Proteomic Analysis Reveals CTSD as Potential Diagnostic Biomarkers for Diabetes

Diabetes has emerged as a significant public health concern. The existing diabetes diagnostic markers, such as fasting glucose and glycated hemoglobin, have limitations in accuracy and sensitivity in epidemiological investigations. This study identified new serum biomarkers for diabetes through a three-step process: screening, validation, and correlation analysis. The study utilized mass spectrometry-based proteomics data from pooled serum acquired in both data-dependent acquisition (DDA) and data-independent acquisition (DIA) modes, along with DIA-mode proteomics data from 107 individuals (54 controls and 53 diabetic patients), leading to the identification of 365 differentially expressed proteins (DEPs). By excluding 20 different expression proteins that exhibit opposing trends in the three groups results, we identified 345 DEPs. ROC-AUC analysis was performed on proteins that were consistently detected across all three datasets and exhibited a consistent expression trend, as well as on DEPs demonstrating a consistent expression trend in the results from two sets of DIA-mode mass spectrometry data, which helped to evaluate the diagnostic potential of these proteins for distinguishing diabetic patients from controls. In the ROC analysis, CTSD exhibited the largest area under the curve (AUC: 0.862). Subsequent Galectin-3-binding protein (LGALS3BP) and Cathepsin D (CTSD) were validated through ELISA in the serum of 38 diabetic patients and 32 controls. And the results showed both of them had significant upregulation in the diabetic group. Notably, LGALS3BP demonstrated a strong association with established diabetes markers and risk factors in the correlation analysis in which the data from the same cohort were analyzed, particularly waist circumference. Additionally, CTSD was significantly correlated with both glycated hemoglobin (HbA1c) and body mass index (BMI). In one word，based on mass spectrometry-based proteomics data, cross-validation, AUC-ROC analysis, and ELISA validation data, Correlation analysis, CTSD exhibits a more robust association with diabetes pathogenesis, rendering it a superior candidate for serum biomarker development in diabetes. Those findings not only contribute to the repertoire of diagnostic biomarkers for diabetes but also provide a robust analytical framework for the identification of specific protein markers in large-scale serum samples.