A Metabonomics Study Of Primary Biliary Cirrhosis And Autoimmune Hepatitis

Primary biliary cirrhosis (PBC) and autoimmune hepatitis(AIH) are two common autoimmune liver diseases in clinic. The characteristic serologic hallmark of PBC is antimitochondrial antibody (AMA). However, AMA is negative in 5% to 10% PBC patients, which often lead to missed diagnosis. At the same time, some AMA-positive people have normal liver function and no obvious change in liver biopsy. Follow-up study has found many of them will finally progress to PBC. The diagnosis of AIH is based on both (a) clinical, serologic and immunologic features and (b) the exclusion of drug-induced hepatitis, alcoholic liver disease, chronic hepatitis C, hereditary liver diseases. Since the differential diagnosis of AIH is complex and difficult, many patients miss the best treatment window. Therefore, utilizing new technologies to look for new biomarkers of PBC and AIH has a wide clinical application prospect.Metabonomics is defined as "the quantitative measurement of the dynamic, multiparametric, metabolic response of living systems to pathophysiologic stimuli or genetic modification" by Jeremy Nicholson in 1999. Metabonomics study focus on the small molecules in a biological sample, including the metabolites and intermediate metabolites, which has a molecular weight less than 1000Da. What distinguishes metabonomics from early studies in the field of bioinformatics is that metabonomics focuses on the complete metabolite profile of samples instead of one or a few metabolites and their associated pathways.Our study has built a metabolomics research platform applying ultrafast liquid chromatography coupled with tandem mass spectrometry (UFLC-MS/MS). Multivariate analysis was utilized to compare the plasma metabolic profiles of PBC patients and control, AIH patients and control, respectively. Finally we find several novel biomarkers of PBC and AIH. 18 PBC patients,13 AIH patients and 14 controls were enrolled in this study. The fasting blood sample of the subjects was taken in the morning. Liver function test, ICG retention rate at 15 min and fibro-scan result were obtained from every patient. Pretreatment of plasma sample including protein sedimentation was conducted before chromatography. The raw file was manipulated through peaking finding, filtering and peak alignment to obtain peak information. The result was exported to a two-dimensional table, which was further processed according to "80% rule" and total peak area was normalized to a fixed level. After imported into SIMCA-P software, the data was scaled by Pareto method. PCA model was built to observe the distribution of patient group and control. OPLS-DA was further referred for classification and biomarker selection.Our research found that the distribution of PBC and control was significantly different in PCA model, which indicate there are variance in the metabolic profine between the two groups. The OPLS model established can explain the variance between PBC group and control very well. However, the predictive power of OPLS model was strong.21 biomarkers of PBC were found in total.16 biomarkers had a higher level in PBC patients, while the rest had a higher level in the control group. The distribution of AIH and control was also noticed to be different in PCA model, which indicate there are variance in the metabolic profine between the two groups. The OPLS model established also can explain the variance between AIH group and control very well. The predictive power of OPLS model was not very well, either.14 biomarkers of AIH were finally identified, all of which had a higher level in AIH patients.