BDNF promotes human neural stem cell growth via GSK-3β-mediated crosstalk with the wnt/β-catenin signaling pathway.

Abstract Brain-derived neurotrophic factor (BDNF) plays important roles in neural stem cell (NSC) growth. In this study, we investigated whether BDNF exerts its neurotrophic effects through the Wnt/β-catenin signaling pathway in human embryonic spinal cord NSCs (hESC-NSCs) in vitro. We found an increase in hESC-NSC growth by BDNF overexpression. Furthermore, expression of Wnt1, Frizzled1 and Dsh was upregulated, whereas GSK-3β expression was downregulated. In contrast, hESC-NSC growth was decreased by BDNF RNA interference. BDNF, Wnt1 and β-catenin components were all downregulated, whereas GSK-3β was upregulated. Next, we treated hESC-NSCs with 6-bromoindirubin-3′-oxime (BIO), a small molecule inhibitor of GSK-3β. BIO reduced the effects of BDNF upregulation/downregulation on the cell number, soma size and differentiation, and suppressed the effect of BDNF modulation on the Wnt signaling pathway. Our findings suggest that BDNF promotes hESC-NSC growth in vitro through crosstalk with the Wnt/β-catenin signaling pathway, and that this interaction may be mediated by GSK-3β.


Introduction
Neural stem cells (NSCs) are a valuable resource for therapies against neurodegenerative diseases as well as neural repair because of their capacity for multipotent differentiation (Suksuphew & Noisa, 2015). However, the molecular mechanisms involved in the differentiation and proliferation of NSCs are still not fully understood. Numerous findings show that brain-derived neurotrophic factor (BDNF) promotes the survival, differentiation and proliferation of NSCs derived from rat tissues (Hsu et al., 2007). For example, BDNF improves the effects of NSCs in a rat model of Alzheimer's disease with unilateral fimbria-fornix lesions (Xuan et al., 2008). Additionally, BDNF enhances synapse development mediated by NSCs (Wang & Kisaalita, 2011). Another study reported that BDNF in dendrites promotes the differentiation and maturation of progenitor cells in the subgranular zone of the gyrus (Waterhouse et al., 2012). BDNF also promotes the mobility of a particular NMDA/GABA-responsive subset of neural progenitor cells (NPCs) (Jansson et al., 2012). Furthermore, transplantation of NSCs with high BDNF expression into traumatically injured brain tissue promotes recovery of motor functions (Ma et al., 2012). Exogenous BDNF added to NSCs transplanted into injured spinal cords promotes the survival, proliferation and differentiation of transplanted cells . Finally, some evidence indicates that BDNF regulates NSC proliferation through truncated TrkB receptor, MAP kinase, AKT and STAT-3 signaling pathways (Islam et al., 2009). However, the molecular mechanism underlying the modulating effect of BDNF on NSCs is still unclear (Islam et al., 2009).
It is well known that the Wnt/b-catenin signaling pathway is the key regulator of NSC survival, proliferation and differentiation (Chuang et al., 2015). Wnt/b-catenin signaling plays an important role in many types of NSCs. For example, Wnt/b-catenin determines stem cell fate and promotes stem cell differentiation during development and in adults (Lange et al., 2006). In addition, a recent study reported that the Wnt/ b-catenin signaling pathway and BDNF regulate NSC development in mice, and BDNF crosstalks with Wnt/b-catenin signaling (Chen et al., 2013). Another study reported that modulation of Wnt/b-catenin signaling improves cell replacement therapy in Parkinson's disease (PD) and plays a vital role in the maintenance and processing of stem cells (L'Episcopo et al., 2014).
The Wnt/b-catenin pathway regulates hippocampal neurogenesis, and Wnt-knockout mice show midbrain loss (Lange et al., 2006). In addition, Wnt/b-catenin signaling regulates adult hippocampal neurogenesis as well as NSC proliferation and differentiation (Lange et al., 2006;Mazemondet et al., 2011). BDNF and Wnt signaling can act together to exert their biological functions. For example, BDNF and Wnt cooperatively regulate dendritic spine formation (Hiester et al., 2013). In addition, BDNF expression is regulated by the Wnt/b-catenin signaling pathway (Yi et al., 2012).
A recently study reported that BDNF cross-talks with the Wnt/b-catenin signaling pathway in rat NSCs . However, whether BDNF and the Wnt/b-catenin signaling pathway act together in human embryonic spinal cord NSCs (hESC-NSCs) has not been studied. Furthermore, the crosstalk point and key crosstalk factor between BDNF and Wnt signaling are also unclear. In this study, we used human NSC spheres to investigate the effect of BDNF on human NSCs and the Wnt signaling pathway, whether there is crosstalk between BDNF and Wnt signaling, and the possible crosstalk point and identity of the key mediating factor. The results showed that BDNF promotes hESC-NSC proliferation and differentiation possibly through crosstalk with the Wnt/ b-catenin signaling pathway. In addition, GSK-3/b might be the key mediating crosstalk factor. The results provide an insight into the molecular mechanism underlying the regulatory effect of BDNF on hESC-NSCs. Furthermore, these findings might assist the development of new therapeutic strategies to modulate the proliferation, differentiation and development of hESC-NSCs.

Human embryos
All experiments conducted in this study were approved by the Medical and Health Research Ethics Committee of Kunming Medical University. Consent was obtained from all patients who voluntarily donated fetal tissues and 38 spontaneously aborted fetal specimens, of between 7 and 9 weeks' gestation, at the First People's Hospital of Yunnan Province. There were no visible developmental abnormalities. Tissues were dissected, and spinal cord tissue samples were collected.

Identification of cultured cells by immunofluorescence
To characterize the cultured cells, they were immunostained for stem cell-specific molecular markers [Nestin and Microtubule-associated protein 2 (p2)]. Briefly, after 7 days of culture, cells were washed three times with 0.01 mol/L PBS for 30 min, blocked in 5% donkey serum for 1 h at room temperature, and then incubated with primary antibodies (mouse-anti-Nestin, 1:500, and mouse-anti-p2, 1:500; Santa Cruz) at 4 C overnight. PBS was used as a negative control. The cells were then washed three times in PBST and incubated with secondary antibodies (mouse MoCy3, 1:1000, and rabbit Ra488, 1:1000; Jackson) and 4,6diamidino-2-phenylindole (DAPI, 1:1,000; Invitrogen) for 2 h at room temperature. Then, the cells were washed three times in PBST, mounted on slides and sealed with nail polish. Images were obtained under a confocal microscope (Leica). Control cells did not exhibit any immunolabeling, confirming the specificity of the antibodies.

RNA interference
The small interfering RNA (siRNA) vectors were constructed in our previous study (Yang et al., 2015). To determine the most effective siRNA construct and concentration, cultured hNSCs were treated with 10, 30, 50, 70 or 100 nM of each siRNA.

RT-pcr
Total RNA was isolated from hNSCs using TRI Reagent (Molecular Research Center). cDNA was generated using the RevertAid First Strand cDNA Synthesis Kit (Fermentas). For BDNF (primer sequences shown in Supplementary  Table 1), PCR was performed for 40 cycles, with a Tm of 52 C. For b-actin (primer sequences shown in Supplementary Table 1), used as an internal control, PCR was performed with a Tm of 52 C. Densitometric analysis following electrophoresis was performed using Quantity One software (Bio-Rad).

Real-time PCR
Quantitect SYBR Green PCR master mix (Qiagen) was used. A total reaction volume of 25 mL containing SYBR Green PCR master mix, 10 pmol of each primer and 1 mL of cDNA were used. A dilution series of a sample containing the gene of interest was used as the standard. Thermocycling parameters were as follows: 95 C for 3 min, followed by 40 cycles of 95 C for 15 s, annealing temperature for 30 s and 60 C for 40 s. To determine the specificity and identity of the PCR products, a melting curve analysis between 65 and 95 C was performed for every reaction following the instructions provided by the manufacturer. Gene expression was measured relative to b-actin expression.

Western blotting
Total cellular protein was extracted as described: the different hNSCs treatment groups were washed in cold PBS and lysed in radioimmunoprecipitation assay buffer (150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate and 50 mM Tris, pH 8.0). After centrifugation (12,000 rpm, 5 min) at 4 C, the supernatant was transferred to new tubes. The protein concentration of the samples was determined with a bicinchoninic acid protein assay (Pierce Chemical Company, Rockford, IL). A 40-mg sample of total protein was resolved using 12.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis and blotted onto polyvinylidene fluoride membranes. The membranes were blocked with 5% nonfat milk at room temperature for 1 h in TBST. Membranes were incubated overnight in primary antibody (Supplementary Table 2) at 4 C, and then incubated with secondary antibody for 2 h, and visualized using an enhanced chemiluminescence system (Pierce).

Neurosphere counting and measurement of the hESC-NSC area
In five randomly selected visual fields, images were acquired at 10Â magnification using a Leica inverted microscope imaging system. The neurosphere areas were counted 10 times. Leica LAS AF Lite software measured five neurosphere areas, the largest of which at 10Â magnification.

Calculation of the hESC-NSC differentiation rate
In five randomly selected visual fields, immunofluorescence staining for neuronal nuclei (NeuN; a neuronal marker) and glial fibrillary acidic protein (GFAP; a glial cell marker) was applied, and images were captured using the Leica confocal microscopy imaging system. Leica LAS AF Lite software was used to count NeuN+ and GFAP+ cells. The number of NeuN+ cells/total cells was the differentiation rate of neurons. The number of GFAP+ cells/total cells was the differentiation rate of glial cells.

MTT assay
MTT was purchased from Bio Protean Test Tech and was used to evaluate cell proliferation. Each compound tested was at a final concentration of 10 mg/mL, and each test was performed in triplicate. After incubation for 48 h in the presence or absence of the compound, 10 ml of MTT (5 mg/mL) was added to the wells, and the plate was then incubated at 37 C for 4 h to ensure the complete reduction of MTT to formazan by mitochondrial dehydrogenase. Subsequently, 100 ml 0.1% HCl and 10% sodium dodecyl sulfate solution was added to the plates and incubated overnight at 37 C to dissolve the formazan crystals. Absorbance at 570 nm was measured using a spectrophotometer (Perkin Elmer, Waltham, MA).

BIO inhibition
To determine the optimal concentration of BIO for GSK-3b inhibition, cultured cells were treated with 0, 2, 4 and 8 mmol/L BIO. Cells were incubated at 37 C prior to assessment of mRNA expression by RT-PCR.

Data analysis
Statistical analysis was performed using SPSS 18 (SPSS Inc., Chicago, IL). Intragroup comparisons were performed by analysis of variance. Post-hoc t-tests were used to evaluate intergroup differences. p Values of less than 0.05 were considered significant.

Characterization of hESC-NSCs
hESC-NSCs were identified by immunolabeling two NSCspecific markers, Nestin and p2, followed by confocal microscopy. Fluorescence signals from Nestin and p2 were present in the cytoplasm of almost all cells in neurospheres ( Figure 1A and B). Only DAPI staining was observed in the nucleus ( Figure 1C). Co-expression of Nestin and p2 was observed in the cells ( Figure 1D).

hESC-NSC growth after BDNF overexpression and RNAi
We investigated the growth of hESC-NSCs after BDNF overexpression and RNAi at various time points ( Figure 3A). The neurosphere number and area were significantly increased after BDNF overexpression compared with the control groups. In contrast, after RNAi, the number and area of neurospheres were reduced significantly compared with the control groups ( Figure 3B and C). The differentiation rate was also significantly increased at 24 and 48 h after BDNF overexpression compared with the control groups ( Figure 3D and E). In contrast, after RNAi, the differentiation rate was reduced significantly ( Figure 3D and E).

BDNF and Wnt signals after BDNF over-expression and RNAi
Expression levels of BDNF and TrkB BDNF mRNA and protein levels were significantly increased at 24 and 48 h after BDNF over-expression compared with normal and control groups. Expression of the BDNF receptor TrkB was also increased significantly at 48 h after BDNF over-expression. After RNAi, BDNF expression was reduced significantly compared with the control groups. The expression of TrkB was slightly increased at 48 h after RNAi compared with the control group (Tables 1 and 2, Figure 4).

Expression level of GSK-3
GSK-3b expression was significantly reduced after BDNF over-expression at 24 and 48 h compared with the control groups. However, the GSK-3b expression level was significantly increased after RNAi (Tables 1 and 2, Figure 4).

Expression levels of Wnt1/-catenin
Wnt1 expression was significantly increased in the BDNF over-expression group at 24 h, but significantly reduced after RNAi at the same time point compared with the control groups. Western blotting showed similar results. Interestingly, both Frizzled1 and Dash receptor levels were increased after BDNF over-expression for 24 h. In contrast, their levels were decreased after RNAi for 24 and 48 h (Tables 1 and 2, Figure 4).

Expression levels of other factors
PI3K expression was significantly reduced in the BDNF over-expression group after 24 and 48 h (Tables 1 and 2, Figure 4).

GSK-3b expression after BIO treatment
After BIO treatment of hESC-NSCs, RT-PCR showed that GSK-3b expression was reduced remarkably compared with the control group, and the optimal BIO concentration was 4 mol/L ( Figure 5).

Growth of hESC-NSCs after BIO co-treatment
hESC-NSCs were divided into normal, BDNF over-expression, BDNF RNAi, BIO, BDNF over-expression + BIO and BDNF RNAi + BIO treatment groups ( Figure 6A-F). The two BIO groups were treated with 4 mol/L BIO. Next, we investigated the number and area of NSC neurospheres. We found that the number and area of the hNSCs were all significantly increased after BDNF over-expression compared with the control groups. In contrast, after BDNF RNAi, the number and area of the neurospheres were reduced significantly compared with the control groups ( Figure 6G and H). The differentiation rate also indicated a significant increase in differentiation after BDNF overexpression compared with the control groups. In contrast, after BDNF RNAi, the differentiation rate was reduced significantly ( Figure 6I and J). When BIO was added to BDNF-over-expressing cells, the neurosphere number and area were increased but still lower than those in the control group. However, in the BDNF RNAi group, only the area was increased. When BIO was added, the differentiation rate was increased ( Figure 6).

Expression levels of BDNF and TrkB
After BDNF over-expression and BIO treatment, BDNF expression was significantly increased at 24 and 48 h. TrkB Figure 1. Identification of hESC-NSCs. Neurospheres cultured for 1 week were Nestin + (red, A) and p2 + (green, B). DAPI labeling is blue (C). Co-expression of Nestin + and p2 + is shown in D. Bar ¼ 250 mm. expression was also significantly increased at 48 h. Interestingly, BDNF over-expression or RNAi plus BIO treatment did not cause significant changes in BDNF or TrkB levels compared with BDNF over-expression or RNAi only. GSK-3b levels were significantly increased after RNAi and BIO treatment at 24 and 48 h (Tables 3 and 4, Figure 7).
Expression levels of Wnt/-catenin signaling mRNA expression of Wnt signaling factors was reduced in the BDNF over-expression + BIO treatment group. Western blotting showed similar results. Protein expression was also reduced significantly after RNAi and RNAi + BIO treatment for 24 or 48 h. However, in the BDNF over-expression + BIO group, Wnt expression showed a significant reduction compared with the BDNF over-expression group. In particular, siRNA transfection and BIO treatment caused a reduction in Wnt signals. In addition, BIO treatment of BDNF overexpression and RNAi groups caused significant and consistent reductions of Dsh levels at the same time points compared with the control groups (Tables 3 and 4, Figure 7).

Expression levels of other factors
PI3K expression was significantly reduced in the BDNF over-expression group and BDNF over-expression + BIO treatment group after 24 and 48 h. RNAi and BIO treatment significantly increased PI3K expression compared with RNAi only. BIO treatment further reduced PI3K expression in the BDNF over-expression group. Real-time PCR also showed a reduction of PLC-g1 expression in the RNAi + BIO treatment group at 24 h after treatment. PLC-g1 expression was reduced in BDNF over-expression and RNAi + BIO treatment groups compared with the control group (Tables 3 and 4, Figure 7). , and interference groups, and analyzed by RT-PCR. The transfection group was transfected with pIRES2-ZsGreen1-BDNF (A), and the interference group was treated with siRNA (B). Strong green fluorescence was seen in transfection and interference groups, indicating that pIRES2-ZsGreen1-BDNF and siRNA transfection were successful. Bar ¼ 50 mm. BDNF and b-actin mRNA levels were measured by RT-PCR after transfection with pIRES2-ZsGreen1-BDNF. As a result, 50 ng was found to be the optimal concentration for pIRES2-ZsGreen1-BDNF transfection (C), and 50 nM was the optimal concentration for RNAi (D). *p50.05, control vs. vector or Lipo2000 groups. #p50.05 vs. prior lower concentration. Results are expressed as the mean ± SEM; n ¼ 5.

Discussion
In this study, we investigated the role of BDNF in hESC-NSCs and tested the hypothesis that BDNF promotes hESC-NSC growth by crosstalk with Wnt signaling, which involves GSK-3b as the key crosstalk factor. There are three major novel findings in this study. First, BDNF promoted the differentiation and development of hESC-NSCs and the volume of neurospheres were increased by BDNF overexpression. Second, BDNF modulated the expression of some Wnt component genes in hESC-NSCs and crosstalk with Wnt signaling. Third, the regulatory effect of BDNF appeared to be suppressed by BIO treatment. When hESC-NSCs were transfected with the BDNF plasmid or siRNA and treated with BIO, RT-PCR and western blot analyses showed consistent downregulation of Wnt signaling, suggesting that the BDNF plasmid could not upregulate Wnt signals upon inhibition of GSK-3b.

BDNF promotes the growth, proliferation and differentiation of hESC-NSCs
NSCs exist widely in the striatum, cerebellum, midbrain, hippocampus, cortex and spinal cord of humans and rodents. They have the characteristics of low immunogenicity, selfrenewal, multipotent differentiation, cell fusion, migration and long-term survival (Rao, 2004). NSCs can proliferate and differentiate into neurons, astrocytes and oligodendrocytes (Davis et al., 2006). The regulation of stem cells can be affected by internal and external factors including intrinsic differentiation programs, growth factors, the extracellular matrix, cell adhesion factors and cellular interactions (Nakano & Kornblum, 2009). As a neurotrophic factor, BDNF plays an important role in various biological functions of the central nervous system, such as neuronal development, differentiation, synapse formation, progenitor maturation, survival, axonal regeneration Figure 6. Growth and differentiation of hESC-NSCs after BIO treatment at 0, 24 and 48 h. hESC-NSC growth after BDNF overexpression or RNAi and the effect of BIO (A-F). The number and soma area of hESC-NSCs were increased and decreased after upregulation and downregulation of BDNF expression, respectively, and reduced by BIO treatment (G, H). The differentiation rate (neurons and glial cells) of hESC-NSCs (I, J). The differentiation rate of neuron and glial cells was increased and decreased after upregulation and downregulation BDNF expression, respectively. The differentiation rate of neurons and glial cells was reduced by RNAi + BIO treatment, while the neuron differentiation rate was not affected by BDNF overexpression + BIO treatment (I, J   and neuronal damage (Chen et al., 2007). BDNF and its specific receptor TrkB are widely expressed in developing and adult brains, and function as key neurotrophic factors for the generation of dopamine neurons, synapse maturation and plasticity, neuronal cell survival, morphogenesis, neural plasticity and NPC migration (Horne et al., 2010). The therapeutic effects of BDNF have been studied in PD, Alzheimer's disease, Huntington's disease, amyotrophic lateral sclerosis and other neurological diseases (Blurton-Jones et al., 2009;Giralt et al., 2010;Sadan et al., 2009;Wang et al., 2002). In addition, it has been shown that transplanted NSCs improve cognitive functions in Alzheimer's disease via BDNF that promotes NSC proliferation and differentiation and protects against neurotoxin-induced NSC apoptosis (Blurton-Jones et al., 2009). The signaling events downstream of BDNF have been widely studied, including the following: (1) phosphorylation of PLC-g, which produces the second messengers IP3 and DAG, and increases Ca 2+ and PKC activation; (2) activation of the MAPK pathway by Ras and Raf; (3) phosphorylation of GSK-3b, which leads to transfer of cytoplasmic b-catenin into the nucleus; (4) activation of the LEF/TCF transcription factor complex (Li et al., 2007). It has also been shown that the effect of BDNF on NSCs mainly occurs during cell mitosis at a relatively mature stage but not in the early stage, which can cause NPCs to differentiate into neurons and oligodendrocytes (Nakamura et al., 2011). Interestingly, in this study, after BDNF plasmid transfection, there were no remarkable changes in PI3K, Ca 2+ or PLC-g levels, suggesting that the signaling pathways related to these factors are not involved in NSC growth.
NSCs have neuroprotective and beneficial effects on cognitive and motor function recovery in aging and spinal cord-injured rats by upregulating the expression of BDNF (Park et al., 2013). BDNF secreted from human NPCs can rescue amyloid b-induced toxicity in cultured rat septal neurons (Kitiyanant et al., 2012). NSCs and astrocytes act in neuroprotection by BDNF/Wnt signaling, cell transplantation and gene therapy (Isacson & Kordower, 2008). However, prior to the present study, direct evidence that hNSCs can be modulated by the expression level of BDNF was not available. In this study, after transfection of BDNF, RT-PCR, western blotting and immunolabeling showed strong positivity for BDNF in cultured cells, suggesting that BDNF plasmid construction was successful and that its transfection into cultured cells was efficient. Quantitative experiments showed that BDNF transfection promoted the growth of hESC-NSCs. After transfection, there was an obvious increase in the expression of stem cell markers. In addition, cell counting and differentiation assays showed significant increases in both the cell number and differentiation rate. These results indicate that increased BDNF levels promote the growth of cultured cells. This finding provides the experimental evidence that the development of hESC-NSCs can be regulated by BDNF in vitro.

BDNF functions via the Wnt signal pathway
The Wnt/b-catenin pathway is a key regulator of growth and proliferation, and it is widely involved in cell development, proliferation, differentiation, migration, adhesion, polarization and tumorigenesis (Lupo et al., 2013;Rudloff & Kemler, 2012). It also plays an important role in neural development. b-Catenin is the central component in the Wnt/b-catenin signaling pathway. It not only transmits information in the cytoplasm, but also translocates to the nucleus to activate target gene transcription (Munji et al., 2011). Wnt signaling also plays an important role in controlling stem cell division and differentiation (Voskas et al., 2014). A previous study reported that Wnt signaling induces NSCs to differentiate into neurons and astrocytes but not oligodendrocytes (Prakash & Wurst, 2007).
Expression of all Wnt genes is found in mouse embryonic stem cells that differentiate into neurons (Nordin et al., 2008). Some studies have indicated that the Wnt signaling pathway and its downstream target genes play an important role in NSC development, differentiation and proliferation (Lange et al., 2006;Munji et al., 2011). In addition, other studies have shown that transfection of Wnt3a and Wnt5a into NSCs of the adult rat subventricular zone induces neuronal differentiation. The effects of Wnt3a are more obvious, and the same effects occur in adult hippocampal NSCs (Hubner et al., 2010;Marinaro et al., 2012). Wnt3a and Wnt3 stimulate NSC proliferation, differentiation and neurogenesis through the canonical Wnt signaling pathway (David et al., 2010). Furthermore, Wnt/b-catenin signaling plays an important role in differentiation of embryonic stem cells into neurons induced by retinoic acid and plays a role in neural crest NSC differentiation into sensory neurons (Otero et al., 2004;Lee et al., 2004). Finally, canonical Wnt signaling transiently stimulates proliferation and enhances neurogenesis in neonatal NPC cultures (Hirsch et al., 2007).
Increased Wnt signaling can rescue neurons from degeneration by b-amyloid fibril insult, and the agonist lithium has a therapeutic effect against acute brain injury and neurodegenerative diseases (Wada et al., 2005). BDNF and the Wnt signaling pathway might act together in related biological activities. Consistently, in this study, we found upregulation of Wnt signaling at 24 and 48 h after BDNF transfection but downregulation after siRNA transfection. These results suggest that BDNF is a strong regulator of Wnt signaling. In addition, b-catenin expression was upregulated after transfection. Therefore, BDNF appears to trigger the Wnt/ b-catenin signaling pathway in cultured cells. This result is consistent with previous reports that Wnt signaling is upregulated by BDNF in cultured mouse NSCs in a dosedependent manner (Chen et al., 2013). There are other reports regarding the functional synergistic interactions between BDNF and Wnt signaling. BDNF and Wnt signaling cooperatively regulate dendritic spine formation (Mazemondet et al., 2011). A previous study also reported that the expression of BDNF is regulated by the Wnt signaling pathway (Hiester et al., 2013). It has been demonstrated that BDNF stimulates neural differentiation of stem cells and the survival of differentiated cells through MAPK/ERK and PI3K/Akt-dependent signaling pathways (Lim et al., 2008). BDNF also participates in the generation of endogenous Wnt ligands and crosstalk with intracellular b-catenin signaling cascades in cultured NSCs (Hutchison, 2012). However, our study is the first report that explores the regulatory effect of BDNF on Wnt signaling in cultured hESC-NSCs, and revealed the underlying molecular mechanism of BDNF regulation in hESC-NSC development.
GSK-3b may be the key factor in crosstalk between BDNF and wnt signaling pathways GSK-3b plays an important role in many neuronal functions, including neurite outgrowth, synapse formation, neurotransmission and neurogenesis (Kim et al., 2009). GSK-3b also plays a crucial role in the differentiation and survival of adult rat NSCs (Maurer et al., 2007). By inhibition of GSK-3b and activation of Wnt signaling, the multilineage differentiation potential can be maintained in NSCs (Munji et al., 2011). For example, progranulin enhances NPC proliferation through GSK-3b phosphorylation (Nedachi et al., 2011). GSK-3b rapidly induces long-term self-renewal of primitive neural precursors from human embryonic stem cells . GSK-3b also regulates the differentiation and proliferation of NSCs from the rat subventricular zone (Shimizu et al., 2008). Furthermore, GSK-3b is a co-downstream signaling pathway node that modulates the interactions between BDNF and Wnt signaling pathways. In this study, BDNF did not upregulate Wnt signals upon inhibition of GSK-3b. Therefore, GSK-3b may be the key factor in crosstalk between BDNF and Wnt signaling pathways.

Conclusions
In this study, we found that BDNF promotes the proliferation, differentiation and development of hESC-NSCs, and upregulates Wnt/b-catenin signaling. In particular, GSK-3b was downregulated and may be the key crosstalk factor between BDNF and Wnt signaling. After BIO treatment and BDNF plasmid transfection of the cultured cells, there was significant downregulation of most factors compared with the control groups at the same time points. These results suggest that GSK-3b plays an important role in the regulatory effect of BDNF on the expression of other factors including Wnt. Furthermore, after inhibition of GSK-3b, there was consistent reduction in Wnt signals, despite BDNF plasmid transfection. Thus, based on the above results, we concluded that GSK-3b might be a critical crosstalk factor in BDNF crosstalk with Wnt/b-catenin signaling. More in-depth study of the molecular mechanisms regulating the effect of BDNF in stem cells and the crosstalk between BDNF and Wnt signaling may provide novel strategies for future NSC therapies.

Declaration of interest
This work was supported by the Natural Science Foundation of China (NO 30960156, NO 31260253, NO 31560295 The authors report no declarations of interest.