Quantitative Proteomics Reveals Neuroprotective Mechanism of Ginkgolide B in A β 1 − 42 -induced N2a Neuroblastoma Cells

Objective : Ginkgolide B (GB) possesses anti-inflammatory, antioxidant, and anti-apoptotic properties against neurotoxicity induced by amyloid beta (A β ), but the potential neuroprotective effects of GB in Alzheimer’s therapies remain elusive. We aimed to conduct proteomic analysis of A β 1 − 42 induced cell injury with GB pretreatment to uncover the underlying pharmacological mechanisms of GB. Methods : Tandem mass tag (TMT) labeled liquid chromatography-tandem mass spectrometry (LC-MS/MS) method was applied to analyze protein expression in A β 1 − 42 induced mouse neuroblastoma N2a cells with or without GB pretreatment. Proteins with fold change > 1.5 and p < 0.1 from two independent experiments were regarded as differentially expressed proteins (DEPs). Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed to analyze the functional annotation information of DEPs. Two key proteins osteopontin (SPP1) and ferritin heavy chain 1 (FTH1) were validated in another three samples using western blot and quantitative real-time PCR. Results : We identified a total of 61 DEPs in GB treated N2a cells, including 42 upregulated and 19 downregulated proteins. Bioinformatic analysis showed that DEPs mainly participated in the regulation of cell death and ferroptosis by down-regulating SPP1 protein and up-regulating FTH1 protein. Conclusions : Our findings demonstrate that GB treatment provides neuroprotective effects on A β 1 − 42 induced cell injury, which may be related to the regulation of cell death and ferroptosis. The research puts forward new insights into the potential protein targets of GB in the treatment of Alzheimer’s disease (AD).


Introduction
Alzheimer's disease (AD) is the commonest neurodegenerative disorder in the older people accompanied by progressive cognitive impairment and behavior disfunction. Approximately 3.2% of the population over 65 years old suffers AD, and the global prevalence will reach 115 million sufferers by 2050 [1]. The typical pathological features are extracellular aggregated senile plaques of amyloid beta (Aβ) and intracellular highly phosphorylated neurofibrillary tau tangles in the cerebrum [2]. Abnormal accumulation of Aβ within the cerebrum is an initial pathological change of AD, which appears years or more than a decade ahead of the onset of clinical symptoms [3]. An imbalance between Aβ generation by neurons and clearance from the interstitial fluid results in extracellular amyloid plaques deposition [4], accompanied by the resultant neuroinflammation, oxidative stress, free radical damage, and extensive neuron death [5]. Membrane-spanning protein amyloid precursor protein (APP) is cleaved to Aβ 1−40 peptide and Aβ 1−42 peptide by β-secretase and γ-secretase [6]. Aβ 1−42 peptide is more hydrophobic and tends to aggregate to form an oligomeric or fibrous structure which is the core component of senile plaques [7]. Cells with neurotoxic Aβ 1−42 peptide intervention could replicate the pathologi-cal hallmarks of AD. Therefore, the in vitro Aβ 1−42 -treated mouse neuroblastoma N2a cell line is considered a classical model to mimic the cell injury of AD, which can be used for exploring the underlying pathological mechanism of AD.
Phytochemicals are bioactive substances of plant origin with various structures and significant pharmaceutical properties [8]. Ginkgolide B (GB, C 20 H 24 O 10 , molecular weight = 424.3986 g/mol) is a kind of diterpenoids isolated from the leaves of Ginkgo biloba, and has various pharmacological effects, such as inhibiting platelet activating [9], scavenging oxygen free radicals [10], and antioxidative stress function. GB could prevent neuronal cell injury caused by oxidative stress in vitro and improve cognitive function in the central nervous system [11]. GB also possesses neuroprotective functions against cerebral ischemic injury and Aβ-induced neurotoxicity through various biological properties of anti-inflammation, antioxidative stress, and anti-apoptosis [12][13][14]. However, its specific pharmacological mechanism in AD remains ambiguous. In the current study, we aim to explore the protective effects of GB on cell injury caused by Aβ 1−42 and uncover its underlying cellular mechanisms.
Mass spectrometry (MS)-based quantitative proteomics is an advanced technology for unbiased protein identification and quantitation on a large scale, which re-lies on precise, high-throughput, and reproducible techniques [15]. Advances in liquid chromatography-tandem MS (LC-MS/MS) have qualified for the identification and quantification of thousands of proteins in biological samples. It not only identifies proteins in normal and pathological states, but also accurately quantifies their abundance [16]. LC-MS/MS technique has significant value in identifying functional modules and pathways, discovering biomarkers, as well as diagnosing and surveilling diseases. LC-MS/MS technology with stable isotope-labeling amino acids has been extensively applied for exploring the targets and mechanisms of multiple diseases and drugs.
Therefore, tandem mass tag (TMT) labeled LC-MS/MS-based quantitative proteomic analysis of proteins from GB treated N2a cells was performed to analyze the differentially expressed proteins (DEPs) that are associated with the neuroprotective mechanism of GB. As a result, 61 proteins were identified as DEPs (fold change >1.5 and p < 0.1). Of these, 42 proteins were upregulated and 19 proteins were downregulated after GB treatment. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses revealed that DEPs mainly involved in the regulation of cell death, immune system process, and ferroptosis. Among the DEPs, two key proteins ferritin heavy chain 1 (FTH1) and osteopontin (SPP1), mainly participate in the regulation of cell death and ferroptosis, were verified using western blot and quantitative real-time PCR (qRT-PCR). The current conclusions provide new insight into the potential therapeutic targets of GB in AD.

Preparation of GB Stock Solutions
GB (HPLC purity >98%) was purchased from Chengdu Pufei De Biotech Co., Ltd. Its molecular weight is 424.3986 g/mol and molecular formula is C 20 H 24 O 10 (Fig. 1c). It was dissolved in dimethyl sulfoxide (DMSO, Sigma, St. Louis, MO, USA) and prepared in a stock solution of 5 mmol/L with culture medium. The stock solution was diluted to required concentrations with the same culture before use. DMSO concentration was kept below 0.1% to prevent cytotoxic reaction [17].

Preparation of Aβ 1−42 Oligomers
The preparation of Aβ 1−42 oligomers was based on a widely recognized method [18,19]. 1 mg lyophilized Aβ 1−42 peptide (AS-20276; AnaSpec, shanghai, China) was dissolved in 221 µL of 100% hexafluoroisopropanol at a concentration of 1 mM and evenly divided into two tubes. Then, the peptide film was dried under vacuum. Approximately 0.5 mg dried peptide film was redissolved in 20 µL of fresh dry DMSO to a concentration of 5 mM and diluted with F12 cell culture medium to 100 µM [17]. The solution was incubated at 4°C for 24 h and centrifugated at 14,000 × g at 4°C for 10 min for elimination of fibrils. The su-pernatant that consists of soluble oligomerized Aβ 1−42 was used for experiments.

Cell Culture and Treatments
The mouse neuroblastoma N2a cells were donated by Ji Jianguo Lab (the State Key laboratory of Protein and Plant Gene Research, College of Life Sciences, Peking University) and cultured in Dulbecco's Modified Eagle Medium (DMEM, Hyclone, Logan, UT, USA) containing 10% fetal bovine serum (Hyclone, Logan, UT, USA). N2a cells were maintained in a humidified atmosphere containing 5% CO 2 at 37°C. The cells were pretreated with GB or vehicle for 2 h, followed by incubation in the presence or absence of 10 µM Aβ 1−42 oligomers for an additional 24 h. The experiments involve the control group (treated with 0.1% DMSO for 24 h), the Aβ group (treated with 10 µM Aβ 1−42 for 24 h) and the Aβ + GB group (pretreated with 100 µM GB for 2 h followed by 10 µM Aβ 1−42 for 24 h).

Detection of Cell Survival Rates
Cell survival rates were tested by a colorimetric 3-4,5-dimethylthiazol-2-yl-2,5-diphenyl-tetrazolium bromide (MTT) assay (Beyotime Biotechnology, Shanghai, China). N2a cells were seeded in 96-well plates and treated according to the experiment designed when the cells reached to 60% confluence. Then the cells were incubated with 10 µL MTT (5 mg/mL) at 37°C for 4 h, and the MTT formazan was extracted with 150 µL DMSO. The absorbance of each well was measured at 570 nm using a Multiskan FC microtiter plate reader (Thermo Fisher Scientific, Waltham, MA, USA).

Flow Cytometry Analysis of Cell Apoptosis
Apoptotic N2a cells were determined using Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) cell apoptosis detection kit (TransGen Biotech Co, Beinjing, China). After treatments, N2a cells were digested with trypsin, washed, and resuspended in Annexin Ⅴ binding buffer. Then, cells were incubated with 5 µL of Annexin V-FITC and 5 µL of PI at room temperature for 15 min in the dark. The number of apoptotic cells were detected using a flow cytometer (FACSVerse, BD Biosciences, Franklin, NJ, USA) and analyzed using FlowJo software (Version 7.6, TreeStar Inc, San Carlos, CA, USA). . Data were presented as mean ± SEM. **p < 0.01 and ***p < 0.001 vs. control group; ##p < 0.01 and ###p < 0.001 vs. Aβ group.

Measurement of Intracellular Reactive Oxygen Species (ROS)
Intracellular ROS levels of N2a cells were tested using ROS probe 2',7'-dichlorofluorescein diacetate (DCFH-DA) (Beyotime Biotechnology, Nanjing, China). After treatment, N2a cells were incubated with 10 µM DCFH-DA for 30 min and then washed with serum-free cell culture medium. The DCF fluorescence was quantified using a multimode microplate reader with an excitation source at 485 nm and an emission at 530 nm. The values were expressed as the fold of control.

Measurement of Superoxide Dismutase (SOD) Activity and Malondialdehyde (MDA) Content
The SOD activity and MDA content were detected using commercially available kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) following the protocols. The SOD activity was tested using the xanthine oxidase method and measured by the absorbance of superoxide anion free radical at 550 nm. The lipid peroxidation level was reflected by MDA concentrations using the thibabituric acid (TBA) method. MDA concentrations were measured by the absorbance of TBA reactive substances at 532 nm.

Database Search Parameters
Protein identification, quantification, and MS/MS original data were analyzed by SEQUEST search algorithm using Proteome Discoverer (version 2.2, Thermo Scientific, Waltham, MA, USA). Raw files were searched with a mouse Uniprot protein database. Search criteria included maximum missed trypsin cleavages of 2; fixed modification on lysine and N-terminus for TMT six-plex tags; dynamic modification for oxidation of methionine residues; carbamidomethyl on cysteines; mass tolerance of 10 ppm; 0.05 Da for MS/MS tolerance; 1% false discovery rate (FDR) [21,22]; and 1% FDR at peptide and protein levels [17]. Reporter ion intensities was normalized using the index of total reporter ion intensity.

Data Analysis and Interpretation
The protein ratios (Aβ + GB/Aβ) were normalized in the transformed Log2 fold change to adjust the unequal protein content. p-value was calculated with an independent student's t-test (Aβ + GB versus Aβ). Proteins with a cutoff of p < 0.1 and fold change >1.5 from two independent experiments were regarded as DEPs.

KEGG and GO Enrichment Analyses of DEPs
The Database for Annotation, Visualization, and Integrated Discovery (DAVID) is an online database that offers systematic and comprehensive functional annotation information of proteins and genes to reveal biological information [23,24]. KEGG pathway enrichment analysis and GO analysis (biological process, molecular function, and cellular component) of DEPS were carried out using DAVID (version 6.8) to analyze the function of DEPs. p < 0.05 was regarded as statistical significance.

Quantitative Real-Time PCR Assay
Total RNA was extracted from N2a cells using Trizol reagent (Invitrogen, Carlsbad, CA, USA) following the supplier's recommendations. Total RNA (1.5 µg) was reversely transcribed to cDNA using a HiFi-Script cDNA Synthesis kit (CW Bio, Beijing, China). Quantitative analysis of FTH1 and SPP1 expressions was performed using CFX96™ Real Time PCR Detection System (Bio-Rad, Hercules, CA, USA). The mRNA levels were calculated using the 2 −∆∆CT algorithm and normalized to GAPDH. The primers in this article were as follows: SPP1,

Statistical Analysis
All data were shown as mean ± SEM and analyzed by one-way analysis of variance (ANOVA) followed by Student-Newman-Keuls test for intergroup comparisons. All analyses were conducted using SPSS statistical software (Version 17.0, SPSS, Chicago, IL, USA). A p value < 0.05 was identified as statistical significance.
The characteristic of Aβ 1−42 oligomers was revealed by TEM (Fig. 1a), and the toxicity of Aβ 1−42 oligomers was tested by MTT assay. The MTT results showed a dose-dependent association between Aβ 1−42 concentration and cell injury. 10 µM Aβ 1−42 treatment markedly decreased cell survival rate compared with the control group (Fig. 1b). However, 20 µM Aβ 1−42 produced no notable changes in cell viability compared with 10 µM Aβ 1−42 (Fig. 1b, p > 0.05). Therefore,Aβ 1−42 oligomers at a concentration of 10 µM was chosen for Aβ- induced cell model. GB, a kind of diterpenoids isolated from the leaves of Ginkgo biloba, provides neuroprotective functions against Aβ-induced neurotoxicity in previous studies [14]. To verify whether GB exerts any effect on N2a cells themselves, the viability of various concentrations of GB intervened N2a cells was detected. GB (20-200 µM) alone caused no obvious effects on cell viability, with the maximal cell viability occurring at 100 µM (Fig. 1d). GB (20-200 µM) pretreatment showed certain inhibitory effect on cell injury caused by Aβ 1−42 oligomers, especially at a concentration of 100 µM (Fig. 1e). Hence, 100 µM GB was used for further experiments.

GB Attenuated Aβ 1−42 -Induced Oxidative Stress and Aβ Peptide Levels in N2a Cells
The levels of oxidative stress in N2a cells in various groups were tested by measuring intracellular levels of ROS, SOD and MDA. Compared with the control group, 10 µM Aβ 1−42 gave rise to a prominent reduction of SOD level and a considerable elevation of ROS and MDA levels (Fig. 2a-c). On the contrary, 100 µM GB pretreatment significantly relieved Aβ 1−42 -induced oxidative injury with increased SOD level and decreased ROS and MDA levels (Fig. 2a-c). Then, we examined whether GB affects Aβ 1−42 peptide expression. Western blot revealed that Aβ 1−42 peptide level significantly increased with Aβ 1−42 intervention, and the uptrend was reversed by 100 µM GB pretreatment (Fig. 2d).

Global Proteome Profiling of N2a Cells with GB Pretreatment
To explore the neuroprotective mechanisms of GB against cell injury stimulated by Aβ 1−42 oligomers, TMTlabeled LC-MS/MS analysis was executed to find out the relative changes in protein abundance. The workflow is exhibited in Fig. 3a. Cells from the control group, Aβ group and Aβ + GB group were labeled with 126 and 127 TMT tags, 128 and 129 TMT tags, and 130 and 131 TMT tags, respectively. In the end, 6969 proteins were identified and 6740 proteins (96.71%) were quantified in two independent biological replicates. The relative abundance ratios (Aβ + GB/Aβ) distribution of the quantified proteins was normality (Fig. 3b). Only proteins with fold change >1.5 and p value < 0.1 were identified to be differentially expressed. 61 proteins were considered as DEPs, including 42 upregulated and 19 downregulated proteins. In the volcano figure, red dots indicate upregulated proteins and blue dots indicate downregulated proteins (Fig. 3c). The DEPs were showed in Table 1.
To identify obviously different proteins in GB treated N2a cells, hierarchical clustering analysis was performed. The heatmap of 61 DEPs revealed that DEPs were distinctly separated into three distinct areas (Fig. 4). The expression levels of DEPs in the Aβ group were apparently different from those in the control group, while GB pretreatment attenuated these differences. These results indicated that protein expression undergoes extensive remodeling during AD pathogenesis and this trend can be partially reversed by GB.

Validation of the Key Proteins in DEPs
To confirm the MS-based quantitative proteomics, we analyzed the expression levels of two key proteins SPP1 and FTH1 in Aβ 1−42 -induced and GB treated N2a cells using western blot and qPCR. FTH1 protein and mRNA levels in the Aβ group decreased markedly, and 100 µM GB treatment upregulated their expressions (Fig. 6b,c). SPP1 protein and mRNA levels in the Aβ group significantly increased, and these upregulations were potently suppressed by 100 µM GB pretreatment (Fig. 6a,c). The change tendency of FTH1 and SPP1 expression levels detected by western blot and qPCR is consistent with mass spectrometric data.

Discussion
Extracellular aggregated amyloid plaques and intracellular highly phosphorylated neurofibrillary tangles are the typical pathological features of AD. Abnormal aggregation of neurotoxic Aβ promotes inflammatory responses, oxidative stress, and apoptotic activation, which results in extensive neuron death and the onset of clinical symptoms [26]. GB has anti-apoptotic and antioxidant functions against cerebral ischemic injury and Aβ-induced neurotoxicity [12,14]. Aβ-induced cell injury model of AD has been widely used for the investigation of AD pathogenesis. In the present study, N2a cells stimulated by Aβ 1−42 oligomers alone revealed significantly decreased cell survival rates and cell apoptosis. Nevertheless, this change was efficiently reversed in cells with GB pretreatment ( Fig. 1e-g). These results showed that GB relieved cell toxicity and apoptosis caused by Aβ 1−42 oligomers, implying the neuroprotective effect of GB against cell injury. Next, we investigated whether GB protects against Aβ 1−42 -induced oxidative injury in N2a cells. GB inhibited intracellular ROS and MDA production and induced an upregulation of antioxidant substance like SOD (Fig. 2ac), demonstrating the antioxidant effect of GB on cell toxicity stimulated by Aβ 1−42 oligomers. In addition, the increased Aβ peptide level in the Aβ group was also reversed by 100 µM GB (Fig. 2d). GB provided anti-apoptotic and antioxidant function on Aβ 1−42 -induced cell injury, and downregulated Aβ peptide expression. However, the potential molecular mechanisms of GB in the progression of AD remain ambiguous.
In order to explore the neuroprotective effects of GB in AD and reveal its underlying molecular mechanisms, TMT labeled proteomics was performed to explore the profile of DEPs associated with GB pretreatment in Aβ-induced cell injury model of AD (Figs. 3,4). We confirmed 61 DEPs in Aβ 1−42 -induced N2a cells with GB pretreatment, including 42 upregulated and 19 downregulated proteins (Table 1).
These DEPs contained several known AD protein biomarkers, like presenilin-2 (PSEN2), SPP1, and SLC7A11. The protein levels of Psen2, SPP1, and SLC7A11 were remarkably increased in Aβ group, and decreased after GB treatment. On the basis of LC-MS/MS analysis results, bioinformatic analysis of DEPs were applied to uncover the underlying signaling pathways and protein targets of GB.
GO enrichment analysis showed that DEPs mainly participated in the regulation of cell death, reminded us that apoptotic signaling pathway has definite function in the pharmacological effects of GB in AD (Fig. 5). Proteins involved in the regulation of cell death pathway included PSEN2 and SPP1. Secreted glycophosphoprotein SPP1 has an extensive range of functions [27], such as prompting neuroinflammation and apoptosis. SPP1 protein expression has been observed to upregulate in the prefrontal cortex and cerebrospinal fluid of AD patients [27,28], as well as in APP/human presenilin-1(PS1) KI mouse model of AD [29]. Our study showed that SPP1 participants in the regulation of apoptotic progress of AD and exacerbates cell apoptosis. We validated obvious up-regulation of SPP1 in Aβ 1−42induced N2a cells, and the increased SPP1 level was significantly reduced after GB treatment (Fig. 6). These results further support the anti-apoptotic effect of GB in AD, which possibly is related to downregulated SPP1 protein.
KEGG enrichment analysis showed that DEPs in GB treated N2a cells were concerned with ferroptosis. Ferroptosis is a newly recognized iron-dependent oxidative cell death that is considered as a crucial pathological process in AD [30,31]. It depends on the regulation of various iron metabolism proteins, such as FTH1 and ferroportin [32]. FTH1, the essential iron storage protein, provides protective effects against neural damage through the regulation of iron metabolism [33,34]. In APP/PS1 mice, neurons in cerebral cortex and hippocampus were vulnerable to ferroptosis, accompanied by downregulated FTH1 protein [30]. GB exerted anti-ferroptosis effects by up-regulating Nrf2 and FTH1 in nonalcoholic fatty liver disease [35]. Consistent with previous conclusions, FTH1 remarkably decreased in the Aβ group and GB pretreatment upregulated its expression in our study (Fig. 6). This result implies that GB may exert anti-ferroptosis effect by up-regulating FTH1 protein in AD.
Taken together, above results confirmed that GB exerts neuroprotective effects against Aβ 1−42 induced cell injury by restoring DEPs involved in the regulation of cell death and ferroptosis. GB provided anti-apoptotic and antiferroptosis roles in AD through down-regulating SPP1 and up-regulating FTH1. Therefore, SPP1 and FTH1 proteins can be considered as potential therapeutic targets of GB for AD.

Conclusions
In conclusion, the current study demonstrates that GB therapy dramatically attenuates Aβ 1−42 -induced cell tox-icity, cell apoptosis and oxidative stress. We identified 42 upregulated and 19 downregulated DEPs in GB treated N2a cells using TMT labeled proteomics. GB provided anti-apoptotic and anti-ferroptosis effects against Aβ 1−42induced cell injury may partly through down-regulating SPP1 protein and up-regulating FTH1 protein. SPP1 and FTH1 proteins are expected to be potential therapeutic targets of GB in AD. Our conclusions were mainly derived from mass spectrometry data and bioinformatic analysis, which must have some limitations. More precise validation and additional study will be performed in the future.

Consent for publication
The manuscript is approved by all authors for publication. We would like to declare on behalf of my co-authors that the manuscript has not been previously published, and is not currently submitted for review to any other journal, and will not be submitted elsewhere before a decision is made by this journal.

Ethics Approval and Consent to Participate
Our research was based on N2a cell line, and did not involve human or animals. No ethical approval is required.

Funding
This study was supported by a Grant-in-Aid from the Scientific Research Project of Hebei Administration of Traditional Chinese Medicine (2020172).