Potential Coagulation-Related Diagnostic Model Associated with Immune Infiltration for Acute Myocardial Infarction

Academic Background

Acute Myocardial Infarction (AMI) is one of the leading causes of death worldwide. Despite significant advancements in diagnosis, treatment, and prognosis in recent years, the incidence and mortality rates of AMI remain high. Percutaneous Coronary Intervention (PCI) combined with anticoagulant therapy is currently the gold standard for the diagnosis and treatment of AMI. However, although PCI and anticoagulant therapy significantly improve myocardial perfusion, up to 50% of patients still experience significant transmural infarction, often accompanied by Microvascular Injury (MVI). Additionally, in some AMI patients with inadequate initial anticoagulation, the risk of myocardial reinfarction is significantly increased. Therefore, the occurrence and prognosis of AMI are closely related to prothrombotic mediators and coagulation factors such as prostaglandin, tissue factor, and von Willebrand factor.

Recent studies have shown that Coagulation-Related Genes (CRGs) play an important role in the prognosis of AMI. Genomic differences, particularly in coagulation and lipid metabolism pathways, have been shown to be prevalent in AMI patients. Furthermore, immune infiltration plays a key role in the hypercoagulable and inflammatory infiltration phases of AMI. However, the role of coagulation factors in AMI patients and the mechanisms regulating patient prognosis are not yet fully understood. Therefore, this study aims to reveal the potential characteristics and clinical value of CRGs in AMI patients, providing new insights for disease risk prediction and clinical treatment, and promoting the management of AMI prognosis.

Source of the Paper

This paper was co-authored by Guoqing Liu, Wang Liao, Xiangwen Lv, Lifeng Huang, Min He, and Lang Li, affiliated with the First Affiliated Hospital of Guangxi Medical University and the Second Affiliated Hospital of Guangxi Medical University. The paper was published online on October 8, 2024, in the journal Genes & Immunity, with the DOI: https://doi.org/10.1038/s41435-024-00298-z.

Research Process and Results

Data Collection and Preliminary Processing

The study first downloaded the GSE66360 dataset from the GEO database, which contains data from 49 AMI patients and 50 normal controls, serving as the training dataset for this study. Additionally, the GSE48060 dataset (31 AMI patients and 21 normal controls) was used as the external validation dataset. The study also downloaded 139 CRGs from the MSigDB team and used Perl software to annotate the dataset and convert it into gene expression data. Subsequently, the obtained gene expression data were normalized to eliminate inter-study differences.

Identification of Differentially Expressed Genes (DEGs) and Correlation Analysis

Using the “limma” package in R, the study analyzed gene expression in AMI and control samples, screening for significantly differentially expressed genes (|logFC| > 1 and p < 0.01). A heatmap of the DEGs was then constructed, and the intersection of DEGs and CRGs was visualized using a Venn diagram, ultimately identifying 10 differentially expressed CRGs. The study also used the “RCircos” package to map the chromosomal locations of these differentially expressed CRGs and assessed the correlations between these genes using Pearson correlation analysis.

Construction of Machine Learning Algorithms

The study employed two machine learning algorithms, Random Forest (RF) and Support Vector Machine (SVM), to identify signature CRGs. By comparing residual boxplots and reverse cumulative distribution plots, the RF algorithm was found to have the smallest residuals and better predictive performance. Subsequently, ROC curves were used to further validate the accuracy of the two algorithms, ultimately determining that the RF algorithm outperformed the SVM algorithm in predicting AMI.

Establishment of the Nomogram

Based on the six signature CRGs, the study constructed a nomogram to predict the incidence of AMI, and its accuracy was assessed using Decision Curve Analysis (DCA), Clinical Impact Curve Analysis, Calibration Curve Analysis, and ROC Curve Analysis.

Molecular Typing and Principal Component Analysis (PCA)

The study used the “ConsensusClusterPlus” package to perform consensus clustering analysis on the candidate CRGs, dividing AMI patients into two distinct subgroups. Principal Component Analysis (PCA) was used to calculate the score for each AMI sample, and a CRG score (PCA score) was established to differentiate between different CRG patterns.

Immune Cell Infiltration Analysis

Single Sample Gene Set Enrichment Analysis (ssGSEA) was used to assess immune cell infiltration levels in different CRG clusters, and a relevant heatmap was generated. The results showed that THBD, SERPINA1, THBS1, MMP9, PLAU, DUSP6, PLEK, and SERPINB2 were positively correlated with major innate immune cells (monocytes, neutrophils, mast cells, eosinophils, NK cells, etc.).

Functional Enrichment Analysis

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analyses were performed on the differentially expressed genes (DEGs) between CRG patterns to explore their potential biological functions. The results showed that these DEGs were significantly enriched in biological processes such as cytokine production, response to external stimuli, secretory granule membrane, tertiary granule, cytoplasmic vesicle lumen, immune receptor activity, carbohydrate binding, cytokine receptor binding, and cytokine activity.

Targeted Drug Prediction Based on Signature Genes

Drugs targeting the proteins encoded by the signature genes were searched using the DGIDB database, and gene-drug network maps were drawn. The results showed that targeted drugs for MMP9 included Celecoxib, Bevacizumab, Prinomastat, etc., while targeted drugs for THBD included Simvastatin, Cilostazol, and Warfarin, and the targeted drug for SERPINA1 was the α-1 proteinase inhibitor.

Conclusions and Significance

The diagnostic model for AMI based on six signature CRGs developed in this study demonstrates good predictive capability, providing new targets and strategies for the diagnosis and treatment of AMI. The study reveals the potential role of CRGs in AMI and suggests that these genes may influence the progression and prognosis of AMI by regulating immune cell infiltration and inflammatory responses. Additionally, the study predicts targeted drugs for these signature genes, offering new insights into personalized treatment for AMI.

Research Highlights

1. Innovative Diagnostic Model: This study is the first to construct a diagnostic model for AMI based on CRGs and validate its accuracy using machine learning algorithms.
2. In-Depth Analysis of Immune Infiltration: The study reveals the association between CRGs and immune cell infiltration, providing new perspectives on the immune mechanisms of AMI.
3. Targeted Drug Prediction: Through gene-drug network analysis, the study predicts several potential targeted drugs for AMI treatment, offering new options for clinical therapy.

Research Limitations

1. Single Source of Samples: The datasets used in this study are from a single database, and future studies with larger and more diverse sample sets are needed to improve the accuracy of the model.
2. Potential Dataset Differences: Differences between datasets may lead to biases in statistical and analytical results, requiring further validation in future studies.
3. Insufficient Clinical Validation: Due to incomplete patient information in public databases, the study was unable to further evaluate the model’s effectiveness in terms of disease progression and severity, necessitating validation through clinical cohort studies in the future.

Summary

This study constructs a diagnostic model for AMI based on CRGs, revealing the potential role of these genes in AMI and providing new targets and strategies for the diagnosis and treatment of AMI. Future research needs to further explore the downstream mechanisms of CRGs and their specific roles in AMI to advance personalized treatment and prognosis management for AMI.