Severe, Early Axonal Degeneration Following Experimental Anterior Ischemic Optic Neuropathy

Abstract

Purpose.: Anterior ischemic optic neuropathy (AION) is the most common acute optic neuropathy in adults older than 50 and leads to axonal degeneration, thinning of the retinal nerve fiber layer and loss of the retinal ganglion cells (RGCs). We used experimental AION model to study early axonal changes following ischemia.

Methods.: We induced optic nerve head ischemia in adult mice using photochemical thrombosis and analyzed retinal changes within 1 week. We used confocal scanning laser ophthalmoscopy (cSLO) and fluorescence microscopy of retinal whole mount preparations to analyze axonal degeneration in Thy1-YFP-H mice and those injected with annexin-V-A488 intravitreally.

Results.: Three days after AION, morphometric analyses in Thy1-YFP-H mice revealed evidence of early axonal changes, including swollen or branched axonal stumps. There was also a beads-on-a-string appearance of YFP expression. The axonal enlargements occurred at an interval of 17 ± 1 μm or 6 ± 0 enlargements/100 μm. At day 7 after AION, the degenerating intraretinal RGC axons exhibited intense annexin-V-A488 staining (P = 0.002). The annexin-V staining pattern was fragmented, with intersegment interval of 20.1 ± 1.4 μm or 5.8 ± 0.4 annexin-V-A488⁺ fragments/100 μm, which were similar to that of degenerating Thy1-YFP⁺ axons.

Conclusions.: Following a photochemical thrombosis model of AION, RGC axons displayed severe degenerative changes within 1 week, suggesting that after ischemia, RGC axons may degenerate in a temporally and spatially distinct fashion from that of the soma. Our findings also further established annexin-V as a useful marker of retinal degeneration because it strongly labeled dying RGC axons.

Introduction

Axons are the information highways that connect the 100 billion neurons and form the 100 trillion synapses in the human brain. Certain neurons and their axons, including those in the optic nerve, hippocampus, and corpus callosum are particularly susceptible to loss in normal aging for reasons we do not understand well, and optic nerve axonal loss in humans has been estimated at 0.3 to 0.6% or approximately 5000 axons per year.^1–3 Common acquired optic neuropathies such as anterior ischemic optic neuropathy (AION) and glaucoma are predominantly axogenic, meaning the initial injury starts at the axons.⁴ Following a variety of insults to axons, accelerated axonal loss occurs via Wallerian-like degeneration, likely involving axonal transport failure, neurotrophin deprivation, mitochondrial dysfunction, and superoxide formation, which are distinct and overlapping compared with mechanisms of somatic degeneration.^5–8 

Despite much research, there is still relatively limited understanding of axonal degeneration following AION, and no effective treatment exists. Based on clinical studies, we know that ischemia starts at the anterior optic nerve near the convergence of the unmyelinated retinal ganglion cell (RGC) axons and the beginning of myelination in posterior ciliary artery territory and near the circle of Zinn-Haller.^9–12 Following ischemia, RGC axons and soma are lost, and both axonal and somatic layer thinning can be measured in AION patients and to a certain extent in animal model using optical coherence tomography (OCT).^13–17 

Many have used a laser-assisted photochemical thrombosis model to study axonal and somatic changes in RGCs after AION.^17–23 One day after AION induction, there is prominent swelling of the anterior optic nerve and abnormality of the visual evoked potential, indicating impaired axonal transport and blockade of signal transduction in the optic nerve.^{17–19,21,22} Over days after AION, there is activation of immediate early genes such as c-fos mRNA that peaks on day 1, in addition to increased RNA expression of different heat shock proteins, which is consistent with endogenous injury responses.^18,19 Optic nerve axonal degeneration, optic nerve oligodendrocytes apoptosis, and decreased myelin basic protein expression occur within the first week after AION.^19,24 By 21 days after AION, up to 60% of RGC axons in the optic nerve are lost.¹⁹ After experimental AION, RGC cell body loss is estimated at 25 to 65%^18,21,24 and occurs in a bimodal fashion, peaking around days 10 and 21.²⁵ In vivo imaging of experimental AION using serial OCT measurements of the inner retinal layers containing RGC axons and cell bodies after experimental AION has supported these histologic findings. There is acute swelling of the ganglion cell complex 1 day after experimental AION and gradual thinning over 3 weeks, which correlate with the time course of axonal changes and somatic loss seen on histology.¹⁷ 

In experimental AION and other optic neuropathy models, RGC somatic degeneration has been measured using annexin-V, a calcium-dependent protein located in synaptic vesicles.²⁶ Applied extracellularly, annexin-V labels cells undergoing apoptosis with nanomolar affinity by binding to the exposed phosphatidylserine in the extracellular membrane following energy failure.^27–29 Annexin-V has been used to label dying RGCs in vitro⁶ and in vivo^30,31 after optic nerve transection,⁶ retinal laser photocoagulation,³² NMDA-induced excitotoxicity,³³ experimental AION,²⁵ and experimental glaucoma.^34,35 However, annexin-V has had limited use in labeling degenerating axons in the peripheral and central nervous system. In this paper, we present our data on evidence of early axonal degeneration and annexin-V labeling of RGC axons following experimental AION. 

Methods

Animals

All animal care and experiments were carried out in accordance with the Association for Research in Vision and Ophthalmology statement for the Use of Animals in Ophthalmic and Vision Research and with approval from the Stanford University Administrative Panel on Laboratory Animal Care. Adult wild-type C57BL/6 mice (Charles River Laboratories, Inc., Wilmington, MA, USA) and Thy1-YFP-H mice (generous gift of Ben Barres) were housed in cages at constant temperature, with a 12:12-hour light/dark cycle, with food and water available ad libitum. The Thy1-YFP-H mice express YFP in a relatively small number of RGCs, which help imaging of individual cell bodies and processes.³⁶ All procedures were performed under sedation, achieved with ketamine 50 to 100 mg/kg (Hospira, Inc., Lake Forest, IL, USA), xylazine 2 to 5 mg/kg (Bedford Laboratories, Bedford, OH, USA), and buprenorphine 0.05 mg/kg (Bedford Laboratories). The pupils of anesthetized mice were dilated with 1% tropicamide (Alcon Laboratories, Inc., Fort Worth, TX, USA) and 2.5% phenylephrine hydrochloride eye drops (Akorn, Inc., Lake Forest, IL, USA). Lubrication of the eyes and custom-made contact lenses ensured health and clarity of the cornea throughout in vivo imaging. 

Induction of AION Using Photochemical Thrombosis

To induce optic nerve head ischemia, we injected rose bengal (1.25 mM in phosphate-buffered saline, 2 ml/kg animal weight) intravenously through the tail vein, then induced ischemia at the optic nerve head^18–20 with a frequency-doubled 532 nm Nd:YAG laser (PASCAL, OptiMedica, Sunnyvale, CA, USA) using 15 laser spots of 400 μm diameter, low power of 50 mW, and 1 second duration per pulse. In each mouse, one eye had AION induction, and the contralateral eye served as controls. The AION induction was randomized. 

Axonal Changes in Thy-1-YFP-H Mice and Imaging Using cSLO and Fluorescence Microscopy

After AION, we performed serial confocal scanning laser ophthalmoscopy (cSLO) imaging of Thy1-YFP-H mice at 488 nm using Spectralis HRA+OCT (Heidelberg Engineering, GmbH, Heidelberg, Germany). The focus of the lens was adjusted for every eye to capture the most robust fluorescence signal in the RGC axons in the inner retina. On average, 100 images at high-resolution mode were averaged per picture. There was preservation of YFP signal over several days after AION to easily track axonal integrity in vivo. Histologic studies were carried out on the last time point of cSLO imaging (days 3–5). The animals were killed through transcardiac perfusion of 4% paraformaldehyde in PBS, and we prepared whole mount retinae, which were immunostained with anti-GFP rabbit polyclonal antibody (1:500–1000 dilutions; Sigma-Aldrich, St. Louis, MO, USA) and secondary goat anti-rabbit-IgG A488-labeled antibody (1:200–400 dilutions, Invitrogen/Life Technologies, Grand Island, NY, USA) and mounted in DAPI-containing medium (Vectashield, Vector Laboratories, Burlingame, CA, USA). Fluorescence microscopy was done using an inverted Nikon Eclipse TE300 microscope (Nikon Corporation, Tokyo, Japan) with 4×, 10×, and 20× objectives, and images were captured using Metamorph software (Molecular Devices, Sunnyvale, CA, USA). 

Quantification of Thy1-YFP⁺ Axons

To quantify the RGC axonal changes, the photographs of the YFP were assessed in a masked fashion by two investigators. We counted the number of total YFP positive axons per eye under higher magnification, the number of axons that exhibited branched or swollen axonal endings, and the number of axons that had a beads-on-a-string pattern. We used ImageJ (http://rsbweb.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA) to analyze the pattern of YFP signal along the axons. Twenty-one axons in nine eyes (approximately 300 μm each) were manually outlined and straightened using the “straighten” plug-in. We then measured the fluorescence signal profile and calculated the interpeak intervals, number of peaks per 100 μm, the width of the axonal enlargements (“beads”), and the width of the axons in the AION and control eyes. The intersegment interval was calculated by dividing the number of peaks by the length of the axon and converted to number per 100 μm of axon. Because the measurements were done on axons in flat-mounted retinae and artificially straightened, our length measurements were an approximation of the axonal length in vivo. 

Annexin-V-A488 Staining of Degenerating Axons and Imaging Using cSLO and Retinal Whole Mount Preparations

At 5 and 7 days following AION, we analyzed annexin-V-A488 labeling of RGCs. We first performed intravitreal injection of 1.5 μL annexin-V-A488 (Invitrogen/Life Technologies) using a 31-gauge insulin syringe. At 2 to 2.5 hours later, the mouse eyes were dilated, lubricated, and covered with contact lens for cSLO imaging using Spectralis HRA+OCT (Heidelberg Engineering, GmbH) in a similar fashion as Thy1-YFP-H in vivo imaging. The cSLO imaging of annexin-V-A488 signal was not as robust as the Thy1-YFP-H mice, probably because the signal was relatively weak. After cSLO imaging, at approximately 3 hours after intravitreal annexin-V-A488 injection, the mice were killed through transcardiac perfusion of 4% paraformaldehyde (Polysciences, Inc., Warrington, PA, USA) in PBS (Sigma-Aldrich), and we prepared whole mount retinae, which were rapidly mounted in DAPI-containing medium (Vectashield, Vector Laboratories) and immediately imaged due to the lability of the annexin-V-A488 signal in a masked fashion. The cSLO images were compared with whole mount retinae from the same eyes to assess ability of in vivo imaging to identify features seen on histology. 

Quantification of Annexin-V-A488⁺ Axons

Annexin-V-A488-labeled retinal whole mounts were imaged using the Nikon Eclipse TE300 inverted microscope (Nikon Corporation) using 4×, 10×, and 20× objectives and Metamorph software (Molecular Devices). The 10× photographs (550 μm wide, 550 μm tall) of the annexin-V-A488 staining of the retina in whole mount preparation were quantified in a masked fashion by two investigators. The number of annexin-V-A488 positive axons was counted at the peripheral 150 μm of each photograph where the axons were more distinguishable. Because we only counted axons with robust annexin-V labeling and distinctively visualized, the actual number of degenerating RGC axons at week 1 were likely underestimated. We did not quantify the total fluorescence signal per area due to the occasional high background of annexin-V-A488 in some eyes. To study the pattern of annexin-V-A488 signal along the axon, we manually outlined the axons, straightened them using the “straighten” plug-in in ImageJ (http://rsbweb.nih.gov/ij/; provided in the public domain by the National Institutes of Health), and measured the fluorescence signal along the straightened processes. We converted a pixel in Image J to micrometer measurements using a scale bar. Because the measurements were done on axons in flat-mounted retinae and artificially straightened, our length measurements were an approximation of the length in vivo. We analyzed 50 to 70 axons, each approximately 300 μm, and quantified the number of annexin-V-A488 signal peaks using a cut off of fluorescence signal over 150 arbitrary fluorescence units (maximal 255). Two investigators quantified the number of annexin-V⁺ fragments per axon, and the counts were averaged. The intersegment interval was calculated by dividing the number of peaks by the length of the axon and converted to number per 100 μm of axon. 

Statistics

We used a nonparametric Wilcoxon matched-pairs signed rank test or a Mann-Whitney U test in Prism (GraphPad, La Jolla, CA, USA). We calculated averages and the standard error of the mean in Microsoft Office Excel (Microsoft, Seattle, WA, USA). The cut-off for statistical significance was set at P < 0.05. There was no difference using either of the two statistical methods. 

Results

Axonal Degeneration After AION in Thy1-YFP Mice

To study axonal changes over large distances, we induced experimental AION in Thy1-YFP-H mice, which express YFP in the soma, axon, and dendrites in a small number of RGCs.^36,37 The YFP-labeled RGC axons in Thy1-YFP mice were easily visualized in retinal whole mount preparations (Figs. 1A–C) and in vivo using serial cSLO (Figs. 1D, 1E, 2), which allowed for comparison of the same axons over time. Fragmentation of the YFP signal along the RGC axons was evident as early as 30 minutes after AION in 1-year-old mice and progressed rapidly over 3 days on cSLO (Fig. 2). This fragmentation was less common in 3- and 7-month-old mice. In vivo, the fluorescence intensity of the YFP protein was qualitatively lower over days after AION (Figs. 1, 2). This was also seen in retinal whole mount, which showed that there were significantly fewer axons that brightly expressed YFP in a uniform pattern 3 to 5 days after AION (AION: 18 ± 3%, control: 29 ± 4%; P = 0.02), which may be related to down-regulation of the Thy-1 promoter after injury.³⁹ 

Figure 1

 

In vitro and in vivo images of 3-month-old Thy1-YFP mouse retinae after AION. (A) Retinal whole mount preparation 5 days post-AION showing severed axonal stumps that appeared enlarged (closed arrow) and irregularly branching (open arrows), consistent with retrograde axonal degeneration away from the optic nerve head (asterisk). Open arrowheads denote same RGC cell bodies seen in (A), (D), and (E). (B) High magnification view of degenerating axons after AION showing YFP expression in enlarged axonal stumps (arrows) and axonal swelling (arrowheads), like beads on a string. (C) Examples of beading (arrowheads) seen on higher magnification view of the retina at 6 o'clock position in (A). (D) In vivo cSLO image of the same retina after AION in (A). Closed arrowheads point to corresponding axonal features seen in (A, E). (E) Baseline cSLO image of the retina in (A, D) before AION.

Figure 2

 

Rapid fragmentation of YFP expression in a 1-year-old Thy1-YFP-H mouse retina after optic nerve head ischemia in serial cSLO images. Thirty minutes after ischemia, irregular YFP expression along the axon was already visible. The fragmented appearance became more striking within days after ischemia. Empty arrows indicate the same cell bodies seen across the four time points.

In retinal whole mount preparations made 3 to 5 days following ischemia, we observed many axons undergoing retrograde degeneration (Fig. 1) (N = 12 mice). In the AION eyes, there was a significant increase in the number of axons that were branched, swollen, or retracting away from the optic nerve head compared with control eyes (P = 0.004) (Fig. 1A). Of the axons expressing YFP, 14 ± 1% appeared to retract away from the optic nerve head in a branched manner, compared to 2 ± 0% of the axons in the control eyes. In addition, 14 ± 2% of the axons appeared to have a swollen ending that resembled a stump, compared to 3 ± 1% of the axons in the control eyes. After AION, the YFP⁺ RGC axons were also conspicuously irregular, and many more had a beads-on-a-string appearance than control (AION: 12 ± 2%; control 4 ± 1%, P = 0.008) (Fig. 1), similar to those seen in Wallerian-like anterograde axonal degeneration after optic nerve crush.³⁸ The average distance between the peaks of the axonal enlargements was 16.7 ± 0.8 μm (range, 9–24 μm, N = 21 axons in nine eyes) or 6.3 ± 0.4 YFP⁺ segments per 100 μm. The axonal enlargements (“beads”) ranged in width from 3.8 to 8.6 μm or 1.9 to 3.8 times the width of the normal axons. The axonal width between the beads (the “string”; 1.6–2.7 μm) was similar to that of control axons (1.6–3.0 μm). 

Annexin-V-A488 Labeling of Degenerating RGC Axons in Wildtype C57BL/6 Mice

To label degenerating axons after AION, we performed intravitreal injection of annexin-V-A488 and performed in vivo and in vitro imaging in wildtype C57BL/6 mice. At day 7 after AION, we observed some cells in the RGC layer labeling with annexin-V-A488 (Fig. 3), consistent with known RGC apoptosis after experimental AION.^27–29 

Figure 3

 

Annexin-V-A488 labeling of degenerating cell bodies and neurites 7 days after AION. (A) Annexin-V-A488 (green). (B) DAPI (red). (C) Combined (A, B). The annexin-V-A488 labelled not only cell bodies (yellow, arrowheads) but also radial processes along the directions of RGC axons in a fragmented pattern.

However, the annexin-V-488⁺ cell body staining was qualitatively less prominent compared with the intense annexin-V-A488⁺ labeling of the radially orienting RGC axons at day 7 (Fig. 4). The annexin-V-A488 staining was most striking near the optic nerve head, at least partly because of the convergence of RGC axons, and the signal intensity decreased slightly within 300 μm from the optic disc, and waned in the peripheral retina (Fig. 4). The annexin-V-A488 staining was easily seen in cSLO but better quantified in retinal whole mount preparations. Compared with the strong axonal staining 7 days after AION, annexin-V-A488-labeling of the axons was relatively sparse and low at day 5 in retinal whole mount preparations and not visualized in cSLO (N = 2; data not shown). Control eyes with no annexin-V-A488 signal were commonly observed, and in some eyes with relatively sparse labeling, we sometimes observed a faint, diffuse blood vessel labeling or punctate background labeling of noncellular debris (Fig. 5B). 

Figure 4

 

In vitro and in vivo images of prominent annexin-V-A488⁺ labeling of radially oriented RGC axons 7 days after AION. (A–C) Representative examples of annexin-V-A488⁺ staining in retinal whole mount preparations. Signals were most intense near the optic nerve head (asterisk). (D) cSLO image of (C).

Figure 5

 

Quantification of the number of annexin-V-A488⁺ axons. (A) Representative example of whole mount retinal preparation 7 days after AION and (B) in control eye from the same animal taken with same imaging parameters. We only counted axons that were easily visible, so the number may be an underestimation. (C) Bar graph showing significantly increased number of annexin-V-A488⁺ axons after AION compared with control eyes, consistent with axonal degeneration after AION (N = 14 eyes per group, P = 0.002).

We quantified the number of annexin-V⁺ axons in AION and contralateral control eyes 1 week after ischemia under masked condition (Fig. 5). We found that there was significantly increased annexin-V-A488 labeling of RGC axons in AION eyes compared with contralateral control eyes (AION: 35 ± 8 annexin-V⁺ processes/eye; control: 6 ± 1 annexin-V⁺ processes/eye; N = 14 for both conditions; P = 0.002) consistent with an increase in axonal degeneration after AION at day 7 (Fig. 5). 

Fragmented Pattern of Axonal Degeneration

Similar to the beads-on-a-string pattern of YFP expression in RGC axons in Thy1-YFP-H mice after ischemia (Figs. 1, 2), there was a strikingly fragmented pattern of annexin-V-A488 labeling of RGC axons at day 7 (Figs. 4, 5). Using ImageJ, we further analyzed the fragmented pattern of 40 annexin-V-A488⁺ processes (Fig. 6). The annexin-V-A488⁺ fragments were short, typically less than 5 μm in length, and occurred at irregular intervals. In contrast to the beads-on-a-string appearance of the YFP⁺ RGC axons, which had some continuous signal along the axons, the annexin-V⁺ fragments were discontinuous. On average, the annexin-V⁺ fragments had intersegment intervals of 20.1 ± 1.4 μm (range, 4–20 μm, N = 40 axons in four eyes) or 5.8 ± 0.4 annexin-V⁺ segments per 100 μm. The intersegment intervals and the number of segments per 100 μm were similar between that of the Thy-1 YFP⁺ axons and the annexin-labeling (P = 0.1 for both). 

Figure 6

 

Analysis of the fragmented pattern of annexin-V-A488 staining. (A) Representative retinal whole mount preparation showing annexin-V-A488-labeling of RGC axons 7 days after AION. There was a fragmented pattern of staining along the axons (arrows). (B, C) show two examples of ImageJ analysis of annexin-V-A488 signal along straightened axons. The fluorescence intensity profile showed short segments of intense labeling at irregular intervals along the axon over 200 μm. (B) Corresponds to the top axon, and (C) corresponds to the bottom axon. The straightened raw data are shown above the traces.

Discussion

Our study of RGC axons after experimental AION using Thy1-YFP-H mice and annexin-V-A488 labeling of cellular energy failure revealed that severe, degenerative axonal changes occurred within 1 week after ischemia. Annexin-V has previously been shown to label dying RGC cell bodies in many conditions,^{6,25,30,32–35,40–42} and our data demonstrated for the first time that annexin-V intensely labeled degenerating RGC axons after AION. Annexin-V staining was well visualized in vivo using cSLO and in vitro in retinal whole mount preparations 7 days after AION, and the staining pattern along degenerating axons was most prominent near the optic nerve head and occurred in a fragmented pattern, consistent with Wallerian-like degeneration. 

After AION, we showed that intraretinal RGC axons undergo degeneration in a matter of days, and previous studies have shown that early axonal damage is also seen in the distal axons in the optic nerve.^19,24 In our study of Thy1-YFP-H retinae, AION led to visible intraretinal axonal stumps within 3 days. This meant some of the axons were essentially severed at the site of ischemia, possibly by a process similar to acute axonal degeneration, which can occur as early as 30 minutes after optic nerve crush, is calcium-dependent, and involves autophagy.⁴³ Acute axonal degeneration is thought to occur suddenly and leads to proximal and distal axonal degeneration,⁴⁴ while Wallerian degeneration occurs in distal, severed axon after an initial delay with beading and swelling of the axon followed by granular disintegration of the axonal cytoskeleton.^7,45,46 Acute axonal degeneration may share a similar mechanism as Wallerian degeneration since acute axonal degeneration is delayed by the Wallerian degeneration slow (Wld^S) mutation.⁴⁴ Three days after AION, the beads-on-a-string appearance of YFP expression in intraretinal RGC axons in Thy1-YFP-H mice was similar to that seen in the optic nerve axons after optic nerve crush, which likely occur in a Wallerian-like degeneration.³⁸ In fact, central and peripheral axons have been shown to undergo Wallerian or Wallerian-like degeneration from various causes and diseases. This type of axonal degeneration is thought to be an active process dependent on calcium and calcium-dependent calpain proteases. It can be delayed by the Wld^S mutation and can occur independently from somatic degeneration.^7,38,47 

The sequence of events in axonal degeneration after AION involved early axonal degenerative changes 3 to 5 days after AION, limited annexin-V labeling of intraretinal RGC axons on day 5, and intense, punctate annexin-V staining of intraretinal RGC axons on day 7. This initial delay and sudden fragmentation and axonal degeneration have been described in Wallerian-like degeneration and, in fact, central axons exhibit more delay than peripheral axons.^7,46 The fragmented pattern of annexin-V staining along the RGC axons we observed has also been seen in Wallerian-like axonal degeneration. A study using Thy1-YFP-H mice documents the fluorescent signal of sciatic-tibial nerves after crush injury and finds a wave of fragmentation progressing retrogradely from the site of injury.⁴⁶ More specifically, the formation of a vacuole devoid of YFP expression precedes fragmentation, suggesting one possible mechanism by which the axons fragment. Several other groups have reported this fragmentation using annexin-V. In hippocampal neuronal culture exposed to β-amyloids, beaded or fragmented neurites as observed by phase photomicrograph are evident in all neurons that have either condensed chromatin or DNA damage, and annexin-V clusters at these beads.⁴⁸ In microscalpel transection of neurites in dorsal root ganglion cultures, Sievers et al.⁴⁹ show that annexin-V-FITC binding occurs within 30 minutes of transection and lasts at least 6 hours. The binding pattern is initially relatively diffuse along the neurites, and, eventually, the pattern of annexin-V label appears more and more fragmented, which correlates with neurite blebbing and fragmentation.⁴⁹ Kim et al.⁵⁰ also report punctate signal of annexin-V derivative pSIVA in NGF-deprived dorsal root ganglion cultures and in transected sciatic nerves in vivo. They attribute this pattern to the irregular expression of phosphatidylserine on the outer leaflet of the cell membranes of apoptotic cells.⁵⁰ 

The prominence of annexin-V staining of the axons compared with that of the soma at day 7 is compatible with the idea that axonal degeneration may occur autonomously and may even slightly precede that of the soma. In rodent photochemical thrombosis model of AION, there is prominent evidence of optic nerve degeneration within the first week,¹⁹ while the first peak of RGC loss occurs around day 10.²⁵ In optic nerve transection, in which the RGC axons are acutely severed, there is also a 6- to 7-day delay between the time of injury and the first visualizations of RGC apoptosis,⁶ so subsequent somatic degeneration appears to involve active signaling possibly initiated by the withdrawal of neurotrophic support.⁷ In a time-lapse imaging study using dorsal root ganglion neurons following NGF-deprivation, Kim et al.⁵⁰ uses an annexin-V-derivative called pSIVA to show that the degenerative process is first seen in the axon and then extends toward and includes the soma over hours. This idea of axonal degeneration prior to somatic degeneration is consistent with the idea that human AION begins with ischemia in the anterior optic nerve just posterior to the lamina cribrosa,^10,12,51 which leads to impaired axonal transport of molecules such as neurotrophins and mitochondria, which are important for RGC survival.⁸ Aside from the temporal separation between axonal and somatic degeneration, several studies have also shown that axonal degeneration can occur as a compartmentalized process distinct from that of the soma^48,52–55 and that somatic degeneration does not automatically occur with axonal degeneration, since Wallerian degeneration of the RGC axons can be suppressed by expression of the Wld^S gene even when severed from the soma.⁴⁷ Taken together, our data are consistent with the idea that after AION, axonal degeneration is an active process distinct from that of the RGC apoptosis. 

Another important finding in our study is the use of annexin-V as a marker for degenerating axons, which adds to the vast literature of the use of annexin-V as a marker for somatic degeneration. Annexin-V staining has been co-localized with other apoptosis markers such as terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL), a marker for apoptosis³⁵; Trypan blue, which labels cells with disrupted membrane^56,57; and propidium iodide (PI),^50,58 a late marker for cell death.^59–62 As a marker for apoptosis, annexin-V is used in various in vivo applications, such as the Detection of Apoptosing Retinal Cells (DARC) with cSLO,^30,31 which has been used to demonstrate the neuroprotective effect of topical CoQ10 following experimental glaucoma.⁴⁰ Because of its low toxicity and capacity to be recombined with radionuclides, recent clinical trial studies have utilized 99m-Technetium-annexin-V (^99mTc-Annexin-V) scintigraphy to label dying cell in vivo in patients with type 1 diabetes mellitus,⁶³ arrhythmogenic right ventricular cardiomyopathy/dysplasia,⁶⁴ and different types of cancer.^65–67 Annexin-V or molecules like it may one day be an important marker of optic nerve injury in patients and help direct therapeutic intervention. 

Further dissection of the timing and mechanism of axonal degeneration will provide important data for future, directed therapy because targeting axonal degeneration may be different from that of the soma, and both axons and cell bodies need to be salvaged since they are both necessary to preserve visual function. Furthermore, better understanding of the changes in the RGC axons, which converge to form the optic nerve, may have implications on white matter tracts throughout the central nervous system and in neurodegenerative diseases such as Alzheimer's disease, Parkinson's disease, multiple sclerosis, and traumatic brain injury. 

Acknowledgments

Supported by the Career Award in Biomedical Sciences from the Burroughs Wellcome Foundation, the Weston Havens Grant, the Center for Biomedical Imaging at Stanford Grant, and the Vice Provost Undergraduate Education Grant from Stanford University (YJL). JHM was supported by the Medical Scholars Program from Stanford University School of Medicine. GHL was supported by the Undergraduate Advising and Research Grant. 

Disclosure: G.H. Lee, None; M.P. Stanford, None; M.A. Shariati, None; J.H. Ma, None; Y.J. Liao, None