Development and validation of a new diagnostic prediction model of ENHO and NOX4 for early diagnosis of systemic sclerosis

• DOI: 10.3389/fimmu.2024.1273559
• Publication Date: 2024-01-29T10:35:12Z
• Authors: Leting Zheng; Qiulin Wu; Shuyuan Chen; Jing Wen; Fei Dong; Ningqin Meng; Wen Zeng; Cheng Zhao; Xiaoning Zhong
• Keywords: prediction model; NOX4; 03 medical and health sciences; machine learning; 0302 clinical medicine; systemic sclerosis; Immunology; macrophage; ENHO; Immunologic diseases. Allergy; RC581-607

Objective Systemic sclerosis (SSc) is a chronic autoimmune disease characterized by fibrosis. The challenge of early diagnosis, along with the lack effective treatments for fibrosis, contribute to poor therapeutic outcomes and high mortality SSc. Therefore, there an urgent need identify suitable biomarkers diagnosis Methods Three skin gene expression datasets SSc patients healthy controls were downloaded from Gene Expression Omnibus (GEO) database (GSE130955, GSE58095, GSE181549). GSE130955 (48 diffuse cutaneous 33 controls) utilized screen differentially expressed genes (DEGs) between normal samples. Least absolute shrinkage selection operator (LASSO) regression support vector machine recursive feature elimination (SVM-RFE) performed diagnostic construct prediction model. results further validated in GSE58095 (61 36 GSE181549 (113 44 datasets. Receiver operating characteristic (ROC) curves applied assessing level ability. Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was used verify tissues out cohort (10 5 controls). Immune infiltration analysis using CIBERSORT algorithm. Results A total 200 DEGs identified Functional enrichment revealed that these may be involved pathogenesis SSc, such as extracellular matrix remodeling, cell-cell interactions, metabolism. Subsequently, two critical (ENHO NOX4) LASSO SVM-RFE. ENHO found down-regulated while NOX4 up-regulated their levels above three our cohort. Notably, differential expressions more pronounced than those limited Next, we developed novel model NOX4, which demonstrated strong predictive power cohorts own Furthermore, immune dysregulated various cell subtypes within specimens, negative correlation observed Macrophages M1 M2, positive M2. Conclusion This study can serve detection play potential role disease.