NSCs were isolated from the SVZ region of newborn (0–2 days) C57BL/6 mice and cultured in NSCs proliferation media as previously describe in our laboratory (Yang et al., 2009, 2012). Briefly, the SVZ regions of the fresh brains were isolated and cut into 1-mm³ pieces and subsequently suspended in 3 mL of 0.25% trypsin-EDTA (Invitrogen) at 37°C for 15 min. After filtration through a 70-μm cell strainer (BD Falcon), the cells (1 × 10⁶/ml) were plated onto poly-L-lysine coated 24-well plates (BD Bioscience, San Jose, CA) and maintained in a humidified atmosphere (5% CO₂-95% air) at 37°C. NSCs were cultured in Dulbecco's modified Eagle's medium (DMEM)/F12 (Gibco, Grand Island, USA) supplemented with 2% B27 (Gibco), 20 ng/mL epidermal growth factor (EGF, Peprotech, Rocky Hill, NJ, USA), and 20 ng/mL basic fibroblast growth factor (b-FGF, Peprotech) with 100 IU/mL penicillin and 100 μg/mL Streptomycin (Sigma) in a humidified atmosphere at 37°C. Neurospheres were formed after 3–5 days of culture. For passaging, free-floating neurospheres were collected and mechanically dissociated into small neurospheres or single cells and reseeded at a density of 10⁶ cells/mL in NSCs proliferation medium. NSCs at passage 3–8 were used in the following experiments.

To generate APP expression construct, we subcloned the human APP_695swe (APP) sequence (AuGCT DNA-SYN Biotechnology Co. Ltd. Beijing, China) into the GFP lentiviral vector pCDH-CMV-MCS-EF1-copGFP (System Biosciences; Mountain View, CA) at XbaI and NotI restriction sites (Invitrogen). The newly generated APP and three other helper plasmids pLP1, pLP2, and pLP/VSV-G (Invitrogen) were isolated from bacteria with the plasmid small kit without endotoxin (Omega Bio-tek; Norcross, GA), and their concentration was adjusted to 1 μg/μl. The 293T cells (Dalian Medical Unversity, Dalian, China) were cultured in DMEM supplemented with 10% FBS and 1%P/S. Ninety percent of confluent 293T cells in DMEM were transfected with plasmid DNA containing 15 μg APP or 15 μg GFP (negative control vector), 6.5 μg pLP1, 2.5 μg pLP2, and 3.5 μg pLP/VSV-G with Lipofectamine 2000 (Life Technologies, USA) in 10-cm dishes. After 6 h, the medium with plasmids was replaced by 10 ml fresh culture medium. The culture media were harvested after 2 days and filtered through a 0.45 μm membrane, then stored at −80°C for further use (Yao et al., 2015; Jiao et al., 2016). Lentiviral particles encoding APP-GFP and GFP were respectively transduced into NSCs. After 3 days, the stably transduced cells were determined by immunocytochemistry staining or cultured for future use.

Ost (structure shown in Figure 1A, purity >98%) was purchased from the National Institute for the Control of Pharmaceutical and Biological products (110822-200407; Beijing, China), dissolved in carboxymethyl cellulose sodium (CMC-Na, 0.05%), and subsequently store at 4°C (Chen et al., 2011; Kong et al., 2015).

The cell counting kitCCK-8 (Dojindo Laboratories; Kumamoto, Japan) was used to measure cell viability according to the manufacturer's instructions. Briefly, 10 μL of CCK-8 solution was added to each well containing 100 μL of cell culture supernatant (10⁵ cells) in 96-well plates. The reaction system was incubated at 37°C for 4 h. Absorbance was measured in a microplate reader (MR-96, Mindray, Shenzhen, China) at 450 nm (Chen et al., 2016; Jiao et al., 2016). The cell viability for each group was calculated by contrasting with the GFP group.

NSCs were seeded onto 6-well plates (2 × 10⁶ cell) and cultured at 37°C for 48 h. Subsequently, cells were collected and washed twice with PBS, resuspended in 500 μL of binding buffer and mixed with 5 μL of Annexin V-Light 650, and 5 μL of Propidium Iodide was added to the cells, followed by incubation in the dark for 15 min (Zou et al., 2013; Ma et al., 2014). Apoptotic cells were analyzed using a flow cytometer.

APP-expressing NSCs were transfected with antisense miR-9 oligonucleotide (GenePharma, Shanghai, China) using Lipofectamine 2000 reagent according to the manufacturer's instructions (Wang et al., 2011). miR-9 inhibitor: 5′-UCAUACAGCUAGAUAACCAAAGA-3′. Subsequently, cells were divided into five groups: GFP (transduced with GFP), APP (transduced with APP), APP+Ost, APP+miR-9 inhibitor, APP+miR-9 inhibitor+Ost. After 72 h of transfection, NSCs were cultured in differentiation medium composed of DMEM/F12 supplemented with 10% FBS plus 1% P/S for 10–14 days (Zhang et al., 2012). The medium was changed every 3 days.

Cryosections of brain were fixed with 4% paraformaldehyde, penetrated with 1% Triton X-100, and then washed three times with phosphate-buffered saline (PBS). Sections were incubated with 10% donkey serum albumin (Abcam) in PBS for 30 min, after which primary antibodies were added and incubated at 4°C overnight. The following primary antibodies were used: mouse anti-Nestin (1:150, Abcam, Cambridge, MA, USA), rabbit anti-Sox2 (1:150, Abcam), anti-neuronal nuclei (NeuN, 1:150, Abcam), anti-galactocerebroside (GalC, 1:150, Chemicon), anti-glial fibrillary acidic protein (GFAP, 1:150, Abcam), anti-NF-M (1:150, StemCell Technologies, Vancouver, BC, Canada), anti-APP (1:150), anti-Aβ (1:150, Abcam) and anti-synaptophysin (SYP, 1:150, Abcam). After washing three times with PBS, sections were incubated with Cy3-conjugated donkey anti-rabbit immunoglobulin lgG secondary antibodies (1:200, Jackson, West Grove, PA, USA) for 1 h at room temperature. All of these were then supplemented with DAPI nuclear dye (1:100, Sigma), washed third with PBS, and viewed using an inverted fluorescence microscope (Gu et al., 2015). Image J (National Institutes of Health, Bethesda, MD, USA) was used for quantitative analysis.

Quantitative real-time polymerase chain reaction (qRT-PCR) was performed as previously described (Li et al., 2016). Total RNA from tissue was extracted using TRIzol reagent (Carlsbad, CA, USA) and quantified using a spectrophotometer. RNA (3 μg) was converted to cDNA by reverse transcriptase using a RevertAidFirst Strand cDNA Synthesis Kit (Thermo Scientific, Lafayette, CO, USA). miR-9 and U6 small nucleolar RNA (Guangzhou RiboBio Co., Ltd, Guangzhou, China) were used for normalization and U6snRNA (U6) served as a control. The qRT-PCR primers used in the present study are described in Table 1. PCR was performed using 2 μL of cDNA and TransStart Top Green qPCR SuperMix (TransGen Biotech, Beijing, China) under the following conditions: 94°C for 30 s; 35 cycles of 94°C for 5 s, 60°C for 15 s, and 72°C for 10 s. The results are expressed as Cq-values. Relative changes in gene expression were determined using the ΔΔCt method, and β-actin was used as an internal control (Ma et al., 2014).

Equal amounts of proteins (50 μg) were loaded on 10% SDS-PAGE and transferred onto polyvinylidenedifluoride membranes. The membranes were blocked for 1 h with a blocking buffer containing 5% BSA in Tris-buffered saline solution and Tween 20 (10 mM Tris-HCl, 150 mM NaCl, 0.05% Tween 20; TBS-T). Membranes were then incubated overnight at 4°C with different primary antibodies diluted in the same blocking buffer. Incubations with HRP-conjugated secondary antibodies (Sigma) were performed for 1 h at room temperature and visualized by quantitative chemiluminescence using ECL Western blotting detection reagents (Millipore). Signal intensity was quantified using Image J. Antibodies used were as follows: anti-APP (1:1,000, Abcam), anti-Aβ (1:1,500, Abcam), anti-NICD (1:1,000, Abcam), and anti-Hes 1 (1:1,000; Santa Cruz Biotechnology). To control for loading, blots were stripped and reprobed with mouse anti β-actin (1:2,000; Santa Cruz Biotechnology; Smrt et al., 2010).

APP/PS1 double transgenic (Tg) mice, overexpressing mutated human APP and PS1 (APP_swe/PS1_ΔE9), which could effectively simulate the pathological features of AD patients (McClean and Holscher, 2014), were purchased from the Model Animal Resource Information Platform (Nanjing, China). Tg mice were housed under a 12-h light/dark cycle, with food and water freely available. Cognitive impairment first appeared in 7-month-old Tg mice, and Aβ deposits in the hippocampus could be detected in 9-month-old Tg mice (Zhou et al., 2015). In the present study, Tg mice at an age of 9 months were used and randomly divided into two groups (n = 8 for each group): an Ost-treated group (Tg+Ost), intragastrically (i.g.) treated with Ost (20 mg/kg, dissolved in 0.05% CMC-Na) daily for 6 weeks (Kong et al., 2015); wild-type (WT) C57BL/6 mice at the same age served as controls, and were treated with 0.05% CMC-Na (i.g.) daily for 6 weeks. All procedures involving animals were reviewed and approved by the Liaoning University of Traditional Chinese Medicine Institutional Animal Care and Use Committee.

After 6 weeks treatment with Ost, Morris water maze (MWM, Chengdu TME Technology Co., Ltd., Chengdu, China) was used to evaluate learning and memory capacity of the mice following methods previously described (Du et al., 2016). The apparatus consisted of a circular pool (120 cm diameter × 60 cm height) with a black inner wall, which was subdivided into four equal quadrants and filled with water (25°C) to the depth of 30 cm. An escape platform (10 cm diameter) was placed in one of the quadrants (the target quadrant) and submerged ~2 cm below the surface of the water. The test contained a platform trial that measured the animal's spatial acquisition ability and a spatial probe test that assessed memory. All the data, including the swim path and the swim time, were measured by a camera and automated analyzing system.

After the MWM test, mice were deeply anesthetized with chloral hydrate and perfused with 1% PBS. Subsequently, 4% paraformaldehyde was perfused to fix the brain. The brains were harvested and cryo-protected in PBS containing 30% sucrose until brains sank to the bottom. Subsequently, the brains were equilibrated to an optimum cutting temperature and placed in the freezer (Zeng et al., 2016). The brains were sectioned into 4–10 μm thick sections using a cryostat (CM1900, Leica). The sections located in the hippocampus were mounted on glass slides for staining and visualized on an OLYMPUS SZX9 and BX51 microscope (Tokyo, Japan) equipped with a digital camera.

In this study, 4 μm slices for Nissl staining. The sections (randomly chosen) were processed through different baths in the following order (and time): 100% ethanol (30 s), 95% ethanol (30 s), 70% ethanol (30 s), distilled water (30 s, three times), cresyl violet (56°C, 1 h), neutral differentiation solution (2 min), 100% ethanol (30 s), xylene (1 min); the samples were then mounted with neutral balata and covered with a coverslip (Zhai et al., 2015). The Nissl staining sections were visualized on a OLYMPUS SZX9 and BX51 microscope (Tokyo, Japan) with a digital camera.

All data are expressed as the means ± standard deviation (SD) and were analyzed using SPSS version 13.0 (SPSS, IL, USA). Differences of other parameters were evaluated using the analysis of variance for multiple groups. Differences were considered significant at P < 0.05.

To identify the immunocytochemical characteristics of NSCs, derived from the SVZ region of newborn C57BL/6 mice, the neural-specific markers Nestin and Sox2 were utilized. Immunocytochemical labeling showed that cultured NSCs were positive for Nestin and Sox2 (Figure 1B). In addition, NSCs changed morphology and developed into neurons (NeuN⁺), oligodendrocytes (GalC⁺), and astrocytes (GFAP⁺) in NSC differentiation medium after 10–14 days (Figure 1C), indicating that NSCs are multipotent (Oh et al., 2015).

We subsequently transfected these cells with APP/GFP or GFP lentiviral vectors. After 3 days of transduction, strong GFP expression was observed in ~82.3% cells transduced with both vectors, while strong APP staining was visible in the APP group (Figure 2A). RT-PCR and Western blot (Figures 2B,C) revealed the abundant expression of APP, which gives rise to the Aβ, as a neurotoxic oligomer (Haass and Selkoe, 2007; Schonrock et al., 2010b). APP mRNA and protein showed the highest expression in APP-expressing cells than in other groups, but the GFP group showed no difference with the Untrans group.

To evaluate whether Ost protects NSCs transduced with APP. APP-expressing NSCs were incubated with Ost (100 μmol/L) for 24 h (Yao et al., 2015), and subsequently, cell viability was determined using a CCK-8 kit. As shown in Figure 3A, APP-expressing NSCs markedly decreased cell viability (67.0 ± 12 vs. 100% in GFP group, n = 3, P < 0.01), while treatment with Ost obviously restored cell survival to 86.0 ± 6.9% compared to the APP group (n = 3, P < 0.01, Figure 3A). No significant differences were obtained between the GFP and GFP+Ost groups (P > 0.05).

To further investigate the protective effects of Ost on APP-expressing NSCs, cell apoptosis was assayed using flow cytometric analysis. The results showed that cell apoptosis reached 37.6 ± 2.2% in the APP group, a level significantly higher than observed in the GFP cells (vs. 4.6 ± 0.4% in GFP group, n = 3, P < 0.01, Figures 3B,C), and the addition of Ost resulted in a significant reduction in cell apoptosis (from 37.6 ± 2.21% to 23.5 ± 1.2%, n = 3, P < 0.01, Figures 3B,C) in the APP+Ost group. However, no change in cell death were observed between the GFP and GFP+Ost groups. Based on the above evidence, we successfully developed an AD cell model and 100 μmol/L Ost had no cytotoxicity (Lee et al., 2013).

At 10–14 days after NSCs were cultured under differentiation conditions, cells were fixed and immunostained with antibodies against Neurofilament M (NF-M; Yang et al., 2014; Xia et al., 2015). Here, immunofluorescence analysis revealed that the proportion of NF-M immunoreactive cells was 24.0 ± 1.6% in the APP group, which was lower than that in the GFP group, at 35.9 ± 2.5%, but Ost treatment significantly increased the percentage of NF-M positive cells (32.2 ± 1.9% vs. APP group, n = 3, P < 0.01, Figures 4A,B). Moreover, the NF-M-positive fluorescence signal was significantly decreased in the APP group compared with that in the GFP group (74.6 ± 3.4% vs. 100%, n = 3, P < 0.01), while Ost (88.0 ± 4.9%) treatment increased NF-M expression (n = 3, P < 0.01, Figure 4C), further confirming that Ost stimulated NSCs differentiation into neurons in an AD cell model.

To investigate whether Ost treatment could regulate the expression of miR-9 in APP-expressing cells, a miRNA qRT-PCR assay was used. In the present study, APP transduction led to the inhibition of miR-9 expression in these cells (0.57 ± 0.03 in the APP group vs. 1.00 ± 0.11 in the GFP group, n = 3, P < 0.01, Figure 4D). By contrast, Ost treatment resulted in the upregulation of miR-9 (0.90 ± 0.03 APP+Ost group vs. 0.57 ± 0.03 APP group, n = 3, P < 0.01, Figure 4D). To further investigate the role of miR-9 in the present study, we used an antisense miR-9 oligonucleotide to inhibit miR-9 (Wang et al., 2011). As shown in Figure 4D, miR-9 inhibitor significantly reduced the expression of miR-9, and neuronal differentiation was significantly decreased in the APP plus miR-9 inhibitor group (n = 3, P < 0.01, Figures 4A–C). Treatment with Ost partially increased the percentage of NF-M positive cells in the APP plus miR-9 inhibitor group (24.3 ± 1.98% vs. 20.43 ± 1.50, n = 3, P < 0.05, Figures 4A,B).

Several studies reported that APP could increase glial differentiation and decrease neuron differentiation of NPCs associated with the activation of Notch signaling (Kwak et al., 2011). The qRT-PCR results showed that APP-expressing NSCs did not change the expression of Notch 1 while the expression of Hes 1 was drastically increased compared with the GFP group (n = 3, P < 0.01, Figures 5A,B), indicating that APP-induced Notch signaling activation does not involve up regulation of Notch 1 expression. Treatment with Ost attenuated the increase of Hes 1 expression compared with the APP group (n = 3, P < 0.01, Figure 5B). Western blot analysis was performed to verify the levels of NICD and Hes 1 protein expression in these cells. These results shown in Figures 5C,D suggested that infection of APP significantly increased the protein levels of NICD and Hes 1 by 155.47 ± 9.69 and 166.68 ± 15.47% (P < 0.01), respectively, compared with the GFP group. Meanwhile, treatment with Ost reduced the protein levels of NICD and Hes 1 by 120.81 ± 10.50 and 130.71 ± 9.98% (n = 3, P < 0.01), respectively.

Inhibition of miR-9 led to increased levels of endogenous Hes 1 mRNA and NICD, Hes 1 proteins compared with the GFP group (P < 0.01, Figure 5), while treatment with Ost significantly reduced the levels of endogenous protein compared with the APP plus miR-9 inhibitor group (n = 3, P < 0.05). These results suggest that Ost stimulates APP-expressing NSCs differentiate to neurons partly through upregulation of miR-9 and inhibition of the Notch signaling pathway in APP-expressing cells.

To investigate the potential protective effects of Ost on the cognitive function of Tg mice, we first examined the effect of Ost on the water maze task. Spatial learning was expressed as the latency of distance and time spent on finding the escape platform in the water maze. During the navigation test, the typical swimming paths of the WT group, Tg and Tg + Ost groups were observed (Figure 6A). In the testing period, Tg mice showed that the latency to find the platform and the escape distance were significantly longer than observed for WT mice (n = 8, P < 0.01, Figures 6B,C). For treated Ost mice, the escape latency and escape distance were significantly decreased compared with Tg mice in the testing period (n = 8, P < 0.01, Figures 6B,C).

The platform was removed in the probe trial to measure the memory function on the 5th day of the test. The results indicated the number of platform crossings was notably reduced for Tg mice compared to WT mice, and the time spent in the target quadrant was significantly shorter for Tg mice than for WT mice (n = 8, P < 0.01 vs. the WT group, Figures 6D,E). In the Tg+Ost group, the number of platform crossings and time spent in the target quadrant were significantly increased compared to the Tg group (n = 8, P < 0.01 vs. the Tg group, Figures 6D,E). These results suggested that Ost treatment improved the learning and memory ability of APP/PS1 Tg mice.

Several studies have reported that Aβ accumulation is the major causative factor in AD development (Schonrock et al., 2010a). Therefore, we examined Aβ accumulation in APP/PS1 Tg mice. We confirmed a higher accumulation of Aβ in the hippocampus of APP/PS1 Tg mice using immunofluorescence (n = 8, P < 0.01 vs. WT group, Figures 6F,G), and Ost treatment reduced Aβ accumulation in the hippocampus compared to Tg mice (n = 8, P < 0.01 vs. the Tg group, Figures 6F,G).

The loss of neurons in the hippocampus area is one of characteristics of AD (Haass and Selkoe, 2007). Nissl staining revealed that the number of neurons was significantly reduced in the hippocampal area, including DG and CA3, compared to the WT group (264.0 ± 30.3 vs. 557.0 ± 40.1 cells in the DG; 313.8 ± 27.4 vs. 513.8 ± 41.4 cells in the CA3, P < 0.01, n = 8, Figures 7A,B). Treatment with Ost increased the number of neuronal cells in the hippocampal area (DG: 365.0 ± 30.2 cells; CA3: 372 ± 37.2 cells, vs. the Tg group, P < 0.01, n = 8, Figures 7A,B).

To further investigate the effect of Ost on neurons in Tg mice, immunofluorescence was used to analyze the expression of SYP (SYP) firstly. In the present study, the Tg group exhibited a significant reduction in the intensity of SYP in the hippocampus (55.3 ± 5.8%, v0.01) compared to the WT group (100%, n = 8, Figures 7C,D), while treatment with Ost attenuated the reduction of SYP and obviously restored the protein (78.4 ± 5.8%, P < 0.01 vs. Tg group, n = 8, Figures 7C,D). Subsequently, we also measured the concentration of SYP and PSD-95 using ELISA. The results showed that Ost increased the levels of SYP and PSD-95 compared to the Tg group (SYP: 12.40 ± 0.89 vs. 10.20 ± 1.49 ng/mL; PSD-95: 127.87 ± 20.20 vs. 101.31 ± 10.72 ng/L, P < 0.01, n = 8, Figures 7E,F). These data showed that Ost rescues the impairment of hippocampal neurons in APP/PS1 transgenic mice.

To investigate the neuroprotective mechanism of Ost in APP/PS1 transgenic mice, qRT-PCR was used to examine miR-9 expression. The results indicated that Ost significantly upregulated miR-9 expression in the DG and CA3 regions (n = 8, P < 0.01, Figure 8), which was consistent with the in vitro result.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer GHD and handling Editor declared their shared affiliation, and the handling Editor states that the process met the standards of a fair and objective review.