FTO promotes post-stroke neuroprotection by m⁶A demethylation of c-Jun

Abstract

N⁶-methyladenosine (m⁶A) is a critical epitranscriptomic regulator of neuronal function. Cerebral ischemia induces m⁶A hypermethylation due to decreased expression of m⁶A demethylase fat mass and obesity-associated (FTO) protein. Previously, we showed that cerebral overexpression of FTO with an adeno-associated virus (AAV) 9 protects the post-stroke brain. We presently evaluated the mechanistic basis for FTO-dependent m⁶A demethylation in post-ischemic neuroprotection using the mice transient middle cerebral artery occlusion model of experimental stroke. Based on the bioinformatic predictions and m⁶A abundance, pro-apoptotic transcription factor Jun proto-oncogene (c-Jun) with 19 m⁶A sites was chosen as an exemplary target. FTO overexpression normalized the post-stroke m⁶A hypermethylation of c-Jun without altering its transcript levels. FTO-dependent m⁶A demethylation suppressed translation of c-Jun. Consequently, several c-Jun target genes are transcriptionally repressed, and the post-ischemic neuronal apoptosis is decelerated, as seen by decreased cleaved caspase-3 levels and TUNEL⁺ neurons in the FTO AAV9 treated group compared to the control AAV9 treated group. Moreover, replenishing c-Jun precluded the FTO-mediated post-stroke neuroprotection and functional recovery. Collectively, this study demonstrated that the FTO/m⁶A/c-Jun axis ameliorates post-stroke neuronal apoptosis and brain damage, leading to better functional outcomes.

Introduction

Stroke remains the fifth leading cause of death in the United States, with limited therapeutic interventions. ¹ Complex post-stroke sequelae necessitate the identification of cerebroprotective targets capable of simultaneously modulating multiple pathological processes. ² The potential of such pleiotropic agents, including transcription factors, noncoding RNAs, and epigenetic mechanisms, in minimizing secondary brain damage and/or promoting functional recovery after ischemic stroke has been extensively investigated. ³ Emerging evidence implicated epitranscriptomic regulators as a new class of druggable targets for minimizing stroke pathophysiology.^4–7

RNAs undergo >175 epitranscriptomic modifications that modulate their stability, splicing, transport, localization, translation and degradation. ⁸ Methylation of adenosine at 6^th position (N⁶-methyladenosine; m⁶A) is the most prevalent epitranscriptomic modification in the CNS. The m⁶A methylation is mediated by well-orchestrated machinery composed of a writer complex (methyltransferase (METTL) 3 and METTL14), erasers (fat mass and obesity-associated (FTO) protein and alkB 5 homology protein) and readers (YTH domain family (YTHDF) 1, YTHDF2 and YTHDF3). ⁹ We and others have previously reported that experimental stroke in rodents downregulates FTO, with a concomitant m⁶A hypermethylation of RNAs in the brain.^10,11 By overexpressing FTO using an adeno-associated virus (AAV) 9, we further showed that FTO directly modulates post-ischemic m⁶A levels. Importantly, FTO overexpression yielded robust post-ischemic neuroprotection in both sexes. ⁴

Although FTO is known to demethylate ∼1,500 RNAs in neurons, our understanding of the functional significance of individual m⁶A hypermethylated transcripts to post-stroke outcomes is limited. ¹² Previously, we demonstrated inflammation and apoptosis as the major pathways enriched for the post-stroke m⁶A hypermethylated transcripts. ¹⁰ As transcripts with more m⁶A sites will have a higher affinity for m⁶A readers, they have a higher chance to influence the functional outcomes. ¹³ The pro-apoptotic transcription factor Jun proto-oncogene (c-Jun) mRNA harbors 19 highly enriched m⁶A sites and is one of the highly m⁶A hypermethylated transcripts in the ischemic brain. ¹⁰ Notably, m⁶A methylation of c-Jun is negatively implicated in endotoxemia-augmented venous thrombosis. ¹⁴ Using c-Jun as the putative target, we currently evaluated the role of FTO/m⁶A signaling in modulating post-stroke neuronal fate. Furthermore, we also assessed the conjoint effect of FTO and c-Jun on post-stroke brain damage and functional recovery.

Methods

Mice

Adult C57BL/6J male mice (12 weeks; 25 ± 3 g) were purchased from The Jackson Laboratory. All mice used in this study were housed 5 per cage in standard facilities with a 12-hour light/12-hour dark schedule in a temperature- and humidity-controlled vivarium and ad libitum access to food and water. All experimental procedures involving mice were approved by the Research Animal Resources and Care Committee of the University of Wisconsin-Madison. The mice were cared for in accordance with the Guide for the Care and Use of Laboratory Animals [U.S. Department of Health and Human Services Publication Number. 86-23 (revised)]. All experiments were conducted in compliance with the “Animal Research: Reporting of In Vivo Experiments” guidelines. ¹⁵ Mice were randomly allocated to study groups using the GraphPad random number generator tool. Neurological deficits were assessed with a modified neurological severity score based on the following criteria: 0 - no neurological deficit, 1 - inability to fully extend the right forepaw, 2 - turning to the right, 3 - circling to the right, 4 - inability to walk spontaneously, and 5 - death due to stroke. Mice that displayed no neurological deficits during the acute phase after surgery (5 males), or those that exhibited subarachnoid hemorrhage (2 males) or intracerebral hemorrhage (1 male) upon euthanasia were excluded from the study. The outcome parameters were assessed by a person who was blinded to the group assignments.

Transient middle cerebral artery occlusion

Focal cerebral ischemia was induced by intraluminal middle cerebral artery occlusion (MCAO) for 1 h using a 6-0 silicon-coated monofilament (Doccol Corporation, USA) under isoflurane anesthesia as described earlier.^16–18 The physiological parameters were monitored, and rectal temperature was maintained at 37°C during the surgery. Our previous study confirmed the lack of regional cerebral blood flow changes in mice treated with AAV9 before, during and after transient MCAO. ⁴ Mice that showed no signs of neurological deficits during the acute phase after surgery and that showed hemorrhage after euthanasia were excluded from the study. Sham-operated mice underwent a similar surgical procedure except for occlusion. Cohorts of mice were euthanized at 12 h, 1 day or 28 days after MCAO as needed in various experiments.

Bioinformatics and luciferase reporter assay

The SRAMP tool was employed to predict the m⁶A motifs (RRACH; R can be adenosine or cytosine or uracil, A is adenosine that can be modified to m⁶A, C is cytosine, and H can be guanosine or adenosine) in the 3′-UTR region of c-Jun. ¹⁹ The existence of m⁶A peaks at these predicted motifs in the mouse brain was confirmed from the previously deposited MeRIP-seq dataset using the REPIC tool. ²⁰ The wild-type c-Jun 3′-UTR luciferase reporter vector, a mutant c-Jun 3′-UTR luciferase reporter vector (with 8 m⁶A sites mutated) and FTO cDNA clone were obtained from GeneCopoeia USA. Rat pheochromocytoma (PC12) cells were co-transfected with either a wild-type reporter or a mutant reporter along with FTO or empty plasmid using Lipofectamine 2000 (Invitrogen). After 2 days, the culture medium was collected, and the luciferase activity and the secreted alkaline phosphatase activity were measured using Secrete-Pair Dual Luminescence Assay Kit (GeneCopoeia) according to the manufacturer’s instructions. Each transfection was conducted in triplicates.

Real-time PCR

One µg of total RNA was reverse transcribed to cDNA with High Capacity RNA to cDNA Kit (Applied Biosystems). The mRNA levels of c-Jun, Hsp72 and c-Jun targets (Bim, Bid, Nfkb1, Casp3, Bcl10, Traf1, FasL, TRAIL, MCP-1 and Tnf-α) were estimated by real-time PCR analysis with SYBR-green method using gene-specific primers (Supplementary Table 1). 18S rRNA was used as housekeeping control, and relative gene expression was calculated by the comparative C_t method (2^−ΔΔCt). Real-time PCR assays were conducted in triplicate.

Western blot

Protein samples (15 µg) were electrophoresed, transferred to a nitrocellulose membrane, blocked and probed with antibodies against c-Jun (1:1,000; Cell Signaling Technology), Cleaved Caspase-3 (1:1,000; Cell Signaling Technology) followed by HRP-conjugated anti-mouse or anti-rabbit IgG (1:3,000; Cell Signaling Technology). Blots were stripped and reprobed with an antibody against GAPDH (1:1,000; Santa Cruz Biotechnology) followed by HRP-conjugated anti-mouse IgG. Blots were visualized with ECL and quantified using Image Studio software (LI-COR Biotechnology).

FTO and c-Jun overexpression

AAV9 was produced by Vector Biolabs as described below. Mouse FTO and c-Jun genes were subcloned into pAAV cis-plasmid, pAAV cis-plasmid was scaled-up using Qiagen Endo-free Mega Prep kit, and the resulting plasmids were transfected into HEK293 cells. Recombinant AAVs were purified through a series of CsCl centrifugation steps and the viral titers were determined by quantitative real-time PCR. These recombinant AAV9 were injected 21 days prior to transient MCAO. Briefly, the mice were anesthetized and fixed into a stereotaxic frame (Harvard Apparatus), and a burr hole was drilled in the pericranium 2.5 mm lateral to the sagittal suture and 0.25 mm posterior to the coronal suture. Two µL of viral suspension containing a total viral load of ∼10¹¹ genome copies of FTO AAV9 or FTO + c-Jun AAV9 was injected into the cortex at a rate of 0.2 µL per minute using a Hamilton syringe positioned 1.5 mm below the cranium at an angle of 37° with sagittal plane. The needle was withdrawn after 10 min, and the wound was closed. The null AAV9 virus was used as a control for all the experiments.

Lesion, atrophy volume and white matter damage estimation

To assess the lesion volume at 7 days of reperfusion following transient MCAO, mice were anesthetized with isoflurane and subjected to T2-weighted MRI using a 4.7‐T small animal system with 205/120/HD/S gradient 210 mm bore varian magnet (Agilent Technologies, USA). Respiration rate was monitored during the imaging. 8–10 equidistant coronal slices/mice were acquired with a slice thickness of 1.0 mm. MRI scans were analyzed using NIH ImageJ software with an FDF plugin by a person blinded to the study groups. On day 28, mice were euthanized by transcardiac 4% paraformaldehyde perfusion. Each brain was post-fixed, cryoprotected and sectioned (coronal, 40 µm thick). Six serial sections (between +2.28 to −5.64 mm from bregma) were stained with 0.1% cresyl violet, scanned and analyzed using the NIH ImageJ software. The brain atrophy volume was calculated as the volume of the contralateral hemisphere – the volume of the ipsilateral hemisphere. For white matter damage evaluation, sections near the coordinate +0.5 mm from bregma were stained with 0.1% luxol fast blue (56°C for 12 h) followed by counterstaining with 0.1% cresyl violet (room temperature for 5 min). Stained sections were scanned, and corpus callosum thickness was computed by NIH ImageJ software.

Statistical analyses

For analyzing data that was collected repeatedly from the same set of mice at different time points (such as rotarod, beam walk and MWM tests), a nonparametric two-way repeated-measures ANOVA (Sidak post hoc correction) was used. The Mann-Whitney U test was used to compare two groups. A one-way ANOVA (Tukey post hoc correction) was used to compare multiple groups. GraphPad Prism 9 software was used for all the statistical analyses. All outcome measures were estimated blindly.

Results

FTO-dependent m⁶A signaling regulates c-jun expression in the post-stroke brain

Prediction of m⁶A sites among post-stroke m⁶A hypermethylated inflammatory and apoptotic transcripts using the SRAMP bioinformatics server identified c-Jun as a putative target with the highest number of RRACH motifs. ¹⁰ Of a total of 19 predicted m⁶A motifs in c-Jun, 9 are within the 5′-UTR, 2 are within the coding region and 8 are within the 3′-UTR region (Figure 1(a)). Moreover, 15 of these are high-confident sites and 3 are moderate-confident sites (Figure 1(a)). Using the MeRIP dataset from the REPIC database, we identified robust m⁶A enrichment at c-Jun loci in the mouse brain, indicating that the predicted m⁶A sites in c-Jun mRNA are bonafide (Figure 1(b)). To ascertain the role of FTO in regulating the fate of c-Jun, we introduced 8 synonymous point mutations (adenine replaced with thymine) at the putative m⁶A sites in the 3′-UTR region of c-Jun. The wild-type and mutant c-Jun 3′-UTR were cloned into Gluc/SEAP reporter system, and the luciferase activity was measured in empty control and FTO plasmid treated PC12 cells (Figure 1(c)). FTO treatment abrogated the luciferase activity of the wild-type c-Jun vector (by ∼30%, p < 0.05; n = 4/group), but not the mutant c-Jun vector, compared to the respective control-treated groups (Figure 1(c)). Thus, FTO negatively regulates the c-Jun expression in an m⁶A-dependent manner.

Figure 1.

Bonafide m⁶A motifs in c-Jun mRNA and its regulation by FTO. (a) c-Jun mRNA harbors 19 consensus RRACH sites within 5′-UTR, coding and 3′-UTR regions. High, low and medium confidence m⁶A sites are shown in red, green and blue, respectively. (b) MeRIP sequencing tracks from the REPIC genome browser show the peaks corresponding to predicted m⁶A sites. Each track represents an independent replicate of a mouse brain sample and (c) we used wild-type and mutant c-Jun 3′-UTR constructs. When cotransfected in PC12 cells, FTO plasmid decreased the luciferase activity of the wild-type, but not the mutant, c-Jun 3′-UTR reporter vector compared with the respective empty plasmid transfected control. Gaussia luciferase activity (Gluc) was normalized to secreted alkaline phosphatase activity (SeAP). Data are mean ± SD (n = 4/group). *p < 0.05 compared with wild-type c-Jun/empty plasmid control by Mann-Whitney U test.

To confirm the m⁶A-mediated regulation of c-Jun in vivo, we investigated the effect of FTO overexpression on the post-stroke c-Jun mRNA and protein levels, and m⁶A abundance. Transient middle cerebral artery occlusion (MCAO) significantly increased c-Jun mRNA expression in the peri-infarct cortex of the control AAV9 treated mice (by 4-fold, p < 0.05; n = 4/group) at 12 h of reperfusion compared with sham, which was not altered by FTO AAV9 treatment (Figure 2(a)). However, FTO AAV9 treatment prevented the post-ischemic m⁶A hypermethylation of c-Jun (by 68%, p < 0.05; n = 4/group) in the peri-infarct cortex at 12 h of reperfusion compared with the control AAV9 group (Figure 2(b)). Consequently, FTO AAV9 treatment abolished the post-ischemic induction of c-Jun protein expression (by ∼40%, p < 0.05; n = 4/group) in the peri-infarct cortex at 1 day of reperfusion compared with the control AAV9 group (Figure 2(c) and (d)). Binding of the m⁶A reader YTHDF1 promotes translation of the m⁶A modified transcripts. ²¹ FTO demethylates m⁶A-modified transcripts, and thus reduces YTHDF1 binding. Hence, FTO AAV9 treatment abrogated YTHDF1 binding to c-Jun mRNA (by ∼91%, p < 0.05; n = 4/group) in the peri-infarct cortex at 1 day of reperfusion compared with the control AAV9 group (Figure 2(e)). This indicates that FTO-mediated m⁶A demethylation of c-Jun decreases its translation without affecting the transcription (Figure 2(f)). Furthermore, it is a specific regulatory effect as the negative control Hprt mRNA did not show alterations in the post-ischemic m⁶A levels (Supplementary Figure 1A). YTHDF1 binding to Hprt mRNA was also not significantly different in either FTO AAV9 or control AAV9 cohorts compared to sham (Supplementary Figure 1B).

Figure 2.

FTO demethylated and decreased c-Jun protein expression after stroke. (a) Focal ischemia-induced c-Jun mRNA expression (estimated by real-time PCR) at 12 h reperfusion compared to sham that was not curtailed by FTO overexpression. (b) Post-ischemic m⁶A enrichment of c-Jun mRNA (assessed by real-time PCR following methylated RNA immunoprecipitation) was increased significantly at 12 h reperfusion compared with sham, which was curtailed by the FTO overexpression. (c and d) Focal ischemia also significantly increased the c-Jun protein levels at 24 h reperfusion compared with sham that was prevented by FTO overexpression. Immunoblots in c are from representative mice of each group. (e) Relative YTHDF1 binding to c-Jun mRNA (estimated by real-time PCR after YTHDF1 immunoprecipitation) was also significantly increased at 1 day of reperfusion following focal ischemia compared with sham, which was curtailed by FTO overexpression and (f) correlation between c-Jun transcription and translation confirmed that FTO overexpression prevents its translation but not transcription. mRNA and protein fold change were plotted as fold over sham. Values are mean ± SD (n = 4/group). *p < 0.05 compared with sham and ^#p < 0.05 compared with control AAV9 ischemia group by one-way ANOVA (Tukey post hoc correction).

FTO overexpression impeded c-jun mediated post-stroke neuronal apoptosis

c-Jun is an activator protein 1 (AP-1) family of the transcription factor, which is considered the master regulator of genes that promote neuronal apoptosis. ²² Specifically, AP1 consensus sequence (TGAC/GTCA) is identified in the promoters (2 kb upstream of transcription start site) of Bim (−519), Bid (−377), Nfkb1 (−992), Casp3 (−151), Bcl10 (−1917), Traf1 (−1756), FasL (−1678), TRAIL (−268), MCP1 (−41) and Tnf-α (−900). FTO AAV9 treatment significantly prevented the post-stroke induction of many c-Jun downstream pro-apoptotic genes, including Bim (by ∼49%), Bid (by ∼37%), Nfkb1 (by ∼47%), Casp3 (by ∼52%), Bcl10 (by ∼48%), Traf1 (by ∼52%), FasL (by ∼50%), TRAIL (by ∼52%), MCP-1 (by ∼61%) and Tnf-α (by ∼56%) in the peri-infarct cortex at 1 day of reperfusion compared to control AAV9 treated group (p < 0.05 in all cases; n = 4/group) (Figure 3). FTO AAV9 treatment had no effect on the post-ischemic induction of positive control Hsp72 compared to the control AAV9 group (Figure 3).

Figure 3.

FTO overexpression decreased the post-stroke transcription of apoptosis-related c-Jun target genes. c-Jun binding modulates its target gene expression. Real-time PCR analysis using peri-infarct cortex at 1 day of reperfusion following transient MCAO showed increased expression of Bim, Bid, Nfkb1, Casp3, Bcl10, Traf1, FasL, TRAIL, MCP1, and Tnf-α in the control AAV9 treated cohort which was significantly curtailed in the FTO AAV9 treated cohort. Expression of HSP72, which was used as a positive control was not significantly different between the control AAV9 and FTO AAV9 cohorts. 18S rRNA was used as housekeeping control. Data are mean ± SEM (n = 4/group). *p < 0.05 compared with sham and ^#p < 0.05 compared with control AAV9 ischemia group by one-way ANOVA (Tukey post hoc correction). AP1, activator protein 1; Bim, Bcl2-interacting mediator of cell death; Bid, BH3 interacting domain death agonist; Nfkb1, nuclear factor-kappa B subunit 1; Casp3, caspase 3; Bcl10, B-cell lymphoma/leukemia 10; Traf1, Tnf receptor-associated factor 1; FasL, Fas ligand; TRAIL, Tnf-related apoptosis-inducing ligand; MCP1, monocyte chemoattractant protein-1; Tnf-α, tumor necrosis factor-alpha; Hsp72, heat shock 10 KDa protein 2.

Furthermore, FTO AAV9 treatment significantly reduced the post-ischemic protein levels of the apoptosis marker cleaved caspase-3 (by ∼49%, p < 0.05; n = 4/group) (Figure 4(a)), which colocalized with the neuronal marker NeuN⁺ in the peri-infarct cortex at 1 day of reperfusion compared to the control AAV9 treated group (Figure 4(b)). FTO AAV9 treatment also reduced total TUNEL⁺ cells (by ∼43%, p < 0.05; n = 4/group) (Supplementary Figure 2(a) and (b)) as well as NeuN⁺ and TUNEL⁺ neurons (by 41%, p < 0.05; n = 4/group) (Supplementary Figure 2A and 2C) in the peri-infarct cortex at 1 day of reperfusion following 1 h transient MCAO compared to the control AAV9 treated group. Thus, FTO might decrease neuronal apoptosis after stroke potentially by blocking the c-Jun guided pro-apoptotic transcriptional cascade.

Figure 4.

FTO overexpression decreased the post-stroke cleaved caspase-3 levels. (a) Representative immunoblots and quantification of the cleaved caspase-3 expression in the cortical peri-infarct region of control AAV9 and FTO AAV9 treated mice at 1 day of reperfusion following transient MCAO compared to sham and (b) Representative images of cleaved caspase-3 immunostaining in the NeuN⁺ neurons. Images were taken from the cortical peri-infarct region indicated by the cresyl violet stained coronal brain sections at 1 day of reperfusion after transient MCAO. Data are mean ± SD (n = 4/group). *p < 0.05 compared with sham and ^#p < 0.05 compared with control AAV9 ischemia group, by one-way ANOVA (Tukey post hoc correction). DAPI, 4′,6-diamidino-2-phenylindole. Scale bars are 15 µm.

c-Jun overexpression abrogates FTO mediated post-stroke neuroprotection

To further validate if FTO mediated post-stroke neuroprotection is primarily due to c-Jun downregulation, we performed a reverse rescue experiment by concurrent overexpression of FTO and c-Jun. Injection of dual AAV9 (FTO+c-Jun) showed a robust and widespread distribution of both FTO and c-Jun in the ipsilateral cortex, striatum and hippocampus at 21 days (Figure 5(a)). Post-stroke motor function recovery was significantly worsened in the dual AAV9 (FTO+c-Jun) treated mice compared with FTO AAV9 treated mice (p < 0.05; n = 7/group) (Figure 5(b) and (c)). The dual AAV9 (FTO+c-Jun) treated mice also showed significantly increased infarct volume (by 2-fold, p < 0.05; n = 7/group) at 7 days of reperfusion following 1 h transient MCAO compared with the FTO AAV9 treated mice (Figure 5(d)). Post-stroke cognitive functional recovery was significantly worsened in the dual AAV9 (FTO+c-Jun) treated mice compared with FTO AAV9 treated mice (p < 0.05; n = 7–8/group) (Figure 6(a) to (c)). The dual AAV9 (FTO+c-Jun) treated mice showed significantly bigger lesion/atrophy volume (by ∼2-fold; p < 0.05) compared with FTO AAV9 vector treated cohort on day 28 of reperfusion (Figure 6(d)). The FTO AAV9 treated cohort showed significantly lower damage of the corpus callosum compared to the control AAV9 vector-treated cohort, which was nullified in the dual AAV9 (FTO+c-Jun) treated cohort on day 28 of reperfusion (Figure 6(e)). Altogether, this data indicates that c-Jun is a major downstream target of FTO in the post-stroke brain.

Figure 5.

c-Jun overexpression counteracted FTO-mediated post-stroke neuroprotection. (a) FTO and c-Jun co-staining illustrate the anatomical distribution of the transgenic dual AAV9 virus in mouse brain (between + 2.28 to −5.64 mm from bregma) 21 days following intracerebral injection (n = 4/group). Injection site is indicated by a white arrowhead and the magnified images show the cellular localization of FTO and c-Jun. DAPI, 4’,6-diamidino-2-phenylindole. Scale bar, 15 µm. (b and c) Motor function assessment by the rotarod (b) and beam walk (c) tests in FTO AAV9 and dual AAV9 (FTO + c-Jun) treated male mice. The performance was assessed the day before occlusion and on days 1, 3, 5, and 7 of reperfusion after 1 h transient MCAO. Data are mean ± SEM (n = 7/group). *p < 0.05 compared with the FTO AAV9 group by two-way repeated-measures ANOVA (Sidak post hoc correction) and (d) representative MRI scans and infarct volumes of FTO AAV9 and dual AAV9 (FTO + c-Jun) treated mice. Infarct volume was measured at 7 days of reperfusion after 1 h transient MCAO. Data are mean ± SD (n = 7/group). *p < 0.05 compared with the FTO AAV9 group by Mann-Whitney U test.

Figure 6.

c-Jun overexpression attenuated FTO-mediated post-stroke functional recovery and prevented gray and white matter protection. (a) Representative swimming tracks of the control AAV9, FTO AAV9 and dual AAV9 (FTO+c-Jun) treated mice during the probe trial of the Morris water maze test. (b) Learning curves based on the escape latency for the control AAV9, FTO AAV9 and dual AAV9 (FTO+c-Jun) treated mice during the training phase of the Morris water maze test. *p < 0.05 compared with the control AAV9 treated group and ^#p < 0.05 compared with the FTO AAV9 group by two-way repeated-measures ANOVA (Sidak post hoc correction). (c) Time spent in the platform quadrant by the control AAV9, FTO AAV9 and dual AAV9 (FTO + c-Jun) treated male mice during the probe trial phase of the Morris water maze test. *p < 0.05 compared with the control AAV9 group by Mann-Whitney U test. #p < 0.05 compared with the FTO AAV9 group by two-way repeated-measures ANOVA (Sidak post hoc correction). (d) Representative cresyl violet-stained coronal brain sections (between +2.28 to −5.64 mm from bregma) and atrophy volume of Control AAV9, FTO AAV9 and dual AAV9 (FTO + c-Jun) treated mice. Atrophy was estimated at 28 days of reperfusion after 1 h transient MCAO and (e) representative luxol fast blue-stained images (+0.5 mm from bregma) and white-matter damage quantification of control AAV9, FTO AAV9 and dual AAV9 (FTO + c-Jun) treated male mice. Corpus callosum thickness (medial) was measured at 28 days of reperfusion after 1 h transient MCAO to assess white-matter integrity. Data are mean ± SD (n = 7–8/group). *p < 0.05 compared with the control AAV9 group by Mann-Whitney U test. #p < 0.05 compared with the FTO AAV9 group by Mann-Whitney U test.

Discussion

The m⁶A RNA methylome undergoes significant remodeling after acute injury to the CNS. ^10,23–30 We previously showed that ischemic stroke induces m⁶A hypermethylation in >100 transcripts due to loss of FTO. ¹⁰ FTO-mediated m⁶A demethylation was shown to be involved in various post-stroke pathophysiological processes, including mitophagy, pyroptosis, oxidative stress, inflammation and vascular repair.^5–7,31,32 Thus, FTO might be responsible for providing the pleiotropic activity to thwart multiple post-stroke pathological events simultaneously.

In the present study, we chose c-Jun as an exemplary target to decipher the functional significance of m⁶A-methylated RNAs in the ischemic brain. The rationale for this choice is that upregulation of c-Jun is long known to mediate post-stroke pathophysiology, but the upstream mechanisms that regulate it remain enigmatic.^33–35 We observed that c-Jun mRNA harbors 19 bonafide m⁶A sites and is also one of the robustly hypermethylated RNAs in the ischemic brain. Following transient MCAO, the mRNA expression, m⁶A methylation and protein expression of c-Jun increased concurrently. However, when mice were treated with FTO AAV9, the m⁶A methylation and protein expression of c-Jun decreased without a change in the mRNA expression. Thus, m⁶A hypermethylation seems to promote either the translation or stability of c-Jun mRNA after stroke. The binding of the m⁶A reader YTHDF1 is known to induce translation of methylated RNAs as it directly interacts with the eukaryotic translation initiation factor 3. ²¹ Our data shows a marked increase in YTHDF1 binding to c-Jun mRNA following focal ischemia in the control AAV9-treated cohort, which was reduced by >90% in the FTO AAV9-treated cohort. Overall, FTO overexpression led to demethylation of c-Jun mRNA and prevention of its binding of YTHDF1, leading to decreased c-Jun translation after stroke. These findings posit m⁶A methylation as a major upstream regulator of c-Jun after stroke.

Post-transcriptional regulation of c-Jun by m⁶A is also implicated in other pathological conditions such as hypoxic stress, nerve injury, tumorigenesis and thrombosis.^14,27,36,37 Interestingly, m⁶A was shown to exert diverse effects on c-Jun, depending on the context. During hypoxia, m⁶A increases the stability of c-Jun without altering its protein expression. ³⁶ Whereas, during thrombosis, m⁶A promotes the translation of c-Jun in a YTHDF1-dependent manner, akin to cerebral ischemia. ¹⁴ This dichotomy might be originating from the differential site-selective methylation of c-Jun. Of note, none of the studies assessed the stoichiometry of m⁶A methylation at each individual site within c-Jun mRNA. Although we demonstrated that mutating 8 out of 19 consensus m⁶A sites located in the 3′-UTR region decreases c-Jun translation, further experiments using techniques such as site-specific cleavage and radioactive labeling followed by purification, exonuclease digestion, and thin-layer chromatography (SCARPET) are needed to determine the stoichiometry of methylation at individual m⁶A sites in the post-stroke brain. ³⁸ In addition to differential site-specific methylation of c-Jun, the binding affinity of other m⁶A binding proteins such as YTHDF 2 and 3 could also be assessed.c-Jun is a classic transcriptional inducer of neuronal apoptosis.^22,39,40 For instance, c-Jun inhibition impedes nerve growth factor withdrawal-induced neuronal apoptosis. ⁴¹ Alternatively, c-Jun overexpression is sufficient to drive apoptosis in primary neurons. ⁴² ChIP-seq data retrieved from the Encyclopedia of DNA Elements project showed ∼826 c-Jun target genes. ⁴³ Out of these, DAVID pathway analysis indicates that 40 genes are associated with apoptotic processes.^44,45 Our studies showed that FTO overexpression decreases post-stroke transcription of several key apoptosis-related c-Jun target genes, including Bim, Bid and Casp3. Furthermore, FTO overexpression decreased post-stroke neuronal apoptosis. These findings support the recent studies that implicated FTO in preventing the apoptosis of post-ischemic neurons and cardiomyocytes.^11,46 Moreover, FTO-mediated post-stroke neuroprotection, and motor and cognitive function recovery perished upon overexpression of c-Jun. This indicates that c-Jun is a major downstream target of FTO-dependent m⁶A signaling in the post-stroke brain and potentiates neuronal apoptosis after stroke. Interestingly, c-Jun also regulates the transcription of genes encoding neuropeptides, neurotrophins and cytokines. ⁴⁷ It will be interesting to explore the non-apoptotic mechanisms of the FTO/m⁶A/c-Jun axis in the post-stroke brain.

Multiple cell death mechanisms, including apoptosis, autophagy, ferroptosis, necroptosis, pyroptosis, phagoptosis, and parthanatos are implicated in post-stroke brain damage. ⁴⁸ Ferroptosis is an iron-dependent cell death pathway culminating in lipid peroxidation and reactive oxygen species accumulation. ⁴⁹ Emerging evidence highlights the crosstalk between m⁶A signaling and ferroptotic regulators after cerebral ischemia.^50–54 Notably, FTO overexpression decreased the Fe²⁺ and enhanced the anti-oxidant enzyme glutathione peroxidase 4 levels, ameliorating ferroptotic cell death in mice subjected to transient MCAO. ⁵¹ In a neonatal hypoxic-ischemic brain injury model, FTO was shown to inhibit neuronal ferroptosis by decreasing the stability of ferritin heavy chain 1 in m⁶A dependent manner. ⁵² Ubiquitin-specific peptidase 18 was found to stabilize FTO, thus causing demethylation and suppression of nuclear receptor coactivator 4, a master regulator of ferritinophagy after cerebral ischemia. ⁵³ Altogether, FTO/m⁶A axis displays moonlighting functions in post-ischemic neuronal death.

This is the first study to systematically dissect the mechanistic link between m⁶A methylation alterations and the post-stroke pathophysiological processes. Overall, we demonstrated that FTO antagonizes the pro-apoptotic milieu in the peri-ischemic cortex by epitranscriptomic remodeling of c-Jun (Figure 7). While these findings provide impetus to the therapeutic utility of FTO in minimizing post-stroke secondary brain damage and accelerating functional recovery, such an avenue still awaits the discovery of specific small molecule activators of FTO.

Figure 7.

FTO/m⁶A/c-Jun axis in post-stroke pathophysiology. Stroke-induced downregulation of FTO elevates m⁶A levels of c-Jun. YTHDF1 binds to m⁶A modified c-Jun and promotes its translation. Upregulated c-Jun enhances the transcription of its target genes involved in neuronal apoptosis, ultimately inducing neurologic deficits. In contrast, FTO overexpression normalizes the post-stroke m⁶A levels of c-Jun, decreases its translation, thereby dampens the apoptotic cascade. As a result, peri-ischemic neurons are salvaged, and secondary brain damage is minimized. Bim, Bcl2-interacting mediator of cell death; FasL, Fas ligand; TRAIL, Tnf-related apoptosis-inducing ligand; Tnf-α, tumor necrosis factor-alpha; MCP1, monocyte chemoattractant protein-1.

CRediT author statement

Anil K Chokkalla: conceptualization, methodology, investigation, formal analysis, writing- original draft. Suresh L Mehta: methodology, investigation, writing- review and editing. Soomin Jeong: methodology, investigation. Hui-Lung Sun: methodology, investigation. Qing Dai: methodology, investigation. Raghu Vemiganti: conceptualization, supervision, writing- review and editing, funding acquisition.

Data availability

All supporting data are available in the article and online-only Supplemental material.

Supplementary material

Supplemental material for this article is available online.