Hypoxia is an important factor in tumor growth. It is associated with resistance to conventional anticancer treatments. Gene therapy targeting hypoxic tumor cells therefore has the potential to enhance the efficacy of treatment of solid tumors. Transfection of a panel of tumor cell lines with plasmid constructs containing hypoxia-responsive promoter elements from the genes, vascular endothelial growth factor (VEGF) and erythropoietin, linked to the minimal cytomegalovirus (mCMV) or minimal interleukin-2 (mIL-2) promoters showed optimum hypoxia-inducible luciferase reporter gene expression with five repeats of VEGF hypoxic-response element linked to the mCMV promoter. Adenoviral vectors using this hypoxia-inducible promoter to drive therapeutic transgenes produced hypoxia-specific cell kill of HT1080 and HCT116 cells in the presence of prodrug with both herpes simplex virus thymidine kinase/ganciclovir and nitroreductase (NTR)/CB1954 prodrug-activating systems. Significant cytotoxic effects were also observed in patient-derived human ovarian cancer cells. The NTR/CB1954 system provided more readily controllable transgene expression and so was used for in vivo experiments of human HCT116 xenografts in nude mice. Subjects treated intratumorally with Ad-VEGFmCMV-NTR and intraperitoneal injection of CB1954 demonstrated a statistically significant reduction in tumor growth. Immunohistochemistry of treated xenografts showed a good correlation between transgene expression and hypoxic areas. Further investigation of these hypoxia-inducible adenoviral vectors, alone or in combination with existing modalities of cancer therapy, may aid in the future development of successful Gene-Directed Enzyme Prodrug Therapy systems, which are much needed for targeting solid tumors.
Most solid tumors develop regions of transiently or chronically hypoxic cells during growth. A variety of different mechanisms contribute to the development of hypoxia in solid tumors and hypoxia is associated with unfavorable prognosis, regardless of the treatment modality applied. These hypoxic regions are known to be more resistant to radiotherapy and chemotherapy than well-oxygenated regions of a tumor.1 Hypoxia is also related to malignant progression, increased invasion, angiogenesis and an increased risk of metastasis formation. This highlights the growing need to develop antitumor therapies, which target hypoxic cells and which might, therefore, be used in combination with existing modalities of cancer therapy to kill both normoxic and hypoxic regions of tumors.
Overexpression of the α-subunit of hypoxia-inducible factor 1 (HIF-1α) is thought to be an essential part of tumor progression and a marker of highly aggressive disease in various tumor types.2 HIF-1 is activated in response to hypoxia and activates transcription by binding to the hypoxic-response elements (HREs) within the promoter regions of genes, which regulate biological processes. Hypoxia-inducible genes such as vascular endothelial growth factor (VEGF) and erythropoietin (EPO) are frequently overexpressed in tumors.3 The HREs of VEGF and EPO have been shown to be very sensitive to hypoxic conditions and have been used in a variety of hypoxia-targeted gene therapy studies.4, 5, 6, 7, 8, 9 These promoter elements therefore offer the prospect of selective therapeutic gene expression in hypoxic tumors and enhanced tumor kill when combined with established treatment modalities to which areas of hypoxic tumor are resistant. Previous studies have placed a therapeutic transgene under the control of conditional promoters such as the carcinoembryonic antigen promoter targeting colorectal cancer cells,10 the human telomerase reverse transcriptase promoter as a cancer-specific target11, 12 or HRE to target hypoxia.4, 5, 6, 7, 8, 9
The use of viral vectors is an attractive and well-established approach for the delivery of gene therapies. The concept of virus-directed enzyme prodrug therapy (VDEPT) involves the delivery of genes encoding a prodrug-activating (‘suicide’) enzyme, which metabolizes a non-toxic prodrug to produce a toxic metabolite. Therapeutic combinations such as the herpes simplex virus thymidine kinase (HSVtk) gene and its prodrug ganciclovir (GCV), or bacterial nitroreductase (NTR) and its prodrug CB1954, produce toxic metabolites that are freely diffusible and able to kill neighboring cells via a ‘bystander effect’.13, 14 Both systems are well characterized and have been investigated in clinical trials,15, 16, 17 but have not been compared in parallel under hypoxic conditions under the control of different transcriptional promoters. The combination of transcriptional regulation, hypoxia-selective HREs and adenoviral delivery of prodrug-activating genes therefore provides great potential for selective targeting of hypoxic regions within solid tumors.
We have constructed plasmid and adenoviral vectors encoding the HSVtk and NTR suicide genes, under the transcriptional control of either VEGF or EPO HREs combined with either the minimal cytomegalovirus (mCMV) or minimal interleukin-2 (mIL-2) promoter, and compared the cytotoxic effects of hypoxia-inducible HSVtk or NTR gene products in established cancer cell lines and in primary cultures of human tumor cells in vitro. As preparatory studies for potential clinical trials, we also examined the efficacy of the NTR adenoviral vectors in human tumor xenograft models in nude mice in vivo.
Materials and methods
Culture of human cell lines
All cultures were incubated in a humidified incubator with 5% CO2 at 37°C. The human embryonic kidney cells, HEK293, were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and 2 mM L-glutamine. For hypoxic conditions, cells were incubated in either a humidified incubator with 1% O2 at 37°C or 150 μM cobalt chloride (CoCl2) was added to the medium.
The human urothelial carcinoma cell line, UMUC3, and the ovarian carcinoma cell lines, SKOV3 and OVCA433, were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (FCS) and 2 mM L-glutamine. The human colon cancer HCT116 and JON bladder carcinoma cell lines were maintained in RPMI 1640 with 10% FCS and 2 mM L-glutamine, whereas HT1080 human fibrosarcoma cells were maintained in α-modified Eagle's medium with 10% FCS, 2 mM L-glutamine and 1% non-essential amino acids. All cell lines were obtained from the American Tissue Culture Collection.
Culture of patient-derived tumor specimens
Primary ovarian cancer cells were isolated from the ascitic fluids of patients undergoing therapeutic drainage procedures. Ethics approval and written informed consent was obtained before collection of ascitic fluid. Briefly, 100 ml ascitic fluid was centrifuged for 10 min at 500 g and the pellet resuspended in 10–15 ml red cell lysis buffer (Sigma-Aldrich, Dorset, UK) for 10 min to remove contaminating erythrocytes. Following a further 5 min centrifugation at 500 g, the supernatant was removed and the cells were resuspended in RPMI containing 15% FCS, 2 mM L-glutamine, 50 IU penicillin and 50 μg ml−1 streptomycin, plus 10–50% autologous ascitic fluid. The cells were plated in culture vessels according to cell density (usually 2 × 25 cm2 culture vessels) and maintained in RPMI plus 15% FCS after initial passage. The primary ovarian cells used for these studies were at passage 3 and are the same cells as described in detail by Ingram et al.20
The luciferase gene-containing plasmid, pGL3-control (Promega, Southampton, UK) was used as the basis for the reporter constructs. The SV40 promoter of pGL3-control was replaced with the mCMV or mIL-2 promoter at the BglII and HindIII restriction sites (Figures 1a and b). The mIL-2 promoter was amplified by polymerase chain reaction using the primers: mIL-2For, 5′-IndexTermagatctaacattttgacacc-3′ and mIL-2Rev, 5′-IndexTermaagcttgtggcaggagttgag-3′ using the pLH-Z12I-PL vector (Ariad Argent now Clontech, Saint-Germain-en-Laye, France) as a template. In a similar way, the mCMV promoter was amplified using the primers: mCMVFor, 5′-IndexTermagatctgagtaggcgtgtacgg-3′ and mCMVRev, 5′-IndexTermaagcttgaggctggatcggtc-3′ and the pTRE2 vector (Clontech) as a template. These vectors were named pGL3-mIL2 and pGL3-mCMV accordingly and used for the insertion of the tandemly repeated HREs at the BglII and HindIII restriction sites (Figures 1a and b).
Five repeats of the VEGF and EPO HREs were derived from the 3′-flanking region of the human EPO and the 5′-flanking region of the human VEGF gene were cloned into a modified pTRE2 vector (Clontech) by analogy to the method described by Post et al.38 Briefly, a modified pTRE2 vector was generated with the enhancer element removed by inserting a second XhoI recognition site downstream of the enhancer and restriction digest with XhoI (vector kindly provided by D Jevremovic, Mayo Clinic, Rochester, MI). The modified pTRE2 vector was then used as a cloning vector after dephosphorylation of the linearized plasmid. EPO and VEGF HRE oligonucleotides (see Table 1) were annealed before insertion into the linearized modified pTRE2 vector. The oligonucleotides were designed with overhanging 5′ and 3′ ends with added XhoI (5′-IndexTermCTCGAG-3′) and SalI (5′-IndexTermGTCGAC-3′) restriction sites, respectively. Insertion of the annealed oligonucleotide into a linearized vector pre-digested with XhoI changes the SalI 3′ recognition site from 5′-IndexTermGTCGAC-3′ to 5′-IndexTermGTCGAG-3′, thereby eliminating the restriction site. The vector with the inserted oligonucleotide is then re-digested with XhoI (using the remaining recognition site at the 5′ end of the inserted oligonucleotide), thus allowing addition of further oligonucleotides. This process is then repeated until five tandem repeats of EPO or VEGF HREs have been incorporated. Correct insertion of oligonucleotides was confirmed by restriction digest and sequencing. Five tandem copies of the respective HRE were then amplified by polymerase chain reaction using the modified pTRE2 vectors containing five HRE tandem repeats as a template with primers incorporating MluI and XmaI restriction sites to allow directional insertion into pGL3-mIL2.
Ad-CMV-NTR (CTL102) was kindly provided by Innovata plc. (Nottingham, UK). Other therapeutic recombinant adenovirus constructs, Ad-mCMV-HSVtk, Ad-VEGFmCMV-HSVtk, Ad-mCMV-NTR and Ad-VEGFmCMV-NTR, were generated using the Ad-easy system (Clontech). Initially, the luciferase gene in the pGL3 constructs was replaced with either the HSVtk or bacterial NTR suicide gene, using the HindIII and XbaI restriction enzyme sites. The HSVtk gene was obtained from the plasmid pBS-sk/MEEP-tk39, 40 (kindly provided by Dr Georges Vassaux, Nantes, France) and the NTR gene was amplified from pd2NTR-Basic141 (a gift from Professor Nicol Keith, CRUK, Glasgow, UK). From these constructs, fragments containing the complete prodrug enzyme gene were amplified by polymerase chain reaction to incorporate NotI sites at each end and inserted into pGEMTeasy for sequencing. The following primers were used: VEGFFor, 5′-IndexTermgcggccgctcgagccacagtgc-3′; mCMVFor, 5′-IndexTermgcggccgcgagtaggcgtgtacggt-3′ and PolyARev, 5′-IndexTermgcggccgctaccacatttgtagaggtttt-3′. The amplicons were inserted at the NotI site in the pShuttle vector (Clontech).
Adenovirus production and infection
The pShuttle constructs were prepared using the Ad-easy system (Clontech), according to the manufacturer's instructions. Briefly, the insert was amplified from the corresponding sequence in the pGL3 constructs, as described above, and the final PacI-linearized plasmids were transfected into 3 × 106 HEK293 cells, previously seeded into a 60-mm culture dish. The recombinant viruses were amplified, purified and quality checked for recombination by ViraQuest Inc. (North Liberty, IA). The viral titer was also provided by ViraQuest.
Western blot analysis for HIF-1α
Six-well plates were seeded with 4 × 105 tumor cells (established cell lines or primary ovarian cultures), and left overnight at 37°C. Cells were then subjected to normoxic or hypoxic conditions (1% O2 or 150 μM CoCl2) for 24 h. Cells were washed with cold phosphate-buffered saline and nuclear extracts were prepared using a modified Dignam method.42 Cells were scraped into 1.5 ml of cold phosphate-buffered saline and then transferred to an Eppendorf tube and pelleted for 20 s at 0.5 g at 4°C. To obtain nuclear extracts, the cell pellets were resuspended in 400 μl cold Buffer A (10 mM HEPES-KOH (pH 7.9), 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol, 1 mM sodium vanadate (Na3VO4) plus a cocktail of protease inhibitors (Sigma)) and incubated on ice for 10 min, followed by vortexing for 10 s. The samples were centrifuged for 10 s at 2.1 g at 4°C and the pellet resuspended in 20–40 μl of cold buffer C (20 mM HEPES-KOH (pH 7.9), 25% glycerol, 420 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM dithiothreitol, 1 mM Na3VO4 plus a cocktail of protease inhibitors) and kept on ice for 20 min. The cellular membrane was pelleted by centrifugation for 2 min at 15.7 g at 4°C and the supernatant was frozen in 10–20 μl aliquots at −80°C. In all, 15 μg of the nuclear extracts were electrophoresed on a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel and the membrane was hybridized with an HIF-1α monoclonal primary antibody (BD Biosciences, Oxford, UK) at a dilution of 1:1000, and then with a goat anti-mouse horseradish peroxidase secondary antibody (Southern Biotech, Birmingham, AL, USA) at a 1:2000 dilution. Bands were then visualized by enhanced chemiluminescence using an ECL Plus kit (GE Healthcare, Little Chalfont, UK).
Transfection and hypoxic transgene expression
For luciferase reporter assays, cells were seeded into 24-well plates at 2 × 105 cells per well and transfected 24 h later with 200 ng per well of plasmid DNA using Effectene (Qiagen, Crawley, UK), according to the manufacturer's instructions. At 24 h post-transfection, cells were exposed to hypoxia (1% O2) for a further 48 h before harvesting for luciferase assay. To control for transfection efficiency between samples, pGL3-lacz was transfected in parallel and visualized by lacz staining. Samples analyzed for luciferase activity were normalized using the protein levels of each sample and the fold induction was calculated as a (H−N)/N ratio for each cell line.
Following transfection and hypoxic transgene expression, cells were washed with phosphate-buffered saline and then 100 μl of reporter lysis buffer was added. The cell lysates were harvested by scraping and stored at −20 °C in 1.5 ml Eppendorf tubes. The lysates were then analyzed for luciferase activity using the Steady-Glo system (Promega) as per the manufacturer's instructions and luminescence was measured using a MicroBeta counter (Perkin Elmer, Cambridge, UK) at 1 s intervals.
Cell proliferation assay
For prodrug enzyme cytotoxicity assays, 1 × 106 cells per ml were infected with adenoviruses in suspension for 1.5 h at 37°C at a multiplicity of infection (MOI) of 100 (1 × 108 PFU), for the cell lines or MOI of 10 (1 × 107 PFU) for the primary cultures. The inoculum was removed and fresh medium added. For the cell lines, cells were diluted and plated out at 3000 cells per well in 200 μl of medium in 96-well plates; for primary ovarian cultures, cells were plated out at 6000 cells per well. At 48 h post-infection, the appropriate prodrug was added at varying concentrations (0, 1, 3, 10, 30, 100 and 300) and the cells were exposed to hypoxic conditions (1% O2 or 150 μM CoCl2) for 3 or 5 days for NTR and HSVtk, respectively. Cell viability was analyzed by WST-1 assay (Roche Diagnostics, Burgess Hill, UK). Briefly, 75 μl of 1 × WST-1 reagent was added and the plate incubated for 2 h at 37°C, and then the absorbance measured at 450 nm using a Thermo Electron multiscan plate reader. The data are presented as a percentage of viable cells compared to the untreated cells, expressed as 100% survival. The absorbance of untreated cells is 100% survival. The half-maximal inhibitory concentration (IC50) values were calculated as the concentration of prodrug giving 50% of this absorbance. The P-values were calculated using the unpaired t-test. Values were considered statistically significant at the P<0.05 level.
Female Balb/C immunodeficient mice (Harlan, Blackthorn, UK) aged 6–8 weeks were used. Mice received Harlan 2018 diet (Harlan) and water ad libitum. Mice were kept in cages in an air-conditioned room with regular alternating cycles of light and darkness. All animal procedures were carried out under a project license issued by the UK Home Office and UKCCCR guidelines43 were followed throughout.
Tumors were excised from a donor animal, placed in sterile physiological saline containing penicillin and streptomycin and cut into small fragments of approximately 2 mm3. Under brief general inhalation anesthesia, HCT116 tumor fragments were implanted into one or both flanks of each mouse, as appropriate, using a trocar. Once the tumors had reached a volume of 100–200 mm3 as measured by calipers, mice were allocated into cages of 8 (treatment groups) or 4 (control groups) by restricted randomization to keep group mean tumor size variation to a minimum.
In vivo tumor reduction experiments
A total of 1.2 × 109 PFU of each adenovirus or saline as control was injected intratumorally at three sites per tumor in 20 μl volumes per site, with the day of injection designated day −3. CB1954 was administered intraperitoneally at 20 mg kg−1 on day 0 to day 5. The effect of therapy was assessed as described previously.44 Briefly, two-dimensional caliper measurements of the tumors were taken daily, with volumes calculated using the formula (a2 × b)/2, where a is the smaller and b the larger diameter of the tumor. Tumor volume was then normalized to the respective volume on day 0, and log plots of relative tumor volume vs time were made. Unpaired t-tests were performed to determine the statistical significance in growth rate (based on tumor volume doubling time, relative tumor volume 2) between all groups.
Tumors were formalin fixed, paraffin embedded and 4-μm sections were cut for the immunostaining of consecutive sections with Glut1 and NTR. Sections were de-paraffinized and rehydrated before immunohistochemistry. Slides were treated in xylene for 3 × 5 min and then rehydrated through absolute alcohol to water, that is, 3 × 1 min in 100% ethanol and then 5 min in running tap water. For Glut1 antibody staining, heat-mediated antigen retrieval was performed by placing the slides in 0.2 M citric acid (pH 6.0) (pre-heated for 2 min in microwave) and heating for 10 min in a microwave (800 W) on full power.
The slides were cooled to room temperature for 20 min and then rinsed briefly in Tris-buffered saline (TBS). Endogenous peroxidase was blocked with 0.3% hydrogen peroxide (Sigma) in methanol for 10 min and rinsed well in running tap water for 5 min. Sections were then incubated with Glut1 antibody (Abcam, Cambridge, UK) at a dilution of 1:100 for 1 h at room temperature. Slides were rinsed in 0.025% Tween-TBS (TBST) twice for 10 min and then TBS for 5 min. An anti-rabbit horseradish peroxidase-conjugated antibody from the Rabbit Envision kit (Dako, Ely, UK) was added for 30 min, followed by 2 × 5 min TBST and 1 × 5 min TBS washes. Antibody staining was developed using DAB Plus (Dako) and slides counterstained with Mayer's hematoxylin for 2 min. The slides were then dehydrated, cleared and mounted for microscopy.
For nitroreductase staining, no antigen retrieval was performed and endogenous peroxidase was blocked as described above. Slides were rinsed twice in TBS for 5 min and then blocked with an avidin/biotin blocking kit (Vector Laboratories, Burlingame, CA) and rinsed in TBS. A final blocking step was performed with 1 × casein (Vector Laboratories) for 20 min. Slides were then incubated with a sheep anti-NTR antibody (kindly provided by Professor Peter Searle) at a dilution of 1:1000 for 1 h at room temperature. All antibodies were diluted in antibody diluent buffer (Zymed, Invitrogen, Paisley, UK). Slides were washed twice between steps with TBST and once with TBS for 10 min each. A rabbit anti-goat biotinylated secondary antibody was incubated for 30 min and then a streptavidin peroxidase complex (Vector Laboratories) was added for 30 min. Antibody staining was developed using DAB Plus (Dako) and slides counterstained with Mayer's hematoxylin for 2 min. The slides were then dehydrated, cleared and mounted for microscopy.
Hypoxic induction of HIF-1α expression in human cancer cell lines
Hypoxic induction of HIF-1α expression was confirmed in a panel of cancer cell lines derived from different human tissues and a primary culture of human ovarian cancer by western blot analysis (Figure 1c). All cell lines tested showed very little or no HIF-1α expression under normoxic conditions, but significantly increased expression under hypoxic conditions (1% O2). Similar results were obtained using 150 μM CoCl2 to induce hypoxia (data not shown). Particularly high levels of HIF-1α expression were seen in the HT1080 (human fibrosarcoma) and HCT116 (colorectal carcinoma) cell lines under hypoxic conditions. Interestingly, primary cultures of ovarian cancer cells obtained from patients’ ascites also demonstrated high levels of HIF-1α expression, suggesting hypoxic induction.
Comparison of VEGF and EPO HREs
The same cancer cell lines used in Figure 1c were transiently transfected with luciferase reporter plasmids, with or without five repeats of VEGF or EPO HREs, inserted upstream of either an mCMV or an mIL-2 promoter (Figure 2). Constructs containing the mCMV promoter demonstrated up to eightfold increased luciferase expression after 48 h under hypoxic conditions with the VEGF HREs, and up to sixfold with EPO HREs (Figure 2a), depending on the cell line, with HCT116 showing particularly high levels of hypoxic induction. Similarly, up to ninefold hypoxic induction was seen with the mIL-2 promoter (Figure 2b). However, the absolute levels of luciferase activity due to hypoxic induction seen with the mIL-2 promoter plasmid constructs were approximately two orders of magnitude lower than those observed with the mCMV promoter (102–104 vs 104–106; data not shown). As the therapeutic efficacy of hypoxia-inducible transgene expression is likely to require high levels of expression as well as a significant differential in expression between normoxia and hypoxia, constructs driven by the mCMV promoter were used for all subsequent experiments.
In all the cell lines tested, constructs containing five repeats of VEGF and the mCMV synthetic promoter demonstrated luciferase activity at least as high as, or higher than, the five repeats of EPO and the mCMV promoter under hypoxic conditions (Figure 2a). Particularly high levels of hypoxic expression were seen with JON and HCT116 cells and moderate levels in HT1080 cells. Five repeats of the VEGF HRE, coupled to an mCMV promoter, were used for further studies in HCT116 and HT1080 cells to investigate the relative cytotoxicity of HSVtk and NTR prodrug-activating enzyme genes, as JON cells demonstrated much lower levels of transfection efficiency.
Comparison of different prodrug-activating enzyme genes
Having identified our optimum hypoxia-inducible HRE-promoter system, we sought to compare the HSVtk and NTR prodrug-activating enzyme systems. We constructed a panel of recombinant replication-defective adenoviral vectors containing the HSVtk or the bacterial NTR therapeutic transgenes. These viral vectors were used to compare the cytotoxicity of these two prodrug enzyme systems under normoxic and hypoxic conditions. HT1080 and HCT116 cells used in previous experiments both showed efficient transduction, as assessed by FACS analysis, using AdGFP virus with 99% and 95% positive cells, respectively, at an MOI of 100.
Both Ad-CMV-HSVtk- and Ad-CMV-NTR-positive control viruses, in which the transgene is driven by the full CMV promoter, showed strong cytotoxic effects under both normoxic and hypoxic conditions in each of these cell lines (Figure 3). Our hypoxia-inducible Ad-VEGFmCMV-HSVtk virus demonstrated hypoxia-specific cytotoxic effects in the HT1080 fibrosarcoma cells in the presence, but not in the absence, of GCV. Induction of hypoxia by either low oxygen or chemical induction using 150 μM CoCl2 produced similar results with an IC50 of 5.9 and 31 μM under hypoxia for CoCl2 and low oxygen, respectively (Figures 3a and c and Table 2). Similarly, Ad-VEGFmCMV-NTR demonstrated a considerable cytotoxic effect under hypoxic conditions in HT1080 cells with an IC50 of 4–18 μM of the prodrug CB1954 (Figures 3b and d and Table 2). In all cases, the IC50 in hypoxic conditions was similar to that for the viruses in which prodrug-activating enzyme expression was driven by the full CMV promoter under either normoxic or hypoxic conditions.
For the HCT116 cells, with the Ad-VEGFmCMV-HSVtk under hypoxic conditions, a strong cytotoxic effect was observed with an IC50 of 1 μg ml−1 (3.9 μM) (Figures 3e and g). A similar level of cell kill was obtained with the Ad-VEGFmCMV-HSVtk under hypoxic conditions to that obtained with the full CMV promoter under either hypoxic or normoxic conditions. A significant cytotoxic effect was also seen with the NTR transgene in Ad-VEGFmCMV-NTR virus under hypoxic conditions, with 50% of cell death occurring at around 15 μM (Figures 3f and h). The hypoxia-specific effect seen in the HCT116 cells was more pronounced than in the HT1080 cell line, reaching statistical significance at the P<0.05 level. The HCT116 cell line was more susceptible under normoxic conditions than the HT1080 cells, but killed at lower GCV concentrations, which is more desirable when considering the cell line to use for in vivo models. The IC50 values observed with both HSVtk/GCV and NTR/CB1954 are comparable to values obtained in previous studies.18, 19
Cytotoxicity of adenovirus-delivered prodrug enzyme systems in human primary ovarian cancer cells
Primary ovarian cancer cells taken directly from patient ascites are not easily transduced by adenovirus. However, an early adaptation to culture is to become adherent, with consequent alterations in phenotype, which results in enhanced infectivity. These adherent cells are a more representative model for ovarian cancer because, as shown by Ingram et al.,20 solid ovarian tumors are highly susceptible to adenovirus infection and are a better representative of the true model of ovarian cancer cells. Thus, the use of human primary tumor cells provides a more realistic assessment of the potential value of VDEPT in the clinical setting than the use of established tumor cell lines. Therefore, the Ad-VEGFmCMV-HSVtk and Ad-VEGFmCMV-NTR viruses were used to transduce primary human ovarian cancer cells, obtained from ascitic fluid from patients, to examine their ability to induce cell death under hypoxic conditions. The primary tumor cells were easily transducible at an MOI of 10 as demonstrated previously.20 Adenoviruses containing the mCMV promoter and the full CMV promoter were included in these experiments as negative and positive controls, respectively. The use of the mCMV promoter alone did not produce any significant cytotoxic effect. The full CMV promoter produced highly efficient cell kill, under both normoxic and hypoxic conditions with both prodrug-activating genes, but greater cytotoxicity was seen with NTR/CB1954 system. A significant hypoxia-specific cytotoxic effect was seen with Ad-VEGF-HSVtk/GCV, which resulted in around 65% cell death at 5 days, comparable to that observed with the positive control Ad-CMV-HSVtk, whereas there was <20% cell kill under normoxic conditions (Figure 4a). Ad-VEGF-NTR demonstrated an even greater differential cell kill under hypoxia with up to 97% cell death, again similar to the positive control with the full CMV promoter (Figure 4b) and considerably greater than the 25% seen under normoxia.
Antitumor efficacy of Ad-VEGFmCMV-NTR with CB1954 in HCT116 tumor xenografts
For in vivo testing of our hypoxia-inducible system, we selected the Ad-VEGFmCMV-NTR virus/CB1954 prodrug-activating system. This was on the basis that it demonstrated a steeper dose:response curve in established lines, and showed a better differential cell kill between normoxia and hypoxia in primary cultures of ovarian cancer cells, both suggesting a potentially beneficial efficacy:toxicity ratio in vivo.
Ad-VEGFmCMV-NTR was injected intratumorally into HCT116 xenografts in nude mice (Figure 5). The Ad-VEGFmCMV-NTR showed a significant delay (P<0.01) in tumor growth over a 15-day time course when CB1954 was administered intraperitoneally compared with Ad-VEGFmCMV-NTR alone. Conversely, Ad-mCMV-NTR, with or without CB1954, did not demonstrate any significant effect on tumor growth. Ad-VEGFmCMV-NTR also showed a significant effect compared with the Ad-mCMV-NTR groups (P<0.05) in the presence of CB1954, but not in its absence.
Localization of adenovirus transgene expression in hypoxic regions of HCT116 xenograft models
Our hypothesis is that the improved cytotoxicity of the Ad-VEGFmCMV-NTR virus seen in vivo is due to its ability to target selectively hypoxic areas of tumor. To test this hypothesis, immunostaining experiments were performed with anti-NTR and anti-Glut1 antibodies to establish whether there is a correlation between tumor areas expressing the virally encoded transgene and areas of hypoxia. Tumors were excised once the animals had been killed and adjacent sections were stained using Glut1 as a marker of hypoxia and an antibody against NTR as a marker of adenoviral transgene expression.
Tumors from mice injected with the Ad-VEGFmCMV-NTR virus and administered with or without CB1954 showed similar staining patterns for Glut1 and NTR. That is, the hypoxic regions within the tumor show Glut1 staining of necrotic cells. The cells surrounding this region show more defined Glut1 staining with characteristic membrane staining for Glut1 (Figures 6a, c, e and g). As expected, Glut1 staining was also seen in the absence of virus (Figure 6i) and in the presence of virus, but in the absence of the prodrug (Figure 6e), demonstrating that the hypoxic regions of tumors are not the result of the presence of virus alone, but rather the virus–prodrug combination. Staining for NTR showed the prodrug-activating enzyme to be present within the same region surrounding the hypoxic cores (Figures 6b, d, f and h), but not in the absence of virus (Figure 6j). Thus, the adenovirus was demonstrated to colocalize to areas of hypoxic cells within the tumors of mice and efficiently express the NTR transgene. No NTR staining was seen in negative controls without the primary antibody (data not shown).
Hypoxia plays a critical role in tumor growth and development. Resistance to conventional chemo- and radiotherapies in hypoxic cells is due to the different survival factors induced in these cell types. Thus, there is a need to establish new therapeutic agents, which selectively target hypoxic regions of tumors. To help overcome this resistance and aid tumor eradication, in combination with conventional therapies, we have developed hypoxia-inducible adenoviral vectors using HREs derived from human VEGF and EPO gene promoters, driving either the HSVtk or NTR prodrug-activating enzyme genes to target the hypoxic areas of solid tumors.
HSVtk/GCV and NTR/CB1954 have been widely used in a number of cytotoxic gene therapy studies, including therapies targeting hypoxia.21, 22, 23, 24 Only a few studies have compared both HSVtk and NTR systems in parallel, but none of these comparisons has been under hypoxic conditions with different transcriptional regulation. In one study, HSVtk/GCV showed the widest therapeutic index, whereas NTR/CB1954 demonstrated a stronger bystander effect in a retroviral vector.25 CB1954 is known to be activated by one-electron reductases selectively in hypoxic conditions; thus, an increased cytotoxic effect may be observed. However, a report by Wilson et al.26 showed no increased or decreased effect on the bystander activity of CB1954 due to hypoxia. Others found NTR/CB1954 to suppress tumor growth more efficiently in vivo than HSVtk/GCV.27 Both prodrug-activating systems demonstrate efficacy in a variety of models. However, the use of HSVtk/GCV in cancer gene therapy may be limited owing to the requirement of target cell division, as HSVtk/GCV is only active in dividing cells. NTR/CB1954, on the other hand, has the potential advantage over the HSVtk/GCV prodrug-activating system that it is effective in both dividing and non-dividing cells.28
Previous studies have shown that CoCl2 is just as efficient as low oxygen levels in inducing a hypoxic environment. In common with several previous studies,29, 30 which used both conditions in parallel, our data using either 1% oxygen or 150 μM CoCl2 demonstrated similar results. This was the case with either HSVtk/GCV or NTR/CB1954 prodrug-activating system. Both the VEGF and EPO HREs are capable of inducing a similar hypoxia-inducible response, but the mCMV promoter induces stronger gene expression than the mIL-2 promoter.
Both therapeutic transgenes showed a significant and selective in vitro response under hypoxia compared with normoxia when driven by HREs. The HSVtk/GCV system demonstrated a greater cytotoxic effect than the NTR/CB1954 in both the HT1080 and HCT116 cells. However, use of these viruses to target hypoxia in human primary ovarian cancer cells found a more marked effect with the NTR/CB1954 system, with 95% cell death observed under hypoxia.
Similar to this study, previous in vivo studies with NTR/CB1954 have shown significant reduction in tumor volume and increases in animal survival.31, 32, 33, 34, 35 Shibata et al.8 used the NTR/CB1954 system in a xenograft model with a VEGF HRE to examine the effects on hypoxia. The NTR gene was stably transfected into fibrosarcoma cells before tumor formation in mice. Mice were exposed to and breathed 10% oxygen to enhance tumor hypoxia in order for cells to exhibit a significant tumor growth delay. Exposure to 10% oxygen is not ideal as this may also cause hypoxia in normal tissues, not just in the tumor. Other in vivo studies targeting hypoxia using alternative prodrug-activated systems found that tumors of 100–200 mm3 display hypoxic conditions without the need for exposure of mice to low oxygen levels, which is a much more physiological approach.36, 37 Indeed, immunostaining of our xenografts showed intrinsic hypoxic regions within the tumors as determined by Glut1 staining. Sections from these xenografts showed adenoviral transgene expression within the hypoxic regions of the tumors, as determined by comparing the NTR staining with the Glut1 staining. These experiments used replication-defective adenovirus, and as tumor xenografts were harvested up to 15 days after treatment with virus, it is almost certain that there would no longer be any detectable adenoviral hexon capsid protein. Furthermore, as no endogenous expression of the bacterial protein NTR would be expected in mammalian cells, and as no staining with anti-NTR antibody was seen in the absence of virus, it seems reasonable to infer that NTR expression is indicative of the presence of adenovirus. Thus, the colocalization of the Glut1 and NTR staining suggests that the positive therapeutic effect of the virus may well be via hypoxia-specific induction of therapeutic transgene expression.
To our knowledge, this is the first time such a detailed comparative study of the performance of prodrug-activating systems HSVtk/GCV and NTR/CB1954 under different promoters has been carried out in hypoxic conditions. Our study is potentially informative for future studies of the effect of hypoxia-targeting VDEPT systems. Further in vivo efficacy studies for this system should include combination of VDEPT with radio- and chemotherapy treatments, with the aim of improving the therapeutic index for the treatment of solid tumors. The successful development of such gene therapy systems targeting hypoxic regions of tumors are urgently needed and warrant further investigation.
herpes simplex virus thymidine kinase
vascular endothelial growth factor
Brown JM . Exploiting the hypoxic cancer cell: mechanisms and therapeutic strategies. Mol Med Today 2000; 6: 157–162.
Vaupel P, Mayer A . Hypoxia in cancer: significance and impact on clinical outcome. Cancer Metast Rev 2007; 26: 225–239.
Fechner G, Muller G, Schmidt D, Garbe S, Hauser S, Vaupel P et al. Evaluation of hypoxia-mediated growth factors in a novel bladder cancer animal model. Anticancer Res 2007; 27: 4225–4231.
Binley K, Askham Z, Iqball S, Spearman H, Martin L, de Alwis M et al. Long-term reversal of chronic anemia using a hypoxia-regulated erythropoietin gene therapy. Blood 2002; 100: 2406–2413.
Chadderton N, Cowen RL, Sheppard FC, Robinson S, Greco O, Scott SD et al. Dual responsive promoters to target therapeutic gene expression to radiation-resistant hypoxic tumor cells. Int J Radiat Oncol Biol Phys 2005; 62: 213–222.
Koshikawa N, Takenaga K, Tagawa M, Sakiyama S . Therapeutic efficacy of the suicide gene driven by the promoter of vascular endothelial growth factor gene against hypoxic tumor cells. Cancer Res 2000; 60: 2936–2941.
Kaliberov SA, Buchsbaum DJ, Gillespie GY, Curiel DT, Arafat WO, Carpenter M et al. Adenovirus-mediated transfer of BAX driven by the vascular endothelial growth factor promoter induces apoptosis in lung cancer cells. Mol Ther 2002; 6: 190–198.
Shibata T, Giaccia AJ, Brown JM . Hypoxia-inducible regulation of a prodrug-activating enzyme for tumor-specific gene therapy. Neoplasia 2002; 4: 40–48.
Lee S, Kim K, Kim HA, Kim SW, Lee M . Augmentation of erythropoietin enhancer-mediated hypoxia-inducible gene expression by co-transfection of a plasmid encoding hypoxia-inducible factor 1alpha for ischemic tissue targeting gene therapy. J Drug Target 2008; 16: 43–50.
Okabe S, Arai T, Yamashita H, Sugihara K . Adenovirus-mediated prodrug-enzyme therapy for CEA-producing colorectal cancer cells. J Cancer Res Clin Oncol 2003; 129: 367–373.
Ji X, Zhang J, Cheng L, Wei F, Li H, Liu X et al. Oncolytic adenovirus delivering herpes simplex virus thymidine kinase suicide gene reduces the growth of human retinoblastoma in an in vivo mouse model. Exp Eye Res 2009; 89: 193–199.
Dorer DE, Nettelbeck DM . Targeting cancer by transcriptional control in cancer gene therapy and viral oncolysis. Adv Drug Deliv Rev 2009; 61: 554–571.
Greco O, Dachs GU . Gene directed enzyme/prodrug therapy of cancer: historical appraisal and future prospectives. J Cell Physiol 2001; 187: 22–36.
Wilson WR, Hicks KO, Pullen SM, Ferry DM, Helsby NA, Patterson AV . Bystander effects of bioreductive drugs: potential for exploiting pathological tumor hypoxia with dinitrobenzamide mustards. Radiat Res 2007; 167: 625–636.
Nasu Y, Saika T, Ebara S, Kusaka N, Kaku H, Abarzua F et al. Suicide gene therapy with adenoviral delivery of HSV-tK gene for patients with local recurrence of prostate cancer after hormonal therapy. Mol Ther 2007; 15: 834–840.
Ahn M, Lee SJ, Li X, Jimenez JA, Zhang YP, Bae KH et al. Enhanced combined tumor-specific oncolysis and suicide gene therapy for prostate cancer using M6 promoter. Cancer Gene Ther 2009; 16: 73–82.
Patel P, Young JG, Mautner V, Ashdown D, Bonney S, Pineda RG et al. A phase I/II clinical trial in localized prostate cancer of an adenovirus expressing nitroreductase with CB1984. Mol Ther 2009; 17: 1292–1299.
Weedon SJ, Green NK, McNeish IA, Gilligan MG, Mautner V, Wrighton CJ et al. Sensitisation of human carcinoma cells to the prodrug CB1954 by adenovirus vector-mediated expression of E. coli nitroreductase. Int J Cancer 2000; 86: 848–854.
Jaberipour M, Vass SO, Guise CP, Grove JI, Knox RJ, Hu L et al. Testing double mutants of the enzyme nitroreductase for enhanced cell sensitisation to prodrugs: effects of combining beneficial single mutations. Biochem Pharmacol 2010; 79: 102–111.
Ingram N, MacCormac LP, Oxley NT, Burns PA, Hall GD . Role of cell surface molecules and autologous ascitic fluid in determining efficiency of adenoviral transduction of ovarian cancer cells. Cancer Gene Ther 2010; 17: 684–693.
Ahn GO, Brown M . Targeting tumors with hypoxia-activated cytotoxins. Front Biosci 2007; 12: 3483–3501.
Greco O, Marples B, Joiner MC, Scott SD . How to overcome (and exploit) tumor hypoxia for targeted gene therapy. J Cell Physiol 2003; 197: 312–325.
Palmer DH, Young LS, Mautner V . Cancer gene-therapy: clinical trials. Trends Biotechnol 2006; 24: 76–82.
Ingram N, Porter CD . Transcriptional targeting of acute hypoxia in the tumour stroma is a novel and viable strategy for cancer gene therapy. Gene Ther 2005; 12: 1058–1069.
Nishihara E, Nagayama Y, Narimatsu M, Namba H, Watanabe M, Niwa M et al. Treatment of thyroid carcinoma cells with four different suicide gene/prodrug combinations in vitro. Anticancer Res 1998; 18: 1521–1525.
Wilson WR, Pullen SM, Hogg A, Helsby NA, Hicks KO, Denny WA . Quantitation of bystander effects in nitroreductase suicide gene therapy using three-dimensional cell cultures. Cancer Res 2002; 62: 1425–1432.
Heinkelein M, Hoffmann U, Lucke M, Imrich H, Muller JG, Meixensberger J et al. Experimental therapy of allogeneic solid tumors induced in athymic mice with suicide gene-transducing replication-competent foamy virus vectors. Cancer Gene Ther 2005; 12: 947–953.
Aghi M, Hochberg F, Breakefield XO . Prodrug activation enzymes in cancer gene therapy. J Gene Med 2000; 2: 148–164.
Hayashi M, Sakata M, Takeda T, Yamamoto T, Okamoto Y, Sawada K et al. Induction of glucose transporter 1 expression through hypoxia-inducible factor 1alpha under hypoxic conditions in trophoblast-derived cells. J Endocrinol 2004; 183: 145–154.
Poulios E, Trougakos IP, Gonos ES . Comparative effects of hypoxia on normal and immortalized human diploid fibroblasts. Anticancer Res 2006; 26: 2165–2168.
Braidwood L, Dunn PD, Hardy S, Evans TR, Brown SM . Antitumor activity of a selectively replication competent herpes simplex virus (HSV) with enzyme prodrug therapy. Anticancer Res 2009; 29: 2159–2166.
White CL, Menghistu T, Twigger KR, Searle PF, Bhide SA, Vile RG et al. Escherichia coli nitroreductase plus CB1954 enhances the effect of radiotherapy in vitro and in vivo. Gene Ther 2008; 15: 424–433.
Benouchan M, Do Nascimento F, Perret GY, Colombo BM . Delivery of the bacterial nitroreductase gene into endothelial cells prolongs the survival of tumour-bearing mice by bystander mechanisms. Int J Oncol 2006; 28: 457–462.
Lukashev AN, Fuerer C, Chen MJ, Searle P, Iggo R . Late expression of nitroreductase in an oncolytic adenovirus sensitizes colon cancer cells to the prodrug CB1954. Hum Gene Ther 2005; 16: 1473–1483.
Chen MJ, Green NK, Reynolds GM, Flavell JR, Mautner V, Kerr DJ et al. Enhanced efficacy of Escherichia coli nitroreductase/CB1954 prodrug activation gene therapy using an E1B-55K-deleted oncolytic adenovirus vector. Gene Ther 2004; 11: 1126–1136.
Post DE, Sandberg EM, Kyle MM, Devi NS, Brat DJ, Xu Z et al. Targeted cancer gene therapy using a hypoxia inducible factor dependent oncolytic adenovirus armed with interleukin-4. Cancer Res 2007; 67: 6872–6881.
Cho WK, Seong YR, Lee YH, Kim MJ, Hwang KS, Yoo J et al. Oncolytic effects of adenovirus mutant capable of replicating in hypoxic and normoxic regions of solid tumor. Mol Ther 2004; 10: 938–949.
Post DE, Van Meir EG . Generation of bidirectional hypoxia/HIF-responsive expression vectors to target gene expression to hypoxic cells. Gene Ther 2001; 8: 1801–1807.
Vassaux G, Hurst HC, Lemoine NR . Insulation of a conditionally expressed transgene in an adenoviral vector. Gene Ther 1999; 6: 1192–1197.
Ring CJ, Harris JD, Hurst HC, Lemoine NR . Suicide gene expression induced in tumour cells transduced with recombinant adenoviral, retroviral and plasmid vectors containing the ERBB2 promoter. Gene Ther 1996; 3: 1094–1103.
Plumb J, Norlin C, Young PC . Exposures and outcomes of children with urticaria seen in a pediatric practice-based research network: a case–control study. Arch Pediatr Adolesc Med 2001; 155: 1017–1021.
Andrews PA, Jones JA . Characterization of binding proteins from ovarian carcinoma and kidney tubule cells that are specific for cisplatin modified DNA. Cancer Commun 1991; 3: 93–102.
Workman P, Twentyman P, Balkwill F, Balmain A, Chaplin D, Double J, Embleton J et al. United Kingdom Co-ordinating Committee on Cancer Research (UKCCCR) Guidelines for the Welfare of Animals in Experimental Neoplasia (Second Edition). Br J Cancer 1998; 77: 1–10.
Shnyder SD, Cooper PA, Pettit GR, Lippert III JW, Bibby MC . Combretastatin A-1 phosphate potentiates the antitumour activity of cisplatin in a murine adenocarcinoma model. Anticancer Res 2003; 23: 1619–1623.
This work and all authors were supported by Cancer Research UK. We thank Sarah Perry for all the help and advice with the immunohistochemistry and to Luci MacCormac for the primary human ovarian cultures used in this study.
The authors declare no conflict of interest.
About this article
Cite this article
Harvey, T., Hennig, I., Shnyder, S. et al. Adenovirus-mediated hypoxia-targeted gene therapy using HSV thymidine kinase and bacterial nitroreductase prodrug-activating genes in vitro and in vivo. Cancer Gene Ther 18, 773–784 (2011). https://doi.org/10.1038/cgt.2011.43
International Journal of Oncology (2017)
Cell physiology regulation by hypoxia inducible factor-1: Targeting oxygen-related nanomachineries of hypoxic cells
International Journal of Biological Macromolecules (2017)
Structural basis of Escherichia coli nitroreductase NfsB triple mutants engineered for improved activity and regioselectivity toward the prodrug CB1954
Process Biochemistry (2015)
Free Radical Biology and Medicine (2015)
Lentivirus-mediated RASSF1A expression suppresses aggressive phenotypes of gastric cancer cells in vitro and in vivo
Gene Therapy (2015)