Rapid Ocular Angiogenic Control via Naked DNA Delivery to Cornea

Abstract

purpose. To determine the efficacy and safety of naked plasmid gene therapy to the corneal stroma and epithelium.

methods. Naked plasmid DNA was injected under pressure into the cornea of mice. The expression of genes coding for beta galactosidase (β-gal), enhanced green fluorescent protein (EGFP), vascular endothelial growth factor (VEGF), and soluble Flt-1 (s-Flt) was recorded and measured with regard to dose, time course, and bioactivity.

results. LacZ gene expression of the protein β-gal was demonstrated as early as 1 hour, with expression persisting for 10 days. Plasmid-injected corneas remained clear and free of inflammation. EGFP was bicistronically expressed with VEGF to demonstrate the practicality of simultaneous in vivo analysis of gene expression and growth factor bioactivity. Corneal injection of a plasmid containing VEGF cDNA induced corneal and anterior chamber neovascularization. Moreover, corneal injection of plasmid containing the cDNA for the soluble form of the VEGF receptor Flt-1 effectively prevented corneal neovascularization.

conclusions. The cornea is readily accessible for gene therapy in the laboratory and in the clinic. The method described is safe, effective, titratable, and easily monitored. Naked DNA delivery to the cornea has the potential to alter the treatment of a wide variety of corneal and anterior segment diseases.

Beginning in 1990, the feasibility of direct gene transfer into musculature was demonstrated in the laboratory and the clinic. Wolff et al. ¹ showed protein expression in mouse skeletal muscle after intramuscular injections and Isner ² and Losordo et al. ³ demonstrated the biological effect and benefit of gene transfer to human skeletal and cardiac tissues. 

An attractive feature of focal naked plasmid gene therapy is the ability to express locally a protein with reduced chance of systemic exposure and inflammation. However, the limitations of naked plasmid gene therapy as described include limited accessibility to tissue, poor transfection efficiency, invasive monitoring techniques, and the inability to control levels of gene expression. Ocular tissues, by contrast, are easily anesthetized, readily accessible, and quickly and noninvasively monitored by way of slit-lamp biomicroscopy. Prior attempts to deliver DNA to ocular tissues have met with limited success. Gold microbeads coated with naked plasmid and delivered via a jet of air to the cornea introduce foreign material into the cornea and do not have high transfection efficiency. ⁴ The repeatability of such a method in clear cornea is also a concern. Lipid transfection reagents, electoporation, and viral techniques can induce inflammation or cell death, effects that may not be tolerable in the delicate transparent tissues of the eye. ^{5 6 7}  

The technique of stromal hydration, that is, forcing saline solution into the corneal stroma, is used routinely at the end of cataract surgery to temporarily swell the cornea and ensure a watertight wound. ⁸ We used stromal hydration to deliver naked plasmid to the cornea and discovered a straightforward and efficient method of safely transfecting cornea in vivo. The level of gene expression is remarkably titratable and is sufficient to induce biological effects not only in the cornea, but also the anterior and posterior segments of the eye. 

Materials and Methods

Plasmid Construction

pLacZ and pIRES2-EGFP were purchased from Clontech (Santa Clara, CA). Plasmids were purified using a Qiagen kit (Santa Clarita, CA) according to the manufacturer’s instructions. Mouse VEGF₁₆₄ was subcloned into the bi-cistronic expression vector pIRES2-EGFP upstream of the internal ribosome entry site and EGFP to make pIRES2-EGFPmusVEGF. The plasmid was sequenced to ensure fidelity and the correct orientation of mouse VEGF₁₆₄. A CMV promoter was used to drive the expression of the LacZ and VEGF genes. 

Murine VEGF₁₆₄ cDNA was cloned into the vector pCMMP to make pCMMPmuVEGF. pCMMP contains a multiple cloning site flanked by a CMV promoter and the bovine growth hormone polyadenylation signal. The plasmid was sequenced to ensure fidelity and the correct orientation of mouse VEGF₁₆₄. 

Murine Flt-1 (1–3) was PCR amplified from Flt-1 cDNA (gift of Shay Soker, Children’s Hospital, Harvard Medical School, Boston, MA) with appropriate primers and cloned into pHIHGAdd2 to make pHIHGAdd2Flt. An EcoRI-SalI fragment containing murine Flt-1 (1–3) with a C-terminal His tag was cloned into the vector pHIHGAdd2 containing a multiple cloning site flanked by a CMV promoter and SV40 polyadenylation signal. 

Animals and Anesthesia

All animal experiments were approved by the Children’s Hospital Animal Care and Use Committee and conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Before all experimental manipulations, CD-1 mice were anesthetized with 70–80 mg/kg intraperitoneal Nembutal sodium solution (Abbott Laboratories, North Chicago, IL). 

Intrastromal Corneal Injections

Under direct microscopic observation, a nick in the epithelium and anterior stroma of a CD-1 mouse cornea was made in the midperiphery with a 1/2-inch 30-gauge needle (Becton Dickinson, Franklin Lakes, NJ). A 1/2-inch 33-gauge needle with a 30° bevel on a 10-μl gas tight syringe (Hamilton, Reno, NV) was introduced into the corneal stroma and advanced 1.5 mm to the corneal center. Two microliters of plasmid solution was forcibly injected into the stroma to separate corneal lamellae and disperse the plasmid. 

Analysis of β-Galactosidase Production

Eyes injected with pLacZ were enucleated at various time points, fixed, and stained with X-gal (Gibco BRL, Rockville, MD). Corneas were removed after staining, and flat mounts were photographed with a MZ FLII fluorescence stereomicroscope (Leica, Heerbrugg, Switzerland) using a CCD digital camera (Dage, Michigan City, IN). The corneal area stained with X-gal was quantified with Improvision Openlab v.2 software (Coventry, England). A masked observer established threshold levels of hue and saturation, and all areas of the cornea above the threshold levels were marked as stained and quantified. The area of staining was divided by the total area of the cornea, as determined by the limbal vascular arcade. Selected corneas were paraffin embedded, sectioned, and stained with hematoxylin and eosin. 

Enhanced Green Fluorescent Protein Expression

Mice injected with pIRES2-enhanced green fluorescent protein (EGFP) were examined and photographed under general anesthesia using the Leica microscope equipped with the Dage CCD digital camera and Improvision Openlab software. 

Analysis of Angiogenesis Induction and Inhibition

Mice were perfused with fluorescein isothiocyanate (FITC)-coupled lectin to label the vasculature. Under deep anesthesia, the chest cavity was carefully opened, and a 16-gauge perfusion cannula was introduced into the left ventricle. Drainage was achieved using a 16-gauge needle placed in the right atrium. After PBS perfusion, fixation with 1% paraformaldehyde and 0.5% glutaraldehyde was achieved at physiologic pressure followed by perfusion with FITC-coupled Con A lectin (20 μg/ml in PBS, pH 7.4, 5 mg/kg BW; Vector Laboratories, Burlingame, CA). The eyes were then enucleated, and the corneas were flat mounted. Using a fluorescent microscope, threshold levels of green saturation and intensity were established by a masked observer and used to quantify all areas of neovascularization within the limbal vascular arcade. The masked observer, using the Leica microscope and Improvision Openlab software image analysis, measured the area of neovascularization. The area of neovascularization was divided by the total corneal area and expressed as a percentage. 

VEGF ELISA

Each cornea was placed into 200 μl of lysis buffer (20 mM imidazole HCl, 10 mM KCl, 1 mM MgCl₂, 10 mM EGTA, 1% Triton, 10 mM NaF, 1 mM Na molybdate, 1 mM EDTA, pH 6.8) supplemented with a protease inhibitor cocktail (Boehringer Mannheim, Indianapolis, IN) followed by sonication. The lysate was cleared of debris by centrifugation at 14,000 rpm for 15 minutes (4°C), and the supernatant was assayed. Total protein was determined using a commercial assay (BCA kit; Bio-Rad, Hercules, CA). Supernatant VEGF levels were determined using a sandwich enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN) and normalized to total protein. 

VEGF Pellet Construction

The technique for mouse corneal pocket pellet construction and placement has been well described. ⁹ Briefly, mouse VEGF₁₆₄ (R&D Systems) was mixed with Hydron (Interferon Sciences, New Brunswick, NJ) and sucralfate (Sigma-Aldrich, St. Louis, MO) so that each pellet contained 50 μg of VEGF₁₆₄ protein. Pellets were implanted into the temporal quadrant of the cornea 24 hours after the area was injected with pIRES2-EGFP or pHIHGAdd2Flt. 

Statistics

Data are presented as mean ± SEM. Significance was tested using the paired two-sided t-test. P values < 0.05 were deemed significant. 

Results

Plasmid DNA Transfection: Dose Response and Time Course

Experiments were performed to determine the feasibility, efficiency, and safety of intrastromal naked plasmid DNA delivery to the cornea. Twelve hundred nanograms of pLacZ was suspended in normal saline and injected into the stroma of albino mice. At 24 hours, the corneas were optically clear and grossly free of inflammation (Fig. 1A) . Staining with X-gal showed widespread protein expression (Fig. 1B) that, according to microscopic observations, was localized to keratocytes and epithelial cells (Fig. 1C) . Further analysis showed normal corneas with little or no inflammation after pLacZ, saline, or sham injection, compared with uninjected controls (Figs. 2A 2B 2C 2D) . In contrast, injection of 1200 ng of the bicistronic VEGF- and EGFP-expressing plasmid pIRES2-EGFPmusVEGF induced inflammation and vascular engorgement at 24 hours, a result indicative of VEGF protein bioactivity (Fig. 2E) . 

Single doses as small as 0.05 ng of pLacZ in 2 μl of saline were injected into corneas. The expression of β-gal protein, measured as a percentage of corneal surface area stained with X-gal, was 0.6% ± 0.1% (n = 3), and doses ≥ 1800 ng produced a maximal effect at 24 hours, with 40.5% ± 7.5% (n = 3) of the corneal surface staining for β-gal (Figs. 3A 3B 3C 3D) . Twelve hundred nanograms of pLacZ in 2 μl of saline was injected into corneas that were stained at 1, 4, 8, 12, and 24 hours and 2, 5, 10, and 12 days. The expression of β-gal protein (Figs. 3E 3F 3G 3H) was observed as early as 1 hour (1.8% ± 1.4%, n = 3), peaked by 24 hours (50% ± 3.3%, n = 3), and was negligible by 12 days (data not shown). 

In Vivo Transfection and Monitoring

Mouse VEGF₁₆₄ was subcloned into the bicistronic ⁹ plasmid vector pIRES2-EGFP to create pIRES2-EGFPmusVEGF. The plasmid contained genes for both EGFP and VEGF. Effective transfection after a 1200-ng injection of the plasmid was assayed via light microscopy for vascular proliferation (Fig. 4A) and via fluorescent microscopy for EGFP expression (Fig. 4B) . In addition to triggering corneal neovascularization, pIRES2-EGFPmusVEGF also induced iris vascular engorgement and anterior hyphema (Fig. 4A) . 

Angiogenesis Induction and Inhibition

Corneas were injected with 1200 ng pCMMPmuVEGF (gift of Daniel Griese, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA), a plasmid vector that expresses murine VEGF₁₆₄ alone. The contralateral corneas received an equal amount of pIRES2-EGFP as a control. As with pIRES2-EGFPmusVEGF, VEGF produced via naked plasmid transfer of pCMMPmuVEGF-induced corneal neovascularization, iris vascular engorgement, and hyphema. After 7 days, 15.8% ± 4.8% of the cornea became vascularized (Fig. 4C) . In contrast, essentially no neovascularization was observed in the pIRES2-EGFP eyes (P < 0.05 vs. pIRES2-EGFP–treated corneas; n = 10 per condition). Surprisingly, both pIRES2-EGFPmusVEGF and pCMMPmuVEGF induced retinal hemorrhages (data not shown), a result suggesting a concentration gradient sufficiently steep to result in VEGF diffusion to the mouse retina. 

To confirm the production of VEGF protein after pCMMPmuVEGF injection, corneal VEGF levels were assayed via ELISA. Seven days after 1200 ng pCMMPmuVEGF injection, 7.07 ± 1.53 pg VEGF/μg total corneal protein was detected. In contrast, the pIRES2-EGFP–injected corneas contained only 0.25 ± 0.14 pg VEGF/μg corneal protein (n = 6, P < 0.001). 

To determine whether VEGF-induced corneal neovascularization could be inhibited, 1200 ng pHIHGAdd2Flt was injected into corneas. pHIHGAdd2Flt shows constitutive expression of a soluble fragment of the VEGF receptor Flt-1 (extracellular domains 1–3), an inhibitor of VEGF bioactivity. The contralateral control corneas received 1200 ng of pIRES2-EGFP. Twenty-four hours later, controlled-release pellets containing 50 μg of mouse VEGF₁₆₄ were implanted in the corneas as described previously. ^{9 10} Seven days later, the pHIHGAdd2Flt treated corneas showed greatly reduced levels of corneal neovascularization when compared with control corneas, 8.7% ± 3.0% (n = 10) vs. 32.3% ± 4.9% (n = 10), respectively (P < 0.04; Figs. 4D 4E 4F ). 

Discussion

The expression of the genes injected into cornea is extraordinarily rapid. We hypothesize that injected naked DNA enters keratocytes and epithelial cells under direct pressure in a manner similar to that shown in vascular endothelium and myocardium. ^{11 12} Additionally, endocytosis may also play a role because naked DNA has been shown to enter cells via this route. ¹³ We speculate that corneal endothelial cells are poorly transfected because of their amitotic state and because of the potential barrier of Descemet’s membrane. 

Plasmid saline solutions induce little to no inflammation, as in the case of pLacZ, or appropriately cause inflammation and neovascularization, as with pIRES2-EGFPmusVEGF and pCMMPmuVEGF. Ongoing work in the larger rabbit eye has validated the results in the mouse eye described here (data not shown). Given the extraordinarily rapid expression of protein with this approach, it should prove useful in the treatment of acute corneal diseases. The inhibition of pathologic corneal vessels, via local gene injection, could theoretically be achieved. Anti-inflammatory gene products could be used to control postoperative inflammation in cataract surgery, without the attendant glaucoma risk associated with topical steroids. Control of immune cell responses with cytokine inhibitors expressed locally and transiently could limit rejection episodes in corneal transplants. Visually handicapping persistent epithelial defects might be effectively treated with plasmids encoding growth factors. Because standard ophthalmologic equipment, that is, the slit-lamp biomicroscope, allows one to visualize the cornea and surrounding tissues under high magnification and determine transfection success and safety, these methods can be readily transferred from the laboratory to the clinic. Protocols for anterior chamber surgery or manipulation often require only one anesthetic drop instilled into the conjunctival cul-de-sac. Quick, safe, and readily available anesthetic technique makes laboratory use and eventually clinical application of direct stromal injection attractive. How broadly these techniques are applied will be determined by ongoing work. 

 Supported by the Roberta W. Siegel Fund, Boston, Massachusetts (APA); Deutsche Forschungsgemeinschaft Grant DFG Jo 324/2–1, Cologne, Germany (AMJ); the Juvenile Diabetes Foundation International, New York, New York (AMJ and APA); National Institutes of Health Grants EY11627 and EY12611 (APA); and the Massachusetts Lions (APA).

 Submitted for publication November 10, 2000; revised April 3, 2001; accepted April 19, 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.

 

Gross appearance of a mouse eye 24 hours after 1200 ng pLacZ injection (A), after staining with X-gal (B), and in cross section (counterstained with hematoxylin and eosin) showing expression in epithelium (asterisks) and keratocytes (arrowheads; C).

Figure 2.

 

Sections of hematoxylin and eosin–stained, normal uninjected cornea (A) and 24 hours after 1200 ng pLacZ plasmid (B), saline (C), sham (needle injury without injection; D), and 1200 ng pIRES2-EGFPmusVEGF (E) injections.

Figure 3.

 

Plasmid transfection dose response and time course. Representative corneas showing LacZ expression 24 hours after injection with 0.5 (A), 25 (B), and 500 ng (C). Collated data show that maximum expression was obtained with 1800 ng (n = 3 per condition; D). Representative corneas showing LacZ expression at 4 hours (E), 12 hours (F), and 10 days (G). Collated data (H) showing time course of expression (n = 3 per condition). The percentage of corneal surface area stained with X-gal is shown. Values represent mean ± SEM.

Figure 4.

 

Corneal neovascularization and inhibition via naked DNA gene therapy. Expression of VEGF leads to corneal neovascularization and hyphema (anterior segment hemorrhage) 7 days after 1200 ng pIRES2-EGFPmusVEGF injection (A). Fluorescent microscopy of the same cornea demonstrates areas of pIRES2-EGFPmuVEGF expression (B). pCMMPmuVEGF injection results in bioactive VEGF-inducing corneal neovascularization (P < 0.05 vs. pIRES2-EGFP–treated corneas, n = 10 per condition; C). Representative corneas expressing EGFP (D) and Flt-1 (E) 7 days after pIRES2-EGFP and pHIHGAdd2Flt injection, respec-tively. The collated data demonstrate the potent inhibition of VEGF pellet-induced corneal neovascularization after plasmid soluble Flt-1 gene therapy (P < 0.05, n = 10 per condition; F). Values represent mean ± SEM.