| Part?A study of metabolic subtypes of pancreatic cancers based on NMR metabolomics of tissueObjective:To explore the establishment of metabolic subtypes of pancreatic cancers based on nuclear magnetic resonance(NMR)metabolomics of tissue,reveal the heterogeneity of pancreatic cancers at metabolomics level and discover differences of clinical progression and prognosis between metabolic subtypes.Methods:One hundred thirteen pancreatic cancer(PCa group)and 56 normal pancreatic tissue(Control group)were collected and detected with ~1H NMR spectroscopy to obtain metabolomic data.Based on metabolic data,hierarchical cluster analysis(HCA)classified PCa groups were into different metabolic subgroups.By using multivariate and univariate statistical analyses,the tissue metabolic alterations and relevant metabolic pathways of PCa subgroups,and the metabolic difference between PCa subtypes were analyzed.Then,metabolic subtypes were determined by on metabolic feature of PCa subgroups.Then,the clinical characteristics of patients in different subtypes were collected and compared.The receiver operating characteristic(ROC)were conducted to search for metabolic biomarkers for distinguishing pancreatic cancer and specific subtype.Results:Based on HCA analysis,PCa could be classified into three subtypes:PCa-1PCa-2A and PCa-2B.Compared with Controls,all PCa subtypes have obvious metabolic alterations.The concentration of sugar group,carboxylic acid group,some of amino acids and catabolites of amino acids were elevated in all PCa subtypes tissue,and tricarboxylic acid cycle(TCA)substrates,pyrimidine and purine groups,nucleoside group,choline group and nicotinamide group were decreased.Only a few metabolites were in different variation trends among PCa subgroups.These findings indicated that the overall metabolic reprogramming among different subgroups were similar.However,the alterative degree of metabolites and the mainly relevant pathway of metabolic alteration in different subgroups were different.The high levels of amino acids and low levels of lactate,taurine,glutathione and adenosine group were the main feature of PCa-1 subgroup,which could be defined as the metabolic subtype of high amino acids enrichment.The relative higher level of pyrimidine and purine groups,nicotinamide group and choline group were the main feature of PCa-2B,which could be defined as the high nucleoside-nicotinamide-choline metabolic subtype.Meanwhile,the PCa-2B could be defined as the intermediate type.Regarding clinical progression and prognosis,the patients with metabolic subtype of high amino acids enrichment were found to have an earlier postoperative recurrence and a worse long-term outcome.These findings indicated a worse biological characteristics of pancreatic cancer in patients with metabolic subtype of high amino acids enrichment.Nine metabolites,such as myo-inositol and malate and six metabolites,such as lysine and tyrosine,were identified as the metabolic biomarkers for diagnosing pancreatic cancer and distinguishing patients with metabolic subtype of high amino acids enrichment from other patients.Conclusion:Different subtypes of pancreatic cancer have a similar trends of overall metabolic alteration in tissue.But the change degrees of different metabolite concentration and major relevant metabolic pathways were heterogeneous between subtypes.The significant differences in patient prognosis between different subtypes intensively indicated a strong relationship between metabolic heterogeneity and tumor biological behaviors,which benefit getting a deep insight into the tumor biology and metabolism of pancreatic cancer.Combined with the metabolic biomarkers identified in this study,metabolic subtypes would help to establishing methods of pancreatic cancer screening and the assessment of postoperative patients with high risk of recurrence.Part? A study of metabolic subtypes of pancreatic cancers based on NMR metabolomics of serum.Objective:To explore the establishment of metabolic subtypes of pancreatic cancers based on nuclear magnetic resonance(NMR)metabolomics of serum,assess the heterogeneity of pancreatic cancers at metabolomics level and discover differences of clinical progression and prognosis between metabolic subtypes.Methods:Serum samples from 39 health volunteers,20 pancreatic cancer patients with jaundice(J-PCa)and 37 pancreatic cancer patients without jaundice(NJ-PCa)were collected,and then detected by using ~1H NMR spectroscopy.The metabolomic difference between NJ-PCa,J-PCa and control groups were compared by using multivariate and univariate statistical analyses.Then,NJ-PCa and J-PCa groups were classified into subgroups by using HCA,respectively.The metabolic alterations,corresponding metabolic pathways and intragroup metabolic difference between subgroups were analyzed.ROC was employed to search for metabolic biomarkers of patients with pancreatic cancer.Then,the clinical characteristics of patients in different subtypes were collected and compared.Results:Pattern recognition models indicated obvious metabolomic differences among control,NJ-PCa and J-PCa groups.Only a few metabolites were in same variation trends among NJ-PCa and J-PCa,indicating that jaundice could trigger a huge influence on serum metabolome which further cover up the specific metabolic alterations caused by pancreatic cancer.Based on HCA,J-PCa and NJ-PCa can be classified into two(J-PCa 1 and 2)and three subtypes(NJ-PCa 1,2 and 3),respectively.Compared with control group,several amino acids,choline groups,sugar group,alcohol group were decreased in J-PCa 1 serum while lactate,lipid and lipoprotein were increased.Only a few metabolites were significantly altered in J-PCa 2 and NJ-PCa 1.Meanwhile,a common increase in amino acids and glycerol and a common decrease in lipid and lipoprotein were detected in NJ-PCa 2 and 3.The difference in lactate and glucose concentration were the main feature to distinguish NJ-PCa 2 and 3.Based on degree of metabolic alteration,J-PCa 2 and NJ-PCa 1 could be defined as the subtypes with low metabolic disorder while the J-PCa 1,NJ-PCa 2 and 3 were defined as the metabolic subtypes with high metabolic disorder.Based on lactate concentration,NJ-PCa 2 and3 could be further defined as metabolic subtypes with low and high lactate.Phenylalanine could act as a metabolic biomarker for distinguish pancreatic caner patients from the health.Regarding clinical progression and prognosis,it was no significant difference between J-PCa 1 and 2 subgroups.Meanwhile,in NJ-PCa groups,patients with metabolic subtypes of high metabolic disorder seem to have a higher proportion of patients with distant metastasis.NJ-PCa 3 have a shorter overall survival compared with other two subtypes while NJ-PCa 2 have a longer overall survival.Conclusion:Jaundice could cause a significant alteration in serum metabolism of pancreatic cancer patients.The inconsistency in the degree of metabolic disorder in serum was high among pancreatic cancer patients and serum metabolomics were highly heterogeneous.The subtypes were different in distant metastasis and prognosis of patients,implying that the serum metabolic alterations caused by pancreatic cancer were associated with tumor biological behaviors and tumor-host interaction,which can provide important clues for the researches targeting corresponding mechanism.These findings provided a foundation to further establish serum metabolomics-based strategy to diagnose pancreatic cancer and assess progression and prognosis of pancreatic cancer.Part ? The study of serum metabolic differences between the pancreatic animal models with different differentiation.Objective:To detect the serum metabolic difference between pancreatic cancer animal models with different differentiation and preliminarily assess the feasibility of preoperative differentiation assessment based on serum metabolomics.Methods:Balb/c nude mice were implanted with three pancreatic cancer cell strains(Panc-1,Bxpc-3 and SW1990)to establish pancreatic cancer models with different tumor differentiations.Serum of these pancreatic cancer models were collected and detected by ~1H NMR.The metabolic difference between different groups were analyzed by using multivariate statistical analyses and the relevant metabolic pathways were analyzed by using MBRole online service.Results:Twelve,thirteen and eleven of Panc-1,Bxpc-3 and SW1990 models were successfully established and pathological examinations confirmed that the tumors in Panc-1,Bxpc-3 and SW1990 models accorded with pathological features of low,low to moderate,and moderate to high differentiation,respectively.Through pair comparison between serum metabolome of three group,pattern recognition models with a high reliability were established,indicating that the metabolic influence from pancreatic cancers with different differentiation were significantly different.In SW1990 group,the concentration of glucose and lipid were relatively lower while the citrate,lactate and several amino acids were relatively higher,implying that the concentration difference of glycolytic and glutaminolytic substrates may be associated with tumor differentiation.Meanwhile,the concentrations of choline group in serum of Panc-1 group were relatively higher,implying the relative higher choline groups in serum may be the metabolic hallmark of pancreatic cancer with low differentiation.Conclusion:In this study,it was found that the serum metabolic alterations caused by pancreatic cancers with different differentiation were obviously different,which could establish pattern recognition models with a high reliability.These findings preliminarily validated the feasibility of metabolomics-based strategy to predict the tumor differentiations. |
PDF Full Text Request