Identification of novel diagnostic protein biomarkers in human plasma for breast cancer by targeted protein assay

Introduction: Breast cancer (BC) is the most common cancer among women worldwide and ranks as the second leading cause of cancer-related deaths. Fortunately, the prognosis of BC is good when detected at an early stage, however, sensitive diagnostic tools for early detection of BC are vital for improving survival rates. This study aims to identify protein biomarkers with diagnostic and prognostic value in human plasma to improve early BC diagnosis.

Methods: An absolute quantification of 131 proteins was performed on 279 well-defined Icelandic biobank-based plasma samples from 135 BC patients and 144 healthy controls by UPLC-MRM-MS assay. Among the BC patients, 45 were BRCA2 mutation carriers. Absolute quantification of the proteins were conducted using a PeptiQuantTM protein human kit containing synthetic light peptide and matching heavy peptide as internal standard for each protein. Prior to analysis the plasma samples were proteolytically cleaved with trypsin, internal standards were added to the peptide mixture and concentrated by solid-phase extraction using liquid handling robot. Data analyses were conducted using Skyline Quantitative Analysis software for pre-processing of the data. SIMCA Pro-17, R studio, and Python were used for statistical analysis, multivariate data analysis, and machine learning. Predictive models including random forest, support vector machine, and logistic regression were developed and validated through cross-validation using the SciKit-Learn library.

Results: The targeted proteomic assay was successfully implemented for absolute quantification of 131 proteins in human plasma samples with precision and accuracy for calibration standards and quality controls within 20% relative standard deviation. Out of the 131 proteins, 99 proteins were quantifiable in the Icelandic bio-bank plasma samples, surpassing the lower limit of quantification. The samples were analyzed in eight batches, each containing 25 cases and 25 controls, showing minimal batch-to-batch variability. The results reveal that protein signatures of the patients with a BRCA2 germline mutation are different from patients with sporadic BC. This difference was found to be significant for few proteins in comparison to control groups as well as between the patient groups. Notably, predictive models could distinguish between patients with sporadic BC from BC patients with BRCA2 mutation with AUC > 0.8, offering valuable insights for targeted screening for BC. Furthermore, seven proteins showed the potential to discriminate BC patients with the luminal B subtype from controls.

Conclusions: Targeted proteomics using UPLC-MRM-MS shows potential for identifying and quantifying protein biomarkers in human plasma for diagnosis of BC in BRCA2 mutation carriers and for the luminal B subtype.