Proteomic study on neurite responses to oxidative stress: search for differentially expressed proteins in isolated neurites of N1E-115 cells

Reactive oxygen species attack several living organs and induce cell death. Previously, we found axonal/dendrite degeneration before the induction of cell death in hydrogen peroxide-treated neuroblastoma: N1E-115 cells and primary neurons. This phenomenon may be connected with membrane oxidation, microtubule destabilization and disruption of intracellular calcium homeostasis. However, its detailed mechanisms are not fully understood. Here, we identified proteins after treatment with hydrogen peroxide using isolated neurites by liquid chromatography-matrix-assisted laser desorption/ionization-time of flight/time of flight analysis. Twenty-one proteins were increased after treatment with hydrogen peroxide. Specifically, 5 proteins which were secretogranin-1, heat shock protein family D member 1, Brain acid soluble protein 1, heat shock 70-kDa protein 5 and superoxide dismutase 1, were identified of all experiments and increased in isolated neurites of hydrogen peroxide-treated cells compared to the controls. Furthermore, secretogranin-1 and heat shock protein family D member 1 protein expressions were significantly increased in normal aged and Alzheimer’s transgenic mice brains. These results indicate that secretogranin-1 and heat shock protein family D member 1 might contribute to reactive oxygen species-induced neurite degeneration. Both proteins have been related to neurodegenerative disorders, so their study may shed light on neurite dysfunction.

Introduction R eactive oxygen species (ROS) attack living tissues and induce cell death. (1,2) To prevent ROS-induced cell death, organisms have antioxidant defense systems such as antioxidant enzymes [superoxide dismutase (SOD), catalase (CAT), glutathione peroxidase (GPx), etc.] and compounds (vitamins, polyphenols, etc.). However, these defense systems gradually attenuate with senescence. Accumulations of age-related oxidative damages derived from the collapse of redox balance increase the risks of development and progression of severe diseases such as Alzheimer's (AD), Parkinson's, Huntington's, and other neurodegenerative disorders. (3,4) Based on the free radical theory of aging hypothesized by Harman in 1956, (5) the relationship between agerelated oxidative damage and brain dysfunction has been a subject of much study. The brain is particularly vulnerable to ROS, given that its ratio of oxygen consumption is about 20%. (6) Polyunsaturated fatty acids, such as docosahexaenoic acid and eicosapentaenoic acid, are present in brain cell membranes and are easily oxidized compared to other fatty acids. (7) If neurons receive sufficient damage from ROS, it induces cell death. The alteration and dis-sociation of neuronal networks by cell death, is difficult to recover. Therefore, to prevent ROS-related neuronal cell death, it is important to find early signs of neuronal changes prior to cell death induction.
Previously, we found that treatment with a low concentration of hydrogen peroxide induced axonal and dendrite degeneration. (8) The neurites showed abnormal morphologies, such as fragmentation, shrinkage, and beading, a well-known marker of axonal degeneration. The number of dendrites significantly decreased in hydrogen peroxide-treated primary granule cells. (9) Axons and dendrites play a crucial role in neurotransmission in the central nervous system, and several lines of evidence demonstrate a relationship between axonal degeneration and neurodegenerative disorders. (10) However, axons and dendrites can contract and expand easily, and the detailed mechanism of neurite degeneration has not yet been elucidated. For these reasons, we are focusing on neurite functions. To clarify the mechanism of hydrogen peroxideinduced neurite degeneration in neuronal cells, changes in cytoskeletal and cytoskeletal-related protein expressions have been measured. The ratios of cleaved and phosphorylated forms of collapsing response mediator protein (CRMP)-2, which is a microtubule-related protein, are significantly enhanced in hydrogen peroxide-treated neurons. (8,11,12) The ratio of microtubule-associated protein-light chain 3 II (MAP-LC3 II), which is marker of autophagy, and the levels of lipid hydroperoxide significantly increase in hydrogen peroxide-treated cells. (8) These changes have also been detected in the hippocampal CA1 region of vitamin Edeficient and normal aged mice brains. (13) Vitamin E is a natural lipophilic substances that acts as an antioxidant, and vitamin Edeficient mice have accelerated oxidative damage in every tissue. These results indicate that ROS-induced microtubule alterations, including membrane oxidations, may contribute to neurite degeneration. However, the detailed mechanism is still not understood. Furthermore, these previous experiments have relied on wholecell lysates. In the present study, we isolated elongated neurites and identified various proteins in hydrogen peroxide-treated neurons.

Methods
Cell culture, animals and reagents. N1E-115 cells (ECACC #88112303), which are derived from a mouse neuroblastoma C1300 tumor, were originally obtained from DS Pharma Biomedical Co., R Ltd. (Osaka, Japan). Cells were grown in Dulbecco's minimum essential medium containing 10% heat-inactivated fetal calf serum (FCS) (Biological Industries, Beit Haemek, Israel), 2 mM glutamine, 50 U/ml penicillin and 50 mg/ml streptomycin, and were plated in wells, dishes, or cell culture inserts of laminin [derived from mouse Engelbreth-Holm-Swarm (EHS) sarcoma]-coated plates at a density of 4.0´10 5 cell/ml. Medium from cells cultured for 48 h was collected and used as conditioned medium. To elicit neurite elongation, after confirming cell adhesion, the medium was changed to conditioned medium containing 1% dimethyl sulfoxide (DMSO) solution. After 48 h, cells were used in experiments (Fig. 1A).
All animal experiments were performed with the approval of the Animal Protection and Ethics Committee of the Shibaura Institute of Technology. Three-month-old control animals (C57BL/ 6, male) were purchased from Japan SLC, Inc. (Hamamatsu, Japan). Twenty-four-month-old (C57BL/6, male) mice were obtained from the Tokyo Metropolitan Institute of Gerontology (Tokyo, Japan). AD transgenic mice [#008730, MMRRC034848, B6.Cg-Tg(APPSwFlLon,PSEN1*M146L*L286V)6799Vas/ Mmjax, Alias/5XFAD] were purchased from the Jackson Laboratory (Bar Harbor, ME). All mice were maintained under controlled conditions of temperature (22 ± 2°C), 12 h light/dark cycle, and had free access to food and water. The aged-mice were acclimated to their new environment for one week before each experiment. Normal diet pellets (Labo MR Stock) were purchased from Nosan Corp (Kanagawa, Japan). After dissection, the cortex regions were homogenized and used as brain samples.
All other chemical agents were obtained from either FUJIFILM Wako Pure Chemical Co. (Osaka, Japan) or Sigma-Aldrich Co. (St. Louis, MO). All tissue culture plates and dishes were purchased from Becton Dickinson and Company (Franklin Lakes, NJ).

Optimization of hydrogen peroxide concentration.
To determine the optimal hydrogen peroxide concentration against N1E-115 cells, cell survival was assessed using the trypan blue dye exclusion assay. After confirming neurite elongation, cells were treated with various hydrogen peroxide concentrations (0, 5, 10, 20, 50, and 100 mM) for 24 h, 0.8% (w/v) trypan blue in phosphate-buffered saline (PBS) was added to each sample, and the plate was incubated for 20 min in a CO 2 incubator at 37°C. Cells were then washed with PBS at least three times. Photomicrographs of the cells were taken using an Olympus IX81 phasecontrast microscope (Olympus Corp., Tokyo, Japan) equipped with a DP71 digital camera (Olympus Corp.), after which the images were stored and processed on a personal computer. Photomicrographs used for analysis were selected randomly for each treatment and incubation condition. The number of living cells per unit of area was counted, and the data were presented as a percentage of the total number of cells. At least four wells were analyzed for all experimental conditions, and each experiment was repeated three times.
Immunocytochemical analysis. To clarify the separation between cell bodies and neurites across insert membranes, we performed immunostaining. After 24 h of treatment with 10 mM hydrogen peroxide, N1E-115 cells were washed and scraped on the upper side of the insert, including cell bodies by a sterilized cotton swab, and the under layer of the insert was stained with b-actin and Hoechst 33258 dye. The cells or neurites were fixed with 4% paraformaldehyde (PFA) in PBS for 15 min at 4°C. Cells were then blocked with 10% goat serum in PBS for 1 h at room temperature (RT) and probed with mouse anti-beta actin monoclonal antibody (#ab8224, Abcam plc., Cambridge, UK) diluted 1:400 in PBS containing 1% goat serum, 1% bovine serum albumin (BSA), and 0.05% Triton X-100 overnight at 4°C. The cells were then incubated with Alexa Fluor 488-conjugated anti- mouse or anti-rabbit IgG secondary antibodies (Life Technologies Japan Corp., Tokyo, Japan) diluted 1:500 in PBS for 1 h at RT. Finally, to observe the status of nuclei, cells were stained with 1 mM Hoechst 33258. Photomicrographs of neurites were taken on a phase-contrast microscope equipped with a digital camera, stored, and then processed on a personal computer. Fluorescently labeled cells were observed and photographed with a fluorescence microscope. Immunocytochemical experiments were performed at least three times.

Extraction of neurite proteins for iTRAQ labeling.
Neurites isolated from cell culture inserts in the presence or absence of hydrogen peroxide were centrifuged at 15,000 rpm for 10 min at 4°C, respectively. After discarding supernatants, extraction buffer (7 M urea, 2% CHAPS, 50 mM HEPES, pH 8.5) was added and the samples were sonicated on ice. After adjustment of pH (from 8.4 to 9.0 by 1 M Tris-HCl sol.), the samples were incubated for 15 min at RT. SDS (20%), deionized water was added, and the samples were incubated for 10 min at 70°C. The samples were centrifuged at 15,000 rpm for 20 min at 4°C and resuspended in ice cold acetone. After incubation for 60 min at 4°C, the samples were centrifuged at 15,000 rpm for 20 min at 4°C. The pellet samples were resuspended in acetone and then centrifuged at 15,000 rpm for 20 min at 4°C. After discarding supernatants, the samples were dried and were resuspended in 0.2% SDS in 0.5 M triethylammonium bicarbonate (TEAB) buffer. Finally, protein concentrations were measured by the Bradford assay, and the samples were diluted to 1 mg/ml by 0.2% SDS in 0.5 M TEAB buffer. These were used as samples for the iTRAQ labeling.
iTRAQ labeling. Extracted neurite proteins in the presence or absence of hydrogen peroxide were labeled with iTRAQ reagent using the iTRAQ Reagent 3-Assay Duplex Trial Kit (AB SCIEX, Framingham, MA) according to the manufacturer's protocol with some modifications. (14) Twenty mg of extracted neurite proteins were reduced by 4 mM tris-(2-carboxyethyl)phosphine (TCEP) for 1 h at 60°C, and were alkylated by 8 mM methyl methanethiosulfonate (MMTS) for 10 min at RT. The samples were trypsinized overnight at 37°C. iTRAQ reagents 114 and 117 were dissolved 70 ml ethanol and mixed with each sample. After mixing with two iTRAQ-labelled samples, mixtures were concentrated by evaporation, and strong cation exchange (SCX) load buffer (10 mM KH 2 PO 4 in 25% acetonitrile (MeCN), pH 3.0) was added. Finally, phosphoric acid was added, the pH confirmed as 2.5 to 3.3, and the peptides were fractionated into seven fraction on Mono-Spin TM SCX spin column (GL Sciences Inc., Tokyo, Japan) using 10 mM KH 2 PO 4 in 25% MeCN containing 10, 25, 50, 100, 175, 350 and 1,000 mM KCl, pH 3.0. After evaporation and reconstitution in 2% MeCN/0.1% trifluoroacetic acid (TFA), the samples were desalted by a Mono-Spin TM C18 spin column (GL Sciences). Eluates were evaporated in vacuo to dryness, reconstituted in 2% MeCN/0.1% TFA, and were subjected to liquid chromatographymatrix-assisted laser desorption/ionization-time of flight/time of flight (LC-MALDI-TOF/TOF) analysis.
Protein identification. All MS/MS data were analyzed by ProteinPilot software ver. 4.5 beta (AB SCIEX). Detailed analytical conditions were as follows: sample type, identification; cysteine alkylation, MMTS; digestion, trypsin; instrument, 5800; species, Mus musculus; ID focus, biological modifications; database, uniprot_sprot_can+iso_20100622+Contams+. Protein identification was based on the following selection criteria protein score (ProtScore) >1.3 (unused, p<0.05, 95% confidence) and at least one peptides with an ion score above 95% confidence. The ratios of each identified protein analyzed by LC-MALDI-TOF/TOF were calculated as the average ratio of 117/114.
Statistical analysis. Data were plotted as means ± SE, and were analyzed by Dunnett's test, with p<0.05 considered significant.

Optimization of hydrogen peroxide concentration.
Before identification of proteins in isolated neurites of hydrogen peroxide-treated N1E-115 cells, we optimized the hydrogen peroxide concentration. As shown in Fig. 1A, treatment with 10 mM hydrogen peroxide induced neurite degeneration such as beadings. Treatment with hydrogen peroxide induced cell death in a concentration-dependent manner in N1E-115 cells (Fig. 1B). At more than 10 mM of hydrogen peroxide, the ratio of cell death significantly increased compared to the control. Because large parts of cells did not remain on the surface of the culture dish, we could not count cell death in 100 mM treated samples. Since we wanted to observe changes in neurite-specific proteins prior to the induction of cell death, 10 mM hydrogen peroxide was used for further experiments.

Isolation of neurites from a cell culture insert system.
To isolate cell bodies and elongated neurites in hydrogen peroxide-treated N1E-115 cells, we used a cell culture insert system containing many small holes with diameters of 8 mm ( Fig. 2A-I). Elongated neurites can pass through the holes and stick to the opposite surface of the insert membranes ( Fig. 2A-IV and 2B). After neurite elongation, we scraped cell bodies from cell culture inserts by sterilized cotton swabs (Fig. 2C) and turned over the insert membranes for staining ( Fig. 2A-V). We could not find histone proteins on the bottom surface (Fig. 3A). On the other hand, nuclear proteins were clearly stained on the top surface of the membranes before scraping. In western blotting assays, we could not detect nuclear protein in isolated neurite samples (Fig. 3B). We used this isolation method for further experiments.
Identification of proteins in isolated neurites of hydrogen peroxide treated N1E 115 cells by LC MALDI TOF/TOF. iTRAQ-labeled peptides samples were analyzed by LC-MALDI-TOF/TOF. Around 200 proteins were identified in each independent experiment (Fig. 4A). Detailed data of protein identification were summarized in Supplemental Table 1*. Figure 4B shows twenty-one proteins detected in at least twice experiments and with an increased expression ratio by hydrogen peroxide treatment. Five proteins represented by arrows in Fig. 4B were selected, and their alterations of expressions were confirmed in isolated neurites treated with hydrogen peroxide and in animal homogenate in subsequent experiments.
Validation by western blot analysis. To determine whether the identified proteins could be detected in isolated neurites, we checked protein expression by western blotting. As shown in Fig. 5A and B, secretogranin-1, 60 kDa heat shock protein (HSPD) 1 and superoxide dismutase 1 were significantly increased in isolated neurites of hydrogen peroxide-treated cells. Brain acid soluble protein 1 (BASP1) and 78 kDa glucose-regulated protein (Hspa5) were not increased by the treatment of hydrogen peroxide.
Identified proteins were increased in aged and AD model mouse brains. To test the relationship between changes in the expressions of these proteins and aging or age-related neurodegenerative disorders, we used normal aged and AD transgenic mice and checked protein expression by western blotting. After genotyping, we separated normal and AD mice, and used as experimental animals (Fig. 6A). Secretogranin-1 and HSPD1 proteins were significantly increased in normal aged mice brains compared to the controls. Secretogranin-1, BASP1 and HSPD1 protein were significantly increased in the AD mice brains compared to the controls ( Fig. 6B and C). cells by a cell culture insert system. In the present study, we aimed to identify neurite-specific proteins in hydrogen peroxidetreated N1E-115 cells. We first needed to establish a new neurite isolation method. Because cell bodies and neurites were not separated in previous studies, including ours, it was not possible to determine what proteins are directly related to the induction of neurite degeneration. (8,9,15) Certainly, many intracellular proteins are translated in the nucleus and transported to distal ends of neurites by the axonal transport system. (16,17) However, neurites can expand and contract easily by external stimuli such as changes in pH, ionic composition, etc. (18,19) Furthermore, we have found that treatment with a low concentrations of hydrogen peroxide induces neurite degeneration prior to cell death. (8,9) It is well known that neurite degeneration starts from the distal end of neurites in Wallerian Degeneration Slow (WLD s ) models. (20) This indicates that neurite degeneration is an early sign of neuronal degeneration.
To determine changes in hydrogen peroxide-derived proteins expressions in neurite region of N1E-115 cells, we used a cell culture insert system to isolate neurites of N1E-115 cells. Cell culture insert systems are commonly used in cell to cell interaction experiments. (21) There are many small pores in the insert of about 8 mm, and elongated neurites being found under the surface of the insert membrane (Fig. 2). To collect neurites, we scraped under the surface of inserts, and performed two different experiments to confirm the absence of cell bodies (Fig. 3). Both fluorescent staining and immunostaining showed no cell body regions at all, and so we used this method to isolate neurite samples for further experiments.
However, the total protein volume of isolated neurites was very low compared to whole cell lysates. In this study, a total of four 6-well plates were prepared per sample. Two plates were used for LC-MALDI-TOF/TOF analysis, and two for western blotting. To eliminate confounding variables as much as possible, we made two N1E-115 cell samples together in the presence or absence of hydrogen peroxide.
Identification of proteins by LC MALDI TOF/TOF analysis in the presence or absence of hydrogen peroxide in isolated neurites of N1E 115 cells. First, we checked the reproducibility iTRAQ labeling. A control sample lysate was divided into two tubes equally, and each sample was labelled with different iTRAQ reagents (114 and 116). Each iTRAQ-labeled sample was mixed and analyzed two times by LC-MALDI-TOF/TOF. In this test analysis, we identified around 300 proteins. The iTRAQ ratio of identified proteins was roughly 0.8 to 1.3. The same 203 proteins were identified, and the matching ratio was to 57.5%. These results showed that our labelling method worked. Next, we studied the changes of protein profiles by the treatment with hydrogen peroxide in triplicate. When there was a difference more than 1.5 or less than 0.6 times, we defined as differentially expressed proteins. After LC-MALDI-TOF/TOF analysis, only proteins with an increased expression ratio by hydrogen peroxide treatment were selected (Fig. 4B). Furthermore, we excluded nuclear proteins, such as histones and ribonucleoprotein. Why nucleusrelated proteins were detected in isolated nerites samples was unknown. It may be technical problem including contamination of cytoplasm. Finally, we selected 5 proteins and conducted further detailed studies by western blotting. On the other hand, the number of proteins with a decreased expression ratio by hydrogen peroxide treatment were 9 (data not shown). Large number of proteins were related to cytoskeletal proteins. Furthermore, we tried to analyze the levels of protein modifications such as phosphorylation, acetylation, polymerization, etc. However, this proved difficult. We were able to identify proteins and those that increased with hydrogen peroxide treatment. Detecting novel proteins and protein modifications are issues of current research.

Secretogranin 1 increased in isolated neurites of hydro gen peroxide treated N1E 115 cells, and in normal aged
and AD mice brains. To determine the reproducibility of the 5 identified proteins, we checked their expressions by western blotting. The secretogranin-1 expression in isolated neurites of hydrogen peroxide-treated cells was significantly higher than that of the controls (Fig. 5). Secretogranin-1, also called chromogranin B, is an acidic secretory soluble protein and is expressed in neuroendocrine and neuronal cells. (22) Secretogranin-1 belongs to a chromogranin family, which plays a crucial role in the sorting and aggregation of secretory products in the trans-Golgi network, (23) and some researchers have used it as a marker of large dense core vesicles. (24) Specifically, secretogranin-1 is highly expressed in excitatory hippocampal CA3 and DG neurons. (25) Recently, it has been reported that a relationship exists between changes in the secretogranin-1 and the progression of some severe diseases.
Monaghan et al. (26) reported that the plasma secretogranin-1 level was increased in human gastroenteropancreatic neuroendocrine tumors. Secretogranin-1 has been identified in biopsy specimens of medullary thyroid carcinoma patients by LC-MS/MS. (27) Kroksveen et al. (28) reported that secretogranin-1 is increased in the cerebrospinal fluid of early multiple sclerosis patients, and may be used as a biomarker of this disease. Furthermore, several lines of evidence demonstrate a relationship between secretogranin-1 and AD. Chromogranin peptides including secretogranin-1 are located in neuritic plaques of the cortical layer in AD patients. (24) The localization of secretogranin-1 is matched with amyloid plaques in AD transgenic mice brains. (29) Mattsson et al. (30) reported that the secretogranin-1 concentration is significantly decreased in early AD patients, and is one possible reason for the reduction of synapses. In our study, the expression of secretogranin-1 was significantly increased in AD transgenic mice brains compared to age-matched controls (Fig. 6). It is well known that the development and progression of AD is strongly correlated with ROS production and ROS-based neuronal injury. (1,31,32) It is . The score of control cells in identified protein of each RUN was set to 1. Asterisks show more than 1.5 fold expression in hydrogen peroxide treated cells compared to controls. Gray box shows that peptides score is less than 1. Blank box shows undetected. Other boxes show that the peptide score is more than 1. Arrows show that protein expressions were verified by further experiment. Each result (RUN) represents three independent experiments. possibility that changes in secretogranin-1 expression in neurites relates to these neurodegenerative disorders in some form. However, further experiments are needed to elucidate the relationship between changes in ROS-related secretogranin-1 expression and AD progression. Additionally, the expression of secretogranin-1 was significantly increased in normal aged mice brains. Considering that secretogranin-1 expression was significantly increased in hydrogen peroxide-treated N1E-115 cells compared to untreated ones, our results may support the free radical theory of aging.
HSPD1 is increased in isolated neurites of hydrogen peroxide treated N1E 115 cells, and normal aged and AD mice brains. HSPD1 protein expression in isolated neurites of hydrogen peroxide-treated neurons was significantly increased compared to untreated ones (Fig. 5). Furthermore, HSPD1 was significantly increased in normal aged and AD transgenic mice brains compared to young controls (Fig. 6). HSPD1 is also known as HSP60, and is localized in the mitochondrial matrix space. (33) In general, HSPs are upregulated in response to hyperthermia, an imbalance of the redox system, chemical stimulation from extracellular regions, etc. (34) Much of the literature on HSP60 is related to immune diseases. (35) However, several studies have demonstrated a relationship between changes in HSP60 and neurodegenerative disorders. (34,36,37) HSP60 binds to amyloid precursor protein and beta-amyloid, and accumulates in the mitochondria of AD transgenic mice brains. (34) HSP60 supports the folding of matrix proteins and plays a crucial role in protection against Parkinson's disease. (37) Additionally, in endothelial cells, the expression of HSP60 is a marker of chemical stress such as from smoking or free radicals, and increases the risk of atherosclerosis. (35) In contrast to our results, Kleinridders et al. (38) reported that mitochondrial dysfunction is induced by downregulation of HSP60 in diabetic mice, and ROS production and insulin resistance are increased. In our previous study, we found that mitochondria accumulate in neurite beading regions (18) and induce superoxide production in hydrogen peroxide-treated neuronal cells. (39) Mitochondria play crucial roles in the branching and elongation of neurites. (40,41) It is possible that changes in mitochondrial function, including in HSP60, contribute to the maintenance of neurite function. We are continuing to study the relationship between changes in mitochondrial function and the induction of neurite degeneration, especially regarding the expression of ROS-related proteins.
In this study, we isolated neurites using a cell culture insert system and identified proteins increased in isolated neurites of hydrogen peroxide-treated N1E-115 cells. In particular, the expressions of secretogranin-1 and HSPD1 were significantly increased not only in isolated neurites with hydrogen peroxide, but also in normal aged and AD transgenic mice brains. More study is needed to elucidate the relationship between these proteins and ROS-related neurite degeneration, such as an overexpression and a knockdown techniques. These proteins and others may serve as new biomarkers prior to the induction of neurite degeneration caused by ROS.  , and gray columns show aged mice (n = 9). White columns show 3 month old control mice (n = 9). The ratios of each protein band intensity to a tubulin intensity are shown, and the ratios of control samples were set to 1. Each column represents the mean of three independent experiments. Data were analyzed using Dunnett's test, * indicates p<0.05 and ** indicates p<0.01. Western blot analysis of each protein (C).