Potential of DosR and Rpf antigens from Mycobacterium tuberculosis to discriminate between latent and active tuberculosis in a tuberculosis endemic population of Medellin Colombia

BACKGROUND Tuberculosis (TB) remains one of the most deadly infectious diseases. One-third to one-fourth of the human population is estimated to be infected with Mycobacterium tuberculosis (Mtb) without showing clinical symptoms, a condition called latent TB infection (LTBI). Diagnosis of Mtb infection is based on the immune response to a mixture of mycobacterial antigens (PPD) or to Mtb specific ESAT-6/CFP10 antigens (IGRA), highly expressed during the initial phase of infection. However, the immune response to PPD and IGRA antigens has a low power to discriminate between LTBI and PTB. The T-cell response to a group of so-called latency (DosR-regulon-encoded) and Resuscitation Promoting (Rpf) antigens of Mtb has been proved to be significantly higher in LTBI compared to active TB across many populations, suggesting their potential use as biomarkers to differentiate latent from active TB. METHODS PBMCs from a group LTBI (n = 20) and pulmonary TB patients (PTB, n = 21) from an endemic community for TB of the city of Medellín, Colombia, were in vitro stimulated for 7 days with DosR- (Rv1737c, Rv2029c, and Rv2628), Rpf- (Rv0867c and Rv2389c), the recombinant fusion protein ESAT-6-CFP10 (E6-C10)-, or PPD-antigen. The induced IFNγ levels detectable in the supernatants of the antigen-stimulated cells were then used to calculate specificity and sensitivity in discriminating LTBI from PTB, using different statistical approaches. RESULTS IFNγ production in response to DosR and Rpf antigens was significantly higher in LTBI compared to PTB. ROC curve analyses of IFNγ production allowed differentiation of LTBI from PTB with areas under the curve higher than 0.70. Furthermore, Multiple Correspondence Analysis (MCA) revealed that LTBI is associated with higher levels of IFNγ in response to the different antigens compared to PTB. Analysis based on decision trees showed that the IFNγ levels produced in response to Rv2029c was the leading variable that best-classified disease status. Finally, logistic regression analysis predicted that IFNγ produced by PBMCs in response to E6-C10, Rv2029c, Rv0867c (RpfA) and Rv2389c (RpfA) antigens correlates best with the probability of being latently infected. CONCLUSIONS The Mtb antigens E6-C10, Rv2029c (PfkB), Rv0867c (RpfA) and Rv2389c (RpfA), may be potential candidates to discriminate LTBI from PTB.