Statistics from Altmetric.com
The course of ganglion cell axons from the retina to the optic nerve is predictable and has been known accurately for over a century. The arcuate arrangement of the axons temporally and the presence of the median raphé were first described by the anatomists Michel1 and Dogiel2 in the latter part of the 19th century. Vogt correctly deduced the arrangement of the superficial nerve fibres in the retina when, in 1913, he observed the retinal nerve fibre layer ophthalmoscopically with the aid of red-free light.3 In contrast, the anatomical orientation of retinal nerve fibres within the optic nerve head has been the subject of considerable debate. Before 1930, the common view was that nerve fibres from the peripheral retina entered the central portion of the optic nerve head.4 Loddoni established that the reverse was true: nerve fibres from the far retina entered the peripheral optic nerve head while more central fibres occupied a central position in the nerve.5 However, the anterior-posterior relation within the retinal nerve fibre layer remained unclear. The work of Wolff and Penman supported the view of the older literature that long retinal axons occupied a scleral position in the nerve fibre layer, while more central axons occupied a vitreal position.6 More recent work has supported this view.7 8 Ogden employed highly localised intraretinal injections of horseradish peroxidase in monkey eyes, and elegantly demonstrated that long retinal fibres actually occupy a more vitreal position in the retina and shorter retinal fibres are more scleral in location.9 Consequently, at the margin of the optic disc, short and long fibres decussate to match the topographic arrangements of the retinal nerve fibre layer and the anterior portion of the optic nerve head.
The work of Morgan and colleagues, which is reported in this issue of the BJO (p 680), provides new information about the anatomy of retinal ganglion cell axons as they course through the lamina cribrosa. In their careful study of two human eyes, they found that while the majority of axons took a direct course through the pores of the lamina cribrosa, approximately 10% deviated from this course and passed between adjacent cribrosal plates. Thus, a substantial minority of axons change trajectory within the lamina and take a more tangential (and perhaps more precarious) course. They found, with the retrograde tracer horseradish peroxidase applied to the freshly cut stump of the retrobulbar optic nerve, that this pattern was present in both central and peripheral portions of the optic nerve head. The existence of such a population of axons highlights the distinction between the apparently precise topographic arrangement of nerve bundles and the rather meandering course that can be taken by individual axons.10 This finding suggests the possibility that axons which deviate from a straight intralaminar course might be particularly susceptible to compressive effects because of their tangential trajectory between laminar plates. These deviant axons could thus represent a subgroup of retinal ganglion cells that are selectively damaged in early glaucoma. Other studies have suggested that the special anatomy of the lamina cribrosa region could make humans particularly susceptible to glaucoma. In the disc periphery, the lamina is thicker, and the course of nerve fibres is more curvilinear than at the disc centre.11 It has been suggested that the microscopic arrangement of fibre bundles within the lamina cribrosa could play an aetiological role in the pathogenesis of low tension glaucoma.12
Apoptosis appears to be a mechanism of retinal ganglion cell death in experimental and human glaucoma.13 14 The molecular triggers for apoptosis in glaucoma are unknown, but may include deprivation of neurotrophic factors, and excitotoxic injury initiated by ischaemia.15 16 If retrograde transport of neurotrophic factors from the central target tissue to the retinal ganglion cells is interrupted at the lamina cribrosa by compression or distortion, it seems reasonable to postulate an apoptotic neuronal cell death. This is a seductive theory, but there is little hard evidence to support this mechanism (or in fact any other mechanism) as a cause of glaucomatous damage. Alternatively, can collapse of laminar plates and distortion of the laminar pores occur secondary to loss of optic nerve fibres within these tissues? The inferior and superior poles of the optic nerve contain fibres that are predominantly damaged in early glaucoma. Are greater numbers of deviant axons located at the inferior and superior poles of the optic nerve head, where early glaucomatous damage selectively occurs? Do these deviant axons belong to any particular class of ganglion cells?
The interesting and important finding reported here by Morgan and colleagues will stimulate discussion about the selective vulnerability of optic nerve fibres. This kind of study is labour intensive, and the results from just two non-glaucomatous eyes have been reported. The authors’ findings support the concept that mechanical axon compression can cause glaucomatous optic nerve damage. However, if only 8–12% of axons show this aberrant trajectory, selective damage to these cells alone is not sufficient to cause glaucomatous visual field loss. Further work will undoubtedly show a great deal of variability among individuals with regard to the proportion of cells that show a deviant intralaminar trajectory. Are there greater numbers of these anomalous fibres in eyes that are susceptible to glaucomatous damage? Histological studies to answer this question may prove difficult to interpret: if fibres are selectively vulnerable, they may appear less frequently in damaged optic nerve heads, or not at all.
The embryological explanation for axon deviation within the optic nerve head becomes relevant in light of this report by Morgan and coworkers. It compels us to reconsider the issue of retinotopic projection of optic nerve fibres and the intermingling of axons from the superficial and deep fibres in the anterior optic nerve head. This intermingling apparently continues throughout the lamina cribrosa, perhaps in an attempt to achieve the correct topographic orientation in the retrolaminar portion of the optic nerve.
This report is a good example of the value of careful observation, even if in only a few specimens. It will undoubtedly stimulate consideration of whether deviant axons represent an anatomical subpopulation that is vulnerable to compressive effects in human glaucoma.