Highly Sensitive Quantification Method for Amine Submetabolome Based on AQC-Labeled-LC-Tandem-MS and Multiple Statistical Data Mining: A Potential Cancer Screening Approach

The relationship between amine submetabolome and cancer has been increasingly investigated. However, no study was performed to evaluate the current methods of amine submetabolomics comprehensively, or to use such quantification results to provide an applicable approach for cancer screening. In this study, a highly sensitive and practical workflow for quantifying amine submetabolome, which was based on 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC)-labeled-HPLC-MS/MS analysis combined with multiple statistical data processing approach, was established and optimized. Comparison and optimization of two analytical approaches, HILIC separation and precolumn derivatization, and three types of surrogate matrices of plasma were performed systematically. The detection sensitivities of AQC-labeled amines were increased by 50-1000-fold compared with the underivatization-HILIC method. Surrogate matrix was also used to verify the method after a large dilution factor was employed. In data analysis, the specific amino-index for each cancer sample was identified and validated by univariate receiver operating characteristic (ROC) curve analysis, partial least-squares discrimination analysis (PLS-DA), and multivariate ROC curve analysis. These amino indexes were innovatively quantified by multiplying the raised markers and dividing the reduced markers. As a result, the numerical intervals of amino indexes for healthy volunteers and cancer patients were provided, and their clinical value was also improved. Finally, the integrated workflow successfully differentiated the value of the amino index for plasma of lung, breast, colorectal, and gastric cancer samples from controls and among different types of cancer. Furthermore, it was also used to evaluate therapeutic effects. Taken together, the developed methodology, which was characterized by high sensitivity, high throughput, and high practicality, is suitable for amine submetabolomics in studying cancer biomarkers and could also be applied in many other clinical and epidemiological research.