A Model to Study Differences between Primary and Secondary Degeneration of Retinal Ganglion Cells in Rats by Partial Optic Nerve Transection

Abstract

purpose. To use a rat model of optic nerve injury to differentiate primary and secondary retinal ganglion cell (RGC) injury.

methods. Under general anesthesia, a modified diamond knife was used to transect the superior one third of the orbital optic nerve in albino Wistar rats. The number of surviving RGC was quantified by counting both the number of cells retrogradely filled with fluorescent gold dye injected into the superior colliculus 1 week before nerve injury and the number of axons in optic nerve cross sections. RGCs were counted in 56 rats, with 24 regions examined in each retinal wholemount. Rats were studied at 4 days, 8 days, 4 weeks, and 9 weeks after transection. The interocular difference in RGCs was also compared in five control rats that underwent no surgery and in five rats who underwent a unilateral sham operation. It was confirmed histologically that only the upper optic nerve had been directly injured.

results. At 4 and 8 days after injury, superior RGCs showed a mean difference from their fellow eyes of −30.3% and −62.8%, respectively (P = 0.02 and 0.001, t-test, n = 8 rats/group), whereas sham-operation eyes had no significant loss (mean difference between eyes = 1.7%, P = 0.74, t-test). At 8 days, inferior RGCs were unchanged from control, fellow eyes (mean interocular difference = −4.8%, P = 0.16, t-test). Nine weeks after transection, inferior RGC had 34.5% fewer RGCs than their fellow eyes, compared with 41.2% fewer RGCs in the superior zones of the injured eyes compared with fellow eyes. Detailed, serial section studies of the topography of RGC axons in the optic nerve showed an orderly arrangement of fibers that were segregated in relation to the position of their cell bodies in the retina.

conclusions. A model of partial optic nerve transection in rats showed rapid loss of directly injured RGCs in the superior retina and delayed, but significant secondary loss of RGCs in the inferior retina, whose axons were not severed. The findings confirm similar results in monkey eyes and provide a rodent model in which pharmacologic interventions against secondary degeneration can be tested.

Glaucoma is the second most frequent cause of blindness in the world. ¹ Injury to retinal ganglion cell (RGC) axons in this disorder and other optic nerve diseases often leads to retrograde degeneration of RGC bodies in the retina. ² In optic neuropathies, one or more contributing factors probably lead to direct or primary neuronal death. It has been suggested that some loss of RGCs in glaucoma and other optic neuropathies is also due to indirect or secondary degeneration. ³ If present, secondary degeneration could be a substantial additive factor in damage caused by glaucoma and other optic neuropathies, and its therapy could represent an additional avenue in treatment. 

In central nervous system (CNS) diseases, there is increasing evidence that death of neurons and glia can be associated with damage to other neurons that were not injured by the primary insult. ^{4 5 6} This secondary degeneration is detectable in a zone surrounding ischemic or traumatic lesions in the brain and appears to be due to the initiation of biochemical events that exacerbate the direct damage. ⁷ Moreover, in CNS injury and ischemia, the secondary (delayed) neuronal death in the penumbra area is believed to occur by apoptosis. 

In separating primary from secondary degeneration in optic nerve disease models, some investigators have used the temporal sequence of cell death. ³ They reasoned that death in an initial (arbitrarily chosen) time window is primary and later death is secondary. Although this concept is logical, it is not without qualifications. Some neurons may take longer to die after primary injury for a variety of reasons. We conducted an initial study in monkeys that allowed more precise segregation of directly injured RGCs from RGCs that died without direct insult. ⁸ The projection of RGCs to their synaptic targets in the brain was organized into a topographic map. RGC bodies in the upper and lower retina pass separately through the optic nerve and tract. ^{8 9} In our monkey experiment, we cut the superior one third of the optic nerve, leading to primary degeneration of most RGCs in the superior retina. In the inferior retina, we found significant loss of RGCs with intact axons. 

Our model allowed more precise separation of directly injured RGCs from RGCs that died without direct insult. Thus, partial optic nerve transection represents a useful model for studying the magnitude, distribution, selectivity, and differential mechanisms of secondary neuronal degeneration. The expense of procuring monkeys precludes large-scale testing of treatments for secondary degeneration. As a result, we developed a similar model in the optic nerve of rats. In the present study, we demonstrate that partial optic nerve transection can be reproducibly performed in the rat. The technical difficulty of reliably cutting the same portion of the superior nerve was overcome through specially designed instrumentation. With a rodent model, large numbers of animals can be studied to allow careful delineation of the nature of secondary degeneration after the death of neighboring neurons. 

Methods

Animals

Wistar rats weighing 400 to 450 g were treated under procedures approved and monitored by the Animal Care Committee of the Johns Hopkins School of Medicine and following the procedures outlined in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Animals were housed with a 14-hour-light/10-hour-dark cycle with standard chow and water ad libitum. 

Partial Transection Model

Unilateral, partial optic nerve transection was performed in animals that were anesthetized with intraperitoneal ketamine (50 mg/kg) and xylazine (5 mg/kg), topical proparacaine 1% eye drops, and with pupils dilated by tropicamide 1% eye drops. An incision in the superior conjunctiva was made, and the eye was gently retracted outward with forceps, exposing the nerve behind the eye. A specially designed diamond knife was then used to incise the optic nerve to a depth of one third of its diameter at a point 2.5 to 4.0 mm behind the eye. The knife was similar to that previously used for radial keratotomy in human eyes, with a blade that could be advanced in 10-μm increments from a twist-handle, and a single metal guard that limited penetration to that measured on the scale. Typically, the blade was set at a depth of 150 μm. The conjunctival incision was self-closing. Sulfacetamide 10% ointment was applied at the end of surgery. Each surgically injured eye was inspected ophthalmoscopically to ensure patency of blood flow to the eye. 

RGC Body Labeling with Fluorescent Gold Dye

For most of the present experiments, RGC were backfilled with a fluorescent dye (Fluorogold; Fluorochrome, Inc., Denver, CO) delivered to the superior colliculus on both sides. We anesthetized rats as for the surgery. The rat was placed under sterile conditions in a stereotactic apparatus (Stoelting, Inc., Wood Dale, IL) and the scalp was incised. The anterior–posterior and lateral location of the superior colliculus on both sides was determined and marked with cautery using coordinates from a standard atlas for the Wistar rat brain. ¹¹ A burr hole was centered over each colliculus using a tool with a 0.75-mm drill bit (Dremel, Inc., Racine WI). Injections were made with a blunt, 30-gauge needle connected to a 50-mL syringe (Hamilton Co., Reno, NV), containing 5% fluorescent gold in 0.9% normal saline, driven by the Stoelting stereotactic injector at 1 mL per minute. On each side of the brain, three injections, each consisting of 1.5 mL, were made at sequentially deeper planes (2.7, 3.2, and 3.6 mm from brain cortex level). The needle was left in position for 2 minutes at each plane before moving deeper. At the end of the procedure, the scalp was closed with staples. 

The following groups of rats were evaluated after fluorescent gold labeling. Five rats served as the bilaterally unaltered control. Five rats underwent unilateral sham optic nerve transection, in which the dural sheath was opened and the knife was placed on the nerve, but no cut was made. In addition, 33 rats underwent partial transection 1 week after fluorescent gold labeling and were killed at the following times after partial transection: eight rats at 4 days, eight rats at 8 days, nine rats at 4 weeks, and eight rats at 9 weeks. 

RGC Counting Procedure

Rats were killed by exsanguination while under deep intraperitoneal ketamine and xylazine anesthesia. They were perfused through the heart for 20 minutes with 4% paraformaldehyde in 0.1 M phosphate buffer [pH 7.2], at a flow rate of 20 mL/min. The 12 o’clock position was marked by a light cautery burn. The globe and optic nerve were removed on each side. Wholemounts of the retina were prepared with notches at their periphery to allow flattening and to mark the horizontal meridian. These were viewed in a fluorescence microscope (Axioskop; Carl Zeiss Meditech, Inc., Thornwood, NY) with appropriate filters to identify fluorescent gold–labeled cells. RGCs in corresponding zones of retina above and below the horizontal were counted in images taken with a 40× objective subtending an area in fixed tissue of 0.096 mm². These images were taken along six radii centered on the position of the optic nerve head, three above and three below. Two of these radii were directly vertical (above and below), and the other four were 20° nasal and temporal to the vertical. Four consecutive, contiguous images were obtained along each radius, beginning 0.67 mm away from the optic nerve head. The number of labeled cells in the ganglion cell layer was counted in each image and converted to a density value. The counting of cells labeled with fluorescent gold was performed by three of the investigators. To determine whether there was significant interobserver or intraobserver variation, 18 fields were counted by all three observers on two separate occasions in masked fashion. The correlation coefficients among the three observers varied from 0.96 to 0.99. Hence, we included data from all three observers interchangeably in the statistical analyses. The observers were masked to the procedure that had been performed on each eye. 

RGC Axon Counting Procedure

The optic nerve from both eyes of each animal was removed 1.5 mm posterior to the eye. This was postfixed in 1% osmium tetroxide in phosphate buffer and processed into epoxy resin. A 1-m thick cross-section was cut from the nerve end that had been closest to the globe and stained with 1% toluidine blue. The total nerve fiber count for both eyes was estimated for each animal from these sections. To accomplish this, we measured the area of the optic nerve in cross section by outlining its outer border at 10× magnification on an image-analysis system with a digital camera and software (Sensys camera with Metamorph software; Universal Imaging Corp., West Chester, PA). ⁸ Three optic nerve area measurements were taken, and the mean was used. To measure the size and density of axons, we captured images with a 100× phase-contrast objective from 10 randomly selected, nonoverlapping areas of each nerve. These were edited to eliminate non-neural objects and the density of axons per square millimeter was calculated for each image and averaged for all images to derive a mean density for each nerve. The mean density was multiplied by the total nerve area to estimate the fiber number for the nerve. The number of axons counted per nerve was approximately an 11% sample (9500 axons) of the total number from each eye. Examination of the nerves was performed by a person unaware of which procedure the eye had undergone. For this study, the number of axons from eyes with partial transactions, sham, and noninjured control were compared with the mean axon estimate from 205 normal Wistar rat nerves collected as part of other studies and counted in identical fashion. 

Evaluation of Axon Topography

Our model is based on knowledge that the axons of the upper optic nerve that are cut in superior, partial transection belong only to RGCs residing in the superior retina. To confirm that this anatomic feature is present in the Wistar rat, we designed an additional experimental procedure. In seven additional animals, no fluorescent gold was injected in the brain, but a unilateral, partial transection was performed in the superior optic nerve. Seven days later, the animals were reanesthetized and rhodamine-dextran crystals (Molecular Probes, Inc., Eugene, OR) were placed on the cut surface of a full transection of the optic nerve closer to the eye than the prior partial transection on the same side. These animals were allowed to survive for 24 hours, then killed as described earlier. The number of RGCs was compared with that in six control rats that had not had either fluorescent gold or partial transection, but has undergone complete transection and placement of rhodamine-dextran 24 hours before death. The retinal wholemounts were studied by fluorescence microscopy with appropriate filter for rhodamine to count the number of RGCs, as was described earlier. Separate control animals were used to avoid the performance of bilateral optic nerve transection in animals that survived for 24 hours. 

Our fluorescent gold data show that most of the degeneration in the superior retina (primary degeneration) is completed by the first week after injury (see the Results section). Hence, we reasoned that the rhodamine-dextran labeling in the inferior retina 1 week after partial transection would show no loss of RGC compared with their controls, because there would be no primary degeneration (due to topographic segregation), and secondary degeneration would not yet have begun. If there were inferior RGC loss in the partial transection eyes of this group, then there might be axons of inferior RGCs in the superior optic nerve, or secondary degeneration might have begun in the first week. 

Results

Control RGC Data

Control eyes and those observed during the first 8 days after partial transection had three types of cells in the RGC layer that were labeled with fluorescent gold. Type 1 cells were clearly large RGCs with full dendritic trees and often a visible axon (Fig. 1) . 

Type 2 cells had smaller cell bodies with no labeling of dendrites, and only an occasionally visible axon, but they met the size and shape criteria of RGCs. Cell types 1 and 2 were intensely labeled with fluorescent gold, and the sum of type 1 and 2 cells was used to estimate the total number of RGCs. Some much less intensely labeled cells (type 3 cells, not shown) were also visible in all fields examined. These cells were of indeterminate lineage, and their vague outline was attributed to autofluorescence. They were not included in estimates of RGC number. The mean, total number of RGCs counted per normal, control retina was 5122 ± 1173 (n = 35 normal eyes from 35 rats, mean ± SD). The mean density of RGCs in these 35 normal eyes was 2223 ± 509 cells/mm². By measuring the total retinal area in fixed tissue (39.5 mm²) and multiplying by the densities, we estimated that the fluorescence-labeled type 1 and 2 cells constitute 87,809 ± 20,106 neurons per eye. In 205 control rats, we estimated the number of RGC axons in the optic nerve immediately behind the eye. Assuming one axon per RGC, the mean number of RGCs was 85,511 ± 11,040 (mean ± SD). Therefore, the number of RGCs identified by fluorescent gold labeling seems to be a reasonable estimate of all RGCs. 

Only type 1, 2, and 3 cells were detected in the retinas that were evaluated within the first 8 days after partial transection (and a total of no more than 18 days after fluorescent gold injection). In contrast, retinas at 5 and 10 weeks after fluorescent gold injection (and 4 and 9 weeks after partial transection) had a fourth cell type (Fig. 1) . These type 4 cells were irregular in cell body shape, were both larger and smaller than those thought to be RGCs, and had intense fluorescent gold labeling. They frequently had multiple cell processes and were clearly similar to microglial or macrophage cells in structure. Indeed, there is considerable evidence that fluorescent gold migrates from RGC into which it first arrives after collicular injection, labeling macrophages, and retinal vascular cells. ¹² The transfer of fluorescent gold from initially labeled neurons to other cells is particularly prominent when neurons containing fluorescent gold degenerate, with dye engulfed by macrophage-like activity of a variety of cells. ¹³ As a result, in these experiments, we included only type 1 and 2 cells in our estimates of the number of RGCs. 

RGC Loss after Partial Transection

The estimated number of RGC in both eyes of normal, uninjured rats differed by 1.6% (Table 1) . 

In rats with one eye that underwent a sham operation, the difference in the number of RGCs between the two eyes was not significant when compared with uninjured control values, with the difference between the two eyes amounting to a mean of only 1.7% (P = 0.74, one sample t-test, Table 1 ). The number of RGCs in the entire retina was lower than the normal number in every group of animals with partial optic nerve transection from 4 days to 9 weeks after surgery (all P < 0.003, t-test, Table 1 ). The estimated loss of RGCs over the whole retina peaked at approximately 40% at 1 month and was similar at 9 weeks. 

The loss of axons in the optic nerve followed a similar time course and magnitude to the cell body count. Sham-operation eyes had an estimated loss of only 10% of axons, which was not statistically significant compared with control optic nerves (P > 0.05, Fig. 2 ). 

The axon loss 4 days after partial transection was substantial and increased at 8 days. There appeared to be no significantly greater loss at 4 and 9 weeks after surgery. All these time points were significantly different from the control (P < 0.001, t-test). 

Superior Compared with Inferior RGC Death

There was a significant difference in the survival of RGC bodies in the superior retina compared with the inferior retina. The loss superiorly was apparent at 4 days and increased 8 days after partial transection of the superior optic nerve (Table 1 ; Fig. 3 ). 

Inferior retinal RGCs had minimal loss at 4 and 8 days that did not achieve statistical significance. By 4 and 9 weeks, there was significant loss inferiorly, but its magnitude was somewhat less than that in the superior retina at each time point. 

Regional Differences

We also compared the RGC loss in the two more peripheral zones with that in the two more central zones, both superiorly and inferiorly. It might be expected that the most peripheral, superior zone would show the most dramatic loss, because the peripheral axons of the optic nerve would have been most likely to be cut. Similarly, the inferior peripheral retina would be expected to show least damage. Although the peripheral, superior zones showed the greatest loss at 4 and 8 days, the comparison of subregions of the retina showed no statistically significant finding at any time point (Fig. 4) . 

Backfilling with Rhodamine-Dextran

In the 7 eyes in which we backfilled RGCs with rhodamine-dextran, there was clear evidence of loss only of superior RGC at 8 days. The superior peripheral retina had 29.4% loss comparing the mean for treated eyes to the mean for control eyes (P = 0.036, t-test), as would be expected from the superior partial transection. The mean values for superior central, inferior central, and inferior peripheral retinal areas were not statistically different from mean control retinas in the same areas (all P ≥ 0.05). The variability in RGC counts with the rhodamine-dextran specimens was substantially higher than for those in which fluorescent gold was used. 

Optic Nerve Topography

To assure that our model produced transection only in the superior segment of the optic nerve, we performed detailed serial cross sections of the optic nerve in 10 specimens. These demonstrated that the rat optic nerve shares features with the monkey and human in maintaining a relative topographic positioning of axons belonging to RGCs that are neighbors in the retina. Near to the area of partial transection, the loss of axons was confined to the superior one third of the optic nerve in a localized area (Fig. 5) . 

As the nerve was sectioned serially moving toward the brain, these superior fibers came to occupy a more central location within the nerve, presumably in preparation for their crossing at the optic chiasm to the opposite side of the brain. 

Discussion

In the central nervous system, injury from various primary lesions, such as ischemia and trauma, can lead to widespread damage to neurons beyond the initial injury site. ^{14 15 16 17} This phenomenon is known as secondary degeneration and can result in greater loss of tissue than that caused by the initial disorder. Yoles and Schwartz ³ reported findings compatible with secondary degeneration after optic nerve injury in rats, involving substantial loss of RGC and continuing for an extended period after termination of the primary injury. ³ The secondary death of neighboring neurons and glia is believed to occur by apoptosis. ¹⁸ A variety of mechanisms for this secondary degeneration have been proposed, including alteration of extracellular ion concentration, release of oxygen free radicals, and high levels of excitatory neurotransmitters. 

Our results provide additional, direct evidence for the occurrence of secondary degeneration in the optic nerve of rats. Although there have been studies in which secondary degeneration of central neurons was assumed, the present model allows explicit delineation of which neurons died after actual injury to their axons and which did so due to indirect toxicity. Previous studies have presumed that the initial injury zone was confined to a particular distribution (due either to the nature of the injury or the distribution of blood vessels that were occluded) or they presumed that primary and secondary degeneration follow different time courses, with the latter taking longer to begin. Because we are observing secondary degeneration in a more isolated form, we can confirm that cells with axons that are not directly injured not only die when found in proximity to other dying neurons, but they do so on a delayed time course. It is known that there can be a time lag of weeks between axonal injury and cell body death in RGCs. ^{2 3} The magnitude of inferior RGC loss in this study is similar to that observed in monkey eyes after partial, superior optic nerve transection. ⁸  

Previous studies of optic nerve injury have used complete transection, nerve crushing, ischemia, or elevated IOP to study the death of RGCs. In each of these models, however, it is not possible to know whether a particular RGC was actually injured by the stimulus. In each of these settings, an RGC could have been subjected to noxious stimuli by the injury itself. Hence, the nature of its degenerative process cannot be assumed to be a secondary one. 

We suspect that there are differences in the mode of cell death that distinguishes primary and secondary degeneration. Optic nerve injury, optic nerve transection, and both experimental and clinical glaucoma lead to apoptotic RGC death in rats, monkeys, and humans. ^{19 20 21 22} In a variety of CNS disorders, primary injury leads to secondary damage by creating a hostile environment in the surrounding tissues. After the initial injury, there are several routes by which consecutive damage to additional tissue might occur. First, when neurons and glia die, their cellular byproducts could produce a detrimental effect on surrounding cells. Second, the initial injury mechanism might lead to breakdown in normal physiological equilibria, perhaps disturbing blood flow, breaking down the blood–brain barrier, or releasing molecules that are not perceived by normal immune surveillance, including effects produced by non-neuronal cells, such as glia, macrophages, and lymphocytes. Finally, neurons whose axons are in the injured area could die by loss of trophic support from the dead, primary neurons (transsynaptic degeneration). We plan further studies to evaluate these possibilities in our model system. 

Our model may be subject to several methodological weaknesses. First, there could have been primary RGC death in the inferior retina if the blade passed too far into the nerve. Our evaluations of the site of injury in the large number of nerves studied rules out this possibility as a cause of our finding. Second, axons of RGCs residing in the inferior retina may be randomly distributed in the optic nerve. However, several prior studies in other mammals ^{9 10} strongly support an orderly arrangement and clinical–pathologic correlations of ocular disease with atrophy in optic nerve cross sections confirm such regional organization. ²³ In addition, both the serial sections of the optic nerve and our rhodamine-dextran data from this study attest to the orderly topographical segregation of fibers from RGCs of superior and inferior retina. Third, damage to the blood supply of the eye during surgery may have caused primary ischemic death of inferior RGCs. There was, however, no clinical sign of such ischemia, and the fact that the cell death followed a different time course in the superior and inferior retina speaks against this explanation. Finally, it may be that toxic effects are exerted on the inferior optic nerve axons after the nerve sheath is opened and the diamond blade is passed through the superior nerve. Clearly, our sham surgery did not lead to inferior RGC death with the manipulations that included opening the sheath. Furthermore, the diamond blade is so sharp that its passage cannot be felt by the surgeon and must be observed to know that it has occurred. Hence, we believe that the model isolates direct injury to superior RGCs. 

We did not find a greater loss of RGCs in the portion of the inferior retina closest to the primary degeneration (inferior central sector), compared with the inferior peripheral sector. There may be no such gradient, and any RGC within an eye that sustains significant superior RGC injury may be equally likely to die. This would be the case if toxic effects were present diffusely throughout the retina or vitreous. Similarly, it would also be the case if secondary damage were dependent on toxic effects at the level of the optic nerve, which represents such a small area in the rat that diffusion of any molecular influence would be nearly instantaneous. We plan more detailed studies of regional influences in this model to determine the effect of proximity to directly injured neuronal cell bodies. 

Fluorescent gold backfilling is intended to identify RGCs reliably. To our knowledge, there are no other methods that stain or immunolabel all RGCs more specifically. However, we recognize that there are disadvantages of fluorescent gold backfilling. By 2 weeks after injection, the label exits the RGCs and is taken up by other resident cells of the retina, particularly glia. This is particularly a problem when RGC death occurs and phagocytosis of fluorescent gold-labeled cells occurs. The type 4 cells illustrated here represent these non-RGC elements that are sometimes heavily invested with fluorescent gold. Work with RGC injury models in which backfilling with this dye is used should take this into account when there are long periods between injection and death or when RGC degeneration occurs as part of the study. Counts of RGCs should include only those fluorescent gold–labeled cells that have morphology appropriate for an RGC. 

In summary, we suggest that secondary degeneration may cause substantial loss of some RGCs as an indirect effect of the injury and death of directly injured RGCs. The model developed in the current study allows investigations into the mechanisms of secondary degeneration, providing a means to distinguish it from primary degeneration, and to prevent it by specific therapies. 

 Supported by National Eye Institute Grants EY02120 (HAQ) and EY01765 (Core Facility Grant, Wilmer Institute); The Wagner Charitable Trust, New York, New York; and the David Worthen Glaucoma Fellowship, the Glaucoma Research Foundation, San Francisco, California; Prevent Blindness America; The TFC Frost Trust, United Kingdom, and the University College, Oxford, United Kingdom.

 Submitted for publication July 1, 2002; revised January 27, 2003; accepted February 17, 2003.

 Disclosure: H. Levkovitch-Verbin, None; H.A. Quigley, None; K.R.G. Martin, None; D.J. Zack, None; M.E Pease, None; D.F. Valenta, None

 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.

 

Fluorescent gold images of retinal cells. (A) Three type 1 RGCs, large and with dendrites and axon, and three type 2 RGCs, smaller and with no visible dendrites. Note that the fluorescent gold is generally clumped in the cytoplasm around the nucleus. (B) Retina at 4 weeks after transection in the superior retina. There are several type 4 cells that show irregular outlines, intense fluorescent gold labeling, and a variety of processes. The size of most type 4 cells is smaller than true RGC. Bars, 20 μm.

Table 1.

 

Percent Loss of Retinal Ganglion Cells after Partial Transection

Group (n)	Whole Retina	P *	Superior Retina	P , †	Inferior Retina	P , †
Control (5)	1.6 ± 14.5		3.2 ± 31.1		3.0 ± 9.8
Sham (5)	1.7 ± 13.9	0.74	3.8 ± 29.1	0.78	1.4 ± 6.8	0.71
4 day (8)	−20.0 ± 8.9	0.0004	−30.3 ± 19.3	0.02	−10.0 ± 10.4	0.054
8 day (8)	−30.2 ± 14.1	0.0005	−62.8 ± 26.8	0.0018	−4.8 ± 11.0	0.29
4 week (9)	−42.7 ± 9.0	0.0001	−46.3 ± 13.5	0.0012	−39.2 ± 16.5	0.0002
9 week (8)	−38.0 ± 23.5	0.003	−41.2 ± 25.6	0.017	−34.5 ± 24.7	0.0096

Figure 2.

 

The loss of axons in the optic nerve at various times after partial optic nerve transection. Data represent fibers across the entire optic nerve, showing rapid loss in the first week, followed by a slower loss at the later time points. Compared with control optic nerves and compared with the sham surgery group, all values 4 days and longer after transection show statistically significant loss (P < 0.001, ANOVA).

Figure 3.

 

The percentage of RGCs lost after partial optic nerve transection. Note that the loss in the superior retina occurs rapidly during the first week and levels off in the later periods. The inferior loss is not present until the 3-week time point, though other times between 8 days and 3 weeks were not studied. Hence, the precise time point at which secondary degeneration might begin is unknown. The difference in time course points to secondary degeneration as the more likely cause of degeneration inferiorly.

Figure 4.

 

Graph shows regional loss of RGCs in retina, divided into four zones. The control and sham-operation groups showed no significant difference between eyes. At the 4- and 8-day time points, the superior peripheral loss was greatest, as would be expected from the greater likelihood that these RGCs would have axons cut by superior partial transection. The inferior RGC loss is slightly, but not significantly, greater in inferior central zone (closer to the primary injury) than in the inferior peripheral zone.

Figure 5.

 

Optic nerve cross section, 8 days after transection superiorly. This section shows that near to the site of transection, the injury was confined to a local zone of the superior nerve, where darker, degenerating axons were found in the area indicated by arrows. Toluidine blue; scale bar, 50 μm.