Introduction

Epithelial ovarian cancer accounts for 90% of all ovarian cancer and is a leading cause of cancer death in the developed world. In the United Kingdom around 7,000 new cases are diagnosed annually and because patients normally present at a late stage, when the disease has spread throughout the abdominal cavity, the 5-year survival rate is only about 30%. Accordingly, epithelial ovarian cancer accounts for more deaths than all of the other gynaecological cancers combined.¹

Ovarian cancer cells usually remain confined to the peritoneal cavity even at late stages of the disease; hence intraperitoneal (IP) delivery provides the opportunity for aggressive local therapy whilst minimizing systemic exposure. Clinical trials have demonstrated a clear survival advantage for IP delivery of platinum-based chemotherapeutic agents,^2,3 confirming a regional advantage observed in animal pharmacokinetic studies.^4,5 IP delivery of gene therapy vectors is particularly attractive, since it facilitates selective transduction of tumor cells; however, promising preclinical results using serotype 5 adenovirus (Ad5) delivery have not been replicated in patients. A recent multicenter phase II/III clinical trial of IP E1-deleted Ad5 expressing wild-type p53 was halted at interim analysis due to inadequate therapeutic benefit and complications in the treatment arm.⁶ There are several factors that are responsible for this disappointing clinical outcome, most notably IP toxicities such as bowel adhesions and obstructions that were also noted in the phase I trial.⁷ In addition, ovarian ascitic fluid usually contains high levels of Ad5-neutralizing antibodies⁸ that may well compromize the efficiency of tumor cell infection. Finally, the coxsackie-adenovirus receptor (CAR) and integrins are expressed at only low levels in many epithelial ovarian cancers, with levels inversely related to tumor grade,^9,10potentially limiting Ad5 transduction efficiency. We therefore set out to develop an adenovirus delivery system capable of more efficient tumor cell infection, simultaneously aiming to avoid the dose-limiting toxicities that arise through indiscriminate infection of non-target cells.

Soluble polymers based on N-(2-hydroxypropyl)methacrylamide (HPMA) have been used as drug carriers in several clinical trials.¹¹ Amino-reactive copolymers of HPMA have now been developed that are able to coat the adenovirus capsid, shielding receptor-binding epitopes and providing steric protection from cell and antibody binding. Normal infection is thereby inhibited, and polymer coated adenovirus (pHPMA-adenovirus) administered intravenously displays complete loss of hepatic infection, extended plasma kinetics and decreased hepatotoxicity compared to the unmodified parental virus.¹² Since adenovirus infection appears to be tractable through a range of internalizing receptors, this tropism-ablated pHPMA-adenovirus provides a versatile platform for retargeting infection by covalent linkage of appropriate ligands. Early studies showed that infection can be retargeted using basic fibroblast growth factor, vascular endothelial growth factor,¹³ and oligopeptides such as SIGYPLP¹⁴ and SIKVAV,¹⁵ and that the presence of the polymer coating provides a degree of protection against neutralization by anti-adenovirus antibodies.^8,13

In selecting a strategy for retargeting pHPMA-adenovirus to infect epithelial ovarian cancer following IP administration, we sought a tumor upregulated (or tumor-specific) endosome-internalizing receptor capable of delivering retargeted pHPMA-adenovirus into its normal intracellular pathway of infection.^16,17 The epidermal growth factor receptor (EGFR or ErbB1) is a promising tumor-associated target, overexpressed (compared to normal cells) in more than 60% of ovarian cancers and associated with the malignant phenotype.¹⁸ Ligand binding stimulates endocytic internalization¹⁹ and previous retargeting strategies using adapter molecules, such as bispecific antibodies,^20,21 suggest that the receptor is permissive for adenovirus infection. Since tumor-selective retargeting of pHPMA-adenovirus has not been attempted before, EGFR represents an ideal candidate to assess the selectivity, infection efficiency and anticancer efficacy of this technology in the context of ovarian cancer.

Results

Conjugation of murine EGF to reactive HPMA copolymer

Reactive conjugates bearing murine epidermal growth factor (mEGF), for subsequent coating of adenovirus, were prepared by reacting the N-terminal -amino group of mEGF (in dimethyl sulfoxide) with HPMA copolymers containing reactive thiazolidone-2-thione groups (pHPMA-TT) under anhydrous conditions (Figure 1a). Amino acid analysis of the resulting mEGF-pHPMA-TT conjugate showed it retained 6.4 mol% of reactive TT groups, with an average 1.8 mEGF molecules per polymer chain, which accounted for 10% of the total mass. Virus size was evaluated by photon correlation spectroscopy (PCS) (Figure 1b). Unmodified virus had an average size of 118 nm, whereas mEGF-pHPMA-adenovirus had an average size of 133 nm, at a coating concentration of 10 mg/ml, and 135 nm, at a coating concentration of 20 mg/ml (data for 10 mg/ml shown). These data are consistent with previous published PCS and electron microscopy data of non-retargeted pHPMA-adenovirus¹³ and suggest that the virus remains as individual particles, rather than the polymer causing cross-linkage between particles.

Figure 1.

Poly-(2-hydroxypropyl)methacrylamide (pHPMA) production and physical properties of pHPMA-adenovirus. (a) Diagrammatic representation of mEGF-pHPMA synthesis. For further details see methods. (b) Photon correlation spectroscopy measurement of virus particle size, before and following pHPMA or mEGF-pPHMA coating. Mean unmodified adenovirus = 118 nm; pHPMA-adenovirus and mEGF-pHPMA-adenovirus coated at 10 mg/ml = 135 nm. Ad, unmodified adenovirus; p-Ad, pHPMA-adenovirus; mEGF-p-Ad, mEGF-pHPMA-adenovirus; mEGF, murine epidermal growth factor.

Full figure and legend (17K)

mEGF retains biological activity following conjugation to pHPMA

The ability of polymer-conjugated mEGF to bind cell surface EGFR was assessed by flow cytometry following labeling of the polymer conjugate with biotin ethylenediamine and removal of residual TT groups. Cell-association of the biotin modified polymer conjugate was visualized using extravidin-R-phycoerythrin. A significant increase in fluorescence was observed in EGFR-positive A431 cells, but not EGFR-negative K562 cells using mEGF-pHPMA-biotin, while the pHPMA-biotin control conjugate lacking mEGF failed to bind either cell type. This implies that mEGF retains EGFR-binding ability following conjugation to pHPMA (Figure 2a).

Figure 2.

Biological activity of epidermal growth factor (EGF) following conjugation with poly-(2-hydroxypropyl)methacrylamide (pHPMA). (a) EGF receptor (EGFR) negative (K562) and positive (A431) cells were incubated with biotinylated mEGF-pPHMA (solid histogram) and a biotinylated pHPMA control (broken line, open histogram) followed by labeling with extravidin phycoerythrin (PE) and analysis by flow cytometry (cells plus extravidin-PE shown with solid line, open histogram). Percentage of positive cells following incubation with mEGF-pHPMA and extravidin-PE are indicated. (b) Western blot of mEGF-pHPMA and murine EGF (mEGF) stimulation of EGFR autophosphorylation (Y1092). A431 cells were serum starved and stimulated for 30 minutes with mEGF or mEGF-pHPMA. Cell lysates were run on a 7.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis, blotted onto nitrocellulose paper, and developed with antibodies against phosphorylated EGFR. Concentrations of mEGF-pHPMA relate to the mEGF content only. (c) mEGF (top row) and mEGF-pHPMA (bottom row) were added to serum-starved A431 cells and EGFR labeled (green) in cells incubated for 0, 20, and 60 minutes to demonstrate the effect of mEGF and mEGF-pHPMA on EGFR internalization. Labeling of mEGF-pHPMA with Texas red (red) in the right hand panels demonstrates co-localization of EGFR and mEGF-pHPMA following internalization (yellow). Scale bars represent 10 m. These data are representative of at least three independent experiments.

Full figure and legend (69K)

To determine whether pHPMA-conjugated mEGF retained the ability to stimulate EGFR tyrosine kinase activity, EGFR positive cells (A431) were incubated with purified mEGF conjugate for 30 minutes. Cellular proteins were then separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis and the EGFR phosphorylation status was determined by means of Western blot analysis (Figure 2b). mEGF-pHPMA mediated phosphorylation of EGFR in a dose–dependent manner, with polymer-bound mEGF showing similar specific activity as unmodified mEGF.

Following stimulation by its ligand, EGFR normally internalizes into endosomes. To assess whether polymer-conjugated mEGF retained the ability to promote internalization, EGFR distribution was evaluated by immunofluorescence and confocal microscopy after exposure of cells to mEGF and mEGF-pHPMA. After 60 minutes of exposure, EGFR had moved from the cell surface into perinuclear areas following incubation with both mEGF and mEGF-pHPMA-biotin (Figure 2c), confirming the ability of the conjugated mEGF to promote receptor internalization. To assess whether the mEGF-polymer conjugate remained associated with the EGFR, the cellular distribution of mEGF-pHPMA-biotin was visualized using Avidin–Texas red. Although a considerable part of the conjugate remained extracellular, EGFR and mEGF-pHPMA-biotin were found to co-localize in perinuclear areas. Taken together, these data suggest that mEGF-pHPMA is able to bind selectively to EGFR-positive cells; that it can stimulate EGFR tyrosine kinase activity and internalization; and following internalization it remains at least partly in the same subcellular compartments as EGFR.

Use of mEGF-pHPMA to retarget adenovirus transduction via the EGFR

To assess whether retargeting to EGFR would provide a cellular uptake route enabling infection by polymer coated adenovirus, transgene expression of mEGF-pHPMA-adenovirus infected cells was compared in two CAR-negative cell types; an EGFR-negative murine fibroblast A9 cell line and a stably transduced EGFR-expressing variant (A9-EGFR, Figure 3a). Tropism-ablated pHPMA-adenovirus mediated less luciferase expression compared to the unmodified virus in both cell types. In contrast, coating the virus with mEGF-pHPMA gave much greater transduction than either the unmodified virus or non-retargeted pHPMA-virus in A9-EGFR cells but not in the parental EGFR-negative A9 cells (Figure 3b). Elevated transduction of A9-EGFR cells by mEGF-pHPMA-adenovirus could be blocked by preincubation with an antibody known to block EGF binding (Figure 3c), which was not seen with an isotype control antibody (data not shown). The same antibody had no significant effect on transduction of a CAR-expressing A9 cell variant (A9-CAR) where the unmodified virus achieved much greater transduction than the mEGF-pHPMA-adenovirus (data not shown). Taken together, these data demonstrate the EGFR-specificity of transduction by mEGF-pHPMA-adenovirus in A9-EGFR cells.

Figure 3.

Epidermal growth factor receptor (EGFR) status of cell lines used. Murine A9 fibroblast cells were stably transfected with a human EGFR containing plasmid, clones isolated and screened for EGFR expression. (a) Flow cytometry for EGFR status of murine A9 fibroblast cells transfected with human EGFR: A549 cells (positive control); parental A9 cells (negative control) and A9-EGFR clone. Broken line histograms represent isotype control antibody labeling and solid line histograms represent anti-EGFR antibody labeling. Percentage of positive cells following incubation with anti-EGFR antibody indicated for each cell line. Data are representative of repeated measurements over a 12-month period. (b) poly-(2-hydroxypropyl)methacrylamide (pHPMA) and mEGF-pHPMA (10 mg/ml) were used to coat AdCMV-LucE1 and transduction efficiency measured by a luciferase assay in the EGFR-positive, coxsackie-adenovirus receptor (CAR)-negative A9 cell line (right panel) (P = 0.0018), and EGFR-negative, CAR-negative parental A9 cells (left panel). (c) murine EGF (mEGF) binding to EGFR was blocked in A9-EGFR cells by preincubating cells for 30 minutes with an anti-EGFR antibody (GR01L). No anti-EGFR blockage (unfilled bars); preincubation with anti-EGFR antibody (black bars); (***P < 0.001). Data in b and c are representative of three independent experiments performed in triplicate. Bars represent mean + SD. Ad, unmodified adenovirus; p-Ad, pHPMA-adenovirus; mEGF-p-Ad, mEGF-pHPMA-adenovirus.

Full figure and legend (20K)

To determine whether EGFR could be used to target virus infection to tumor cells, transgene expression by mEGF-pHPMA-adenovirus was assessed in a range of cancer cell lines with varying expression of EGFR, mainly of ovarian origin (Figure 4a). Cellular EGFR expression was quantified by enzyme-linked immunosorbent assay (Figure 4b) and flow cytometry (data not shown). Whereas non-retargeted pHPMA-adenovirus achieved only low levels of transduction when assessed by luciferase expression, mEGF-pHPMA-adenovirus showed greater levels of transduction, proportional to cellular EGFR levels, than the unmodified virus. EGFR levels did not affect transduction by unmodified adenovirus (Figure 4c). Variation in the ability of pHPMA to inhibit infection may be related to the CAR and integrin expression in different cell lines, and the differential distribution of reactive amine groups (which are the target for pHPMA modification) across binding sites on the fiber and penton base.

Figure 4.

In vitro transduction efficiency of mEGF-pHPMA modified adenovirus (a) mEGF-pHPMA-adenovirus transduction of a range of epidermal growth factor (EGFR)-positive cell lines. Unmodified AdCMV-LucE1 (unfilled bars); poly-(2-hydroxypropyl)methacrylamide (pHPMA)-adenovirus (hashed bars); mEGF-pHPMA-adenovirus (black bars).For all cell lines the level of transgene expression following transduction with pHPMA-adenovirus was compared to mEGF-pHPMA-adenovirus (***P < 0.001). (b) Coxsackie-adenovirus receptor (CAR) and EGFR status of each cell line was assessed by flow cytometry (data not shown) and EGFR levels were formally quantified using a commercial enzyme-linked immunosorbent assay, which gave a similar pattern of results. (c) Levels of luciferase gene expression following transduction with unmodified AdCMV-Luc E1 and mEGF-pHPMA-adenovirus were compared with cellular EGFR levels. (d) Specific activity of mEGF-pHPMA-adenovirus in EGFR-positive cells (luciferase activity per virus particle, as measured by quantitative polymerase chain reaction (QPCR)) was assessed by transduction of A431 cells with AdCMV-LucE1. After 24 hours, cell lysate was analyzed for luciferase activity and DNA extracted for viral particle estimation by QPCR. Data are representative of at least two independent experiments performed in at least triplicate. Bars represent + SD. Ad, unmodified adenovirus; p-Ad, pHPMA-adenovirus; mEGF-p-Ad, mEGF-pHPMA-adenovirus. (e, f) Cytotoxicity analysis of mEGF-pHPMA-Ad5WT. pHPMA and mEGF-pHPMA (10 mg/ml) were used to coat Ad5WT. Known concentrations of adenovirus in plaque forming units (pfu) were incubated with A2780 cells (e) and SKOV-3 cells (f). After 6 days an MTS assay was performed to assess cell survival. Bars represent + SD. Dashed line with open squares, unmodified Ad5WT; dotted line with crosses, pHPMA-Ad5WT; solid line with filled diamonds, mEGF-pHPMA-Ad5WT. mEGF, murine EGF; RLU, relative light units.

Full figure and legend (24K)

We estimated the intracellular efficiency of EGFR-mediated infection in A431 cells by comparing levels of transgene expression with the number of virus genomes per cell (24 hours postinfection). The specific transduction activity of mEGF-retargeted virus (relative light units/genome) was the same as, or slightly greater than, unmodified virus infecting via its normal pathways (Figure 4d). This suggests that polymer-coated virus particles retargeted through the EGFR can uncoat efficiently and mediate the same level of transgene expression as particles entering via the normal infection pathways. To assess the ability of mEGF-retargeted virus to enter a lytic replication cycle, two ovarian cancer cell lines (SKOV-3 and A2780) were exposed to unmodified wild type adenovirus (Ad5WT), pHPMA-Ad5WT or mEGF-pHPMA-Ad5WT in a cell survival study (Figure 4e and f). mEGF-pHPMA-Ad5WT killed both ovarian cancer lines with similar or greater efficiency than unmodified Ad5WT, suggesting that mEGF-pHPMA was able to mediate a lytic replication cycle. Non-retargeted pHPMA-Ad5WT showed greatly decreased cell killing, correlating with the reduction in transgene expression already demonstrated. A replication incompetent reporter adenovirus (AdCMV-LucE1) did not mediate cell killing under these conditions (data not shown).

Taken together, all of these data indicate that mEGF-pHPMA is able to retarget adenovirus via the EGFR in vitro and that the majority of CAR-mediated uptake is inhibited.

Therapeutic efficacy of mEGF-pHPMA retargeted oncolytic adenovirus in a murine xenograft model

Adenovirus can be used in cancer therapy either to deliver a therapeutic transgene, as in the case of p53 replacement, or as a directly oncolytic replicating virus. Although genetic engineering has been used to increase cancer-selectivity,²² even wild-type adenovirus (Ad5WT) replicates more quickly in many cancer cells than in normal cells and has already shown clinical promise as an oncolytic "virotherapy" for cervical cancer.^23,24 We therefore used mEGF-pHPMA conjugates to retarget Ad5WT for IP delivery as an oncolytic virotherapy of human ovarian cancer in a murine xenograft model. Pilot data (not shown) were used in an a priori power calculation (G*Power software, Germany) to determine required group size and an adequately powered randomized and blinded survival study was then performed, with 12 mice/group. Mice were injected IP with luciferase-expressing SKOV-3 cells and were treated after 4 days with 5 10¹⁰ virus particles (v.p.) of Ad5WT, either unmodified, coated with pHPMA or mEGF-pHPMA, or a phosphate-buffered saline (PBS) control; treatment was repeated twice at 72 hour intervals. The activity of unmodified virus and EGF-pHPMA-adenovirus were compared at equal particle numbers, not equitoxic doses. In vivo bioluminescence was then measured weekly for 5 weeks to monitor tumor load. Importantly, in vivo bioluminescence enabled refinement of humane end-points, by identifying mice with high tumor loads before they began showing signs of illness, in addition to providing an on-going measure of tumor burden. Mice were sacrificed and post-mortem examination performed; adhesions were scored and tumors dissected. Only mice that had shown evidence of initial tumor establishment were included in the analysis.

mEGF-pHPMA-Ad5WT mediated a significantly longer median survival compared with pHPMA-Ad5WT or PBS control [34 days (PBS) versus 64.5 days (mEGF-pHPMA-Ad5WT); P < 0.0001; hazard ratio 0.19, 95% confidence interval 0.005–0.1] and [46 days (pHPMA-Ad5) versus 64.5 days (mEGF-pHPMA-Ad5WT); P = 0.0014; hazard ratio 0.3, 95% confidence interval 0.06–0.5] (Figure 5a). There was no significant difference between the PBS and pHPMA-Ad5WT groups, although most of the mice receiving pHPMA-Ad5WT did survive longer than controls. It is possible that residual infectivity of the pHPMA-Ad5WT mediates some activity in vivo, or even that pHPMA-Ad5WT entering cells by fluid phase endocytosis might mediate infection, as has recently been shown following intravenous injection.²⁵

Figure 5.

In vivo efficacy of mEGF-pHPMA retargeted wild-type Ad5. Female nude mice implanted with SKOV-luc cells intraperitoneally and treated with 3 doses of Ad5WT (5 10¹⁰ virus particles of unmodified Ad5WT [solid line, open inverted triangles); pHPMA-Ad5WT (dashed line, pink triangles); or mEGF-pHPMA-Ad5WT (dashed line, blue diamonds)] or a phosphate-buffered saline (PBS) control (solid line, closed squares). (a) Kaplan–Meier survival curves. (b) Bioluminescence measurements of tumor load 16 days after tumor inoculation. Representative merged luminescence images for each group are shown. Bars represent median group values and points are individual mouse luminescence. (**P < 0.01; ***P < 0.001; ns, not significant (P > 0.05)) (c) Adhesion scores for each group; mEGF-pHPMA-adenovirus treated mice had significantly fewer adhesions than mice treated with unmodified adenovirus by Chi-squared analysis (P = 0.00039). Bars represent the percentage of each group: no adhesions (open bars); mild adhesions (hashed bars); moderate adhesions (dotted bars); severe adhesions (black bars). (**P < 0.01, ***P < 0.001) (d) Representative images of mice from each group are shown. Ad, unmodified Ad5WT; p-Ad, pHPMA-Ad5WT; mEGF-p-Ad, mEGF-pHPMA-Ad5WT. Data are consistent with data from a randomized and blinded tumor load pilot study (data not shown), which were used in an a priori power calculation for this survival experiment. mEGF, murine epidermal growth factor; ns, not significant; pHPMA, N-(2-hydroxypropyl)methacrylamide.

Full figure and legend (41K)

Bioluminescence at 16 days postimplantation demonstrated a significantly lower tumor load for unmodified Ad5WT and mEGF-pHPMA-Ad5WT compared to either pHPMA-Ad5WT or PBS control treatments (Figure 5b); (a 150-fold lower median flux was found for mEGF-pHPMA-Ad5WT versus PBS, P < 0.001) (a 100-fold lower median flux was found for pHPMA-Ad5WT versus mEGF-pHPMA-Ad5WT, P < 0.05). There was a linear inverse relationship between mean group flux levels at 16 days postimplantation and median survival, indicating that bioluminescence, in this model, is a useful surrogate end-point (data not shown). It is also noteworthy that some animals treated with mEGF-pHPMA-Ad5WT were sacrificed because of subcutaneous tumor growth at the site of injection, and were free of macroscopic peritoneal disease at the time of death.

Mice treated with the unmodified Ad5WT were at increased risk of developing peritoneal adhesions and evidence of subacute bowel obstruction, sufficient to require early sacrifice in 25% of cases. This is an interesting parallel to the observed clinical toxicities and could reflect related processes. In contrast, mice treated with mEGF-pHPMA-Ad5WT showed markedly less adhesion formation, as determined by Chi-squared analysis (unmodified Ad5WT versus mEGF-pHPMA-Ad5WT; P = 0.00039) (Figure 5c) and none required early sacrifice through toxicity. It is possible that the polymer layer, or the changed virus tropism, prevents the inflammatory effects that lead to adhesion formation in the case of unmodified Ad5WT.

Discussion

Viruses offer several treatment modalities for cancer, ranging from supplementation of damaged tumor suppressor genes (such as p53) and immune stimulation, through to direct killing by virus replication. So-called "virotherapy" relies on the ability of lytic viruses to replicate and kill cancer cells while sparing normal tissues. The first clinical study of this approach used wild-type adenoviruses,²³ showing few side effects against normal tissue, although latterly attention has been focused on using genetically modified, often attenuated, virus strains.

Experimental IP treatments of clinical ovarian cancer have included both non-replicating and conditionally replicating Ad5, but all approaches have encountered significant issues of inflammation, leading to formation of peritoneal adhesions and bowel obstructions. These symptoms were observed in 14% of patients in an early phase IP clinical trial of a replication incompetent Ad5 expressing p53 (SCH 58500),⁷ and appear to have contributed to cessation of a large phase II/III study.⁶ In addition, abdominal pain and vomiting, symptoms often caused by peritoneal inflammation and adhesion formation, were common side effects in a phase I clinical trial of a conditionally replication-competent adenovirus (ONYX015), although no episodes of bowel obstruction were observed, probably reflecting the lower doses of Ad5 used in this trial (5 doses of 1 10¹¹ v.p. of ONYX015²⁶ c.f. 5 doses of 7.5 10¹³ v.p. of SCH 58500 ref. 7).

In this study we have shown that coating of Ad5 with mEGF-pHPMA can shield normal receptor-binding epitopes, inhibiting CAR-mediated infection and producing an infective virus retargeted via the EGFR. The specific cellular activity of the retargeted virus (defined as reporter gene activity per cell-associated virus genome) was the same as that of the unmodified virus, which showed that the polymer coating did not inhibit unpackaging and expression of the virus following cell entry. We suggest that the polymer coating is shed from the virus together with capsid proteins (to which the polymer remains covalently bonded), probably as a result of endosomal activation of the virus protease.²⁷ Although we have previously shown that reporter viruses retargeted in this way can mediate transgene expression, the observation here that polymer-mediated delivery of wild-type adenovirus causes cell lysis leading to virotherapy is the first demonstration that retargeted polymer-coated adenovirus can enter a full replication cycle.

Using an IP tumor model xenograft of human SKOV-3 ovarian cancer we found that IP treatment with unmodified wild-type Ad5 (3 doses of 5 10¹⁰ v.p.) leads to a significant frequency of peritoneal adhesions and subacute bowel obstructions, requiring euthanasia in about 20% of cases. Our preliminary studies (not shown) indicated the effect was also mediated by E1-deleted Ad5, and may be dose–dependent. No adhesion formation was reported in preclinical trials of ONYX015 (maximum 5 doses of 1 10⁹ v.p.²⁸). Use of the coating technology to shield normal infection pathways and retarget infection of Ad5 via EGFR (3 doses of 5 10¹⁰ v.p.) gave significant amelioration of all toxicities, including peritoneal adhesion formation and bowel obstruction. Anticancer efficacy was unchanged from that of the unmodified virus, suggesting that the polymer-coating approach may combine efficacy with improved toxicology, giving clinical advantages that could enable effective IP virotherapy of ovarian cancer. Another benefit of the decreased toxicity is the possible administration of higher Ad5 doses, potentially leading to further improved therapeutic outcomes.

The polymer-coating approach is not heritable and will only affect the biodistribution and infectivity profile of the input virus. Progeny replication-competent virus produced in situ would be expected to infect cells via their normal tropisms. This is an important safety feature, removing the possibility of creating a new pathology and also maximizing the likelihood of neutralization of the virus that spreads from the tumor site into the bloodstream or ascitic fluid. Given this scenario, the dramatic difference in inflammatory toxicities between unmodified and mEGF-pHPMA coated wild-type virus was not anticipated. The difference may be because much of the virus spread following primary infection is cell-to-cell, rather than transcoelomic, or possibly because the secondary burst of progeny virus is below a threshold required to stimulate severe adhesions. In a clinical setting, potential side effects could be further abrogated, if required, by the use of a conditionally-replicating adenovirus, although the increased selectivity often comes at the cost of decreased potency.

Improving the efficiency of gene delivery and avoiding the toxic side effects caused by delivery vectors is critical to the future success of gene- and virotherapy. EGFR is highly expressed on many cancer cells, underlying its value as a target for other anticancer therapies such as tyrosine kinase inhibitors²⁹ and monoclonal antibodies;³⁰ hence an EGFR-retargeted virotherapy may find a broad spectrum of applications. Other groups have previously reported strategies to retarget adenovirus via the EGFR, including the use of bifunctional antibodies^20,21 or engineering an adenovirus that secretes its own fiber-EGFR adapter molecule.³¹ However, none of these approaches efficiently de-targets normal adenovirus tropism in vivo, offers protection from neutralizing antibodies or significantly protects the adenovirus capsid from the innate immune system. In contrast, the complete de-targeting and well-defined retargeting achieved by this covalent capsid modification approach,¹² coupled with the possibility of shielding the adenovirus from the immune system and the absence of safety concerns arising from introducing new, heritable, virus tropisms, provides a feasible strategy that can enable the development of clinically useful agents.³²

Materials and Methods

All reagents, Chemicals and equipment were obtained from Fisher Scientific, UK or Sigma-Aldrich, Gillingham, UK unless otherwise stated.

Synthesis of pHPMA-TT. Copolymers of HPMA containing side chains bearing pendent reactive thiazolidone (TT) groups (pHPMA-TT) were synthesized as described by V.S. and K.U.³³ In short, HPMA and GG-TT were polymerized by free radical polymerization, initiated with AIBN in anhydrous dimethyl sulfoxide, 12.5 wt/vol%, followed by precipitation into 20 vol anhydrous acetone: ether (1:3). The average molecular weight was 114,000 g/mol and the copolymer contained 8.7 mol% TT-bearing monomers, determined by UV spectroscopy ( = 10280, = 305 nm).

Conjugation of mEGF to pHPMA-TT. mEGF (Serotec, UK) was mixed with 4 wt/wt monomethoxy poly(ethylene glycol) (MW = 2,000 g/mol) followed by lyophilization. Under anhydrous conditions the mixture was resolubilized, in anhydrous dimethyl sulfoxide and added to pHPMA-TT at a final concentration of 50 mg/ml polymer and 10 mg/ml of mEGF. After reaction for 2 hours the conjugate was isolated by precipitation into anhydrous acetone:ether (1:3) (20 vol). Following additional purification the degree of conjugation was determined using amino-acid analysis (OPT method). In brief, the samples were derivatized with OPT reagent (27 mg O-phthaldialdehyde, 500 l ethanol, 20 l 2-mercaptoethanol, 4.5 ml borate buffer - pH 9.3, 400 mmol/l) prior to analysis (HPLC (Shimadzu, UK) equipped with a Luna 5 m C18(2) column, 2 150 mm (Phenomenex, UK)), and eluted with an acetonitrile;phosphate buffer (10 mmol/l, pH 5.5) gradient. Amino acid standard curves were generated from derivatized amino acids using the above method.

Modification of pHPMA-TT polymers with biotin ethylenediamine. pHPMA-TT based polymer solution (5 mg/ml, 100 mmol/l HEPES, pH 7.8) was added with a tenfold excess of biotin ethylenediamine to amino-reactive groups and agitated overnight. The solution was purified on a PD10 Column (GE Healthcare, Little Chalfont, UK) and lyophilized, yielding a high performance liquid chromatography pure conjugate.

Coating adenovirus with pHPMA-TT. Polymers based on pHPMA-TT were dissolved in 18 M water at 100 mg/ml and diluted to give final polymer concentrations of 20 mg/ml or 5 mg/ml when added to virus. Virus stock in 40 l storage buffer (50 mmol/l HEPES, pH 7.8, 0.1 g/l Ca and MgCl₂, 10% Glycerol)¹³ was added to 10 l polymer and incubated at room temperature for 40 minutes. The virus was then transferred onto ice and incubated at 4 °C overnight.

Photon correlation spectroscopy. PCS measurements were performed on a Malvern Instruments, UK 3000 Hs equipped with a 50 mW laser at 532 nm. Samples for PCS were prepared using 5 l of virus in 600 l of 0.2 m filtered PBS, and data averaged over three runs (10 measurements per run).

Cell lines and culture. A9, SKOV-3, IGROV-1 and A2780 cell lines (ECACC, Salisbury, UK) were maintained in high glucose Dulbecco's modified Eagle's Media (DMEM) containing 2 mmol/l glutamine (PAA Laboratories, Yeovill, UK), supplemented with 1% penicillin–streptomycin solution (10,000 U/ml penicillin and 10 mg/ml streptomycin (Sigma-Aldrich, UK), 1% sodium pyruvate (Invitrogen, Paisley, UK) and 10% fetal bovine serum (FBS) (PAA Laboratories, Yeovill, UK). OVCAR-3 cells were maintained in Rosewell Park Memorial Institute 1640 media (PAA Laboratories, Yeovill, UK), supplemented with 1% penicillin–streptomycin solution and 10% (FBS) as before, with the addition of 1 g/ml bovine insulin (Sigma-Aldrich, Gillingham, UK) and 2 mmol/l glutamine. A9-CAR (a kind gift of Bob Newbold, Brunel University, UK) and A9-EGFR cell lines were maintained in DMEM, containing 2 mmol/l glutamine supplemented with 1% penicillin–streptomycin solution, 1% sodium pyruvate, 10% FBS and 800 g/ml G418 (Invitrogen, Paisley, UK). SKOV-luc cells (a kind gift of Iain McNeish, St Bartholomew's Hospital, London) were maintained in DMEM as above, but with 20% FBS and 500 g/ml G418.

mEGF-pHPMA cell binding, internalization and EGFR stimulation. For assessment of mEGF-pHPMA cell binding, cells were incubated with biotinylated polymer, and then extravidin-R-phycoerythrin (Sigma-Aldrich, Gillingham, UK) for 40 minutes at 4 °C, fixed with 4% paraformaldehyde/PBS and analyzed by flow cytometry (FACSCalibur, BD Biosciences, Oxford UK). Internalization of EGFR and mEGF-pHPMA was determined by stimulating serum-starved cells on chamber slides (VWR, Lutterworth, UK) with either 100 nmol/l mEGF (Serotec, UK) or mEGF-pPHMA for 2–60 minutes. Cells were fixed, permeabilized and labeled with mouse anti-EGFR (Santa Cruz Biotechnology, Santa Cruz, CA) and antimouse–fluorescein isothiocyanate (Vector, Laborateries, Peterborough, UK). Cells were mounted with Vectashield and nuclei counterstained with TO-PRO-3 (Invitrogen, Paisley, UK) prior to confocal microscopy (Nikon, Kingston, UK, Optiphot-2 upright microscope, fitted with a 60 oil immersion objective; Bio-Rad MRC1024 with Krypton/Argon lasers). EGFR phosphorylation was assayed by Western blotting. Cells were stimulated with 300 ng/ml mEGF or mEGF-pHPMA for 30 minutes and lysed (1 RIPA lysis buffer (Sigma-Aldrich, Gillingham, UK) and complete tablets (protease inhibitor cocktail; Roche Diagnostics, Burgess Hill, UK)). Equal amounts of protein (20 g) were loaded onto a 7.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis gel for electrophoresis, blotted onto nitrocellulose paper, and developed with antibodies (goat anti-pEGFR (Y1092) (Santa Cruz Biotechnology, Santa Cruz, CA); antigoat–horseradish peroxidase (Vector, Laborateries, Peterborough, UK); rabbit anti- actin (Sigma-Aldrich, Gillingham, UK); antirabbit–horseradish peroxidase (Abcam, Cambridge, UK)) and Lumiglo (Insight Biotechnologies, London, UK).

Production of stably transduced EGFR-positive, CAR-negative cell line and flow cytometry assessment of EGFR and CAR status. A9 cells (3 10⁵ cells/well) were seeded into 6-well plates (Corning, Schiphol-Rijk, The Netherlands). After 24 hours, the media was replaced with serum-free DMEM (PAA Laboratories, Yeovill, UK) and cells transfected with 2.5 g of plasmid pUSEampEGFR (Upstate, Lake Placid, NY) complexed with 1,2-dioleoyl-3-trimethylammonium-propane (1 mg/ml; Sigma-Aldrich, Gillingham, UK) (5:1 mass/mass ratio of 1,2-dioleoyl-3-trimethylammonium-propane to DNA) in HEPES buffer, at a final concentration of 10 mmol/l, for 4 hours at 37 °C in 5% CO₂. After 20 hours in normal growth media (DMEM with 10% FBS), cells were split and replated in media, supplemented with 800 g/ml G418. Colonies were grown up and positive clones screened by flow cytometry: primary antibodies, anti-EGFR mouse monoclonal (Santa Cruz Biotechnology, Santa Cruz, CA) and anti-CAR mouse monoclonal (a kind gift from Vivien Mautner, University of Birmingham, UK); secondary antibody, goat antimouse conjugated to R-phycoerythrin (Dako, Ely, UK). Cells were analyzed by flow cytometry at 585 nm.

Virus production, particle determination, and transduction. Wild-type Ad5 and a luciferase reporter adenovirus (AdCMV-LucE1) were amplified as previously described.¹³ Adenovirus particle number was determined by a modified version of the PicoGreen (Invitrogen, Paisely, UK) assay,^34,35 and infectious particles by plaque assay. Adenovirus infection and expression was determined using a luciferase reporter virus (AdCMV-LucE1). Cells were plated out 24 hours prior to transduction in DMEM with 10% FBS. Cells were washed with PBS and AdCMV-LucE1 (1,000 v.p./cell) added in DMEM supplemented with 2% FBS and incubated at 37 °C in 5% CO₂ for 90 minutes. Cells were washed and DMEM with 10% FBS added to wells. At 48 hours post-transduction the media were removed, cells washed and lysed with lysis buffer (Promega, Southampton, UK). Luciferin (Promega, Southampton, UK) were added to cell lysate and their luminescence was measured (Lumat LB 9507, Berthold Technologies, Redbourn, UK). EGFR blockade was performed by preincubating cells at 37 °C in 5% CO₂ for 30 minutes with an anti-EGFR antibody known to block the EGF binding site (GR01L; Calbiochem, Merck, Nottingham, UK) (10 g/ml in DMEM supplemented with 2% FBS).

Assessment of cellular EGFR status. A commercial enzyme-linked immunosorbent assay (Dako, Ely, UK) was used to assess cellular EGFR status, according to the manufacturer's specifications. Concentrations of EGFR in samples, from known numbers of cells, were measured and the number of EGFR molecules calculated assuming an EGFR molecular weight of 110 kd.

Evaluation of specific activity of adenovirus. A431 cells (1 10⁴) were transduced with AdCMV-Luc E1 (1,000 v.p./cell) as described above. After 24 hours, cells were lysed and assayed for luciferase expression. Aliquots of cell lysate were taken for DNA extraction (Genelute DNA mini-prep; Sigma-Aldrich, Gillingham, UK) and adenovirus copy number assessed by quantitative polymerase chain reaction as previously described.¹²

Assessment of oncolytic activity of viruses. SKOV-3 and A2780 cells (1 10⁴) were plated out 24 hours prior to transduction in DMEM with 10% FBS. Next day, cells were infected with Ad5WT at the range of a multiplicity of infection of plaque forming units per cell in low-serum media (DMEM with 2% FCS). Cells and virus were incubated at 37 °C, 5% CO₂ for 90 minutes. After this time the virus media was removed, cells were rinsed PBS, and low-serum media was added to each well and incubated at 37 °C, 5% CO₂ for 6 days. Cells were fed with fresh low-serum media every 2–3 days. After 6 days, cell viability was assessed by the MTS assay (Cell Titer 96^® Aqueous Non-Radioactive Cell Proliferation Assay (MTS) Promega, Southampton, UK) and percentage survival compared to non-infected control cells. Sigmoid dose–response curves were generated using Prism 4 software (GraphPad Software, San Diego, CA) in order to calculate the IC₅₀ values.

In vivo model and bioluminescence. All procedures were performed humanely under license according to UK Home Office regulations using appropriate analgesia and anesthesia. Nude mice (6–8 weeks old), were housed in environmentally enriched, individually ventilated cages, were implanted by IP injection with 1 10⁷ SKOV-luc cells. After 96 hours mice were treated with 5 10¹⁰ v.p. Ad5WT (unmodified, pHPMA-adenovirus, mEGF-pHPMA-adenovirus) and a PBS control by IP injection and treatments repeated twice at 72 hour intervals. Previous studies have demonstrated that a non-replicating adenovirus or a UV-inactivated oncolytic adenovirus have no beneficial effect on survival in an IP ovarian cancer tumor model and were therefore not included as additional negative controls.^22,28 Treatment vials were randomized and blinded independently, so that investigators were blinded to treatment groups. Mice were imaged weekly five times using a Xenogen IVIS 100 imaging system (Xenogen, MA) as previously described²² and imaged supine and prone for 1 minute (f/stop 1, medium binning). Exposure times for saturated images were reduced accordingly. Flux levels were analyzed with Living Image Software (Xenogen, MA). At post-mortem examination, mice were scored for adhesion formation: none; mild (few filmy adhesions); moderate (more extensive adhesions); severe (extensive adhesions with evidence of bowel matting, distension and subacute obstruction).

Statistical analysis. Virus transduction data were log transformed and analyzed by analysis of variance using Prism 4 software (GraphPad, San Diego, CA), with Bonferroni and Dunnet corrections for multiple analyses as required. Non-parametric data was analyzed by Kruskal–Wallis analysis with Dunn's multiple comparison test. Kaplan–Meier survival analysis was performed using Prism4 software (GraphPad, San Diego, CA). Chi-square analysis was used to compare adhesion scores using Excel software (Microsoft, Redmond, WA).

References

REFERENCES

1. CRUK(2007). UK Ovarian cancer statistics, 2002. Cancer Research UK.
2. Alberts, DS, Liu PY, Hannigan EV, O'Toole R, Williams SD, Young, JA et al. (1996). Intraperitoneal cisplatin plus intravenous cyclophosphamide versus intravenous cisplatin plus intravenous cyclophosphamide for stage III ovarian cancer. N Engl J Med 335:1950–1955. | Article | PubMed | ISI | ChemPort |
3. Armstrong, DK, Bundy, B, Wenzel, L, Huang, HQ, Baergen, R, Lele, S et al. (2006). Intraperitoneal cisplatin and paclitaxel in ovarian cancer. N Engl J Med 354:34–43. | Article | PubMed | ISI | ChemPort |
4. DeGregorio mW, Lum BL, Holleran WM, Wilbur BJ and Sikic, BI (1986). Preliminary observations of intraperitoneal carboplatin pharmacokinetics during a phase I study of the Northern California Oncology Group. Cancer Chemother Pharmacol 18:235–238. | Article | PubMed | ChemPort |
5. Pretorius RG,Petrilli ES, Kean CK, Ford LC, Hoeschele JD and Lagasse, LD(1981). Comparison of the iv and ip routes of administration of cisplatin in dogs. Cancer Treat Rep 65:1055–1062. | PubMed | ChemPort |
6. Zeimet AG and Marth, C (2003). Why did p53 gene therapy fail in ovarian cancer? Lancet Oncol 4:415–422. | Article | PubMed | ISI | ChemPort |
7. Buller RE,Runnebaum IB, Karlan BY,Horowitz JA, Shahin M, Buekers, T et al. (2002). A phase I/II trial of rAd/p53 (SCH 58500) gene replacement in recurrent ovarian cancer. Cancer Gene Ther 9:553–566. | Article | PubMed | ISI | ChemPort |
8. Stallwood, Y, Fisher, KD, Gallimore, PH and Mautner, V (2000). Neutralisation of adenovirus infectivity by ascitic fluid from ovarian cancer patients. Gene Ther 7:637–643. | Article | PubMed | ChemPort |
9. Zeimet, AG, Muller-Holzner, E, Schuler, A, Hartung, G, Berger, J, Hermann, M et al. (2002). Determination of molecules regulating gene delivery using adenoviral vectors in ovarian carcinomas. Gene Ther 9:1093–1100. | Article | PubMed | ChemPort |
10. Kelly FJ, Miller CR, Buchsbaum DJ, Gomez-Navarro J, Barnes MN, Alvarez, RD et al. (2000). Selectivity of TAG-72-targeted adenovirus gene transfer to primary ovarian carcinoma cells versus autologous mesothelial cells in vitro. Clin Cancer Res 6:4323–4333. | PubMed | ISI | ChemPort |
11. Thomson, AH, Vasey, PA, Murray, LS, Cassidy, J, Fraier, D, Frigerio, E et al. (1999). Population pharmacokinetics in phase I drug development: a phase I study of PK1 in patients with solid tumours. Br J Cancer 81:99–107. | Article | PubMed | ChemPort |
12. Green NK, Herbert CW, Hale SJ, Hale AB, Mautner V, Harkins, R et al. (2004). Extended plasma circulation time and decreased toxicity of polymer-coated adenovirus. Gene Ther 11:1256–1263. | Article | PubMed | ISI | ChemPort |
13. Fisher, KD, Stallwood, Y, Green, NK, Ulbrich, K, Mautner, V and Seymour LW (2001). Polymer-coated adenovirus permits efficient retargeting and evades neutralising antibodies. Gene Ther 8: 341–348. | Article | PubMed | ISI | ChemPort |
14. Parker, AL, Fisher, kd, Oupicky, D, Read, ML, Nicklin, SA, Baker, AH et al. (2005). Enhanced gene transfer activity of peptide-targeted gene-delivery vectors. J Drug Target 13: 39–51. | Article | PubMed | ChemPort |
15. Stevenson M, Hale AB, Hale SJ, Green NK, Black G, Fisher, KD et al. (2007). Incorporation of a laminin-derived peptide (SIKVAV) on polymer-modified adenovirus permits tumor-specific targeting via alpha6-integrins. Cancer Gene Ther 14:335–345. | Article | PubMed | ISI | ChemPort |
16. Meier O and Greber UF (2004). Adenovirus endocytosis. J Gene Med 6(suppl. 1):S152–S163. | Article |
17. Wiethoff, CM, Wodrich, H, Gerace, L and Nemerow, GR (2005). Adenovirus protein VI mediates membrane disruption following capsid disassembly.J Virol 79:1992–2000. | Article | PubMed | ISI | ChemPort |
18. Alper, O, Bergmann-Leitner, ES, Bennett, TA, Hacker, NF, Stromberg, K and Stetler-Stevenson WG (2001). Epidermal growth factor receptor signaling and the invasive phenotype of ovarian carcinoma cells. J Natl Cancer Inst 93: 1375–1384. | Article | PubMed | ChemPort |
19. Carter, RE and Sorkin A (1998). Endocytosis of functional epidermal growth factor receptor-green fluorescent protein chimera. J Biol Chem 273:35000–35007. | Article | PubMed | ISI | ChemPort |
20. Blackwell, JL, Miller, CR, Douglas, JT, Li, H, Reynolds, PN, Carroll, WR et al. (1999). Retargeting to EGFR enhances adenovirus infection efficiency of squamous cell carcinoma. Arch Otolaryngol Head Neck Surg 125:856–863. | PubMed | ISI | ChemPort |
21. Grill J, Van Beusechem VW, Van Der Valk P, Dirven CM, Leonhart A, Pherai, DS et al. (2001). Combined targeting of adenoviruses to integrins and epidermal growth factor receptors increases gene transfer into primary glioma cells and spheroids. Clin Cancer Res 7:641–650. | PubMed | ISI | ChemPort |
22. Lockley, M, Fernandez, M, Wang, Y, Li, NF, Conroy, S, Lemoine, N et al. (2006). Activity of the adenoviral E1A deletion mutant dl922-947 in ovarian cancer: comparison with E1A wild-type viruses, bioluminescence monitoring, and intraperitoneal delivery in icodextrin. Cancer Res 66:989–998. | Article | PubMed | ChemPort |
23. Smith RR, Huebner RJ, Rowe WP, Schatten WE and Thomas,LB (1956). Studies on the use of viruses in the treatment of carcinoma of the cervix. Cancer 9:1211–1218. | Article | PubMed | ISI |
24. Georgiades J, Zielinski T, Cicholska A and Jordan, E (1959). Research on the oncolytic effect of APC viruses in cancer of the cervix uteri; preliminary report. Biul Inst Med Morsk Gdansk 10:49–57. | PubMed | ChemPort |
25. Fisher KD, Green NK, Hale A, Subr V, Ulbrich K and Seymour, LW (2007). Passive tumour targeting of polymer-coated adenovirus for cancer gene therapy. J Drug Target 15:546–551. | Article | PubMed | ChemPort |
26. Vasey PA, Shulman LN, Campos S, Davis J, Gore M, Johnston, S et al. (2002). Phase I trial of intraperitoneal injection of the E1B-55-kd-gene-deleted adenovirus ONYX-015 (dl1520) given on days 1 through 5 every 3 weeks in patients with recurrent/refractory epithelial ovarian cancer. J Clin Oncol 20:1562–1569. | Article | PubMed | ISI | ChemPort |
27. Greber, UF, Webster, P, Weber, J andHelenius A (1996). The role of the adenovirus protease on virus entry into cells. EMBO J 15:1766–1777. | PubMed | ISI | ChemPort |
28. Heise, C, Ganly, I, Kim, YT, Sampson-Johannes, A, Brown, R and Kirn D (2000). Efficacy of a replication-selective adenovirus against ovarian carcinomatosis is dependent on tumor burden, viral replication and p53 status. Gene Ther 7:1925–1929. | Article | PubMed | ChemPort |
29. Baselga J, Rischin D, Ranson M, Calvert H, Raymond E, Kieback, DG et al. (2002). Phase I safety, pharmacokinetic, and pharmacodynamic trial of ZD1839, a selective oral epidermal growth factor receptor tyrosine kinase inhibitor, in patients with five selected solid tumor types. J Clin Oncol 20:4292–4302. | Article | PubMed | ISI | ChemPort |
30. Xiong, HQ, Rosenberg, A, LoBuglio, A, Schmidt, W, Wolff, RA, Deutsch, J et al. (2004). Cetuximab, a monoclonal antibody targeting the epidermal growth factor receptor, in combination with gemcitabine for advanced pancreatic cancer: a multicenter phase II Trial. J Clin Oncol 22:2610–2616. | Article | PubMed | ISI | ChemPort |
31. Hemminki A, Dmitriev I, Liu B, Desmond RA, Alemany R and Curiel, DT (2001). Targeting oncolytic adenoviral agents to the epidermal growth factor pathway with a secretory fusion molecule. Cancer Res 61:6377–6381. | PubMed | ISI | ChemPort |
32. De Geest B, Snoeys J, Van Linthout S, Lievens J and Collen, D (2005). Elimination of innate immune responses and liver inflammation by PEGylation of adenoviral vectors and methylprednisolone. Hum Gene Ther 16:1439–1451. | Article | PubMed | ISI | ChemPort |
33. Subr V and Ulbrich, K (2006). Synthesis and properties of new N-(2-hydroxypropyl)-methacylamide copolymers containing thiazolidine-2-thione reactive group. Reactive and Functional Polymers 66:1525–1538. | Article | ChemPort |
34. Singer, VL, Jones, LJ, Yue, ST and Haugland, RP (1997). Characterization of PicoGreen reagent and development of a fluorescence-based solution assay for double-stranded DNA quantitation. Anal Biochem 249:228–238. | Article | PubMed | ISI | ChemPort |
35. Mittereder, N, March, KL and Trapnell, BC (1996). Evaluation of the concentration and bioactivity of adenovirus vectors for gene therapy. J Virol 70:7498–7509. | PubMed | ISI | ChemPort |

Acknowledgments

This work was carried out in The University of Oxford, Oxon, UK and supported by the Oxfordshire Health Services Research Council, WellBeing of Women, Cancer Research UK and Macmillan Cancer Relief. We are grateful to Vivien Mautner (University of Birmingham, UK) for the generous donation of adenoviruses, Iain McNeish (St. Bartholomew's Hospital, London) for the SKOV-luc cell line, Bob Newbold (Brunel University, UK) for the A9-CAR cell line and to Andrew Jefferson and Nick Alp, Wellcome Trust Centre for Human Genetics, Oxford, for assistance with confocal microscopy.