Pathophysiology of Ganglion Cell Death and Optic Nerve Degeneration Workshop [NEI Strategic Planning]

Pathophysiology of Ganglion Cell Death and Optic Nerve Degeneration Workshop

November 12-13, 2002

On November 12-13, 2002, the National Eye Institute sponsored the Pathophysiology of Ganglion Cells and Optic Nerve Degeneration Workshop at the Airlie Conference Center, Warrenton, VA. The purpose of the workshop was to hold a dialogue on glaucoma-in the context of retinal ganglion cell compromise-in order to identify research needs and opportunities. The specific goals of the meeting were: to review the state of knowledge of neurodegeneration as it applies to glaucoma, to identify common mechanisms and technologies from other areas of neurodegeneration and neuroprotection that could be applied to research on glaucoma, and to summarize the workshop discussions and post the collective ideas on the NEI Web Site.

Scientists and clinicians, representing a number of areas of medicine, were invited to speak on their area of expertise and brainstorm about the needs and opportunities as they pertain to retinal ganglion cell research. Even though the focus was glaucoma, it was recognized that this disease can productively be considered as part of the broad scope of other neurodegenerative diseases. This was emphasized in the open discussion session in which many of the remarks focused on the need for a broader, multi-disciplinary effort.

The workshop is part of NEI’s Phase II Program Planning. However, the workshop is not inclusive of all the important basic and clinical research being conducted on glaucoma. Future meetings will incorporate additional topics and will reflect the evolving research as the scientific community works to enhance understanding of a serious disease that affects over 3 million Americans and millions more world-wide.

The report that follows summarizes presentations by the participants on: (1) retinal ganglion cell biology and physiology, (2) current hypotheses of glaucoma pathophysiology, including the role of vascular compromise, immune dysfunction, and glutamate excitotoxicity, (3) animal models of glaucoma and other chronic neurodegenerative disorders, (4) common mechanisms between glaucoma and other neurodegenerative conditions, and (5) endogenous and pharmacological molecules that promote axon survival and regeneration.

The report contains the following sections:

Background Information on Glaucoma

Major Topics of Discussion by Workshop Participants

A Summary of Needs and Opportunities

Workshop Agenda

List of Workshop Participants

Presentation Summaries

Section I: Current Hypotheses of Glaucoma Pathophysiology

What is Glaucoma? What is the Role of Intraocular Pressure in the Etiology of Glaucoma?
What is the Vascular Component of Glaucoma?
What Damages the Optic Nerve Head in Glaucoma?

Section II: Animal Models Contribute to the Study of Glaucoma

Primate and Rat Models of Ocular Hypertension
Mouse Models of Inherited Optic Nerve and Ganglion Cell Degeneration

Section III: The Biology of the Inner Retina is affected in Glaucoma

Retinal Ganglion Cell Interactions and Survival
Retinal Ganglion Cell Development and Neurodegenerative Events
Glutamate and Synaptic Transmission

Section IV: Mechanisms of Chronic Neurodegeneration Relevant to Glaucoma Etiology

Amyotrophic Lateral Sclerosis
Parkinson’s Disease and Stroke

Section V: Degeneration and Regeneration

Growth Factor Deprivation and Neuronal Cell Death
Factors Promoting Retinal Ganglion Cell Survival and Axon Regeneration
The Optic Nerve as a Model of CNS White Matter Damage

Section VI: The Microenvironment in the Optic Nerve Head

Role of Nitric Oxide in Retinal Ganglion Cell Degeneration


Background Information on Glaucoma

Glaucoma is a degenerative disorder of the optic nerve. It is characterized by a loss of ganglion cells and their axons. There are several types of glaucoma (e.g., open-angle, normal tension, and early onset-to name a few); however, all of these types have the same common feature, ganglion cell loss.

Millions of people have glaucoma. In the United States, alone, there are approximately 3 million people with the disease, of which about 120,000 are now blind. Most of these cases are attributed to primary open angle glaucoma (POAG), an age-related form of the disease. In addition, glaucoma is the number one cause of blindness in African-Americans. Blindness from all forms of glaucoma is estimated in excess of $1.5 billion annually in U. S. Social Security benefits and health care expenditures.

Elevated intraocular pressure (IOP) is frequently but not always associated with glaucoma. Because some people exhibit no association between increased IOP and glaucoma, it is thought that other mechanisms probably play a role in this blinding disease.

For patients who present with high IOP, treatment may involve topical drugs or surgical intervention to lower pressure. New approaches to treatment are essential, because not all patients respond to current treatments, and vision that is lost cannot currently be restored. Since ganglion cell death is the common feature of glaucoma, it has become increasingly important to understand the fundamental pathophysiology of the retina. This report is intended to stimulate the development of new paradigms and neuroprotective methodologies for the clinical treatment of glaucoma.


Major Topics of Discussion by Workshop Participants

Neurons exist in a highly complex relationship with other cells and with the extracellular environment. Studying the neurons of the normal retina and optic nerve will help identify key factors that are responsible for the health of CNS neurons. Trophic factor responsiveness, cell receptors, and influences from other cells like amacrine and Muller cells, are important areas to explore, as well as the effect of age on retinal ganglion cell numbers and robustness.

Studying the pathophysiology of glaucoma requires a multidisciplinary approach, including research on cell and retinal tissue structure, vulnerability of specific retinal cells or cell components, trophic and mechanical factors and cell signaling, synaptic transmission, and apoptosis.

The optic nerve head is unmyelinated and has different functional and mechanical properties from other regions of the optic nerve. Knowing the trophic factors, cell types, cell interactions, receptors, channels, and additional principles that govern the optic nerve head are undoubtedly important for understanding glaucoma. Studies on the optic nerve in rats, mice, and nonhuman primates are necessary, but it is important that these projects are not purely descriptive.

Evidence suggests that axons of retinal ganglion cells destruct in a unique way, independent of the soma. Studying the vulnerability of the axon, especially in the brief unmyelinated section of the optic nerve head, will address whether glaucoma is a problem of the axon or soma. It would appear that injury to axons occurs first and that events in the soma follow. Experiments should look at the order of events in retinal ganglion cell death.

Altered synaptic transmission and glutamate excitotoxicity are areas of intense study. There is debate on whether excess glutamate has a positive or negative influence on cells. It is possible that there are different sub-classes of retinal ganglion cells that respond differently to glutamate.

Description of the death pathway for retinal ganglion cells will determine whether a truly apoptotic mechanism is involved. Nerve cell death may have both caspase-dependent and -independent components. Understanding this cascade will be important for developing strategies for identifying neuroprotective molecules.

Animal models (genetic and experimentally-induced) help in understanding how retinal ganglion cells die and how insults, such as intraocular pressure changes, affect the retina. The number of axons of retinal ganglion cells in the optic nerve among rodent strains varies widely; special care must be taken to select controls and to use adequate numbers of animals.

Therapeutic intervention relies heavily on fundamental observations that characterize the pathophysiology of the specific disease. However, there are general principles among similarly-related disorders that hold promise for a common therapeutic approach. Neuroprotective molecules being studied in amyotrophic lateral sclerosis, Parkinson’s disease, and stroke are prime candidates for testing in glaucoma. Already, there are numerous drugs being screened for neurological disorders that could be tested in cell and animal models of glaucoma.


A Summary of Needs and Opportunities

  • Define mechanisms of survival and death in the retinal ganglion cell axon and soma, in vivo and in vitro
  • Define distinct subsets of ganglion cells, their unique roles, and the possible selective vulnerability of these subtypes in glaucoma
  • Enhance understanding of glial cells in the retina and optic nerve. Discover their response to pressure and their role in degeneration of the retinal ganglion cells
  • Enhance understanding of the blood-brain barrier at the optic nerve head. Understand the effects on retinal ganglion cells and glial cells of exposure to serum proteins that might leak from a defective blood-brain barrier
  • Develop animal models and standard methodologies to dissect pathophysiology, to identify novel and effective therapeutic agents, and to share with the research community
  • Hold additional meetings to address needs and opportunities for glaucoma and for other important diseases of the eye.


Section I: Current Hypotheses of Glaucoma Pathophysiology

What is Glaucoma? What is the Role of Intraocular Pressure in the Etiology of Glaucoma?
(Don Zack, Johns Hopkins University School of Medicine)

Glaucoma is not a single disease but rather a heterogenous group of disorders which share in common a slow progressive loss of ganglion cells and their axons of the optic nerve that resulting in a distinct pattern of visual loss. The major types of glaucoma are open-angle, normal tension glaucoma, glaucoma secondary to other medical conditions, and glaucoma with onset in early childhood.

Clinically, increased intraocular pressure (IOP) has been at the center of diagnosis, and epidemiological data have supported the contention that IOP is important. IOP is controlled by a balance between the secretion and drainage of aqueous humor. The aqueous humor is secreted by the ciliary body and follows a course from behind the iris to in front of it (carrying nutrients to the iris, lens, and cornea) and then out of the eye into the venous circulation via a trabecular network and the canal of Schlemm. Traditionally, increased IOP has been thought to be due to excess fluid in the eye that was most likely caused by a decrease in its outflow.

The relationship between IOP and a clinical diagnosis of glaucoma notwithstanding, many people have glaucomatous optic neuropathy with normal IOP, and conversely, many people who have elevated intraocular pressure do not necessarily develop optic nerve damage. For these reasons, it is recognized that other mechanisms must work in conjunction or in addition to IOP to cause optic nerve degeneration.

The characteristic anatomical change in the optic nerve in glaucoma is a cupping of the optic disk where ganglion cell axons have been lost. The death of the axons is associated with a loss of ganglion cell bodies in the retina and ganglion cell axon terminals in the dorsal lateral geniculate body. Unfortunately, a problem with vision is often not detected until it is quite advanced. Thirty to fifty percent of ganglion cells must be lost before their absence is detected in a visual field test and, even then, a patient is unlikely to realize that a problem exists.

Research into the pathophysiology of glaucoma has been based on the clinical manifestations of the disease and with the use of the limited animal models available. Clues have come from the specific pattern of optic nerve degeneration observed using visual fields. Results suggest that not all axons are equally susceptible to IOP, and there is evidence that increased IOP may affect even side-by-side ganglion cell fibers differentially. Prominent among the theories to explain the unique patterning observed clinically is one that invokes a mechanical compression due to increased IOP. Others include vascular compromise, neurotropic withdrawal, glutamate neurotoxicity, a nitric oxide effect, and immune dysfunction.

To summarize, the belief until recently has been that the damage in glaucoma is at the level of the optic disk and that it is frequently caused by an increase in intraocular pressure. It is true that lowering IOP often slows the progress of optic nerve disease and that gross damage occurs at the level of the optic disk, but there is reason to believe that other mechanisms are at work and that by uncovering these mechanisms the eyesight of millions of people may be preserved.

Facts about the optic nerve:

  • The human optic nerve contains approximately one million myelinated fibers (axons)
  • The axons arise from ganglion cells of the retina, which converge at the optic disc and run through the optic nerve to synapse in the lateral geniculate nucleus of the brain
  • The optic disc (also known as the optic nerve head) is approximately 1.5 millimeters in diameter. The convergence of the axons forms a depression in the disc, known as the cup. The size of the cup increases as axons degenerate. A cup-to-disc (C/D) ratio is used to describe the severity of glaucoma
  • Just before entering the optic nerve, the ganglion cell axons in bundles, unmyelinated at this point, traverse the lamina cribrosa, a thin sheet of connective tissue.


Current Hypotheses of Glaucoma Pathophysiology

What is the Vascular Component of Glaucoma?
(Alan Laties, University of Pennsylvania)

The retina is dependent on its blood supply for meeting its high metabolic needs. The pathophysiology and loss of axons of ganglion cells in glaucoma may be related to a compromised blood supply of the eye. Evidence supporting a vascular role in glaucoma pathophysiology is the presence of hemorrhaging sometimes seen in patients with optic nerve atrophy, which can take the form of flame-shaped hemorrhages on or near the disc. In addition, anatomical constraints imposed on the optic nerve head suggest a contribution of the vasculature to optic nerve pathology.

Two vascular systems contribute to the nutrition of the retina-the retinal artery and the ciliary arteries. The central retinal artery of the ophthalmic artery enters the optic nerve about 1.25 centimeters behind the eyeball and runs parallel to the nerve fibers. At the optic disc, it branches into superior and inferior branches, which subdivide into nasal and temporal arteries. These form the vascular network of the inner layers of the retina and supply portions of the nerve fiber layer of the optic nerve head.

The small ciliary arteries, 10-20 of them, penetrate the sclera at the posterior pole of the eye to form the choroidal capillaries, a dense single-layered vascular network. The lamina cribrosa at the optic nerve head is supplied by branches of the short posterior ciliary arteries.

The optic nerve head is subject to several distinctive pathologies that stem from its anatomy. Tightly fitted into a narrow scleral canal over most of its extent, the anterior-most part of the optic nerve has dual responsibilities: it must provide safe conduct from the eye for the unmyelinated axons of ganglion cells and simultaneously provide space for the central retinal blood vessels to enter the eye. To complicate matters, two distinct pressure domains exist, one to either side of a sieve-like membrane in structural continuity with the sclera, the lamina cribrosa. In separating two pressure domains, pore size of the lamina cribrosa is crucial. Pores must be large enough to permit unimpeded egress of groups of nerve fiber bundles, yet, be small enough to fulfill the task of containing the pressure of the eye. To the degree the neural structures within it exceed a critical diameter, the hydrostatic pressure differential across the lamina cribrosa could impose destructive shear forces on them. Moreover, the basket-like, perforated form of the lamina cribrosa dictates that it be more compliant than the surrounding sclera. Yet coaxial translation, again, risks injury to nerve fiber bundles in transit.

Raised intraocular pressure, altered ocular perfusion, loss of autoregulation, and characteristics of the blood-brain barrier at the optic disc are also possible candidates in the pathogenesis of glaucoma. The vascular tone of the blood vessels of the retina and optic nerve head are autoregulated and depend on contraction of the smooth muscle cells of the vascular endothelium. The smooth muscle cells are controlled by a variety of influences including neurotransmitters, hormones, myogenic and metabolic factors (for example, PO2, PCO2), and endothelium-derived factors that include the powerful vasodilator nitric oxide.

Retinal vessels typically contain tight junctions that form an impermeable seal between cells. An unusual feature of the optic disc is a small zone where there is no blood-brain barrier. When fluorescein is injected for angiographic studies, it is possible to detect its flow in the extracellular space of the optic nerve head.

One of the impediments to identifying a role for vasculature in the pathophysiology of glaucoma is that the flow of blood in the eye is difficult to assess clinically and experimentally. Techniques include: scanning laser fluorescein angiography with image analysis, color doppler imaging, laser doppler velocimetry, scanning laser doppler flowmetry, and pulsatile ocular blood flow systems. Animal studies indicate that the blood flow of the optic nerve head responds to a change in IOP. Studies also indicate a response in oxidative metabolism to varied IOP and mean arterial blood pressure, and that low blood pressure leads to loss of autoregulation and ischemia of the optic nerve head.

To conclude, anomalies or abnormalities in the blood flow of the retina and optic disc may be affecting the integrity of retinal ganglion cell axons. Blood vessel compression, loss of autoregulation, or leaking blood vessels may be involved. Just like their cell bodies, the axons of ganglion cells need a vigorous blood supply for normal metabolic activity. Any blood vessel dysfunction at the optic disc could cause axonal malfunction and atrophy.


Current Hypotheses of Glaucoma Pathophysiology

What Damages the Optic Nerve Head in Glaucoma?
(M. Rosario Hernandez, Washington University in St. Louis)

Primary open angle glaucoma is an optic neuropathy characterized by the loss of axons of the retinal ganglion cell fibers at the level of the lamina cribrosa in the optic nerve head resulting in the clinical manifestation of cupping. Astrocytes are the major glial cell type in the nonmyelinated optic nerve head in most mammalian species providing cellular and metabolic support to the axons and forming the interface between connective tissue surfaces and surrounding blood vessels. As such, it is of great interest to determine if they play a role in the pathophysiology of the disease.

In the lamina cribrosa, astrocytes line lamellar connective tissue plates that run horizontal and perpendicular to the axons of the retinal ganglion cells. Cellular components of the cribriform plates include: astrocytes, lamina cribrosa cells, microglia and smooth muscle cells, and endothelial cells of blood vessels. Components of the extracellular matrix include: elastic fibers, collagen fibers, microfibrils, proteoglycans and glycoproteins. The axons run through the lamina cribrosa in bundles. In humans and non-human primates there is a normal gradient in hydrostatic pressure between inside the eye and the retrolaminar optic nerve across the optic nerve head. When intraocular pressure increases above physiological levels, the pressure gradient also increases submitting the lamina cribrosa and the retinal ganglion cell axons to deformation and mechanical stress. In glaucoma, cupping of the optic disc and compression, stretching and rearrangement of the lamina cribrosa occurs in response to elevated IOP.

Astrocytes react vigorously to injury and stress. In glaucoma as with other CNS injuries, reactive astrocytes change shape, migrate out into the extracellular matrix into the nerve bundle area, express new proteins, and are involved in massive destruction and remodeling of the extracellular matrix. The expression of various collagen mRNAs and glycoproteins by reactive astrocytes is increased and so is de novo expression of elastin mRNA. Induction of expression of elastin leads to the formation of pathological irregular shaped elastotic fibers. Changes in the extracellular matrix and in the elastic component of the lamina cribrosa leads to loss of compliance and resiliency needed to adapt to changes in intraocular pressure even within the normal range. Matrix metalloproteinases (MMPs) are selectively expressed in reactive astrocytes in glaucoma and may be key to the progress of the remodeling of the extracellular matrix. MMPs may play a role in the transition of a quiescent astrocyte to the reactive phenotype by activating growth factors and disrupting the attachment of the cells to their substrate allowing changes in shape and migration. Interestingly, expression of MMPs was not increased in reactive astrocytes of the optic nerve head of monkeys with loss of retinal ganglion cells after optic nerve transection compared to a monkey model of laser induced glaucoma.

Despite the obvious involvement of astrocytes of the optic nerve in glaucoma, it is not clear what their role is in retinal ganglion cell death. Normal optic nerve astrocytes in culture respond to pressure-related stress adopting many features of the reactive phenotype. Reactive astrocytes cause large amounts of oxidative damage by synthesizing enzymes that lead to reactive oxygen species. The intermediate filament cytoskeleton reorganizes around the center of the cell to protect the nucleus and the cell’s synthetic machinery, whereas the actin and microtubules prepare the cell to migrate. Mechanical stress increases astrocyte motility and decreases cell adhesion, both characteristic of reactive astrocytes. Recently, using oligonucleotide microarray technology, astrocytes grown from human glaucomatous optic nerve heads were compared with normal, age-matched cells. Many interesting genes and pathways appeared differentially expressed in glaucomatous astrocytes including enzymes involved in steroid metabolism such as 3-alpha-hydoxysteroid dehydrogenases and the synthesis of a variety of growth factors, cytokines, and receptors. Glaucomatous astrocytes exhibited many functional and biochemical characteristics of reactive astrocytes.

Most of the work in the field of glaucoma has focused on the axons of the retinal ganglion cells. It is possible that the astrocytes (activated by elevated intraocular pressure or by other mechanisms) alter the environment of the axons and producing a milieu that may cause axonal degeneration or that may prevent survival of healthy retinal ganglion cells. Future work might suggest mechanisms for altering astrocytes or the environment to protect ganglion cells.


Section II: Animal Models Contribute to the Study of Glaucoma

Primate and Rat Models of Ocular Hypertension
(John C. Morrison, Casey Eye Institute)

Two compelling reasons exist for developing animal models in glaucoma. One is to understand the mechanism of pressure-induced optic nerve damage. The other is to develop therapies for protecting the optic nerve in glaucoma. An effective model should allow for the obstruction of aqueous humor outflow in order to raise IOP and for monitoring of intraocular pressure.

Nonhuman primates have been the animal of choice for studying glaucoma. Monkeys have been shown to develop glaucomatous changes in the optic nerve when their intraocular pressure is elevated experimentally. On histological examination, ganglion cell nerve fiber degeneration and debris can be seen in the extracellular matrix of the optic nerve. Monkeys can be trained to perform visual field tests, and those with experimental glaucoma show deficits indicative of glaucoma.

Because nonhuman primates are costly to acquire and raise, dangerous to handle for testing of IOP, and of uncertain age if they are caught in the wild, a rat model would be preferable. Validation of the rat model is facilitated by the following observations: 1) Frequent IOP measurements using a tonometer are easy to obtain in rats. 2) The composition of the rat optic nerve head is similar to that of the primate, including the connective tissue and cellular (astrocyte) support structures of the optic nerve axon bundles, and 3) Some etiological manifestations such as extracellular matrix material deposition, fluctuations in IOP, cupping, and ganglion cell axon and soma injury are similar to those seen in glaucoma patients. The fact that rats are inexpensive and easy to handle, yield ample tissue for analysis, and through breeding, exhibit less genetic variability are real advantages. In the near future, IOP monitoring of small mammals may be made easier through a transducer chip that would sit on the animal’s sclera or through an indwelling catheter in the anterior chamber of the eye.

A number of techniques for inducing increases in intraocular pressure in the rat are available. One that we have used successfully involves sclerosing the trabecular meshwork using hypertonic saline. By cannulating the veins and injecting the corrosive saline retrogradely into the canal of Schlemm, it is possible to scar and close up the trabecular meshwork. Intraocular pressure elevates in approximately a week and lesions begin to develop. The lesions seen in the rat optic nerve following this procedure includes a decrease in axon density and an increase in vesicular bodies and vacuoles that is most prominent in the superior region of the optic nerve. The patterning of damage seen in this model lends itself to developing hypotheses related to damage seen in human disease.


Animal Models Contribute to the Study of Glaucoma

Mouse Models of Inherited Optic Nerve and Ganglion Cell Degeneration
(Simon John, The Jackson Laboratory, Howard Hughes Medical Institute)

Inherited models of optic nerve and ganglion cell degeneration are another way to study glaucomatous changes to cell bodies and axons of the eye. Several mouse models-some in which the genetic mutation occurs spontaneously and others in which the mutation is induced-show progressive and gradual onset of disease. Two mouse models that develop spontaneous glaucoma characterized by progressive IOP elevation, RGC death, and optic nerve cupping (DBA/2J and AKXD-28/Ty) have been very valuable models for dissecting pathways of cell death in spontaneous glaucoma. The difference between the more progressive milder onset of pressure insult in these mice and the more sudden insult in experimentally-induced models may be very important. Since exposure to milder insults affects a cell’s susceptibility to subsequent more severe insults, there may be meaningful differences in the relative importance of specific death mechanisms/pathways between spontaneous and induced diseases.

Inbred mice allow researchers to study larger numbers of genotypically identical animals in highly controlled environments. The DBA/2J mouse has an iris disorder where pigment from the iris washes into the aqueous humor and plugs the drainage system of the eye. IOP rises and, after a few months, the animals’ retinal ganglion cells and axons die. The optic disk in this mouse strain shows cupping and retinal ganglion cell bodies appear apoptotic. The phenotype is suppressed in a variant of the DBA/2J mouse, which makes no iris pigment. This variant develops no pressure increase or glaucomatous lesions, which argues strongly in favor of increased IOP as a cause of glaucoma in the DBA/2J model. The AKXD-28/Ty is so susceptible to pressure changes in the eye that its entire retina is affected by increased intraocular pressure. This mouse model could be useful for detecting modifier genes that regulate the influence of pressure.

The ability to produce transgenic and gene-targeted mice, and the large number of existing mouse mutants, offers a unique opportunity to study the contribution of various factors to retinal ganglion cell degeneration. As unpublished data support involvement of an axonal degeneration pathway, mice with mutations in various genes that affect specific cell death pathways in the soma and axon will be invaluable in understanding glaucomatous neurodegeneration.

In summary, genetic strategies have proven very effective when studying variable, asynchronous disease processes such as ganglion cell death in glaucoma. In fact, it is likely that genetic experiments will be essential for understanding the triggers of complex asynchronous neurodegenerati