Erythropoietin promotes the differentiation of fetal neural stem cells into glial cells via the erythropoietin receptor‐β common receptor/Syne‐1/H3K9me3 pathway

Abstract Aims To investigate the effect of erythropoietin (EPO) on the differentiation of neural stem cells (NSCs)/neural progenitors (NPs) in the treatment of hypoxic–ischemic injury and its potential mechanisms. Methods Fetal NSCs/NPs were treated with EPO after oxygen and glucose deprivation/reoxygenation (OGD/R). Cell viability, proliferation, and differentiation of NSCs/NPs were detected by CellTiter‐Glo, Edu assay, flow cytometry, and quantitative real‐time PCR (qPCR). Immunofluorescence staining, co‐immunoprecipitation (Co‐IP), and western blotting were used to test the existence of EPO receptor/β common receptor (EPOR/βCR) heterodimer on NSCs/NPs and the possible pathway. Results EPO treatment at different time points increased cell viability without affecting proliferation. EPO treatment immediately after OGD/R promoted oligodendrocyte and astrocyte differentiation, while decreasing neuronal differentiation of NSCs/NPs. EPOR/βCR heterodimer existed on the cell surface of the fetal cortical NSCs/NPs, EPO treatment significantly increased the mRNA expression of βCR and elevated the correlation between EPOR and βCR levels. In addition, mass spectrometry analysis identified Syne‐1 as a downstream signaling molecule of the EPOR/βCR heterodimer. Immunofluorescence staining and western blotting indicated that the βCR/Syne‐1/H3K9me3 pathway was possibly involved in the differentiation of fetal neural stem cells into the glial cell effect of EPO. Conclusion EPO treatment immediately after OGD/R could not facilitate fetal NSCs/NPs neurogenesis but promoted the formation of the EPOR/βCR heterodimer on fetal NSCs/NPs, which mediates its function in glial differentiation.


Erythropoietin (EPO) is a cytokine mainly induced under hypoxic
conditions, which primarily acts on erythroid progenitor cells in the bone marrow. Compelling evidence has revealed EPO's secondary functions, with a crucial focus on the central nervous system (CNS).
High-dose EPO administration prompts newly differentiating neurons in the adult mouse brain. 1 However, the regulatory effects of EPO on neuronal differentiation at different stages of brain development remain unclear.
EPO may interact with up to four distinct isoforms of its receptor (homodimers EPOR/EPOR or hetero-oligomers EPOR/ βCR), activating different signaling cascades with roles in neuroprotection and neurogenesis. High-dose EPO administration amplifies autocrine/paracrine EPO/EPOR signaling, prompts the emergence of new CA1 neurons, and enhances dendritic spine densities. 1 Although the presence of these different isoforms paves the way for new interesting mechanisms through which EPO could exert its function in the brain, many questions remain unanswered. Certainly, it is necessary to investigate how EPO's binding to each isoform activates a specific intracellular pathway, modulating the expression of target genes. Once these mechanisms have been elucidated, it will be possible to develop new isoform-selective drugs, resulting in a more specific therapy. However, the direct interaction between EPOR and βCR to form an innate repair receptor remains controversial, 2 and whether EPOR/βCR heterodimer is present on NSCs/NPs and whether it mediates the neurogenesis effect of EPO remain unknown.
This study used primary cortical NSCs/NPs from fetal mice to explore the effect of EPO intervention on neurogenesis in vitro at different time points after reoxygenation and the possible mechanisms to provide a novel scientific basis for future experimental studies and clinical trials of cerebral ischemic diseases.

| Culture of primary neural stem cells (NSCs) and neural progenitors (NPs)
Female C57BL/6J mice (25- 1% GlutaMAX, and 1% penicillin/streptomycin, plated in ultra-low attachment six well-plates, and maintained at 37°C in a humidified atmosphere containing 5% CO 2 . NSCs and NPs were subcultured by Accutase every 2-3 days. The third passage of cells was plated onto PDL-coated cell culture plates with complete medium.

| Oxygen and glucose deprivation (OGD) operation
The cell medium was replaced with glucose-free DMEM (Gibco; Thermo Fisher Scientific, Inc.) in a three-gas incubator containing 95% nitrogen and 5% CO 2 at 37°C for 1, 2, 4, 6, and 8 h. The cells were then switched to glucose-containing medium and maintained at 37°C in a 5% CO2 incubator for 1, 2, 4, and 7 days.

| Cell viability assay
Cell viability was assessed using the CellTiter-Glo® Luminescent Cell Viability Assay. Briefly, the plate and its contents were equilibrated at room temperature for approximately 30 min, followed by the addition of a volume of CellTiter-Glo®Reagent equal to the volume of cell culture medium present in each well and mixed for 2 min on an orbital shaker to induce cell lysis. The plate was incubated at room temperature for 10 min to stabilize the luminescent signal, and luminescence was recorded using a luminometer.

| EdU incorporation assay
Proliferation of NSCs and NPs was assessed using Click-iT® EdU Imaging Kits according to the manufacturer's instructions. Isolated NSCs and NPs were seeded in 48-well-plates (4 × 10 5 cells/well) and cultured in medium without EGF and bFGF. After OGD operation, EdU was added to the medium, and the cells were cultured for another 4 days. Cultured cells were fixed with 4% paraformaldehyde (PFA) for 15 min, followed by permeabilization using 0.5% Triton X-100. After permeabilization, the cells were incubated with Click-iT® reaction cocktail for 30 min in a dark room. The total number of cells and Edu-positive cells were counted using highcontent analysis.

| Flow cytometry for epidermal growth factor receptor and doublecortin assessment
Neurospheres were dissociated using Accutase (Sigma-Aldrich). A fraction of cells from each tube was selected to prepare a negative control tube (unmarked cells). Cells were fixed with 4% PFA for 30 min, washed with 1 ml phosphate-buffered saline (PBS) 0.15% bovine serum albumin (BSA), and incubated with 0.5% Triton X-100 for 30 min. Cells were washed once with 1 ml PBS 0.15% BSA and incubated with EGF receptor (EGFR) antibody (Alexa Fluor® 750, Novus) or doublecortin (DCX) antibody (PE, Biorbyt) for 30 min in the dark. The cells were washed once and resuspended in PBS. Cells were then analyzed on a BD FACSCalibur flow cytometer.

| Quantitative real-time reverse transcription PCR
Purified RNA from NSCs and NPs was used as a template to synthesize cDNA using oligo-d (T) primers and SuperScript III / RNaseOUT Enzyme Mix (Invitrogen). Relative gene expression was calculated using the 2 −ΔΔCT method, normalized, and expressed as fold change relative to U6 or β-actin. Real-time polymerase chain reaction was performed in triplicate: primers for myelin basic protein (Mbp): F: 5'-TCCGACGAGCTTCAGACCA-3′ and R:

| Co-immunoprecipitation
Co-immunoprecipitation (Co-IP) assays were performed to identify the proteins. Briefly, 1 × 10 6 cells were collected and lysed in 300 μl buffer containing non-denaturing lysis buffer, protease inhibitor, and phosphatase inhibitors. For Co-IP using antibodies, before being added to the cell lysates, the antibodies were incubated with Protein A/G Magnetic Beads (Sigma-Aldrich) and IgG for 3 h at 4°C to eliminate nonspecific binding. Subsequently, the cross-linked Protein A/G Magnetic Beads were added directly to the cell lysates and incubated overnight at 4°C. The magnetic beads were washed with IP wash buffer and collected. The protein complexes were eluted from the beads using 50 mM glycine (pH 2.8) and analyzed by western blotting.

| Western blotting
Cell samples were processed for western blotting analysis, as previ-

| β common receptor siRNA transfection
NSCs and NPs were transfected with βCR siRNA using Lipofectamine Stem Transfection Reagent (Invitrogen) for 48 h according to the manufacturer's instructions, and the transfection efficiency was validated by western blotting.
The images were digitized using an Olympus Fluoview FV1000 microscope (Olympus).

| Statistical analysis
Statistical analysis was performed using SPSS version 23.0. Data were tested for normality using the Shapiro-Wilk test (p < 0.05).
Numerical data are presented as the mean standard deviation.
Student's t-test was used for two-group comparisons. One-way analysis of variance (ANOVA) with the Tukey-Kramer post hoc test was used for comparisons among several quantitative variables.
The correlation between the two variables was assessed using the Pearson's correlation test. Data for neurobehavioral tests were analyzed using two-way repeated measures ANOVA followed by the Bonferroni post hoc correction. The correlation between two variables was determined using Pearson's correlation test. p < 0.05 was considered statistically significant.

| OGD time-dependently depressed neurogenesis of NSCs/NPs by decreasing cell proliferation and viability
To mimic the pathological changes in hypoxic-ischemic injury, we

| EPO treatment at different time point after OGD/R rescued cell viability rather than proliferation of NSCs/NPs
Next, we explored the effect of EPO on the proliferation rate and cell viability of cortical NSCs/NPs after OGD 2 and 6 h, representing mild and severe hypoxic injury. We showed that EPO administration at 1 to 6 h after OGD for 2 or 6 h could increase the survival of NSCs/NPs (Figure 2A, B, p < 0.05), with no effect on the proliferation rate ( Figure 2C, D). Moreover, we detected EGFR% by flow cytometry to compare the ratio of actively proliferating cells among the 2 h OGD, immediate EPO treatment after OGD (EPO immediately), and delayed EPO treatment (EPO 1 h) groups. The results indicated that there was almost no difference in the ratio of EGFR-positive cells among the three groups 24 h after OGD ( Figure 2E). Therefore, immediate or delayed delivery of EPO could promote neural regeneration after hypoxic injury by increasing cell viability without influencing cell proliferation.

| EPO treatment immediately after OGD/R boosted the differentiation of NSCs/NPs toward oligodendrocytes and astrocytes
Differentiation of NSCs/NPs into neurons, oligodendrocytes, and astrocytes is another key process during neurogenesis. First, we explored differentiation toward neurons. We detected DCX% by flow cytometry to compare the differentiation to neurons in the 2 h OGD, immediate EPO treatment after OGD (EPO immediately), and delayed EPO treatment (EPO 1 h) groups 1 day after OGD 1, 2, 4, 6, and 8 h ( Figure 3A). Unexpectedly, when treated immediately, EPO decreased the DCX%, indicating the inhibition of EPO on differentiation toward neurons ( Figure 3A, p < 0.05), whereas EPO treatment 1 h after OGD did not alter the ratio of DCX-positive cells, suggesting that delayed delivery of EPO exerts no effect on the differentiation of NSCs/NPs toward neurons. Furthermore, by detecting the relative mRNA levels of MAP-2 and β-tubulin, two markers for mature neurons, at 7 days after OGD, we found that OGD and immediate EPO treatment did not significantly alter the potential of NSCs/ NPs to differentiate into neurons ( Figure 3B, C).
We further investigated the effect of immediate EPO treatment on the glial differentiation of NSCs/NPs. We demonstrated that OGD from 1 to 8 h did not significantly change the relative mRNA levels of MBP and CNPase, two markers of mature oligodendrocytes, 7 days after OGD, suggesting that the differentiation toward oligodendrocytes may not be directly affected by hypoxia. However, the relative mRNA levels of both MBP and CNPase were significantly elevated after EPO administration, suggesting that EPO could promote oligodendrocyte regeneration ( Figure 3D, E, p < 0.05). The mRNA level of GFAP, a marker for mature astrocytes, was significantly increased by EPO treatment. Another marker for astrocytes, S100β, was also slightly increased by EPO treatment in the OGD 1, 4, 6, and 8 h groups, indicating that EPO could also promote the differentiation toward astrocytes ( Figure 3F, G, p < 0.05). In summary, EPO treatment immediately after reoxygenation promoted the differentiation of NSCs/NPs toward oligodendrocytes and astrocytes, but not neurons.

| EPOR/β CR heterodimer exists on NSCs/NPs
To determine the existence of EPOR homodimers or heterooligomers on the surface of NSCs/NPs, immunofluorescence, confocal microscopy, and Co-IP were used. We first identified that both EPOR/EPOR homodimers and EPOR/βCR hetero-oligomers were present on the NSCs/NPs ( Figure 4A). Furthermore, Co-IP indicated that EPOR and βCR were bound together in NSCs/NPs in the control, OGD, and EPO treatment groups ( Figure 4B). Moreover, EPO treatment did not influence the mRNA levels of EPOR ( Figure 4D Figure 4E). In summary, we preliminarily showed the existence of an EPOR/βCR heterodimer on the surface of NSCs/NPs, and EPO treatment significantly increased the mRNA expression of βCR and elevated the correlation between EPOR and βCR levels.

| Correlation between EPOR/β CR and neural markers in NSCs/NPs cells
To establish a preliminary association between EPOR/βCR heterooligomers and neural cell markers, we performed a correlation analysis between the mRNA levels of cell markers and the two receptors of EPO. The results indicated that EPOR was positively correlated with the expression of CNPase, MBP, and GFAP ( Figure 5B, C, p < 0.05), whereas no significant trends were observed between EPOR and MAP-2, βIII-tubulin, and S100β ( Figure 5A, C). βCR was positively correlated with the expression of MAP-2 and βIII-tubulin ( Figure 5D, p < 0.05), and no significant correlation was found between βCR and CNPase, MBP, GFAP, and S100β ( Figure 5E, F). Our results suggest that EPOR might be involved in the differentiation of NSCs/NPs into mature oligodendrocytes and astrocytes, whereas βCR is possibly related to the differentiation of NSCs/NPs into mature neurons. Therefore, the expression of EPOR/βCR heterodimer was correlated with the expression of oligodendrocyte markers.
In addition, there were five potential proteins that were oxidized ( Table 2) and another five potential proteins that were phosphorylated ( Table 3) in the EPO-treated cells. In sum, we identified 37 types of potential proteins that were expressed or chemically modified by EPO treatment, with one potential protein, Syne-1, which was among the 28 proteins found only in EPO-treated cells and phosphorylated by EPO treatment (Figure 6D, E). Among them, we hypothesized that Syne-1 is a key molecule for the effect of EPO on the oligodendrocyte differentiation of NSCs/NPs based on a previous study. 3 Since Syne-1 was found to bind to EPOR, confocal immunofluorescence showed that Syne-1 was colocalized with βCR on the NSCs/NPs ( Figure 6F), suggesting that Syne-1 is a potential downstream signaling molecule of the EPOR/βCR heterodimer.
According to previous studies, H3K9me3 is one of the downstream molecules of Syne-1 in oligodendrocyte progenitors, and the methylation of H3K9 exerts important effects on myelination and cell differentiation. [3][4][5] Thus, we investigated whether Syne-1/H3K9me3 is downstream of the EPOR/βCR heterodimer. Immunofluorescence staining showed that Syne-1 expression was decreased after OGD, whereas EPO treatment increased Syne-1 expression, which was reversed by βCR siRNA ( Figure 6G). Furthermore, western blotting demonstrated that H3K9me3 was upregulated after OGD and was reduced by EPO treatment. Similarly, βCR siRNA reversed the effect of EPO on H3K9me3 ( Figure 6H, p < 0.05). Thus, EPO activates the EPOR-βCR/Syne-1/H3K9me3 signaling pathway and controls the cell fate switch of NSCs/NPs.

| DISCUSS ION
The present study investigated the effect of EPO on post-hypoxic neurogenesis using primary cortical NSCs/NPs in fetal mice. We F I G U R E 2 EPO increases the cell viability while exerting no effect on proliferation of neural stem and progenitor cells. (A, B) Cell viability after 2 or 6 h of OGD plus EPO treatment was determined by CellTiter-Glo®Luminescent assay at 2 days after OGD. n = 6/group. *p < 0.05 vs. the OGD group. (C, D) Cell proliferation rate after 2 and 6 h of OGD plus EPO treatment was determined by Edu assay at 4 days after OGD. EPO (immediately) denotes cells were given EPO (50 U/ml) immediately after OGD treatment. EPO (1, 2, 4, 6 h) denotes cells were given EPO (50 U/ml) 1, 2, 4, or 6 h after OGD treatment. n = 6/group. *p < 0.05 vs. the control group. (E) EGFR% was detected by flow cytometry 1 day after 2 h of OGD treatment. OGD denotes cells were not given EPO treatment after OGD. EPO (immediately) denotes cells were given EPO (50 U/ml) immediately after OGD treatment. EPO (1 h) denotes cells were given EPO (50 U/ml) 1 h after OGD treatment. n = 6/group Enhancement of neurogenesis is generally an effective therapeutic strategy for brain injury, especially hypoxia/ischemia-induced brain damage. As a prominent neuroprotective agent, many preclinical experiments have shown that EPO can promote neurogenesis and oligodendrogenesis following hypoxic-ischemic cerebral insults.
However, the present study showed that EPO treatment increased the viability of fetal cortical NSCs/NPs and promoted their proliferation and differentiation toward oligodendrocytes and astrocytes rather than neurons. After cerebral stroke, endogenous neurogenesis was insufficient to restore the damaged neurological function, which was partly due to the high apoptosis rate of NSCs/NPs. 6 One mechanism of EPO on the generation of mature erythrocytes is to increase cell viability by activating the PI3K signaling pathway or the Bcl-2 family. 7,8 Previous studies have suggested that EPO administration could increase the proliferation and neuronal differentiation of normal NSCs/NPs. 9 Conversely, our results indicated that instead of cell proliferation and neuronal differentiation, EPO could promote cell viability and differentiation toward oligodendrocytes and astrocytes of NSCs/NPs following OGD/R, which is partly consistent with previous results of Hassouna that EPO may be able to increase neurogenesis without entering the cell cycle. 10 The white matter of the CNS comprises myelin and is derived from oligodendrocytes. 11 Hypoxia produces a rapid and significant loss of axons in both the acute and subacute periods. 12 Spontaneous axonal regeneration is fundamental but inadequate for restoring function. Therefore, our results show that EPO treatment immediately after OGD/R cannot directly promote the differentiation of NSCs/NPs into neurons but can reduce the axonal damage of neurons by promoting the differentiation of oligodendrocytes and astrocytes. Our data showed that EPO treatment immediately after OGD/R was not a good time point F I G U R E 3 EPO treatment increases the differentiation of neural stem and progenitor cells toward oligodendrocytes and astrocytes at 7 day after OGD/R. (A) DCX% was detected by flow cytometry 1 day after 2 h of OGD treatment. OGD denotes cells were not given EPO treatment after OGD. EPO (immediately) denotes cells were given EPO (50 U/ml) immediately after OGD treatment. EPO (1 h) denotes cells were given EPO (50 U/ml) 1 h after OGD treatment. n = 6/group. *p < 0.05 vs. the OGD group. Gene expression of neuronal markers MAP-2 (B) and βIII-tubulin (C), the oligodendrocyte markers MBP (D) and CNPase (E), astrocyte markers GFAP (F) and S100β (G) in different groups were determined by RT-PCR. Control denotes cells without OGD treatment. OGD 1, 2, 4, 6 and 8 h denote cells were treated with 1, 2, 4, 6 and 8 h OGD. OGD 1, 2, 4, 6 and 8 h + EPO denote cells were given 50 U/ml EPO after OGD operation for 1, 2, 4, 6 and 8 h. n = 6/group. *p < 0.05 vs. the control group for neurogenesis, further demonstrating that the effects of EPO are dose-and time-dependent in the hypoxic-injured brain. EPOR/βCR exists in primary human renal epithelial cells, endothelial progenitor cells, and macrophages. [13][14][15] At least three versions of the EPO receptor are now known to exist within the brain, including the βCR. 16 Although the heteromeric EPO receptor involving the βCR was initially hypothesized to confer neuroprotection, evidence now supports a role for this heteromeric EPO receptor and the homodimeric EPO receptor in the protection of neurons and glia. 16 The nonpeptidyl compound STS-E412, a type of EPO that selectively binds to the EPOR/βCR heterodimer, could act as a neuroprotective agent in the CNS, suggesting the neuroprotective role of the EPOR/βCR heterodimer. 13 Our results demonstrated for the first time that EPOR/βCR heterodimers also exist on primary NSCs/NPs. Moreover, EPO treatment increased the mRNA level of βCR and the correlation between EPOR and βCR. In contrast to our current findings, it was previously shown that EPOR expression is induced by hypoxia and EPO, and its distribution corresponds to that of EPO, suggesting that brain EPO works in a paracrine/autocrine manner in response to hypoxia. Our data suggest that EPO could promote the formation of the EPOR/βCR heterodimer, and the effect of EPO on NSCs/NPs was mediated by the EPOR/βCR heterodimer.
Additionally, βCR was elevated by EPO treatment, whereas the differentiation toward neurons was not altered by EPO, demonstrating that there may be ligands other than EPO that mediate the effect of βCR on neural differentiation, which needs further studies.
We also performed a correlation analysis to further identify the possible association between EPOR/βCR and neural cell differentiation. EPOR was demonstrated to be associated with the oligodendrocyte markers MBP and CNPase and the astrocyte marker GFAP, suggesting that it might be involved in oligodendrocyte and astrocyte differentiation. βCR was correlated with the neuronal markers β-tubulin and MAP-2, suggesting that βCR is possibly involved in neuronal differentiation. EPO treatment immediately after OGD/R boosted the differentiation of NSCs/NPs toward oligodendrocytes and astrocytes. Therefore, EPOR/βCR heterodimer expressions were correlated with oligodendrocyte marker expression.
A better understanding of the progression of NSCs/NPs in the developing cerebral cortex is important for modeling neurogenesis.
To further identify the downstream signaling of the EPOR/βCR heterodimer, we analyzed the protein bands isolated by Co-IP with EPOR antibody and found several types of protein fragments that were likely F I G U R E 4 EPOR/βCR heterodimer exists on primary neural stem and progenitor cells. (A) Confocal microscopy was used to determine the colocalization of EPOR (green) and βCR (red) on neural stem and progenitor cells. Nuclei were stained with DAPI (blue). Scale bar: 5 μm. (B) Cellular lysates immunoprecipitated (IP) with anti-EPOR antibody (Co-IP) or mouse lgGb2 (lgG) and cell lysates without IP treatment (input) were immunoprobed with anti-βCR antibody. (C, D) Gene expression of the EPOR and βCR in different groups of cells was measured by RT-PCR. Control denotes cells without OGD treatment. OGD 1, 2, 4, 6 and 8 h denote cells were treated with 1, 2, 4, 6 and 8 h OGD. OGD 1, 2, 4, 6 and 8 h + EPO denote cells were given 50 U/ml EPO after OGD operation for 1, 2, 4, 6 and 8 h. n = 6/group. *p < 0.05 vs. the control group. (E, F) Correlation between EPOR and βCR after OGD or plus EPO treatment to be involved in the effect of EPO on NSCs/NPs, among which Syne-1 (nesprins) might be related to oligodendrocyte differentiation. 3 Syne-1 is a family of multi-isomeric scaffolding proteins that form the linker of nucleoskeleton-and-cytoskeleton with SUN (Sad1p/UNC84) proteins at the nuclear envelope. Syne-1 is extensively expressed in the adult murine CNS and regulates axon termination and synapse formation during neurodevelopment in the nervous system. 17,18 The N-terminal cytosolic domain of Syne could act as a connection with the C-terminal

F I G U R E 5 Correlation between EPOR/βCR and neural markers in primary neural stem and progenitor cells. (A) Correlation between EPOR and neuronal markers MAP-2 and βIII-tubulin. (B) Correlation between EPOR and oligodendrocyte markers MBP and CNPase. (C)
Correlation between EPOR and astrocyte markers GFAP and S100β. (D) Correlation between βCR and neuronal markers MAP-2 and βIIItubulin. (E) Correlation between βCR and oligodendrocyte markers MBP and CNPase. (F) Correlation between βCR and astrocyte markers GFAP and S100β  In addition, H3K9 demethylation is related to cell differentiation. 5 We found that the expression of H3K9me3 in the NSCs/NPs increased after OGD treatment, whereas EPO treatment decreased the level of H3K9me3, which might be associated with the effect of EPO on the oligodendrocyte differentiation of NSCs/NPs.
The current study has some limitations. A recent study showed that cell-cell interactions may exert neuroprotective effects on ischemic stroke in in vitro experiments. [20][21][22] Moreover, some studies have strongly suggested that glial cells, such as microglia, protect against neuronal cell death in cultures. 23,24 Our previous study investigated the neuroprotective effects of EPO by shifting microglial polarization and inhibiting excessive gliogenesis after cerebral ischemia in mice. 25 Certainly, this study also missed opportunities to explore the crosstalk between microglia and NSCs/ NPs because only pure NSCs/NPs were used in the current experiments. Therefore, more stereoscopic studies should be conducted in future studies.
In summary, our results demonstrated that EPO treatment immediately after OGD/R could increase cell viability and differentiation toward oligodendrocytes of primary fetal mouse cortical NSCs/NPs.
The EPOR/βCR heterodimer and its downstream signaling pathway, Syne-1/H3K9, were possibly involved in the effect of EPO on the oligodendrocyte differentiation of NSCs/NPs.

AUTH O R CO NTR I B UTI O N S
Z.Y., S.Z., H.Z., F.L., and Z.T. conducted the study design, experiments, data analysis, and manuscript preparation. Y.L. and R.W. designed and managed the study.

ACK N OWLED G EM ENTS
This work was supported by the National Natural Science Foundation of China (81801149, 81471340, 81771412, and 81971222).

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.