Discovery of molecular therapeutics for glaucoma: Challenges, successes, and promising directions

Abstract

Glaucoma, a heterogeneous ocular disorder affecting ~60 million people worldwide, is characterized by painless neurodegeneration of retinal ganglion cells (RGCs), resulting in irreversible vision loss. Available therapies, which decrease the common causal risk factor of elevated intraocular pressure, delay, but cannot prevent, RGC death and blindness. Notably, it is changes in the anterior segment of the eye, particularly in the drainage of aqueous humor fluid, which are believed to bring about changes in pressure. Thus, it is primarily this region whose properties are manipulated in current and emerging therapies for glaucoma. Here, we focus on the challenges associated with developing treatments, review the available experimental methods to evaluate the therapeutic potential of new drugs, describe the development and evaluation of emerging Rho-kinase inhibitors and adenosine receptor ligands that offer the potential to improve aqueous humor outflow and protect RGCs simultaneously, and present new targets and approaches on the horizon.

Graphical Abstract

INTRODUCTION

Glaucoma: a brief background

Glaucoma, a group of neurodegenerative diseases of the optic nerve in the posterior eye, with varying etiologies, is a leading cause of irreversible blindness^{1, 2} and affects more than 60 million people worldwide.³ Glaucoma can be divided into two major subtypes, based on whether the iridocorneal angle (Figure 1) is closed (closed-angle glaucoma (CAG)), namely, in which the iris physically comes in contact with the trabecular meshwork (TM), or open (open-angle glaucoma (OAG)). Primary open-angle glaucoma (POAG), the most common form overall, refers to presentation of glaucoma symptoms without attribution to other disease, injury, or closed iridocorneal angle.⁴ Secondary forms of glaucoma are those attributable to other causes; for example, secondary open angle glaucoma can be a result of pseudoexfoliation (PEX) syndrome⁵ or prolonged use of steroids.⁶ Glaucoma can be further divided based on age of onset into congenital, juvenile (JOAG, for those with onset at less than 35 years of age), and the more common, adult onset.

Figure 1. Overview of the anterior segment of the eye.

Key anatomical structures discussed in this Perspective are labeled. The conventional pathway for aqueous humor outflow is highlighted with blue arrows.

Patients with glaucoma are typically asymptomatic until vision loss, which is caused by damage to the nerve fiber layer surrounding retinal ganglion cells (RGCs) followed by RGC death.⁷ A clinically described feature referred to as ‘cupping’ differentiates glaucoma from other optic neuropathies involving RGC death.⁸ Namely, a change is apparent in the bundled optic nerve fibers at the center of the optic disc (i.e. the ‘cup’); the extent of cupping is used to monitor disease progression in high-risk patients before any changes in visual field are obvious.¹ A more readily monitored causal risk factor is elevated intraocular pressure (IOP), but it is not a necessary precondition for RGC death and vision loss in glaucoma, as in the case of a third glaucoma subtype, normal tension glaucoma (NTG).^{2, 9} As discussed in this Perspective, both for patients exhibiting elevated pressure as well as those with NTG, the standard of care is to slow loss of visual field by reduction of IOP.^9–11 This approach has proven useful in reducing the loss of visual field over time, but IOP reduction alone does not halt disease progression.¹² As described below, IOP-reducing methods currently used in the clinic, as well as emerging therapies, are not developed de novo to target specific molecular underpinnings of glaucoma but rather are derived from other similar disorders, such as those of the cardiovascular system.

Efforts to improve our understanding of glaucoma pathogenesis, which remains unclear due to inherent disease heterogeneity and late age onset, involve identifying genetic susceptibility. The first, best-studied, and to date most strongly glaucoma-associated gene encodes for myocilin.¹³ Initially named the trabecular meshwork inducible glucocorticoid response (TIGR) gene due to its link with steroid-induced secondary glaucoma,¹⁴ numerous mutations in myocilin with autosomal dominant Mendelian inheritance patterns have since been linked with POAG and JOAG, in nearly every population and country throughout the world.^{15, 16} Other associated genes include those for (OPTN) and WDR36, which are seen in both inherited and sporadic forms of adult onset OAG.^15–17 The Glu50Lys variant of OPTN is also associated with NTG.^{18, 19} Together, the known genetic variants of MYOC, OPTN, and WDR36 account for about 10% of glaucoma cases,^{15, 16} but relatives of all POAG patients are 20% more likely to develop glaucoma in their lifetime than controls, pointing to a wider genetic link.²⁰ Genome-wide association studies (GWAS), particularly the worldwide consortium with the acronym NEIGHBOR,²¹ continue to report susceptibility loci for glaucoma.^{22, 23} For example, a gene product that has received some attention among the over 20 genes identified by GWAS is CYP1B1, associated with primary congenital glaucoma as well as POAG.¹⁷ With the exception of myocilin discussed below, however, how these proteins contribute to pathogenesis remains yet unknown.

Relevant anatomical changes with age/glaucoma

Aqueous humor (AH), the fluid found in the anterior segment, supplies nutrients to nonvascular eye tissues such as the lens and cornea. The balance between AH production and outflow is intimately tied to IOP maintenance.²⁴ AH is produced in the non-pigmented epithelial (NPE) cells of the ciliary body, then flows to the anterior segment, and eventually drains through the (1) so-called conventional pathway via the trabecular meshwork (TM) to Schlemm’s canal (SC) and finally into the episcleral veins (Figure 1), or via (2) the uveoscleral pathway through the ciliary body and iris root, which bypasses the TM and SC.²⁴ Active transport of chloride and sodium ions from the stroma to the AH, across both the pigmented epithelium and NPE of the ciliary body, is responsible for the majority of AH formation.²⁵ Control of ion movement across the ciliary epithelium is regulated at least in part by cyclic AMP (cAMP) modulation. Current and emerging therapeutic targets include those in signaling cascades that ultimately affect cAMP levels (see below), but studies to date do not yet converge on whether increased levels of cAMP increase²⁶ or decrease²⁷ AH formation.

AH outflow facility is altered in both normal aging and glaucoma.²⁸ In the healthy aging eye, IOP is maintained as both inflow and uveoscleral outflow tend to decrease; conventional outflow facility is unchanged.²⁹ By comparison, glaucoma-associated IOP increases are believed to be a result of reduced outflow predominantly through the conventional pathway, not increased AH formation or decreased outflow through the uveoscleral pathway.²⁸ Documented anatomical alterations include increased extracellular matrix (ECM) stiffness in the TM,³⁰ increased cell stiffness of the SC, decreased SC epithelial cell pore formation, a loss of TM cells,³¹ and increased sheath-derived plaque formation in the TM and SC.³² Nevertheless, the extent to which these are primary culprits for pathogenesis is complicated by the fact that changes to the TM and SC due to glaucoma versus the normal aging process are difficult to distinguish.³³

Model systems used to study glaucoma

A major impediment to developing new glaucoma therapeutics is the intrinsic heterogeneity of the disease. There is no established model system e.g. tissue or organ culture, animal model, or human postmortem tissue, to study the disease, develop, or systematically evaluate potential new treatments. Cell culture and organ models offer a controlled environment and are useful for initial screening of new drugs, but are derived from human postmortem tissues that vary widely³⁴. Often the medical history of a donor eye is not well documented and the eyes may be affected by undiagnosed disease, as well as long-term drug treatments for glaucoma and/or other disorders.

The use of cultures lacks a robust clinical translation such as that seen with use of animal models.³⁴ Yet, animal eyes, including those of birds, rodents, rabbits, dogs, cats and primates, vary greatly in their anatomical similarity to those of humans. Glaucoma in animal models may be induced by various methods including genetic mutation, treatment with steroids or photocoagulation, or in some species, occurs spontaneously.³⁵ Induced glaucoma offers quick, but non-natural, disease progression, while spontaneous glaucoma may take years for symptoms to appear.³⁵ Rat and mouse glaucoma models are the most commonly used animals due to cost and access, but their eyes are not as similar to humans as other animal models and their small size can make testing difficult.^{34, 35} Primate eyes are the most similar to those of humans, but such animals are limited in availability and expensive to study.^{34, 36}

Standard experiments used to evaluate molecules of therapeutic potential for glaucoma

Typical drug discovery methods, namely high throughput screening of chemical libraries using biochemical or cell-based assays, cannot be used to detect reduction in IOP or changes in outflow resistance. Rather, IOP changes are typically assessed directly in animal models using rebound tonometry, a non-invasive method. Rebound tonometry uses a small probe that is rebounded off of the cornea, where the inverse of the deceleration speed of the probe correlates well with IOP.^{37, 38} Several measurements can be taken back-to-back or on different points of the cornea without harm. However, high variance is inherent in this measurement, and can be greater than the reductions in IOP afforded by treatment, thus complicating data interpretation when limited numbers of specimens or animals can be investigated.³⁹ To assess outflow facility, perfusion of mounted post-mortem eye tissues is commonly employed. In this experiment, solutions containing small molecules of interest are perfused through the sample, and changes in eye tissue with increased perfusion pressures are monitored.⁴⁰ Perfusion experiments are limited by availability of specimens, and variations in handling of these ex vivo samples will affect measurements.

TREATMENTS CURRENTLY USED IN THE CLINIC

Early detection is critical for preventing vision loss, as any loss of visual field due to retina degeneration is not reversible, but diagnosis is complicated by the fact that the age of onset and rate of glaucoma pathogenesis varies for different forms of the disease and varies widely among patients.⁴¹ Early signs of the disease can be missed if only IOP or variations in visual field are monitored, so ophthalmologists typically also look for changes or damage to the optic disk.¹ Once diagnosed, options to treat glaucoma fall into several classes all of which target lowering IOP as a method to slow disease progression.¹⁰ IOP is modulated by decreasing AH production or by increasing outflow. Even though drugs are generally applied topically to the eye, the five classes of current therapies (Table 1A) are not selective to eye tissues and were originally developed for other diseases. In addition, there are surgical options (Table 1B). Below we briefly summarize the five classes of currently marketed drugs and surgery used; for comprehensive descriptions and other clinical considerations, we refer the reader to a number of excellent review articles.^42–45

Table 1.

Current glaucoma therapies used in the clinic

A. IOP Lowering Drugs
Class	Example	Method of IOP Reduction
Cholinomimetics (Cholinergic Agonists)	Pilocarpine	Increase AH outflow
Adrenergic receptor agonists (sympathomimetics)	Brimonidine	Reduce AH production, increase outflow
β-Adrenergic receptor antagonists (β-Blockers)	Timolol	Reduce AH production
Carbonic Anhydrase inhibitors	Dorzolamide	Reduce AH production
Prostaglandin analogues	Latanoprost	Increase uveoslceral outflow
B. Surgery
CAG
Type	Purpose
Laser peripheral iridotomy (LPI)	Widen iridocorneal angle
Iridoplasty	Widen iridocorneal angle
Argon laser peripheral iridoplasty (ALPI)	Contraction burns to create space between the anterior iris surface and the TM
Phacoemulsification cataract extraction (phaco/IOL)	Removal of lens alleviates pressure
POAG
Type	Purpose
Glaucoma filter surgery/trabeculectomy	Partial removal of drainage angle allowing AH to bypass TM and alleviate pressure
Trabeculectomy with Shunt/Stent	Trabeculectomy with addition of shunt or stent to alleviate pressure
Trabeculoplasty	Lasers open drainage channels to alleviate pressure

Cholinomimetics

The first class of drugs used to treat glaucoma, cholinergic agonists (also known as cholinomimetics, miotics, parasympathomimetics, acetylcholine receptor agonists), have been in use as glaucoma therapeutics since the 1870s.⁴⁶ Cholinergic agonists increase AH outflow through the conventional pathway by causing contraction of the ciliary body. Cholinergic agonists are divided into two classes, those that act directly by binding to muscarinic acetylcholine receptors in the ciliary body, or indirectly by inhibiting acetylcholinesterase.^{1, 47} Physostigmine, the first cholinergic agonist, was used during eye surgery to constrict the pupil for over a decade before its IOP reducing effects were realized.⁴⁶ Pilocarpine, the most common⁴⁴ and oldest⁴⁶ cholinergic agonist glaucoma drug, is a nonselective muscarinic acetylcholine receptor inhibitor.⁴⁴ A third molecule, carbachol, works by both direct and indirect inhibition.⁴⁸

Adrenergic receptor agonists (sympathomimetics) and antagonists (β-blockers)

Adrenergic receptors are G-protein coupled receptors (GPCRs) that regulate the sympathetic nervous system. Epinephrine, the natural agonist of adrenergic receptors, was used as a glaucoma therapy in the 1950s.⁴⁶ Currently, the most commonly prescribed adrenergic receptor agonists are selectively targeted to the α2 subclass (Table 1A),⁴⁹ and are sometimes coupled with a β-adrenergic receptor antagonist (next paragraph), which together are particularly effective for patients who do not tolerate prostaglandin analogs (see below). The α2-adrenergic receptors are involved in a cascade that maintains vascular constriction by activating G_i proteins, which inhibit adenylyl cyclase⁵⁰ and prevents protein kinase A (PKA) from phosphorylating myosin light chain (MLC) kinase, which remains active and prevents vascular relaxation.^{51, 52} Agonists of α2-adrenergic receptors are believed to decrease AH production by decreasing blood flow to the ciliary body,^{1, 52–54} and increase uveoscleral outflow.⁵⁵

The topical administration of β-adrenergic receptor antagonists, or β-blockers, which have been used to treat cardiovascular hypertension since the 1960s,^{56, 57} was a breakthrough in glaucoma treatment, since they significantly lowered IOP with fewer side effects than older glaucoma drugs. Before the introduction of prostaglandin analogs (see below), β-blockers were the most commonly prescribed treatment for glaucoma.^{58, 59} β-adrenergic receptors are found within the ciliary body⁶⁰ and early studies showed β-blocker propranolol decreased IOP in glaucoma patients,⁶¹ but its usefulness was limited due to side effects.⁴⁶ In the late 1970s, a topical timolol treatment became the first FDA approved β-blocker for the treatment of glaucoma.⁴⁶ Timolol decreases the formation of AH,^{62, 63} perhaps via the aforementioned mechanisms involving chloride transport in ciliary epithelium and cAMP,^{25, 64} or by cAMP-independent mechanisms,^{65, 66} but does not affect outflow facility.^{62, 63}

Carbonic Anhydrase Inhibitors

Carbonic anhydrase inhibitors (CAIs) are the best understood glaucoma therapeutic class on a molecular level. The production of bicarbonate by the Zn-dependent carbonic anhydrase enzyme is necessary for chloride transport, and thus AH formation (see above).²⁴ Inhibition of carbonic anhydrase decreases AH formation and in turn decreases IOP. The original CAIs were sulfonamides developed as diuretics and antimicrobials.⁴⁶ Current CAIs include the systemic acetazolamide and the topical dorzolamide.⁴⁸ While CAIs can offer decreases in IOP similar to β-blockers (see above), this effect is seen mostly with systemic treatment, which, with prolonged use, can cause side effects such as paresthesia of the hands and feet. Topical CAIs have fewer adverse effects, but are also less potent.⁵⁹

Prostaglandin Analogs

Prostaglandin analogs, the newest class of glaucoma drugs, have become the first line of treatment.⁶⁷ Naturally occurring prostaglandins bind a variety of cell surface receptors and are involved in mediating smooth muscle contraction and inflammation.⁶⁸ Since their discovery,⁶⁹ prostaglandins and their analogs have been studied as potential therapeutics in a range of diseases. For glaucoma, the finding that prostaglandin F2α lowered IOP in monkeys bolstered the interest in further evaluating prostaglandins for therapeutic intervention.⁴⁶ Prostaglandin analogs have been shown to lower IOP mainly via increasing uveoscleral outflow,^{70, 71} along with conventional AH outflow.⁷² While the specific mechanism of action of increasing outflow remains unknown, studies indicate that prostaglandin analogs affect ECM remodeling^73–75 and may relax the ciliary body.⁷⁶ Prostaglandin analogs have short half-lives and can be administered locally to the eye at low doses, which reduces the likelihood of occurrence of side effects.⁶⁷

Surgery and Implants

When drugs fail to lower IOP, more invasive measures such as laser treatment, surgery or shunt implant may be employed (Table 1B). Improvements in surgical techniques have lowered the risk of such procedures and are thus becoming more prevalent, including alongside drug treatments, for both CAG and POAG. Argon laser trabeculoplasty (ALT), or the newer, lower energy alternative, selective laser trabeculoplasty (SLT) mechanically increase outflow by damaging TM tissue, termed photocoagulation.⁷⁷ How photocoagulation of the TM leads to increased AH outflow is not understood, but the effect is thought to be due to either scarring of the TM or an increase in macrophages that clear debris via phagocytosis.^{78, 79} SLT is generally considered a safe treatment for IOP reduction as most complications are only present for a short time after the procedure; however, additional medication may need to be used and over time repeated treatment may be required to lower IOP to clinically acceptable levels.^{77, 78, 80} Trabeculectomy, also called glaucoma filtration surgery, is a third alternative. Due to a high 5-year failure rate⁸¹, and complications such as cataract, this surgery is reserved until other treatments have been unsuccessful.⁸² The surgery creates a fistula for AH to bypass the damaged TM, enabling increased outflow.⁷⁹ A shunt may be added after trabeculectomy to maintain the opening created by surgery.^{83, 84} Thus far, shunts appear to decrease the complications associated with trabeculectomy, but these devices are relatively new and lack long-term studies.⁸⁴ Finally, specific to CAG, a hole in the iris can be created either by a laser iridotomy⁸⁵ or surgical iridectomy,⁸⁶ to enable AH outflow despite the closed iridocorneal angle.

NEW THERAPIES IN LATE-STAGE DEVELOPMENT

New drugs under development attempt to leverage our increasingly better understanding of the role eye tissues, particularly the TM⁸⁷ and optic nerve head (ONH),⁸⁸ play in glaucoma pathogenesis. As described below, Rho-associated protein kinase (ROCK) inhibitors and adenosine receptor (AR) ligands modulate the pathways involved in both ECM and cytoskeletal restructuring. In addition to lowering of IOP, these molecules appear to increase ONH blood flow or protect against RGC loss.⁸⁹ The neuroprotective benefits of these molecules are exciting because they are orthogonal to IOP reduction, which, from a clinical perspective, already has inexpensive and successful options (see prior section). The lack of clinical tools with sufficient sensitivity to unequivocally demonstrate neuroprotection is an impediment, however. Currently available retina/optic nerve imaging tools and visual field tests cannot detect the subtle changes that occur at a cellular level early in the disease process, when neuroprotection would be most impactful.⁹⁰ Delivery of neuroprotective agents to the posterior eye is also an important area of consideration (see also Outlook).

ROCK inhibitors

ROCKs are serine/threonine kinases that are downstream effectors of Rho GTPases. ROCKs phosphorylate substrates involved in regulation of F-actin stress fiber formation, focal adhesion,^{91, 92} as well as cell-cell adhesion and migration properties.^{93, 94} Downstream effects on actin cytoskeleton organization and smooth muscle contractility are relevant to many cellular processes, and dysregulation of the ROCK-Rho signaling system is associated with diseases of tissues throughout the body.^{95, 96} Known ROCK substrates are numerous and include MLC kinase, Lin11, Isl1 and Mec3 Kinase (LIM Kinase), and adducin, among others.⁹³ ROCKs are composed of an N-terminal kinase domain (Figure 2a), a coiled-coil domain with a Rho-binding site, and a pleckstrin homology (PH) domain.^{96, 97} Upon binding of GTP-activated Rho GTPase, the ROCK autoinhibitory complex between the N-terminal kinase domain and the C-terminal Rho-binding and PH domains is disrupted to enhance ROCK phosphorylation activity.⁹³ There are two ~160 kDa ROCK isozymes, ROCK1 and ROCK2, which share 92% sequence similarity in their kinase domain.^{97, 98} ROCK1, predominantly expressed in non-neuronal tissues such as the skeletal muscles, heart and lung, and ROCK2, primarily expressed in the brain, phosphorylate the same substrates, suggesting they have at least partially overlapping functions.⁹⁹ ROCK1 is required for stress fiber and focal adhesion formation, whereas ROCK2 is involved in phagocytosis and formation of the fibronectin fibrillar matrix.^{100, 101}

Figure 2. ROCK inhibition.

(a) Cartoon representation of ROCK-1 bi-lobed kinase domain in complex with R-1 (PDB code 2F2U). Canonical domain features highlighted: Small N-terminal lobe; larger C-terminal lobe; P-loop, phosphate binding loop containing glycine-rich motif; C-loop, catalytic loop; A-loop, activation loop. (b) Zoom into active site and residues within hydrogen bonding distance of fasudil presented in ball-and-stick. (c) Summary of proposed signaling cascade for ROCK inhibition leading to reduced IOP.

The ROCK-Rho signaling system is pertinent to AH outflow and glaucoma.^{102, 103} Both ROCK isoforms are expressed in tissues throughout the eye including in the TM, ciliary body, corneal epithelium and the retina.¹⁰⁴ The selective role of ROCK1 and ROCK2 in glaucoma remains unknown, however. While ROCK1 is highly expressed in the ciliary epithelium, ROCK2 is expressed in the retina,¹⁰⁴ and both ROCK1 and ROCK2 knockout mice have lower IOP than wild type mice.¹⁰⁵

Development of ROCK inhibitors to treat glaucoma

Modulation of ROCK is an active therapeutic direction to treat cancers,¹⁰⁶ cardiovascular disorders,⁹⁵ metabolic disease,¹⁰⁷ neurodegenerative disorders,⁹⁶ as well as glaucoma (Table 2). As such, considerable effort has been placed in developing selective ROCK inhibitors, and then testing them in potential therapeutic contexts; review articles detailing the extended history of the discovery, improvement, and validation of the ROCK kinases have been written.^{96, 108–111} Briefly, ROCK inhibitors predate the discovery of the ROCK enzyme and were derived from naphthalene-based calmodulin (CaM) antagonists, namely, isoquinoline, which serendipitously inhibits protein kinases at high concentrations.¹¹² The isoquinoline derivative R-1 (HA-1077/fasudil), which has been used clinically in Japan as a vasodilator to prevent cerebral vasospasm,¹¹³ and the related R-2 (K-115), have been tested in clinical trials for glaucoma (see below). Like other kinase inhibitor targets, these molecules competitively bind in the ROCK adenosine triphosphate (ATP) binding pocket and thus selectivity of ROCK over other kinases is difficult. For example, R-1 has been shown to inhibit protein kinase C-related protein kinase 2 (PRK2),¹¹⁴ cyclic GMP-dependent protein kinase G (PKG) and protein kinase C (PKC).¹¹¹ Structures of ROCK1 and ROCK2 in complex with R-1 (Figure 2b) and other inhibitors have revealed nuances in pocket shape and physicochemical properties compared to other kinases,⁹⁷ leading to compounds with more potent inhibition and specificity than R-1,^{111, 115} including some that have been tested in glaucoma models¹¹⁶ and are in clinical trials (Table 2).¹¹⁷ Many other newer-generation isoquinoline derivatives with nanomolar IC₅₀ values, better selectivity for ROCK, and promising pharmacokinetics properties^118–123 have yet to be tested for treating glaucoma, however. Pyridine is a second well-studied ROCK inhibitor scaffold and has likewise been subjected to extensive optimization for potency and selectivity.^124–126 R-3 (Y-27632), the first selective inhibitor for ROCKs¹²⁷ and derivatives such as R-4 (Y-39983),¹²⁸ R-5,¹²⁶ R-6, and R-7¹²⁹ have been examined for effects on IOP. Newer scaffolds identified via high throughput screening, such as pyrazole, indazole, aminofurazans, 2,4-diaminopyrimidines, benzodioxanes, urea derivatives and benzothiophenes, have expanded the chemical space for ROCK inhibition.^{108–110, 130, 131} Inhibitors based on these new scaffolds appear promising in other contexts, including overcoming issues of bioavailability,^{124, 132–139} which motivates their consideration for glaucoma. Thus far, treatment with R-8 (SR-3677), a derivative of the pyrazole scaffold,^{109, 140, 141} improves outflow facility in perfused porcine eyes;¹⁴¹ R-9, a potent and selective ROCK2 inhibitor derived from additional structure-activity-relationship (SAR) studies of the urea scaffold¹³⁰, lowers IOP in a rat model; R-10, an SAR-based derivative of the benzothiphene scaffold, yielded a comparable reduction in IOP to R-4.¹³¹

Table 2.

ROCK inhibitors

Compound (Other Names or Identifiers)	Structure	Potency (nM) a,b	Clinical Trials for IOP reduction
		ROCK1 ROCK2
R-1 (Fasudil, HA- 1077)		260 320
R-2 (K-115, Ripasudil)		51 19	Phase II (Approved for treating glaucoma and ocular hypertension in Japan)
R-3c (Y-27632)		200 140
R-4 c (Y-39983, RKI-983, SNJ-1656)		N/A 3.6	Discontinued after Phase II
R-5		300 78
R-6		346 75
R-7		105 24
R-8c SR-3677		56 3
R-9		N/A <1
R-10		20 N/A
R-11 Scaffold for AMA-0076		N/A 2.5 (reported for compound where X=F R= CH2-oxolan-2-yl)	Phase II - completed
R-12 (AR-12141 Precursor to AR-12286)		6 1.5	AR-12286 in Phase II
R-13 (AR-12432 Precursor to AR-13324)		1.2 1.1	AR-13324 in Phase III

^aIC₅₀ values unless otherwise noted.

^bReferences R-1¹¹⁵, R-2²⁷², R-3¹⁰⁹, R-4,¹¹⁹ R-5¹²⁶, R-6,¹²⁹ R-7,¹²⁹ R-8,^{109, 141} R-9,¹³⁰ R-10,¹³¹ R-11,¹⁵⁶ R-12, ¹⁵⁷ R-13¹⁵⁷.

^cTested in perfusion studies. N/A: not available.

Proposed IOP-lowering mechanisms (Figure 2c)

ROCK inhibitors have been tested in perfusion studies using porcine, bovine, and rabbit eyes, and in vivo in both ocular normotensive and hypertensive rabbit eyes.⁹⁴ These studies have repeatedly demonstrated that ROCK inhibitors decrease IOP by increasing AH drainage through the TM via cytoskeletal restructuring.^{94, 142, 143} For example, treatment of perfused bovine eyes with R-3 leads to changes in cellular morphology of TM and SC cells, decreases the number of focal adhesions, increases permeability of SC cells,¹⁴⁴ and widens extracellular spaces in the TM of perfused eyes.¹⁴² Such changes have been attributed to a decrease in MLC phosphorylation upon ROCK inhibition.^{142, 145} Consistent with this model, MLC kinase inhibitors have also been shown to affect the actin cytoskeleton of cultured human TM cells and reduce IOP of rabbit eyes,¹⁴⁶ similar to the effects seen with ROCK inhibitors. In addition to inhibitors of MLC kinase, those targeted to LIM kinase have also been shown to reduce IOP.^{143, 147} ROCK activation of LIM kinase leads to inhibition of actin depolymerization by cofilin.¹⁴² Conversely, LIM kinase inhibition, either directly or through upstream inhibition of ROCK, appears to induce TM relaxation via depolymerization of actin filaments.¹⁴³ Thus, ROCK inhibitors have not only shown promise as glaucoma therapeutics, but studies of their downstream effects have opened up new potential IOP lowering targets, namely, MLC kinase and LIM kinase.

Other potential therapeutic benefits

Besides lowering IOP, ROCK inhibitors are active in the posterior eye. R-1 increases ONH blood flow in rabbits by restoring induced blood flow impairment and related optic nerve damage.¹⁴⁸ Similar results were obtained for the R-4 in restoring ONH blood flow, also in rabbits.¹⁴⁹ R-1, R-3 and R-4 have been further shown to protect RGCs from apoptosis. R-1 protects the retina of rats from both ischemia/reperfusion induced damage¹⁵⁰ and N-methyl-D-aspartate (NMDA) induced excitotoxicity.¹⁵¹ In rats treated with R-4, the number of RGCs with regenerating axons was increased.¹⁴⁹ R-3 protected RGCs from ischemia induced damage in rat eyes ex vivo¹⁵² and increased RGC regeneration after optic nerve crush both in rats and RGC culture.¹⁵³ Finally, R-3 prevented postoperative scarring in a glaucoma surgery rabbit model by inhibiting collagen deposition and fibroproliferation, increasing positive surgery outcomes over controls.¹⁵⁴

Clinical Trials

With the consistent success of lowering IOP upon treatment in animal models and their neuroprotective effects, several ROCK inhibitors have entered clinical trials (clinicaltrials.gov).^{94, 109, 155} R-2 (from Kowa Pharmceuticals), is currently in Phase II trials as is the pyridine derivative R-11 (AMA-0076, Amakem).¹⁵⁶ R-4 (RKI-983 or SNJ-1656 from Senju and Novartis, respectively) was discontinued during Phase II. Aerie Pharmaceuticals has put forth several isoquinoline derivative ROCK inhibitors R-12 (AR-12286), R-13 (AR-13324), and R-13 combined with the commonly prescribed IOP-reducing prostaglandin analog latanaprost (PG-324). At the time of writing this Perspective, Aerie Pharmaceuticals is recruiting subjects for Phase III trials for R-13.¹⁵⁷

Adenosine Receptor Ligands

Adenosine receptors (ARs), GPCRs that regulate adenylyl cyclase and thus affect the production of cAMP, are expressed in tissues throughout the eye including the ciliary body,¹⁵⁸ the TM,¹⁵⁹ SC,¹⁶⁰ and the retina¹⁵⁸ and have been implicated in maintenance of IOP.¹⁶¹ At the cellular level, AR classes A₁ and A₃ deactivate adenylyl cyclase via G_i proteins, and A_2A and A_2B activate adenylyl cyclase through G_s proteins.¹⁶² A₁AR and A₃AR activate phospholipase C (PLC), which is part of the mitogen-activated protein kinase (MAPK) and the extracellular signal related kinase (ERK1/2) pathway. A₁AR activation of ERK1/2 has been shown to lead to secretion of matrix metalloproteinases (MMPs) and accelerate ECM turnover.¹⁶³ A_2AAR and A_2BAR-mediated stimulation of cAMP production also modulates ERK1/2; however, this activates connective tissue growth factor (CTGF), increasing ECM deposition.¹⁶⁴ Thus, depending on the AR class, agonists or antagonists should elicit the desired effect of lowered IOP (Figure 3a).¹⁶⁵ For example, to promote IOP maintenance, enhancement of ECM turnover via an A₁AR agonist or a decrease in ECM deposition with an A_2AAR antagonist should lower IOP.

Figure 3. AR ligands.

(a) Summary of proposed signaling cascade for AR agonist and antagonist leading to reduced IOP. Dashed arrows imply the possibility of additional intervening steps in the signaling cascade. (b) Left: Cartoon representation of A_2AAR with bound AR-2 colored in rainbow from N-terminus (blue) to C-terminus (red). Perpendicular helix 8 (red) may be involved in protein-protein interactions. Right: Zoom into ligand binding pocket and residues within hydrogen bonding distance of AR-2 presented in ball-and-stick. (c) Summary of SAR efforts using adenosine and xanthine scaffolds.

Development of AR ligands to treat glaucoma (Table 3)

Table 3.

AR agonists, antagonists

Compound	Function	Structure	Affinity/Potencya,b (nM)	Clinical Trials
AR-1 (Adenosine)	Non-selective agonist		A1 310 A2A 730 A2B 23500 A3 290 (EC50 values)
AR-2 (NECA)	Non-selective agonist		A1 14 A2A 20 A2b 140 A3 25
AR-3 (CHA)	A1 agonist		1.3–2.3
AR-4 (R-PIA)	A1 agonist		A1 2 A3 16
AR-5 (INO-8875)	A1 agonist		0.97	Phase I/II
AR-6 (CPA)	A1 agonist		5.9
AR-7 (CF-101, IB-MECA)	A3 agonist		A3 1.8 A1 51	Phase II
AR-8 (CGS-21680)	A2A agonist		A2A 27 A3 67
AR-9 (ATL-313)	A2A agonist		0.7 nM (Kd)	Phase I
AR-10 (2-O-Ado)	A2A agonist		31.8
AR-11 (2-CN-Ado)	A2A agonist		49.3
AR-12 (OPA-6566)	A2A agonist	N/A	N/A	Phase II
AR-13 (ZM241385)	A2A antagonist		1.6
AR-14 (MRS-1097)	A3 antagonist		108
AR-15 (MRS-1191)	A3 antagonist		31.4
AR-16 (MRS-1523)	A3 antagonist		19
AR-17 (LJ-1251)	A3 antagonist		4.16
AR-18 (MRS-1292)	A3 antagonist		42 (IC50)
AR-19 (LJ-979)	A3 antagonist		15.5
AR-20 (MRS-3826)	A3 antagonist		910
AR-21 (MRS-3827)	A3 antagonist		N/A
AR-22 (MRS-3771)	A3 antagonist		29
AR-23 (OT-7999)	A3 antagonist		0.61

^aK_i values unless otherwise noted.

^bK_i/IC₅₀/EC₅₀/K_d value references: AR-1,²⁷³ AR-2²⁷⁴, AR-3²⁷⁵, AR-4¹⁵⁵, AR-5¹⁵⁵, AR-6²⁷⁶, AR-7¹⁵⁵, AR-8^{274, 275} AR-9²⁷⁷, AR-10¹⁷⁶, AR-11¹⁷⁶, AR-13¹⁷³, AR-14²⁷⁸, AR-15²⁷⁸, AR-16¹⁸⁹, AR-17^{180, 211, 279}, AR-18¹⁸⁹, AR-19¹⁸⁰, AR-20¹⁸⁰, AR-22¹⁸⁰, AR-23²⁷⁴ N/A: not available.

AR-1 (adenosine) was linked to heart rate modulation and arterial blood pressure more than 80 years ago¹⁶⁶ and soon after was used clinically to treat arrhythmia.¹⁶⁷ Ligands for particular AR classes have been developed for decades,¹⁶⁸ for a variety of diseases beyond cardiovascular, such as rheumatoid arthritis and Parkinson’s disease,^{162, 169, 170} by taking advantage of their varying natural affinities for adenosine.¹⁶⁹ Because ARs are expressed throughout the body, including in the brain, heart, lung and blood vessels,¹⁷¹ drug development for any particular disorder is fraught with significant side effects. Extensive SAR studies have been completed, guided by the available A_2AAR structures (Figure 3b), including those in complex with AR-1 (PDB ID 2YDO),¹⁷² as well as synthetic agonists such as 5′-N-ethylcarboxamidoadenosine (AR-2, NECA, Figure 3b) (PDB ID 2YDV)¹⁷² and antagonists¹⁷³.

Most AR agonists shown to decrease IOP are derivatives of AR-1 (Table 3, Figure 3c). In general, compounds with N⁶-cycloalkyl substitutions favor A₁AR (Figure 3c),^{155, 161, 168} and several including N⁶-cyclohexyladenosine (AR-3, CHA), (R)-N⁶-(2-phenyl-1-methylethyl) adenosine (AR-4, R-PIA), AR-5 (INO8875), and N^6-cyclopentyladenosine (AR-6, CPA) have been tested for their effect on IOP. A₃ARs are better targeted by N⁶-benzene or N⁶-substituted benzene ligands (Figure 3c), but to date, only AR-7 (CF-101/IB-MECA) has shown promise for lowering IOP.^{155, 161, 169, 174, 175} Agonists selective for A_2AAR and A_2BAR are those typically modified at the C² position, and agonists selective for A_2AARs are further substituted at the 5′ C within the ribose portion of the scaffold (Figure 3c).^{169, 174, 176} A_2AAR agonists demonstrated to lower IOP include AR-8 (CGS-21680), AR-9 (ATL-313), AR-10 (2-O-Ado) and AR-11 (2-CN-Ado). The fewest number of described agonists are those selective for A_2BAR, which is less tolerant of substitutions at positions that confer selectivity to other ARs; to the best of our knowledge, available A_2BAR agonists have not been studied in the context of glaucoma.¹⁶⁵

AR antagonists shown to reduce IOP include one targeted to A_2AAR and several to A₃AR (Table 3). To date, A₁AR antagonists have shown little promise as IOP lowering drugs as they increase IOP in both rabbits and mice, and effects of A_2BAR antagonists in the context of IOP and glaucoma have not yet been reported.¹⁶⁵ A₁, A_2A and A_2BAR antagonists are largely derivatives of xanthine (Figure 3c), including the natural antagonist caffeine. A₃ARs have low affinity for caffeine and other xanthine derivatives, leading to a range of examined scaffolds, which vary in their core structure and include purine and pyridine derivatives, quinazoline derivatives, tricyclic xanthines, and nuceloside derivatives.^177–180 A_2AAR antagonists include xanthine derivatives modified at C⁸ with styryl groups and nitrogen polyheterocyclic compounds,¹⁷⁹ such as AR-13 (ZM241385), which reduces IOP in mice.¹⁶⁵. A₃AR antagonists tested for their effect on IOP include pyridine derivatives AR-14 – AR-16, nucleoside derivatives AR-17–AR-22, and the tricyclic xanthine derivative AR-23.

Proposed IOP-lowering mechanisms (Figure 3a)

The molecular role(s) of the AR classes in IOP regulation is not well established, as there appears to be both issues of species specificity, as in the case of the A₂AR agonist AR-8,^{161, 181} as well as selectivity issues, as in the case of A₁AR agonists AR-3 and AR-4 respectively,¹⁶⁸ and is thus an important area for further investigation. Available evidence from studies conducted in bovine and monkey eyes indicates that A₁AR agonists lower IOP by increasing outflow^{182, 183} through the conventional pathway.¹⁸² At the molecular level, decreased outflow resistance from activation of A₁AR with agonists has been shown to result from enhanced signaling leading to elevated levels of relevant matrix metalloproteases 2 and 9 (MMP-2, MMP-9), ^{163,182, 184} and thus ECM remodeling.^{185, 186} Consistent with this finding, outflow increase observed with A₁AR activation can be reversed by MMP inhibition.¹⁸² A₁AR,¹⁸⁷ A_2AAR, and A₃AR agonists have all also been shown to reduce TM cell volume,¹⁵⁹ which in turn, independently increases outflow facility.¹⁸⁸ A₃AR antagonists AR-14, AR-15, and AR-16 prevent chloride release and reduce AH production in NPE cells in the ciliary body.^{189, 190} Decreased IOP has been documented in studies of mice after treatment with A₃AR antagonists as well as after A₃AR knockout,^{190, 191} but not all studies of A₃AR converge.¹⁶⁵ For example, in human clinical trials as a treatment for dry eye syndrome, the A₃AR agonist AR-7, i.e. not an antagonist, was found to lower IOP.¹⁹² It is suspected that chronic administration of AR-7 in this study led to lower IOP via the documented “effect inversion” pharmacological mechanism seen for other AR ligands in other contexts,¹⁹³ or due to inadvertent cross-activation of A₁ARs.¹⁶⁹

Other potential therapeutic benefits

RGC death in glaucoma is associated with ATP-binding P2X7 purinergic receptors at the cell surface due to an influx of Ca^2+194–197 and, more controversially, due to prolonged activation of glutamate-binding NMDA and non-NMDA receptors.^198–200 Activation of A₁ARs and A₃ARs by adenosine can inhibit the effects of both glutamate- and ATP- induced Ca²⁺ influx to prevent RGC death.^201–203 Treatment with adenosine has further been shown to increase ONH blood flow in healthy human eyes,²⁰⁴ and RGC loss has also been attributed to reduced ONH blood flow, as progression of the disease correlates with reduced ONH blood flow in humans^{205, 206} and in a rhesus monkey model of glaucoma.²⁰⁷ A₁ and A_2A subtypes were both specifically shown to be involved in ONH blood flow increase in rabbit eyes²⁰⁸ but involvement of A₃AR may also be relevant, as in one study, treatment of gerbils with A₃AR agonist AR-7 improved post-ischemic neuronal blood flow.²⁰⁹ Finally, there may be a role for A_2AAR in neuroprotection by controlling inflammation, analogous to reports related to Alzheimer’s and Parkinson’s disease.²¹⁰

Clinical trials

As of the writing of this Perspective, selected compounds are in preclinical (AR-17)²¹¹, Phase I (AR-9)^{169, 174}, and Phase II (AR-5 ¹⁵⁵, AR-12, and AR-7) trials (Table 3).^{155, 165}

SELECTED PROMISING NEW DIRECTIONS AND EMERGING TARGETS

Numerous targets are actively being pursued for IOP reduction with potential neuroprotective benefits.²¹² These include other targets to (a) increase ONH blood flow²¹³, (b) increase TM outflow via ECM remodeling or altering TM cell contractility, (c) increase uveoscleral outflow, and/or (d) decrease AH production. Here we highlight just a few examples of such targets, others are listed in Table 4.

Table 4.

Other therapeutic strategies under consideration.

Drug/target/pathway	Proposed Therapeutic Mode of Action
Angiopoietin-like molecules	Increase TM outflow through reorganization of the ECM.280
Cannabinoids	Decrease AH production and increases TM outflow, also may be neuroprotective.213
Cochlin	Regulation may modulate TM outflow.281
CTGF	Outflow increase by modification of TM actomyosin cytoskeleton.282, 283
Endothelin-1	Increase TM outflow by relaxation of TM, also shown to be neuroprotective of RGCs.284
Ghrelin	Increase TM outflow by relaxation of the TM.285
Latrunculins	Increase TM outflow by reducing cell contractility and cell-cell and cell-matrix adhesions. Drugs have reached human trials.286
MMP-2, MMP-9	Increase TM turnover, possible neuron regeneration in retina (see text).
Melatonin	Increase TM outflow through adrenoceptor and cholinoceptor modulation.287
Myocilin	Decrease of TM cell stress and death (see text).
NGF	Protect RGCs from neurodegeneration (see text).
Nitric oxide (NO)	Increase TM outflow by relaxation of the TM and ciliary muscle.288 A NO-donating prostaglandin drug has entered clinical trials.289
Transforming Growth Factor-β (TGF-β)	Increase TM outflow through ECM remodeling or modification of the actin cytoskeleton.290

Nerve Growth Factor (NGF)

NGF is an emerging treatment for glaucoma²¹⁴ based on its ability to prevent neural apoptosis, and its clinical evaluation in a number of non-ocular diseases including Alzheimer’s, Parkinson’s, diabetes, and HIV-related neuropathies.²¹⁵ NGF binds to the cell surface receptor tropomyosin kinase receptor A (TrKA) (Figure 4a), activating a signaling cascade which includes ERK1/2 and, through subsequent upregulation of B-cell lymphoma 2 (Bcl-2), prevents apoptosis.^{216, 217} Treatment with exogenous NGF has been shown to protect RGCs from death in rats after optic nerve sectioning²¹⁸ and ischemic injury,²¹⁹ to reduce retinal damage due to increased IOP in rabbits,²²⁰ and to protect RGCs in a rat glaucoma model.²²¹ Interestingly, in a small human trial, three patients with advanced glaucoma treated with NGF-containing eye drops showed improvements in visual field and optic nerve function that lasted for several weeks after treatment, with minimal side effects.²²¹ Results from Phase I trials indicate NGF is safe with limited side effects with ocular administration,²²² and, as of the writing of this Perspective, Phase II trials are ongoing for NGF as a dry eye treatment.

Figure 4. Emerging new potential targets.

(a) NGF in complex with TrkA extracellular domain (PDB code 2IFG).²⁶⁸ (b) Full-length pro-MMP-2 (PDB code 1CK7), with domains highlighted.²⁶⁹ (c) myoc-OLF 5-bladed propeller. Top: cartoon representation. Bottom: surface representation with bound glycerol molecules on surface presented as magenta sticks.²³¹ (d) Grp94 overall domain structure (PDB code 2O1U)²⁷⁰ and zoom into N-terminal ATP binding pocket with radamide inhibitor bound (PDB code 2GFD).²⁷¹

Matrix Metalloproteases

Zinc-dependent matrix metalloproteases are involved in a wide range of functions including bone remodeling, blastocyst implantation, nerve growth, and their namesake, ECM remodeling.²²³ ECM remodeling, in particular, plays a role in maintaining IOP as MMPs 1, 2, 3, 9 and 14 are all upregulated under high IOP conditions,²²⁴ inhibition of MMPs increase outflow resistance,¹⁸⁶ and, as discussed earlier, increased activity of both MMP-2 (Figure 4b) and MMP-9 are involved in decreased outflow resistance in the TM.^{182, 183, 185, 186} Substrates of MMP-1, MMP-2 and MMP-9 include collagens, fibronectin and elastin; whereas MMP-1 is upregulated, MMP-2 and MMP-9 have been found to be reduced in POAG eyes.¹⁸⁵ In principle, increasing MMP activity in the TM could be useful as a glaucoma therapy to restore normal outflow through the TM by increasing turnover of the ECM, particularly in steroid induced glaucoma where MMPs are downregulated with prolonged steroid treatment.²²⁵ In a gene therapy-based study, induced MMP-1 expression successfully compensated for MMP-1 downregulation caused by steroid administration, and both prevented and reversed steroid-induced ocular hypertension in sheep.²²⁶ MMP-1 was secreted to the ECM, degraded collagen type 1 in the TM, and thereby increased outflow.²²⁵ There is also some evidence supporting a role for MMPs in neurogenesis,²²⁷ neurodegeneration,^{227, 228} death after neural injury,²²⁹ and repair,²²⁷ germane to reversing retina degeneration, but molecular details are still forthcoming. Thus, further studies are needed to determine the potential of modulating MMPs for glaucoma in both capacities of lowering IOP and neuroprotection.

Myocilin

Myocilin was the first gene linked with inherited forms of glaucoma¹³ and is also upregulated in steroid induced glaucoma.¹⁴ Missense mutations in myocilin account for ~5% of all POAG and ~10% of all JOAG cases^{4, 230}, and are clustered within its olfactomedin (myoc-OLF) domain.^{230, 231} Myocilin is expressed in anterior segment tissues including the TM,^{232, 233}. Recent studies have implicated myocilin in cell proliferation and survival through the ERK1/2 pathway, which involves the aforementioned MMPs,²³⁴ as well as in programmed RGC cell death.²³⁵ While interacting partner(s) of myocilin and precise biological function remains elusive, it is an attractive de novo therapeutic target for glaucoma patients with inherited mutations.

The prevailing mechanism proposed for myocilin pathogenesis caused by mutations within myoc-OLF is that of a gain of function, namely, instead of being secreted to the TM, mutant myocilins aggregate within the ER of TM cells, causing TM cell death. TM cells are highly phagocytotic and clear debris in surrounding extracellular tissue for proper AH filtration.²³⁶ Thus, TM cell death is thought to bring about dysregulation of outflow through the TM, leading to elevated IOP, and then glaucoma.⁴ Disease-causing myoc-OLF variants are less stable than wild type,²³⁷ and grow degradation-resistant amyloid-like fibrils in vitro, which have also been detected in cells.^{238, 239} Notably, the absence of myocilin, either due to the presence of a premature stop codon in humans, or due to its knock-out in mice, does not result in glaucoma,^{240, 241} suggesting that viable therapeutic strategies would prevent or alleviate aggregation and its toxic effects.

One avenue is to inhibit myocilin misfolding as a means to prevent aggregation and escape detection by ER-associated degradation (ERAD). The result would be the secretion of mutant myocilin, which could restore partial function or take advantage of TM cell phagocytosis for facile degradation. Early studies demonstrated that mutant myocilin secretion could be rescued by culturing TM cells at a lower temperature,^242–244 presumably because aggregation rates are slowed at temperatures below the unfolding temperature of mutant myocilin. Thus far, attempts to rescue secretion using small molecules have included both chemical and pharmacological chaperone approaches. For the former, chemical chaperones trimethylamine-N-oxide (TMAO)²⁴⁵ and 4-phenylbutyrate (4PBA)²⁴⁶, which is approved to treat urea cycle disorders and is under consideration for other ER-stress inducing misfolding diseases like cystic fibrosis²⁴⁷, improved secretion of mutant myocilin in cell culture. Related, a series of osmolytes were shown to increase thermal stability of mutants to better than wild-type levels in vitro,^{237, 248} with simple compounds like sucrose exhibiting the most potent stability enhancement. The high concentrations of osmolytes required to enhance the stability and secretion of mutant myocilin limits their usefulness as therapeutics, however. With no known function for myocilin, there is no biochemical assay for the discovery of tailored molecules, so-called pharmacological chaperones, and secondary assays such as cell secretion and inhibition of aggregation are low-throughput, making their use as a primary screening method laborious. One high-throughput, target-independent stability assay was developed recently for myoc-OLF in which the protein is first slightly destabilized with denaturant, and potential ligand scaffolds are then identified by monitoring subsequent re-stabilization using a straightforward fluorescence read-out. From this assay, two compounds were found that restored wild type myoc-OLF thermal stability, inhibited amyloid formation in vitro and increased secretion in a cell model.²⁴⁹ The recent crystal structure of myoc-OLF (Figure 4c) should accelerate the discovery of new selective ligands with similar properties through in silico and rational drug design methods. To this end, several small molecules were found serendipitously to bind to the myoc-OLF surface in the crystal, suggesting the surface is druggable (Figure 4c).²³¹

An alternative method to prevent the intracellular build up of mutant myocilin is to manipulate ERAD-specific molecular chaperones, an approach also under consideration for a number of other diseases.²⁵⁰ For glaucoma, treatment of a Y437H-mutant myocilin mouse model with the aforementioned chemical chaperone 4PBA reduces IOP.²⁵¹ The targets of 4PBA are most likely chaperones, such as Hsc70,²⁵² Hsp70,²⁵³ and glucose regulated protein 78 (Grp78),²⁵⁴ but not myocilin²⁴⁸. The lack of a clear molecular target for 4PBA, however, is a disadvantage for SAR studies to generate more potent and selective molecules. A more directed approach may be to target Grp94 (Figure 4d), which has been shown to recognize mutant myocilin in a tight interaction that abrogates retrotranslocation for degradation by the proteasome.²⁵⁵ Although the reasons for the unproductive interaction are not yet clear, Grp94 increases the rate of myoc-OLF aggregation and co-precipitates with myoc-OLF in vitro.²⁵⁶ In cell culture, if levels of Grp94 are reduced by treatment with RNAi, or activity is inhibited using molecules directed to the N-terminal ATP binding site (Figure 4d), mutant myocilin clearance is promoted by autophagy.²⁵⁵ Incubation of Grp94 with a more selective inhibitor than radamide (see Figure 4d) reestablishes baseline aggregation rates and prevents co-aggregation in vitro.²⁵⁶ From a medicinal chemistry perspective, this proteostasis approach is attractive for several reasons: (1) degradation of myocilin should not cause glaucoma (2) co-crystal structures of Grp94 with inhibitors are tractable, enabling a pathway for development of new molecules and SAR, (3) straightforward secondary assays to assess aggregation in vitro and in cells are accessible, and (4) myocilin-glaucoma mouse models are available for testing optimized compounds. Nevertheless, given that Grp94 levels are investigated in cancer^{257, 258} and are correlated with apoptosis,^{259, 260} this approach will require considerable work to prevent such activity and ensure selectivity over other similar chaperones.

OUTLOOK

Progress critical to the development of new glaucoma drugs will continue to require highly interdisciplinary teams composed of individuals spanning fields of expertise from physics to medicine, to biology, biochemistry, and engineering. New targets will emerge as our understanding of normal and pathological aspects of TM biomechanics and molecular details of the mechanisms regulating AH improves. Susceptibility genes found through GWAS and family studies worldwide will also continue to identify new potential targets and research directions. One area of growth for the field is to lower the barrier to testing potential glaucoma drugs with a high likelihood of translation to humans, for example, a three dimensional model of the TM in which changes correlate with IOP reduction;⁸⁷ until a broadly accepted and robust system is developed, testing molecules for IOP reduction will remain extremely low throughput, relying on perfusion studies or animal models. Besides small molecules, gene therapy with viral vectors and small interfering RNAs (siRNAs),^{261, 262} analogous to other diseases such as Alzheimer’s and Parkinson’s,²⁶³ are also in clinical trials. Facile and accurate methods to evaluate neuroprotection efficacy need to be developed. Finally, patient compliance is critical for therapeutic outcomes in the clinic. This issue is being addressed through advanced delivery methods such as implants, contact lenses, microneedles, and gels,^264–267 which are being tested with current therapies. In the long term, such systems will offer localized, sustained, delivery of drugs, and better bioavailability in particular tissues, to improve the effectiveness of all classes of glaucoma therapies.

ABBREVIATIONS

4PBA

4-phenylbutyrate

AH

aqueous humor

ALT

alternative laser trabeculoplasty

AR

adenosine receptor

Bcl-2

B-cell lymphoma 2

CAG

closed-angle glaucoma

CAI

carbonic anhydrase inhibitors

CaM

calmodulin

CTGF

connective tissue growth factor

ECM

extracellular matrix

ERAD

ER-associated degradation

ERK1/2

extracellular signal related kinase

GWAS

genome-wide association studies

IOP

intraocular pressure

JOAG

juvenile open-angle glaucoma

LIM Kinase

Lin11, Isl1 and Mec3 kinase

MLC

myosin light chain

myoc-OLF

myocilin olfactomedin domain

NGF

nerve growth factor

NPE

non-pigmented epithelial

NTG

normal tension glaucoma

OAG

open-angle glaucoma

ONH

optic nerve head

OPTN

optineurin

PEX

pseudoexfoliation

PH

pleckstrin homology

PKG

cyclic GMP-dependent protein kinase G

POAG

primary open-angle glaucoma

PRK2

protein kinase 2

RGC

retinal ganglion cell

ROCK

Rho-associated protein kinase

SAR

structure-activity-relationship

SC

Schlemm’s canal

siRNAs

small interfering RNAs

SLT

selective laser trabeculoplasty

TGFβ

Transforming growth factor-β

TM

trabecular meshwork

TMAO

trimethylamine-N-oxide

TrKA

tropomyosin kinase receptor A

Biographies

Rebecca K. Donegan earned her B.Sc. in chemistry from Middle Tennessee State University and defended her Ph.D. dissertation in April 2015. Her doctoral work under the supervision of Raquel Lieberman included solving the crystal structure and characterizing the calcium binding of the olfactomedin domain of myocilin.

Raquel L. Lieberman received her B. Sc. In Chemistry from Massachusetts Institute of Technology in 1998 and her Ph. D. in Chemistry from Northwestern University in 2005 where she worked with Prof. Amy C. Rosenzweig on biophysical and structural studies of an intramembrane metalloenzyme. After conducting postdoctoral work with Prof. Michael S. Wolfe (Center for Neurological Diseases, Harvard Medical School) and Prof. Gregory A. Petsko (Brandeis) where she expanded to molecular aspects of protein misfolding, she joined the faculty in the School of Chemistry and Biochemistry at Georgia Institute of Technology and is currently an Associate Professor. Research projects in Dr. Lieberman’s lab focus on molecular aspects of proteins implicated in misfolding disorders.