Blocking FOXO4 confers neuroprotection against oxidative stress and ischemia-reperfusion caused neuronal injury

Abstract

Previous studies suggest that FOXO4 facilitates inflammation and oxidative stress in non-brain tissues under stress or disease conditions, indicating that downregulation of FOXO4 may be neuroprotective in ischemia-reperfusion-induced brain injury. However, this possibility has not been tested in a cerebral ischemia-reperfusion context. Here, we exposed FOXO4 knockout (KO) primary neuronal cultures to oxidative stress or oxygen-glucose deprivation (OGD), and subjected KO mice to transient middle cerebral artery occlusion (tMCAO). Furthermore, we used a structure-based approach combined with cell-based assays to identify a FOXO4 inhibitor and evaluated its effects in tMCAO-treated wild-type (WT) mice. Our results showed that FOXO4 KO reduced oxidative stress- and OGD-induced neuronal death, attenuated tMCAO-induced infarct volume, improved animal survival, decreased neurological deficits, and enhanced functional recovery compared with WT cells or mice. Immunohistochemical staining and Western blot analyses further indicated reduced neuroinflammation in the KO brain. The identified FOXO4 inhibitor, Actinomycin D (ActD), suppressed FOXO4 activity at a low dose, reduced brain injury, and improved functional recovery in WT mice following tMCAO. Together, these findings indicate that FOXO4 is a potential therapeutic target and that suppression of its activity may promote neuronal survival after ischemic stroke-induced brain injury.

Introduction

Stroke is a major health problem: it is the fifth leading cause of death and the leading cause of disability in the United States. Ischemic stroke is the most frequent type of stroke, affecting millions of people across the globe (1). Despite intensive research, there is still no effective treatment for this condition, and tissue plasminogen activator remains the only Food and Drug Administration-approved medication for treating ischemic stroke (2). While it is necessary to restore cerebral blood flow, reperfusion itself induces the production of large amounts of reactive oxygen species that further damage proteins and other intracellular macromolecules to exacerbate brain injury (3). Moreover, ischemia-reperfusion (I/R) also disrupts the blood-brain barrier, resulting in infiltration of leukocytes and inflammatory responses to aggravate brain injury (4). To develop effective therapeutics, it is necessary to identify and validate additional therapeutic targets.

Transcriptional gene regulation, governed primarily by numerous transcriptional factors, is one of the upstream parts potentially modulating I/R-induced brain injury. The Forkhead Box Proteins O (FOXOs), consisting of four structurally and functionally related proteins, FOXO1 (also referred to as FKHR), FOXO3 (also known as FOXO3a), FOXO4 (also known as AFX1), and FOXO6, are one family of transcription factors, representing mammalian homologs of daf-16 in C. elegans and playing a crucial role in cell survival, cell proliferation, metabolism, response to oxidative stress, apoptosis, and aging (5). Under oxidative stress or the absence of the cellular survival drive of growth factors, FOXOs translocate to the nucleus and upregulate a series of target genes, thereby promoting cell growth arrest and apoptosis (6, 7).

Despite the highly structural and functional similarities of FOXOs, previous data have suggested that the physiological roles of FOXOs are functionally diverse in mammals (8, 9). For instance, FOXO1 deficiency is embryonically lethal, and FOXO3 deficiency exhibits age-dependent infertility in females in mice; however, the loss of FOXO4 does not cause any notable changes in the mouse, suggesting that FOXO4 may play a different role from other FOXO members (8, 10). FOXO4 was found to promote cell death in the heart and the liver following I/R, while downregulation of FOXO4 confers protection against I/R-induced tissue injuries in these organs (11, 12). Moreover, following cardiac ischemia, FOXO4 enhances the interaction of leukocytes with the endothelial cells of blood vessels to promote early tissue inflammation (13). In contrast, downregulation of FOXO4 suppresses oxidative stress-induced cell death in proangiogenic cells and promotes neovascularization in ischemic limbs (14). Functionally, FOXO activity is negatively regulated by phosphorylation via the phosphoinositide 3-kinase-Akt pathway, a well-known cell survival pathway (6, 7). Ischemic preconditioning upregulates Akt activity, leading to FOXO inhibition and promotion of neuronal survival against a subsequent severe ischemic insult (15). These prior studies suggest that FOXO4 may be a therapeutic target for ischemic stroke, yet this has not been tested in ischemic stroke neuronal cells or animals. In this study, we examined the role of FOXO4 in neuronal injury in vitro and in vivo following oxidative stress, oxygen-glucose deprivation (OGD), and I/R, using the primary neuronal cultures and FOXO4 knockout (KO) mice combined with different approaches, including pharmacological methods.

Materials and Methods

Mice

All animal-related experiments and procedures were approved by the Institutional Animal Care and Use Committee of the Texas Tech University Health Sciences Center and were in compliance with the National Institute of Health Guide for the Care and Use of Laboratory Animals. Animals were maintained on a 14-/10-hour light/dark cycle with a maximum of four mice per cage, able to access water and food ad libitum. FOXO4 global knockout (KO) mice in the background of FVB have been described before (13) and were used for ischemic stroke-induced brain infarct studies, while the same strain of wild-type (WT) mice was used as the control. For animal behavioral studies, the KO and WT male mice with a hybrid background (F1 generation) produced from the heterozygous KO FVB females crossed with WT C57BL/6 males were used, because of the retina degeneration problem in FVB mice. As FOXO4 is an X chromosome-linked gene, in this breeding strategy, 50% of males would be FOXO4 KO males, and 50% would be WT males that were used as controls for behavioral studies following ischemic stroke. More than one litter of animals was used in each experiment. To minimize the variability in experimental data caused by fluctuating hormones in female mice, only the male animals between 8–12 weeks of age with a body weight of 25–30 g were used in the study. Sample size calculations and power analysis were performed according to our previously described methods (16) using the statistical software Stata (StataCorp LP, College Station, TX, USA).

Primary cortical neuronal cultures

Primary cortical neuronal cultures were prepared from wild-type and FOXO4 KO mice at postnatal day 0 according to our previously described methods (17). Briefly, the cerebral cortex was isolated and digested with 5 ml of 0.25% trypsin/EDTA supplemented with 75 μl 0.1% DNase (2000 UI/mg). After 15–20 minutes of digestion at 37°C, the digested tissues were centrifuged at 1000 × g for 3 minutes at room temperature. The supernatant was discarded, and the resulting tissue pellets were resuspended with fetal bovine serum (FBS) to stop the digestion. After centrifugation, the tissues were resuspended with primary neuronal culture medium (Neurobasal medium supplemented with 2% B27, 2 mM L-glutamine, and penicillin/streptomycin) and then pipetted up and down several times using serological pipets to disrupt tissues. After filtering through a 70 μm nylon cell strainer, the cell suspension was adjusted and plated in poly-DL-lysine-coated 12-well plates. After 7 days of incubation at 37°C in a 5% CO2 incubator, the cells were treated with 20 μM menadione (MD) for 24 h or oxygen-glucose deprivation (OGD) for 3 h and then in normal culture conditions for 21 h before cell viability was assessed.

Oxygen-glucose deprivation (OGD)

WT and FOXO4 KO primary neuronal cultures were subjected to OGD at 7–10 in vitro days according to previous methods (17). Briefly, the neuronal culture medium was replaced with a glucose-free Hank’s Balanced Salt Solution (HBSS), and then the culture plates were placed into a hypoxic chamber where the air was replaced with 5% N₂ and 95% CO₂. After 3 h of OGD condition, the neuronal cultures were returned to the normal culture condition, and after 21 h cell viability was measured via ATP and MTT assays.

ATP assay

We used an ATP detection kit (Cayman Chemical, Ann Arbor, Michigan, USA) to measure ATP levels in cultured neurons according to the manufacturer’s instructions.

MTT assay

Cell viability in cell cultures was assessed using an MTT assay kit (R&D Systems, Minneapolis, MN, USA) based on the company’s guide.

Transient middle cerebral artery occlusion (tMCAO)

The tMCAO procedure was performed according to previously described methods (18, 19). Briefly, anesthesia of mice was induced with 5% isoflurane and then maintained with 2% isoflurane, and body temperature was maintained at 37°C. The left hemisphere was subjected to tMCAO using a silicon-coated monofilament (RWD, Sugar Land, TX, USA), and after 1 hour, the monofilament was removed. A Laser Speckle Imaging System (RWD, Sugar Land, TX, USA) was used to monitor cerebral blood flow through the MCA. Only the mice with successful occlusion of MCA were included in the studies, as reflected by reduced blood flow by over 80% and reperfusion with more than 75% recovery of blood flow in the MCA (Fig. S1). Following the surgery, mice were placed on a heating pad to maintain body temperature at 37°C until they completely recovered. Mouse neurological severity score (NSS) was assessed using a 5-point scale as previously described (16, 18). Animal genotypes were blind to the researchers who performed the surgery and data analysis. Only the animals with an NSS score of 2 or 3 were included in the experiments (18).

TTC staining and measurement of infarct volumes

Mice were euthanized 24 h after I/R to isolate the brains. After being sectioned into 2 mm slices, the brain sections were incubated with 2% of 3,5-triphenyl tetrazolium chloride (TTC, Sigma, Saint Louis, MO, USA) at 37°C for 20 minutes. Subsequently, the sections were fixed in 4% paraformaldehyde and then imaged. The infarcted volume was measured using ImageJ software and calculated as previously described (20).

Assessment of modified neurological deficits

At 1, 3, 5, and 7 days following I/R, a modified neurological severity score (mNSS) system was applied to assess motor and sensory functions in mice (21). Based on this system, a scale of 0 to 14 was given to an animal, with 0 for normal neurological behavior and 14 for the maximal neurological deficit (22).

Novel object recognition (NOR) test

A box (50 × 50 × 30 cm) containing two objects was used for the NOR test at 7 days following I/R. On the first day of the test, mice were placed in the box and allowed to move freely for 10 minutes. On the second day, one of the objects was replaced with a new one (novel object), and the other remained as a familiar object. Then, mice were placed into the box to move and explore the objects. By using a camera, the total time spent exploring each object was recorded, and the ratio of the new object exploration was calculated as the discrimination index (DI = (Tnovel − Tfamiliar)/(Tnovel + Tfamiliar) (19).

Y-maze test

In the training session, each mouse was placed in an arm of the Y maze (start arm) with one of the arms blocked with an opaque door (novel arm) and allowed to explore the start arm and the remaining arm (other arm) for 5 minutes. The mouse was then removed and returned to its home cage, and the novel arm was unblocked. After one hour, in the test trial, the mouse was replaced on the start arm and allowed to explore all three arms freely for 2 minutes. The number of arm entries was recorded by the experimenter, blind to the genotype of the mice.

Immunofluorescent staining

Brains were coronally cut into 3 parts, and the frontal and occipital parts were used for TTC staining to define the penumbral area. The middle part was washed and fixed in 4% paraformaldehyde. Then, the 16 μm thickness sections were prepared on a Cryostat (Leica, Buffalo Grove, IL, USA), and the brain sections were processed based on the prior method (23). The primary antibodies used were anti-GFAP (1:50, EMD Millipore, #MAB360), anti-Iba1 antibody (1:50, Cell Signaling Technology, #31659), anti-CD45 (Invitrogen #MA180090), anti-Ly6G (Cell Signaling #87048), and anti-CD11 (Developmental Studies Hybridoma # M1/70.15.11.5.2). The secondary antibodies were Texas red-conjugated anti-rabbit and FITC-conjugated anti-mouse antibodies. DAPI (IHC-tek#IW-1404) was used to stain the nuclei.

The images were captured with a fluorescence microscope (Echo Revolve, San Diego, USA) and quantified with ImageJ software (24). The ratio of the red- or green-positively stained cells in each field was calculated by the total red or green cells divided by the total number of cells (DAPI positively stained cells) (25, 26).

Western blot

The cortex of the left (i.e., the I/R side) hemisphere of each mouse brain was collected for lysate preparation in the RIPA buffer supplemented with a protease cocktail. The brain tissues were sonicated in the lysate buffer, and the lysates were centrifuged at 12,000 × g for 4 min at 4°C. The supernatant was collected, and then the total protein concentration was measured before being subjected to the SDS-PAGE and transferred to the nitrocellulose membrane according to previous methods (17). The antibodies used were anti-IL-1β (1:1000, Cell Signaling, #63124), anti-IL-6 (1:1000, Cell Signaling, #12912), anti-TNFα (1:1000, Cell Signaling, #11948), and anti-β-actin (1:1000, Santa Cruz Biotechnology, #sc-1616). The secondary antibodies used were anti-rabbit IgG, HRP-conjugated antibody (Cell Signaling, #7074), and anti-mouse IgG, HRP-conjugated antibody (Cell Signaling, #7076). The western blot results were documented using an imaging system (C400 Azure Biosystem), and protein band intensities were measured using the NIH ImageJ software.

Structure-based drug screening

We utilized the Autodock Vina in the PyRx 0.9.8 software (https://pyrx.sourceforge.io/) to perform the docking-based virtual screening of several small compound libraries against FOXO4 DNA-binding domain (DBD, Protein Data Bank, https://www.rcsb.org/; PDB: 3L2C), with a grid box of 20 × 20 × 20 Å positioned within the box center coordinates −1.0, 13.0 and −15.0, focusing on the DBD of FOXO4. The FOXO4 DBD was modified by removing the DNA chains during computational analyses. We screened 7,742 compounds from four different chemical libraries, including the NIH Clinical Collections https://pubchem.ncbi.nlm.nih.gov/source/NIH%20Clinical%20Collection), the FDA-approved and clinical drugs (https://www.targetmol.com/library-collection-2/sorted-by-clinical-phase and https://www.apexbt.com/discoveryprobetm-clinical-fda-approved-drug-library.html), and a natural product library (https://www.apexbt.com/discoveryprobetm-natural-product-library-plus-1.html). The lead compounds were then docked onto FOXO1 DBD (PDB: 3COA) and FOXO3 DBD (PDB: 2UZK), and any candidates that also showed a high affinity to FOXO1 or FOXO3 were eliminated from the list.

Cell-based reporter assay

HEK293 cells expressing a human FOXO4 plasmid (Addgene #17549)(27) and a luciferase reporter plasmid (Addgene, # 34571) driven by a FOXO4 target gene Arginase-1 promoter(13, 28), were treated with triplates with 100 nM of identified compounds or vehicle (DMSO) for 8 h and the treated cells were then subjected to a luciferase-based reporter assay using a luciferase reporter substrate kit (Abcam, Waltham, MA, USA). Moreover, to eliminate the compounds that also showed inhibitory effect on FOXO1 and FOXO3, we performed additional reporter assays with HEK293 cells co-transfected with the 3XIRS luciferase (Luc) reporter that contains 3 copies of the insulin-responsive sequence(29) or FHRE-Luc(30) together with FOXO3 plasmid (Addgene #8360)(30).

Molecular docking

We used SMILES(31) notation for ActD (CHEMBI: 27666) ligand against the FOXO4 (PDB: 3L2C) to predict the protein-ligand binding sites in SwissDock 2024(32) and Protenix (https://protenix-server.com/add-prediction), keeping multiple sequence alignment and seed 95326 as a parameter. AutoDock Vina has applied a method for SwissDock, a grid box of 20 × 20 × 20 Å positioned within the box center coordinates −1.0, 13.0, and −15.0, focusing on the DNA-binding domain of FOXO4 for accurate docking predictions of actinomycin D. The PDB model (3L2C) was modified by removing the DNA chains during computational analyses. A separate analysis was executed with the matching coordinates, keeping Attracting Cavities for higher confidence predictions in SwissDock. The results were visualized using UCSF ChimeraX (33) on a Macintosh System.

Drug treatment of mice

For drug treatment studies, WT males were randomly separated into the vehicle group and drug treatment group using an online tool (http://www.graphpad.com/quickcalcs/). Either 0.25 or 0.5 mg/kg of the identified inhibitor, actinomycin D (ActD), or vehicle was intraperitoneally injected into WT 1 h before tMCAO. After 24 h, mice were euthanized to collect the brains for TTC staining. Alternatively, the mice were allowed to survive for 1–10 days to assess functional recovery.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism version 9.0 statistical software. Differences between the two groups were assessed using an unpaired t-test. For comparisons between more than two groups of animals, repeated measures and two-way ANOVA with Tukey’s post hoc test were utilized. All numerical data were presented as mean ± SD or SEM. P < 0.05 was regarded as statistically significant.

Results

KO of FOXO4 reduces neuronal death caused by oxidative stress or OGD in primary cortical neuronal cultures

To assess the effect of KO of FOXO4 on oxidative stress-induced cell death, we treated the primary cortical neuronal cultures with an oxidative stress inducer, menadione (MD)(18), and then measured the viability of the cells via ATP and MTT assays. KO of FOXO4 was confirmed by Western blot analysis of FOXO protein level (Fig. 1a). Following MD treatment, KO neurons showed increased cell survival as reflected by higher ATP levels in KO cells compared to WT cells (Fig. 1b). Increased viability in the KO cells was also supported by the MTT assay (Fig. 1c). However, in those neuronal cultures without MD treatment, the cellular viability did not differ between WT and KO neurons (Fig. 1b, 1c). To validate these results, we further treated the two types of neuronal cultures with an in vitro model of ischemic stroke, OGD, and then examined cell viability. In both the ATP assay (Fig. 1d) and MTT assay (Fig. 1e), KO neurons showed reduced cell death compared to WT neurons. These results indicate that KO of FOXO4 is neuroprotective against oxidative stress- or OGD-induced neuronal death in primary neuronal cultures.

Figure 1. Effect of KO of FOXO4 on oxidative stress- and OGD-induced neuronal death in primary neuronal cultures.

a. Western blot analysis of FOXO4 protein levels in primary neuronal cultures to validate FOXO4 KO. b. ATP assay results of KO and WT neurons following MD treatment. c. MTT assay results of KO and WT neurons following MD treatment. d. ATP assay results of KO and WT neurons following OGD treatment. e. MTT assay results of KO and WT neurons following OGD treatment. Data are shown as mean ± SD. N = 10–14, ***p < 0.001. ns, no significant difference (Student’s t-test).

KO of FOXO4 reduces the infarct volume and enhances functional recovery after tMCAO

To determine whether FOXO4 influences the outcome of I/R-induced brain injury in vivo, we performed tMCAO in FOXO4 KO and WT mice. Selective KO of the gene was validated by Western blot analysis (Fig. 2a). After 24 h following the reperfusion, mice were euthanized to assess brain injury by the TTC staining (Fig. 2b). Our results showed a significant decrease in infarct volume in the KO group compared to the WT group (Fig. 2c, 2d), suggesting that disrupting the FOXO4 gene is neuroprotective against I/R-induced neuronal injury.

Figure 2. KO of FOXO4 reduces brain infarct volume, improves animal survival, and enhances functional recovery following I/R.

a. Western blot analysis of FOXO4 protein levels in mouse brain to validate FOXO4 KO. b. A schematic of experimental design. c. Representative TTC staining of mouse brains from the sham surgery, WT, and KO animal groups after MCAO. d. Quantitation of TTC staining results. e. KO mice increased the animal survival rate compared to WT mice after tMCAO. f. KO mice improved mNSSs compared to WT mice after tMCAO. g. KO mice showed better memory compared to WT mice in the NOR test after tMCAO. Data are shown as mean ± SD. N = 10–14 for (d), ****p < 0.0001. N = 14– 15 for (e)-(g), * p < 0.05, **p < 0.01, ***p < 0.001 (one- or two-way ANOVA).

To determine whether KO of the FOXO4 influences animal survival following I/R, animal death and survival were recorded on days 1, 2, 3, 4, 5, and 7 after tMCAO. As shown in Fig. 2e, KO of FOXO4 increases the animal survival rate. Moreover, we also assessed the neurological deficits and found that KO of FOXO4 facilitated animal functional recovery, as reflected by the reduced mNSSs compared to the WT mice following the tMCAO procedure on days 1, 3,5, and 7 (Fig. 2f). Improved functional recovery in the KO mice was also supported by the cognitive function test, including the memory capabilities that were evaluated 10 days after MCAO. We observed better learning and memory in KO mice compared to WT mice in the NOR test (Fig. 2g). Therefore, KO of FOXO4 reduces neuronal death, improves animal survival, and enhances functional recovery following I/R.

KO of FOXO4 attenuates neuroinflammation in the brain following I/R

Neuroinflammation is a hallmark following I/R in the brain. To assess whether KO of FOXO4 alters the inflammatory responses, brain tissues on the ipsilateral side were isolated from KO and WT mice 48 h after I/R. Western blot analysis of the pro-inflammatory cytokines, including TNF-α, IL-1β, and IL-6, revealed a significant reduction in the KO brains compared to WT brains (Fig. 3a–3d). These data indicate that KO of FOXO4 attenuates I/R-induced neuroinflammation.

Figure 3. KO of FOXO4 attenuates neuroinflammation following I/R.

a. Representative western blot results of the indicated proteins. b-d. Quantified results of the indicated proteins. Data are shown as mean ± SD, * p < 0.05, **p < 0.01 (Student’s t-test).

KO of FOXO4 suppresses activation of astrocytes and microglia in the brain following I/R

To assess the reactive astrocytes and microglia, we performed immunostaining of the astrocyte marker, GFAP (glial fibrillary acidic protein), and microglia marker, Iba1 (ionized calcium-binding adaptor molecule 1), in the mouse brain 48 h following I/R. As shown in Fig. 4, the number of GFAP (Fig. 4a, 4c) and Iba1 (Fig. 4b, 4d) positively stained cells was significantly reduced in the KO brain compared to the WT brain, suggesting that FOXO4 deletion suppresses gliosis following I/R in the brain.

Figure 4. KO of FOXO4 reduces gliosis two days after I/R compared to WT mouse brains.

a. Representative images showing GFAP staining in the WT and FOXO4 KO brains. Scale bar, 50 μm. b. Representative images showing Iba1 staining in the WT and FOXO4 KO brain sections following I/R. Scale bar, 50 μm. c. Quantitation of GFAP positively stained cells (green). d. Quantitation of Iba1 positively stained cells (red). All numeric data are shown as mean ± SD; n = 7 for each group. *p < 0.05, **p < 0.01 (Student’s t-test).

KO of FOXO4 reduces leukocyte infiltration in the brain following I/R

Another pathological feature following I/R is leukocyte infiltration into the brain. To determine the leukocyte infiltration, the brain sections were immunohistochemically stained either with the CD45 (a marker for the total hematopoietic cells), CD11b (a marker for macrophages, granulocytes, and NK cells), or Ly6G (a marker for monocytes, granulocytes, and neutrophils). As shown in Fig. 5a and 5b, the number of CD45-positive-stained cells in the KO brain showed a reduction trend compared to the WT brain, despite the lack of statistical difference between the two groups. In contrast, the number of positively stained CD11b was significantly decreased (Fig. 5c and 5d) in the KO brain compared to the WT brain. Similarly, the number of positively stained Ly6G (leukocytes) was also significantly decreased (Fig. 5e and 5f) in the KO brain compared to the WT brain. These data indicate that KO of FOXO4 attenuates leukocyte infiltration.

Figure 5. KO of FOXO4 attenuates leukocyte infiltration following I/R.

a. Representative image showing CD45 staining. b. Quantitation of CD45-positively stained cells in the brain. c. Representative images showing CD11b staining. d. Quantitation of CD11b-positively stained cells. e. Representative image showing Ly6G staining. f. Quantitation of Ly6G-positively stained cells. Scale bar, 50 μm. Data are shown as mean ± SD; n = 5–7 for each group. ns, no significant difference; *p < 0.05, ***p < 0.001 (Student’s t-test).

Identification of Actinomycin D as a FOXO4 inhibitor

The results demonstrated above strongly suggest that disrupting FOXO4 function is beneficial for neuronal survival following oxidative stress or I/R. To translate the results into potential clinical treatments, we identified compounds that could directly inhibit the interaction between FOXO4 and its DNA substrate. The crystal structure of the FOXO4 protein is unavailable; however, the 1.9 Å resolution 3D crystal structure of the DNA-binding domain (DBD, amino acid sequence 93–196) of human FOXO4 bound to a 103 bp DNA duplex has been reported (Fig. 6a) (34). Additionally, the human FOXO4 DBD is identical to the mouse FOXO4. We used this FOXO4 DBD and performed virtual docking of several small-molecule libraries against FOXO4 DBD (Protein Data Bank, PDB: 3L2C). The top-scoring compounds were then docked onto FOXO1 DBD (PDB: 3COA) and FOXO3 DBD (PDB: 2UZK), and any candidates that also showed a high affinity to FOXO1 or FOXO3 were eliminated from the list. This led to Actinomycin D (ActD), an FDA-approved antibiotic, as one of the top candidates (Fig. 6b). ActD is used for chemotherapy to treat many different types of cancer(35) and was the first natural antibiotic found to have anti-cancer activity (36).

Figure 6. Identification of ActD as a FOXO4 inhibitor.

a. Structure of the FOXO4 DBD–DNA complex. The FOXO-DBD is shown in blue ribbon (upper), and the DNA double helix is shown as orange and green (lower) (https://www.rcsb.org/structure/3L2C). b. Structure of ActD. c. The result of a luciferase reporter assay shows that ActD inhibited FOXO4 activity. d. The result of a luciferase reporter assay shows that ActD did not inhibit FOXO3 activity. e. SwissDock-based AutoDock Vina execution exhibited active residues Arg78, His79, and Ser82 interacting with ActD. f. AlphaFold3-based Protenix predicted similar heavy atom contacts with ligand to H2 (Leu47), H3 (Asn75, Arg78, His79, Ser82 and Leu83) and S3 (Lys89) domains of FOXO4. All numerical data are shown as mean ± SD. N = 3, *p <0.05, ns, no significant difference (Student’s t-test).

In a cell-based reporter assay, ActD showed significant inhibitor activity against FOXO4 (Fig. 6c). Structurally, FOXO4 DBD is more similar to FOXO3 DBD than to FOXO1 DBD(37). However, our reporter assay revealed that ActD did not inhibit FOXO3 activity at the same concentration (0.1 μM) (Fig. 6d).

The results of computational protein-ligand docking advocated helix H3 as the major site of inhibition (Fig. S2). SwissDock-based AutoDock Vina execution exhibited active residues Arg78, His79 and Ser82 interacting with ActD (Fig. 6e). AlphaFold3-based Protenix(38) predicted similar heavy atom contacts with ligand to H2 (Leu47), H3 (Asn75, Arg78, His79, Ser82, and Leu83), and S3 (Lys89) domains of FOXO4 (Fig. 6f). It is important to note that the sequential numbering of these receptor residues is based on the active DBD of FOXO4 and does not correspond to the entire protein from the N to C terminal. Interaction of ActD with Arg78 is crucial as two hydrogen bonds are indicated (Fig. 6e) with donor-acceptor distances of 3.480 Å (NH₂) and 3.451 Å (NE), whereas the distance between ActD and Ser82 and Lys89 is 2.735 Å and 3.334 Å (Fig. 6e, 6f), respectively. The corresponding seven fundamental amino acid residues bound to the ligand are L124, N152, R155, H156, S159, L160, and K166 for FOXO4.

Pharmacological blockage of FOXO4 reduces tMCAO-induced brain injury and promotes functional recovery

To further assess the therapeutic effect of ActD on I/R-induced brain injury in mice following tMCAO, we treated WT mice with ActD (0.25 or 0.5 mg/kg, i.p.) or vehicle 1 h before tMCAO (Fig. 7a). Following 1 h of ischemia and 24 h reperfusion, mice were sacrificed for TTC staining (Fig. 7b). Our data showed that ActD treatment significantly reduced I/R-induced brain injury at 0.5 mg/kg (Fig. 7b, 7c). Moreover, ActD treatment also enhanced the animal function recovery, as shown by reduced mNSS scores (Fig. 7d) and increased cognitive function reflected by the Y-maze test compared to the vehicle treatment (Fig. 7e). These data indicate that ActD is neuroprotective against I/R-induced brain injury.

Figure 7. ActD reduces I/R-indued brain injury in the WT mice following tMCAO.

a. A schematic of experimental design. b. Representative TTC staining of mouse brains from the sham surgery, vehicle-treated, and ActD (0.25 or 0.5 mg/kg)-treated mouse brains after MCAO. c. Quantitation of TTC staining results. N = 12; **p < 0.01, ****p < 0.0001 (one-way ANOVA). d. ActD improved mNSSs compared to vehicle-treated mice after tMCAO. N = 14–15; *p < 0.05 (two-way ANOVA). e. ActD-treated mice showed better memory compared to vehicle-treated mice in the Y-maze test after tMCAO. N = 14–15; *p < 0.05 (Student’s t-test). All data are shown as mean ± SD.

Discussion

In this study, we investigated the role of FOXO4 in cerebral I/R-induced neuronal injury and validated that FOXO4 is a therapeutic target. Moreover, we identified ActD as a FOXO4 inhibitor that exhibited therapeutic effects against I/R injury. While the protective role of FOXO4 in non-neuronal cells and tissues has previously been reported in vitro and in vivo (39–41), to our knowledge, this is the first study to investigate the role of FOXO4 in I/R-induced neuronal injury in neurons and animal brains.

Unlike FOXO1 KO mice, which are embryonic lethal (42), FOXO4 KO mice do not differ from their WT littermates under normal conditions. Our findings revealed that FOXO4 KO mice showed improved outcome following cerebral I/R, with a significantly reduced infarct volume 24 h after I/R and improved functional recovery thereafter. FOXO4 KO also enhanced learning and memory abilities in mice one week after I/R. Furthermore, we observed that FOXO4 KO mice exhibited a significant reduction in inflammatory cytokines in the brain cortex 48 h after I/R.

Inflammatory cytokines originate from two sources: immune cells recruited from the systemic circulation and locally activated glial cells (43). We observed a significant decrease in CD11- and Ly6G-positive cells in the KO brain following I/R, indicating reduced leukocyte infiltration in the FOXO4 KO brain. Because CD11b is highly expressed by myeloid cells, which represent the first line of the immune response (44), the significant decrease in CD11b in the FOXO4 KO brain likely reflects reduced leukocyte infiltration through the microvessels.

Additionally, our results support reduced activation of local microglia and astrocytes in KO mice compared with WT mice. These findings are consistent with previous reports demonstrating the anti-inflammatory effect of FOXO4 deletion under ischemic conditions in another organ in vivo (45) and in cerebral endothelial cells in an in vitro I/R model (39). After I/R, ROS, cell debris, mitochondrial dysfunction, excitotoxicity, and a disruption of the blood-brain barrier (BBB) collectively trigger inflammatory responses involving both local and systemic immune factors. As a transcription factor, FOXO4 regulates multiple pathological processes, including apoptosis, endothelial function, ROS production, and BBB integrity, thereby facilitating cell death. In contrast, FOXO4 deletion leads to neuroprotective effects and slows the progression of neuronal death and infarction, ultimately preserving functional performance, as supported by our findings.

DBDs of transcription factors have traditionally been considered “undruggable” due to their relatively flat and similar surface areas (46). However, recent studies have challenged this notion by developing improved in silico screening approaches to eliminate potential non-selective inhibitors (47). For instance, top-scoring compounds targeting the STAT3 DBD identified through an in silico screening of a chemical library were subsequently docked onto the STAT1 DBD (48, 49). Compounds predicted to bind to the STAT1 DBD were excluded due to potential lack of selectivity, and this strategy enabled the identification of a STAT3-selective hit molecule (49).

Using a similar strategy, we identified a FOXO4 inhibitor, ActD, which is considered a potent, broad-spectrum transcription inhibitor that binds to double-stranded DNA at GC-rich sites, thereby blocking RNA polymerase progression and inhibiting RNA chain elongation (50). However, it may exhibit relative specificity to FOXO4, as our results suggest that it did not suppress FOXO3 activity at the same concentration in vitro. This finding indicates that ActD may exert a stronger inhibitory effect on FOXO4 than on other transcription factors. Importantly, ActD reduces infarct volume and enhances functional recovery following cerebral I/R in mice. These results suggest that ActD may have therapeutic potential for the treatment of ischemic brain injury.

Our studies have some limitations. First, only young adult male mice were used in our studies. Future research should consider including female and older mice in the studies. Second, the animals used here are genetically modified mice, and when tMCAO is applied, genetic alteration already exists. Third, the treatment of ActD in the tMCAO mice was done before I/R, and we did not test the therapeutic effects of the drug on tMCAO-induced neuronal injury after I/R. To translate the studies from bench to bedside, it is necessary to determine whether treating the tMCAO mice within the critical therapeutic window, 3–6 h after tMCAO, is still neuroprotective. Finally, as ActD is a broad-spectrum transcription inhibitor, identification of other FOXO4-specific inhibitors is needed.

In conclusion, using a genetically modified mouse model, we validated FOXO4 as a therapeutic target for I/R-induced brain injury. Our data suggest that KO of FOXO4 reduces neuronal injury, promotes functional recovery, suppresses leukocyte infiltration, and attenuates neuroinflammation. Moreover, we identified ActD as a potential therapeutic agent for I/R-induced neuronal death, which may be used at low doses in the clinic to treat acute ischemic stroke or other neurological disorders.