The Rodent Model of Glaucoma and Its Implications

Abstract

Glaucoma is a group of progressive optic neuropathies, characterized by the degeneration of retinal ganglion cells related to the level of intraocular pressure and other factors. The exact pathogenesis of glaucoma is not known, and current therapeutic options are not sufficient to prevent or recover vision loss in glaucoma patients. Functional, repeatable, and easy-to-use animal models are therefore needed. Because of their inherent advantages, rodent animals, including mice and rats, have been widely developed as models to study various aspects of glaucoma and to evaluate possible novel therapies. However, no single model has been shown to emulate all aspects of glaucoma. In this review, we discuss currently available rodent animal models of glaucoma, their strengths and weaknesses, and the possible implications for current glaucoma research.

Glaucoma, a group of diseases characterized by progressive optic nerve degeneration resulting in visual field loss and irreversible blindness, is a leading cause of permanent vision loss worldwide. The death of retinal ganglion cells (RGCs) is a major element in the pathophysiology of these diseases, relating to the level of intraocular pressure (IOP).^1,2 However, because of the complex pathogenesis of glaucoma, its molecular etiology is not completely understood. Clinically, reduction of IOP is the only proven treatment method, but this approach cannot completely arrest the progress of the disease and offers no prospect of improved prognosis.¹

Animal models of glaucoma that simulate RGC and optic nerve loss in humans are of considerable importance in elucidating the disease's mechanisms and in evaluating novel therapies. To date, a wide range of animal models have been developed for glaucoma research, including monkeys, dogs, rabbits, rats, and mice; of these, the rodent model is the most widely developed and used. All of these animal models fall into 2 categories: IOP-dependent or IOP-independent. These animal models not only help to identify targets for therapeutic drugs but also improve our understanding of the causes and progression of glaucoma.^3,4 However, owing to the complexity of this disease, there is as yet no ideal model that can completely simulate human glaucoma. In this review, we will compare multiple rodent animal models of glaucoma and their current use in translational science, including the elucidation of molecular mechanisms of glaucoma, potential drug screening, and functional evaluation of the disease.

MOUSE MODELS

In general, mouse glaucoma models fall into 2 categories: spontaneous (most are transgenic mouse) and induced (nontransgenic) models.

Spontaneous Models

Many transgenic models have been created to emulate the development of ocular anterior segment abnormality or RGC loss that captures certain features of glaucoma. It has been found that several genes, including myocilin (MYOC/GLC1A), optineurin (OPTN/GLC1E), and WD repeat domain 36 (GLC1G) mutations, are involved in the pathogenesis of primary open-angle glaucoma (POAG).^2,5 Myocilin is expressed in many ocular tissues, including trabecular meshwork cells and the ciliary body, which are capable of secreting myocilin into the aqueous humor.⁶ Although the exact role of myocilin is unknown, it is suggested that mutant myocilin which accumulates in the intracellular space may be toxic to trabecular meshwork cells.² Approximately 4% of cases of adult-onset POAG and more than 10% of juvenile-onset cases are associated with MYOC mutations, and these cases feature elevated IOP.^7,8 It has been demonstrated that the Tyr437His mutation is associated with the relatively early onset of glaucoma in humans, and this mutation corresponds to the Tyr423His mutation in mouse myocilin.^2,9

Several transgenic mouse glaucoma models with the MYOC gene mutation have been created, and it has been demonstrated that overexpression or knockout of wild-type mouse MYOC in eye tissue did not induce IOP elevation or any signs of degeneration in the retina and optic nerve.^10,11 However, in transgenic mice, expression of the Tyr423His point mutation MYOC protein in iridocorneal angle tissue showed similar pathologic changes to human glaucoma. At 8 months of age, there was about 20% loss of RGCs in the peripheral retina and axonal degeneration in the optic nerve. However, starting at 12 to 18 months of age, this mouse model can develop moderate IOP elevation (∼2 mm Hg higher than the wild-type), indicating that the loss of RGCs may not be related to increasing IOP.⁹ Similar pathologic changes were observed in transgenic mice expressing the Tyr437His mutant of human myocilin protein, which displayed progressive age-related changes in RGC electrical responsiveness and was associated with marked astrogliosis and axonopathy without increasing IOP.^12,13

Recently, a new transgenic mouse model was developed, carrying a mutation of human MYOC (Tyr437His) under the control of the cytomegalovirus (CMV) promoter. These mice showed normal structure of the anterior segments but developed elevated IOP, starting at 3 months of age and lasting for 14 months. IOP exhibited diurnal rhythms and was highest at night (14.1 mm Hg in wild-type vs 20.3 mm Hg in transgenic mice). Furthermore, the increase in IOP was found to cause specific and progressive structural and functional loss of RGCs and degeneration of the optic nerve.¹⁴ Using this model, Sheffield et al^14,15 found that chronic and persistent endoplasmic reticulum (ER) stress was linked to POAG pathogenesis and that reduction of ER stress with phenylbutyric acid or sodium 4-phenylbutarate could prevent glaucoma phenotypes in the mouse model. All of these data suggest that expression of mutated MYOC is toxic to RGCs and the optic nerve and is necessary for the development of open-angle MYOC-related glaucoma. Moreover, MYOC is not required for the physiological regulation of IOP, and glaucoma is the result of a gain-of-function or novel property of mutant MYOC. The mutated MYOC within the trabecular meshwork not only accumulated in the cell cytoplasm but also prevented normal MYOC secretion into the extracellular space, which is related to the malfunction of aqueous outflow.

In another transgenic model created for open-angle glaucoma, Mabuchi et al¹⁶ demonstrated that transgenic mice with a targeted type I collagen mutation developed ocular hypertension after 16 weeks, which gradually increased until 36 weeks, with higher expression of collagen I in the sclera and associated structures, and progressive optic nerve axon loss. One possible mechanism is that the mutation in the gene for α1 subunit of collagen type I blocks MMP-1 hydrolysis, which reduces aqueous outflow facility.

Another important spontaneous secondary glaucoma model is the DBA/2J mouse, which is the model most commonly used to simulate chronic IOP elevation and age-related glaucoma. These mice developed iris defects with pigment dispersion into the anterior chamber and accumulation of pigment granules in the trabecular meshwork, blocking the aqueous outflow pathway and leading to IOP elevation.^17,18 Using this model, numerous studies have explored the pathogenesis of glaucoma and evaluated novel therapies.^19–22 However, these mice exhibit a high degree of individual variability and asymmetry in disease development, and the difficult and lengthy process of introducing genetic mutations or applying genetic technologies to this mouse line further limits their application.²³

Normal tension glaucoma is another type with glaucomatous optic neuropathy independent of IOP, but in such patients, the optic nerve may be more vulnerable to normal range IOP.²⁴ Based on the theory that excessive glutamatergic system activity, especially of the N-methyl-d-aspartate (NMDA) subtypes, is toxic and contributes to the death of RGCs in glaucoma, and that glutamate transporter (glutamate transporter 1, excitatory amino acid carrier 1, glutamate aspartate transporter) is the only mechanism for removing glutamate from extracellular fluid in the retina,²⁵ Harada et al^26,27 found that glutamate aspartate transporter and excitatory amino acid carrier 1 knockout mice showed spontaneously occurring RGC death and typical glaucomatous damage of the optic nerve without elevated IOP, which is similar to normal tension glaucoma,^26,27 and that brimonidine could prevent neurodegeneration by stimulating the production of several neurotrophic factors including brain derived neurotrophic factor (BDNF) and basic fibroblast growth factor (bFGF) in this model.²⁸

An interesting mouse model simulating acute primary angle-closure glaucoma has been developed, in which overexpression of calcitonin-like receptors in mice enhances signaling of smooth muscle-relaxing peptide adrenomedullin in the pupillary sphincter muscle, resulting in pupillary palsy, with abrupt transient rises in IOP up to a mean level of 48.2 ± 7.3 mm Hg between 30 and 70 days of age. In these mice, decreasing endogenous adrenomedullin could prevent the elevation of IOP; this model suggests that adrenomedullin levels and its receptor in the iris sphincter may provide a novel target for the treatment of angle-closure glaucoma.²⁹

It is known that current first-line IOP reduction drugs (topical prostaglandin analogs) are ineffective in mice lacking the prostaglandin F receptor.³⁰ This transgenic model is useful for clarifying the molecular basis of ocular response to IOP-lowering medications. With our improving understanding of the genetic features of glaucoma, increasingly functional, repeatable, and accurate spontaneous mouse (transgenic mouse) models can be developed.

Induced Mouse Model

The course of genetic glaucoma models is often quite long (in the order of several months) and generally not amenable to experimental manipulation. Induced mouse glaucoma models are shorter term and offer greater control over the extent of pathology, potentially offering an ideal model for advancing the understanding of the pathogenesis of glaucoma and for screening novel therapies.

Laser photocoagulation of the trabecular meshwork, either alone or in concert with episcleral veins, can elevate IOP and cause dependent RGC and optic nerve degeneration.^31,32 Recently, Yun et al³³ used laser photocoagulation on the trabecular meshwork, which caused immediate IOP elevation 1 day after laser application lasting for 24 weeks, inducing RGC loss and optic nerve axon loss. One important finding in this laser-induced model was that the induced ocular hypertension not only caused the loss of RGCs, but also damaged the normal function of surviving RGCs with the alteration of retrograde axonal transport,^34,35 so these models can be used for finding early intervention to prevent the progression of the disease before the loss of RGCs.

However, all of these laser-induced models exhibit inflammatory infiltration in the anterior chamber³³ and are also limited by the absence of a repeatable internal control, in which an equivalent insult to the fellow eye does not result in extended elevated IOP. These models showed large interanimal variability, and there remains a need for a stable and accurate inducible glaucoma model.

The “microbead glaucoma model” may be an excellent inducible glaucoma model for inducing stable high IOP.^36–38 Chen et al^39,40 injected 2-μL polystyrene microbeads with a glass micropipette connected to a Hamilton syringe into the anterior chamber of C57BL/6J mice; 81 of 87 mice receiving the injection exhibited consistent IOP elevation, and the control eyes, which received an equal volume of saline, showed rare IOP elevation. A single injection of small microbeads (10 μL) resulted in a 4-week elevation of IOP, extending to another 4 weeks after a second injection in week 4, with approximately 50% loss of RGC bodies and axons after an 8-week elevation of IOP.^39,40 These models showed IOP elevation to be highly consistent both in duration and magnitude, with interanimal variability of approximately 5% of the mean, representing an excellent model for the assessment of IOP-induced pathology.

With the extensive use of glucocorticoids in clinics for the treatment of various ocular diseases, glucocorticoid-related ocular hypertension has attracted much attention. Recently, Zode et al⁴¹ developed a glucocorticoid-induced glaucoma; using topical dexamethasone phosphate eye drops 3 times per day for 6 weeks, IOP began to increase at 2 weeks, and withdrawal of the glucocorticoid led to a reduction in IOP levels, which resembles glucocorticoid-induced glaucoma in human patients. In so doing, they were able to demonstrate that chronic ER stress of the trabecular meshwork contributes to glucocorticoid-induced ocular hypertension.

Advantages and Disadvantages of Mouse Models

Advantages

The mouse eye has a similar anterior segment organization to humans, with well-developed ciliary processes behind the iris, Schlemm canal, and ciliary muscle, as well as a similar uveoscleral outflow pathway; aqueous humor production and aqueous humor turnover rate are also similar to the human eye, but with a greater proportion of uveoscleral outflow.⁴² The mouse eye also has a vascularized retina and an unmyelinated retinal nerve fiber layer, and its retinal morphology is similar to the human eye, comprising 10 layers. In addition, there are high degrees of conservation between the mouse and human genomes, enabling genetic manipulation by altering the mouse genome, with the ability to breed the animals as desired. The mouse model is also a relatively quick and inexpensive form of experiment to use.

Disadvantages

The mouse eye has no macula, a reduced number of identifiable types of RGCs, and no lamina cribrosa in the optic nerve, and the eye globe is much smaller than in humans, making it difficult to access and manipulate. The small size of the mouse eye has made both induction and measurement of IOP elevation more challenging. It is also important to note that no mouse model can perfectly emulate the pathophysiology of human glaucoma, with its relatively different physiological and biological processes.

RAT MODELS

Only rarely can transgenic rat models spontaneously develop ocular hypertension or normal IOP glaucoma; however, it was found that the P23H-1 rat strain which carries a rhodopsin mutation has progressive RGC loss and RGC axonal compression as the rat ages.⁴³ Most rat glaucoma models are experimentally induced. Models of experimentally elevated IOP in rats provide valuable opportunities to discover and study mechanisms of pressure-induced optic nerve damage. Furthermore, these IOP-dependent models cause RGC dysfunction before cell loss, providing an opportunity for early intervention.⁴⁴

Laser-induced Ocular Hypertension

Reducing aqueous humor outflow by laser photocoagulation to the tissue involved in drainage, including the trabecular meshwork and limbal and episcleral blood vessels, causes IOP elevation. Ueda et al⁴⁵ first applied laser to the trabecular meshwork to create localized scarring, causing a high IOP greater than 20 mm Hg over 12 weeks with weekly laser treatment and resulting in glaucomatous optic neuropathy. Subsequently, modified laser photocoagulation of the trabecular meshwork, either alone or in concert with episcleral veins, also elevated IOP in rats; this could last for 3 weeks, accompanied by RGC loss and axonal degeneration after only a single laser treatment.^46,47 Recent research found that laser-induced ocular hypertension in adult rats not only caused the loss of RGCs, but also resulted in severe damage to the outermost retinal layers.⁴⁸ These laser-induced models are a good way to study neuroprotection in glaucoma. It has been demonstrated that brimonidine promotes neuroprotective activity unrelated to its effect on ocular hypotension in the chronic laser-induced ocular hypertension model, which decreased the expression of glial fibrillary acidic portein (GFAP).⁴⁷ However, the laser-induced rat model needs specialized equipment, and laser energy uptake varies when applied to different rat species, requiring modification of laser parameters when used in different rats.⁴⁹ The neuroprotective role of retinal stem cells has been demonstrated using this model.⁵⁰

Episcleral Vein Injection of Hypertonic Saline

Sclerosis of the trabecular meshwork by injecting a hypertonic saline solution into the episcleral veins can also induce ocular hypertension in rats.⁵¹ This method can induce ocular hypertension for a relatively longer time, sometimes persisting for as long as 200 days, and can also produce progressive RGC loss and optic nerve degeneration.^52,53 Using this model, Tehrani et al⁵⁴ found that optic nerve head astrocyte process reorientation occurs early in response to elevated IOP, and an inflammatory response of astrocytes was demonstrated, providing support for the exploration of astrocyte-based treatments for glaucoma.⁵⁵ However, IOP varies dramatically in magnitude between individual animals in this model, and repeat injections are required when IOP fails to increase after the first treatment. Additionally, the insertion of a microneedle into the episcleral vein is difficult and requires considerable practice.

Episcleral Vein Cauterization

Cauterizing the episcleral vein to interfere with the drainage of aqueous humor from the collecting ducts of Schlemm canal is another commonly used method for causing IOP elevation; the amplitude of IOP increase depends on the number of episcleral veins cauterized.⁵⁶ One study found that episcleral vein cauterization inducing ocular hypertension exhibited a great increase in IOP when compared with hypertonic saline episcleral vein sclerosis⁵³; this model induced RGC loss throughout the periphery of the retina, which differs from other rat models showing axonal degeneration in the superior region of the optic nerve.⁵⁷ Many neuroprotective drugs have been tested using this model.^58,59

Microbead Injection Model

Ocular hypertension in rats can also be induced by injecting polystyrene microbeads into the anterior chamber to block the trabecular meshwork. When compared to the microbead occlusion model in mice, the rat model requires a greater volume of microbeads to be injected, and IOP elevation can also last for 8 weeks with an additional injection.⁴⁰ Recently, Smedowski et al⁶⁰ modified the method for this “bead glaucoma model” by rapid injection of polystyrene microbeads of 2 different sizes in a viscoelastic suspension into the anterior chamber. This method achieved an abrupt elevation of IOP to 55 mm Hg, and a decline to 30.9 ± 3.2 mm Hg that could last for 6 weeks, accompanied by loss of RGC and optic nerve axons. This model simulated a severe glaucoma attack and achieved longer-lasting IOP elevation with chronic damage of RGCs.

Acute IOP Elevation Model

Sellés-Navarro et al⁶¹ first demonstrated the effect of an acute increase of IOP on RGC. They concluded that the duration of the initial transient period of hypertension and the duration of the survival period influence the degree of RGC death, which provided clues for early intervention for acute IOP elevation. To better and simply simulate the acute increase of IOP in acute angle-closure glaucoma, a rat model was developed with acute IOP elevation. The anterior chamber of the rat eye was cannulated with a 30-gauge infusion needle connected to a normal saline reservoir, which was elevated to maintain a high IOP for 60 minutes. This model also included some mechanisms of retinal ischemia-reperfusion.⁶² Using this model, 1 study demonstrated a critical role for caspase-8 in transducing toll like receptor 4-mediated IL-1β production and RGC death in acute glaucoma, suggesting a new strategy for the treatment of acute glaucoma.⁶³ However, this model can only be used to explore the effect of an acute increase of IOP on ocular structure.

Advantages and Disadvantages of Rat Models

Advantages

The predominance of rats as glaucoma models is based on the structure and vasculature of the rat optic nerve head, which is similar to that in humans. The ultrastructural relationship between astrocytes, RGC axons, and the connective tissue of the optic nerve head also appear quite similar to the primate form and exhibit high potential for revealing cellular mechanisms of axonal injury.⁴⁴ Furthermore, the rat eye is larger than the mouse eye and relatively easy to manipulate, which largely accounts for the wider use of rats in the induced ocular hypertension model involving alteration of aqueous fluid dynamics in the anterior segment.

Disadvantages

The lamina cribrosa is relatively sparse in the rat, and in addition, there is a relatively thin sclera. These anatomical characteristics probably contribute to our impression that the rat's optic nerve is more sensitive to chronic IOP elevation.⁶⁴ The central retinal artery and vein of the rat lie beneath the neural tissue, which may contribute to the observation of early damage from elevated IOP in rats, as it appears first in the superior regions of the optic nerve and retina.^44,65 Moreover, all these inducible rat glaucoma models show large interanimal variability, necessitating relatively large numbers of experimental cohorts, and such models inevitably present complications, such as inflammation in the anterior chamber.

MODELS FOR BOTH MOUSE AND RAT (INDEPENDENT OF INTRAOCULAR PRESSURE)

Some other models have been developed for both mice and rats, using the same methods to cause RGC loss without increasing IOP. These include the optic nerve crush model,^66–68 the retinal ischemia-reperfusion model,^69–72 and the NMDA-induced retinal damage model.^73–75 Although these models have also been used to explore the pathogenesis of glaucoma and to screen neuroprotective drugs for glaucoma, they all simulate only 1 aspect of glaucoma.

INTRAOCULAR PRESSURE MEASUREMENT OF RODENT ANIMALS

Because most rodent glaucoma models are IOP-dependent, it is critical to accurately measure IOP in mice and rats.^35,76 There are a few feasible methods which have been developed for the measurement of rodent IOP. The direct method is through inserting a glass-tipped micropipette which is connected to a pressure transducer into the anterior chamber. This method requires anesthesia and cannot be repeated within the same day.⁷⁷ With this method, Savinova et al⁷⁸ compared the IOP in 30 genetically distinct mice and found that there was a broad range of IOP from approximately 10 to 20 mm Hg. Age, time of day, obesity, and diabetes all had some effect on mouse IOP, which was similar to in humans. The noninvasive rebound tonometer is an indirect method, which can obtain repeated IOP measurements, but is less accurate when compared with direct methods.⁷⁹ The modified Goldmann tonometer can be used without anesthesia to obtain repeated IOP measurements, but it has greater measurement variation, and a custom prism is needed.⁸⁰ The most widely used methods are the commercially available TonoPen and Tonolab, which are convenient, repeatable, and noninvasive; however, they need mutiple readings to get an average value.⁸¹

Lindsey and Weinreb⁷⁶ reviewed articles with IOP measurement by different methods, and they found that the normal IOP of mice varies by mouse strain type, methods of anesthesia, and methods of IOP measurement. When using different measurement methods, there should be some modifications for mice and rats.⁶⁴ Although it is difficult to obtain valid IOP values in rodent animals, what we can do is use average values with repeated readings and always measure IOP under the same conditions (the same time, the same anesthetic drugs if necessary, and with the same measurement methods), setting the baseline for comparison.

IMPROVING RODENT ANIMAL MODELS OF GLAUCOMA

An accurate glaucoma animal model should recapitulate the morphological, biochemical, and pathophysiological changes of human glaucoma; it should have similar underlying biological mechanisms; and it should respond comparably to therapeutic agents. Rodent species offer greater advantages as model systems for creating ocular hypertension or IOP-independent glaucomatous optic neuropathy. However, there is still much that can be done to improve rodent animal models of glaucoma. First, there is a need for accurate, repeatable, nonanesthetic, and noninvasive IOP measurement methods with real-time monitoring. Second, better transgenic mouse models can be developed with the advance and maturation of technologies, such as the Cre-Lox recombination system and the CRISPR/Cas9 system, which can either increase or eliminate transcription of particular genes in particular tissues at particular times and evaluate altered gene regulation due to targeted mutations within regulatory components of gene promoters. Finally, with a better understanding of the pathogenesis of glaucoma, improving clinical technologies will help in the diagnosis and clinical classification of glaucoma, which will then feed back into the animal models.

CONCLUSIONS

Glaucoma is a complex disease with a complicated pathophysiology that is far from being completely understood. Current therapy is insufficient to prevent the progress of the disease, and animal models are of prime importance to help in elucidating the pathophysiology of glaucoma and in developing novel interventions to halt or reverse its progression. Rodent models offer many advantages and have been extensively developed for the study of glaucoma. However, no single model has been shown to recapitulate all aspects of glaucoma, and each rodent model has particular advantages and disadvantages that make it suited to answering specific research questions. With new developments in biotechnology, basic science research tools, and better understanding of the pathogenesis of glaucoma, more and better animal glaucoma models will be developed, helping to better treat the disease.