GDF11 Regulates PC12 Neural Stem Cells via ALK5-Dependent PI3K-Akt Signaling Pathway

Growth differentiation factor 11 (GDF11), belonging to the transforming factor-β superfamily, regulates anterior-posterior patterning and inhibits neurogenesis during embryonic development. However, recent studies recognized GDF11 as a rejuvenating (or anti-ageing) factor to reverse age-related cardiac hypertrophy, repair injured skeletal muscle, promote cognitive function, etc. The effects of GDF11 are contradictory and the mechanism of action is still not well clarified. The objective of the present study was to investigate effects of GDF11 on PC12 neural stem cells in vitro and to reveal the underlying mechanism. We systematically assessed the effects of GDF11 on the life activities of PC12 cells. GDF11 significantly suppressed cell proliferation and migration, promoted differentiation and apoptosis, and arrested cell cycle at G2/M phase. Both TMT-based proteomic analysis and phospho-antibody microarray revealed PI3K-Akt pathway was enriched when treated with GDF11. Inhibition of ALK5 or PI3K obviously attenuated the effects of GDF11 on PC12 neural stem cells, which exerted that GDF11 regulated neural stem cells through ALK5-dependent PI3K-Akt signaling pathway. In summary, these results demonstrated GDF11 could be a negative regulator for neurogenesis via ALK5 activating PI3K-Akt pathway when it directly acted on neural stem cells.


Introduction
Around the world, advances in medicine technology and public health measures and improvements in living standards have greatly prolonged human life span [1]. With the number of aged persons precipitously increasing, more and more people suffer from age-related cognitive decline [2,3]. Age-related cognitive decline is a common degenerative disease of the central nervous system, of which Alzheimer's disease (AD) is the most well-known form. In the United States, an estimated 6.2 million people over 65 years old are living with AD, which makes AD presently the 6th leading cause of death [4]. There are approximately 10 million AD patients in China, and AD is the fifth leading cause of death among urban and rural residents [5,6]. Age-related cognitive decline, with the features of long disease duration, slow progress, and complex etiology, remains incurable, and there are even a few drugs that can prevent or delay its progress. AD that brings heavy burdens on the patients, their families and society, has become an important social issue. A looming question is to develop new strategies for treating, preventing, or delaying age-related cognitive decline.
Recent studies suggested that the blood of young mice could fight against ageing of old mice. They found young blood were with effects of reversing age-related cardiac

GDF11 Mitigated Cell Viability and Proliferation
We first analyzed the effects of GDF11 on cell viability of PC12 cells by enhanced Cell Counting Kit-8 (CCK-8) assay. PC12 cells were cultured using starved medium with various concentrations of GDF11 (0, 12.5, 25, 50 and 100 ng/mL) for a 72 h period. GDF11 with 25 and 100 ng/mL slightly yet significantly increased (less than 10%, p < 0.05) cell viability after 24 h cultivation. Meanwhile, other concentrations of GDF11 showed no effects on cell viability. When treated for 48 h, GDF11 did not affect the cell viability. Only 100 ng/mL GDF11 a small magnitude reduction on cell viability (less than 10%, p < 0.05) when assessed at 72 h ( Figure 1A).
Results of xCELLigence RTCA indicated both GDF11-and vehicle-treated cells significantly proliferated after 120 h cultivation ( Figure 1B). At approximately the 50th h, compared to the control group, 12.5 ng/mL GDF11 showed no effect on PC12 cell proliferation, whereas 25, 50, and 100 ng/mL GDF11 significantly suppressed cell growth (p < 0.01). From the 70th h, the addition of GDF11 resulted in obvious suppression of proliferation (p < 0.05) at any of the time points compared with the control. Of note, there was a slight dose-dependent suppressing effect of GDF11 on cell proliferation.
We performed the assay of cell population doubling times to investigate the long-term effect of GDF11 on the proliferation of neural stem cells. PC12 cells, with or without GDF11 treatment, all displayed significant proliferative effects for the five-passage duration ( Figure 1C). GDF11 showed no substantial effect on PC12 proliferation at P1 and P2, though 50 ng/mL GDF11 marginally decreased proliferation (p < 0.05). From P3, higher concentrations of GDF11 (50 and 100 ng/mL) significantly inhibited PC12 cell proliferation (p < 0.05), whereas low concentrations of GDF11 did still not affect the cell growth yet. At P5, the accumulative population doubling times were 29.47 (control), 26.93 (12.5 ng/mL), 26.02 (25 ng/mL; p < 0.05 compared with control), 22.92 (50 ng/mL; p < 0.05 compared with control), and 22.04 (100 ng/mL; p < 0.05 compared with control), respectively. That meant 100 ng/mL GDF11 resulted in the lowest accumulative population doubling times (~25.2% lower than control; p < 0.05) after a five-passage cultivation.
Taken together, GDF11 showed limited effects on the viability of PC12 cells during a 48 h treatment, whereas higher concentrations of GDF11 significantly attenuated cell proliferation in a long-term cultivation. The viability of PC12 cells treated with various c trations of GDF11 or vehicle for 24 h, 48 h, or 72 h was measured by CCK-8 assay. * p < 0.05. ( viability was real-time monitored by a real time cell analyzer (xCELLigence RTCA TP) for a period. * p < 0.05, ** p < 0.01. (C) Accumulative population doubling times (PDTs) of PC1 treated with different GDF11 concentrations for a total period of 5 passages. * p < 0.05. (D) P levels of nestin, βIII-tubulin and GFAP were altered when PC12 cells were treated with v concentrations of GDF11 (0, 12.5, 25, 50 and 100 ng/mL) for 72 h. (E) Quantitative analyses of p expressions of (D) in relation to β-actin expression. * p < 0.05, *** p < 0.001. (F) Transcriptiona of nestin, βIII-tubulin and GFAP when PC12 cells were incubated with or without GDF11 fo p < 0.05, ** p < 0.01, *** p < 0.001.
Results of xCELLigence RTCA indicated both GDF11-and vehicle-treated cel nificantly proliferated after 120 h cultivation ( Figure 1B). At approximately the 5 compared to the control group, 12.5 ng/mL GDF11 showed no effect on PC12 cell eration, whereas 25, 50, and 100 ng/mL GDF11 significantly suppressed cell grow 0.01). From the 70th h, the addition of GDF11 resulted in obvious suppression of pr ation (p < 0.05) at any of the time points compared with the control. Of note, there slight dose-dependent suppressing effect of GDF11 on cell proliferation.
We performed the assay of cell population doubling times to investigate the term effect of GDF11 on the proliferation of neural stem cells. PC12 cells, with or w GDF11 treatment, all displayed significant proliferative effects for the five-passage tion ( Figure 1C). GDF11 showed no substantial effect on PC12 proliferation at P1 a though 50 ng/mL GDF11 marginally decreased proliferation (p < 0.05). From P3, (B) Cell viability was real-time monitored by a real time cell analyzer (xCELLigence RTCA TP) for a 120 h-period. * p < 0.05, ** p < 0.01. (C) Accumulative population doubling times (PDTs) of PC12 cells treated with different GDF11 concentrations for a total period of 5 passages. * p < 0.05. (D) Protein levels of nestin, βIII-tubulin and GFAP were altered when PC12 cells were treated with various concentrations of GDF11 (0, 12.5, 25, 50 and 100 ng/mL) for 72 h. (E) Quantitative analyses of protein expressions of (D) in relation to β-actin expression. * p < 0.05, *** p < 0.001. (F) Transcriptional levels of nestin, βIII-tubulin and GFAP when PC12 cells were incubated with or without GDF11 for 5 h. * p < 0.05, ** p < 0.01, *** p < 0.001.

GDF11 Promoted Differentiation of PC12 Cells
As shown in Figure 1D,E, western bolt results revealed GDF11 slightly decreased protein levels of the neural progenitor cell marker, nestin. Conversely, protein levels of βIII-tubulin (neuronal biomarker) and GFAP (astrocytic biomarker) of the GDF11-treated groups were increased when compared to the control. Similar with protein levels, the mRNA expression of nestin declined and that of βIII-tubulin and GFAP increased when exposed to GDF11 ( Figure 1F, File S1). These results suggested that GDF11 promoted neural stem cells to neuronal and astrocytic differentiation.

GDF11 Stimulated Apoptosis of PC12 Cells
The flow cytometery results of Annexin V-FITC/PI dual staining illuminated that the apoptotic and necrotic cells following GDF11 treatment were significantly increased (Figure 2A,B). Only 14.4% of PC12 cell were apoptotic in the control group, whereas 17.8%, 23.5% (p < 0.05), 27.2% (p < 0.05) and 30.5% (p < 0.05) cells displayed apoptosis when treated with 12.5, 25, 50 and 100 ng/mL GDF11 for 72 h, respectively. Compared to the control, exposure to 100 ng/mL GDF11 stimulated approximately double apoptosis cells. Furthermore, GDF11 showed a slight dose-dependent effect to induce apoptosis. βIII-tubulin (neuronal biomarker) and GFAP (astrocytic biomarker) of the GDF11-treated groups were increased when compared to the control. Similar with protein levels, the mRNA expression of nestin declined and that of βIII-tubulin and GFAP increased when exposed to GDF11 ( Figure 1F, File S1). These results suggested that GDF11 promoted neural stem cells to neuronal and astrocytic differentiation.

GDF11 Stimulated Apoptosis of PC12 Cells
The flow cytometery results of Annexin V-FITC/PI dual staining illuminated that the apoptotic and necrotic cells following GDF11 treatment were significantly increased (Figure 2A,B). Only 14.4% of PC12 cell were apoptotic in the control group, whereas 17.8%, 23.5% (p < 0.05), 27.2% (p < 0.05) and 30.5% (p < 0.05) cells displayed apoptosis when treated with 12.5, 25, 50 and 100 ng/mL GDF11 for 72 h, respectively. Compared to the control, exposure to 100 ng/mL GDF11 stimulated approximately double apoptosis cells. Furthermore, GDF11 showed a slight dose-dependent effect to induce apoptosis.
We also evaluated transcriptional levels of the key proteins of cell cycle (i.e., cyclin D1, cyclin D2, cyclin A, and cyclin B1). The results of qRT−PCR revealed that GDF11 hardly affected the mRNA levels of cyclin D1 and cyclin A, whereas GDF11 significantly decreased transcription of cyclin D2 and cyclin B1 ( Figure 3A). These might, in part, be the underlying mechanism of GDF11-induced cell cycle G2/M arrest. incubated with GDF11 (0, 12.5, 25, 50 and 100 ng/mL) or vehicle for 72 h and analyzed by Flow cytometry. (D) Histogram of cell cycle distribution as in (C). * p < 0.05, ** p < 0.01.
We also evaluated transcriptional levels of the key proteins of cell cycle (i.e., cyclin D1, cyclin D2, cyclin A, and cyclin B1). The results of qRT−PCR revealed that GDF11 hardly affected the mRNA levels of cyclin D1 and cyclin A, whereas GDF11 significantly decreased transcription of cyclin D2 and cyclin B1 ( Figure 3A). These might, in part, be the underlying mechanism of GDF11-induced cell cycle G2/M arrest.

TMT-Based Protein Identification and Quantification
The SDS-PAGE results of the control and 100 ng/mL GDF11-treated group showed good consistency of the samples in each group ( Figure S1A). The results of the MS data analysis (including the distributions of peptide length and protein mass) indicated that the sample preparation procedures matched the requirement and were reliable ( Figure S1G,H). A total of 5 209 proteins were quantifiable (Table 1), and the distribution information of the identified proteins (including isoelectric point, molecular weight, and peptide coverage) were shown in Figure S1. Results showed that 165 proteins were differentially expressed, of which 90 were up-regulated and 65 were down-regulated in GDF11-treated group when compared with control ( Figure 4A,B and Table S2).

Annotation Analysis of the Differentially Expressed Proteins
GO enrichment analysis was performed to analyze the biological significance of the aforementioned 165 differentially expressed proteins (DEPs). The detailed GO annotation of the DEPs was shown in Figure 4D and Table S3. Briefly, the analysis of biological process revealed the DEPs were mainly enriched in phosphatidylinositol phosphorylation,

Annotation Analysis of the Differentially Expressed Proteins
GO enrichment analysis was performed to analyze the biological significance of the aforementioned 165 differentially expressed proteins (DEPs). The detailed GO annotation of the DEPs was shown in Figure 4D and Table S3. Briefly, the analysis of biological process revealed the DEPs were mainly enriched in phosphatidylinositol phosphorylation, cell cycle arrest, phosphatidylinositol-mediated signaling, glycerophospholipid metabolic process, and so on. And for the molecular function analysis, the DEPs were mainly involved in transcription cofactor activity, transcription co-activator activity, and protein kinase C activity.
To further understand the functional differences of the dysregulated proteins, KEGG pathway analysis was implemented. The top 20 enriched KEGG pathways were displayed in Figure 4C, including the pathways of Ras signaling, PI3K-Akt signaling, HIF-1 signaling, phosphatidylinositol signaling system, and sphingolipid metabolism. From the GO and KEGG enrichment analysis, we found the DEPs were closely involved in phosphatidylinositol signaling system, including phosphatidylinositol 3-kinase (PI3K)-Akt pathway, which implied PI3K-Akt pathway was participated in the modulating effects of GDF11 on PC12 cells.
Protein-protein interaction (PPI) network analysis via STRING database and Cytoscape software obviously revealed that at least three crucial cross-talk signaling clusters were displayed in the complex PPI network, and the key nodes were Pik3ca (coding for the p110α catalytic subunit of PI3K), Trp53bp1, and Cep290 ( Figure 5A). After simplifying the PPI network, Pik3ca was the visible key node ( Figure 5B), which further suggested PI3K-Akt was closely associated with GDF11 effects on PC12 cells. dylinositol signaling system, including phosphatidylinositol 3-kinase (PI3K)-Akt pathway, which implied PI3K-Akt pathway was participated in the modulating effects of GDF11 on PC12 cells. Protein-protein interaction (PPI) network analysis via STRING database and Cytoscape software obviously revealed that at least three crucial cross-talk signaling clusters were displayed in the complex PPI network, and the key nodes were Pik3ca (coding for the p110α catalytic subunit of PI3K), Trp53bp1, and Cep290 ( Figure 5A). After simplifying the PPI network, Pik3ca was the visible key node ( Figure 5B), which further suggested PI3K-Akt was closely associated with GDF11 effects on PC12 cells.

PI3K-Akt Pathway Was Enriched by CSP100 PLUS Microarray
Detailed information of the 304 unphosphorylated proteins and phosphorylated sites in the CSP100 PLUS microarray was displayed in Figure S2A,B and Table S4. Treatment with 100 ng/mL GDF11 altered 42 unphosphorylated protein expressions, of which 19 down-regulated and 23 up-regulated ( Figure S2C). Compared to the control, the phos-

PI3K-Akt Pathway Was Enriched by CSP100 PLUS Microarray
Detailed information of the 304 unphosphorylated proteins and phosphorylated sites in the CSP100 PLUS microarray was displayed in Figure S2A,B and Table S4. Treatment with 100 ng/mL GDF11 altered 42 unphosphorylated protein expressions, of which 19 downregulated and 23 up-regulated ( Figure S2C). Compared to the control, the phosphorylation levels of 25 proteins were up-(21/25) or down-regulated (4/25) in the GDF11-treated group ( Figure 6A). The significant up-regulation of phosphorylation of PI3K, GSK3, p70S6K, HSP27, Smad2, and Smad3 were the key proteins in PI3K-Akt and TGFβ-Smad2/3 signal pathways. We performed an integrated analysis of the 16 signaling pathway of the microarray to explore the core signal-net ( Figures S3A and 6B), and it illuminated PI3K and Akt2 were the key proteins involved in the core signal-net. Further analysis of pathway cross-talk revealed PI3K-Akt pathway was closely related with cell cycle, apoptosis and AMPK pathway after GDF11 treatment ( Figure S3B). And PI3K-Akt was the fundamental signaling pathway involved in the effects of GDF11 on neural stem cells ( Figure 6C).

GDF11 Activated PI3K-Akt Signal Pathway
Because both TMT-based proteomic analysis and CSP100 PLUS microarray indicated PI3K-Akt was significantly altered when PC12 cells were treated with GDF11, we investigated PI3K-Akt pathway to reveal the underlying mechanism of the effects of GDF11 on PC12 cells. The results of western blot displayed 100 ng/mL GDF11 treatment significantly increased the phosphorylation levels of both PI3K and Akt, when compared with the con- We performed an integrated analysis of the 16 signaling pathway of the microarray to explore the core signal-net ( Figure S3A and Figure 6B), and it illuminated PI3K and Akt2 were the key proteins involved in the core signal-net. Further analysis of pathway cross-talk revealed PI3K-Akt pathway was closely related with cell cycle, apoptosis and AMPK pathway after GDF11 treatment ( Figure S3B). And PI3K-Akt was the fundamental signaling pathway involved in the effects of GDF11 on neural stem cells ( Figure 6C).

GDF11 Activated PI3K-Akt Signal Pathway
Because both TMT-based proteomic analysis and CSP100 PLUS microarray indicated PI3K-Akt was significantly altered when PC12 cells were treated with GDF11, we investigated PI3K-Akt pathway to reveal the underlying mechanism of the effects of GDF11 on PC12 cells. The results of western blot displayed 100 ng/mL GDF11 treatment significantly increased the phosphorylation levels of both PI3K and Akt, when compared with the control (Figure 7A,B, File S1). These results showed GDF11 activated PI3K-Akt signaling pathway in neural stem cells and were consistent with the results of the TMT-based proteomic analysis and CSP100 PLUS microarray.

GDF11 Affected Neural Stem Cells through ALK5-Dependent PI3K-Akt Pathway
Given that members of TGF-β family bind to TGF-β receptors to form complex which phosphorylates the downstream proteins to regulate cellular behaviors. We then measured whether GDF11-induced PI3K-Akt pathway depended on its binding to the TGF-β type 1 receptor (TGFβRI or ALK5), as it was previously suggested [22]. In the present study, we first inhibited ALK5 with SB431542, a specific ALK5 inhibitors. We found SB431542 showed no cytotoxicity on PC12 cells ( Figure 7C). Results of western blot displayed that inhibiting ALK5 by SB431542 strongly diminished phosphorylation levels of Smad2/3, PI3K and Akt ( Figure 7D,E, File S1). Furthermore, blocking ALK5 abolished the effects of GDF11 on PC12 cells. Evaluating in a detailed manner, the effects of GDF11induced suppression of proliferation and migration ( Figure 8A-C), promotion of differentiation and apoptosis ( Figures 8D,E and 7F-H), and arrest of cell cycle ( Figure 8F) were reversed by inhibiting ALK5 with SB431542. These results demonstrated that GDF11 triggered downstream signaling pathways via an ALK5-dependent manner to regulate the (G,H) were Quantitative analysis of (F). * p < 0.05, ** p < 0.01, *** p < 0.001.

GDF11 Affected Neural Stem Cells through ALK5-Dependent PI3K-Akt Pathway
Given that members of TGF-β family bind to TGF-β receptors to form complex which phosphorylates the downstream proteins to regulate cellular behaviors. We then measured whether GDF11-induced PI3K-Akt pathway depended on its binding to the TGF-β type 1 receptor (TGFβRI or ALK5), as it was previously suggested [22]. In the present study, we first inhibited ALK5 with SB431542, a specific ALK5 inhibitors. We found SB431542 showed no cytotoxicity on PC12 cells ( Figure 7C). Results of western blot displayed that inhibiting ALK5 by SB431542 strongly diminished phosphorylation levels of Smad2/3, PI3K and Akt ( Figure 7D,E, File S1). Furthermore, blocking ALK5 abolished the effects of GDF11 on PC12 cells. Evaluating in a detailed manner, the effects of GDF11-induced suppression of proliferation and migration ( Figure 8A-C), promotion of differentiation and apoptosis ( Figure 8D,E and Figure 7F-H), and arrest of cell cycle ( Figure 8F) were reversed by inhibiting ALK5 with SB431542. These results demonstrated that GDF11 triggered downstream signaling pathways via an ALK5-dependent manner to regulate the life activities of neural stem cells. Subsequently, we explored the effects of GDF11-induced PI3K-Akt pathway on neural stem cells by blocking PI3K with a PI3K inhibitor LY294002. Treatment with LY294002 decreased the GDF11-induced Akt phosphorylation, without affecting the phosphorylation levels of Smad2/3 ( Figure 7D,E, File S1). These indicated that co-treatment with GDF11 and LY294002 inhibited the PI3K-Akt pathway, while it activated TGF-β/Smad2/3 pathway. Moreover, LY294002 attenuated GDF11-mediated inhibition of proliferation and migration ( Figure 8A-C), increase of differentiation and apoptosis ( Figure 8D-E and Figure 7F-H), and cell cycle arrest ( Figure 8F), which suggested PI3K-Akt signaling pathway was crucial for the effects of GDF11 on neural stem cells.
Taken together, GDF11 regulated cellular behaviors of neural stem cells via ALK5dependent PI3K-Akt signaling pathway.

Discussion
In this study, we demonstrated, in vitro, GDF11 suppressed cell proliferation and Subsequently, we explored the effects of GDF11-induced PI3K-Akt pathway on neural stem cells by blocking PI3K with a PI3K inhibitor LY294002. Treatment with LY294002 decreased the GDF11-induced Akt phosphorylation, without affecting the phosphorylation levels of Smad2/3 ( Figure 7D,E, File S1). These indicated that co-treatment with GDF11 and LY294002 inhibited the PI3K-Akt pathway, while it activated TGF-β/Smad2/3 pathway. Moreover, LY294002 attenuated GDF11-mediated inhibition of proliferation and migration ( Figure 8A-C), increase of differentiation and apoptosis ( Figure 8D-E and Figure 7F-H), and cell cycle arrest ( Figure 8F), which suggested PI3K-Akt signaling pathway was crucial for the effects of GDF11 on neural stem cells.
Taken together, GDF11 regulated cellular behaviors of neural stem cells via ALK5dependent PI3K-Akt signaling pathway.

Discussion
In this study, we demonstrated, in vitro, GDF11 suppressed cell proliferation and migration, promoted differentiation and apoptosis, and induced cell cycle arrest in the G2/M phase. And the effects of GDF11 were diminished when blocking ALK5 by SB431542 or inhibiting PI3K by LY294002. GDF11 regulated the life activities of neural stem cells via ALK5-dependent activation of downstream PI3K-Akt signaling pathway.
At present, our understanding on the function of GDF11 in ageing and neurogenesis was still quite limited and controversial. GDF11, as a member of TGF-β superfamily, modulates anterior-posterior patterning during mammalian embryonic development [15,23]. In the initial stages, GDF11 was recognized as a negative regulator of neurogenesis [16], which was consistent with other studies [21,24], and our results showing that GDF11 inhibited PC12 cell proliferation and promoted differentiation. However, GDF11 attracted much attention for its rejuvenating effect on neurogenesis in recent years [8,25,26], which was contradictory with the negative regulation on neurogenesis. The recent study by Mayweather et al. [21] explained this phenomenon. They revealed endogenous GDF11 in brain suppressed hippocampal neurogenesis through directly contacting with neural stem cells. In contrast, the circulatory GDF11 in blood, which was recognized as a pro-youthful factor to increases hippocampal neurogenesis, was perhaps to play indirect roles on neural stem cells without crossing the blood-brain barrier. In our present study, we used the in vitro model to investigate the effects of GDF11 on neural stem cells. GDF11 in the medium acted directly on PC12 cells, just like the manner of endogenous GDF11 directly acting on hippocampal neural stem cells. Therefore, our findings of the effects of GDF11 on suppression of proliferation and promotion of differentiation are reliable. It appears that GDF11 is a "pro-ageing factor" when it directly acts on neural stem cells, whereas GDF11 is a "rejuvenating factor" when it indirectly acts on neural stem cells. However, it is still unclear whether the effects of endogenous GDF11 and circulatory GDF11 on other cells (e.g., skeletal muscle cells and cardiomyocytes) are the same, and this requires further investigation. It also raises an interesting suggestion for further studies that GDF11 or other proteins may have different (or even totally opposite) effects on the same cell, depending on the direct or indirect action. Furthermore, we found GDF11 not only promoted apoptosis but also suppressed migration and cell cycle of neural stem cells. In line with our results, both Wu et al. [16] and Williams et al. [24] showed that GDF11 arrested cell cycle of neural stem cells. And Williams et al. [24] also supported our results that GDF11 diminished migration of neural stem cells in vitro. However, Katsimpardi et al. [8] found the circulatory GDF11 in blood promote migration of neural cells, which was in contrast with our in vitro results. It should be noted that circulatory GDF11 showed indirect effects on neural stem cells whereas GDF11 directly acted on neural stem cells in our present study. Therefore, the results of Katsimpardi et al. [8] cannot be simply compared with our results.
Given that cell signaling interaction is not a simple linear regulation, and it is impossible that GDF11 regulates neural stem cells solely through PI3k-Akt pathway, other factors or pathways may also be involved in GDF11 regulatory network. Seong and Kang [34] revealed that miR-1260b directly targeted the 3 UTR of GDF11 to mediate GDF11-Smaddependent signaling to regulate proliferation of vascular smooth muscle cell. circUCK2 inhibited miR-125b-5p activity to increase GDF11 expression, resulting in amelioration of neuronal injury in mice, which suggested that the circUCK2/miR-125b-5p/GDF11 axis was an essential signaling pathway involved in ischemia stroke [35]. GDF11 significantly attenuated the severity of skin inflammation by suppressing NF-κB signaling pathway in both imiquimod-induced mice model and IL-23-induced mice model [36]. As described above, GDF11 regulates cell behaviors through complex signaling regulatory networks. More studies should be performed to further refine the GDF11-mediated regulatory network.

Cell Viability/Proliferation Assay
Cell viability was evaluated by CCK-8 (Beyotime Biotechnology, China) at time points of 24, 48, and 72 h. PC12 cells were cultured in 96-well plates at a density of 3000 cells/well in 200 µL complete medium overnight. Thereafter, complete medium was replaced with starved medium for 6 h to synchronize growth. Various concentrations of GDF11 were supplemented. According to the manufacturer's instructions, 20 µL of CCK-8 agent was added to each well 2 h before the termination of the experiment. The absorbance was detected at 450 nm using a SpectraMax M2 e (Molecular Devices, San Jose, CA, USA). Then cell viability was calculated by comparing the optical density (OD) values of GDF11-treated with untreated cells.
Proliferation of PC12 cells was also measured by a real time cell analyzer (xCELLigence RTCA TP, ACEA Biosciences, San Diego, CA, USA), which is an impedance-based technology that can be used for real-time monitoring and noninvasively quantifying cell proliferation. PC12 cells were cultured in an E-16 plate at a density of 2000 cells/well, and then the plate was inserted into the xCELLigence station at 37 • C with 5% CO 2 for 120 h. Data were collected every 10 min, and then the proliferation curve was drawn.
Cell population doubling times (PDTs) were performed to assay cell proliferation as well. PC12 cells were seeded in a 96-well plate at a density of 3000 cells/well. PC12 cells were passaged every 48 h, and cells collected at the end of each passage were counted using an automatic image cytometer (Cellometer K2; Nexcelom Bioscience, Lawrence, MA, USA). Cell population doubling times was calculated using the following formula: PDTs = Pn/P0, where P0 and Pn represent the cell number at the initial seeding passage (0) and at a specific passage (n), respectively.

Wound Healing/Migration Assay
For wound healing/migration assay, Culture-Insert 4 Well in µ-Dish 35 mm (Ibidi GMBH, Germany) was used. PC12 cells were employed to each well of the Culture-Insert at a density of 5000 cells/well with complete medium. When reaching 70-80% confluence, the medium was replaced by starved medium for another 6 h. Then the Culture-Insert was removed to generate cell-free gaps (i.e., wound) with a width of 500 µm, and various concentrations of GDF11 diluted with starved medium (0 ng/mL, 12.5 ng/mL, 25 ng/mL, 50 ng/mL, and 100 ng/mL, respectively) were added. Images were captured at 0 h, 12 h and, 36 h with an inverted microscope to measure the wound closure. The percent of wound closure of various treatments at different time points were calculated with Wimasis software (Onimagin Technologies, Córdoba, Spain).

qRT-PCR Analysis
Total RNA was extracted from the cultured cells using TRIZOL reagent (Invitrogen, Waltham, MA, USA) according to the manufacturer's standard procedure. Total RNA (1 µg) was used for reverse transcription in a final volume of 20 µL. Subsequently, qRT-PCR was performed using the Quantitect SYBR Green PCR kit (Qiagen, Germantown, MD, USA) with an ABI StepOnePlus Real-Time PCR System (Applied Biosystems, Waltham, MA, USA). The relative expressions of target genes were calculated using the 2 −∆∆Ct method by normalizing with the housekeeping gene (GAPDH or β-actin) expression and presented as fold changes relative to control. The PCR primers were synthesized by Beijing Genomics Institute (Shenzhen, China), and the sequences of primers are displayed in Table S1.

Western Blot Analysis and Validation
Cells were lysed by RIPA buffer (Beyotime Biotechnology, China) supplemented with 1× protease inhibitor cocktail and 1× phosphatase inhibitor cocktail. Total protein was measured by using a BCA protein assay kit (Beyotime Biotechnology, China). Then, equal amount proteins (25 µg) of each sample were denatured with NuPAGE 4× LDS loading buffer (Invitrogen, USA), and were separated by SDS-PAGE (NuPAGE 4-12% Bis-Tris gel; Invitrogen, USA). After transferred to PVDF membranes, proteins were incubated with the following rabbit antibodies at 4 • C overnight: anti-nestin (Cell Signal Technology (Cell Signal Technology, USA), anti-p-PI3K p85 (Cell Signal Technology, USA), anti-p-Akt (Cell Signal Technology, USA), and β-actin (Cell Signal Technology, USA), respectively. Subsequently, the PVDF membranes were incubated with horseradish peroxidase-labeled goat anti-rabbit IgG (ZSGB-BIO, Beijing, China) at room temperature for 2 h, and then were visualized by ECL reagents (Foregene Co., Ltd., Chengdu, China).

Sample Preparation, TMT-Labeling, and LC-MS/MS
PC12 cells were cultured in 6-well plates with complete medium overnight. Then the medium was replaced by starved medium for another 6 h. The group of C1, C2, and C3 was treated with 100 ng/mL GDF11 supplemented with starved medium, while the controls of A1, A2, and A3 were treated with starved medium. The cells were lysed with lysis buffer (8 M urea, 1% protease inhibitor). Protein concentration was determined by a BCA kit (Beyotime Biotechnology, China), and the consistency of each group was detected by SDS-PAGE.
The detailed methodology of protein digestion, TMT-labeling and LC-MS/MS were performed as we previously described [3,37,38]. Briefly, at the beginning, proteins of each sample (100 µg) were reduced, alkylated, and digested according to the protocols. Subsequently, the digested peptides were labeled with the TMT label reagents according to the instructions of 6-plex TMT labeling kit (Thermo Fisher Scientific, Waltham, MA, USA). Sample labeling was as follows: group A1: 126, group A2: 127, group A3: 128, group C1: 129, group C2: 130, and group C3: 131.
The labeled samples were subsequently fractionated into 60 fractions by high pH reverse-phase high-performance liquid chromatography (HPLC). Then, the 60 fractions were combined into 18 fractions, and vacuum-dried. The dried samples were subsequently reconstituted and detected by LC-MS/MS on a Q Exactive Plus mass spectrometer (Thermo Scientific) that was coupled to an EASY-nLC 1000 (Thermo Fisher Scientific, USA) ultraperformance liquid chromatography (UPLC) system.

MS/MS Data Analysis and Bioinformatics Analysis
The raw MS/MS data were searched using the MASCOT engine (version 2.2) embedded into Proteome Discoverer 1.4 software with a false discovery rate (FDR) < 1% at protein, peptide, and PSM levels and minimum score for peptides > 40. The peptide mass tolerance for precursors was set as 20 ppm in the first search and 5 ppm in the main search, and the 0.02 Da was for fragment ions. The requirements of all identified proteins were with at least 2 peptides with ≥1 unique peptide, and only the proteins at p < 0.05 were considered to be accurately quantified. For comparisons between GDF11treat groups and controls, a protein with a fold change of > ±1.2 (p < 0.05) was presented as a DEP. DEPs were analyzed according to GO terms for biological process, cellular component, and molecular function using the UniProt-GOA database (http://www.ebi.ac.uk/GOA/ accessed on 12 June 2020). Domain functional descriptions of the DEPs were predicted by using InterProScan soft (http://www.ebi.ac.uk/interpro/ accessed on 12 June 2020). Pathway enrichment analysis was assessed using KEGG pathway database (http://www.genome.jp/kegg/ accessed on 13 June 2020). Protein−protein interaction (PPI) networks were analyzed by the STRING database (http://string-db.org accessed on 12 June 2020), and visualized by Cytoscape software (version 3.6.0), respectively.

Profiles of Key Signaling Pathway by an Antibody Array
The key signaling pathways induced by GDF11 in neural stem cells were profiled using a phosphorylation-profiling antibody microarray CSP100 PLUS (Full Moon Microsystems, Charlotte, MI, USA) performed by Wayen Biotechnology (China) according to the protocols. A total of 304 unphosphorylated proteins and phosphorylated sites in 16 signaling pathways were analyzed. The obtained data were first normalized by housekeeping protein, and then the ratio of protein expression and phosphorylation changes were calculated by comparing them to the control. Differentially expressed proteins or phosphorylation sites were set with a fold change > ±1.2. The differentially phosphorylated proteins were analyzed using Ingenuity Pathway Analysis (IPA; QIAGEN, Hilden, Germany) to reveal the core signal-net and pathways.

Statistical Analysis
The statistical analysis was performed using GraphPad Prism 8.0.2 and the data were presented as the mean ± standard error of the mean (SEM) otherwise specified. Hierarchical clustering heat map of the DEPs was conducted by using HemI 2.0 (https://hemi.biocuckoo. org/ accessed on 4 March 2022). Multi-group comparisons were conducted by one-way ANOVA followed by Tukey's post hoc test. Paired analysis of control and GDF11-treated groups was performed using two-tailed paired Student's t test. In all analyses, p < 0.05 was considered statistically significant.

Conclusions
Our in vitro study found that GDF11 showed substantially effects on PC12 neural stem cells, which was suppression of proliferation and migration, promotion of apoptosis and differentiation, and arrest of cell cycle. Further proteomic and microarray analysis revealed PI3K-Akt pathway was significantly enriched when PC12 cells were treated with GDF11. By blocking ALK5 or PI3K, we demonstrated that GDF11 regulated the life activities of neural stem cells via binding ALK5 to activate PI3K-Akt signaling pathway. These results implied that GDF11 could be a potential target for pharmacologic blockade, not a rejuvenative factor, when GDF11 directly acted on neural stem cells, and could develop new therapeutic strategy for neurogenesis.