Network-Based Analysis Reveals Novel Biomarkers in Peripheral Blood of Patients With Preeclampsia

Jing Lin, Meng Yu, Mengfan Song, Wei GuFrontiers in Molecular Biosciences 2022 · DOI

Abstract

WGCNA is a potent systems biology approach that explains the connection of gene expression based on a microarray database, which facilitates the discovery of disease therapy targets or potential biomarkers. Preeclampsia is a kind of pregnancy-induced hypertension caused by complex factors. The disease's pathophysiology, however, remains unknown. The focus of this research is to utilize WGCNA to identify susceptible modules and genes in the peripheral blood of preeclampsia patients. Obtain the whole gene expression data of GSE48424 preeclampsia patients and normal pregnant women from NCBI's GEO database. WGCNA is used to construct a gene co-expression network by calculating correlation coefficients between modules and phenotypic traits, screening important modules, and filtering central genes. To identify hub genes, we performed functional enrichment analysis, pathway analysis, and protein-protein interaction (PPI) network construction on key genes in critical modules. Then, the genetic data file GSE149437 and clinical peripheral blood samples were used as a validation cohort to determine the diagnostic value of these key genes. Nine gene co-expression modules were constructed through WGCNA analysis. Among them, the blue module is significantly related to preeclampsia and is related to its clinical severity. Thirty genes have been discovered by using the intersection of the genes in the blue module and the DEGs genes as the hub genes. It was found that HDC, MS4A2, and SLC18A2 scored higher in the PPI network and were identified as hub genes. These three genes were also differentially expressed in peripheral blood validation samples. Based on the above three genes, we established the prediction model of peripheral blood markers of preeclampsia and drew the nomogram and calibration curve. The ROC curves were used in the training cohort GSE48424 and the validation cohort GSE149437 to verify the predictive value of the above model. Finally, it was confirmed in the collected clinical peripheral blood samples that MS4A2 was differentially expressed in the peripheral blood of early-onset and late-onset preeclampsia, which is of great significance. This study provides a new biomarker and prediction model for preeclampsia.

Introduction: The Challenge of Preeclampsia

Preeclampsia (PE) is a leading cause of maternal and fetal mortality worldwide, characterized by sudden-onset hypertension and organ damage during pregnancy. Despite its prevalence, the underlying pathophysiology remains a "black box." The clinical challenge is two-fold: early detection before symptoms appear and the development of non-invasive testing. While placental studies provide clues, reliable biomarkers in peripheral blood are needed to make screening practical and repeatable.

Methodology: Systems Biology and Network Analysis

To move beyond simple lists of individual genes, the researchers employed a multi-step bioinformatics pipeline centered on Weighted Gene Co-expression Network Analysis (WGCNA). Unlike standard analysis, WGCNA groups genes into "modules" based on similar expression patterns, allowing researchers to find clusters that correlate specifically with clinical traits like disease severity.

FIGURE 1: Flowchart of data preparation, analysis, and validation steps used in the study.

The study utilized three distinct cohorts:

  • Training Cohort (GSE48424): Microarray data from 36 subjects used to build the initial network.
  • Validation Cohort (GSE149437): 442 whole blood samples used to verify the diagnostic value of the hub genes.
  • Clinical Test Cohort: 30 peripheral blood samples collected for experimental validation via quantitative RT-PCR.
Figure 2A: Initial sample clustering of the training cohort to detect outliers before WGCNA analysis.

Key Results: The Three-Gene Signature

By intersecting the most relevant WGCNA module (the "blue module") with traditional Differentially Expressed Genes (DEGs) and utilizing Protein-Protein Interaction (PPI) networks, the team identified a core trio of biomarkers: HDC, MS4A2, and SLC18A2. All three showed significantly reduced expression in PE patients.

  • High Predictive Power: The combined three-gene model achieved an Area Under the Curve (AUC) of 0.849 in the training set and 0.867 in the validation set, indicating strong diagnostic potential.
  • Clinical Stratification: MS4A2 expression was significantly lower in early-onset PE compared to late-onset PE. Since early-onset cases are typically more severe, this gene could serve as a critical marker for patient stratification.

Discussion: The Mast Cell Connection

A striking insight from the study is that the top hub genes—specifically MS4A2 and HDC—are heavily involved in mast cell function. MS4A2 encodes a subunit of the high-affinity IgE receptor. This suggests that preeclampsia may be driven by an immune-inflammatory response involving the reduction or dysfunction of mast cells in the maternal-fetal interface.

Limitations and Future Work

While promising, the authors acknowledge that the clinical validation cohort was relatively small (n=30). Larger, multi-center studies are needed to confirm the model's robustness. Furthermore, future research should involve longitudinal tracking of these biomarkers throughout pregnancy to determine exactly when the diagnostic signal first appears and to investigate the causal role of mast cell reduction in placental remodeling.

Summary updated

Abstract

WGCNA is a potent systems biology approach that explains the connection of gene expression based on a microarray database, which facilitates the discovery of disease therapy targets or potential biomarkers. Preeclampsia is a kind of pregnancy-induced hypertension caused by complex factors. The disease's pathophysiology, however, remains unknown. The focus of this research is to utilize WGCNA to identify susceptible modules and genes in the peripheral blood of preeclampsia patients. Obtain the whole gene expression data of GSE48424 preeclampsia patients and normal pregnant women from NCBI's GEO database. WGCNA is used to construct a gene co-expression network by calculating correlation coefficients between modules and phenotypic traits, screening important modules, and filtering central genes. To identify hub genes, we performed functional enrichment analysis, pathway analysis, and protein-protein interaction (PPI) network construction on key genes in critical modules. Then, the genetic data file GSE149437 and clinical peripheral blood samples were used as a validation cohort to determine the diagnostic value of these key genes. Nine gene co-expression modules were constructed through WGCNA analysis. Among them, the blue module is significantly related to preeclampsia and is related to its clinical severity. Thirty genes have been discovered by using the intersection of the genes in the blue module and the DEGs genes as the hub genes. It was found that HDC, MS4A2, and SLC18A2 scored higher in the PPI network and were identified as hub genes. These three genes were also differentially expressed in peripheral blood validation samples. Based on the above three genes, we established the prediction model of peripheral blood markers of preeclampsia and drew the nomogram and calibration curve. The ROC curves were used in the training cohort GSE48424 and the validation cohort GSE149437 to verify the predictive value of the above model. Finally, it was confirmed in the collected clinical peripheral blood samples that MS4A2 was differentially expressed in the peripheral blood of early-onset and late-onset preeclampsia, which is of great significance. This study provides a new biomarker and prediction model for preeclampsia.
BibTeX
@inproceedings{lin2022network,
  author = {Jing Lin and Meng Yu and Mengfan Song and Wei Gu},
  title = {{Network-Based Analysis Reveals Novel Biomarkers in Peripheral Blood of Patients With Preeclampsia}},
  booktitle = {Frontiers in Molecular Biosciences},
  year = {2022},
  doi = {10.3389/fmolb.2022.757203},
}
Figure