Detection of Mycobacterium tuberculosis Complex Bacilli and Nucleic Acids From Tongue Swabs in Young, Hospitalized Children

Abstract

Diagnosis of tuberculosis in pediatric patients remains challenging due to inherent difficulties associated with obtaining respiratory samples for molecular and culture-based testing. To address this, recent studies have highlighted the utility of tongue swabs to detect Mycobacterium tuberculosis genomic DNA in the oral epithelia of tuberculosis infected adults. It is unknown whether tongue swabs have similar utility for diagnosis of childhood tuberculosis and if the presence of DNA in these swabs was associated with whole bacilli. We therefore sought to conduct a preliminary assessment of the utility of tongue swabs to detect tubercle bacilli and their associated genetic material in young children. For this, we recruited hospitalized children with clinically diagnosed tuberculosis (n = 26) or lower respiratory tract infection (LRTI, n = 9). These categories were blinded for downstream laboratory tests, which included PCR, spoligotyping, smear microscopy, and culture. Mtb genomic DNA was detected by PCR only in clinically diagnosed TB cases [11/26 (31.4%)] and not in cases with LRTI. Of these, 5/11 [45.5%] were associated with a spoligotype. Spoligotyping also detected an additional six specimens that were negative by PCR. Using smear microscopy, 19/26 [73.1%] and 4/9 [44.4] were Mtb positive in the tuberculosis or LRTI categories respectively. We noted positive results on all three tests in 5/26 [19.2%] in the tuberculosis category and 0/9 in the LRTI category. All specimens were culture negative. Collectively, these preliminary data present a compelling case for broader testing of tongue swabs to diagnose tuberculosis in children where obtaining standard sputum specimens is not easy.

Keywords: auramine staining; pediatric tuberculosis; qPCR; spoligotyping; tongue swabs.

Figure 1

Participant disposition flow chart. A total of 35 patients were analyzed in this study. These children were hospitalized for either tuberculosis (TB) or lower respiratory tract infection (LRTI). The clinical diagnoses were blinded for downstream tests which included detection of genomic DNA (gDNA, using qPCR and spoligotyping), smear microscopy, and culture using detection methods to quantify culturable [colony forming unit (CFUs)] and differentially culturable tubercle bacteria [most probable number (MPN)]. Upon unblinding, 26 patients were grouped into the TB category. Of these, 11 (42.3%) had been detected using PCR genome equivalents (GE) ≥10. Fourteen swabs were positive by spoligotyping. Of the nine patients that were diagnosed with LRTI, a single Mtb strain-type and two Mycobacterium africanum strains were detected by spoligotyping. TB, tuberculosis; LRTI, lower respiratory tract infection; MPN, most probable number; CFU/ml, colony forming units per ml; AFB, acid-fast bacilli will present as Auramine-positive; FOVs, fields of view; IS6110, insertion element only found within the members of the Mycobacterium tuberculosis complex (MTBC); qPCR, quantitative real-time PCR; DR, direct repeats are separated by spacers that polymorphic—unique in the different members of the MTBC.

Figure 2

Auramine smear microscopy from tongue swab samples. Ten microliters of each sample was aliquoted onto a microscope slide and stained with Auramine-O. (A) Representative images for positive samples from patients clinically diagnosed with TB; containing ≥10 genome equivalents as detected by qPCR and a spoligotype. Boxes in red represent positive controls from sputum (adult patient diagnosed with TB) and the lab strain of Mtb (H37Rv). (B) Samples from patients clinically diagnosed with LRTI, but their smears were detected as positive for Mtb. Sample 29 contained <10 genome equivalents of Mtb gDNA, ‘scanty’ acid-fast bacilli count and was assigned a strain-type by spoligotyping. Sample 32 was scored as ‘1+’ for acid-fast bacilli but negative for all other tests. In all cases, slides were viewed using a Zeiss Observer Z1 using two channels, i.e. DIC (bright-field) and FITC (green) with exposure times of 100–150 and 3,000 ms, respectively. Scale bar represents 5 µm.