Spatiotemporal Characterization of Optic Nerve Degeneration after Chronic Hypoperfusion in the Rat

Abstract

Purpose.: Permanent, bilateral occlusion of the common carotid arteries (2VO) is an established model of chronic hypoperfusion. Previous studies have noted the vulnerability of the optic nerve (ON) to 2VO; however, little information is available regarding the spatiotemporal pattern of axonal degeneration and the accompanying glial cell responses. The present study was conducted to investigate these topics.

Methods.: At various times after surgery, ONs were removed for mRNA or Western blot analysis or to be processed for histology and immunohistochemistry.

Results.: 2VO precipitated an infarct within the proximal ON, whereas the retinal ganglion cells, unmyelinated axons of the optic nerve head, and the distal portion of the ON were initially preserved. The onset of degeneration was rapid, with disturbances in fast axonal transport occurring by 6 hours, damage to the axonal cytoskeleton in the proximal ON detectable by 24 hours, and complete axonal loss within the infarcted area manifest within 3 days. Wallerian degeneration of the distal segment of the ON proceeded thereafter, with almost complete loss of the ON axonal cytoskeleton evident by 30 days. Degradation of the axonal cytoskeleton was accompanied by increasing microglial activation and proliferation and a delayed infiltration of macrophages into the lesion site. Robust and persistent upregulation of stress proteins by astrocytes and oligodendrocytes, which correlated with axonal damage, was found throughout the ON after 2VO. Extracellular matrix remodeling was evident in the optic nerve head and proximal ON.

Conclusions.: 2VO causes rapid degeneration of the ON, with some similarity to rodent ischemic optic neuropathy.

Disruption in blood supply to the retina and/or optic nerve (ON) is a leading cause of blindness worldwide, manifesting in a variety of ischemia-like diseases, including arterial and venous occlusions, diabetic retinopathy, and ischemic optic neuropathies. ¹ Furthermore, an increasing body of evidence indicates that prolonged vascular insufficiency at the optic nerve head (ONH) is one of the key triggers of the cascade of events that leads to progressive retinal ganglion cell death in glaucomatous optic neuropathy. ^2,3 Experimentally, analysis of the myriad effects of ischemia on the retina and ON has mainly been explored in acute models of ischemia, of which the most common method involves elevating intraocular pressure above systolic blood pressure for a defined period. ^1,4 Although data gained from these studies have shed much light on the detrimental processes that occur during ischemia-reperfusion, the acute, global ischemia created is of most relevance clinically to central retinal artery or ophthalmic artery occlusions. Chronic ischemia/oligemia has received much less attention experimentally, despite its ostensibly greater significance in several clinical conditions. 

The principal means by which ocular researchers have induced a chronic, ischemic-like injury is via permanent, bilateral occlusion of the common carotid arteries (2VO) in rats. Investigation of the changes in cerebral blood flow during 2VO have indicated the existence of three successive phases: an acute, ischemia/hypoxia-like phase lasting 2 to 3 days; a chronic, oligemic phase lasting up to 3 months; and, finally, a return to normal baseline flow. ^5,6 The extent of hypoperfusion in the retina and ON during 2VO is unclear. In the only study performed to date, Spertus et al., ⁷ using fluorescein angiography, showed marked changes in the retinal and ONH vasculature 2 days after 2VO. These observations are consistent with the acute, ischemia/hypoxia-like phase in cerebral blood flow, but later time points were not analyzed and no actual blood flow measurements were made. With regard to tissue damage, initial results suggest that 2VO-induced neuronal injury in the retina is of a nature similar to that in the brain, comprising glial cell activation, but minimal neurodegeneration. ^8–10 Subsequent studies, however, have reported extensive death of ganglion cells by 7 days, with other neuronal classes affected only at much later time points. ^11–14 The primary site of injury of ganglion cells during 2VO is unresolved: Yamamoto et al. ¹⁴ suggested that ganglion cell loss is triggered by retinal ischemia, whereas Stevens et al. ¹² contended that ON ischemia is the main causative factor. Irrespective of the primary site of injury, the procedure appears to provide a valid model for investigating processes involved in ischemia-like ganglion cell death and ON degeneration. Indeed, it can be argued that the model is, in one respect, superior to high intraocular pressure–induced ischemia and even to chronic paradigms of raised intraocular pressure, since these techniques elicit direct deformation of the ONH in addition to causing vascular insufficiency, hence making it problematic to differentiate whether the resultant pathologic effect arises from mechanical or vascular effects. It follows that 2VO may have greater similarity to normotensive glaucoma. 

Recent work, most notably that of Yamamoto et al., ¹⁴ has provided information concerning 2VO-induced histopathologic and immunohistochemical changes in the retina, but relatively little is known about the sequence of events in the ONH and ON. As such, the principal purpose of the present study was to explore the effect of 2VO on axonal, glial, and extracellular matrix (ECM) elements of the ONH and ON. The results will facilitate a greater understanding of the 2VO model of injury and more generally of the effect of vascular hypoperfusion on these important structures. 

Materials and Methods

Animals and Procedures

All experiments conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. This project was approved by the Animal Ethics Committee of the Institute of Medical and Veterinary Science, Adelaide, and conformed to the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes, 2004. Adult Sprague-Dawley rats (250–350 g) were housed in a temperature- and humidity-controlled room with a 12-hour light/12-hour dark cycle and provided with food and water ad libitum. Experimental procedures were performed with the animals under isoflurane anesthesia. Anesthesia was initiated with 3% isoflurane, then maintained throughout the operation with 1.5% to 2% isoflurane. Body temperature was maintained at 37 ± 0.5°C with a thermoregulatory heating unit connected to a rectal probe. A ventral incision was made. The common carotid arteries were identified via a midline incision and blunt dissection of neck muscles. They were carefully separated from the carotid sheath and vagus nerve and ligated with silk sutures. Sham-surgery animals underwent the same operation without occlusion of the vessels. At various times after surgery, the rats were killed by cardiac perfusion with physiological saline under terminal anesthesia (intraperitoneal injection of 100 mg/kg ketamine and 10 mg/kg xylazine). Eyes and optic nerves were then removed for mRNA analysis, Western blot analysis, or immunohistochemistry. For the purposes of evaluation, ONs were essentially divided into three segments: the ONH, comprising unmyelinated axons in the retina and neck region; the proximal ON (nearer to the eye); and the distal ON (nearer to the brain). These latter two segments were each approximately 3 mm in length. The mortality rate during and after surgery was 32%. The total number of rats analyzed at each time point was as follows: n = 5 (6 hours), n = 6 (1 day), n = 12 (3 days), n = 12 (7 days), n = 4 (21 days), n = 5 (30 days). 

Immunohistochemistry and Histology

After enucleation, the tissues were fixed in 10% neutral-buffered formalin and processed for routine paraffin-embedded sectioning. ONs were embedded longitudinally or transversely. Some globes with ONs attached were embedded sagittally. In all cases, 5-μm serial sections were cut and stained for routine histologic analysis with hematoxylin and eosin (H&E), in a standard protocol. For immunohistochemistry, the tissue sections were deparaffinized, rinsed in 100% ethanol, and treated for 30 minutes with 0.5% H₂O₂ in methanol, to block endogenous peroxidase activity. Antigen retrieval was achieved by microwaving the sections in 10 mM citrate buffer (pH 6.0). For localization of the ECM proteins collagen I, collagen VI, and laminin, the sections were digested an additional 3 minutes with trypsin (0.25 g/L), to further unmask the antigen sites. The sections were then blocked in PBS containing 3% normal horse serum, incubated overnight at room temperature in primary antibody (containing 3% normal horse serum), and incubated consecutively with biotinylated secondary antibody (1:250; Vector Laboratories, Peterborough, UK) and streptavidin-peroxidase conjugate (1:1000; Pierce, Rockford, IL). Color development was achieved with 3′-,3′-diaminobenzidine. The sections were counterstained with hematoxylin, dehydrated, and mounted. The specificity of antibody staining was confirmed by incubating adjacent sections with isotype controls (mouse IgG1 and IgG2a, 50878 and 553454; BD Australia, North Ryde, NSW, Australia) for monoclonal antibodies or normal rabbit/goat serum for polyclonal rabbit/goat antibodies. In addition, positive control labeling was performed in appropriate rat brain tissue sections. Primary antibody details are given in Table 1. 

Table 1.

 

Antibodies Used for and Immunohistochemistry Western Blot Analysis

Target	Host	Clone/Cat No.	Dilution	Source*
Actin	Mouse	AC-15	1:20000 (w)	Sigma
βAPP	Mouse	22C11	1:1250	Gift, C. Masters
BDNF	Rabbit	N-20	1:5000	Santa Cruz
T cells, CD3	Rabbit	A0452	1:3000	Dako
Collagen I	Goat	AB758	1:200	Millipore
Collagen VI	Rabbit	Ab6588	1:1000	Abcam
αB-crystallin	Mouse	SPA-222	1:4000, 1:1000 (w)	Stressgen
ED1	Mouse	MCA341	1:500	Serotec
Hsp-27	Rabbit	SPA-801	1:2500, 1:1000 (w)	Stressgen
GFAP	Rabbit	Z0334	1:5000 (f)	Dako
	Mouse	M761	1:200 (f)	Dako
Hsp-32	Rabbit	SPA-895	1:10,000	Stressgen
Hsp-72	Mouse	SPA-810	1:200	Stressgen
Iba1	Rabbit	019-19741	1:50,000	Wako
Laminin	Rabbit	AT-2404	1:3000	EY Labs
MBP	Rabbit	A-0623	1:5000	Dako
MHC II	Mouse	OX-6	1:400	Serotec
nestin	Mouse	MAB353	1:1000	Millipore
npNFH	Mouse	SMI-32	1:15,000	Sternberger
Olig-2	Rabbit	AB9610	1:16,000	Millipore
pNFH	Mouse	SMI-31	1:200,000	Sternberger
NFL	Mouse	NR4	1:3000, 1:20000 (w)	Sigma
Tau	Goat	C-17	1:2000, 1:250 (f)	Santa Cruz
β3-Tubulin	Mouse	TU-20	1:1000, 1:2500 (w)	Millipore

 (f), two-step fluorescent staining; (w), Western blot analysis.

*  Sigma-Aldrich, Sydney, NSW, Australia; Colin L. Masters, Department of Pathology, The University of Melbourne, Melbourne, VIC, Australia; Santa Cruz Biotechnology, Santa Cruz, CA; Dako, Sydney, NSW, Australia; Millipore, North Ryde, NSW, Australia; Abcam, Cambridge, UK; Stressgen, Victoria, BC, Canada; Serotec, Oxford UK; Wako Chemicals, Osaka, Japan; EY Labs, San Mateo, CA; Covance-Sternberger, Princeton, NJ.

To verify which glial cell type(s) in the optic nerve was responsible for expression of the stress proteins nestin, Tau, heat shock protein (Hsp)-27, and αB-crystallin after 2VO, double labeling of these proteins was performed with markers for astrocytes (GFAP), oligodendrocytes/oligodendrocyte precursor cells (Olig-2), and microglia (Iba1). Visualization of one antigen was achieved with a three-step procedure (primary antibody, biotinylated secondary antibody, and streptavidin-conjugated AlexaFluor 488), whereas the second antigen was labeled by a two-step procedure (primary antibody and secondary antibody conjugated to AlexaFluor 594). In summary, the sections were prepared as described earlier, except for the omission of the endogenous peroxidase block, then incubated overnight at room temperature in various combinations of mouse anti-nestin, goat anti-Tau, mouse anti-αB crystallin, rabbit anti-Hsp-27, rabbit or mouse anti-GFAP, rabbit anti-Olig-2, and rabbit anti-Iba1. On the next day, sections were incubated with the appropriate biotinylated secondary antibody (1:250) for the three-step procedure plus the correct secondary antibody conjugated to AlexaFluor 594 (1:250; Invitrogen, Mulgrave, VIC, Australia) for the two-step procedure for 30 minutes, followed by streptavidin-conjugated AlexaFluor 488 (1:500; Invitrogen) for 1 hour. Both of these steps were performed in a darkened environment. Sections were then mounted in antifade medium (ProLong Gold; Invitrogen) and examined under a confocal fluorescence microscope. 

For histologic analysis of optic nerve cross sections, intraorbital (termed proximal) and intracranial (termed distal) optic nerves from saline-perfused sham-surgery and 2VO rats were fixed by immersion in 2.5% glutaraldehyde with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 24 hours at 4°C. Tissues were then placed in 2% osmium tetroxide in saline overnight and washed with cacodylate buffer at room temperature. Subsequently, they were dehydrated in graded alcohols and embedded in epoxy resin for transverse sectioning. Transverse sections (0.5 μm) were cut on an ultramicrotome, mounted on glass slides, and stained for myelin with 1% toluidine blue. 

Terminal Deoxynucleotidyl Transferase–Mediated, dUTP Nick-End Labeling Assay

TUNEL assays were performed essentially as described previously. ¹⁵ In brief, the sections were deparaffinized, rehydrated, treated with proteinase K, and washed in distilled water. Endogenous peroxidases were inactivated by incubation in H₂O₂. The sections were equilibrated in TdT buffer, before incubation in the same buffer containing TdT (0.15 U/μL) and biotin-16-dUTP (10 μM) for 60 minutes at 37°C. The reaction was terminated by two washes in saline sodium citrate solution. The sections were rinsed, and nonspecific binding sites were blocked by washing in 2% bovine serum albumin. The sections were incubated with streptavidin-peroxidase conjugate (1:1000) for 30 minutes at 37°C, rinsed in PBS, and developed for color with 3′-,3′-diaminobenzidine, as for histology. 

Evaluation of Immunohistochemistry

The presence or absence of an infarct in the proximal ON was documented in each rat. In those rats with infarcts, the presence of ED1-positive macrophages within the lesion was also noted. Axonal precursor protein (APP) accumulation in the ONH as a result of disrupted axonal transport was assessed by basic semiquantitative grading of immunostaining (0, minimal APP accumulation; 1, presence of swollen, bulbous APP-positive axons; and 2, intense, widespread APP positivity). 

For quantification of immunostaining in the distal ON, immunostained sections, each expressing a homogenous and representative level of immunoreactivity, were photographed at 200× magnification. They were then imported into Image-J 1.42q (http://rsb.info.nih.gov/ij/; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD), where they underwent color deconvolution to separate diaminobenzidine reaction product from hematoxylin counterstain. ¹⁶ The images were subsequently imported into analysis software (analySIS FIVE; Olympus Soft Imaging Solutions, Münster, Germany). After determination of background levels, the tissue was analyzed with regard to the specifically stained area in pixels, by using the built-in function of the software. Statistical analysis of immunoreactivity in the distal ON was performed by ANOVA with post hoc Student's t-test and Bonferroni correction for multiple comparisons. 

Retinal ganglion cell loss in the retina 7 days after 2VO was assessed by measurement of the number of Brn-3- and islet-1-immunolabeled cells in the ganglion cell layer. Evaluations were performed under a light microscope equipped with an ocular micrometer bar. To minimize sampling errors, all analyses were performed on sections taken at the level of the optic nerve head. The cells were counted to a distance of 2 mm either side of the optic nerve head. Statistical analysis was performed by Student's t-test (sham versus 2VO). 

Electrophoresis and Western Blot Analysis

Proximal ONs from sham, 3-day 2VO, and 7-day 2VO rats (n = 4–6) were sonicated in freshly prepared 20 mM Tris/HCl buffer (pH 7.4) containing 2 mM EDTA, 0.5 mM EGTA, 1 mM dithiothreitol and the protease inhibitors, phenylmethyl-sulfonyl fluoride (0.1 mM), leupeptin (50 μg/mL), aprotinin (50 μg/mL), and pepstatin A (50 μg/mL), as well as a phosphatase inhibitor cocktail. An equal volume of sample buffer (62.5 mM Tris/HCl [pH 7.4], containing 4% SDS, 10% glycerol, 10% β-mercaptoethanol, and 0.002% bromophenol blue) was added, and the samples were boiled for 3 minutes. An aliquot was taken at this stage for determination of protein content. Electrophoresis of samples was performed on 12% polyacrylamide gels containing 0.1% SDS. Samples were transferred onto PVDF membranes overnight and then blocked with Tris-buffered saline containing 0.1% (vol/vol) Tween-20 and 5% (wt/vol) nonfat dried skimmed milk. Blots were probed with antibodies to actin, NFL, β₃-tubulin, αB-crystallin, and Hsp-27 (Table 1) for 3 hours at room temperature, and appropriate secondary antibodies conjugated to biotin and streptavidin-peroxidase conjugate were subsequently and sequentially used. Blots were developed with a 0.016% solution of 3-amino-9-ethylcarbazole in 50 mM sodium acetate (pH 5) containing 0.05% Tween-20 and 0.03% H₂O₂. Images were acquired from labeled blots with a flatbed scanner (CanoLide; Canon, Tokyo, Japan) and analyzed for densitometry with image-analysis software (Eastman Kodak, Rochester, NY). Densitometry values were then normalized for actin. Statistical analysis was performed by ANOVA followed by a post-hoc Student's t-test with Bonferroni correction for multiple comparisons. 

Real-Time RT-PCR

Reverse transcription–polymerase chain reaction (RT-PCR) studies were performed as described previously. ¹⁷ In brief, ONs were carefully dissected, total RNA was isolated, and first-strand cDNA was synthesized from 0.5 μg DNase-treated RNA. To assess genomic DNA contamination, we performed additional reactions in which the reverse transcriptase enzyme was omitted. PCR products were not observed for any of the primer pairs tested with these samples. 

The primer pairs were designed from sequences contained in the GenBank database (http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information [NCBI], Bethesda, MD) using the primer design software Primer3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi ¹⁸ ) and were selected to amplify sequences that spanned at least one intron. Primer sequences were analyzed for T_m (melting temperature), secondary structure, and primer-dimer formation with primer-analysis software (NetPrimer; Premier Biosoft, Palo Alto, CA) and verified for their specificity to the target sequence by using the BLAST database search program (www.ncbi.nlm.nih.gov/BLAST; provided in the public domain by NCBI). Some primer pairs were taken from previous publications—specifically those for GAPDH ¹⁹ and for IL-1β, IL-6, and TNF-α. ^20,21 All primers were optimized for annealing temperature and MgCl₂ concentration before they were used in assays. The oligonucleotide primer sequences, their annealing temperatures, and MgCl₂ concentrations are shown in Table 2. 

Table 2.

 

Primer Sequences for mRNAs Amplified by Real-Time RT-PCR

mRNA	Primer Sequences	Product Size (bp)	Mg2+ Conc.	Annealing Temp. (°C)	GenBank Accession No.
GAPDH	5′-TGCACCACCAACTGCTTAGC-3′	87	mM	63	M19533
	5′-GGCATGGACTGTGGTCATGAG-3′
NFL	5′-ATGGCATTGGACATTGAGATT-3′	105	4 mM	63	AF031880
	5′-CTGAGAGTAGCCGCTGGTTAT-3′
Thy1	5′-CAAGCTCCAATAAAACTATCAATGTG-3′	83	3.5 mM	63	X03150
	5′-GGAAGTGTTTTGAACCAGCAG-3′
IL-1β	5′-CACCTCTCAAGCAGAGCACAG-3′	79	3.5 mM	63	M98820
	5′-GGGTTCCATGGTGAAGTCAAC-3′
IL-6	5′-TCCTACCCCAACTTCCAATGCTC-3′	79	3.5 mM	64	EO2522
	5′-TTGGATGGTCTTGGTCCTTAGCC-3′
iNOS	5′-CTGGAGGTGCTGGAAGAGTT-3′	226	3.5 mM	63	L12562
	5′-CTTCGGGCTTCAGGTTATTG-3′
TNF-α	5′-AAATGGGCTCCCTCTCATCAGTTC-3′	111	3.5 mM	63	X66539
	5′-TCTGCTTGGTGGTTTGCTACGAC-3′

Real-time PCR reactions were performed in 96-well optical reaction plates using the cDNA equivalent of 20 ng total RNA for each sample in a total volume of 25 μL containing 1× SYBR Green PCR master mix (Bio-Rad, Regents Park, NSW, Australia) and forward and reverse primers at a final concentration of 400 nM. The thermal cycling conditions were 95°C for 3 minutes and 40 cycles of amplification comprising 95°C for 12 seconds, annealing temperature for 30 seconds, and 72°C for 30 seconds. After the final cycle of the PCR, primer specificity was checked by the dissociation (melting) curve method according to the manufacturer's protocol (Bio-Rad). In addition, specific amplification was confirmed by electrophoresis of PCR products on 3% agarose gels. PCR assays were performed (IQ5 icycler; Bio-Rad), and all samples were run in duplicate. 

To allow a comparison to be made between the levels of expression of target mRNAs in the ONs of sham-surgery and 2VO animals, results obtained from the real-time PCR experiments were quantified with the Relative Expression Software Tool (REST; produced by a consortium and available in the public domain at (http://www.gene-quantification.net/rest.html) and statistical significance was determined using the Pair-wise Fixed Reallocation Randomization Test. ²² All values were normalized to that of the endogenous reference gene GAPDH, and threshold cycles were calculated (IQ5 icycler Software; Bio-Rad). The results are expressed as the mean ± SEM. 

Results

Degenerative Changes in the ON in the First 24 Hours after 2VO

In sham-surgery rats, APP immunoreactivity was localized to the cell bodies of ganglion cells in the retina (data not shown), but minimal staining was associated with ganglion cell axons (Figs. 1a, 1c). At 6 hours after 2VO, a striking accumulation of APP was visible within the ganglion cell axons in the ONH (Fig. 1b) and the transition region of the proximal ON (Fig. 1d), indicative of disrupted fast axonal transport. No alterations of axonal cytoskeletal components, including neurofilaments and tubulins, were noted in the ONH or ON (Figs. 1e, 1f) at this time point. Moreover, there was no consistent indication of microglial or astroglial activation in the ONH or ON (data not shown). 

Figure 1.

 

Changes in the ONH and ON at 6 hours after 2VO. In sham-surgery rats, low-intensity APP immunoreactivity was observed within the ONH (a) and ON (c). At 6 hours after 2VO, a striking accumulation of APP was evident within swollen axons in the ONH (b, arrows) and proximal ON (d). In contrast, no alteration in NP-NFH was observed at this time point (e, f). Scale bar: (a, b), 60 μm; (c–f), 30 μm.

At 24 hours after 2VO, the ganglion cell somata were still essentially functional at 24 hours after 2VO, as indicated by the normal expression of Thy1 mRNA in the retina (see Fig. 7f), yet intense, widespread APP immunoreactivity was observed in the ONH (Fig. 2a). Accumulation of brain-derived neurotrophic factor (BDNF) was also evident in the prelaminar ONH (data not shown), together with subtle abnormalities in neurofilament heavy (NFH), particularly of the nonphosphorylated isoform (NP-NFH, Figs. 2c, 2d). No TUNEL-positive cells (Fig. 2b) or phagocytic microglia were visible in the ONH, but ED1-positive perivascular macrophages were evident in the vitreous humor adjacent to the ONH (Fig. 2f). Astrocytes in the ONH showed upregulated expression of the intermediate filaments GFAP (data not shown), nestin (Figs. 2g, 2h) and Hsp-27 (Figs. 2i, 2j). No astroglial expression of two other highly inducible Hsps, Hsp-32 and -70, was detected, although a strong upregulation of Hsp-32 (Figs. 2k, 2l), as well as GFAP (data not shown) and nestin (Figs. 2g, 2h), was seen in retinal Müller cells. 

Figure 2.

 

Changes in the ONH at 24 hours after 2VO. Intense APP staining was observed in the ONH at 24 hours after 2VO (a, arrow), accompanied by subtle beading of NP-NFH (d, arrows) compared with the sham-surgery control (c). No TUNEL-positive cells were apparent (b). ED1-positive microglia were absent within the ONH of sham-surgery (e) and 2VO (f) rats; however, 24 hours after 2VO, macrophages were observed in the vitreous humor adjacent to the vasculature of the ONH (f, arrows). Astrocytes within the ONH of sham-surgery rats expressed relatively low levels of the stress proteins nestin (g) and Hsp-27 (i). At 24 hours after 2VO, a marked increase in immunoreactivity to both antigens was noted in the ONH, which was limited to astrocytes for Hsp-27 (j), but also included Müller cells for nestin (h, arrows). Hsp-32 immunoreactivity was absent in the ONH of sham-surgery rats (k). At 24 hours after 2VO, Hsp-32 was upregulated in Müller cells, but not astrocytes (l, arrows). Scale bar, 60 μm.

No axonal cytoskeletal changes or glial cell activation was noted in the distal ON (data not shown) at 24 hours after 2VO, but considerable disruption of the axonal cytoskeleton was noted in the proximal ON (Fig. 3). Damage varied between animals, but was always concentrated in the transition region. In some rats, injury was limited to axon transport disruption and minor neurofilament abnormalities (Figs. 3d–f). In other rats, small infarcts were observed (Figs. 3g–l) that featured loss of neurofilaments and tubulins and TUNEL-labeled cells, presumably glial cells. In the vicinity of the lesion, Iba1-labeled microglia displayed some retraction of processes indicative of activation (Fig. 3k), whereas astrocytes displayed a moderate increase in nestin expression (Fig. 3l). No consistent induction of oligodendroglial stress proteins, specifically the small Hsp αB-crystallin and the microtubule-associated protein Tau was observed at 24 hours (data not shown), and no change in expression of Hsp-32 or -70 was found in the proximal or distal ON at any point after 2VO (data not shown). 

Figure 3.

 

Changes in the proximal ON at 24 hours after 2VO. In sham-surgery rats, the axonal cytoskeleton (a), Iba1-labeled ramified microglia (b), and nestin-positive astrocytes (c) appeared normal. After 2VO, considerable damage was visible, which was concentrated in the transition region of the ON. The extent of injury varied among the animals. In some rats, injury was limited to APP-immunoreactive axonal swelling (d) with disruption of neurofilament-labeled axon bundles (e) and the occasional TUNEL-positive cell (f, arrows). In other rats, small infarcts were observed (g–l, arrows) that featured APP accumulation (g), neurofilament (h), and β₃-tubulin (i) loss; frequent TUNEL-labeled cells (j); and loss of microglial (k) and astrocytic (l) immunolabeling. Activated glial cells were apparent at the boundaries of the infarct (k, l). Scale bar, 30 μm.

Real-time PCR performed on ON samples demonstrated that 24 hours after 2VO, there was significant upregulation in the mRNA levels of three proinflammatory genes, interleukin (IL)-1β, IL-6, and inducible nitric oxide synthetase (iNOS), compared with that in sham-surgery ONs (Table 3). An increase in the level of tumor necrosis factor (TNF)-α mRNA was also measured, but did not reach statistical significance. 

Table 3.

 

Effect of 2VO on Expression of Inflammatory Marker mRNAs in the Optic Nerve

mRNA Analyzed	mRNA Expression Relative to Shams
	1 Day after Surgery	7 Days after Surgery
IL-1β	3.25 ± 0.82*	1.59 ± 0.26
IL-6	82.93 ± 43.77*	1.29 ± 0.45
TNF-α	1.61 ± 0.38	1.27 ± 0.38
iNOS	14.28 ± 7.99*	3.80 ± 2.94

 In all cases, values (represented as the mean ± SEM) are normalized for GAPDH, where n = 6.

*   P < 0.01 by Pair-wise Fixed Reallocation Randomization Test. ²²

Degenerative Changes in the ON 3 Days after 2VO

Real-time PCR showed that approximately 50% of ganglion cell somata were still viable at 3 days after 2VO, as indicated by the retinal expression levels of Thy1 and NFL mRNAs (see Fig. 7f). In the proximal ON, toluidine blue-stained transverse sections exposed disorganization, axonal swelling, and some changes in myelin compared with sham-surgery ONs (Figs. 4a, 4b), whereas H&E and immunostaining techniques revealed the presence of an extensive infarct (Figs. 4c–o), just distal to the ONH, which featured numerous TUNEL-positive cells (Figs. 4d, 4e) and degradation of the axonal cytoskeleton (Figs. 4f–i). Degenerative axonal changes were apparent throughout the remainder of the proximal segment (Figs. 4g–i). Astroglial, microglial, and oligodendroglial survival within the necrotic lesion was negligible. In the ONH (data not shown) and at the penumbra, however, intense glial activity was observed (Figs. 4j–o). The astrocytes were strongly labeled by antibodies directed against the intermediate filament nestin and Hsp-27, oligodendrocytes displayed intense Tau staining and were positive for αB-crystallin, and numerous Iba1-positive microglia with an amoeboid morphology were visible. Of interest, relatively few ED1-positive microglia were seen at the border of the infarct, and no ED1-labeled macrophages had migrated into the ischemic area. In the distal ON (Fig. 5), axonal cytoskeleton abnormalities and glial cell activity were much reduced compared with the proximal segment, but were still markedly higher than in sham-surgery rats. TUNEL-labeled cells were only infrequently detected. 

Figure 4.

 

Degenerative changes in the proximal ON at 3 days after 2VO. Analysis of toluidine blue–stained, transverse sections of ONs from 2VO rats revealed the presence of axonal swelling, large vacuoles, myelin unraveling (arrowhead), and general disorganization in the proximal (b) compared with the control (a) ON. H&E staining (c) of the ONH and proximal ON typically revealed the presence of a substantial infarct (arrows). TUNEL-positive cells were abundant at the proximal (d, arrow) and distal (e, arrowheads) boundaries of the lesion, but not within the ONH (d). In longitudinal sections of the proximal ON, the infarct was clearly visible (e–o, arrows: infarct boundary) and featured some regions of intense APP staining (f), with obvious loss of neurofilaments (g, h) and β₃-tubulin (i). Swelling, vacuolization, and disruption of the axonal cytoskeleton was present in the remainder of the proximal segment (g–i). Microglia, astrocytes, and oligodendrocytes were largely absent from the infarct, as evidenced by staining for Iba1 (j), ED1 (k), nestin (l), Hsp-27 (m), Tau (n), and αB-crystallin (o). At the boundary of the infarct, however, numerous Iba1- and several ED1-labeled microglia were present (j, k). Astrocytes showed strong labeling for nestin (l) and Hsp-27 (m), and oligodendrocytes stained intensely for Tau (n, arrowheads) and αB-crystallin (o). Scale bar: (a, b) 15 μm; (c) 300 μm; (e–o) 60 μm.

Figure 5.

 

Degenerative changes in the distal ON at 3 days after 2VO. Degenerative changes appeared throughout the distal segment (a, arrowhead), but were less pronounced than in the proximal ON, and TUNEL-positive cells were rarely detected (b). APP-positive axons were absent (c). NP-NFH was particularly affected with widespread, intense, beaded immunostaining (e). β₃-Tubulin (f) showed less evidence of pathologic change than did P-NFH (d). Numerous, activated Iba1-positive microglia (g) were evident, but no ED1-labeled cells (h). Astroglial expression of nestin (i) and Hsp-27 (j) and oligodendroglial expression of αB-crystallin (l) were all elevated compared with the sham control, but were substantially lower than in the proximal ON, whereas oligodendroglia were not intensely stained by Tau (k). Scale bar, 30 μm.

To definitively identify which glial cell types express the stress proteins nestin, Tau, Hsp27, and αB-crystallin, we performed double-labeling experiments with markers specific for astrocytes (GFAP), oligodendrocytes (Olig-2), and microglia (Iba1). Selected results are presented in Figure 6. As expected, nestin and Tau co-localized with GFAP and Olig-2, respectively, indicating their association with astrocytes and oligodendrocytes (data not shown). The most striking observation was that the small Hsps, Hsp27 and αB-crystallin, showed only slight co-localization (Figs. 6a–c). Hsp27 immunoreactivity overlapped with that of GFAP (Figs. 6d–f) and nestin, indicating its presence within the astrocytes, whereas αB-crystallin was expressed in Tau- (Figs. 6g–i) and olig-2-positive oligodendrocytes. 

Figure 6.

 

Delineation of ON cell types expressing the stress proteins Hsp-27 and αB-crystallin at 3 days after 2VO by double-labeling immunofluorescence with glial cell–specific markers. Hsp-27 (a, green) and αB-crystallin (b, red) were predominantly localized to discrete population of cells, as shown in the merged image (c), where minimal overlapping fluorescence is visible. Hsp-27 immunoreactivity (e, red) co-localized with GFAP (d, green) as revealed in the merged image (f). αB-Crystallin (h, red) expression was observed in many of the same cells as Tau (g, green), as shown in the merged image (i). Scale bars, 20 μm.

Degenerative Changes in the ON 7 Days after 2VO

In the retina, a dramatic loss of ganglion cells had occurred by 7 days after 2VO, as indicated by the number of immunolabeled ganglion cells (Figs. 7a–e) and the levels of mRNAs encoding ganglion cell–specific markers (Fig. 7f). 

Figure 7.

 

Effect of 2VO on retinal ganglion cell (RGC) survival. In sham-surgery retinas, Brn-3 immunoreactivity (a) was localized specifically with the RGCs, whereas islet-1 (c) immunoreactivity was associated with RGCs and bipolar cells. Seven days after 2VO, the majority of the RGCs were lost, as indicated by the reduced number of cells immunoreactive for Brn-3 (b) and islet-1 (d). Scale bar: 30 μm. (e) Quantification of Brn-3 and islet-1 immunolabeling in sham-surgery and 2VO retinas. For each retina, cells were counted to a distance of 2 mm either side of the ONH. Data are expressed as the mean ± SEM (n = 12; **P < 0.01, by unpaired t-test). (f) Expression of RGC-specific NFL and Thy1 mRNAs in 2VO retinas relative to the sham-surgery control. Data (expressed as mean ± SEM) are normalized for GAPDH (n = 6–12; *P < 0.05, **P < 0.01 by Pair-wise Fixed Reallocation Randomization Test. ²² ).

In the ONH, there was no indication of necrosis (Fig. 8a), but robust microglial (Fig. 8b) and astroglial (Fig. 8c) activity was visible, together with a reduction in neurofilaments and tubulins (data not shown). Transverse sections taken through the proximal ON of sham-surgery (Figs. 8d–f) and 2VO (Figs. 8g–i) rats demonstrated that the entire cross section of the nerve was affected, with minimal residual axonal cytoskeleton (Fig. 8g) and an absence of viable astrocytes (Fig. 8i). ED1-positive macrophages had accumulated throughout the necrotic area (Figs. 8b, 8h, 8k). No myeloperoxidase-labeled neutrophils and very few CD3-positive T cells were found in this area (data not shown). At the distal boundary of the infarct (Figs. 8j–l), neurofilament retraction bulbs and swelling characteristic of Wallerian degeneration were visible (Fig. 8j). These effects were not present at 3 days. 

Figure 8.

 

Degenerative changes in the ONH (a–c) and proximal ON (d–l) at 7 days after 2VO. Unlike the proximal ON, the morphology of the ONH remained largely intact, with no evidence of necrosis (a); nevertheless, activated microglia (b) were present within the ONH, and dense astroglial nestin labeling was apparent (c). The beginning of the infarct was clearly visible within the transition region of the proximal ON (a–c). Transverse sections taken through the proximal ON of sham-surgery (d–f) and 2VO (g–i) rats demonstrated that the entire cross section of the nerve was affected, with minimal residual neurofilament labeling (g), numerous ED1-positive macrophages (h), and an absence of viable astrocytes (i, blood vessels were nestin-positive, however). (j–l) Distal boundary of the infarct. On the infarct side, macrophages were abundant (k). Distal to the lesion, neurofilament retraction bulbs and swelling, characteristic of Wallerian degeneration, were visible (j), together with intense astroglial nestin expression (l). Scale bar: (a–i) 60 μm; (j–l) 30 μm.

In the distal ON, axonal degenerative changes were more advanced than at 3 days. A clear loss of neurofilament (Fig. 9a) and β₃-tubulin (Fig. 9b) axonal fibers was observed. MBP immunolabeling was abundant, but displayed abnormalities (Fig. 9d) compared with sham-surgical animals (Fig. 9c). Microglia were numerous and, unlike at 3 days, were ED1-positive and showed presentation of MHC class II antigen (Figs. 9e–g). Astrocytes and oligodendrocytes displayed intense staining for stress proteins (Figs. 9i–l). Real-time PCR demonstrated that the mRNA levels of the proinflammatory genes IL-1b, IL-6, and iNOS in the ON had returned to baseline by 7 days after 2VO (Table 3). 

Figure 9.

 

Degenerative changes in the distal ON at 7 days after 2VO. By 7 days after 2VO, an obvious loss of axonal cytoskeletal components had occurred (a, b). Decreased labeling for P-NFH (a) and β₃-tubulin (b) was clearly evident. Myelin basic protein (d, arrow) immunoreactivity showed disorganization and swellings compared with sham rats (c). MBP (d, arrow) immunoreactivity showed disorganization and swelling. Numerous, Iba1-, ED1-, and MHC class II-positive microglia were present (e–g), but only a few T cells, labeled by CD3, were detectable (h). Astrocytes throughout the distal ON labeled strongly for nestin (i) and Hsp-27 (j). Some, but not all, oligodendrocytes displayed intense Tau staining (k), but expression of αB-crystallin by oligodendrocytes was uniformly robust (l). Scale bar, 30 μm.

Degenerative Changes in the ON 14 to 30 Days after 2VO

The major feature of ON degeneration between 7 days and 1 month after 2VO was a gradual loss of the axonal cytoskeleton in the ONH and distal ON. At 14 days, some axonal fibers were still detectable in the nerve fiber layer and laminar region of the ONH (Figs. 10a, 10b). This was not the case at 1 month (data not shown). Likewise, there was a systematic reduction in the axonal cytoskeleton in the distal ON, such that by 30 days, only sparse fragments of β₃-tubulin and neurofilament labeling were visible (Figs. 10e, 10f). Microglial activity throughout the ON increased over time: ED1-positive macrophages were still visible in the infarct area at 30 days, and, compared with earlier time points, there was a dense concentration of phagocytic microglia in the distal ON (Figs. 10g, 10h). There was some reduction and a marked abnormality in MBP staining by 30 days (Fig. 10i). No obvious astroglial loss was detected during this period; however, by 30 days, TUNEL-positive nuclei were noted (Fig. 10l). 

Figure 10.

 

Degenerative changes in the ONH and proximal ON 14 days (a–d) and distal ON 30 days (e–l) after 2VO. Fourteen days after 2VO, some residual NFL (a) and β₃-tubulin (b) immunoreactivity was still detectable within the retinal nerve fiber layer and ONH neck (arrows), but minimal axons persisted in the remainder of the, largely necrotic, proximal ON, which was filled with ED1-positive macrophages (c, arrow). Nestin immunoreactivity revealed viable astrocytes within the ONH, but only blood vessels in the infarct zone (d, arrow). Thirty days after 2VO, the distal ON was almost devoid of axonal cytoskeleton, as shown by the presence of a few immunoreactive fragments of P-NFH (e) and β₃-tubulin (f). Iba1 (g) and ED1 (h) immunostaining revealed a high density of activated, amoeboid microglia. MBP (i) immunoreactivity was reduced in quantity and clumped and swollen in appearance. A significant number of astrocytes and oligodendrocytes persisted, as shown by labeling for nestin (j) and Tau (k), respectively, although the presence of TUNEL-positive nuclei (l) indicated ongoing glial cell death. Scale bar: (a–d), 60 μm; (e–l), 30 μm.

Quantification of ON Changes after 2VO

Several approaches were used to quantitatively evaluate the spatiotemporal pattern of changes in the ON that occur after 2VO, the results of which are shown in Table 4 and Figure 11. The overall findings reflect the qualitative observations documented in the previous sections and reveal a good consistency between animals at each time point. 

Table 4.

 

Summary of Optic Nerve Data at Various Times after 2VO

	Sham	6 Hours	1 Day	3 Days	7 Days	30 Days
APP grade (ONH)	0.0 ± 0.0	1.2 ± 0.4	1.5 ± 0.5	1.8 ± 0.4	0.3 ± 0.5	0.0 ± 0.0
Presence of infarct (proximal ON)	0%	0%	40%	92%	92%	100%
Macrophage infiltration (proximal ON)	0%	0%	0%	0%	92%	100%
NFL immunoreactivity (distal ON)	149.0 ± 8.9	ND	ND	141.1 ± 9.5	74.6 ± 33.4*	1.3 ± 0.9*
ED1 immunoreactivity (distal ON)	0.1 ± 0.1	ND	ND	0.2 ± 0.1	2.5 ± 1.2*	16.4 ± 9.0*
Nestin immunoreactivity (distal ON)	1.7 ± 0.9	ND	ND	7.0 ± 4.4†	95.0 ± 21.5*	83.4 ± 33.2*
MBP immunoreactivity (distal ON)	172.9 ± 3.1	ND	ND	ND	158.3 ± 6.2*	110.8 ± 37.7*

 APP immunoreactivity in the ONH was assessed by basic semiquantitative grading: 0, minimal APP accumulation; 1, presence of swollen, bulbous APP-positive axons; and 2, intense, widespread APP positivity. Immunoreactivity in the distal ON was quantified and expressed as area in pixels (×10⁴). All values are presented as mean ± SD, where n = 6 (sham), n = 5 (6 hours), n = 5 (1 day), n = 6 (3 days), n = 6 (7 days), and n = 5 (30 days). Statistical analyses of immunoreactivity in the distal ON were performed by ANOVA followed by a post hoc Student's t-test with Bonferroni correction for multiple comparisons. ND, not determined.

 Significant differences against shams are indicated by * P < 0.01 and

†   P < 0.05.

Figure 11.

 

Expression of NFL, β₃-tubulin, Hsp27, and αB-crystallin in the proximal ON of sham-surgery rats, and 3 and 7 days after 2VO, as evaluated by Western blot analysis. (a) Representative blots from three animals in each group are shown together with the molecular masses of the bands. (b) Densitometry measurements normalized to actin. ***P < 0.01, by one-way ANOVA followed by a post hoc Student's t-test with Bonferroni correction (n = 4–6).

ECM Remodeling after 2VO

In the ONH of normal rats, the ECM components collagen I, collagen VI, and laminin were restricted to laminar beams and pial septa (Figs. 12a, 12c, 12e, 12f). After 2VO, deposition of all three ECM components occurred within the ONH and proximal ON, in regions of axonal cytoskeleton denudation (Figs. 12b, 12d, 12g–l). Abnormal ECM labeling was minimal before 1 week, but increased progressively thereafter and formed a lattice across the ON. Also of note, was a reduction in volume of the proximal ON, which was clearly evident after 21 days (Figs. 12k, 12l). 

Figure 12.

 

ECM remodeling in the ON after 2VO. In sham-surgery rats, antibodies directed against the ECM components collagen VI (a, e, col6) and laminin (c, f, lamin) showed strong labeling associated with the laminar beams and pial septa of the ONH. After 2VO, there was a gradual increase in deposition of collagens I and VI and laminin within the ONH and proximal ON, which coincided with complete degeneration of the axonal cytoskeleton (g, h, arrows). This was first evident at 7 days (b, d), was pronounced by 14 days (g–j), and was extensive by 21 days (k, l). Scale bar: (a–d) 30 μm; (e–h) 150 μm; (i–l) 60 μm.

Discussion

Previous studies have noted the vulnerability of the ON to chronic hypoperfusion induced by 2VO ^5,12,23–28 ; however, little information has been available regarding the spatiotemporal pattern of axonal degeneration in the ON and the accompanying glial cell responses. In the present study, we sought to examine the pattern and the response. The major finding is that the ON was not uniformly affected. 2VO precipitated an infarct within the proximal ON, just distal to the ONH, whereas the unmyelinated axons of the ONH and the distal portion of the ON were relatively preserved, at least initially. The onset of degeneration was rapid, with disturbances in fast axonal transport occurring as soon as 6 hours after 2VO. Induction of proinflammatory cytokines and damage to the axonal cytoskeleton in the proximal ON were detectable by 24 hours, and complete loss of axons within the infarcted area manifested within 3 days. Essentially, 2VO caused a vascular axotomy of the ON, resulting in Wallerian degeneration of the ON. 

It has been known for some time that degenerative events after 2VO are more advanced in the ON than in other central nervous system (CNS) white matter tracts ^24,27–29 and that they are apparent within the first few days. ^12,25,29 The use of immunohistochemistry in the present study provided a more sensitive way of observing the early stages of pathogenesis than did histologic stains, such as Klüver-Barrera and silver impregnation, which were used in the other studies. The earliest indicator of ON injury was APP accumulation in the ONH and proximal ON. APP is synthesized by retinal ganglion cells ³⁰ and is transported in an orthograde fashion along the ON by fast axonal transport. Abnormal APP localization is routinely used in brain studies as a surrogate marker of disrupted fast axonal transport and early axonal injury, but surprisingly, the methodology has rarely been exploited by researchers studying ON damage, although in a recent study Goldblum et al. ³¹ showed accumulation of APP in the ONH of rats with an inherited form of glaucoma. It is reasonable to infer that APP accumulation in the prelaminar ONH merely reflects a build-up of the protein within the axons before the obstruction. Two observations, however, suggest that the ONH itself is also a site of axonal disruption. First, APP immunostaining along the ON was not uniform, but was concentrated at two locations: the prelaminar ONH and the incipient lesion; second, APP-labeled axons in the ONH featured bulbs and swelling, which are indicative of injury. It can be concluded that 2VO also caused energy-dependent axon transport failure in the ONH, a deduction supported by the build-up of another axonally transported protein, BDNF, in the same location. The lack of intense APP immunoreactivity in the remainder of the ON, at both early and late time points, is consistent with published data ²⁹ and indicates an absence of distal lesions. 

Prior work has identified the utility of SMI-32, which labels nonphosphorylated/dephosphorylated NFH, for early visualization of degenerating axons. ^32–35 In those studies, the pattern of immunolabeling changed from one consisting of light, uniform staining within axons to one featuring axonal beading, swelling, and spheroids. The results prompted the suggestion that dephosphorylation of neurofilaments is an early step in axonal degeneration. We found similar results in the present study, with marked alterations in SMI-32 staining in the distal ON evident before obvious loss of the axonal cytoskeleton. Degeneration of the axonal cytoskeleton in the distal ON after 2VO was effectively complete by 30 days, as evidenced by a lack of immunolabeling for neurofilaments and tubulins, at which point only limited loss of MBP had occurred. The persistence of myelin after 2VO is consistent with findings in other studies describing Wallerian degeneration in the ON. ^36,37 Clearance of myelin debris during Wallerian degeneration in the CNS is protracted compared with that in peripheral nerves, because of a combination of minimal phagocytosis by oligodendrocytes, failure of microglia to develop into fully phagocytic cells, and restricted infiltration of hematogenous macrophages. ³⁸ Both of these latter factors were evident in the present study. First, despite the presence of a large, necrotic infarct in the proximal ON by 3 days after 2VO, most of the microglia in the ischemic penumbra and the distal ON expressed neither MHC class II nor the phagolysosomal antigen ED1 at this time point; second, there was no macrophage infiltration into the infarct until 7 days after 2VO, and no macrophage, and negligible T cell, penetration into the distal ON at any time point. Nevertheless, by 7 days after 2VO, microglia in the distal ON were immunopositive for both ED1 and MHC class II, and throughout the period of axonal cytoskeletal degradation they continued to proliferate and adopt a more phagocytic morphology. 

Axonal degeneration is accompanied by ECM remodeling. Work by Johnson et al. ^39,40 has shown that in a rat model of glaucoma, deposition of collagens and laminin occurred within the neck and transition regions of the ONH in areas normally occupied by nerve fiber bundles. They showed that the higher the intraocular pressure and the longer its duration, the greater the loss of axonal fibers and the more extensive the deposition of ECM components. Broadly similar results were obtained in the present study; however, laminin and collagen deposition after 2VO occurred earlier than in experimental glaucoma and extended beyond the ONH into the myelinated region of the ON. This result is not unexpected, because of the more rapid loss of the axonal cytoskeleton after 2VO compared with experimental glaucoma. Conversely, less ECM deposition occurred in the neck region of the ONH after 2VO than in the glaucoma model. Again, this most likely reflects the relative sparing of nerve fibers in the ONH after 2VO, compared with glaucoma. 

In contrast to the substantial body of evidence that has accumulated concerning macroglial changes in models of retinal degeneration, the macroglial response during ON injury has received little attention. In the present study, we analyzed glia in the ONH and the proximal and distal ON for indications of cellular stress, with special attention paid to the induction of Hsps. Hsps are a family of highly conserved proteins that function as molecular chaperones. In conditions of cellular stress, synthesis of certain Hsps is subject to dynamic regulation, during which time they play a key role in preventing denaturation and aggregation of proteins and in facilitating refolding of abnormally folded proteins. ⁴¹ In effect, Hsps serve as biomarkers to identify stress specificity. Information on Hsp expression by glial cells in the ON is currently lacking. Using single- and double-labeling immunohistochemistry, we evaluated the cellular expression of four inducible Hsps: Hsp-70, Hsp-32 (also termed heme oxygenase-1), and the closely related small Hsps, Hsp-27 and αB-crystallin. Hsp-70 was not detected in glial cells in the ONH or ON after 2VO, a result that is consistent with the predominantly neuronal induction of the protein in the brain after ischemia. ⁴² Of interest, Hsp-32 also was not observed in the ON glial cells, although it was rapidly induced in Müller glial cells of the retina. In contrast, robust and persistent upregulation of both Hsp-27 and αB-crystallin was found throughout the ON after 2VO. Induction of both Hsps correlated with axonal damage, such that labeling was strongest and observed soonest at the margins of the infarct. Of interest was the finding that astroglia in the retina, ONH and ON persistently expressed Hsp-27, but seemingly not αB-crystallin, whereas the reverse was true of oligodendroglia. Differential expression of small Hsps by astrocytes and oligodendrocytes has been noted previously in the brain. ^43–45  

One function of small Hsps in stressed cells is to stabilize the cytoskeleton. This effect occurs through their interaction with microtubules, microfilaments, and intermediate filaments, ⁴⁶ and obvious changes in astroglial intermediate filaments and oligodendroglial microtubules were noted in the ON after 2VO. The most striking alteration concerned the type VI intermediate filament nestin, which is highly expressed in immature, dividing cells, is largely absent in terminally differentiated cells, but can be re-expressed during neurodegeneration. ⁴⁷ The present results show that nestin was persistently upregulated by astrocytes in the ONH and throughout the ON. As with Hsp-27, astroglial nestin expression was most intense and induced earliest, in the proximal ON, and then was gradually upregulated through the distal ON. Unlike astrocytes, oligodendrocytes are devoid of intermediate filaments, but do contain a network of microtubules and express low levels of the microtubule-associated protein Tau. ⁴⁸ Rapid accumulation of Tau within oligodendrocytes has been demonstrated in different models of middle cerebral artery occlusion. ^33,49,50 It is unclear whether increased Tau immunoreactivity is an early pathologic event or an endogenous survival factor, but it is considered a specific marker of oligodendrocyte injury and has been used to assess oligodendrocyte status in neuroprotection studies. ^51–53 The present results show that increased Tau immunolabeling is also a hallmark of oligodendrocyte injury within the ON. Tau staining was intense at the margins of the infarct at 3 days after 2VO, but oligodendrocytes in the distal ON were only strongly Tau-positive at later time points, as Wallerian degeneration progressed. 

In their characterization of the retinal response to 2VO, Yamamoto et al. ¹⁴ showed a drastic loss of ganglion cells by 7 days. They argued that the primary trigger was ischemia at the level of the cell body, since retinal stress markers were upregulated within 24 hours. Results from the present study advocate damage to the axon as a major contributory factor to death. Essentially, somatodendritic injury, such as that caused by excitotoxicity, results in complete dysfunction of the ganglion cell body with an acute reduction in mRNA synthesis. ^54,55 In contrast, axonal injury, induced for example by optic nerve transection, causes delayed death of ganglion cells, with an initial preservation in the levels of ganglion cell–specific markers, ⁵⁶ and an induction of stress proteins, such as Hsp-27. ⁵⁷ In the present study, we showed a gradual reduction in the mRNA levels of Thy1 and NFL after 2VO, continuing synthesis of APP, as evidenced by a gradual build-up of the protein in the ONH in the first 3 days and an induction of Hsp27 expression by ganglion cells (data not shown). The combined data indicate that, in many ganglion cells, the somata, unlike the axons, remain functional for some time after 2VO and that the model has greater similarity to optic rather than retinal neuropathies. Nevertheless, there was some reduction in ganglion cell mRNA synthesis 1 day after 2VO, and we have visualized retinal neuronal death using a fluorescent neuronal cell viability marker (Fluoro-Jade C; Millipore) at 3 days after 2VO ¹⁵ ; thus, retinal ischemia may well play a significant role. 

Glaucomatous optic neuropathy (glaucoma) represents a family of neurodegenerative diseases characterized by structural damage to the ON at the level of the ONH and the slow, progressive death of retinal ganglion cells. The pathogenesis of glaucoma is poorly understood, with debate centering on the relative contribution of mechanical and vascular influences. The vascular hypothesis proposes that insufficient perfusion of the ONH is the principal trigger for ON degeneration and is most persuasively advocated in regard to normotensive glaucoma. ^3,58 We had anticipated that 2VO, by reducing ONH perfusion without causing mechanical injury via raised intraocular pressure, would prove a valid experimental model of normotensive glaucoma; however, the combined results of the present study indicate that 2VO produces an injury phenotype with arguably greater similarity to the experimental rodent models of anterior ischemic optic neuropathy developed by Bernstein et al., ^59–61 albeit with a vastly greater lesion size. Changes analogous to those witnessed in the anterior ischemic optic neuropathy models include rapid axonal damage in the proximal ON followed by microglial activation and macrophage infiltration, delayed retinal ganglion cell loss, and oligodendrocyte stress, demyelination, and Wallerian degeneration distal to the lesion. ^59–61  

Footnotes

 Supported by Grant 565202 from the National Health and Medical Research Council (NH&MRC) and by the Ophthalmic Research Institute of Australia.

Footnotes

 Disclosure: G. Chidlow, None; M.C. Holman, None; J.P.M. Wood, None; R.J. Casson, None

The authors thank Andreas Ebneter, Mark Daymon, and Jim Manavis for helpful discussions and expert technical assistance.