Inducible Adeno-Associated Virus Vector–Delivered Transgene Expression in Corneal Endothelium

Abstract

purpose. To investigate whether recombinant adeno-associated virus (rAAV) vector–mediated transgene expression is induced by inflammation in corneal endothelial cells in vivo.

methods. The ocular anterior chamber of New Zealand White rabbits was injected with rAAV-LacZ (10⁷ units of infection). Transient ocular anterior segment inflammation was induced by an intravitreal injection of lipopolysaccharide (LPS). The effect of inflammation on LacZ gene expression in corneal endothelial cells was evaluated by histochemical staining and reverse transcription–polymerase chain reaction (RT-PCR). The influence of rAAV on endothelial cell function was monitored by measuring corneal thickness.

results. Inflammatory reaction peaked at 1 day after LPS treatment and, at the same time, most of the endothelial cells (91.3% ± 7.2%) showed prominent LacZ gene expression. The transgene expression gradually diminished to basal level (3.4% ± 2.1%) when the inflammation subsided at 15 days after LPS treatment. The diminished transgene expression was efficiently reactivated to a high level (86.1% ± 8.7%) by a second LPS injection 60 days later. Moreover, the transgene expression remained low for a long period (60 days) in the absence of LPS treatment, but was increased to high levels (87.3%± 8.1%) 1 day after LPS treatment. Throughout the observation period, endothelial cell function remained intact.

conclusions. The rAAV vector can deliver genes into endothelial cells, and transgene expression is dramatically induced by inflammation. The rAAV-delivered transgene is stable and does not compromise endothelial cell function. Inducible rAAV-mediated transgene expression in corneal endothelial cells is a potential strategy in the treatment and prevention of ocular diseases.

The cornea is a transparent tissue that provides the principal refractive surface of the eye. ¹ Corneal endothelial cells possess a pumping function and play a critical role in the maintenance of corneal clarity. ² Loss of corneal endothelium cell function may result in loss of corneal transparency and visual acuity. ³ Corneal endothelium damage may result from age, trauma, inflammation, and inherited diseases. ^{4 5 6 7} Progressive endothelial cell loss results in endothelial decompensation, corneal edema, cloudy cornea and ultimate blindness. ⁸ At present, corneal transplantation is the mainstay approach for the treatment of blindness due to cloudy cornea. ⁹ However, the endothelium itself is a critical target of corneal graft rejection. ¹⁰ Recent developments in gene therapy may provide an alternative strategy to treat and prevent damage to corneal endothelial cells. ^{11 12} The potential applications of gene therapy include correction of the genomic anomalies responsible for inherited corneal endothelium diseases, modulation of immune response to prolong corneal graft survival, and introduction of an appropriate therapeutic gene to protect the endothelium itself from acquired damage such as intraocular inflammation and graft rejection. 

It has been reported that adenoviral-based vectors are able to deliver transgenes into endothelial cells, but short-term expression has been noted. ^{13 14} It has also been reported that a lentivirus-based vector can delivery transgenes into corneal endothelial cells, with long-term expression of 60 days. However, the biosafety of the lentivirus vector remains to be determined. ¹⁵ Adeno-associated virus (AAV) is a single-stranded, nonpathogenic virus. The recombinant (r)AAV vector represents a promising alternative to current viral delivery systems. Removal of all viral coding sequences (96% of the genome) eliminates the possibility of immune response to residual viral gene expression. ^{16 17} The rAAV genome can integrate into the host chromosome, facilitating long-term expression. ¹⁸ Recent studies have shown that gene therapy with the rAAV vector results in efficient delivery and long-term expression in a variety of tissues in vivo, such as brain, retina, and optic nerve. ^{19 20 21} Hudde et al. ²² have also reported that the rAAV vector can deliver a transgene into corneal endothelium for at least 30 days ex vivo, but with limited transgene expression. However, this may not reflect the potential of rAAV as a vector for in vivo gene therapy in corneal endothelial cells. rAAV-mediated transgene expression in corneal endothelial cells in vivo remains to be determined. 

Among the challenges in developing a gene therapy for the corneal endothelium is the achievement of an efficient, prolonged, and yet regulated gene expression in vivo. In our previous study, we observed that rAAV-delivered transgene expression regulated by the human cytomegalovirus (hCMV) promoter has a striking correlation with the inflammation process in joint tissues. ²³ Goater et al. ²⁴ have also reported that rAAV-mediated transgene expression can be activated by inflammation. Based on this, a rat model of disease-inducible gene therapy approaches for arthritis has been established. ²⁵ Corneal endothelial cells have been exposed to inflammatory cytokines in various ocular conditions, such as keratitis, anterior uveitis, and corneal graft rejection. ^{6 10 26} Therefore, we were interested in testing whether rAAV-mediated transgene expression regulated by the hCMV promoter is also inducible, so that a disease-inducible gene therapy for corneal diseases can be developed. 

In this study, the rAAV vector encoding the Escherichia coliβ -galactosidase (LacZ) gene, driven by a CMV promoter, was introduced into the anterior chamber of rabbit eyes by an intracameral injection. rAAV-mediated transgene expression in corneal endothelial cells was evaluated in the presence and absence of lipopolysaccharide (LPS)-induced inflammation. 

Materials and Methods

Animals and Experimental Ocular Anterior Segment Inflammation

New Zealand White rabbits weighing 2 to 3 kg were handled in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Ocular anterior segment inflammation was induced in the rabbits with an intravitreal injection of LPS (Sigma, St. Louis, MO). ²⁶ Rabbits were anesthetized by intramuscular injection of 35 to 50 mg/kg ketamine. LPS was dissolved in distilled water with gentle sonication and diluted to 2.0 ng/μL in phosphate-buffered saline (PBS). LPS (100 ng) was injected into the central vitreal space through the pars plana using a 30-gauge needle. After injection, gentamicin ointment was applied. 

In Vivo AAV Gene Delivery

The rAAV-LacZ viruses were synthesized as previously described. ^{17 23} Rabbits were prepared and anesthetized as usual. The eyelid was retracted, and 25 μL rAAV-LacZ virus (10⁷ units of infection) was injected into the anterior chamber by using a 1-cm beveled 30-gauge needle and a 50-μL syringe (Hamilton, Reno, NV). The injection was performed in a biosafety cabinet by inserting the needle tangentially into the chamber at the limbus. After rAAV-LacZ administration, gentamicin ointment was applied. 

Histochemical Detection of β-Galactosidase

Rabbits were killed with an overdose of pentobarbital sodium. The eyes were enucleated and prefixed by immersion in 4% paraformaldehyde-PBS (pH 7.4) on ice for 15 minutes. The eyeballs were rinsed in PBS two times and incubated in a dark room for 12 hours at 37°C with 5-bromo-4-chloro-3-indolyl-β-d-galactosidase (X-gal; Calbiochem, La Jolla, CA) in a solution containing 10 mM K₃Fe(CN)₆, 10 mM K₄Fe(CN)₆, 2 mM MgCl₂, 0.01% deoxycholate, and 0.02% NP40 in PBS (pH 7.8). Eyes were postfixed for 6 hours in 2% glutaraldehyde and 4% formaldehyde-PBS (pH 7.0), cryoprotected by sequential soaking in 10% and 30% sucrose solution, placed in optimal temperature cutting (OCT) compound (Miles, Elkhart, IN), snap frozen in liquid nitrogen and cut into 8-μm sections. The sections were counterstained with eosin and examined for the E. coli β-galactosidase (LacZ) gene signal. 

Histology

The rabbits’ eyes were enucleated and fixed in 4% paraformaldehyde for 24 hours. The fixed tissues were embedded in paraffin, cut into 5-μm sections and stained with hematoxylin and eosin. 

Reverse Transcription–Polymerase Chain Reaction

For RT-PCR, The corneas were isolated and frozen immediately in liquid nitrogen. The samples were homogenized in guanidine thiocyanate, and total RNA was extracted by phenol-chloroform. One microgram of the extracted RNA was purified by incubating with RNase-free DNase I, to eliminate vector DNA contamination, and then used for reverse transcription. cDNA was synthesized by using oligo(dT) primer and 200 IU transcriptase (SuperScript II; Gibco BRL, Gaithersburg, MD), according to the manufacturer’s instructions. The PCR procedure was performed as previously described. ²⁷ PCR amplification was performed with two primers 5′-CTGGATCAAATCTGTCGATCCTTCCCGCCC-3′ and 5′-CTGCTGCTGGTGTTTTGCTTCCGTCAGCGC-3′, which are expected to generate a 457-bp LacZ gene DNA fragment. The amplification products were separated by agarose gel electrophoresis, stained with ethidium bromide, and photographed. The housekeeping gene, β-actin, served as the control to ensure that equal amounts of RNA were analyzed from each sample. The sequence of upstream primer for β-actin was 5′-AGGCCAACCGCGAGAAGATGACC-3′, and the reverse primer was 5′-GAAGTCCAGGGCGACGTAGCAC-3′, which are expected to produce a 350-bp DNA fragment. 

Measurement of Central Corneal Thickness

Corneal endothelial cell function was monitored by measurement of central corneal thickness, using an ultrasonic corneal pachymeter (SonoGage, Cleveland, OH). After application of 1 drop of 0.5% proparacaine (Alcon, Fort Worth, TX), central corneal thickness was measured and recorded to the nearest hundredth of a millimeter. Triplicate measurements were performed at the indicated time for 3 months. Paired t-tests were used to determine the difference in central corneal thickness between experimental and control groups at each time point. All data are expressed as the mean ± SD. P < 0.05 was considered to be significant. 

Results

Induction of Ocular Anterior Segment Inflammation

In this study, the expression of rAAV-mediated transgenes in corneal endothelial cells was monitored by determining the percentage of cells that expressed the LacZ transgene. Because previous reports have indicated that the LacZ gene expression delivered by AAV vector is suppressed in the absence of inflammation in joint tissue, ^{23 28} establishment of experimental ocular inflammation is necessary for us to evaluate transgene expression in the presence of inflammation. In this study, transient ocular anterior segment inflammation was induced by an intravitreal injection of LPS (100 ng). The inflammation was evaluated by the presence of leukocyte infiltration and exudate accumulation in the tissues of the ocular anterior segment, such as the ciliary body (Figs. 1A 1C 1E 1G ) and corneal endothelial cells (Figs. 1B 1D 1F 1H) . Immediately before LPS injection, no leukocyte infiltration or exudate accumulation was observed in the ciliary body (Fig. 1A) and corneal endothelial cells (Fig. 1B) . At 1 and 5 days after LPS injection, inflammation was evident from the presence of leukocyte infiltration and exudate accumulation (Figs. 1C 1D 1E 1F) . The inflammation peaked 1 day after LPS injection (Figs. 1C 1D) and subsided gradually. At 5 days after LPS injection, the inflammation had partially subsided (Figs. 1E 1F) . At 15 days after LPS injection, no sign of inflammation was identified in the ocular anterior segment (Figs. 1G 1H) . 

rAAV-Mediated Transgene Expression in Corneal Endothelium

We then investigated rAAV-mediated gene expression in endothelial cells by using LacZ as a reporter gene under the condition of ocular anterior segment inflammation. To rule out the possibility that LacZ-positive signals were from lysosomal galactosidase in endothelial cells, control experiments without rAAV-LacZ injection was included. Briefly, 25 μL rAAV-LacZ (10⁷ units of infection) was injected into the anterior chamber of the right eye, and the same amount of PBS was injected into the anterior chamber of the left eye as a control (n = 12). Twenty-four hours later, inflammation was induced by an intravitreal injection of LPS in both eyes. One day after LPS injection, the eyeballs were removed, fixed, and reacted to X-gal. Transgene expression was determined by counting blue endothelial cells under a microscope in eight randomly selected high-power fields. Our observations indicate that LPS treatment resulted in an accumulation of LacZ-positive endothelial cells (91.3% ± 7.2%) in those eyes injected with rAAV-LacZ (Figs. 2A 2B ) but not in those eyes injected with PBS (Figs. 2C 2D) . Thus, the possibility that LacZ-positive signals resulted from the activation of endogenous β-galactosidase was excluded. 

In a separate group of animals (n = 12), we injected rAAV-LacZ into the eyes without inducing prior inflammation by LPS injection, and only a few LacZ-positive cells (3.4%± 2.1%) were observed (Figs. 2E 2F) . This finding was drastically different from observations in those eyes that received both rAAV-LacZ and LPS treatment (Figs. 2A 2B) . Our result suggests that the rAAV-mediated transgene was suppressed in the absence of inflammation. Again, no LacZ-positive cells were observed in the eyes without rAAV-LacZ and LPS injection (Figs. 2G 2H) . In addition to corneal endothelial cells, we found that inflammation enhanced LacZ transgene expression in the cells of trabecular meshwork and iris epithelial cells. However, noβ -galactosidase activity was detected in lens epithelial cells (data not shown). 

To further characterize the induction of transgene expression by LPS treatment, a time-course analysis was performed. Both eyes of each rabbit (n = 12) were treated with rAAV-LacZ. Twenty-four hours later, the right eye was injected with LPS, and the contralateral eye was injected with PBS as a control in each animal. At the indicated time, the eyes were removed, fixed, and reacted to X-gal. Immediately before LPS injection, only 2.4% ± 1.7% corneal endothelial cells showed LacZ-positive staining (Fig. 3A ). At 1 day after LPS injection, the inflammation reached a peak, and most of the corneal endothelial cells (91.3 ± 7.2%) showed LacZ-positive staining (Fig. 3C) . At 5 days after injection, the inflammation subsided partially and LacZ-positive endothelial cells were reduced to 47.6% ± 9.8% (Fig. 3E) . At 15 days after injection, no sign of inflammation was identified in the ocular anterior segment, and transgene expression decreased to 4.3% ± 3.1% (Fig. 3G) . In the eyes without inflammation, only approximately 1% to 5% of endothelial cells showed LacZ-positive stain before or at 1, 5, and 15 days after PBS injection (Figs. 3B 3D 3F 3H) . Our results indicate that rAAV-mediated transgene expression can be induced by inflammation and that expression is closely correlated with LPS-induced intraocular inflammation. 

Reactivation of rAAV-Mediated Transgene Expression Driven by CMV Promoter by Repeated Inflammation

In this study, we observed diminished numbers of blue-stained cells in accordance with the subsidence of LPS-induced inflammation. There are two possibilities to explain this phenomenon: One is that the transgene may be deleted or transduced cells may undergo programmed cell death. Under this mechanism, re-exposure of cornea to LPS would not increase LacZ-positive cells. Another possibility is that the transgene remains stable, but gene expression is suppressed. Under this mechanism, re-exposure of the cornea to LPS would reactivate gene expression and therefore increase LacZ-positive signals. To test these possibilities, another episode of intraocular inflammation was induced by a second LPS injection 60 days after rAAV-LacZ injection. Briefly, both eyes of each rabbit (n = 12) were treated with rAAV-LacZ (10⁷ units of infection). Twenty-four hours later, intraocular inflammation was induced by LPS injection in both eyes. At 60 days after rAAV-LacZ injection when the transient inflammation had already subsided, LPS reinjection was performed in the right eye, and the contralateral eye was injected with PBS as a control. Before LPS reinjection, only 3.4% ± 2.1% of endothelial cells showed a LacZ-positive signal (Fig. 4A ). At 1 day after reinjection, most of the endothelial cells (86.1% ± 8.7%) showed a positive LacZ signal (Fig. 4C) . At 5 days after reinjection, LacZ-positive endothelial cells were moderately reduced to 43.4% ± 7.6% (Fig. 4E) . At 15 days after reinjection, no sign of inflammation was identified in the ocular anterior segment, and transgene expression decreased to 3.7% ± 2.2% (Fig. 4G) . In those eyes without a second inflammation, the LacZ-positive endothelial cells remained low (2%–5%) before or at 1, 5, and 15 days after PBS injection (Figs. 4B 4D 4F 4H) . Our results suggest that the diminished number of blue-stained cells, in accordance with the subsidence of inflammation, was due to suppression of the gene expression rather than to loss of transduced cells. 

Activation of rAAV-Mediated Transgene Expression Driven by CMV Promoter by Delayed Inflammation

To further confirm the stability of the rAAV-delivered transgene in the corneal endothelial cells, the induction of gene expression was delayed for 60 days after rAAV-LacZ injection. In this study, both eyes of each rabbit (n = 12) were treated with rAAV-LacZ. At 60 days after rAAV-LacZ injection, delayed inflammation was induced by LPS intravitreal injection in the right eye; the contralateral eye was injected with PBS. Before LPS injection, LacZ-positive endothelial cells remained ay 3.3%± 2.4% (Fig. 5A ). In the eyes that received LPS injection, 87.3% ± 8.1% endothelial cells had dark blue transgene stains (Fig. 5C) at 1 day after LPS injection. At 5 days after injection, the percentage of LacZ-positive endothelial cells was moderately reduced to 44.6% ± 7.1% (Fig. 5E) . At 15 days after injection, no sign of inflammation was identified in the ocular anterior segment, and the percentage of endothelial cells with transgene expression decreased to 3.6% ± 2.7% (Fig. 5G) . In those eyes that did not received LPS injection, only approximately 2% to 5% of the endothelial cells showed LacZ-positive staining immediately before or at 1, 5, and 15 days after PBS injection (Figs. 5B 5D 5F 5H) . Our results indicate that the rAAV vector delivered the transgene into the corneal endothelial cells and expression remain low without inflammation. The transgene was stably maintained in corneal endothelial cells for at least 60 days and still could be activated by delayed inflammation. 

RT-PCR Analysis

To further confirm the effect of inflammation on LacZ transgene expression in corneal endothelial cells, RT-PCR analysis was performed. In this study, the anterior chamber of each rabbit eye was injected with rAAV-LacZ (10⁷ units of infection). Twenty-four hours later, LPS was injected into the vitreal space in each rabbit eye. The mRNA in the corneal endothelial cells was extracted for RT-PCR analysis immediately before or at 1 and 5 days after LPS injection. Immediately before LPS injection, no LacZ gene expression was found (Fig. 6 , lane 2). However, gene expression was observed at 1 and 5 days after LPS insult (Fig. 6 , lanes 3, 4). In a separate group, the anterior chamber of each rabbit eye was injected with rAAV-LacZ. Sixty days later, delayed intravitreal injection with LPS was performed in each rabbit. The mRNA in the corneal endothelial cells was extracted for RT-PCR analysis immediately before or at 1 and 5 days after delay LPS injection. Immediately before the delayed LPS injection, only weak LacZ gene expression was observed (Fig. 6 , lane 5). However, prominent gene expression was found at 1 and 5 days after delayed LPS insult (Fig. 6 , lanes 6, 7). Our results further confirm that expression of an rAAV-mediated transgene can be induced by inflammation and that the transgene expression remains stable in corneal endothelial cells for at least 60 days. 

The Effect of rAAV on Corneal Endothelial Cell Function

In this study, the pumping function of corneal endothelial cells was evaluated by examining central corneal thickness. The anterior chamber of the right eye was injected with 25 μL rAAV-LacZ (10⁷ units of infection), and the contralateral eye was injected with the same amount of PBS as a control in each rabbit (n = 12). Throughout the experimental period, the central corneal thickness was determined by pachymeter at the indicated times. In those eyes that received an injection of rAAV-LacZ, the central corneal thickness was 407 ± 17, 421 ± 15, 418 ± 19, 417 ± 11, 412 ± 14, 413 ± 18, and 412 ± 16 μm before and at 1, 5, 15, 30, 60, and 90 days after rAAV-LacZ injection, respectively. In those eyes that received an injection of PBS, the central corneal thickness was 411 ± 10, 422 ± 13, 416 ± 16, 419 ± 13, 415 ± 22, 410 ± 13, and 414 ± 13 μm before and at 1, 5, 15, 30, 60, and 90 days after PBS injection, respectively. No significant difference in corneal thickness was found between the eyes injected with rAAV-LacZ and vehicle (PBS) at each time point (Fig. 7) . Our results suggest that the pumping function of endothelial cells is not compromised by the rAAV vector. 

Discussion

In this study, we observed a striking correlation between the transgene expression and the severity of the ocular anterior segment inflammation. At the peak of the inflammation (1 day after LPS injection), most of the endothelial cells (91%) had high level transgene expression (Fig. 3C) , which diminished to the basal level of 3% when inflammation subsided at 15 days after LPS treatment (Fig. 3G) . Re-exposure of endothelial cells to a second injection of LPS after recovery from the initial LPS-induced inflammation led to a dramatic reactivation of transgene expression (Fig. 4C) . The pattern of reactivated transgene expression induced by a second LPS injection was similar to that induced by the first treatment (Fig. 3C) . We also observed that the LacZ gene could be delivered by rAAV and remain inactive in endothelial cells for 60 days (Fig. 5A) , but still could be efficiently induced by LPS treatment (Fig. 5C) . The findings of RT-PCR analysis further confirmed these observations. 

Although the exact mechanism of activation of transgene expression by inflammation remains unclear, there are a couple of possible mechanisms to explain our observations. It has been established that the rate-limiting step of rAAV-mediated gene expression is the second-strand synthesis that can be facilitated by DNA repair. ^{28 29} It is also known that inflammation leads to DNA damage of the host cell genome, which results in the active production of DNA repair enzyme and cofactor. ^{30 31} In this study, we observed that rAAV-mediated transgene expression was activated by LPS-induced inflammation. Besides, the induction of transgene expression in the presence of inflammation (within 3 days after rAAV vector delivery) was faster than that which occurred in the absence of inflammation (7–14 days after rAAV vector delivery). DNA repair may thus be one of the mechanisms leading to this rapid induction. However, this possibility has not been confirmed, because we did not measure DNA repair in the corneal endothelial cells. Moreover, the DNA repair mechanism is not enough to explain the facilitation of gene expression by inflammation that was induced at 60 days after the rAAV-LacZ injection (Fig. 5 ; Fig. 6 , lanes 5, 6, 7). In another plausible mechanism, genes could be delivered into endothelial cells, but the gene expression may be suppressed in the absence of inflammation, and inflammation enhances gene expression through CMV promoter activation. In this study, we observed that LacZ gene expression driven by CMV promoter was induced by LPS injection at 60 days after rAAV-LacZ infection (Fig. 5 ; Fig. 6 , lanes 5, 6, 7). This finding indicates that corneal endothelial cells are transduced by the rAAV vector, and the transgene driven by the CMV promoter is activated by LPS insult. 

Because this group of rabbits were not injected with LPS at 24 hours after rAAV-LacZ infection, this result further suggests that the rAAV vector delivers transgenes into the corneal endothelial cells in the absence of inflammation and the transgene then remains inactive without LPS insult. The very low number of endothelial cells that expressed β-galactosidase activity (Fig. 5A ; Fig. 6 , lane 5) in the corneas of eyes that received the rAAV-LacZ but did not receive LPS injection may, therefore, be attributable to the low activity of the CMV promoter, not the absence of infection by the rAAV vector. Recent studies have also reported that the CMV promoter is activated by LPS, and the LPS-induced NFκB-signaling pathway has been proposed to be the mechanism of CMV promoter activation. ^{23 32} NFκB is a principle transcription factor responsible for CMV promoter activation. ³³ In general, NFκB, which is constitutively expressed in cytoplasm, is bound to the inhibitor IκB and remains inactive. Only when IκB is degraded can NFκB be released and become functional for transcription activation. ^{34 35} It also has been reported that degradation of IκB can occur in the presence of LPS or proinflammatory cytokines (IL-1, TNF), and then NFκB can be released from the NFκB–IκB complex for CMV promoter activation. ³⁶ Therefore, we observed that inflammation resulted in the activation of transgene expression. 

The corneal endothelium is unique in its nonproliferative and nonreplenishable nature, ² and preservation of cell viability is therefore of the utmost importance. The injection of rAAV vector into the anterior chamber did not appear to lead to endothelial cell damage, as was evident from corneal thickness measurements throughout the 90-day observation period (Fig. 7) . Our results indicate that the rAAV vector is a relatively safe gene delivery system for corneal endothelial cells. In addition, the inducible rAAV-mediated transgene expression that we observed in this study may offer a means to minimize potential side effects. In our study, we also observed that transgene expression was enhanced and reactivated by IL-1-induced uveitis (data not shown). A therapeutic gene product can thus be synthesized only when cells are experiencing insults involving inflammatory cytokines such as IL-1 and TNF. In the future, appropriate therapeutic transgenes could be activated in endothelial cells to protect the cells from damage induced by various diseases, such as anterior uveitis. 

To sum up, our results suggest the rAAV vector is capable of delivering transgenes into corneal endothelial cells. The stable gene delivery and inducible gene expression indicates that gene therapy with the rAAV vector encoding the appropriate gene is a potential strategy in the treatment and prevention of ocular diseases. 

 Supported by the Department of Health Grants DOH89-TD-1144 and DOH90-TD-1029.

 Submitted for publication June 11, 2001; revised September 25, 2001; accepted October 2, 2001.

 Commercial relationships policy: N.

 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.

 

Establishment of ocular anterior segment inflammation. Ocular anterior segment inflammation was induced in rabbit eyes by intravitreal injection of LPS (100 ng). The eyes were enucleated immediately before or at 1, 5, and 15 days after LPS injection. The inflammation was evaluated by the presence of leukocyte infiltration and exudate accumulation in the tissues of the ocular anterior segment, such as the ciliary body (A, C, E, G) and corneal endothelial cells (B, D, F, H). Immediately before LPS injection, no leukocyte or exudate was observed in the ciliary body (A) and corneal endothelial cells (B). At 1 and 5 days after LPS injection, inflammation was evident from the presence of leukocyte infiltration (arrowhead in C and E) and exudate accumulation (arrow in D and F). The inflammation peaked at 1 day after LPS injection (C, D) and subsided gradually. At 5 days after LPS injection, the inflammation had partially subsided (E, F). At 15 days after LPS injection, no sign of inflammation was identified in the ocular anterior segment (G, H). Bar, 100 μm.

Figure 2.

 

rAAV-mediated LacZ gene expression in corneal endothelial cells. In each rabbit of the experimental group, the anterior chamber of the right eye was injected with 25 μL rAAV-LacZ (10⁷ units of infection) and the contralateral eye with the same amount of PBS. Twenty-four hours later, LPS was injected into those eyes that had been previously treated with rAAV-LacZ (A, B) or PBS (C, D). In each rabbit of the control group, the anterior chamber of the right eye was injected with 25 μL rAAV-LacZ (10⁷ units of infection) and the contralateral eye with the same amount of PBS. Twenty-four hours later, PBS was injected into those eyes that had been treated with rAAV-LacZ (E, F) or PBS (G, H). Twenty-four hours after the second injection, all eyes in the experimental and control groups were removed, fixed, reacted with X-gal, and counterstained with eosin. In the eyes treated with rAAV-LacZ and LPS (A, B), most of the corneal endothelial cells showed a LacZ-positive signal (arrowheads), but no significant LacZ transgene expression was found in the epithelial and stromal tissue (arrow). Magnification: (A, C, E, G) ×100; (B, D, F, H) ×400. Bar, 50 μm.

Figure 3.

 

Correlation between the inflammation process and rAAV-mediated transgene expression. The anterior chambers of rabbit eyes were injected bilaterally with 25 μL rAAV-LacZ (10⁷ units of infection). Twenty-four hours later, LPS was injected into the vitreal space in the right eye (A, C, E, G), and the contralateral eye was injected with the same amount of PBS as a control (B, D, F, H) in each rabbit. The eyes were enucleated, reacted with X-gal, and counterstained with eosin, immediately before (A, B) or at 1 (C, D), 5 (E, F), and 15 (G, H) days after LPS injection. The LacZ transgene expression in corneal endothelial cells was closely correlated with the LPS-induced inflammation process (A, C, E, G). Bar, 50 μm.

Figure 4.

 

rAAV-mediated transgene expression was reactivated by a second inflammation. The anterior chambers of rabbit eyes were injected bilaterally with rAAV-LacZ (10⁷ units of infection). Twenty-four hours later, LPS was injected into the vitreal space in both eyes. At 60 days after rAAV-LacZ injection, a second LPS injection was performed in the right eye (A, C, E, G), and the contralateral eye was injected with PBS as a control (B, D, F, H). The eyes were enucleated, reacted with X-gal, and counterstained with eosin, immediately before (A, B) or at 1 (C, D), 5 (E, F), and 15 (G, H) days after the second LPS injection. LacZ transgene expression in corneal endothelial cells was reactivated by a second inflammation (A, C, E, G). Bar, 50μ m.

Figure 5.

 

rAAV-mediated transgene expression was activated by delayed inflammation. The anterior chambers of rabbit eyes were injected bilaterally with rAAV-LacZ (10⁷ units of infection). At 60 days after rAAV-LacZ injection, delayed inflammation was induced by LPS intravitreal injection in the right eye (A, C, E, G), and the contralateral eye was injected with PBS as a control (B, D, F, H). The eyes were enucleated, reacted with X-gal, and counterstained with eosin, immediately before (A, B) or at 1 (C, D), 5 (E, F) and 15 (G, H) days after LPS injection. LacZ transgene expression in corneal endothelial cells was activated by delayed inflammation (A, C, E, G). Bar, 50 μm.

Figure 6.

 

RT-PCR analysis. The anterior chambers of rabbit eyes were injected with 25 μL rAAV-LacZ (10⁷ units of infection). Twenty-four hours later, LPS was injected into the vitreal space in each rabbit. The mRNA in corneal endothelial cells was extracted immediately before (lane 2) or at 1 (lane 3) and 5 (lane 4) days after LPS injection and subjected to RT-PCR analysis. In another group, the anterior chambers of rabbit eyes were injected with rAAV-LacZ (10⁷ units of infection). At 60 days after rAAV-LacZ injection, delayed inflammation was induced in the eyes by intravitreal injection of LPS. The mRNA in corneal endothelial cells was extracted immediately before (lane 5) or at 1 (lane 6) and 5 (lane 7) days after LPS injection and subjected to RT-PCR analysis. Lane M: marker; lane 1: naive rabbit eye.

Figure 7.

 

The effect of the rAAV vector on corneal endothelial cell function. The anterior chamber of the right eye was injected with 25 μL rAAV-LacZ (10⁷ units of infection), and the contralateral eye was injected with the same amount of PBS as a control in each rabbit (n = 12). The effect of the rAAV vector on corneal endothelial cell function was evaluated by determining the difference in central corneal thickness between experimental and control eyes. Central corneal thickness was independently measured by pachymeter in rAAV-LacZ – and PBS-injected eyes of each rabbit (n = 12) at the indicated time. No statistically significant difference was observed between the two groups at each time point. Day 0: measurement immediately before injection. Error bars, SD.

The authors thank Ru-Yu Pan, Dai-Wei Lu, and Jyh-Horne Wang for technical support.