Abstract

Accurate diagnosis of complete inactivation of tuberculosis lesions is still a challenge with respect to sputum-negative tuberculosis. RNA-sequencing was conducted to uncover potential lncRNA indicators of metabolic activity in tuberculosis lesions. Lung tissues with high metabolic activity and low metabolic activity demonstrated by fluorine-18-fluorodeoxyglucose positron emission tomography/computed tomography were collected from five sputum-negative tuberculosis patients for RNA-sequencing. Differentially-expressed mRNAs and lncRNAs were identified. Their correlations were evaluated to construct lncRNA-mRNA co-expression network, in which lncRNAs and mRNAs with high degrees were confirmed by quantitative real-time PCR using samples collected from 11 patients. Prediction efficiencies of lncRNA indicators were assessed by receiver operating characteristic curve analysis. Bioinformatics analysis was performed for potential lncRNAs. 386 mRNAs and 44 lncRNAs were identified to be differentially expressed. Differentially-expressed mRNAs in lncRNA-mRNA co-expression network were significantly associated with fibrillar collagen, platelet-derived growth factor binding, and leukocyte migration involved in inflammatory response. Seven mRNAs (C1QB, CD68, CCL5, CCL19, MMP7, HLA-DMB, and CYBB) and two lncRNAs (ENST00000429730.1 and MSTRG.93125.4) were validated to be significantly up-regulated. The area under the curve of ENST00000429730.1 and MSTRG.93125.4 was 0.750 and 0.813, respectively. Two lncRNAs ENST00000429730.1 and MSTRG.93125.4 might be considered as potential indicators of metabolic activity in tuberculosis lesions for sputum-negative tuberculosis.