A biomechanical paradigm for axonal insult within the optic nerve head in aging and glaucoma

Abstract

This article is dedicated to Rosario Hernandez for her warm support of my own work and her genuine enthusiasm for the work of her colleagues throughout her career. I first met Rosario as a research fellow in Harry Quigley's laboratory between 1991 and 1993. Along with Harry, John Morrison, Elaine Johnson, Abe Clark, Colm O'Brien and many others, Rosario's work has provided lamina cribrosa astrocyte cellular mechanisms that are biomechanically plausible and in so doing provided credibility to early notions of the optic nerve head (ONH) as a biomechanical structure. We owe a large intellectual debt to Rosario for her dogged persistence in the characterization of the ONH astrocyte and lamina cribrosacyte in age and disease. Two questions run through her work and remain of central importance today. First, how do astrocytes respond to and alter the biomechanical environment of the ONH and the physiologic stresses created therein? Second, how do these physiologic demands on the astrocyte influence their ability to deliver the support to retinal ganglion cell axon transport and flow against the translaminar pressure gradient? The purpose of this article is to summarize what is known about the biomechanical determinants of retinal ganglion cell axon physiology within the ONH in the optic neuropathy of aging and Glaucoma. My goal is to provide a biomechanical framework for this discussion. This framework assumes that the ONH astrocytes and glia fundamentally support and influence both the lamina cribrosa extracellular matrix and retinal ganglion cell axon physiology. Rosario Hernandez was one of the first investigators to recognize the implications of this unique circumstance. Many of the ideas contained herein have been initially presented within or derived from her work (Hernandez, M.R., 2000. The optic nerve head in glaucoma: role of astrocytes in tissue remodeling. Prog Retin Eye Res. 19, 297-321.; Hernandez, M.R., Pena, J.D., 1997. The optic nerve head in glaucomatous optic neuropathy. Arch Ophthalmol. 115, 389-395.).

Figure 1. Glaucoma, cupping and axonal insult within the optic nerve head (ONH)

The ONH is made up of prelaminar, laminar and retrolaminar regions (A). Within the clinically visible surface of the Normal ONH (referred to as the optic disc) (B), central retinal vessels enter the eye and RGC axons appear pink due to their capillaries (which are principally supplied by branches from the posterior ciliary arteries (PCA) in (C). The primary site of RGC axon insult in glaucoma is within the lamina cribrosa (schematically depicted with axon bundles in (D), isolated by trypsin digest in a scanning electron micrograph in (E) and drawn with stippled extracellular matrix (ECM), central capillary (red) and surrounding astrocytes (yellow with basement membranes in black) (F). Blood flow within the ONH, while controlled by autoregulation, can be affected by non IOP-related effects such as systemic blood pressure fluctuation and vasospasm within the retrobulbar portion of the PCAs. Additional IOP-induced effects may include compression of PCA branches within the peripapillary sclera (due to scleral stress and strain) and compression of laminar beam capillaries reducing laminar capillary volume flow (C and F)(Langham, 1980). There is no direct blood supply to the axons within the laminar region. Axonal nutrition within the lamina (F) requires diffusion of nutrients from the laminar capillaries, across the endothelial and pericyte basement membranes, through the ECM of the laminar beam, across the basement membranes of the astrocytes, into the astrocytes, and across their processes to the adjacent axons (vertical lines). Chronic age-related changes in the endothelial cell and astrocyte basement membranes, as well as IOP-induced changes in the laminar ECM and astrocyte basement membranes may diminish nutrient diffusion to the axons in the presence of a stable level of laminar capillary volume flow. The clinical manifestation of IOP-induced ONH structural change is most commonly “deep cupping” (G) but in some eyes cupping can be shallower accompanied by pallor (H). Z-H = circle of Zinn-Haller; PCA= posterior ciliary arteries; NFL = nerve fiber layer; PLC = prelaminar region; LC = lamina cribrosa; RLC = retrolaminar region; ON = optic nerve; CRA = central retinal artery. (A) Reprinted with permission from Arch Ophthalmol (Anderson, 1969); (C) reprinted with permission from The Glaucomas. St. Louis: Mosby; 1996:177–97 (Cioffi and Van Buskirk, 1996); (D) reprinted with permission from Optic Nerve in Glaucoma. Amsterdam: Kugler Publications; 1995:15–36 (Quigley, 1995b); (E) reprinted with permission from Arch Ophthalmol (Quigley et al., 1990); (F) reprinted with permission from Arch Ophthalmol (Morrison et al., 1989b)

Figure 2. Principle distribution of forces, pressures and the translaminar pressure gradient within the optic nerve head (ONH)

A. Cut-away diagram of intraocular pressure (IOP)-induced mechanical stress in an idealized spherical scleral shell with a circular scleral canal spanned by a more compliant lamina cribrosa. In this case, the majority of the stress generated by IOP/orbital pressure difference (red arrows on the inner surface of the sclera) is transferred into a hoop stress borne within the thickness of the sclera and lamina (blue arrows) that is concentrated circumferentially around the scleral canal (green arrows). B. Note that the pressure behind the lamina is not simply CSF pressure but is retrolaminar tissue pressure (RLTP) which has been demonstrated to be approximately 0.82 × CSF + 2.9 mm Hg by Morgan, et al in dogs (Morgan et al., 1998) C. The difference between IOP and the retrolaminar tissue pressure is the translaminar pressure difference which generates both a net posterior (outward) force on the surface of the lamina (the red arrows over the lamina) and a hydrostatic pressure gradient (the translaminar pressure gradient - schematically shown in green) within the neural and connective tissues of the pre-laminar and laminar regions. Note that the in-plane hoop stress transferred to the lamina from the sclera is much larger than the stresses induced by the translaminar pressure difference. CSF directly influences laminar position through its effect on the translaminar pressure difference. CSF may also effect scleral flange position within the region it projects to the sclera (Figure 2), but in most eyes, because the projection of the CSF space is minimal this is not likely important (the CSF space within Panels B and C in this figure is greatly expanded due to perfusion fixation). IOP has a similar direct effect on laminar position, but has an additional (and potentially more important) effect on laminar position through the peripapillary sclera. However, while the magnitude of the translaminar pressure difference may be small relative to the stresses within the sclera and lamina, the axons experience it as the translaminar pressure gradient the steepness of which is influenced by the thickness of the tissues over which it is experienced. The translaminar pressure gradient, as such, may serve as a primary barrier to axon transport and flow within this region and an important physiologic determinant for the ONH axons and cells. (Panel A is modified from Downs et al, 2009).

Figure 3. The volume flow of blood within the posterior ciliary arteries should be affected by IOP-related stress and strain within the peripapillary sclera and scleral flange

The posterior ciliary arteries pass through the peripapillary sclera (yellow, left and center panel) immediately adjacent to the scleral portion of the neural canal. We refer to this portion of the sclera as the scleral flange (yellow, middle panel) (Yang et al., 2005. Neural Canal and Peripapillary Scleral Alterations Within Three-Dimensional Reconstructions of Early Glaucoma Monkey Optic Nerve Heads. Invest. Ophthalmol. Vis. Sci. 46, ARVO Abstract# 3511). The sclera thins here to accommodate an expansion of the neural tissues that occurs in a highly eye-specific fashion. While the large penetrating vessels to the choroid are outside of the flange, the circle of Zinn-Haller and the penetrating branches that pass to the pre-laminar, laminar and retrolaminar nerve pass directly through these tissues and are therefore subject to the compressive effects of their contained mechanical stress and strain. Within the lamina cribrosa, there is no direct blood supply to the axons or the axon bundles. Each laminar beam contains a capillary (Panel F, Figure 1) which are here shown in a vascular casting of a monkey eye (Cioffi and Van Buskirk, 1996). (Left and right panels reprinted with permission from The Glaucomas. St. Louis: Mosby; 1996:177–97.(Cioffi and Van Buskirk, 1996)

Figure 4. Hayreh (Hayreh et al., 1970) demonstrated sensitivity of the peripapillary choroidal circulation (green) to acute IOP elevation in the monkey ONH

Fluorescence fundus angiogram of the right eye of a cynomologus monkey after experimental central retinal artery occlusion at normal (A) and 70 mm Hg IOP (B and C). The non perfused region of the peripapillary choroid is highlighted in green in (C). We and others have hypothesized that IOP-related stress and strain within the scleral flange (Figure 3) may contribute to this phenomenon and that similar effects may occur within the laminar capillary beds (red) (C and D). Panels A and B are reproduced with permission from the British Journal of Ophthalmology (Hayreh et al., 1970). (Panel D reprinted with permission from The Glaucomas. St. Louis: Mosby; 1996:177–97. (Cioffi and Van Buskirk, 1996)

Figure 5. Histologic sections from the superior scleral canal of both eyes of a normal monkey perfusion fixed with one eye at IOP 10 mm Hg (middle left and above) and one eye at IOP 45 mm Hg (middle right and below)

Higher magnification demonstrates patent prelaminar, laminar, and retrolaminar capillaries in the IOP 10 eye (middle left). However in the IOP 45 eye (middle right), the prelaminar and anterior laminar capillaries do not appear patent, with only the posterior laminar and retrolaminar capillaries open.

Figure 6. ONH neural and connective tissue changes at the onset of CSLT-detected ONH surface change

These findings were initially reported in n=3 monkeys (Yang et al., 2007a) and have recently been expanded to n=9 animals (Yang et al., 2010). Normal lamina cribrosa (unhatched), scleral flange (hatched), prelaminar tissue (beneath the internal limiting membrane - brown line), Bruch’s membrane (solid orange line), BMO zero reference plane (dotted orange line), Border tissue of Elschnig (purple line), choroid (black circles) are schematically represented in the upper illustration. Changes in EG are depicted below. We have previously reported the bowing of the lamina and peripapillary scleral flange and thickening of the lamina in these same EG eyes (grey shading) (Yang et al., 2007b). These connective tissue changes underlie posterior deformation of the ONH and peripapillary retinal surface (dotted brown to solid brown ILM) that occurs in the setting of thickening (arrows) not thinning of the prelaminar neural tissues (brown shading) in EG. EG eye expansion of all three volumetric parameters is depicted below. The interactions between these parameters are important. While expansion of the cup and deformation of the surface are clinically detectable at this early stage of the neuropathy, because they occur in the setting of prelaminar tissue thickening, (not thinning), we believe that expansion of Post-BMO Total Prelaminar Volume (due to ONH connective tissue deformation) drives these findings. Thus, cupping in these nine EG eyes is “laminar” in origin, without a significant “prelaminar” component (Yang et al., 2007a).

Figure 7. The range of ONH connective tissue deformation (cupping) at the onset of CSLT-detected ONH surface change in nine monkeys with unilateral experimental glaucoma (Yang et al., 2010)

Minimal (Monkey 1, left) to maximum deformation (Monkey 9, right) within the 9 animals is depicted. The following ONH landmarks are delineated within representative superior temporal (ST) to inferior nasal (IN) digital sections from the normal (upper panel) and early experimental glaucoma (EEG) (middle panel) eye of both animals: anterior scleral/laminar surface (white dots), posterior scleral/laminar surface (black dots), neural boundary (green dots), NCO reference plane (red line) and NCO centroid (vertical blue line). Delineated points for the normal (solid lines) and EEG eye of each eye are overlaid relative to the NCO centroid in the lower panel. For each animal, the area under the NCO reference plane and above the anterior laminar surface Post-NCO Total Prelaminar Volume is outlined in both the normal (darker green) and EEG (light blue) eye for qualitative comparison. The overlaid delineations for both eyes of each animal (lower panels) suggest that expansion of this area is due to posterior laminar deformation and neural canal expansion from its onset (Monkey 1) and that these two phenomena continue through more advanced stages of connective tissue deformation and remodeling (Monkey 9). As such,“glaucomatous” deformation of the ONH connective tissues is present early and progresses early in the neuropathy (Burgoyne and Downs, 2008; Yang et al., 2007a). These phenomena are accompanied by regional laminar thickening and outward bowing of the peripapillary sclera in both of these animals and most of the 9 EEG eyes. In addition, in both monkeys laminar migration to the point of partial pialization has occurred on the inferior side (right side of each schematic) of the canal (see Figure 8). (Yang et al., 2010. Optic Nerve Head Lamina Cribrosa Insertion Migration and Pialization in Early Non-Human Primate Experimental Glaucoma. Invest. Ophthalmol. Vis. Sci. 51, ARVO Abstract# 1631)

Figure 8. The pathophysiology of early experimental glaucomatous damage to the monkey ONH includes not only “thickening” but regional “migration” of the laminar insertion away from the sclera to the point that “complete pialization” of the laminar insertion is achieved in a subset of eyes

Neural canal landmarks (Red – Neural Canal Opening (end of Bruch’s Membrane); Blue – Anterior Scleral Canal Opening; Yellow – Anterior Laminar Insertion; Green – Posterior Laminar Insertion; Purple – Posterior Scleral Canal Opening) and segmented connective tissue (dark grey - lamina cribrosa; purple - peripapillary sclera; light green - pial sheath) within digital section images from the inferior region of the normal (top) and the contralateral early experimental glaucoma (bottom) ONH of a representative monkey. Note that in most normal monkey eyes, the lamina inserts into the sclera as is demonstrated in this monkey’s normal eye (top). However at an identical location in the early experimental glaucoma eye of this animal (bottom) in addition to the lamina being thickened and posteriorly deformed, the laminar insertion has migrated outward such that both the anterior and posterior lamina effectively insert into the pial sheath. While regions of laminar insertion into the pia have been reported in normal human eyes (Sigal et al., 2010a), these findings are the first to suggest that active remodeling of the laminar insertion from the sclera into the pia is part of the pathophysiology of “glaucomatous” ONH damage. This phenomenon when present has important implications for the mechanism of axonal insult within these regions. (These data were reported by Hongli Yang, PhD at the 2010 ARVO meeting and are currently in preparation for publication, Yang et al., 2010. Optic Nerve Head Lamina Cribrosa Insertion Migration and Pialization in Early Non-Human Primate Experimental Glaucoma. Invest. Ophthalmol. Vis. Sci. 51, ARVO Abstract# 1631).