Steroid-eluting contact lenses for corneal and intraocular inflammation

Abstract

Ocular inflammation is one of the leading causes of blindness worldwide, and steroids in topical ophthalmic solutions (e.g. dexamethasone eye drops) are the mainstay of therapy for ocular inflammation. For many non-infectious ocular inflammatory diseases, such as uveitis, eye drops are administered as often as once every hour. The high frequency of administration coupled with the side effects of eye drops leads to poor adherence for patients. Drug-eluting contact lenses have long been sought as a potentially superior alternative for sustained ocular drug delivery; but loading sufficient drug into contact lenses and control the release of the drug is still a challenge. A dexamethasone releasing contact lens (Dex-Lens) was previously developed by encapsulating a dexamethasone-polymer film within the periphery of a hydrogel-based contact lens. Here, we demonstrate safety and efficacy of the Dex-Lens in rabbit models in the treatment of anterior ocular inflammation. The Dex-Lens delivered drug for 7 days in vivo (rabbit model). In an ocular irritation study (Draize test) with Dex-Lens extracts, no adverse events were observed in normal rabbit eyes. Dex-Lenses effectively inhibited suture-induced corneal neovascularization and inflammation for 7 days and lipopolysaccharide-induced anterior uveitis for 5 days. The efficacy of Dex-Lenses was similar to that of hourly-administered dexamethasone eye drops. In the corneal neovascularization study, substantial corneal edema was observed in rabbit eyes that received no treatment and those that wore a vehicle lens as compared to rabbit eyes that wore the Dex-Lens. Throughout these studies, Dex-Lenses were well tolerated and did not exhibit signs of toxicity. Dexamethasone-eluting contact lenses may be an option for the treatment of ocular inflammation and a platform for ocular drug delivery.

Graphical Abstract

1. Introduction

Anterior ocular inflammation can be caused by trauma, ocular surgery and by numerous infectious and non-infectious diseases [1–3]. Corneal neovascularization (CNV) and inflammation is a significant cause of diminished corneal transparency and subsequent reduction of vision [1, 4]. Diseases and conditions associated with CNV include inflammatory disorders, infectious keratitis, contact lens-related hypoxia, alkali burns, stromal ulceration, and limbal stem cell deficiency [5]. CNV is also a major risk factor for graft rejection after corneal transplantation [4, 6]. Anterior uveitis, which is the inflammation of the iris, ciliary body and aqueous humor is a common form of anterior ocular inflammation representing 30–90% of all uveitis cases [7].

Corticosteroid eye drops are the mainstay of therapy for the prevention and treatment of corneal transplant rejection, postoperative inflammation and anterior uveitis [2, 4, 8–12]. Following a corneal transplant, steroid eye drops are used for extended periods (1 to 2 years) and in some cases, for the remainder of a patient’s lifetime [13–15]. Unfortunately, eye drops are an inherently inefficient delivery system with very low bioavailability [16, 17]. As a result, the frequency of drop administration is increased to provide tissues with therapeutic drug levels. For non-infectious inflammatory conditions like acute uveitis, steroid eye drops are commonly prescribed hourly [2, 10]. Understandably, patient adherence with such intensive treatment regimens can be difficult. Also contributing to non-adherence, eye drops frequently sting, burn, and cause a transient blurring of vision upon application [16, 18, 19]. Due to the challenges associated with steroid eye drops, patient friendly methods of sustained steroid delivery like sustained release formulations, devices and implants, nanoparticles and dissolvable wafers have been proposed as a means of improving the treatment of anterior ocular inflammation [12, 20–23].

Drug-eluting contact lenses can potentially provide sustained drug levels to the cornea and improve treatment efficacy and patient adherence [19, 24, 25]. Unmodified contact lenses soaked in drug solutions release the drug within minutes or hours and the amount of drug absorbed by the lens may not be therapeutically significant [19, 26–28]. Thus, contact lenses need to be modified such that they can control drug release as well as load a therapeutically relevant drug amount. We have incorporated a drug-polymer film within the periphery of a standard contact lens so that it provides sustained release of drug to the eye [29, 30]. We have previously demonstrated the ability to provide sustained release of latanoprost in rabbits and monkeys [31, 32] as well as dexamethasone in rabbits [33]. To the best of our knowledge, there have been no studies investigating the efficacy and safety of sustained dexamethasone delivery from dexamethasone-eluting contact lenses (Dex-Lens) in experimentally induced anterior ocular inflammation.

We had previously shown that the Dex-Lens released dexamethasone in vitro and in vivo in a sustained manner for at least 7 days and was able to treat VEGF-induced retinal vascular leakage in a rabbit model [33]. The aim of this study is to evaluate the safety and efficacy of Dex-Lenses on experimentally induced CNV and anterior uveitis in rabbit models and to compare the efficacy to that of intensive dexamethasone eye drop therapy. [33]Studies of cytotoxicity and ocular irritation were performed to evaluate the safety and biocompatibility of the Dex-Lens. As an animal model, rabbits were chosen because they can wear contact lenses designed for human use [34].

2. Materials and Methods

2.1. Fabrication of dexamethasone-eluting contact lenses

Pharmaceutical-grade dexamethasone (Spectrum Chemical, New Brunswick, NJ) was dissolved in hexafluoro isopropanol (HFIP) (Sigma Aldrich), along with GMP grade poly (lactic-co-glycolic) acid (PLGA 85-15, Evonik Degussa) at a concentration of 60 mg/mL for each component. A total of 40 μl of the combined solution was then pipetted onto a concavity that had been lathed into a cylinder of dry polymerized methafilcon (Kontur Kontact Lens Company). Methafilcon is a commonly used contact lens hydrogel with a high water content (55%) and consists of consists of poly(hydroxyethyl methacrylate) and methacrylic acid. After rotation on a spin coater (Model SC100B, Best Tools, LLC, St. Louis, MO) for 6 minutes, the HFIP evaporated and only a drug-polymer film remained. A central aperture was cut from the film using a 6-mm biopsy punch (Sklar) (Fig. 1). The film was dried by desiccation and lyophilization. 9 mg/ml of a UV photo-initiator (Irqacure 2959, Sigma Aldrich) was dissolved into liquid monomer for methafilcon (Kontur Kontact Lens Company). The drug-polymer film was encapsulated by liquid monomer using ultraviolet photopolymerization (400 W metal halide bulb, Loctite Corporation, Rocky Hill, CT) to recreate a hydrogel cylinder block of methafilcon with the drug-polymer film encapsulated inside in the form of a ring. The methafilcon block was then lathed into a contact lens that consisted of the drug-PLGA film fully encapsulated in methafilcon. The lenses were stored in airtight glass vials. Finally, the glass containers that were holding the lenses were placed in a temperature-controlled container and terminally sterilized by irradiation in a Gamma Cell 220E Cobalt 60 Irradiation Unit (Atomic Energy of Canada Ltd., Ottawa, Canada) with a total dose administration of 25 kGy.

Fig. 1:

A) Schematic of Dexamethasone contact lens (Dex-Lens) that illustrates the drug-polymer film completely encapsulated within the periphery of the contact lens. B) Ocular Coherence Tomography (OCT) cross-section of Dex-Lens demonstrates the drug-polymer film within the contact lens hydrogel.

2.2. Cytotoxicity

Cytotoxicity of the Dex-Lens was tested using a minimum essential media (MEM) elution assay. L929 murine fibroblasts were purchased from ATCC (Manassas, VA) and grown in Dulbecco’s Modified Eagle’s Medium (DMEM) with 10% FBS, 1% penicillin-streptomycin, and 4 mM l-glutamine. Cells were plated in 96 well plates at a density of 1×10⁵ cells/ml. Dex-Lenses were immersed in DMEM at a SA/V ratio of 6 cm²/mL for 24 hours at 37°C. 100 μL of this media (extract) was applied neat to the cells. 11 lenses were used for each group, with 4 to 6 wells used for each lens. Similarly, vehicle lenses were extracted in DMEM media and the extracted media was applied to the cells. To test the effects solely attributed to the drug dexamethasone alone was tested in additional wells. Dexamethasone was dissolved in DMSO (Sigma Aldrich) and spiked at concentrations of 180, 90, and 18 μg/ml. 6 wells were used for each concentration. Cells were incubated with extract or components for 24 hours. DMSO alone was added to the cells as a control. MTT Assay (Abcam, Cambridge MA) was used to measure cell viability. Results were compared to cells that received no extract or dexamethasone.

2.3. Animal studies

Female New Zealand White (NZW) rabbits, weighing 2–3 kg, were used in all studies. All the study protocols were approved by the Institutional Animal Care and Use Committee of Massachusetts Eye and Ear (pharmacokinetics) or the Schepens Eye Research Institute (all others). All animals were treated according to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research.

2.3.1. Ocular irritation study of Dex-Lens

An ocular irritation study (Draize test) was used to further assess potential Dex-Lens adverse effects [35]. Dex-Lenses were immersed in sterile saline or cottonseed oil (Spectrum Chemical, New Brunswick, NJ) at a SA/V ratio of 6 cm²/mL (~ 0.8 ml) and incubated at 37°C for 72 hours. Two types of extraction liquids were used to capture both polar and non-polar leachables. 100μL extract was instilled in the conjunctival cul-de-sac of one eye in 3 rabbits. The contralateral eye remained untreated and served as a control. A physical exam was conducted daily, and both eyes were graded according to the Organization for Economic and Cooperative Development (OECD) grading scale for ocular irritation [36]. The scale evaluates the cornea, iris, conjunctiva, and chemosis separately on a 0–4 scale. Slit lamp photos taken 1, 24, and 72 hours after extract instillation. 3 lenses were used for each extraction liquids, and each rabbit had extract applied from one lens.

2.3.2. In vivo pharmacokinetics

NZWs (n=4) received a permanent lateral tarsorrhaphy in the right eye. NZWs were put under general anesthesia (30 mg/kg ketamine, 5mg/mL xylazine, and 1mg/mL acepromazine). General anesthetic was augmented with Lidocaine 1% with Epinephrine 1:100,000. A thin strip of skin was removed along the lid margin of the outer 1/3 of the upper and lower lids, and the lid margins were sutured using 6-0 Vicryl. Tarsorrhaphies healed for one week prior to additional procedures.

A Dex-Lens soaked in sterile saline for one hour for hydration was placed in the eye that received the tarsorrhaphy. Under general anesthesia at predetermined time points, anterior chamber paracentesis was performed. To accomplish this, the contact lens was slid to the side of the cornea and a 31-gauge (G) needle was inserted in the superior limbus and 100 μL aqueous humor (AH) was withdrawn. Time points were arranged such that NZWs did not undergo anterior chamber paracentesis more often than every other day. Four sets of lenses were used on each NZW to obtain all time points and hence it was not a single lens that was utilized at all time points. There was a minimum three-day washout period between lenses. Results were compared to NZWs that received dexamethasone drops hourly for 3 hours and had aqueous humor sampled within an hour of the last drop. AH samples were diluted in methanol to precipitate proteins, centrifuged, and filtered. Dexamethasone in AH samples was quantified by LC/MS-MS with an Agilent 6410 triple quad LC/MS-MS that used ESI as the source. The gradient mobile phase was 0.1% acetic acid (Sigma Aldrich) and 0.1% acetic acid in acetonitrile (Sigma Aldrich). The column was an Agilent C18 2.1 × 150 mm (3 μM), with a flow rate of 0.25 mL/min. The polarity was positive and scan type was MRM, with dwell energy of 200. Collision energy was 4eV and fragmentation voltage was 100V.

2.3.3. In vivo efficacy in a CNV model

A modification of a previously described suture technique was used to induce CNV [37, 38]. Two 7-0 silk sutures were placed in the inferior and superior cornea of the right eye so that the sutures made a horizontal “figure 8” pattern (Fig. 3A). The two corneal bites ran parallel to the superior limbus. The bites were 3 mm in length and were placed 2 and 4 mm from the limbus.

Fig. 3:

A) Photograph of sutures that were placed along the superior and inferior cornea to induce cornea neovascularization (CNV). B) Representative red free slit-lamp photographs from the four study groups 7 days after suture placement. The CNV invasion area was demarcated by a yellow line traced using ImageJ. C) The CNV invasion area on Days 5 and 7, which was calculated from the tracings generated by two masked cornea specialists (n = 6 per group). Dexamethasone drops (Dex Drops) and dexamethasone contact lens (Dex-Lens) had significantly less CNV invasion area compared to eyes that received no treatment or the vehicle lens. Data are mean ± standard deviation. P-values on day 7 as compared to no treatment on day 7 (Tukey’s HSD post hoc test).

Each rabbit was then exposed to one of the following regimens for one week: 1) no treatment, 2) 0.1% topical dexamethasone sodium phosphate drops (1 drop every hour for 8 hours daily; Dex Drops), 3) one Dex-Lens, 4) one vehicle contact lens with the polymer film but without dexamethasone (n = 6 per group). The Dex-Lenses and the vehicle contact lenses were soaked in sterile saline for one hour before insertion. We placed a temporary lateral tarsorrhaphy using 5-0 Nylon to improve lens retention. Unlike a permanent tarsorrhaphy, the temporary tarsorrhaphy involves no incision and can be completely undone by removing the eyelid suture. In the interest of avoiding confounding results, the tarsorrhaphy was performed on all animals. All corneal sutures were performed under the operating microscope by the same investigator. The left eye was untreated.

2.3.3.1. CNV study analysis

All animals were evaluated by slit-lamp biomicroscopy (Topcon Medical Europe BV, Capelle a/d IJssel, Netherlands) and spectral-domain ocular coherence tomography (OCT; Spectralis, Heidelberg Engineering, Heidelberg, Germany) under general anesthesia. Superior and inferior corneas were evaluated by slit-lamp biomicroscopy and OCT on days 5 (120 hr) and 7 (168 hr). At each measurement, the tarsorrhaphy and contact lenses were removed. On days 5 and 7, red-free calibrated slit-lamp biomicroscopy photographs and anterior segment OCT measurements were taken to quantify the CNV and to measure the width of the cornea suture, respectively. At least three consecutive sets of measurements with the two devices were performed by a single experienced examiner. On day 7 of the study the animals were euthanized using 120/mg/kg IV Euthasol. The corneas were harvested and cut horizontally into two parts. The superior cornea was analyzed for the expression of vascular endothelial growth factor (VEGF)-A and VEGF receptor (KDR) by real time PCR, VEGF-A and VEGF-KDR by qRT-PCR and inferior cornea was analyzed for the frequency of CD45+ immune cells by flow cytometry as described below.

2.3.3.2. CNV image analysis

Red free slit-lamp biomicroscopy photographs at baseline, day 5 and day 7 were randomized. A research technician visually selected the best image with respect to focus and contrast in each measurement set. CNV invasion area of superior and inferior cornea was evaluated by two corneal specialists who were masked to the treatment groups. All slit-lamp biomicroscopy images were exported from the system as a portable network graphics image file into ImageJ 1.38X (National Institutes of Health, Bethesda, Maryland, USA) software for analysis using a previously described method (50). The invasion area of anterior or inferior CNV was outlined and converted to region of interest (ROI) file. The invasion area of CNV was calibrated using the ROI file based on the length of 7-0 silk suture by OCT. The total CNV invasion area was calculated in the superior and inferior cornea for each rabbit.

2.3.3.3. RNA isolation and real-time PCR

Corneal tissues were harvested and lysed to isolate total RNA using the RNeasy Micro Kit (Qiagen, Germantown, MD, USA). Purified RNA was then reverse transcribed into single-stranded cDNA using random hexamer primers and Superscript III Kit (Invitrogen, Carlsbad, CA, USA). Real time PCR was performed in Mastercycler Realplex2 (Eppendorf, Hamburg, Germany) using TaqMan Universal PCR Master-mix (Qiagen, Germantown, MD, USA) and preformulated TaqMan primers. Primers for GAPDH (Oc03823402_g1), VEGF-A (Oc03395999_m1) and VEGF receptor KDR (Oc03395666_m1) were obtained from ThermoFisher Scientific (Waltham, MA, USA). To evaluate VEGF A and KDR expression, amount of VEGF-A and KDR mRNA were normalized with that of GAPDH (internal control) in each sample and results were analyzed with the comparative threshold cycle method [39, 40].

2.3.3.4. Corneal tissue digestion

Corneas were digested to prepare single cell suspensions as described previously [41, 42]. Briefly, corneal tissues were incubated in RPMI media (Lonza, Walkersville, MD, USA) that contained 2 mg/mL collagenase type IV (Sigma-Aldrich Corp.) and 2 mg/mL DNase I (Roche, Basel, Switzerland) for 45 minutes at 37°C, and then filtered through a 70-μm cell strainer for flow-cytometric analysis.

2.3.3.5. Flow cytometry to assess CD45+ cell frequency

To evaluate infiltration of immune cells, single cell suspensions of corneas were prepared and stained with phycoerythrin (PE) conjugated anti-CD45 monoclonal antibodies (clone MEM55) or matched isotype (negative control) [43]. As previously described [44, 45] cells were incubated with antibody or isotype control for 45 minutes in dark at 4℃ and then washed with FACS buffer (2% FBS in PBS). Stained cells were acquired and analyzed using LSR II flow cytometer (BD Biosciences, San Jose, CA, USA) and Summit software (Dako Colorado, Inc., Fort Collins, CO, USA). Antibody and isotype control were purchased from BioLegend (San Diego, CA, USA).

2.3.3.6. Safety evaluation of Dex-Lens

The safety of the Dex-Lens in the CNV study was assessed by slit lamp examination and by IOP measurement. The rabbit eyes were initially stained with fluorescein at day 7 and red free slit lamp photographs were taken. A masked cornea specialist analyzed the photographs for staining by using a standardized grading system [46] where the cornea is divided into 5 regions and each region is assigned a grade from 0 (no staining) to 3 (high staining). The grades were totaled to give the final value.

The intraocular pressure (IOP) was measured with a Tono-Pen XL applanation tonometer (Reichert Inc., Depew, NY, USA) at baseline, day 5, and day 7.

2.3.4. Efficacy assessment in a rabbit model of uveitis

Lipopolysaccharide (LPS) from Escherichia coli was diluted in sterile saline solution at a concentration of 100 ng LPS per 10 microliters of sterile saline. NZWs were injected with 10 microliters of LPS solution through the cornea into the aqueous humor using a 30G needle connected to a Hamilton syringe. Prior to the LPS injection, rabbit eyes were examined under an operating microscope and IOP was measured.

The rabbits were randomized into one of three treatment arms for 5 days (n=5): 1) 1 drop of artificial tears every hour for 8 hours daily, 2) 1 drop of 0.1% dex drops every hour for 8 hours daily, 3) one Dex-Lens. At day 1 (24 hr), 3 (72 hr), and 5 (120 hr) the eyes of anesthetized animals were examined under the operating microscope. While anesthetized, a 31G needle was inserted through the superior limbus and used to sample 100 mL of aqueous humor during the same time points. To serve as an objective surrogate biomarker for ocular inflammation, the protein concentration within the aqueous humor sample was measured using a Bradford Assay. After the 5^th day of treatment, examination, and aqueous humor removal, the rabbits were euthanized.

2.4. Statistical analysis

All continuous data was expressed as mean ± standard deviation and statistical analysis was done by either Student t-test, ANOVA or multiple comparison.

3. Results

3.1. Dex-Lens was manufactured with a clear aperture

A ring-shaped dexamethasone-polymer film (Fig. 1A and 1C) was encapsulated within the periphery of a methafilcon contact lens hydrogel as described in Methods (Fig. 1A). The drug-polymer film was composed of a 1:1 ratio of dexamethasone and PLGA 85:15 and contained approximately 1.5 mg of dexamethasone [33]. When hydrated, the Dex-Lens had an outer diameter of 15.4 mm. The ring of drug-polymer film had an outer diameter of 11.7 ± 0.4 mm and an inner diameter of 7.4 ± 0.2 mm. The center of the Dex-Lens was clear to allow for unobstructed light transmission (Fig. 1B). The Dex-Lens released dexamethasone in vitro in a sustained manner for at least 7 days [33].

3.2. Cytotoxicity of dexamethasone solutions and the Dex-Lens

Minimum essential media (MEM) elution assay was employed to determine the cytotoxicity of the Dex-Lens [47]. Dex-Lenses were soaked in DMEM for 24 hours to extract any leachables and L929 murine cells plated in a 96-well plate were exposed to the Dex-Lens extract and incubated for 24 hours. Dexamethasone concentration in the Dex-Lens extract was measured by HPLC to be 180 μg/ml by method described previously [33]. The viability of cells exposed to Dex-Lens extract was 74.5±21.6% (n=11) relative to untreated cells. For the vehicle lens, the viability of cells was found to be 83.3±13%. The finding of steroids resulting in cell viability loss is consistent with the literature [48–50]. To verify the hypothesis that the loss in cell viability is in part due to dexamethasone, solutions of different concentrations of dexamethasone (18, 90, 180 μg/ml) were added to the cells directly. From the concentrations tested, an inverse dose related response was observed: higher dexamethasone concentrations resulted in lower cell survival rates of 68.7±10.4%, 60±10% and 38±18% for 18, 90, 180 μg/ml of dexamethasone respectively after 24 hours of incubation, similar to results elsewhere [49, 50] confirming the hypothesis that the loss in cell viability with the Dex-Lens extracts is due to dexamethasone itself and the other leachables from Dex-Lens are not leading to cytotoxicity. The cell viability with DMSO was 97.7±9.4% indicating that the loss of viability associated with dexamethasone was not due to the DMSO that was used as the solvent to make dexamethasone stock solution.

3.3. Dex-Lens extract did not lead to ocular irritation in healthy rabbit eyes

An ocular irritation study (Draize test) was performed to investigate potential ocular adverse effects from extracts released from the Dex-Lens in normal rabbit eyes [35]. According to the Draize eye test protocol, both polar and non-polar leachables were extracted with different liquids. 100 μL of the extract was instilled in the conjunctival cul-de-sac of the right eye of the rabbit (n=3 per group) and the eye was closed manually for 30 s. The left eye remained untreated. Both eyes (treated [right] and untreated [left]) were examined by slit lamp biomicroscopy by an ophthalmologist and graded using the OECD grading scale for ocular irritation at 1, 24, and 72 hours after extract instillation [36]. The scale evaluates the cornea, iris, conjunctiva, and chemosis separately on a 0–4 scale where 0 = normal and higher grades indicate complications. The rabbits did not develop any corneal or conjunctival lesions or abnormalities. For all time points and tissues, the scoring remained unchanged from the baseline indicating that both Dex-Lens extracts (polar and non-polar) did not cause ocular irritation.

3.4. Dex-Lenses provided sustained delivery of dexamethasone for 1 week to aqueous humor

The release of dexamethasone in vivo was measured by placing Dex-Lenses on the right eyes of rabbits and sampling the aqueous humor at predetermined time points. Separately, rabbits also received 1–2 drops of 0.1% dexamethasone phosphate ophthalmic solution (dex drops) every hour for 3 hours. Within an hour after the 3^rd drop was instilled, the aqueous humor was sampled. Dexamethasone concentration was measured by LC/MS-MS in the sampled aqueous humor.

Eyes that wore the Dex-Lens maintained dexamethasone levels in the aqueous humor that were significantly greater than that from dexamethasone drops at most time points (Student t-test, Fig. 2). Peak concentration of dexamethasone in the aqueous humor from the Dex-Lens was 3500±2100 ng/mL at 3 hours after lens insertion. In comparison, the peak concentration of dexamethasone in the aqueous humor from dexamethasone drops was 28±10 ng/ml after placement of the 3^rd hourly drop. The peak aqueous humor concentration observed in this study from dexamethasone drops was comparable to previous reports [33, 51, 52].

Fig. 2:

Dexamethasone concentration in the aqueous humor of rabbits after 3 sets of hourly dexamethasone 0.1% drops (left) and during 7 days of dexamethasone contact lens (Dex-Lens) wear (right). At most time points, the dexamethasone levels were significantly higher with a Dex-Lens than dexamethasone drops (n = 4 per time point). Data are mean ± standard deviation. p -values as compared to hourly drops by student t-test.

3.5. Dex-Lens inhibited neovascularization and inflammation in a suture-induced cornea neovascularization (CNV) rabbit model

Silk sutures were used to induce corneal neovascularization. Immediately after placement of the sutures, the study eye had a temporary lateral tarsorrhaphy and underwent one of the following regimens for one week (n=6 per group): 1) no treatment, 2) hourly 0.1% dexamethasone sodium phosphate drops (dex drops; 8 drops per day), 3) one Dex-Lens, 4) one vehicle contact lens with the polymer film but without dexamethasone. Calibrated red-free slit lamp photographs were taken at baseline following suture placement and then at five and seven days later (Fig. 3B shows representative examples of corneal photographs in each group on day 7). In the study, the CNV invasion area on the cornea was analyzed on days 5 and 7 by two independent and masked cornea experts using ImageJ. The CNV invasion area was reported as % CNV inversion area by dividing the CNV invasion area by the total corneal area. The difference in the measured CNV area between the treatment groups was significant at day 5 and day 7 (p<0.0001 by ANOVA). The Dex-Lens group had a significantly smaller CNV invasion area compared to the corneas that received no treatment and also less than corneas that received the vehicle lens (p<0.0001, multiple comparison using Tukey’s HSD post hoc test). The Dex-Lens group had almost 75% smaller CNV invasion area (2.8±1.1%) than the no treatment group (11.2±1.5%, p<0.001, Tukey) at the end of 5 days and 70% smaller CNV invasion area (4.6±2.3%) than the no treatment group (14.5±2.1%, p<0.001, Tukey) at the end of 7 days. There was no difference between no treatment and vehicle lens groups on either day (p=0.26 and 0.06 on Days 5 and 7, respectively). The CNV invasion area in the Dex-Lens group was comparable to the dex drops group (Fig. 3C). The intra-class correlation coefficient (95% confidence interval) for CNV invasion area was 0.96 and 0.96 in inter-observer and intra-observer measurements indicating good agreement between the two observers and the measurements repeatability as calculated using established methods [53, 54].

3.5.1. Biological markers of inflammation and neovascularization

After euthanization, the NZW corneas were harvested to investigate effect of Dex-Lens on infiltration of inflammatory cell and angiogenic mediators. Each excised cornea was equally divided into two halves horizontally. The inferior half of corneas was digested to prepare single cell suspension for flow cytometry analysis of CD45+ cells, a pan-leucocyte marker for inflammatory cells. The CD45+ infiltration is reported as frequency (% of cells in cornea). The infiltration of CD45+ cells was statistically significantly reduced in the Dex-Lens treated group (5.0±4.5%) compared to the no-treatment group (14.8±4.1%, p=0.0015, Tukey). Moreover, the frequencies of CD45+ cells in dex drops treated group (4.8±2.7%) were significantly lower relative to untreated controls (p=0.0013, Tukey), but were comparable to Dex-Lens treated group (p=0.99).

The superior cornea was lysed to isolate total RNA and expression of angiogenic mediator VEGFA and its receptor VEGFR2 (KDR) KDR, that are indicators of neovascularization were quantified by qRT-PCR real time PCR (Fig 4C). Dex-Lens treated group showed a significant reduction in expression of KDR (compared to no-treatment (p<0.001, Tukey) and a reduction compared to vehicle-lens group (p=0.065, Tukey). Moreover, KDR expression in Dex-Lens was comparable to dex drops treated group (p=0.67, Tukey). The no-treatment group showed a higher expression of VEGF-A compared to dex drops treated group (p=0.04, Tukey) and Dex-Lens treated group (p=0.5, Tukey) (Fig 4C). VEGF-A expression was comparable between Dex-Lens and dex drops treated groups (p=0.45, Tukey). Thus, the reduced CD45+ cell infiltration as well as the reduced KDR and VEGF-A expression indicate supression of inflammation and neovascularization.

Fig. 4:

Corneas from various treatments were harvested 7 days after suture placement and cut in two halves for flow cytometry and real-time PCR. A) Single-cell suspensions of cornea were stained with anti-CD45 antibodies and flow cytometry was performed. A) Representative flow cytometry dot plot and histograms showing the gating strategies and frequencies of CD45+ cells (% in cornea) in different groups. B) Quantification of CD45+ cell frequencies (% in cornea) in indicated groups. (C) Bar chart showing VEGF-A and VEGFR-2 mRNA expression (normalized to GAPDH) in different groups as quantified by real-time PCR. Data are mean ± standard deviation. P-values as compared to no treatment (Tukey’s HSD post hoc test) (n = 6 per group).

3.5.2. Dex-Lens inhibited corneal edema caused by extended contact lens wear and tarsorrhaphy

Contact lens wear can cause temporary corneal edema; an increase in corneal thickness that can be exacerbated with overnight lens wear or continuous wear of lens over several days [55–58]. It is important to measure the extent of corneal edema as it can hinder vision by impairing transmission of light. Thus, the central corneal thickness (CCT) was measured at baseline and day 7 of the treatments by analyzing the OCT images of the cornea. CCTs increased by 35.5±15.8% (no treatment), 17±1.4% (dex drops), 8.9±11.8% (Dex-Lens) and 89.2±9.9% (vehicle lens) (mean±std. dev.) (Fig. 5). One of the NZWs wearing the Dex-Lens had an increase in corneal thickness of 32.51%, that can be considered a statistical outlier (greater than [mean + 2 × std. dev.]) and thus the average increase in Dex-Lens group was 4.1±2.6%. Table 1 gives the p-values for multiple comparisons between different groups (ANOVA in MATLAB). NZWs receiving no treatment had an average increase in CCT of 35.5% may be caused by suture placement and temporary tarsorrhaphy. The vehicle lens group had an average increase of 89.2% (p<0.001 compared to no treatment), which indicates that contact lens placement in this model can worsen corneal edema. The increase of CCT in the group treated with dex drops and Dex-Lens was significantly lower when compared with the eyes in the no treatment or vehicle lens groups (Table 1).

Fig.5:

A) Representative corneal OCT images for different treatment groups on day 0 and day 7. Dexamethasone contact lens (Dex-Lens) resulted in less of an increase in cornea thickness as the vehicle lens. B) Box plots of percentage change in central corneal thickness after 7 days of suture and treatment initiation. The red plus is outlier as calculated by MATLAB (n = 6 per group). Data are mean ± standard deviation. P-values as compared to no treatment (multiple comparison using Tukey’s HSD post hoc test).

Table 1:

p-values from Multiple Comparison (Tukey’s HSD post hoc test) for change in central corneal thicknesses for different groups (n = 6 per group) in the corneal neovascularization study.

	Dex Drops	Dex-Lens	Vehicle Lens
No treatment	0.003	0.003	<0.0001
Dex Drops		0.2092	<0.0001
Dex-Lens			<0.0001

3.5.3. CNV study: ocular surface health and intraocular pressure measurements

Mechanical forces from extended contact lens wear or toxicity from drugs can compromise corneal health and the integrity of the corneal epithelium [59] and thus a masked cornea expert graded the photographs for the amount of corneal staining using the mean NEI corneal fluorescein staining score [46]. There was minimal pinpoint punctate staining of the cornea in all study groups at baseline and 7 days after treatment (Fig. 6A; ANOVA, p=0.058).

Fig. 6:

Determination of safety of dexamethasone contact lens (Dex-Lens). A) NEI corneal fluorescein staining scores at Day 7 after suture placement. No significant staining was observed. B) Intraocular pressure (IOP) measurements on Day 0 and Day 7 after suture placement (n = 6 per group). Data are mean ± standard deviation. No significant difference was observed between various groups (Multiple Comparison using Tukey’s HSD post hoc test).

Intraocular pressure (IOP) was measured by an applanation tonometer at baseline and day 7 because ocular hypertension is a potential side effect of ocular steroids [60, 61]. Among all the treatment groups, there was no significant change in IOP at day 7 compared to the baseline (Fig.6B) (ANOVA, p=0.067).

3.6. Dex-Lens inhibited anterior uveitis induced by LPS

Anterior uveitis is characterized by an increase in aqueous humor protein, that can be measured as an objective and quantifiable marker of inflammation since a normal eye lacks protein in the aqueous humor [62]. In NZWs we tested the ability of Dex-Lenses to prevent intraocular inflammation within the anterior segment using an established model of non-infectious acute anterior uveitis induced by injection of LPS into the anterior chamber of one eye [63–65]. Aqueous humor protein levels peaked on day 1 after LPS injection in the NZWs treated with saline drops (Fig. 7). The aqueous humor protein levels in the saline group remained significantly higher than baseline throughout the study with p-values <0.001 on days 1, 3, and 5 (Student t-test) indicating continued effect of the LPS injection. Animals treated with Dex-Lenses did not have an increase in aqueous humor protein concentration following the LPS injection (Student t-test compared to baseline: day 1, p=0.88; day 3, p=0.55; and day 5, p=0.67) demonstrating the efficacy of Dex-Lens treatment.

Fig. 7:

In vivo efficacy for treatment of induced anterior uveitis. Aqueous humor protein concentration after endotoxin (lipopolysaccharide) injection into the aqueous humor followed by treatment with: 1) sterile 0.9% saline for 5 days, 2) 0.1% dexamethasone drops (Dex Drops) for 5 days, or 3) dexamethasone-eluting contact lenses (Dex-Lens) worn for 5 days (n = 5 per group). The aqueous humor protein concentration shows that the Dex-Lens was effective in inhibiting anterior uveitis over 5 days. Data are mean ± standard deviation. P-values are compared to baseline (or between treatments if indicated) using Student t-test.

NZWs treated with dexamethasone drops also had significantly greater aqueous humor protein concentrations on day 1 (p=0.04) compared to baseline, but not on day 3 (p=0.77) or day 5 (p=0.18). Previous studies that used LPS intracameral injections to induce inflammation found similar trends with inflammatory markers peaking during the first 24 hours and decreasing thereafter [63–65]. Animals treated with Dex-Lenses did not have an increase in aqueous humor protein concentration following the LPS injection (Student t-test compared to baseline: day 1, p=0.88; day 3, p=0.55; and day 5, p=0.67). When compared to animals treated with saline drops, NZWs treated by Dex-Lenses had significantly less protein concentrations on day 1 (p=0.028) and day 3 (p=0.016), but not on day 5 (p=0.25). NZWs treated with dex drops also had significantly less inflammation than those treated with saline drops on day 1 (p=0.034) and 3 (p=0.015), but not on day 5 (p=0.25).

4. Discussion

Contact lenses have long been sought as an ocular drug delivery route with the goal of improving ocular drug flux and patient compliance. A Dex-Lens, containing a dexamethasone-PLGA film, has been shown to deliver dexamethasone to ocular tissues at therapeutically relevant levels [33]. In this study, we demonstrate that the Dex-Lens is capable of inhibiting corneal and intraocular inflammation for an extended period of time. The Dex-Lens was found to be biocompatible and had similar efficacy to intensive eye drop therapy in multiple in vivo rabbit models. Clinically, the Dex-Lens could be placed by an eye care provider and worn for a week, thus replacing the need for topical steroids. The Dex-Lens could potentially replace the bandage contact lens typically placed by the eye care provider after several ocular procedures. The Dex-Lens could be particularly helpful for patients who require intensive (e.g. hourly) steroid drops. If steroids are needed for a longer period of time, the Dex-Lens could be replaced by the eye care provider weekly.

In the CNV study, the Dex-Lenses achieved a 70% reduction in CNV invasion area over 7 days as compared to no treatment. Typically, corneal neovascularization is accompanied by increased tissue infiltration of immune cells and higher expression of angiogenic factors including VEGF-A and its receptor VEGFR-2 (KDR). Our data demonstrate that the Dex-Lens suppresses infiltration of CD45+ cells and decreases the expression of angiogenic factors VEGF-A and KDR compared to untreated groups, which clearly suggest that the Dex-Lens exerted the therapeutic effect on inhibition of CNV invasion area by suppressing tissue inflammations and mediators of angiogenesis. This is consistent with the literature, which reports that both the anti-inflammatory and anti-lymphangiogenic properties of corticosteroids like dexamethasone are responsible for preventing corneal graft rejection [66, 67]. There are several other studies that have investigated treatment of CNV by sustained release drug delivery modalities. These include a single subconjunctival administration with dexamethasone sodium phosphate loaded PLGA nanoparticles over 2 weeks [68], axitinib in a dissolvable nanowafer repeatedly placed daily on cornea for 11 days [69], or by delivering decorin 4 times a day via a gellan gum for 16 days [70]. In comparison, only a single topically applied Dex-Lens was needed over a period of 1 week to effectively inhibit CNV. Additionally, all of the above studies were conducted in murine models and in general murine or rodent models are not suitable for testing ocular drug delivery systems owing to their small size [71]. We tested the Dex-Lens in rabbit models that are more suitable for testing ocular drug delivery devices since rabbit eyes have similar size to human eyes [71, 72].

Uveitis is a vision-threatening condition that may be caused by infectious or non-infectious etiologies, which result in ocular inflammation. Correspondingly, a rabbit model of LPS induced anterior uveitis was used to test the potential of Dex-Lens as a viable treatment option. The Dex-Lens was able to effectively inhibit anterior uveitis that was estimated by measuring aqueous humor protein concentration. Previously, a sustained release dexamethasone implant containing 700 μg of dexamethasone (Ozurdex^®; Allergan, Inc., Irvine, CA) had been evaluated for its effectiveness in the treatment of anterior uveitis in rabbits [73]. The implant was successful in reducing clinical and biological signs of uveitis; however, the implant carries a risk of migrating to the anterior chamber and the treatment may not be easily discontinued in case of steroid-related hypertension [74–76]. Another topical sustained release device (DSP-Visulex, [77]) was effective in reducing experimental uveitis in rabbits. This device needed to be applied on the eye for a few minutes (10 to 15 minutes) at least once a week. In contrast, the Dex-Lens has some advantages over DSP-Visulex including the fact that Dex-Lens requires almost no application time or any special apparatus unlike the DSP-Visulex.

The success of the Dex-Lens can potentially be attributed to its ability to sustain high drug concentration and to increase the drug residence time on ocular surface. In contrast, eye drops have to immediately overcome precorneal factors like tear dynamics and blinking that limit residence time of the drug in an eye drop on the ocular surface. Additionally, the various layers of the cornea and conjunctival vasculature act as further barriers to ocular drug delivery and reduce the amount of drug that penetrates the eye. As a result of these factors and others, the bioavailability of eye drops is likely limited. For instance, only 1–7% of the drug in a drop enters the eye [16, 18]. A drug-eluting contact lens with sustained drug release may increase the residence time of a drug on the ocular surface [78, 79] and overcome the tear film turnover leading to higher bioavailability. The exchange of fluid between the contact lens and the cornea (post-lens tear film) has been shown to be slower than that present on the ocular surface [80, 81]. With regards to ocular drug delivery, this is advantageous as the drug released from the contact lens into the post-lens tear film likely has prolonged contact time with the cornea; this can lead to increased diffusion into the eye. However, previous research on such therapeutic contact lenses has shown that it is challenging to either load the therapeutic amount of drug or extend the drug release or both, and hence no such contact lenses are currently available [19, 26, 82, 83]. In our case, the Dex-Lens utilizes a drug-polymer film that contains a substantial amount of the drug that is delivered over a prolonged period at therapeutically relevant levels.

The effect of dex drops and Dex-Lens on the prevention of corneal edema was notable. In the present study, placement of corneal sutures and temporary tarsorrhaphy induced corneal edema as measured by an increase in CCT (35.5 ±15.8%) that was compounded by placement of a vehicle contact lens (89.2±9.9%). This increase in CCT was found to be similar to those observed in other studies that induced corneal edema using contact lens and a tarsorrhaphy [84, 85]. When the placement of the corneal sutures and a temporary tarsorrhaphy was combined with either dex drops or Dex-Lens, there was significantly less corneal edema (17±1.4% and 4.1±2.6% respectively). In general, decreased oxygen transmission to the cornea results in corneal hypoxia leading to the upregulation of proangiogenic factors such as vascular endothelial growth factor (VEGF) [86]. An indicator of corneal hypoxia is corneal edema, where 8% or more increase in corneal thickness is concerning [55]. In this CNV study, several compounding factors could have caused corneal edema including CNV caused by the corneal sutures, the temporary tarsorrhaphy and the contact lens. CNV is a known factor associated with corneal edema [1]. The tarsorrhaphy partially closed the eyelids over the contact lens; this can further limit the oxygenation of the cornea and can also induce substantial edema in rabbit eyes [84, 85]. Commercial contact lenses have been used in rabbits to induce temporary corneal edema [87, 88] and the lenses may have caused additional corneal edema in this study. There is conflicting evidence in the literature as to the effect corticosteroids have on corneal edema and on the corneal endothelial cells (that maintain corneal transparency by pumping water out of the tissue) [33, 89–92]. Additional research is needed to better understand the effect of steroids on corneal hypoxia and corneal endothelial cell function.

One of the limitations of this study is that the models of inflammation used were not severe enough to allow us to determine if the Dex-Lens was more effective than dex drops, which were effective in preventing inflammation when given every hour. We chose a suture induced model of CNV because it is aligned with the clinical scenario of corneal suture placement during full thickness corneal transplants. In the uveitis model, the inflammatory response was not as robust as some other methods of inducing uveitis including injection of a tuberculosis antigen [93, 94] and it was not possible to differentiate between the efficacy of hourly dex drops and Dex-Lens treatments. In general, the model of uveitis that we used generates elevated protein levels within the aqueous humor without causing significant pain or distress to the animals thereafter [63–65]. Consistent with previous studies that used the same study design in rabbits, the clinical inflammatory response evident on slit lamp exam during this study was mild and transient for all the treatment groups [63–65]. For example, there were no significant slit lamp exam findings indicative of a robust inflammatory response (e.g. conjunctival redness or aqueous humor cell) in any of the treatment groups. There were no safety concerns, such as an increase in IOP or ocular surface toxicity, noted for any of the animals studied. We hypothesize that because the Dex-Lens provides significantly more drug to the aqueous humor than dex drops, the Dex-Lens may be able to reduce inflammation more than dex drops in uveitis models that are associated with more intraocular inflammation.

5. Conclusion

We have previously developed a topical dexamethasone delivery system capable of extended drug delivery. The Dex-Lens contains a therapeutically relevant amount of the drug and provided controlled release of dexamethasone for at least 7 days, and it is safe and effective in controlling the inflammation of corneal CNV and anterior uveitis. The Dex-Lens can improve patient compliance by avoiding intensive eye drop administration, and potentially provide a more convenient and superior treatment for inflammatory-mediated eye conditions.