Neovascularization (NV) of the cornea can complicate multiple visual pathologies. Utilizing a controlled, alkali-burn injury model, a quantifiable level of corneal NV can be produced for mechanistic study of corneal NV and evaluation of potential therapies for neovascular disorders.
Under normal conditions, the cornea is avascular, and this transparency is essential for maintaining good visual acuity. Neovascularization (NV) of the cornea, which can be caused by trauma, keratoplasty or infectious disease, breaks down the so called ‘angiogenic privilege’ of the cornea and forms the basis of multiple visual pathologies that may even lead to blindness. Although there are several treatment options available, the fundamental medical need presented by corneal neovascular pathologies remains unmet. In order to develop safe, effective, and targeted therapies, a reliable model of corneal NV and pharmacological intervention is required. Here, we describe an alkali-burn injury corneal neovascularization model in the mouse. This protocol provides a method for the application of a controlled alkali-burn injury to the cornea, administration of a pharmacological compound of interest, and visualization of the result. This method could prove instrumental for studying the mechanisms and opportunities for intervention in corneal NV and other neovascular disorders.
Corneal blindness is the fourth most common cause of blindness, responsible for approximately 4% of all cases1. Corneal neovascularization (NV) plays a significant role in many of these pathologies, including herpetic keratitis (the leading infectious cause of blindness in the West) and trachoma (the leading cause of infectious blindness worldwide)2. Current therapies include steroids, nonsteroidal anti-inflammatory drugs (NSAIDs), anti-VEGF therapies, and cyclosporin A as well as conventional or laser surgical techniques3. However, the severely debilitating nature of corneal NV based pathologies, the paucity of surgical facilities capable of treating corneal NV, and the lack of a strongly performing pharmacological option led a recent expert roundtable to conclude that, despite the extant therapies, the fundamental medical need presented by these pathologies remains unmet4.
The human cornea consists of 5 layers, 3 cellular layers (epithelial, stromal and endothelium) and 2 interface (Bowman membrane and Descemet membrane). It functions as a mechanical barrier and refractive surface for the eye. Its transparent nature is the consequence of a delicate balance of its components and is integral to its proper function5. Normally avascular, the cornea receives blood from microvessels running along its outer edge which are fed from the ciliary and ophthalmic arteries. Corneal NV occurs when a stimulus promotes angiogenesis of these vessels allowing them to grow towards the center of the cornea and thus limit vision6. Corneal angiogenesis includes hemangiogenesis and lymphangiogenesis, which result in the ingrowth of blood vessels and lymphatic vessels from the limbal vascular arcade towards the center of the cornea. This leads to a breakdown of corneal "angiogenic privilege", an increase in corneal opacity and fibrosis, disruption of the corneal layers, and edema7. The precise triggers of corneal NV are numerous, ranging from a response to infectious disease such as trachoma to a chemically induced state caused traditional medicines, industrial chemicals, or even chemical warfare agents.
The molecular mechanisms of this process are not, as yet, fully characterized; however, a few key players have been identified. Under normal conditions the cornea possesses a unique ‘angiogenic privilege' maintained by a redundant array of anti-angiogenic factors (such as soluble VEGF-R1)8. However, in response to an external stimulus (such as an injury), there will be a local upregulation of pro-angiogenic factors (e.g. VEGF-A). This tips the balance of pro and anti- angiogenic factors that underlies the cornea's angiogenic privilege, and leads to hemangiogenesis, lymphangeogenesis, and inflammation, therefore causing corneal pathology and even blindness9.
Given the unmet medical need of this highly debilitating pathology, it is of interest to the field to have a reliable animal model of corneal NV. Here we present such a model: controlled alkali-burn injury. Various eye-injury models based on using filter paper rings have been used since 1970s10. In 1989, a group of Harvard Medical School ophthalmologists characterized a standard model of a central corneal alkali-burn injury in rabbit based on soaking a piece of circular filter paper with sodium hydroxide (NaOH) and applying it to the cornea at a specific range of concentrations11. Since then, this technique has been adapted for use in the mouse12-14. Recently, the Wang lab studied the therapeutic effects of the histone deacetylase (HDAC) inhibitor SAHA in the pathogenesis of corneal NV using a mouse corneal alkali-burn injury model15. The methodology of the mouse corneal alkali-burn injury model presented here was built mainly on the prior work of two other papers14,16.
Note: The following protocol and representative results use the HDAC inhibitor SAHA as an example compound. However, this protocol is by no means limited to the use of SAHA, and is recommended as a general method to test the effects of soluble compounds on corneal neovascularization. Minor modifications will need to be made for degree of dilution as well as frequency and duration of application. Additionally, compounds that are easily soluble in water will be able to be administered in the absence of DMSO.
Ethical Statement: All animal experiments should only be performed in compliance with national law and institutional regulations. This protocol was approved for use by Tulane University Institutional Animal Care and Use Committee.
1. Preparation of Materials (in Order of Use)
2. Alkali-Burn Injury & Compound Treatment
3. Clinical Assessment
4. Corneal Staining and Flat Mounts
After alkali-burn injury, corneal NV occurs in a predictable, time-dependent fashion. Figure 1 demonstrates the stark difference both in neovascularization and corneal opacity between an untreated animal (Figure 1A) and an animal treated with the HDAC inhibitor SAHA (Figure 1B) at the 7 day time point.
Figures 2A and 2B demonstrate a corneal flat mount of an untreated control eye with primary PECAM-1 and LYVE-1 staining and secondary Alexa Fluor 488 and 594 staining (respectively). Figures 2C and 2D show the same staining on an eye treated daily with the HDAC inhibitor SAHA, note the dramatic decrease in both hemangiogenesis and lymphangiogenesis.
Figures 3A and 3B provide a detailed look at the two stains. PECAM-1 serves a marker for the blood vessels (Figure 3A), while LYVE-1 binds specifically to the lymphatic vessels (Figure 3B). An overlay of the two fields is shown in Figure 3C, allowing comparison of hemangiogenesis vs. lymphangiogenesis as well as a comparison of the differing cell morphology.
Following PFA fixation (step 4.1), you can use conventional sectioning protocols (not detailed in the protocol above) to generate either frozen or paraffin embedded sections of the eye. While this does not allow the same level of quantification of invasion that a flat mount does, sagittal sections of the cornea can show you corneal thickness and relative depth of the tubes within the eye. Figures 3D-F shows sagittal, frozen sections of the cornea and either F4/80 (macrophage staining), DAPI (nuclear staining), or a merged image.
Figure 1. Progression of corneal neovascularization seven days after alkali-burn injury. (A) Representative image of an untreated eye. (B) Representative image of an eye treated three times per day with our compound of interest (the HDAC inhibitor SAHA). Note the difference in corneal opacity and neovascularization. Please click here to view a larger version of this figure.
Figure 2. Representative images of an untreated control (A and B) and a SAHA treated (C and D) cornea seven days after alkali-burn injury. (A and C) Wide field image of vascular endothelial cell staining with PECAM-1. (B and D) Wide field image of lymphatic endothelial cell staining with LYVE-1. Please click here to view a larger version of this figure.
Figure 3. (A) Vascular endothelial cell staining with PECAM-1. (B) Lymphatic endothelial cell staining with LYVE-1. (C) Merged PECAM-1/LYVE-1 staining. (D) F4/80 staining of macrophages in a sagittal frozen section. (E) DAPI staining of cell nucleus in a sagittal cut frozen section. (F) Merged F4/80 and DAPI staining. Please click here to view a larger version of this figure.
The protocol presented here results in reproducible levels of hemangiogenesis, lymphangiogenesis, and inflammation, making it an ideal system to study these three (interrelated) processes. While this method produces centralized corneal NV, several methods that have been developed to cause more directed NV, namely suturing of the cornea17 and implanted growth-factor expressing pellets18, might also be of interest. Our protocol is designed for use in the adult mouse, providing an easy to use animal model while also allowing a lab to take full advantage of a host of molecular and transgenic techniques not yet available for larger mammals. The above protocol and representative results detail the use of mice on a C57bl/6J background. Albino mice would also be suitable and may provide for easier imaging; however, a recent study indicates that the neovascular response of albino mice may not be as dramatic19. Further, unlike several other neovascularization models, corneal NV can be scored with examination through a surgical microscope or even by the naked eye. If a more rigorous quantification of the extent of neovascularization is required, digital images of the stained flat mount can be analyzed via a number of software packages, an example of how to do so with photoshop CS4 can be seen in Conner, et al. recent Nature Protocols paper "Quantification of oxygen-induced retinopathy in the mouse" 20.
Proparacaine hydrochloride is an amino ester analgesic applied as a topical solution (it is possible that other members of this family would be equally acceptable). Even though the animal is placed under general anesthesia for the procedure, we deem additional topical analgesia an ethical necessity to prevent undo pain to the cornea. It is imperative that the PBS used to dilute your compound and flush the eye be filtered and kept clean throughout the procedure (check for visible signs of contamination before each use). PBS is called for based on individual lab tradition; any equivalent balanced salt solution should achieve the desired result. Likewise, revisions of the immunohistochemistry protocol presented here may be called for if antibodies from other sources are used (i.e. the degree of dilution required).
The most technically challenging portions of this procedure are the initial placement of the NaOH soaked filter paper and the dissection of the cornea. We recommend that both techniques be practiced prior to the actual procedure. Filter paper must be placed in the center of the corneal in order to promote an equal level of neovascularization from all sides. Any degree of offset will create a high level of variability from mouse to mouse. Corneal dissection requires a good deal of manual dexterity. The eye is likely to deform in response to an attempt to perforate the pericorneal region. We recommend using a sharp, small gauge needle to make the initial cut and release the pressure inside. After an initial puncture is made, a pair of surgical scissors can be inserted into the hole and used to cut along an imaginary line separating the anterior of the eye from the posterior eye cup. Working slowly and carefully should yield an intact cornea.
It should be noted that while the primary purpose of this protocol is to assay the efficacy of various compounds in treating corneal alkali-burn injury it also has the potential to be used to study corneal wound re-epitheliaztion, corneal fibrosis, and limbal epithelial stem cell renewal. Furthermore, it serves as a general model to explore the mechanisms of pathological hemangiogenesis, lymphangiogenesis, and inflammation. The relative ease with which the protocol can be performed, as well as its noninvasive nature, presents an appealing opportunity for in vivo compound screening, efficacy tests, and characterization of transgenic mouse models with respect to pathological angiogenesis.
The authors have nothing to disclose.
We are grateful for Dr. Xinyu Li’s help in preparing the manuscript. S.W. was supported by a Startup fund from Tulane University, President’s Research Council New Investigator Award from UT Southwestern Medical Center, NIH Grant EY021862, a career development award from the Research to Prevent Blindness foundation, and a Bright Focus Award in Age Related Macular Degeneration Research.
1 mL Syringe | BD | 309659 | |
18 Guage Needle | BD | 305918 | |
10 mL Syringe | BD | 306575 | |
25 Guage Needle | BD | 305916 | |
Anti-F4/80 (rat anti-mouse) | AbD Serotech | MCA497RT | |
Anti-LYVE-1 (rabbit anti-mouse) | Abcam | ab14917 | |
Anti-PECAM-1 (rat anti-mouse) | BD | 553370 | |
Anti-IgG Alexa488 (goat anti-rat) | Invitrogen | A11006 | |
Anti-IgG Alexa594 (goat anti-rabbit) | Invitrogen | A11012 | |
Camera | Tucsen | TCC 5.0 ICE | |
Coverslips | Fisher | 12-548-B | |
DMSO | Sigma | D4540-1L | Caution: Mutagenic, Toxic |
Forceps (Blunt), Iris | WPI | 15915 | |
Forceps (Sharp), Dumont #4 | WPI | 500340 | |
KCl | Fisher | P217-500 | |
Ketamine Solution | MedVet | RXKETAMINE | Controlled substance, proper license required for use. |
Light Source for Microscope | AmScope | LED-14M-YA | |
Microscope (Stereo 7X-45X) | AmScope | SM-1B | |
Mounting Medium, Vectashield | Vector | H-1000 | |
NaCl | Fisher | S271-10 | |
NaH2PO4 | Fisher | S397-500 | |
NaOH | Fisher | S318-1 | Caution: Corrosive |
Paraformaldehyde | P6148-500G | Caution: Allergenic, Carcenogenic, Toxic | |
Proparacaine Hydrochloride | Sigma | P4554-1G | |
Scissors (5mm blade), Vanas | WPI | 14003 | |
Goat Serum | MPBio | 92939249 | |
Microscope Slides | Fisher | 12-550-15 | |
Triton X-100 | Sigma | T8787-100ML | |
Whatman Grade 1 Filter Paper | Whatman | 1001-6508 | |
Xylazine Solution | MedVet | RXANASED-20 |