Introduction

Bcl-2 gene was identified at the breakpoint of the t(14;18) (q32;q21) chromosomal translocation that is associated with human follicular lymphoma, and since then, it has been shown that the Bcl-2 protein is mainly located on the mitochondrial outer membrane and is responsible for the regulation of programmed cell death (Pegoraro et al. 1984; Hockenbery et al. 1990). Currently, it is recognized that Bcl-2 can not only modulate apoptosis by interacting with other pro- or anti-apoptotic Bcl-2 family members, but it also has an anti-autophagy function by directly binding with Beclin-1 (Adams and Cory 1998; Pattingre et al. 2005). Consequently, Bcl-2 is considered to be a arbiter of cell survival for development, tissue homeostasis, protection against diverse insults, and various diseases including cancer (Adams and Cory 1998).

In mice, Bcl-2 is highly expressed in the organs related to the immune system (such as the thymus and the spleen) and the brain (Hockenbery et al. 1991; Abedohmae et al. 1993). During brain development, the mRNA levels of Bcl-2 are the highest on embryonic day 15, which is the stage of neurogenesis, indicating that Bcl-2 may have a considerable impact on the process of neuronal development in the CNS (central nervous system) (Abedohmae et al. 1993; Merry et al. 1994). According to previous reports, overexpression of Bcl-2 prevents neurotrophic factor deprivation-induced neuronal apoptosis in embryonic neurons (Allsopp et al. 1993). Moreover, Bcl-2 transgenic mice are protected against middle cerebral artery occlusion-induced ischemia (Martinou et al. 1994), and Bcl-2-deficient mice exhibit progressive degeneration of motor, sympathetic, and sensory neurons during early postnatal development (Michaelidis et al. 1996). In addition to these neuroprotective functions, Bcl-2 has also been shown to influence neurogenesis in pheochromocytoma and neuroblastoma cell lines (Sato et al. 1994; Hanada et al. 1993). Zhang et al. (1996) reported that Bcl-2 directly modulated neuronal differentiation in human neural crest cells, and Hilton et al. (1997) suggested that Bcl-2 had an impact on the axonal growth rate of the embryonic sensory neurons. Moreover, Bcl-2 promotes regeneration of injured axons in the mammalian CNS (Chen et al. 1997), indicating that Bcl-2 may have a direct function on neurogenesis in addition to its well-documented role as a suppressor of apoptosis. Nevertheless, the precise non-apoptotic functions of Bcl-2 during neuronal differentiation still remain unclear.

Neurogenesis is the differentiation process for neuronal lineage commitment of multipotent neural stem cells either during the development of the fetal nervous system or during the regeneration of an impaired adult nervous system, which is spatiotemporally coordinated by a variety of intrinsic and extrinsic factors in stem cell niches (Gage 2000; Paridaen and Huttner 2014). Proneural bHLH proteins, including NeuroDs and neurogenins, are essential transcriptional regulators for cell fate specification in neural stem cell niches (Paridaen and Huttner 2014; Ross et al. 2003). Moreover, deregulated expression or mutation of the bHLH genes occurs in brain cancers, as well as neurodevelopmental, neuropsychiatric, and neurodegenerative disorders (Dennis et al. 2019). Their expression is also regulated by various signaling networks such as Notch, Wnt, and BMP (Dennis et al. 2019). However, the signaling pathway that couples Bcl-2 to the regulation of bHLH transcription factors during neuronal differentiation is not yet fully understood.

Here, we aimed to clarify the non-apoptotic functions of Bcl-2 during neuronal differentiation and to explain the mechanism by which Bcl-2 regulates neurite formation. Using gain- or loss-of-function studies, we showed that the neurite lengths of differentiating H19-7 cells were affected by the modulation of the Bcl-2 gene. Moreover, we observed that Bcl-2 overexpression influenced various neurogenic factors by analyzing whole-genome expression patterns obtained through microarrays. In conclusion, we demonstrated that Bcl-2 overexpression promoted neurite outgrowth through the Bmp4/Tbx3/NeuroD1 cascade in the H19-7 cells, indicating that Bcl-2 may have a direct role in neuronal development in addition to its well-known anti-apoptotic function.

Materials and Methods

Materials

The cell culture materials, such as Dulbecco’s modified Eagle’s medium (DMEM) with l-glutamine and low glucose, fetal bovine serum (FBS), phosphate-buffered saline (PBS), penicillin/streptomycin, and trypsin/EDTA, were purchased from WISENT (St. Bruno, QC, Canada). G-418 was obtained from Duchefa (Haarlem, Netherlands). RNAiso Plus for RNA extraction was purchased from Takara Bio (Otsu, Japan), Reverse Transcription Master Mix for cDNA synthesis was obtained from Elpis Bio (Daejeon, Korea), and GoTaq® DNA polymerase kits were purchased from Promega (Madison, WI, USA). SensiFAST™ SYBR® No-ROX kits were purchased from Bioline (London, UK). The antibodies used were purchased as follows: Bcl-2 (SC-7382) from Santa Cruz Biotechnology (Santa Cruz, CA, USA), Calnexin (ADI-SPA-860) from Enzo Life Sciences (Farmingdale, NY, USA); Bmp4 (ab39973), Tbx3(ab99302), and NeuroD1 (ab16508) from Abcam (Cambridge, UK); GFP (11814460001) from Roche (Indianapolis, IN, USA), β-tubulin type III (TUJ1, 802001) from BioLegend (San Diego, CA, USA), Alexa Flour® 488-conjugated goat anti-rabbit IgG (A11008) from Thermo Fisher Scientific (Rockford, IL, USA), and Cy™ 3-conjugated AffiniPure goat anti-rabbit IgG (H + L) (111-165-144) from Jackson ImmunoResearch (West Grove, PA, USA). All other chemicals were of analytical grade.

Cell Culture

H19-7 cells, which are conditionally immortalized rat hippocampal precursor cells with a temperature-sensitive SV40 large T antigen, were maintained in DMEM supplemented with 10% (v/v) heat-inactivated FBS, 100 units/mL penicillin, and 100 µg/mL streptomycin under 0.2 mg/mL G-418 selection at 33 °C (the temperature at which the SV40 large T antigen is functional) in a humidified incubator with 5% CO2. To induce differentiation, the medium was changed to N2 medium supplemented with 20 ng/mL bFGF (R&D Systems, Minneapolis, MN, USA), and the cells were incubated at 39 °C for 2 days. Neural precursor cells (NPCs) were obtained from pregnant Sprague–Dawley rats. Protocols were reviewed and approved by the Institutional Animal Care and Use Committee of Hanyang Laboratory Animal Research Center (2014-0126A). The embryonic brain cortices from E14 rat embryos were mechanically triturated in a Ca2+/Mg2+-free Hank’s balanced salt solution (HBSS; Gibco, Grand Island, NY, USA), seeded at 2 × 105 cells in 100-mm culture dishes (Corning Life Sciences, Acton, MA, USA), pre-coated with 15 µg/mL poly-l-ornithine (Sigma-Aldrich, St. Louis, MO, USA) and 1 µg/mL fibronectin (Invitrogen, Carlsbad, MO, USA), and cultured for 5 to 6 days in N2 medium supplemented with bFGF. The cell clusters generated were dissociated in HBSS, plated on fresh culture dishes, and induced for additional proliferation in N2 + bFGF up to 70 to 80% cell confluency before the cells were subjected to differentiation for 2 days.

Gene Expression Profiling Using cDNA Microarrays

Microarray analysis was performed using tools from the commercial microarray service Ebiogen (Seoul, Korea) and deposited in the NCBI Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo) repository (accession number GSE130481). In brief, the gene expression profiles of rat NPCs from the vector + bFGF group (proliferation control), the vector − bFGF group (differentiation control), and the Bcl-2 + bFGF group (Bcl-2 overexpression) were compared using cDNA microarrays. NPCs derived from the E14 rat embryos were transiently transfected with 4 µg of either MSCV-IRES-EGFP or rBcl-2-MSCV-IRES-EGFP using a Nucleofector™ kit (Lonza, Walkersville, MD, USA) according to the manufacturer’s protocol. Total RNA was then isolated and amplified. The amplified cRNA was hybridized to an Agilent Rat Oligo Microarray (44 K) (Agilent Technology, Palo Alto, CA, USA) at 65 °C for 17 h in an Agilent DNA microarray hybridization oven and then washed. The hybridized images were scanned using an Agilent DNA microarray scanner and quantified using Feature Extraction software (Agilent Technology). The data normalization and selection of the genes that had undergone fold changes were performed using GeneSpring GX software, version 7.3 (Agilent Technology). The gene classification was performed using BioCarta (http://www.biocarta.com), GenMAPP (http://www.genmapp.org), and DAVID (http://david.abcc.ncifcrf.gov). To obtain insights into the biological processes related with neural development, Gene Ontology (GO) analysis was performed. Using GO terms corresponding to ‘nervous system development,’ the neural development-related genes were selected for further analysis. All microarray analyses and visualization were conducted using R software, version 3.4.1 (www.r-project.org).

Retroviral Construction and Infection

Rat Bcl-2 cDNA was amplified from NPCs of E14 Sprague–Dawley rat embryos using PCR and cloned into the retroviral vector MSCV-IRES-EGFP. The construct was obtained using a forward primer containing a BglII site (5′-GAAGATCTATGGCGCAAGCCGGGAGA-3′) and a reverse primer containing an EcoRI site (5′-GGAATTCTCACTTGTGGCCCAGGTATGCAC-3′). PCR mixtures contained 1 µL of cDNA from NPCs, 5 µL of 10X PCR buffer, 5 µL of dNTP mix (2.5 mM), 1 µL of Taq polymerase (10 units/µL, Pyrobest, Takara Bio) adjusted to 50 µL with distilled water. The PCR amplification consisted of denaturation for 5 min at 95 °C followed by 30 cycles of 30 s at 95 °C, 30 s at 60 °C, and 90 s at 72 °C, with a final extension for 7 min at 72 °C. The PCR product was separated via electrophoresis on a 1.5% agarose gel and visualized using a 3UV™ transilluminator (UVP, Upland, CA, USA). The DNA band of the expected size was excised, purified using a FavorPrep™ GEL/PCR purification mini kit (Favorgen Biotech, Pingtung, Taiwan), and digested with the appropriate restriction enzymes (BglII and EcoRI; New England Biolabs, Beverly, MA, USA) at 37 °C. The digested sample was then ligated into the MSCV-IRES-EGFP retroviral vector. The cloned rat Bcl-2 was confirmed using sequencing (Cosmo Genetech, Seoul, Korea), and the sequence was aligned with the NCBI rat Bcl-2 mRNA sequence (NM_016993.1) using the Genomatix pairwise comparison program (http://www.genomatix.de/cgi-bin/dialign/dialign.pl). The viral particles were produced by transfecting the retrovirus packaging cell line 293GPG with the cloned vector using polyethylenimine (PEI; Sigma-Aldrich), and supernatants containing viral particles were harvested after incubation for 48 h. For viral infection, H19-7 cells were incubated with a viral suspension (4 × 106 particles/mL) containing polybrene (1 µg/mL; Sigma-Aldrich) for 4 h followed by transfer to fresh medium.

Transient Transfection with Small Interfering RNAs

Bcl-2, Bmp4, Tbx3, and NeuroD1 siRNAs, as well as control siRNA (Bioneer, Daejeon, Korea), were transiently transfected into H19-7 cells for loss-of-function studies. The siRNA sequences were as follows: Bcl-2 (5′-GAAUCAAGUGUUCGUCAUA-3′ and 5′-UAUGACGAACACUUGAUUC-3′), Bmp4 (5′-GACUACUGGACACCAGACU-3′ and 5′-AGUCUGGUGUCCAGUAGUC-3′), Tbx3 (5′-CUGGAUAAAAAGGCCAAGU-3′ and 5′-ACUUGGCCUUUUUAUCCAG-3′), and NeuroD1 (5′-GAAACAUGACCAAAUCAUA-3′ and 5′-UAUGAUUUGGUCAUGUUUC-3′). Each siRNA (100 nM) was introduced into the cells using PEI transfection reagent.

Immunostaining and Measurement of Neurite Outgrowth

Cells were washed with PBS and fixed with 4% (w/v) paraformaldehyde (PFA; Sigma-Aldrich) for 15 min followed by washing three times with 0.1% bovine serum albumin (BSA; Sigma-Aldrich) in PBS at 25 °C. The cells were permeabilized with 0.3% Triton X-100 in PBS for 20 min and blocked with 10% normal goat serum in PBS for 1 h at room temperature. Next, the cells were incubated with anti-TUJ1 antibody (1:2000) and anti-GFP antibody (1:500) at 4 °C followed by incubation with Cy™ 3-conjugated AffiniPure goat anti-rabbit IgG (H + L) (1:200) and Alexa Flour® 488-conjugated goat anti-rabbit IgG (1:500) for 1 h before mounting with Vectashield (Vector Laboratories, Burlingame, CA, USA) containing 4, 6-diamidino-2-phenylindone (DAPI). The immunoreacted cells were imaged using a fluorescence microscope (Leica Microsystems, Wetzlar, Germany), and the morphological characteristics were quantified using ImageJ software (NIH; http://rsb.Info.nih.gov/ij/). The neurite length was defined as the distance from the soma to the tip of the branch. The neurite length data were obtained from ten randomly selected areas from at least three independent experiments, and more than 100 cells were counted for each group in each experiment.

Reverse Transcription (RT)-PCR and Real-Time PCR

Total RNA was isolated using RNAiso Plus according to the manufacturer’s protocol. Next, 1 µg of isolated total RNA was incubated with Reverse Transcription Master Premix kit for 60 min at 37 °C. The cDNA was mixed with specific primers and GoTaq® DNA polymerase, and the mixtures were amplified using a thermocycler (Takara Bio). The primer sequences used were as follows: Bcl-2 (5′-CCTGGCATCTTCTCCTTCCA-3′ and 5′-CACCCCATCCCTGAAGAGTT-3′), Gapdh (5′-GGCATTGCTCTCAATGACAA-3′ and 5′-AGGGCCTCTCTCTTGCTCTC-3′), Bmp4 (5′-GAGATCGCCACCTACAGGAA-3′ and 5′-GCTCCTGCTTCGACTCCTTA-3′), Tbx3 (5′-TTTGGCCATGTCGCCTTTTG-3′ and 5′-GAGCGGCTATTCAGTTCCGA-3′), NeuroD1 (5′-CTCAGTTCTCAGGACGAGGA-3′ and 5′-TAGTTCTTGGCCAAGCGCAG-3′), NeuroD2 (5′-TTCTCGCTCAAGCAGGACTC-3′ and 5′-GCGTTGAGCTCCTCGTACAT-3′), Mash1 (5′-GGCTCAACTTCAGTGGCTTC-3′ and 5′-TGGAGTAGTTGGGGGAGATG-3′). The PCR conditions were as follows: denaturation at 95 °C for 30 s, annealing at 60 °C (NeuroD1, 58 °C) for 30 s, and extension at 72 °C for 40 s (Bcl-2, 28 cycles; Gapdh, 25 cycles; Bmp4, Tbx3, NeuroD1, NeuroD2, and Mash1, 33 cycles). The PCR products were analyzed on a 1.5% agarose gel. For real-time PCR, 1 µL of cDNA was amplified in duplicate in a final volume of 20 µL using the SensiFAST™ SYBR NO-ROX® kit and the RT-PCR primers listed above. Thermocycling conditions were 95 °C for 10 min followed by 40 cycles of 95 °C for 15 s and 58 °C (NeuroD1) or 60 °C (all other genes) for 1 min. The primer sequences for real-time PCR were the same as for RT-PCR.

Western Blot Analysis

Protein expression was analyzed by immunoblotting as described previously (Zhang et al. 2019; Pan et al. 2017). The cells were lysed in 20 mM Tris/HCl at pH 7.5 with 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 2.5 mM sodium pyrophosphate, 1% Triton X-100, and 1 mM PMSF. Samples containing 30 µg of total protein were loaded onto 12% SDS-PAGE gels, electrophoresed, and transferred onto nitrocellulose membranes (Amersham Biosciences, Piscataway, NJ, USA). After blocking with 5% nonfat dried milk for 1 h, the membranes were incubated with primary antibodies followed by a horseradish peroxidase (HRP)-conjugated secondary antibody (1:2000; Jackson ImmunoResearch). Specific bands were detected using enhanced chemiluminescence (ECL) (Thermo Fisher Scientific), and band intensities were quantified using ImageJ software.

Statistical Analysis

Statistical analysis was performed with SPSS software, version 18.0 (SPSS Inc., Chicago, IL, USA). The data were expressed as the mean ± SD of at least three independent experiments. Comparisons between groups were made using the unpaired Student’s t test. p values < 0.05 were considered statistically significant.

Results

Bcl-2 Affects Neuronal Differentiation of Rat Hippocampal Progenitor Cells

We previously reported that phospholipase D1 (PLD1), a membrane-anchored enzyme that hydrolyzes phosphatidylcholine into phosphatidic acid, which acts as a lipid second messenger, and the polar head group choline (Park and Han 2018), increases Bcl-2 expression via the PLA2/Cox2/EP4/PKA/p38MAPK/CREB cascade in rat NPCs and via the JNK/STAT3 pathway in rat hippocampal progenitor cells during neuronal differentiation (Park et al. 2014; Yoon et al. 2012). Next, we sought to determine how Bcl-2 influences the process of neuronal differentiation. To investigate whether the modulation of Bcl-2 impacts neurite outgrowth, control or Bcl-2 siRNA was transfected into H19-7 cells and then stimulated using N2 medium supplemented with 20 ng/mL bFGF for neuronal differentiation. As shown in Fig. 1a and b, bFGF-induced neurite outgrowth was repressed by Bcl-2 depletion compared to the bFGF-treated control siRNA group. Inversely, Bcl-2 was overexpressed through retroviral transduction with MSCV-IRES-EGFP or rBcl-2-MSCV-IRES-EGFP followed by stimulation with bFGF. As shown in Fig. 1c and d, the neurite length of GFP+ cells was significantly potentiated compared with the bFGF-treated vector group. Moreover, these phenomena could be completely reproduced in E14 rat cortical NPCs (Fig. S1). These results suggest that Bcl-2 might be directly involved in neuronal differentiation.

Fig. 1
figure 1

Effect of Bcl-2 gene modulation during neuronal differentiation in H19-7 cells. a H19-7 cells were transfected with control or Bcl-2 siRNA for 2 days, after which the medium was changed to N2 medium supplemented with 20 ng/mL bFGF and the temperature shifted to 39 °C for 2 days. The cells were immunostained with TUJ1. Scale bar, 100 µm. b Immunoblotting was performed to verify the knockdown efficiency, and the neurite lengths of the TUJ1+ cells were measured from 10 randomly selected areas from at least three independent experiments and presented as the mean ± SD. ***p < 0.001. c MSCV-IRES-EGFP or rBcl-2-MSCV-IRES-EGFP vectors were introduced into the cells through retroviral transduction for 2 days, followed by stimulation with bFGF and shifting the temperature to 39 °C for 2 days. d Cells were then immunostained with TUJ1 and GFP. Scale bar, 100 µm. The results are presented as the mean ± SD of three independent experiments. **p < 0.01

Full size image

Bcl-2 Overexpression Increases Expression of Various Neurogenic Factors in NPCs

To investigate whether Bcl-2 overexpression regulates the expression of neurogenic factors, we examined the gene expression patterns of NPCs derived from the cerebral cortex of E14 rat embryos by using microarrays. We found that the expression of 4762 genes was upregulated by Bcl-2 overexpression, and of these, 3537 genes overlapped with the genes upregulated by the differentiation stimuli (Fig. 2a). On the other hand, out of the 6116 genes that were decreased by Bcl-2 overexpression, 4989 of them overlapped with the genes downregulated by differentiation stimulation (Fig. S2). Additionally, 706 genes were identified as genes involved in the nervous system development through GO term analysis (Fig. 2b). Among them, Bcl-2 overexpression upregulated 372 genes and downregulated 334 genes (Table S1). As in the differentiated control group (− bFGF), a variety of neurogenic marker genes, such as Map2, Mapt, Synj1, and Snap25, were increased by Bcl-2 overexpression (Fig. 2b). Moreover, neurogenic bHLH transcription factors, including NeuroD2 and Ascl1 (Mash1), as well as Bmp4 and Tbx3 were increased by Bcl-2 overexpression (Fig. 2b). Next, we verified whether the microarray results could be reproduced in H19-7 cells. As shown in Fig. 3a and b, Bcl-2 overexpression induced a significant increase in neurite outgrowth. Moreover, the genes identified in the microarrays (Bmp4, Tbx3, NeuoD1, NeuroD2, and Mash1) showed significantly increased mRNA levels (Fig. 3c, d). The expression of the Bmp4, Tbx3, and NeuroD1 proteins also showed a similar tendency (Fig. 3e, f). In summary, Bcl-2 overexpression upregulates the expression of various neurogenic factors, such as Bmp4, Tbx3, and NeuroD1, and increases neurite outgrowth.

Fig. 2
figure 2

Microarray analysis of the genes upregulated by Bcl-2 overexpression in rat NPCs. a Whole-genome expression profiles were analyzed using cDNA microarrays of rat NPCs. The number of genes upregulated 1.2-fold by differentiation stimuli or by Bcl-2 overexpression is shown in the Venn diagram. b Upregulated genes were categorized using Gene Ontology (GO) analysis based on the biological processes. The genes corresponding to the nervous system development are expressed in the heat map (right panel). The columns represent the individual groups, and the rows represent each upregulated gene. The relative expression value for each gene is depicted by Z-score, with red indicating high expression and blue indicating low expression

Full size image
Fig. 3
figure 3

Bcl-2 overexpression induces neurite outgrowth as well as expression of several genes. a H19-7 cells were infected with MSCV-IRES-EGFP or rBcl-2-MSCV-IRES-EGFP using retroviral transduction for 2 days. Immunofluorescence images for TUJ1 and GFP were used to determine neurite outgrowth. Scale bar, 50 µm. b The neurite length of TUJ1+ cells was measured from ten randomly selected areas from at least three independent experiments and presented as the mean ± SD. ***p < 0.001. c, d Total RNA was isolated using Trizol reagent, and mRNA levels of Bcl-2, Bmp4, Tbx3, NeuroD1, NeuroD2, and Mash1 were determined using RT-PCR (c) and real-time RT-PCR (d). The results are presented as the mean ± SD of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001. e, f Cells were lysed and analyzed using immunoblotting with anti-Bcl-2, anti-Bmp4, anti-Tbx3, anti-NeuroD1, and anti-Calnexin antibodies. The band intensity was quantified with ImageJ software. The results are presented as the mean density (as a fold increase) ± SD from three independent experiments. *p < 0.05, **p < 0.01

Full size image

Bmp4 is Involved in the Process of Neurite Outgrowth Induced by Bcl-2 Overexpression

Bone morphogenetic proteins (BMPs) are extracellular secretory proteins belonging to the transforming growth factors-β (TGF-β) superfamily and mediate developmental processes, including morphogenesis, proliferation, lineage commitment, differentiation, and apoptosis in the nervous system (Mehler et al. 1997). In particular, Bmp4 is highly expressed in the embryonic cortex of the developing dorsal forebrain (Furuta et al. 1997) and triggers neuronal differentiation of the NPCs in the neocortical ventricular zone (Li et al. 1998; Moon et al. 2009). However, many reports suggest that BMPs and their signaling pathways are involved in the commitment of the astroglial fate during neuronal development in a subclass-specific manner (Mehler et al. 1997; Gross et al. 1996). To investigate the relationship of Bcl-2 overexpression-induced Bmp4 with other genes, we transfected 100 nM of control or Bmp4 siRNA into Bcl-2-overexpressing H19-7 cells. Bcl-2-induced neurite elaboration was significantly repressed by Bmp4 depletion (Fig. 4a, b). Additionally, the Bcl-2-induced expression of Tbx3, NeuroD1, NeuroD2, and Mash1 was strikingly downregulated by Bmp4 knockdown at both the mRNA (Fig. 4c, d) and protein levels (Fig. 4e, f). These results indicate that Bmp4 participates in Bcl-2-mediated neurite outgrowth by controlling the expression of Tbx3 and neurogenic bHLH transcription factors.

Fig. 4
figure 4

Effect of Bmp4 gene depletion in Bcl-2-overexpressing H19-7 cells. a Cells were infected with MSCV-IRES-EGFP or rBcl-2-MSCV-IRES-EGFP using retroviral transduction for 2 days followed by transfection with control or Bmp4 siRNA. The cells were then immunostained for TUJ1 and GFP. Scale bar, 50 µm. b Neurite length was measured. The results are presented as the mean ± SD of three independent experiments. ***p < 0.001. c, d Total RNA was isolated using Trizol reagent, and mRNA levels of Bcl-2, Bmp4, Tbx3, NeuroD1, NeuroD2, and Mash1 were determined using RT-PCR (c) and real-time RT-PCR (d). The results are presented as the mean ± SD of three independent experiments. *p < 0.05, **p < 0.01. e, f Cells were lysed, and immunoblotting was performed using anti-Bcl-2, anti-Bmp4, anti-Tbx3, anti-NeuroD1, and anti-Calnexin antibodies. The results are presented as the mean density (as a fold increase) ± SD from three independent experiments. ***p < 0.001

Full size image

The Upregulation of Bmp4 Induced by Bcl-2 Overexpression is Independent of Canonical Bmp/Smad Pathway

BMPs are secreted extracellularly and bind to their specific receptor leading to the phosphorylation of the Smad1/5/8 protein complex and translocation into the nucleus. The Smad1/5/8 protein complex then acts as a transcription factor and induces the expression of several target genes (Wrana 2013). One of those target genes, Id, represses neuronal differentiation of the NPCs and induces astroglial lineage commitment by preventing DNA binding to neurogenic bHLH transcription factors or by promoting their degradation through ubiquitin-mediated proteolysis (Miyazono and Miyazawa 2002). In addition to these canonical pathways, BMPs also can activate LIMK or the small GTPase Cdc42 instead of phosphorylating Smad. These Smad-independent pathways stimulate cell migration and axon growth (Wrana 2013; Yam and Charron 2013). To examine whether Bcl-2-mediated neurite outgrowth is dependent on Bmp/Smad signaling, we used anti-Bmp4 antibody to neutralize the secreted Bmp4. As shown in Fig. S3, Bcl-2-induced upregulation of Tbx3, NeuroD1, NeuroD2, and Mash1 was reduced by Bmp4 neutralization in a dose-dependent manner. However, neither Bcl-2 overexpression nor Bmp4 neutralization affected the expression of both Id1 and Id2 in H19-7 cells. These results indicate that while Bmp4 induced by Bcl-2 overexpression can act in an autocrine or paracrine manner after being secreted into the extracellular space, the Smad-dependent canonical pathway is not associated with Bcl-2-mediated neurite formation.

Tbx3 is Also Associated with the Process of Neurite Outgrowth Induced by Bcl-2 Overexpression

The T-box family of transcription factors, which are characterized by a common DNA binding motif termed as the T-box, are involved in embryonic developmental processes by regulating cell differentiation and migration (Naiche et al. 2005; Papaioannou 2014). In particular, several subfamilies of T-box genes are essential for brain development and their mutations are associated with neurodevelopmental disorders such as autism spectrum disorders (Mihalas and Hevner 2017). Among them, Tbx3 is involved in various signaling pathways related to the pluripotency network and has a critical role in the maintenance of the embryonic stem cells during early cell fate specification (Russell et al. 2015). In order to investigate how Tbx3 impacts other genes during Bcl-2-mediated neurite outgrowth, we transfected 100 nM of control or Tbx3 siRNA into Bcl-2-overexpressing H19-7 cells. We observed that Bcl-2-induced neurite outgrowth was significantly repressed by Tbx3 depletion (Fig. 5a, b). Moreover, NeuroD1, NeuroD2, and Mash1 that were induced by Bcl-2 overexpression were downregulated by Tbx3 knockdown, while there was no significant change in the expression of Bcl-2 and Bmp4 (Fig. 5c, d). These alterations also coincided with changes in protein levels (Fig. 5e, f). We demonstrated that Tbx3 could mediate the neurite outgrowth by controlling the expression of neurogenic bHLH genes downstream of Bmp4 during Bcl-2-mediated neuronal differentiation.

Fig. 5
figure 5

Effect of Tbx3 gene depletion in Bcl-2-overexpressing H19-7 cells. a Cells were infected with MSCV-IRES-EGFP or rBcl-2-MSCV-IRES-EGFP via retroviral transduction for 2 days followed by transfection with control or Tbx3 siRNA. The cells were then immunostained for TUJ1 and GFP. Scale bar, 50 µm. b Neurite length was measured. The results are presented as the mean ± SD of three independent experiments. ***p < 0.001. c, d Total RNA was isolated using Trizol reagent, and mRNA levels of Bcl-2, Bmp4, Tbx3, NeuroD1, NeuroD2, and Mash1 were determined using RT-PCR (c) and real-time RT-PCR (d). The results are presented as the mean ± SD of three independent experiments. *p < 0.05, **p < 0.01. e, f Cells were lysed, and immunoblotting was performed using anti-Bcl-2, anti-Bmp4, anti-Tbx3, anti-NeuroD1, and anti-Calnexin antibodies. The results are presented as the mean density (as a fold increase) ± SD from three independent experiments. *p < 0.05

Full size image

NeuroD1 is a Key Downstream Regulator of Bcl-2-Mediated Neurite Outgrowth in H19-7 Cells

bHLH proteins possess an HLH domain that enables dimerization and a basic domain for DNA binding (Dennis et al. 2019). bHLH transcription factors are hetero-dimerized with E proteins, which in turn bind to the E-box DNA elements to promote the expression of neurogenic target genes, and their transcriptional activities are inhibited by interaction with the Hes or Id families (Ross et al. 2003). As mentioned previously, we established that Bcl-2 regulated neurite outgrowth by increasing the levels of the proneural bHLH transcription factors, including NeuroD1, NeuroD2, and Mash1, without changing the expression of Id. NeuroD1 is highly accumulated in late development stage of the nervous system and is an essential participant in terminal differentiation, neuronal maturation, and survival (Gao et al. 2009). NeuroD1 also triggers neuronal specification by reprogramming the transcription factor and chromatin landscapes at regulatory elements of neurogenic target genes (Pataskar et al. 2016). To verify the function of NeuroD1, we transfected 100 nM of NeuroD1 siRNA into Bcl-2-overexpressing H19-7 cells. Bcl-2-induced neurite elaboration was significantly suppressed by NeuroD1 knockdown (Fig. 6a, b). However, NeuroD1 depletion did not influence the expression of other genes (Bcl-2, Bmp4, Tbx3, NeuroD2, and Mash1) at the mRNA level (Fig. 6c, d). This indicates that NeuroD1 functions independently from other bHLH genes during Bcl-2-mediated neurite outgrowth. Moreover, we found that the protein expression levels of Bcl-2, Bmp4, and Tbx3 did not show any significant changes by the NeuroD1 knockdown (Fig. 6e, f). In conclusion, we demonstrated that Bcl-2 overexpression promotes neurite formation via the Bmp4/Tbx3/NeuroD1 pathway in H19-7 cells.

Fig. 6
figure 6

Effect of NeuroD1 gene depletion in Bcl-2-overexpressing H19-7 cells. a Cells were infected with MSCV-IRES-EGFP or rBcl-2-MSCV-IRES-EGFP through retroviral transduction for 2 days followed by transfection with control or NeuroD1 siRNA. The cells were then immunostained for TUJ1 and GFP. Scale bar, 50 µm. b Neurite length was measured. The results are presented as the mean ± SD of three independent experiments. **p < 0.01. c, d Total RNA was isolated using Trizol reagent, and mRNA levels of Bcl-2, Bmp4, Tbx3, NeuroD1, NeuroD2, and Mash1 were determined using RT-PCR (c) and real-time RT-PCR (d). The results are presented as the mean ± SD of three independent experiments. *p < 0.05. e, f Cells were lysed, and immunoblotting was performed using anti-Bcl-2, anti-Bmp4, anti-Tbx3, anti-NeuroD1, and anti-Calnexin antibodies. The results are presented as the mean density (as a fold increase) ± SD from three independent experiments. *p < 0.05

Full size image

Discussion

Multipotent neural stem cells can give rise to all types of neural cells through asymmetric cell division, including neurons, astrocytes, and oligodendrocytes in the mammalian nervous system (Gage 2000; Paridaen and Huttner 2014). They are spatiotemporally controlled by intrinsic or extrinsic factors found in the neural stem cell niche (Paridaen and Huttner 2014). In particular, bFGF is a necessary stem cell mitogen required for proliferation and neurogenesis during the development of the nervous system (Raballo et al. 2000). It interacts with receptor tyrosine kinase FGFRs (FGF receptors) on the cell surface and triggers various cellular processes by activating downstream molecules including Ras, phospholipase C, and phosphoinositide 3-kinase via tyrosine phosphorylation of the adaptor protein FRS2 (FGF receptor substrate 2) (Schlessinger 2004). We previously reported that PLD1 plays an important role in bFGF signaling involved in neuronal differentiation in rat hippocampal progenitor cells. PLD1 is activated by Src and Ras during bFGF-induced neurite outgrowth and regulates neurotrophin-3 expression through the RhoA/ROCK/JNK/Elk1 pathway in H19-7 cells (Oh et al. 2007; Choi et al. 2012). Furthermore, PLD1 upregulates Bcl-2 expression through the PLA2/Cox2/EP4/PKA/p38MAPK/CREB cascade in rat NPCs and through the JNK/STAT3 pathway in H19-7 cells during neuronal differentiation (Park et al. 2014; Yoon et al. 2012). This indicates that PLD1-mediated Bcl-2 expression is required in bFGF signaling for neuronal fate determination of NPCs. However, the mechanism by which Bcl-2 influences neuronal differentiation is not fully understood.

Bcl-2 is generally regarded as an anti-apoptotic protein that prevents the cytosolic release of apoptogenic factors, including cytochrome c, in the mitochondrial outer membrane (Adams and Cory 1998; Pan et al. 2014, 2015). Many previous reports have shown the neuroprotective effects of Bcl-2 against environmental insults such as ischemia (Martinou et al. 1994). Moreover, the interaction of Bcl-2 with calcineurin can inhibit Ca2+ release by suppressing IP3R (inositol 1,4,5-triphosphate receptor) through the phosphatase cascade of DARPP-32 (dopamine and cAMP-regulated phosphoprotein-32)/PP1α (protein phosphatase 1 alpha) on the endoplasmic reticulum (ER) (Tang et al. 2003; Erin et al. 2003). These mechanisms may also be involved in neuropsychiatric disorders. Bcl-2 rs956572 polymorphism has been found to be associated with dysfunction of intracellular Ca2+ homeostasis and the cortical glutamatergic system in the pathophysiology of type I bipolar disorder (Uemura et al. 2011; Soeiro-de-Souza et al. 2013). Moreover, mood-stabilizing agents, such as lithium and valproate, increase Bcl-2 expression in the rat CNS (Chen et al. 1999). Bcl-2 family proteins are involved in the development of resistance against therapy-induced apoptosis during the treatment of several types of brain cancers such as malignant glioblastoma (Strik et al. 1999; Fels et al. 2000). Taken together, this strongly indicates the possibility that Bcl-2 can be a potential therapeutic target. Recently, the use of the BH3-mimetic drug Venetoclax (ABT-199) is on the rise as a Bcl-2 inhibitor in various fields including clinical lymphoma therapy (King et al. 2017). However, it still needs to be confirmed whether this drug can also be used to treat brain cancer. In addition, studies using various cancer models have suggested that Bcl-2 expression is intimately regulated by PLD1, indicating that targeting PLD1 might have robust anticancer effects via the inhibition of Bcl-2 function (Cho et al. 2008). The small-molecule PLD1 inhibitor FIPI shows remarkable preclinical efficacy for the treatment of colorectal cancer (Kang et al. 2015), and the peptide drug TVTSP, a PLD1-specific inhibitor, was recently verified as an efficient therapeutic agent for several types of cancer and other inflammatory conditions (Cho et al. 2011; Lee et al. 2018). Therefore, detailed research needs to be done in the future to evaluate if these drugs can also be considered as potential therapeutics for several brain disorders.

Apart from its neuroprotective function, Bcl-2 mediates neurogenesis for the developing CNS and the regenerating adult nervous system (Michaelidis et al. 1996; Zhang et al. 1996; Hilton et al. 1997; Chen et al. 1997). However, it is unclear whether Bcl-2 exerts its neuroprotective effect by merely providing sufficient time for neurogenesis of NPCs by protecting them from cell death. Loss of Bcl-2 function represses neurite elaboration in response to differentiation stimuli, and inversely, gain of Bcl-2 function enhanced neurite outgrowth (Fig. 1), indicating that Bcl-2 is strongly involved in the neuronal differentiation process. Next, we sought to determine how Bcl-2 impacts neurite outgrowth. According to previous reports, Bcl-2 overexpression increases synaptosomal-associated protein 25 (SNAP-25) in a dopaminergic neuronal cell line (Oh et al. 1996). Moreover, Suzuki and Tsutomi (1998) suggested that Bcl-2 controls neuronal differentiation by accelerating polymerization of neurofilament chains, and Chang et al. (2007) reported that Bcl-XL overexpression induces proneural bHLH genes (Ngn1, NeuroD1, and Mash1) and suppresses anti-neuronal bHLH genes (Hes1/5 and Id1/2/3) without changing the expression of the Notch signaling genes. Therefore, we analyzed the whole-genome expression profiles of rat NPCs to compare the genes that were altered by differentiation stimuli and the genes affected by Bcl-2 overexpression. Similar to the differentiated control group, Bcl-2 overexpression increased many neurogenic marker genes (Fig. 2b). In particular, proneural bHLH genes, including NeuroD2, Neurog2, and Mash1, were upregulated. When Bcl-2 was overexpressed in H19-7 cells, neurite length was significantly increased, and NeuroD1, NeuroD2, and Mash1 were also upregulated (Fig. 3). However, the expression of Neurog1 and Neurog2 was not affected (data not shown). These results imply that Bcl-2 overexpression has a direct impact on the neuronal differentiation through upregulation of neurogenic bHLH genes. Additionally, we found that Bcl-2 overexpression significantly increased Bmp4 and Tbx3 expression, indicating that Bmp4, Tbx3, and several bHLH transcription factors are included in Bcl-2-mediated neurite outgrowth.

To investigate the hierarchy of these molecules, we selectively depleted Bmp4, Tbx3, or NeuroD1 in Bcl-2-overexpressing H19-7 cells. When Bmp4 was depleted, Bcl-2-mediated neurite formation was significantly repressed, and Tbx3 and NeuroD1 were downregulated (Fig. 4), indicating that Bmp4 is required for Bcl-2-mediated neurite outgrowth by regulating Tbx3 and NeuroD1. Bmp4 induces neuronal differentiation in the neocortical ventricular zone (Li et al. 1998; Moon et al. 2009). However, many reports suggest that BMPs might induce astroglial fate specification (Mehler et al. 1997; Gross et al. 1996). It is thought that this discrepancy occurs due to the dependency of Bmp4 on Smad. Extracellularly secreted Bmp4 activates Smad1/5/8 protein by binding to BMP receptors on the cell surface, which facilitates Smad1/5/8 to act as a transcriptional regulator and induce the expression of the target genes including Id (Wrana 2013; Miyazono and Miyazawa 2002). Conversely, Bmp4 can induce axon growth by activating LIMK or the small GTPase Cdc42 without Smad phosphorylation (Wrana 2013; Yam and Charron 2013). In this study, Bmp4 neutralization downregulated its downstream genes (Tbx3, NeuroD1, NeuroD2, and Mash1) without affecting the expression of Id1 and Id2 (Fig. S3). In addition, Smad1 was not phosphorylated by Bcl-2 overexpression (data not shown). These results indicate that although the Smad-dependent canonical pathway might not be associated with Bcl-2-mediated neurite formation, Bmp4 functions in an autocrine or paracrine manner after being secreted into the extracellular space. Bcl-2 also increases stroke-induced striatal neurogenesis in adult brains by modulating Bmp4 function via activation of β-catenin signaling (Lei et al. 2012). The expression of Bmp4 induced by Bcl-2 overexpression has been thought to be mediated by Wnt/β-catenin signaling. We also found that Tbx3 acts downstream of Bmp4 and controls neurite outgrowth by increasing the expression of the neurogenic bHLH genes during Bcl-2-mediated neuronal differentiation (Fig. 5). Bmp4 has been shown to induce Tbx3 expression in the pathophysiology of glomerular diseases (Wensing and Campos 2014), and β-catenin reinforces the pluripotency by activating Oct-4 via a TCF-independent mechanism, which induces the expression of Tbx3 as a Oct-4 target gene (Kelly et al. 2011). Moreover, Tbx3 mediates TGF-β1-induced anti-proliferative and pro-migratory effects in breast epithelial cells by repressing Tbx2 transcription, which allows for the derepression of p21 and a G1 cell cycle arrest (Li et al. 2013). Finally, while Bcl-2-mediated neurite outgrowth was significantly suppressed by NeuroD1 knockdown, the expression of other genes was not altered (Fig. 6). This indicates that NeuroD1 functions independently of other bHLH genes during Bcl-2-mediated neurite outgrowth and that Bcl-2 overexpression promotes neurite elaboration via the Bmp4/Tbx3/NeuroD1 pathway. Taken together, this study shows that Bcl-2 may have a direct role in neuronal differentiation in addition to its well-known anti-apoptotic function in response to environmental insults.

Conclusions

We demonstrated that Bcl-2 overexpression induces neurite outgrowth through the Bmp4/Tbx3/NeuroD1 cascade, thereby contributing to understanding of the non-apoptotic function of Bcl-2 during neuronal differentiation.