Individuals with latent tuberculosis infection (LTBI) were regarded as an enormous reservoir of cases with active tuberculosis (TB). To strengthen LTBI management, biomarkers and tools are urgently required for identifying and ruling out active TB in a fast and effective way. Based on an open-label randomized controlled trial aiming to explore short-course LTBI treatment regimens, DNA methylation profiles were retrospectively detected to explore potential biomarkers, which could discriminate active TB from LTBI. The Infinium MethylationEPIC BeadChip array was used to analyze genomewide DNA methylation levels for 15 persons with LTBI who later developed active TB and for 15 LTBI controls who stayed healthy. The differentially methylated CpGs (dmCpGs) located in the promoter regions pre- and post-TB diagnosis were selected (P < 0.05 and |Δβ|>0.10) and evaluated by receiver operating characteristic (ROC) analysis. Eight dmCpGs were identified to be associated with TB occurrence; six were located in hypermethylated genes (cg02493602, cg02206980, cg02214623, cg12159502, cg14593639, and cg25764570), and two were located in hypomethylated genes (cg02781074 and cg12321798). ROC analysis indicated that the area under curve (AUC) of these eight dmCpGs ranged from 0.72 to 0.84. Given 90% sensitivity, the specificity was highest for cg14593639 at 66.67%. The combination analysis indicated that "cg02206980 + cg02214623 + cg12159502 + cg12321798" showed the best performance, with an AUC of 0.88 (95% confidence interval [CI]: 0.72, 0.97), a sensitivity of 93.33% (95% CI: 70.18%, 99.66%), and a specificity of 86.67% (95% CI: 62.12%, 97.63%). Our preliminary results indicate the potential value of the DNA methylation level as a diagnostic biomarker for discriminating active disease in LTBI testing. This finding requires further verification in independent populations with large sample sizes. IMPORTANCE Approximately a quarter of the world population had been infected with Mycobacterium tuberculosis, and about 5 to 10% of these individuals might develop active disease in their lifetimes. As a critical component of the "end TB strategies," preventive treatment was shown to protect 60 to 90% of high-risk LTBIs from developing active disease. Developing new TB screening tools based on blood-based biomarkers, which could identify and rule out active TB from LTBI, are prerequisite before initialing intervention. We tried to explore potential DNA methylation diagnostic biomarkers through retrospectively detected DNA methylation profiles pre- and post-TB diagnosis. Eight dmCpGs were identified, and the combination of "cg02206980 + cg02214623 + cg12159502 + cg12321798" showed a sensitivity of 93.33% and a specificity of 86.67%. The preliminary results provided new insight into detecting the DNA methylation level as a potential tool to distinguish TB from LTBI.

Tuberculosis (TB) remains one of the deadliest communicable diseases. Biomarkers predicting the risk of active disease development from latent tuberculosis infection (LTBI) are urgently needed for precise intervention. This study aimed to identify potential circulating microRNAs (miRNAs) playing such a role in Chinese population. Based on a prospective study aiming to track the development of active TB among rural residents with LTBI, the baseline levels of circulating miRNAs were retrospectively compared between those who developed TB (case group) and those age-gender matched controls remain free of TB (contraol group) during the follow-up. Agilent human miRNA microarray were used to select differently expressed circulating miRNAs and verified by subsequent real-time quantitative PCR (RT-qPCR). Six candidate miRNAs were expressed at statistically significant levels between the two groups at the baseline, as determined by microarray. Following verification among 150 study participants by RT-qPCR, the levels of hsa-miR-16-5p (P < 0.001) and hsa-miR-451a (P < 0.001) were found to be significantly lower in case group compared to control group. The combined areas under curves (AUCs) and precision-recall curves (PRCs) were 0.84, 0.86 and 0.85, 0.87 for hsa-miR-16-5p and hsa-miR-451a, respectively. hsa-miR-451a combined with body mass index (BMI) and prior history of TB presented the best performance, with a sensitivity of 80.82% and an acceptable specificity of 79.22%. After adjusting the two co-variables, the AUC of hsa-miR-451a was 0.78. Circulating levels of hsa-miR-451a showed potential to predict development of active TB from LTBI in a Chinese population. Further studies are warranted to verify these findings in varied study settings. IMPORTANCE Approximately a quarter of the world population are infected with M. tuberculosis and about 5% to 10% of these might develop active disease in their lifetime. Preventive treatment could effectively protect individuals at a high risk of developing active disease from LTBI, and is regarded as a critical component of End TB Strategies. Biomarkers which could accurately identify high-risk population and predict the risk of disease development are urgently needed for developing local guidelines of LTBI management and precise intervention. A nested case-control study was designed to explore possible microRNAs related with TB occurrence based on a previous prospective study, which aimed to track the development of active TB among rural residents with LTBI. The baseline circulating levels of hsa-miR-16-5p and hsa-miR-451a were significantly lower in TB cases compared to those in LTBI controls. Further receiver operator characteristic (ROC) curve analysis found that hsa-miR-451a showed considerable potential to predict the development of active TB from LTBI.