Back
Previous studies suggested that treatment with low-dose arsenic trioxide (ATO) might reduce the risk of flares in patients with active systemic lupus erythematosus (SLE), but the underlying pharmacological mechanisms have not been investigated.
The potential differentially expressed genes (DEGs) targets were identified using machine learning and network pharmacology analysis. The expression of characteristic genes and its association with immune cells was further examined and validated.
Of twelve intersection DEGs, five immunoregulatory genes in SLE were identified by three machine learning models (RF, SVM and XGB), in which MMP9 demonstrated the highest ROC AUC (092) in predicting disease development (ROC AUC 0.942). KEGG analysis indicated strong associations with IL-17 signalling pathway (p=1.67E-18). MMP9 also showed positive correlations with macrophages and neutrophils in ssGSEA analysis (correlation coefficient 0.88 and 0.66 respectively). Our in vitro validation concurred with the bioinformatics results, which showed that ATO treatment in PBMC isolated from LN patients (n=5) could downregulate MMP9 and IL17A expression.
ATO could attenuate MMP/IL-17A pathways in PBMCs in LN patients and hence may serve as a novel therapeutic option.