Adenovirus-mediated Human Topoisomerase IIα Gene Transfer Increases the Sensitivity of Etoposide-resistant Human Breast Cancer Cells¹ Free

Drug resistance continues to be a serious problem in cancer therapy. Cellular resistance has been attributed to several different mechanisms, including alteration of the cell’s topo IIα³ (1). The topo IIα nuclear enzyme is essential to the survival of eukaryotic cells (2, 3, 4, 5). It is involved in DNA replication, chromosome segregation, and other essential cellular processes. Topo II is also a target for a variety of clinically important antineoplastic drugs, including etoposide (VP-16) and Adriamycin. Topo II-targeted drugs act as topo II poisons by stabilizing the enzyme-DNA cleavable complex. This stabilization initiates a biochemical cascade, leading to cell death. The resistance to topo II-targeted drugs involves quantitative and/or qualitative alterations in topo II gene expression. Mutations in the gene are associated with the production of a mutated topo II protein that is less able to cleave DNA in the presence of antineoplastic agents (6). Low cellular levels of topo II lead to reduced formation of drug-stabilized topo II-DNA cleavable complexes, which quantitatively correlate with cell death (7). The cytotoxicity of topo II-targeted drugs tends to increase as cellular topo II content increases. MDA-VP human breast cancer cells, which are resistant to VP-16, express lower amounts of topo IIα than do MDA parent cells (8). Therefore, elevation of the topo IIα level could theoretically result in increased drug sensitivity.

Our previous studies have shown that transient transfection with a vector containing either the Drosophila or human topo IIα gene into drug-resistant tumor cells enhanced their drug sensitivity (8, 9). Transient transfection of MDA-VP cells with a glucocorticoid-inducible vector carrying the human topo IIα gene increased human topo IIα mRNA and protein levels, increased cleavable complex formation, and enhanced sensitivity to the topo II-reactive drugs VP-16, doxorubicin, and amsacrine (8). Although this vector was suitable for in vitro testing, highly efficient gene transfer is mandatory for in vivo testing. Adenovirus vectors have a number of advantages for gene delivery, including high titer production and high transfection efficiency. Adenovirus can infect and direct high levels of protein expression in both proliferating and quiescent cells, an important feature for use in vivo (10). Recombinant adenoviral vectors containing various genes are already being used in clinical trials (11).

We constructed a recombinant adenovirus containing the human topo IIα gene (Ad-hTopoIIα). Infection of this adenovirus into MDA-VP cells significantly enhanced their sensitivity to VP-16. Our data also indicated that the drug-sensitizing effect of Ad-hTopoIIα infection was selective and that infection itself was not toxic to normal human cells. The sensitizing effect of Ad-hTopoIIα appears to depend upon the original reduced expression level of the gene in the resistant target cells.

Human embryonic kidney cells transformed with adenovirus type 5 (293 cells) were obtained from American Type Culture Collection (Manassas, VA). The etoposide (VP-16)-resistant human breast cancer cell line MDA-VP was initially derived and cloned from the MDA-MB-231 parent cell line by Dr. T. Fojo (National Cancer Institute, Bethesda, MD; Ref. 12). HEL and HFC were described previously (13). All cells were free of Mycoplasma, as screened by Gen-Probe (Gen-Probe Co., San Diego, CA).

Nuclei were isolated, and extracts were prepared by incubation in 350 mm NaCl, as described previously (3). Serial dilutions of nuclear extracts from MDA and MDA-VP cells were used to measure the catalytic activity of topo IIα in decatenation assays, as described previously (3). The concentration of nuclear extract that resulted in 50% decatenation of kDNA was defined as 1 unit of decatenating activity.

VP-16-induced DNA double-strand cleavage activity was measured as reported previously (14). Briefly, SV40 form I DNA (Life Technologies, Inc., Gaithersburg, MD) was incubated with nuclear extract containing either 0.5 or 2.5 units of decatenating activity in the presence of either 10 μl of DMSO or 50 μm VP-16. The generation of form II (nicked) or form III (linearized) DNA, as visualized on ethidium bromide-stained agarose gels, is an indication of DNA-cleaving activity.

Cells (5 × 10⁶) were radiolabeled with [³H]thymidine (ICN Biomedicals Inc., Irvine, CA) and [¹⁴C]leucine (Amersham, Arlington Heights, IL) for 24 h at 37°C. The cells were then incubated for 90 min to 2 h with varying doses of VP-16. DNA-protein complexes were covalently linked by the addition of SDS, and then KCl was added to precipitate proteins as described previously (3). The extent of cleavable complex was determined by calculating the ³H-labeled DNA:¹⁴C-labeled protein ratio in the precipitated DNA-protein complexes.

SV40 DNA was uniquely 3′-end labeled with [α-³²P]dATP (Amersham). Nuclear extracts containing 2.5 units of decatenating activity were incubated with 0.01 μg of end-labeled SV40 DNA and varying doses of VP-16, as described previously (14). Data were expressed as cpm of drug-treated samples minus cpm of untreated samples.

The human topo IIα gene was excised from pBS-hTopII plasmid (American Type Culture Collection). To subclone this cDNA, we performed PCR using forward primer 5′-GAAGATCTTCGCCGCCACCATGGAAGTGTCACCATT-3′, which contains a BglII site and Kozak sequence (15), and reverse primer 5′-ACAAGACATTTTTTGGGTCCCT-3′. After PCR, the products (0.3 kb) were digested with BglII/Eco01091. The second fragment of human topo IIα gene was obtained by Eco01091/ClaI digestion of pBS-hTopII. The adenoviral vector pAvCvSv, constructed using plasmid pXCJL-1 by addition of the human cytomegalovirus promoter and the SV40 early polyadenylation signal (16, 17), was digested with BglII/ClaI. The shuttle vector pAvCvSv-hTopIIα was constructed by subcloning the human topo IIα fragment into the pAvCvSv vector using a Fast-Link DNA ligation and screening kit (Epicentre Technologies Co., Madison, WI). Every junction and fragment was sequenced to confirm correct ligation. The packaging vector pBHG10 (Microbix Biosystem Inc., Ontario, Canada; Ref. 18) and the shuttle vector pAvCvSv-hTopIIα were then cotransfected into 293 cells by DOTAP liposome transfection-mediated reagent (Boehringer Mannheim Corp., Indianapolis, IN). Infectious recombinant adenovirus plaques were picked 10–14 days after transfection and then propagated and screened. The viral DNA was purified, and the presence of the human topo IIα gene was confirmed by PCR and restriction enzyme digestion (19). PCR was performed using primers 5′-GTGTGGAACTAGAAGGC-3′ and 5′-GGAGGTGGAAGACTGAC-3′ (for the topo IIα gene) and 5′-TCGTTTCTCAGCAGCTGTTG-3′ and 5′-CATCTGAACTCAAAGCGTGG-3′ (for the adenovirus E2B fragment). PCR products were resolved in 1% agarose gel, stained with ethidium bromide, and visualized under UV light.

Ad-hTopoIIα virus was propagated in 293 cells and twice purified by cesium chloride gradient centrifugation (10). The virus titers were determined by plaque assay. Ad-β-gal was used as a control. Cells in logarithmic-growth phase were infected with Ad-hTopoIIα or Ad-β-gal at various multiplicities of infection (1–1000 pfu/cell) for 48 h and then treated with VP-16 or harvested for various assays.

Total RNA was extracted by Trizol Reagent (Life Technologies, Inc., Grand Island, NY). RNA (20 μg) was electrophoresed on 1% formaldehyde/agarose gel and transferred to a Hybond-N⁺ membrane (Amersham). A human topo IIα gene probe (ZII69), a generous gift of Dr. L. Liu (Robert Wood Johnson Medical School, University of Medicine and Dentistry of New Jersey, Piscataway, NJ; Ref. 20), and β-actin and GAPDH probes were labeled using the Rediprime DNA labeling system (Amersham).

Cells (2 × 10⁶) were seeded 1 day before treatment and then incubated with Ad-hTopoIIα or Ad-β-gal for 48 h. Cells were washed with cold PBS and lysed with buffer containing the protease inhibitors aprotinin (2 μg/ml), leupeptin (2 μg/ml), pepstatin A (1 μg/ml), and phenylmethysulfonyl fluoride (100 μg/ml). Lysates were passed 10 times through a 25-gauge needle; 50 μg of protein were then solubilized in SDS sample buffer (8), boiled for 5 min before loading onto a 7.5% SDS-polyacrylamide gel, and then transferred to a nitrocellulose membrane. Specific protein detection was performed with a human topo IIα polyclonal antibody (TopoGEN Inc., Columbus, OH) and β-actin monoclonal antibody (Amersham) using the ECL Western blotting analysis system (Amersham) according to the manufacturer’s instructions. Densitometric analysis was performed, and values were normalized with β-actin densities.

Cells (3000–5000 per 100 μl) were seeded into 96-well cell culture plates and allowed to adhere overnight. Cells were infected with the designated concentration of adenovirus for 48 h. Various doses of VP-16 were then added in triplicate. The antiproliferative activity was determined by MTT assay 48 h later, as described previously (21).

The band depletion-immunoblotting assay was performed as described previously (3). Cells (10⁶) were infected with 100 pfu/cell of Ad-hTopoIIα or Ad-β-gal for 48 h before the addition of 200 μm VP-16 and further incubation at 37°C for 1 h. Cells were then lysed and boiled for 2 min. The lysate proteins were electrophoresed in SDS-polyacrylamide gel and immunoblotted using the human topo IIα polyclonal antibody.

To verify that the newly constructed Ad-hTopoIIα adenovirus contained both the human topo IIα gene and the appropriate adenoviral sequences, two pairs of primers were chosen for PCR. Use of pBS-hTop II (Lane 1) and pBHG10 (Lane 2) as positive control templates resulted in 530- and 860-bp bands identifying topo IIα and adenoviral sequences, respectively (Ref. 22; Fig. 1). Both bands are present in Lane 3, with Ad-hTopoIIα viral DNA as the template for PCR. Water served as the negative control (Lane 4). Restriction mapping further verified this result (data not shown). To exclude the contamination of wild-type adenovirus in the generation of recombinant Ad-hTopoIIα virus, we used a specific E1A primer to detect wild-type adenovirus for PCR. The negative PCR result (data not shown) indicated that no wild-type adenovirus was present in the Ad-hTopoIIα virus preparations (23).

MDA and MDA-VP cells were characterized previously with respect to their sensitivity to VP-16, ADR, amsacrine and cisplatin ⁽⁸⁾. Cross-resistance to the three topo II inhibitors was observed, with no cross-resistance to the DNA-damaging agent cisplatin. MDA and MDA-VP cells do not express mdr-1 or P95, and they express equivalent levels of MRP and bcl-2 (data not shown). Intracellular drug accumulation was found to be similar between the two cell lines (8).

The most significant difference between the two cell lines was the level of topo IIα expression (Fig. 2). MDA-VP cells had a >85% reduction in steady-state mRNA expression levels compared with the MDA parent cells using densitometric analysis (Fig. 2,B). Western blot analyses confirmed these results, also showing a >85% reduction in the topo IIα protein level in MDA-VP cells, compared with the MDA parent cells (Fig. 3). Topo II-DNA complex formation in the presence of VP-16 was quantified using an SDS-KCl assay (3). As shown in Fig. 4, a direct relationship was observed between complex formation in the two cell lines and their levels of topo IIα enzyme. This assay demonstrated that, in the MDA-VP cells, VP-16 produced ∼20% of the amount of cleavable complex detected in MDA parent cells. Considering that the value of 1 in Fig. 4 represents no effect, at 100 μm VP-16, the MDA-VP cells had a value of 2 or 1 above baseline, an 80% decrease compared with the value seen with the MDA parental cells (5 above baseline). The data in Figs. 2 and 3 indicate an 85% decrease in topo IIα mRNA and protein, very comparable with the 80% decrease in cleavable complex formation.

A decatenation assay using equal amounts of nuclear protein from the two cell lines was performed. This assay showed the conversion of catenated kDNA, a form that cannot enter the gel, into circular decatenated DNA, which can enter the gel. The amount of nuclear extract required to decatenate 50% of the kDNA was defined as 1 unit (Fig. 5). Using this assay, we determined that MDA-VP cells required 0.54 μg of nuclear extract protein to decatenate 50% of the kDNA, whereas MDA parent cells required only 0.31 μg. Therefore, the topo IIα activity per total microgram of protein in the MDA-VP cells was ∼43% of that in the parent cells. This finding of decreased topo IIα activity per unit of nuclear protein was consistent with the mRNA and protein data.

To rule out the possibility that MDA-VP cells contained a drug-resistant form of the topo IIα protein, we used equal units of topo IIα decatenating activity from MDA and MDA-VP cells to quantify VP-16-induced DNA double-strand cleavage of SV40 DNA. This cleavage was detected using an agarose gel system to demonstrate the linearization of the supercoiled and single-nicked SV40 DNA (forms I and II) by topo IIα-containing nuclear extracts in the presence of 50 μm VP-16. Comparable amounts of DNA double-strand cleavage (form III) were seen when equal decatenating units of topo IIα-containing nuclear extracts from MDA and MDA-VP cells were incubated in the presence of VP-16 (data not shown). This qualitative assay demonstrated that topo IIα from MDA parent and MDA-VP cells cleaved DNA similarly in the presence of VP-16. To quantitate the activity of drug-induced DNA double-strand cleavage and protein cross-linking by topo IIα from the two cell lines, we used [α-³²P]dATP end-labeled SV40 DNA. In Fig. 6, 3′ end-labeled SV40 DNA was used to perform a modified SDS-KCl assay to quantitate the amount of radioactive DNA that coprecipitated with the protein in the presence of varying doses of VP-16. This assay showed that drug-induced DNA cleavage by equal decatenating activities of topo IIα-containing nuclear extracts from the MDA and MDA-VP cells was approximately the same. These findings suggest that there are no mutations or alterations in the MDA-VP topo IIα that affect its ability to interact with VP-16.

To determine whether the expression of human topo IIα mRNA in etoposide-resistant human breast cancer cells could be increased, MDA-VP cells were infected with Ad-hTopoIIα or Ad-β-gal at a multiplicity of infection of 100 pfu/cell for 48 h. The level of topo IIα expression was significantly increased in MDA-VP cells after infection with Ad-hTopoIIα but not Ad-β-gal (Fig. 2,A, Lanes 3 and 4). The expression of topo IIα in the MDA-VP cells increased ∼7.4-fold, from 11.9 to 87.9%. The increased expression level was dose-dependent up to 100 pfu per cell. No further increase was seen using >100 pfu per cell (data not shown). Infection of MDA parent cells with Ad-hTopoIIα resulted in no significant increase in topo IIα mRNA (Fig. 2, C and D).

As shown in Fig. 3, the M_r 170,000 human topo IIα protein signal (24, 25) for MDA-VP cells is only 13.7% of that expressed in MDA parent cells. The protein level was significantly elevated in MDA-VP cells after infection with Ad-hTopoIIα. Relative density analysis indicated a 5.9-fold increase (Lane 3), compared with uninfected MDA-VP cells (Lane 2). Ad-β-gal did not significantly increase the topo IIα protein level in MDA-VP cells (Lane 4).

To determine whether the exogenous human topo IIα protein produced after Ad-hTopoIIα infection was sensitive to VP-16, the band depletion-immunoblotting assay was performed using a specific human topo IIα antibody. After SDS denaturation of VP-16-treated cells, drug-sensitive human topo IIα protein forms a covalently linked complex with cellular DNA. Therefore, exposure of the cellular topo IIα to VP-16 should result in depletion of the immunologically detectable band if the protein is drug sensitive because the binding of topo II to the DNA prevents it from migrating through the SDS-PAGE system (3). As shown in Fig. 7 (Lanes 1–4, left to right), human topo IIα protein from both MDA parent and MDA-VP cells was sensitive to VP-16. The level of topo IIα protein was elevated 6-fold in the MDA-VP cells after infection with Ad-hTopoIIα (Fig. 7, Lane 5), compared with either control cells (Lane 3) or Ad-β-gal-infected cells (Lane 7). Depletion of this band from Ad-hTopoIIα-infected cells treated with VP-16 indicated that the exogenous human topo IIα protein produced after infection was drug sensitive.

MDA-VP cells were less sensitive to the cytotoxic actions of VP-16 than were MDA parent cells, with IC₅₀ of 45 and 3 μm, respectively. As shown in Fig. 8, infection with Ad-hTopoIIα significantly increased the sensitivity of the MDA-VP cells in a dose-dependent manner. Treatment with 10 μm VP-16 yielded only 13% cytostasis in uninfected MDA-VP cells. This cytostatic activity was increased 4.5-fold to 58% by 100 pfu/cell Ad-hTopoIIα. No further increase was seen using 1000 pfu/cell. In contrast, Ad-β-gal had no effect on cell sensitivity to VP-16. There was no significant difference in cytopathic effect after exposure to Ad-hTopoIIα alone, compared with Ad-β-gal alone at viral doses below 100 pfu/cell. Thus, transfer of the human topo IIα gene using the Ad-hTopoIIα vector increased not only human topo IIα mRNA expression and topo IIα protein production but also cellular sensitivity to the cytotoxic actions of VP-16.

To determine the effect of Ad-hTopoIIα on normal human cells, we infected HFC and HEL cells with 100 pfu/cell Ad-hTopoIIα or Ad-β-gal, and cell sensitivity to 10 μm VP-16 was subsequently quantified. Both cell lines were relatively resistant to VP-16, with 9 and 17% cytostasis, respectively, after exposure to 10 μm alone (Fig. 9). No significant cytopathic effect was observed after infection with either Ad-hTopoIIα or Ad-β-gal alone. In contrast to our finding using MDA-VP cells, neither Ad-hTopoIIα nor Ad-β-gal increased the sensitivity of HEL or HFC cells to VP-16 (Fig. 9).

Total RNA and protein were extracted from HEL and HFC cells 48 h after infection with Ad-hTopoIIα or Ad-β-gal. The expression of human topo IIα mRNA and protein was analyzed by Northern blot and Western blot analyses. Averages from three independent experiments are shown in Fig. 10. Expression of human topo IIα mRNA and protein in HEL cells infected with Ad-hTopoIIα was only slightly elevated. There was also a small increase in human topo IIα protein in the HFC after infection with Ad-hTopoIIα. However, Ad-β-gal infection also resulted in slightly increased human topo IIα mRNA and protein production, indicating that the lack of increased sensitivity could result from failure of the vector to substantially increase topo IIα mRNA and protein production in the normal cells. Staining with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside in Ad-β-gal-infected cells confirmed viral uptake in both HEL and HFC. Therefore, lack of enhanced topo IIα expression and protein production was not due to the inability of the adenoviral vector to infect these normal cells.

This study demonstrated that the sensitivity of VP-16-resistant breast cancer cells can be increased by infection with a recombinant adenovirus expressing the human topo IIα gene. The adenoviral system provides a method of delivering a wild-type topo IIα gene to tumor cells with a mutated drug-resistant enzyme or to cells with decreased enzyme levels. Our studies provided evidence that drug resistance to topo II inhibitors can be circumvented by topo IIα gene transfer.

MDA-VP cells are resistant to the cytotoxic actions of VP-16 with an IC₅₀ of 45 μm compared with 3 μm for the MDA parent cells and are also cross-resistant to doxorubicin and amsacrine (8). Neither MDA-VP nor MDA parent cells express mdr-1 or P95 (8). Drug accumulation studies using [³H]VP-16 revealed no difference in the accumulation of intracellular VP-16 (8), indicating that resistance is unlikely to be mediated by altered drug transport. Rather, resistance is presumed to result from the decreased expression of topo IIα (Fig. 3). VP-16-induced double-strand cleavage of SV40 DNA and band depletion-immunoblotting studies using lysates from VP-16-treated cells demonstrated that the topo IIα protein from MDA-VP cells is drug sensitive. Thus, it is unlikely that mutation in the protein resulting in defective VP-16 interaction is the etiology of the resistance.

We previously demonstrated that topo IIα gene transfer using a pMAM vector system increases VP-16-mediated cytotoxicity (8, 9, 26). The pTOP₂-MAMneo vector, constructed and transfected into cells by calcium phosphate coprecipitation is not, however, suitable for in vivo investigations. High-efficiency gene transfer using a vector that can enter the tumor cells spontaneously is needed for animal studies to determine whether topo IIα gene transfer can alter drug resistance in vivo and, thus, be a potential strategy for clinical gene therapy. Recombinant adenoviral vectors efficiently infect cells and are one method used to accomplish gene transfer. Adenoviral vectors containing the p53 gene have been successful in altering tumor cells in patients with lung and head and neck cancer (27, 28). Our goal was to design a recombinant adenoviral vector that would increase intracellular topo IIα protein, the molecular target of VP-16 and other topo II inhibitors (29, 30, 31, 32).

In this study, we successfully constructed a recombinant adenovirus containing the normal human topo IIα gene, verified by both PCR and restriction mapping. Because human topo IIα is a large gene (5.6 kb), we selected the pBHG10 packaging vector, which has a larger capacity for gene insertion. Use of packaging vectors like pJM17 with smaller capacities resulted in truncation of the topo IIα gene after homologous recombination (data not shown). The strategy described here was successful because infection with this adenoviral construct containing the human topo IIα gene led to the expression of topo IIα in MDA-VP cells and production of drug-sensitive topo IIα protein. Sensitivity to VP-16 was enhanced 4.5-fold in cells infected with this construct.

Our data indicated that adenovirus-mediated gene transfer provides a novel way to circumvent drug resistance in cells with decreased topo IIα. The topo II activity in primary breast tumors is considerably lower than that found in cervix, colon, and lung tumors (33). The absence of detectable topo II activity in 10% of 29 tumors evaluated implies a level of intrinsic resistance in some tumors. In addition, large variations in individual cellular topo II expression within each tumor have been described (33), helping to explain the heterogeneous response to topo II-directed therapy of some patients and the emergence of resistance after previously successful drug therapy. Increasing the cellular level of normal topo II in the tumor may offer a way to increase the sensitivity of the emerging resistant cells to the cytotoxic action of topo II-reactive agents.

A 4–5-fold change in resistance may be regarded by some as low. However, when one considers that chemotherapeutic drugs are normally used at or near maximum tolerated doses, a 4–5-fold change in the effectiveness of a fixed dose may, indeed, be clinically relevant. This increase is more than that achieved with high-dose protocols followed by bone marrow or stem cell rescue. High-dose methotrexate or ifosfamide showed increased efficacy and changes in tumor sensitivity at a 3-fold increase over the standard dose.

Our approach using topo IIα gene therapy also holds promise for clinical application because there appeared to be selective topo IIα up-regulation in tumor cells expressing low levels of topo IIα. The vector did not stimulate an increase in topo IIα mRNA or protein in the two normal cell lines tested nor was sensitivity to VP-16 altered after infection of normal cells. In the host, tumor cells grow side-by-side with normal cells. The ability to selectively sensitize tumor cells decreases the potential for normal cell toxicity, thus raising the therapeutic index. It is unclear why exposure of normal cells to Ad-hTopoIIα did not lead to enhanced topo IIα expression because the Ad-β-gal vector was able to infect these cells and cause production of β-gal protein. The tight control of topo IIα activity in normal cells may block sustained increases in protein production. Indeed, we have previously shown down-regulation of the endogenous human topo IIα gene in brain tumor cells with normal topo IIα levels after transfection with the Drosophila topo IIα gene (9). In those studies, we were able to distinguish exogenous from endogenous gene product by using the Drosophila gene. The mechanism of this down-regulation appeared to be a feedback mechanism via an effect on the endogenous topo IIα promoter (26).

In this study, we used the human topo IIα gene and, thus, were unable to determine whether the exogenous gene was not expressed in the normal cells or whether the endogenous gene was being down-regulated so that the total cellular topo IIα level (endogenous and exogenous) was maintained at the same level seen before infection. In contrast, the expression level of endogenous topo IIα in the MDA-VP cells was extremely low de novo. Therefore, it may be difficult to detect small additional decreases in topo IIα mRNA with strong exogenous gene transcription and translation.

Our data suggest that, regardless of the mechanisms, increased gene expression and cytotoxicity will be sustained only in tumor cells with decreased topo IIα. Indeed, transduction of MDA parent cells with Ad-hTopoIIα resulted in no significant increase in topo IIα mRNA (Fig. 2, C and D). Low levels of topo IIα protein, indeed, appear to be clinically relevant and to correlate with disease response to topo II inhibitors (34, 35). The data presented support our hypothesis that topo IIα gene transfer may circumvent drug resistance by increasing the target for VP-16 and other topo II-reactive antineoplastic agents.

We thank Joyce Benjamin of The University of Texas M. D. Anderson Cancer Center for manuscript preparation.