MTHFD2 modulates neural stem cell proliferation and apoptosis after intracerebral hemorrhage by regulating mitochondrial NADPH homeostasis

Abstract

Background

Intracerebral hemorrhage (ICH) represents one of the most severe forms of cerebrovascular injury, characterized by high mortality and lasting neurological dysfunction. Neural stem cells (NSCs), which are indispensable for neurogenesis, are mainly distributed in the hippocampus, a brain region essential for learning and memory. However, alterations in hippocampal NSCs following ICH and the mechanisms that mediate these changes remain poorly defined. The present study investigates the effects of ICH on hippocampal NSCs, focusing on the regulatory role of MTHFD2 in maintaining mitochondrial redox homeostasis through NADPH metabolism.

Methods

Both in vivo and in vitro ICH models were used, including a collagenase-induced mouse model and a hemin-treated NSC model, to examine molecular and cellular responses of hippocampal NSCs to hemorrhagic injury. Western blotting, RNA sequencing, CUT&Tag profiling, and ChIP-qPCR were employed to analyze the related pathways.

Results

We found that NSCs exhibited a time-dependent response after ICH, showing initial activation followed by gradual functional exhaustion and increased apoptosis. ICH induced persistent endoplasmic reticulum stress and significantly upregulated MTHFD2, a mitochondrial enzyme essential for one-carbon metabolism. Mechanistically, MTHFD2 was required to maintain mitochondrial integrity and redox homeostasis by regulating NADPH levels. Knockdown of MTHFD2 reduced NSC proliferation, increased apoptosis, and worsened cognitive impairment in ICH mice. Exogenous NADPH supplementation partially restored these changes, emphasizing the importance of redox balance in NSC survival and function.

Conclusions

Our findings identify MTHFD2 as a key metabolic regulator supporting NSC adaptation to ICH through NADPH-dependent mitochondrial mechanisms. Targeting mitochondrial redox metabolism may provide a potential strategy for preserving NSC function and improving cognitive recovery after ICH.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-026-07961-1.

Introduction

Intracerebral hemorrhage (ICH) represents a severe cerebrovascular event, with a 30-day mortality rate ranging from 9.3% to 50%, and survivors frequently experience persistent cognitive dysfunction [1–4]. Approximately 70% of these survivors show varying degrees of cognitive decline, including impairments in learning, memory, and executive function [1, 5–7]. Despite the clinical significance of such deficits, most ICH research has concentrated on motor dysfunction, while the mechanisms underlying cognitive decline have received comparatively limited attention [8]. The hippocampus, a principal brain region responsible for learning and memory, contains neural stem cells (NSCs) that are indispensable for neurogenesis [9–11]. It possesses a distinctive ability for continuous neurogenesis within the dentate gyrus, encompassing the birth, maturation, and integration of new neurons into existing neural circuits [12]. This ongoing process is fundamental to learning, memory, and cognitive flexibility [12]. Given the importance of the hippocampus in these cognitive processes, it is reasonable to infer that ICH-induced injury to this region, particularly to its resident NSCs, contributes significantly to post-ICH cognitive impairments [5, 13].

Stress responses triggered by the post-ICH environment act as key drivers of NSC activation and proliferation. However, this increased metabolic state is accompanied by excessive generation of mitochondrial reactive oxygen species (mROS), making NSCs highly vulnerable to oxidative damage. Mitochondria, functioning simultaneously as the source and target of mROS, are particularly susceptible to such stress, which in turn compromises NSC survival and function. Under physiological conditions, cellular antioxidant defenses, primarily dependent on reduced glutathione (GSH) and nicotinamide adenine dinucleotide phosphate (NADPH), counteract ROS accumulation and maintain redox homeostasis [14, 15]. Notably, NADPH provides the reducing power required for GSH regeneration and the activity of the thioredoxin system. Despite its fundamental importance in redox regulation, the specific contribution of NADPH to determining NSC fate and behavior after ICH remains largely unexplored.

MTHFD2, an essential enzyme in the mitochondrial one-carbon metabolic pathway, has emerged as an important mediator of cellular responses to stress [16–18]. Traditionally, one-carbon metabolism has been considered mainly as a pathway providing one-carbon units for nucleotide synthesis and other biosynthetic processes, rather than a major source of intracellular NADPH. However, growing evidence indicates that the loss of key enzymes in this pathway, such as MTHFD2, reduces the cellular NADPH/NADP+ ratio and decreases the GSH/GSSG ratio. Recent studies have also demonstrated that MTHFD2 participates in maintaining mitochondrial function and regulating energy metabolism, both of which are fundamental for the survival and proliferation of NSCs [2, 19, 20]. MTHFD2 is known to be highly expressed in embryonic and pluripotent stem cells, where it contributes to redox balance, DNA repair, and the maintenance of self-renewal capacity [19, 20]. Despite these findings, the function of MTHFD2 in hippocampal NSCs following ICH remains unclear. Whether MTHFD2 affects NSC proliferation and apoptosis through modulation of mitochondrial redox homeostasis and NADPH availability has yet to be determined. The potential role of MTHFD2 in shaping NSC behavior and its contribution to post-ICH cognitive impairment therefore requires further investigation.

This study integrated both in vivo and in vitro approaches, including primary hippocampal NSC cultures and collagenase-induced ICH models, to investigate the alterations in NSC proliferation and apoptosis following hemorrhagic injury. Through a combination of immunofluorescence, Western blotting (WB), RNA sequencing, and CUT&Tag sequencing, we comprehensively examined the molecular mechanisms involved. Notably, although ATF4-dependent transcriptional regulation of MTHFD2 has been reported in other cellular systems, this study confirmed this regulatory pathway in hippocampal NSCs under ICH-related stress conditions. Therefore, we aimed to elucidate the role of MTHFD2 in regulating mitochondrial NADPH homeostasis, redox balance, and NSC survival after ICH. We hypothesized that MTHFD2 promotes NSC proliferation and prevents apoptosis after ICH by maintaining mitochondrial NADPH levels and redox equilibrium.

Materials and methods

Animal ICH models

C57BL/6 mice aged 8–12 weeks and weighing 20–25 g were purchased from GemPharmatech, China. All experimental procedures were conducted following the guidelines of the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University. Briefly, the animals were anesthetized with isoflurane and secured in a stereotactic device (RWD, Shenzhen, China). A midline scalp incision was made, and the injection needle was positioned 0.5 mm anterior and 2.0 mm lateral to the bregma, then advanced to a depth of 3.0 mm. A total of 0.4 µL collagenase IV (Sigma, USA) was injected at a rate of 0.4 µL/min. To prevent backflow, the needle was kept in place for one to 2 min before being withdrawn. After surgery, mice were randomly assigned to different experimental groups.

Next, the mice were sacrificed using deep anesthesia at 1, 3, and 30 days after ICH. Transcardial perfusion was performed with PBS followed by 4% paraformaldehyde (PFA). The brains were post-fixed in 4% PFA and dehydrated in 30% sucrose. Coronal Sect.  (25 μm thick) were prepared using a cryostat and stored at -20 °C until immunofluorescence (IF) analysis.

For WB analysis, the hippocampal tissues were carefully dissected and homogenized in RIPA lysis buffer containing protease and phosphatase inhibitors. The extracted proteins were collected for subsequent WB analysis.

Isolation and primary culture of hippocampal NSCs

To isolate and culture NSCs, pregnant mice at embryonic day 16 (E16) were used. Hippocampi from E16 embryos were dissected under a microscope and transferred into Petri dishes containing fresh DMEM/F12 medium. The tissues were enzymatically digested with 0.125% trypsin-EDTA at 37 °C for 10 min, after which the resulting suspension was filtered through a 40 μm cell strainer before being cultured in NSC medium. The culture medium consisted of DMEM/F12 supplemented with B27, 20 ng/mL epidermal growth factor, and 20 ng/mL basic fibroblast growth factor. Neurospheres developed after 3–5 days of incubation and were subsequently collected and dissociated into single cells using Accutase for 10 min at 37 °C, with gentle pipetting to ensure complete separation. The isolated cells were then resuspended in fresh NSC medium for further culture. For adhesion experiments, culture vessels were precoated with Poly-L-ornithine at a concentration of 10 µg/mL, and the cells were seeded onto these treated surfaces to promote attachment and proliferation under adherent conditions.

Immunofluorescence assay

Cells were fixed with 4% PFA for 15 min at room temperature, followed by permeabilization with PBS containing 0.3% Triton X-100 for another 15 min. After permeabilization, the cells were blocked with 10% bovine serum albumin in PBS for 1 h at room temperature to minimize non-specific antibody binding. Primary antibodies were incubated overnight at 4 °C in blocking buffer, followed by three washes with PBS. Secondary antibodies were then applied for 1–2 h at 37 °C, after which the samples were washed again with PBS. Coverslips were mounted on slides using antifade mounting medium, and fluorescence images were acquired using a fluorescence microscope. Detailed information regarding the antibodies used is provided in Tables S1 and S2.

TUNEL staining

To assess NSC apoptosis, the TUNEL assay (Beyotime, China) was performed according to the manufacturer’s instructions. NSCs were first fixed with 4% paraformaldehyde for 15 min at room temperature and then permeabilized with 0.3% Triton X-100 for an additional 15 min. Following fixation and permeabilization, the samples were thoroughly washed with PBS before TUNEL staining. The specimens were examined using fluorescence microscopy, and high-resolution images were captured for subsequent analysis.

EdU assay

To evaluate cell proliferation in vivo, EdU was dissolved in PBS and administered intraperitoneally at 50 mg/kg, 24 h prior to analysis. For in vitro labeling of NSC proliferation, EdU was added to the culture medium at a final concentration of 10 µM and incubated for 2 h. To assess EdU incorporation, frozen brain sections or cultured cells were fixed with 4% PFA, permeabilized with 0.5% Triton X-100, and stained following the instructions provided in the EdU Cell Proliferation Kit (Beyotime, China). The nuclei were counterstained with DAPI, and fluorescence images were acquired using a fluorescence microscope. The proportion of EdU-positive cells was quantified using ImageJ software.

Western blot

Protein extraction from NSCs or mouse brain tissues was performed and quantified using the BCA kit. Equal amounts of protein were separated by SDS-PAGE and transferred onto PVDF membranes. The membranes were incubated with primary antibodies overnight at 4 °C, washed with TBST, and then incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Protein bands were detected and visualized using ECL reagent and the eBlot Touch Imager. β-Actin or GAPDH was used as a loading control. Detailed information about the antibodies used is provided in the Supplementary Tables S1 and S2.

Electron microscopy

Primary fixation of NSCs was performed in 2.5% glutaraldehyde at 4 °C, followed 6 h later by secondary fixation with 1% osmium tetroxide. The samples were then dehydrated through a graded series of ethanol solutions and embedded in Araldite resin. Ultrathin sections approximately 60 nm thick were prepared and sequentially stained with 2.0% uranyl acetate and lead citrate. The stained sections were subsequently examined using an FEI electron microscope.

RNA sequencing

NSCs with MTHFD2 knockdown were collected, and total RNA was extracted using TRIzol reagent. High-quality RNA libraries were prepared and sequenced on the Illumina NovaSeq 6000 platform (Biotree Biotech, China), generating 150 bp paired-end reads. To ensure data quality, raw reads were processed with CutAdapt software to remove low-quality sequences and adapter contaminants, resulting in clean reads for subsequent analysis.

Mitochondrial features, superoxide, and membrane potential measurements

For the assessment of mitochondrial characteristics, including superoxide levels and membrane potential, MitoSOX (Yeasen, China) was used to quantify mitochondrial mROS in NSCs. In addition, total cellular ROS levels were measured by flow cytometry following staining with dihydroethidium (Meilunbio, China). The mitochondrial membrane potential (MMP) was evaluated using the JC-1 detection kit (Beyotime, China). Fluorescence intensity and mitochondrial morphology were examined under a confocal laser scanning microscope to obtain high-resolution images.

Viral and plasmid-based genetic manipulation

To reduce the expression of target genes in NSCs, a lentiviral packaging system was used to achieve stable knockdown of MTHFD2 and ATF4 (OBiO Biotech, China). The shRNA sequences used were as follows: MTHFD2, 5′-GCTCATGAAGAACACCATTAT-3′; ATF4, 5′-GCGAGTGTAAGGAGCTAGAAA-3′; and the negative control shRNA, 5′-TTCTCCGAACGTGTCACGT-3′. In parallel, ATF4 overexpression was established using lentiviral vectors (pcSLenti-EF1-EGFP-CMV-Atf4-3xFLAG-WPRE). For MTHFD2 overexpression, NSCs were transiently transfected with a pcDNA3.1-MTHFD2 plasmid. For in vivo experiments, stable MTHFD2 knockdown at the injection site was achieved by adeno-associated virus (AAV) infection using the construct pAAV-nestin-EGFP-3×flag-miR30shRNA(Mthfd2)-WPRE, following the manufacturer’s protocol.

Evaluation of lentiviral transfection efficiency

For reliable quantification, multiple approaches were employed to evaluate the efficiency of lentiviral transfection. Initially, transcriptional efficiency was assessed using lentiviral empty vectors containing green fluorescent protein (EGFP). NSCs were infected at different multiplicities of infection (MOI), and the transfection efficiency was determined by calculating the proportion of EGFP-positive cells among the total cell population, as observed under a fluorescence microscope. At an MOI of 5, the transfection efficiency in NSCs reached approximately 95%, as shown in the supplementary materials (Supplementary Fig. S3A).

WB analysis was subsequently performed to confirm the knockdown efficiency at the protein level. Overall, these assessments provided comprehensive validation of lentiviral transfection efficiency, ensuring the robustness and reliability of the experimental results.

Morris water maze test

To assess spatial learning and memory, the Morris water maze test was conducted. Adult mice were trained in a circular pool filled with water maintained at 20–22 °C. The pool was divided into four quadrants, with a hidden platform placed in one of them. Over five consecutive days, each mouse underwent four training trials per day, each lasting a maximum of 60 s. Mice that failed to locate the platform within 60 s were gently guided to it and allowed to remain there for 15 s. On the sixth day, a probe trial without the platform was conducted to assess memory retention by recording the time spent in the target quadrant and the number of crossings over the previous platform position.

Novel object recognition assay

To assess short-term recognition memory, the novel object recognition assay was performed. Mice were first acclimated in a 40 × 40 cm open-field arena for 10 min on day 1. On day 2, during the training session, they were exposed to two identical objects for 10 min. Following a 1 h intertrial interval, each mouse was reintroduced to the arena containing one familiar object and one novel object for a 5-min test session. The time spent exploring each object was recorded using a video tracking system. The discrimination index was computed as the proportion of time spent exploring the novel object relative to the total exploration time.

Statistical analysis

All statistical analyses were conducted using GraphPad Prism 10. Continuous variables between two groups were compared using the Student’s t-test, while comparisons among more than two groups were performed using one-way ANOVA followed by Tukey’s multiple comparisons test to determine statistical significance. Data are presented as mean ± SEM, and a p-value less than 0.05 was considered statistically significant.

Results

ICH-induced biphasic response of hippocampal NSCs: acute overactivation followed by long-term exhaustion and apoptosis

To investigate the changes in NSCs in the hippocampal dentate gyrus following ICH, we performed immunofluorescence staining in brain slices from mice. Compared to the sham group, the total number of NSCs (Nestin+) in the dentate gyrus markedly decreased at 3 days post-ICH (Fig. 1A). Nevertheless, the number of proliferating NSCs (Nestin + EdU+) increased markedly at this time (Fig. 1A). By 30 days post-ICH, both the total number of NSCs and the number of proliferating NSCs were significantly decreased compared with the sham group (Fig. 1A).

Fig. 1.

ICH induces an early increase in NSC proliferation, followed by a sustained depletion of the NSC pool, both in vivo and in vitro. (A) Immunofluorescence staining for EdU and Nestin in the dentate gyrus of the hippocampus from sham and ICH mice at 3 and 30 days post-injury, along with quantification of total neural stem cells (Nestin⁺) and proliferating neural stem cells (Nestin⁺EdU⁺). (B) Immunofluorescence staining for EdU and Sox2 in primary NSCs treated with or without 40 µM hemin for 6–24 h, along with quantification total NSCs (Sox2+) and proliferating NSCs (Sox2 + EdU+). (C) Representative images and quantification of neurosphere diameter after treatment with or without 40 µM hemin for 24 h. (D) TUNEL staining showing apoptotic changes in primary NSCs treated with or without 40 µM hemin for 24 h. (E) Western blot analysis of cleaved caspase-3, Bax, and Bcl2 in hippocampal tissues from sham and ICH mice at 3, 7, and 30 d post-ICH. (F) Western blot analysis of cleaved caspase-3, Bax, and Bcl-2 in primary NSCs treated with 40 µM hemin for 24 h. Data are presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001

Since ICH leads to the release and degradation of hemoglobin into heme and its oxidized form hemin, which induces oxidative stress, mitochondrial dysfunction and apoptosis in neural cells, hemin exposure is often used to mimic the cellular microenvironment of ICH in vitro [21]. To further investigate the mechanisms underlying the alterations in NSCs observed in vivo, in vitro experiments were performed using primary cultures of mouse NSCs. The optimal concentration of hemin for evaluating NSC proliferation and apoptosis was first determined. Results from the CCK8 assay showed that treatment with 40 µM hemin markedly reduced NSC viability after 24 h (Supplementary Fig. S1A), and therefore, this concentration was selected for subsequent analyses. The effects of hemin on NSC proliferation and survival were then assessed at different time points. After 6 h of hemin exposure, there was no noticeable difference in the total number of NSCs (Sox2+) between the control and hemin groups (Fig. 1B). Conversely, the number of proliferating NSCs (Sox2 + EdU+) was significantly increased in the hemin-treated group compared with the control group (Fig. 1B). After 24 h of hemin treatment, both the total number of NSCs (Sox2+) and the number of proliferating NSCs (Sox2 + EdU+) were significantly decreased in the hemin group (Fig. 1B).

Also, a comparable temporal pattern could be observed in the in vitro model, where 6 h of hemin exposure corresponded to the early proliferative phase and 24 h reflected the later stage characterized by decreased proliferation and increased apoptosis, consistent with the changes seen in vivo (Fig. 1B). These findings indicate that ICH triggers an early transient rise in NSC proliferation, followed by a sustained reduction in the NSC population. Consistently, the sphere-forming assay demonstrated that the neurosphere diameter was markedly smaller in the hemin-treated group than in the control group after 24 h of treatment (Fig. 1C).

To investigate whether the hemin-induced changes in NSCs were associated with apoptosis, TUNEL staining was performed in hemin-treated NSCs. After 24 h of hemin treatment, the number of TUNEL-positive NSCs showed a marked increase compared with the control group (Fig. 1D), indicating enhanced apoptotic cell death. WB analysis of cleaved caspase-3, Bax, and Bcl-2 expression in the hippocampus of ICH mice demonstrated a considerable upregulation of cleaved caspase-3 and Bax at 3 d post-ICH, which remained elevated until 30 d post-ICH (Fig. 1E). The anti-apoptotic protein Bcl-2 was significantly reduced at 3 days post-ICH and remained suppressed at 7 days, with a partial recovery observed by day 30 (Fig. 1E). These findings indicate that ICH induces a sustained increase in apoptosis in the hippocampus, particularly within the dentate gyrus. Consistent with these in vivo findings, similar patterns were observed in hemin-treated NSCs. WB analysis revealed that cleaved caspase-3 and Bax levels were significantly increased in the hemin-treated group compared with the control group, while Bcl2 expression was decreased (Fig. 1F).

Overall, our results demonstrate that ICH induces a biphasic response in NSCs within the dentate gyrus of the hippocampus. During the acute phase, NSC proliferation initially increased, whereas in the chronic phase, both the total number of NSCs and the number of proliferating NSCs were markedly reduced, suggesting long-term depletion of the NSC pool. To elucidate the mechanisms underlying these changes and to directly examine how hemorrhagic conditions affect NSC proliferation and survival, an in vitro model was established using hemin-treated primary NSCs.

PERK/ATF4 signaling transcriptionally activates MTHFD2, and knockdown of MTHFD2 inhibits proliferation and promotes apoptosis in NSCs

To investigate the functional significance of MTHFD2 in hippocampal NSCs post-ICH, we first examined temporal changes in MTHFD2 expression within the hippocampus. WB analysis of hippocampal lysates revealed persistent upregulation of MTHFD2 protein at 3, 7, and 30days post-ICH compared with the sham group (Fig. 2A). Similarly, in primary cultured NSCs, a time-course analysis of MTHFD2 expression following 40 µM hemin exposure (12, 24, 36, and 48 h) revealed a progressive increase in MTHFD2 levels, reaching a peak at 24 h and declining thereafter (Fig. 2B), which supports the use of the 24 h time point as a representative stage for assessing hemin-induced NSC dysfunction. These results indicate that MTHFD2 may dynamically respond to hemin exposure in vitro in response to hemorrhagic or hemin-induced stress.

Fig. 2.

MTHFD2 knockdown inhibits the proliferation and promotes apoptosis of NSCs in vitro. (A) Western blot analysis showing MTHFD2 protein expression in hippocampal tissues at 3, 7, and 30days after ICH. (B) Western blot analysis showing dynamic changes in MTHFD2 expression in NSCs after hemin treatment for 12, 24, and 48 h, revealing an initial increase followed by a gradual decrease. (C) Representative images and quantification of neurosphere formation at Day 1 and Day 3 showing that MTHFD2 knockdown reduces neurosphere diameter. (D) CCK-8 assay showing that MTHFD2 knockdown inhibits NSC proliferation over 72 h. (E) EdU staining and quantification, including control, sh-MTHFD2, oe-MTHFD2, hemin, sh-MTHFD2 + hemin, and oe-MTHFD2 + hemin groups, demonstrating that MTHFD2 overexpression rescues the hemin-induced reduction in NSC numbers. (F) Western blot analysis of cleaved caspase-3, Bax, and Bcl-2 expression confirming that MTHFD2 knockdown enhances pro-apoptotic protein levels. (G) TUNEL staining showing that overexpression of MTHFD2 reverses hemin-induced apoptosis in NSCs. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001

Previous studies have reported that the PERK/ATF4 branch of the endoplasmic reticulum stress (ERS) pathway transcriptionally activates MTHFD2. In this study, we validated this regulatory relationship in NSCs by CUT&Tag, dual-luciferase reporter assays, ChIP-qPCR, and WB analyses, confirming that PERK/ATF4 signaling promotes MTHFD2 expression under stress conditions (Supplementary Fig. S2A-J).

Given these changes, we assessed proliferation and self-renewal. Neurosphere assays revealed a significant decrease in both neurosphere number and diameter in the sh-MTHFD2 group (Fig. 2C), and decreased proliferative capacity on CCK-8 assays over 72 h (Fig. 2D). In contrast, EdU staining across six groups showed that MTHFD2 overexpression rescued the hemin-induced reduction in Sox2 + EdU+ NSCs (Fig. 2E).

Next, apoptosis in NSCs following MTHFD2 knockdown was analyzed. WBs showed that MTHFD2 knockdown elevated cleaved caspase-3 and Bax and reduced Bcl-2 (Fig. 2F-G). Consistently, TUNEL staining demonstrated increased apoptosis with hemin, which was mitigated by MTHFD2 overexpression (Fig. 2F).

Together, these data indicate that MTHFD2 is transcriptionally activated in stress conditions, supports NSC proliferation and self-renewal, and protects against apoptosis.

MTHFD2 is essential for preserving mitochondrial function, morphology, and redox homeostasis in NSCs

Considering that MTHFD2 predominantly localizes in the mitochondria and may influence NSC proliferation and apoptosis, we investigated its role in maintaining mitochondrial morphology and function. MitoTracker staining showed that hemin exposure led to an increase in mitochondrial fragmentation and loss of network structure, which became more pronounced after MTHFD2 knockdown (Fig. 3A). WB analysis demonstrated that MTHFD2 knockdown markedly reduced the mitochondrial fusion proteins MFN1 and MFN2 and increased phospho-Drp1 (Ser616), indicating enhanced mitochondrial fission (Fig. 3B). In contrast, MTHFD2 overexpression restored MFN1 and MFN2 expression and suppressed the hemin-induced elevation of p-Drp1(Ser616), suggesting rebalanced mitochondrial dynamics (Fig. 3C). JC-1 staining showed a substantial decline in the red/green fluorescence ratio in the sh-MTHFD2 and Hemin groups, indicating a loss of mitochondrial membrane potential (MMP). Although MTHFD2 overexpression appeared to mitigate this reduction under hemin treatment, the effect did not reach statistical significance (Fig. 3D).

Fig. 3.

MTHFD2 is essential for maintaining mitochondrial function, morphology, and redox homeostasis in NSCs. (A) MitoTracker confocal images showing mitochondrial morphology in Control and sh-MTHFD2 NSCs, with quantification of network metrics (right). (B) Western blots and quantification of MFN1, MFN2, and p-Drp1(Ser616) in Control, sh-MTHFD2, Hemin, and sh-MTHFD2 + Hemin groups. (C) Western blot extension after MTHFD2 overexpression: oe-MTHFD2 and oe-MTHFD2 + Hemin, showing rescue of MFN1/MFN2 reduction and attenuation of p-Drp1(Ser616) under hemin stimulation. (D) JC-1 staining and quantification of the red/green fluorescence ratio in Control, sh-MTHFD2, oe-MTHFD2, Hemin, sh-MTHFD2 + Hemin and oe-MTHFD2 + Hemin groups. (E) ATP content in Control, sh-MTHFD2, Hemin, and sh-MTHFD2 + Hemin groups. (F) Intracellular NADPH levels in Control, sh-MTHFD2, Hemin, and sh-MTHFD2 + Hemin groups. (G) Flow-cytometric analysis of ROS, including Control, sh-MTHFD2, oe-MTHFD2, Hemin, sh-MTHFD2 + Hemin, and oe-MTHFD2 + Hemin. Data are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001

To evaluate mitochondrial energy metabolism, ATP levels were measured and revealed a significant decrease following MTHFD2 knockdown, which was further exacerbated by hemin treatment. While MTHFD2 overexpression partially restored ATP production under hemin stimulation, the effect did not reach statistical significance compared to the sh-MTHFD2 + Hemin group (Fig. 3E). Consistent with its role in redox regulation, MTHFD2 deficiency led to a marked reduction in the NADPH/NADP⁺ ratio, an effect that was significantly reversed by MTHFD2 overexpression, particularly in the presence of hemin (Fig. 3F). Flow-cytometric analysis of ROS using DHE indicated strong ROS accumulation in sh-MTHFD2 and hemin-treated cells, whereas MTHFD2 overexpression reduced ROS under hemin stimulation (Fig. 3G).

Taken together, these findings show that MTHFD2 preserves mitochondrial morphology, maintains redox balance, and sustains energy production in NSCs. Loss of MTHFD2 results in mitochondrial fragmentation, depolarization, ATP depletion, and oxidative stress, whereas its overexpression effectively counteracts these deleterious effects under hemorrhagic stress.

ICH impairs mitophagy in NSCs and MTHFD2 is essential for maintaining mitophagy

To investigate the function of MTHFD2 in NSCs, we performed RNA-seq sequencing following MTHFD2 knockdown. The volcano plot from the RNA-seq data revealed significant alterations in gene expression profiles (Fig. 4A). KEGG pathway enrichment analysis showed that the “Phagosome” pathway was significantly enriched among the differentially expressed genes (Fig. 4B). Additionally, GSEA analysis demonstrated significant enrichment of the “Lysosome” pathway (Fig. 4C). These findings suggest that MTHFD2 may be involved in regulating autophagy-related processes in NSCs.

Fig. 4.

MTHFD2 maintains mitophagy in NSCs and can be partially rescued by exogenous NADPH. (A) Volcano plot of RNA-seq data showing differentially expressed genes following MTHFD2 knockdown in NSCs. (B) KEGG pathway enrichment analysis of differentially expressed genes revealed significant enrichment of the “Phagosome” pathway. (C) GSEA analysis showing enrichment of the lysosome-related gene set upon MTHFD2 knockdown. (D) Western blot analysis demonstrating that MTHFD2 knockdown, with or without hemin treatment, increased the levels of PINK1, p62, and the LC3BII/LC3BI ratio, indicating impaired mitophagy. (E) Exogenous NADPH supplementation reduced the accumulation of PINK1, p62, and LC3BII/LC3BI in MTHFD2-knockdown NSCs treated with hemin, suggesting partial restoration of mitophagy. (F) Schematic illustration of the Mt-Keima lentiviral system used to monitor mitophagy. (G) Reduced Mt-Keima red/green colocalization, quantified by Pearson’s correlation coefficient, indicates impaired mitophagy in hemin-treated and MTHFD2-deficient NSCs, partially rescued by NADPH. (H) Diminished MitoTracker–LysoTracker colocalization, quantified by Pearson’s correlation coefficient, reveals defective mitochondrial–lysosomal contact in NSCs exposed to hemin or MTHFD2 knockdown, with partial restoration by NADPH. Data are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001

To validate these findings, we conducted in vitro experiments using primary NSCs. WB analysis revealed that MTHFD2 knockdown, with or without hemin stimulation, led to increased levels of PINK1 and p62, and elevated LC3BII/LC3BI ratio compared to the control and hemin-treated groups (Fig. 4D). These results suggest that MTHFD2 depletion impairs mitophagy in NSCs. Given that MTHFD2 contributes to intracellular NADPH generation, which is essential for various redox reactions, we hypothesized that NADPH may influence mitophagy through metabolic regulation. To examine this, exogenous NADPH was supplemented in MTHFD2-knockdown NSCs under hemin treatment. WB analysis revealed that exogenous NADPH reduced the accumulation of PINK1, p62, and LC3BII/LC3BI (Fig. 4E), indicating that exogenous NADPH partially rescues the impaired mitophagy caused by MTHFD2 knockdown.

To directly assess mitophagic flux, we employed a lentiviral Mt-Keima reporter system (Fig. 4F), Mt-Keima is a pH-sensitive fluorescent protein that emits green fluorescence (488 nm) in mitochondria and red fluorescence (550 nm) upon lysosomal degradation. Quantitative analysis of Mt-Keima red/green colocalization using Pearson’s correlation coefficient revealed a significant reduction in mitophagic flux following hemin treatment and MTHFD2 knockdown, as evidenced by decreased red-green overlap (Fig. 4G). This impairment was further exacerbated by combined MTHFD2 deficiency and hemin exposure, but was partially rescued by NADPH supplementation (Fig. 4G). MitoTracker and LysoTracker staining showed reduced mitochondrial-lysosomal colocalization in hemin-treated cells, worsened by MTHFD2 knockdown, but partially restored by NADPH (Fig. 4H).

Collectively, our findings demonstrate that MTHFD2 is essential for maintaining mitophagy in NSCs. Loss of MTHFD2 disrupts mitophagic processes, leading to mitochondrial dysfunction, while exogenous NADPH supplementation partially reverses these impairments.

Exogenous NADPH supplementation partially rescues mitochondrial dysfunction, proliferation inhibition, and apoptosis in MTHFD2-knockdown NSCs

To determine whether exogenous NADPH could restore the MMP impaired by MTHFD2 knockdown and hemin treatment, JC-1 staining was performed. The ratio of red (aggregate) to green (monomer) fluorescence was significantly decreased in MTHFD2-knockdown NSCs following hemin stimulation, indicating a loss of MMP. Remarkably, exogenous NADPH supplementation partially restored this ratio, suggesting improved mitochondrial function (Fig. 5A). To further assess the effect of NADPH on mitochondrial mROS, Mitosox staining was applied. The results demonstrated a pronounced increase in mROS production in MTHFD2-knockdown NSCs after hemin exposure, whereas exogenous NADPH administration significantly reduced Mitosox fluorescence intensity, indicating that NADPH effectively mitigated oxidative stress under these conditions (Fig. 5B).

Fig. 5.

Exogenous NADPH supplementation partially rescues mitochondrial dysfunction, proliferation inhibition, and apoptosis in MTHFD2 knockdown-induced NSCs. (A) JC-1 staining showing that exogenous NADPH increased the red/green fluorescence ratio in MTHFD2-knockdown NSCs following hemin treatment, indicating partial restoration of MMP. (B) Mitosox staining indicating that exogenous NADPH reduced mROS levels in MTHFD2-knockdown NSCs after hemin treatment. (C) Western blot analysis demonstrating that NADPH supplementation partially reversed the altered expression of mitochondrial dynamics-related proteins, including decreased MFN1 and MFN2 and increased p-Drp1(S616). (D) Western blot results showing that exogenous NADPH partially restored apoptotic markers, reducing cleaved caspase-3 and Bax levels while increasing Bcl2 expression. (E) TUNEL staining indicating that NADPH supplementation reduced the number of apoptotic cells in MTHFD2-knockdown NSCs treated with hemin. (F) EdU assay showing that exogenous NADPH partially rescued the proliferation defect in MTHFD2-knockdown NSCs. Data are presented as mean ± SEM, *p < 0.05, **p < 0.01

WB analysis showed that in the sh-MTHFD2 group, exogenous NADPH partially restored the expression pattern of mitochondrial dynamics–related proteins, including elevated levels of the mitochondrial fission protein p-Drp1(S616) and reduced levels of the mitochondrial fusion proteins MFN1 and MFN2 (Fig. 5C). Similarly, in the sh-MTHFD2 group, NADPH supplementation partially reversed the increased expression of cleaved caspase-3 and Bax, along with the decreased expression of the anti-apoptotic protein Bcl2 (Fig. 5D). TUNEL staining further confirmed that exogenous NADPH supplementation reduced the number of apoptotic TUNEL+ NSCs in the sh-MTHFD2 group (Fig. 5E). Consistent with these findings, EdU staining demonstrated that exogenous NADPH significantly restored the proliferation capacity of sh-MTHFD2-treated NSCs under hemin stress, as evidenced by a marked increase in the percentage of Sox2⁺EdU⁺ cells compared to the sh-MTHFD2 + hemin group (Fig. 5F).

Together, these results demonstrate that exogenous NADPH supplementation can partially ameliorate mitochondrial dysfunction, proliferation inhibition, and apoptosis induced by MTHFD2 knockdown in NSCs, underscoring the essential role of NADPH in maintaining redox homeostasis and supporting NSC function.

MTHFD2 knockdown exacerbates neurological function and cognitive impairment in mice following ICH

To validate our in vitro findings, we established an in vivo mouse model to investigate the effects of MTHFD2 knockdown on neurological function and cognitive performance after ICH. Adeno-associated virus was stereotactically injected into the hippocampus to induce MTHFD2 knockdown specifically in hippocampal NSCs. We first examined the influence of MTHFD2 knockdown on NSC proliferation and survival in the hippocampus following ICH. Brain slice staining for EdU and nestin revealed a marked reduction in both proliferating (EdU+) and total (Nestin+) NSCs in the sh-MTHFD2 + ICH-Day3 group compared with the ICH-Day3 group without MTHFD2 knockdown (Fig. 6A). WB analysis of hippocampal tissue collected 3 d after ICH showed that cleaved Bax and caspase-3 expression levels were significantly higher in the sh-MTHFD2 + ICH-Day3 group than in the ICH-Day3 group, whereas Bcl2 expression was markedly decreased (Fig. 6B).

Fig. 6.

Knockdown of MTHFD2 exacerbates neurological functions and cognitive impairment in mice following ICH. (A) Brain slice staining for EdU and nestin showing the effect of MTHFD2 knockdown on NSC proliferation in the hippocampus 3 d post-ICH. (B) Expression levels of cleaved caspase-3, Bax, and Bcl-2 in hippocampal tissue collected 3 d after ICH. (C) The Morris water maze test assessing spatial learning and memory performance. (D) Novel object recognition test evaluating recognition memory. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01

To evaluate the functional effects of MTHFD2 knockdown on cognitive performance, behavioral assessments were performed using the Morris water maze and novel object recognition tests. In the Morris water maze, mice in the sh-MTHFD2 + ICH group spent significantly less time in the target quadrant than ICH mice without MTHFD2 knockdown, indicating impaired spatial learning and memory (Fig. 6C). Similarly, in the novel object recognition test, sh-MTHFD2 + ICH mice displayed a lower preference for the novel object compared with ICH mice without MTHFD2 knockdown, as reflected by a reduced discrimination ratio (Fig. 6D). These results suggest that MTHFD2 knockdown leads to deficits in recognition memory in ICH mice.

Collectively, our in vivo findings demonstrate that MTHFD2 knockdown aggravates hippocampal NSC loss and enhances apoptosis following ICH, resulting in significant impairments in learning and memory. Collectively, these findings highlight the essential role of MTHFD2 in preserving NSC function and cognitive health after ICH.

Discussion

ICH is a severe type of stroke, and survivors frequently experience long-term cognitive impairment that profoundly affects their quality of life [5]. The hippocampus, a key brain region responsible for learning and memory, is highly susceptible to ICH-induced injury [22]. NSCs located in the hippocampal dentate gyrus are essential for sustaining neurogenesis and supporting cognitive function [12, 23]. Our study demonstrates that mitochondrial dysfunction represents a pivotal factor contributing to ICH-induced NSC impairment, with MTHFD2 identified as a central mediator of this process (Fig. 7). Importantly, exogenous NADPH supplementation partially reversed the proliferation inhibition and apoptosis resulting from MTHFD2 knockdown. These findings underscore the significance of maintaining cellular homeostasis and redox equilibrium as potential strategies to protect NSCs and improve cognitive outcomes in patients with ICH.

Fig. 7.

Schematic illustration of MTHFD2-mediated NADPH homeostasis in maintaining mitochondrial function and mitophagy in NSCs after ICH. Following ICH, hippocampal NSCs experience endoplasmic reticulum stress (ERS), which activates the PERK/ATF4 signaling pathway. Activation of ATF4 leads to the transcriptional upregulation of MTHFD2, a key mitochondrial enzyme essential for NADPH production. MTHFD2 maintains intracellular NADPH levels and supports redox homeostasis within NSCs. By sustaining adequate NADPH supply, MTHFD2 preserves mitochondrial integrity and morphology, promotes mitophagy, and enhances NSC survival under ICH-induced stress conditions. This schematic highlights the crucial role of MTHFD2-mediated metabolic regulation in maintaining NSC function and mitigating cognitive deficits following brain injury

Our study revealed a dynamic temporal response of NSCs to ICH. At the initial stage, NSCs exhibited enhanced activation, likely as a compensatory mechanism against ICH-induced damage. This early response, however, was followed by a prolonged phase of functional decline and increased apoptosis, ultimately leading to NSC exhaustion and long-term depletion. This pattern indicates that ICH induces an acute adaptive activation of NSCs that becomes unsustainable over time. The initial overactivation may serve to compensate for cellular loss but subsequently results in depletion of the NSC pool. These findings are in agreement with previous reports showing that acute brain injuries can transiently stimulate NSC activity, yet persistent activation leads to their exhaustion and impaired neurogenic potential [9]. Our results are also consistent with studies demonstrating that aberrant neurogenesis compromises the self-renewal capacity of hippocampal NSCs, contributing to their depletion and disrupted adult hippocampal neurogenesis [24, 25]. Therefore, maintaining NSC homeostasis and sustaining optimal autophagic flux are essential for preserving long-term neurogenic capacity and cognitive function.

In this study, we observed that ICH activates the PERK/ATF4 signaling pathway in hippocampal NSCs. Although ATF4-dependent transcriptional regulation of MTHFD2 has been previously reported under different stress conditions, our results confirm the presence of this regulatory axis in NSCs and demonstrate its functional importance in the context of ICH-induced brain injury [26–28]. Activation of the PERK/ATF4 pathway leads to upregulation of MTHFD2, which plays a vital role in maintaining mitochondrial homeostasis and supporting NSC survival. MTHFD2, a key enzyme involved in mitochondrial one-carbon metabolism, is critically important for cellular function in the setting of ICH [16]. Previous research has shown that MTHFD2 is highly expressed in embryonic stem cells and induced pluripotent stem cells, where it contributes to maintaining redox balance and promoting rapid proliferation [19, 20]. Consistent with these findings, our study revealed that MTHFD2 regulates stem cell pluripotency, as evidenced by the enrichment of signaling pathways associated with the “ Phagosome” pathway following MTHFD2 knockdown [20]. Beyond its metabolic role demonstrated in this study, MTHFD2 has also been implicated in additional cellular regulatory processes in other systems. While these functions were not examined here, they may contribute to broader MTHFD2-mediated mechanisms that warrant future investigations. Our findings extend current understanding of the molecular mechanisms governing NSC behavior and highlight MTHFD2 as a promising therapeutic target for promoting neurogenesis and cognitive recovery after ICH.

MTHFD2 knockdown has been reported to disrupt the mitochondrial electron transport chain [20, 29]. In our study, MTHFD2 knockdown in NSCs led to a reduction in MMP, increased ROS accumulation, and impaired mitochondrial dynamics and ATP production. These disruptions in mitochondrial function and redox balance have profound effects on NSC survival and neurogenesis. During the transition from quiescence to an active state, NSCs may shift their primary energy source from glycolysis to oxidative phosphorylation [30, 31]. This metabolic shift can lead to increased ROS generation, rendering NSCs more vulnerable to oxidative stress following ICH [32]. The accumulation of ROS can activate apoptotic signaling pathways, leading to cell death. The resulting loss of mitochondrial function compromises the energy supply necessary for NSC proliferation and differentiation, thereby diminishing overall neurogenic capacity [33, 34]. This dual impact, increased oxidative stress and impaired energy production, highlights the susceptibility of NSCs to ICH-induced mitochondrial injury and emphasizes the necessity of maintaining mitochondrial integrity for sustaining NSC viability and promoting functional recovery [33–35]. Our findings reaffirm the essential role of MTHFD2 in preserving redox balance and energy metabolism, highlighting its importance in NSC function and the detrimental consequences of its dysregulation following ICH.

One of the most compelling findings of our study is that exogenous NADPH can partially restore the proliferative capacity and reduce the apoptosis caused by MTHFD2 knockdown in NSCs. This observation suggests that the detrimental effects of MTHFD2 deficiency on NSC function may be mediated, at least in part, through its regulatory role in NADPH production and redox balance. NADPH plays a pivotal role in regenerating reduced glutathione (GSH), the principal intracellular antioxidant [36]. By maintaining sufficient levels of GSH, NADPH enables the neutralization of ROS and protects essential cellular components, including lipids, proteins, and DNA, from oxidative injury [37, 38]. When NADPH availability is insufficient, cells become increasingly vulnerable to oxidative stress, leading to mitochondrial dysfunction, DNA damage, and eventual apoptosis [38–40]. MTHFD2 contributes substantially to NADPH generation, which is indispensable for maintaining redox homeostasis and defending cells against oxidative stress [39, 41]. In NSCs, reduced NADPH availability resulting from MTHFD2 knockdown likely disrupts anabolic and antioxidant processes, thereby suppressing proliferation and enhancing apoptosis. The partial restoration of NSC function by exogenous NADPH supplementation indicates that replenishing NADPH levels may represent a feasible therapeutic approach to counteract the adverse effects of MTHFD2 deficiency. This conclusion is in agreement with previous studies showing that intracellular NADPH concentration critically influences cell proliferation and apoptotic regulation [42–44]. Therefore, strategies aimed at maintaining or restoring NADPH balance could enhance NSC survival and proliferative capacity, thereby facilitating neurogenesis and improving cognitive outcomes following ICH. Future investigations could further evaluate the long-term effects of NADPH supplementation on NSC function and determine the optimal dosing and delivery approaches for potential clinical application.

To evaluate the functional significance of MTHFD2 in vivo, we examined the effects of MTHFD2 knockdown on cognitive performance in an ICH mouse model. Our findings demonstrate that MTHFD2 knockdown aggravates neurological deficits and cognitive impairment in ICH mice, as evidenced by diminished performance in both the Morris water maze and novel object recognition tests. These results indicate that MTHFD2 plays an important role in cognitive recovery following ICH, most likely by supporting the survival and functional integrity of hippocampal NSCs. The cognitive decline observed in MTHFD2-knockdown mice corresponds with the reduced neurogenesis and increased apoptosis of NSCs. Given that the hippocampus is a critical region responsible for learning and memory, the loss of NSCs substantially compromises its functional capacity, contributing to the observed cognitive deficits.

In conclusion, our study characterizes the dynamic response of hippocampal NSCs following ICH, revealing a pattern of initial activation followed by progressive exhaustion and apoptosis. We identified MTHFD2 as a stress-inducible mitochondrial enzyme that is significantly upregulated in NSCs after ICH. Functionally, MTHFD2 plays an essential role in maintaining mitochondrial integrity, morphology, and redox homeostasis by sustaining NADPH levels under hemorrhagic stress. Loss of MTHFD2 results in impaired NSC proliferation, increased apoptosis, and aggravated cognitive deficits in ICH mice. Notably, exogenous NADPH supplementation partially alleviated these impairments, underscoring the therapeutic potential of targeting NADPH-dependent pathways to preserve NSC function and enhance cognitive recovery after brain injury.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ERS

Endoplasmic reticulum stress

ICH

Intracerebral hemorrhage

MMP

Mitochondrial membrane potential

ROS

Reactive oxygen species

NSC

Neural stem cell

UPR

Unfolded protein response

WB

Western blot

Author contributions

Yikui Liu: Writing – original draft, Investigation, Methodology, Data curation, Conceptualization. Canxin Xu: Writing – review & editing, investigation. Baofeng Wang: Investigation, Methodology, Data curation. Fengzhen Cui: Methodology, Data curation, Funding acquisition. Aoqian Xu: Methodology, Investigation. Yuxiao Ma: Methodology, Investigation. Qixiang Zhang: Methodology. Qingfang Sun: Investigation. Yongtao Zheng: Funding acquisition, Supervision, Resources, Project administration. Yuhao Sun: Resources, Funding acquisition, Supervision, Project administration, Conceptualization. Liuguan Bian: Supervision, Funding acquisition, Resources, Project administration, Conceptualization.

Funding

This work was supported by the National Natural Science Foundation of China (grant numbers 82171292, 82171283, 82501560, 82471318, and 82404301). Additional support was provided by the China Postdoctoral Science Foundation under Grant Number 2025M772143, the Fundamental Research Funds for the Central Universities (YG2024Q NB04), and the National Research Center for Translational Medicine at Shanghai (NRCTM(SH)- 2023-02).

Declarations

Ethics approval and consent to participate

This study was approved by the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University, and all experimental procedures were performed in accordance with its ethical guidelines.

Consent for publication

All authors have reviewed and approved this work for publication.

Competing interests

The authors declare that they have no competing interests.