Detection of M. tuberculosis DNA in peripheral blood mononuclear cells of tuberculosis contacts does not associate with blood RNA signatures for incipient tuberculosis

Detection of M. tuberculosis DNA in peripheral blood mononuclear cells of tuberculosis contacts does not associate with blood RNA signatures for incipient tuberculosisTo the Editor:Human exposure to Mycobacterium tuberculosis (Mtb) is thought to result in a spectrum of outcomes, including bacillary clearance, quiescent Mtb infection, incipient tuberculosis (TB), subclinical TB and active TB [1].Incipient TBdefined as a prolonged asymptomatic phase of early disease preceding clinical presentation as active disease [2]may be distinguished from quiescent Mtb infection by detection of host gene expression signatures in blood, whose presence associates with increased risk of progression to active TB [3].We and others have previously reported detection of Mtb DNA in CD34-positive peripheral blood mononuclear cells (PBMCs) of asymptomatic TB-exposed adults with normal chest radiographs [4][5][6].In one such study conducted in Ethiopia [5], administration of isoniazid preventive therapy (IPT) reduced the proportion of HIV-infected individuals in whom this signal was detectable from 95% to 53% ( p<0.0001), suggesting that detection of Mtb DNA in PBMCs may represent the presence of viable bacteria.Whether these bacteria provoke a host transcriptional response associated with subclinical disease is not known.To investigate, we conducted a cross-sectional study with a nested prospective component, as previously described [5].Briefly, three groups of asymptomatic adults at risk of TB infection (household TB contacts, farm workers with occupational exposure to bovine TB and HIV-infected people attending an outpatient clinic) and living in or around Addis Ababa, Ethiopia, were invited to give a blood sample for processing as described below.Principal exclusion criteria were age <18 years; symptoms of active TB; haemoglobin concentration <10 gdL -1 ; and presence of any chest radiograph abnormality consistent with active TB.HIV-infected participants who did not have any contraindication to IPT were offered a 6-month course of 300 mg isoniazid and pyridoxine 50 mg daily and invited to return to give a second blood sample for the same laboratory tests following treatment completion.The study was approved by the Ethiopian National Research Ethics Review Committee, Addis Ababa, Ethiopia (ref 310/253/2017).Written informed consent was obtained from all participants.PBMCs were isolated and sorted into CD34-positive and -negative subsets prior to DNA extraction and detection of Mtb DNA, as previously described [5].For determination of host gene expression signatures, 3 mL blood was collected into Tempus Blood RNA Tubes (Thermo Fisher Scientific) and processed with Tempus Spin RNA Isolation Kits (Thermo Fisher Scientific) to obtain high-quality total RNA.For genome-wide mRNA sequencing, complementary DNA libraries were generated using Kappa HyperPrep kits (Roche) and sequenced on the Illumina NextSeq 550 system using NextSeq 500/550 High Output 75 Cycle Kits (Illumina) RNA sequencing data were mapped to Ensembl Human GRCh38 release 104 using Kallisto (v0.46.1).Scores for eight incipient TB signatures were calculated as reported elsewhere [7][8][9][10][11][12][13]. Unpaired t-tests were used to compare gene expression Z-scores of HIV-uninfected participants in whom Mtb DNA was detected versus undetected at baseline.In this analysis, a sample size of n=24 per group achieved 92% power to identify a difference in Z-score means of 1 (the minimum biologically meaningful difference) with 5% alpha, assuming standard deviation of 1.0 for Z-score values.Paired t-tests were used to compare gene expression Z-scores of HIV-infected participants before versus after IPT, with stratification by Mtb PCR responder status.In this analysis, a sample size of n=17 responders achieved 97% power, and a sample size of n=7 non-responders achieved 60% power, to identify a difference in Z-score means of 1 with 5% alpha.Between-group differential gene expression was investigated using Deseq2 and false discovery rate (FDR) <0.05.