Proteomic Mass Spectrometry Analysis And The Prediction Model Construction Associated With Prognosis Of Severe Hepatitis

Objective: To obtain the plasma protein fingerprints of HBV-related severe hepatitis patients with different clinical outcomes and healthy people by using SELDI-TOF-MS?surface enhanced laser desorption/ionization time-of flight mass spectrometry?,screen out differential proteins associated with the prognosis of severe hepatitis,and establish the prediction model using these differential proteins,provide the reliable and sensitive index for the prognosis of severe hepatitis.Methods: Plasma specimens of 35 patients with HBV-related severe hepatitis in the infectious diseases department of the Affiliated Hospital of Southwest Medical University?Sichuan,China?were selected as the experimental group.Plasma specimens of 15 healthy volunteers were collected as the control group.Demographic data and clinical data of patients?including liver function,coagulation function,etc.?were collected at the same time.Experimental group patients followed up for at least 12 weeks.According to the outcomes,the patients were divided into death group and survival group.Plasma specimens of patients with HBV-related severe hepatitis and normal healthy human were detected by SELDI-TOF-MS,and the plasma protein fingerprints were obtained with CM10?an weak cation exchange?.The proteomic profiles of plasma were analyzed and identified by ProteinChip Software 3.2 and Biomarker Wizard Software.The collected data were statistically analyzed by SPSS 16.0 software.The preliminary diagnosis prediction model for patients with HBV-related severe hepatitis was established using decision tree model by Biomarker Patterns Software?BPS 5.0?,and the sensitivity and specificity to the prognosis of patients with HBV-related severe hepatitis were calculated.Measurement data described as mean±Standard deviation.comparisons of experimental group and health group were made by two independent samples t test.Statistical significance was accepted at P<0.05.Results: Comparing the protein peaks levels in plasma samples in the range from 1000 Da to 50000 Da of the most discriminative mass/charge between HBV-related severe hepatitis patients and healthy controls,57 differently expressed protein peaks were detected by the analysis of Biomarker Wizard software.There were 25 protein peaks with statistical difference between survival group and healthy control group?P<0.05?,6 protein peaks increased significantly in survival group,including M2017₅7?M7568₈6?M7937₈7?M15109₀?M15268₂?M15845₀,19 protein peaks decreased significantly in survival group,including M2196₇1?M2235₇8?M2747₂3?M2796₈5?M2881₄8?M2913₂6?M3224₁7?M3322₈4?M4309₄1?M6437₀1?M6634₈4?M6839₆4?M8601₁2?M9185₅6?M28057₇?M33395₂?M34404₉ ? M44842₁ ? M45146₈.There were 24 protein peaks with statistical difference between death group and healthy control group?P<0.05?,6 protein peaks increased significantly in death group,including M2017₅7?M6634₈4?M7568₈6?M7937₈7?M15268₂?M15845₀,18 protein peaks decreased significantly in death group,including M2235₇8?M2747₂3?M2881₄8?M2913₂6?M3224₁7?M3322₈4?M3376₄9?M4309₄1?M5908₂6?M6437₀1?M6584₆3?M6839₆4?M8601₁2?M14670₈?M15109₀?M28057₇?M33395₂?M34404₉.There were 16 protein peaks with statistical difference between survival group and death group?P<0.05?,including M3376₄9,M3479₆5,M3492₇9,M3936₅6,M5649₁1,M7568₈6,M7766₄6,M7937₈7,M8601₁2,M9290₄4,M11692₆,M 14670₈,M15109₀,M15268₂,M15845₀,M45146₈.The decision tree classification model were composed of 5 discriminative protein peaks selecting from survival group and death group by Biomarker Patterns Software.These differential protein peaks were M 3492₇9,M 5649₁1,M 3936₅6,M 7568₈6,M 8601₁2.The sensitivity of the decision tree model to the prognosis of patients with HBV-related severe hepatitis was 72.00% and the specificity was 100.00%.Conclusion: 1.SELDI-TOF-MS is used to detect the discriminative protein expression in the course of the development of HBV-related severe hepatitis,and there are significant differences in plasma protein expression in patients with different clinical outcomes.2.It is important to use those protein peaks to establish the diagnosis prediction model for the clinical outcomes of HBV-related severe hepatitis.it is suggested that the expression of these proteins may be related to the prognosis of HBV-related severe hepatitis,but further study are needed to comfirm.