Neutralisation of adenovirus infectivity by ascitic fluid from ... - Nature

Gene Therapy (2000) 7, 637–643  2000 Macmillan Publishers Ltd All rights reserved 0969-7128/00 $15.00 www.nature.com/gt

VIRAL TRANSFER TECHNOLOGY

RESEARCH ARTICLE

Neutralisation of adenovirus infectivity by ascitic fluid from ovarian cancer patients Y Stallwood, KD Fisher, PH Gallimore and V Mautner CRC Institute for Cancer Studies, University of Birmingham, Birmingham B15 2TA, UK

Animal models and phase I clinical trials have shown that repeat virus delivery and subsequent transgene expression is limited by the generation of humoral and cellular immune responses directed towards the therapeutic vector. The presence of a pre-existing immune response may even prevent initial delivery. In order to determine the presence of pre-existing anti-adenovirus humoral immunity we analysed ascitic fluid, collected from the peritoneal cavity of patients with advanced ovarian cancer. Twelve ascitic fluid and four matched serum samples were examined. The titre and isotype of anti-adenovirus antibodies was determined by

ELISA, and Western blotting identified the molecular basis of the immune response, which was primarily directed towards fibre and penton base. Neutralisation of virus infectivity was assessed in vitro by measurement of green fluorescent protein reporter gene expression. We found that the ascitic fluid samples contain antibodies that recognise both adenovirus types 2 and 5, were predominantly IgG and directed towards the viral antigens responsible for cell adhesion, and had virus neutralising activity. Gene Therapy (2000) 7, 637–643.

Keywords: adenovirus; antibodies; ascitic fluid; isotype; neutralisation

Introduction Adenovirus is a powerful tool for the treatment of cancer by gene therapy, in particular because of its ability to infect a broad range of tissues including both dividing and non-dividing cells.1,2 These properties can be further exploited by utilising regional delivery of adenovirus vectors in order to ensure high efficiency of gene delivery to the target tumour. Ovarian cancer remains confined to the peritoneal cavity for much of its history and as such offers an opportunity for local delivery of a therapeutic recombinant adenovirus vector via the intraperitoneal route. Ovarian carcinoma which has spread to the peritoneum can be accompanied by the formation of large amounts of ascitic fluid which is known to be associated with adverse outcome and is thought to arise by leakage from the circulation and lymphatic system, although the precise mechanism of ascites production is not well understood.3–5 The composition of ascites can vary, ranging from an acellular fluid with a protein content similar to serum, to a fluid with large numbers of tumour cells and blood cells. Due to the relationship of ascites to serum, the possibility exists that prior exposure to adenovirus may lead to the presence of neutralising antibodies in ascites, which could compromise the delivery of therapeutic adenovirus vectors by this route. Because the peritoneum is not a natural target for adenovirus infection, the immune status of peritoneal fluid has not been extensively studied. This study was undertaken to examine the anti-adenovirus antibody status of patients

Correspondence: V Mautner Received 31 August 1999; accepted 22 December 1999

with ascites arising from ovarian cancer, in order to anticipate the likely efficacy of this delivery route for adenovirus-mediated gene therapy. Adenovirus is a common human pathogen: antibodies to serotypes 1, 2 and 5 are present in 40–60% of children, and adults are usually immune to infection with these serotypes.6 Ad5 has been the vector of choice for adenovirus-mediated gene therapy, in part because of the availability of suitable mutants and vector constructs, and also because its molecular properties have been extensively characterised. However, the prevalence of adenovirus infection and resultant immunity is a potential limitation to the success of adenovirus-mediated gene therapy. Within the ascitic fluid in the peritoneal cavity the presence of pre-existing antibodies directed towards the input adenovirus vector could block the successful first delivery of a therapeutic gene and compromise therapy. The human adenoviruses are classified into six subgroups based on their haemagglutination properties7 and within each subgroup, individual serotypes are defined by the characteristics of neutralisation of the virus infection. Neutralisation as defined in vitro is determined by epitopes present on the external variable domains of the virion proteins hexon and fibre.6,8,9 In order to investigate the presence of pre-existing antiadenovirus antibodies in ascitic fluid and their possible effect upon adenovirus infection, ascitic fluid obtained from ovarian cancer patients during routine peritoneal drainage was examined for the presence of antibodies recognising adenovirus from subgroups B and C. In addition, the isotype was determined to indicate the status of the immune response. The adenovirus antigens recognised by antibodies present in the ascitic fluid were

Adenovirus neutralisation by ascitic fluid Y Stallwood et al

638

identified by Western blot and ELISA. The effect of ascites on the infectivity of adenovirus types 5 and 3 was examined in vitro using a neutralisation assay.

Results Detection of anti-adenovirus antibodies in ascitic fluid Twelve cell-free ascitic fluid samples (1asc–12asc), four of which had matched serum samples (1ser–-4ser), were examined for the presence of anti-adenovirus antibodies. To detect the presence of antibodies recognising adenovirus serotypes 2 and 5 (Ad2 and Ad5), an enzyme linked immunosorbent assay (ELISA) was developed using heat-inactivated Ad2 or Ad5 purified virions as antigen; bound antibodies were detected with antihuman Ig peroxidase conjugate. All the samples contained antibodies which recognised Ad2 and Ad5, but the binding characteristics varied: Figure 1a shows profiles

for eight ascitic samples, and a hyperimmune rabbit antiserum raised against Ad5wt virus (R1/89). While samples 5asc and 12asc showed a clear difference in recognition of Ad2 and Ad5, and conventional sigmoid binding curves, the other samples showed much less discrimination between Ad2 and Ad5, and displayed rather shallow profiles. Figure 1b compares the ELISA profiles for the matched sera and ascites; overall there is a close similarity within each pair, in terms of the shape of the curve, and the response to Ad2. For example, ascites 4asc and serum 4ser show conventional binding profiles, and recognise Ad2 better than Ad5, whereas 2asc and 2ser show a minimal response, and hardly distinguish between the two serotypes. Because of these variations, it was difficult to identify a precise 50% maximal end point titre; instead, the reciprocal of the highest dilution giving an OD490 reading of 0.5 was selected as a measure of titre. This is summarised in Figure 2, which also shows the median values for each sample group. The titre of R1/89 is 64 for

Figure 1 ELISA for anti-adenovirus antibodies in ascitic fluid. Plates coated with heat-inactivated Ad2 or Ad5 virus (20 ng/well) were incubated with 100 ␮l of serial two-fold dilutions of ascitic fluid or serum initially diluted 10−2 in 0.1% Tween in PBS. After incubation at 20°C for 2 h, bound antibodies were detected using peroxidase conjugated anti-human Ig (whole molecule) and absorbance measured at 490 nm. 䊏, Ad2 and 䊉, Ad5. (a) Unmatched ascites and control rabbit serum R1/89; (b) matched ascites and serum. Results are the mean of triplicate experiments. Gene Therapy


639

Figure 2 Anti-adenovirus antibody titre. Titres were determined from the binding profiles shown in Figure 1. The reciprocal of the highest dilution giving an OD490 reading of 0.5 was chosen arbitrarily as a measure of titre. Titres of antibodies in ascitic fluid and serum are plotted separately for Ad2 and Ad5 virus. 䊏, ascites versus Ad2; 왔, serum versus Ad2; 왖, ascites versus Ad5; and 쏆, serum versus Ad5. Matched ascites and serum are identified by their sample number. Median values are denoted by a solid black horizontal line. The titre of rabbit hyperimmune serum is represented by *.

Ad5, and two of the patient sera and one matched ascites equal or exceed this value. In general the matched pairs had similar titres, with the notable exception of sample 1ser, where the serum had the highest titre (51200 versus Ad2), 30-fold greater than the matched ascites.

Isotype characterisation of anti-adenovirus antibodies The samples were further characterised by ELISA to determine the predominant isotype of antibody present. Samples contained detectable levels of IgA, IgM and IgG when tested against both Ad2 and Ad5. Figure 3 shows the titration of matched ascites and serum 1asc/1ser and ascites 4asc on Ad5 virus coated plates. The profile for 4asc, which is typical of the majority of ascites and sera examined, shows a relatively high level of IgG compared to IgM and IgA. Serum sample 1ser is unique in having an enhanced level of IgM; this serum had the highest overall titre against Ad2 and Ad5; it should be noted that the ascites had a 32-fold lower overall Ig titre and did not contain similar high levels of IgM. Characterisation of adenovirus antigens recognised by antibodies The ascites and serum samples were used in Western blotting to determine which adenovirus capsid antigens they recognised. For Ad2, the major virus structural proteins hexon, penton base and fibre were recognised by all the samples tested. Recognition of fibre and penton base was much stronger than that of hexon, although all the proteins transferred with equal efficiency. Figure 4 demonstrates the pattern of Ad2 and Ad5 antigen recognition seen with ascites 4asc and serum 4ser. It is clear that both ascites and serum give a stronger signal with the Ad2 proteins, concordant with the ELISA pattern. A similar correlation of Western blot recognition and ELISA titre was seen for all the samples (data not shown).

Figure 3 Isotype characterisation of antibodies present in ascitic fluid and serum. Two-fold dilutions of ascitic fluid or serum initially diluted 10−2 in PBS/Tween were added to 96-well plates coated with heat-inactivated purified Ad2 or Ad5 virus (20 ng/well). Anti-human IgA, IgG and IgM secondary antibodies were added to the plates (see Materials and methods) and following addition of OPD the reaction was measured by colorimetry at OD 490nm. (a) Ascites 4asc; (b) ascites 1asc and (c) serum 1ser. 왖, IgG; 䊏, IgA and 왔, IgM. Results are the mean of triplicate experiments.

Neutralisation of adenovirus infection To examine the effect of antibodies present in ascitic fluid or serum on the infectivity of adenovirus in vitro, a neutralisation test based on a rapid assay for expression of green fluorescent protein (GFP) was developed. An E1deleted replication-defective recombinant adenovirus expressing GFP was used to infect A549 cells, and virus neutralisation assessed by the reduction in GFP fluorescence obtained at 72 h after infection in the presence of ascites or serum. Virus was used at 104 particles per cell, a titre known to give a signal in the linear range of fluorescence. All the samples inhibited GFP expression at 72 Gene Therapy


640

Figure 4 Recognition of adenovirus antigens by antibodies present in serum and ascitic fluid. Dilutions of purified Ad2 and Ad5 virus were subject to SDS-PAGE. Following semi-dry blotting and blocking in non-fat milk, membranes were incubated with (a) serum 4ser and (b) ascites 4asc diluted 10−3 in PBS/Tween. The proteins indicated are the major virus structural proteins hexon (120 kDa), penton base (87 kDa) and fibre (62 kDa).

h after infection (examples are shown in Figure 5). The dilution of ascites at which 50% neutralisation was seen varied between samples. Most ascites showed 50% inhibition of virus infection when diluted by a factor of 102– 103 (as in Figure 5a). In contrast, matched pair 4asc/4ser had to be diluted 104 to obtain 50% inhibition of infectivity (Figure 5b). Serotype specificity of neutralisation was confirmed using rabbit anti-Ad2 and anti-Ad5 sera (Figure 5b): Complete inhibition of virus infectivity was achieved at a dilution of 10−2 for the anti-Ad5 serum, and greater than 50% inhibition was obtained with a dilution of 10−4. In contrast, the Ad2 serum blocked only 20% of infectivity at 10−2 and failed to neutralise at 10−4. To ensure that complement mediated effects played no part in inhibition of virus infection, samples were subject to heat-inactivation at 56°C for 30 min. Heat-inactivated ascitic fluid retained the neutralising capacity of untreated material (Figure 6a). To show that inhibition of virus infection was caused by neutralising antibodies, samples were depleted of antibodies by treatment with protein A conjugated to Sepharose beads. After treatment with protein A, the ascitic sample had lost 80% of its neutralising activity (Figure 6a). Neutralisation of Ad3, a subgroup B adenovirus, by ascites and serum was examined using an assay based on viral cytopathic effect (cpe). Complete neutralisation was seen with ascites 4asc and serum 4ser in A549 cells infected with 102 particles per cell; ascites neutralised at a dilution of 10−2, whereas serum was able to neutralise when diluted 10−5. A full cpe was seen with all samples when using virus titres of 103–105 particles per cell. Thus the samples were unable to neutralise Ad3 at comparable levels to Ad5, where neutralisation was seen when cells were infected with 104 particles per cell.

Discussion The use of recombinant adenovirus vectors in a number of gene therapy phase 1 trials has suggested that its success may ultimately be limited by the generation of humoral and cellular immune responses.10–14 Cytotoxic T cells can destroy virus-infected cells10–12 whereas neutralising antibodies may prevent successful virus infection and, should repeated delivery be required, delivery of the transgene could be blocked and therapy comproGene Therapy

Figure 5 Neutralisation of adenovirus infectivity by ascites and serum. A549 cells were infected with AdCMV-GFP⌬E1 at an MOI of 104 particles per cell, in the presence of serial 10-fold dilutions of serum or ascites, and the fluorescence emission at 538 nm recorded 72 h after infection. Titres are expressed as the % gfp fluorescence relative to the total recorded in the absence of serum or ascitic fluid. (a) Ascitic samples with low neutralisation titres; (b) matched serum and ascites 4ser and 4asc, anti-Ad5 serum R1/89 and anti-Ad2 serum R2/73. Results are the mean of triplicate experiments.


Figure 6 Neutralisation of virus infection is due to antibodies in ascites. A549 cells were incubated with AdCMV-GFP⌬E1 virus plus ascites and the relative fluorescence intensity of the expressed gfp was measured at 72 h after infection. Antibody-depleted ascites were prepared by preincubating a 50 ␮l aliquot of ascites 4asc with three sequential aliquots of protein A Sepharose beads at 20°C over a 3 h period. Heat treated ascites were preincubated at 56°C for 30 min before use in neutralisation assays.

mised. Previous exposure to adenovirus may block even the first delivery of recombinant adenovirus, although this was not observed in a clinical trial of intrapleural virus delivery for localised mesothelioma.15 Our study examines ascitic fluid from the peritoneal cavity for the presence of antibodies which could prevent or modulate the local delivery of an adenovirus vector for the treatment of ovarian cancer. The samples came from patients with advanced ovarian cancer, who were producing an extensive ascites. ELISA assays using Ad2 and Ad5 virus as antigen showed that all samples examined contained antibodies recognising both serotypes, which presumably is a measure of the presence of group-specific epitopes found on all human adenoviruses.7 Where there are marked differences in the titres for Ad2 and Ad5, for example ascites 5asc and 12asc, it is likely that this reflects different levels of type-specific recognition. Whether this correlates with neutralising activity remains to be confirmed. There is no clear correlation between the overall antibody titre as measured by ELISA, and the neutralising activity, when all the samples are taken into consideration. This is consistent with the recent report of high IgG levels to Ad5 in a majority of normal subjects, but only 55% of IgG positive sera capable of neutralisation.16 The ability of our samples to neutralise Ad5 and, to a lesser extent, Ad3 infectivity suggests that either there has been exposure to multiple serotypes, or that neutralisation with human sera and ascites is less type-specific than has previously been appreciated. The presence in the ascitic fluid of antibodies to the respiratory viruses Ad2 and Ad5 is most unlikely to have arisen there as the result of peritoneal infection, and is probably the result of leakage from the local microvasculature supporting the tumour and microvessels lining the peritoneal cavity.3–5 As a large percentage of the population are infected with adenovirus during childhood,6 it was predicted that the predominant isotype would be IgG and this was the

case for all the ascitic samples tested, indicating a secondary immune response. IgA and IgM were also detectable, albeit at much lower levels, except in one serum (1ser) where the IgM level was elevated, signifying that this patient had a very recent adenovirus infection, such that a primary immune response was detected in the serum. The lower level of IgM in the matching ascites 1asc suggests it may not yet have passed out of the systemic circulation into the ascitic fluid. There is limited evidence regarding the relationship between the site of infection or virus administration and antibody response: a study of BAL fluid from patients with a range of lung diseases17 found pre-existing IgA, as well as IgG and IgM, and BAL samples capable of neutralising virus infectivity were all IgA positive (although the level of IgA did not correlate with the neutralising ability). In a separate study of mesothelioma patients, IgG and IgM were detected in serum (IgA was not reported) following intrapleural recombinant adenovirus delivery, and comparable levels of neutralising antibody were found in serum and pleural fluid.15 A recent publication where adenovirus was administered to skin, airway epithelium, metastatic liver tumours and heart muscle showed that serum levels of neutralising antibody depended on the route of infection and the level of preexisting serum antibody, but not on the virus dose.18 Western blotting, allowing recognition of individual virus capsid proteins, indicated the molecular basis of the humoral immune response. The predominant antigens recognised were fibre and penton base, with relatively less signal corresponding to hexon, the major capsid protein. A similar pattern was seen in two separate studies in the serum of patients both before and after treatment with recombinant adenovirus administered by the intrapleural15 and intratumoral routes.19 In the intratumoral study all patients recognised fibre, but some had no antihexon antibody before vector delivery and some failed to recognise penton base.19 The ascitic fluids recognised Ad2 and Ad5 antigens, but gave a stronger signal in the Western blot with Ad2 compared with Ad5, even though equal amounts of virus protein were present. The ascitic fluids can also recognise Ad3 antigens in a Western blot indicating that patients may recognise and potentially be immune to infection with a range of adenovirus serotypes (data not shown). Virus delivery and subsequent transgene expression could be compromised if antibodies in ascitic fluid possess neutralising activity. Neutralisation is serotypespecific and is determined by the hexon and fibre, which carry the type-specific determinants.6–9 Neutralising antifibre antibodies block the initial attachment stage, whereas anti-hexon antibodies are thought to act at later stages of internalisation.20 In Ad5 neutralisation assays all ascites and serum samples (tested at 100-fold dilution) neutralised the infectivity of 104 particles per cell by at least 50%, comparable to the reduction seen with BAL fluid.17 The effects of heat inactivation and protein A treatment on ascitic fluid and serum have shown that neutralisation of virus infection is due to antibodies alone. Attempts to inhibit higher titres of virus were compromised because at 109 particles per 104 cells the input virus was cytotoxic, and the GFP signal was not linear with dose beyond 108 particles. Addition of ascites to 109 virus particles in fact resulted in an increased GFP signal,

641

Gene Therapy


642

attributable to protection from the cytotoxic effect, but gave no net neutralisation. Several groups have sought to circumvent type-specific neutralisation, for example by using an alternative serotype for repeat administration21 or by replacing the hexon of an Ad5 vector with Ad12 hexon22 thus removing Ad5 serotype specific epitopes. However, for this to be effective the patient must not have neutralising antibodies directed towards these alternative serotypes. We show here that Ad3 can also be neutralised by ascites, indicating that ascites and serum samples contain antibodies capable of neutralising virus from two different subgroups. In conclusion, we have shown the presence of adenovirus-neutralising antibodies in ascitic fluid from patients with advanced ovarian cancer; this is an important factor which will have to be taken into consideration in the design of intraperitoneal gene therapy strategies. As delivery will likely be planned to succeed peritoneal drainage, the opportunity exists to monitor for antibody status, and also to reduce the level at the time of treatment by lavage. It is encouraging that a recent extensive study indicates that immune response to virus challenge can be predicted by the presence of pre-existing antibodies, for a variety of delivery routes;18 it remains to be determined whether the peritoneum falls within this category, and to what extent this immune response actually compromises effective gene delivery.

Materials and methods Patients Twelve patients with advanced stage ovarian cancer were randomly selected and consent was given for ascitic fluid obtained during routine peritoneal drainage procedures and blood to be used for research purposes. Ascitic fluid samples were obtained from all patients, and blood samples from four. Ascitic fluid and serum preparation Ascitic fluid was collected by intraperitoneal catheter drainage and centrifuged at 900 g (10 min, 4°C, Beckman GS-200R, High Wycombe, UK). Blood was incubated for 12 h at 4°C, the clot removed and serum clarified by centrifugation at 900 g (10 min, 10°C). Cells and virus A549 cells, a cell line non-permissive for replication of E1-deleted recombinant adenovirus, were passaged twice weekly in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% foetal calf serum (FCS) and 2 mm glutamine and incubated at 37°C, 5% CO2. 911 cells, an Ad E1 complementing cell line23 were passaged as for A549 cells. AdCMV-GFP⌬E1 is an E1, E3 deleted Ad5 virus expressing green fluorescent protein under the control of the CMV promoter inserted in the E1 region (unpublished data), and was propagated in 911 cells following infection in DMEM supplemented with 2% FCS and 2 mm glutamine (infection media). Cells showing a full cytopathic effect were harvested, centrifuged at 900 g (10 min, 10°C, Beckman GS-200R) and resuspended in infection media. Purified virus was obtained as described.24 Briefly, the cell suspension was subject to three cycles of freeze/thawing, centrifugation as before, Gene Therapy

and then layered on to CsCl gradients. Purified virus was dialysed against phosphate buffered saline (PBS)/10% glycerol in Slide-A-Lyzer cassettes (Pierce and Warriner, Cheshire, UK).

Virus particle count Virus particle number was calculated based on the protein concentration of CsCl banded virus and was determined using the BioRad protein assay (Basingstoke, UK) with BSA as standard. Virus particle number was estimated using the equation 1 mg/ml protein = 3.45 × 1012 particles/ml.25 Detection of whole immunoglobulin (Ig), IgA, IgM and IgG anti-adenovirus antibodies Adenovirus types 2 and 5 were diluted to 200 ng protein/ml in PBS and heat-inactivated (56°C, 30 min). 100 ␮l aliquots were applied to each well of a 96-well flat bottom plate (3912, Falcon, Becton Dickinson, Oxford, UK) and incubated at 37°C for 2 h. Subsequent steps were all performed at room temperature. Plates were washed three times with 0.1% Tween (Sigma, Poole, UK) in PBS, and three times in PBS. Ascitic fluid and serum samples were initially diluted 10−2 in 0.1% Tween in PBS, and subsequently diluted two-fold. 100 ␮l of each dilution was added to the plate, incubated for 2 h and washed as before. A hyperimmune rabbit anti-adenovirus type 5 serum was used as positive control for recognition of adenovirus antigens. For the detection of whole Ig, 100 ␮l of a 10−4 dilution of rabbit anti-human Ig (whole molecule) antibody conjugated to horseradish peroxidase (HRP) (Sigma) was applied to the wells and incubated for 2 h. 10 ml of substrate, 10 mg OPD (Sigma) in 75 ␮m citrate-phosphate buffer, sterile filtered through a 0.2 ␮m filter, was mixed with 20 ␮l 30% H2O2. 100 ␮l substrate was added to each well, incubated for 20 min, and the reaction quenched by the addition of 50 ␮l 4 m H2SO4. Absorbance was measured at 490 nm on an ELISA plate reader (Bio-Tek Kontron Instruments, Watford, UK). For measurement of isotypes, antibodies were used at the recommended dilutions for optimal detection: 6 × 10−3 of anti-human IgA-HRP, 3 × 10−3 of anti-human IgM-HRP and 10−4 of anti-human IgG-HRP (The Binding Site, University of Birmingham, UK). Western blotting of adenovirus antigens Major antigens recognised by anti-adenovirus antibodies were determined by Western blotting. Ad2 and Ad5 virus, diluted in PBS to concentrations of 125, 250, 500 and 1000 ng protein/ml, were subject to SDS-PAGE (10% Protogel, National Diagnostics, Hull, UK). Following electrophoresis, proteins were transferred by semi-dry electroblotting on to nitrocellulose membrane (Gelman Sciences, Northampton, UK) which was blocked in 5% non-fat dried milk/0.1% Tween in PBS for 90 min at room temperature. Ascitic fluid and serum samples diluted 10−3 in 0.1% Tween/PBS were added to the membranes and incubated for 90 min at room temperature. Membranes were washed 3 × 10 min in 0.1% Tween in PBS. A 10−4 dilution of anti-human Ig (as before) in 0.1% Tween in PBS was added to the membranes and incubated for 90 min. After washing, membranes were incubated with ECL reagents (Amersham, Amersham, UK) for 30 s, exposed to X-OMAT AR film (Kodak, Amersham, UK) for 30 s and developed.


Quantitation of virus neutralisation A fluorescent assay using Ad CMV-GFP⌬E1 was developed to measure virus neutralisation by ascitic fluid and serum. 96-well tissue culture plates (Nunc, Paisley, UK) were seeded with 1 × 104 A549 cells, 24 h before infection. Medium was removed and replaced with 100 ␮l phenol red free DMEM (Sigma) supplemented with 2% FCS + 2 mm glutamine. Cells were infected with purified Ad CMV-GFP⌬E1 at 107 and 108 particles per well. Ascitic fluid and serum samples were diluted in phenol red free DMEM (as above) and added to the wells at final dilutions of 10−2 to 10−5. Plates were incubated at 37°C, 5% CO2. At 72 h after infection fluorescence levels were measured in a fluorescent plate reader (Fluoroskan, Cambridge, UK) using wavelengths of 485 nm (excitation) and 538 nm (emission). To show that virus neutralisation is serotype specific, controls using rabbit hyperimmune anti-Ad5 serum and rabbit anti-Ad2 serum were included. Heat-inactivated ascitic fluid was prepared by incubating ascites at 56°C, 30 min and used in neutralisation assays as described above. Ascites was depleted of antibodies by incubating with protein A-conjugated Sepharose. 50 ␮l aliquots of ascites were incubated with three sequential 50 ␮l aliquots of protein A Sepharose at 20°C, 3 h, before use in neutralisation assays. Neutralisation of Ad3 was determined using a viral cytopathic effect (cpe) assay. A549 cells were seeded in 24-well plates at a density of 105 cells per well. 24 h later media was removed and 100 ␮l Ad3wt added (at multiplicities of infection of 101 to 105 particles per cell) along with 100 ␮l ascites and serum (final dilutions 10−2 to 10−5). Two hours later, a further 500 ␮l DMEM supplemented with 2% FCS and 2 mm glutamine was added. At 72 h after infection cells were fixed with 4% paraformaldehyde, stained with crystal violet and the cpe scored.

Acknowledgements We would like to thank Drs Moira Gilligan and Rachael Barton for the collection of ascitic fluid and serum samples and Dr Peter Searle for construction of Ad-CMVgfp⌬E1. We are indebted to Dr Leonard Seymour for stimulating discussions and critical reading of the manuscript. This work was supported by the Cancer Research Campaign and the Medical Research Council.

References 1 Searle P, Mautner V. Adenoviral vectors: not to be sneezed at. Gene Therapy 1998; 5: 725–727. 2 Zhang W. Development and application of adenoviral vectors for gene therapy of cancer. Cancer Gene Therapy 1999; 6: 113–138. 3 Runyon BA. Editorial: malignancy-related ascites and ascitic fluid ‘Humoral tests of malignancy’. J Clin Gastroenterol 1994; 18: 94–98. 4 Puls LE et al. The prognostic implication of ascites in advancedstage ovarian cancer.Gynecol Oncol 1996; 61: 109–112. 5 Nagy JA, Herzberg KT, Dvorak JM, Dvorak HF. Pathogenesis of

6 7

8

9

10

11

12

13

14 15

16 17

18

19

20 21

22

23

24

25

malignant ascites formation: initiating events that lead to fluid accumulation. Cancer Res 1993; 53: 2631–2643. Horwitz MS. Adenoviruses In: Fields BN et al (eds). Virology. Lippincott-Raven: Philadelphia, 1996, pp 2149–2169. Shenk T. Adenoviridae: The viruses and their replication In: Fields BN et al (eds). Virology. Lippincott-Raven: Philadelphia, 1996, pp 2111–2148. Crawford-Miksza L, Schnurr DP. Analysis of 15 adenovirus hexon proteins reveals the location and structure of seven hypervariable regions containing serotype-specific residues. J Virol 1996; 70: 1836–18445. Toogood CI, Crompton J, Hay RT. Antipeptide antisera define neutralizing epitopes on the adenovirus hexon. J Gen Virol 1992; 73: 1429–1433. Yang Y, Li Q, Ertl HCJ, Wilson JM. Cellular and humoral immune responses to viral antigens created barriers to lungdirected gene therapy with recombinant adenoviruses. J Virol 1995; 69: 2004–2015. Yang Y, Ertl HCJ, Wilson JM. MHC class I-restricted cytotoxic T lymphocytes to viral antigens destroy hepatocytes in mice infected with E1-deleted recombinant adenoviruses. Immunity 1994; 1: 433–442. Van Ginkel FW et al. Adenoviral gene delivery elicits distinct pulmonary-associated T helper cell responses to the vector and to its transgene. J Immunol 1997; 159: 685–693. Michou AI et al. Adenovirus-mediated gene transfer: influence of transgene, mouse strain and type of immune response on persistence of transgene expression. Gene Therapy 1997; 4: 473–482. Wilson JM. Vehicles for gene therapy. Nature 1993; 365: 691–692. Molnar-Kimber KL et al. Impact of pre-existing and induced humoral and cellular immune responses in an adenovirus-based gene therapy phase I clinical trial for localized mesothelioma. Hum Gene Ther 1998; 9: 2121–2133. Chirmule N et al. Immune responses to adenovirus and adenoassociated virus in humans. Gene Therapy 1999; 6: 1574–1583. Bastian A, Bewig B. Inhibition of adenovirus-mediated gene transfer by bronchoalveolar lavage fluid. Gene Therapy 1999; 6: 637–642. Harvey BG et al. Variability of human systemic humoral immune responses to adenovirus gene transfer vectors administered to different organs. J Virol 1999; 73: 6729–6742. Gahery-Segard H et al. Immune response to recombinant capsid proteins of adenovirus in humans: antifiber and anti-penton base antibodies have a synergistic effect on neutralizing activity. J Virol 1998; 72: 2388–2397. Wohlfart C. Neutralisation of adenoviruses: kinetics, stoichiometry, and mechanisms. J Virol 1988; 62: 2321–2328. Mastrangeli A et al. ‘Sero-switch’ adenovirus-mediated in vivo gene transfer: circumvention of anti-adenovirus humoral immune defenses against repeat adenovirus vector administration by changing the adenovirus serotype. Hum Gene Ther 1996; 7: 79–87. Roy S, Shirley PS, McClelland A, Kaleko M. Circumvention of immunity to the adenovirus major coat protein hexon. J Virol 1998; 72: 6875–6879. Fallaux FJ et al. Characterization of 911: a new helper cell line for the titration and propagation of early region 1-deleted adenoviral vectors. Hum Gene Ther 1996; 7: 215–222. Mautner V. Methods for growth and purification of enteric adenovirus type 40. In: Wold WSM (ed). Adenovirus Methods and Protocols. Humana Press: New Jersey, 1999, pp 283–294. Lemay P, Boudin M, Milleville M, Boulanger P. Human adenovirus type 2 protein IIIa.I. Purification and characterisation. Virology 1980; 101: 131–143.

643

Gene Therapy