Ex vivo adenovirus-mediated gene transfer and ... - Semantic Scholar

Gene Therapy (1997) 4, 639–647  1997 Stockton Press All rights reserved 0969-7128/97 $12.00

Ex vivo adenovirus-mediated gene transfer and immunomodulatory protein production in human cornea HB Oral1, DFP Larkin1,2, Z Fehervari1,3, AP Byrnes4,5, AM Rankin6, DO Haskard7, MJA Wood4, MJ Dallman6 and AJT George1 1

Department of Immunology, Royal Postgraduate Medical School, Hammersmith Hospital, London; 2Department of Pathology, Institute of Ophthalmology, London; 4Department of Human Anatomy, University of Oxford; 6Infection and Immunity Section, Department of Biology, Imperial College of Science, Technology and Medicine, London; 7Department of Medicine, Cardiovascular Medicine, Royal Postgraduate Medical School, Hammersmith Hospital, London, UK

One attractive strategy to prevent or control allograft rejection is to genetically modify the donor tissue before transplantation. In this study, we have examined the feasibility of gene transfer to human corneal endothelium, using a number of recombinant adenovirus constructs. Ex vivo infection of human corneas with adenoviral vectors containing lacZ, under transcriptional control of either cytomegalovirus (CMV) or Rous sarcoma virus (RSV) promoters, provided high-level gene expression, which was largely restricted to endothelium. Expression of the reporter gene persisted at relatively high levels for up to 7 days, followed by a decline to indetectable levels by 28 days. RT-PCR analysis of lacZ transcription showed a similar picture with a short period (3–7 days) of RNA

transcription after infection. In contrast, adenoviral DNA persisted for at least 56 days. Subsequently, we examined the expression of a potential therapeutic gene, CTLA-4 Ig fusion protein. Following infection of human corneas with adenoviral vectors encoding CTLA-4 Ig protein, high levels of the fusion protein were detected in corneal culture supernatants for up to 28 days. This protein was functionally active, as determined by binding to B7.1 (CD80)expressing transfectants. This study suggests that genetic alteration of donor cornea before transplantation is a feasible approach for preventing or controlling allograft rejection. Similar gene-based strategies might also be feasible to prevent rejection of other transplanted tissues or organs.

Keywords: transplantation; CTLA-4 Ig; b-galactosidase; RT-PCR; endothelium

Introduction Gene therapy offers the exciting possibility of expressing genes in vivo with therapeutic potential in target cells. However, targeting novel genes to specific organs in vivo is problematic. One of the strategies to achieve efficient therapeutic gene expression is ex vivo genetic modification of the target cells or organs, followed by their transplantation. Cornea is a particularly suitable candidate tissue for such gene-based approaches, because unlike other transplanted tissues, it can be maintained in standard culture conditions for periods of up to 1 month, allowing time for its genetic alteration before transplantation. The corneal endothelium has distinct advantages as a target for gene therapy. It is a cell monolayer on the internal surface of the cornea with well-defined anatomy,

Correspondence: AJT George, Department of Immunology, RPMS, Hammersmith Hospital, Du Cane Road, London W12 0NN, UK Current addresses: 3Department of Immunology, Glaxo Wellcome Medicines Research Centre, Gunnels Wood Road, Stevenage SG1 2NY, UK; 5 Department of Molecular Microbiology and Immunology, Johns Hopkins University, School of Public Health, 615 N Wolfe Street, Baltimore, MD 21205-2179, USA Received 8 January 1997; accepted 7 March 1997

easy accessibility and critical importance in maintenance of corneal transparency, which itself allows direct observation of the effects of gene transfer. In addition, modification of endothelial cell gene expression with the objective of local modulation of immune response or alteration of cellular physiology could have widespread clinical application, as the negligible replication capacity of corneal endothelium is insufficient to compensate for endothelial damage which may follow surgical trauma, inflammation (such as allograft rejection) or degenerative disease. At present, the only available treatment for such corneal injury is replacement of the cornea with an allograft. Primary endothelial disease is the indication for approximately 50% of corneal transplants,1,2 and the endothelium is itself the critical target of allograft rejection. In addition, short-term and chronic endothelial cell loss following transplantation, in the absence of overt rejection, is a frequent cause of corneal graft failure.3 We have previously examined gene transfer to whole rabbit corneas using a recombinant adenovirus vector carrying the lacZ gene driven by a human cytomegalovirus (CMV) promoter. In these studies, rabbit corneas were incubated with adenovirus ex vivo, and the efficiency and time course of gene expression following periods of corneal culture were investigated.4 Transgene

Ex vivo adenovirus-mediated gene transfer to human cornea HB Oral et al

640

expression was comparatively restricted to the corneal endothelium with expression in almost 100% of cells at day 1, diminishing to expression in 10% by day 14. Using the same experimental conditions, it has been found that the lacZ transgene was expressed in rat corneas in up to 50% of endothelial cells for 21 days ex vivo,5 suggesting a disparity in expression between the species. If techniques of ex vivo gene transfer are to be of clinical utility, several key issues require resolution. Studies using animal tissues may not accurately predict the efficacy and safety of gene transfer to human tissues. One more fundamental problem is the loss of transgene expression, which has characterised E1-deleted adenoviruses and other gene delivery systems such as lipofection and receptor–DNA conjugate mediated gene delivery. 6 Here we describe efficient marker gene expression in the endothelium of human corneas, with high levels of b-gal production for 7 days following ex vivo infection with recombinant adenoviruses under transcriptional control of either Rous sarcoma virus (RSV) or CMV promoters. It has long been recognised that T cells, both CD4+ and CD8+, play a critical role in rejection of allografts.7–9 One of the approaches to prevent graft rejection is to prevent T cell activation as a result of interaction with graft antigens. Antigen-specific T cell activation requires not only recognition of TCR by the MHC molecules, but also secondary stimulation provided by other receptor–ligand interactions.10 One of the costimulatory pathways is the interaction of B7.1 (CD80) and B7.2 (CD86) molecules on antigen presenting cells with the CD28 molecule on T cells.11,12 The inhibition of CD28 costimulation has been shown to cause T cell anergy and hyporesponsiveness.13–15 CTLA-4 is an alternative receptor on T cells for B7.1 and B7.2, with high homology to CD28.16,17 In comparison to CD28, it binds B7.1 with a 20fold higher avidity.12 Furthermore, CTLA-4 Ig has been shown to extend graft survival in a number of rat and mouse transplant models.18–20 This indicates the possibility of gene-based immunomodulation of the allogeneic response to a corneal allograft by ex vivo virus-mediated gene transfer to the donor cornea before transplantation.

Results Efficiency and localisation of lacZ expression in corneas In preliminary experiments conditions were optimised for infecting human corneas. Incubation of corneal samples with various concentrations of the AdCB2 or the AdRL virus constructs allowed us to determine the optimal conditions for efficient gene transfer. Maximal gene expression, with 90–100% of endothelial cells expressing b-gal, was obtained with 4 × 108 p.f.u. per cornea of the AdRL, whereas 1.5 × 108 p.f.u. per cornea of the AdCB2 provided a maximal transgene expression (data not shown). These viral numbers were used for all further experiments. b-Gal expression was almost entirely localised in the corneal endothelium following infection with either AdRL or AdCB2 (Figure 1a and b). However, some blue staining was observed in both keratocytes and epithelial cells, especially at the cut edge of the specimens as has been seen in the rabbit and rat (Figure 1c).4,5 Control, nontransduced corneas did not stain positively with Xgal (data not shown). These results indicate that the general expression pattern of b-galactosidase following

adenoviral infection is similar to that seen in the rabbit and rat corneas.4,5

Kinetics of b-gal expression Using a b-gal colorimetric assay, we examined the time course of the recombinant protein expression in corneal samples, infected by either AdRL or AdCB2. In these experiments, b-galactosidase expression was assayed following 1, 3, 7, 14, 21 and 28 days of ex vivo culture. Following infection with AdCB2, highest expression was seen at day 7, whilst AdRL yielded maximal b-gal activity at day 3 (Figure 2a and b). b-Gal activity was not detectable 28 days after infection with either virus. The RSV promoter-driven expression was quantitatively lower at all time-points, other than day 3. Control, mockinfected, corneas did not show any b-gal activity. The variability between samples was greater than that previously reported for rabbit or rat corneas.4,5 This may be due to the heterogeneity of the human tissue samples, particularly the differences in the number of endothelial cells present. Reverse transcriptase PCR for RNA analysis Analysis of the lacZ gene transcription was performed using RT-PCR of corneal samples previously infected with either AdCB2 or AdRL. RNA was extracted from these samples at various times after infection (from day 1 to day 28). Following infection with AdRL, the 1036 bp sequence of the lacZ gene was amplified from the samples at day 1, 3 and 7 by RT-PCR (Figure 3). However, RT-PCR analysis performed in AdCB2-infected cornea samples revealed that transcription of lacZ gene was not detectable after day 3 after 29 cycles of RT-PCR (Figure 3). No such band was seen in cDNA prepared from cornea segments at later time-points and from uninfected cornea. In addition, direct PCR of the RNA samples did not lead to appearance of the 1036 bp band, suggesting that DNA contamination was not responsible for the results seen (data not shown). Persistence of adenoviral DNA and the lacZ gene in gene-modified human corneas To determine the presence of adenoviral DNA and the lacZ gene, PCR analysis was performed on the DNA samples extracted from AdCB2 and mock-infected cornea specimens at various time-points, from 1 to 56 days, following infection. The 250 bp band corresponding to the adenoviral E2A sequence was seen at all time-points. The PCR analysis amplifying the 1036 bp internal sequence of the lacZ gene also revealed the presence of the reporter gene up to day 56. No such band was demonstrable in DNA extracted from mock-infected corneal specimens (Figure 4). These results indicate that adenoviral DNA remains present for long periods after expression of the gene has terminated. Similar results were obtained with AdRL vector, with DNA seen up to 56 days (data not shown). Time course of CTLA-4 Ig secretion by gene-modified corneas CTLA-4 Ig protein secretion from two whole cornea samples infected with AdCTLA was determined by sandwich ELISA. Both corneas were washed three times after infection to prevent the possibility of carryover of CTLA4 Ig from the purified viral preparation. The corneas showed similar kinetics for the production of CTLA-4 Ig


641

Figure 1 b-Gal expression in infected human corneas. (a) Human cornea photographed following infection with AdRL, culture ex vivo for 24 h and incubation in X-gal solution. The endothelial surface is uppermost. Diffuse macroscopically visible blue staining is seen, indicating b-gal expression. (b) Diffuse blue staining of corneal endothelium following incubation in AdRL (original magnification × 400). (c) Blue staining at the cut edge of a corneal specimen following infection with AdRL and 3 days of culture ex vivo. Keratocytes (scattered cells in the stroma of the cornea) adjacent to the edge and migrated epithelial cells (the sheet of cells migrating down from the top surface of the cornea, along the cut edge) express b-gal (original magnification × 100).


642

Figure 2 Persistence of b-galactosidase expression in two sets of human cornea samples maintained ex vivo following infection with AdCB2 (a) or AdRL (b). b-Gal activity in cornea lysates was determined by soluble b-galactosidase assay at the indicated days after infection. Activity is expressed as milliunits (mU) per 20 mg of total protein in the cornea extracts. Detection threshold of the assay is 5 mU/20 mg.

Figure 3 RT-PCR analysis of lacZ gene transcription. Total RNA was prepared from an uninfected cornea sample (negative control) and corneal samples at various times after AdRL or AdCB2 infection. RT-PCR analysis was carried out to amplify the 1056 bp sequence of the lacZ gene and the 540 bp sequence of b-actin gene. Direct PCR product from pCMV/b-gal vector was used as a positive control for the lacZ gene.

fusion protein. Following high-level initial protein production, the secretion of the CTLA-4 Ig fusion protein into culture supernatants was continued over 28 days (Figure 5). The cumulative CTLA-4 Ig production reached approximately 2 mg over 28 days. These data demonstrate that adenovirus-mediated gene delivery approaches are feasible for expression of therapeutic genes in corneal endothelium.

Flow cytometric analysis of CTLA-4 Ig binding The binding activity of the CTLA-4 Ig fusion protein, which blocks T cell costimulation, was tested by assessing

its ability to recognise its specific ligand B7.1 (CD80) by flow cytometry. The flow cytometry profiles in Figure 6c demonstrate specific binding of the fusion protein found in the supernatant of AdCTLA-infected corneas to B7.1 (CD80) transfected DAP-3 cells, in a similar manner to that seen with the purified CTLA-4 Ig fusion protein (Figure 6d). The small peak at around 10 fluorescent units was seen in both samples, and is presumably due to lower expression of the transfected gene in the same cells (Figure 6c and d). No binding appeared with either of the negative controls, PBS or human IgG1 (Figure 6a and b, respectively). In addition, incubation of the samples


643

Figure 4 Presence of adenoviral DNA and lacZ DNA in infected corneas. Total DNA was isolated from an uninfected cornea (negative control) and corneal tissues at various times following infection with AdCB2. PCR analysis was performed to amplify the 250 bp adenoviral DNA corresponding to the adenoviral E2A sequence, and the 1036 bp sequence of the lacZ gene at each time-point.

Figure 5 Time course of CTLA-4 Ig production by gene-modified corneas. Levels of CTLA-4 Ig protein in supernatants from AdCTLA-infected corneas were determined by sandwich ELISA. Corneal culture supernatants were collected from AdCTLA-infected corneas and replaced with fresh culture medium on days 2–36. The graph represents cumulative amounts of CTLA-4 Ig produced by two infected whole corneal samples. No secretion of CTLA-4 Ig was seen in mock-infected corneas (data not shown).

with nontransfected DAP-3 cells did not show any binding activity (data not shown).

Discussion Gene transfer to the cornea before its transplantation offers the opportunity to modify the alloresponse directed against the graft by local expression of immunomodulatory proteins, such as CTLA-4 Ig. The most significant target for gene modification will be corneal endothelial cells, because unlike other corneal cell types, damage during rejection is irreversible on account of its negligible replication capacity. The feasibility of marker

gene transfer to the cornea using recombinant viral vectors has been demonstrated by ourselves and others.4,5,21–23 Previously we have shown that the lacZ transgene expression following ex vivo adenovirus-mediated gene transfer to both rabbit and rat corneas was possible, and also entirely restricted to the endothelium.4 In the rabbit, the lacZ gene expression was maximal at day 1 diminishing to undetectable levels by day 21.4 Using the same experimental conditions, it has been found that the reporter gene was expressed in rat corneas in up to 50% of endothelial cells for 21 days ex vivo,5 suggesting a disparity in expression among the species. In the current experiments, we have demonstrated that replicationdefective adenoviruses have capacity to infect human cornea ex vivo, and that the transgene expression is largely restricted to the corneal endothelium. The reason for this lack of transgene expression in the epithelium is not clear, especially as in clinical corneal disease caused by wild-type virus adenoviruses, virus can be cultured from epithelial cells, indicating that these cells can internalise adenovirus.24 Absence of expression of the transduced lacZ gene in corneal stromal cells suggests that the endothelium or underlying Descemet’s membrane play a role as an effective barrier to adenoviral infection of the stroma. Similarly, it has been shown that the vascular endothelium and the underlying elastic lamina in blood vessels can act as a barrier, preventing adenoviral access to deeper layers of the vessel wall.25 In this study, the time-course experiments performed with two adenoviral constructs driven by either the RSV or the CMV promoter have shown that the reporter gene expression persists at relatively high levels for up to 7 days, followed by gradual decline to indetectable levels by 28 days. Analysis of lacZ transcription, using RT-PCR, gives a broadly similar picture, with RNA transcription limited to a short period after infection within the detection limits of the RT-PCR used. These data suggest that short-term transgene expression is not restricted to the CMV promoter used in the earlier studies. The reason for the short-term gene expression is not clear. The comparative lack of endothelial cell replication in the human cornea suggests that dilution of viral DNA during cell proliferation is unlikely to be the cause. In this study, we


644

Figure 6 Demonstration of the binding activity of CTLA-4 Ig protein using flow cytometry. The culture supernatant from AdCTLA-infected human cornea 1 day after infection (c) was removed and incubated with B7.1 (CD80) transfected DAP-3 fibroblasts (DAP3-B7.1). The binding activity was determined by flow cytometry following incubation with FITC-conjugated mouse anti-human IgG1. Purified CTLA-4 Ig protein (d) was employed as a positive control, whilst PBS (a) and a human IgG1 standard (b) were used as negative controls.

demonstrated that adenoviral DNA existed for at least 56 days after infection, suggesting that loss of viral DNA is not responsible for the short-term gene expression. However it is possible either that both the CMV and RSV promoter are inactivated over this time, or that the viral or recombinant DNA is modified or sequestered in some way as to prevent gene transcription. Corneal endothelium is a critical cell both in primary corneal disease, allograft rejection and allograft failure. 1,2,26 Therefore, the alteration of the endothelial cell genome may have therapeutic potential in diverse corneal diseases and vector-mediated local expression of immunomodulatory molecules may have widespread application in clinical transplantation. In this study, we have examined the feasibility of using the adenoviral constructs for the expression of CTLA-4 Ig, which is a protein that has been shown to block T cell costimulation mediated by interactions between CD28 and the B7 family (CD80 and CD86). This protein has been shown in a range of models to block allograft rejection.18–20 This report also shows that corneas modified by ex vivo adenovirus-mediated gene transfer can produce CTLA-4 Ig for up to 30 days, and this potential therapeutic protein, obtained from the culture supernatants of genemodified corneas, is capable of binding B7.1 molecules. It is of interest that the CTLA-4 Ig is produced for a longer period than b-gal, suggesting that a secreted protein may have a different kinetic of expression than a cytoplasmic protein possibly due to its slow release from secretory vesicles. In summary, this study has established a model for successful reporter and therapeutic gene transfer to corneal endothelium. We have also demonstrated that marker gene mRNA is detectable for only 3–7 days following infection with recombinant virus, but the adenoviral DNA can be detected for 56 days. Finally, we have

shown that functionally active CTLA-4 Ig can be secreted by human cornea which has been transduced ex vivo with a recombinant adenovirus vector containing the gene encoding this protein.

Materials and methods Reagents and antibodies RPMI 1640, minimal essential medium (MEM), Dulbecco’s modified Eagle’s medium (DMEM), Hanks’ balanced salt solution (HBSS), foetal calf serum (FCS), l-glutamine, penicillin and streptomycin were purchased from Gibco BRL (Paisley, UK). Caesium chloride and X-gal were from NBL Gene Sciences (Cramlington, UK). Ortho-phenylenediamine (OPD) tablets were obtained from Dako (High Wycombe, UK). All other chemicals were purchased from Sigma Chemical (Poole, UK), unless otherwise specified. Monoclonal antibodies against different epitopes of the Fc portion of human g1 immunoglobulin (SB7E6 and SB7H2) were kindly donated by Dr Martin Glennie (Tenovus Research Laboratories, Southampton, UK). Cell lines EA.hy-926, an endothelial cell hybridoma,27 was maintained in growth medium composed of RPMI 1640, 5% FCS, 2 mm l-glutamine, 100 U/ml penicillin and 100 mg/ml streptomycin at 37°C and 5% CO2. DAP-3 human fibroblast cells and DAP-3 transfected with B7.1 (CD80) (DAP-3-B7.1) were cultured in DMEM supplemented with 5% FCS, 2 mm l-glutamine, 100 U/ml penicillin and 100 mg/ml streptomycin.28 293 Embryonic human kidney cells, that contain the E1 region of the adenovirus genome, were used in propagation and titration of adenoviral vectors.29 This cell line was cultured in DMEM con-


taining 10% FCS, 2 mm l-glutamine, 100 U/ml penicillin and 100 mg/ml streptomycin.

Corneal samples Corneas were removed at surgery from patients with keratoconus, a disease in which corneal shape is abnormal but endothelial and epithelial cells normal. These corneas were cut into four segments before infection with recombinant adenovirus, which commenced in each case within 4 h of surgical removal. Viral vectors The recombinant E1-deleted type 5 adenoviral vectors AdCB2, AdRL and AdCTLA were used in this study. AdRL and AdCB2 contain the E. coli lacZ gene, encoding b-galactosidase under the transcriptional control of a Rous sarcoma virus (RSV) and a human cytomegalovirus (CMV) promoter, respectively. The AdRL construct has previously been described.30 AdCB2 was made using a 642 bp fragment from pCDNAI (Invitrogen, NV Leek, The Netherlands) containing the CMV promoter. The virus was produced by recombination of the plasmid encoding the CMV promoter and the lacZ gene (pXCXCB2) with pJM17, containing the d1309 genome,31 in 293 cells as previously described.30 AdCTLA encodes a mouse CTLA-4 human Fcg1 fusion protein driven by the RSV promoter. The extracellular region of mouse CTLA-4 gene was amplified from BALB/c splenic RNA by RT-PCR, using the following primers: mCTLA-4E: 5′UACGAATTCCTGAGGACCTCAGGCACATT-3′ (containing 5′ untranslated sequence and EcoRI site); and mCTLA-4J: 5′ UACGGATCCACTTACCTGGTAGAATCCGGGCATGGTTCT-3′ (containing membrane proximal extracellular sequence, a BamHI site and a splice donor sequence). The CTLA-4 PCR product was directionally cloned into pBluescript (KS+), then subcloned into pIg vector,32 which contains the gene encoding for human Fcg1, immediately preceded by a splice acceptor site. This complete CTLA-4 Ig insert was further subcloned into the pXCXRSV downstream of the RSV promoter,30 and the virus was produced by recombination with pJM17 as above. Viruses from the resulting plaques were plaque-purified three times and the DNA structure was verified by restriction analysis. Viruses were grown in 293 cells until the cytopathic effect occurred. The virus particles were released from the pellet of these cells by four freeze–thaw cycles, and purified by two rounds of caesium chloride (CsCl) density gradient centrifugation, as described previously.33 After dialysis against PBS to remove CsCl, glycerol was added to a final concentration of 10% (v/v). The viral preparations were passed through a 0.2 mm filter and stored at −70°C. Viral titers were assessed by plaque assay on 293 cells as described.33 The viral titres were 1.9 × 109 p.f.u./ml, 5 × 109 p.f.u./ml, and 7.9 × 109 p.f.u./ml for AdCB2, AdRL and AdCTLA4, respectively. Determination of optimum conditions for infecting endothelial cells Optimum infection conditions for adenoviral vectors were determined in EA.hy-926 cells as described previously.4,34 To determine the optimum conditions for AdCTLA virus infection, the CTLA-4 Ig fusion protein was detected using an ELISA, as described below.

Infection and culture conditions for corneas Each human cornea segment was incubated with 1 ml of 2% FCS in MEM containing 4 × 108 p.f.u. AdRL, 1.5 × 108 p.f.u. AdCB2 or 1 × 108 p.f.u. AdCTLA for 3 h at 37°C in 5% CO2. Control mock-infected corneas were incubated in virus-free medium. Following incubation, infected and control specimens were washed three times with HBSS to avoid prolonged viral exposure and the possible recombinant protein contamination in the viral preparations, which might have been produced during cultivation in host 293 cells. The cornea samples, were maintained in MEM supplemented with 10% FCS, 2 mm lglutamine, 100 U/ml penicillin and 100 mg/ml streptomycin at 37°C and 5% CO2 for variable periods. Culture medium was changed every 3 days. Histochemical analysis of lacZ expression For morphological studies to determine the tissue localisation of transgene expression, corneal specimens were examined by histochemical X-gal staining at day 1 following infection by the AdRL or the AdCB2 recombinant adenovirus constructs.4 Corneal sections were examined for blue staining of cells by light microscopy. Assay for b-gal activity b-Gal activity in corneal extracts was determined using o-nitrophenyl-b-d-galactopyronidase (ONPG) as a substrate.4 The total protein content of the corneal lysates was also estimated using a modified Lowry method.35 bGal activity was presented by normalizing to the total protein content of the corneal sample. PCR analysis of lacZ mRNA Total RNA was extracted from infected or mock-infected human corneas using the RNAzol B RNA Isolation Solvent (Cinna/Biotecx Laboratories, Houston, TX, USA) according to the manufacturer’s instructions. The cDNA was prepared using the Pharmacia First-Strand cDNA Synthesis kit (Pharmacia, Uppsala, Sweden) and stored at −20°C. PCR amplification of the 1036 bp segment of the lacZ gene was performed as described below. A primer pair to amplify a 540 bp segment of the housekeeping b-actin gene (5′ primer: 5′-GTGGGGCGCCCCA GGCACCA-3′; 3′ primer: 5′-CTCCTTAATGTCACGCAC GATTTC-3′ was used for testing the presence of total cDNA in each sample. 36 The PCR reactions prepared as described below were run under the following cycling conditions: 29 cycles of denaturation at 94°C for 1 min, annealing at 65°C for 1 min and extension at 72°C for 2 min. DNA isolation and analysis by polymerase chain reaction (PCR) Tissue debris remaining from corneal lysate preparation was used for total DNA extraction and purification. Briefly, corneal tissues were washed with ice cold PBS, incubated in a digestion buffer (100 mm NaCl, 10 mm Tris Cl, 25 mm EDTA, 0.5% SDS and 0.1 mg/ml proteinase K) for 18 h at 50°C, followed by phenol/chloroform extraction, ethanol precipitation and resuspension in TE buffer (pH 8.0).37 This yielded approximately 200 ng/ml of DNA. A primer pair was designed to amplify a 250 bp segment of adenoviral DNA corresponding to the E2a sequence (5′ primer, 5′-ACTGGCAGGGACACGTTGCGA-3′; 3′ primer, 5′-AGGAGCGTGCTGGCCAGCGT-

645


646

3′). The PCR reaction mixture was prepared in a total volume of 50 ml containing 8 ml of corneal DNA (approximately 5 ng/ml), 1 mm of deoxynucleoside triphosphate mixture, 10 mm of 5′ and 3′ oligonucleotide primers, 10 × PCR buffer (Bioline, London, UK) and 2.5 U Taq polymerase (Bioline). The PCR was performed under the following conditions: 94°C for 1 min; followed by addition of Taq polymerase, 29 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min. PCR products obtained in this manner were run on a 2% agarose gel and visualised with UV light. The oligomeric primers flanking an internal 1036 bp sequence of the lacZ gene have been previously published (primer 1: 5′-GCCGACCGCACGCCGCATCC AGC-3′, primer 2: 5′-CGCCGCGCCACTGGTGTGGGC3′).38 These primers were used for detection of the lacZ gene. Fifty microlitres of PCR reaction mixture was prepared as described above and subjected to the following cycling conditions; 30 cycles of denaturation at 94°C for 1 min, annealing and extension at 72°C for 2 min. PCR products were analyzed by gel electrophoresis as above.

Time course of CTLA-4 Ig fusion protein production Levels of CTLA-4 Ig fusion protein in the supernatants from AdCTLA-infected corneas were measured by sandwich ELISA. Corneal culture supernatants were removed from human corneas incubated with AdCTLA or virusfree medium, and replaced with fresh culture medium on days 2–36 following infection. 96-Well ELISA plates were coated with 10 mg/ml of SB7E6 mAb recognising the Fc portion of human IgG1. After incubation for 1 h at 37°C, the plates were washed with PBS/0.1% Tween-20 and blocked with 100 ml of 0.1% milk for 30 min at room temperature. Standard calibration curves were constructed using human IgG1 . Sixty microlitres of corneal culture supernatant or standard was added to the appropriate wells in triplicate and incubated for 1 h at 37°C. After the primary incubation, the plates were washed four times with PBS/0.1% Tween-20. Sixty microlitres of HRP-conjugated SB7H2 mAb (anti-human IgG1 Fc) was added to each well at a dilution 1:2000 and incubated for 1 h at 37°C. After an additional four washes, 50 ml of OPD substrate, prepared according to the manufacturer’s instructions, was added to the wells and incubated in the dark at room temperature for 20–30 min. The reaction was stopped by adding 50 ml of 5% H2 SO4, and the plates were read at 491 nm using the microplate reader. Flow cytometric analysis of CTLA-4 Ig activity The binding activity of the CTLA-4 Ig fusion protein was tested by flow cytometry. Briefly, CD80 transfected DAP3 fibroblast (DAP3-B7.1) cells were adjusted to 1 × 106 cells per ml and aliquoted in ice cold microcentrifuge tubes. One hundred microlitres of supernatant from AdCTLA-infected corneal samples, removed after ex vivo culture for 24 h following infection, were incubated with these cells on ice for 30 min. PBS and a human IgG1 standard were employed as negative controls, while purified CTLA-4 Ig protein was used as a positive control. After incubation, cells were washed once with PBS/azide (0.1% azide in PBS) and then incubated with 100 ml of FITC-conjugated mouse anti-human IgG1 mAb (Dako) on ice for 30 min. Finally, the cells were washed once and resuspended in 100 ml of PBS/azide in poly-

styrene round-bottom analysis tubes before analysis with a Flow cytometer (Coulter, Luton, UK).

Acknowledgements The authors thank Professor Robert Lechler for kindly providing the DAP-3 and the DAP-3-B.7.1 cells and Dr Jon Friedland for giving the b-actin primers and the protocol. We also thank Dr Martin Glennie for providing SB7E6 and SB7H2 antibodies. This work was supported by the Iris Fund for Prevention of Blindness, the Special Trustees of Moorfields Eye Hospital (DFPL), Uludag˘ University, Turkey (HBO), and Medical Research Council (MJD). DOH is the recipient of a British Heart Foundation Professorial award.

References 1 Price FW, Whitson WE, Collins KS, Marks RG. Five-year corneal graft survival. Arch Ophthalmol 1993; 111: 799–805. 2 Vail A et al. Corneal transplantation in the United Kindom and Republic of Ireland. Br J Ophthalmol 1993; 77: 650–652. 3 Bourne FW, Hodge DO, Nelson LR. Corneal endothelium five years after transplantation. Am J Ophthalmol 1994; 118: 185–196. 4 Larkin DFP et al. Adenovirus-mediated gene delivery to the corneal endothelium. Transplantation 1996; 61: 363–370. 5 Fehervari Z et al. Gene transfer to ex vivo stored corneas. Cornea (in press). 6 Schofield JP, Caskey CT. Non-viral approaches to gene therapy. Br Med Bull 1995; 51: 56–71. 7 Khodadoust AA, Silverstein AM. Transplantation and rejection of individual cell layers of the cornea. Invest Ophthalmol 1969; 8: 180–195. 8 Williams KA et al. Patterns of corneal graft rejection in the rabbit and reversal of rejection with monoclonal antibodies. Transplantation 1992; 54: 38–43. 9 Larkin DFP, Calder VL, Lightman SL. Identification and characterisation of cells infiltrating the graft and aqueous humour in rat corneal allograft rejection. Clin Exp Immunol 1997; 107: 381–391. 10 Mueller DL, Jenkins MK, Schwartz RH. Clonal expansion versus functional clonal inactivation: a costimulatury signalling pathway determines the outcome of T cell antigen receptor occupancy. Annu Rev Immunol 1989; 7: 445–480. 11 Linsley PS et al. Binding of the B cell activation antigen B7 to CD28 costimulates T cell proliferation and interleukin 2 mRNA accumulation. J Exp Med 1991; 173: 721–730. 12 Gimmi CD et al. B-cell surface antigen B7 provides a costimulatory signal that induces T cells to proliferate and secrete interleukin 2. Proc Natl Acad Sci USA 1991; 88: 6575–6579. 13 Harding FA et al. CD28-mediated signalling co-stimulates murine T cells and prevents induction of anergy in T cell clones. Nature 1992; 356: 607–609. 14 Tan P et al. Induction of alloantigen-specific hyporesponsiveness in human T lymphocytes by blocking interaction of CD28 with its natural ligand B7/BB1. J Exp Med 1993; 177: 165–173. 15 Boussiotis VA et al. B7 but not intracellular adhesion molecule1 costimulation prevents the induction of human alloantigenspecific tolerance. J Exp Med 1993; 178: 1753–1763. 16 Linsley PS et al. CTLA-4 is a second receptor for the B cell activation antigen B7. J Exp Med 1991; 174: 561–569. 17 Harper K et al. CTLA-4 and CD28 activated lymphocyte molecules are closely related in both mouse and human as to sequence, message expression, gene structure, and chromosomal location. J Immunol 1991; 147: 1037–1044. 18 Lenschaw DJ et al. Long-term survival of xenogeneic pancreatic islet grafts induced by CTLA4Ig. Science 1992; 257: 789–792. 19 Lin H et al. Long-term acceptance of major histocompatibility complex mismatched cardiac allografts induced by CTLA-4 Ig plus donor specific transfusion. J Exp Med 1993; 178: 1801–1806.


20 Pearson TC et al. Transplantation tolerance induced by CTLA4 Ig. Transplantation 1994; 57: 1701–1706. 21 Mashhour B, Couton D, Perricaudet M, Briand P. In vivo adenovirus-mediated gene transfer into ocular tissues. Gene Therapy 1994; 1: 122–126. 22 Feldman ST et al. Gene transfer to the anterior segment of the eye by infection with recombinant viral vector. Invest Opthalmol Vis Sci 1994; 35 (Suppl.): 1383 (Abstr.). 23 Budenz DL, Bennett J, Alonso L, Maguire A. In vivo gene transfer into murine corneal endothelial and trabecular meshwork cells. Invest Opthalmol Vis Sci 1995; 36: 2211–2215. 24 Chodosh J, Miller D, Stoop WG, Pflugfelder SC. Adenovirus epithelial keratitis. Cornea 1995; 14: 167–174. 25 Guzman RJ et al. Efficient and selective adenovirus-mediated gene transfer into vascular neointima. Circulation 1993; 88: 2838–2848. 26 Larkin DFP. Corneal allograft rejection. Br J Ophthalmol 1994; 78: 649–652. 27 Edgell C-JS, McDonald CC, Graham JB. Permanent cell line expressing human factor VIII-related antigen established by hybridization. Proc Natl Acad Sci USA 1983; 80: 3734–3737. 28 Hargreaves R, Logiou V, Lechler R. The primary alloresponse of human CD4+ T cells is dependent on B7 (CD80), augmented by CD58, but relatively uninfluenced by CD54 expression. Int Immunol 1995; 7: 1505–1513. 29 Graham FL, Smiley J, Russell WC, Nairn R. Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J Gen Virol 1977; 36: 59–74.

30 Byrnes AP, Rusby JE, Wood MJA, Charlton HM. Adenovirus gene transfer causes inflammation in the brain. Neuroscience 1995; 66: 1015–1024. 31 McGrory WJ, Bautista DS, Graham FL. A simple technique for the rescue of early region I mutations into infectious human adenovirus type 5. Virology 1988; 163: 614–617. 32 Simmons DL. Cloning cell surface molecules by transient expression in mammalian cells. In: Hartley D (ed). Cellular Interactions in Development. Oxford University Press: Oxford, 1993, pp 93–127. 33 Green M, Wold WSM. Human adenoviruses: growth, purification and transfection assay. Meth Enzymol 1979; 58: 434–435. 34 Sanes JR, Rubinstein JLR, Nicholas J-F. Use of recombinant retroviruses to study post-implantation cell image in mouse embryos. EMBO J 1986; 5: 3133–3142. 35 Krezse G-B. Methods for protein determination. In: Bergmeyer J, Grassl M (eds). Methods of Enzymatic Analysis: Samples, Reagents, Assessment of Results, vol 2. Verlag Chemie: Weinheim, 1983, pp 89–92. 36 Zhang M et al. Interleukin-12 at the site of disease in tuberculosis. J Clin Invest 1994; 93: 1733–1739. 37 Moore DD. Preparation of genomic DNA. In: Ausubel FM et al (eds). Current Protocols in Molecular Biology, vol 1. John Wiley: New York, NY, 1992, pp 2.2.1–2.2.3. 38 Csete ME et al. Efficient gene transfer to pancreatic islets mediated by adenoviral vectors. Transplantation 1995; 59: 263– 268.

647