AUTHOR=Yang Xi-Hu , Jing Yue , Wang Shuai , Ding Feng , Zhang Xiao-Xin , Chen Sheng , Zhang Lei , Hu Qin-Gang , Ni Yan-Hong 

TITLE=Integrated Non-targeted and Targeted Metabolomics Uncovers Amino Acid Markers of Oral Squamous Cell Carcinoma

JOURNAL=Frontiers in Oncology

VOLUME=Volume 10 - 2020

YEAR=2020

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2020.00426

DOI=10.3389/fonc.2020.00426

ISSN=2234-943X

ABSTRACT=Purpose
It is essential to explore potential molecular involved in oral squamous cell carcinoma (OSCC) malignant transformation and progression. The purpose of this study was to delineate the amino acids metabolic characteristics of OSCC patients and test their diagnostic value. 
Experimental Design 
Eight pairs of matched tumor and normal samples were collected for Gas chromatography-mass spectrometry (GC-MS) high throughput untargeted analysis. Another 20 cases (each case including Tumor and Normal tissues) were also enrolled for Ultra-high performance liquid chromatography-tandem mass spectrometer (UHPLC-MS) amino acids targeted quantitative analysis. T-test and ROC curve analysis were used to determine candidate markers. PCA, PLS-DA, and heatmap analysis were used to verify the ability of candidate markers to distinguish tumors from normal tissues. 
Results
A Total of 10 amino acids biomarker were selected as OSCC candidate diagnostic biomarkers by GC-MS high throughput untargeted metabolomics analyses (AUC＞0.80). We further measured the specific concentration of these candidate amino acids biomarkers in another batch of 20 cases by UHPLC-MS/MS targeted quantitative analysis. Results showed 9 out of 10 were consistent with untargeted metabolomics group, which had significant differences (t test, P＜0.05). Moreover, 3 out of 9 amino acid markers (glutamate, aspartic acid and proline) displayed high sensitivity and specificity (AUC＞0.90) by ROC curves analysis and obtained optimal sensitivity and specificity by binary logistic regression in the Glmnet package (AUC=0.942).
Conclusions
In conclusion, a panel including 3 amino acids (glutamate, aspartic acid and proline) was identified as potential diagnostic biomarkers of OSCC by a combination of non-targeted and targeted metabolomics methods.