Prodrug-mediated ablation of transduced target cells is now a widely investigated approach for the selective chemotherapy of malignant tumors. Currently, one of the most efficient and extensively studied prodrug activation systems is the bacterial cytosine deaminase (CDase) in combination with 5-fluorocytosine (5-FC).1, 2, 3, 4, 5, 6, 7 CDase, an enzyme present in fungi and bacteria8, 9 but absent in mammalian cells,10 deaminates the nontoxic prodrug 5-FC to its highly toxic derivative 5-fluorouracil (5-FU).

The cytotoxicity of 5-FU is determined by its conversion into 5-fluorouridine triphosphate (5-FUTP) and 5-fluoro-deoxyuridine monophosphate (5-FdUMP). 5-FUTP inhibits RNA synthesis whereas 5-FdUMP inhibits the enzyme, thymidylate synthase, thus depleting replicating cells of thymidine nucleotide precursors during DNA synthesis. The rate-limiting step in the generation of these active species is the formation, via a series of enzymatically catalyzed reactions, of an intermediary metabolite, 5-fluorouridine monophosphate (5-FUMP).

In Escherichia coli and Saccharomyces cerevisiae 5-FU is converted to 5-FUMP by uracil phosphoribosyltransferase (UPRTase)11, 12 and it has been reported that adenovirus-mediated transduction of the E. coli upp gene that encodes UPRTase markedly sensitized tumor cells to 5-FU in vitro and in vivo.13

We had previously reported the efficacy of a new suicide gene derived from a fusion of the S. cerevisiae CDase (FCY1) and UPRTase genes (FUR1).14 This suicide gene, designated FCU1 (GenBank Accession Number AF312392), encodes a bifunctional chimeric protein that combines the enzymatic activities of FCY1 and FUR1 and efficiently catalyzes the direct conversion of 5-FC into the toxic metabolites 5-FU and 5-FUMP.14 The UPRTase activity of FCU1 is comparable to that encoded by the parental FUR1 gene and the CDase activity is increased at least 100-fold over native yeast CDase.14 Furthermore, this strategy generates a bystander effect, since 5-FU generated within the cells diffuses from the transduced cells into the nontransduced cells.14, 15 This is distinct from some other prodrug systems which rely on gap junctions for cell-to-cell transfer of the active drug.16, 17

Initially, replication-deficient adenovirus was used as the gene transfer vector for FCU1 and we showed that adenovirus-mediated transfer of FCU1 to tumor masses derived from human colon carcinoma cells in nude mice suppresses the growth of the tumors when the animals are exposed to systemic administration of 5-FC.14 Nevertheless, relatively large doses of 5-FC were required for this effect. In an effort to improve the efficacy of the FCU1/5-FC strategy for cancer gene therapy, the functionality of the FCU1 gene was also assessed in a propagation-deficient vaccinia viral context. Modified vaccinia virus Ankara (MVA) is a highly attenuated vaccinia virus, which has been used safely as a smallpox vaccine during the end stage of the smallpox eradication program.18, 19 MVA is largely nonpropagative in human and other mammalian cells20 but infection with MVA results in the rapid replication of viral DNA, and therefore rapid amplification of the FCU1 transgene. In addition there are now several vaccinia viral promoters with varying kinetics of activity and transcriptional strengths which are available to drive the transgene.21 Three vaccinia virus promoters were assessed for their ability to express the FCU1 fusion protein in the MVA context and the synthetic p11K7.5 promoter was found to drive production of the highest concentration of FCU1 in vitro. Intratumoral injections of MVA expressing FCU1 into human tumors implanted subcutaneously into mice, with concomitant systemic administration of 5-FC, led to substantial delays in tumor growth and lower doses of the prodrug 5-FC were required to obtain a tumor growth control in comparison with Ad-FCU1. In addition we present data showing in vivo that 5-FC was converted to 5-FU in tumors by MVA-FCU1 and that 5-FU production was localized to the tumors. Replication competent vaccinia virus have previously been explored as vectors for suicide gene transfer to tumor cells22 but to our knowledge, this is the first report to assess the therapeutic efficacy of MVA expressing a suicide gene for cancer gene therapy and these results demonstrate that MVA-FCU1/5-FC provides a solution for generating therapeutic concentrations of 5-FU locally, while avoiding excessive toxicity in normal tissues.

Materials and methods

Cell culture

Primary chicken embryo fibroblasts (CEF) were prepared from chicken embryo obtained from fertilized eggs (Charles River SPAFAS, Germany) previously incubated 11 days at 37 °C in a humid atmosphere. Chicken embryos were dissected and treated with a solution of trypsin 2.5% (w/v). CEF cells were maintained in MBE supplemented with 10% fetal bovine serum. Human colon cancer cell lines SW480, WiDr and LoVo, human breast cancer cell lines MCF7, T-47D and MDA-MB-231, human pancreas cancer cell line Capan-2, human hepatocellular carcinoma cell line Hep G2 and human lung cancer cell line A549 were obtained from the ATCC (Rockville, MD). All human cell lines were grown in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum.

Plasmid constructions, virus production and MVA transduction

The different MVA shuttle plasmids contain the early-late vaccinia promoters p7.5,23 pH5R24 or p11K7.5 (GenBank Accession Number CS054492, kindly provided by R Wittek, University of Lausanne) surrounded by the flanking sequences of the deletion III of the MVA (DelIII-R and DelIII-L sequences). The shuttle plasmids also contain the E. coli xanthine-guanine phosphoribosyltransferase gene (GPT gene) under the control of the pH5R vaccinia virus early-late promoter. Synthesis of xanthine-guanine phosphoribosyltransferase enables GPT+ recombinant MVA to form plaques in a selective medium containing mycophenolic acid, xanthine and hypoxanthine.25 A restriction fragment HindIII-KpnI containing FCU1 from plasmid pCI-neoFCU114 was inserted downstream of the different vaccinia promoters. The resulting shuttle plasmids are described in Figure 1. In these constructions the selection marker GPT is placed between two homologous sequences in the same orientation (DelIII-L sequence). Generation of MVA was performed by infection of CEF with a subclone of MVA-null named MVAN33 isolated in our laboratory, followed by transfection of the different shuttle plasmids by CaCl2 precipitation. Double recombination occurs between homologous regions (DelIII-R and DelIII-L) in the shuttle plasmid and the virus, resulting in the insertion of the gene cassette into the so-called deletion III of MVA. Recombinant MVA viruses are isolated by mycophenolic acid selection and multiple purification steps. The selection marker is easily eliminated by several passages without selection allowing the growth of gpt recombinant MVA obtained after intragenic homologous recombination between the two DelIII-L sequences flanking the GPT gene. The same methods were used to generate the MVAp11K7.5-GFP, which expresses GFP under the control of the p11K7.5 promoter. Virus structures were confirmed by multiple PCRs. Final recombinant MVA viruses were amplified in CEF and virus stocks were titrated on CEF by plaque assay. The control vector, MVA-null is defined as the MVAN33 without any inserted transgene.

Figure 1
figure 1

Generation of MVA expressing the FCU1 gene and evaluation of FCU1 protein expression. (a) MVA shuttle plasmids containing the FCU1 gene. The FCU1 gene is under the control of three different early-late promoters (p7.5, pH5R or p11K7.5). The GPT gene is driven by the vaccinia virus pH5R early-late promoter. DelIII-R and DelIII-L: flanking sequences surrounding deletion III of MVA. AmpR: β-lactamase coding region. Numbers refer to base pairs. (b) Western blot analysis. WiDr cells were infected with MVA at a MOI of 0.01 for 24 h. Protein was extracted and electrophoresed in a 10% SDS–PAGE gel. The expression level was analyzed by western blot using rabbit anti-FCU1 peptide antiserum. Lane 1 (left to right), mock-infected WiDr cells; Lane 2, WiDr cells infected with MVA-null; Lane 3, WiDr cells infected with MVAp7.5-FCU1; Lane 4, WiDr cells infected with MVApH5R-FCU1; Lane 5, WiDr cells infected with MVAp11K7.5-FCU1. Molecular weight standards are shown in kDa on the left. The presence of FCU1 (Mr 42 000) is indicated. (c) Relative amount of the immunoreactive FCU1 protein expression. The relative density at each point was calculated by dividing that value by the density of WiDr cells infected with MVAp11K7.5-FCU1 (100%).

The Ad-FCU1 vector is an E1/E3-deleted vector derived from the human adenovirus serotype 5 with an expression cassette in the E1 region containing the FCU1 gene driven by the CMV immediate-early enhancer/promoter.14 The control vector Ad-null is similar except that it contains no transgene.14

To determine the in vitro transduction efficiency of MVA, cells were infected with MVAp11K7.5-GFP and 24 h later, single-cell suspensions were analyzed by flow cytometry using a FACScan instrument (Becton Dickinson, Le Pont de Claix, France).

Western blot analysis

WiDr tumor cells were infected with each MVA vector at a multiplicity of infection (MOI) of 0.01 and incubated for 24 h. Cell lysate proteins (30 μg) (determined using a Bio-Rad protein assay) were run on a 10% SDS–PAGE gel under reducing conditions and transferred onto a nitrocellulose membrane. The membrane was incubated with rabbit anti-FCU1 peptide (codons 112–126) antiserum at a 1:500 dilution, washed and incubated with secondary antibody coupled horseradish peroxidase (Amersham, Les Ulis, France). Signal detection was done by enhanced chemiluminescence (Amersham). The density of the bands was quantified using a densitometer (Imaging Densitometer, Bio-Rad, Marnes-la-Coquette, France).

Enzymatic assays

CDase activity, UPRTase activity and CDase-UPRTase activities in WiDr cells were determined using 5-FC (Toronto Research Chemicals Inc., North York, Canada) and 5-FU (Sigma, Saint Quentin Fallavia, France) as substrates. WiDr human tumor cells (5 × 106 cells) were infected with each MVA vector at a MOI of 0.01. Twenty four hours later, enzymatic assays were determined as previously described.14 5-FC, 5-FU and 5-FUMP were separated isocratically using HPLC (Hewlett Packard HP 1100 liquid chromatograph with UV detection at 260 nm and 280 nm). A Supelco supelcosil LC-18-S (5 μm packing; 4.6 × 250 mm) column and a guard cartridge (10 × 3 mm; Varian, Les Ulis, France) were used with a flow rate of 1 ml min−1. The mobile phase was 20 mM KH2PO4, 5 mM tetrabutylammoniumsulfate, 5% methanol adjusted to pH5 with potassium hydroxyde.

In vitro cell sensitivity to 5-FC

Human tumor cells were transduced by the respective recombinant MVA at a MOI of 0.01. A total of 2 × 105 transduced cells per well were incubated in six-well culture dishes in 2 ml of medium containing various concentrations of 5-FC. Cells were cultured at 37 °C for 6 days, and the viable cells were counted by trypan blue exclusion. Results are expressed as the ratio between the number of viable cells in plates containing the drugs versus the number of viable cells in the corresponding drug-free controls.

Analysis of the in vitro bystander effect

To determine whether transduced cells exposed to 5-FC mediated the production of 5-FU in the media to affect neighboring cells, WiDr cells, infected with the different vectors at a MOI of 0.01, were incubated in 12-well culture dishes (106 cells per well). After 24 h, 1 mM 5-FC was added to the cultures. After 6, 24 and 48 h, the concentrations of 5-FC and 5-FU in the media were measured using HPLC. Media (50 μl) were quenched with 1 ml of ethyl acetate/2-propanol/0.5 M acetic acid solution (84:15:1). The samples were vortexed and centrifuged. The organic supernatant was evaporated to dryness under a stream of nitrogen at 60 °C and reconstituted in 50 μl of water and analyzed by HPLC as described above using a mobile phase of 50 mM phosphoric acid adjusted to pH 2.1 with ammonium hydroxide.

Animal experiments

Female Naval Medical Research Institute (NMRI) nude mice were obtained from Charles River Laboratories (Saint Aubin-les-Elbeuf, France). Animals used in the studies were uniform in age (6 weeks) and body weights ranged from 23–26 g. NMRI nude mice were injected subcutaneously (s.c.) into the flank with 5 × 106 LoVo cells. When tumors reached a diameter of 60–80 mm3 (this size of tumor contains approximately 5 × 107 to 1 × 108 cells), the mice were randomized in a blinded manner and treated with the indicated vectors for the in vivo experiments.

Determination of tumor and blood concentration of 5-FC and converted 5-FU

Using human tumor LoVo cells implanted s.c. in nude mice, control vehicle (100 μl PBS) or 5.106 PFU of MVAp11K7.5-FCU1 (resuspended in 100 μl PBS) were intratumorally injected on day 0. The prodrug 5-FC was given by oral gavage at 100 mg kg−1 per day (0.5 ml 5-FC 0.5% in water) twice daily for 2 weeks. Animals were killed 1 h post-5-FC gavage on days 1, 3, 6, 8, 10 and 14, and blood samples and tumors were collected. Plasma was separated by centrifugation from blood collected via the retro-orbital sinus in heparinized tubes. Tumors were homogenized using a Polytron homogenator. Tumor or plasma samples were quenched with 1 ml of ethyl acetate/2-propanol/0.5 M acetic acid solution (84:15:1). The organic supernatant was reconstituted in 50 μl of water and analyzed by HPLC as described above using 50 mM phosphoric acid adjusted to pH 2.1 as mobile phase.

Antitumor activity of MVA-FCU1 versus Ad-FCU1

Human tumors were treated with the MVAp11K7.5-FCU1 or Ad-FCU1 vectors followed by per os 5-FC administration at different concentration. Animals were treated intratumorally with the indicated vectors at dose of 5.108 IU (Ad-null, Ad-FCU1) or 5.106 PFU (MVA-null, MVAp11K7.5-FCU1). The vectors or the control vehicle (PBS) were directly injected into the tumor at days 11, 13 and 15 after tumor implantation. From day 11 on, 5-FC was given by oral gavage at 100 mg kg−1 per day (0.5 ml 5-FC 0.5% in water), 250 mg kg−1 per day (0.6 ml 5-FC 1%) or 500 mg kg−1 per day (0.6 ml 5-FC 2%) twice a day for 2 weeks. Tumor volumes were calculated in mm3 using the formula (π/6)(length × width2).

Statistical analysis

Statistical analyses were performed using the nonparametric Mann–Whitney U-test and STATISTICA 7.1 software (StatSoft Inc.). A P<0.05 was considered to be statistically significant.


In vitro infection of human tumor cells by MVA

Nine different human cancer cell lines, representing five different human tumor types, were tested for their susceptibility to infection in vitro with an MVA expressing GFP marker protein under the control of the p11K7.5 promoter (MVA p11K7.5-GFP). The cell lines were infected with MVA p11K7.5-GFP at an MOI of 0.01 and 0.1 PFU and the transduction efficiency was assessed by flow cytometry 24 h post infection. The results are summarized in Table 1. The 50% transduction efficiency of this vector in human cancer cell lines varied, but all these cell lines showed 25–75% transduction efficiency at an MOI of 0.1. The observed infection frequency which was slightly higher than expected can be explained by the previous observation26 that one plaque forming unit of MVA is often an aggregate of several viral particles. The transduction efficiencies of human tumor cells by MVA at an MOI of 0.1 were comparable to those obtained with an E1/E3 deleted, serotype 5 adenovirus vector at an MOI of 1.14

Table 1 Susceptibility of human tumor cell lines to MVA infection

Expression of FCU1 fusion protein in WiDr cells infected with MVA-FCU1

Western blot analysis of the FCU1 protein, using FCU1 specific rabbit antiserum, was done to compare the levels of FCU1 produced by the three vaccinia virus promoters. An FCU1 specific signal was observed in cellular extracts isolated from WiDr cells infected by MVA-FCU1 (Figure 1b, Lanes 3–5). The apparent molecular size of FCU1 protein was 42 kDa, in agreement with the expected size of an S. cerevisiae FCY1:FUR1 fusion protein. The relative density of the FCU1 protein level indicated a 2.7-fold increase in WiDr cells after infection with MVApH5R-FCU1 (Figure 1c, Lane 4), compared to MVAp7.5-FCU1-infected cells (Lane 3). The level of FCU1 protein was elevated 6.7 and 2.5-fold in cells infected with MVAp11K7.5-FCU1 (Lane 5) compared to MVAp7.5-FCU1-infected cells (Lane 3) and MVApH5R-FCU1-infected cells (Lane 4), respectively. These results show that the p11K7.5 promoter led to the highest level of FCU1 expression.

Functionality of the FCU1 fusion gene

In order to confirm the strongest promoter driving expression of FCU1 protein and to determine whether FCU1 is functional when produced under the direction of each of the three promoters, WiDr human tumor cells were infected with each vector at an MOI of 0.01 and the specific CDase, UPRTase and CDase/UPRTase enzymatic activities were then determined 24 h later, as described in Materials and methods. As shown in Table 2, elevated CDase activity was found in cells infected with MVAp11K7.5-FCU1, while no CDase activity was detectable in mock-infected cells or in MVA-null-infected cells. No endogenous UPRTase activity was detected in uninfected WiDr cells (Table 2). Consistent with the western blot analysis (Figure 1b), higher UPRTase activity was recovered using the p11K7.5 construct. Analysis of the direct conversion of 5-FC to 5-FUMP, indicative of the existence of a combined CDase-UPRTase activity, confirmed that FCU1 does encode a bifunctional enzyme (Table 2). Moreover, this combined enzymatic activity was found to be 5.8- and 2.3-fold higher in cells infected with MVAp11K7.5-FCU1 than in cells infected with MVAp7.5-FCU1 and MVApH5R-FCU1, respectively. Taken together, these in vitro enzymatic assays confirm that the p11K7.5 promoter is the most active promoter driving the FCU1 bifunctional enzyme.

Table 2 Specific CDase, UPRTase and CDase/UPRTase activities in WiDr cell line

Enhanced sensitivity of human tumor cells to 5-FC after infection with MVA-FCU1 via the 5-FC-mediated bystander effect

To test whether MVA-mediated FCU1 gene expression in the cells conferred sensitivity to 5-FC, nine different human tumor cell lines were infected with MVA-null, MVAp7.5-FCU1, MVApH5R-FCU1 or MVAp11K7.5-FCU1 at a MOI of 0.01 and then exposed to 5-FC at various concentrations for six subsequent days. As shown in Figure 2, all cell lines infected with the MVA-FCU1 became sensitive to 5-FC. Consistent with the results from the enzymatic analyses (Table 2), infection with MVAp11K7.5-FCU1 confers to the cells the highest sensitivity to 5-FC when compared with the MVAp7.5-FCU1- or MVApH5R-FCU1-infected cells. For example, infection of MDA-MB-231 and Capan-2 cells with MVAp11K7.5-FCU1 followed by 5-FC treatment resulted in a 10- and 100-fold shift in IC50 compared with MVApH5R-FCU1 and MVAp7.5-FCU1 infections, respectively (Figure 2). As expected, infection with MVA-null did not modify the sensitivity of the cells to 5-FC. Given the relatively weak efficiency of infection at 0.01 PFU/cell (Table 1), these results suggest that a bystander effect occurs under these treatment conditions. To investigate this 5-FC-mediated bystander effect, analysis of the cell culture supernatant by HPLC revealed a progressive increase in the amount of 5-FU in the extracellular milieu of WiDr cells transduced with MVA-FCU1 at an MOI of 0.01 and incubated with 1 mM 5-FC (Figure 3a). This result indicates that when only 4% of WiDr cells were transduced (Table 1), 5-FC was efficiently converted by MVA-FCU1 infected cells and that 5-FU was freely diffusible across cell membranes. As expected, MVAp11K7.5-FCU1-infected cells were much more effective at transforming 5-FC into 5-FU, with a mean conversion ratio of 72% after 2 days of incubation.

Figure 2
figure 2

In vitro sensitivities of MVA-transduced human tumor cells to 5-FC. Cells transduced at a MOI of 0.01 with the indicated vectors were exposed to various concentrations of 5-FC and cell survival was measured 6 days later as described in Materials and methods. Results are expressed as the percentage of surviving cells in presence and absence of the drug. Values are represented as mean±s.d. of four individual determinations.

Figure 3
figure 3

5-FU generated in vitro and in vivo by MVA-FCU1 prodrug system. (a) Conversion of 5-FC to 5-FU and release of 5-FU in the cell culture supernatant. WiDr cells were infected with the indicated vectors at a MOI of 0.01 and then incubated with 1 mM 5-FC. The relative concentrations of 5-FC and 5-FU in the media were measured by HPLC, 6, 24 and 48 h later. The data are expressed as the percentage of 5-FU in the media relative to the total amount of 5-FC+5-FU. Each data point represents the mean±s.d. of triplicate determinations. (b) Intratumoral concentration of 5-FU in human tumors. Tumors were established by s.c. implantation of 5 × 106 LoVo cells into nude mice. When tumors had reached 60–80 mm3, the animals were intratumorally inoculated with MVAp11K7.5-FCU1 (5.106 PFU) and treated for 2 weeks with per os administrations of 5-FC. At 1, 3, 6, 8, 10 and 14 days after viral injection, tumor tissues were collected 1 h after a single dose of 5-FC (100 mg kg−1) and concentration of 5-FU in the tumor was measured by HPLC as described in Materials and methods. Three mice were used for each group.

In vivo 5-FC/5-FU conversion in human tumors by MVAp11K7.5-FCU1

LoVo tumors were established in nude mice and the concentrations of 5-FU in tumor tissue at different days post infection were determined after a single intratumoral injection of MVA p11K7.5-FCU1 (Figure 3b). The highest levels of 5-FU were detected in tumor tissues 6 days post infection. At 14 days post infection, 5-FU was still detected in the tumor tissues. In mice treated with 5-FC only, the concentration of 5-FU in tumors was undetectable (Figure 3b) indicating that no endogenous conversion of 5-FC to 5-FU occurs in the tumors. The concentrations of 5-FU in serum in mice receiving 5-FC, with and without MVA-FCU1 in tumors, were also determined. No 5-FU was detected in serum of the group treated with 5-FC alone. Similarly, 5-FU could not be detected in serum of mice treated with MVA-FCU1/5-FC (data not shown) suggesting that 5-FU does not diffuse from the tumor tissues. In contrast, in the mice treated with a single injection of 5-FU at the maximum tolerated dose (10 mg kg−1 administered intraperitoneally), the blood concentration of 5-FU reached a maximum of 5 μg ml−1 5 min after injection of 5-FU and the highest level of 5-FU in tumor tissues was detected 10 min post injection with a concentration of 0.6–0.7 ng mg−1 of tumor. When a single intraperitoneal injection of 5-FU at 200 mg kg−1 (20-fold the maximum tolerated dose) was given, the 5-FU level reached a maximum of 90 μg ml−1 of serum and 10 ng mg−1 of tumor 10 min post-5-FU administration. One hour after 5-FU injection, 5-FU became undetectable in tumors. The intratumoral 5-FU level resulting from MVA-FCU1 was equivalent to the maximum level attained with systematically administered 5-FU at 20-fold the maximum tolerated dose.

In vivo efficacy of MVA-mediated transfer of FCU1

To examine the therapeutic effects of MVA-FCU1 in vivo, nude mice bearing established LoVo colon tumors were injected intratumorally with MVA-null, MVAp11K7.5-FCU1 or mock (vehicle alone) with a concomitant per os administration of either 5-FC (at 1000, 500 or 200 mg kg−1 per day) or water for 14 days. As shown in Figure 4a, administration of MVAp11K7.5-FCU1/5-FC resulted in a statistically significant suppression of tumor growth whatever the dose of 5-FC, whereas no modification in tumor growth was observed in the control groups. At 5 weeks after delivery of MVAp11K7.5-FCU1/5-FC, the LoVo tumors were 75% smaller, respectively, compared to the controls (all controls averaged; Figure 4a). In parallel, control experiments were also performed to determine the in vivo antitumor effect of 5-FU in nontransduced LoVo tumor cells. Despite the administration of doses of 5-FU that were at the maximum tolerated concentrations (i.p. injection of 10 mg of 5-FU per kg twice daily during 2 weeks), no statistically significant inhibition of tumor growth was observed (data not shown). These results indicate that the significant antitumor effect was due to the local production of high concentrations of 5-FC derivates via the FCU1 gene product expressed from the p11K7.5 promoter within the MVA vector. MVAp11K7.5-FCU1/5-FC was also compared to replication-deficient adenovirus expressing FCU1. Using adenovirus, administration of FCU1 with a concomitant per os administration of 5-FC at 1000 mg kg−1 per day resulted in a statistically significant suppression of tumor growth compared to the control groups, whereas no modification in tumor growth was observed with administration of 5-FC at 500 and 200 mg kg−1 per day (Figure 4b). The data in Figure 4 indicate that the transfer of the FCU1 gene to tumors is much more efficient using the MVA system than the adenovirus system previously described.14 MVAp11K7.5-FCU1 allows the same therapeutic effect at one-fifth of the daily dose of 5-FC in comparison to Ad-FCU1.

Figure 4
figure 4

Antitumor effect of the MVAp11K7.5-FCU1/5-FC (a) and Ad-FCU1/5-FC (b) therapies on established human LoVo colon carcinoma (n=14 per group). 5 × 106 LoVo cells implanted s.c. into nude mice were injected at days 11, 13 and 15 with the vehicle alone or with either 5 × 106 PFU of MVA (a) or 5 × 108 IU of Adenovirus (b). The animals were then treated twice daily with per os administrations of 5-FC for two weeks at 1000, 500 or 200 mg kg−1 per day. In MVA groups, differences in tumor sizes between the group treated with MVAp11K7.5-FCU1+5-FC and the other groups was shown to be statistically significant (P<0.01), whatever the dose of 5-FC. In groups treated with adenovirus, differences in tumor sizes between the group treated with Ad-FCU1+5-FC and the other groups was shown to be statistically significant (P<0.01) only with 5-FC at 1000 mg kg−1 per day.


MVA, a member of the poxvirus family, was derived from the vaccinia virus strain Ankara by over 570 serial passages in CEF cells.18 The resulting MVA strain was unable to productively grow in cell cultures of human origin.20 MVA vaccine was used during the end stage of the smallpox eradication program in Germany and Turkey in more than 120 000 humans without documentation of any of the complications associated with other vaccinia viruses.19 Furthermore recent preclinical studies have shown that MVA is avirulent upon inoculation of various animals.27, 28 Because of its high degree of attenuation, MVA has appeared as an attractive alternative to standard vaccinia strains for the development of viral vectors to be used in vaccination or immunotherapy.29 Like other attenuated poxvirus vectors used as recombinant vaccines (ALVAC and NYVAC), MVA will not propagate in most mammalian cells. However, unlike ALVAC30 and NYVAC,31 MVA infection results in the replication of viral DNA.20 This, in turn, results in the high level production of recombinant proteins.20 MVA also has many other advantages as a live viral vaccine vector: (1) large foreign gene capacity; (2) long-term stability in frozen or lyophilized state; (3) replication exclusively in the cytoplasm eliminating any risk of integration.

In this report, we assessed the ability of MVA to transfer the FCU1 genetic sequence to tumor cells in vitro and in vivo, and compared it to the previously reported Ad-FCU1.14 We first tested altogether nine human cancer cell lines representing five different tissues for transduction efficiency. Conditions have been established for each vector where some preclinical efficacy has been demonstrated. Our findings demonstrate that an MVA vector can efficiently transduce human tumor cells in vitro, allowing 25–75% gene transfer efficiency. Similar gene transfer efficiencies were obtained with an E1/E3 deleted adenovirus vector.14 It should be emphasized that it would be nonproductive to compare poxvirus- and adenovirus vectors directly because their titering methods are different (technique of titration, target cell line, transduction volume, time of incubation) and therefore the real concentrations of biologically active virus may not be equal.

We have described a comparison of the efficacy of MVA vectors expressing FCU1 under the control of three different promoters. In vitro experiments unambiguously showed that the p11K7.5 promoter constitutively drives the highest level of FCU1 as assessed by protein levels and enzymatic assays. Similarly, pH5R promoter strength was intermediate between that of p11K7.5 and p7.5. This synthetic 110-bp p11K7.5 promoter, which consists in a fusion between the p11K late promoter and the p7.5 early-late promoter, has been successfully used in our laboratory for the expression of GFP and other foreign proteins by MVA and vaccinia virus.

Our results indicated that MVA-FCU1+5-FC generates a very strong killing effect. The amount of 5-FU produced by recombinant viruses in the media essentially eradicated >95% of tumor cells in vitro. We also confirmed that 5-FU diffuses in and out of cells and does not require cell-to-cell contact for bystander activity. This bystander effect should enhance the antitumor efficacy of the suicide gene therapy approach by eliminating uninfected tumor cells.

Early work with bacterial CDase in combination with 5-FC was promising; however, later studies revealed that due to poor catalytic efficiency this system is likely to have limited clinical applicability.32, 33 Recent studies34, 35 have demonstrated that S. cerevisiae CDase deaminates 5-FC more efficiently than E. coli CDase, and that human and murine cancer cells transduced with the yeast FCY1 gene were significantly more sensitive in vitro and in vivo to 5-FC than tumor cells transduced with the bacterial codA gene. This was attributed to the more favorable Km and Vmax of yeast CDase for 5-FC, which was 22-fold lower and 6-fold higher than those of bacterial CDase, respectively.34, 35 In our previous report,14 we also noted an unexpected increase in the deaminase activity of the fusion protein FCU1 which was over 100-fold higher than the native CDase. Similarly, a study recently reported an increase in the catalytic activity of the S. cerevisiae CDase when expressed as a fusion protein with Haemophilus influenzae UPRTase.36 This increased CDase activity is a consequence of the enhanced thermal stability of the fusion protein compared to native CDase allowing it to function for a greater period of time.36

In this study, we have evaluated the conversion of 5-FC to 5-FU in tumors after a single intratumoral injection of MVA p11K7.5-FCU1 in combination with a per os regimen of 5-FC. After a single dose of 100 mg kg−1 5-FC, the converted 5-FU in tumors reached the highest level (5.5 ng mg−1 of tumor) 3–8 days post infection. At 14 days post infection, 5-FU was still detected in the tumor tissues indicating the persistence of gene expression under the control of the p11K7.5 promoter and/or the high stability of FCU1 protein in the tumor mass. In our study, the intratumoral 5-FU level resulting from MVA-FCU1 was equivalent to the maximum level attained with systematically administered 5-FU at 20-fold the maximum tolerated dose. In addition we showed that after the direct production of 5-FU at the tumor site using MVA-FCU1, no 5-FU could be detected in serum resulting in tumor cell death without systemic toxicity. This high intratumoral concentration of 5-FU and its undetectable level in serum could be attributable to the retention (‘trapping’) of 5-FU in the tumor. Such a ‘trapping’ of 5-FU within tumor cells has been noted in a number of previous studies of 5-FU therapy37, 38 and has been rationalized on the basis that the 5-FU half-life is significantly longer in tumors than in blood.37, 39

We compared the efficiency of FCU1 gene transfer by MVA vector and nonreplicative-adenovirus, both of which are unable to proliferate in human tumors. Our finding demonstrates that both types of vectors transduce proliferating tumor cells in vivo with comparable efficiency. After a single intratumoral injection of 5.108 IU of adenovirus expressing GFP14 or 5.106 PFU of MVA-GFP, only a few percent (1–5%) of tumor cells are GFP+ and these cells were localized along the needle track (data not shown). At these doses of MVA or adenovirus, when high concentration of 5-FC were used (1000 mg kg−1 per day), both MVA-FCU1 and Ad-FCU1 vectors suppressed human colon tumor growth. These Ad-FCU1 results were consistent with our previously reported experiments.14 At relatively lower 5-FC doses (500 and 200 mg kg−1 per day) MVA-FCU1 could also suppress tumor growth, whereas no modification of tumor growth was observed using Ad-FCU1. These results suggest an increased transgene expression in MVA context in comparison to Ad-FCU1 and further investigations are needed to determine level and persistence of gene expression after direct injection of MVA in solid human tumors. 5-FC is used clinically for antifungal therapy and in this context is typically administered daily for 6 weeks. Nontoxic systemic serum levels of 50–100 μg ml−1 are achievable with oral administration of 100–200 mg kg−1 per day in four divided doses.40 Our findings, of effective cell killing with tolerable equivalent human doses of 5-FC suggest significant advantages in using MVA-mediated transfer of FCU1 for molecular therapy since in previous investigations involving 5-FC/CDase a dose of at least 500 mg kg−1 5-FC was required for therapeutic effects in tumor-bearing animals.2, 3, 4, 5, 6, 7, 34, 35, 36, 41 Attempts to reduce the dose of 5-FC to less than 500 mg kg−1 resulted in no effect on tumor growth.42

In summary, this study represents, to our knowledge, the first study to use MVA vector as gene transfer vehicle for suicide gene therapy. It demonstrates the efficacy of intratumoral 5-FU produced via intratumoral injection of MVA-FCU1 and a regimen of systemic 5-FC at well-tolerated doses. The results of the present study also demonstrate the superior efficacy of MVA in comparison to the traditionally used adenovirus in delivery of a suicide gene in human tumor cells.

We believe that MVA-FCU1 is highly promising for clinical development for the local control of cancers, such as colorectal liver metastases, hepatocarcinoma, glioma or prostate cancer and could also be combined with other therapies, such as radiotherapy, thus exploiting 5-FU radiosensitizing potential.41