POU3F4 up-regulates Gli1 expression and promotes neuronal differentiation and synaptic development of hippocampal neural stem cells

Abstract

Background

Neural stem cells (NSCs) are considered to be the most promising cell type for cell replacement therapy in neurodegenerative diseases. However, their low neuronal differentiation ratio impedes their application in such conditions. Elucidating the molecular mechanism of NSC differentiation may provide the necessary experimental basis for expanding their application. Previous studies have indicated that POU3F4 can induce neuronal differentiation of NSCs, this study aims to underly the possible exact mechanism of POU3F4 on the NSC differentiation and development.

Methods

NSCs were isolated and cultured from the hippocampus of neonatal mice. The frozen hippocampal sections were prepared for immunohistochemical staining. Synaptic development was assessed using electron microscopy. High-throughput sequencing was employed to analyze the gene expression profile following the overexpression of Brn4. Gene expression levels were determined through Western blotting and qRT-PCR. Cell cycle and differentiation were evaluated using flow cytometry and immunofluorescent staining.

Results

It was found that POU3F4 promoted the neuronal differentiation of hippocampal NSCs and synapse development, and inhibited NSC proliferation. POU3F4-deficient mice exhibited impairments in learning and memory. RNA sequencing and ChIP assays confirmed that Gli1 was downstream of POU3F4. Loss and gain function experiments indicated that Gli1 mediated POU3F4 promoting neuronal differentiation and synapse development. Forced expression of Gli1 in hippocampus improved learning and memory function of animal models.

Conclusions

The results suggest that POU3F4 and Gli1 promote neuronal differentiation and synaptic development of NSCs, and that Gli1 partially mediates the effects of POU3F4.

Graphical abstract

Introduction

Alzheimer’s disease (AD) is a degenerative disease of the central nervous system that is characterized by memory impairment, aphasia, agnosia, spatial skill impairment, and personality and behavior changes. According to a report from the World Health Organization on AD in 2012, the number of patients with AD is expected to reach 76 million by 2030, and AD has become one of the major diseases affecting human health [1]. The onset of AD is hidden and the pathogenesis is unknown, which has so far led to a lack of effective prevention and treatment measures. Studies have shown that in the early stages of AD a large number of neurofibrillary tangles appear in the hippocampal formation. These are known to affect learning and memory functions in mammalian brains, eventually leading to hippocampal neuronal degeneration [2]. There are neural stem cells (NSCs) in the subgranular zone (SGZ) of the dentate gyrus (DG) in the mammalian hippocampus. After migration and differentiation, they eventually mature to granular cells in the DG. NSCs are the most direct source of neurons. Increasing evidence shows that cell replacement therapy using NSCs is an effective method for the treatment of nerve injury and has attracted extensive attention from researchers [3, 4]. Therefore, repairing or reconstructing the neural network of brain degeneration in patients with AD by promoting the differentiation of endogenous hippocampal NSCs or NSCs transplanted in vitro into specific neurons brings new hope for its prevention and treatment.

Previous studies have shown that under natural conditions most NSCs differentiate into glial cells, with a low proportion of neuronal differentiation. Even under some intervention conditions, such as the application of glial-derived neurotrophic factor (GDNF) [5], Nng2, or Mash1 [6, 7], the proportion of neurons that differentiate from NSCs improves, but with an insufficient number of neurons to meet the needs of treatment. During the pathogenesis of AD, the formation of harmful factors is not conducive to neuronal differentiation of NSCs in the hippocampus [8]. Therefore, it is particularly important to explore the factors and molecular mechanisms that affect the differentiation of NSCs so as to guide NSCs to differentiate into a sufficient number of neurons.

The differentiation of NSCs in adult mammals is a complex and complete biological process that is regulated by a network of transcription factors and signaling pathways [9, 10]. The transcription factor POU3F4 (also known as Brain4, Brn4, etc.) belongs to the third family of POU proteins (POU III) together with Brn1 and Brn2. Studies have shown that POU3F2, POU3F3, and POU3F4 appear in the nervous system in the early embryo and are widely distributed [11,12,13], suggesting that these factors may participate in the development of the early nervous system. Recent reports have shown that POU3F4 plays an important role in the reprogramming of somatic cells into neural cells. Kim et al. [14,15,16] used POU3F4 in combination with classical reprogramming factors Sox2, KLF4, and Myc to directly reprogram fibroblasts into neural precursor cells. Potts et al. overexpressed POU3F4 in astrocytes that then transdifferentiated into neurons [17]. Shimazaki et al. found that POU3F4 promoted the differentiation of striatal NSCs into neurons [18]. In our previous studies, we found that POU3F4 promoted neuronal differentiation of NSCs from the hippocampus in vitro [19]. These results suggest that POU3F4 plays an important role in the differentiation of NSCs, but the specific mechanism of action is still unknown.

The Gli family has three members, Gli1, Gli2, and Gli3, all of which have five conserved zinc finger structures [20]. Studies have shown that among the three members, Gli1 is the only factor directly initiated by Shh signal and is an important member of the Shh pathway. Therefore, a change in Gli1 expression can directly indicate whether the Shh signaling pathway is activated [21, 22]. Shh signaling is crucial for neurogenesis in the embryonic or adult brain [23]. It has been reported that in the neonatal rat model of ischemia and hypoxia, transplantation of exogenous umbilical cord blood monocytes can promote the differentiation of endogenous NSCs into neurons and reduce the rate of their differentiation into glial cells. During this process, the expression of Shh and Gli1 is significantly increased [24]. These studies suggest that Shh and its downstream factor Gli1 may be involved in neurogenesis. The relationship between POU3F4 and Gli1 and their roles in regulating NSC differentiation, proliferation, and development is still unclear.

In this study, we elucidated the regulatory effect of POU3F4 on Gli1 and verified the role of POU3F4 and Gli1 in the proliferation and differentiation of hippocampal NSCs into neurons, as well the development of synapses. Our results provide new insights into hippocampal neurogenesis and an experimental basis for the application of NSCs in cell transplantation therapy.

Materials and methods

Animals and surgery

Neonatal and adult Institute of Cancer Research (ICR) mice were all purchased from the experimental animal center of Nantong University. POU3F4 knockout (POU3F4^−/−) mice were gifts from E. Bryan Crenshaw III from the Children’s Hospital of Philadelphia. It has been described in the report [25]. All animals were kept in a controlled environment on a 12 h:12 h light/dark cycle in cages that provided free access to food and water. All animal experiments were performed in accordance with the National Institutes of Health Laboratory Animal Care and Use Guidelines. The research protocol was approved by the Ethics Committee of Nantong University Affiliated Hospital. All efforts were made to minimize the number and suffering of the animals used in this study. The animals were finally euthanized using CO₂ inhalation. Briefly, place the animals in a chamber with up to 70% carbon dioxide, ensuring the gas flow replaces at least 20% of the chamber volume per minute. Maintain this level for at least 3 min. The work has been reported in line with the ARRIVE guidelines 2.0.

The adult male ICR mice were anesthetized with enflurane, oxygen, and nitrous oxide (1:33:66). They were then placed in a stereotaxic device, and a wire knife was used to make a transverse cut in the dorsal layer of the hippocampal CA1 at specific bregma coordinates: AP = 0.7, ML = 0.8 and 2.0 on the right; AP = 0.7, ML = − 0.8 and − 2.0 on the left, with a depth of 3.0 mm. Fourteen days post-FF transection, 5 µL of lentivirus was microinjected into the hippocampal DG at two sites (AP = 1.9, ML = 1.0 and AP = 2.3, ML = 1.5, with depth of 2.0 mm) at 0.5 μL/min. The needle was held in place for 10 min before being slowly withdrawn.

NSC culture and identification

The cerebral hemispheres were carefully removed from postnatal day 1 ICR or POU3F4^−/− and the litter wild type (WT) mice under aseptic conditions. The hippocampi were separated and mechanically dissociated in Dulbecco’s Modified Eagle’s Medium and Ham F12 (DMEM/F12, Gibco, Grand Island, NY, USA). The cell suspension was then passed through a 200-mesh nylon filter and single cells were cultured in 25 cm² flasks at a density of 1 × 10⁴ cells/ml, with NSC expansion medium composed of DMEM/F12, 2% B27 (Gibco), 20 ng/ml epidermal growth factor (EGF, Gibco), 20 ng/ml basic fibroblast growth factor (bFGF or FGF-2, Sigma-Aldrich, St. Louis, MO, USA), and 100 U/ml penicillin/streptomycin (Beyotime Institute of Biotechnology, Shanghai, China) at 37 °C in a humidified incubator of 95% air and 5% (v/v) CO₂. About 5‒7 days later, neurospheres were formed and suspended in the medium. Accutase enzyme (Sigma) was used to digest spheres into single cells and these cells were cultured in the flasks continuously. After three generations, some single cells were planted in 24- or 48-well plates and cultured with NSC expansion medium to form spheres. Other single cells were seeded in 48-well plates that were precoated with poly-L-lysine (Sigma) at a density of 1 × 10⁴ cells/well and cultured with the differentiation medium composed of DMEM/F12 and 2% fetal bovine serum (FBS, Gibco). Seven days later, immunocytochemistry was used to detect the expression of some stem cell markers and various neural cell markers, to determine the purity, proliferation and differentiation.

Behavior tests

For the Morris water maze (MWM), a total of 13 POU3F4 knockout mice and 13 wild type littermates from same colony were tested at 1, 2 and 5-month age. Because the POU3F4 is X-linked allele, the animals were randomly selected among hemizygous males and homozygous females. In the learning experiment during the first four days, the platform was put in the first quadrant of the maze, and the mice were placed into the water from fixed positions in the four quadrants. Then, the time (in seconds) for the mice to find the underwater platform was recorded. If the time exceeded 90 s, the mice were guided to the platform and left on the platform for 10 s. Each mouse was trained 4 times a day with an interval of 1 h. On the fifth day, the platform was removed and the 90 s exploration training began. The mice were placed in the third quadrant, and the number of times the mice passed through the target area was recorded by a camera connected to a computer.

For the passive avoidance task, a total of 8 POU3F4 knockout mice and 8 wild type littermates were tested at 2-month age. The mice were allowed a 10-min acclimatization period with free access to either the light or dark compartment of the avoidance training box. animals were placed into the light compartment, and the guillotine door was opened after 30 s. Upon entering the dark compartment, the door was closed. The second habituation trial was carried out 30 min later. For the training trial, a foot shock (1 mA, 50 Hz, 2 s) was delivered to the grid floor of the dark part. After 20 s, the mouse was removed from dark compartment placed into the home cage. The test trial was performed 24 h after training. Each animal was placed in the light compartment again and 30 s later the door was raised. Then, the latency of entering the dark compartment was recorded (as step-through latency). This procedure ended when the animal entered the dark compartment or when it remained in the light compartment for 300 s. In these sessions, no electric shock was given to the animals.

For the active avoidance task, the above animals were placed in the left compartment, and facing the end wall. After 20 s, a trial started with conditioned stimulus (CS, light of 7 W and sound of 2400 Hz at 40 dB, presented simultaneously) onset and, 5 s later, followed by a 1.0 mA intensity foot shock (unconditioned stimulus, US). Running to the opposite compartment within the CS-US interval, the experiment was immediately terminated and recorded as avoidance response. After the US onset, the experiment was scored as an escape response.

BrdU incorporation assay

BrdU Staining Kit from Abcam (USA) was used for incorporation assay according to the manufacture’s protocol. In brief, mice were intraperitoneally injected with BrdU at a dosage of 50 mg/kg body weight daily for 5 consecutive days. Subsequently, brain tissue samples were fixed and permeabilized using 0.01% Triton X-100, followed by blocking in 2% bovine serum albumin (BSA). The samples were then stained overnight at 4 °C using a monoclonal anti-BrdU antibody and incubated with Alexa Fluor568-conjugated goat anti-mouse IgG for 2 h. Finally, the samples were counterstained with Hoechst 33,342 (1:1000; Sigma) and observed under a fluorescence microscope (Zeiss, Germany).

Preparation of brain sections and immunohistochemical staining

Mice were anesthetized using intraperitoneal injection of compound anesthetics (100 ml in volume containing chloral hydrate 4.25 g, pentobarbital sodium 0.886 g, magnesium sulfate 2.12 g, propylene glycol 33.8 ml, absolute ethanol 14.25 ml) at a dose of 0.4 ml/100 g body weight and then underwent surgery to expose the heart. The infusion needle was quickly inserted into the left ventricle and then the right atrial appendage was cut open. About 50 ml of 0.9% NaCl was rapidly injected into the left ventricle. Subsequently, 50 ml of 4% paraformaldehyde (PFA) was perfused per mouse. After perfusion, the mice were killed, and the brains were removed and placed in 4% PFA for 24 h. Then, to dehydrate them, the fixed brains were immersed in 20% and 30% sucrose solution successively. After precipitation, the brains were prepared for cryosectioning. Coronal Sects. (15 μm) of the hippocampus were prepared using a Leica cryostat (Leica CM1900, Germany). A set of about 8 brain slices were subjected to the immunofluorescent staining. Briefly, the sections were blocked in 10% goat serum in PBST (0.01 M sodium phosphate buffer, pH 7.4, 0.05% Tween-20) for 1 h at room temperature (RT) and incubated with guinea pig anti-doublecortin (DCX) antibody (1:1000, Millipore, USA) overnight at 4 °C, followed by incubation with Alexa Fluor 488-conjugated goat anti-guinea pig IgG (Invitrogen, USA). Immunofluorescence signals were visualized under a fluorescence microscope (Leica DMIRB, Germany). The total number of DCX-positive neurons with typical neuronal morphology in the hippocampal DG of the 8 slices was counted for statistics analysis. Some positive dot signals without typical neuronal appearance were not included in the statistics.

Construction of recombinant lentiviruses

The POU3F4-overexpressing lentivirus (LV-POU3F4; NM_000307-HA) was constructed based on the lentiviral vector (pLenti-EF1a-EGFP-F2A-Puro-CMV-POU3F4-HA). The Gli1-overexpressing lentivirus (LV-Gli1; NM_005269) was constructed based on the lentiviral vector H4276 (pLenti-CMV-Gli1-3FLAG). A lentivirus that interferes with the expression of POU3F4 (POU3F4-shRNA, TargetSeq: GCAGCCACACGAGGTTTAT) was constructed on the vector pLKD-CMV-mcherry-2A-Puro-U6-shRNA. The lentivirus that interferes with the expression of Gli1 (Gli1-shRNA, TargetSeq: GCCCTGTGTTCCACATGAT) was constructed on the vector pLKD-CMV-G&PR-U6-shRNA. Among the four viruses used, LV-POU3F4 and the corresponding control lentiviruses (LV-Ctrl) was generated by Genechemical Biotechnology Co. (Shanghai, CHN), and the other three and corresponding LV-Ctrl were all generated by Heyuan Biotechnology Co. (Shanghai, CHN).

Differentiation of NSCs after infection with lentiviruses

Single-cell NSCs (passage 3) were seeded in a 48-well plate at a density of 1 × 10⁴ cells/well and were transfected with lentiviruses [virus concentrations: LV-POU3F4: 1 × 10⁵ transduction units (TU); LV-Gli1, POU3F4-shRNA and Gli1-shRNA: 2 × 10⁵ TU] for 12 h. Then, the medium containing the viruses was removed and cells were continuously cultured with the differentiation medium (DMEM/F-12 medium supplemented with 2% B27 and 2% FBS). After 7 d, cells were examined using immunocytochemical analysis.

Immunocytochemical analysis

Cells were fixed with 4% paraformaldehyde at room temperature for 20 min. After washing with 0.01 M PBS (Sangon, Shanghai, CHN), cells were incubated overnight at 4 °C with primary antibodies, followed by incubation with secondary antibodies at room temperature for 2 h. The nuclei were stained with Hoechst 33,342 (1:1000, Pierce, Rockford, IL, USA) for 15 min. Immunofluorescence was visualized under a fluorescence microscope (Leica, Wetzlar, GER). The primary antibodies were as follows: mouse anti-microtubule associated protein 2 (MAP2, 1:200), mouse anti-nestin (1:100, Millipore, Billerica, MA, USA), rabbit anti-glial fibrillary acidic protein (GFAP, 1:1,000), rabbit anti-β-III-tubulin (Tuj1, 1:1000), rabbit anti-2ʹ,3ʹ-cyclic nucleotide 3ʹ-phosphodiesterase (CNPase, 1:1000, Abcam, Cambridge, UK), rabbit anti-Ki67 (1:200, Sigma). The secondary antibodies were Alexa Fluor 568-conjugated goat anti-mouse IgG (1:1000) and goat anti-rabbit IgG (1:1000), Alexa Fluor 488-conjugated goat anti-rabbit IgG (1:1000, Invitrogen, Carlsbad, CA, USA). The positive cells were counted in 10 randomly selected microscopic fields per group. Briefly, under a fluorescence microscope with a magnification of 200 times, 10 fields were randomly selected on two perpendicular lines through the center of the well. The total positive cells of the 10 selected fields were counted and analyzed statistically. The mean perimeter of per positive cell were analyzed in the randomly selected microscopic fields.

Western blot (WB) analysis

Proteins were extracted from cells using mammalian Protein Extraction Reagent (Pierce) according to the manufacturer’s instructions. Then, the protein concentration was determined with an Enhanced BCA Protein Assay kit (Beyotime Institute of Biotechnology). Equal amounts of protein were separated on a 10% polyacrylamide gel in the presence of sodium dodecyl sulfate and then transferred to a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA, USA). Subsequently, the membrane was blocked with 5% non-fat milk in tris-buffered saline (TBS, Sangon) with 0.1% Tween-20 buffer (Sangon) and incubated overnight at 4 °C with the primary antibodies. After being washed with TBS, the membrane was incubated with the secondary antibodies at room temperature for 2 h and the complexes were visualized using enhanced chemiluminescence (Pierce). Finally, immunoreactivity was detected with the ChemiDoc XRS system (Bio-Rad). The primary antibodies were as follows: rabbit anti- POU3F4 (1:1000, Abcam), rabbit anti-Gli1 (1:1000, Abcam), rabbit anti-Post Synaptic Density 95 (PSD-95, 1:100, Abcam), and mouse anti-β-actin (1:1000, Abcam). The secondary antibodies were as follows: horseradish peroxidase-conjugated goat anti-mouse or -rabbit IgG (1:3000, Pierce). The expression levels of POU3F4, Gli1, and PSD-95 protein were determined relative to that of β-actin.

Chromatin immunoprecipitation (ChIP) assay

The ChIP assay was carried out according to the instructions in the SimpleChIP® Enzymatic Chromatin IP Kit (Agarose Beads, #9002, Cell Signaling Technology, Boston, MA, USA). First, 37% formaldehyde was added to the cells for cross-linking, and then glycine was used to terminate the cross-linking. The chromatin was digested into 150‒900 bp DNA–protein fragments using Micrococcal Nuclease and ultrasound treatment. Normal rabbit IgG antibody (#2729) was added to the negative control group, histone H3 (d2b12) XP® Rabbit mAb antibody (#4620) was added to the positive control group, and HA antibody was added to the LV-POU3F4 infection group. After incubation with the immunoprecipitating antibodies overnight at 4 °C, the protein/DNA complexes were co-precipitated and captured by protein G beads. Then, the chromatin was eluted from antibody/protein G agarose beads, and the DNA was purified using a centrifugation column and detected using reverse-transcription polymerase chain reaction (RT-PCR) analysis.

RT-PCR analysis

RT-PCR was performed on a Corbett RG-6000 PCR system (Dusseldorf Chagan, GER) using Faststart Universal SYBR Green Master Mix (Roche Diagnostics, Basel, SUI) according to the manufacturer’s protocol using the following PCR conditions: 95 °C for 10 min, followed by 40 cycles of 95 °C for 30 s, 60 °C for 30 s and 72 °C for 40 s. The primers (synthesized by Sangon) for Gli1 were as follows: Forward primer: 5ʹ-GCAGCAGGAATTGTTGTGGG-3ʹ; Reverse primer: 5ʹ-GCCTGATTTGTGATTGGCCG-3ʹ. PCR products were visualized by electrophoresis using a 1% agarose gel and scanned using the Q550I W image analysis system (Leica).

Establishment of mouse model injected with LV-POU3F4 in the hippocampus

Adult ICR mice (8 weeks old) were anesthetized using intraperitoneal injection of compound anesthetics at 0.1 ml/100 g body weight. Then, the mice were fixed on the TAXIC-653 stereotactic instrument (Zhenghua, Anhui, CHN). The skin above the skull was disinfected using 75% ethanol and gently cut, and the co-ordinates of the anterior fontanel on the sagittal axis (A), coronal axis (L), and vertical axis (V) were recorded. According to the mouse atlas, the co-ordinates of the injection points on the left and right DG were determined as: A1 = A – 0.2 cm, L1 = L – 1.2 cm, L2 = L + 1.2 cm, V1 = V + 2.0 cm. After drilling on the skull at the corresponding sites, the lentiviruses were injected into the DG with a micro-injection pump at a rate of 0.5 µl/min. The left DG was injected with LV-CTRL, and the right DG was injected with LV-POU3F4 (virus concentration: 1 × 10⁸ TU/ml; virus volume: 2 µl). After the injection, the needle was kept for 10 min, then withdrawn slowly. After ensuring the mice did not have an intracranial hemorrhage, the skin was sutured and the mice were returned to the cage according to their sex.

Transmission electron microscopy examination

Two weeks after lentivirus injection, the hippocampi were harvested from the mice after intraperitoneal anesthesia and trimmed into 1 cm³ tissue blocks, immediately put into glutaraldehyde fixed solution (containing 4% paraformaldehyde and 2.5% glutaraldehyde) for 2 h. Then, tissue blocks were fixed in 2.5% glutaraldehyde (stored overnight at 4 °C), and post-fixed in 1% OsO4 for 2 h. After fixation, samples were washed three times (5 min each time) with 0.1 M PBS. Samples were then dehydrated in a graded ethanol concentration (30, 50, 70, 95, and 100%), treated with propylene oxide, and embedded in Epon812. The ultrathin Sects. (70 nm) were produced using a Leica UC6 cryo ultramicrotome and collected on 100-mesh copper grids and sequentially stained with uranyl acetate and lead citrate. A Hitachi H7700 transmission electron microscope was used to examine the ultrastructure of synapses including the thickness of the PSD and the number of synapses that were statistically analyzed using the image-Pro Plus 6.0 Image analysis system.

RNA-sequencing

In this experiment, we injected LV-POU3F4 into the DG of the hippocampus in adult mice after anesthesia as above. The method for virus injection was similar to the method described above. The co-ordinates of the injection points were as follows: A1 = A – 0.2 cm, L1 = L – 1.2 cm, L2 = L + 1.2 cm, V1 = V + 2.0 cm. The left DG was injected with LV-CTRL, and the right DG was injected with LV-POU3F4 (virus concentration: 1 × 10⁸ TU/ml; virus volume: 5 µl). Ten days after virus injection, animials were euthanized using CO₂ inhalation and hippocampal tissues were collected for RNA extraction. After RNA quality evaluation, qualifying samples were stored on dry ice and submitted to the Sangon (Shanghai, China) for high-throughput sequencing. Differentially expressed genes were selected according to the following criteria: fold-change not less than 2, false-discovery rate value less than 0.01. Then, cluster and pathway analyses were performed on the differentially expressed genes.

Flow cytometry

The evaluation of cell apoptosis was performed using the PE Annexin V Apoptosis Detection Kit I (BD, USA) and flow cytometry. Briefly, single cells were stained with the PE Annexin V Apoptosis Detection Kit I. The apoptotic ratio was measured using BD FACSCalibur Flow cytometry.

CCK8 assay

NSCs from WT and POU3F4^−/− hippocampus were seeded in a 24-well plate at a density of 1 × 10⁴ cells per well. Each well was supplemented with 50 μl of CCK8 solution (Beyotime, China) and thoroughly mixed before incubating at 37 ℃ for 2 h in a Synergy2 multi-mode microplate reader to measure absorbance at 450 nm. The data was then collected, analyzed, and interpreted.

Statistical analysis

Data are expressed as mean ± standard error of the mean (SEM) based on at least three independent samples in each group and analyzed using Prism v.6.0 software (GraphPad Inc., San Diego, CA, USA). Experiments with two experimental groups were evaluated using an unpaired Student’s two-tailed t-test. In experiments with more than two experimental groups, one-way analysis of variance was used. P < 0.05 was considered statistically significant.

Results

Culture and identification of hippocampal NSCs

The method to isolate and purify the NSCs from hippocampus has been used for many years in our lab, the purity of NSCs has been described in our previous report [26]. In this study, the third generation of single cells isolated from the hippocampi of postnatal day 1 mice were cultured in NSC expansion medium. Seven days later, visible neurospheres from the WT mice were formed and suspended in the medium (Supplementary Fig. 1A). Immunocytochemical analysis showed that most cells were nestin-positive (Supplementary Fig. 1B), indicating that these neurospheres were embryogenic, and Ki67-positive (Supplementary Fig. 1C), suggesting that the active proliferation of the cultured hippocampal cells. The neurospheres were digested to single cells and plated onto the glass slide. Immunotaining indicated that about 95.8% cells were positive to Nestin and 94.2% both to Nestin and Ki67 (Supplementary Fig. 1D–G). In order to further determine whether these cells had multiple differentiation potential, third generation single cells were cultured in the differentiation medium for 7 d and immunocytochemical analysis showed that some cells expressed Tuj1 (Supplementary Fig. 1H), GFAP (Supplementary Fig. 1I), or CNPase (Supplementary Fig. 1J), indicating that the cultured hippocampal cells had the potential to differentiate into neurons, astrocytes, and oligodendrocytes. Therefore, it was believed that the cultured cells were NSCs.

In vitro effects of POU3F4 on differentiation and proliferation of hippocampal NSCs

Our previous reports have proven that POU3F4 possess the ability to promote hippocampal NSCs differentiating into neurons [19, 27]. To further determine the effects of POU3F4 on neuronal differentiation the hippocampal NSCs were cultured from postnatal day 1 POU3F4^−/− and WT littermates (genotype shown in Supplementary Fig. 2A). After 7 days of induced differentiation, the cells were subjected to immunocytochemical staining. The results indicated that the proportion and perimeter of MAP2-positive or Tuj1-positive cells in the WT hippocampal NSCs with forced expression of POU3F4 were more than that infected with the control lentivirus (Fig. 1A and B). Compared with the wild type, POU3F4 knockout significantly reduced the proportion and neurite length of neurons (Fig. 1C and D). Both flow cytometry and CCK8 assay demonstrated that POU3F4 knockout did not lead to decreased cell viability or apoptosis (Supplementary Fig. 2B and C). These results suggest that POU3F4 plays an important role in the differentiation of hippocampal NSCs into neurons. The rescue assay was conducted on the POU3F4^−/− NSCs that were infected with POU3F4 lentivirus (LV-POU3F4) or vector lentivirus (LV-Ctrl). The infection efficiency was shown in Supplementary Fig. 3A and B. Compared with the LV-Ctrl group, after incubation in the differentiation medium for 7 d the decreased neuronal differentiation was markedly reversed by POU3F4 overexpression (Fig. 1C and D). The results suggest that POU3F4 knockout reduced the differentiation of NSCs into neurons, but overexpression of POU3F4 could at least partially reverse the reduction. Together with our previous reports [19, 27], we concluded that the transcription factor POU3F4 could promote neuronal differentiation of hippocampal NSCs.

Fig. 1

Overexpression of POU3F4 facilitated neuronal differentiation of hippocampal NSCs. A, B Representative images and quantitative analysis of the number and perimeter of MAP2 (A) and Tuj1 (B) positive cells (green) following induced differentiation of wild-type hippocampal neural stem cells (NSCs) with or without POU3F4 infection. C, D Representative images and quantification of MAP2 (C) and Tuj1 (D) positive cells (red) in WT and POU3F4.^−/− hippocampal NSCs post-differentiation. Nuclei were stained with Hoechst (blue). n = 3. Scale bar = 100 µm. * P < 0.05, ** P < 0.01 and *** P < 0.001

In order to determine the effect of POU3F4 on the proliferation of hippocampal NSCs, the WT cells were infected with the vector (LV-Ctrl) and LV-POU3F4 for 12 h in the NSC expansion medium. Then, 48 h later, Ki67 immunostaining was performed on the cells. The results showed that the number of Ki67-positive cells in the NSCs overexpressing POU3F4 was significantly less than in the cells infected with the negative control virus (Fig. 2A and B). The size of the neurospheres generated from WT and POU3F4^−/− mice were analyzed on the 3rd, 7th, and 14th day to assess the proliferation rate of the NSCs. The results showed that on the 3rd day, the maximum diameter of the neurospheres in the two groups was between 50 and 100 µm, and the number of neurospheres with a diameter between 50 and 100 µm in the POU3F4^−/− group was greater than that in the WT group (Fig. 2C and D). On the 7th day, some neurospheres were found to be more than 100 µm in diameter. The number of neurospheres with a diameter between 100 and 200 µm in the POU3F4^−/− group was significantly greater than that in the WT group (Fig. 2C and E). On the 14th day, there was no difference in the number of neurospheres of various sizes between the two groups (Fig. 2C and F). The in vitro experiment demonstrated that overexpression of POU3F4 reversed the increased proliferation of POU3F4^−/− hippocampal NSCs, as observed in the POU3F4 KO model (Supplementary Fig. 4A and B). These findings suggest that POU3F4 may exert an inhibitory effect on hippocampal NSC proliferation, particularly during early stages of development.

Fig. 2

Effect of POU3F4 on proliferation of hippocampal NSCs. A Immunostaining for Ki67 (red) in WT mouse hippocampal NSCs which were infected with LV-POU3F4 or LV-Ctrl. Nuclei were stained with Hoechst (blue). B Quantification of Ki67/GFP-positive cells in the LV-Ctrl and LV-POU3F4 groups in A. C Morphology of neurospheres of hippocampal NSCs obtained from the WT and POU3F4^−/− mice at different days of culture. D–F Quantification of the number of neurospheres with different diameters in the 3rd d (D), 7th d (E) and 14th d (F) of culture. n = 3. Scale bar = 100 µm in (A). Scale bar = 400 µm in (C). * P < 0.05, *** P < 0.001 vs LV-Ctrl or WT

Differentiation and Proliferation in the DG of POU3F4 deletion mice

The expression of DCX serves as a marker for newborn neurons. We conducted DCX immunofluorescence staining to investigate the impact of POU3F4 on neurogenesis in the DG in vivo (Fig. 3A). Our findings demonstrate that the number of DCX-positive cells in the DG of POU3F4^−/− mice was significantly lower than that observed in WT mice, regardless of their age at either 3 or 5 weeks-old age (Fig. 3B). Furthermore, DCX-positive neurons exhibited typical cell bodies and processes in the WT DG of hippocampus, while there were only dot signals and abnormal processes in the POU3F4 KO hippocampal DG (Fig. 3A). However, we did observe a slight increase in Ki67 positive cells within the POU3F4^−/− DG compared to WT mice (Supplementary Fig. 5A and B). Additionally, the BrdU incorporation assay revealed a significant increase in labeled cells within the DG following POU3F4 knockout when compared to WT mice (Supplementary Fig. 5C and D). To illustrate the link between DCX-positive cells and proliferation, Ki67 and DCX dual immunofluorescence staining was conducted on adult brain slices. Results indicated that POU3F4-deletion mice had fewer and less developed DCX-positive neurons in the hippocampal DG compared to the WT group. Few new neurons in the DG exhibited both DCX and Ki67 markers (Supplementary Fig. 5E), likely because these neurons had lost their proliferative ability and no longer expressed Ki67. These results indicate an inhibitory effect of POU3F4 on cell proliferation in vivo, and, on the contrary, the potential influence on neuronal differentiation within the DG region.

Fig. 3

Effects of POU3F4 on the DG neurogenesis and behavior of learning and memory. A Immunofluorescence staining against DCX in the DG of WT and POU3F4^−/− mice (3 and 5-week-old). The arrows showed DCX positive cells. The boxes in the corners of the figures showed locally enlarged images. B Statistical analysis of the number of DCX positive cells in each group. n = 3. Scale bar = 400 µm. experiment respectively. The tested indexes were as follows: the latency time that mice took to find the target platform, the latency time that mice first arrived at the target area, and the number that mice passed through the target area. C, D, E The WT and POU3F4^−/− mice at 1 (C), 2 (D) and 5 (E)-month old age were subjected to Morris water maze. Left, middle and right panels are respectively the statistical analysis of the time to find platform, latency time to cross the target area and the number of times mice entered the target quadrant. F, G The WT and POU3F4^−/− mice at 5-month old age were subjected to passive (F) and active (G) avoidance task. C, n = 13; D, F, G, n = 8; E, n = 5. *P < 0.05, **P < 0.01, ***P < 0.001 vs WT

POU3F4 deletion declined learning and memory ability of mice

To test the effect of POU3F4 on learning and memory behavior in mice, WT and POU3F4^−/− mice were exposed to the MWM and avoidance tasks. Results from MWM indicated that the time WT mice took to find the target platform gradually decreased during the first four days of learning, while the time POU3F4^−/− mice took did not decrease significantly at either 1-, 2- or 5-month age. On the fifth day, the platform was removed, and the 90 s exploration training was started. The number of times the mice entered the target area and the latency time when they first arrived at the target area were recorded. Compared with WT mice, the latency time of POU3F4^−/− mice to cross the target area was longer, and the number of times they entered the target quadrant was also less (Fig. 3C–E). The passive and active avoidance tasks showed that POU3F4^−/− mice had lower latency to avoid the shock stimulus (Fig. 3F and G). These results indicated that knockout of POU3F4 impairs the learning and memory ability of mice, which may result from the negative effects on neurogenesis and synapse development.

POU3F4 promoted synaptic development

In order to determine the effect of POU3F4 on neuronal synaptic development, we first infected the WT hippocampal NSCs with LV-POU3F4 or POU3F4-shRNA for 12 h (the respective negative lentivirus, LV-Ctrl, was used as a control, efficiency shown in Supplementary Fig. 3 and Supplementary Fig. 6) and subsequently cultured cells in the differentiation medium for 7 d. WB analysis showed that, compared with controls, POU3F4 overexpression significantly increased the level of the synapse-specific marker PSD-95 (Fig. 4A), while POU3F4 knockdown reduced PSD-95 expression (Fig. 4B). Second, LV-POU3F4 was used to induce POU3F4 expression in the right DG of adult WT mice, and the contralateral side was injected with LV-Ctrl (LV-GFP). Two weeks later, the changes in synapses were determined by transmission electron microscopy. The results showed that after overexpression of POU3F4, there were more synapses in the hippocampus and the thickness of the PSD increased significantly (Fig. 4D–F). WB analysis also showed that the level of PSD-95 protein in the hippocampus increased significantly after overexpression of POU3F4 (Fig. 4C). To further determine the effect of POU3F4 on synaptic development, hippocampal tissues from 8-week-old WT and POU3F4^−/− mice were examined using electron microscopy. The results showed that, compared with WT mice, the number of synapses and the thickness of PSD in the hippocampus of POU3F4^−/− mice were significantly reduced (Fig. 4G-I). At the same time, the results of Western Blot detection also showed that the expression of PSD95 in the hippocampus of POU3F4^−/− mice was also decreased (Supplementary Fig. 10). These results indicate that POU3F4 can promote neuronal synaptic development in vivo and in vitro.

Fig. 4

POU3F4 promoted neuronal synaptic development during NSC differentiation. A, C upper panels are the WB analysis about the expression of PSD95 protein in WT hippocampal NSCs infected with LV-POU3F4 (A) or POU3F4-shRNA (B) and subsequently induced to differentiate. Lower panels are the statistical analysis of PSD95 protein in upper panels. C WB detection and statistical analysis of PSD95 protein in the hippocampus of WT mice after lentiviruses injection. D Transmission electron microscopy detection of synaptic morphology in the hippocampus of mice which were injected with LV-Ctrl or LV-POU3F4. The arrows showed synapses. E Quantification of the number of synapses in D. F Quantification of thickness of PSD in D. G The synaptic morphology in the hippocampus of WT and POU3F4.^−/− mice were detected by transmission electron microscopy. The arrows showed synapses. H Quantification of the number of synapses in G. I Quantification of thickness of PSD in G. The level of PSD95 in the LV-Ctrl group was set to 1. n = 3. Scale bar = 200 µm. * P < 0.05, ** P < 0.01 vs control or WT. Full-length western blots are presented in Supplementary material 1

Identification of Gli1 as a downstream gene of POU3F4

The above results and the data in our previous study [19, 27] confirmed that POU3F4 could promote neuronal differentiation and synapse development but inhibit proliferation in hippocampal NSCs. To explore the effects downstream of POU3F4 underlying the impact on the hippocampus, we constructed a mouse model that overexpressed POU3F4 in the hippocampus and used RNA-sequencing to examine the possible downstream factors regulated by POU3F4. The RNA-seq data were analyzed with FDR < 0.001 and fold change ≥ 2 as screening criteria, and the results showed that the transcription level of 638 genes elevated after POU3F4 overexpression. The heat map of some differentially expressed genes is shown in Fig. 5A. Pathway analyses suggested that chemokine signaling pathway, cholinergic and serotonergic synapse were most enriched among the differentially expressed genes (Fig. 5B). Gli1 was identified as a downstream gene of POU3F4 in the hippocampal NSCs, which was a member of the Shh signaling pathway that has been reported to be closely related to NSC development. Then, we used JASPAR software to predict the sites on the mouse Gli1 promoter capable of binding to POU3F4. The result suggested that there were three binding sites on the Gli1 promoter, and the correlation scores were all greater than 0.8 (Fig. 5C). Following that, we used a ChIP assay to determine whether POU3F4 could bind to the Gli1 promoter. The results showed that a Gli1 gene fragment could be precipitated by the HA antibody (Fig. 5D), indicating that POU3F4 could bind to the promoter region of the Gli1 gene. Then, we used NSCs to further verify whether POU3F4 could influence the expression of the Gli1 protein. NSCs were induced to overexpress POU3F4 via lentivirus infection for 12 h and subsequently cultured for 3 d in expansion medium. The result indicated that the expression of the Gli1 protein (Fig. 5E and G) was obviously upregulated with an increase in POU3F4 levels (Fig. 5E and F). The level of Gli1 in the hippocampus of POU3F4^−/− mice was also decreased (Supplementary Fig. 10). This indicated that Gli1 was a downstream gene from the transcription factor POU3F4.

Fig. 5

Identification of Gli1 as a downstream gene of POU3F4 in the hippocampal NSCs. A Cluster Heatmap showing the differentially expressed genes after overexpression of POU3F4 in hippocampus. B Pathway analysis of these differentially expressed genes. C Predicted sites of Gli1 promoter which may be bound with POU3F4 using JASPAR software. D Detection of binding effect of POU3F4 and Gli1 promoter by ChIP. E WB detecting the expression of POU3F4 and Gli1 in NSCs after POU3F4 overexpression. Full-length blots are presented in Supplementary material 1. F Quantification of POU3F4 protein in the LV-POU3F4 group normalized to that in the LV-Ctrl group in D. G Quantification of Gli1 protein in D. The amount of POU3F4 and Gli1 protein in the LV-Ctrl group was set to 1 respectively. n = 3. *P < 0.05, *** P < 0.001 vs LV-Ctrl

Gli1 promoted hippocampal NSCs to differentiate into neurons.

In order to identify the role of Gli1 in the neuronal differentiation of NSCs, cells were transfected with lentiviruses coding Gli1 (LV-Gli1), shGli1 and vector control (LV-Ctrl) to up- or down-regulate level of Gli1 (infection efficiency shown in Supplementary Fig. 7 and Supplementary Fig. 8), and subsequently induced to differentiate for 7 d. Immunocytochemical staining showed that the numbers of Tuj1 (Fig. 6A and C) and MAP2-positive cells (Fig. 6A and D) were significantly increased after overexpression of Gli1, while the proportion of GFAP-positive cells (Fig. 6A and B) in the LV-Gli1 group was lower than that in the LV-CTRL group. On the contrary, downregulation of Gli1 markedly inhibited the differentiation of hippocampal NSCs into Tuj1 (Fig. 6E and G) and MAP2 (Fig. 6E and H) positive neurons, meanwhile, the proportion of differentiated astrocyte increased markedly (Fig. 6E and F). These results indicated that Gli1 could promote hippocampal NSCs to differentiate towards neurons, but not glial cells.

Fig. 6

Gli1 promoted hippocampal NSCs to differentiate into neurons. Single hippocampal NSCs were seeded in a 48-well plate at a density of 1 × 10⁴ cells/well and followed with infection of LV-Ctrl, LV-Gli1, shCtrl or shGli1 for 12 h and subsequent culture in differentiation medium for 7d. A Immunocytochemical staining against GFAP, Tuj1, and MAP2 antibodies (all are red) after induced differentiation of hippocampus NSCs transfected lentivirus-Gli1. B–D Quantification of the GFAP (B), Tuj1 (C) and MAP2 (D) positive cells in A. (E) Immunocytochemical staining against GFAP, Tuj1 and MAP2 antibodies (all are red) after induced differentiation of hippocampus NSCs transfected shGli1. F, G and H Quantification of the GFAP (F), Tuj1 (G) and MAP2 (H) positive cells in E. Nuclei were stained with Hoechst (blue). n = 3. ** P < 0.01, *** P < 0.001, **** P < 0.0001 vs control. Scale bar = 200 µm

Gli1 rescued the POU3F4 knockout-induced decrease in neuronal differentiation

Hippocampal NSCs from POU3F4^−/− mice differentiated into fewer neurons than those from WT mice (Fig. 7A and B, upper two panels, both infected with LV-Ctrl). Overexpression of Gli1 in POU3F4^−/− hippocampal NSCs via infection with lentivirus-Gli1 significantly increased the proportion of MAP2-positive cells compared with POU3F4^−/− NSCs infected with control lentivirus (Fig. 7A and D). The same result was also found following Tuj1 immunocytochemical staining (Fig. 7B and E). These results illustrate that Gli1 not only promoted differentiation of NSCs into neurons but also at least partially rescued POU3F4-deficiency-induced neuronal differentiation inhibition.

Fig. 7

Gli1 reversed proliferation increase and neuronal differentiation inhibition of the POU3F4-knockout hippocampal NSCs. Hippocampal NSCs derived from WT and POU3F4^−/− mice were infected with lentiviruses and induced to differentiate, and then subjected to immunocytochemical analysis. A, B, C Immunostaining showing the MAP2 (A), Tuj1 (B), Ki67 (C)-positive cells (red). D, E, F Quantification of the proportion of the above corresponding positive cells in A, B and C. Nuclei were stained with Hoechst (blue). n = 3. * P < 0.05, *** P < 0.001. Scale bar = 100 µm

Gli1 suppressed the increase of NSC proliferation after POU3F4 knockout

The previous results and Fig. 7C and F showed more proliferation of NSCs with POU3F4 deletion. To determine the effects of Gli1 on the hippocampal NSCs, the POU3F4^−/− hippocampal NSCs were infected with lentivirus-Gli1. The results indicated that overexpression of Gli1 counteracts the proliferative increase observed following POU3F4 deletion, suggesting that Gli1 exerts an inhibitory effect on the proliferation of hippocampal NSCs (Fig. 7C and F).

Gli1 promoted synaptic development during differentiation of NSCs

To determine the role of Gli1 in neuronal synaptic development, hippocampal NSCs were transfected with Gli1 and vector lentiviruses for 12 h. After 7-day induced differentiation, WB was used to determine the expression of PSD-95. It was found that overexpression of Gli1 could increase the expression of PSD-95 (Supplementary Fig. 9A and B). Furthermore, Gli1 in the NSCs was knocked down using the shRNA lentivirus (targeting Gli1). After 7-day induced differentiation, the results showed that the expression of PSD-95 decreased significantly (Supplementary Fig. 9C and D). These results indicated that Gli1 could promote neuronal synaptic development during the differentiation of hippocampal NSCs.

Gli1 promoted the improvement of cognitive function and neuronal development

To investigate Gli1's impact on cognitive function, we injected lentivirus into the hippocampus of animal models. After a period of 35 days, we employed the Morris water maze to evaluate the animals' learning and memory capabilities. The results demonstrated a significant decline in the learning and memory capabilities of the animals following denervation injury. However, a notable restoration of cognitive function was observed subsequently to Gli1 overexpression by decreasing escape latency and path length, and increasing platform crossings (Fig. 8A–D). Additionally, immunofluorescence staining revealed the presence of DCX-positive neurons in the DG of the hippocampus. In the hippocampus of adult mice, DCX-positive neurons were predominantly localized within the subgranular layer of DG, with their dendrites extending into the granular layer. Following denervation injury, there was a notable reduction in the number of DCX-positive neurons. However, subsequent Gli1 overexpression led to a restoration in both the number of DCX-positive neurons and their dendritic projections (Fig. 8E and F). These findings suggest that Gli1 could hold potential therapeutic value for the treatment of AD.

Fig. 8

Effect of Gli1 on the improvement of cognitive function and neuronal development. The lentivirus with or without Gli1 were injected into model or sham animal hippocampi. 35 days later, the animals were subjected to learning and memory behavior test and immunofluorescent staining. (A) representative trajectory diagrams to reach the platform. (B) Statistical analysis of the escape latency between groups in A. (C) representative trajectory diagrams to cross the platform location during a single 120 s probe trial. (D) Statistical analysis of the number crossing target platform between groups in C. (E, F) Immunofluorescence analysis of DCX-positive (green) cells in hippocampal DG. Nuclei were stained with Hoechst. The bar represents 400 μm. The data are presented as mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001

Discussion

POU3F2, POU3F3, and POU3F4 belong to the POU protein family of transcription factors. They appear at an early embryonic stage and are widely distributed in the nervous system [11,12,13], suggesting that these factors may be involved in the development of the early nervous system. Zhang et al. established a denervated hippocampus model by cutting the fimbria fornix of adult rats to block the cholinergic fibers projecting from the septum to the hippocampus, and then transplanted NSCs from the subventricular zone (SVZ) into the hippocampus. The results showed that in the hippocampus of the denervated adult rats the implanted NSCs were more likely to survive and differentiate into neurons, and POU3F4 expression was upregulated [27, 28]. In this study, we found that POU3F4 could influence neurogenesis in the hippocampal DG of mice in vivo. If POU3F4 was absent, neurogenesis in the DG would be damaged, and learning and memory functions of mice would also be impaired. All of these results indicate that POU3F4 is involved in the differentiation of hippocampal NSCs into neurons and the maintenance of learning and memory functions. However, the mechanism of POU3F4-mediated hippocampal neurogenesis is still unknown. So, first of all, we used NSCs derived from the hippocampus of postnatal day 1 WT and POU3F4^−/− mice to study the relationship between POU3F4 and NSC differentiation. The results showed high purified NSCs were generated from the hippocampus, of which only a few are other cells, which is convenient for subsequent differentiation experiments. Overexpression of POU3F4 promoted the differentiation of NSCs into neurons. However, POU3F4 knockout inhibited the neuronal differentiation of hippocampal NSCs, and POU3F4 re-overexpression could partially rescue this condition. This result further verified that POU3F4 could promote the neuronal differentiation of hippocampal NSCs.

Then, we constructed a mouse model overexpressing POU3F4 in the hippocampus. Through RNA-sequencing and pathway analysis, we found that the expression of Gli1, a member of the Shh signaling pathway that is closely related to NSC development, was upregulated after POU3F4 overexpression. Studies have reported that there are three members of the Gli family, Gli1, Gli2 and Gli3, all of which have five conserved zinc finger structures [20]. Among the three members, Gli1 is the only factor that directly initiates transcription by Shh signal. It has been confirmed that transplantation of exogenous umbilical cord blood mononuclear cells to neonatal rodents with ischemia and hypoxia can promote the differentiation of endogenous NSCs into neurons and reduce the proportion of them differentiating into glial cells. During this process, the expression of Shh and Gli1 is significantly increased [24]. One of the polyphenol monomers in green tea, epigallocatechin-3-gallate (EGCG), can promote neurogenesis in the hippocampus of adult mice. At the same time, the expression of Shh and Gli1 has also been upregulated [29]. Learning can activate the Shh signaling pathway in the amygdala of adult mice, and increase the expression of the Shh receptor Ptc1 and its downstream factor Gli1, resulting in the emergence of new neurons in the amygdala and the production of growth memory [30]. These studies suggest that Shh and Gli1 may be involved in the process of neurogenesis. Based on these studies and our RNA-sequencing results, we suppose that Gli1 might be a downstream gene of POU3F4 in rodents and that POU3F4 could promote NSCs to differentiate into neurons by increasing the expression of Gli1.

Next, we further used JASPAR software to predict the POU3F4-binding sites on Gli1, and the result showed that there were at least three possible POU3F4 binding sites in the promotor domain of Gli1. Our ChIP experiment confirmed that POU3F4 protein could bind to the Gli1 promotor. Subsequently, we verified that POU3F4 could increase the protein expression of Gli1 in hippocampal NSCs. So far, we believe that, as a downstream gene of POU3F4, the protein expression of Gli1 could be regulated by POU3F4. However, this result does not lead to the conclusion that POU3F4 promotes NSCs to differentiate into neurons just by upregulating Gli1. Therefore, we cultured NSCs derived from the hippocampus of POU3F4^−/− mice. We found that the proportion of neuronal differentiation in NSCs derived from the POU3F4^−/− mice was lower than that from the WT mice, while overexpression of Gli1 could rescue this condition partially in the POU3F4^−/− NSCs. These results indicated that Gli1 promoted the differentiation of NSCs into neurons, even without POU3F4 in vitro. Based on these results, we concluded that Gli1, as a downstream gene of POU3F4, could participate in the process of POU3F4 promoting NSCs to differentiate into neurons. However, in vivo rescue experiments are necessary to verify the role of Gli1 as a downstream POU3F4 gene in promoting neuronal differentiation. Of course, we could not rule out the possibility that other factors downstream of POU3F4 participated in this process.

The synapse is the place of functional connection between neurons and the key part of information transmission. Whether the synapse develops well or not can indicate whether the neurons are mature. Therefore, we further examined the effects of POU3F4 and Gli1 on neuronal synaptic development during NSC differentiation. PSD-95 is the most important and abundant scaffold protein on the postsynaptic membrane. It mainly exists in mature excitatory glutamatergic synapses and plays an important role in synaptic plasticity [31]. Synapsin I is abundant in most neural cells and can regulate the release of neurotransmitters [32]. Therefore, synapsin I and PSD-95 are markers of synaptic remodeling [33]. Our study showed that during the differentiation of hippocampal NSCs, overexpression of POU3F4 or Gli1 could increase the expression of PSD-95, while after knockdown of POU3F4 or Gli1, the expression of PSD-95 was significantly reduced. We also detected the effect of POU3F4 and Gli1 on the expression of synapsin I. We found that no matter whether POU3F4 or Gli1 levels were up- or downregulated in the NSCs, the expression of synapsin I did not change significantly (data not shown). These results indicated that POU3F4 could promote the development of neuronal synapses during the differentiation of NSCs into neurons. Once POU3F4 was absent, the development of neuronal synapses would be affected. Together with the findings from the transmission electron microscopy examination, our results imply that the effect of POU3F4 on synaptic development mainly occurred in the postsynaptic membrane, and Gli1 could also participate in this process. Whether POU3F4 and Gli1 can affect the development and function of all neuronal synapses remains to be further studied.

There is no clear conclusion about the effect of POU3F4 on cell proliferation. In this study, we found that overexpression of POU3F4 inhibited the proliferation of NSCs derived from WT mice. Furthermore, the proliferation rate of NSCs derived from POU3F4^−/− mice was higher than that of cells obtained from WT mice during an early stage of development, while at a later stage this change was not significant. We believe that POU3F4 inhibits the proliferation of NSCs at an early stage of their development, which is not necessarily a good phenomenon. The more NSCs proliferate during the early stage of their development, the more neurons they differentiate into. Therefore, POU3F4 inhibits the proliferation of NSCs at an early stage, which does not seem to meet the needs of biological neural development. We speculate that at the early stage of neural development, there should be multiple factors involved in regulating the proliferation of NSCs, not just POU3F4. Perhaps, the inhibitory effect of POU3F4 on the proliferation of NSCs only plays a small part in regulating their proliferation and may not even play a decisive role in their proliferation. Therefore, we think that the effect of POU3F4 on NSCs is mainly related to the regulation of their differentiation, rather than their proliferation. In fact, some studies have used POU3F4 for cell reprogramming or cell trans-differentiation toward neurons and achieved good results [34, 35], signifying the importance of POU3F4 in the decision of stem cell fate.

We provide new evidence that Gli1 overexpression inhibits the proliferation increase seen after POU3F4 deletion, indicating Gli1's role in suppressing hippocampal NSC proliferation. While Gli1 is a downstream component of the Shh signaling pathway, known for promoting proliferation in certain progenitor and stem cells, there are still ambiguous reports about the effect of Gli1 on stem cell proliferation. Specifically, within the nervous system, Gli1 boosts neural precursor proliferation and is key in cellular transformation and tumorigenesis [36]. However, Gli1 knockdown in oligodendrocyte precursor cells (OPCs) significantly increases their proliferation [37]. Additionally, IL-1β has been shown to upregulate Gli1 expression, thereby mediating the inhibition of OPC proliferation. Furthermore, Knockout of Gli1 in postnatal hippocampal NSCs greatly reduces their self-renewal and proliferation [38]. This suggests that Gli1's role in cell proliferation varies with cell type and developmental stage, and the evidence in this study indicated that it may partially inhibit the proliferation of these specific NSCs. The in vivo data suggest that Gli1 has the potential to enhance cognitive function in Alzheimer's disease (AD) models subjected to mechanical injury, indicating its possible therapeutic application for AD.

In summary, we confirmed that POU3F4 could promote the neuronal differentiation and synaptic development of hippocampal NSCs by increasing the expression of Gli1. Our study provides a necessary experimental basis for elucidating the mechanism of neuronal differentiation of NSCs induced by POU3F4. Of course, further research is required to determine how POU3F4 affects Gli1 expression and whether there are other downstream factors involved in promoting NSC differentiation. In addition, we also unexpectedly found that POU3F4^−/− mice showed abnormal movement with one-direction turning (data not shown). The reasons for this need further investigation. It is uncertain whether POU3F4 is involved in the regulation of behavior or central motor movements. POU3F4 deletion might cause anxiety and insecurity in mice, leading to abnormal movement. However, we had no conclusive evidence to explain this phenomenon. Therefore, the function of POU3F4 is worthy of further study. A major limitation is that we used a wholemount POU3F4 knock out mouse model in this manuscript. Analysis of conditional knock out model with temporal or regional control of POU3F4 deletion would enable to make stronger conclusions about the role of POU3F4 in hippocampal adult neural stem cells, which has been included in our next plan.

Conclusions

POU3F4 promotes neuronal differentiation and synaptic development of hippocampal NSCs by increasing the expression of Gli1.