Novel RNA Biomarkers Improve Discrimination of Children with Tuberculosis Disease from Those with non-TB Pneumonia after In-Vitro Stimulation

The diagnosis of pediatric tuberculosis (TB) poses a challenge for clinical teams worldwide. TB-mediated changes in the expression of host genes in the peripheral blood can serve as diagnostic biomarkers and can provide better insights into the host immune mechanisms of childhood TB. Peripheral blood mononuclear cells (PBMCs) from children (n=102) with microbiologically confirmed TB disease, ΤΒ infection (ΤΒΙ), pneumonia, and healthy controls (HC) were stimulated with either the Purified Protein Derivative (PPD) or the Early Secretory Antigen 6kDa-Culture Filtrate Protein 10 (ESAT6-CFP10) complex of Mycobacterium tuberculosis (Mtb). RNA was extracted and quantified using gene expression microarrays. Differential expression analysis was performed comparing microbiologically confirmed TB to the other diagnostic groups for the stimulated and unstimulated samples. Using variable selection, we identified sparse diagnostic gene signatures; one gene (PID1) was able to distinguish TB from pneumonia after ESAT6-CFP10 stimulation with an AUC of 100% in the test set, while a combination of two genes (STAT1 and IFI44) achieved an AUC of 91.7% (CI95% 75.0%-100%) in the test set after PPD stimulation. The number of significantly differentially expressed (SDE) genes was higher when contrasting TB to pneumonia or HC in stimulated samples, compared to unstimulated ones, leading to a larger pool of candidate diagnostic biomarkers. Our approach provides enlightened aspects of peripheral TB-specific responses and can form the basis for a point of care test meeting the World Health Organization (WHO) Target Product Profile (TPP) for pediatric TB. 102 children were enrolled in the study and were classified into 4 different groups, microbiologically confirmed TB disease, TBI, pneumonia and HC. In the patient groups the whole pheripheral blood was collected before treatment for TB disease or TBI was initiated. PBMCs were isolated according to the Histopaque 1077 (Sigma-Aldrich) cell isolation procedure, cultured in 0.2 ml growing medium, and were exposed for 24 hours in triplicate to a) recombinant Mtb antigens, ESAT-6 and CFP-10, b) Tuberculin PPD , or c) remained in the nutrient agent without stimulation. The supernatant was removed from each well and the remaining culture medium was triturated to resuspend the attached cells, transferred to a new vial, and centrifuged. The supernatant was discarded and Trizol reagent was added to the cell pellet. After washing the cell monolayer with PBS, Trizol reagent was added to the culture plate, incubated for 5 minutes at RT, before being added to the cell pellet. The combined Trizol cell lysate was stored at -80oC. Τotal RNA was isolated from pooled replicates for each condition per patient using a modified RNeasy protocol (Qiagen). After QC, RNA was converted to biotinylated cRNA using Illumina TotalPrep RNA Amplification kits (Applied Biosystems) and the labelled cRNA hybridised to Human HT-12 V4 Expression BeadChip arrays (Illumina) following manufacturer’s instructions. Data were analysed in R.