N6-methyladenosine-modified miR-873 promotes receptor-interacting protein kinase 3-mediated necroptosis after intracerebral hemorrhage in mice and HT22 cells

Abstract

Background

Intracerebral hemorrhage (ICH) causes a severe form of stroke characterized by high morbidity, mortality, and long-term disability. Neuronal cell death is influenced at the posttranscriptional level. Certain microRNAs influence neuronal cell death at the posttranscriptional level by regulating receptor-interacting protein kinase 3 (RIPK3), a key mediator of necroptosis. The specific mechanism by which miR-873 mediates neuronal necroptosis following ICH remains unclear. Epigenetic abnormalities, particularly N6-methyladenosine (m6A) modification, are increasingly recognized as critical contributors to ICH pathophysiology.

Methods

ICH mice model was established, followed by intracerebroventricular injection for gene manipulation. Brain water content was measured to assess cerebral edema. Neurological function was evaluated using the Morris water maze and neurological deficit scoring. Molecular and cellular analyses included Western blotting, quantitative real-time PCR, immunofluorescence, and luciferase reporter assays. Primary neuronal cultures, plasmid construction, and m6A RNA methylation quantification were performed to investigate underlying mechanisms. Differential gene expression was analyzed using microarray profiling, and data were statistically evaluated with appropriate analytical methods.

Results

The m6A modification is upregulated and positively involved in the functional role of miR-873 in ICH. miR-873 rescued necroptosis in ICH. miR-873 targets RIPK3. RRACH (R = G or A; H = A, C, or U) m6A sequence motifs predominantly contribute to the m6A modification of miR-873. The m6A modification regulates necroptosis in ICH. Knockdown of methyltransferase-like 3 improved the neurological function prognosis of ICH in mice.

Conclusion

m6A modification modulates miR-873 expression, thereby influencing RIPK3-mediated necroptosis in ICH. These findings provide potential therapeutic targets for mitigating neuronal injury after hemorrhagic stroke.

Video Abstract

Video Abstract.

Download video file ^{(78.3MB, mp4)}

Introduction

Intracerebral hemorrhage (ICH) is the cause of 10–15% of all strokes, which are usually because of cerebral vessel pathology. This condition exhibits a high mortality and disability rate [1,2]. Currently, treatment for ICH concentrates on the perihemorrhagic edema and secondary injury. Despite these efforts, survivors often experience prolonged and severe neurological impairment [3–5]. The intricate process of ICH development is closely linked to N6-methyladenosine (m6A) modification [6].

The m6A modification is more extensive in the nervous system than in other organs, is essential to brain development and adult brain function [7], and is one of the most prevalent alterations found in eukaryotic messenger RNAs (mRNAs), constituting over 80% of RNA base methylations [8–10]. This modification is a reversible process, orchestrated by a set of enzymes comprising m6A ‘writers’ (methyltransferases), m6A ‘erasers’ (demethylases), and m6A ‘readers’ (m6A-binding proteins). These enzymatic actors exhibit the capability to add, remove, or recognize m6A-modified sites [11–13].

Methyltransferase-like 3 (METTL3) is one of the most critical component proteins of the methylation transferase complex, mainly responsible for catalyzing the m6A modification of RNA molecules on N6-methyladenine [14]. Deficiency or abnormal expression of METTL3 inevitably disrupts the m6A levels within intracellular RNA, thereby precipitating various diseases [15]. m6A modification leads to increased expression of N-methyl-D-aspartate receptor 1 and elevated oxidative stress, thereby contributing to dopaminergic neuronal death [16]. METTL3 conditional gene knockout causes cerebellar hypoplasia [17], and METTL3 is capable of mediating m6A modification of the adenylate within the GGAC motif on primary microRNA (pri-miRNA). Expression of METTL3 is essential for the proper processing of most pri-miRNAs to mature miRNAs [18]. And heterogeneous nuclear ribonucleoproteins A2B1 (HNRNPA2B1) is one of the members of the HNRNP family that are known to participate in the processing of precursor mRNA [19,20]. HNRNPA2B1 plays a crucial role in the alternative splicing of precursor mRNA. Its overexpression promotes pri-miRNA processing within the nucleus, thereby increasing the levels of precursor miRNA (pre-miRNA) and miRNA [21].

miR-873 suppresses the translation of receptor-interacting serine/threonine-protein kinase 1 (RIPK1)/receptor-interacting protein kinase 3 (RIPK3), thereby regulating necroptosis and mitigating myocardial injury following ischemia and reperfusion injury [22]. At the posttranscriptional level, some miRNAs mediate neuronal cell death by regulating the expression of RIPK3, which is a key molecule in necroptosis. The mechanism of miR-873-mediated neuronal necroptosis after ICH injury is unclear. Our research identified an increase in m6A modification expression after cerebral hemorrhage. In addition, we found a set of aberrantly expressed miRNAs in the brain tissue of mice after ICH, with miR-873 being one of the most significantly down-regulated miRNAs. And RRACH (R = G or A; H = A, C, or U) m6A sequence motifs predominantly contribute to the m6A modification of miR-873. We explored the mechanism of miR-873 mediating necroptosis after ICH injury. Inhibition of m6A regulators, METTL3 and HNRNPA2B1, restored miR-873 expression and attenuated RIPK3-mediated necroptosis. These findings uncover a novel m6A–miRNA axis contributing to neuronal injury in ICH.

We aimed to investigate the involvement of m6A and miR-873 in neuronal necroptosis following ICH injury. By identifying molecular targets within this pathway, we strive to discover novel interventions for brain hemorrhage injury and contribute to the advancement of ICH treatment.

Materials and methods

Animals

All animal experiments were performed according to the European Union Directive for animal experiments and approved by the Animal Ethics Committee of Wannan Medical College (LLCS-2021-081). Eight-week-old male C57BL/6 mice (about 22 g) were maintained in a temperature-controlled environment with a 12-h light/dark cycle (five mice in a cage).

Mice intracerebral hemorrhage model and intracerebroventricular injection

Mice were anesthetized with isoflurane. The cranium was exposed until the bregma was visible. A 0.5-mm cranial burr hole was drilled 2 mm lateral to the right and 0.5 mm posterior to the bregma. A microliter syringe needle was inserted 4 mm from the skull surface, and 0.5 μl of saline containing 0.075 U collagenase IV was injected into the right basal ganglia. Next, the drill hole was sealed with bone wax. Ten days before ICH, 2 μl (5 nmol/2 μl) of HNRNPA2B1 in-vivo small interfering (siRNA) (RiboBio, Guangzhou, China) or 2 μl of METTL3 short hairpin (shRNA) (Hanbio, Shanghai, China; 1 × 109 TU/ml) was injected into the right cerebral ventricle at the another place (1.0 mm lateral to the bregma and 2.7 mm in depth) within 10 min.

Measurement of brain water content

Brain edema was measured in mice using the wet/dry method. Mice were euthanized and decapitated under deep anesthesia at 24 h after ICH. Then the brains were removed at once and divided into three parts: right brain (ipsilateral), left brain (contralateral), and cerebellum. Brain tissues were quickly weighed on an electronic analytical balance to obtain the wet weight, then the dry weight was obtained after drying for 24 h at 100 °C. The following formula was used for calculating brain water content: [(wet weight − dry weight)/wet weight] × 100%.

Morris water maze

Twenty days after ICH, the mice’s learning and spatial memory were tested via the Morris water maze. A circular pool (76 cm in depth and 122 cm in diameter) with a submerged platform (1.5 cm beneath the water surface) whose location could be identified only using various visual cues (different colors and sizes) on the swimming pool wall. These visual cues were kept constant during the whole experiment. Animals (n = 5, per group) were placed in the pool to use visual cues to find the submerged platform within 90 s and were then subjected to 5 consecutive days of training. On the sixth day, the platform was removed, and mice were released randomly from one of the four compass locations. All data were collected by use of Smart 3.0 software (Panlab, Spain) and calculating the target crossing times and percent distance in the target quadrant.

Neurological deficit evaluation

Garcia test: the modified Garcia test was used to assess the neurological function of experimental mice. This test (normal score: 0; maximal deficit score: 21) includes spontaneous movement tests, sensory tests, mustache touch test, extension test, lateral rotation test, and height climbing tests. The mice (n = 10, per group) were evaluated before ICH and 12, 24, and 48 h after ICH. Corner turn test: mice were guided into a 30° corner, and the test was repeated 20 times for each mouse. The score of left turns/total turns was then calculated. The mice (n = 10, per group) were evaluated before ICH and 12, 24, and 48 h after ICH.

Western blot analysis

The concentration of relevant proteins was quantified with a bicinchoninic acid protein assay kit (Beyotime Biotechnology, Shanghai, China). 10% SDS-PAGE separated proteins from each sample, and then transferred to nitrocellulose filter membranes. These membranes were then blocked with 5% skimmed milk for 2 h at room temperature and incubated overnight at 4 °C with the following primary detection antibodies: rabbit anti-mouse METTL3 mAb (1 : 1000; Abcam, Shanghai, China), rabbit anti-mouse HNRNPA2B1 mAb (1 : 1000; Affinity, Jiangsu, China), rabbit anti-mouse RIPK3 mAb (1 : 1000; ABclonal, Wuhan, China), and rabbit anti-mouse RIPK1 mAb (1 : 1,000; ABclonal) at 4 °C overnight. Species-matched secondary antibodies were subsequently applied and incubated for 2 h at room temperature. The results were analyzed by use of ImageJ software.

Quantitative real-time PCR analysis

Total RNA from cells or mice perihematomal brain tissue used in this study was extracted with TRIzol reagent (Invitrogen, Shanghai, China) according to the manufacturer’s instructions. First-strand cDNA was synthesized using the First-Strand cDNA Synthesis Kit (TIANGEN, Beijing, China). Then, according to the SYBR Green method, quantitative real-time PCR for mRNA and miRNA was performed on a Real-Time Fluorescence PCR Instrument (#ABI Thermo Fisher Scientific, Massachusetts, USA). U6 and glyceraldehyde-3-phosphate dehydrogenase were used as an endogenous control for miRNA and mRNA detection. The relative changes in RNA expression were calculated using the 2^−∆∆CT method.

Immunofluorescence

Frozen sections of brain tissue or HT22 cells (which were seeded into the 6-well culture plate) were used for immunofluorescence staining. After treatment, tissue sections and cells were incubated overnight with rabbit anti-mouse METTL3 mAb (1 : 1000; Abcam) and rabbit anti-mouse HNRNPA2B1 mAb (1 : 1000; Affinity). The tissue sections and cells were washed with phosphate-buffered saline and incubated for 2 h with Cy3-conjugated goat anti-rabbit IgG (1 : 100, GB21303; Servicebio, China). Then, after washing in phosphate-buffered saline, the samples were stained with 4',6-diamidino-2-phenylindole (G1012; Servicebio). All staining results were observed using a confocal microscope (FV3000; Olympus, Japan).

Luciferase reporter assay

HT22 cells were seeded into 24-well plates and incubated for 24 h until reaching 50–60% confluency. The wild-type m-RIPK3-3′UTR (m-RIPK3-3′UTR wt) and mutant m-RIPK3-3′UTR (m-RIPK3-3′UTR mut) containing the putative binding region of miR-873-5p were subcloned into Psi-Check2 Dual-luciferase vectors (Shanghai, China). The PC-m-RIPK3-3′UTR wt and PC-m-RIPK3-3′UTR mut were designed and purchased from Hanbio. The luciferase reporter plasmids were co-transfected into HT22 cells with either miR-873-5p mimics or the negative control. The luciferase signals were detected with the Dual-Luciferase Reporter Assay System (Promega, Beijing, China).

Cell culture and transfection

Murine hippocampal HT22 cells were purchased from Shanghai Biological Technology Co., Ltd. (Shanghai, China), for enzyme research. And the HT22 cells were cultured at 37 °C in 5% CO₂ atmosphere. Hemin stimulation in HT22 cells was used as an ICH model in vitro. MiR-873-5p mimics and miR-873-5p inhibitor, negative control, and negative control inhibitor were purchased from Hanbio. siRNA specific targeting HNRNPA2B1 (HNRNPA2B1 siRNA) and its NC were also purchased from Hanbio. siRNA specific targeting METTL3 (METTL3 siRNA) and its negative control were purchased from RiboBio. Then, the Lipofectamine 3000 (Thermo Fisher Scientific) was used and transfected into the cells following the manufacturer’s protocol.

Plasmid construction and N6-methyladenosine RNA methylation quantification

To construct the m6A-knockdown cell line, m6A modification sites (RRACH) on miR-873 were mutated. Mmu-miR-873 plasmid, mmu-miR-873 mut plasmid, and its NC were purchased from Hanbio.

According to the manufacturer’s instructions, the levels of total m6A were measured by the m6A RNA Methylation Quantification Kit (Colorimetric; Ab185912; Shanghai, China). Each sample analysis was performed using 200 ng of poly-A-purified RNA.

Microarray analysis

miRNA microarray assays were performed with the help of KangChen Bio-tech (Shanghai, China). Acquired array images were analyzed by the use of Agilent Feature Extraction software (version 11.0.1.1). Data normalization and subsequent processing were performed with the GeneSpring GX v12.1 software package (Agilent Technologies). Heat maps and volcano plots were performed using the R scripts.

Statistical analysis

Data were expressed as the mean ± SD. All data were tested for normal distribution. Student’s t-test or one-way analysis of variance test was performed to analyze differences between groups. Differences were considered to indicate statistical significance when the P value was less than 0.05.

Results

The N6-methyladenosine modification is upregulated and positively involved in the functional role of miR-873 in intracerebral hemorrhage

Increased expression of m6A and METTL3 were found in ICH (Fig. 1a). The global m6A level of the ICH mice was higher than that of the normal mice (Fig. 1b). PCR results demonstrated that the METTL3 level of the ICH patients was higher than that of the normal patients (Fig. 1b). METTL3 levels peaked at 24 h following ICH, while HNRNPA2B1 levels peaked at 12 h post-ICH (Fig. 1c and d).

Fig. 1.

The m6A modification is upregulated and positively involved in the functional role of miR-873 in ICH. (a) Immunofluorescent images of m6A (red) and METTL3 (green) in the Sham and Hemin groups (n = 3) of the HT22 cells at 24 h. Scale bar = 50 μm. (b) The global m6A level in the brain mRNAs of the normal group (n = 6) compared with that in ICH mice (n = 6). qRT-PCR analysis of METTL3 mRNA expression in ICH patients (n = 15) compared with healthy controls (n = 15). GAPDH served as an internal control. (c) and (d) The relative level of METTL3 and HNRNPA2B1 at 12, 24, and 48 h in the Sham and ICH groups (n = 3) of mice. (e) Heatmap showing differentially expressed genes in ICH mice (n = 3) and healthy controls (n = 3) at 24 h. (f) qRT-PCR of miR-873 and pri-miR-873 expression in the HT22 cells (n = 6) compared with that in healthy controls (n = 6) at 24 h. (g) qRT-PCR of miR-873 and pri-miR-873 expression when knocking down of METTL3 in the HT22 cells at 24 h. (h) qRT-PCR of miR-873 and pri-miR-873 expression when knocking down of HNRNPA2B1 in mice at 24 h. P values are from Student’s t-test or one-way ANOVA test. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05 compared with Sham. ^####P < 0.0001, ^###P < 0.001, ^##P < 0.01 compared with Hemin + NC or ICH + NC. ANOVA, analysis of variance; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HNRNPA2B1, heterogeneous nuclear ribonucleoprotein A2B1; ICH, intracerebral hemorrhage; m6A, N6-methyladenosine; METTL3, methyltransferase-like 3; mRNAs, messenger RNAs; qRT-PCR, quantitative real-time PCR.

The microarray analysis revealed significant differences in the expression levels of 21 miRNAs between ICH and normal brain tissue across three pairs of samples. miR-873 emerged as one of the most markedly down-regulated miRNAs in ICH (Fig. 1e). Comparative analysis revealed a substantial increase in pri-miR-873 levels alongside a significant decrease in miR-873 levels in cells exposed to Hemin (Fig. 1f). These results indicate that knockdown of METTL3 in cells significantly decreased pri-miR-873 levels while simultaneously increasing miR-873 levels. Similarly, knockdown of HNRNPA2B1 in mice led to reduced pri-miR-873 levels and elevated miR-873 levels. (Fig. 1g and h).

miR-873 rescued necroptosis in intracerebral hemorrhage

After ICH in mice, the protein levels of RIPK1 and RIPK3 increased significantly compared with those in the sham group. RIPK1 levels and RIPK3 levels peaked at 24 h (Fig. 2a and b). PCR analysis further showed that miR-873 mimics and inhibitors exerted significant effects in in-vitro cell experiments (SI 1c). The expression of RIPK1 and RIPK3 decreased following miR-873 mimics transfection, whereas it increased after miR-873 inhibitor transfection in cells exposed to hemin (Fig. 2c–f).

Fig. 2.

miR-873 rescued necroptosis in ICH. (a) and (b) The relative level of RIPK1 and RIPK3 at 12, 24, and 48 h in the Sham and ICH groups of mice (n = 3). (c) and (d) The relative level of RIPK1 and RIPK3 in the Sham, Hemin, Hemin + NC, and Hemin + miR-873 mimics groups of the HT22 cells (n = 3) at 24 h. (e) and (f) The relative level of RIPK1 and RIPK3 in the Sham, Hemin, Hemin + NC, and Hemin + miR-873 inhibitors groups of the HT22 cells (n = 3) at 24 h. P values are from Student’s t-test or one-way ANOVA test. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05 compared with Sham. ^###P < 0.001, ^##P < 0.01 compared with Hemin + NC. ANOVA, analysis of variance; RIPK1, receptor-interacting serine/threonine-protein kinase 1; RIPK3, receptor-interacting protein kinase 3.

miR-873 targets receptor-interacting protein kinase 3

We identified binding sites for miR-873 in the mRNA sequences of RIPK3 and RIPK1 by use of the miRBase database (Fig. 3a). The luciferase reporter assay revealed that co-transfection of miR-873 mimics with the m-RIPK3-3′UTR wt plasmid significantly decreased luciferase activity. Conversely, there was no significant difference in luciferase activity between co-transfection of miR-873 mimics and m-RIPK3-3′UTR mut plasmids (Fig. 3b). Red-labeled miR-873 and green-labeled RIPK3 were observed in a punctate pattern around the nucleus. The merged images demonstrated that red-labeled miR-873 colocalized with green-labeled RIPK3, suggesting a potential interaction between the two (Fig. 3c).

Fig. 3.

miR-873 targets RIPK3. (a) Binding sites of miR-873 in RIPK3 and RIPK1 mRNA in miRTarbase database. (b) Luciferase activity in the co-transfection of miR-873 mimics and the m-RIPK3-3′UTR wt plasmid compared with that in the co-transfection of miR-873 mimics and m-RIPK3-3′UTR mut plasmids of the HT22 cells (n = 3). (c) Immunofluorescent images of miR-873 (red) and RIPK3 (green) in the Sham and ICH groups of mice (n = 3) at 24 h. Scale bar = 50 μm. P values are from Student’s t-test or one-way ANOVA test. ***P < 0.001 compared with NC + RIPK3 wt, NC + RIPK3 mut and miR-873 mimic + RIPK3 mut. ANOVA, analysis of variance; mut, mutant; RIPK1, receptor-interacting serine/threonine-protein kinase 1; RIPK3, receptor-interacting protein kinase 3; wt, wild-type.

RRACH N6-methyladenosine sequence motifs interfered with the function of miR-873

We constructed expression vectors containing miR-873 point mutants, wherein the four methylated adenine residues within the m6A motifs were replaced with guanine (A–G transition mutation), while utilizing the wild-type miR-873 RNA molecule as a control. These constructs were then transfected into HT22 cells (Fig. 4a and b). Mutation of the m6A motifs led to increased expression levels of RIPK1 and RIPK3 compared with the unmutated group (Fig. 4c and d).

Fig. 4.

RRACH m6A sequence motifs interfered with the function of miR-873. (a) Schematic representation of miR-873 point mutant plasmid construction. (b) Four rach (R = G or A; H = A, C or U) Schematic representation of m6A sequence motif mutations. (c) and (d) The relative level of RIPK1 and RIPK3 in the Hemin, Hemin + pcDNA, Hemin + miR-873, and Hemin + miR-873 mut groups of the HT22 cells (n = 3) at 24 h. P values are from Student’s t-test or one-way ANOVA test. ****P < 0.0001 compared with Hemin + pcDNA. ^###P < 0.001, ^#P < 0.05 compared with Hemin + miR-873. ANOVA, analysis of variance; pcDNA, plasmid DNA; RIPK1, receptor-interacting serine/threonine-protein kinase 1; RIPK3, receptor-interacting protein kinase 3.

The N6-methyladenosine modification regulates necroptosis in intracerebral hemorrhage

Western blot analysis revealed that knockdown of METTL3 led to decreased expression levels of HNRNPA2B1, RIPK3, and RIPK1 post ICH, both in vitro (Fig. 5a and b) and in vivo (Fig. 5d and e). These findings were further confirmed by PCR (Fig. 5c). Moreover, knockdown of HNRNPA2B1 led to decreased expression levels of RIPK3 and RIPK1 post ICH, both in vitro (Fig. 5f and g) and in vivo (Fig. 5h and i).

Fig. 5.

The m6A modification regulates necroptosis in ICH. (a) and (b) The relative level of HNRNPA2B1; RIPK1 and RIPK3 in the Sham, Hemin, Hemin + NC, and Hemin + METTL3 siRNA groups of the HT22 cells (n = 3) at 24 h. (c) qRT-PCR of HNRNPA2B1, RIPK1, and RIPK3 expression in the Sham, Hemin, Hemin + NC, and Hemin + METTL3 siRNA groups of the HT22 cells (n = 3) at 24 h. (d) and (e) The relative level of METTL3, HNRNPA2B1, RIPK1, and RIPK3 in the Sham, ICH, ICH + NC, and ICH + METTL3 shRNA groups of mice (n = 3) at 24 h. (f) and (g) The relative level of RIPK1 and RIPK3 in the Sham, Hemin, Hemin + NC, and Hemin + HNRNPA2B1 siRNA groups of the HT22 cells (n = 3) at 24 h. (h) and (i) The relative level of RIPK1 and RIPK3 in the Sham, ICH, ICH + NC, and ICH + HNRNPA2B1 in-vivo siRNA groups of mice (n = 3) at 24 h. P values are from Student’s t-test or one-way ANOVA test. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05 compared with Sham. ^#### P < 0.0001, ^###P < 0.001, ^##P < 0.01, ^#P < 0.05 compared with Hemin + NC or ICH + NC. ANOVA, analysis of variance; HNRNPA2B1, heterogeneous nuclear ribonucleoprotein A2B1; ICH, intracerebral hemorrhage; m6A, N6-methyladenosine; METTL3, methyltransferase-like 3; mRNAs, messenger RNAs; qRT-PCR, quantitative real-time PCR; RIPK1, receptor-interacting serine/threonine-protein kinase 1; RIPK3, receptor-interacting protein kinase 3; shRNA, short hairpin RNA; siRNA, small interfering RNA.

Knockdown of methyltransferase-like 3 improved the neurological function prognosis of intracerebral hemorrhage in mice

We evaluated hematoma volume and behavioral outcomes following ICH (SI 1a). Mice in the ICH + METTL3 shRNA groups exhibited significantly shorter latencies compared to the ICH groups over 5 consecutive days of acquisition training. During the test on day 6, the ICH + METTL3 shRNA groups crossed the platform more times and covered a greater distance to locate the original platform position within 90 s compared to the ICH groups. These results strongly suggest that knockdown of METTL3 mitigated spatial learning and memory deficits following ICH (Fig. 6a–d). The modified neurological Garcia score indicated a marked improvement in neurological function in the ICH + METTL3 shRNA groups compared to the ICH groups (Fig. 6e and f). In addition, brain water content in the ipsilateral hemisphere significantly increased in the ICH groups. However, the ICH + METTL3 shRNA groups exhibited a significant reduction in brain water content in the ipsilateral hemisphere compared to the ICH groups (Fig. 6g).

Fig. 6.

Knockdown of METTL3 improved the prognosis of ICH in mice. (a) Swim traces in probe trials. (b) Escape latency in training trials in the Sham, ICH, ICH + METTL3 shRNA groups of mice (n = 5). (c) and (d) Target crossing and distance in the target quadrant in probe trials in the Sham, ICH, and ICH + METTL3 shRNA groups of mice (n = 5). (e) and (f) Garcia test and Corner turn test in the Sham, ICH, ICH + NC, and ICH + METTL3 shRNA groups of mice (n = 5). (g) Brain edema at 24 h in the Sham, ICH, ICH + NC, and ICH + METTL3 shRNA groups of mice (n = 5). P values are from Student’s t-test or one-way ANOVA test. ****P < 0.0001, ***P < 0.001, *P < 0.05 compared with Sham. ^####P < 0.0001, ^#P < 0.05 compared with ICH or ICH + NC. ANOVA, analysis of variance; ICH, intracerebral hemorrhage; METTL3, methyltransferase-like 3; shRNA, short hairpin RNA.

Discussion

m6A modification plays a vital role in the nervous system. However, the mechanisms underlying its effects in ICH remain to be elucidated. Previous studies have demonstrated that METTL3 and HNRNPA2B1 can interact with pri-miRNAs at m6A modification sites, thereby influencing pre-miRNA processing and subsequently impacting the levels of mature miRNAs [21]. Our results showed an increase in the expression of pri-miR-873 but a decrease in the expression of miR-873 in ICH. Knockdown of METTL3 or HNRNPA2B1 resulted in increased expression of miR-873 but decreased the expression of pri-miR-873. The paradoxical phenomenon we observed suggests a dysregulation in the miRNA biogenesis pathway, potentially driven by competitive inhibition at the pre-miRNA stage. We hypothesize that this phenomenon may be attributed to the elevated levels of pre-miR-873 caused by the overexpression of pri-miR-873 following ICH, leading to competitive inhibition of mature miR-873 within the nucleus. It is plausible that this competitive inhibition is alleviated when the expression of pre-miR-873 decreases within a specific range, thereby facilitating the exit of miR-873 from the nucleus.

miRNA dysregulation often correlates with disease progression in diseases; pre-miRNA molecules directly compete with mature miRNAs for binding to target mRNAs [23]. Elevated pre-miR-21 levels were shown to sequester mature miR-21, disrupting its repressive effects on transforming growth factor beta receptor 2 mRNA translation and contributing to epileptogenesis [23]. In neuronal contexts, pre-miRNAs may act as ‘decoys’, occupying RNA-binding proteins or target sites that would otherwise facilitate mature miRNA processing or function. This competition could explain why pri-miRNA accumulation does not translate into increased mature miRNA activity. The Drosha/DGCR8 complex is critical for pri-miRNA cleavage. QKI5 binds to a pri-miR-124-1 enhancer element ~300 nt upstream of the stem-loop, promoting efficient processing during erythropoiesis [24]. This complex is highly sensitive to structural and sequence changes. Dysregulation of such RNA-binding proteins could lead to aberrant pri-miRNA retention and stalled processing. These results suggesting that competitive inhibition at the pre-miRNA stage could contribute to miRNA network dysfunction.

The inverse correlation between pri-miRNA and mature miRNA levels highlights a critical layer of miRNA regulation. By further clarifying the competitive mechanism and processing inhibition mechanism of pre-miRNA, it may provide a direction for understanding the dysregulation of miRNA in neurological diseases. The pre-miRNA processing is not merely a passive step but involves active competition between binding proteins and processing enzymes. The alteration in the proportion of pre-miRNAs may have a dominant effect on the function of mature miRNAs. Perturbations in pri-miRNA processing could underlie miRNA dysregulation in neurological disorders. The reciprocal regulation between epigenetic modifiers and processing machinery underscores a critical layer of posttranscriptional control, where dynamic equilibrium between transcriptional activation and processing efficiency determines functional miRNA output.

Current studies have found that the sites modified by m6A have specific structural features, and a significant loss of these particular structural features will lead to the loss of m6A modification. The RRACH m6A motif structure domain is recognized by writer and reader in cells [25]. Since miRNAs in ICH act within the RRACH m6A motif, this effect may alter the stability and structural conformation of these specific structures, making the m6A site easier to recognize and bind. The increase in miRNA target transcription of m6A site recognition may lead to a higher association of these transcripts with necroptosis.

We aimed to determine whether the potential functions of m6A-modified miR-873 are related to RRACH motifs. We constructed expression vectors containing miR-873 point mutants. The expression levels of RIPK1 and RIPK3 increased upon mutation of the m6A motifs compared to the unmutated group. However, we found that compared to no mutations, the expression levels of RIPK1 and RIPK3 were lower than expected. We, therefore, suspect either that other miRNAs regulate necroptosis or that there are additional methylation sites involved in the regulation of necroptosis.

We demonstrate that METTL3 and HNRNPA2B1 collaborate to positively regulate the processing of pri-miR-873 in an m6A-dependent manner, resulting in decreased miR-873 expression and subsequently increased expression of downstream necroptosis. This discovery may offer potential prognostic and therapeutic targets for the treatment of ICH.

Acknowledgements

This research was supported by the National Natural Science Foundation of China (82471358, 82171329, 81701161, and 81874280), research project supported by the Shanxi Scholarship Council of China (2024-144), Outstanding Youth Scientific Research Project of Anhui Universities (2023AH020051 and 2024AH010046), Clinical and Translational Research Project of Anhui Province (202427b10020002), Natural Science Foundation of Zhejiang Province (LQ21H170002), and the Project of Wuhu Health Commission (grant no.: WHWJ2023y006). The central government guides local funds for science and technology development (project no.: YDZJSX2024D069).

All authors contributed to the study conception and design, commented on previous versions of the manuscript, and read and approved the final manuscript. J.W., S.W., and X.W. performed animal model establish, Western blot, and cell culture. B.L. and J.H. performed neurobehavior test. D.Y. performed high sequencing analysis and verification. S.Z. and Z.C. conceived and designed the experiments. Y.Y. and Y.X. analyzed the data and wrote the manuscript.

All animal experiments were in accordance with the European Union Directive for animal experiments and approved by the Animal Ethics Committee of Wannan Medical College (LLCS-2021-081). Informed consent was obtained from all individual participants included in the study.

Conflicts of interest

There are no conflicts of interest.