Comparative Proteome Analysis Of Tuberculous Pleural Effusion And Malignant Pleural Effusion

ObjectiveIt is difficult to distinguish tuberculous pleural effusion from malignant pleuraleffusion. It has been known that pleural effusion, which is usually closer to theaffected organ and thus more specific for lung disorders than other body fluids,always contain plasma proteins as well as proteins released by inflammatory,epithelial or even cancer cells. Some of them may have potential diagnostic value. Atpresent, proteomic techniques offer a new way to discover specific biomarkers. In thisstudy we performed label free quantitative proteomics approach to identify proteinsdifferentially expressed in tuberculous pleural effusion and malignant pleural effusionand selected multiple candidates for preclinical validation, in order to seek novelbiomarkers.Methods1. Label free quantitative proteomics was applied in this study to detect the proteinsdifferentially expressed in5tuberculous pleural effusion and5malignant pleuraleffusion. The identified differential proteins were analyzed by bioinformatics withGene Onology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG).2. To validate the results of label free quantitative proteomics, Western blot was usedto examine the expression of a selected panel of proteins.3. Enzyme-linked immunosorbent assay (Elisa) was utilize to validate candidateproteins in50tuberculous pleural effusion and30malignant pleural effusion. Support vector machine was used to further screen candidate proteins to establish diagnosticmodel, meanwhile, diagnostic value of this model was estimated.Results1. A total of92differently expressed proteins (P＜0.05) were identified by nano liquidchromatography electrospray ionisation tandem mass spectrometry(NanoLC-ESI/MS/MS). In term of ProfileAnalysis2.0software, there were63proteins that differed significantly between tuberculous pleural effusion group andmalignant pleural effusion group. Among those proteins,31proteins wereup-regulated significantly in tuberculous pleural effusion group compared withmalignant pleural effusion group, while32proteins were down-regulated significantly.GO annotation terms indicated that the identified differential proteins were mainly inextracellular part and vesicle; main molecular functions were enzyme inhibitorfunction and protein binding function; main biological processes were immuneresponse, homeostasis process, defense reponse, coagulation, inflammatory response,protein maturation processing, complement activation, cell apoptosis, and host cellprocess. KEGG analysis indicated that complement and coagulation cascades,non-small cell lung cancer, HIF-1signaling pathway, ErbB signaling pathway, andp53signaling pathway were the top ranked pathways. These findings suggest that thedifferential proteins might be implicated in initiation and progression of tuberculouspleural effusion or malignant pleural effusion.2. The relative expression level of15proteins (immunoglobulin kappa constant,complement9, vitamin D-binding protein, ceruloplasmin, alpha-2-HS-glycoprotein,alpha2macroglobulin, alpha-1-acid glycoprotein, orosomucoid2, complement4,haptoglobin, kininogen1, fibrinogen gamma chain, fibrinogen beta chain,apolipoprotein A-I and stress-induced-phosphoprotein1) were verified by Westernblot. The results showed that the relative expression level of12proteins coincided with the results of label free quantitative proteomics, and only3proteins showedincompatible change (accuracy,80.0%), which demonstrates the accuracy of labelfree quantitation.3. The levels of immunoglobulin kappa constant, complement9,alpha-2-HS-glycoprotein, vitamin D-binding protein, ceruloplasmin, alpha2macroglobulin, kininogen1, orosomucoid2and alpha-1-acid glycoprotein intuberculous pleural effusion (n=50) and malignant pleural effusion (n=30) weremeasured by Elisa system. The levels of immunoglobulin kappa constant,complement9, alpha-2-HS-glycoprotein, vitamin D-binding protein, ceruloplasminand alpha2macroglobulin were significantly higher in tuberculous pleural effusiongroup than in malignant pleural effusion group (P＜0.05); kininogen1wassignificantly lower in tuberculous pleural effusion group than in malignant pleuraleffusion group (P＜0.05); orosomucoid2and alpha-1-acid glycoprotein were nostatistically significant difference between the two groups (P＞0.05). To evaluate thediagnostic performance of the7statistically differential proteins, receiver operatingcharacteristic (ROC) curve was constructed. ROC curve analysis indicated that thediagnostic efficacy of single protein was not excellent (the area under the ROC curve＜0.9). Therefore, immunoglobulin kappa constant, complement9,alpha-2-HS-glycoprotein and kininogen1were selected to established diagnosticmodel by support vector machine. The sensitivity, specificity, accuracy and the areaunder the ROC curve were96.0%,100.0%,97.5%and0.965, respectively, whichdemonstrated that the diagnostic accuracy of this model was high.ConclusionsUsing label free quantitative proteomics approach to investigate differential proteinsbetween tuberculous pleural effusion and malignant pleural effusion is a feasiblestrategy to seek valuable biomarkers. Immunoglobulin kappa constant, complement9, alpha-2-HS-glycoprotein and kininogen1could be used to identify tuberculouspleural effusion and malignant pleural effusion as a set of potential biomarkers.