Herpes Simplex Virus–Mediated Gene Delivery to the Rodent Visual System

Abstract

purpose. To determine the types of cells in the visual system of the mouse and rat that can express a transgene delivered by an attenuated replication competent Herpes simplex virus-1 (HSV-1) vector.

methods. C57/BL6 × BALB/C mice and Albino rats were treated with 1 × 10⁷ pfu of the HSV-1 ribonucleotide reductase mutant (hrR3) expressing the Escherichia coli lacZ gene. The hrR3 virus was delivered by topical application to the cornea, intravitreal (IV) injection, intracameral injection (IC), or stereotactic injection into the visual cortex (VC). At specified times postinfection, animals were killed and tissues were removed, fixed, sectioned, and stained with X-gal or hematoxylin and eosin for histochemical and histopathologic examination.

results. Topical delivery after corneal scarification in both mouse and rat resulted in lacZ expression in 25% of the corneal epithelial cells and 25% of the retinal pigment epithelium (RPE) cells. Topical application without concurrent scarification also resulted in transgene delivery to 20% of the RPE cells of the rat but not the mouse. IV injection resulted in expression primarily in RPE cells, with up to 100% of the cells expressing lacZ in the mouse and rat. Other cells expressing the transgene included ciliary body (CB) and optic nerve cells. Up to 25% of the retinal ganglion cells in the rat expressed lacZ, but only rarely in mice. IC delivery in rats resulted in expression in trabecular meshwork, CB cells, RPE, and iris epithelium. Injection into area 17 of the rat VC resulted in efficient labeling of the VC neurons and neurons in the lateral geniculate nucleus without any evident pathology or inflammation. Neither inflammation nor disease pathology was observed in either the mouse or rat after any route of delivery.

conclusions. It was demonstrated that the hrR3 HSV-1 virus can deliver a functioning gene to several cell types in the eye and neurons in the VC and that the virus can move via retrograde transport to nuclei that project to the VC.

Gene delivery to cells of the eye is a promising technology for the treatment of a number of pathologic conditions. Viruses can efficiently transfer genes to cells, and several viral based vector systems have been used. Adenoviruses (AVs) have been widely used for transgene delivery because of their ease of preparation and ability to infect a wide range of cells. Delivery of recombinant AV to the eye via subretinal injection results in reporter gene expression in the retinal pigment epithelium (RPE) cells with limited expression in photoreceptor cells at the site of injection. ^{1 2 3 4} Delivery via intravitreal injections with AV vectors results in reporter gene expression in retinal ganglion cells (RGCs), photoreceptor cells, and RPE cells ^{3 5 6} and delivery via intracameral injection can deliver a reporter gene to the trabecular meshwork, iris pigment epithelium. and corneal endothelium. ^{7 8} Delivery of an AV vector to the superior colliculus of mice also results in reporter gene expression in the RGCs. ⁹ Therapeutically, recombinant AV vectors have been used to deliver genes to the retina or optic nerve after intravitreal and subretinal injections. ^{1 10 11 12 13 14} However, AV transgene delivery has been limited because of the host immune response to viral expressed proteins. ^{15 16 17} Although the eye is a relatively immunologically protected site (reviewed in ^{Ref. 18)} intracameral injection of a recombinant AV in the rabbit elicited a pathogenic immune response that led to tissue destruction of the corneal endothelium in the anterior chamber. ⁸ Similar results have also been reported in a transfused human eye model. ¹⁹  

Adeno-associated virus (AAV) has also been used to deliver transgenes because of its ability to infect a wide range of human cell types and lack of pathogenicity of wild-type virus. However, because of the small genome size, AAV vectors are limited to delivery of less than 8 kB of transgene, leaving little room for inclusion of large genes and promoters or upstream regulatory elements. Gene delivery to the eye using AAV via subretinal or intravitreal injection has resulted in transgene delivery and expression in the RPE cells and to the photoreceptors, RGCs, and Müller cells; however, the delivery was limited to the area of the subretinal injection. ^{20 21 22 23} These subretinal injections also result in transient retinal detachment that usually resolves within a few days; however, damage can persist at the site of injection. In addition, AAV vectors have been used to deliver therapeutic genes to photoreceptor cells and cells in the optic nerve via subretinal ¹³ or intravitreal ²⁴ injection. 

Murine retroviral based vectors (mRV) have been used to deliver genes to cells of the eye ex vivo ²⁵ and in vivo ^{26 27} ; however, the use of mRVs is limited because of their inability to infect terminally differentiated cells. Recently, human immunodeficiency virus (HIV)–based retroviral vector systems have been used to deliver genes to the retina via subretinal injection. ^{28 29} HIV vectors are able to efficiently deliver a transgene to photoreceptor and RPE cells of neonatal rats; however, delivery to cells in adult rats was much less efficient. In addition, subretinal injections of the HIV-based vector produced a local retinal detachment and resulted in infiltration of inflammatory cells. Therapeutically, retroviral delivery of the HSV thymidine kinase (TK) gene together with ganciclovir has been used to inhibit proliferative vitreoretinopathy in a rabbit model. ²⁷  

Several herpes simplex virus (HSV) vectors have been developed for gene delivery; however, to date, there are no published reports of their use for gene delivery in the eye. The two most frequently used HSV-based vectors include the following: replication defective HSV, with deletions in immediate-early genes such as ICP0, ICP4, or ICP27, ^{30 31 32 33} and amplicon vectors, which are deleted for all essential HSV genes. ^{34 35} Amplicons require a helper virus to complement HSV gene expression, and, like replication defective viruses, require a helper cell line for propagation. Some amplicon systems suffer from contamination by helper virus, ³⁶ and the production of high titer stocks with amplicon vectors is difficult and requires multiple passages in vitro, adding to the time of production. Because of the defective nature of these vectors, they are also unlikely to cross synapses in the nervous system. 

One potential HSV-based vector system, which would avoid the use of helper virus and helper cell lines, is a replication competent, attenuated virus. Such viruses are deleted for an essential in vivo virulence gene but retain the ability to be propagated in vitro. Thus, these viruses can easily be grown and concentrated to high titer (>10⁹ pfu/ml), don’t require helper cell lines or helper viruses; and because these viruses may retain limited replication capacity in vivo, they may be able to cross neuronal synapses. Limited replication by these vectors also may allow delivery to adjacent cells, thereby expanding the number of cells expressing the transgene. Studies with replication competent attenuated viruses with deletions of the TK ³⁷ or US3 ³⁸ gene showed transgene expression when injected into the central nervous system (CNS); however, this was accompanied by cytopathology at the site of injection. Injection of a replication attenuated HSV-1 virus with a deletion of the γ₁34.5 gene caused severe inflammation in the CNS. ³⁹ Although these viruses have not been tested in the eye, these results suggest that TK, US3, andγ ₁34.5 mutants may not be sufficiently attenuated for use as gene delivery vectors. 

The HSV ribonucleotide reductase (RR) gene is an important virulence gene. The RR protein is composed of a heterodimer of the UL39 and UL40 gene products and functions to increase the cellular pool of dNTPs by catalyzing the conversion of ribonucleotides to deoxyribonucleotides. ⁴⁰ Although HSV RR has been shown to be nonessential in tissue culture, it is important for viral growth in slowly growing and/or terminally differentiated cells. ^{41 42} In addition we and others have shown that RR null viruses are unable to cause disease after corneal or intravitreal administration. ^{43 44 45} Ribonucleotide reductase mutants are also impaired in their ability to cause death in mice after intracerebral or intraperitoneal injection. ^{44 46 47} Moreover, although the virus can establish a latent state in sensory nerve ganglia, RR mutants are unable to reactivate from latency. ^{45 47 48} These findings suggest that an RR mutant HSV-1 virus might be a useful vector for gene delivery to cells of the visual system. 

The work presented here shows that delivery of the RR mutant HSV-1 virus hrR3 to the rodent eye results in efficient gene delivery and expression in the RPE cell layer regardless of the route of administration. Surprisingly, delivery of the lacZ gene to RPE cells was achieved by topical application to the intact cornea, suggesting delivery via eye drops and patient self-administration ration may be feasible. Efficient gene delivery and expression also occurred in the ciliary body, trabecular meshwork, and corneal epithelium. Other ocular cell types also expressed the lacZ transgene but at low efficiency. In particular, photoreceptors were refractory to gene delivery. Injection of the hrR3 virus by various routes into the mouse or rat eye showed no signs of ocular pathology and little or no inflammation. In addition, injection of the hrR3 virus into the visual cortex of rats resulted in lacZ expression in neurons at the site of injection (area 17) and in the lateral geniculate nucleus (LGN), which projects to the VC, with little or no cytopathology or inflammation. These results demonstrate the feasibility of using attenuated HSV vectors lacking RR for visual system gene delivery. 

Methods

Tissue Culture and Virus Preparation

African green monkey kidney cells (Vero) were grown in Dulbecco’s modified Eagle’s medium supplemented with 5% serum (1:1 mixture of fetal bovine serum and defined, supplemented calf serum), 100 U/ml penicillin, and 100 μg/ml streptomycin sulfate as described previously. ⁴⁹ The construction of the hrR3 virus has been described previously ⁴¹ ; briefly, the E. coli lacZ gene was cloned in frame with UL40 (RR large subunit) so that expression of lacZ is controlled by the UL40 promoter with subsequent deletion of most of the UL40 gene reading frame. High titer stocks of the virus were prepared on Vero cells as previously described. ⁴⁹ Virus stocks were purified for in vivo delivery by high-speed centrifugation through a 36% sucrose cushion in RSB (10 mM Tris, pH 7.4, 10 mM NaCl, 3 mM MgCl₂) and then resuspended in phosphate-buffered saline (PBS). ⁵⁰ Titers were determined by plaque assay on Vero cells. 

In Vivo Gene Delivery

For ocular delivery of hrR3, C57/BL6 × BALB/C adult mice (gift of Dr. Daniel Albert, University of Wisconsin, Madison, WI) and albino adult rats (350 g; Harlan Sprague–Dawley, Indianapolis, IN) were used. Mice were anesthetized with 5% halothane by inhalation and the rats with intramuscular injection of xylazine (9 mg/kg) and ketamine–HCl (90 mg/kg). Corneal scarification was performed as previously described. ⁴⁹ Briefly, the cornea was scarified with a 30-gauge needle, and then 5 μl of the hrR3 virus suspension containing 1 × 10⁷ pfu in PBS was placed on the cornea. Direct corneal delivery of the hrR3 virus was also performed by placing a 5 μl drop of virus suspension containing 1 × 10⁷ pfu directly on the cornea without concurrent scarification. After virus delivery to the cornea, the eyelid was held closed for 30 seconds, and the animals were returned to their cages. 

For intravitreal delivery, mice or rats were anesthetized as described above, and a specified volume of the hrR3 virus in PBS was injected into the vitreal cavity through a transcleral approach 2 mm posterior to the cornea–sclera junction with a 30-gauge needle and Hamilton syringe, taking care not to disturb the lens or the retina. The needle was left in the eye for 10 seconds to allow ocular pressure to stabilize. For intracameral administration, animals were anesthetized as above and then injected with the hrR3 virus in PBS into the anterior chamber via a transcorneal approach with a 30-gauge needle and Hamilton syringe, taking care not to damage the lens or iris. All animal procedures conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and NIH guidelines for responsible care and use of animals and were approved by the institutional IACUC. 

Tissue Processing

At specified times after gene delivery, the animals were killed by cervical dislocation under halothane anesthesia for mice or by CO₂ asphyxiation for rats. Eyes were removed and prefixed in 10% formalin in PBS for 1 hour at room temperature and then incubated in X-gal solution (1 mg/ml X-gal, 16 mM potassium ferrocyanide, 16 mM potassium ferricyanide, 2 mg/ml MgCl₂) for 48 hours as we described previously. ⁵¹ The eyes were then postfixed in 2.5% gluteraldehyde and 1% formaldehyde for 1 hour and then stored in a solution of 1% formaldehyde and 6% sucrose in PBS at 4°C. The eyes were then embedded in paraffin, sectioned (10-μm-thick) along the pupil–optic nerve axis and counterstained with nuclear fast red. 

Quantification of X-gal Staining

Sections were observed microscopically, and tissue or cell types were scored for the amount of X-gal staining as follows: 0 = 0% of cells staining blue, 1 = less than 25% of cells staining blue, 2 = greater than 25% but less than 50% of cells staining blue, 3 = greater than 50% but less than 75% of cells staining blue, and 4 = greater than 75% of the cells staining blue. 

Inflammation and Ocular Pathology

Comparable sections were stained with hematoxylin–eosin and then analyzed histopathologically. Briefly, inflammatory infiltrate was scored as follows: 0, none; 1, inflammatory cells present on low-power (40×) but less than 10 per field on highest power (400×); 2, 10 to 100 inflammatory cells per high-power field in the area of densest infiltration; and 3, more than 100 inflammatory cells per high-power field or epithelioid cell predominance. Tissue inflammation was scored as follows: 0, none; 1, mild infiltrate; 2, marked infiltrate with preservation of normal tissue structure; 3, marked infiltration with some loss of tissue structure; and 4, obliteration of normal tissue structure. ⁴³ Scoring of the slides was done in a masked manner with at least two scorers rating each infection route for each animal. 

Intracranial Injection

For delivery to the visual cortex, rats were anesthetized with ketamine/xylazine as described above. Using aseptic precautions, a bone flap was removed from the skull, and a unilateral injection was made in the visual cortex (area 17 ⁵² ). A 1-μl injection, 1 mm deep into the cortex was made using a Hamilton syringe mounted on a micromanipulator on a stereotax. A total of 2 × 10⁶ pfu of hrR3 was delivered over a 2-minute period. Six rats were used; one received PBS, and five were given hrR3. Three days later, the rats were deeply anesthetized with 60 mg/kg pentobarbitol sodium and perfused transcardially with 4% paraformaldehyde and 0.25% gluteraldehyde. Brains were postfixed overnight, and transverse sections (50-μm-thick) were cut on a cryostat for X-gal staining and analysis. 

Results

Gene Delivery in the Rat

Corneal Delivery.

After corneal scarification and topical delivery of the hrR3 virus, we observed transgene expression in the corneal epithelium concentrated at the sites of corneal scarification (Table 1) . Expression of lacZ was also seen in cells resembling keratocytes in the corneal stroma directly beneath the site of scarification (Fig. 1A ). Occasional cells of the iris pigment epithelium, trabecular meshwork, and the ciliary body also expressed the lacZ transgene after corneal application. Surprisingly, the most efficient gene delivery after corneal scarification was to the RPE with greater than 25% of these cells displaying transgene expression. The expression pattern was not uniform with greater numbers of lacZ expressing cells near the ora serata and fewer near the optic nerve. We also observed rare RGCs expressing the lacZ transgene; however, these were few and randomly distributed. 

Delivery of the hrR3 virus to the eye without prior scarification of the cornea resulted in a similar pattern of lacZ transgene delivery (Table 1) . We observed lacZ gene expression in the iris epithelium, ciliary body, and trabecular meshwork (Fig. 1F) . Transgene expression was observed in approximately 10% of the RPE cells (Fig. 1E) and in an occasional RGC, photoreceptor cell inner segment, and cells of the optic nerve. Although, lacZ transgene expression was distributed across the retina, greater transgene expression was observed near the ora serata. 

Intracameral Delivery.

After intracameral delivery with either 2 × 10⁶ pfu (1 μl) or 1 × 10⁷ (5 μl) pfu of the hrR3 virus, we observed lacZ expression in 75% of the RPE cell layer (Table 1) . A few RGCs also expressed lacZ, and these cells were randomly distributed across the whole retina. In addition, intracameral delivery resulted in transgene expression in 10% to 20% of stromal and pigmented epithelial cells of the ciliary body and cells of the trabecular meshwork as well as the corneal endothelium (Fig. 1B) . 

Intravitreal Delivery.

We delivered either 2 × 10⁶ pfu (1 μl) or 1 × 10⁷ (5 μl) pfu of the hrR3 virus to the vitreal chamber. Regardless of the amount or volume of virus delivered, we observed efficient labeling of the RPE, with nearly every cell expressing the lacZ transgene (Table 1) . We also observed lacZ expression in approximately 25% of the RGCs, which were spread throughout the retina (Fig. 1C) . Rarely, we observed lacZ expression in the inner nuclear layer, indicating possible transgene delivery to horizontal, bipolar, or amacrine cells (Fig. 1D) . In one eye, we observed a column of RPE cells, photoreceptor cells, bipolar cells, and RGCs expressing the lacZ transgene. Müller cells did not appear to express the transgene. We also observed a few photoreceptor cells that expressed lacZ in the cell body, but staining was not seen in the outer segments. The lacZ reaction product was also seen at the tips of the photoreceptor outer segments where there was retinal detachment during tissue processing, but this appeared to be due to RPE material that had remained attached to the outer segments. Transgene expression was also observed in cells of the optic nerve, which are presumably glial cells. We also noted transgene expression in 10% to 20% of the iris pigment epithelium, trabecular meshwork, ciliary body stroma, and pigmented epithelium as well as the corneal endothelium after intravitreal virus delivery. 

Pathology and Length of Transgene Expression

Parallel hematoxylin and eosin–stained sections from the rat were examined for signs of inflammation or virus-induced pathology at 3, 7, or 14 days postdelivery. We did not observe cellular infiltration in the vitreal or anterior chambers, nor did we observe destruction of tissue structure in the eye after any of the delivery routes even at 14 days postdelivery. Inflammatory responses were only seen when there was obvious needle damage to the lens capsule during the injection procedure. We did observe waning of the lacZ gene expression beginning at 7 days, with little gene expression evident at 14 days postdelivery. 

Gene Delivery in the Mouse

Corneal Delivery.

Topical application of the hrR3 HSV-1 virus to the cornea after corneal scarification of the mouse resulted in transgene expression similar to that observed in the rat (Table 2) . Expression of lacZ was observed in the corneal epithelium and underlying stroma as well as the RPE cells; however, in contrast to corneal scarification delivery of hrR3 in the rat, we did not observe lacZ expression in the iris pigmented epithelium, trabecular meshwork, or ciliary body epithelium of the anterior chamber. Topical delivery of the hrR3 virus to the mouse cornea without scarification did not result in transgene expression in any cell type except for the rare RPE cell (Table 2) . 

Intracameral Delivery.

After intracameral injection with 1 × 10⁷ pfu of the hrR3 virus, the mouse eye displayed lacZ transgene expression primarily in the ciliary body nonpigmented epithelium (Fig. 2C ) and the RPE cell layer; however, we did not observe transgene delivery to other cells of the anterior chamber, nor did we observe any RGC delivery. 

Intravitreal Delivery.

Intravitreal injection of 1 × 10⁷ pfu hrR3 virus in the mouse resulted in lacZ expression in 100% of the RPE cells (Figs. 2A 2B) and occasional cells in the inner and outer nuclear layers (Fig. 2A , arrow). Figure 2B shows a partially tangential section of the retina, with extensive lacZ expression in RPE cells showing the X-gal product distributed throughout the cytoplasm of the RPE cells. In addition, intravitreal delivery resulted in transgene delivery to the ciliary body nonpigmented epithelium (Fig. 2C) ; however, no other cells of the anterior chamber expressed lacZ after this route of delivery in the mouse. 

Pathology and Length of Transgene Expression.

Parallel sections of the eyes were also stained with hematoxylin and eosin at 3, 7, and 14 days postdelivery so that we could observe potential inflammation or pathology induced by the hrR3 HSV-1 virus. Similar to our results in the rat, we did not observe cellular infiltrate into the vitreal or anterior chambers, nor did we observe destruction of tissue architecture or cytopathology in the eye after any route of delivery even at 14 days postdelivery. As was observed with the rat, lacZ gene expression began to decline at 7 days postdelivery, with little gene expression evident at 14 days postdelivery. 

Delivery to Visual Cortex

To assess transgene delivery to the visual cortex, the hrR3 virus was stereotactically delivered to area 17 of the rat brain. Representative results are shown in Figure 3 . Expression of lacZ occurred in numerous cells at the site of delivery (Figs. 3A and 3C) and extended approximately 1 mm on either side of the injection track and 1 mm deeper in the visual cortex than the tip of the needle used for delivery. Thus, the cells in a total volume of 6.3 mm³ were able to take up the vector and express the transgene. The cells expressing lacZ were large and had characteristic neuronal morphology (Fig. 3C) , including some dendritic extensions that were positive for lacZ (Fig. 3C , arrow). Analysis of a control PBS-injected brain revealed lacZ staining only in endothelial cells of blood vessels and in the red nucleus as has been reported previously. ⁵³ Thus, lacZ expression in neurons in area 17 of the visual cortex was due to virus-mediated transgene delivery. 

To determine whether the virus could be transported to other regions of the CNS, we also analyzed brain sections containing the LGN. As shown in Figure 3B , numerous LGN cells with the size and morphology of neurons were positive for lacZ expression in brains injected with hrR3. Based on the volume of the LGN and thickness of the sections, ⁵⁴ we estimate that, on average, approximately 300 neurons in each LGN were able to express the lacZ transgene. In contrast, no LGN lacZ expression was seen in PBS-injected brains, indicating that lacZ expression was due to the virally delivered gene. There was little, if any, evidence of an inflammatory or glial cell reaction in areas where lacZ was expressed, and we saw no evidence for cytotoxic effects, suggesting that the hrR3 virus is safe for use in CNS delivery. However, we should note that only day 3 animals were evaluated, and this may be too early to observe some pathologic responses. ³⁶  

Discussion

In this study, we examined transgene delivery and expression using a replication competent HSV vector in the mouse and rat after corneal scarification, intracameral injection, and intravitreal injection as well as intracranial injection to the visual cortex in the rat. Overall, the most prominent cell type expressing the transgene in the eye was the RPE cell layer regardless of the route of delivery. Other cells of the eye that expressed the lacZ transgene, although to a much lesser extent, included the ganglion cells of the neural retina, ciliary body, trabecular meshwork, and iris pigment epithelial cells. Corneal delivery after scarification, although highly efficient, was primarily restricted to the area damaged by scarification and adjacent underlying keratocytes in the corneal stroma. Surprisingly, corneal delivery without concurrent scarification also resulted in transgene expression in several cell types of the anterior and vitreal chambers (RPE in particular) in the rat but not in the mouse, suggesting that specific species differences exist between ocular HSV infections in the rat and mouse. In addition, in a few eyes of the rat, intravitreal injection resulted in gene delivery and expression in an occasional column of ganglion, bipolar, photoreceptor, and RPE cells. As has been observed for other viruses inoculated into the eye, the lens was refractory to infection by hrR3 by all infection routes. Even when the lens capsule was accidentally damaged by the needle, the exposed cells of the lens did not express lacZ, suggesting that the lens cannot be efficiently infected by the hrR3 HSV-1 virus. 

Intravitreal injections of AV and AAV appear to be much less efficient at delivery to the neural retina than direct subretinal injections. ^{3 4 6 14 21} With intravitreal delivery, AAV can deliver genes to RPE cells and ganglion cells, whereas AV vectors can efficiently deliver a gene to the iris pigment epithelium, corneal epithelium, and RGCs with intravitreal delivery, but the RPE cells are not efficiently transduced and those RPE cells expressing the transgene are limited to the area surrounding the injection site. In our studies, the attenuated hrR3 virus was capable of efficiently delivering the lacZ gene to the RPE cells across the entire retina and was able to deliver the transgene to up to 25% of the RGCs, but only occasional cells of the inner nuclear layer and photoreceptor cells after intravitreal injection. Because of the intense X-gal staining in these sections, we were unable to determine the specific cells that express the lacZ transgene in the inner nuclear layer. In a companion study examining hrR3 gene delivery in the monkey eye, ⁵¹ we also found that intravitreal injection resulted in efficient delivery of the lacZ transgene to cells of the RPE, suggesting that the efficient HSV-mediated RPE delivery is not specific to just rodent species. Because the RPE cell layer provides trophic support for the neural retina including the photoreceptor cells, the RR mutant virus may be ideal for delivering survival-promoting genes to the retina without the damage caused by subretinal injections. 

Gene delivery to trabecular meshwork or ciliary body cells may prove useful in gene therapy strategies for glaucoma. Injection of recombinant AV to the anterior chamber of mice results in gene delivery to the corneal endothelium, iris pigmented epithelium, and trabecular meshwork, with approximately 20% to 30% of the cells expressing the delivered transgene. ⁶ Anterior chamber inoculation of AV in rabbits also results in efficient delivery to the ciliary body and other cell types, but this was accompanied by severe inflammation. ^{8 10} We observed only limited transgene expression in the anterior chamber in the mouse after intracameral injection possibly due to technical difficulties related to injecting into the restricted space; however, we did observe lacZ gene expression in the RPE cell layer. Nearly identical results were observed in the rat, where intracameral delivery of the hrR3 virus resulted in lacZ expression in tissues of the anterior chamber as well as the RPE cell layer. Recently, we showed that intracameral injection of the hrR3 virus in the monkey eye resulted in a similar pattern of lacZ expression, with efficient delivery of the lacZ transgene to the ciliary body epithelium and trabecular meshwork of the anterior chamber and also the RPE cells, although delivery was accompanied by a transient inflammatory response. ⁵¹ Thus, our results extend trabecular meshwork and ciliary body gene delivery to the rat and suggest that HSV vector systems may be useful for glaucoma-related gene delivery. 

We found that delivery of the hrR3 HSV-1 virus by corneal scarification resulted in efficient gene delivery to the corneal epithelium only in areas that had been damaged by the scarification. Topical application of the hrR3 virus to the cornea without prior scarification did not result in lacZ transgene expression in the corneal epithelium or underlying keratocytes, indicating that damage to the surface must be required for access of the virus to the corneal cells. Corneal delivery of the AV and AAV do not appear to efficiently deliver genes to the corneal cells in the absence of corneal injury. These results suggest that restricted delivery to sites of corneal damage can be achieved with the HSV-1 virus and may, thus, be useful for modulating corneal wound healing. 

There are a number of reports of intraocular gene delivery using AAV-, AV-, and HIV-based vectors by subretinal injection. ^{2 3 4 20 21 22 23 55} Injections by this route typically deliver genes to the RPE cells and the photoreceptors, but delivery is usually restricted to areas where retinal detachment occurs. The most efficient gene delivery by this route appears to occur either in young pup mice or adult mice undergoing photoreceptor degeneration. These results suggest that the photoreceptor cells need to be undergoing a growth or repair in order for efficient delivery to occur. In addition, this delivery route resulted in localized retinal detachment that, although resolving within a few days, did result in damage to photoreceptor cells in the immediate area. For this reason, we chose not to attempt subretinal injections in our studies. 

The hrR3 virus was capable of delivering the lacZ transgene to cells of the retina regardless of the route of administration to the eye. The route by which the vector virus reaches these cells is not clear because the vitreal chamber was presumably not disturbed with either the corneal delivery or the intracameral injection. Because intracameral or intravitreal delivery of the hrR3 virus in a volume of only 1 μl also resulted in efficient gene delivery to the retina, it is unlikely that an increase in pressure resulting from the injection forced the virus into the vitreous or the anterior chamber. We also noted lacZ expression in both chambers of the eye after delivery of the hrR3 virus by corneal scarification and corneal drops, neither of which should increase intraocular pressure. It is of particular interest to note that topical delivery of the hrR3 virus to the cornea of rats and mice could deliver the lacZ transgene to the RPE of both animals and the RGCs, ciliary body epithelium, and trabecular meshwork of the rat without any disruption of the eye. Scarification of the cornea before application of the virus enhanced this transfer, but simply dropping the virus suspension onto the cornea without prior scarification was sufficient to result in transgene expression in these intraocular tissues. Liposome gene delivery has been shown to deliver a transgene to the retina after topical application to the cornea ⁵⁶ ; however, no other report has demonstrated vector-mediated gene delivery to the retina after corneal application. Because the hrR3 virus is capable of delivering a transgene to cells of the retina and angle without the need for injection into the vitreous or anterior chamber of the eye, corneal delivery allows for a noninvasive and presumably less traumatic procedure for delivering transgenes to these cells. This may also provide for gene delivery by self-administration by the patient, which would be a considerable advantage. 

Evaluation of gene delivery using reporter vectors depends on two factors: the ability of the vector to enter the specific cell type and the ability of the promoter to be expressed in a given cell. Thus, one potential explanation for the lack of delivery to photoreceptors would be a lack of expression of the marker gene. Therefore, in most gene delivery studies, delivery may be underestimated. In our studies, the lack of promoter expression as an explanation for poor delivery to photoreceptors is unlikely for two reasons. First, we did observe rare expression in cells in both the inner and outer nuclear layers. Second, and more importantly, the lacZ gene was expressed from the ICP6 viral promoter whose expression is enhanced by the virally encoded ICP4, ICP0, and ICP27 genes, which are present in the hrR3 virus. Thus, expression of the ICP6 lacZ marker gene is not entirely dependant on cell-specific transcription factors in this vector system. We, therefore, believe that the “poor” lacZ gene expression is due to poor vector delivery and is probably related to a lack of accessibility of the virus to the photoreceptors. 

The eye is an immune-privileged site (reviewed in ^{Ref. 18)} due to its anatomic isolation as well as its unique immunomodulatory mechanisms, and this may explain the lack of inflammation observed with the hrR3 virus as well as most other intraocular recombinant viral vector infections. In fact, we did not observe immunologic cellular infiltrate even 14 days postinfection, indicating that there is no immediate or late response to the hrR3 virus in the eye. This is consistent with our previous results showing that multiple intravitreal injections of an RR null virus into the mouse eye does not cause immunologic pathology even 2 months postinfection. ⁴³ The lack of an ocular immune response to injection of other recombinant viruses is not uniform because one study of AV showed that injection into the anterior chamber of rabbits resulted in a severe inflammatory response in 3 of 8 animals such that ciliary and iris epithelium and most corneal epithelium cells were swollen and detached ⁸ ; however, this may be due to species differences because most studies of virus-mediated gene delivery to date have used rodent models. 

Contrary to results in the rat and mouse, we observed inflammation and cellular infiltrate after both intracameral and intravitreal injections of the hrR3 HSV-1 virus in the monkey eye. ⁵¹ The inflammatory response was characterized by an infiltration of lymphocytes, polymorphonuclear cells, and macrophages into the anterior and vitreal chambers; however, this inflammation appeared to be transient and had waned by day 10 of the study. The immune response in the monkeys was most severe in the virus-injected eyes but was also seen in the control PBS-injected eye. One potential caveat to our primate studies is that these monkeys had previously been used in other ocular experiments, and it may be that the immune privilege of the eye had already been compromised, contributing to a general inflammatory response as a result of the needle injection alone. The inflammatory response observed in the monkey may also reflect species differences in response to the HSV virus and suggests that testing of recombinant virus vectors in primates at least for pathology and toxicity is essential because this model will presumably more closely mimic human responses. 

Injection of the hrR3 virus into area 17 of the rat visual cortex resulted in efficient lacZ expression at the site of injection as well as at the LGN, which sends afferent projections to the visual cortex. Although previous reports have described endogenous lacZ expression and staining in the rat brain, ⁵³ the pH of our fixing and staining solution (pH 7.6) should have eliminated most of the background staining. In addition, endogenous lacZ expression was seen only in endothelial cells and neurons in the red nucleus of PBS-injected rat brain as observed previously ⁵³ ; thus, the visual cortex and LGN staining was due to the delivered transgene. Transgene expression in the visual cortex and LGN appeared to occur in both neuronal and nonneuronal cells, indicating that the hrR3 virus is capable of infecting many different cells of the CNS as has been described previously for other HSV vectors. ^{32 35 37 38} Delivery to the LGN likely resulted from anterograde transport of the virus from LGN axonal termini located in the visual cortex. Transport of virus with associated gene delivery has also been observed for CNS injection of TK and ICP0 HSV null viruses. ^{36 57} Transport of the virus from the initial injection site may provide a means for gene delivery not only to injected tissues but to innervating tissues of the CNS as well. This phenomenon may be site-specific in the brain; however, because delivery with HSV amplicon vectors can remain localized in certain areas (see ^{Refs. 38} and ⁵⁸ and Howard Federoff, personal communication, May 1999). Previous investigators have observed extensive tissue destruction and inflammation in the CNS after injection of wild-type HSV ^{37 59 60} and even modest tissue destruction at the site of injection with TK and ICP0 mutant HSV vectors. ^{36 57} We did not observe virus-induced pathology or immune cell infiltrate in any of the rats in our study 3 days after injection of the RR mutant virus, suggesting that vectors attenuated by mutations of RR may be more suitable for CNS delivery; however, analysis of sections at 3 days after delivery may not have allowed sufficient time for pathologic responses to become manifest. ³⁶  

We also examined expression of the vector delivered lacZ gene 3, 7, and 14 days after vector delivery. Expression from the ICP6 promoter was strong at 3 days postinfection; however, 7 and 14 days after vector delivery, expression began to decline in all transduced tissues of the eye. Because the ICP6 promoter is an immediate early/early promoter, we had not anticipated long-term expression from this construct. The goal of this project was to identify the tissues of the eye that were receptive to transgene delivery. In some clinical situations, short-term delivery may be more desirable, such as treatment of transient retinal ischemia or traumatic optic nerve injury. However, promoters that express long term will need to be identified before treatment of many eye diseases can be realized. 

Previous investigations have examined the use of defective viral vectors in the eye; however, these viruses have been limited in their ability to efficiently deliver transgenes to cells across the whole eye and have also been ineffective at delivering genes to the photoreceptors except by the use of subretinal injection. One goal of our study was to test the hypothesis that attenuated viruses may allow for more efficient gene delivery to photoreceptor cells and neuronal cells distant to the site of injection because of the possibility that limited replication may allow the virus to cross synapses. We found that the hrR3 HSV-1 virus can deliver a transgene to many cells of the eye after corneal delivery, intracameral injection, and intravitreal injection; however, we were not able to deliver genes efficiently to the photoreceptor cells despite efficient delivery to the RPE cells, which are in close contact to the photoreceptors. It is not clear why photoreceptors are refractory to delivery, but possible explanations include a lack of viral receptors on the outer segment membranes, a lack of viral transport from the outer segments to the cell body, or the presence of an extensive extracellular matrix surrounding the photoreceptors. The success of gene delivery by subretinal injection with concomitant retinal detachment suggests that access to the photoreceptor cell body may be a critical factor. 

 Supported by a grant from the Retina Research Foundation (Houston, Texas; CRB). CRB is the Retina Research Foundation Alice McPherson Professor.

 Submitted for publication June 30, 1999; revised November 30, 1999; accepted December 30, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Summary of Ocular Gene Delivery in the Rat

Cell Type or Tissue	Route of Administration
		CD	CS	AC	VC
Corneal epithelium	0	1	0	0
Corneal endothelium	0	0	1	1
Iris pigment epithelium	+	1	0	1
Trabecular meshwork	+	1	1	1
Ciliary body	1	1	1	1
Lens	0	0	0	0
Ganglion cells	+	1	1	2
Photoreceptors	+	0	+	+
RPE	1	2	3	4
Optic nerve	+	0	0	1
Sclera	1	1	1	1

 CD, corneal drops; CS, corneal scarification; AC, anterior chamber; VC, vitreal chamber, +, occasional (less than 1%) of the cells labeled; 0, 0% of cells labeled; 1, less than 25% of cells labeled; 2, 25%–50% of cells labeled; 3, 50%–75% of cells labeled; 4, greater than 75% of cells labeled.

Figure 1.

 

Gene delivery in the rat eye with hrR3 by cornea scarification (A), intracameral injection (B), intravitreal injection (C, D), and corneal drops (E, F). Lac Z expression can be observed in the corneal epithelium and underlying stroma (A). Transgene expression in the anterior chamber was observed at the angle (B, F) in the iris pigment epithelium (B, black arrows), ciliary body stromal cells (B and F, red arrows), and pigmented epithelial cells (black arrowheads). Transgene delivery was also observed in the retina (C, D, E) in the RPE (C and E, arrows) and RGCs (C, arrowheads) and in occasional cells of the inner nuclear layer (D, arrows). Scale bars, (A, D) 20 μm; (B, C, E, F) 70 μm.

Table 2.

 

Summary of Ocular Gene Delivery in the Mouse

Cell Type or Tissue	Route of Administration
		CD	CS	AC	VC
Corneal epithelium	0	2	0	0
Corneal endothelium	0	0	0	0
Iris pigment epithelium	0	0	0	0
Trabecular meshwork	0	0	0	0
Ciliary body	0	0	3	2
Lens	0	0	0	0
Ganglion cells	0	0	0	1
Photoreceptors	0	0	0	+
RPE	+	1	3	4
Optic nerve	0	0	0	1
Sclera	0	1	1	1

 CD, corneal drops; CS, corneal scarification; AC, anterior chamber; VC, vitreal chamber, +, occasional (less than 1%) of the cells labeled; 0, 0% of cells labeled; 1, less than 25% of cells labeled; 2, 25%–50% of cells labeled; 3, 50%–75% of cells labeled; 4, greater than 75% of cells labeled.

Figure 2.

 

Gene delivery in the mouse after intravitreal and intracameral delivery of the hrR3 virus as described in the Methods section. After intravitreal delivery, a section through the pupil–optic nerve axis (A) shows lacZ expression localized to the RPE cell layer (arrowhead) and occasional lacZ expression in the inner nuclear layer (arrow), and a partially tangential section (B) through the posterior of the globe shows transgene delivery across the RPE cell layer, with lacZ expression evident throughout the cytoplasm of the RPE cells. Intracameral delivery of hrR3 (C) shows some lacZ expression in the nonpigmented epithelial cells of the ciliary body (arrow). Scale bars, (A) 25 μm; (B, C) 50 μm.

Figure 3.

 

Gene delivery in the rat brain after injection of the hrR3 virus into area 17 of the visual cortex. lacZ expression was observed in the visual cortex at the site of injection (A, 4×) and in the LGN (B, arrow; 10×). Higher magnification of the visual cortex (C, 20×) shows distinctive neuronal morphology in lacZ-expressing cells (arrows).

The authors thank Janice Lokken for preparing the tissue sections and Inna Larsen for preparation of the manuscript.