Targeted transcriptomic signature for monitoring anti-tuberculosis treatment response

Abstract

Abstract The long duration of anti-tuberculosis treatment is associated with poor adherence and treatment outcomes leading to higher relapse rates. Shortening treatment duration has been suggested to improve treatment outcome by increasing adherence. Sputum conversion, the gold standard for monitoring treatment, has low sensitivity and specificity for predicting relapse. Better markers are therefore needed and blood-based biomarkers have been suggested as potential alternatives. The aim of this study was to identify transcriptional signatures to predict anti-tuberculosis treatment outcomes, relapse, and treatment failure. We extracted RNA from whole blood PAXgene samples from 37 confirmed TB cases at diagnosis (baseline), 14 days and two months of treatment and measured gene expression using targeted sequencing on the NanoString nCounter MAX platform using the 799-gene host response panel. To identify genes that predict treatment response and relapse, we identified differentially expressed genes (DEGs) at Day 14 and two months of treatment and compared expression between fast (sputum converters by 2 months) and slow responders (sputum converters after 2 months). Using a cut-off of 1.5-fold change and a p value of 0.05, we observed 58 DEGs at Day 14 and 175 DEGs at two months relative to baseline. When we compared slow and fast responders, 14 genes were differentially expressed but these were not significant when adjusted for multiple comparisons. We have shown that DEGs can identify different stages of disease or treatment. Further analysis is ongoing to include other treatment timepoints (four months and six months) and importantly to include participants with known relapse.