Metabolic Characteristics Of Large And Small Extracellular Vesicles From Pleural Effusion Reveal Biomarker Candidates For The Diagnosis Of Tuberculosis And Malignancy

Rational:Pleural effusion is a common respiratory disease worldwide;however,rapid and accurate diagnoses of tuberculosis pleural effusion(TPE)and malignancy pleural effusion(MPE)remain challenging.Although extracellular vesicles(EVs)have been confirmed as promising sources of disease biomarkers,little is known about the metabolite compositions of its subpopulations and their roles in the diagnosis of pleural effusion.Objectives:we performed metabolomics and lipidomics analysis to investigate the metabolite characteristics of two EV subpopulations derived from pleural effusion by differential ultracentrifugation,namely large EVs(lEVs,pelleted at 20,000g)and small EVs(sEVs,pelleted at 110,000g),and assessed their metabolite differences between tuberculosis and malignancy.Methods:A total of 80 pleural effusion specimens,including 20 in the discovery set and 60 in the validation set,were collected from 40 patients with pulmonary tuberculosis and 40 patients with lung cancer.The age and sex of subjects in the two groups were matched as much as possible.30mL of pleural fluid were centrifuged at 20000g and 110000g to obtain lEVs and sEVs respectively.The morphological characteristics of EVs were observed by transmission electron microscopy(TEM),the particle size and density distribution of lEVs or sEVs were analysed with a nanoparticle tracking analysis(NTA)instrument.In addition,we used western blotting analysis of EVs to identify protein expression.The EVs and the original pleural effusion were frozen and thawed together to prepare an analysis sample,and then the metabolites and lipids in the sample were extracted with chemical reagent.To obtain high-sensitivity and high-coverage of the EVs in metabolic profiling analysis,pseudotargeted metabolomics and lipidomics methods were used to acquire the LC-MS/MS spectra.Peak alignment of the acquired raw data was performed using the software tool Analyst V1.6 software.Multivariable analysis was conducted using SIMCA-P software.Univariate analysis of clinical and metabolic profiling data was performed by Multi Experiment Viewer software.The Wilcoxon Mann-Whitney test with benjamini-Hochberg-based false discovery rate(FDR)correction was utilized to evaluate the statistical significance,with p<0.05 and FDR<0.05 defined as statistically significant.Pearson correlation analysis among the clinical parameters and differential metabolites was implemented with SPSS software,and its network was displayed by Cytoscape software.Main results:First of all,we obtained pleural effusion from different clinical diseases and extracted extracellular vesicles.Two subgroups of EVs,lEVs and sEVs,were identified by TEM,NTA and WB analysis.TEM showed that the size and morphology of lEVs were more heterogeneous than those of sEVs,and the size of most lEVs(>150 nm)was larger than that of sEVs(50-150 nm).Additionally,the size distribution analysed by NTA showed that more than 90%lEVs showed a wide range of 100-400 nm in diameter,and their size distribution between MPE and TPE did not show a significant difference.The membrane proteins CD9,CD63(common EVs markers),and TSG101 were detected in both EV subgroups by western blotting.Most markers showed higher levels in sEVs than in lEVs.A total of 579 metabolites,including amino acids,acylcarnitines,organic acids,steroids,amides and various lipid species,were detected.The results showed that the metabolic profiles of lEVs and sEVs overlapped with and difference from each other but significantly differed from those of pleural effusion.Additionally,different type of vesicles and pleural effusion showed unique metabolic enrichments.Furthermore,lEVs displayed more significant and larger metabolic alterations between the tuberculosis and malignancy groups.In both sEVs and lEVs,most differential metabolites were closely associated with pCEA and pADA.The overlapped differential metabolites mainly displayed inverse relationships between pCEA and pADA.Further,the levels of pCEA and pADA show remarkable correlations with the levels of various amino acids in lEVs,which showed weak relationships in sEVs.Finally,a panel of four biomarker candidates,including phenylalanine,leucine,phosphatidylcholine 35:0,and sphingomyelin 44:3,in pleural lEVs was defined based on the comprehensive discovery and validation workflow.This panel showed high performance for distinguishing TPE and MPE,particularly in patients with delayed or missed diagnosis,such as the area under the receiver-operating characteristic curve(AUC)>0.95 in both sets.Conclusions:We described the metabolite characteristics of sEVs and lEVs in pleural effusion.In addition,the metabolite profiles obtained from lEVs and sEVs overlap and differ from each other and the original pleural effusion samples,indicating the potential of EVs-derived metabolomics analysis and discovery of biomarkers.In addition,we also explored the metabolic reprogramming of tuberculosis and malignany at the level of lEVs and sEVs instead of the level of conventional pleural effusion,providing new insights into the mechanism of pleural effusion.Finally,we evaluated the potential of EVs metabolites in the diagnosis of TPE and MPE.In the diagnosis of TPE and MPE,especially for patients with delayed or missed diagnosis,the combined use of lEVs metabolites and clinical parameters may be effective.