Introduction & Objective: Cathepsin L (CTSL) has been demonstrated to play a significant role in the pathogenesis and progression of Diabetic Nephropathy (DN). However, there are currently no CTSL inhibitors available for clinical use. This study utilizes artificial intelligence models to discovering novel CTSL inhibitors and explores their application in DN.
Methods: We developed a neural network model to predict the inhibitory potential of molecules against CTSL activity. This model was then applied to a library of biologically active compounds and conducted activity assays on the top-ranked molecules. We then co-treated HK2 cells with selected CTSL inhibitors and free fatty acids to observe changes in epithelial-mesenchymal transition and inflammation levels between the medicated and normal groups.
Results: We identified five small molecules significantly reduced the activity of CTSL in a live-cell system at the nanomolar level. Notably,10μM of S7391 significantly reduced CTSL expression levels in HK2 cells, while restoring the levels of epithelial-mesenchymal transition-related genes and inflammatory factors to those of the control group.
Conclusion: We successfully developed a deep learning model capable of predicting whether small molecule compounds can inhibit the enzymatic activity of CTSL and identified five small molecules that can inhibit CTSL's enzymatic activity at the nanomolar level. Among these, S7391 was able to reverse renal tubular damage induced by free fatty acids, indicating its potential as a therapeutic drug for diabetic kidney disease through the inhibition of CTSL activity.
Q. Li: None.
National Natural Science Foundation of China (81930019)
