Selective ablation of human cancer cells by telomerase-specific adenoviral suicide gene therapy vectors expressing bacterial nitroreductase

Abstract

Reactivation of telomerase maintains telomere function and is considered critical to immortalization in most human cancer cells. Elevation of telomerase expression in cancer cells is highly specific: transcription of both RNA (hTR) and protein (hTERT) components is strongly upregulated in cancer cells relative to normal cells. Therefore, telomerase promoters may be useful in cancer gene therapy by selectively expressing suicide genes in cancer cells and not normal cells. One example of suicide gene therapy is the bacterial nitroreductase (NTR) gene, which bioactivates the prodrug CB1954 into an active cytotoxic alkylating agent. We describe construction of adenovirus vectors harbouring the bacterial NTR gene under control of the hTR or hTERT promoters. Western blot analysis of NTR expression in normal and cancer cells infected with adenoviral vectors showed cancer cell-specific nitroreductase expression. Infection with adenoviral telomerase–NTR constructs in a panel of seven cancer cell lines resulted in up to 18-fold sensitization to the prodrug CB1954, an effect that was retained in two drug-resistant ovarian lines. Importantly, no sensitization was observed with either promoter in any of the four normal cell strains. Finally, an efficacious effect was observed in cervical and ovarian xenograft models following single intratumoural injection with low doses of vector, followed by injection with CB1954.

Introduction

A major problem with conventional anticancer cytotoxic therapies is their low therapeutic index: many conventional drugs display a lack of selectivity for malignant cells leading to dose-limiting toxicity. Recent insights into tumour cell biology have provided a wealth of possibilities for the development of novel mechanism-based therapeutics (Garrett and Workman, 1999; Karamouzis et al., 2002; Keith et al., 2002; Scapin, 2002).

An interesting target for the development of novel anticancer strategies is telomerase, a ribonucleoprotein reverse transcriptase that extends human telomeres by a terminal transferase activity (White et al., 2001; Keith et al., 2002; Mergny et al., 2002). Human telomerase activity is present during embryonic development, but is repressed in most adult somatic tissues with low levels of activity detected in some tissues with self-renewal capacity. By contrast, telomerase is highly active in the vast majority of human tumours (Kim et al., 1994; Shay and Bacchetti, 1997; Holt and Shay, 1999). Reactivation of telomerase expression in immortal cell populations compensates for cell division-associated telomere attrition and is considered to be a critical step in immortalization of human cancer cells in vitro and in vivo (Counter et al., 1992). It is increasingly clear that telomerase activity is controlled on multiple levels, with the possibility of many strategies for therapeutic intervention (White et al., 2001; Keith et al., 2002; Mergny et al., 2002), including inhibitors that lead to telomere shortening and apoptosis or senescence in vitro and decreased tumourigenic potential in vivo (Damm et al., 2001; Hahn et al., 1999; Naasani et al., 1999; Zhang et al., 1999; Pascolo et al., 2002). Transcription of the core RNA (hTR) and reverse transcriptase (hTERT) components constitutes the major level of differential regulation between normal and cancer cells. Therefore, the differential activities of the hTR and hTERT promoters may provide a sound basis for the development of transcriptionally directed cytotoxic gene therapy approaches (Plumb et al., 2001).

Several recent studies have used hTERT and hTR promoters to drive expression of therapeutic transgenes in tissue culture models of gene therapy (Abdul-Ghani et al., 2000; Gu et al.2000, 2002; Koga et al., 2000,2001; Boyd et al., 2001; Komata et al., 2001; Majumdar et al., 2001; Plumb et al., 2001). Additionally, hTERT-specific constructs have been administered to xenograft models using naked DNA injections, liposomes and adenovirus vectors, although no study has yet addressed in vivo transfer of hTR-specific therapeutic constructs. A number of therapeutic transgenes are available for use in cytotoxic gene therapy, and prodrug activating systems such as herpes simplex thymidine kinase/gancyclovir or cytosine deaminase/5-FC have been widely used in preclinical gene therapy models (Aghi et al., 2000). Both systems have been shown to be efficacious in a variety of models, but their use in cancer gene therapy may be limited by the requirement for division of target cells. Additionally, the multistep enzymatic conversion of these prodrugs may present a kinetic bottleneck limiting their utility. Using plasmid vectors, we previously demonstrated that hTR- and hTERT-restricted expression of bacterial nitroreductase (NTR) in stable cell lines selectively directs the toxic activation of the weak alkylating agent CB1954 to cancer cells. The single enzyme activation of CB1954 produces a powerful bifunctional alkylating agent capable of inducing p53-independent apoptosis in both dividing and nondividing cells (Anlezark et al., 1992; Cui et al., 1999; Plumb et al., 2001). Thus, the NTR/CB1954 system may have some advantages over other suicide gene systems.

In that study, we showed that the efficiency of telomerase-directed gene therapy relies largely on the activity of hTR and hTERT promoters in individual cell lines, with cytotoxicity restricted exclusively to highly expressing cells. Here we extend these observations with the construction of first-generation adenoviral (Ad) vectors harbouring hTR-NTR and hTERT-NTR expression constructs (Ad-hTR-NTR and Ad-hTERT-NTR).

Results

Construction and characterization of adenoviral vectors

The Adeasy system (He et al., 1998) was used to construct E1/E3-deleted adenovirus constructs harbouring the NTR coding sequence under the transcriptional control of hTR and hTERT promoters. Characteristics of the telomerase gene therapy adenoviruses Ad-hTR-NTR and Ad-hTERT-NTR are given in Table 1. The ratio of viral particles (VP) at A260 relative to infectious plaque forming units (PFU) was low for both vectors (15.6 for Ad-hTR-NTR and 6.5 for Ad-hTERT-NTR), indicating that the preparations were of high quality. The 876 bp hTR and 541 bp hTERT promoter fragments incorporated in the vectors have previously been characterized in transfection and cell survival assays and are able to drive sufficient expression of NTR to sensitize a range of human tumour cell lines to the cytotoxic effects of CB1954 in vitro and in vivo (Zhao et al., 1998,2000; Plumb et al., 2001).

Table 1 Characteristics of adenoviral vectors

In order to compare the efficiency and selectivity of NTR expression between normal and cancer cell lines, human cervical carcinoma cells (C33a) and human foetal lung fibroblasts (WI38) were infected for 1 h with 50 PFU/cell of either Ad-hTR-NTR, Ad-hTERT-NTR or the reporter virus Ad-CMV-LacZ. NTR expression was analysed by Western blotting (Figure 1a), 48 h postinfection. Strong 24 kDa NTR signals were evident in the lanes corresponding to Ad-hTR-NTR- and Ad-hTERT-NTR-infected cervical carcinoma cells, with the hTR promoter generating a significantly stronger expression than hTERT. No signal was detected in the Ad-hTERT-NTR-infected WI38, but a weak signal was detected in WI38 infected with Ad-hTR-NTR although the differential in expression from the hTR promoter between the two cell lines was extremely large. X-gal staining of Ad-CMV-LacZ infected cells 48 h postinfection indicated that both cell lines were efficiently infected by adenovirus at this multiplicity of infection (Figure 1b,c). The proportion of cells infected was estimated by counting approximately 500–1000 cells per experiment. For the representative experiment shown, the proportions of infected cells were 83%±3% (C33a) and 98%±1.0% (WI38). Therefore, Ad-hTR-NTR and Ad-hTERT-NTR constructs drive efficient expression of NTR specifically in cancer cells, but not in normal cells.

Figure 1
figure1

Expression of NTR in adenovirus-infected cell lines. (a) C33a (lanes 2–4) and WI38 cells (lanes 5–7) were mock infected (lanes 4 and 7), or infected for 1 h with 50 PFU/cell either Ad-hTR-NTR (lanes 2 and 5) or Ad-hTERT-NTR (lanes 3 and 6). Protein, 48 h post infection, was extracted from infected cells and 20 μg was used for Western blot detection of NTR protein expression using the rabbit anti-NTR antibody R36 and an HRP-conjugated anti-rabbit secondary. For comparison, lane 1 shows the expression of NTR in C33a cells stably harbouring an hTR-NTR expression plasmid (Plumb et al., 2001). (b) and (c) show C33a (b) and WI38 cells (c) infected with 50 PFU/cell Ad-CMV-LacZ and stained for expression as described in Materials and methods

One concern regarding adenovirus-mediated gene transfer is that human cellular promoters incorporated into first-generation adenovirus backbones may lose their tissue specificity because of strong virus-specific transcriptional regulatory elements (Ring et al., 1996; Shi et al., 1997; Vassaux et al., 1999). In order to confirm that the telomerase promoters retain their predicted cellular transcriptional characteristics in the context of the Ad vectors, we defined the initiation sites of NTR transcripts expressed in Ad-hTR-NTR- and Ad-hTERT-NTR-infected cells. cDNA libraries were prepared from infected C33a cells and the 5′ ends of the hTR-NTR and hTERT-NTR transcripts were amplified and sequenced (Figure 2).

Figure 2
figure2

Sequences of the 5′ ends of hTR-NTR and hTERT-NTR transcripts in virus-infected cervical carcinoma cells. 5′ ends of NTR transcripts expressed in C33a cervical carcinoma cells were amplified from l μg cDNA using the Clontech SMART-RACE kit. The figure shows genomic sequences of (a) hTR and (b) hTERT promoters, in addition to the published transcriptional start sites and the start sites defined from sequence analysis of the 5′ ends of NTR transcripts. For the hTR promoter, a single transcriptional start site 46 bp upstream of the template start was established by Feng et al (1995). In contrast, several sites have been reported for the hTERT promoter. The white arrows indicate the transcriptional start region (TSR) identified by Wick et al (1999), while the black (TSS1) and grey (TSS2) arrows represent the start sites defined by Takakura et al (1999) and Horikawa et al (1999), respectively. The start site of hTERT-NTR is indicated by the hatched arrow (TSS3). Nucleotides corresponding to transcriptional starts are underlined, while important structural and regulatory features are indicated by bold lettering. The translational start site of NTR is indicated by bold, enlarged characters

The transcriptional start site for the hTR promoter defined here confirms that transcription of NTR from the Ad-hTR-NTR vector initiates 46 bp upstream of the start of the hTR template sequence, at the site that was previously described (Feng et al., 1995). Additionally, we place the hTERT transcriptional start site (TSS3) 64 bp upstream of the hTERT start codon. Several other transcriptional start points have been defined for hTERT and it is not uncommon for weak, TATA-less sequences such as the hTERT promoter to show complex transcriptional initiation patterns (Clark et al., 1998; Rudge and Johnson, 1999; Dong et al., 2000). The transcriptional start site defined here is within the transcriptional start region (TSR) 60–112 bp upstream of the hTERT translation start codon defined by Wick et al (1999) and lies close to the sites defined by Takakura et al (1999) (TSS1) and Horikawa et al (1999) (TSS2). Therefore, our data are consistent with previous reports and, together with the protein expression data presented in Figure 1 indicate that the hTR and hTERT promoters retain their native characteristics when the constructs are delivered to human cells in an adenovirus background.

Selective sensitization of human cancer cells to CB1954 in vitro following infection with telomerase-specific gene therapy viruses

To determine whether Ad-hTR-NTR and Ad-hTERT-NTR selectively sensitize human cancer cells to CB1954 in vitro, we infected a panel of normal mortal and cancer cells with Ad-hTR-NTR, Ad-hTERT-NTR and exposed cells to CB1954. Preliminary virus titration experiments indicated that high infectious doses had a nonspecific vector-induced cytotoxic effect in some cell lines (data not shown); therefore, throughout this study we have used multiplicities of infection of 10 and 50 PFU/cell at which no nonspecific toxicity was observed. All experiments included equivalent infectious doses of the reporter vector Ad-CMV-LacZ that has been widely used as a control for transduction efficiency in studies of adenovirus-mediated gene transfer.

To directly quantify the differential activity of the viral constructs between normal and cancer cells, we performed MTT assays on cells exposed to CB1954 and calculated IC50 values from the cytotoxicity curves in mock-infected cells and in cells infected with 10 and 50 PFU/cell of Ad-hTR-NTR and Ad-hTERT-NTR (Table 2). Based on the IC50 data, values were assigned for the sensitization of virus-infected cells to CB1954 relative to mock-infected cells (Figure 3). Sensitization values are taken to be the fold difference between IC50 for mock- and virus-infected cells of an individual cell strain. Additionally, the Ad-CMV-LacZ vector was included in all experiments in order to estimate the proportion of cells transduced at each infectious dose (Table 2).

Table 2 IC50 values for CB1954 cytotoxicity in cell lines infected with Ad-hTR-NTR and Ad-hTERT-NTR gene therapy viruses
Figure 3
figure3

Mean sensitization to CB1954 and infectivity in adenovirus-infected cancer and normal cell lines. Cell lines were infected with 50 PFU per cell Ad-hTR-NTR or Ad-hTERT-NTR and subjected to CB1954 challenge and MTT assay. IC50 values for individual experiments in each cell line were taken to be the concentration of CB1954 required to reduce cell densities in MTT assays to 50% of those of untreated controls. Sensitization values were the fold difference between the IC50 of the mock-infected cells and those of cells infected with the gene therapy vectors. Black bars correspond to Ad-hTR-NTR-infected cells and grey bars to Ad-hTERT-NTR-infected cells. Data are means and s.e. calculated from three experiments

A clear and efficient promoter and dose-dependent cytotoxic effect was observed in C33a cervical carcinoma, DU145 prostatic carcinoma and A2780 ovarian adenocarcinoma cells on administration of both virus and CB1954. C33a cells showed the most pronounced cytotoxic effect with a high dose administration (50 PFU/cell) of Ad-hTR-NTR resulting in an 18-fold sensitization to CB1954, while infection with Ad-hTERT-NTR resulted in a fivefold enhancement of cytotoxicity. In all cell lines that showed sensitisation, 50 PFU/cell was more effective than 10 PFU/cell. Thus, adenovirus vectors harbouring the telomerase–NTR expression cassettes sensitized C33a cells to the effects of CB1954 in a promoter- and dose-dependent fashion. Similarly, the prostatic carcinoma cell line DU145 showed strong sensitization to CB1954 with the hTR-specific construct (fivefold), although Ad-hTERT-NTR did not have a powerful effect (twofold).

A2780 cells were very poorly transduced with adenoviral vectors, reaching only 34% infectivity at the high infectious dose of Ad-CMV-LacZ (Table 2). Nevertheless, cells infected by Ad-hTR-NTR were eightfold more sensitive to CB1954 than mock-infected cells. Infection with the hTERT-NTR virus resulted in threefold enhancement of cytotoxicity. The low efficacy observed for the hTERT construct in DU145 and A2780 cells presumably reflected both the low efficiency of infection in A2780 (Table 2) and the lower activity of the hTERT promoter (Plumb et al., 2001). Therefore, Ad-hTR-NTR was the more efficacious vector, while Ad-hTERT-NTR resulted only in a modest effect. The data indicate that it is possible to achieve a cytotoxic effect when only a minority of cells are transduced, providing that the promoter activity is of sufficient strength in infected cells.

In contrast, and despite their high permissiveness for adenovirus infection (98% at 50 PFU), WI38 cells were not sensitized to CB1954 either by Ad-hTR-NTR or by Ad-hTERT-NTR. An important consideration for telomerase-directed gene therapy must be its specificity for a broad range of cancer cells over a broad range of normal cells. The previous studies of telomerase-directed gene therapy have demonstrated lack of activity in several normal human fibroblast cell strains including WI38, in addition to normal retinal pigmented epithelial cells and telomerase negative immortal cells that utilize an alternative pathway for maintenance of telomere stability (Gu et al., 2000,2002; Koga et al., 2000,2001; Komata et al., 2001; Majumdar et al., 2001; Plumb et al., 2001). In order to extend these findings, we tested normal adult mammary (HMEC), prostate (HPrEC) and bronchial (HBrEC) epithelial cell strains.

Despite the reasonable permissiveness of both HMEC and HBrEC for adenovirus infection (Table 2), none of the normal cell strains tested showed more than 1.4-fold reduction in IC50 with either promoter, indicating that the sensitivity of cell lines to CB1954 remained unchanged after administration of the vectors. The normal human prostatic epithelial cell strain HPrEC by contrast were relatively refractory to the virus, reaching only 37% infectivity. However, it is clear from the A2780 data (Table 2) that infection of the total cell population is not a prerequisite for significant enhancement of CB1954-induced cytotoxicity providing that significant transgene expression can be achieved in infected cells. Taken together, these data give a strong indication that the hTR and hTERT promoters did not drive sufficient expression of NTR in normal human cell lines to result in enhancement of cytotoxicity and suggests that the system has a selective effect for cells with high promoter activity. Further evidence of the requirement for a high promoter activity in addition to efficient infectivity can be seen from the 5637 bladder carcinoma cell line that was not effectively targeted by either promoter, having sensitization values under 2 for all treatments. We previously correlated the low hTR and hTERT promoter activity of 5637 cells with a low sensitivity to CB1954 in stable cell line models harbouring hTR-NTR and hTERT-NTR constructs (Plumb et al., 2001). Therefore, the data presented here are consistent with high promoter activity as a prerequisite for efficient transcriptional targeting using the telomerase promoters, and we speculate that the low sensitization values observed in UVW cells most likely reflect similar low promoter activities.

Telomerase–NTR vectors sensitize drug-resistant ovarian carcinoma cells to CB1954

A major limitation to current therapies is the development of drug-resistant phenotypes. The active metabolite of CB1954 is a bifunctional alkylating agent, thus it is important to evaluate the potential of the telomerase gene therapy vectors to overcome alkylating agent resistance. The A2780 ovarian adenocarcinoma derivatives, A2780–CP70 (selected by cisplatin exposure) and A2780–ADR (selected by adriamycin exposure), show crosstolerance to a range of DNA-damaging agents, including the purine analogue 6-thioguanine and the alkylating agent N-methyl-N-nitrosourea. Resistance in these cell lines is because of loss of the mismatch repair protein hMLHl. (Anthoney et al., 1996; Drummond et al., 1996; Brown et al., 1997; Strathdee et al., 1999). To assess whether NTR expression from telomerase-specific vectors is sufficiently high to allow local activation of CB1954 to overcome the drug-resistant phenotype in hMLHl negative cells, we analysed the effects of the vectors in these cell lines (Figure 3).

A2780–CP70 cells showed a slightly higher basal tolerance of CB1954 than the parental A2780 cells, with an IC50 of 60 μ M for CP70 compared with 29 μ M for A2780 (Table 2). A2780–ADR had a higher basal tolerance with an IC50 of 238 μ M. Interestingly, however, neither derivative showed resistance when transduced with the gene therapy vectors, exhibiting similar cytotoxicity profiles to the parental cells (Figure 3, Table 2). Thus, infection with 50 PFU per cell of Ad-hTR-NTR sensitized CP70 cells to CB1954 some 10-fold and ADR cells 14-fold, while the same infectious dose of Ad-hTERT NTR resulted in modest sensitizations of two-and threefold. Additionally, the cisplatin-resistant cell line had a similar permissiveness for Ad-CMV-LacZ infection as the parental cell line, although ADR cells were more permissive, indicating that hTR promoter-restricted expression of NTR is sufficiently strong to result in significant cell death in a drug-resistant cell line despite transduction of a minority of cells.

Ad-hTR-NTR and Ad-hTERT-NTR vectors sensitize human cancer cells to CB1954 in vivo

In order to assess the sensitization of model human tumours to Ad-hTR-NTR and Ad-hTERT-NTR vector administration followed by CB1954 in vivo, C33a or A2780 xenografts were established in the flanks of athymic female nude mice. Randomly distributed groups of six xenograft-bearing animals recieved a single intratumoural injection of either Ad-hTR-NTR, Ad-hTERT-NTR or Ad-CMV-LacZ. A single injection of 80 mg/kg CB1954 was administered by tail vein injection, 24 h later. Daily calliper measurements were performed for a further 7 days and tumour volumes were estimated from the measurements (volume=d3×π/6). LacZ-expressing tumours were harvested for staining on the day of drug injection.

The C33a tumours in flanks of control animals increased in volume at a similar rate, approximately doubling in size over 7 days. By contrast, a single injection of telomerase–NTR gene therapy vectors coupled to a single CB1954 administration resulted in a 40% reduction in relative tumour volume for the hTERT-NTR virus and a 43% reduction for the hTR-NTR virus. The difference in tumour volumes between animals injected with virus only and those injected with virus and drug was highly significant at day 7 (P<0.001). We note with interest that the infectious dose and the number of drug and vector administrations used in this experiment are efficacious, though substantially lower than those utilized in other adenovirus-based reports of telomerase gene therapy (Gu et al., 2000; Majumdar et al., 2001). Figure 4b shows the results of an additional experiment in A2780 xenograft-bearing mice. In this experiment, control tumours increased in size approximately fourfold, while a single intratumour injection with 1×l09 PFU, coupled to a single intravenous injection of 80 mg/kg CB1954 resulted in reductions of tumour volume of 54% for the hTR-NTR construct and 57% for the hTERT-NTR construct. Again, the difference in tumour volume between virus-only and virus/CB1954 groups was significant at day 7 (P<0.001). It should also be noted that one animal showed complete tumour regression.

Figure 4
figure4

Reduction of tumour volume of C33-A and A2780 xenografts after intra tumoural injection of telomerase gene therapy viruses followed by systemic administration of CB1954 (a) and (b) 107 C33a or A2780 cells per mouse were injected subcutaneously into the flanks of six groups of six female athymic nude mice and allowed to develop for 14 days until tumour diameters were approximately 5 mm. At this time (day 0), four groups of mice were injected intra tumourally with a total of 4×l08 PFU for C33a xenografts (panel a) or 1×109 PFU for A2780 (panel b) of Ad-hTR-NTR or Ad-hTERT-NTR per tumour (two groups for each virus). The following day (day 1), three groups (one-drug-only group and 1 each of the virus-injected groups) were intravenously injected with 80 mg/kg CB1954 and the mean tumour volumes of all groups were monitored daily for 7 days. Results given are the mean tumour volumes and standard errors at each time point derived from six mice per group by the formula volume=d3×π/6. A clear reduction in tumour volume over the course of the experiment is evident in the groups injected with both virus and drug, but not in any of the control groups. The differences in tumour volume between virus-only and virus/CB1954 groups were analysed using Students' t-test and found to be highly significant for both experiments (P<0.001). Filled box: hTR virus with CB1954, open box: hTERT virus with CB1954. Closed triangle: hTR virus alone, open triangle: hTERT virus alone. Closed circle: diluent alone, filled circle: CB1954 alone. (c) and (d) Photomicrographs of vector-infected frozen tissue sections at low magnification (original magnification ×2.5). C33a xenografts (panel c) were injected with a single dose of 4×108 PFU Ad-CMV-LacZ in parallel with tumour reduction experiments. Tumours were excised and cryosections were stained for LacZ expression and counterstained with eosin, 24 h later. A2780 xenografts (panel d) were infected with Ad-hTR-NTR or Ad-hTERT-NTR at 1×109 PFU. After 24 h, tumours were excised for immunohistochemical analysis of NTR expression and counterstained with haematoxylin. The figure shows a section infected with Ad-hTERT-NTR

Distribution of adenovirus vectors in xenograft models

An important factor for consideration in the evaluation of the efficacy of any virus-mediated gene transfer model is the efficiency of virus penetration of the tumour mass. In most cases using replication defective vectors, this is likely to be the major limiting factor since no reliable systems exist that can transduce all cells of a tumour. To address this issue, parallel experiments were performed to determine the distribution of cells infected and, thus, to assess indirectly the potential of the bystander effect in our models. Control C33a and A2780 xenografts were injected with infectious doses of Ad-CMV-LacZ (C33a) or Ad-hTR-NTR and Ad-hTERT-NTR (A2780) equivalent to those administered in the tumour reduction experiments. Tumours were excised 24 h later, corresponding to the time of CB1954 administration in the tumour volume reduction experiments and patterns of LacZ or NTR expression were examined in frozen sections derived from four to five separate tumours for each cell type. Representative in vivo expression data are shown in Figure 4c,d.

In keeping with the low dose, single administrations of vector employed in these experiments, in most of the xenograft sections analysed for Ad-CMV-LacZ infection or NTR expression, only a minority of cells showed transgene expression, suggesting limited xenograft penetration. Excised tumours showed similar heterogeneous patterns of staining whether LacZ or NTR was detected, with many of the transduced cells restricted mainly to the area surrounding the needle track or to the peripheral edge of the section. In some cases patchy internal staining was observed, indicating that in some instances the vectors had penetrated the tumour mass, with low efficiency, although the proportion of LacZ- or NTR-expressing cells was still considerably lower than the proportion of untransduced cells and staining was less intense than in peripheral areas. This pattern most likely reflects the initial exposure of cells to virus supernatant, as it is injected directly into the cell mass and bathing of the cell mass in virus on needle withdrawal. It is encouraging that significant therapeutic effect can be achieved in these models despite low-efficiency tissue penetration. These data point to the activity of a bystander effect mediated by diffusion of the active lipophilic 4-hydroxylamino derivative of CB1954 into the cell mass following initial activation in a low proportion of virus-tranduced cells with high hTR and hTERT promoter activity.

Discussion

A major limitation of current cancer therapeutics is a lack of real selectivity for tumour cells in vivo and, hence, a major aim for the future is the development of tumour-specific targeting systems. The hTR and hTERT promoters are interesting candidates for the development of gene therapy systems since they should display activity over a range of malignancies. Therefore, in order to model delivery of telomerase-specific suicide gene therapy vectors, we have constructed adenovirus vectors expressing hTR-NTR and hTERT-NTR constructs.

In the present study, the hTR promoter was stronger than hTERT in all the cancer cell lines assayed, yet it retained a selective effect. It is apparent from this and other studies that although highly efficient cell killing can be directed by both promoters the efficiency of hTERT-promoter directed constructs may be restricted in cancer cells with low promoter activity (Abdul-Ghani et al., 2000; Gu et al., 2000,2002; Koga et al., 2000,2001; Komata et al., 2001; Plumb et al., 2001). However, the ability of telomerase–NTR vectors to efficiently sensitize some very poorly transduced cancer cell lines to CB1954 is encouraging. Thus, it is unnecessary to infect all telomerase positive cells in a population, providing that the promoter activity is of sufficient strength to drive high-level expression leading to a bystander effect in those cells that harbour the construct. This view is supported by the poor distribution of infection, yet efficient growth retardation in xenografts injected with low doses of telomerase–NTR vectors.

Although the ability to effectively target cytotoxicity to cancer cells is paramount in the development of gene therapy strategies, it is highly likely that the development of drug-resistant phenotypes such as loss of mismatch repair function will still represent an important obstacle to effective therapy. In this regard, we note with interest that telomerase-directed NTR expression in hMLHl-deficient drug-resistant derivatives of A2780 results in efficient killing of cells that are crossresistant to both alkylating agents and other common agents that elicit more diverse effects. This raises the intriguing notion that the use of telomerase-specific molecular therapy to generate high intracellular concentrations of active alkylating agents can induce sensitivity to cell death in cancer cells hitherto thought to have a drug-resistant phenotype.

We conclude from the present data that the sensitizing effect of telomerase-specific adenoviral gene therapy vectors is dependent largely on promoter activity; thus, cell lines with low promoter activity are not sensitized to CB1954 even in circumstances where essentially all of the cells are infected. While a minority of tumour cells, such as the 5637 bladder carcinoma line and the UVW malignant glioma line, may display low hTR and hTERT promoter activity, from previous expression studies it is clear that a number of tumour types are good candidates for telomerase-directed therapies (Soder et al., 1997,1998; Sarvesvaran et al., 1999; Wisman et al., 2000; Downey et al., 2001; Hiyama et al., 2001). Importantly, this situation is also expected to apply to the majority of normal adult somatic cells in vivo, suggesting that Ad-hTR-NTR and Ad-hTERT-NTR/CB1954 will be selectively toxic to cancer cells. Indeed, we have not observed any significant enhancement of toxicity following vector transduction in four mortal human cell strains suggesting that telomerase-directed molecular therapeutics may prove to be a highly effective, yet safe class of anticancer agent.

A growing body of literature supports the development of telomerase-directed therapeutics as selective, yet wide-ranging antitumour agents (Keith et al., 2002). A number of very different systems have been proposed to target telomerase expression or activity, but gene therapy remains an attractive approach. Taken together, the present data confirm that the differential promoter activities of the hTR and hTERT promoters between normal and cancer cells that we and others have observed are retained in viral models of gene transfer and help to validate the general principle underlying a telomerase-directed approach in gene therapy. Moreover, the NTR/CB1954 system is an interesting candidate for development under the direction of telomerase promoter constructs, since the safety of pharmacologically relevant doses of CB1954 in humans has already been demonstrated in clinical trials (Chung-Faye et al., 2001). In summary, it is clear that the addition of Ad-hTR-NTR and Ad-hTERT-NTR to a potential telomerase-specific anticancer armoury is an exciting prospect, although optimal systems for telomerase-directed gene therapy will presumably require targeted delivery systems in addition to a clear understanding of the regulation of hTR and hTERT genes in target tumours (Keith et al., 2002).

Materials and methods

Cell lines and viruses

The human cancer cell lines 5637 (bladder carcinoma), DU145 (prostatic carcinoma), C33a (cervical carcinoma) and UVW (malignant glioma) were obtained from ATCC and maintained in appropriate media. A2780 (ovarian adenocarcinoma) cells were originally obtained from Dr RF Ozols (Fox Chase Cancer Centre, PA, USA). Selection of the drug-resistant A2780 variants A2780–CP70 and A2780–ADR has been described elsewhere (Anthoney et al., 1996). WI38 normal human foetal lung fibroblasts were obtained from Coriell Cell Repository (USA) and the normal adult human mammary epithelial cells (HMEC), prostatic epithelial cells (HPrEC) and bronchial epithelial cells (HBrEC) were obtained from Clonetics (USA) and were subcultured and maintained according to the instructed culture systems in each case.

The viruses Ad-hTR-NTR and Ad-hTERT-NTR were generated using the Adeasy system supplied by Qbiogene (Middlesex, UK), according to the manufacturer's instructions. Ad-hTR-NTR contains a previously characterized 876 bp fragment of the hTR proximal promoter (Zhao et al., 1998; Plumb et al., 2001) and Ad-hTERT-NTR contains a 541 bp fragment of the hTERT promoter (Plumb et al., 2001). Both vectors contain the Escherichia coli NTR coding sequence. All ligations into transfer vectors were performed using a rapid ligation kit obtained from Roche Diagnostics Ltd (East Sussex, UK) and electro-cotransformation of adenovirus backbone and transfer vectors were performed in a Hybaid (Middlesex, UK) cell shock electroporator. Expression cassettes were inserted in a left-to-right orientation in the adenovirus backbone and sequence and orientation of the promoter–NTR constructs in the transfer vectors and adenovirus genomic plasmids was confirmed by restriction digest and sequencing. Large-scale amplification was performed by Qbiogene. Since virus-mediated gene transfer studies commonly include infectivity controls, we obtained the reporter vector Ad-CMV-LacZ from Qbiogene for use as a positive control allowing an estimate of infection efficiency during cytotoxicity and tumour reduction experiments. Small-scale amplifications on 293 cells were purified using the Adenoprep kit (Virapur, USA). Viral titre was checked for all constructs by plaque assay on 293 cells and by measurement of optical density (A260) for Ad-hTR-NTR and Ad-hTERT-NTR.

Western blotting

Purification of the cytoplasmic protein fraction for Western blotting was achieved using the NE-PER differential nuclear and cytoplasmic protein extraction kit (Pierce and Warriner UK Ltd, Chester, UK). Quantification of cytoplasmic protein extracts was accomplished by BCA/Cu (II)SO4 assay and colorimetric changes were quantified using a Dynex Technologies (West Sussex, UK) MRX II microplate reader. Protein equivalents (20 μg) were electrophoresed in a 12% SDS–polyacrylamide gel, then blotted onto nitrocellulose filter and blocked overnight at 4°C in TBS-T (0.7% Tween 20) containing 5% nonfat dried milk. The following day, filters were probed for 2 h with a 1 : 50 dilution of rabbit anti-NTR primary antibody R36, then with an HRP-conjugated anti-rabbit secondary. Bound HRP was detected using ECL Western blotting HRP detection reagents (Amersham Pharmacia, Buckinghamshire, UK). R36 is a kind gift from Dr Steve Hobbs (CRC Centre for Cancer Therapeutics, Institute of Cancer Research, Surrey).

PCR and sequencing

Sequence and orientation of NTR expression cassettes in pAdeasy-1 were determined by dideoxy chain termination sequencing reactions performed using PE Biosystems (Cheshire, UK) Big Dye Terminator system and reagents according to the manufacturer's instructions, using the primers Shuntlf (5′-ggcgtaaccgagtaagatttgg-3′), Shuntlr (5′-tgctggatgggctgtattgc-3′) and AdNTseq5a (5′-cattccactaaggcatttgatg-3′) for 3′ end sequencing. Sequence analyses were performed on ABI-PRISM 377. Generation of cDNA and amplification of transcript 5′ ends from infected C33a cells was accomplished using the Stratagene Europe (Amsterdam, Netherlands) Smart RACE kit. RNA (1 μg) was used in cDNA synthesis reactions and PCR steps used the primers included in the kit, together with the gene-specific primer Shuntlr for 5′ amplification. Reaction products were subcloned for sequencing in an Invitrogen (Renfrewshire, UK) TOPO-TA™ cloning vector and were sequenced using the primer set provided in the kit.

LacZ reporter assay

For monolayer experiments, cells were seeded in six-well plates 24 h prior to infection. On the day of infection, cells were mock infected or were infected with Ad-CMV-LacZ for 1 h at 37°C at either 10 or 50 PFU/cell. Following infection, cells were incubated in fresh growth medium for a further 48 h, corresponding to the day of drug addition. The cell layers were rinsed and fixed in 0.2% glutaraldehyde, 5 mM EGTA, 2 mM MgCl2 in ice-cold PBS for 20 min. Next, the cells were rinsed and incubated with 20 μg/ml X-gal, 2.5 mMK3Fe(CN)6, 2.5 mM K4Fe(CN)6 for 24 h in the dark. Proportions of blue cells were assessed by counting five random fields (500–1000 cells). All experiments were repeated at least three times with the exception of HBrEC cells, which were analysed only once because of limited replicative potential, and data presented are the means of three repeats. To determine the tissue distribution of Ad-CMV-LacZ in cervical carcinoma xenografts, frozen sections of xenografts infected with a single administration of 4×l08 PFU Ad-CMV-LacZ in parallel with the tumour reduction experiments were fixed and stained as described above. Photomicrographs were obtained using Axiovert digital video recording equipment (Axiomatic Technology Ltd, Nottingham, UK).

MTT assay

Cells were initially treated in six-well plates with 10 or 50 PFU/cell Ad-hTR-NTR or Ad-hTERT-NTR for 1 h, then trypsinized and plated in 96-well plates for drug treatment and MTT assay (Plumb et al., 2001). On the day of the drug challenge, fourfold serial dilutions of CB1954 were prepared in cell growth medium. For an individual experiment, each data point on kill curves was plotted as the mean percentage of the untreated control, calculated across triplicate plates for each independent drug concentration. Mean IC50 values were calculated from triplicate plates using Softmax 2.42 analysis package (Molecular Devices Corporation, CA, USA). Sensitization values are the fold difference between the IC50 values for mock-infected cells and those which were infected by Ad-hTR-NTR or Ad-hTERT-NTR. All experiments were repeated at least three times with the exception of HBrEC cells, which were analysed once in triplicate. Final sensitization values and IC50 values presented are the means and s.e. derived from all three independent experiments.

In vivo tumour reduction experiments

C33a and A2780 cells were harvested and resuspended in PBS. Cells (107) were injected subcutaneously into the flanks of athymic female mice. When mean tumour diameters were at least 5 mm, the animals were randomized into six groups of six animals. At this time (day 0) 4×108 PFU (C33a) or 1×l09 PFU (A2780) Ad-hTR-NTR and Ad-hTERT-NTR were administered to animals by direct intratumoural injection. The following day, 80 mg/kg CB1954 was administered to all groups except controls. For general estimation of toxicity, animals were weighed daily and tumour volumes were estimated from calliper measurements (volume=d3×π/6). For statistical analysis, unpaired, two-tailed Student's t-tests were used. P values of <0.05 were considered significant. All animal experimentation was performed according to United Kingdom Home Office regulations and UKCCR guidelines were adhered to at all times. At 80 mg/kg CB1954, animals lost at most 10% of body weight, but had recovered to weight of controls by day 6. The combination of CB1954 and virus prolonged the effect of CB1954 alone on body weight loss. However, it should be noted that weight measurement includes the tumour burden. Thus, where a treatment is effective, body weight would be expected to be less than untreated controls.

Preparation of NTR-specific rabbit polyclonal antibody and immunohistochemistry

The NTR-B rabbit polyclonal antibody to NTR, which was used for all immunohistochemistry, was produced using a peptide consisting of the terminal 16 amino acids of the full-length NTR protein with an additional cysteine amino acid at the NH2 terminus. The peptide NH2–CTPLKSRLPQNITLTEV–CO2H was then conjugated to a carrier protein using the Pierce (Chester, UK) Imject® Maleimide-activated mcKLH kit. The peptide was mixed either with complete Freund's adjuvant (for the initial injection) or incomplete Freund's adjuvant (for subsequent booster injections) at a concentration of 100 μg/ml of peptide-KLH. All animal experimentation was performed to United Kingdom Home Office regulations and UKCCR guidelines were adhered to at all times. For immunohistochemical analysis of NTR expression, A2780 ovarian adenocarcinoma xenografts were established and, when tumours reached a minimum diameter of 5 mm, were injected once with 1×109 PFU. Ad-hTR-NTR or Ad-hTERT-NTR. After 24 h, tumours were excised and frozen sections of uninfected or of Ad-hTR-NTR- and Ad-hTERT-NTR- infected xenografts were prepared and stained using a Vectastain kit (Vector Laboratories Ltd, Peterborough, UK) according to the manufacturer's instructions. A 1 : 1000 dilution of NTR-B was used for all sections analysed. Photomicrographs were obtained using Axiovert digital video recording equipment (Axiomatic Technology Ltd, Nottingham, UK).

References

  1. Abdul-Ghani R, Ohana P, Matouk I, Ayesh S, Ayesh B, Laster M, Bibi O, Giladi H, Molnar-Kimber K, Sughayer MA, de Groot N and Hochberg A . (2000). Mol. Ther., 2, 539–544.

  2. Aghi M, Hochberg F and Breakefield XO . (2000). J. Gene. Med., 2, 148–164.

  3. Anlezark GM, Melton RG, Sherwood RF, Coles B, Friedlos F and Knox RJ . (1992). Biochem. Pharmacol., 44, 2289–2295.

  4. Anthoney DA, McIlwrath AJ, Gallagher WM, Edlin AR and Brown R . (1996). Cancer Res., 56, 1374–1381.

  5. Boyd M, Mairs RJ, Cunningham SH, Mairs SC, McCluskey A, Livingstone A, Stevenson K, Brown MM, Wilson L, Carlin S and Wheldon TE . (2001). J. Gene Med., 3, 165–172.

  6. Brown R, Hirst GL, Gallagher WM, McIlwrath AJ, Margison GP, van der Zee AG and Anthoney DA . (1997). Oncogene, 15, 45–52.

  7. Chung-Faye G, Palmer D, Anderson D, Clark J, Downes M, Baddeley J, Hussain S, Murray PI, Searle P, Seymour L, Harris PA, Ferry D and Kerr DJ . (2001). Clin. Cancer Res., 7, 2662–2668.

  8. Clark MP, Chow CW, Rinaldo JE and Chalkley R . (1998). Nucleic Acids Res., 26, 1801–1806.

  9. Counter CM, Avilion AA, LeFeuvre CE, Stewart NG, Greider CW, Harley CB and Bacchetti S . (1992). EMBO J., 11, 1921–1929.

  10. Cui W, Gusterson B and Clark AJ . (1999). Gene Ther., 6, 764–770.

  11. Damm K, Hemmann U, Garin-Chesa P, Hauel N, Kauffmann I, Priepke H, Niestroj C, Daiber C, Enenkel B, Guilliard B, Lauritsch I, Muller E, Pascolo E, Sauter G, Pantic M, Martens UM, Wenz C, Lingner J, Kraut N, Rettig WJ and Schnapp A . (2001). EMBO J., 20, 6958–6968.

  12. Dong S, Lester L and Johnson LF . (2000). J. Cell Biochem., 77, 50–64.

  13. Downey MG, Going JJ, Stuart RC and Keith WN . (2001). J. Oral Pathol. Med., 30, 577–581.

  14. Drummond JT, Anthoney A, Brown R and Modrich P . (1996). J. Biol. Chem., 271, 19645–19648.

  15. Feng J, Funk WD, Wang SS, Weinrich SL, Avilion AA, Chiu CP, Adams RR, Chang E, Allsopp RC, Yu J, Le S, West MD, Harley CB, Andrews WH, Greider CW and Villeponteau B . (1995). Science, 269, 1236–1241.

  16. Garrett MD and Workman P . (1999). Eur. J. Cancer, 35, 2010–2030.

  17. Gu J, Andreeff M, Roth JA and Fang B . (2002). Gene Ther., 9, 30–37.

  18. Gu J, Kagawa S, Takakura M, Kyo S, Inoue M, Roth JA and Fang B . (2000). Cancer Res., 60, 5359–5364.

  19. Hahn WC, Stewart SA, Brooks MW, York SG, Eaton E, Kurachi A, Beijersbergen RL, Knoll JH, Meyerson M and Weinberg RA . (1999). Nat. Med., 5, 1164–1170.

  20. He TC, Zhou S, da Costa LT, Yu J, Kinzler KW and Vogelstein B . (1998). Proc. Natl. Acad. Sci. USA, 95, 2509–2514.

  21. Hiyama E, Hiyama K, Yokoyama T and Shay JW . (2001). Neoplasia, 3, 17–26.

  22. Holt SE and Shay JW . (1999). J. Cell Physiol., 180, 10–18.

  23. Horikawa I, Cable PL, Afshari C and Barren JC . (1999). Cancer Res., 59, 826–830.

  24. Karamouzis MV, Gorgoulis VG and Papavassiliou AG . (2002). Clin. Cancer Res., 8, 949–961.

  25. Keith WN, Bilsland A, Evans TRJ and Glasspool RM . (2002). Exp. Rev. Mol Med, http://www.expertreviews.org/02004507h.htm.

  26. Kim NW, Piatyszek MA, Prowse KR, Harley CB, West MD, Ho PL, Coviello GM, Wright WE, Weinrich SL and Shay JW . (1994). Science, 266, 2011–2015.

  27. Koga S, Hirohata S, Kondo Y, Komata T, Takakura M, Inoue M, Kyo S and Kondo S . (2000). Hum. Gene Ther., 11, 1397–1406.

  28. Koga S, Hirohata S, Kondo Y, Komata T, Takakura M, Inoue M, Kyo S and Kondo S . (2001). Anticancer Res., 21, 1937–1943.

  29. Komata T, Kondo Y, Kanzawa T, Hirohata S, Koga S, Sumiyoshi H, Srinivasula SM, Barna BP, Germano IM, Takakura M, Inoue M, Alnemri ES, Shay JW, Kyo S and Kondo S . (2001). Cancer Res., 61, 5796–5802.

  30. Majumdar AS, Hughes DE, Lichtsteiner SP, Wang Z, Lebkowski JS and Vasserot AP . (2001). Gene Ther., 8, 568–578.

  31. Mergny JL, Riou JF, Mailliet P, Teulade-Fichou MP and Gilson E . (2002). Nucleic Acids Res., 30, 839–865.

  32. Naasani I, Seimiya H, Yamori T and Tsuruo T . (1999). Cancer Res., 59, 4004–4011.

  33. Pascolo E, Wenz C, Lingner J, Hauel N, Priepke H, Kauffmann I, Garin-Chesa P, Rettig WJ, Damm K and Schnapp A . (2002). J. Biol. Chem., 277, 15566–15572.

  34. Plumb JA, Bilsland A, Kakani R, Zhao J, Glasspool RM, Knox RJ, Evans TR and Keith WN . (2001). Oncogene, 20, 7797–7803.

  35. Ring CJ, Harris JD, Hurst HC and Lemoine NR . (1996). Gene Ther., 3, 1094–1103.

  36. Rudge TL and Johnson LF . (1999). J. Cell Biochem., 73, 90–96.

  37. Sarvesvaran J, Going JJ, Milroy R, Kaye SB and Keith WN . (1999). Carcinogenesis, 20, 1649–1651.

  38. Scapin G . (2002). Drug Discov. Today, 7, 601–611.

  39. Shay JW and Bacchetti S . (1997). Eur. J. Cancer, 33, 787–791.

  40. Shi Q, Wang Y and Worton R . (1997). Hum. Gene Ther., 8, 403–410.

  41. Soder AI, Going JJ, Kaye SB and Keith WN . (1998). Oncogene, 16, 979–983.

  42. Soder AI, Hoare SF, Muir S, Going JJ, Parkinson EK and Keith WN . (1997). Oncogene, 14, 1013–1021.

  43. Strathdee G, MacKean MJ, Illand M and Brown R . (1999). Oncogene, 18, 2335–2341.

  44. Takakura M, Kyo S, Kanaya T, Hirano H, Takeda J, Yutsudo M and Inoue M . (1999). Cancer Res., 59, 551–557.

  45. Vassaux G, Hurst HC and Lemoine NR . (1999). Gene Ther., 6, 1192–1197.

  46. White LK, Wright WE and Shay JW . (2001). Trends Biotechnol., 19, 114–120.

  47. Wick M, Zubov D and Hagen G . (1999). Gene, 232, 97–106.

  48. Wisman GB, De Jong S, Meersma GJ, Helder MN, Hollema H, de Vries EG, Keith WN and van der Zee AG . (2000). Hum. Pathol., 31, 1304–1312.

  49. Zhang X, Mar V, Zhou W, Harrington L and Robinson MO . (1999). Genes Dev., 13, 2388–2399.

  50. Zhao JQ, Glasspool RM, Hoare SF, Bilsland A, Szatmari II and Keith WN . (2000). Neoplasia, 2, 531–539.

  51. Zhao JQ, Hoare SF, McFarlane R, Muir S, Parkinson EK, Black DM and Keith WN . (1998). Oncogene, 16, 1345–1350.

Download references

Acknowledgements

This work was supported by the Cancer Research UK, the Fifth Framework Program of the European Commission and Glasgow University. We would like to thank Dr Steve Hobbs for the initial gift of NTR antisera.

Author information

Correspondence to W Nicol Keith.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • telomerase
  • hTERT
  • hTR
  • gene therapy
  • nitroreductase
  • GDEPT
  • adenovirus

Further reading