Delivery of Antisense Oligonucleotides to the Cornea

Abstract

Antisense oligonucleotides (ASOs) are synthetic nucleic acids that recognize complementary RNA sequences inside cells and modulate gene expression. In this study, we explore the feasibility of ASO delivery to the cornea. We used quantitative polymerase chain reaction to test the efficacy of a benchmark ASO targeting a noncoding nuclear RNA, Metastasis-Associated Lung Adenocarcinoma Transcript 1 (MALAT1), in a human corneal endothelial cell line, ex vivo human corneas, and in vivo in mice. In vivo delivery was via intravitreal or intracameral injections as well as topical administration. The anti-MALAT1 ASO significantly reduced expression of MALAT1 in a corneal endothelial cell line. We achieved a dose-dependent reduction of target gene expression in endothelial tissue from ex vivo human donor corneas. In vivo mouse experiments confirmed MALAT1 reduction in whole corneal tissue via intravitreal and intracameral routes, 82% and 71% knockdown, respectively (P < 0.001). Effects persisted up to at least 21 days, 32% (P < 0.05) and 43% (P < 0.05) knockdown, respectively. We developed protocols for the isolation and analysis of mouse corneal endothelium and observed reduction in MALAT1 expression upon both intravitreal and intracameral administrations, 64% (P < 0.05) and 63% (P < 0.05) knockdown, respectively. These data open the possibility of using ASOs to treat corneal disease.

Introduction

Antisense oligonucleotides (ASOs) are small synthetic nucleic acids that can alter gene expression by binding to complementary RNA through Watson-Crick base pairing [1]. Although most first generation ASOs failed to succeed in clinical trials due to poor pharmacokinetics and efficacies, newer generation ASOs have shown much greater success. For example, Spinraza is a recent U.S. Federal Drug and Administration (FDA) approved ASO that acts by increasing splicing of the SMN2 gene to produce an active variant that can partially correct the molecular defect that causes spinal muscular atrophy [2].

Spinraza is administered by intrathecal injection with saline and has shown potent effects in multiple clinical trials [3,4]. Side effects are mild and as few as four injections are required per year. Patient outcomes are dramatically improved, with enhanced survival and patients reaching motor milestones not observed in the natural course of the disease [4]. The success of Spinraza along with other recently approved drugs and ongoing clinical programs has suggested that ASOs may have substantial unexplored potential for treating eye disease.

Numerous corneal diseases have been linked to abnormal gene expression such as Fuchs' endothelial corneal dystrophy (FECD) [5], corneal fibrosis [6], corneal neovascularization [7], keratoconus [8], and several corneal dystrophies [9]. The advantages of ASOs for treating these diseases include (1) Generality, Watson-Crick base-pairing allows recognition of any gene to modulate its expression; (2) the potential for local delivery; and (3) the potential suggested by experience with Spinraza and other recent drugs, for potent and long-lasting effects accompanied by mild side effects.

Limited studies have pursued ASO feasibility in relieving pathologic phenotypes related to corneal disease. Clinically, only topical Aganirsen, an ASO designed to temper corneal neovascularization by providing therapeutic inhibition of insulin receptor substrate-1, has advanced to the clinical trial stage [10–13]. Published reports did not provide the detailed mechanistic insight or efficacy data necessary to make accurate predictions for the impact of ASOs or design optimal experiments. ASOs have shown striking success targeting mutant ataxin-7 expression in mouse retina [14].

FECD is a leading cause of vision loss and its treatment accounts for over 70,000 corneal transplantations per year world-wide [15]. While effective, corneal transplantation is not perfect and failure rates are variable depending on clinical center. Corneal transplants are limited by the supply of donor cornea, require a significant recovery period, and are usually performed on patients who have suffered from decline in vision for many years before vision loss becoming sufficiently disabling to justify surgery [16]. Effective nonsurgical approaches to treating FECD would have the potential to reach a large population and would be especially beneficial to those who do not have access to transplant surgery or poor candidates for surgery or wish to be treated before major vision loss has occurred.

The molecular cause for two-thirds of FECD cases is pathologic CTG trinucleotide repeat expansions in the TCF4 and DMPK genes [17–20]. The expanded repeat within intron 2 of the TCF4 gene is one of the most common disease-associated mutations, with a prevalence of 3% within the Caucasian population [18]. Repeat-associated FECD is caused by toxic mutant RNA and is not related to expression of TCF4 protein. The toxic CUG repeat RNA has been shown to accumulate and manifest as foci within the nuclei of corneal endothelial cells of patients. Several studies have shown that ASOs can effectively target these toxic foci, the root cause of FECD, to reduce detrimental downstream effects in both in vitro and ex vivo patient tissue [21–23].

These anti-CUG ASOs have the potential to be lead compounds for the development of drugs that might provide a nonsurgical intervention for FECD. However, before development can begin, it was necessary to determine whether ASOs could be used to control gene expression in the cornea and, specifically, to tissue most affected by FECD—the corneal endothelium. One previous study had reported that an ASO or siRNA targeting connexin43 was delivered by an intracameral injection to accelerate corneal endothelium healing in a rat corneal scrape injury model [24]. We note, however, that both the ASO and siRNA were not chemically modified, the ASO was 30 bases long, and no data were shown to confirm decreased expression of the target gene. In our investigation, we aimed to elucidate the efficacy of different ASO delivery routes to the cornea in vivo.

ASOs targeting the CUG repeat do not directly affect TCF4 protein or RNA levels, making the CUG repeat a poor target for initially testing the hypothesis that ASOs could be effective in the cornea. Therefore, as a proof of principle, we selected Metastasis-Associated Lung Adenocarcinoma Transcript 1 (MALAT1) as a target. Like the expanded CUG repeat that causes FECD, MALAT1 is a noncoding nuclear RNA target. MALAT1 is highly expressed in normal human corneal tissue and is known to be conserved globally in murine ocular tissue [25,26]. We used an anti-MALAT1 “gapmer” ASO with a central DNA “gap” that serves to recruit RNAse H to degrade the target RNA flanked by chemically modified bases [in this case 2′-methoxyethyl (MOE)] that increase binding affinity [27]. MALAT1 is widely used as a benchmark target for developing ASO designs and delivery because it has little or no function in most cells and there are potent and well characterized ASOs that reduce its expression.

In this study, we reduce MALAT1 expression using an anti-MALAT1 ASO in vitro in both human and mouse cell lines and in ex vivo whole human corneal tissue. We administered the anti-MALAT1 gapmer to mice eyes via various delivery routes, including intracameral and intravitreal injections and topical application. We discovered knockdown of MALAT1 in both whole cornea and isolated corneal endothelium, suggesting that use of ASOs can be a strategy for treating corneal diseases.

Materials and Methods

ASO treatment of human and mouse cell cultures

The F35T corneal endothelial cell line derived from FECD patient was a generous gift of Dr. Albert Jun (Johns Hopkins). Cells were grown in modified Eagle's minimal essential media (OptiMEM; ThermoFisher) supplemented with 8% fetal bovine serum, 5 ng/mL human epidermal growth factor (ThermoFisher), 20 ng/mL nerve growth factor (Fisher Scientific), 100 μg/mL bovine pituitary extract (ThermoFisher), 20 μg/mL ascorbic acid (Sigma-Aldrich), 200 mg/L calcium chloride (Sigma-Aldrich), 0.08% chondroitin sulfate (Sigma-Aldrich), 50 μg/mL gentamicin (ThermoFisher), and antibiotic/antimycotic solution (diluted 1/100; Sigma-Aldrich). Cultures were incubated at 37°C in 5% CO₂ and passed when confluent. The mouse NMuLi cells were obtained from ATCC. Cells were cultured in Dulbecco's modified Eagle's medium (Sigma-Aldrich) supplemented with 10% fetal bovine serum at 37°C. Single-strand ASOs were synthesized by Ionis Pharmaceutics (Carlsbad, CA). Anti-MALAT1 duplex RNA was purchased from IDT. ASOs were transfected into cells with lipid RNAiMAX (Life Technologies) as previously described [20].

Ex vivo ASO treatment of human donor cornea

Human donor corneas were obtained from the eye bank of UT Transplant Services. The donor corneas were incubated in 1 mL of Optisol-GS corneal storage media with ASOs at different concentrations at 37°C for 6 days. The Descemet's membrane-endothelium monolayers were dissected as previously reported [20].

Intraocular injections in mice

All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of UT Southwestern Medical Center. C57BL6J mice were purchased from the UT Southwestern Mouse Breeding Core and equal number of age-matched litter-mate male and female mice were used when possible. Mice were provided unlimited access to food and water and were on 12-h light/12-h dark cycle.

Mice were anesthetized with ketamine/xylazine (120, 16 mg/kg, respectively) followed by pupil dilation with one drop of phenylephrine (Alcon) and one drop of tropicamide (Alcon) 2 min after. Upon full dilation, lubricant ointment eye gel (Henry Schein) was applied to prevent corneal desiccation. Both intravitreal and intracameral injections were guided using Zeiss microscope. All intraocularly delivered anti-MALAT1 ASOs were preformulated with 1× phosphate-buffered saline (PBS) to reach a final concentration of 25 μg/μL. Intraocular procedures began with delivery of the ASOs to the right eye. A 30-gauge beveled needle was introduced at either 1 mm posterior to the supero-lateral limbus at a 45° angle toward the vitreous cavity or 1 mm anterior to the supero-lateral limbus at a 45° angle toward the aqueous cavity, depending on whether intravitreal or intracameral injection was performed, respectively. Needle was inserted 1.5 mm deep and then gently removed. Egressed chamber fluid was removed using a cotton tip. Then, 2 μL of anti-MALAT1 ASO solution preloaded in a Hamilton microsyringe fitted to a 33-gauge beveled needle was slowly injected in the fluid cavity over a course of 1 min. Needle was held in place for an additional minute after complete injection to allow cavity fluid equilibration with injected solution. Postinjection AK-poly-bac antibiotic ointment (Perrigo) was applied to the needle wound followed by a drop of lubricant gel. For the left eye, an identical injection procedure was performed but a sham injection of 2 μL of 1× PBS was delivered instead of ASO. Mice were kept on heating pad while recovering from anesthesia.

Ocular topical delivery in mice

Preformulated anti-MALAT1 ASO mixed with 1× PBS to reach a concentration of 800 μM (43 μg) was stored in −20°C. Aliquots of 7.5 μL of ASO solution were thawed daily and delivered to right mouse eyes twice a day at 9 AM and 5 PM for 7 days, followed by ocular tissue harvest on the eighth day as described below. 1× PBS was concurrently administered to the contralateral eye as a control. To administer, mice bodies were stabilized by grasping neck tissue, while delivering full aliquot to ocular surface via micropipette without contacting corneal surface with pipette tip. While grasping mice, eyes were held proptosed for 2 min to prevent immediate blinking and wiping of solution from eye, while enhancing and standardizing solution contact duration with the ocular surface.

Ocular tissue harvest

Following mouse euthanization, whole treated globes were enucleated at the optic nerve, rinsed with balanced salt solution (BSS) and then submerged in a small droplet of BSS in a 100 × 15 mm Petri dish. Globes were then punctured at the corneal edge with a 30-gauge needle and then circumferentially bisected along the limbus, separating the anterior and posterior chambers. Neuroretinal tissue from the posterior chamber was carefully separated from retinal pigmented epithelium and scleral tissue and was rinsed with BSS before storage. Whole corneal tissue from the anterior chamber was preserved after removal of lens and pigmented iris tissue.

Due to the delicate and microscopic nature of mouse corneal endothelium compared to human corneal endothelium, routine endothelial extraction protocols performed in human eye banks could not be easily applied to our animal model. Instead, we developed a new standardized protocol to handle mouse endothelial corneal tissue isolation and to consistently extract sufficient tissue for experimental analysis as followed.

Proceeding whole mouse corneal acquisition, tissue was incubated in trypan blue (Stephens Instruments) for 2 min before washed and resubmerged in BSS, allowing for enhanced visualization of the corneal endothelium, which was apparent by the appearance of a thin translucent blue layer within the concave surface of the cupped cornea. Next, a surgical scalpel number 10 curved blade was used to bisect the cornea into two even halves. Halves were each anchored to Petri dish by using forceps to grasp the epithelial and stromal tissue from the tissue's convex surface and avoiding contact with the stained endothelium. Second forceps were used to grasp the stained endothelium edge followed by gentle separation of the thin blue endothelial layer from the remainder of the cornea. Endothelial tissue was collected as detailed above, from four to five treated mouse globes, and was combined to ensure adequate RNA yield for analysis.

All collected tissues were inspected under high-power microscopy for any signs of congenital, pathological, or procedure-related defects such as cataracts, opacities, fibrosis or obvious signs of edema, which may indicate congenital malformations, procedure-related trauma or endophthalmitis, or abnormal wound healing. Any tissue with apparent defects were discarded and not used for quantitative analysis. Harvested corneal and neuroretinal tissues were used to assess efficacy of MALAT1 expression knockdown.

RNA extraction and quantitative polymerase chain reaction analysis of MALAT1 expression

Total RNA was extracted by TRIzol (ThermoScientific) for F35T or NMuLi cell lines. RNA from human endothelial tissue samples or mouse cornea and endothelial tissue samples were extracted using NucleoSpin RNA XS kit (Fisher Scientific) per the manufacturer's protocol.

After reverse transcription, MALAT1 expression was analyzed by quantitative polymerase chain reaction (qPCR) on a 7500 real-time PCR system (Applied Biosystems) using iTaq SYBR Green Supermix (Bio-Rad). Data were normalized relative to levels of RPL19 mRNA. Primers specific for human MALAT1 are as follows: F 5′-CGGGTGTTGTAGGTTTCTCTT-3′; R 5′-CCCACAAACTTGCCATCTACTA-3′. RPL19: F 5′-AGCCTGTGACGGTCCATTCC-3′; R 5′-CGGCGCAAAATCCTCATTCT-3′. For mouse mMalat1: F 5′-AGCTTTTGAGGGCTGACTGC-3′; R 5′-CCATTCATTCCCCTCTGAGC-3′; mRPL19: F 5′-GTATGCTCAGGCTACAGAAGAG-3′; R 5′-GAGTTGGCATTGGCGATTT-3′. The final data were analyzed using GraphPad Prism 8 software.

Statistical analysis

Two-tail Student's t-test was used for statistical analysis. For each treatment group, MALAT1-treated samples were compared with PBS-treated ones. P < 0.05 means that the difference is considered significant.

Results

Potent MALAT1 knockdown in human corneal endothelial and mouse epithelial cell lines

Experiments using ASOs are notoriously prone to “off-target” effects on gene expression that are unrelated to recognition of the intended target gene [28]. To control these effects, we used two noncomplementary controls in addition to the benchmark anti-MALAT1 gapmer (Fig. 1A). The anti-MALAT1 gapmer used 2′-MOE nucleotides in the flanking regions [27].

FIG. 1.

Anti-MALAT1 MOE-gapmer reduced MALAT1 expression in human and mouse cell lines. (A) Table of ASOs and siRNA used in this study. (B) Relative levels of MALAT1 RNA after treatment with 25 nM concentrations of ASOs in F35T FECD patient-derived corneal endothelial cell line. (C) Result of mouse mMALAT1 knockdown after treatment with 5 or 25 nM of ASOs in NMuLi mouse epithelial cells. NT: no ASO added. t-Test: *P < 0.05; **P < 0.01; ***P < 0.001 relative to treatment with MC. ASO, antisense oligonucleotide; FECD, Fuchs' endothelial corneal dystrophy; MALAT1, metastasis-associated lung adenocarcinoma transcript 1; MOE, methoxyethyl.

We first tested the anti-MALAT1 gapmer in F35T human corneal endothelial cells using the cationic lipid as a transfection reagent. Compared with no treatment sample or negative control ASOs, the anti-MALAT1 gapmer knocked down MALAT1 expression by 90% at 25 nM concentration (Fig. 1B). The MALAT1 gapmer showed better inhibition relative to the analogous duplex RNA targeting MALAT1 (siMALAT1) (Fig. 1B). The anti-MALAT1 gapmer may be superior to the anti-MALAT1 siRNA because the gapmer was the result of a screen for a highly effective compound that can act as a general benchmark. Since the anti-MALAT1 gapmer is complementary to a sequence shared by human and mouse, we also tested the ASO and controls in a mouse NMuli epithelial cell line and observed efficient reduction in expression (Fig. 1C).

Reduced MALAT1 expression in ex vivo human cornea tissues

To test the use of ASOs in human cornea, donor cornea tissues were incubated with ASOs in Optisol corneal storage media without any transfection agent. After 6 days of incubation at 37°C, we dissected the donor corneas and collected the endothelial tissue monolayer, the tissue affected in FECD. After RNA extraction and qPCR analysis, we found that the anti-MALAT1 gapmer effectively reduced MALAT1 expression by about 70% at 2 μM and 80% at 5 or 10 μM concentrations (Fig. 2). The noncomplementary control oligonucleotide MC had no inhibition. These results demonstrate that the anti-MALAT1 ASO can enter human corneal endothelial tissue in the absence of any formulation and reduce target gene expression.

FIG. 2.

Anti-MALAT1 MOE-gapmer effectively reduced MALAT1 expression in human corneal endothelial tissue ex vivo. Human donor corneas were incubated with ASOs without transfection agent for 6 days. t-Test: **P < 0.01; ***P < 0.001 relative to treatment with MC.

Sustained MALAT1 knockdown in whole cornea

After establishing the ability of ASOs to effectively control gene expression in human corneal tissue ex vivo, we investigated whether ASOs could control gene expression in mouse eyes in vivo. We delivered ASOs both intravitreally (IV) and intracamerally (IC) into the aqueous and vitreous humor of mice eyes, respectively (Fig. 3A). Intravitreal injection is a common route of administration, especially for treatment of retinal disease, and has been the method of delivery for previous FDA-approved ASO for cytomegalovirus (CMV) retinitis [29,30].

FIG. 3.

Comparison of delivery methods in mouse eyes using anti-MALAT1 MOE-gapmer. (A) Diagram of eye outlining strategies for administering ASOs. (B) Analysis of MALAT1 RNA level using whole corneas. Fifty micrograms of ASO was injected one-time. Whole corneas were dissected after 3, 7, and 21 days. RPL19 was used as qPCR internal control. For day 3, IV, IC, n = 6; day 7, IV, n = 6; IC, n = 5; day 21, IV, IC, n = 6. (C) Analysis of MALAT1 RNA level in mouse retina tissues after 7 days ASO treatment. IV, n = 6; IC, n = 5. t-Test: *P < 0.05; **P < 0.01; ***P < 0.001 relative to PBS treatment in each group. IC, intracameral injection; IV, intravitreal injection; Naive, no treatment samples; PBS, phosphate-buffered saline; qPCR, quantitative polymerase chain reaction.

Fifty micrograms of anti-MALAT1 ASO was injected into the mouse right eye, while PBS was used in the left eye as a control. Mouse eye globes were harvested at 3, 7, and 21 days after injection. Five to six mouse eyes were used for each group. RNAs from whole corneal tissues were used for qPCR analysis. RPL19, a ribosomal RNA was used as an internal control. We observed ASO-mediated knockdown of MALAT1 regardless of mode of administration (Fig. 3B). The left graph shows data point distribution for each treatment group, whereas the right shows averaged numbers. For IV injection, ASO achieved 82% inhibition for day 3, 65% for day 7, and 32% for day 21. For IC injection, percentage of inhibition was similar, with 72% for day 3, 54% for day 7, and 43% for day 21. The MALAT1 expression in PBS treatment groups is similar with that in naive control group.

We also quantified MALAT1 expression in neurosensory retinal tissue after day 7 of ASO treatment to further compare route efficacy (Fig. 3C). Potent inhibition was observed, with 89% inhibition for IV and 80% for IC injection. The results suggest that ASO delivery to the anterior chamber after intracameral injection can effectively diffuse into the posterior segment and reduce gene expression in the retina.

Topical delivery of ASO to mouse eye achieved no inhibition

We also applied ASOs as eye drops to the mice to test the hypothesis that topical administration could be effective. Such administration would require that the oligonucleotide to penetrate the outer surface of the eye. We had modeled our dosage concentration based on previous ASO topical administration experiments on murine eyes [11]. No inhibition of MALAT1 expression was observed after a 7 day twice-daily course of ASO application (Fig. 4A), suggesting that further innovation would be necessary to elucidate whether topical application could be achievable with our ASO.

FIG. 4.

Comparison of delivery methods in mouse eye in vivo using anti-MALAT1 MOE-gapmer. (A) Result of topical delivery of ASO to mouse eye. RNA from whole cornea was analyzed. (B) Analysis of MALAT1 RNA level using EDM tissue. ASO was injected one time. EDM monolayers were dissected for analysis after 7 days. Each data point represents a pool of RNA from four to five mouse eyes. t-Test: *P < 0.05; **P < 0.01 relative to PBS treatment in each group. EDM, endothelium-Descemet's membrane.

Efficacy in corneal endothelium

In FECD, corneal endothelium and its underlying basement membrane (Descemet's membrane) is the major affected tissue. The corneal endothelium is the inner surface of the cornea and is a relatively homogeneous single-cell layer. It is ∼1 mm from the outer surface of the eye. Because the corneal endothelium is a single layer of cells, it is difficult to obtain sufficient material for analysis from donor human specimens and obtaining material from the much smaller mouse eyes presented even greater challenges.

After developing suitable protocols for obtaining corneal endothelial tissue from mice for analysis, we tested the anti-MALAT1 ASO for inhibition at 7 days after IV or IC administration. From our preliminary quantification (data not shown), total RNA extraction yield was inconsistent even in the control-treated eye, indicating that tissue collection was not enough to confidently extrapolate information. To increase RNA yields for analysis, we pooled together total endothelial tissue from four to five treated mouse globes to ensure adequate RNA yields for quantification. We observed efficient inhibition of MALAT1 expression in corneal endothelium. Inhibition efficiencies were 64% and 63% for IV and IC administration, respectively. The PBS-treated or untreated groups showed no inhibition of MALAT1 expression (Fig. 4B). These data are consistent with the knockdown of MALAT1 expression in the whole cornea (Fig. 3B).

Discussion

Intravitreal and intracameral injections are routine routes of drug delivery for ocular pathologies such as diabetic retinopathy, age-related macular degeneration, and endophthalmitis [29–31]. In this study, we have demonstrated that these routes of administration can also be used to effectively deliver ASOs to the cornea. Despite the anatomical distance differences between aqueous and vitreous chambers relative to the cornea (Fig. 3A), the efficacy of both injection routes was similar at all time points (Fig. 3).

MALAT1 expression levels in whole cornea were diminished >80% after 3 days following intraocular administration and persisted up to 21 days postinjection. This long duration of effect is consistent with previous animal studies delivering ASOs with 2′-MOE modifications (similar to our construct) through intravitreal injection to treat CMV retinitis in monkey eyes, which showed an ASO half-life of 2 months in the retina [32] and also with observation of ASOs that are delivered to nondividing cells in other organs.

MALAT1 expression begins to recover during that 21-day period. Drug elimination from the cornea differs greatly from retinal tissue clearance and can be affected by multiple factors, including rate of drug metabolism within corneal cells, rate of turnover of corneal cell layers, and clearance of drug from fluid chambers [33]. Basal corneal epithelial cells in rats are replaced every 3 days and complete turnover of the corneal epithelium occurs within 12–15 days [34]. Epithelial turnover may be a major contributing factor to ASO elimination given the constant plasticity of the superficial corneal layer, which explains the partial temporal recovery of MALAT1 expression after 3 weeks in our animal model.

We also tested topical administration, modeling our protocol on that used to test Aganirsen, a topical ASO in Phase III clinical trials for patients with corneal neovascularization [10–13]. Topical drug application is the most widely used treatment in ophthalmology and is an ideal drug delivery strategy, given its simplicity and noninvasive nature. It minimizes the risk of infection, bleeding, and patient discomfort [35].

We did not observe MALAT1 knockdown after topical delivery. Unlike intraocular injections that directly deliver ASO into defined ocular cavities, naked negatively charged ASO topical solution must traverse across a heterogeneous tear film [consisting of a meibum (oil), aqueous and mucin layers] and several corneal epithelial cell lipid bilayers [36], leading to an ocular bioavailability of <5% [37]. In addition, dwell time on ocular surfaces is hindered by frequent blinking, as opposed to intraocular delivery, in which a drug can incubate within confined chambers, conferring a much slower turnover rate [37].

Balanced against the advantages of topical administration, intracameral or intravitreal administration would eliminate issues with patient compliance, avoid the need to penetrate the outer surface of the eye, and ensure effective drug delivery within the eye. For nondividing cells such as those in the corneal endothelium, the infrequent dosing used for other antisense drugs may be sufficient. Because diseases such as FECD lead to loss of vision and the only other currently effective treatment is corneal transplantation, infrequent intracameral or intravitreal administration may be an acceptable option for patients.

In addition to MALAT1 RNA reduction in mouse whole cornea, we demonstrated significant knockdown in isolated mouse corneal endothelium postinjection. The number of corneal endothelial cells in mice is small and these experiments demonstrated that analysis of gene expression after ASO knockdown could be achieved. Our results suggest ASOs have the potential to be developed for treating patients with corneal endothelial disease, such as FECD. The root cause of FECD is an expanded trinucleotide repeat and our laboratory and others have shown that targeting the trinucleotide expansion using ASOs block the mutant repeat RNA and reverse splicing defects associated with FECD.

In summary, these data suggest that in vivo delivery of ASOs to corneal tissue through intraocular injection is a feasible and effective method for regulating gene expression. Topical delivery experiments were unsuccessful but might benefit from further investigations into formulations that might optimize drug penetration. Modulating gene expression in the corneal endothelium with ASOs may be a viable method of treating genetic corneal disease such as FECD.

Author Disclosure Statement

T.P.P. and F.R. are employees of Ionis Pharmaceuticals. The other authors report no conflicts of interest.

Funding Information

This study was supported by grants R01EY022161 (V. Vinod Mootha), and R35GM118103 (David R. Corey) from the National Institutes of Health, Bethesda, MD, an unrestricted grant from Research to Prevent Blindness (RPB), a Core Grant for Vision Research (P30EY030413), Harrington Scholar-Innovator Award from Harrington Discovery Institute (V. Vinod Mootha), the Alfred and Kathy Gilman Special Opportunities in Pharmacology Fund (David R. Corey), and the Robert A. Welch Foundation I-1244 (David R. Corey). JDH was supported by a Career Development Award from RPB and a National Eye Institute R01 grant (EY027785). V. Vinod Mootha is the Paul T. Stoffel/Centex Professor in Clinical Care. David R. Corey is the Rusty Kelley Professor of Biomedical Science.