METTL4-Mediated Mitochondrial DNA N6-Methyldeoxyadenosine Promoting Macrophage Inflammation and Atherosclerosis

Abstract

BACKGROUND:

Mitochondrial dysfunction is a key factor in the development of atherogenesis. METTL4 (methyltransferase-like protein 4) mediates N6- methyldeoxyadenosine (6mA) of mammalian mitochondrial DNA (mtDNA). However, the role of METTL4-mediated mitoepigenetic regulation in atherosclerosis is still unknown. This study aims to investigate the potential involvement of METTL4 in atherosclerosis, explore the underlying mechanism, and develop targeted strategies for treating atherosclerosis.

METHODS:

Expression levels of mtDNA 6mA and METTL4 were determined in atherosclerotic lesions. We explored the mechanism of METTL4 involvement in atherosclerosis using Mettl4^Mac-KO-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice and cell models, as well as bone marrow transplantation. Natural compound libraries were screened to identify potent METTL4 antagonists. In addition, bioinspired proteolysis targeting chimera technology targeting macrophages within plaques was used to increase the efficacy of the METTL4 antagonist.

RESULTS:

The expression levels of mtDNA 6mA and METTL4 were significantly increased in plaque macrophages. Mettl4^Mac-KO-Apoe^-/- mice displayed suppressed mtDNA 6mA levels and atherosclerotic progression, which were reversed by METTL4 restoration through bone marrow transplantation (n=6). Mechanistically, elevated METTL4 expression reduces mitochondrial ATP6 (MT-ATP6) expression by suppressing its transcription, thereby impairing the activity of mitochondrial respiration chain complex V. This disruption leads to the accumulation of excess protons in the mitochondrial intermembrane space, causing mitochondrial dysfunction. Consequently, mtDNA is released into the cytoplasm, ultimately triggering inflammasome activation. All results were reversed by the mutation in the METTL4 methyltransferase active site. Mettl4^MUT-Apoe^-/- mice showed suppressed mtDNA 6mA levels and atherosclerotic progression and repaired mitochondrial function of macrophage, which were reversed by METTL4 restoration through bone marrow transplantation (n=6). Pemetrexed was identified as the first METTL4 antagonist to effectively alleviate atherosclerotic progression. Furthermore, we generated a proteolysis targeting chimera drug based on pemetrexed that specifically targeted METTL4 in macrophages within plaques, showing a promising therapeutic effect on atherosclerosis.

CONCLUSIONS:

This study revealed a novel mechanism by which mtDNA 6mA orchestrated mitochondrial function–related gene expression in macrophages, thereby promoting atherosclerosis. Through various experimental techniques, such as gene manipulation, pharmacological inhibition, and proteolysis targeting chimera, this study demonstrated that mtDNA 6mA and its specific enzyme METTL4 hold potential as therapeutic targets for atherosclerosis.

Clinical Perspective.

What Is New?

• We first revealed a novel mechanism by which the mitochondrial DNA N6-methyldeoxyadenosine regulates mitochondrial function in macrophages and atherosclerosis.

• The MT-ATP6 N6-methyldeoxyadenosine mediated by METTL4 (methyltransferase-like protein 4) in macrophages led to the accumulation of excess protons in the mitochondrial intermembrane space, reactive oxygen species burst, mitochondrial DNA leak, and inflammasome activation, contributing to atherosclerosis.

• We identified pemetrexed as the first METTL4 antagonist with a potent inhibitory effect on METTL4 methyltransferase and alleviation of atherosclerosis.

What Are the Clinical Implications?

• The mitochondrial DNA N6-methyldeoxyadenosine modification by METTL4 may be a potential therapeutic target for the treatment of atherosclerotic cardiovascular diseases.

• The administration of pemetrexed and a proteolysis targeting chimera drug based on pemetrexed targeting METTL4 in macrophages effectively decreased mitochondrial DNA N6-methyldeoxyadenosine levels and prevented atherosclerosis progression.

• Overexpression of METTL4 in macrophages within atherosclerotic plaques may be an independent risk factor for atherosclerosis.

As the metabolic signaling centers, mitochondria play a vital role in cellular metabolism, stress responses, and homeostasis maintenance, thereby dictating the fate of cells. Mitochondrial disorders contribute to diverse and complex diseases, including cancer, Alzheimer disease, aging, and cardiovascular diseases.¹ Emerging research has revealed that mitochondrial-to-nuclear communication through mitochondrial metabolites produced by energy metabolism reprogramming orchestrates the genomic epigenetic landscape, balancing cellular homeostasis and disease process.² Therefore, mitochondria-dependent pathways may represent an attractive therapeutic target for ameliorating human diseases.

Although lipid-lowering therapy and surgery are the most common therapeutic approaches for atherosclerotic cardiovascular disease, they can still lead to serious sequelae, resulting in a lower quality of life and a shorter lifespan. The pathogenesis of atherosclerosis is complicated and multifactorial, and its specific mechanisms remain unclear. Mitochondrial dysfunction plays a significant role in initiation and progression of atherosclerosis by exacerbating oxidative stress, inflammatory response, apoptosis, and cholesterol accumulation.^3,4 Targeting mitochondrial dysfunction is a novel therapeutic strategy for atherosclerotic cardiovascular disease.^5,6

Human double-stranded circular mitochondrial DNA (mtDNA) has been assembled independently of the nuclear genome, encompassing 37 genes responsible for encoding 13 proteins involved in oxidative phosphorylation, humanin micropeptide, 22 transfer RNAs, and 2 ribosomal RNAs.⁷ Recent evidence accumulated over the last decade has identified that as an independent risk factor, a reduction in the mtDNA copy number and mtDNA damage are associated with atherosclerotic cardiovascular disease outcomes.^8,9 Therefore, protection against mitochondrial dysfunction from the perspective of mtDNA is of great significance for atherosclerosis therapeutics.

Epigenetic regulation of mtDNA (mitoepigenetics) represents an emerging and rapidly advancing field of research.¹⁰ Although RNA N6-methyladenosine has been extensively studied and plays a significant role in chromatin architecture, transcriptional regulation, and genome stability, investigations of DNA N6-methyldeoxyadenosine (6mA), particularly mtDNA 6mA, are rare.¹¹ The existence of the mtDNA 6mA was identified only in 2020 with the development of sequencing technology. Hao et al first discovered that 6mA modification exhibited a significant enrichment in mtDNA at an 8000-fold higher level than that in nuclear DNA content.¹² The investigation of the function and regulatory mechanisms of 6mA in the mitochondrial genome is currently in its early stages. Furthermore, the involvement of mtDNA 6mA in human diseases, including atherosclerosis, remains poorly understood.

METTL4 (methyltransferase-like protein 4), widespread in eukaryotes, belongs to the MT-A70 (AdoMet-binding subunit of mRNA N6-adenosine-methyltransferase) family along with the traditional RNA N6-methyladenosine methyltransferases METTL3 and METTL14, with a conserved methyltransferase structural domain.¹³ It catalyzes U2 small nuclear RNA N6,2'-O-dimethyladenosine methylation to regulate pre-mRNA splicing.¹⁴ Furthermore, METTL4 can mediate nuclear DNA 6mA methylation, thereby promoting tumor metastasis.¹⁵ METTL4 has recently been acknowledged as predominantly localized in the mitochondria and demonstrated to act as the 6mA methyltransferase within mammalian mitochondria.¹² METTL4 is currently the only known enzyme that could initiate mtDNA 6mA methylation. However, whether METTL4 plays a role in atherosclerosis through mtDNA 6mA remains unclear.

In this study, we found that METTL4 induced mitochondrial dysfunction promoting atherosclerotic progression through mtDNA 6mA. Meanwhile, we screened out the first METTL4-specific antagonist pemetrexed (PEM), a common antitumor agent, which compensated the defects of potent antagonist targeting METTL4 to date. Further, we develop the proteolysis targeting chimera (PROTAC)–PEM targeting METTL4 degradation more selectively for macrophages alleviating atherosclerosis. Through these investigations, the study aimed to advance the understanding of atherosclerosis progression from the perspective of mitoepigenetics and to identify potential therapeutic avenues for targeting METTL4 in the treatment of atherosclerosis.

METHODS

The data that support the findings of this study are available from the corresponding author upon reasonable request. We used the ARRIVE (Animal Research: Reporting of In Vivo Experiments) checklist when writing our report.¹⁶ More detailed materials and methods are available in the Supplemental Material.

Experimental Animals

All animal experiments were approved by the Animal Care and Use Committee of Nanjing Medical University.

Statistical Analysis

All experiments were repeated in at least 3 independent experiments. Data are expressed as average±SEM. Statistical analyses were performed using the GraphPad Prism Software Version 9.0. When the data were normally distributed using the Shapiro-Wilk normality test, the differences among groups were evaluated using the unpaired 2-sided Student t test and 1-way ANOVAs followed by Tukey multiple comparisons test. In case of unequal variances, unpaired 2-sided Student t test with Welch correction and Brown-Forsythe and Welch ANOVA test followed by Dunnett T3 multiple comparisons test was used. When the data did not pass the normality test, nonparametric Mann-Whitney U test and Kruskal-Wallis test followed by Dunn multiple comparisons test were used. To test the effect of 2 independent factors, 2-way ANOVA followed by Tukey multiple comparisons test was performed. P<0.05 was considered statistically significant.

RESULTS

METTL4 Is Involved in the Progression of Atherosclerosis

To investigate the potential involvement of mtDNA 6mA in the development of atherosclerosis, we stimulated human aortic endothelial cells, human aortic smooth muscle cells, and human monocyte–derived macrophages (HMDMs) with ox-LDL (oxidized low-density lipoprotein) in vitro to establish atherosclerotic cellular models. Our findings revealed a significant increase of mtDNA 6mA levels in ox-LDL–stimulated HMDMs compared with the control group (Figure 1A and 1B; Table S1). However, there was no change in ox-LDL–stimulated human aortic smooth muscle cells compared with the control group, as well as human aortic endothelial cells (Figure S1A). These provide the evidence of the potential role of macrophage mtDNA 6mA in atherosclerosis pathogenesis. In our study, we found that 6mA was more enriched in endogenous mtDNA than in total DNA derived from ox-LDL–stimulated HMDMs (Figure 1C; Figure S1B). Therefore, ox-LDL effectively increased mtDNA 6mA in macrophages.

Figure 1.

METTL4 (methyltransferase-like protein 4) is involved in the progression of atherosclerosis. A and B, N6-methyldeoxyadenosine (6mA) dot blot (A) and MethylFlash m6A DNA Methylation ELISA Kit (B) analysis of mitochondrial DNA (mtDNA) 6mA levels in human monocyte–derived macrophages (HMDMs) treated with or without oxidized low-density lipoprotein (ox-LDL, 50 μg/mL, 24 hours). n=6 per group. C, 6mA dot blot analysis of mtDNA and total DNA in ox-LDL–stimulated HMDMs. n=6 per group. D, Western blot analysis of the subcellular localization of METTL4 protein in HMDMs treated with or without ox-LDL. VDAC (mitochondria) and H3 (nucleus) were selected as organelle-specific marker proteins. n=6 per group. E, Super-resolution fluorescence imaging the colocalization of METTL4 (red) with mitochondria (TOMM20, green) in ox-LDL–stimulated HMDMs with Scr or siMETTL4. n=6 per group. F, 6mA dot blot analysis of mtDNA 6mA levels in HMDMs transfected with Scr or siMETTL4 followed by ox-LDL stimulation. n=6 per group. G, Western blot analysis of the levels of METTL4 in nonatherosclerotic (Non-AS) and atherosclerotic (AS) arteries derived from patients. n=6 per group. H, Western blot analysis of the levels of METTL4 in the arteries derived from Apoe^-/- mice fed with normal chow (NC) or high-fat diet (HFD). n=6 per group. I, RT-qPCR (quantitative reverse transcription polymerase chain reaction) analysis of Mettl4 and inflammatory factors (MCP-1, IL-1β, and TNF-α) in the athero-prone (lower curvature, LC) and athero-protective (greater curvature, GC) regions of atherosclerotic lesions derived from HFD-fed Apoe^-/- mice. n=6 per group. J, Western blot analysis of METTL4 in human aortic smooth muscle cells (HASMCs), human aortic endothelial cells (HAECs), HMDMs, and bone marrow–derived macrophages (BMDMs) treated with or without ox-LDL. n=6 per group. K, Immunofluorescence analysis of METTL4 (green) and macrophage marker (CD68, red) in the HMDMs treated with or without ox-LDL. n=6 per group. L, Immunofluorescence analysis of METTL4 (green) and macrophage marker (CD68, red) in the aortic root from Apoe^-/- mice fed with a HFD for 8 and 12 weeks. n=6 per group. M, RT-qPCR analysis of METTL4 levels in peripheral blood mononuclear cells (PBMCs) of healthy individuals and PBMCs of symptomatic and asymptomatic patients with carotid atherosclerosis. n=21 per group. N, Linear regression analysis of mRNA levels of METTL4 with inflammatory factors in PBMCs from patients with carotid atherosclerosis. n=42 per group. O, Nuclear run-on experiments coupled with RT-qPCR analysis of the global nascent METTL4 transcript in HMDMs treated with or without ox-LDL. n=6 per group. Data represent the mean±SEM. **P<0.01, ***P<0.001 by unpaired 2-sided Student t test (B and H through J), unpaired 2-sided Student t test with Welch correction (I and O), Brown-Forsythe and Welch ANOVA test followed by Dunnett T3 multiple comparisons test (M), Mann-Whitney test (G), and Pearson correlation analysis (N).

Currently, METTL4 is the only known methyltransferase that mediated mtDNA 6mA in mammalian cells. However, the subcellular distribution of METTL4 remains controversial.^12,14 In our study, we found that METTL4 was mainly enriched in the mitochondria of macrophages (Figure 1D and 1E; Figure S1C and S1D). Moreover, METTL4 deficiency significantly decreased the ox-LDL–stimulated mtDNA 6mA in HMDMs, other than mitochondrial RNA (Figure 1F; Figure S1E and S1F). Therefore, ox-LDL stimulation increased mtDNA 6mA in HMDMs through METTL4 methyltransferase.

To determine the clinical importance of our findings, we evaluated the expression of METTL4 in human and mouse atherosclerotic lesions. The protein and mRNA levels of METTL4 were elevated in the vascular tissues of patients and mice with atherosclerosis compared with those of nonatherosclerotic tissues (Figure 1G and 1H; Figure S1G and S1H; Table S2). Furthermore, the expression of both Mettl4 and markers of inflammation (MCP-1, IL-1β, and TNF-α) was higher in the athero-susceptible lesser curvature of the aorta relative to the athero-resistant greater curvature aortic region (Figure 1I). Subsequently, we found that the expression of METTL4 was elevated significantly in HMDMs and bone marrow–derived macrophages in response to ox-LDL, rather than in human aortic endothelial cells and human aortic smooth muscle cells (Figure 1J and 1K; Figure S1I and S1J). METTL4 expression in atherosclerotic lesions was further analyzed by immunofluorescence staining. In high-fat diet (HFD)–fed mice, METTL4 expression increased predominantly in CD68⁺ cells in atherosclerotic lesions in a time-dependent manner (Figure 1L). Thus, our data demonstrate that METTL4 was overexpressed in macrophages within plaques. Furthermore, METTL4 mRNA levels were consistently higher in circulating peripheral blood mononuclear cells derived from asymptomatic and symptomatic patients with carotid atherosclerosis than in those from healthy controls (Figure 1M; Table S1). METTL4 levels were higher in symptomatic patients than in asymptomatic patients. The METTL4 mRNA levels in peripheral blood mononuclear cells derived from patients with atherosclerosis were positively correlated with IL-1β, TNF-α, MCP-1, and IL-6 (Figure 1N). Further, we carried out run-on transcription technology to detect gene transcription and expression in real time and found that Bromouridine-trisphosphate–labeled nascent transcripts of METTL4 increased in ox-LDL–stimulated HMDMs (Figure 1O). Taken together, increased METTL4 levels in macrophages may be causally related to plaque inflammation.

Myeloid-Specific Deletion of METTL4 Reduces Atherosclerosis

To further validate the role of macrophage METTL4 in atherosclerosis, we successfully generated myeloid-specific deletion of METTL4 mice on Apoe^-/- background (Mettl4^Mac-KO-Apoe^-/- mice) through crossing Mettl4^flox/flox-Apoe^-/- mice with LyzM-Cre mice (Figure 2A; Figure S2A through S2D). Then, the Mettl4^flox/flox-Apoe^-/- mice and Mettl4^Mac-KO-Apoe^-/- mice (8 weeks old) were fed with a HFD for 12 weeks to establish an atherosclerotic model (Figure 2B). Oil Red O staining of the aortas revealed that HFD-fed Mettl4^Mac-KO-Apoe^-/- mice exhibited a significant reduction in atherosclerotic lesions in the whole aorta compared with HFD-fed Mettl4^flox/flox-Apoe^-/- mice (Figure 2C). Moreover, aortic root analysis of HFD-fed Mettl4^Mac-KO-Apoe^-/- mice showed decreased lipid deposition, lesion area, and necrotic core compared with HFD-fed Mettl4^flox/flox-Apoe^-/- mice (Figure 2D and 2E). We then conducted a more detailed analysis of aortic root components. Compared with HFD-fed Mettl4^flox/flox-Apoe^-/- mice, the macrophage content in atherosclerotic plaque decreased significantly in HFD-fed Mettl4^Mac-KO-Apoe^-/- mice (Figure 2F and 2G). HFD-fed Mettl4^Mac-KO-Apoe^-/- mice showed significantly increased collagen content, fibrous cap thickness, and smooth muscle cell content, suggesting enhanced atherosclerotic plaque stability (Figure 2H and 2I). No statistically significant differences were found in the metabolic parameters between the 2 groups, including body weight, plasma cholesterol, plasma glucose, blood pressure, or heart rate (Table S3). Further, the level of mtDNA 6mA decreased significantly in the thioglycolate-elicited peritoneal macrophages (TEPMs) derived from HFD-fed Mettl4^Mac-KO-Apoe^-/- mice compared with those from HFD-fed Mettl4^flox/flox-Apoe^-/- mice (Figure 2J). These data implied that myeloid-specific deletion of METTL4 effectively alleviates atherosclerotic progression.

Figure 2.

Myeloid-specific deletion of METTL4 reduces atherosclerosis. A, Strategy for the generation of Mettl4^flox/flox mice. B, Schematic diagram of atherosclerotic model establishment in Mettl4^flox/flox-Apoe^-/- mice and Mettl4^Mac-KO-Apoe^-/- mice. C, En face Oil Red O staining of the aortas of HFD-fed Mettl4^flox/flox-Apoe^-/- mice and Mettl4^Mac-KO-Apoe^-/- mice. n=6 per group. D and E, The Oil Red O (D) and HE staining (E) in the aortic roots of HFD-fed Mettl4^flox/flox-Apoe^-/- and Mettl4^Mac-KO-Apoe^-/- mice. n=6 per group. F, RT-qPCR (quantitative reverse transcription polymerase chain reaction) analysis of CD11b within the aortic root plaques derived from HFD-fed Mettl4^flox/flox-Apoe^-/- and Mettl4^Mac-KO-Apoe^-/- mice. n=10 per group. G, Representative immunofluorescence staining images of macrophages (CD68, red) and DAPI (blue) in the aortic roots from HFD-fed Mettl4^flox/flox-Apoe^-/- and Mettl4^Mac-KO-Apoe^-/- mice. n=6 per group. H, The Masson and Sirius red staining in the aortic roots of HFD-fed Mettl4^flox/flox-Apoe^-/- and Mettl4^Mac-KO-Apoe^-/- mice. n=6 per group. I, Representative immunofluorescence staining images of smooth muscle cell (SMA, green) and DAPI (blue) in the aortic roots of HFD-fed Mettl4^flox/flox-Apoe^-/- and Mettl4^Mac-KO-Apoe^-/- mice. n=6 per group. J, Dot blot analysis of mtDNA 6mA levels in thioglycolate-elicited peritoneal macrophages (TEPMs) derived from HFD-fed Mettl4^flox/flox-Apoe^-/- and Mettl4^Mac-KO-Apoe^-/- mice. n=6 per group. Data represent the mean±SEM. **P<0.01, ***P<0.001 by unpaired 2-sided Student t test (C through E and H) and unpaired 2-sided Student t test with Welch correction (C and F through I).

To further validate the aforementioned findings, lethally irradiated Mettl4^Mac-KO-Apoe^-/- mice were reconstituted with bone marrow cells (BMCs) from either Mettl4^flox/flox-Apoe^-/- or Mettl4^Mac-KO-Apoe^-/- mice and fed a HFD for 12 weeks, followed by an analysis of aortic root lesions (Figure S3A). Meanwhile, we also performed Western blot, and quantitative reverse transcription polymerase chain reaction (RT-qPCR) to detect METTL4 expression in mice after transplantation. The results confirmed the engraftment was successful during the bone marrow transplant experiments (Figure S3B and S3C). Compared with Mettl4^Mac-KO-Apoe^-/- mice reconstituted with Mettl4^Mac-KO-Apoe^-/- mice bone marrow, Mettl4^Mac-KO-Apoe^-/- mice reconstituted with Mettl4^flox/flox-Apoe^-/- mice bone marrow showed increased lesions area, lipid deposition, necrotic core, and macrophage content (Figure S3D through S3H), as well as decreased collagen content, fibrous cap thickness, and smooth muscle cell content (Figure S3I and S3J; Table S4). Taken together, the myeloid-specific deletion of METTL4 reduced atherosclerosis.

METTL4 Activated Macrophage Inflammasome by Cytoplasmic mtDNA Released Through Mitochondrial Permeability Transition Pore Opening

To explore the underlying mechanism of METTL4 involved in atherosclerotic development, RNA sequencing was performed to identify the differences between scramble small interfering RNA (siRNA)+ox-LDL–stimulated and siMETTL4+ox-LDL–stimulated HMDMs. Overall, 885 upregulated and 738 downregulated genes were identified in METTL4-deficient HMDMs followed by ox-LDL stimulation (Figure 3A). Gene Ontology enrichment analyses revealed that METTL4 deficiency affected the inflammatory responses (Figure 3B and 3C). To verify the RNA sequencing results, we selected inflammatory genes for RT-qPCR. The expression of these genes in HMDMs and bone marrow–derived macrophages was increased upon ox-LDL stimulation and was downregulated when METTL4 was knocked down (Figure 3D; Figure S4A through S4C). Given that METTL4 regulates mitochondrial homeostasis, which is crucial for inflammasome activation, we speculated that METTL4 may affect ox-LDL–stimulated macrophage inflammation by regulating mitochondrial function. Compared with the ox-LDL–stimulated HMDMs and bone marrow–derived macrophages, the decreased mitochondrial membrane potential and oxygen consumption rate, and the increased extracellular acidification rate as well as the increased levels of intracellular and mitochondrial reactive oxygen species (ROS) were reversed by METTL4 deficiency (Figure 3E and 3F; Figure S4D through S4F). The ultrastructure analysis showed that mitochondria swelling, cristae fracture, and dissolution in ox-LDL–stimulated HMDMs were inhibited by METTL4 deficiency (Figure 3G). The formation of mitochondria with fragmented and granular forms in ox-LDL–stimulated HMDMs was also inhibited by METTL4 deficiency (Figure 3H). Taken together, METTL4 deficiency significantly restored mitochondrial activity in ox-LDL–stimulated HMDMs.

Figure 3.

METTL4-activated macrophage inflammasome through cytoplasmic mtDNA released through mitochondrial permeability transition pore (mPTP) opening. A, Volcano plot of RNA-Seq data (GSE280434) from HMDMs treated with siMETTL4+ox-LDL and Scr+ox-LDL. B, Bubble diagram illustrating the Gene Ontology (GO) enrichment of differentially expressed genes. Each circle corresponds to the number of genes assigned to each category. C, Enrichment of genes involved in GO biological processes. D, RT-qPCR (quantitative reverse transcription polymerase chain reaction) analysis of inflammatory factors in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation is presented as a heatmap. n=6 per group. E, Flow cytometry analysis of mitochondrial membrane potential using a JC-1 probe in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation. n=6 per group. F, Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation were monitored using a Seahorse XFe24 analyzer. n=6 per group. G, The ultrastructure of mitochondria in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation was examined by transmission electron microscopy. LD indicates lipid droplets; M, mitochondria; and N, nucleus. n=6 per group. H, The fluorescence imaging of mitochondrial morphology in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation. n=6 per group. I, Quantitative PCR (qPCR) analysis of cytoplasmic mtDNA (MT-ND1 and MT-ND2), nuclear LINE1 elements (L1ORF1 and L1ORF2), and ribosomal gene (18SRRNA) in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation. n=6 per group. J, The METTL4-deficient HMDMs were transfected with or without METTL4 overexpression plasmid, which were then pretreated with 5 μM CsA for 2 hours, followed by ox-LDL stimulation. Then, qPCR was conducted to analyze cytoplasmic DNA content of these HMDMs. n=6 per group. K, Western blot analysis of inflammasome-associated proteins (caspase-1 and IL-1β) in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation. n=6 per group. L, The METTL4-deficient HMDMs were transfected with or without METTL4 overexpression plasmid, which were then treated with mtDNA or DNase (deoxyribonuclease) I. Then, Western blot was conducted to analyze the expression of inflammasome-associated proteins in the cells. n=6 per group. M, OCR and ECAR in TEPMs from HFD-fed Mettl4^Mac-KO-Apoe^-/- and Mettl4^flox/flox-Apoe^-/- mice were monitored using a Seahorse XFe24 analyzer. n=6 per group. N, Representative immunofluorescence staining images of macrophage (CD68, red) and IL-1β (green) in the aortic roots derived from HFD-fed Mettl4^flox/flox-Apoe^-/- and Mettl4^Mac-KO-Apoe^-/- mice. n=6 per group. O, RT-qPCR analysis of inflammatory factors in plaque macrophages from HFD-fed Mettl4^flox/flox-Apoe^-/- and Mettl4^Mac-KO-Apoe^-/- mice. The results are presented as a heatmap. n=6 per group. Data represent the mean±SEM. *P<0.05, **P<0.01, ***P<0.001 by 2-way ANOVA followed by Tukey multiple comparisons test (D and I), Brown-Forsythe and Welch ANOVA test followed by Dunnett T3 multiple comparisons test (J), unpaired 2-sided Student t test (O), and unpaired 2-sided Student t test with Welch correction (O).

Recently, numerous evidence has reported that mitochondrial damage could contribute to release of mitochondrial DNA (mtDNA) to cytoplasm through mitochondrial permeability transition pore (mPTP) activating inflammasome.¹⁷ In our study, we found that mPTP opening and cytosolic DNA content significantly elevated in ox-LDL–stimulated HMDMs (Figure S4G and S4H). The principal source of cytosolic double-stranded DNA in ox-LDL–treated HMDMs was mtDNA (Figure 3I). The aforementioned phenotypes could be reversed by METTL4 deficiency. In addition, in METTL4-deficient HMDMs, ox-LDL could induce the elevation of cytosolic levels of MT-ND1 and MT-ND2 when METTL4 was restored through overexpression, which was effectively prevented by mPTP inhibitor cyclosporin A (Figure 3J; Figure S4I and S4J). It is interesting that mPTP opening and mtDNA cytoplasmic release in ox-LDL–stimulated HMDMs were significantly inhibited by the mitochondrial ROS scavenger 2,2,6,6-tetramethylpiperidine N-oxyl (TEMPO) (Figure S4K and S4L). Therefore, our data confirmed that METTL4 leads to the mtDNA release through mPTP opening, depending on ROS, in ox-LDL–stimulated HMDMs. Then, METTL4 deficiency effectively inhibited the activation of caspase-1 and IL-1β in ox-LDL–stimulated HMDMs (Figure 3K; Figure S4M). To further confirm that mtDNA is the direct mediator of inflammasome activation in ox-LDL–stimulated HMDMs, mtDNA was isolated from ox-LDL–stimulated HMDMs and then transfected into the naive HMDMs. As shown in Figure 3L and Figure S4N, in METTL4-deficient HMDMs, mtDNA led to an apparent activation of caspase-1 and IL-1β after restoring METTL4 through overexpression, which was suppressed by deoxyribonuclease I. Altogether, these data suggested that in ox-LDL–stimulated macrophages, the activation of inflammasome was triggered by the cytoplasmic mtDNA released through mPTP opening, which was mediated by ROS in a METTL4-dependent manner.

We verified the identified mechanism in vivo. Compared with HFD-fed Mettl4^flox/flox-Apoe^-/- mice, the TEPMs from HFD-fed Mettl4^Mac-KO-Apoe^-/- mice showed significant amelioration of mitochondrial energy metabolism, decreased mitochondrial ROS and mPTP opening, and decreased cytosolic mtDNA (Figure 3M; Figure S4O through S4R). Immunofluorescence staining showed a significant decrease of IL-1β levels in plaque macrophages from HFD-fed Mettl4^Mac-KO-Apoe^-/- mice compared with HFD-fed Mettl4^flox/flox-Apoe^-/- mice (Figure 3N). In addition, the levels of inflammatory factors in isolated plaque macrophages and serum were significantly decreased in HFD-fed Mettl4^Mac-KO-Apoe^-/- mice compared with those in HFD-fed Mettl4^flox/flox-Apoe^-/- mice (Figure 3O; Figure S4S). Subsequently, the rescue experiments were performed. As shown in Figure S4T, the decreased levels of inflammatory factors in plaque macrophages isolated from HFD-fed Mettl4^Mac-KO-Apoe^-/- mice were elevated after receiving bone marrow from Mettl4^flox/flox-Apoe^-/- mice. Collectively, these findings demonstrate that METTL4 promotes macrophage inflammasome activation through cytoplasmic mtDNA released through mPTP opening, thereby promoting atherogenesis.

mtDNA 6mA Is the Critical Factor for METTL4 in the Regulation of Mitochondrial Dysfunction in ox-LDL–Stimulated Macrophages

To further test the essential role of the methyltransferase activity of METTL4 in mediating mtDNA 6mA modification in ox-LDL–stimulated HMDMs, we constructed the enzymatic site mutation plasmid for METTL4 (DPPW was replaced with APPA). Overexpression of wild-type METTL4 (WT), but not the catalytically inactive METTL4 mutant (MUT), increased mtDNA 6mA levels in ox-LDL–stimulated HMDMs (Figure 4A). Furthermore, the inactive METTL4 mutant inhibited the mitochondrial ROS generation and mtDNA release in ox-LDL–stimulated HMDMs (Figure 4B; Figure S5A). To elucidate whether the cytoplasmic release of mtDNA increases macrophage inflammation depending on METTL4 methyltransferase, we examined mitochondrial function in METTL4-deficient HMDMs transfected with METTL4-WT or METTL4-MUT plasmid followed by ox-LDL stimulation.

Figure 4.

The mtDNA 6mA is the critical factor for METTL4 in regulating mitochondrial dysfunction in ox-LDL–stimulated macrophages. A, MethylFlash m6A DNA Methylation ELISA Kit analysis of mtDNA 6mA levels in HMDMs transfected with pcDNA, METTL4-WT, and METTL4-MUT followed by ox-LDL stimulation. n=6 per group. B, qPCR (quantitative polymerase chain reaction) analysis of cytoplasmic mtDNA content in HMDMs transfected with pcDNA, METTL4-WT, and METTL4-MUT followed by ox-LDL stimulation. n=6 per group. C, Flow cytometry analysis of the mitochondrial membrane potential using the JC-1 probe in METTL4-deficient HMDMs transfected with METTL4-WT and METTL4-MUT followed by ox-LDL stimulation. n=6 per group. D, ECAR and OCR of the METTL4-deficient HMDMs transfected with METTL4-WT and METTL4-MUT followed by ox-LDL stimulation were monitored using a Seahorse XFe24 analyzer. n=6 per group. We constructed mtDNA-depleted METTL4-deficient HMDMs (ρ0) using Etbr. Then, the ρ0 cells and METTL4-deficient HMDMs were transfected with METTL4-WT or METTL4-MUT, followed by ox-LDL stimulation. E, Transmission electron microscopy analysis of the ultrastructure of HMDMs. LD indicates lipid droplets; M, mitochondria; and N, nucleus. n=6 per group. F, Western blot analysis of inflammasome-associated proteins in HMDMs. n=6 per group. G, RT-qPCR (quantitative reverse transcription polymerase chain reaction) analysis of inflammatory factors in HMDMs. The results are presented as a heatmap. n=6 per group. H, Flow chart of experimental procedure for mitochondrial transplantation in HMDMs. I, Western blot analysis of the purity of extracted mitochondria. n=6 per group. J, Flow cytometry analysis of the internalization of MitoTracker Green–labeled mitochondria in HMDMs. n=3 per group. The mitochondria were extracted from HMDMs transfected with METTL4-WT followed by ox-LDL stimulation (Mit^WT+ox-LDL) or HMDMs transfected with METTL4-MUT followed by ox-LDL stimulation (Mit^MUT+ox-LDL). Then, the METTL4-deficient HMDMs were transfected with pcDNA followed by ox-LDL stimulation, which were then transplanted with Mit^WT+ox-LDL. The METTL4-deficient HMDMs were transfected with METTL4 overexpression plasmid followed by ox-LDL stimulation, which were then transplanted with Mit^MUT+ox-LDL. K, The effects of mitochondrial transplantation on ECAR and OCR in HMDMs were monitored using a Seahorse XFe24 analyzer. n=6 per group. L, TMRM fluorescence staining analysis showing the effects of mitochondrial transplantation on mitochondrial membrane potential. n=6 per group. M, MethylFlash m6A DNA Methylation ELISA Kit analysis of the effect of mitochondrial transplantation on mtDNA 6mA levels in HMDMs. n=6 per group. N, qPCR analysis of the effects of mitochondrial transplantation on the cytoplasmic mtDNA content in HMDMs. n=6 per group. O, RT-qPCR analysis of the effects of mitochondrial transplantation on expressions of inflammatory factors in HMDMs. n=6 per group. Data represent the mean±SEM. *P<0.05, **P<0.01, ***P<0.001 by 1-way ANOVA followed by Tukey multiple comparisons test (A, G, M, and O), and Brown-Forsythe and Welch ANOVA test followed by Dunnett T3 multiple comparisons test (B, L, N, and O).

Meanwhile, we also challenged METTL4-deficient HMDMs with mtDNA-depleting drug ethidium bromide to generate ρ0 cells, which were then transfected with METTL4-WT or METTL4-MUT plasmid followed by ox-LDL stimulation. In METTL4-deficient HMDMs, METTL4-MUT transfection followed by ox-LDL stimulation (Figure S5B) restored mitochondrial membrane potential (Figure 4C) and improved mitochondrial energy metabolism dysfunction (Figure 4D; Figure S5C) compared with those transfected with METTL4-WT followed by ox-LDL stimulation. Meanwhile, METTL4-MUT transfection followed by ox-LDL stimulation decreased mitochondrial ultrastructure damage and the activation of inflammasome and the expression of inflammatory genes other than METTL4-WT transfection, which was similar to ethidium bromide (Figure 4E through 4G; Figure S5D).

Furthermore, we performed mitochondrial transplantation to explore the effect of METTL4 on inflammasome activation by regulating mitochondrial function in ox-LDL–stimulated HMDMs (Figure 4H through 4J; Figure S5E). We effectively extracted mitochondria from HMDMs transfected with the METTL4-WT or METTL4-MUT followed by ox-LDL stimulation (Mit^WT+ox-LDL, Mit^MUT+ox-LDL) and subsequently transplanted them into recipient HMDMs. Compared with METTL4-deficient HMDMs treated with the empty vector pcDNA3.1 and ox-LDL, Mit^WT+ox-LDL transplantation resulted in significant mitochondrial energy metabolism dysfunction, a reduction in mitochondrial membrane potential, an increase in mtDNA 6mA levels, cytosolic mtDNA accumulation, elevated mitochondrial ROS, and increased production of inflammatory cytokines, similar to the effects of exogenous supplementation with METTL4 followed by ox-LDL stimulation (Figure 4K through 4O; Figure S5F and S5G). However, compared with METTL4-deficient HMDMs treated with METTL4 overexpression and ox-LDL, these changes were significantly inhibited when transplanted with Mit^MUT+ox-LDL. Therefore, we confirmed that the overexpression of METTL4 in ox-LDL–stimulated HMDMs increased mtDNA 6mA modification and impaired mitochondrial function, leading to cytoplasmic mtDNA accumulation and inflammasome activation, depending on its 6mA methyltransferase catalytic activity.

METTL4-Mediated Mitochondrial ATP6 (MT-ATP6) 6mA Modification Mediated Excess Proton Accumulation in the Mitochondrial Intermembrane Space

It is well known that the mtDNA encodes 13 components of mitochondrial respiratory chain complexes, and altered mitochondrial transcription drives various human pathologies.¹⁸ As shown in Figure 5A and 5B, METTL4 deficiency effectively increased the expression and activity of mitochondrial respiratory chain complex V (ATPase) in ox-LDL–stimulated HMDMs, compared with HMDMs transfected with scramble siRNA followed by ox-LDL stimulation. Furthermore, METTL4 deficiency markedly increased the expression of mitochondrial ATP synthase F0 subunit 6 (MT-ATP6) in ox-LDL–stimulated HMDMs, as determined upon screening for mtDNA-encoded gene expression (Figure 5C and 5D; Figure S6A). In addition, in METTL4-deficient HMDMs, transfecting METTL4-WT followed by ox-LDL decreased MT-ATP6 expression compared with pcDNA+ox-LDL group, other than METTL4-MUT+ox-LDL group (Figure 5E and 5F; Figure S6B). In addition, MT-ATP6 has an mtDNA 6mA motif, consistent with the literature (Figure 5G).¹² Immunoprecipitation analysis showed that transfecting METTL4-WT followed by ox-LDL stimulation led to the increased MT-ATP6 6mA levels and decreased binding of TFAM (mitochondrial transcription factor A) to the MT-ATP6 in METTL4-deficient HMDMs compared with the pcDNA+ox-LDL group, which was reversed by METTL4-MUT (Figure 5G and 5H). Therefore, we propose that METTL4 promotes MT-ATP6 6mA modification and decreases TFAM-mediated MT-ATP6 transcription, leading to mitochondrial dysfunction.

Figure 5.

METTL4-mediated MT-ATP6 6mA caused excess protons accumulated in the mitochondrial intermembrane space. A, Western blot analysis of mitochondrial respiratory chain complex I-V in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation. n=6 per group. B, Mitochondrial respiration chain complex activity assay kit measures the activity of mitochondrial respiratory chain complex I-V in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation. n=6 per group. C, RT-qPCR (quantitative reverse transcription polymerase chain reaction) analysis of mtDNA encoded genes in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation, as shown in the heat map. n=6 per group. D, Western blot analysis of MT-ATP6 expression in HMDMs transfected with or without siMETTL4 followed by ox-LDL stimulation. n=6 per group. E, RT-qPCR analysis of MT-ATP6 in METTL4-deficient HMDMs transfected with pcDNA, METTL4-WT, or METTL4-MUT followed by ox-LDL stimulation. n=6 per group. F, Western blot analysis of MT-ATP6 in METTL4-deficient HMDMs transfected with pcDNA, METTL4-WT, or METTL4-MUT followed by ox-LDL stimulation. n=6 per group. G, The mtDNA 6mA motif of MT-ATP6 (left). Immunoprecipitation (IP) analysis of the binding of 6mA and MT-ATP6 in METTL4-deficient HMDMs transfected with pcDNA, METTL4-WT, or METTL4-MUT followed by ox-LDL stimulation (right). n=6 per group. H, IP analysis of the binding of TFAM and MT-ATP6 in METTL4-deficient HMDMs transfected with pcDNA, METTL4-WT, or METTL4-MUT followed by ox-LDL stimulation. n=6 per group. I through L, The analysis of the ECAR and OCR (I), cytoplasmic mtDNA content (J), the expression of inflammasome-associated proteins (K), and the inflammation (L) in METTL4-deficient HMDMs transfected with METTL4 or ATP6 overexpression plasmid followed by ox-LDL stimulation. n=6 per group. M, Fluorescence analysis of the submitochondrial localization of protons in METTL4-deficient HMDMs transfected with METTL4 or ATP6 overexpression plasmid followed by ox-LDL stimulation. Mitochondrial (TOMM20, green) and protons (red) are shown. n=6 per group. Data represent the mean±SEM. **P<0.01, ***P<0.001 by unpaired 2-sided Student t test (A through C), unpaired 2-sided Student t test with Welch correction (B), 1-way ANOVA followed by Tukey multiple comparisons test (E and H), Brown-Forsythe and Welch ANOVA test followed by Dunnett T3 multiple comparisons test (G, J, and L), and Kruskal-Wallis test followed by Dunn multiple comparisons test (M).

To confirm the pivotal role of MT-ATP6, we examined the effects of ATP6 supplementation on mitochondrial function in ox-LDL–stimulated HMDMs. The results demonstrated that in METTL4-deficient HMDMs, restoring METTL4 through overexpression followed by ox-LDL stimulation significantly contributed to a decrease in ATPase activity (Figure S6C) and an increase in mitochondrial ROS generation (Figure S6D), mitochondrial metabolism dysfunction (extracellular acidification rate and oxygen consumption rate; Figure 5I; Figure S6E), cytosolic mtDNA accumulation (Figure 5J), inflammasome activation (Figure 5K; Figure S6F), and expression of inflammatory factors (Figure 5L), which were reversed by additional ATP6 supplementation.

ATP6 is a core subunit of mitochondrial respiratory chain complex V generating ATP through exploiting proton gradient.¹⁹ Yang et al have reported that reduced ATP synthase activity could slow down the reflux of translocated protons triggering a burst of ROS and premature programmed cell death of the tapetal cells, contributing male sterility in plants.²⁰ To explore the detailed mechanism underlying the effect of METTL4 on mitochondrial dysfunction, we first examined the submitochondrial localization of protons. In METTL4-deficient HMDMs conducted by shMETTL4 annotated in the left upper of Figure 5M, ox-LDL had no effect on the proton subcellular distribution and mitochondrial morphology (Figure 5M). However, restoring METTL4 through overexpression followed by ox-LDL stimulation significantly increased proton accumulation in the mitochondrial intermembrane space, compared with the ox-LDL–stimulated METTL4-deficient HMDMs transfected with pcDNA (Figure 5M). The phenomenon was reversed by additional ATP6 supplementation (Figure 5M). Therefore, METTL4 mediates the accumulation of excess protons in the mitochondrial intermembrane space, triggering a burst of ROS and inflammasome activation through inhibiting MT-ATP6 expression by mtDNA 6mA modification in ox-LDL–stimulated HMDMs.

Myeloid-Specific Mutation in METTL4 Methyltransferase Active Site Reduced Atherosclerosis

To explore whether the effect of macrophage METTL4 on atherogenesis depended on its methyltransferase activity in vivo, we successfully constructed a knock-in mouse in which the DPPW catalytic site was mutated to APPA using CRISPR/Cas9 technology (Figure S7A and S7B). Mettl4^WT-Apoe^-/- mice and Mettl4^MUT-Apoe^-/- mice (8 weeks old) were fed with a HFD for 12 weeks to establish an atherosclerotic model. HFD-fed Mettl4^MUT-Apoe^-/- mice exhibited significant reduction of atherosclerotic lesions compared with HFD-fed Mettl4^WT-Apoe^-/- mice (Figure 6A). The level of mtDNA 6mA modification in TEPMs was significantly reduced in HFD-fed Mettl4^MUT-Apoe^-/- mice compared with that in HFD-fed Mettl4^WT-Apoe^-/- mice (Figure 6B). In addition, the staining of aortic root sections revealed a substantial decrease in the lesion area and necrotic core in HFD-fed Mettl4^MUT-Apoe^-/- mice compared with those in HFD-fed Mettl4^WT-Apoe^-/- mice, along with a significant reduction in lipid deposition (Figure 6C). Compared with HFD-fed Mettl4^WT-Apoe^-/- mice, the macrophage content decreased significantly, along with a remarkable increase in the collagen content, fibrous cap thickness, and smooth muscle cell levels within the aortic root plaques observed in HFD-fed Mettl4^MUT-Apoe^-/- mice (Figure 6D and 6E; Figure S7C; Table S5). The levels of inflammatory factors in plaque macrophages isolated from HFD-fed Mettl4^MUT-Apoe^-/- mice were lower than those isolated from HFD-fed Mettl4^WT-Apoe^-/- mice (Figure 6F).

Figure 6.

Myeloid-specific mutation in METTL4 methyltransferase active site reduced atherosclerosis. A, En face Oil Red O staining of the aortas of HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=6 per group. B, MethylFlash m6A DNA Methylation ELISA Kit analysis of mtDNA 6mA levels in TEPMs from HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=10 per group. C, Oil Red O and HE staining of aortic roots of HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=6 per group. D, Representative immunofluorescence staining images of macrophage (CD68, red) and DAPI staining (blue) in the aortic roots of HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=6 per group. E, Masson, Sirius red, and immunofluorescence staining images of smooth muscle cells (SMA, green) and DAPI (blue) in the aortic roots of HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=6 per group. F, RT-qPCR (quantitative polymerase chain reaction) analysis of inflammatory factors in plaque macrophages from HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=6 per group. G, IP analysis of 6mA and MT-ATP6 binding in TEPMs from HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=10 per group. H and I, RT-qPCR (quantitative reverse transcription polymerase chain reaction) and Western blot analysis of the expression of MT-ATP6 in TEPMs from HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=6–10 per group. J, Mitochondrial respiration chain complex activity assay kit analysis of the activity of mitochondrial respiratory chain complex I-V in TEPMs from HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=6 per group. K, ECAR and OCR in TEPMs from HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice were monitored using a Seahorse XFe24 analyzer. n=6 per group. L, qPCR analysis of cytoplasmic mtDNA content in TEPMs from HFD-fed Mettl4^WT-Apoe^-/- mice and Mettl4^MUT-Apoe^-/- mice. n=10 per group. M, Western blot analysis of inflammasome-associated proteins in TEPMs from HFD-fed Mettl4^WT-Apoe^-/- and Mettl4^MUT-Apoe^-/- mice. n=6 per group. N, En face Oil Red O staining of the aorta of Mettl4^Mac-KO-Apoe^-/- mice received bone marrow from Mettl4^Mac-KO-Apoe^-/- (KO), Mettl4^WT-Apoe^-/- (WT), or Mettl4^MUT-Apoe^-/- (MUT) mice followed by a HFD. n=6 per group. O, RT-qPCR analysis of CD11b within the aortic root plaques from HFD-fed Mettl4^Mac-KO-Apoe^-/- mice receiving bone marrow cell transplantation treated as in N. n=10 per group. P, RT-qPCR analysis of inflammatory factors in plaque macrophages from Mettl4^Mac-KO-Apoe^-/- mice receiving bone marrow cell transplantation treated as in N. n=10 per group. Data represent the mean±SEM. **P<0.01, ***P<0.001 by unpaired 2-sided Student t test (A through D, F, H, and J), unpaired 2-sided Student t test with Welch correction (A, F, G, I, and L), Brown-Forsythe and Welch ANOVA test followed by Dunnett T3 multiple comparisons test (N), Kruskal-Wallis test followed by Dunn multiple comparisons test (O and P), and 1-way ANOVA followed by Tukey multiple comparisons test (P).

To elucidate the mechanism observed in vitro, TEPMs derived from HFD-fed Mettl4^MUT-Apoe^-/- mice and HFD-fed Mettl4^WT-Apoe^-/- mice were collected and subjected to further analysis. The results revealed a decrease in MT-ATP6 6mA modification and an increase in MT-ATP6 expression in HFD-fed Mettl4^MUT-Apoe^-/- mice compared with that in HFD-fed Mettl4^WT-Apoe^-/- mice (Figure 6G through 6I). In addition, mitochondrial respiratory chain complex V activity and the mitochondrial energy metabolism dysfunction were significantly improved in HFD-fed Mettl4^MUT-Apoe^-/- mice compared with that in HFD-fed Mettl4^WT-Apoe^-/- mice (Figure 6J and 6K; Figure S7D). In addition, mitochondrial ROS, the cytosolic accumulation of mtDNA, and inflammasome activation were inhibited in HFD-fed Mettl4^MUT-Apoe^-/- mice compared with that in HFD-fed Mettl4^WT-Apoe^-/- mice (Figure 6L and 6M; Figure S7E). These data suggested that a myeloid-specific mutation in the METTL4 methyltransferase active site reduced atherosclerosis by ameliorating mitochondrial dysfunction and inflammasome activation.

To further validate the above findings, we lethally irradiated Mettl4^Mac-KO-Apoe^-/- mice and reconstituted them with BMCs from Mettl4^Mac-KO-Apoe^-/- mice, Mettl4^WT-Apoe^-/- mice, or Mettl4^MUT-Apoe^-/- mice (Figure S8A and S8B; Table S6). The mice receiving BMCs transplantation were then fed with a HFD for 12 weeks to establish an atherosclerotic model. In comparison with HFD-fed Mettl4^Mac-KO-Apoe^-/- mice receiving BMCs transplantation from Mettl4^Mac-KO-Apoe^-/- mice, HFD-fed Mettl4^Mac-KO-Apoe^-/- mice receiving BMCs transplantation from Mettl4^WT-Apoe^-/- mice exhibited significant progression of plaques and a notable increase in macrophage content within the plaques (Figure 6N and 6O). Furthermore, the HFD-fed Mettl4^Mac-KO-Apoe^-/- mice receiving BMCs transplantation from Mettl4^WT-Apoe^-/- mice exhibited increased MT-ATP6 6mA modification, decreased MT-ATP6 expression and mitochondrial respiratory chain complex V activity, upregulation of mitochondrial ROS, cytoplastic accumulation of mtDNA, impaired mitochondrial energy metabolism, and increased inflammation levels compared with HFD-fed Mettl4^Mac-KO-Apoe^-/- mice receiving BMCs transplantation from Mettl4^Mac-KO-Apoe^-/- mice (Figure 6P; Figure S8C through S8I). However, all these alterations were reversed in Mettl4^Mac-KO-Apoe^-/- mice receiving BMCs transplantation from Mettl4^MUT-Apoe^-/- mice. The findings indicate that METTL4 exacerbates atherosclerosis by impairing mitochondrial function and activating inflammasome by increasing MT-ATP6 6mA modification in macrophages.

Pemetrexed Was Identified as the First METTL4 Antagonist Effective in Mitigating Atherosclerosis Progression

Currently, no selective pharmacological antagonist is available against METTL4. To investigate the therapeutic potential of targeting the enzymatic activity of METTL4 as an antiatherogenic strategy, high-throughput screening of diverse drug-like compounds was performed. METTL4 is a homologous member of the MT-A70 family eukaryotic methyltransferase with a conserved methyltransferase domain including a DPPW motif, as the same as METTL3.¹² Nevertheless, the crystal structure of human-derived METTL4 is unknown. Therefore, we performed homology modeling and molecular docking based on the crystal structure of the MT-A70 domain of Arabidopsis METTL4 and the human-derived METTL3 sequence to identify the compound (Figure 7A).^13,21 Using a structure-based drug discovery approach, we identified that xanthinol nicotinate, pemetrexed, and linagliptin, among the 18 compounds screened from TargetMol, could effectively inhibit the methyltransferase activity of METTL4 (Figure 7B and 7C). By investigating the cytotoxicity effect with different concentrations and action times on HMDMs, the optimum dosages (pemetrexed 0.1 μM, xanthinol nicotinate 125 μM, and linagliptin 0.1 μM) and action time (24 hours) were selected for subsequent experiments (Figure S9A and S9B). We found that the decrease in the mRNA level of MT-ATP6 and increase in the levels of inflammatory factors induced by ox-LDL were inhibited by pemetrexed rather than by xanthinol nicotinate and linagliptin (Figure 7D and 7E). Furthermore, we identified 75 nM as the optimal concentration of pemetrexed for the ox-LDL–stimulated HMDMs (Figure 7F). Pemetrexed effectively reduced MT-ATP6 6mA modification, increased MT-ATP6 expression, and decreased mitochondrial ROS, accumulation of cytosolic mtDNA, and expression of inflammatory factors in ox-LDL–stimulated HMDMs (Figure 7G; Figure S9C through S9F).

Figure 7.

Pemetrexed (PEM) was identified as the first METTL4 antagonist effective in mitigating atherosclerosis progression. A, Procedure for 3D mathing and ensemble docking-based virtual screening to identify METTL4 inhibitors. B, Enzymatic inhibitory activity of the indicated compounds against purified METTL4 using a bioluminescence assay. n=6 per group. C, The molecular docking of Pemetrexed, Xanthinol Nicotinate, and Linagliptin with METTL4. D, RT-qPCR (quantitative reverse transcription polymerase chain reaction) analysis of MT-ATP6 expression in ox-LDL–stimulated HMDMs treated with PEM, Xanthinol Nicotinate (XN), and Linagliptin. n=6 per group. E, RT-qPCR analysis of inflammatory factors in ox-LDL–stimulated HMDMs treated with PEM, XN, and Linagliptin. The results are presented as a heatmap. n=6 per group. F, RT-qPCR analysis of MT-ATP6 in ox-LDL–stimulated HMDMs treated with different concentrations of PEM. n=6 per group. G, IP analysis of 6mA and MT-ATP6 binding in ox-LDL–stimulated HMDMs treated with or without PEM. n=6 per group. H, En face Oil Red O staining of the aortas of HFD-fed Apoe^-/- mice treated with different dosages of PEM. n=6 per group. I, Oil Red O staining of the aortic roots of HFD-fed Apoe^-/- mice treated with different dosages of PEM. n=6 per group. J, The HE, Masson, and Sirius red staining of the aortic roots of HFD-fed Apoe^-/- mice with or without PEM. n=6 per group. K and L, Dot blot analysis of mtDNA 6mA levels (K) and Western blot analysis of MT-ATP6 expression (L) in TEPMs from HFD-fed Apoe^-/- mice with or without PEM. n=6 per group. M, qPCR analysis of cytoplasmic mtDNA content in TEPMs from HFD-fed Apoe^-/- mice with or without PEM.. n=10 per group. N, The ECAR and OCR in TEPMs from HFD-fed Apoe^-/- mice with or without PEM were monitored using a Seahorse XFe24 analyzer. n=6 per group. O, RT-qPCR analysis of inflammatory factors in plaque macrophages from HFD-fed Apoe^-/- mice with or without PEM. n=6 per group. Data represent the mean±SEM. *P<0.05, **P<0.01, ***P<0.001 by 1-way ANOVA followed by Tukey multiple comparisons test (B, D through F, and H), unpaired 2-sided Student t test (G, J, L, M, and O), and unpaired 2-sided Student t test with Welch correction (O).

To verify the therapeutic effect of pemetrexed in vivo, different dosages of pemetrexed (1, 5, 25, and 125 mg/kg per 3 days) were administered to HFD-fed Apoe^-/- mice for 8 weeks after 4 weeks of HFD. Compared with HFD-fed Apoe^-/- mice with dimethyl sulfoxide pemetrexed significantly reduced atherosclerotic lesions in the aorta and lipids accumulation in the aortic root when the dosage exceeded 1 mg/kg per 3 days (Figure 7H and 7I; Table S7). However, the therapeutic effect plateaued with no discernable changes upon increasing dosage beyond 5 mg/kg per 3 days. Therefore, pemetrexed (5 mg/kg per 3 days) was selected for subsequent experiments. Decreased lesion area and necrotic core, as well as increased collagen content and fibrous cap thickness, were observed in the aortic root plaques in HFD-fed Apoe^-/- mice treated with pemetrexed compared with those in the control group (Figure 7J). To further elucidate the therapeutic mechanism of pemetrexed in atherosclerotic, TEPMs derived from HFD-fed Apoe^-/- mice treated with or without pemetrexed were collected and subjected to further analyses. Compared with HFD-fed Apoe^-/- mice with DMSO, pemetrexed effectively contributed to a decrease in mtDNA 6mA modification, an increase in MT-ATP6 expression, and a decrease in cytoplasmic content of mtDNA (Figure 7K through 7M). Pemetrexed improved mitochondrial dysfunction and decreased the levels of inflammatory factors in HFD-fed Apoe^-/- mice with DMSO (Figure 7N and 7O; Figure S9G and S9H). Collectively, we identified pemetrexed as the first METTL4 antagonist that can be used as an effective therapeutic agent for atherosclerosis through alleviating mitochondrial dysfunction and inflammation by reducing mtDNA 6mA.

PROTAC-Pemetrexed (PROTAC-PEM) Could Effectively Alleviate Atherosclerosis

PROTAC is now emerging as a novel therapeutic technology that harnesses the ubiquitin-proteasome system to selectively induce targeted protein degradation with the potential to modulate traditional undruggable targets characterized by high selectivity, low side effects, and low drug resistance.²² Based on structural chemistry, METTL4 was predicted to be a potential drug target suitable for PROTAC-mediated protein degradation.²³ We first designed METTL4 degrader (dMETTL4, a selective PROTAC-PEM reagent capable of degrading METTL4), which is a chimera composed of PEM, a linker, and an E3 ligase. To improve the bioavailability and specificity of PROTAC-PEM for M1 macrophages in atherosclerotic plaques, dMETTL4 was loaded with Poly (lactic acid-glycolic acid) (PLGA) nanoparticles and further coated with M2 macrophage membranes derived from IL4-stimulated RAW264.7 (Figure 8A). We first identified that 0.05 μM PROTAC-PEM incubated for 24 hours could efficiently degrade METTL4 in HMDMs without any effect on cell viability, which was abolished by the proteasome inhibitor MG132 (Figure 8B; Figure S10A through S10D). Meanwhile, PROTAC-PEM specifically degrade the METTL4 in macrophages, other than smooth muscle cells, endothelial cells, or fibroblasts (Figure S10C through S10F). In addition, PROTAC-PEM effectively decreased the MT-ATP6 6mA modification and increased MT-ATP6 expression, improved mitochondrial metabolism dysfunction, and reduced cytosolic accumulation of mtDNA and inflammatory factors in ox-LDL–stimulated HMDMs (Figure 8C through 8G). These data suggest that PROTAC-PEM exerts its therapeutic effects through a proteasomal pathway in vitro.

Figure 8.

PROTAC-PEM could effectively alleviate atherosclerosis. A, The preparation procedure for PROTAC-PEM targeting macrophages. B, Western blot analysis of METTL4 expression in HMDMs pretreated with MG132 (10 μM, 4 hours), followed by ox-LDL stimulation combining with PROTAC-PEM. n=6 per group. C, IP analysis of the binding of 6mA and MT-ATP6 in ox-LDL–stimulated HMDMs treated with or without PROTAC-PEM. n=10 per group. D, RT-qPCR (quantitative reverse transcription polymerase chain reaction) analysis of MT-ATP6 expression in ox-LDL–stimulated HMDMs treated with or without PROTAC-PEM. n=6 per group. E, ECAR and OCR in ox-LDL–stimulated HMDMs treated with or without PROTAC-PEM were monitored using a Seahorse XFe24 analyzer. n=6 per group. F and G, qPCR (quantitative polymerase chain reaction) analysis of cytoplasmic mtDNA content (F), RT-qPCR analysis of inflammatory factors (G) in ox-LDL–stimulated HMDMs treated with or without PROTAC-PEM. n=6 per group. H, En face Oil Red O staining of the aorta of NC or HFD-fed Apoe^-/- mice treated with or without PROTAC-PEM. n=6 per group. I, RT-qPCR analysis of CD11b in aortic root plaques derived from NC or HFD-fed Apoe^-/- mice treated with or without PROTAC-PEM. n=10 per group. J, RT-qPCR analysis of inflammatory factors in macrophages within atherosclerotic plaques of NC or HFD-fed Apoe^-/- mice treated with or without PROTAC-PEM. n=6 per group. K, Dot blot analysis of mtDNA 6mA levels in TEPMs from NC or HFD-fed Apoe^-/- mice treated with or without PROTAC-PEM. n=6 per group. L and M, RT-qPCR analysis of MT-ATP6 (L), and qPCR analysis of the cytoplasmic mtDNA content in TEPMs (M) from NC or HFD-fed Apoe^-/- mice treated with or without PROTAC-PEM. n=10 per group. N, Representative immunofluorescence staining images of macrophage (CD68, red) and IL-1β (green) in the aortic root derived from HFD-fed Apoe^-/- mice treated with or without PROTAC-PEM. n=6 per group. O, A mechanistic diagram of METTL4-mediated mtDNA 6mA modification in macrophage induced atherosclerosis development. Data represent the mean±SEM. *P<0.05, **P<0.01, ***P<0.001 by 2-way ANOVA followed by Tukey multiple comparisons test (B through D, F through J, and L and M).

To verify the therapeutic effects of pemetrexed-PROTAC-PEM in vivo, PROTAC-PEM with different dosages was administered to HFD-fed Apoe^-/- mice. As shown in Figure S10G, compared with HFD-fed Apoe^-/- mice, PROTAC-PEM significantly reduced atherosclerotic lesions in the aorta when the dosage exceeded 1 mg/kg per 3 days. However, the therapeutic effect plateaued with no further changes upon increasing dosage beyond 3 mg/kg per 3 days. Therefore, PROTAC-PEM (3 mg/kg per 3 days) was selected for subsequent experiments. We also confirmed that PROTAC was more selective for macrophages (Figure S10H and S10I). No statistically significant differences were noted in the metabolic parameters between the 2 groups, including body weight, plasma cholesterol, plasma glucose, blood pressure, or heart rate (Table S8). PROTAC-PEM significantly reduced atherosclerotic lesions in the aorta, arterial arch, and thoracic aorta compared with those in HFD-fed Apoe^-/- mice treated with vehicle (Figure 8H). PROTAC-PEM significantly decreased the macrophage content and inflammatory factors expression in macrophages within the aortic root plaques (Figure 8I and 8J). We isolated TEPMs from HFD-fed Apoe^-/- mice with or without PROTAC-PEM treatment to verify the mechanism observed in vitro. We found that TEPMs from HFD-fed Apoe^-/- mice treated with PROTAC-PEM showed decreased levels of mtDNA 6mA, increased levels of MT-ATP6 expression, reduced mitochondrial ROS, and cytosolic accumulation of mtDNA compared with those from HFD-fed Apoe^-/- mice treated with vehicle (Figure 8K through 8M; Figure S10J). In addition, immunofluorescence staining showed a significant decrease in IL-1β levels in plaques from HFD-fed Apoe^-/- mice treated with PROTAC-PEM compared with that in the vehicle group (Figure 8N). The PCR primers, siRNA sequences, and antibodies used in this study is shown in Tables S9–S11. Overall, our data confirmed for the first time that PROTAC-PEM technology can be successfully applied to atherosclerotic treatment by targeting the degradation of METTL4 in macrophages.

DISCUSSION

We first discovered that the mtDNA 6mA, mediated by METTL4 in macrophages, facilitates mitochondrial damage, contributing to inflammation and atherosclerosis. Overexpression of METTL4 decreased mitochondrial MT-ATP6 transcription by increasing the MT-ATP6 6mA and inducing mitochondrial complex V deficiency. Then, excess protons accumulated in the mitochondrial intermembrane space and triggered a burst of ROS, leading to the opening of the mPTP and the release of mtDNA into the cytoplasm, subsequently activating the inflammasome. Moreover, pemetrexed was identified as the first METTL4 methyltransferase antagonist. Bioinspired PROTAC was constructed by coating the PROTAC degrader (dMETTL4)–loaded PLGA nanoparticles with M2 macrophage membranes targeting plaque macrophages for atherosclerosis treatment. We found that pemetrexed and corresponding PROTAC-PEM effectively inhibited macrophage inflammation and atherosclerosis (Figure 8O). Therefore, we established the essential role of METTL4 in resolving atherosclerotic progression by regulating mitoepigenetics.

Mitochondrial dysfunction plays a critical role in the progress of atherosclerosis.²⁴ In particular, improving mitochondrial function by balancing mitochondrial fusion and fission and increasing mitophagy could effectively delay atherosclerosis.⁵ In a recent cohort and Mendelian randomization study, Luo et al found that low leukocyte mtDNA abundance was associated with an increased risk of incident CVD.²⁵ Thus, mtDNA plays a pivotal role in atherosclerosis progression by regulating mitochondrial function.

The epigenetic regulation of mtDNA (mitoepigenetics), a rapidly emerging field, has recently garnered significant interest.²⁶ Hao et al identified for the first time that mtDNA could also be methylated at position N6 in mammals. 6mA modification exhibits enrichment in mtDNA, which contained an 8000-fold higher level of 6mA compared with nuclear DNA content.¹² However, the function of mtDNA 6mA modification in pathological processes is still poorly understood. In our study, we reported aberrant elevated mtDNA 6mA modification in the macrophage mitochondria of atherosclerosis plaques, which eventually caused mitochondrial dysfunction and inflammation. Therefore, the mtDNA 6mA modification could be identified as a new risk factor in atherosclerosis.

METTL4 is a mammalian methyltransferase conserved from yeast to human. However, the function of METTL4 has remained highly controversial in recent years. Chen et al reported that METTL4 localized in the nucleus induced N6-methylation of N6,2'-O-dimethyladenosine on U2 snRNA and regulated RNA splicing.¹⁴ However, Hao et al showed that METTL4 primarily resided in the mitochondrial matrix (Integrated Mitochondrial Protein Index score, 0.748), therefore mediating mtDNA 6mA modification in mammals responsible for attenuating mtDNA transcription and function.¹² Hence, the role of METTL4 seems to be context-dependent. In our study, we first identified METTL4 as a mitochondrial protein in macrophages that decreased MT-ATP6 transcription and mitochondrial complex V expression through MT-ATP6 6mA. Excess protons accumulated in the mitochondrial intermembrane space and triggered a burst of ROS, the opening of mPTP, and the release of mtDNA into the cytoplasm, leading to the activation of inflammasome. Furthermore, myeloid-specific deletion of METTL4 or myeloid-specific mutation of METTL4 at the methyltransferase active site could effectively reduce atherosclerosis through ameliorating mitochondrial dysfunction and inflammasome activation. Therefore, METTL4 may be a promising therapeutic target for atherosclerosis from the perspective of mitoepigenetics.

Currently, no inhibitor targeting METTL4 has been reported. By conducting a screening of natural compound libraries, we identified for the first time that pemetrexed functions as a METTL4 methyltransferase antagonist worldwide. Pemetrexed effectively binds the catalytic domain of METTL4 by forming several hydrogen bonds and salt bridges. Our results indicate that pemetrexed can effectively reduce the methyltransferase activity of METTL4, thereby inhibiting MT-ATP6 6mA modification, alleviating the transcription and translation blockade of MT-ATP6, and subsequently mitigating macrophage mitochondrial dysfunction, inflammation, and atherosclerosis. Although pemetrexed (100 mg/kg), an antitumor agent, could potentially induce weight loss and reduce cholesterol levels in mice,^27,28 our study did not observe any alteration in body weight or cholesterol levels. Furthermore, pemetrexed did not cause any toxic effect in vivo. This contradiction may be attributed to the low dosage of pemetrexed (5 mg/kg per 3 days) used in our investigation, which may be insufficient to elicit weight loss or reductions in cholesterol levels. Taken together, our data have identified, for the first time, that the methyltransferase inhibitor pemetrexed, which targets METTL4, may be an effective drug for the treatment of atherosclerosis.

There is an urgent need to explore targeted therapeutic approaches for atherosclerosis. The application of PROTAC molecules is a new therapeutic strategy with great potential and clinical application prospects because of its advantages of high selectivity, small adverse reactions, and low drug resistance.²⁹ METTL4 has been identified as a promising candidate as a novel PROTAC target based on PROTAC tractability assessment.²³ However, PROTACs face challenges because of their low bioavailability and cell specificity.³⁰ Therefore, we developed a bioinspired PROTAC-PEM targeted METTL4 degradation in macrophages in plaques based on the M2 membrane packing technique, which is well used in cancer research overcoming the lower cell specificity and bioavailability.^31–33 The PROTAC-PEM complex inhibited the progression of atherosclerosis by effectively reducing MT-ATP6 6mA levels and alleviating mitochondrial dysfunction and inflammatory response in vitro and in vivo. Therefore, targeting METTL4 by PROTAC-PEM provides a novel strategy for the treatment of atherosclerosis. Further exploration is required to verify the safety and therapeutic efficacy of PROTAC-PEM for the clinical treatment of atherosclerosis.

In conclusion, our study confirmed that overexpressed METTL4 in macrophages promoted atherosclerosis through exacerbating mitoepigenetics-mediated mitochondrial injury and inflammation. In addition, we identified that pemetrexed functions as a METTL4 methyltransferase inhibitor and that PTORAC-PEM could be an effectively therapeutic agent for atherosclerosis targeting macrophages. Collectively, we elucidated a novel mitoepigenetic mechanism of atherogenesis and demonstrated a new therapeutic strategy for atherosclerosis.

ARTICLE INFORMATION

Acknowledgments

H.C., Xuesong Li, and X.W. designed and supervised the study. Xuesong Li, L. Zheng, X.C., and X.H. performed data analysis and drafted the article. L. Zheng, X.C., X.H., Y.T., J.M., Xinyu Li, H.W., and Q.Y. performed the in vivo experiments. L. Zheng, X.C., X.H., M.C., Y.Z., M.D., M.X., and L. Zhang performed the in vitro experiments. L. Zheng collected the clinical data. H.J. performed the informatics analysis. D.H., T.L., F.L., and X.W. supervised the in vivo and in vitro study.

Sources of Funding

This work was supported by the National Natural Science Foundation of China (grants 82270421, 81970428, 31771334, 81800385, and 82270484), major research plan of the National Natural Science Foundation of China (grant 91649125), the British Heart Foundation (grant PG/22/11217), the major project of the Natural Science Foundation of the Jiangsu Higher Education Institution of China (grant 21KJA310006), Postgraduate Research Innovation Program of Jiangsu Province (grant KYCX23_1973), Natural Science Foundation for Higher Education Institutions in Jiangsu Province (grant 23KJB310004), and the Nanjing Health Science and Technology development special fund (grant YKK21259).

Disclosures

None.

Supplemental Material

Expanded Methods

Tables S1–S11

Figure S1–S10

References 34–47