Nerve Fiber Layer Splaying at Vascular Crossings

Abstract

purpose. To examine the anatomic relationships of retinal blood vessels to the nerve fiber layer.

methods. Retinal flat preparations from monkey and cat were stained with methylene blue dye and examined by light microscopy.

results. A focal transition in nerve fiber bundle patterns occurs just above major retinal blood vessels. Bundles abandon a compact form and splay open just as they reach a major blood vessel. Maximum separation of axon fascicles occurs immediately above the blood vessel and the bundles reform after the crossing of the blood vessel is complete. In so doing, nerve fiber bundles temporarily thin and broaden, assuming the shape of an inverted hull. In contrast, at the retinal periphery, minor blood vessels and axon fascicles have no special relationship to each other.

conclusions. At major blood vessel crossings, a focal alteration in nerve fiber bundle anatomy takes place under which local deformability may be enhanced. This adaptation probably lessens the risk of injury to ganglion cell axons from vascular compression. In addition, deflection of the nerve fiber pathway obviates the need for major blood vessels to bend or kink to achieve a crossing, thus avoiding turbulent flow.

The nerve fiber layer of the retina has long been a focus of special interest, especially in relation to the pathologic events in glaucoma. ^{1 2 3} Depending on the research methods used, attention has largely been directed to the effects of an increase in intraocular pressure on either nerve fibers or blood vessels. In both cases, remarkable insights have been gained. For instance, the demonstration of a focal build-up of axoplasmic constituents at the optic nerve head defines a mechanical component in the pathologic course of acute glaucoma. ^{4 5} And the finding of diminished blood flow to the optic nerve head validates long-held speculation that inadequate vascular perfusion contributes to the pathologic progress of chronic simple glaucoma. ^{6 7}  

Our interest in the nerve fiber layer was sparked by two recent publications that graphically illustrate a downward traction on nerve fiber layer axons caused by aberrant blood vessels formed during the course of hereditary retinal degeneration in the RCS rat and the rd mouse. ^{8 9} In each case, nerve fiber axons were deformed and rendered dysfunctional, and ganglion cell death ensued. Even if these reports are based on an unusual pathologic sequence, they nevertheless raise questions about the normal relationship of retinal blood vessels to nerve fiber layer axons. Despite careful scholarly work on neurovascular relationships, the nature of this relationship has yet to be completely defined. The present report, based on a review of archival material originally produced for studies on topographic projections from retina to brain, highlights two neurovascular relationships present in cat and monkey. It documents differential stratigraphic patterns of nerve fiber layer axons in relation to the retinal region, as well as a remarkable splaying of axon bundles that overlie or cross large blood vessels. Considering its crowded nature and the conflicting tasks required within the nerve fiber layer, each appears to represent a favorable adaptation. 

Methods

The methylene blue staining technique used for the demonstration of nerve fiber layer pathways in retinal flat preparations has been previously described. ^{10 11} Experiments were performed in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and in compliance with National Institutes of Health Guidelines. In brief, one adult rhesus macaque monkey and two adult female cats were killed by overdose of intraperitoneally administered pentobarbital sodium. Enucleated eyes were bisected in a coronal plane and the vitreous gel removed. Two whole monkey and the four whole cat retinas were stained in situ supravitally by incubation of intact posterior segments in a staining solution consisting of 0.10 mL 1% methylene blue and 20 mL 0.9% saline at 37°C for 15 minutes. Thereafter, the retinas were carefully dissected in a saline bath. When after several more minutes the degree of staining was considered optimal, fixation with a 4% solution of ammonium molybdate for 1 hour was followed by dehydration in graded ethanols and mounting on glass slides, nerve fiber layer up. Peripheral relaxing incisions were made to achieve a low level of tissue distortion in the mounting procedure. In the flattened retinas the nerve fiber pathways can be readily discerned. Advantage can be taken of the greatly dye-enhanced birefringence property of methylene blue–stained nerve fiber layer axons by viewing the preparations in polarized light. 

Results

A schematic (Fig. 1) , drawn from methylene blue wholemounts, defines the nerve fiber architecture of the cat retina. Simple inspection discloses that the nerve fiber layer architecture of the cat is substantially similar to that of the primate. Although the central area of the cat retina is tilted upward by more than 20°, both retinas exhibit a temporal raphe. In both cat and primate, large nerve fiber bundles follow an arcuate path from the temporal raphe to the optic disc. Although at first the distribution of nerve fiber bundles and major blood vessels looks different from that of the primate, if in the mind’s eye one rotates the temporal raphe of the cat clockwise to meet the horizontal, the dissimilarity between the two largely disappears. 

The retinal flat preparations reveal distinct relationships of blood vessels to nerve fibers by region. At the periphery, the two show no precise radial stratification and act for practical purposes as independent operators. Blood vessels are of small diameter and proceed along somewhat variable paths. In many instances, ganglion cell axons can be traced from their origins until they join a fascicle. Singly and in small groups, ganglion cell axons can be observed as they begin their centripetal journey. When they encounter a blood vessel, they pass either just above or just below it in no apparent pattern. In fact, newly formed fascicles are seen on occasion successively to exhibit first one and then the other behavior (Fig. 2 , arrows). Farther along their paths, individual axons and fascicles ascend to lie immediately beneath the internal limiting membrane to form the definitive nerve fiber layer. In turn, they group together into large nerve fiber bundles. At this point, a definitive stratification has been established. Arteries and veins are uniformly situated beneath the nerve fiber layer. By their greater diameter, they indent both the ganglion cell and nerve fiber layers. Concomitantly, blood vessels and nerve fiber layer show an obvious interaction, as focal spreading of nerve fiber fascicles becomes readily apparent (Fig. 3) . The splaying phenomenon is enhanced as blood vessels attain a larger size. 

A dramatic increase in focal alteration of nerve fiber bundle structure occurs just above major blood vessels. The nerve fiber layer bundles thin and splay open widely (see right arrow, Fig. 4 , right arrows). Separation into their constituent fascicles begins as the nerve fibers ascend to pass over the blood vessel, and their separation is at a maximum at the midpoint of their traverse. 

However, axon fascicles can be seen to regroup as they descend to continue their path to the optic nerve. Once again, they join to form large, intact bundles. In short, at major vascular neural crossings, an open fan shape of separated nerve fiber layer constituents temporarily replaces the compact, tightly packed form of an intact nerve fiber bundle. The splaying phenomenon occurs above both artery and vein. It is most striking when major blood vessels issue large branches that run at sharp angles to nerve fiber bundles. The larger the blood vessel, the more extensive the upward deflection and the broader the temporary separation of nerve fiber bundle constituents. 

At times, given their generally parallel nature, large blood vessels and the nerve fiber layer run in parallel, and crossing is delayed. The nerve fibers simply overlie the blood vessel for a distance. When this occurs, nerve fiber splaying is again enhanced. In fact, with greater separation it becomes more difficult to visualize nerve fiber layer constituents (compare Figs. 3 and 4 ). At this point, it is not possible at the light microscopic level to differentiate single axons from fascicles. 

As well as can be judged from our preparations, it is always the nerve fiber layer that deflects at a crossing; major blood vessels are never seen to flex or detour around nerve fibers. The loss of tissue and disturbance of tissue relationships caused by the transection of nerve fibers and blood vessels at the optic nerve head required for tissue preparation of flatmounts precludes a careful description of neural vascular relationships in this region. 

Discussion

As yet, downward dragging with ligation of nerve fiber layer fascicles comparable to that described in the RCS rat and rd mouse has not been observed in humans with hereditary retinal degeneration. Although this may represent an inadequate search up to now for a comparable pathologic sequence, it could as well derive from a differential location of the major retinal blood vessels. For in the rodent, as depicted by Villegas-Perez et al. ⁸ and Wang et al., ⁹ large blood vessels appear to be superficial to the nerve fiber layer, whereas in cat, monkey, and human they are mainly deep to it. However, such an anatomic difference, even if true, does not rule out downward dragging, because capillaries would probably suffice. The closer the approach to the optic nerve head, the greater the thickness of the nerve fiber layer and, as noted by Snodderly et al., ¹² where the capillaries are more numerous, in the juxtapapillary region, a second planar level of capillaries appears within the thickened nerve fiber layer. With a clearly defined stratification, elements of this bilevel capillary network bestride major components of the nerve fiber layer. The intertwined neurovascular relationships of these small blood vessels would probably provide a suitable framework for downward dragging of nerve fiber layer fascicles, were radially oriented, newly formed blood vessels similar to those observed in diseased rodent eyes to come into contact with them. 

If nerve fiber splaying is accepted as a valid picture of local anatomic variation when nerve fiber bundles overlie or cross major blood vessels, questions arise as to genesis and consequences. It is well documented that ganglion cell axons proceed across the retina during retinal development, pursuing a centripetal course in the nerve fiber layer to join in the formation of the optic nerve. It is also known that a continuing central-to-peripheral sequence of retinal neuropil development takes place so that there is an overlap with vasculogenesis. Specifically, newly formed axons are still being produced well after major retinal blood vessels are in place. ^{13 14 15} Thus, the genesis of splaying of nerve fiber bundles is not so much a question of which comes first, blood vessel or axon bundle, as it is of local regulatory signals between the two. In recent years, both the behavior and the interactions of retinal axonal growth cones have become increasingly clarified. ^{15 16 17} As a result, it is now recognized that retinal axonal pathfinding is a response to a complex set of guidance molecules presented in some instances by cells, both neural and nonneural, and in others by extracellular matrix. ^{15 16 17}  

As regards focal splaying of nerve fiber bundles, the recent report by Tuttle et al. ¹⁷ describing region-specific axonal patterning later in the visual pathway is relevant. They investigated a comparable transition from compact retinal axonal bundle to focal splaying along the dorsal–ventral axis of the rat diencephalon, demonstrating a direct relation to an equally sharp transition in the level of axon-specific repulsive molecules. Where repulsive extracellular matrix molecules exhibited a higher intensity of immunostaining, axon bundles were compact and splaying is evident over regions of lower intensity of immunostaining. On this basis, the form of axon bundles in the retina, whether compact or splayed clearly represents a local variation in the complex chronotopical guidance sequence described by Thanos and Mey ¹⁶ that determines spatial distribution of ganglion cell axons in the x, y, and z axes. Whatever the regulatory influences, focal splaying of nerve fiber bundles is probably a favorable adaptation. Certainly, for retinal axons to deflect upward to pass over large blood vessels appears advantageous, given that any bend or kink in a blood vessel leads to turbulent flow. Because turbulent flow so induced would engender intimal disease, a deflection of nerve fiber around blood vessel rather than blood vessel around nerve fiber is physiologically preferable. 

Perhaps more important, focal separation of axon fascicles leading to a characteristic inverted fan shape (a convex hull in geometry) may be protective in a second way, because it favors both deformability and widened tissue distribution of kinetic energy. Assumption of an open, widely spaced deformable distribution at major vascular crossing points probably obviates the risk of vascular point compression of nerve fiber bundles. Although not presently recognized in the retina, there is increasing evidence of risk to the optic nerve, as well as to several cranial nerves from local compression by blood vessels. ^{18 19} In fact, since the concept of microvascular compression of cranial nerves was first proposed by Dandy ²⁰ it has gained increasing currency. It is currently thought in many instances to represent the essential pathologic cause of trigeminal and glossopharyngeal neuralgias. ²¹ By assuming at major vascular crossing points a readily deformable and displaceable form, retinal fiber bundles would be less subject to injury either from repetitive pulsatile forces deriving from the drumbeat of pulse waves or, more rarely, from sudden large increases in intraocular pressure, such as occur from blunt trauma to the eye. 

Within the eye, special considerations apply just around the regions of the optic nerve head. For here, as the central retinal artery over a length of a few millimeters enters the eye, breaks into its branched divisions, and begins its retinal path, its branches pass obliquely through crowded nerve fiber bundles to attain a final position deep to the nerve fiber layer. To the degree that the paths of blood vessels and nerve fiber bundles are not strictly parallel and that packing of both is tight in this region, a concept of vascular compression analogous to that observed for cranial nerves can be considered. On this basis, if for any reason microvascular compression were to occur, it would probably be selective for nerve fiber bundles at the upper and lower poles of the optic nerve head, especially at crossing points for those immediately beneath the path of branching blood vessels. Moreover, were compression to occur as major blood vessels pass obliquely through the nerve fiber layer to reach a position just below it, effects on vision would then be noted first from dysfunction of ganglion cells at the retinal periphery, in keeping with observations by Radius and Anderson ² that axons originating from the retinal periphery lie deep near the nerve head. 

The rapid shift in the form of the nerve fiber layer bundle at major vascular crossings has implications as well for imaging by scanning laser polarimetry. Their forward displacement, in combination with divergence as axon fascicles first separate to begin their ascent over a large blood vessel and their subsequent convergence as they descend on the other side clearly predicts a local disturbance in nerve fiber birefringence signal. This notion is confirmed in scanning laser polarimetric images taken over major retinal blood vessels. A focal deficit in mean retardation is commonly observed in this location. ²² With partial but incomplete correctness the focal disturbance in birefringence signal is attributed solely to thinning of the nerve fiber layer, ignoring the added effect of an obligatory focal alteration in orientation. 

 Supported by National Institutes of Health Grants EY10009-09 and EY13134-01, the Paul and Evanina Bell Mackall Foundation, and a Senior Scientist Award (AML) and an unrestricted grant from Research to Prevent Blindness.

 Submitted for publication November 5, 2001; revised January 29, 2002; accepted February 11, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

A schematic drawing illustrates the pathways followed by major blood vessels and ganglion cell axons in the cat retina. Adapted from Laties AM, Sprague JM. The projection of optic fibers to the visual centers in the cat. J Comp Neurol. 1966;127:1:35–70.

Figure 2.

 

In the peripheral retina of a macaque dye-enhanced birefringent axons cross above or below small blood vessels in no apparent pattern. Arrows: small fascicle that does both. Magnification, ×300.

Figure 3.

 

Nerve fiber fascicles fan out as they overlie a medium sized blood vessel (appearing as a black void) in the equatorial region of a cat retina. Magnification, ×300.

Figure 4.

 

Nerve fiber layer constituents in a macaque retina are widely separated as they pass over a branch temporal vein (right arrow) near the optic nerve head. Note the integrity of axon fascicles as they pass over a minor branch (left arrow). Magnification, ×190.