Ex Vivo and In Vivo Animal Models for Mechanical and Chemical Injuries of Corneal Epithelium

Summary

Here, animal models based on mouse and rabbit are developed for mechanical and chemical injury of corneal epithelium to screen new therapeutics and the underlying mechanism.

Abstract

Corneal injury to the ocular surface, including chemical burn and trauma, may cause severe scarring, symblepharon, corneal limbal stem cells deficiency, and result in a large, persistent corneal epithelial defect. Epithelial defect with the following corneal opacity and peripheral neovascularization result in irreversible visual impairment and hinder future management, especially keratoplasty. Since the animal model can be used as an effective drug development platform, models of corneal injury to the mouse and alkali burn to rabbit corneal epithelium are developed here. New Zealand white rabbit is used in the alkali burn model. Different concentrations of sodium hydroxide can be applied onto the central circular area of the cornea for 30 s under intramuscular and topical anesthesia. After copious isotonic normal saline irrigation, residual loose corneal epithelium was removed with corneal burr deep down to the Bowman's layer within this circular area. Wound healing was documented by fluorescein staining under Cobalt blue light. C57BL/6 mice were used in the traumatic model of murine corneal epithelium. The murine central cornea was marked using a skin punch, 2 mm in diameter, and then debrided by a corneal rust ring remover with a 0.5 mm burr under a stereomicroscope. These models can be prospectively used to validate the therapeutic effect of eye drops or mixed agents such as stem cells, which potentially facilitate corneal epithelial regeneration. By observing corneal opacity, peripheral neovascularization, and conjunctival congestion with stereomicroscope and imaging software, therapeutic effects in these animal models can be monitored.

Introduction

The human cornea consists of five major layers and plays a pivotal role in ocular refraction to maintain visual acuity and structural integrity for protecting intraocular tissues¹. The outermost part of the cornea is the corneal epithelium, composed of five to six layers of cells that sequentially differentiate from the basal cells and move upward to shed from the ocular surface¹. Compared to the cornea in humans and New Zealand rabbits, mouse cornea has a similar corneal structure, but thinner periphery than the central part due to a reduced thickness in the epithelium and the stroma². Because of its unique position in the ocular optic system, many external insults such as mechanical injury, bacterial inoculation, and chemical agents may easily endanger epithelial integrity and further lead to vision-threatening epithelium defect, infectious keratitis, corneal melting, and even corneal perforation.

Although various therapeutic agents, such as lubricants, antibiotics, anti-inflammatory agents, auto-serum products, and amniotic membrane have already been used to improve re-epithelialization and reduce scarring, other potential treatment modalities that can enable wound healing, reduce inflammation, and suppress scar formation are still being developed and tested on different platforms. Various animal models for corneal epithelial wound healing have been proposed, including corneal epithelium removal with a corneal rust ring remover in diabetic mouse³, linear scratches over mouse corneal epithelium by a sterile 25 G needle for bacterial inoculation⁴, trephine-assisted removal of the corneal epithelium by corneal rust ring remover⁵, epithelial cautery over half of the cornea and limbus⁶, trephine-facilitated rabbit corneal abrasion by a dulled scalpel blade⁷, and bovine cornea injury by flash freezing in liquid nitrogen⁸.

Other than mechanical injury to the corneal epithelium, chemical agents are also common insults to the ocular surface, especially acidic and alkali agents. Sodium hydroxide (NaOH, 0.1-1 N for 30-60 s) is one of the commonly used chemicals in murine and rabbit models of corneal chemical burn^{9,10,11,12,13}. 100% ethanol had also been applied to the cornea in the rat chemical burn model, followed by additional mechanical scrapping using a surgical blade¹⁴. Since maintenance of a healthy ocular surface relies on functional units, including the eyelids, Meibomian glands, lacrimal system, the conjunctiva, and the cornea, in vivo animal models have some merits over ex vivo cultured cornea epithelial cells or corneal tissues. In this article, the mouse model of corneal abrasion wound, and the rabbit model of corneal alkali burn are demonstrated.

Protocol

All of the experimental procedures in animal studies were approved by the Research Ethics Committee at the Chang Gung Memorial Hospital and adhered to the ARVO statement for use of animals in ophthalmic and vision research.

1. Ex vivo wound healing model of the mouse corneal epithelium

1. Preparation of the mice
   1. Administer general anesthesia to C57BL/6 mice by intraperitoneal delivery of ketamine hydrochloride (80-100 mg/kg of body weight) and xylazine (5-10 mg/kg of body weight). 
   2. Ensure that general anesthesia is working by confirming loss of movement to a noxious stimulus and loss of righting reflex in mice.
   3. Fix the mice head by hand and apply topical anesthesia on both eyes with one drop of 0.5% proparacaine hydrochloride ophthalmic solution (Figure 1A). Disinfect the ocular surface and lids three times with 5% betadine. 
2. Creation of murine corneal epithelial wound model
   NOTE: Perform the below procedures under a stereomicroscope. The corneal wound was created in vivo, not ex vivo, to manipulate the eyeball better and close to the real-world situation. This is a terminal procedure; therefore, only clean instruments (not sterile technique) are required. 
   1. Mark the central cornea of the mouse using a skin biopsy punch (2 mm in diameter) to confirm a well-circumscribed and well-measurable area of the wound.
   2. Gently indent the punch over the central cornea to leave a circular mark (Figure 1B). Utilizing a hand-held corneal rust ring remover with a 0.5 mm burr, debride the corneal epithelium down to the Bowman's layer ensuring not to damage the latter (Figure 1C). Remove the residual, loose tissues inner to the wound margin with corneal forceps.
   3. Confirm the area of debridement with fluorescein staining (Figure 1D). To perform fluorescein staining, put a drop of normal saline onto a fluorescein paper to dissolve fluorescein, and then place the fluorescein-containing drop onto murine epithelial defect for visualizing it under Cobalt blue light.
3. Ex vivo culture of the murine corneal abrasion wound model.
   1. For harvesting the murine eyeballs, proceed as follows.
   2. Sacrifice the mice by cervical dislocation after inducing anesthesia with 5% isoflurane in an induction chamber. Ensure that anesthesia is working by confirming the loss of movement to a noxious stimulus and loss of righting reflex in mice.
   3. Gently press with the tip of forceps at the superior and inferior orbital rims to push the eyeball out. Introduce the tip of the closed corneal scissors into the retrobulbar space along the inferior orbital wall, ensuring not to penetrate the eyeball.
   4. Hold the eyeball steady with 0.3 mm corneal forceps, and then cut off the optic nerve and periorbital soft tissue with corneal scissors to isolate the eyeball.
   5. For ex vivo culturing of murine eyeballs, proceed as follows.
   6. Prepare a 48-well plate with melted wax inside the well and wait for solidification. With the tip of conjunctiva forceps, create a round hole on the surface of solidified wax for accommodating the eyeballs.
   7. Place the harvested eyeballs directly onto the 48-well plate (Figure 1E) with wax-covered bottoms and sidewalls to establish stabilization (Figure 1F).
   8. Culture the eyeballs with Dulbecco's modified eagle medium (DMEM) containing 1% fetal bovine serum (FBS) in a humidified atmosphere of 5% CO₂ at 37 °C with or without antibiotics, depending on the purpose of the study.
      NOTE: If the model is used for studying corneal epithelial wound healing, antibiotics would be required to prevent infection. However, if this model is used to evaluate the efficacy of antibiotics or mixed agents, prophylactic antibiotics would not be necessary.
   9. Immerse the ocular surface with the culture medium without causing the eyeball to float.
   10. Document the course of wound healing by fluorescein staining (step 1.2.3) and collecting photographs with a digital camera under Cobalt blue light.
       ​NOTE: In prospective experiments with mice models of mechanical corneal injury, those receiving corneal abrasion and tested further for the efficacy of therapeutic agents are viewed as an experimental group, and those receiving corneal abrasion without further treatment are regarded as a negative control group.

2. In vivo rabbit model of corneal alkali injury

NOTE: In this model, an alkali burn injury is induced followed by mechanical debridement of the corneal epithelium to generate a well-defined and even wound area for subsequent quantification. Sterilize all instruments before use.

1. Preparation of the rabbit with pre-operative analgesia, including intramuscular injection of systemic analgesics and topical eye drops.
   1. Administer general anesthesia to New Zealand white rabbits by intramuscular injection of ketamine hydrochloride (35-44 mg/kg of body weight), mixed with xylazine (5-10 mg/kg of body weight) at the hind leg.
   2. After positioning the rabbit and covering it with a towel, apply topical anesthesia over the right eye with a drop of 0.5% proparacaine hydrochloride ophthalmic solution (Figure 2A) under a stereomicroscope. Disinfect the ocular surface and lids three times with 5% betadine.
2. Inducing alkali burn injury over the cornea
   1. Place circular filter papers with a diameter of 8 mm (cut using an 8 mm punch) in a Petri dish. Using a dropper, add 0.5 N sodium hydroxide (NaOH) into the Petri dish to soak the filter papers. Drain excess NaOH solution from the filter papers before placing them onto the rabbit cornea.
      CAUTION: 0.5 N NaOH may cause severe erosive injury to human tissues. Wear gloves when handling. If the skin or the eyes come in contact with NaOH droplets, irrigation with copious amounts of normal saline and medical help are required to reduce further damage.
   2. After opening the eyelids with a lid speculum and confirming that the rabbit nictitating membrane is not interfering with the insertion of filter paper (Figure 2B), place the circular filter paper soaked in 0.5 N NaOH onto the central cornea for 30 s, and then remove it with forceps (Figure 2C).
   3. After removing the filter paper, rinse the ocular surface with 10 mL of normal saline to wash out alkali material.
3. Completing corneal epithelial defect
   1. Debride the corneal epithelium within the opacified area down to the Bowman's membrane using a corneal rust ring remover with a 0.5 mm burr (Figure 2D).
   2. Confirm the area of debridement with fluorescein staining under the Cobalt blue light and remove residual corneal epithelium using corneal forceps (Figure 2E).
4. Secure wound condition with tarsorrhaphy
   1. Confirm that the nictitating membrane smoothly covers the ocular surface and corneal epithelial defect at the nasal side. Ensure that the nictitating membrane is not folded or distorted too much to interfere with the process of wound healing and the experiment.
   2. Perform a temporary tarsorrhaphy with or without topical agents using a 6-0 suture to protect the ocular surface and prevent the rabbit from scratching it (Figure 2F). Ensure that the suture for tarsorrhaphy is at 3-4 mm from upper and lower lid margins with 4-5 ties and longer knots to prevent the rabbit from breaking the sutures.
      NOTE: If the experiment is not involved in an antibiotics study, topical agents with antibiotics could be considered.
   3. In this rabbit model, those receiving alkali burn and removal of corneal epithelium were regarded as the control group.
      NOTE: In prospective experiments, the rabbits receiving alkali burn corneal injury and further treated with therapeutic agents is viewed as an experimental group. The rabbits receiving alkali burn treatment only, without further treatment, are regarded as a negative control group.
5. Post-operative analgesia and pain control
   1. Assess the physiological condition and USDA pain levels for 7 days after the procedure, by monitoring pain and distress in the animals. Consider the use of tobramycin ointment and one drop of 0.5% proparacaine hydrochloride ophthalmic solution according to the result of the assessment. Administer Buprenorphine HCl (0.03 mg/kg) every 6-8 hours for 3 days.
      NOTE: For daily measurement of defect area and observation after surgery, the procedure belongs to USDA category D.

Representative Results

Ex vivo wound healing model of the mouse corneal epithelium:
After in vivo debridement of mouse corneal epithelium with hand-held corneal rust ring remover, a mildly depressed central corneal area with positive fluorescein stain can be found in the central 2 mm area (Figure 3A-B). After harvesting the mouse eyeball, it was easily fixed onto a wax-coated 48-well culture plate without significant rotating. Following the protocol, ex vivo culture of the murine eyeballs can be examined and documented daily within a 48-well culture plate under a stereomicroscope (Figure 3C). A day after debriding the murine corneal epithelium, one circular fluorescein-stained epithelial defect measured 2 mm in diameter can be revealed in digital photographs obtained under Cobalt blue light (Figure 3D). Initial irregularly stained wound margin or negative fluorescein staining means incomplete or failed removal of the corneal epithelium. In the normal process of wound healing, the corneal epithelial defect will heal with reduced fluorescein-stained area in 2-3 days.

In vivo rabbit model of corneal alkali injury:
Before any procedure, intact rabbit corneal epithelium cannot be stained with fluorescein staining. After creating alkali injury to the rabbit corneal epithelium, positive fluorescein staining can be observed with or without Cobalt blue light over the central cornea with a clear and complete circular margin (Figure 4A-B and Figure 5B). Incomplete stain with unfilled area represents residual corneal epithelial tissues or failed staining. During regular follow-up, the corneal epithelial wound re-epithelializes with ingrowth of pannus from the limbus, followed by reduced stained area (Figure 5C). The epithelial defect heals within 3-4 weeks. If corneal ulcer, dellen, large epithelial defect, or massive whitish or mucous discharge develops abruptly, insecure tarsorrhaphy, exposed sutures, a mispositioned nictitating membrane, or a foreign body within palpebral conjunctiva should be considered.

Figure 1: Procedures to set up a mouse model of corneal mechanical injury. (A) Topical anesthesia is applied before the procedure. (B) Gentle indentation is done over the central cornea with a 2 mm skin biopsy trephine. (C) Corneal rust ring remover is used to remove the central corneal epithelium. (D) An epithelial defect is stained with fluorescein to confirm the defect area and compare it with the region marked in B. (E,F) The mouse eyeball is harvested and transferred onto a 48-well plate covered with wax beforehand. Please click here to view a larger version of this figure.

Figure 2: Steps to build up a rabbit model of alkali corneal injury. (A) Topical anesthesia is applied to the ocular surface. (B) A lid speculum is used to open the upper and lower eyelids, without folding or squeezing the nictitating membrane. (C) A NaOH-soaked trephined filter paper (8 mm in diameter) is placed onto the central cornea. (D) Corneal rust ring remover is used to debride the 8 mm central epithelium down to the Bowman's layer. (E) The epithelium defect is stained with fluorescein. (F) After the procedure, tarsorrhaphy is performed to protect the wound from scratches. Please click here to view a larger version of this figure.

Figure 3: Positive and negative results in a mouse model of corneal mechanical injury. (A) Intact mouse corneal epithelium without any staining before the procedure. (B) In vivo positive staining with fluorescein on the murine corneal wound without Cobalt blue light. (C) Ex vivo culture of the mouse eyeball without adding fluorescein stain before adding culture medium. (D) Positive staining with fluorescein on corneal epithelial defect in ex vivo mouse model. The 2 mm epithelial defect generally heals within 2-3 days. Please click here to view a larger version of this figure.

Figure 4: Results of fluorescein staining in a rabbit model of corneal alkali injury. (A) Positive fluorescein staining under Cobalt blue light. The photograph was taken just after the mechanical corneal injury. (B) Positive staining with fluorescein dye could also be observed on rabbit ocular surface without Cobalt blue light. (C) Negative staining on the healed ocular surface. Please click here to view a larger version of this figure.

Figure 5: Time course of wound healing and appearance of re-epithelialization in a rabbit model of corneal alkali injury. (A) Re-epithelialization in mouse and rabbit models takes 2-3 days and 3-4 weeks, respectively. (B) An 8 mm epithelial defect stained with fluorescein after alkali burn in a rabbit model. Cobalt blue light was used as the light source. The photograph was taken just after the alkali injury. (C) A healed epithelial defect in the rabbit eye 3 weeks after alkali injury, showing reduced stain area. Please click here to view a larger version of this figure.

Discussion

Mouse and rabbit models of corneal injury provide a useful ex vivo and in vivo platform for monitoring wound healing, testing new therapeutics, and studying underlying mechanisms of wound healing and treatment pathways. Different animal models can be used for a short-term or long-term experiment, depending on the purpose of the research. For instance, after creating an epithelial defect on mouse cornea in vivo, a confined epithelial defect could be used to monitor liquid therapeutic agents in a small volume. At the same time, surrounding functional units, such as eyelids, lacrimal system, and the conjunctiva, can be evaluated under in vivo conditions, as opposed to cell culture or ex vivo culture conditions. Tarsorrhaphy may be still required in this situation if mice movements may affect the experimental condition. It is easier to observe and quantify a circular wound than a simple linear scratch wound⁴. However, skin punch to create a demarcation line on the cornea should be done carefully without cutting through the Bowman's membrane, which will otherwise leave a deep corneal injury or a penetrating wound. The mechanical corneal wound could also be created by an 8 mm corneal trephine and a scalpel blade in a rabbit model¹⁵, wherein a deeper wound down to anterior stroma rather than the corneal epithelium was presented.

Corneal rust ring removal is another pivotal issue that is worth mentioning. Since the mouse eyeball is small, over-removal or under-removal of the corneal epithelium can occur thus affecting the accuracy of the research. Marking cornea with a skin biopsy punch and fluorescein-guided operation will help reduce these mistakes. Although cautery and scalpel blades have been proposed as tools to remove corneal epithelium in animal models^6,7, the damage over the ocular surface may not be easily controlled and reproduced in the same way, which potentially leads to inconsistent results in further experiments.

Compared to the in vivo condition, ex vivo cultured mouse eyeballs in 48-well plates are easier to manipulate due to a larger working space on the plates and can be used to test complex agents in various culture mediums at the same time, such as drug-eluting contact lenses and cell therapies. When mouse eyeballs are being harvested and transferred onto a 48-well plate, meticulous protection of the cornea is important to avoid additional artificial damage to the ocular surface and rupture of the eyeballs. For the following study, the eyeball can be fixed within the paraffin-coated well with the cornea facing upward and immersed within a culture medium. Floating or rotating eyeballs or dehydrated cornea will hinder the results. Since this ex vivo mouse model focuses on changes over the ocular surface, other functional units, such as lacrimal gland and eyelids, are not discussed in this ex vivo model. Ex vivo mouse model also reduces the cost of breeding and housing mice and saves experimental space, compared to in vivo animal model. This model is suitable for a short-term study, rather than a long term one since potential tissue infection and organ failure may develop in the long run.

Although the mouse model costs less and can be scaled up in the laboratory, a small surface area potentially limits the observation of detailed changes of the cornea such as lipid deposition and neovascularization by stereomicroscopes. Instead, a rabbit model of corneal injury with a larger diameter of the eyeball can generally compensate for this disadvantage. By adjusting the concentration of NaOH and the soaking time, different extents of the severity of corneal alkali burn can be created. In rat chemical burn models, 1 N NaOH was used to soak the cornea with 3 mm filter paper for 40 s and 4 mm filter paper for 20 s, providing a similar but smaller area for observation^16,17. To keep the alkali burn consistent in size and concentration, a brand new and sharp punch is suggested in preparation of 8 mm filter paper to avoid any fibre or unfilled corner at the paper margin. Sufficient irrigation to wash out chemical agents from the ocular surface and conjunctival sac is required to reduce continuous damage outside the wound. Since rabbit nictitating membrane may interfere with experimental procedure and induce pain when corroded by alkali agents, it must be carefully protected and put back to physiological position after the procedure to reduce additional inflammation over the ocular surface. After the procedure of alkali burn, rabbit corneal wound can be protected by tarsorrhaphy or other material such as contact lens to secure quality and consistence of experiments.

Materials

Name	Company	Catalog Number	Comments
6/0 Ethicon vicryl suture	Ethicon	6/0VICRYL	tarsorrhaphy
Barraquer lid speculum	katena	K1-5355	15 mm
Barraquer needle holder	Katena	K6-3310	without lock
Barron Vacuum Punch 8.0 mm	katena	K20-2108	for cutting filter paper
C57BL/6 mice	National Laboratory Animal Center	RMRC11005	mouse strain
Castroviejo forceps 0.12 mm	katena	K5-2500
Corneal rust ring remover with 0.5 mm burr	Algerbrush IITM; Alger Equipment Co., Inc. Lago Vista, TX	CHI-675	for debridement of the corneal epithelium
Filter paper	Toyo Roshi Kaisha,Ltd.	1.11
Fluorescein sodum ophthalmic strips U.S.P	OPTITECH	OPTFL100	staining for corneal epithelial defect
Ketamine hydrochloride	Sigma-Aldrich	61763-23-3	intraperitoneal or intramuscular anesthetics
New Zealand White Rabbits	Livestock Research Institute, Council of Agriculture,Executive Yuan		Rabbit models
Normal saline	TAIWAN BIOTECH CO., LTD.	100-120-1101
Proparacaine	Alcon	ALC2UD09	topical anesthetics
Skin biopsy punch 2mm	STIEFEL	22650
Sodium chloride (NaOH)	Sigma-Aldrich	1310-73-2	a chemical agent for alkali burn
Stereomicroscope	Carl Zeiss Meditec, Dublin, CA	SV11	microscope for surgery
Westcott Tenotomy Scissors Medium	katena	K4-3004
Xylazine hydrochloride 23.32 mg/10 mL	Elanco animal health Korea Co., LTD.	047-956	intraperitoneal or intramuscular anesthetics