Retinal ganglion cell death in glaucoma

Clinical background

Glaucoma is a collection of optic neuropathies that exhibit similar clinical phenotypes of thinning of the nerve fiber layer and excavation or cupping of the optic nerve head. Collectively, the glaucomas are a relatively common, but serious, blinding disease that is anticipated to affect nearly 80 million individuals worldwide by the year 2020. The most prevalent form of this disease is known as primary open-angle glaucoma (POAG). It is often associated with elevated intraocular pressure (IOP) and has no distinguishing pathology in the angle of the eye, which is defined as the junction between the iris and cornea, and is the location of the aqueous outflow channels that are critical to IOP homeostasis. In addition to POAG, there are several other varieties of this disease. These include normal-tension glaucoma (NTG), which is associated with IOPs at or below the population average, angle closure glaucoma (ACG), and secondary glaucomas resulting from pigment dispersion or pseudoexfoliation, which are associated with the accumulation of debris in the angle leading to obstruction of the outflow pathways and the elevation of IOP. Clearly, there are a variety of factors at play that affect an individual’s susceptibility to the effects of IOP. Not surprisingly, family history of glaucoma is a major risk factor and glaucoma is often considered a complex genetic chronic neurodegenerative disease of the central nervous system (CNS).


The common feature of all forms of glaucoma is the progressive degeneration of the optic nerve and retinal ganglion cells (RGC) in the retina. This loss of RGCs occurs through an apoptotic-like pathway. In this chapter, we will discuss the current knowledge of ganglion cell death in the context of elevated IOP and damage to the optic nerve. The relationship between IOP and initiation of the disease is not well understood. Although increased IOP is the most important risk factor, the majority of individuals with ocular hypertension will never develop the disease. Alternatively, lowering IOP, which is the only current treatment for glaucoma, is an effective therapy for most forms of the disease, even for people with NTG. Recent findings may suggest new ways to help early diagnosis of the disease and provide therapeutic options in addition to lowering IOP.


RGCs are CNS neurons that transmit visual signals processed in the retina to the visual centers of the brain. Although the mechanism of insult that initiates apoptosis in the RGCs of a glaucomatous eye is not well understood, there are two predominant theories – mechanical damage of RGC axons in the region of the optic nerve head and vascular disturbances in the optic nerve head leading to ischemia in the retina that directly affects RGCs. In humans, it is thought that mechanical damage occurs at the lamina cribrosa where the axons pass through the laminar plates. Pressure on this series of collagenous plates may cause conformational changes in the pores through which bundles of axons pass, thus leading to compression of the axons. The compression could compromise the transport of small molecules, such as neurotrophins, through the axonal process. The loss of neurotrophic support to neuronal soma likely plays a role in the response of the cell body to axonal damage. A similar pattern of damage is seen in mouse models of glaucoma, but the mouse laminar region does not contain collagenous plates. Instead, bundles of murine axons in the optic nerve are surrounded by sheaths of glial cells. Focal damage to these discrete bundles leads to wedge-shaped sectors of RGC loss in the retina ( Figure 27.1 ).

Figure 27.1

Patterns of ganglion cell somal and axonal loss in DBA/2J mice. (A–D) Correlation between patterned cell and axon loss in the DBA/2J mouse model of glaucoma. The retinas were stained for ganglion cell-specific βGEO enzyme activity and the optic nerves were silver-stained after being cross-sectioned immediately posterior to the laminar region. (A, B) Uniform staining throughout the retina is matched by even staining of the axonal bundles in an unaffected eye. (C, D) Two wedge-shaped regions of soma loss (* in C) correlate with two regions of axonal loss in the optic nerve head (* in D).

(E) Dieback degeneration of optic nerves in a mouse model of glaucoma. Optic nerves were labeled postmortem with crystals of 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI) placed at the optic nerve head. These are whole-mount preparations of the left optic nerve from five individual mice. The orientation of each nerve is arranged so that the chiasm is at the top of the panel and the globe is at the bottom. In this analysis, optic nerve (ON) integrity is scored from 1 (most normal) to 5 (complete degeneration). Progressive retrograde degeneration is evident as a loss of DiI labeling as the nerve deteriorates.

(F) Longitudinal, silver-stained section of an optic nerve of a young DBA/2J mouse. Note, the retina has been removed from this preparation. Region A begins at the ON head. Regions B and C contain the laminar region of the ON and clearly show discrete bundling of the axons. Region D represents the postlaminar, myelinated section of the ON. Axons are no longer organized in bundles at this point. Sections shown in (B) and (D) of this figure were cut through region C of the ON. Wedge-shaped loss of somas in the retina is indicative that discrete bundles of axons are being damaged in the glaucomatous DBA/2J mouse. Since discrete axon bundles only occur in the laminar region of this ON, it suggests that this is the initial site of damage in glaucoma in this animal model.

In (B) and (D), the central retinal artery (A) and vein (V) are indicated. In E, bar = 0.5 mm

(A–E: modified from Schlamp CL, Li Y, Dietz JA, et al. 2006. Progressive ganglion cell loss and optic nerve degeneration in DBA/2J mice is variable and asymmetric. BMC Neurosci 2006;7:66.)

Both mechanical damage and ischemia could lead to activation of the supporting glial cells in the optic nerve. It is well established that glia in the CNS become activated in response to damaging stimuli, and several studies have shown that this is also the case in glaucoma. For example, Hernandez et al have shown that at least 150 genes are upregulated in astrocytes from a glaucomatous optic nerve head. More recently, Johnson et al have shown that glial changes in the early glaucomatous optic nerve head include the upregulation of genes involved in cell proliferation, suggesting that this behavior of glia is one of the earliest events associated with optic nerve head pathology. Although it is possible that glial cells are exerting a protective effect on adjacent axons, several studies have suggested that they are triggering the damage to these axons. There are several theories as to how this is accomplished, ranging from direct transmission of neurotoxic compounds, such as nitric oxide, to a passive role, such as reducing energy available to the axons due to decreased glycogen breakdown or stimulating vasoconstriction of capillaries surrounding axons in the lamina. The stimulation of vasoconstriction by the glia could be a link between mechanical damage and ischemic damage during glaucoma.


Compartmentalized self-destruct pathways and degeneration

Whitmore and colleagues have proposed the idea of addressing glaucoma as a neurodegenerative disease in which destruction of the neuron occurs compartmentally. Specifically, compartmental degeneration of the axon, synapse, and dendrites can occur independently of somal loss ( Box 27.1 ). Autonomous axonal degeneration has already been shown to occur in a mouse model of progressive motorneuronopathy. In a murine model of inherited glaucoma, Libby et al have shown that axonal loss occurs independently of somal loss, not just spatially but also via a distinct molecular pathway. In this latter model, which will be discussed in more detail below, the loss of the proapoptotic protein BAX prevents somal loss following increased IOP, but does not prevent axonal degeneration. In addition, experiments using a primate model of experimental glaucoma have shown that damage to the dendritic arbor sometimes precedes damage to ganglion cell bodies. Currently, no studies have shown definitively that synaptic degeneration plays a role in glaucoma pathophysiology, but this cannot be ruled out as a possibility.

Box 27.1

Degeneration pathology

  • Retinal ganglion cell (RGC) degeneration occurs compartmentally, with the different regions of the nerve responding to the initiating insult in a semi-independent manner

  • As demonstrated in Bax knockout mice, axonal degeneration and somal degeneration are autonomous processes. The destruction of one compartment of the neuron does not guarantee degradation of another

  • Damage to the axon precedes damage to the cell soma in models of experimental glaucoma

  • The pattern of damage in DBA/2J mice, which shows wedge-shaped regions of ganglion cell loss, suggests that the initial site of damage is at the optic nerve head where axons for these cells are bundled. This is consistent with early speculation that the lamina cribrosa is the initial site of damage in glaucoma

  • Selective loss of small- versus large-body RGCs is controversial

Wallerian degeneration versus die-back

Damaged axons usually degenerate in one of two basic patterns: wallerian degeneration or “die-back.” Each pattern of degeneration appears to be dependent on the severity of damage to the axons. Wallerian degeneration generally occurs in severely damaged axons and is characterized by a rapid loss of axonal structure throughout the length of the axon. Die-back occurs in axons with more moderate injury and is characterized by a slower retrograde degeneration that proceeds from the synapse to the soma. Although it is not known how axons in a glaucomatous human eye degenerate, clues to this process have come from recent studies in DBA/2J mice. These mice develop an iris atrophy, which leads to an accumulation of pigment in the trabecular meshwork causing an increase in IOP and an optic neuropathy that is similar to human pigmentary glaucoma. It appears that damage to the axon is relatively mild in this chronic model of ganglion cell loss, as the observed axonal degeneration exhibits a pattern similar to die-back ( Figure 27.1 ). In contrast, a more severe insult to the optic nerve, such as axotomy, causes the axons to degenerate in a wallerian pattern. If ganglion cells in a glaucomatous retina undergo this same type of slow axonal degeneration as in the DBA/2J mice, it may provide an explanation for the presence of visual field defects in individuals with no detectable loss of ganglion cells. That is, the axon has already begun to degenerate, so there is no connection to the visual centers of the brain, even though there appear to be unaffected ganglion cell bodies in the retina.

Selectivity of ganglion cell loss

Some controversy exists over whether or not some ganglion cells are more susceptible to apoptosis in glaucomatous conditions ( Box 27.2 ). Early studies indicated that large ganglion cell and nerve fibers were selectively lost in experimental glaucoma of nonhuman primates and human glaucoma. In support of these observations, another study showed a selective loss of anterograde axonal transport to the magnocellular layer of the dorsal lateral geniculate nucleus, which is the target area for the largest RGCs. A caveat of these early studies is that the conclusions were based only on size comparisons between average cell diameters in unaffected and glaucomatous eyes. A potential confounder of this phenomenon is that damaged RGCs atrophy prior to succumbing to apoptosis. Two different studies compared midget and parasol cell soma size and dendritic features to determine if the larger parasol cells were selectively destroyed. These studies found that both RGC populations underwent shrinkage and loss in approximately the same proportions. In addition to this, a study by Jakobs et al used several labeling methods to identify different subtypes of RGCs in the DBA/2J mouse retina and found that the loss of RGCs was not limited to any particular subtype. A more complete discussion of the process of cell shrinkage is made in a following section.

Box 27.2

Rules of apoptosis

  • Changes in gene expression are an important early event in apoptosis. These changes, which are marked by the downregulation of normal ganglion cell gene expression, take place before the committed stage of apoptosis

  • Loss of retinal ganglion cells in mouse models of glaucoma occurs via the intrinsic apoptotic pathway, which must proceed through a stage of mitochondrial dysfunction

  • BH3-only proteins mediate the activation of the proapoptotic protein BAX. BAX plays a critical role in mitochondrial dysfunction during apoptosis

  • Activation of the caspase cascade enables the cell to autodigest itself and complete the apoptotic process

Intrinsic versus extrinsic apoptosis

Apoptosis occurs via one of two major pathways. The first pathway occurs when the cell senses intracellular stress, hence the name intrinsic apoptosis. The intrinsic pathway relies on control by the members of the Bcl2 gene family and involves deregulation of mitochondrial function leading to activation of a cascade of proteases called caspases, initially triggered by caspase-9. In contrast, the extrinsic pathway is initiated by cell surface signaling following the binding of an extracellular ligand to a “death receptor.” This signal leads directly to the caspase cascade via activation of caspase-8, without involvement from the mitochondria. Although the two pathways can operate independently, there is some crossover. For example, caspase-8 can process the Bcl2 family member, Bid, into its active form, tBid, causing amplification of the cell death process by activation of the intrinsic pathway. Like the majority of neurons, it is believed that RGCs die using the intrinsic pathway of apoptosis ( Figure 27.2 ). The main evidence for this comes from Bax knockout mice. The prevalent theory of Bax function in an apoptotic cell is the creation of a pore in the mitochondrial outer membrane by oligomerized BAX proteins. This pore allows the release of small molecules, such as cytochrome c, that are critical for downstream events in the apoptotic pathway. The absence of Bax function prevents the loss of RGCs in both acute and chronic models of optic nerve damage.

Aug 26, 2019 | Posted by in OPHTHALMOLOGY | Comments Off on Retinal ganglion cell death in glaucoma

Full access? Get Clinical Tree

Get Clinical Tree app for offline access