Bioinformatics study and experimental validation of METTL3 regulation of immune-related genes affecting placental vascular development

doi:10.21203/rs.3.rs-3379965/v1

Download PDF

Article

Bioinformatics study and experimental validation of METTL3 regulation of immune-related genes affecting placental vascular development

https://doi.org/10.21203/rs.3.rs-3379965/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The proper development of the placental vascular system is a crucial factor in ensuring fetal health. m6A modification is a key pathophysiological mechanism in placental vascular development. However, the specific mechanism by which m6A influences placental vascular development remains unclear. Here, we explored the role of 21 m6A regulators in placental development based on the Gene Expression Omnibus (GEO) database. Following a series of machine learning techniques, METTL3 was recognized as the pivotal m6A regulator. We subsequently employed consensus clustering analysis to delineate two distinct m6A isoforms, and investigated their correlation with immune cells. Further, through weighted gene co-expression network analysis (WGCNA) coupled with correlation analysis, we pinpointed METTL3-associated placental development genes. These genes were notably enriched in immune-related categories. Furthermore, we uncovered immune-related differentially expressed genes that were associated with differentially expressed m6A regulators. Additionally, we performed an immune infiltration analysis to gain a deeper understanding of how these genes interact with immune cells. Ultimately, to validate our findings, we carried out animal experiments. In conclusion, our study found that targeting METTL3 could affect placental vascular development, which may provide guidance for the clinical treatment of placental-like diseases.

Biological sciences/Immunology

Health sciences/Biomarkers

As a key pivot connecting mother and fetus, the placenta is an important organ for maintaining the growth and development of the fetus, with many functions such as material exchange, defense, synthesis, and immunity(1). Placental vessels are essential components of the placenta. Developmental disorders of the placental vascular system can cause placental dysfunction thus resulting in adverse pregnancy outcomes such as preeclampsia, intrauterine growth restriction, and placental trophoblastic disease(2, 3). The development of the placental vessels is a complex and delicate process that is not only controlled by a variety of signaling molecules, including pro-angiogenic and anti-angiogenic factors but also influenced by the immune system(4). During pregnancy, the immunoregulatory factor Heme oxygenase-1 is highly expressed in the placenta, which improves the development of placental blood vessels and regulates vascular tension by inhibiting immune inflammatory reactions(5). Pregnant malaria patients Activation of the maternal immune system during infections, such as malaria during pregnancy, can trigger the production of large amounts of pro-inflammatory cytokines in the placenta, which can negatively affect placental angiogenesis and result in adverse fetal outcomes, such as pre-eclampsia and congenital heart disease(6–8).

N6-methyladenosine (m6A) is the most common internal modification found in mammalian mRNAs which is dynamically regulated by various factors(9). RNA methyltransferases (writers), like METTL3, METTL14, and WTAP, catalyze the installation of m6A on mRNAs. Demethylases (erasers), including ALKBH5 and FTO, remove m6A modifications from mRNA. Binding proteins (readers), including YTHDC1/2, YTHDF1/2/3, CBLL1, HNRNPC, HNRNPA2B1, LRPPRC, and IGF2BP1, specifically recognize methylation sites(10). Through its regulation of mRNA splicing, localization, translation efficiency, and stability, m6A modification plays a crucial role in various normal and pathological processes(11). There is mounting evidence for the involvement of abnormal m6A methylation modification in placental dysfunction. As Wang et al. found(12, 13), YTHDC1 in placental tissues induces abnormal methylation of circMPP1, which inhibits placental development. In clinical studies, pre-eclampsia has been shown to be associated with the downregulation of m6A modification mediated by the WTAP/IGF2BP1/HMGN3 axis(9). At the maternal-fetal interface, FTO downregulation impairs immune tolerance and angiogenesis, resulting in abnormal methylation, oxidative stress, and spontaneous abortion(14). In recent years, m6A modifications have been found to play important roles in immune regulation(15, 16). However, it remains unclear how m6A methylation modifications regulate placental vascular development through the immune system.

During the last decade, the cross-fertilization of taxonomy with genomics, bioinformatics, and big data has created a new frontier in the life sciences. In view of this advantage, we analyzed the role of m6A modification in placental development and found that m6A regulator METTL3 can affect five immune-related DEGs (THBS1, PDGFRB, JAG1, CXCL12, and ANGPT1) and thus placental vascular development. This discovery can provide a better understanding of the role of m6A modification in placental development and guide the treatment of clinical placenta-related diseases.

The landscape of m6A regulators

In this study, we curated and analyzed a set of 21 recognized m6A regulators, which comprised 14 writers, 5 readers, and 2 erasers. Through expression profiling, we identified 13 differentially expressed m6A regulators (DEm6AR) that showed significant variation between the first and third trimesters (Fig. 1A). Of these, seven genes (FMR1, YTHDF2, YTHDF2, METTL3, IGFBP2, CBLL1, and YTHDC1) were upregulated, while six genes (RBMX, WTAP, RBM15, ALKBH5, HNRNPA2B1, and HNRNPC) were downregulated. To explore the transcriptome relationships among the m6A regulators, we conducted correlation analysis and found a close association between writers, readers, and erasers (Fig. 1B).

Machine learning

Seven potential diagnostic biomarkers were identified using the LASSO regression algorithm (Fig. 1C), while the RF algorithm identified four diagnostic genes (Fig. 1D). These four genes were identified as robust diagnostic biomarkers by using a Venn diagram (Fig. 1E). A ROC analysis of the GSE12285 dataset found that METTL3, RBMX, YTHDC1, and YTHDF2 were effective in discriminating between first-trimester and third-trimester samples (Fig. 1F).

GSVA and GSEA analysis and experimental validation

METTL3, an m6A-modified methylesterase, was found to exhibit significant differences in GSEA analysis across different expression groups, specifically in the calcium signaling pathway, intestinal immune network for IgA production, and N glycan biosynthesis (Fig. 1G). Further analysis through GSVA revealed that the high-expression group of METTL3 was highly enriched in primary immunodeficiency, lysine degradation, and intestinal immune network for IgA production, while the low-expression group was mainly enriched in maturity-onset diabetes of the young, proteasome, and regulation of autophagy (Fig. 1H).

To verify the importance of METTL3 in placental development, the expression of METTL3 in placentas from the first-trimester group and third-trimester group was examined. It was discovered that the first-trimester placentas exhibited significantly lower expression of METTL3 compared to the third-trimester group (Fig. 1I, J).

Subtype analysis of m6A regulators

In order to delve deeper into the role of m6A regulators in placental development, the consensus clustering method was used. The third-trimester samples were classified into two distinct subgroups, namely m6A clusters C1 and C2, using a k value of 2 as illustrated in Fig. 2A. The t-SNE results underscored a pronounced differentiation between these two m6A clusters, as depicted in Fig. 2B. Subsequently, we employed single-sample gene set enrichment analysis (ssGSEA) to quantitatively assess variations in infiltrating immune cells between the two m6A clusters and establish correlations between the expression of DEm6AR and immune cells (Fig. 2C). Significant discrepancies in various immune cell types were observed, including Activated B cell, Activated CD4 T cell, CD56 bright natural killer cell, Eosinophil, Gamma delta T cell, Immature dendritic cell, Macrophage, Mast cell, Monocyte, Type 2 T helper cell, Central memory CD8 T cell, and Effector memory CD8 T cell, all of which exhibited distinct distributions between the two clusters. Notably, a negative correlation was identified between METTL3 expression and a range of diverse immune cell types. Moreover, we meticulously examined variations in immune cells with either high or low METTL3 expression (Fig. 2D), revealing a substantial enrichment of immune cells characterized by lower levels of METTL3 expression. This observation strongly suggests a close association between reduced METTL3 expression and immune responses.

Furthermore, our investigation brought to light the upregulation of METTL3, WTAP, YTHDC1, HNRNPA2B1, and HNRNPC in m6A cluster C1, while RBMX, IGFBP2, CBLL1, and RBM15 exhibited higher expression levels in m6A cluster C2 (Fig. 2E). In our pursuit to comprehend the potential biological implications of the m6A-related cluster, we identified differentially expressed genes (DEGs) within this cluster by comparing m6A cluster C1 with m6A cluster C2 (Fig. 2F). Importantly, GSEA analysis revealed a significant enrichment of immune-related terms within these DEGs (Fig. 2G).

Construction of the m6A gene signature

We categorized the third-trimester samples into two distinct gene clusters, designated as gene clusters C1 and C2, based on the differential expression of genes associated with the m6A cluster. This partitioning strategy allowed for a targeted exploration of specific regulatory mechanisms (Fig. 3A). In the heatmap, a clear pattern emerged: gene cluster C1 encompassed 237 up-regulated genes and 462 down-regulated genes (Fig. 3B). Moreover, these gene clusters exhibited significant variations in the expression levels of DEm6AR (Fig. 3C).

Furthermore, the utilization of single-sample gene set enrichment analysis (ssGSEA) illuminated discernible discrepancies in the distribution of immune cells between the two gene clusters (Fig. 3D).

Correlations between m6A subgroups

Principal component analysis (PCA) was employed to calculate the m6A score. Notably, there were significant differences in the m6A score between the m6A clusters and gene clusters (Fig. 3E, F), with both m6A cluster C2 and gene cluster C2 displaying elevated m6A scores. This observation suggests a potential connection between the m6A score and immune activation. This trend was consistently mirrored in the levels of infiltration by various immune cell types, including Activated B cells, CD56 bright natural killer cells, Eosinophils, Gamma delta T cells, Immature dendritic cells, Macrophages, Type 2 T helper cells, Central memory CD8 T cells, and Effector memory CD8 T cells, in both m6A cluster C2 and gene cluster C2. To provide a visual representation, the distribution of the two m6A clusters, two gene clusters, and two m6A scores was depicted using a Sankey diagram (Fig. 3G).

WGCNA co-expression analysis

Firstly, WGCNA analysis of clinical trait correlations was performed to screen out the top 25% genes with the largest fluctuation for further WGCNA analysis, with outliers removed. Figure 4A shows that when power = 8, scale independence reached 0.9 and mean concordance was high. To detect outliers, we constructed trees using the module’s feature values, and then trees belonging to the same branch with very close distances are merged with the intercept value set to 0.5 (Fig. 4B). Figure 4C shows co-expression modules built by merging similar modules. Correlation analysis was performed to identify modules associated with specific traits based on the eigenvalues of the samples in each module and the characteristics of the samples, with genes represented in red being highly positively correlated in mature placentas. (Fig. 4D).

We followed the same approach as our WGCNA analysis of clinical trait correlations to analyze the correlations after typing, and we excluded outlier samples. Using a power = 7, we achieved scale independence of 0.9 and identified modules by color (Fig. 4E). We then discovered post-typing features and merged similar modules (Fig. 4F, G), which resulted in the identification of genes highly positively correlated in m6A cluster C1, represented by the dark red color (Fig. 4H).

Identification and functional enrichment analysis of METTL3-related genes

Through the intersection of clinical traits and gene modules with specific common traits, a total of 94 genes were pinpointed (Fig. 4I). Subsequently, by conducting correlation analysis, 42 genes closely associated with METTL3 was identified (Fig. 4J). Following this, functional enrichment analysis was performed (Fig. 4K), revealing that genes linked to METTL3 exhibited functional enrichment in processes including positive regulation of chemokine secretion, negative regulation of cytoskeleton organization, regulation of type 2 immune response, and type 2 immune response.

Identification of m6A-targeted immune genes and PPI analysis

We identified DEGs between the first-trimester and third-trimester groups (Fig. 5A). Subsequently, through the integration of genes identified using a three-step process, we successfully identified m6A modification-related immune genes (Fig. 5B). In order to examine the interactions between these immune genes related to m6A modification, we conducted a PPI analysis (Fig. 5C). Degrees, DMNC, MCC, and MNC were used in our analysis to identify genes. From their intersecting genes, we identified 5 hub genes (Fig. 5D). Hub genes are functionally enriched in the MAPK signaling pathway, focal adhesion, rap1 signaling pathway and ras signaling pathway (Fig. 5E).

Immune cell infiltration analysis

We used ssGSEA to quantify immune infiltration and obtain the enrichment levels of 27 immune cells and immune-related functions in the samples (Fig. 6A). The correlation between the 27 immune cells is shown in Fig. 6B. In addition, the correlation between these immune cells and hub genes was analyzed (Fig. 6C).

Validation of the hub genes

Then, five hub genes in the dataset GSE12588 were verified, and all these five genes were found to be upregulated in the dataset (Fig. 7A). The relevant pathways from the five gene pathway enrichment were then called up for visual analysis (Fig. 7B). And to further verify their efficacy, we validated the 5 hub genes in GSE12558 (Fig. 7C) and GSE9986 (Fig. 7D). ROC results indicated that all hub genes can discriminate between the first-trimester and third-trimester samples effectively in both datasets. We found that METLL3 could target all these genes, and among these gene-related pathways, the focal adhesion pathway could target all four genes at the same time. What’s more, the FAK signaling pathway may have an impact on the early development and function of the IVF-ET placental villi.

Moreover, our in vivo animal experiments have found that the mRNA expression of all these 5 genes in the first-trimester placentas was significantly lower than that in the third-trimester placentas, which is consistent with our results above (Fig. 7E).

Experimental validation of the Pregnant Rat Models of Fear Stress

Our sequencing study found that the expression of METTL3 was significantly downregulated in fear stress group (Fig. 8A). Thus, we utilized the pregnant rat model of fear stress to investigate whether METTL3 affects placental vascular development by targeting the five immune-related genes.

Compared to the control group, the third-trimester rats with fear stress exhibited heightened anxiety, as evidenced by a significant preference for the periphery of the arena in the open field test (Fig. 8B, C), as well as an increased duration spent and entries in the closed arm in the elevated plus maze test (Fig. 8D). These findings indicate the successful establishment of the fear stress model. As shown in Fig. 8E, the size of placentas was smaller in the fear stress group than in the control group. Additionally, the weight of the offspring and placentas was lower in the fear stress group (Fig. 9A). The mRNA expression of five hub genes in the fear stress group was significantly down-regulated compared to the control group (Fig. 9B). Transmission electron microscopy results showed that in the control group, placental vascular endothelial cells were tightly connected, with abundant microvilli on the cell surface, and intact nuclear membranes. In contrast, the fear stress group exhibited evident signs of vascular damage in the placentas, with increased proliferation of endothelial cells protruding into the lumen, and a reduction in surface microvilli (Fig. 9C).

As an essential organ during pregnancy, the placenta provides oxygen, nutrition, and protection to the fetus(17). Any changes in placental vascular development and function may impact pregnancy outcomes as well as child health, such as pre-eclampsia and fetal growth restriction(18, 19). m6A is an mRNA modification mediated by a complex of multiple proteins(20, 21). Among them, METTL3, an active transferase that catalyzes methyl transfer from the cofactor S-adenosylmethionine to the N6 position of adenosine, plays a significant role in placental development and function(22). It has been shown that METTL3 expression is significantly down-regulated in pre-eclamptic placentas when compared to normal placentas, whereas the expression of METTL3 in the villi of spontaneously aborted pregnancies was up-regulated (23, 24). It has also been found that pregnant rats with fear stress can affect the expression level of METTL3 in their placentas, causing placental dysfunction(12). Furthermore, METTL3 is involved in processes such as proliferation, migration, and lumen formation of vascular endothelial cells(25). Our study found that METTL3 expression was upregulated in the third-trimester samples, which indicated that METTL3 is essential for placental development.

The third-trimester samples were grouped into two m6A clusters and two gene clusters by applying the unsupervised clustering analysis. In addition, both m6A cluster C2 and gene cluster C2 showed a trend of higher-level immune cell infiltration. Next, the m6A scores were calculated to quantify the clusters of m6A with higher m6A scores found in m6A cluster C2 and gene cluster C2. Then by WGCNA analysis, we further identified genes that were associated with both m6A modification and placental development. We further screened for METTL3-related genes and analyzed them for functional enrichment, finding that these genes were also significantly enriched in immune-related entries, suggesting that METLL3 could regulate immune-related genes and thus affect placental development. Successful pregnancy depends on the maintenance of maternal-fetal immune homeostasis. Innate and adaptive immune cells are vital for supporting placental development, maintaining maternal-fetal immune tolerance, and regulating the inflammatory response during labor(26, 27). Transcriptomic studies of placentas delivered at term and preterm have revealed differential expression of multiple immune cells, including a significant increase in macrophage numbers(28). There is growing evidence for the important role of m6A in immune cell differentiation and placental function(14, 29). It remains unclear, however, how m6A regulates the placental function and immune responses.

To further explore how m6A regulators affect immune-related functions, we further screened for target genes of differential m6A regulators and performed PPI analysis to further screen hub genes. Based on the above findings, THBS1, PDGFRB, JAG1, CXCL12, and ANGPT1 were identified as the five most significant genes. Mohammed et al. found that THBS1 could participate in vascular growth by inhibiting endothelial vascularization(30). In addition, clinical studies have found that THBS1 is upregulated in placental trophoblast cells of preeclampsia pregnant women, and subsequent animal experiments have shown that overexpression of THBS1 inhibits the cAMP signaling pathway and affects syncytialization of trophoblast cells, thereby inducing preeclampsia(31, 32). Lehtoranta et al. found that decreased expression of PDGFRB caused abnormal placental hemodynamics leading to gestational diabetes(33). Based on the GEPIA database, JAG1 and VEGFA exhibited a positive correlation(34), and overexpressing JAG1 promoted the proliferation of endothelial cells(35, 36). By regulating angiogenesis and cell viability, CXCL12-CXCR4 signaling ensures normal placental development(37). Clinical studies have shown that malaria during pregnancy impairs placental vascular remodeling by lowering Angpt-1 levels in the placenta(38). Furthermore, Haneda et al.(39) found that high levels of estradiol during pregnancy increased placental angiogenic activity and maintained normal placental function by increasing ANGPT1 and VEGF expression. High-fat diets during pregnancy increase the prevalence of adverse fetal outcomes by increasing ANGPT1 protein expression, causing local placental hypoxia and vascular thrombosis of the chorionic vasculature(40). Our animal experiment results indicated that all these five hub genes were significantly downregulated in the placentas of the first-trimester group compared to the third-trimester group.

All five hub genes were found to be targets of METTL3, and they were significantly enriched in focal adhesion. A key component of vascular development and angiogenesis is focal adhesion kinase (FAK), which regulates endothelial cell migration, proliferation, and survival(41, 42). The conditional knockout of FAK in Ecs may result in aberrant abnormal vessel growth through lamellipodial extensions, cytoskeleton changes, and reduced EC tubulogenesis(43). Moreover, a study using endothelium-specific inducible FAK knockout mice showed a close correlation between FAK deficiency and tumor-induced decreased angiogenesis(44). The transfer of nutrients and oxygen provided by the placenta is essential for normal fetal growth and development. It is well known that placental vascular dysplasia leads to adverse fetal outcomes(45). Recent studies have demonstrated a crucial regulatory role for the immune response in placental vascular development(46). A balance must be maintained between pro- and anti-inflammatory activity within the placenta during pregnancy. Disruption of this balance is associated with abnormalities in the development and function of the placental vasculature, which can lead to recurrent spontaneous abortion (RSA) and other complications(47). It has been found that adverse pregnancy outcomes in miscarriage-prone mice are associated with decreased placental hemodynamics caused by placental inflammation(48).

To further verify our results, the pregnant rat model of fear stress was utilized. We analyzed the sequencing data, examined the expression of five key genes, and obtained the development of placental vascular. We found that fear stress leads to the downregulation of METTL3, resulting in adverse pregnancy outcomes. Additionally, all five immune-related hub genes were also downregulated. Thus, we propose that METTL3 may mediate placental vascular development by targeting five immune related genes.

Our study unveiled a noteworthy role for m6A methylation modification in placental development. Notably, the pivotal regulator METTL3 exhibited the ability to impact immune-related genes, consequently influencing placental vascular development. Nevertheless, certain limitations persist within our study. Our conclusions were derived from the analysis of datasets sourced from public databases, and the absence of clinical and animal studies hinders the verification of the relationship between METTL3 and placental vascular development. Subsequent investigations will delve into the role of METTL3 in placental development as well as its involvement in placental-related disorders.

Data collection and processing

Using the Gene Expression Omnibus (GEO) database, we obtained data on placental gene expression from the GSE28551 dataset, which contains data for the first-trimester and third-trimester samples. And there were 21 m6A regulators identified by previous studies.

Detection and interactions of m6A RNA methylation regulators

In total, 21 m6A regulators were selected according to a recent publication, including five writers (METTL3, METTL14, RBM15, WTAP and CBLL1), two erasers (FTO, ALKBH5), and 14 readers (YTHDC1, YTHDC2, YTHDF1, YTHDF2, YTHDF3, IGFBP2, IGFBP3, HNRNPA2B1, HNRNPC, FMR1, LRPPRC, ELAVL1, RBM15 and RBMX). GSE28551 datasets were used to extract the expression profiles of 21 m6A regulators. The regulation of m6A was analyzed using Pearson's correlation methods. To identify m6A regulators that were significantly differentially expressed between the first-trimester and third-trimester placentas, Wilcoxon's test was used. And to visualize the expression pattern of DEm6A, the heatmap was used.

Machine learning

Recently, machine learning has become increasingly popular as a tool for analyzing algorithms. Three ML algorithms were used - The least absolute shrinkage and selection operator (LASSO), Support vector machine-recursive feature elimination (SVM-RFE), and Random forest (RF). LASSO is a regularized regression algorithm using the “glmnet” package. We can rank features based on recursion using SVM-RFE. RF analysis can be used to select features, calculate the mean decrease gini (MDG) for each gene and rank them, and determine the optimal number of features by adding differential genes one by one from the largest to the smallest MDG value to determine the accuracy of the classification results. These algorithms allowed us to screen for biomarkers more efficiently.

GSVA and GSEA analyses

Using the GSE28551 dataset, samples were divided into two groups (high and low expression groups) according to their median expression levels. Gene set variation analyses (GSVA) and gene set enrichment analyses (GSEA) were performed using the gene set "c2 cp.kegg.v7.4. symbols.gmt” as a reference to identify the enriched KEGG pathways. To obtain significantly enriched gene sets, we selected gene sets with P < 0.05 in this study.

Unsupervised clustering analysis

Using the "ConsensusClusterPlus" software package, we performed unsupervised clustering analyses on the third-trimester samples, dividing them into two m6A clusters based on their expression profiles. We set the following parameters: max cluster number (maxK) = 10, the proportion of items to sample (pItem) = 0.8, and distance = Pearson. In order to identify the optimal number of clusters, the empirical cumulative distribution function plot was applied. For dimensionality reduction, tSNE was conducted.

DEGs between the two m6A clusters and functional analysis

The differentially expressed genes (DEGs) between the two m6A clusters were analyzed by the "limma" software package with thresholds of |log2FC|≥2 and P＜0.05. To explore the potential biological functions of these DEGs, GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) functional enrichment analyses were performed using "clusterProfiler".

Construction of the m6A gene signature

In order to quantify m6A patterns, each sample was scored using the PCA algorithm. Calculated the m6A score as Σ (Expi * PCi), where PCi and Expi represent the principal component and expression level of each gene, respectively.

Weighted co-expression network construction

To build a co-expression network of DEGs, the Weighted co-expression network construction (WGCNA) package in R was used. Then we performed Pearson's correlation matrix on pairs of genes. To detect the outliers, we calculated the module features genes’ dissimilarity, selected tangents to the module dendrogram, and merged some modules. Under the dendrogram, colored rows showed module assignments based on dynamic tree cuts. After conducting an analysis of the module-trait relationships between module trait genes and clinical traits, two modules associated with a specific trait were identified. Subsequently, by employing the same approach, we clustered the two modules associated with the trait and then Venn diagrammed the most significant trait modules to identify significant clinical traits and genes.

Correlation analysis

To identify genes that may be associated with METTL3, a Pearson's correlation analysis on METTL3 and clinical traits was performed. For this experiment, our thresholds were as follows: correlation coefficient (|r|) > 0.5 and P < 0.001. Genes associated with METTL3 were then analyzed using Functional Enrichment Analysis.

Identification of m6A-targeted immune DEGs

In order to screen out m6A-targeted immune DEGs, four methods were integrated as follows:

(1) To determine the DEGs between the first-trimester and third-trimester samples, the “limma” R package was used with the cutoff criteria of P < 0.05 and | log2fold change (FC)|>1.

(2) To analyze the DEGs between m6A cluster C1 and C2, the “limma” R package was used with the cutoff criteria of P < 0.05 and | log2fold change (FC)|>1.

(3) To predict the DEm6A target genes, the RM2Target database (http://rm2target.canceromics.org/#/home) was utilized. This database contains numerous valid and potential m6A targets which have been verified through high-throughput analyses of m6A RNA methylation regulators, including RIP-seq, ChIP-seq, RNA-Seq, and m6A-Seq.

(4) We utilized InnateDB (https://www.innatedb.com/) and Immport (https://immport.org/shared/home) databases to identify immune-related genes.

To identify m6A-targeted immune DEGs, we performed an intersection analysis of DEGs, DEm6A-related genes, DEm6A-targeted genes, and immune-related genes based on the four methods. Further, we employed the ClusterProfiler R package to perform GO and KEGG enrichment analysis.

Protein-protein interaction analysis

The STRING database (https://string-db.org/) and Cytoscape (version 3.8.2) software were used to observe interactions between immune-related DEGs, and then we used the Pearson correlation function of the R software package “corrplot” to determine their correlation. Furthermore, Protein-protein interaction (PPI) hub genes were obtained from the CytoHubba plug-in in Cytoscape through Degree, DMNC, MCC, and MNC.

Assessment of immune infiltration

We calculated 28 immune cell types from 28 published immune gene sets by using the single-sample gene set enrichment analysis (ssGSEA) method from the R package.

Download of placental sequencing data from Pregnant Rat Models of Fear Stress

Numerous studies have indicated that fear stress can affect the expression of METTL3(17, 18). In our previous work, we conducted placental sequencing studies using pregnant rat models subjected to fear stress. The sequencing data were obtained from NCBI's SRA database (https://submit.ncbi.nlm.nih.gov/subs/sra/) under accession number PRJNA837736. To ensure uniformity and facilitate analysis, the RNA-seq data were converted to TPM (Transcripts per million reads) format and log2 transformed based on the FPKM (Fragments Per Kilobase per Million) values. This transformation allowed for consistency and comparability in the subsequent analysis of RNA-seq data.

Animals and Pregnant Rat Models of Fear Stress

According to the results of MeRIP-seq data analysis, we found the level of METTL3 gene was significantly down-regulated in the placentas from pregnant rats of fear stress. We developed the gestational fear stress model using a modified standard electroshock method(12), which was repeated daily from GD1 to GD19.

Female Wistar rats were mated with Wistar males in a 2:1 ratio (both purchased from Bejing Vital R1ver Laboratory Animal Technology CO., Ltd). and the presence of a vaginal plug indicated a gestation day of 0.5 (GD0.5). The dams were randomly divided into three groups: the first-trimester group (GD10), the third-trimester group (GD19) (control)(19) and the fear stress group (the third-trimester rats with fear stress) (n = 12/each group). Animal experimentation in this study complied with strict ethical guidelines and was approved by the Animal Ethics Committee of Henan University of Traditional Chinese Medicine, China.

Open field test

The open field test (OFT) was performed on the third-trimester group and the fear stress group rats at 19 days of conception. The rats were placed in a black square field (100cm × 100cm × 40cm) and acclimated for at least 1 hour. After acclimation, each rat was allowed to explore freely for 5 minutes in the open field. Any-Maze video tracking software and analysis system were used to monitor the rats' activity and analyze the results of the open field test (Stoelting Co, Wood Dale, IL, USA). The field was cleaned with 75% alcohol between each subject. The time spent in the central area, the total distance traveled, and the entries across the center were recorded.

Elevated Plus Maze test

The Elevated Plus Maze test (EPM) test was conducted on the third-trimester group and the fear stress group rats at 19 days of conception. There are two open arms (43cm length x 11cm width) and two closed arms (43cm length x 11cm width) connected by a central platform 75 cm above the ground in the EPM. Rats were allowed to explore freely for 5 minutes in the EPM. Any-Maze video tracking software was used to monitor the rats' activity (Stoelting Co, Wood Dale, IL, USA). The time spent and the number of entries in both the open and closed arms were quantified.

Western blot

The protein of placentas was extracted using RIPA lysis buffer (Servicebio, G2002-100 ML) containing PMSF (Servicebio, G2008-1 ML), and subsequently, the protein was denatured. Following separation of the samples by 12% SDS-PAGE gel electrophoresis and transfer to PVDF membranes (Servicebio, WGPVDF45), the membranes were blocked with 5% defatted milk for 1 hour. Next, the membranes were incubated overnight at 4°C with primary antibodies (1:2000 diluted METTL3, Servicebio, GB114688; 1:2000 diluted ACTIN, Servicebio, GB11001), followed by three TBST washes the next day and incubation with a secondary antibody (1:5000 diluted) for 1 hour. After the incubation and washing steps, the high-sensitivity ECL luminescent liquid (Servicebio, G2014-50 ML) was uniformly and dropwise added. The bands were detected using the Bio-RadChemiDoc MP luminescence imaging system. The gray value of the images on the film was quantified using ImageJ.

Quantitative real-time PCR

Placentas were obtained on day 10 of pregnancy for the first-trimester group, while placentas were collected on day 20 for both the third-trimester group and the fear stress group. We extracted the total RNA from placental tissue with an RNA extraction solution (Servicebio, Wuhan, China) and measured RNA concentration and purity (Thermo Fisher Scientific, Waltham, USA) with an ultra-micro spectrophotometer (Thermo Fisher Scientific, Waltham, USA). RNA was reverse transcribed to cDNA according to the instructions of the SweScript RT I First Strand cDNA Synthesis Kit (Servicebio, Wuhan, China). The qRT-PCR assay was performed using a Bio-rad CFX fluorescence quantitative PCR instrument (Hercules, USA). In this experiment, the reaction conditions were: pre-deformation at 90°C in 10 minutes, denaturation at 95°C in 10 minutes, and annealing extension at 60°C in 30 seconds. We calculated relative expressions using 2-△△Ct. The selected genes and primers are shown in Table 1.

Table 1

The primer sets for associated mRNA.
Rat	Forward Primer (5′-3′)	Reverse Prime (5′-3′)	Fragment length (bp)
METTL3	GATGGCTTTCTTTTCCTCTGGG	GGTTGAAGCCTTGGGGATTT	214
THBS1	ACTGTGACCCTGGACTTGCTGT	TGAGTATCCCTGAGCCCTTGTG	206
PDGFRB	GAAGCAGCCATGAACCAGGAT	GTCACTCGGCATGGAATCGT	242
JAG1	GTAATCGCATCGTACTGCCTTTC	CAGGCATAATGTCCAAAGAAGTCA	286
CX3CL12	AGTGACGGTAAGCCAGTCAGC	GGGCACAGTTTGGAGTGTTGAG	115
ANGPT1	AAGTCGGAGATGGCCCAGATA	GTTGGATTTCAAGACGGGATGTT	163
GAPDH	CTGGAGAAACCTGCCAAGTATG	GGTGGAAGAATGGGAGTTGCT	138

Transmission electron microscopy

After fixing 2 mm samples of the placenta in 2.5% glutaraldehyde (Merck, Germany), the samples were post-fixed in 1% osmium tetroxide (OsO4) (Merck, Germany) for 1 hour. The tissues underwent gradient dehydration using ethanol and acetone before being embedded in epoxy resin. Ultrathin sections were stained with uranyl acetate and lead citrate. The transmission electron microscopy (TEM) study was conducted at Henan University of Traditional Chinese Medicine.

Our study offers valuable insights into the mechanisms underlying m6A modification during placental development. Our findings suggest that the methylation enzyme METTL3 may selectively target immune-related genes, thereby exerting a significant influence on placental vascular development.

Acknowledgments

The authors would like to thank GEO (http://www.ncbi.nlm.nih.gov/geo) and R for providing the software for free.

Author contributions statement

JJZ and ZXD designed this study, collected the data and performed the bioinformatic and statistical analysis, figure visualization, and manuscript writing. QS, XG and PC conducted animal and qRT-PCR experiments. YYW and YHW managed the literature searches. LPY, ZQZ and JLH help to revise the manuscript. All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.

Funding

This work was supported by the National Natural Science Foundation of China (No.81973596) and College Students' Innovative Entrepreneurial Training Plan Program (202210471031).

Availability of data and materials

The dataset analyzed during the current research period can be obtained at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE28551.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Ethics approval and consent to participate

Study protocols involving rats were approved by Animal Ethics Committee of Henan University of Traditional Chinese Medicine (permit number: DWLL202108003).

Declaration of interests

The authors declare no competing interests.

Ni G. et al. Three-dimensional morphological revealing of human placental villi with common obstetric complications via optical coherence tomography. Bioeng Transl Med. 8. e10372. (2023).
de Vos E.S. et al. Assessment of first-trimester utero-placental vascular morphology by 3D power Doppler ultrasound image analysis using a skeletonization algorithm: the Rotterdam Periconception Cohort. Hum Reprod. 37. 2532-2545. (2022).
Brosens I., Brosens J.J., Muter J., Puttemans P. & Benagiano G. Preeclampsia: the role of persistent endothelial cells in uteroplacental arteries. Am J Obstet Gynecol. 221. 219-226. (2019).
Ma H. et al. Conditioned medium from primary cytotrophoblasts, primary placenta-derived mesenchymal stem cells, or sub-cultured placental tissue promoted HUVEC angiogenesis in vitro. Stem Cell Res Ther. 12. 141. (2021).
Ozen M., Zhao H., Lewis D.B., Wong R.J. & Stevenson D.K. Heme oxygenase and the immune system in normal and pathological pregnancies. Front Pharmacol. 6. 84. (2015).
Raguema N., Moustadraf S. & Bertagnolli M. Immune and Apoptosis Mechanisms Regulating Placental Development and Vascularization in Preeclampsia. Front Physiol. 11. 98. (2020).
Weckman A.M., Ngai M., Wright J., McDonald C.R. & Kain K.C. The Impact of Infection in Pregnancy on Placental Vascular Development and Adverse Birth Outcomes. Front Microbiol. 10. 1924. (2019).
Brown A.S. & Meyer U. Maternal Immune Activation and Neuropsychiatric Illness: A Translational Research Perspective. Am J Psychiatry. 175. 1073-1083. (2018).
Bian Y. et al. WTAP dysregulation-mediated HMGN3-m6A modification inhibited trophoblast invasion in early-onset preeclampsia. Faseb j. 36. e22617. (2022).
Guo R. et al. The diagnostic significance of integrating m6A modification and immune microenvironment features based on bioinformatic investigation in aortic dissection. Front Cardiovasc Med. 9. 948002. (2022).
Miyake K. et al. A cancer-associated METTL14 mutation induces aberrant m6A modification, affecting tumor growth. Cell Rep. 42. 112688. (2023).
Wang Q. et al. Fear Stress During Pregnancy Affects Placental m6A-Modifying Enzyme Expression and Epigenetic Modification Levels. Front Genet. 13. 927615. (2022).
Wang D., Guan H. & Xia Y. YTHDC1 maintains trophoblasts function by promoting degradation of m6A-modified circMPP1. Biochem Pharmacol. 210. 115456. (2023).
Qiu W. et al. RNA Demethylase FTO Mediated RNA m(6)A Modification Is Involved in Maintaining Maternal-Fetal Interface in Spontaneous Abortion. Front Cell Dev Biol. 9. 617172. (2021).
Li C. et al. Integrating machine learning and bioinformatics analysis to m6A regulator-mediated methylation modification models for predicting glioblastoma patients' prognosis and immunotherapy response. Aging (Albany NY). 15. 4051-4070. (2023).
Li C. et al. Role of m6A modification in immune microenvironment of digestive system tumors. Biomed Pharmacother. 164. 114953. (2023).
Wu S., Liu K., Zhou B. & Wu S. N6-methyladenosine modifications in maternal-fetal crosstalk and gestational diseases. Front Cell Dev Biol. 11. 1164706. (2023).
Arruda Correia M.L., Peixoto Filho F.M., Gomes Júnior S.C. & Peixoto M.V.M. Effects of intra-abdominal hypertension on maternal-fetal outcomes in term pregnant women: A systematic review. PLoS One. 18. e0280869. (2023).
Freise L. et al. Three-Dimensional Histological Characterization of the Placental Vasculature Using Light Sheet Microscopy. Biomolecules. 13. (2023).
Zhang Y. et al. circRNA N6-methyladenosine methylation in preeclampsia and the potential role of N6-methyladenosine-modified circPAPPA2 in trophoblast invasion. Sci Rep. 11. 24357. (2021).
Liu H., Zheng J. & Liao A. The regulation and potential roles of m6A modifications in early embryonic development and immune tolerance at the maternal-fetal interface. Front Immunol. 13. 988130. (2022).
Wang J., Gao F., Zhao X., Cai Y. & Jin H. Integrated analysis of the transcriptome-wide m6A methylome in preeclampsia and healthy control placentas. PeerJ. 8. e9880. (2020).
Gu Y., Chu X., Morgan J.A., Lewis D.F. & Wang Y. Upregulation of METTL3 expression and m6A RNA methylation in placental trophoblasts in preeclampsia. Placenta. 103. 43-49. (2021).
Zhao J. et al. The m(6)A methyltransferase METTL3 promotes trophoblast cell invasion by regulating MYLK expression. Placenta. 129. 1-6. (2022).
Yao M.D. et al. Role of METTL3-Dependent N(6)-Methyladenosine mRNA Modification in the Promotion of Angiogenesis. Mol Ther. 28. 2191-2202. (2020).
Weng J., Couture C. & Girard S. Innate and Adaptive Immune Systems in Physiological and Pathological Pregnancy. Biology (Basel). 12. (2023).
Chen Y. et al. Recent insight into autophagy and immunity at the maternal-fetal interface. J Reprod Immunol. 155. 103781. (2023).
Pique-Regi R. et al. Single cell transcriptional signatures of the human placenta in term and preterm parturition. Elife. 8. (2019).
Sui X., Klungland A. & Gao L. RNA m6A modifications in mammalian gametogenesis and pregnancy. Reproduction. 165. R1-r8. (2023).
Mohammed S.A. et al. The BET Protein Inhibitor Apabetalone Rescues Diabetes-Induced Impairment of Angiogenic Response by Epigenetic Regulation of Thrombospondin-1. Antioxid Redox Signal. 36. 667-684. (2022).
Zhou Z. et al. The cAMP-responsive element binding protein (CREB) transcription factor regulates furin expression during human trophoblast syncytialization. Placenta. 35. 907-918. (2014).
Duan F.M. et al. THBS1 regulates trophoblast fusion through a CD36-dependent inhibition of cAMP, and its upregulation participates in preeclampsia. Genes Dis. 8. 353-363. (2021).
Lehtoranta L. et al. Fetal cardiovascular hemodynamics in type 1 diabetic pregnancies at near-term gestation. Acta Obstet Gynecol Scand. 100. 263-271. (2021).
Liu J. et al. [JAG1 promotes migration, invasion, and adhesion of triple-negative breast cancer cells by promoting angiogenesis]. Nan Fang Yi Ke Da Xue Xue Bao. 42. 1100-1108. (2022).
Hunkapiller N.M. et al. A role for Notch signaling in trophoblast endovascular invasion and in the pathogenesis of pre-eclampsia. Development. 138. 2987-2998. (2011).
Balahmar R.M. et al. Analyses of selected tumour-associated factors expression in normotensive and preeclamptic placenta. Pregnancy Hypertens. 29. 36-45. (2022).
Runyan C.L. et al. CXCR4 signaling at the ovine fetal-maternal interface regulates vascularization, CD34+ cell presence, and autophagy in the endometrium†. Biol Reprod. 101. 102-111. (2019).
Tran V. et al. The Angiopoietin-Tie2 axis contributes to placental vascular disruption and adverse birth outcomes in malaria in pregnancy. EBioMedicine. 73. 103683. (2021).
Haneda S. et al. Estrogens Regulate Placental Angiogenesis in Horses. Int J Mol Sci. 22. (2021).
Zhu X., Chen W. & Wang H. Effects of swimming before and during pregnancy on placental angiogenesis and perinatal outcome in high-fat diet-fed mice. PeerJ. 11. e14562. (2023).
Xia J.B. et al. The CXCL10/CXCR3 axis promotes cardiac microvascular endothelial cell migration via the p38/FAK pathway in a proliferation-independent manner. Exp Mol Pathol. 100. 257-265. (2016).
Dama A., Baggio C., Trevisi L., Bolego C. & Cignarella A. Regulation of human endothelial cell migration by oral contraceptive estrogen receptor ligands. Eur J Pharmacol. 945. 175591. (2023).
Shen T.L. et al. Conditional knockout of focal adhesion kinase in endothelial cells reveals its role in angiogenesis and vascular development in late embryogenesis. J Cell Biol. 169. 941-952. (2005).
Zhao X., Peng X., Sun S., Park A.Y. & Guan J.L. Role of kinase-independent and -dependent functions of FAK in endothelial cell survival and barrier function during embryonic development. J Cell Biol. 189. 955-965. (2010).
Wu A., Yu R., Li D. & Tuo Y. Expression and mechanism of micro-ribonucleic acid-29a in chorionic villi obtained at curettage with recurrent abortion. Gynecol Endocrinol. 38. 944-948. (2022).
Wang M. et al. Diallyl Trisulfide Promotes Placental Angiogenesis by Regulating Lipid Metabolism and Alleviating Inflammatory Responses in Obese Pregnant Mice. Nutrients. 14. (2022).
Mu Z. et al. Hyperin promotes proliferation, migration, and invasion of HTR-8/SVneo trophoblast cells via activation of JAK1/STAT3 pathway in recurrent spontaneous abortions. Heliyon. 9. e12958. (2023).
Hua R. et al. CCR2 mediates the adverse effects of LPS in the pregnant mouse. Biol Reprod. 102. 445-455. (2020).

No competing interests reported.

Graphabstract.png

Download PDF

Version 1

posted

You are reading this latest preprint version

Bioinformatics study and experimental validation of METTL3 regulation of immune-related genes affecting placental vascular development

Status:

Version 1

Abstract

Figures

Introduction

Results

The landscape of m6A regulators

Machine learning

GSVA and GSEA analysis and experimental validation

Subtype analysis of m6A regulators

Construction of the m6A gene signature

Correlations between m6A subgroups

WGCNA co-expression analysis

Identification and functional enrichment analysis of METTL3-related genes

Identification of m6A-targeted immune genes and PPI analysis

Immune cell infiltration analysis

Validation of the hub genes

Experimental validation of the Pregnant Rat Models of Fear Stress

Discussion

Methods

Data collection and processing

Detection and interactions of m6A RNA methylation regulators

Machine learning

GSVA and GSEA analyses

Unsupervised clustering analysis

DEGs between the two m6A clusters and functional analysis

Construction of the m6A gene signature

Weighted co-expression network construction

Correlation analysis

Identification of m6A-targeted immune DEGs

Protein-protein interaction analysis

Assessment of immune infiltration

Download of placental sequencing data from Pregnant Rat Models of Fear Stress

Animals and Pregnant Rat Models of Fear Stress

Open field test

Elevated Plus Maze test

Western blot

Quantitative real-time PCR

Transmission electron microscopy

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1