Axonal Protection via Modulation of the Amyloidogenic Pathway in Tumor Necrosis Factor–Induced Optic Neuropathy

Abstract

Purpose.: To examine the changes in and localization of phosphorylated presenilin1 (p-PS1) and amyloid precursor protein (APP) in the optic nerve after intravitreal injection of TNF and to investigate the role of γ-secretase in the cleavage of APP in optic nerve degeneration.

Methods.: Groups of rats were euthanatized at 1 or 2 weeks after intravitreal injection of TNF. Levels of p-PS1 protein in the optic nerve were determined by immunoblotting and immunohistochemistry. The localization of APP was determined by immunohistochemistry, and its downstream cleavage was determined by immunoprecipitation using 6E10 antibody followed by immunoblotting with an APP intracellular domain (AICD) antibody. The effect of a γ-secretase inhibitor on TNF-induced optic nerve degeneration was determined by counting the number of axons.

Results.: p-PS1 was increased in the optic nerve after TNF injection and was found to colocalize with vimentin and glial fibrillary acidic protein, markers of astrocytes. Immunoprecipitation using 6E10 antibody followed by immunoblotting with AICD antibody revealed an increase in γ-secretase activation in the optic nerve after TNF injection, which was inhibited by treatment with the γ-secretase inhibitor. Moreover, γ-secretase inhibition significantly prevented the loss of axons in the optic nerve after TNF injection.

Conclusions.: The increase in p-PS1 and activation of γ-secretase in the optic nerve may be associated with TNF-induced axonal degeneration. Modulation of γ-secretase activity may be useful for the treatment of TNF-related optic neuropathy.

Introduction

Previous studies suggested that glaucoma may be correlated with Alzheimer's disease (AD), ^1,2 although this remains controversial. ³ Decreased levels of β-amyloid (Aβ1-42) and increased levels of tau in cerebrospinal fluid (CSF) from AD patients compared with levels in healthy individuals were reported. ⁴ A recent study has demonstrated that CSF with low levels of the Aβ1-42 group had higher tau and phosphorylated (p)-tau and significantly higher whole-brain loss, ventricular expansion, and hippocampal atrophy rate. ⁵ In the eye, it was reported that decreased levels of Aβ1-42 and increased levels of tau were observed in the vitreous from glaucoma patients compared with the control group. ⁶ These decreased levels of Aβ1-42 in fluid support the idea that the deposition of Aβ in neuronal cells is involved in their apoptotic death. In addition, Guo et al. ⁷ demonstrated that Aβ colocalizes with apoptotic retinal ganglion cells (RGCs) in experimental glaucoma and induces significant RGC apoptosis. 

Amyloid precursor protein (APP) can be proteolytically cleaved by β-secretase, generating a short C-terminal fragment (CTFβ) of 99 amino acids. The CTFβ fragment of APP is then cleaved by γ-secretase into an Aβ peptide and a cytosolic APP intracellular domain (AICD) in the amyloidogenic pathway. ⁸ γ-Secretase is a multiprotein complex consisting of presenilins (PS1 and PS2), nicastrin, Aph-1, and Pen-2. ⁹ A previous study demonstrated that PS is responsible for γ-secretase activity and that inhibition of PS1 activity is a potential target for anti-amyloidogenic therapy in AD. ¹⁰ In addition, a recent study has shown that increased expression of PS1 is sufficient to increase γ-secretase activity. ¹¹ Although a beneficial effect of γ-secretase inhibitor has been demonstrated in neurotoxic peptide Aβ-induced RGC-5 cell death, ¹² the role of γ-secretase in optic nerve degeneration remains unclear. 

Tumor necrosis factor (TNF), a cytokine that is synthesized and released from astrocytes and microglia in the central nervous system, has been implicated in the pathogenesis and progression of AD. ^13,14 In the eye, TNF-mediated neurotoxicity has been linked to optic nerve degeneration in glaucoma patients. ^15–17 More recently, a study of the aqueous humor has shown that a significantly higher percentage of patients in a glaucoma group were positive for TNF compared with a cataract group. ¹⁸ In addition, a recent study of the proteomic data from human glaucoma has shown a prominent up-regulation of TNF/TNFR1 signaling in the glaucomatous retina. ¹⁹ An earlier study reported that intravitreal injection of TNF induces optic nerve degeneration in mice. ²⁰ Madigan et al. ²¹ demonstrated that intravitreal injection of TNF causes optic nerve axonal degeneration with little involvement of neural retinal damage in rabbits. We and other groups also demonstrated that intravitreal injection of TNF induces progressive optic nerve degeneration with slow RGC body death in mice ²² and rats. ²³ These observations suggest that the TNF-induced neurodegeneration model may be useful in understanding the mechanism of axonal degeneration of the RGCs. 

The aim of present study is to examine the changes in and localization of p-PS1 and APP in the optic nerve after intravitreal injection of TNF and to investigate the role of γ-secretase in cleavage of APP in optic nerve degeneration. 

Materials and Methods

Animals

Experiments were performed on 50- to 55-day-old male Wistar rats. All studies were conducted in accordance with the Association for Research in Vision and Ophthalmology (ARVO) statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Ethics Committee of the Institute of Experimental Animals of St. Marianna University Graduate School of Medicine. The animals were housed in controlled conditions, with temperature of 23°C ± 1°C, humidity of 55% ± 5%, and light from 6:00 AM to 6:00 PM. 

Administration of TNF

Intravitreal injection of TNF (Sigma-Aldrich, St. Louis, MO) was performed with 120 rats as described previously. ^23–25 Briefly, rats were anesthetized with an intramuscular injection of a mixture of ketamine-xylazine (10 and 4 mg/kg, respectively). A single 2-μL injection of 10 ng TNF in 0.01 M phosphate-buffered saline (PBS, pH 7.40) or simultaneous injection of TNF and BMS299897 (a γ-secretase inhibitor that is dissolved in 1% dimethylsulfoxide and diluted with PBS to 10⁻⁴ M or 10⁻³ M, i.e., 0.102 μg or 1.023 μg) was administered intravitreally into the right eye of an animal under a microscope to avoid lens injury. PBS or vehicle was injected in the contralateral left eye as a control. In the simultaneous injection group, an additional intravitreal injection of BMS299897 alone was administered 3 days after the first injection. BMS299897 alone group was also conducted. The rats were euthanized 1 and 2 weeks after the injection by intraperitoneal injection of an overdose of pentobarbital sodium, followed by enucleation of the eyes. 

Immunoblotting

Ninety-six rats were used for immunoblot analysis as described previously. ^23–25 Briefly, 1 or 2 weeks after intravitreal injection, optic nerves (4-mm lengths) were collected, homogenized in protein extraction buffer (50 mM Tris-HCl, pH 7.4, 1 mM EGTA, 0.001% leupeptin), and then centrifuged at 15,000g for 15 minutes at 4°C. Two optic nerve specimens were pooled for one sample. Protein concentrations were determined using the Bio-Rad Protein Assay kit (Bio-Rad, Hercules, CA). Protein samples (5 μg per lane) were subjected to SDS-PAGE on 10% polyacrylamide gels and transferred to enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech, Piscataway, NJ) membranes. Membranes were blocked with Tris-buffered saline (TBS)–0.1% Tween-20 containing 5% skim milk. Membranes were first reacted with anti-p-PS1 antibody (1:200; Santa Cruz Biotechnology, Santa Cruz, CA), anti-PS1 antibody (1:200; Cell Signaling, Danvers, MA), glial fibrillary acidic protein (GFAP; a marker of astrocyte; 1:1000; DAKO Corporation, Carpinteria, CA), neurofilament-L (a marker of neurons; 1:200; Chemicon International, Temecula, CA), or β-actin (1:5000; Sigma-Aldrich) in TBS containing 5% skim milk. Membranes were then exposed to peroxidase-labeled anti-rabbit IgG antibody (Cappel, Aurora, OH) or peroxidase-labeled anti-mouse IgG antibody (Cappel) diluted 1:5000 in Tween-20 in TBS. Western blots were visualized with an ECL detection system (ECL Plus Western Blotting Detection Reagents, Amersham Pharmacia Biotech). 

Combined Immunoprecipitation–Immunoblotting

Optic nerves (4-mm lengths) were collected 1 week after intravitreal injection. Five optic nerve specimens were pooled for each sample and a total of 20 independent optic nerves were used for each group. The membrane fraction of the optic nerves was extracted using a ProteoExtract Subcellular Proteome Extraction Kit (Calbiochem, La Jolla, CA), according to the manufacturer's protocol. Briefly, the optic nerves were homogenized in extraction buffer and centrifuged at 1000g for 10 minutes at 4°C. The pellets were resuspended in cellular material extraction buffer and centrifuged at 6000g for 10 minutes. The supernatants were stored as membrane protein portions at −80°C until use. Immunoprecipitation was carried out with the Immunoprecipitation Starter Pack (GE Healthcare Bio-Sciences AB, Uppsala, Sweden), according to the manufacturer's instructions. Briefly, the supernatants were incubated with 6E10 antibody (Covance, Emeryville, CA) for 1 hour at 4°C, followed by incubation with Protein G Sepharose 4 Fast Flow for another 1 hour at 4°C. After the samples were washed three times with ice-cold washing buffer and the supernatants were removed by centrifugation at 12,000g for 1 minute, proteins were precipitated. The proteins were isolated from the beads using sample buffer for 4 minutes at 95°C, separated on a 4% to 20% SDS-PAGE gel (Bio-Rad), and then subjected to immunoblotting with AICD antibody (Covance). Duplications of the same gels were also processed for silver staining to confirm that the same amount of protein was present in each lane according to the manufacturer's instructions (Thermo Scientific, Rockford, IL). 

Axon Counting in Optic Nerves

Morphometric analysis of each optic nerve was performed as described previously ^25,26 in samples from 16 rats. Eyes were obtained from the animals 2 weeks after intravitreal injection. Four-millimeter segments of the optic nerves were obtained starting 1 mm behind the globe. These segments of optic nerve were fixed by immersion in Karnovsky's solution for 24 hours at 4°C, processed, and embedded in acrylic resin. Cross sections (1 μm thick) were cut beginning 1 mm from the globe and stained with a solution of 1% paraphenylene-diamine (Sigma-Aldrich) in absolute methanol. For each section, images at the center and at each quadrant of the periphery (approximately 141.4 μm from the center) were acquired with a light microscope (BX51; Olympus, Tokyo, Japan) with a 100× coupled digital camera (MP5Mc/OL, Olympus) and associated QCapture Pro software (version 5.1, QImaging, Surrey, Canada). The acquired images were quantified using the Aphelion image processing software (version 3.2, ADCIS SA and AAI, Inc., Hérouville Saint Clair, France). The number of axons was determined in five distinct areas of 1446.5 μm² each (each quadrant of the periphery in addition to the center; total area of 7232.3 μm² per eye) from each eye. The number of axons per optic nerve was averaged and expressed as number per square millimeter. A minimum of four eyes per experimental condition was used for analysis. 

Immunohistochemistry

Eyes collected 1 week after intravitreal injection were fixed by immersion in 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer for 24 hours, dehydrated, embedded in paraffin, and sectioned (4 μm thick) through the optic disc. For cross sections, 4-mm segments of the optic nerves were obtained starting 1 mm behind the globe. These segments of optic nerve were also fixed by immersion in 4% PFA for 24 hours, dehydrated, and embedded in paraffin. Cross sections (1 μm thick) were cut beginning 1 mm from the globe. Deparaffinized sections were incubated with 1% bovine serum albumin (BSA) and then reacted with primary antibodies against p-PS1 antibody (1:100; Santa Cruz Biotechnology), APP antibody (1:50; Cell Signaling), APP antibody (1:50; Abbiotec, San Diego, CA), TNF receptor 1 antibody (1:50; Novus Biologicals, Littleton, CO), TNF receptor 2 antibody (1:50; Novus Biologicals), vimentin (a marker of astrocyte; 1:50; Chemicon International), GFAP (a marker of astrocyte; 1:100; Sigma-Aldrich), GFAP (1:100; DAKO Corporation), or neurofilament-L (a marker of neurons; 1:100; DAKO Corporation) diluted in 1% BSA overnight at 4°C. Sections were then exposed to the following secondary antibodies: FITC-labeled anti-rabbit antibody (1:100; Cappel), FITC-labeled anti-rat antibody (1:100; Cappel), FITC-labeled anti-mouse antibody (1:100; Cappel), rhodamine-labeled anti-rabbit antibody (1:100; Cappel), or rhodamine-labeled anti-mouse antibody (1:100; Cappel). The samples were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Vectashield with DAPI; Vector Laboratories, Burlingame, CA). Negative controls were performed by replacing the primary antibody with PBS or serum. 

p-PS1-positive cells in optic nerve cross sections were analyzed in the images captured by a fluorescence confocal microscope. Three images were obtained randomly, and the three sections were used for quantification in one optic nerve (nine images in each eye). The acquired images were quantified using Aphelion image processing software, which can also calculate the total image intensity. Data from three sections of each eye were averaged for one eye, and five eyes were examined for each experimental condition. 

Statistical Analysis

Data are presented as mean ± SEM. Differences among groups were analyzed using one-way ANOVA, followed by Scheffe's method or Mann-Whitney's method. A P value of less than 0.05 was considered to represent a statistically significant difference. 

Results

Changes in p-PS1 Protein Level in the Optic Nerve after TNF Injection

Immunoblot analysis showed a band of p-PS1 at 47 kDa in the optic nerve sample. Compared with PBS-treated eyes, an increase in the p-PS1 protein level was observed 1 week after TNF injection (Fig. 1A) when axonal loss was not apparent morphologically. ²³ There was also an increase in the level of p-PS1 protein in the optic nerve at 2 weeks (Fig. 1B) when axonal loss was significant. ²³ Densitometric analysis showed statistically significant increases in levels of p-PS1 protein in the optic nerve both 1 and 2 weeks after TNF injection compared with levels after PBS injection (Fig. 1). 

Figure 1. 

 

Immunoblot analysis of p-PS1 protein levels in the optic nerve (A) 1 week and (B) 2 weeks after intravitreal TNF injection. Bands at 47 kDa corresponding to the molecular weight of p-PS1 were detected in samples from the optic nerve. Data are normalized to total PS1 levels in the same sample and expressed as a percentage of control. Each column represents mean ± SEM; n = 5 per group. *P < 0.05.

Localization of p-PS1 in the Optic Nerve and Increase in p-PS1 Immunoreactivity after TNF Injection

p-PS1 immunoreactivity was localized predominantly within the nuclei in the optic nerve, which means that it is located in glial cells (Figs. 2A–D). Vimentin immunoreactivity was observed around p-PS1-positive nuclei (Fig. 2B, inset). There was an increase in p-PS1 immunoreactivity 1 week after TNF injection (Figs. 2A–E). Quantitative analysis of the intensity showed a significant increase in p-PS1 immunoreactivity in the optic nerve 1 week after TNF injection compared with that after PBS injection (Fig. 2E). GFAP immunoreactivity was also observed around p-PS1-positive nuclei (Fig. 2D, inset). Immunoblot analysis showed an increase in the GFAP protein level 1 week after TNF injection (Fig. 2F). 

Figure 2. 

 

Localization of p-PS1 in an optic nerve cross section 1 week after intravitreal injection of PBS (A, C) or TNF (B, D). (A, B) Double staining for p-PS1 and vimentin revealed that p-PS1-positive nuclei are colocalized in vimentin-positive astrocytes in optic nerves. (C, D) Double staining for p-PS1 and GFAP revealed that p-PS1-positive nuclei are colocalized in GFAP-positive astrocytes in optic nerves. Scale bar, 25 μm (A–D) and 13.5 μm for inset images (B, D). (E) Intensity of p-PS1 immunoreactivity in an optic nerve cross section 1 week after intravitreal injection. Data are expressed as pixel volume. Each column represents mean ± SEM; n = 5 per group. *P < 0.05. (F) Immunoblot analysis of GFAP protein levels in the optic nerve 1 week after intravitreal injection. Data are normalized to β-actin levels in the same sample and expressed as a percentage of control. Each column represents mean ± SEM; n = 4 per group. *P < 0.05.

Double immunostaining showed that p-PS1-immunopositive cells were not associated with axons in cross sections (Figs. 3A, B). Although axonal loss was not apparent 1 week after TNF injection (Fig. 3B), a decrease in the neurofilament L protein level was observed at 2 weeks (Fig. 3C). 

Figure 3. 

 

Localization of p-PS1 in an optic nerve cross section 1 week after intravitreal injection of PBS (A) or TNF (B). (A, B) Double staining for p-PS1 and neurofilaments revealed no colocalization in optic nerves. Scale bar, 25 μm. (C) Immunoblot analysis of neurofilament-L protein levels in the optic nerve 2 weeks after intravitreal injection. Data are normalized to β-actin levels in the same sample and expressed as a percentage of control. Each column represents mean ± SEM; n = 4 per group. *P < 0.05.

Localization of APP in the Optic Nerve

Since some reports demonstrated that APP colocalizes in RGCs, ²⁷ we expected the colocalization of APP and RGC axons, which are neurofilaments. However, double immunostaining showed no association between APP and neurofilaments in optic nerve cross sections (Fig. 4A). Instead, APP immunoreactivity was mainly colocalized with vimentin-positive glial cells in optic nerve cross sections (Fig. 4B). In addition, some colocalization with APP and p-PS1 was observed after TNF injection (Fig. 4C). 

Figure 4. 

 

Localization of APP in the optic nerve. (A) Double staining for APP and neurofilaments revealed no colocalization in cross sections of optic nerves in PBS-treated eyes. (B) Double staining for APP and vimentin revealed substantial colocalization in cross sections of optic nerves in PBS-treated eyes. (C) Double staining for APP and p-PS1 revealed some colocalization in cross sections of optic nerves in TNF-treated eyes. Scale bar, 25 μm.

Localization of TNF Receptors in the Optic Nerve

To investigate whether TNF receptors exist in glial cells in the rat optic nerve, we performed immunohistochemical examination. TNF-R1 immunoreactivity was observed in GFAP-positive cells in longitudinal sections of optic nerves in TNF-treated eyes (Fig. 5A), consistent with previous results showing that TNF-R1 is primarily located in GFAP-positive astrocytes. ¹⁶ A similar result was obtained for TNF-R2 immunoreactivity after TNF injection (Fig. 5B). 

Figure 5. 

 

Localization of TNF receptors in the optic nerve. (A) Double staining for TNF-R1 and GFAP revealed some colocalization in longitudinal sections of optic nerves in TNF-treated eyes. (B) Double staining for TNF-R2 and GFAP revealed some colocalization in longitudinal sections of optic nerves in TNF-treated eyes. Scale bar, 25 μm.

Acceleration of CTFβ Cleavage in the Amyloidogenic Pathway in the Optic Nerve after TNF Injection

A predominant high molecular-weight protein pulled down by 6E10 antibody was detected in silver staining. Immunoprecipitated proteins using 6E10 antibody followed by immunoblotting using AICD antibody revealed cleavage of the CTFβ fragment of APP, showing a decrease in CTFβ in the optic nerve 1 week after TNF injection compared with vehicle injection (Fig. 6A). This decrease was markedly prevented by treatment with BMS299897, a γ-secretase inhibitor (Fig. 6A). Silver staining confirmed that almost the same amount of protein was present in each lane (Fig. 6B). Densitometric analysis showed a significant decrease in levels of CTFβ protein in the optic nerve 1 week after TNF injection compared with that after vehicle injection and this decrease was prevented by treatment with BMS299897 (Fig. 6C). There was no significant difference in CTFβ cleavage in vehicle and with BMS299897 treatment alone. 

Figure 6. 

 

Immunoprecipitation of optic nerve proteins using 6E10 antibody followed by immunoblotting using AICD antibody. (A) CTFβ protein levels in the optic nerve 1 week after intravitreal injection. (B) Silver staining was used to confirm that the same amount of proteins was present in each lane. (C) Densitometric analysis of CTFβ protein levels. Data are normalized to silver staining and expressed as a percentage of control. Each column represents mean ± SEM; n = 4 per group. *P < 0.05 compared with vehicle injection; †P < 0.05 compared with TNF injection.

BMS299897 Prevented TNF-Induced Optic Nerve Degeneration

We previously demonstrated myelin fragmentation and other degenerative changes involving large-caliber axons 2 weeks after TNF injection. ^23–25 Consistent with our previous results and compared with control eyes (Fig. 7A), substantial degenerative changes were seen in the optic nerve axons 2 weeks after TNF injection (Fig. 7B) in histologic examination. In contrast, BMS299897-treated eyes showed noticeably attenuated effects with better-preserved nerve fibers (Figs. 7C, D). Morphometric analysis showed that there was an approximately 40% loss of axons in TNF-induced optic nerve degeneration (Fig. 7E). On the other hand, BMS299897 treatment significantly prevented TNF-induced axonal loss in a dose-dependent manner (Fig. 7E). BMS299897 10⁻⁴ M and 10⁻³ M exerted 38.85% and 60.18% protection, respectively (Fig. 7E). The decrease in axon number was more prominent in axons with diameters less than 2.0 μm (Fig. 7F). Although 10⁻⁴ M BMS299897 exerted protection only in axons with diameters less than 1.0 μm, 10⁻³ M BMS299897 exerted protection in all axons with diameters less than 2.0 μm (Fig. 7F). 

Figure 7. 

 

Effect of BMS299897 in TNF-induced optic nerve degeneration. Histologic examination findings 2 weeks after (A) vehicle injection, (B) 10 ng TNF injection, (C) 10 ng TNF + 10⁻⁴ M (0.102 μg) BMS299897 injection, or (D) 10 ng TNF + 10⁻³ M (1.023 μg) BMS299897 injection. Scale bar, 10 μm. Effect of BMS299897 on axon numbers (E) and distribution (F) of axon diameters in the optic nerves. Each column represents mean ± SEM. *P < 0.05 compared with vehicle injection; †P < 0.05 compared with TNF injection.

Discussion

The present study demonstrated the presence of p-PS1 in optic nerve glial cells and found that intravitreal TNF injection increased the p-PS1 level in the optic nerve. The immunoprecipitation data showed an increase in γ-secretase activity induced by TNF in the optic nerve. Furthermore, a γ-secretase inhibitor significantly prevented the loss of axons in the optic nerve after TNF injection. These results suggest that modulation of γ-secretase activity can be useful for the treatment of TNF-related optic nerve degeneration. 

We observed a significant increase in p-PS1 in the optic nerve 1 and 2 weeks after TNF injection. It was reported that the dual-specificity tyrosine(Y)-phosphorylation-regulated kinase 1A (Dyrk1A) phosphorylates PS1 and that this phosphorylation increases γ-secretase activity, suggesting that up-regulated Dyrk1A may accelerate AD pathogenesis through PS1 phosphorylation. ²⁸ In addition, a recent study has shown that PS1 phosphorylation increased the Aβ 42/40 ratio and that PS1 phosphorylation was enhanced in the human AD brain. ²⁹ These findings suggest that PS1 phosphorylation in optic nerve astroglial cells may play a pivotal role in TNF-induced optic nerve degeneration. Most recently, TNF/TNFR signaling has been reported to mediate immune/inflammatory responses in astrocytes during glaucomatous neurodegeneration. ³⁰ In the present study, activation of astrocytes was observed in the optic nerve after TNF injection. The TNF injection model has a limitation in mimicking the glaucoma model as it shows axonal degeneration in 2 weeks without further degeneration over longer periods of time, ²³ while the hypertensive glaucoma model shows progressive axonal degeneration. ³¹ A previous review demonstrated that glial production of TNF is increased in the glaucomatous optic nerve and TNF-mediated neurotoxicity is a component of the neurodegeneration in glaucoma. ³² Recently, we have also observed an increase in the TNF level in the optic nerve in the hypertensive glaucoma model. ³³  

APP was reported to accumulate in the optic nerve in rat ³⁴ and mouse ^35,36 glaucoma models. In the retina, APP immunostaining was observed in Müller glia, ³⁷ RGCs, horizontal cells, cone bipolar cells, and amacrine cells, ³⁸ suggesting that APP is present in both glial cells and neurons in the retina. In the present study, double immunostaining showed that APP is colocalized with vimentin-positive glial cells but not with neurofilaments in the normal optic nerve. These findings are consistent with previous result demonstrating that APP-immunopositive cells were astrocytes in the subventricular zone in the adult rat forebrain. ³⁹ APP is a transmembrane protein, whereas neurofilament L is an axoplasmic protein, and therefore the fact that APP does not colocalize with neurofilament L does not exclude the possibility that APP is present in axons. Therefore, it is possible that relatively little axonal transport of APP occurs under normal conditions but that some APP may be transported into the optic nerve axons. ⁴⁰  

Our present immunoprecipitation data showed that there is a decrease in CTFβ in protein pulled down by 6E10 in the optic nerve after TNF injection. This means that cleavage of the CTFβ fragment of APP into a fragment that binds to amino acids 3 to 8 of APP (EFRHDS) and the AICD occurs in TNF-treated samples. Therefore, it is reasonable to hypothesize that TNF injection leads to PS1 phosphorylation and subsequently increases γ-secretase activity in the optic nerve. In contrast, BMS299897, a γ-secretase inhibitor, substantially increased the CTFβ level, indicating that it inhibited the cleavage of the CTFβ fragment of APP. These findings are consistent with previous reports demonstrating that increases in β-cleaved CTF were observed in the brains of BMS299897-treated mice. ^41,42  

In the present study, BMS299897 administration significantly prevented the axonal loss induced by TNF. Protective effects of several γ-secretase inhibitors have been demonstrated against certain types of cell death. For example, it was shown that a γ-secretase inhibitor produced by EMD (Darmstadt, Germany) provided significant protection from apoptosis induced by the overexpression of AβPP in CHO cells. ⁴³ L-685,458, another γ-secretase inhibitor, was also shown to recover cell viability and inhibit caspase-3/7 activation in SH-SY5Y cells after exposure to staurosporine, thapsigargin, or H₂O₂. ⁴⁴ In vivo, the γ-secretase inhibitors N-[N-(3,5-difluorophenacetyl)-l-ananyl]-S-phenylglycine t-butyl ester (DAPT) and ((S,S)-2-[2-(3,5-difluorophenyl)acetylamino]-N-(5-methyl-6-oxo-6,7-dihydro-5H-dibenzo[b,d]azepin-7yl)propionamide reduced brain damage and neurological deficits in a mouse cerebral ischemia–reperfusion model. ⁴⁵ In addition, it was reported that DAPT accelerates axon regeneration after axotomy in mature Caenorhabditis elegans neurons. ⁴⁶ These findings and our current results suggest that γ-secretase inhibitors have protective effects not only against cell body death but also against axonal degeneration. 

The current immunohistochemical study demonstrated that both p-PS1, which is the main component of γ-secretase, and APP locate in glial cells in the optic nerve. p-PS1 is localized in the nuclear membrane and is suggested to play a role in DNA binding activity and to interact with several proteins. ⁴⁷ Since we separated cytosol (including axoplasm) and membrane proteins, the cleavage of APP by the activation of γ-secretase occurs mostly in glial membranes. It is interesting to note that a previous study demonstrated that the inhibition of glial γ-secretase stimulates the myelination of RGC axons in a coculture system. ⁴⁸ Taken together, these findings suggest that the inhibition of glial γ-secretase may contribute to axonal protection. Further study will be needed to clarify the detailed mechanism by which the inhibition of γ-secretase alters axon–glia interactions in the optic nerve. 

In summary, our findings suggest that p-PS1 is present in optic nerve glial column nuclei, and the increase in p-PS1 and activation of γ-secretase in the optic nerve may be associated with TNF-induced axonal degeneration. The inhibition of γ-secretase may have a potential protective effect in TNF-related optic neuropathy. 

Footnotes

 Supported by Grants-in-Aid No. 24592683 (YK) and No. 23792016 (YM) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

Footnotes

 Disclosure: K. Kojima, None; Y. Kitaoka, None; Y. Munemasa, None; S. Ueno, None