1. Differential gene expression and function analysis of METTL in AS plaque
Seven and 17 differentially expressed METTL genes were screened from GE43292 and GSE100927, respectively. Among them, there were 2 differentially up-regulated genes and 5 down-regulated genes in GSE43292, 8 differentially up-regulated genes and 9 down-regulated genes in GSE100927. The differential gene thermogram and boxplot were shown in Figure 2A-D. Correlation analysis showed that the difference METTL gene in GE43292 and GSE100927 had good correlation, with correlation coefficient between 0.1-0.6 (Figure 2E-F). GO enrichment analysis showed that the differential genes were mainly involved in the methylation biological processes of RNA, ncRNA, mRNA, macromolecules, proteins, and so on. They mainly existed in the complex with mRNA modification, methyltransferase complex and lipid droplets, and mainly played a variety of methylation transfer activities (Fig. 3A-B). KEGG was not enriched in differential gene-related signal pathway.
2. Screening of critical signatures
Utilize RF algorithm and LASSO logistic regression algorithm analysis to screen the critical marker genes. The results showed that 7 genes were identified with RF algorithm in GSE43292 (Fig. 4A-B). ROC curve showed that the AUC value of the model was 0.812 (Fig. 4C), and the AUC values of the 7 marker genes were between 0.664-0.784. Fig. 4D&E showed the nomograms of the model. Because GSE100927 has a large number of samples, we first use machine learning to screen the differential METTL marker genes. The results showed that RF model has a better evaluation function. The AUC values of the 4 models were 0.990, 0.970, 0.965 and 0.970 for RF, SVM, XGB and GLM respectively (Fig 4A-C). The top 10 METTL genes in RF model were METTL13, METTL10, METTL21A, METTL12, METTL11A, METTL8,METTL5, METTL17, METTL7B and METTL21C (Fig 4D). METTL7B and METTL5 were overlapping marker genes of GSE43292 and GSE100927 (Fig 4E).
3. Verification of METTL7B and METTL5 in atherosclerosis
To verify whether METTL7B and METTL5 play a role in the development of atherosclerosis, we first analyzed their expressions in GSE28829 and GSE41571. The results showed that the expression of METTL7B was significantly higher in both advanced AS (GSE28829) and ruptured plaque (GSE41571) than that in the control groups (Fig 5A, p<0.05), but there was no significant difference in the expression of METTL5 (Fig 5A, p>0.05). ROC curves showed the AUC for METTL7B in GSE28829 and GSE41571 were 0.9372 and 0.9333 respectively (Fig 5B). Figure 5C shows the role of METTL7B in the development of AS. Therefore, in the follow-up, we will focus on analyzing the functions of METTL7B.
For further support, we measured the expression of METTL7B protein in human atherosclerosis. Immunohistochemistry showed that the expression of METTL7B in AS was significantly higher than that in normal group, and it was mainly located in foam cells (Fig 6A-B). Immunoblotting showed that the expression of METTL7B in vascular endothelial cells in ox-LDS-induced foaming was significantly higher than that in the control group (Fig. 6C-D). To sum up, the results further support that METTL7B, as a potential diagnostic marker, was located in macrophages in atherosclerosis.
4. Correlation analysis between METTL7B and lipid metabolism
Firstly, the bioGPS database was used to analyze the organ distribution of METTL7B mRNA in human tissues. It was found that the distribution of METTL7B mRNA in the coronary artery was very low, while the distribution of METTL7B mRNA in the liver was the highest (Figure 7A). Further use of The human protein atlas database (https://www.proteinatlas.org/) to analyze the protein expression of METL7B in human tissues, we found that it was the most expressed in the liver (Fig. 7B). After isolating various tissues and organs of mice, Western detection found that METTL7B was mainly expressed in the liver, but very low or no expression in other organs, such as kidneys, skin, etc. (Fig. 7C). In addition, we also used Human protein atlas to analyze the expression of METTL7B in various liver cells. It was found that METTL7B was mainly expressed in hepatocytes, and only a small amount was expressed in kupffer cells and immune cells (Fig. 7D).
Since AS is closely related to lipid metabolism, and liver is the most important organ for lipid metabolism, and GO analysis also shows that METTL7B is located in lipid, it is speculated that METTL7B may participate in lipid metabolism balance. Therefore, we analyzed the differential expression of lipid metabolism genes in GSEA database in GSE43292 and GSE28829 (Fig. 7D), and selected 50 differentially expressed genes related to lipid balance. Correlation analysis shows that METTL7B has a good correlation with the above 50 genes related to lipid metabolism, among which there was a significant correlation with APOE, ADCK1, ABCG1, CAV1, FUNDC2, SCARB1, POLD1, PPARG, PPKAA1, MED13, etc. (Figure 8).
In this study, we also selected blood samples from people with hyperlipidemia and normal blood lipids, and found that the blood level of METTL7B in people with hyperlipidemia was significantly higher than that in the normal control group, and METTL7B was positively correlated with blood lipidsTG and HDL-C (Figure 9A). In addition, we also interfered and over-expressed the LO2 METTL7B gene in human hepatocytes and used oleic acid to induce the formation of fat droplets. It was found that the over-expression of METTL7B could promote the formation of fat droplets compared with the wild-type group, while its knockdown could inhibit the formation of fat droplets (Fig. 9B), indicating that METTL7B could affect fat metabolism.
We collected three AS and three normal blood vessel specimens respectively. Immunohistochemical detection showed that METTL7B expression in plaque vessels was significantly higher than that in normal vessels, and it was mainly distributed in vascular endothelium and foam areas (Figure 1C). This result suggests that METTL7B may be closely related to AS, so we have downloaded the data of different progression stages of AS from the GEO database, namely GSE28829 (including 13 early plaque and 16 late plaque tissue chips), GSE41571 (including 6 stable plaque and 5 ruptured plaque tissue chips), GSE41571 (including 6 non-atherosclerotic and 5 atherosclerotic blood sample chips), and analyzed the expression of METTL7B. It was found that METTL7B was more expressed in advanced plaque and ruptured plaque. In addition, METTL7B was also found to be more expressed in the blood of patients with AS (Fig. 2). These results indicate that METTL7B may participate in the occurrence and development of AS.
5. Correlation analysis between METTL7B and immune cell infiltration
Because METTL7B is also expressed in blood vessels, and immuno -histochemistry also shows that it is mainly located in foam cells of AS plaque, the single cell sequencing analysis of The human protein atlas database was performed showing that METTL7B was mainly located in macrophages (Figure 10A) in blood vessels. In order to further analyze whether METTL7B is related to immune cell infiltration, it was found that METTL7B was significantly correlated with Macrophages M0, B cell memory, T cells CD4 memory resting, and T cells CD (Fig. 10B-C, P<0.05).
6. Correlation analysis between METTL7B and eferocytosis
There are 5 and 6 differentially expressed genes of 14 eferocytosis related genes in GSE43292 and GSE100927, respectively, including CXCL1, CALR, ICAM3 and MFGE8. Correlation analysis showed that METTL7B had a significant positive correlation with CALR, ICAM3 and CXCL1 in GSE43292 and GSE100927 databases, and a significant negative correlation with MFGE8 in GSE100927 (Figure 11B). These results suggest that METTL7B may be related to macrophage-mediated cell burial.
7. Screening small molecule compound targeting MTTL7B
Through the docking analysis of the 3D structure of METTL7B and the small molecular compounds in the zinc small molecular structure database, we found that the top 10 small molecular compounds with the binding ability of METTL7B were shown in Table 3, namely, ZINC000008215434, ZINC000001536109, ZINC000022448696, ZINC00000200505305, ZINC000038144598, ZINC000001529323, ZINC000002036915, ZINC000053073961, ZINC000004099104 and ZINC000030731084. Their corresponding molecular names are FAD, Pralatrexate, Indinavir, 5-Methyltetrahydrolic acid disodium salt, Alexidine dihydrochloride, Methodrexate, Aminopterin, Antrafenine, SN38 glucuronide, APC.