Journal Article

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

Zhao, Cheng, Zheng, Leting, Meng, Ningqin, Wu, Qiulin, Chen, Shuyuan, Zeng, Wen, Wen, Jing, Dong, Fei, Zhong, Xiaoning

2024

Overview

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

• Publisher: Frontiers Media SA,Frontiers Media S.A
• Subject: Artificial intelligence / Atrophy / Autoimmune diseases / Biomarkers / Cell interactions / Clinical trials / Connective tissue diseases / Cytokines / Data processing / Datasets / Diagnosis / ENHO / Extracellular matrix / Fibrosis / Gene expression / Gene set enrichment analysis / Immunology / Infiltration / Machine learning / macrophage / Macrophages / NOX4 / NOX4 protein / Ontology / Pathogenesis / prediction model / Prediction models / Programming languages / Regression analysis / Reverse transcription / Scleroderma / Skin / Support vector machines / Systemic sclerosis