal of tetracycline started to regress and reached a size of 20 Ϯ 3.5 mm 2 (Fig. 4B, dotted  lines). This tumor regression was significant (p Ͻ 0.05, analysis and anti-HER-2/neu-targeted Rz3 ribozyme-transfected cells grown in the absence (lane 2) and presence (lane 3) of tetracycline (tet) was analyzed by Northern analysis. HER-2/neu mRNA was quantified relative to glyceraldehyde-3-phosphate dehydrogenase (G3PDH) mRNA by densitometry. The inset shows a representative Northern blot. The positions of HER-2/neu-and glyceraldehyde-3-phosphate dehydrogenase-specific transcripts are indicated. The HER-2/neu Northern blot was exposed to film for 16 h, whereas the glyceraldehyde-3-phosphate dehydrogenase Northern blot was exposed for only 1 h. of variance for repeated measures) and suggests that HER-2/ neu expression is important for continuous growth of ovarian cancer cells in vivo. Furthermore, disruption of HER-2/neu not only arrested tumor growth, but induced regression.
To test if SK-OV-3 tumor growth correlates with the level of HER-2/neu expression, we used SK-OV-3/Rz4 cells, in which, in contrast to SK-OV-3/Rz3 cells, HER-2/neu protein levels are reduced by ϳ50% (see Fig. 3B). When injected subcutaneously at 2 ϫ 10 6 cells/injection site in the presence of tetracycline, SK-OV-3/Rz4 cells grew to a mean tumor size of 54 Ϯ 7.2 mm 2 within 4 weeks (Fig. 4C, ç). Inactivation of ribozyme expression by tetracycline significantly inhibited tumor growth by 43% (Fig. 4C, É; p Ͻ 0.05, Student's t test), which correlated very well with the level of HER-2/neu reduction in these cells.
The overall faster tumor growth in this experiment compared with the previous study (see Fig. 4A) was due to the higher tumor cell inoculum (2 ϫ 10 6 versus 1 ϫ 10 6 ). This result further supports the thesis that HER-2/neu expression is ratelimiting for SK-OV-3 tumor growth in a gene dose-dependent manner. DISCUSSION The phenomenology of HER-2/neu overexpression in human cancer has been well studied, and particularly the easily accessible cell-surface localization made it an excellent target for antibody-based immunotherapies. HER-2/neu-specific monoclonal antibodies (20) and recombinant immunotoxins (21), both of which can inhibit in vitro and in vivo growth of transformed cells, have been described. Despite the clinical usefulness of these approaches, it remains unclear how expression of an apparently normal gene product affects tumorigenesis and tumor progression in vivo. Potentially more specific genetic targeting strategies such as antisense oligonucleotides have been used with limited success (22), probably due to their relatively low efficacy and specificity.
In this study, we expressed HER-2/neu-targeted hammerhead ribozymes under the control of a tetracycline-regulated promoter system in human SK-OV-3 ovarian cancer cells, which express HER-2/neu spontaneously at high levels due to gene amplification. SK-OV-3 cells provide an attractive model since the epidermal growth factor receptor is the only other human epidermal growth factor receptor expressed in these cells, which therefore proliferate in response to epidermal growth factor, but not to new differentiation factor/heregulinlike growth factors (data not shown).
To achieve a conditional HER-2/neu down-regulation, we employed a binary tetracycline-regulated gene expression system in which hammerhead ribozyme expression can be inactivated in vitro and in vivo by the addition of the nontoxic antibiotic tetracycline. The major advantage over a constitutive ribozyme expression system is that the effects of ribozymemediated down-regulation of HER-2/neu expression can be evaluated in one genetically identical cell line, thus preventing interferences of different genetic backgrounds between various cell lines. Furthermore, HER-2/neu can be down-regulated at different time points in tumor growth in animals, which allows the identification of the tumor stages that are most susceptible to anti-HER-2/neu-targeted therapies.
In stably mass-transfected SK-OV-3 cells, Rz3 ribozyme expression depleted HER-2/neu mRNA and protein levels by Ͼ90%, and inactivation of ribozyme expression by tetracycline reversed the effects. This inhibition is even more remarkable taking into account that mass-transfected cells (and not clonal subpopulations) were used and that HER-2/neu is expressed at very high levels in SK-OV-3 cells. This suggests that ribozymes can be more effective than antisense oligonucleotides that reduced HER-2/neu expression by only 50% in selected clones of SK-BR-3 breast cancer cells (22), which express HER-2/neu at comparable levels. This dramatic inhibition enabled us for the first time to study the effects of a functional and conditional HER-2/neu knockout on in vivo tumor growth of ovarian cancer cells. Tumor growth of ribozyme-expressing cells was almost completely abrogated, and inactivation of ribozyme expression by tetracycline partially reversed SK-OV-3 tumor growth. Inactivation of ribozyme expression by tetracycline was not as effective as in the cell culture experiments, which can potentially be explained by an insufficient tetracycline delivery to the tumor cells in vivo. Alternatively, in vivo tumor growth might be more susceptible to residual ribozyme activity and hence small reductions in HER-2/neu expression.
In summary, our data demonstrate that HER-2/neu has a FIG. 4. Subcutaneous tumor growth in nude mice of SK-OV-3 ovarian cancer cells producing different levels of HER-2/neu. A, 1 ϫ 10 6 cells were inoculated subcutaneously (three mice/group and two injection sites/mouse), and tumor growth was monitored for 6 weeks in SK-OV-3/tTA control cells (q) or pTET/Rz3 ribozyme-transfected cells in the presence (Rz3-off; ç) and absence (Rz3-on; É) of tetracycline treatment. B, after 44 days, growth curves of the individual tumors in the absence (rz-on; dotted lines) and continued presence (rz-off; solid lines) of tetracycline were monitored for another 2 weeks. C, SK-OV-3 cells transfected with the pTET/Rz4 ribozyme expression construct were inoculated subcutaneously at 2 ϫ 10 6 cells/injection site (five mice/group and two injection sites/mouse), and tumor growth was monitored for 4 weeks in the presence (Rz4-off; ç) and absence (Rz4-on; É) of tetracycline. rate-limiting role in ovarian cancer and that ribozyme targeting can cause regression of established tumors. The conditional HER-2/neu depletion in vitro and in vivo will enable further elucidation of the role of HER-2/neu in the pathogenesis of human cancer.