LC-MS Based Metabolomics Serum Research In Pancreatic Carcinoma

Pancreatic carcinoma(PC) is known as one of the most highly malignancies in digestive system and has a high mortality rate due to its difficulty of early diagnosis. Therefore, it becomes critical to screen effective early diagnostic biomakers and treatment targets for PC.As the end products of gene expression, metabolites in vivo objectively reflect the real phenotype and their abnormal alterations are closely associated with the initiation and progression of many diseases including cancer. In this study, we compare the level of serum metabolites between PC patients, healthy individuals and postoperative samples, in order to provide reliable data base for PC diagnosis model.200 sera of PC patients, 156 sera of healthy control and 116 sera of post-operative PC patients were employed to characterize the metabolic alterations with a LC-MS based metabolomics approach. By using the principal component analysis(PCA), the obvious differences were obtained in serum metabolic profiles between PC and healthy control groups. To further explore the differential metabolites responsible for the metabolic grouping, multivariate random forest, MS/MS identification and independent validation methods were used to obtain the differential metabolites, finally 6 metabolites identified as independent diagnostic factors for PC discrimination were selected, incuding 8,15-Di HETE,(R)-3-hydroxy-hexadecanoic acid, 9-octadecenoic acid, lipoyllysine, indoxylsulfuric acid and S-prenyl-Lcysteine. The combination of discriminative metabolites together with CA19-9 in the prediction of PC had an accuracy of 0.994.Because of the complexity of the clinical samples, we further explored metabolic profiles based on different influential factors in order to revealing the regulation related. The results indicated that factors of age, gender and differentiation were independent of PC metabolic profiles alterations while factors of serum color, TNM staging and pre-postoperation showed obvious differences. PLS-DA score plot were conducted to further screen metabolites of value. As a results, bilirubin glucuronide, spongipregnoloside C and 26-O-b-D-glucosyl ester were found responsible for the color grouping and indicated jaundice occurrence. 3 metabolites could differentiated TNM-I and TNM- II, III, IV groups, which were small peptides and glycine derivatives. The results illustrate d abnormal protein metabolism during PC progression. At last, in pre-pos metabolomic analysis, two molecules LysoPC(18:0) and LyPC(15:0) exhibited a call back and one kind of dipeptide showed the downtrend after operation.In the end, network and pathway analysis were conducted to explore the latent relationship among differential metabolites. Metabolites mapped in pathways of linoleic acid metabolism revealeda significant downtrend. In network analysis, metabolites of diverse categories reflected the correlation between metabolite categories and fatty acids, glycerophospholipids and carnitine were found gathering together in the biggest subnet. Such crossing linkages might reflect their combined action for fulfilling the aberrant energy requirement during tumor proliferation.