ABSTRACT
Introduction Respiratory pathogens such as influenza viruses may play an influential role in the pathogenesis of TB by negatively affecting immunity against Mycobacterium tuberculosis (MTB). We discovered and validated a transcriptomic signature of risk (SOR), based on mRNA expression of 11 IFN stimulated genes (ISG), which prospectively differentiates between incident TB cases and healthy controls. The SOR score is computed from expression abundance of multiple ISG transcript pairs in peripheral blood, whereby each pair functions as a ?vote? for or against TB risk. We aimed to identify respiratory pathogens other than MTB that might also associate with this SOR score and test whether the SOR score differentiates between individuals with and without respiratory pathogens. Methods We conducted a nested cross-sectional study of the upper respiratory tract microbiome. Upon consent, participants were consecutively enrolled into the study and provided one nasopharyngeal, one oropharyngeal and a PAXgene blood sample. Host blood SOR scores were computed from Ct values for each of the 11 genes, measured by microfluidic qRT-PCR. We used multiplex real-time PCR to detect 33 pathogens including bacteria, viruses and fungi in the nasopharyngeal and oropharyngeal samples Multivariate linear regression was used to identify pathogens associated with, and estimate their effect on, the SOR score. Wilcoxon rank sum tests and receiver-operating-characteristic (ROC) curves were used to differentiate participants with and without respiratory pathogens. Results 1,000 HIV-negative volunteers aged between 18 and 60 years were enrolled. 13 viral and nine bacterial pathogens were detected. Overall prevalence of respiratory pathogens was 43%: 4% were viruses only, 35.8% bacteria only, and 3.2% were a combination of viruses and bacteria. Influenza C, rhinoviruses, coronavirus OC43, adenoviruses, and mycoplasma pneumoniae were significantly associated (Table 1) with a high SOR score. In ROC curve analysis the SOR score differentiated participants as follows;virus vs no-pathogen (area under the curve;AUC) AUC=0.72, 95% CI: 0.63-0.81, virus vs bacteria AUC=0.72, 95% CI: 0.63-0.81, bacteria vs no-pathogen AUC=0.51, 95% CI: 0.48-0.55 (no difference). Participants with either viruses only or both viruses and bacteria had significantly higher (P=0.001) SOR scores (median 47% and 43%, respectively) compared to participants without pathogens (median 14%) or participants with bacteria only (median 13%). Conclusion Participants with upper respiratory tract viral colonization or infection had elevated SOR scores, suggesting induction of interferon signalling. Infection or colonization with respiratory viruses is likely to result in false positive results for other transcriptomic signatures of tuberculosis based on ISGs (Table Presented) .