New insights into the central sympathetic hyperactivity post‐myocardial infarction: Roles of METTL3‐mediated m6A methylation

Abstract Ventricular arrhythmias (VAs) triggers by sympathetic nerve hyperactivity contribute to sudden cardiac death in myocardial infarction (MI) patients. Microglia‐mediated inflammation in the paraventricular nucleus (PVN) is involved in sympathetic hyperactivity after MI. N6‐methyladenosine (m6A), the most prevalent mRNA and epigenetic modification, is critical for mediating cell inflammation. We aimed to explore whether METTL3‐mediated m6A modification is involved in microglia‐mediated sympathetic hyperactivity after MI in the PVN. MI model was established by left coronary artery ligation. METTL3‐mediated m6A modification was markedly increased in the PVN at 3 days after MI, and METTL3 was primarily located in microglia by immunofluorescence. RNA‐seq, MeRIP‐seq, MeRIP‐qPCR, immunohistochemistry, ELISA, heart rate variability measurements, renal sympathetic nerve activity recording and programmed electrical stimulation confirmed that the elevated toll‐like receptor 4 (TLR4) expression by m6A modification on TLR4 mRNA 3'‐UTR region combined with activated NF‐κB signalling led to the overwhelming production of pro‐inflammatory cytokines IL‐1β and TNF‐α in the PVN, thus inducing the sympathetic hyperactivity and increasing the incidence of VAs post‐MI. Targeting METTL3 attenuated the inflammatory response and sympathetic hyperactivity and reduced the incidence of VAs post‐MI.


| INTRODUC TI ON
Sudden cardiac death (SCD) continues to be a major public health challenge, accounting for approximately 20% of all mortality in industrialized countries, 1 while ventricular fibrillation (VF) is considered the final underlying mechanism of deaths. [2][3][4] In myocardial infarction (MI) patients, 5,6 ventricular arrhythmias (VAs), primarily ventricular tachycardia (VT)/VF, are the leading causes of SCD.
Driven by this high-societal impact, profound investigation and understanding of the pathophysiology of ischaemia-induced VAs has attracted scientific efforts, aiming to further reduce the incidence of SCD.
Sympathetic activation, been known for decades, 7 plays a vital role in the pathogenesis of VAs during acute MI (AMI) by overwhelming local norepinephrine (NE) release, which contributes to the high mortality by lowering ventricular fibrillation thresholds in AMI [8][9][10][11] and is highly arrhythmogenic in the peri-infarct area. 12,13 Such local NE release from sympathetic nerve terminals is always elicited by central activation following ischaemia 13 by stimulating efferent myocardial nerve fibers. 12,14 The paraventricular nucleus (PVN), the key and advanced cardiovascular regulatory region of the brain, 15,16 is located in the hypothalamus of the central nervous system (CNS), receives and integrates incoming cardiac reflex information and regulates peripheral sympathetic activities via feedback of related neural signalling pathways. 17 Recent studies by our group and others have shown that microglia-mediated inflammation in the PVN stimulates sympathetic activation, thus leading to myocardial intrinsic sympathetic hyperactivity after MI. 11,[18][19][20] We have previously demonstrated that toll-like receptor 4 (TLR4) is highly expressed in microglia and induces sympathetic hyperactivity through the NF-κB signalling pathway post-MI. 21 However, little is known about the precise mechanisms of the post-transcriptional regulation of TLR4 expression regulation in the PVN post-MI.
N6-methyladenosine (m 6 A) is the most abundant mRNA and epigenetic modification 22,23 that occurs at the post-transcriptional level in mammals. m 6 A is installed onto target mRNA through m 6 A regulators which include the m 6 A methyltransferase complex (METTL3, METTL14, WTAP, VIRMA, RBM15 and ZC3H13) 24 and is enriched near the stop codon, 3′-untranslated region (3′UTR) 25 and long internal exon, primarily occurring in the RRACH sequence (where R = G or A; H = A, C or U) of mRNA. m 6 A was removed by demethylase (FTO and ALKBH5). 26 Emerging evidence has shown the biological and pathophysiological significance of m 6 A in human physiology and cancers by endowing controllable protein production 22,[27][28][29] and mRNA stabilization, 30 and by regulating RNA degradation, splicing, output and translation. 31 Recently, studies have reported the key role of m 6 A in mediating the inflammatory response in vitro and in vivo. For instance, METTL3 inhibited inflammation in response to lipopolysaccharides (LPS) by MAPK and NF-κB pathways in human dental pulp cells (HDPCs) 32 ; m 6 A attenuated inflammation of rheumatoid arthritis (RA) through the NF-κB signalling pathway. 33 However, whether m 6 A is involved in the inflammatory response and regulates the TLR4/NF-κB signalling pathway in the PVN post-MI is unknown so far.
This study demonstrated the pathophysiological roles of m 6 A methyltransferases, METTL3, involved in central sympathetic activation via promoting m 6 A installation post-MI. The molecular mechanism of m 6 A modification mediated by METTL3 was explored by identifying downstream target genes and signalling pathways.
Resultingly, the METTL3 gene in the PVN was knocked down by infection with adeno-associated virus (AAV)-mediated short hairpin RNA (shRNA). The effects of METTL3-mediated m 6 A modification on sympathetic activation were investigated. The results of this study further elucidate the related pathological processes and provide a novel therapeutic target for reducing VAs and SCD post-MI. The rat MI model was established as previously described, 34 with prior anaesthesia by intraperitoneal injection of 3% sodium pentobarbital (30 mg/kg). A fully anaesthetized state (no response to toe pinching) was confirmed, a tracheotomy was performed, and ventilation was achieved with a small-animal ventilator (RWD Life Science).

| Animal and MI model
The heart was exposed via left thoracotomy at the third and fourth intercostal space, and the left anterior descending coronary artery was ligated at 2-3 mm distance from the origin place between the pulmonary artery conus and the left atrium. The ligation material was a 6-0 polypropylene ligature. Electrocardiogram (ECG) data showed ST segment elevation, decreased of anterior wall motion of the left ventricle (LAD), and a mottled and pale appearance at the infarct areas, confirming the success of the infarction (Figure S1A, Figure 1B). In the sham group, the rat heart was exposed, but only threading, without ligation of the left coronary artery.

| PVN microinjection
PVN microinjection followed the previously reported procedures. 21 In brief, rats were secured in a stereotaxic apparatus (RWD Life Science) and cannulas were implanted in the bilateral PVN. The PVN coordinates were determined at 1.8 mm posterior, 0.4 mm lateral to the midline and 7.9 mm beneath the skull surface. A total volume of 1.0 μl microinjector was connected to the cannula with a PE-10 tube, and 50 nl control shRNA or METTL3 shRNA was injected into the PVN using a microinjector as designed. The injection rate was set at 0.1 μl/min by using an infusion pump. The microinjector was F I G U R E 1 Total m 6 A level and expression of m 6 A enzymes in the PVN of sham and MI groups. (A) Diagram of the experiment. (B) Quantification of total m 6 A was determined by an antibody-based colorimetric method in the PVN of the sham group and MI groups (at 1, 3, 5 and 7 days). m 6 A enzymes METTL3 was analysed at mRNA (C) and protein (D) levels in the PVN of the sham group and MI groups (at 1, 3, 5 and 7 days) by quantitative real-time PCR and Western blot analysis. m 6 A enzymes FTO was analysed at mRNA (E) and protein (F) levels in the PVN of the sham group and MI groups (at 1, 3, 5 and 7 days) by quantitative real-time PCR and Western blot analysis. GAPDH served as an internal control. (G) Heatmap of Pearson correlation analysis between total m 6 A levels and m 6 A-related enzymes using qRT-PCR data. Red and blue indicate the positive and negative correlations, respectively. The plots were plotted in the R package heatmap. n = 9 per group. Data are presented as mean ± SD. **p < 0.01 and ***p < 0.001 versus sham group. Abbreviations: MI, myocardial infarction; PVN, paraventricular nucleus; SD, standard deviation kept on site for 15 min to allow shRNA diffusion and avoid shRNA spillover. The microinjector was then removed. PVN location in rats was determined by microinjection of methylene blue 35 ( Figure S2A).

| Experimental design
Three independent experiments were conducted as following described.

| Protocol 1
Forty-five of fifty surviving rats were divided into five groups (sham, MI after 1, 3, 5 and 7 days; n = 9 per group). The temporal expression of m 6 A quantification in the total m 6 A level of PVN was detected using antibody-based colorimetric methods. The PVN total RNA and proteins were extracted for m 6 A-related enzymes detection using quantitative real-time PCR (RT-PCR) and Western blotting.

| Protocol 2
Twenty-four rats were randomly classified into two groups (sham, MI after 3 days; n = 12 per group). Subsequently, the PVNs were collected for RNA sequencing and m 6 A MeRIP sequencing.

| Protocol 3
One hundred ninety rats were randomly divided into four groups: sham + AAV2 with a green fluorescent protein (GFP) reporter and scrambled RNA, which served as the blank con-

| Heart rate variability measurements and cardiac echocardiography
A telemetry electrocardiograph transmitter (TR50B; ADInstrument) was positioned in the abdominal cavity of rats of as previously described. 21 Two leads were placed on the dorsal surface of the xiphoid process and into the anterior mediastinum, close to the right atrium. Twenty-four-hour dynamic electrocardiography was performed, and the electrocardiogram (ECG) data were continually recorded by a PowerLab physiology system and analysed using LabChart Pro software (ADInstruments). A 15 min ECG recording was selected to analyse HRV. Low-frequency (LF; 0.05-0.75 Hz) and high-frequency (HF; 0.75-2.5 Hz) bands were recorded, and the ratio of LF to HF bands was calculated. Telemetric recordings were conducted for 3 days before the devices were removed. After the rats were anaesthetized, cardiac function was measured by echocardiography machine (Fujifilm Vevo 3100). The left ventricular ejection fraction (EF%) and fractional shortening (FS%) were calculated using the M-mode recording of the parasternal long-axis view.

| RSNA recording
Renal sympathetic nerve activity (RSNA) recording was performed as previously described. 34 In brief, the left renal sympathetic nerves were separated using fine glass needles, and the distal terminus of the renal nerve was displaced. The central part of the nerve was placed on a pair of platinum electrodes, and the RSNA was then recorded using power-lab (ADInstruments). Background noise was detected after the renal sympathetic proximal section, and the experimental data baseline was corrected. The baseline level of RSNA was defined as 100% from the absolute value after subtracting the noise level. The data were analysed using LabChartPro software (AD Instruments).

| Programmed electrical stimulation
The rats underwent ventricular programmed electrical stimulation (PES) to acquire VA susceptibility. The PES protocol was as described in our previous study. 36 The rats were anaesthetized, and electrocardiography was performed with three electrodes placed on the upper limb and the right legs. A specially modified electrode was used to stimulate the left ventricle to study the incidence of VAs.
Standard PES protocols were performed as follows: burst (cycle length 100 ms, S0), single (S1), double (S2) and triple (S3) additional stimuli. The total PES protocol lasted for 10 minutes. The arrhythmia scoring was measured according to a previous report. 37

| Tissue collection
The rats were sacrificed via an overdose injection of 3% sodium pentobarbital. The heart and the whole brains were removed.
Extending the myocardium from the infarction scar to 0.5-1.0 mm was considered to represent the infarcted myocardium. The heart was excised, frozen in liquid nitrogen, then laterally sliced into 2 mm thick slices and incubated in 2% 2,3,5-triphenyltetrazolium chloride (TTC; Solarbio, Beijing, China) at 37°C for 30 min. After TTC staining, the surviving myocardium was red, and the infarct area was white. Image J software was performed for calculating infarct size as a percentage of LV. Rats with infarct size greater than 30% were selected based on clinical importance. Two different methods were used to prepare brain tissue for further study: (1) The PVN was collected from the entire brain tissue and stored at −80°C for biochemical analysis. (2) Entire brain tissues were fixed in 4% paraformaldehyde at 4°C for 24 h, then embedded in paraffin, and cut into 5 μm thick sections for immunohistochemistry and immunofluorescence. Blood samples were collected, and the serum was collected by centrifugation (1,200 × g, 20 min) at 4°C, and immediately stored in a −80°C freezer for future ELISA tests.

| m 6 A quantification
RNA-easy™ Isolation Reagent (Vazyme Biotech Co.) was used to extract total RNA from the PVN. m 6 A methylated RNA was measured by using an m 6 A RNA Methylation Quantification Kit (Colorimetric; EpiGentek). In brief, a sample of total RNA (200 ng) was used to the specifically capture and detect antibody binding. The m 6 A signal was quantified at a wavelength of 450 nm. Co.). The 2 −ΔΔCt method was used to calculate the relative targeted gene expression and normalized to GAPDH expression. Primer sequences were listed in Table 1.

| RNA-seq and m 6 A MeRIP-seq
RNA sequencing and MeRIP sequencing were performed simultaneously (Sinomics Corporation). The PVN tissues were collected, and total RNA was extracted by TRIzol reagent (Invitrogen).
RNA-seq: Strand-specific libraries were prepared using a TruSeq ® Stranded Total RNA Sample Preparation kit (Illumina, CA, USA) according to the manufacturer's instructions. An Illumina HiSeq X Ten machine (Illumina) was used to perform RNA-seq. Data analyses were performed as previously described. 38 MeRIP-seq: The RNA was segmented and incubated with m 6 A antibody (1:1000,202003, Synaptic Systems, Germany). Sequencing was conducted on an Illumina Novaseq™6000 machine (LC-Biotechnology Co.). Data analyses were performed as previously described. 39

| Methylated immunoprecipitation-qPCR (MeRIP-qPCR)
Experiments were performed as previously described. 40 Total RNA was extracted from PVNs and then conjugated in IP buffer with protein A/G magnetic beads, containing m 6 A antibody (1:1,000, Synaptic Systems) and anti-IgG, supplemented with RNase inhibitor and protease inhibitor overnight at 4°C. Relative enrichment was calculated using the 2 −ΔΔCt method compared with the input sample.
Primer sequences for MeRIP-qPCR were presented in Table 1.

| Western blotting
Nuclear and cytoplasmic protein isolation from the PVN was performed using a nuclear and cytoplasmic extraction kit (Beyotime Biotechnology). Nuclear proteins were used for NF-κB p65 expression analysis, and cytoplasmic proteins were used for METTL3, METTL14, WTAP, FTO, ALKBH5, TLR4 and IκBα analysis. The protein lysate concentration was measured using a BCA kit (Elabscience).
Proteins (approximately 50 µg) were subjected to 6%-15% SDSpolyacrylamide gels and transferred onto polyvinylidene difluoride membranes (PVDF). After blocking with 5% non-fat milk diluted in TBST for 1 h, and the membranes were then incubated with dif-

| Immunohistochemistry
The expression of central sympathetic proteins was detected by immunohistochemistry. Central sympathetic activity was measured by

| Immunofluorescence staining
Immunofluorescence staining was performed by using a multiplex immunofluorescence staining kit (abs50012, Absin) as previously described. 41 The tissue sections were dewaxed in xylene, dehydrated in alcohol, repaired by microwave and blocked with 5%

| Enzyme-linked immunosorbent assay (ELISA)
The levels of IL-1β, TNFα and the excitatory neurotransmitter norepinephrine (NE) were detected using an ELISA kit (Elabscience). The OD value at 450 nm was used to determine protein concentrations by the end. NE in serum was recorded in pg/ml, and myocardium tissue concentrations of NE were recorded in pg/mg.

| Statistics analysis
Data are presented as the mean ±SD. GraphPad Prism 8.0 (GraphPad Software) was used to analyse the data. Two-tailed unpaired Student's t tests were used for two groups. Two-way analysis of variance (ANOVA) followed was performed to analyse the three or more groups. p < 0.05 was considered statistically different.  Figure 1A showed the diagram of the experiment. High inflammatory response can still be detected 2-3 days after AMI, which gradually decreases to normal low level after about 2 weeks. 42

| RE
In summary, we chose 3 days after MI for further study.

| TLR4/NF-κB signalling pathway as the downstream target of m 6 A modulation in the PVN
To define the potential downstream targets and associated pathways for METTL3-mediated m 6 A modification in the PVN at day 3 of MI, RNA-seq and m 6 A MeRIP-seq were performed. Figure 2A showed a diagram of the experiment. A total of 11265 genes were found and plotted in the volcano graph ( Figure 2B) using RNA-seq, in which 101 and 100 genes were found to be significantly upregulated and downregulated, respectively, comparing with the sham group (log2 Fold Change >1 and p < 0.05; log2 Fold Change ≤1 and p < 0.05). Regarding to the methylation level of genes, 100 genes were significantly hypermethylated (log2 Fold Change >1 and p < 0.05) and 99 genes were significantly hypomethylated (log2 Fold Change ≤1 and p < 0.05) by m 6 A MeRIP-seq analysis ( Figure 2C). To find out methylation state of these highly expressed genes, we plotted a Venn diagram to demonstrate the cross-section of highly expressed genes and hypermethylated genes with m 6 A ( Figure 2D). Seven genes were highly expressed and hypermethylated, accounting for 3.6% of the total analysed genes.
The seven genes were listed in Figure 2E. Fold change of TLR4 ranked the first in the seven genes and m 6 A modification occurred at 3'UTR ( Figure 2E) indicating the pathophysiological significance of TLR4 gene.
To define the biological function of m 6 A modification, KEGG pathway was analysed ( Figure 2F). The enriched hypermethylated genes were predominantly related to the NF-κB signalling pathway.

| METTL3-mediated m 6 A modification activated TLR4/NF-κB signalling pathway in the PVN
To explore the pathophysiological significance of METTL3-mediated

| METTL3 was mainly expressed in microglia of PVN post-MI
Double staining of Iba 1 (red) and METTL3 (green) using immunofluorescence was performed to locate METTL3 in microglia (Figure 4). Iba 1-positive staining was measured to quantify the degree of inflammatory microglial infiltration after MI. Higher Iba 1 expression was observed in the MI groups ( Figure 4C, D, E). However, Iba 1 was slightly attenuated in the MI + shMETTL3 group compared with the MI + shCtrl group (p > 0.5, Figure 4D, E, not significant). But METTL3 was increased significantly in the MI + shCtrl group and decreased in the MI + shMETTL3 group ( Figure 4D, F). In the initial stages of MI inflammation in rats, almost all METTL3 was expressed in microglia in the PVN region.

| METTL3 inhibition attenuated sympathetic hyperactivity and improved cardiac function post-MI
To investigate the functional roles of METTL3 in sympathetic nervous activity, the relevant measurements of the central and peripheral nervous systems were quantified. Figure

| DISCUSS ION
In the present study, we studied regulatory effects of m 6 A methylation modification in sympathetic hyperactivity and the con-  human dental pulp cells (HDPCs). 33 We will also investigate whether METTL3 acts on other molecules or inflammatory pathways to Note: Data are presented as mean ± SD.
F I G U R E 7 Schematic figure showing METTL3-mediated m 6 A modification was discovered in the microglia of PVN in MI rats and enhanced m 6 A installation on TLR4 mRNA 3'UTR region was mediated by METTL3. The elevated TLR4 expression by m 6 A modification on TLR4 mRNA 3'-UTR region combined with activated NF-κB signalling led to the overwhelming production of pro-inflammatory cytokines IL-1β and TNFα in the PVN, thus inducing the central and peripheral sympathetic activation, and increasing the incidence of VAs after MI. Abbreviations: MI, myocardial infarction; PVN, paraventricular nucleus; VAs, ventricular arrhythmias affect sympathetic activation after MI in future. In addition, GAPDH was not a reliable normalizer for qPCR validation of genomic copy number variants because it overlaps highly homologous segmental duplications. We should be more cautious in future study. All the data sets were obtained from acute MI animal model, which differs from the natural myocardial ischaemia, we should be cautious when taking the configuration into human. Targeting METTL3 of human PVN would be challenging in future. If not, delivering shRNA-METTL3 systemically would be out of target since METTL3 is commonly expressed in mammal cells.

| CON CLUS IONS
We demonstrated that METTL3/m 6 A was upregulated in microglia of PVN in MI rats, m 6 A installation on TLR4 mRNA 3'UTR region was increased mediated by METTL3 and enhanced TLR4 expression in the PVN combined with activated NF-κB signalling led to the overwhelming production of pro-inflammatory cytokines during MI, consequently, sympathetic activation initiated arrhythmia after MI ( Figure 7). Targeting METTL3 attenuated VAs and protected cardiac function. METTL3 could be a potential therapeutic candidate for reducing inflammation post-MI.

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.