Integrative Analysis of Key Genes and Signaling Pathways By Bioinformatics in Patients with Type 1 Diabetes Mellitus Combined with Acute Myocardial Infarction


 Although the pathogenesis of type 1 diabetes mellitus (T1D) and acute myocardial infarction (AMI) remains unclear. We investigated the key genes and signaling pathways common to T1D and AMI. First, we screened differentially expressed genes (DEGs) co-expressed by T1D and AMI through gene expression synthesis (GEO) database and text mining. David database was used for enrichment and functional analysis of selected genes. The interaction between proteins (PPI) was created using STRING and Cytoscape software. MCODE is used for module analysis of PPI network. A total of 74 human genes that met the criteria were found in T1D and AMI. The first 10 central genes include STAT3, ITGAM, MMP9, ERBB2, MAPK3, FOS, MYD88, MAPK1, TFRC and TNFRSF1A.The establishment of the aforementioned key genes might serve as novel biomarkers for precision diagnosis and providing medical treatment for the occurrence of AMI in T1D patients in the future.


Introduction
Type 1 diabetes is a chronic disease caused by an autoimmune attack on beta cells in the pancreas [1,2,3] . Patients with acute myocardial infarction have an acute onset, and the mortality rate is high if not treated in time [4,5] . At present, the mortality rate is greatly reduced through timely percutaneous coronary intervention (PCI), stent implantation and drug therapy [6] . Poor blood sugar control is associated with most myocardial infarction [7] . Despite signi cant advances in the treatment of acute myocardial infarction, there is still a lack of effective predictive methods for patients at high risk for type 1 diabetes. Therefore, by studying the correlation between T1D and AMI, we provide evidence for the early diagnosis and prevention of high-risk T1D patients.
The birth of high-throughput sequencing technology is a milestone in the eld of genomics research, making it possible to conduct large-scale whole genome resequencing. The GEO database is a gene expression database created and maintained by the NATIONAL Center for Biotechnology Information (NCBI), which contains storage chips, second-generation sequencing, and other high-throughput sequencing data [8] . Although the pathogenesis of IMA is currently related to T1D, its molecular mechanism remains unclear. Therefore, we retrieved AMI gene expression chips GSE60993 and GSE34198 from geo database, and screened differentially expressed genes (DEG) for analysis by R software (version 3.6.1) [9] . T1D related genes were obtained through text mining. DEG and text mining gene sets were analyzed using Venny, and interactions between proteins (PPI) were created using STRING and Cytoscape software. In this study, we explored the key genes and signaling pathways of AMI associated with T1D, providing prediction and therapeutic targets for the possibility of ACS in T1D patients in the future.

Methods
Data abstraction. We abstracted the gene expression chip data GSE60993 and GSE34198 from the NCBI Gene Expression Comprehensive (GEO) web resource (https://www.ncbi.nlm.nih.gov/geo/)19,23 [10] . The GSE60993 cohort contains seven normal control and seven ST-elevation myocardial infarction samples, while the GSE34198 dataset includes seven normal control and seven AMI samples.
Identi cation of DEGs. The core R package was used to process the downloaded matrix les. After normalization, the differences between AMI and the control group were determined by truncation criteria |log2 fold change (FC)|≥2, adjusted P<0.05), and we selected the remarkable DEGs for downstream analyses [11] .
Text mining. We based on web services GenCLIP3 platform (http://ci.smu.edu.cn/genclip3/analysis.php/) for text mining. After operation, GenCLIP3 retrieves and searches the names of all genes associated with type 1 diabetes found in the existing literature. The gene set obtained by text mining and the differential gene set obtained before were screened for further analysis [12] .
Gene ontology analysis of DEGs and KEGG pathway analysis. Use David v. 6.8 (https://david.ncifcrf.gov/) to obtain the DEGs were analyzed. GO annotation and KEGG pathway enrichment were performed. GO terms include biological processes (BP), Cellular composition (CC) and molecular function (MF). The function description of gene products obtained from various databases is consistent [13] . The Kyoto Encyclopedia of Genes and Genomes (KEGG) is a bioinformatics resource for understanding biological function from a genomic perspective. It is a multi-species, comprehensive resource composed of genomic, chemical and network information, cross-referencing numerous external databases, and containing a complete set of building blocks (genes and molecules) and wiring diagrams (biological pathways) to represent cellular function [14] . The above genes were expressed as P 0.05 was the signi cant threshold for analysis.
PPI network and module analysis. We used the STRING online search tool for protein-protein interaction (PPI) analysis of screened DEGs [15] . When the PPIs comprehensive score was > 0.6, the gene was considered to be selected. The PPI network was visualized by Cytoscape and the key genes were classi ed by MCODE [16] . In order to P 0.05 was used as the standard to analyze the functional enrichment of each DEGs.

DEGs screening
Firstly, 3842 DEGs were screened from AMI samples and normal controls in GSE6099 dataset by limma package built-in R software. 1962 upregulated genes and 1880 downregulated genes were selected.
Meanwhile, 1263 differentially expressed genes, including 569 up-regulated genes and 594 downregulated genes, were obtained by analyzing ACS samples from GSE34198 dataset and the normal control group. Then, the two data sets and the overall distribution of the top 100 differential genes were represented by volcano map and heatmap respectively ( Figure 1A-D). Sample using |log2 fold change (FC)|≥2 criteria and adjusted P<0.05. Through text mining, 2115 human genes associated with T1D were screened out. DEGs in microcolumn data were hybridized to obtain the intersection of three selected genes, resulting in 74 genes ( Figure. 2A).
Function and signal pathway enrichment analysis DEGs obtained above were introduced into DAVID for GO and KEGG enrichment analysis to study the biological functions of DEGs in AMI with T1D. In the GO analysis results, 10 biological process items (BP), 15 cell component items (CC) and 10 molecular function items (MF) were found. P<0.05 signi ed threshold signi cance. (Figure 3A) The BP item "in ammatory response" is mainly rich in 18 genes, the CC item "plasma membrane" is mainly rich in 32 genes, and the MF item "protein binding" is mainly rich in 54 genes. KEGG enrichment evaluation showed that integrated DEGs were signi cantly enriched in Prolactin signaling Pathway, Hepatitis B, and Chemokine signaling Pathway ( Figure 3B). Module screening from the PPI network.
Based on 74 DEGs, the PPI network was developed using Cytoscape public platform and STRING resources for module analysis and visualization. Therefore, we developed a PPI network with 158 crosstalk based on 60 integrated DEGs associated with AMI ( Figure 4A). Based on the degree, The main hub genes extracted from AMI group include STAT3 (signal and Activator of transcription 3), ITGAM (Integrin Subunit alpha M) and MMP9 (matrix Metallopeptidase 9), ERBB2 (ErB-B2 receptor Tyrosine kinase 2) and MAPK3 (Mitogen-activated protein kinase 3). On the other hand, in the AMI group, we use the MCODE algorithm to identify highly interconnected subnets, which are usually protein complexes, and by calculation, we nd 2 highly clustered modules. (Figure 4B ,C)

Discussion
High blood sugar in type 1 diabetes can lead to microvascular complications [17] . Hyperglycemia in blood accelerates oxidative stress in blood vessels [18] , of which cardiovascular disease is an important part [19] . Previous studies suggest that intensive hypoglycemia can reduce complications and improve prognosis in patients with type I diabetes [20] . A recent study of type 1 diabetes followed for 30 years found that intensive glucose reduction was associated with a 30% reduction in the incidence of cardiovascular disease compared with a control group [21] . Timely screening of high-risk type I diabetes patients to prevent the occurrence of myocardial infarction is very important to the health of patients. When we performed PPI analysis on DEGs, we also found many strongly associated genes, such as STAT3, ITGAM, MMP9, ERBB2, MAPK3, FOS13, MYD88, MAPK1, TFRC and TNFRSF1A.
Transcriptional regulator STAT3 plays a key role in in ammation and immune regulation [22] . Moreover, it plays an important role in the eld of tumor growth and immunity [23] , which can be observed through several important KEGG signaling pathways. Xilan Yang et al. 's study suggested that STAT3 pathway also plays a role in atherosclerosis [24] , which may play a role in myocardial infarction. When blood perfusion to the heart is reduced, transcriptional activator 3 (STAT3) may be activated to improve blood supply by promoting angiogenesis [25] , indicating its important role in cardiac blood supply. Meijing Wang et al. found that STAT3-de cient mice had signi cant effects on myocardial function and in ammation [26] . Studies have suggested that TICAM2 plays a key role in promoting neutrophil depletion [27] , which may be related to the in ammatory stimulation of blood vessels by hyperglycemia.
Matrix metalloprotease 9 (MMP9) can be activated after myocardial infarction, which intensi es cardiac ischemia and eventually leads to chronic heart failure (CHF) [28] . Timely interruption of MMP9 appears to reduce infarct size after acute myocardial infarction [29] . Cardiac protection by blocking MMP9 is especially true in diabetics [30] Yadav SK et al. 's study suggested that MMP9 might accelerate the apoptosis of hCSCs cells, reduce the vitality of hCSCs, and accelerate cell death through hyperglycemia [31] . In the future, it may be possible to improve patient outcomes by inhibiting this pathway.
ErbB2 (Her2/ NEU) has been extensively studied in the development of breast cancer. ErbB2 can accelerate mitochondrial apoptosis through the molecular mechanism of Bcl-XL and -XS. The occurrence of myocardial infarction in patients with type I diabetes may be related to mitochondrial dysfunction, so erbB2 is also a highly correlated gene [32] .
The MAPK3 pathway is involved in the repair of myocardial ischemia [33,34,35] . This may be related to the body's self-repair after acute myocardial infarction, which may provide a target for future treatment.
Inhibition of the MyD88 cardiac in ammatory pathway reduces obesity-induced cardiac damage [36] , But others have suggested that MyD88 promotes heart repair after myocardial infarction [37] . In conclusion, in our study, MyD88 pathway plays an important role in myocardial infarction in patients with type 1 diabetes.

Conclusions
In summary, we analyzed two datasets, GSE60993 and GSE34198, and conducted gene expression mining to establish Pathways in cancer and Hepatitis B molecular modulation networks of core functions and enriched signal cells. It provides a prediction and clinical treatment direction for ACS in T1M patients in the future. However, further mechanism needs to be veri ed by more experimental data.

Declarations ETHICAL APPROVAL
This article does not contain any studies with human participants or animals performed by any of the authors.  Venn diagram of DEGs from microarray data and genes list from text mining. DEGs, differentially expressed genes.