Generation of an Adenovirus-Parvovirus Chimera with Enhanced Oncolytic Potential

Abstract

In this study, our goal was to generate a chimeric adenovirus-parvovirus (Ad-PV) vector that combines the high-titer and efficient gene transfer of adenovirus with the anticancer potential of rodent parvovirus. To this end, the entire oncolytic PV genome was inserted into a replication-defective E1- and E3-deleted Ad5 vector genome. As we found that parvoviral NS expression inhibited Ad-PV chimera production, we engineered the parvoviral P4 early promoter, which governs NS expression, by inserting into its sequence tetracycline operator elements. As a result of these modifications, P4-driven expression was blocked in the packaging T-REx-293 cells, which constitutively express the tetracycline repressor, allowing high-yield chimera production. The chimera effectively delivered the PV genome into cancer cells, from which fully infectious replication-competent parvovirus particles were generated. Remarkably, the Ad-PV chimera exerted stronger cytotoxic activities against various cancer cell lines, compared with the PV and Ad parental viruses, while being still innocuous to a panel of tested healthy primary human cells. This Ad-PV chimera represents a novel versatile anticancer agent which can be subjected to further genetic manipulations in order to reinforce its enhanced oncolytic capacity through arming with transgenes or retargeting into tumor cells.

INTRODUCTION

Adenoviruses (Ads) are nonenveloped, icosahedral viruses with a 30- to 38-kbp DNA genome. As of today, over 50 different human serotypes have been described, with most of them infecting the respiratory or gastrointestinal tracts and the eye (33). Ad infections are very common and generally not associated with any serious pathogenicity. Ads represent the most popular gene therapy vectors and were used in about 25% of approved phase I to III clinical trials for vaccine and therapeutic gene transfer during the last 2 decades (9). This is largely due to the ability of these vectors to efficiently deliver transgenes to the nucleus of a wide range of different cell types and mediate high levels of expression of the transgene of interest (33). Ads transduce both proliferating and resting/differentiated cells and remain episomal, which minimizes the risk of insertional mutagenesis (33). Furthermore, Ads are very versatile tools with remarkable DNA packaging capacity, offering a plethora of possibilities for genetic manipulations. The Ad genome can be modified in different ways in order to restrict transgene expression to specific tumor cells (22). Furthermore, it is possible to redirect Ad entry and render it more specific for cancer cells, through the use of molecular adaptors or genetic engineering of the Ad capsid (11, 12, 29). Importantly, Ads can be produced and purified at high titers and quality under good manufacturing practice (GMP) conditions (29).

Autonomous rodent parvoviruses (PVs) are small icosahedral, nonenveloped single-stranded DNA viruses. Their genome is about 5.1 kb long and contains two promoters, P4 and P38, that control the expression of the nonstructural (NS1 and NS2) and structural (VP1 and VP2) proteins, respectively (31). Several PVs, including the minute virus of mice (MVM) and the rat H-1PV, have also oncolytic and oncosuppressive properties, as demonstrated in various cellular and animal cancer models (32). Additionally, PVs are nonpathogenic and show low prevalence in humans, favoring their use as therapeutics (5). H-1PV is currently being evaluated in a phase I and IIa clinical trial for the treatment of patients with recurrent glioblastoma multiforme (32). The antineoplastic property of these PVs is due, at least in part, to preferential viral DNA replication and gene expression in malignant cells. This is caused by the virus dependence on the cell cycle S phase for its replication and, specifically, on cellular factors such as E2F, CREB, ATF, and cyclin A, which are overexpressed and/or activated in cancer cells (32). In addition, PVs may counteract the ability of malignant cells to mount an efficient antiviral innate immune response (13). It has been shown that PVs have the ability to induce cell cycle arrest (16) and different death pathways, including necrosis (27), apoptosis (16, 26), and lysosome-dependent cell death (8), in cancer cells. Although preclinical studies highlight the anticancer potential of PVs (32), this property must be further reinforced in view of the clinical application of these agents. One major hindrance lies in the fact that PVs bind and enter into a variety of healthy human cells, resulting in the sequestration of a large portion of the administered viral dose away from the tumor target cells. Retargeting PV entry to tumor cells would thus increase the efficacy of PV-based treatments and provide additional protection against eventual side effects on healthy tissues. It should be also noted that the difficulty of large-scale production of PVs, as required for clinical applications, remains a major limitation.

We envisioned that it would be of great benefit to generate an Ad-PV chimera combining the unique properties of both vectors. Similarly to any other recombinant adenovirus vector, the chimera should be produced at high titers, solving the problem of the difficulty related to the production of parvoviruses. Furthermore, we recently reported that expression of adenovirus genomic elements boosted the production of recombinant parvovirus in different cell lines (10). Therefore, we speculated that the Ad-PV chimera may enhance PV replication in cancer cells through the concomitant expression of Ad helper functions. In addition, the principle could be extended to include (i) the specific delivery of PV genomes to cancer cells by means of retargeted Ads and (ii) Ad genome arming with therapeutic transgenes that potentiate the PV-killing activity. On the other hand, the PV component of the chimera would confer antineoplastic activity by (i) expressing the cytotoxic NS1 protein under the control of its natural PV promoter and (ii) amplifying the antitumor effect through PV excision from the vector and autonomous replication and spreading through the tumor.

In the present report, we describe the first generation of Ad-PV chimeras and their very promising anticancer properties.

MATERIALS AND METHODS

Cells.

HEK293 (transformed human embryonic kidney), NB324K (transformed newborn human kidney), and A549 (lung carcinoma) cell lines were obtained from ATCC (LGC Standards GmBH, Wesel, Germany) and T-REx-293 from Invitrogen (Darmstadt, Germany). HeLa and SiHa cervical carcinoma cell lines, positive for human papillomavirus (HPV) types 18 and 16, respectively, were a gift from Angel Alonso (DKFZ, Heidelberg, Germany). The ME-180 cell line, positive for HPV 68, was obtained from Elisabeth Schwarz (DKFZ, Heidelberg, Germany). The colon cancer HCT-15 and HCC-2998 and the melanoma Lox-IMVI cell lines were from the National Cancer Institute (NCI) (Bethesda, MD). Low-passage melanoma PMelL cells (purified from skin metastases) have been previously described (23). The human primary oral fibroblasts and foreskin fibroblasts were a gift from Massimo Tommasino (IARC, Lyon, France). Human primary adult melanocytes, lightly pigmented (HEMa-LP), were from Invitrogen (Carlsbad, CA). Human astrocytes were obtained from ScienCell Research Laboratories (San Diego, CA).

HEK293, HeLa, SiHa, A549, and human primary oral fibroblast cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Gibco, Invitrogen, Karlsruhe, Germany). T-REx-293 cells were grown in DMEM containing 10% tetracycline-free certified FBS (PAA, Cölbe, Germany). ME-180 cells were grown in McCoy's 5a modified medium supplemented with 10% FBS. HCT-15, HCC-2998, and Lox-IMVI cells were grown in Roswell Park Memorial Institute (RPMI) 1640 medium complemented with 10% FBS. PMelL cells were cultivated in RPMI 1640 (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum, 10 mM HEPES (Lonza, Basel, Switzerland), and 250 ng/ml amphotericin B and 100 μg/ml gentamycin (both from Invitrogen). Human primary foreskin fibroblasts and NB324K cells were grown in minimum essential medium (MEM) supplemented with 10% and 5% FBS, respectively. Primary human adult melanocytes were grown in medium 254 supplemented with HMGS (Invitrogen, Carlsbad, CA). Human astrocytes were cultivated in astrocyte medium (ScienCell Research Laboratories, San Diego, CA). All media, except the media used to cultivate pMelL and normal melanocytes and astrocytes, contained 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM l-glutamine. Cells were grown at 37°C in 5% CO₂ and 92% humidity.

Viruses.

hH-1 (17) and hH-1-TO viruses were produced in T-REx-293 cells. The cells were cultivated in 10-cm-diameter dishes in tetracycline-free medium and transiently transfected at 12.5% confluence with 10 μg/dish of either phH-1 or phH-1-TO viral constructs. At 4 h, 3 days, and 6 days posttransfection, doxycycline (DOX) (1 μg/ml) was added to the medium. At day 7, cells were harvested within their medium and lysed by 3 freeze-and-thaw cycles and cellular debris was removed by centrifugation. Produced viruses were further amplified by infecting NB324K cells and purified through iodixanol gradient centrifugation.

Ad-hH-1-TO and an Ad control (Ad5ΔE1ΔE3) were produced through 3 rounds of production in T-REx-293 cells cultivated in medium containing 5% tetracycline-free FBS. In the first round, cells in 12-well plates were transfected at 75% confluence with 1 μg/well of pAd-hH-1-TO or pAd5ΔE1ΔE3 plasmids predigested with PacI. After transfection, cells were induced with 1 μg/ml of DOX or left uninduced. At 5 days posttransfection, to ensure optimal growth conditions, half of the culture medium was replaced. At 7 days posttransfection, cells were harvested into their medium and lysed with 3 freeze-and-thaw cycles and cellular debris was removed by centrifugation. In the second round, 25% of the crude viral extract produced in the first round was used to infect T-REx-293 cells grown in 75-cm² flasks. At 5 days postinfection (p.i.), half of fresh medium was added to maintain optimal growth conditions. At 7 days postinfection, cells were harvested and lysed as previously described. The third round was comparable to the second round but was carried out using 175-cm² flasks. The final viral batches were purified twice through cesium chloride (CsCl) gradient ultracentrifugation.

DNA cloning.

The phH-1-TO parvovirus plasmid was constructed by inserting two tetracycline operator 2 (TetO₂) repressor elements into the P4 promoter of phH-1 (17), through PCR cloning. In a first step, two parallel PCRs were carried out, using phH-1 as a template, with the following primers: primer pair PCR1-For (5′-AAACTCGAGGCGGTTCAGGGAGTTTAAACC-3′) and PCR1-Rev (5′-AACTGACTTCTCTCTATCACTGATAGGGAGATCTCTATCACTGATAGGGAAGTAGTTGCTTATATACTTTAAACC-3′) and primer pair PCR2-For (5′-AGCAACTACTTCCCTATCAGTGATAGAGATCTCCCTATCAGTGATAGAGAGAAGTCAGTTACTTATCTTTTCTTTC-3′) and PCR2-Rev (5′-AAAAAGCTTCCATCCGATATCTTTTCCATTCAG-3′). In a second step, a third PCR was carried out using a stochiometric mix of the 2 previous purified PCR products as the template with PCR1-For and PCR2-Rev as primers. The DNA product obtained was digested with PmeI and EcoRV and used to replace the corresponding fragment in phH-1.

pShuttle-cytomegalovirus-free (pShuttle-CMV-free) was constructed as follows: pShuttle-CMV (Qbiogen, MP Biomedicals, Heidelberg, Germany) was digested with BglII, dephosphorylated with calf intestine phosphatase, and subjected to homologous recombination, in Escherichia coli BJ5183, with annealed oligonucleotides 5′-GTTCATAGCCCATATATGGAGTTCAGATCTGGTACCG-3′ and 5′-CGGTACCAGATCTGAACTCCATATATGGGCTATGAAC-3′. pShuttle-hH-1 was generated in 3 steps. (i) In pShuttle-CMV-free, the unique EcoRI site was changed into a SwaI site, through insertion of the annealed oligonucleotides 5′p-AATTATTTA-3′ and 5′p-AATTTAAAT-3′ at the EcoRI location. (ii) The 5′ region of hH-1 was amplified from phH-1 by PCR using the primers 5′-AAGGAAAAAAGTCGACTTTTGTGATGCTCGTCA-3′ and 5′-AGGAAAAAAGATATCTTTTCCATTCAGTTGA-3′. The PCR product was digested by SalI and EcoRV (633 bases) and ligated into the previously modified pShuttle-CMV-free vector, predigested with the same enzymes, resulting in pShuttle-5′hH-1. (iii) The remaining 3′ end of the hH-1 genome (4,730 bases) was excised from the phH-1 plasmid using the EcoRV-NdeI enzymes, subjected to blunting using a Klenow fragment, and ligated in EcoRV-digested pShuttle-5′hH-1, generating pShuttle-hH-1.

pShuttle-hH-1-STOP was cloned by inserting 3XSTOP codons into the unique EcoRV site of the pShuttle-hH-1, located at the beginning of the parvovirus NS coding sequence. The 3XSTOP DNA duplex was generated by the self-annealing of the oligonucleotide 5′-TAATAGTGAGAATTCTCACTATTA-3′.

pShuttle-hH-1-TO was obtained by replacing the AleI-EcoRV fragment of pShuttle-hH-1 with the corresponding 393-base-long DNA fragment from phH-1-TO.

pAd-hH-1, pAd-hH-1-STOP, and pAd-hH-1-TO were generated by recombination of pShuttle-hH-1, pShuttle-hH-1-STOP, and pShuttle-hH-1-TO, respectively, with pAd5ΔE1ΔE3 into E. coli BJ5183, according to the AdEasy adenoviral vector system instruction manual (Agilent Technologies; Stratagene Products, Waldbronn, Germany).

pAd5ΔE1ΔE3, in which the region encompassing the E1 (nucleotides [nt] 459 to 3228) and E3 (nt 27897 to 30463) genes (Ad5 Refseq accession no. AC_000008) was deleted, was constructed as follows. A shuttle plasmid containing E3 flanking sequences (18) was digested using MluI and XbaI, subjected to blunting with a Klenow fragment, and ligated, generating pLeft-Right. This plasmid was digested with SalI and NotI, dephosphorylated, and used for homologous recombination in E. coli BJ5183 with SpeI-linearized pTG3622 (4) from which the E1 region has been deleted.

All the constructs described here were clonally isolated and their full-length sequences verified.

Transfections.

DNA transfections were carried out using Fugene HD (Roche Diagnostics & Applied Sciences, Mannheim, Germany) according to the manufacturer's instructions with minor modifications. Plasmids were diluted in serum-free medium to a final concentration of 20 ng/μl. Fugene HD was then added at a 1:2.5 ratio (expressed as the ratio of micrograms of DNA to microliters of Fugene), and the mixture was incubated at room temperature (RT) for 30 to 60 min. Subsequently, the mixture was added to the cells in a dropwise manner.

Protein extractions and Western blot analysis.

Cellular pellets were lysed on ice for 30 min in 5 volumes of protein extraction buffer (50 mM Tris [pH 8], 200 mM NaCl, 0.5% NP-40, 1 mM dithiothreitol [DTT]) containing protease inhibitors (Complete EDTA-free; Roche, Mannheim, Germany) and 10% glycerol. Cell debris was removed by centrifugation at 10,000 rpm for 10 min at 4°C. A 20-μg volume of total protein extract was resolved by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto a Hybond-P membrane (GE Healthcare). The following antibodies were used for the analysis: mouse monoclonal anti-β-tubulin (clone TUB 2.1; Sigma-Aldrich, Saint Louis, MO), mouse monoclonal anti-actin (clone C4; MP Biomedicals, Illkirch, France), polyclonal anti-NS1 SP8 antiserum (provided by Nathalie Salomé, ATV-DKFZ, Heidelberg, Germany) (3), and polyclonal anti-VP2 antiserum (a gift from Christiane Dinsart, ATV-DKFZ, Heidelberg, Germany) (17).

Parvovirus titration: plaque assay.

NBK cells grown at a density of 20,000 cell/cm² were infected with serial dilutions of crude virus extracts for 1 h, followed by replacement of the inoculum with an overlay of 0.68% Bacto agar (Becton, Dickinson GmbH, Heidelberg, Germany) in MEM (Gibco, Invitrogen) supplemented with 5% FBS, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. At 7 days postinfection, living cells were stained for 18 h with an overlay of neutral-red (0.2 mg/ml)-containing Bacto agar (0.85%) diluted in phosphate-buffered saline (PBS). Plaques were counted, and titers were expressed as PFU/milliliter.

Real-time qPCR.

Crude virus extracts were digested with 50 U/ml of Benzonase nuclease (ultrapure grade; Sigma-Aldrich Chemie GmbH, Steinheim, Germany) for 30 min at 37°C to remove free viral genomic DNA. To release viral DNA from viruses, 10 μl of each sample was lysed in a total of 40 μl of alkaline lysis buffer (1 M NaOH—Tris-EDTA [TE] buffer) at 56°C for 30 min. Lysis was stopped by adding 960 μl of 40 mM HCl. Quantification of viral DNA was carried out by real-time quantitative PCR (qPCR) with an NS1-specific TaqMan probe (Applied Biosystems, Darmstadt, Germany), as previously described (10). With this method, we calculated that 1 PFU of hH-1-TO and 1 Adeno IU (see below) of Ad-hH-1-TO corresponded to approximately 500 and 100 viral genome-containing particles, respectively.

Adenovirus titration.

Recombinant adenovirus titers were determined using an Adeno-X Rapid Titer kit (Clontech, Saint-Germain-en-Laye, France), 3 days after infection of T-REx-293 cells (Invitrogen), and are expressed as infectious units (IU)/ml. Adenovirus IU were measured using an antibody specific for the adenovirus hexon protein. The yield values calculated with this method are in good agreement with those obtained by plaque and gene transduction assays (2).

The concentration of adenovirus physical particles was estimated from the DNA content of the purified viruses, measured by the absorbance at 260 nm. An absorbance value of 1 corresponds to 1.1 × 10¹² adenovirus particles/ml (21).

Electron microscopy.

Carbon-coated 300-mesh copper grids were placed face down onto 5-μl aliquots of virus suspension for 2 min, stained with 2% uranylacetate for 30 s, and dried for approximately 1 min. Micrographs were taken at 38,000-fold magnification with a Zeiss 10A transmission electron microscope (Zeiss, Oberkochen, Germany) using an acceleration voltage of 80 kV. The magnification indicator was routinely controlled by comparison with a grating replica.

Quantification of infection efficiency.

Cells were seeded in a 12-well plate and infected with equal amounts of Ad-green fluorescent protein (Ad-GFP) or H-1-GFP (10 GFP transduction units [TU]/cell) as previously quantified using the HEK293T reference cell line. At 48 h and 72 h p.i., infected cells were harvested by trypsinization and pelleted by centrifugation for 5 min at 1,500 rpm. Samples were washed with PBS, pelleted again, and then resuspended in 750 μl PBS. Cells were analyzed by flow cytometry (FACSort; Becton, Dickinson, Franklin Lakes, NJ), and percentages of GFP-positive cells were determined using FCS Express version 3 (De Novo Software, Los Angeles, CA).

LDH and MTT assays.

Human cells were first seeded in 96-well plates at densities of 4,000 cell/well for cancer cells and 8,000 to 10,000 cell/well for primary cells. The respective culture media were as described above, except for melanocytes and astrocytes, which were seeded in DMEM supplemented with 10% heat-inactivated FBS instead of the artificial media in which they are routinely grown. After infection, all cells were kept in their corresponding basal medium supplemented with 5% of heat-inactivated bovine serum (200 μl/well). Cancer cells and primary cells were incubated for 4 and 5 days, respectively, and then subjected to lactate dehydrogenase (LDH) and MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide] assays as previously described (16).

Released LDH was measured according to the CytoTox 96 nonradioactive cytotoxicity assay (Promega Biotech, Madison, WI), using an enzyme-linked immunosorbent assay (ELISA) reader at 492 nm. After subtraction of the background value determined with nonconditioned medium, the fraction of lysed cells in infected or noninfected cultures was calculated from the ratio of the LDH activity in the conditioned medium to the total LDH activity of the corresponding culture. The total LDH activity was determined after cell lysis by the addition of 10× buffer containing 9% (vol/vol) Triton X-100. The same cell cultures were used to determine LDH release and MTT activity. Activity of MTT (Sigma-Aldrich Chemie GmbH, Steinheim, Germany) was read with an ELISA reader at 595 nm. The viability of infected cells was expressed as the ratio of the corresponding absorbance to that of noninfected cells taken arbitrarily as 100%.

Real-time detection of cell proliferation and viability.

Cells were seeded on a 96-well E-Plate (Roche Diagnostics Deutschland GmbH, Mannheim, Germany) at a density of 4,000 cells/well (or, for HCC-2998, 8,000 cells/well). At 24 h to 72 h later, cells were infected with Ad-hH-1-TO, Ad5ΔE1ΔE3 (Ad control), or hH-1-TO or the combination of the Ad control and hH-1-TO viruses. Cellular proliferation, reflecting virus-mediated cytopathic and cytostatic effects, was monitored in real time, every 30 min, using an xCelligence System (Roche Diagnostics Deutschland GmbH, Mannheim, Germany). The growth curves shown represent the averages of the results of at least three replicate experiments and include relative standard deviations.

RESULTS

Generation of parvovirus hH-1-TO carrying a tetracycline-inducible P4 promoter.

Our first attempts to produce adenovirus carrying the hybrid Ad-PV genome failed, most likely due to the interference of the parvovirus NS proteins (data not shown). Indeed, the introduction of stop codons within the NS open reading frame could rescue Ad production to standard titers (more than 4 × 10¹² Ad particles/ml were obtained after production and purification according to the protocol described in Materials and Methods). Attempts to rescue Ad-PV chimera production by silencing NS1 expression by means of the use of specific antisense oligonucleotides, small interfering RNAs (siRNAs), or short hairpin RNAs (shRNAs) (used singly or in combination) allowed viral particles to be produced, and yet the particles were produced in very small amounts (fewer than 400 Ad particles were produced by 1 × 10⁷ cells), suggesting that even at low levels, NS1 exerted a negative effect on Ad production (data not shown). In view of these results, we decided to modify the parvovirus early P4 promoter, which controls the expression of the NS gene, in order to tightly control its activity during chimera production. We took advantage of the T-REx technology (15, 36) and engineered the P4 promoter to make it inducible by inserting two tetracycline operator 2 elements (TetO²) (36) between the TATA box and the NS starting codon (Fig. 1). With this modification, we expected to repress the P4 promoter in T-REx-293 cells that constitutively express the tetracycline repressor (TetR) in the absence of doxycycline (DOX) and induce the promoter in the presence of the drug (36). In contrast, in cancer cells which do not express TetR, the P4 promoter should be fully functional. This modified parvovirus was generated and named phH-1-TO.

Fig 1.

Construction of an inducible parvovirus P4 promoter (P4-TO). (A) Schematic view of the P4-TO promoter generated by inserting two tetracycline operator 2 (TetO₂) elements into the P4 promoter of the hH-1 genome. TF, transcription factors. (B) DNA sequence of the P4-TO promoter. The TATA box, the two TetO₂ elements, and the NS translation ATG start codon are highlighted.

T-REx-293 cells were transfected with phH-1-TO or parental phH-1, incubated for 2 days, harvested, and processed for Western blotting detection of parvoviral NS1 and VP proteins. In cells transfected with the parental phH-1 viral vector, the addition of DOX did not affect total NS1 protein levels (Fig. 2A). In contrast, in phH-1-TO-transfected cells, expression of NS1 was induced in the presence of DOX, demonstrating that the activity of the P4-TO promoter is under the control of an on-off switch mechanism. Under induction conditions, NS1 protein levels were similar to those observed using the parental viral vector, indicating that the insertion of the TetO₂ did not impair NS production when DOX was supplied to the cell medium (Fig. 2A). It is known that the NS1 protein transactivates the parvoviral p38 promoter, which controls the transcription of the VP genes coding for the capsid proteins. In agreement with the repression of NS1 production, only slight expression of VP1 and VP2 was observed in phH-1-TO-transfected T-REx-293 cells grown under DOX-free conditions. In contrast, VP proteins accumulated in these cells when DOX was added to the medium (Fig. 2A).

Fig 2.

Characterization of the hH-1-TO virus. (A) Inducible expression of the phH-1-TO plasmid containing the P4-TO promoter. T-REx-293 cells, constitutively expressing the Tet repressor, were transfected with either phH-1 or phH-1-TO molecular clones and grown in medium supplemented with DOX (1 μg/ml) (+) or were left unsupplemented (−). After 48 h, total protein cell extracts were prepared from these cultures and analyzed by Western blotting for the presence of viral proteins (NS1 and VP) and actin (used as a loading control). (B) Infectiousness of the hH-1-TO virus. T-REx-293 cells were transfected with either phH-1 or phH-1-TO viral plasmids and grown for 1 week in the presence or absence of DOX. Cells were harvested within their medium and lysed. Produced viruses were further amplified by infecting NB324K cells. Cell lysates from these cultures were then analyzed for the presence of parvovirus particles by a plaque assay using NB324K indicator cells. Representative images (5-cm-diameter areas) from the plaque assay are shown. (C) DOX dependence of hH-1-TO virus replication in T-REx-293 cells. T-REx-293 cells, grown in 6-well plates, were infected with hH-1 or hH-1-TO viruses, at an MOI of 2,500 viral genomes (Vg) per cell, and further grown in the presence or absence of DOX. After 4 days, cells were harvested within their medium and lysed through 3 freeze-and-thaw cycles. After elimination of cellular debris by centrifugation, crude virus preparations were treated with Benzonase to remove free viral DNA and processed for parvovirus-specific qPCR. Titers of hH-1 or hH-1-TO parvoviruses are expressed in Vg/milliliter.

To assess whether transfection with phH-1-TO also resulted in the production of infectious progeny virions, T-REx-293 cells were transfected with phH-1-TO or parental phH-1 constructs and grown in the presence or absence of DOX for a total of 7 days. Cell lysates from these cultures were then used for the infection of NB324K cells. After an additional 7 days, crude cellular extracts were tested for the presence of full virions able to infect, kill, and spread in NB324K indicator cells, as measured by a plaque assay. As expected, the parental hH-1 virus was produced at similar levels irrespective of the presence or absence of DOX. In contrast, DOX was required during the T-REx-293 transfection phase for hH-1-TO virions to be produced at a significant level (Fig. 2B), yielding virus titers comparable to the ones obtained with the parental virus (data not shown). These results show that TetO₂ insertions into the P4 promoter region make PV production in the T-REx-293 cells dependent on DOX induction. Moreover, the insertions are fully compatible with the entire course of parvovirus life cycle, as the de novo-generated hH-1-TO viral particles were fully infectious and capable of autonomously replicating in cells like NBK324 that do not express the TetR.

The propagation of hH-1-TO viruses was further investigated in T-REx-293 cells. Cells were inoculated with either hH-1-TO or hH-1 viruses and grown in the presence or absence of DOX for 4 days with one medium change at 24 h p.i. in order to eliminate unbound viral particles. Cells were then lysed into their respective media, and the parvovirus production was evaluated by a parvovirus-specific qPCR. In agreement with the results presented above, production of hH-1-TO virus in T-REx-293 cells was efficient only when cultures were grown in the presence of DOX, with a 60-fold reduction of virus titers in the absence of the inducer (Fig. 2C). It should be noted, however, that the production of hH-1-TO in the presence of DOX was about 6-fold lower than that of hH-1, suggesting that the modifications introduced into the P4 promoter region slightly reduced the fitness of the virus in these cells.

Generation of the Ad-PV chimera.

The results presented above prompted us to insert the entire parvovirus hH-1-TO genome into the DNA backbone of a replication-deficient adenovirus vector (Ad5ΔE1ΔE3), thus generating the plasmid pAd-hH-1-TO containing the chimeric vector genome. We first tested whether the P4-TO promoter region kept its TetR sensitivity once inserted into the Ad genome. For this purpose, T-REx-293 cells were transfected with the pAd-hH-1-TO chimeric plasmid and grown for 5 days with or without DOX. Total protein extracts from these cells were then analyzed by Western blotting for the presence of the parvoviral NS1 and VP proteins. As illustrated in Fig. 3A, significant production of NS1 was detected only when DOX was provided to the cells. Consistent with previous results (Fig. 2A), NS1 expression correlated with an induction of VP1 and VP2 capsid protein expression. These results confirmed the TetR sensitivity of P4-TO gene expression in an Ad context.

Fig 3.

Generation of Ad-PV chimeras. (A) Inducible gene expression from the chimeric pAd-hH-1-TO plasmid. T-REx-293 cells were transfected with pAd-hH-1-TO plasmid and further grown in medium with or without DOX for 5 days. Cells were then lysed and total protein extracts analyzed by Western blotting for the presence of parvovirus NS1 and VP proteins and β-tubulin (loading control). No Trans., no transfection. (B) Ad-PV chimera production. In a first round of production, T-REx-293 cells were transfected with the chimeric pAd-hH-1-TO plasmid or either of the parental pAd (pAd5ΔE1ΔE3) and pAd-hH-1 plasmids and grown in medium supplemented with DOX or without supplementation. Cell lysates from these cultures were used for infection of fresh T-REx-293 cells in a second round of production, and the procedure was repeated a third time by scaling up the volume of the culture flasks as described in Materials and Methods. Viral stocks were purified twice through CsCl gradient ultracentrifugation and titrated using an Adeno-X Rapid Titer kit (Clontech), and yields were expressed as Ad infectious units/milliliter (IU/ml). (C) Electron microscopy (EM) analysis of produced virions. EM images of the purified Ad-hH-1-TO (Ad-PV), Ad, and hH-1-TO viruses are shown. Bars, 100 nm.

We next investigated whether it was possible to produce Ad-PV chimeric virions from the pAd-hH-1-TO construct. As a negative control, we used the pAd-hH-1 vector containing the wild-type P4 promoter (from which we previously failed to generate the chimeric virions), and, as a positive control, we used the parental Ad plasmid whose E1 gene deletion is complemented by the 293 cells used as producers. In a first round of production, T-REx-293 cells were transfected with PacI-linearized pAd-hH-1-TO, pAd-hH-1, or pAd plasmids and grown in medium with or without DOX for 7 days before being lysed. Crude cell extracts were then used to reinfect fresh T-REx-293 cells in two successive rounds of virus amplification. The final cell lysates were then purified by cesium chloride (CsCl) gradient centrifugation and the viral stocks titrated using an adenovirus replication assay. Virus preparations were also subjected to electron microscopy analysis to control their purity. Analysis of the crude cell lysates before Ad purification showed that, in agreement with previous results, the pAd-hH-1 chimeric vector failed to generate any detectable viral particles. In contrast, chimeric viruses were produced in T-REx-293 cells by transfecting the cells with pAd-hH-1-TO vector, unless DOX was added to the cell culture medium (data not shown). These results are in line with the evidence indicating that parental NS proteins (most likely NS1) were responsible for the inhibition of adenovirus chimeric virus replication. Remarkably, the production of the Ad-hH-1-TO (Ad-PV) chimeric viruses in the presence of functional TetR was very efficient, as it yielded titers similar to Ad control titers after purification (Fig. 3B). Electron microscopy analysis showed neither differences between Ad-PV and Ad control particles nor parvovirus contamination in the produced Ad-PV viral stocks (Fig. 3C). Taken together, these results demonstrate that, by transiently blocking the parvoviral NS transcription unit, it is possible to produce Ad-PV chimeric viruses at high titers in the T-REx-293 packaging cell line.

Generation of infectious parvovirus particles from the Ad-PV chimera.

We then investigated whether fully infectious parvoviruses were produced after infection of transformed target cells with the Ad-PV chimeric viral particles. We used two cell lines permissive for parvovirus production, namely, simian virus 40 (SV40)-transformed NB324K and cervical carcinoma-derived SiHa cells. After infection with purified Ad-PV, crude cellular extracts were analyzed for the presence of infectious parvovirus particles by a plaque assay. In accordance with our initial working hypothesis, autonomously replicating infectious parvoviruses were produced upon infection of both NB324K and SiHa cells with Ad-PV chimeric virions (Fig. 4A). H-1PV production is routinely carried out in two steps, first by transfecting HEK293 cells with a plasmid harboring the viral genome, to produce an initial virus batch, and then by amplifying this batch through infection of NB324K cells for 3 to 4 days. We compared the yields of PV particles produced in NB324K cells following their infection with equivalent genomic amounts of H-1PV and Ad-PV chimera. As shown in Fig. 4B, similar PV titers were obtained irrespective of whether producer cells were infected with H-1PV or Ad-PV chimeras. In order to verify whether PV particles generated by Ad-PV chimera-infected cells in a cancer cell population are indeed able to infect neighboring cells and multiply therein, we have carried out “virus spread” assays in HeLa cells. In order to distinguish between PV production during the first round of infection (Ad-PV → PV conversion) and total PV production, cells were treated with neuraminidase (NA) or left untreated. At 10 h postinfection, NA is known to prevent H-1PV from binding to cell plasma membrane, by catalyzing the hydrolysis of sialic acid, an important component of the H-1PV receptor (1), and therefore does not interfere with virus replication in preinfected cells while inhibiting spreading of progeny virus and further amplification. As shown in Fig. 4C, the PV yield at 72 h postinfection was significantly higher in HeLa cells cultured in the absence of NA, providing evidence of PV spreading in these cultures. It should also be stated that PV yields in the presence of NA were similar for PV- and Ad-PV-infected cells (data not shown). Altogether, these results indicated that the PV component of the Ad-PV chimera was efficiently rescued from the chimera, resulting in primary PV production which was followed by secondary rounds of PV amplification. This was confirmed by measuring the capacity of the PV particles generated and released by Ad-PV-infected cells to kill neighboring cancer cells. HeLa cells were infected with Ad-PV at low multiplicities of infection (MOIs) (0.2 and 0.4 IU/cell) and then grown in the presence or absence of NA for 72 h before being processed for LDH assays. As shown in Fig. 4D, only a small fraction of cells was killed by the chimera when NA was added to the culture medium after Ad-PV infection. This is consistent with the fact that NA treatment prevents second rounds of infection, thereby restricting killing to the fraction of cells initially hit by the chimera. In contrast, the whole population of cells grown in the absence of NA died within 72 h postinfection, indicating that PV particles produced by chimera-infected cells were able to spread to neighboring cancer cells and kill them efficiently (Fig. 4D). These results provide a proof of concept that the Ad-PV chimera can be used as a novel tool for the delivery of autonomous parvoviruses to target cells and that, once brought into cells by the chimera, the PV genome is released and initiates the de novo synthesis of fully infectious parvovirus progeny particles.

Fig 4.

Production, spreading, and cytotoxicity of progeny parvoviruses in cells infected with the Ad-PV chimera. (A) Parvovirus production. NB324K or SiHa cells were infected with Ad-PV chimeras used at an MOI of 10 (NB324K) or 1 (SiHa), and culture media were renewed after 1 day to remove unbound viruses. After further incubation for 4 to 5 days, cells were harvested within their medium and lysed. Crude virus preparations were analyzed for the presence of parvoviruses by plaque assays. Titers of produced parvovirus are expressed in PFU/milliliter. (B) Parvovirus production and spreading. NB324K cells were infected with equivalent genomic amounts of Ad-PV chimera or H-1PV wild-type virus (input, 10 Vg/seeded cell). After 96 h, cells were collected in their medium and subjected to three freeze-thaw cycles. After treatment with 50 U/ml of Benzonase for digesting cellular DNA and nonencapsidated viral DNA, crude cell extracts were analyzed for their content of full viral particles by real-time qPCR, as described in Materials and Methods. (C) Parvovirus spreading. HeLa cells were infected with Ad-PV chimeras (25 Vg/cell). At 10 h postinfection, one set of dishes was treated with 0.1 U/ml of neuraminidase (+ NA) in order to prevent second rounds of parvovirus infection, while another set was left untreated (− NA). Benzonase-pretreated total cell lysates were analyzed by real-time qPCR for virus particle content. (D) Parvovirus spreading and cytotoxicity. HeLa cells, grown in 96-well plates, were infected with Ad-PV chimeras, treated with NA as described for panel C or left untreated, and processed for LDH assay after 72 h.

Enhanced oncotoxic potential of the Ad-PV chimera.

It is known that H-1PV induces cytopathic effects on a large number of cancer cells (reference 10 and our unpublished results). However, there are also cancer cell lines which are weakly susceptible or completely refractory to H-1PV cytotoxicity (our unpublished results). Preliminary experiments showed that the hH-1 and hH-1-TO parvoviruses exerted similar cytotoxic activities against HeLa cells as measured by LDH assays, indicating that the modification introduced into the P4 promoter region did not modify the oncolytic activity of the parvovirus (data not shown). The cytotoxic activities of Ad-hH-1-TO (Ad-PV) chimera and both parental viruses (Ad control [Adc] and hH-1-TO [PV]) were compared using a panel of human cancer cell lines differing in their sensitivities to H-1PV infection, including highly sensitive cells (cervical carcinoma-derived HeLa, melanoma pMelL), poorly sensitive cells (lung cancer-derived A549), and resistant cells (cervical carcinoma-derived ME-180, melanoma Lox-IMVI, colon cancer HCT-15 and HCC-2998). In a first step, the susceptibility of these cells to infection with Ad5 and H-1PV was determined by using recombinant viruses harboring the GFP reporter gene (Ad-GFP or H-1-GFP). As illustrated in Fig. 5, fluorescence-activated cell sorter (FACS) analysis performed 48 and 72 h after infection revealed that Ad-GFP and H-1-GFP transduced HeLa and A549 in similar manners whereas Ad-GFP was much more efficient than H-1-GFP in transducing the cell lines previously identified as resistant to H-1PV cytotoxicity. As expected, pMelL cells (which lack the Ad receptors on their surface) were efficiently transduced by H-1-GFP but not Ad-GFP (Fig. 5).

Fig 5.

Transduction efficiency of recombinant Ad-GFP and H-1-GFP. The indicated cells were infected with recombinant Ad or PV (1 TU/cell as quantified by using HEK-293T cells) carrying the GFP reporter gene. After 48 and 72 h, cells were harvested and analyzed by flow cytometry. Values represent the percentages of GFP-positive cells.

In a second step, viral cytotoxicity was evaluated by an LDH assay (analysis of cell lysis), an MTT assay (analysis of cell viability), and xCelligence (real-time analysis of cell growth). (i) The cytotoxic activity of the Ad-PV chimera in HeLa cells was first evaluated in comparison with that of the parental viruses (Fig. 6A). Cells were infected with equal amounts of the different viruses. In agreement with previous results, LDH and MTT assays revealed that PV used singly or in combination with the Ad control killed HeLa cells very efficiently. A similar cytotoxic effect was observed with the Ad-PV chimera, indicating that, under these experimental conditions, the PV cytotoxic potential is preserved upon PV delivery through the Ad vector. As the Adc alone was much less toxic than PV (as expected from the low MOI used and from the fact that the Adc is a replication-deficient virus), we concluded that the cytotoxicity of the chimeric Ad-PV in these cells was mainly due to the PV component (Fig. 6A). In agreement with these results, analysis of the cell growth curves using the xCelligence system showed that the Adc virus had a limited effect on the proliferation of HeLa cells whereas the PV was able to efficiently suppress the proliferation and induce death of these cells (Fig. 7A). The results showing this striking HeLa cell growth suppression and killing were achieved to a similar extent when the PV was delivered by the chimera. It is worth noting that the cytotoxic effect of the Ad-PV chimera was delayed by approximately 24 h in comparison to the one induced by PV, which is consistent with the time needed for PV rescue from the chimeric vector in these cells. A549 cells were also found to be sensitive to both PV and chimeric Ad-PV cytotoxicity, confirming that the oncolytic potential of H-1PV is kept by the chimeric virus (Fig. 6B and Fig. 7B). In keeping with the fact that A549 cells are less susceptible to H-1PV infection than HeLa cells, a concentration of PV or Ad-PV chimera that was 10 times higher was required in order to efficiently kill these cells. (ii) We then tested whether the chimera is able to kill tumor cells, namely, ME-180, Lox-IMVI, HCT-15, and HCC-2998, previously identified as being resistant to direct H-1PV infection (Fig. 6 and 7). LDH and MTT assays showed that infections with Ad and PV viruses (used singly or in combination) had little cytotoxic effect on these cells. In striking contrast, the Ad-PV chimera was much more toxic and efficiently killed all tumor cell lines analyzed, indicating that the chimera has improved oncolytic activity compared to that of the parental viruses from which it originated (Fig. 6C to F). Kinetic analyses showed that the growth of the cell lines described above was only marginally disturbed by PV, confirming their significant resistance to the parvoviruses (Fig. 7C to F). On the other hand, the Adc virus had various toxic effects in these cells, ranging from full growth suppression (HCC-2998; see Fig. 7F) or growth retardation (ME-180; see Fig. 7C) to full resistance (Lox-IMVI and HCT-15; see Fig. 7D and E, respectively). Similar results were obtained with the Adc used in combination with PV. Interestingly, the Ad-PV chimera was found to be much more toxic than the parental viruses, already fully suppressing the growth of all cells and inducing strong cytotoxic effects at 20 to 40 h postinfection (Fig. 7C to F). (iii) As a control, the Ad-5 receptor-null pMelL cell line (30) was also included in this analysis. As shown in Fig. 6G and 7G, these cells were sensitive to PV cytotoxicity but completely resistant to both the Adc and Ad-PV chimera, indicating that the chimera exerts its cytotoxicity only in cells competent for Ad uptake (Fig. 6G and 7G).

Fig 6.

Improved cytotoxic activity of Ad-PV chimera toward cancer cell lines. LDH and MTT assays were used to measure infected cell killing and viability, respectively. HeLa (A), A549 (B), ME-180 (C), Lox-IMVI (D), HCT-15 (E), HCC 2998 (F), and pMelL (G) human cancer cells were seeded in 96-well plates and infected with Ad-hH-1-TO chimera (Ad-PV), Ad control (Adc) (MOI for both Ads were expressed as IU/cell), hH-1-TO (PV; expressed as PFU/cell), or hH-1-TO in combination with Adc (PV + Adc) viruses at the indicated MOIs (expressed as IU/cell for Ad-PV and Adc or PFU/cell for PV). Untreated cells (−) were used as the control. Values represent the percentages of lysed (LDH assay) or viable (MTT) cells calculated as described in Materials and Methods.

Fig 7.

Enhanced toxicity of the Ad-PV chimera for cancer cell lines. HeLa (A), A549 (B), ME-180 (C), Lox-IMVI (D), HCT-15 (E), HCC 2998 (F), and pMelL (G) human cancer cells were seeded in 96-well E-plates (xCelligence Roche) and infected at the indicated MOIs with Ad-hH-1-TO chimera (Ad-PV), Ad control (Adc), hH-1-TO (PV), or hH-1-TO in combination with Adc (PV + Adc) viruses. The proliferation curves of virus- versus mock-infected cells were monitored in real time using the xCelligence system. Cell index values are proportional to cell numbers and were recorded every 30 min for a maximum of 1 week. Results are presented as average values with relative standard deviation bars from triplicate measurements in a typical experiment. Arrows indicate the time of infection.

In a third step, it was important to verify that the tumor specificity of PV cytotoxicity was kept when the PV was delivered by the Ad vector. To this end, the cytotoxicity of the chimera was compared with that of its parental viruses in primary healthy human cells of different origins, namely, foreskin fibroblasts, oral fibroblasts, melanocytes, and astrocytes, by LDH and MTT assays. Despite the fact that Ad-GFP was more efficient than H-1-GFP at transducing all the cell cultures tested (Fig. 8A), even after a longer virus incubation compared with incubation of cancer cells (5 versus 4 days), healthy cells were found to be fully resistant (foreskin fibroblasts) or only minimally sensitive (oral fibroblasts, melanocytes, and astrocytes) to the cytotoxic activity of the Ad-PV chimera, in striking contrast with the above-mentioned high sensitivity of cancer cells (Fig. 8B).

Fig 8.

Limited cytotoxicity of the Ad-PV chimera for healthy human primary cells. (A) Flow cytometry. Human foreskin fibroblasts, oral fibroblasts, melanocytes, and astrocytes were infected with Ad-GFP or H1-GFP (1 GFP TU/cell). After 72 h, cells were harvested and subjected to flow cytometric analysis for the quantification of GFP-positive cells. Mock-treated cells were used for setting fluorescence background values; HEK293T cells were used as a positive control. (B) LDH and MTT assays. Human foreskin fibroblasts, oral fibroblasts, melanocytes, and astrocytes were seeded in 96-well plates and infected with chimeric Ad-hH-1-TO (Ad-PV), Ad control (Adc), hH-1-TO parvovirus (PV), or a mixture of Ad control and hH-1-TO viruses (PV + Adc) or left uninfected (−). After 5 days of incubation, percentages of lysed or viable cells were assessed by LDH and MTT assays, respectively, as described in Materials and Methods.

Altogether, these results show the improved oncolytic activity of the Ad-PV chimera compared with Ad and PV parental viruses.

DISCUSSION

In this study, we constructed the first adenovirus-autonomous parvovirus (Ad-PV) chimera by inserting the complete genome of hH-1PV into the Ad5 genome deleted of the E1 and E3 regions (Ad5ΔE1ΔE3). Our aim was to develop a system that combines the most favorable features of adenovirus (high-efficiency gene delivery, large packaging capacity, high titer production) and parvovirus (oncolytic and oncosuppressive properties and the absence of pathogenicity in humans). The project faced two major challenges: (i) the production of chimeric adenovirus containing a potentially interfering PV insert and (ii) the complete recovery of the PV genome from the Ad backbone with the generation of fully infectious PV particles exerting cytotoxic activity in cancer cells but not in healthy cells. These two requirements proved to be difficult to reconcile. Indeed, our first attempts to develop Ad-PV chimeras failed at the production stage due to the strong negative interference of parvoviral NS proteins with hybrid adenovirus vector replication. This inhibitory effect has been previously reported in another study where the production of adenovirus type 5 containing a parvovirus-expressing cassette (including the NS transcription unit under the control of its natural P4 promoter and the late parvoviral P38 promoter driving the expression of a heterologous transgene) was possible only after silencing of parvovirus NS1 expression by means of antisense oligonucleotides (28). However, this strategy was not successful in the present system, as neither the use of antisense oligonucleotides nor the use of siRNAs or shRNAs directed against NS1 (transiently transfected as well as constitutively expressed in stable cell lines) allowed the chimeras to be produced at significant titers (data not shown). The explanation for this failure may reside in the greater complexity of our vector in comparison to the one previously published. Indeed, the present chimera can be distinguished by the fact that it comprises the entire PV genome, including NS1-specific nicking sites at its extremities (17), and can thus be expected to be sensitive to NS1-endonuclease activity. The presence of these sites is essential for NS1-mediated excision and release of the PV genome from the adenovirus backbone in target cancer cells but may also preclude hybrid adenovirus genome replication in producer cells by allowing NS1 to disrupt the adenovirus backbone. Similar technical limitations were also encountered in the generation of adenovirus-AAV chimeras, where the expression of the Rep 78 gene precluded adenovirus replication (34) and maturation by colocalizing to the Ad replication centers (35). Our initial failure to produce the Ad-PV chimera suggested that even very low levels of NS1 were sufficient to interfere with Ad vector replication. We circumvented this problem by inserting tetracycline operator sequences (15, 36) into the PV P4 promoter in order to tightly control NS1 expression. In T-REx-293 cells, which constitutively express the Tet repressor (TetR), the activity of P4-TO was drastically suppressed. As a consequence, in parvovirus H-1TO-transfected and -infected T-REx-293 cells in the absence of DOX inducer, no or little expression of NS1 protein and virus production were detected. In contrast, in tumor cells not expressing TetR, expression of NS1 and the overall fitness of the modified parvovirus were only slightly reduced in comparison with wild-type H-1PV results. Thus, we identified a position within the H-1PV P4 promoter at which a foreign sequence (in our case, sequences of Tet-responsive elements) can be inserted to modify the functionality of the promoter without jeopardizing the overall replication and fitness of the virus in permissive cells. This finding may pave the way for further engineering of the P4 promoter through insertion into the same site of other sequences that could improve cancer-specific transcription and replication of H-1PV. As the present Tet-responsive element addition makes NS1 expression inducible, the H-1PV-TO virus could also be used in studies aiming to further characterize the role of NS proteins in the viral life cycle and associated cytopathic effects.

The results we obtained prompted us to introduce the modified parvovirus genome into the Ad genome, resulting in the construction of the Ad-hH-1-TO chimera (Ad-PV). Consistent with our initial hypothesis, the Ad-PV chimera was produced at high titers in T-REx-293 cells. Furthermore, this chimera efficiently delivered its PV component to cancer cells in which the parvoviral genome was excised from the vector and replicated autonomously, yielding infectious progeny PV particles. Rescue from the Ad backbone was probably facilitated by the fact that the inserted H-1 genome (hH-1) contains an extra consensus NS1 nick site at the left-hand viral terminus (17). Most remarkably, the Ad-PV chimera was more efficient in killing various cancer cell lines than the parental PV or Ad (used alone or in combination). The mechanisms underlying this improved cytotoxicity are still a matter of speculation and may differ from one cell line to another. In some tumor cells, the greater efficiency of Ad-PV may reside, at least in part, in the higher capacity of Ad (and therefore of the chimera) for initiating infection compared to that of the PV. Indeed, experiments performed with recombinant Ad and PV viruses expressing GFP showed that some PV-resistant cells could be more efficiently transduced by the Ad. The Ad chimera may be more competent than the PV for delivering the parvoviral genome into the nucleus of such cells as the result of using a distinct mechanism of particle uptake. If so, it may be due to differences between Ad and PV in their recognition of critical cellular factors involved in virus cell-binding and entry (i.e., receptors or coreceptors) or required for the trafficking of the virus from the cell surface to the nucleus. A more efficient Ad-PV chimera-mediated delivery of the PV genome into the nucleus of infected cells can be expected to result in enhanced PV gene expression, leading to an increase in the production of cytotoxic NS1 protein and in the induction of oncolysis. The Ad component of the chimera could also assist parvovirus replication in cancer cells at levels other than entry and nuclear trafficking. We have recently shown that specific Ad genomic elements stimulate recombinant PV production by more than 100-fold (10). Similarly, Ad enhances the production of the human B19 parvovirus (14). Ad also has the ability to counteract the innate immune response (19). However, the fact that Ad and PV coinfection did not kill cancer cells as effectively as infection by the chimera argues against these mechanisms being major contributors to the higher cytotoxic activity of the chimera. Further studies are needed to unravel the mechanism(s) behind the improved oncolytic potential of the chimera. It is also worth mentioning that, although the Ad virus was transducing healthy cells (e.g., human primary astrocytes) more efficiently than the PV, the Ad-PV chimera was of low toxicity to these cells, in keeping with the specificity of PV cytotoxicity for transformed cells (25).

The present Ad-PV chimera is a prototype which is open to further development. (i) It allows new approaches to be considered in order to increase the cancer specificity of PV-based treatments, taking advantage of the Ad-retargeting technology. Indeed, tumor-retargeted Ads can be obtained by inserting cancer-specific retargeting ligands or adaptors (peptides, single-chain antibodies, Affibody molecules, etc.) into the Ad capsid (20). These retargeted Ads could be used for the development of a second generation of Ad-PV chimeras that would act as vehicles for the delivery of PV genomes into cancer cells only. This would be of great benefit to circumvent the uptake of PVs by healthy cells, which results in the sequestration of a major fraction of the PV inoculum (7). (ii) The chimera might represent a valid solution for treatment of tumors that are heterogenous with respect to virus receptors. In this way, cells less susceptible to PV infection could be targeted by the Ad-PV chimera and the released PV progeny particles could in their turn infect cells poorly expressing the Ad receptor(s) at their cell surface. (iii) Beside its use to improve the specificity of infection, the adenoviral component of Ad-PV chimeras gives the possibility of inserting large transgenes in the Ad backbone, which cannot be done with PVs without making them replication incompetent (6). It should be possible in this way to arm the chimera with therapeutic transgenes that reinforce the intrinsic PV cytotoxic activity and/or increase the PV ability to replicate in cancer cells. (iv) Another development of the Ad-PV chimera may take advantage of the availability of Ad serotypes (33). This should allow serologically distinct Ad-PV chimeras to be generated and used sequentially in clinical protocols in order to escape neutralization by antiviral antibodies. (v) In addition to providing the versatility mentioned above, the use of adenoviral vectors as PV carriers may help to improve the production of PVs and PV-based vectors. Indeed, as with any other Ad-based vectors, it should be possible to produce the Ad-PV chimeras in the large amounts and under the GMP conditions compatible with clinical needs.

On the other hand, the chimera was shown to possess strong cytotoxic activity. This is likely due to (i) the expression of the cytotoxic parvoviral NS1 protein, which is known to be especially deleterious to oncogene-transformed cells (24), and (ii) the generation of progeny PV particles, which spread in the tumor cell population and induce secondary rounds of lytic infection, amplifying the initial cytotoxic activity of the chimera.

In conclusion, this report adds a promising new tool to the arsenal of oncolytic viruses. Extension of the present in vitro investigation to animal tumor models is to be conducted in the future with the intention of producing a proof of concept that may justify the further clinical assessment of oncolytic virus chimeras.