Local Gene Transfer to Modulate Rat Corneal Allograft Rejection

Abstract

purpose. Allograft rejection is the leading cause of corneal graft failure. CD4⁺ T cells control the allograft response and represent targets for antirejection therapy. The purpose of this study was to transfer cDNA encoding a monomeric anti-CD4 antibody fragment to donor corneal endothelium, to attempt to modulate orthotopic corneal allograft rejection in the rat.

methods. A replication-deficient adenoviral vector (AdV) encoding anti-CD4 single-chain, variable-domain antibody fragment (scFv) and enhanced green fluorescent protein (eGFP) was constructed (AdCD4GFP). AdV encoding eGFP alone (AdGFP) was used as a control. Transgenic product was detected by reverse transcription–polymerase chain reaction (RT-PCR), Western blot, flow cytometry, and fluorescence microscopy. The alloinhibitory capacity of anti-rat CD4 scFv was measured in the one-way mixed lymphocyte reaction (MLR). The survival of Wistar-Furth corneas transduced with AdV either immediately or 3 days before orthotopic transplantation in Fischer 344 recipients was examined.

results. ScFv and eGFP mRNAs were detected in rat corneas transduced in vitro, and active scFv secreted in corneal supernatants peaked at days 4 to 5 after transduction at 23 ± 4 ng of protein per cornea per day. Antibody and scFv against rat CD4 blocked alloproliferation in MLR. However, transduction of corneas with AdCD4GFP ex vivo, immediately before transplantation, or in vivo, 3 days before transplantation, did not significantly prolong corneal allograft survival (P > 0.05).

conclusions. Anti-CD4 scFvs were capable of blocking allostimulation, but their local expression within the eye did not prolong corneal allograft survival, suggesting that sensitization may still occur.

The leading cause of corneal graft failure is irreversible immunologic rejection, ¹ primarily mediated by CD4-positive T cells. The CD4 molecule is involved in the antigen-specific interaction between the T cell receptor (TCR) and the MHC class II-bound allopeptide on antigen-presenting cells (APCs). It plays a role in stabilizing and prolonging the T cell-APC interaction, ^{2 3} and in recruiting intracellular molecules within the T cell to the site of antigen presentation. ⁴  

CD4 represents an attractive target for therapeutic intervention at an early stage of the alloinflammatory cascade. Anti-CD4 antibodies can delay or prevent allograft rejection and, in some cases, can promote allograft tolerance. ^{5 6 7 8} In rodents, anti-CD4 antibodies prolong corneal allograft survival. ^{9 10 11 12} However, systemic monoclonal antibody treatment remains an unrealistic clinical option for corneal transplantation, ^{13 14} whereas local administration of intact antibody by intracameral injection causes damaging inflammation elicited by the Fc tail of the molecule. ¹⁵ Antibodies delivered topically do not penetrate the ocular surface. ¹⁶ An alternative approach to the delivery of immunosuppressive proteins, such as antibodies to the ocular environs, is by gene therapy. Human donor corneas can be maintained in culture for several weeks before transplantation, making them attractive targets for ex vivo gene transfer. Adenovirus (AdV)-based vectors have been used to transfer genes to corneal tissue from many species, including rabbits, ¹⁷ rats, ¹⁸ mice, ¹⁹ sheep, ²⁰ monkeys, ²¹ and humans. ²²  

An alternative to inflammatory whole antibodies is single-chain variable domain (scFv) antibody fragments. As single antigen-binding sites, they are small (∼28 kDa compared with ∼146 kDa for IgG) and do not contain the inflammatory Fc region. Therefore, local expression of an anti-CD4 scFv by the transplanted cornea may prolong corneal allograft survival, while avoiding systemic complications. In this study, the transduction of rat corneal endothelium with a replication-deficient AdV carrying cDNA encoding an anti-rat CD4 scFv was optimized. The kinetics of scFv secretion by transduced donor corneas, the immunomodulatory effect of scFv in vitro, and the effect of transplanting anti-CD4 scFv gene-modified donor corneas was investigated in a rat model of corneal allograft rejection. 

Materials and Methods

Animals

Adult inbred Fischer 344 (F344; RT^1vl) rats were used as recipients of orthotopic corneal grafts from F344 or inbred Wistar-Furth (WF; RT^1u) donors. Approval for all experimentation was obtained from the institutional Animal Welfare Committee and was performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Adenoviral Vector Construction and Purification

A replication-deficient E1-, E3-deleted serotype 5 adenovirus encoding enhanced green fluorescent protein (eGFP) under the transcriptional control of a CMV promoter (AdGFP) was the kind gift of Bert Vogelstein (John Hopkins University, Baltimore, MD). Construction of a scFv gene derived from the anti-rat CD4 hybridoma line OX38 in the bacterial expression vector pHB400 has been described. ²³ cDNA encoding the mammalian secretory leader sequence of human complement factor H (fHSS) was kindly provided in the fHSS-v3-BUC plasmid by David Gordon (Flinders University, Adelaide, Australia). ²⁴ AdV encoding anti-rat CD4 scFv with the mammalian secretory leader sequence (fHSS) and a 6-histidine tag (his-tag) under the transcriptional control of a cytomegalovirus (CMV) promoter (AdCD4GFP) was generated with a commercial system (pAdEASY; Qbiogene, Carlsbad, CA), based on methods described by He et al. ²⁵ AdCD4GFP also encoded eGFP under the control of a separate CMV promoter. AdV were propagated in E1A, E1B-transcomplementing human embryonic kidney 293 (HEK-293) cells ²⁶ and purified from sonicated cell pellets by double cesium chloride density gradients. ²⁷ Purified AdV was extensively dialyzed, filter-sterilized at 0.2 μm, and stored at −80°C in 10% (vol/vol) glycerol in phosphate-buffered saline (PBS). AdV titer was determined by the tissue culture infectious dose method (TCID₅₀) on HEK-293 cells. ²⁸ Titers of individual batches of virus ranged from 4 × 10⁹ to 6 × 10¹⁰ plaque-forming units (pfu)/mL. 

In Vitro Transduction of Rat Corneas with AdV

The donor rat was killed by overdose of inhalation anesthetic. Eyes were removed and decontaminated in 10% wt/vol povidone iodine (Faulding Pharmaceuticals, Salisbury, Australia) for 2 minutes and rinsed twice in ophthalmic balanced saline solution (Cytosol Ophthalmics, Lenoir, NC). Corneas were dissected with a 1- to 2-mm scleral rim. The iris was removed and the corneas placed in HEPES-buffered RPMI 1640 medium (ICN, Costa Mesa, CA) supplemented with 2% vol/vol heat inactivated (56°C, 30 minutes) fetal calf serum (FCS), 100 IU/mL penicillin, 100 μg/mL streptomycin sulfate, and 2 mM l-glutamine (all from Invitrogen-Gibco, Gaithersburg, MD; RPMI-2% FCS). Corneas were transduced in sterile 96-well, round-bottomed plates (Nalge Nunc International, Rochester, NY), endothelial side up, with AdV diluted in a total volume of 100 μL RPMI-2% FCS for 2 hours at 37°C in 5% CO₂ in air. Donor corneas treated with AdV or control medium were rinsed twice in HEPES-buffered RPMI 1640 medium and placed in balanced saline before being used within 1 hour for grafts or cultured for up to 14 days for in vitro expression experiments in 2 mL HEPES-buffered RPMI 1640 medium supplemented with 100 IU/mL penicillin, 100 μg/mL streptomycin sulfate, 2 mM l-glutamine, 10% FCS (RPMI-10% FCS), and 2.5 μg/mL amphotericin B (Invitrogen-Gibco) at 37°C in 5% CO₂ in air. 

Detection of Transgene mRNA

Corneas were snap frozen in liquid nitrogen and pulverized with a mortar and pestle. RNA was extracted with a kit (Total RNA Isolation Reagent; Advanced Biotechnologies Ltd., Surrey, UK). Up to 10 μg total RNA was treated with 10 U DNase I (Roche, Basel, Switzerland), to remove contaminating genomic and adenoviral DNA before cDNA was prepared with a synthesis kit (FNA; Amersham Pharmacia Biotech, Sydney, NSW, Australia). cDNA encoding scFv (1 kb) was detected by PCR, using the primers 5′-GGGTCGACCAAAAAATGAGACTTCTAGCAAAGATT-3′ and 5′-CATCTAGACTAATGATGGTGATGATGGTGATC-GGCC-3′. Detection of a 300 base pair (bp) product of the eGFP gene with primers 5′-CTGACCCTGAAGTTCATCTGCAC-3′ and 5′-TGGCTGTTGTAGTTGTACTCCAGC-3′ served as a transduction control. PCR on total RNA (before reverse transcription) was used to control for DNA contamination in samples. Products were detected on a 1.5% agarose gel against a 2 log DNA ladder (New England Biolabs, Beverly, MA). 

Detection of eGFP Expression by Fluorescence Microscopy

Corneas were fixed in buffered formalin for 10 minutes and rinsed twice in PBS. Cell nuclei were stained with 50 μL of 10 μg/mL Hoechst 33258 (Sigma-Aldrich, St. Louis, MO) in balanced saline for 30 minutes at room temperature and rinsed twice in PBS. Corneas were flatmounted on glass slides with glycerol medium (Bartels Buffered Glycerol; Trinity Biotech, Wicklow, Ireland) and examined under a fluorescence microscope (model BX50; Olympus Optical Co., Tokyo, Japan) equipped with a digital camera (CoolSNAP high-resolution cooled charge coupled device [CCD], 1.0× tube) and image-analysis software (RS Image, ver. 1.0.1; both from Roper Scientific Inc., Tucson, AZ). Hoechst 33258-stained nuclei were detected in flatmounted rat corneas under ultraviolet (UV) light. Different layers of the corneal flatmount were observed by traversing through different planes within the same optical field. Fields were examined under blue light (to visualize eGFP-positive cells) and UV light (to visualize total endothelial nuclei). Efficiency of transduction was calculated from five central fields (each 0.15 mm²) as the percentage of corneal endothelial cells expressing eGFP. 

Detection of Anti-CD4 scFv Protein by Western Blot

His-tagged anti-CD4 scFv protein was purified from transduced HEK-293 culture supernatant by immobilized metal affinity chromatography. Culture supernatant (18 mL) was incubated with 1 mL Ni-NTA resin (Superflow; Qiagen, Clifton Hill, Victoria, Australia) and 2 mL 10× dilution buffer (500 mM NaH₂PO₄, 1.5 M NaCl, and 100 mM imidazole) for 2 hours at 4°C. The mixture was loaded onto chromatography columns (Bio-Rad, Hercules, CA), washed twice with buffer (50 mM NaH₂PO₄, 300 mM NaCl, and 20 mM imidazole), and bound protein was eluted in 5 mL buffer containing 250 mM imidazole. The eluate was concentrated approximately 15 times by size exclusion on a 10-kDa membrane (Centricon YM-10 columns; Millipore, Billerica, MA). Samples were separated on a 12% SDS-polyacrylamide gel. Gels were stained with Coomassie blue for total protein detection or transferred to nitrocellulose membrane (Hybond ECL; Amersham Pharmacia Biotech, Buckinghamshire, UK) for detection by Western blot with an anti-polyhistidine monoclonal antibody (Clone HIS-1; Sigma-Aldrich). 

Demonstration of Anti-CD4 scFv Function by Flow Cytometry

Active anti-rat CD4 antibody fragment protein was detected by flow cytometry on CD4-positive rat thymocytes. Briefly, thymocytes were incubated with 50 μL of the test samples for 30 minutes at 4°C, followed sequentially by anti-polyhistidine monoclonal antibody, biotinylated anti-mouse antibody (DakoCytomation, Carpinteria, CA), and streptavidin-R-phycoerythrin conjugate (Molecular Probes, Eugene, OR). Fluorescence was measured on a flow cytometer (FACScan; BD Biosciences, Franklin Lakes, NJ). Serial dilutions of purified anti-CD4 scFv were used to create a standard curve. ScFv concentrations in test samples were estimated by using the linear portion of the curve. 

Antibodies and Purified Antibody Fragments

Murine anti-rat CD4 IgG was obtained from hybridoma culture supernatant (OX38, IgG_2a; no. 88051303; European Collection of Cell Cultures, Porton Down, UK) and stored at 4°C, sterile or with 0.05% sodium azide as a preservative. IgG-depleted hybridoma control supernatant was generated by binding antibodies to a protein G matrix (FastFlow Protein G Sepharose 4; Amersham Pharmacia Biotech, Uppsala, Sweden). The extent of depletion was measured in a mouse Ig ELISA, with a sheep anti-mouse Ig capture antibody and a sheep anti-mouse Ig horseradish-peroxidase–conjugated detection reagent (both from Silenus, Hawthorne, Victoria, Australia). Anti-CD4 scFv were purified at CSL Laboratories (Melbourne, Australia), as previously described ²³ to 2.6 mg/mL in buffer (20 mM HEPES and 50 mM NaCl) and stored at 4°C. 

One-Way Rat Mixed Lymphocyte Reaction

For mixed lymphocyte reaction (MLR) studies, rat mixed lymphocyte cultures were prepared as described elsewhere. ²⁹ Stimulator cells were prepared from F344 (syngeneic) or WF (allogeneic) thymus, irradiated to 25 Gy by an x-ray source, washed twice in 0.2% wt/vol bovine serum albumin in PBS (BSA-PBS) and stored for 18 hours at 4°C in RPMI 1640 supplemented with 20% freshly prepared heat-inactivated (56°C, 30 minutes) normal rat serum, 100 IU/mL penicillin, 100 μg/mL streptomycin sulfate, and 2 mM l-glutamine (RPMI-NRS). Responder cells were harvested from the lymph nodes and spleen of F344 rats and washed three times in BSA-PBS. Antibodies and scFv were dispensed into 96-well, round-bottomed plates. Responder (2 × 10⁵) and stimulator (4 × 10⁵) cells were added to appropriate wells in a final volume of 200 μL RPMI-NRS supplemented with 2 × 10⁻⁵ M 2-mercaptoethanol. Triplicate wells were set up for each treatment. Plates were cultured at 37°C in 5% CO₂ in air for 5 days, pulsed with 1 μCi [³H]thymidine (Amersham Biosciences, Buckinghamshire, UK) per well for 18 hours, harvested (Harvester 96 Mach III M; Tomtec, Hamden, CT), and counted in a liquid scintillation counter (1450 Microbeta; PerkinElmer Life Sciences, Boston, MA). 

Orthotopic Corneal Transplantation in the Rat

Rat corneal transplantation was performed as previously described, ³⁰ with the following modifications. The donor corneal rim was placed endothelium-side up on a sterile Teflon block and the central cornea was punched with a 3.1-mm diameter trephine. A 2.9-mm diameter trephine was used to score a partial thickness disc on the recipient cornea. The first full-thickness cut was made with a diamond knife, the dissection was completed with corneal scissors, and the recipient button was removed. The donor button was transferred endothelium down to the recipient bed and sutured in place. Chloramphenicol ointment (1%; Parke Davis, Caringbah, NSW, Australia) was applied to the eye directly after grafting and daily for 3 days after surgery. 

Injection of AdV into the Rat Anterior Chamber

Rats were anesthetized with halothane (Fluothane; Zeneca, Macclesfield, UK). The pupil was dilated with 10 mg/mL tropicamide (Mydriacyl; Alcon Laboratories, Frenchs Forest, NSW, Australia) and 10% phenylephrine hydrochloride (Neo-Synephrine; Abbott Laboratories, North Chicago, IL) and the eye immobilized with two conjunctival sutures. Paracentesis was performed in the anterior chamber with a 31-gauge noncoring needle (Hamilton Co., Reno, NV), and fluid was allowed to drain from the eye. Conjunctival sutures were removed to release pressure on the eye. With a syringe with a 31-gauge noncoring needle, 1 to 5 μL of AdV or control solution was injected slowly into the anterior chamber through the original opening. To prevent backflow, the needle was held in place for 30 seconds before being carefully removed. Chloramphenicol ointment was applied to the eye and the lids sutured shut for 18 to 24 hours. 

Postoperative Assessment and Statistical Analysis

After surgery, animals were examined daily by operating microscope. Grafts were scored for clarity and inflammation on a 0 to 4 scale in units of 0.1. Grafts with a clarity score >1.5 at day 7 (for ex vivo-treated corneas) or >1.0 at day 5 (for in vivo-treated corneas) were regarded as failures and excluded from analysis. Rejection was defined as the day on which graft clarity was ≥2.0 (when iris vessels were no longer clearly visible), and remained above 2.0 for 2 consecutive days. Statistical comparison between groups was performed with the Mann-Whitney test corrected for ties with adjustment for multiple comparisons. 

Results

ScFv Production by Rat Corneas Transduced with AdV In Vitro

Rat corneas (n = 43) were transduced ex vivo with AdV, organ cultured in vitro, and assayed for reporter gene expression by fluorescence microscopy. Hoechst 33258 staining allowed the visualization of all cell nuclei. Epithelial cells, stromal cells, and the endothelium were clearly distinguishable by nuclear morphology and size (Fig. 1) . 

Transduction of corneas with an optimized dose of 5 × 10⁷ pfu AdV induced eGFP reporter gene expression in 69% ± 23% of rat corneal endothelial cells after 4 days in organ culture, without decreasing endothelial cell density or affecting cell morphology. No eGFP-positive cells were noted in epithelial or stromal layers at any dose tested. At higher vector loads, cell morphology was affected and the number of apoptotic bodies increased. Reporter protein expression was detectable as early as day 1 after transduction and eGFP-positive cells were easily detected by day 4 after transduction. Fluorescence was still present after 7 days of organ culture; however, corneas did not tolerate extended organ culture, and tissue morphology of both treated and untreated corneas had deteriorated substantially by then. 

ScFv-encoding mRNA transcripts were detected in rat corneal tissue by RT-PCR at days 2 to 4 in corneas transduced with AdCD4GFP, but not in corneas transduced with control AdV (AdGFP; Fig. 2A ). ScFv protein (∼28 kDa) was detected in concentrated, affinity-purified cell supernatant of AdCD4GFP-transduced HEK-293 cells at day 7 by Western blot (Fig. 2B) . 

Supernatants from untreated corneas or corneas transduced with control vectors (AdGFP) showed no reactivity against the CD4-positive thymocytes by flow cytometry. In contrast, supernatants from corneas transduced with AdCD4GFP showed positive reactivity against thymocytes by flow cytometry (Fig. 3A) . ScFv specificity against CD4 was demonstrated by binding of scFv-containing supernatants to rat CD4 transfected Jurkat cells, without binding to untransfected Jurkat cells or rat erythrocytes (data not shown). Secretion of active scFv from AdCD4GFP-transduced rat corneas into culture supernatants peaked at 23 ± 4 ng scFv per cornea per day at days 4 to 5 after transduction, and declined to 8 ± 4 ng per cornea per day at 8 to 14 days (Fig. 3B) . After 14 days in organ culture, the tissue architecture was almost completely destroyed. 

Blockade of Alloproliferation in MLR by Antibody and Monomeric scFv against CD4

Whole anti-CD4 monoclonal antibody (OX38; 566 ng of IgG per well) blocked proliferation in MLR (Fig. 4A) . This activity was lost after Ig-depletion of the hybridoma supernatant (which reduced IgG concentration to 93 ng per well), indicating that the inhibition was specifically due to the antibody. Purified anti-CD4 scFv protein also inhibited proliferation at levels down to 26 ng per well (130 ng/mL) compared with matched volumes of buffer control (Fig. 4B) . 

Transplantation of Corneas Expressing Anti-CD4 scFv

To determine whether anti-CD4 scFv-expressing corneas would enjoy prolonged allograft survival, donor corneas were transduced ex vivo with 5 × 10⁷ pfu AdV for 2 hours immediately before grafting. Residual donor corneal rims were cultured for 3 days in vitro and processed for detection of transgene mRNA by RT-PCR or examined by fluorescence microscope for eGFP expression, to confirm successful transduction by the viral vector. In addition, supernatants from cultured corneal rims were tested for anti-CD4 activity by flow cytometry, when appropriate. Corneas treated with AdCD4GFP underwent rejection at the same time as AdGFP-treated corneas, with median survival times of 15 days and 16 days, respectively (Fig. 5A) . Medium-treated corneas were rejected earlier than all other treatment groups, at a median of day 12 after grafting (P < 0.05). In addition, three isografts that had been transduced ex vivo with AdCD4GFP remained thin and clear for 60 days. 

Grafted corneas and contralateral normal corneas from four rats that had received isografts transduced ex vivo with AdGFP before transplantation were retrieved at 7 days after transplantation and examined for eGFP mRNA. No transcripts for eGFP were found in the normal corneas, but all grafted corneas expressed eGFP transcripts (data not shown). These data indicate that ex vivo transduction of the donor cornea had not compromised corneal viability and that the transgene was actively being expressed in vivo for at least 7 days after transplantation. 

Given that scFv secretion by modified corneas did not peak until days 4 to 5 after transduction (Fig. 3B) , transplantation was delayed until 3 days after modification of the donor cornea. Because rat corneas do not tolerate extended organ culture, the donor cornea was transduced in vivo by injections of AdV into the anterior chamber of the donor rat, and corneas remained in situ until the time of grafting. Injection of AdV into the anterior chamber resulted in a mild to severe ocular inflammatory response exemplified by the presence of fibrin and cells and reduction in corneal clarity in some rats. Decreasing the number of AdV particles injected reduced the frequency of severe inflammatory events (suggesting that the inflammation was at least in part due to a reaction against the AdV preparations). Some corneas showed a cellular infiltrate at day 3 after injection, as shown by Hoechst 33258 staining of corneal flatmounts. Anterior chamber injection of 4 to 6 × 10⁷ pfu AdCD4GFP in vivo resulted in eGFP expression in 76% ± 15% of endothelial cells in the donor cornea by day 3 (Table 1) . Reducing the amount of AdV injected to 2 to 3 × 10⁷ pfu/eye did not significantly decrease the number of positive cells (56% ± 22%). Hence, to ensure a high level of scFv-expression while limiting the infiltration of inflammatory cells into the graft, donor corneas were transduced by anterior chamber injection of 2 to 5 × 10⁷ pfu AdV per eye in vivo and transplanted into recipients 3 days later. 

AdCD4GFP in vivo-treated corneas did not exhibit a survival advantage and were rejected at a median of 15 days, compared with AdGFP in vivo-treated control corneas (median survival time, 12 days; P = 0.05, Fig. 5B ). Groups receiving AdGFP in vivo-treated corneas rejected their grafts more swiftly than groups receiving corneas transduced ex vivo with the same treatment (median survival time, 16 days; P < 0.05). As an additional control, two isografts that had been transduced in vivo with AdCD4GFP remained thin and clear for 60 days. 

Discussion

This study demonstrated that gene-modified rat donor corneas secreted correctly folded, active scFv protein. Whereas secreted scFv fragments successfully blocked T-cell activation in vitro, expression of anti-CD4 scFv by the donor cornea alone was not sufficient to prolong corneal allograft survival in vivo. 

An important finding in this study was that monomeric anti-CD4 scFv blocked lymphocyte activation in MLRs. As previously described, ³¹ whole anti-CD4 antibodies blocked alloproliferation in MLRs. Anti-CD4 antibodies ameliorate disease in various T cell-mediated animal models including autoimmune disease ³² and allograft rejection. ⁷ The mechanisms through which anti-CD4 antibodies modulate disease are not clearly defined but include depletion, ^{5 33} maintenance of regulatory T-cell subsets, ^{34 35 36 37 38} steric hindrance, ^{39 40} negative signaling, ^{41 42} and modulation of cell surface expression. ⁴³ The complete blockade of MLR by anti-CD4 scFv showed that neither the dimeric nature nor the Fc region of the whole antibody was necessary to block alloproliferation in vitro. Only 26 ng of anti-CD4 scFv was needed to block allostimulation of 2 × 10⁵ responder cells in MLRs. This level is approximately equal to the level of anti-CD4 scFv produced daily by modified rat corneas in culture (23 ± 4 ng of scFv protein per day). 

The transfer of genes encoding anti-CD4 scFv to donor corneas immediately before transplantation had no effect on corneal allograft outcome. We thought that this was possibly a question of timing, because donor corneas take 4 or 5 days to begin secreting high levels of scFv (Fig. 3B) , whereas antigen presentation can occur within hours of corneal transplantation. ⁴⁴ Therefore, we transduced donor corneas in vivo and transplanted AdCD4GFP-treated corneas 3 days after modification (when they would be expressing high levels of anti-CD4 scFv). However, deferring the transplantation until the donor corneas secreted active scFv did not prolong survival to a significant extent compared with that of AdGFP-treated corneas (P = 0.05). In vivo modification of donor corneas by anterior chamber injection of AdV was necessary, because organ culture of normal, unmodified rat corneas, in contrast to human corneas, results in destruction of the tissue after a week in culture. Injection of control AdV (AdGFP) into the donor eye caused an increased tempo of rejection of the graft, compared with donor corneas transduced ex vivo with AdGFP (P = 0.003). Indeed, the flare of inflammation after AdV injection may have increased APC recruitment, making the graft more immunogenic. This acceleration in the rejection rate was not found in AdCD4GFP-treated grafts (P = 0.6). Possibly, local production of anti-CD4 scFv modified an inflammatory response to the injection, vector, or eGFP transgene. It is conceivable that the increased immunogenic status of the donor corneas swamped the potential action of secreted anti-CD4 scFv to some extent. Taken together, our data indicate that ocular expression of an alloinhibitory scFv is not sufficient to prevent corneal allograft rejection, indicating that the major site of allosensitization during corneal transplantation may be elsewhere. 

When considering a gene transfer-based intervention that specifically targets the afferent phase of T-cell activation, the anatomic site of sensitization is of importance. During corneal allograft rejection, sensitization may occur within the eye, ⁴⁵ the conjunctiva, ⁴⁶ and/or the regional lymph nodes. ⁴⁷ Recent evidence has shown the existence of functional ocular lymphatic drainage through uveoscleral and conjunctival routes and the importance of draining lymph nodes in corneal allograft rejection. ^{48 49} Bilateral removal of the superficial cervical lymph nodes increased corneal allograft survival from 50% to 100% in a murine model. ⁵⁰ Although the normal central cornea is avascular and essentially devoid of MHC class II molecule expression, populations of immature MHC class II-negative dendritic cells ⁵¹ or macrophages ⁵² have recently been described in the central cornea. These cells were shown to migrate to local lymph nodes, upregulate MHC class II expression, and participate in allosensitization very quickly after transplantation. ⁴⁹ It is possible that antigen presentation may occur at more than one site. Ocular expression of reagents that target the T cell-APC interaction during sensitization may be futile if these interactions occur primarily in the lymph nodes. 

Other studies have successfully prolonged corneal allograft survival by the ocular expression of immunomodulatory cytokines ⁵³ (Pleyer U, et al. IOVS 2000;41:ARVO Abstract 2882) or molecules against costimulatory molecules on APCs. ⁵⁴ At least some of these treatments may have succeeded because of their effects on resident ocular APCs (whether donor- or recipient-derived) before their migration to the lymph node. We suggest that therapies directly targeting the sensitization of the T cell during corneal allograft rejection may require expression at the regional draining lymph nodes, possibly in conjunction with ocular expression. 

 Supported by the National Health and Medical Research Council of Australia and the Ophthalmic Research Institute of Australia. CFJ is supported by the South Australian Joyner Scholarship in Medicine.

 Submitted for publication September 26, 2004; revised December 23, 2004, and January 28, 2005; accepted January 29, 2005.

 Disclosure: C.F. Jessup, CSL, Ltd. (F, I); H.M. Brereton, CSL, Ltd. (F, I); P.J. Sykes, None; M.A. Thiel, CSL, Ltd. (I); D.J. Coster, CSL, Ltd. (F, I); K.A. Williams, CSL, Ltd. (F, I)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Detection of corneal endothelial cells in whole rat cornea flatmounts. (A–C) Hoechst 33258–stained epithelial, stromal, and endothelial cells were detectable in flatmounted rat corneas under UV light. The different cell layers were observed by traversing through different planes within the same optical field. Efficiency of transduction was determined by examining the same field under (D) blue light (to visualize eGFP-positive cells) and (E) UV light (to visualize total endothelial nuclei). Original magnification, ×20.

Figure 2.

 

Detection of transgene product in AdV-transduced cells. (A) scFv PCR product (1 kb) was detected in 10-fold dilutions of cDNA (10⁻¹–10⁻³) prepared from corneas transduced with AdCD4GFP but not in corneas transduced with control AdV (AdGFP), at 2 days after transduction. eGFP PCR product (300 bp) was detected in both samples. PCR reactions performed on nontranscribed RNA were included as controls. Products were separated on a 1.5% agarose gel. (B) scFv protein (arrow; 28 kDa) was detected in affinity-purified, concentrated supernatant from HEK-293 cells transduced with AdCD4GFP by SDS-PAGE (SDS) and Western blot with anti-histidine antibody (WB). Supernatant from nontransduced cells did not contain scFv protein.

Figure 3.

 

scFv production by AdV-transduced rat corneas in vitro. (A) scFv concentration was determined in supernatants from corneas transduced with AdCD4GFP (▪) by flow cytometry on rat thymocytes. Supernatant from an AdGFP-transduced cornea (□) is shown as a negative control. (B) The production of scFv by rat corneas (n = 3) transduced with AdCD4GFP and cultured in vitro peaked at days 4 to 5 after transduction (solid line). Supernatant from a nontransduced cornea did not show any activity on any day (dashed line). Points show the mean ± SD.

Figure 4.

 

Inhibition of MLR with anti-CD4 antibody and scFv. (A) Whole anti-CD4 antibody (OX38; 566 ng/well) inhibited proliferation (counts per minute of incorporated [³H]thymidine) compared with no-treatment controls. Inhibitory activity was removed on Ig depletion. (B) Anti-CD4 scFv inhibited proliferation to negligible levels down to 26 ng/well. Volume-matched quantities of buffer had no effect on proliferation. Data are the mean of results in triplicate wells. Error bars, ±SD. Data were transformed (log₁₀) and analyzed by two-way ANOVA, followed by multicomparison post hoc tests on simple main effects. *P < 0.05 (with Bonferroni adjustment).

Figure 5.

 

Corneal allograft survival in rats receiving AdV-modified donor corneas. (A) Donor corneas modified ex vivo with AdCD4GFP immediately before graft did not experience prolonged survival (median survival time, 15 days) compared with control AdGFP-treated corneas (median survival time, 16 days). Median survival time of medium-treated control corneas was 12 days. (B) Transplantation of corneas secreting anti-CD4 scFv 3 days after in vivo transduction by anterior chamber injection of AdCD4GFP (median survival time, 15 days) did not significantly prolong corneal allograft survival compared with AdGFP-treated controls (median survival time, 12 days; P > 0.05).

Table 1.

 

Side Effects and Efficiency of Endothelium Transduction Resulting from Injection of the Adenoviral Vectors AdGFP and AdCD4GFP into the Anterior Chamber of Rats

AdV Injected (pfu/eye)	n	AdGFP: AdCD4GFP*	Day 1 Postinjection			Day 3 Postinjection
						Inflammatory Events, †	Corneal Clarity (range), ‡	% eGFP* Endothelial Cells, §	Endothelial Cell Density (/mm2), §	Other Nuclei (/mm2), § , ∥
5–8 × 106	5	2:3	2/5	0.0–0.5	48 ± 30	2818 ± 387	159 ± 145
2–3 × 107	5	2:3	2/5	0.2–0.8	56 ± 22	2375 ± 315	153 ± 122
4–6 × 107	7	1:6	5/7	0.6–1.5	76 ± 15	2059 ± 583	507 ± 465

*  Number of eyes injected with AdGFP or AdCD4GFP expressed as a ratio.

†  Number of eyes with significant amounts of fibrin in the anterior chamber detected by operating microscope.

‡  Corneal clarity scored on a 0–4 scale in 0.1 increments.

§  As calculated from five central fields (0.15 mm²) of Hoechst 33258-stained corneal flatmounts by fluorescence microscopy (mean ± SD).

∥  Other nuclei include apoptotic endothelial cells and infiltrating inflammatory cells.

The authors thank Kirsty Marshall for expert technical support, Ray Yates for animal husbandry, and Shirley Taylor, CSL Ltd., for providing highly purified anti-CD4 antibody fragment.