Study On Serum Metabolic Profiling In Different Stage Of Liver Cirrhosis Patients With Hepatitis B Virus Infection

Hepatitis B virus (HBV)-infected cirrhosis is a worldwide chronic liverdiseases and its pathogenesis is incompletely understood. During the earlycirrhosis, liver is able to compensate the changes resulting from bridgingfibrosis, and most patients did not show specific symptoms until they enterthe stage of decompensated cirrhosis.There is no effective treatment for this disease and early diagnosis is veryimportant. The diagnostic confirmation of cirrhosis is based on thehistological examination and imaging examinations. However, the proposedmethods could not be satisfactorily applied in clinical diagnose. Firstly, theimaging studies could not provide sensitive diagnoses. Secondly, thehistological examination requires the puncturing of liver tissue and bringsgreat pain and risk to the patients. The transition from compensated todecompensated cirrhosis is so insidious that many patients lost the bestopportunity for treatment.Thus in the present study, we employed quantitative1H nuclear magneticresonance (NMR) to analyze the serum metabolic profiling in HBV-infectcirrhosis patients, combined with multivariate statistics, such asunsupervised principal component analysis (PCA), partial least-squares-discriminant analysis (PLS-DA) and orthogonal partialleast-squares-discriminant analysis (OPLS-DA). Our purpose was to searchfor the underlying metabolites that correlated with the pathogenesis ofHBV-infected cirrhosis, and establish diagnostic methods for HBV-infectedliver cirrhosis patients in different stage, furthermore prospective tests ofclinical samples were applied to validate the discriminatory power ofestablished models.PART ICOMPARISON OF THE METABOLIC PROFILING OFHEPATITIS B VIRUS-INFECTED CIRRHOSIS ANDALCOHOLIC CIRRHOSIS PATIENTS WITH USING1HNMR-BASED METABONOMICSObjective The aims of the present study were to depict the serummetabolic characteristics of hepatitis B virus (HBV)-infected cirrhosis andalcoholic cirrhosis patients, and to find the specific serum biomarkersassociated with the diseases. Methods A pilot metabolic profiling study wasconducted using three groups: HBV-infected cirrhosis patients (N=21),alcoholic cirrhosis patients (N=20) and healthy controls (N=20).1H nuclearmagnetic resonance (NMR)-based metabonomics was used to obtain theserum metabolic profiles of the samples. The acquired data were processedby multivariate principal component analysis (PCA), partialleast-squares-discriminant analysis (PLS-DA) and orthogonal partial least-squares-discriminant analysis (OPLS-DA). The discriminatorymetabolites between HBV-infected cirrhosis and alcoholic cirrhosis werefurther validated by classical biochemical assays. Results The OPLS-DAmodel was capable of distinguishing between HBV-infected and alcoholiccirrhosis patients. HBV-infected cirrhosis patients and healthy controlsubjects were discriminated with an R2X of36.4%, an R2Y of91.5%, and aQ2Y of0.887. Alcoholic cirrhosis patients were separated from the healthycontrols with an R2X of44.5%, an R2Y of92.6%, and a Q2Y of0.903.OPLS-DA modeling revealed a clear separation between alcoholic andHBV-infected cirrhotic patients with an R2X of33.8%, an R2Y of87.3%, anda Q2Y of0.826. Five metabolites, creatine, acetoacetate, isobutyrate,glutamine and glutamate, were identified as the most influential factors tocompare HBV-infected cirrhosis and alcoholic cirrhosis. The validation testsusing enzyme-linked immunosorbent assay (ELISA) and high performanceliquid chromatography (HPLC) showed the changes of five metabolites werecoincident well with the results of NMR. Conclusions NMR spectracombined with pattern recognition analysis techniques may provide a newway to distinguish HBV-infected and alcoholic cirrhosis patients andexplore their pathogenesis. PART II1H NMR-BASED SERUM METABOLIC PROFILING INCOMPENSATED AND DECOMPENSATED LIVERCIRRHOSISObjective The aim of the present study was to assess whether themetabolic profiling of serum samples would be helpful to diagnosecompensated and decompensated hepatic cirrhosis. Methods A pilotmetabolic profiling study was conducted using three groups: compensatedcirrhosis patients (N=30) and decompensated cirrhosis patients (N=30).1Hnuclear magnetic resonance (NMR)-based metabonomics was used to obtainthe serum metabolic profiles of the samples. The acquired data wereprocessed by multivariate principal component analysis (PCA), partialleast-squares-discriminant analysis (PLS-DA) and orthogonal partialleast-squares-discriminant analysis (OPLS-DA). A prospective study wasperformed to validate the power of the model using20clinical samples.Results The OPLS-DA model was capable of distinguishing betweendecompensated and compensated cirrhosis patients, with an R2X of24.6%,an R2Y of78.4%and a Q2Y of59.8%. Sixteen metabolites were identified asthe most influential factors for the difference of two groups. The validationof predicting diagnosis showed the accuracy of the OPLS-DA model was85%(17/20). Conclusions1H NMR spectra combined with patternrecognition analysis techniques offer a new way to diagnose compensated and decompensated cirrhosis in the future. PART IIISTUDY ON METABOLIC SIGNATURE OF SERUM INPATIENTS WITH MINIMAL HEPATICENCEPHALOPATHYObjective The aims of the present study were to investigate themetabonomics in the serum samples from minimal hepatic encephalopathy(MHE) patients and establish a diagnostic model for MHE. MethodsNuclear magnetic resonance spectroscopy (NMR) was applied to analyzethe serums metabonomics from25simple1iver cirrhosis patients and25cirrhotic patients with minimal hepatic encephalopathy (MHE). Principalcomponent analysis (PCA), partial least-squares-discriminant analysis(PLS-DA) and orthogonal partial least-squares-discriminant analysis(OPLS-DA) were employed to establish the pattern recognition analysis ofNMR spectra. A prospective study was performed to validate the power ofthe model using10MHE samples. Results The OPLS-DA model wascapable of distinguishing patients with and without MHE, with an R2X of30.4%, an R2Y of85.1%and a Q2Y of0.817. Compared with cirrhoticpatients without MHE, patients with MHE showed increased glutamine, phenylalanine, acetone and decreased valine and isoleucine. The validationinvestigation displayed the accuracy ratio of the OPLS-DA model for MHEwas80%(8/10). Conclusions Metabonomics spectra based on1H-NMRcombined with pattern recognition analysis techniques can explore variousmetabolites in serum, and offer a new way to study the pathogenesis ofliver cirrhosis and to diagnose MHE in the future.