Whole-genome sequencing of presumptive MDR-TB isolates from a tertiary healthcare setting in Mumbai

Abstract

Objectives: Whole-genome sequencing (WGS) of Mycobacterium tuberculosis (MTB), proven to be a better alternative when compared with the combined sensitivity and specificity of all other modalities for diagnosis of tuberculosis (TB), aids epidemiological surveillance investigations by combining the current research with diagnostics. This study was conducted to identify and resolve operational challenges in performing WGS-based drug resistance testing (DRT) for MTB in a TB culture and drug susceptibility testing (DST) laboratory. Three critical, non-redundant steps for WGS-based DRT were tested: viz. DNA extraction, high-throughput paired-end next-generation sequencing (NGS), and genomic analysis pipeline for automated reporting of WGS-based DRT.

Methods: DNA was extracted from 100 liquid culture isolates on a mycobacterial growth indicator tube (MGIT) using DNEASY Ultraclean Microbial Kit (Qiagen, USA) as per the manufacturer's instructions. Illumina paired-end sequencing was performed. All analysis steps were automated using custom python scripts, requiring no intervention. Variant calling was performed as per the World Health Organization (WHO) technical guide.

Results: The number of cultures resistant to rifampicin, isoniazid, pyrazinamide, ethambutol, and streptomycin was 89, 88, 35, 67, and 73, respectively. Resistance to amikacin, kanamycin, and capreomycin was found in 15, 17, and 15 cultures, respectively. Seventy cultures were resistant to fluoroquinolones, four were resistant to ethionamide, and 12 were resistant to linezolid. Six cultures were resistant to only one of the 18 drugs tested. Seventy-five cultures were resistant to more than three anti-TB drugs. One culture was resistant to 13 of the 18 anti-TB drugs tested for this study. The maximum number of variants were observed in the rpoB gene (n = 93, 93%), wherein the Ser450Leu was the predominant mutation (n = 68, 73%). Ser315Thr was the most common variant (n = 86, 97%) that encoded resistance to isoniazid. The Lys43Arg variant encodes resistance to streptomycin and was the third most predominant variant (n = 65, 89%). In addition to the high levels of resistance observed in the dataset, we also observed a high proportion of Beijing strains (n = 63, 63%).

Conclusion: Compared with results from routine diagnostics based on the 'Guidelines on Programmatic Management of Drug-Resistant TB (PMDT) in India', none of the samples had DST available for all 18 drugs. This represents a gap in PMDT guidelines. The WGS-DRT must be considered as the primary DST method after a sample is flagged rifampicin-resistant by cartridge-based nucleic acid amplification testing (CBNAAT). With several research studies currently underway globally to identify novel variants associated with drug resistance and classifiy their minimum inhibitory coefficients, WGS-DRT presents a scalable technology that updates analytical pipelines, relegating the need for changing microbiological protocols.

Keywords: Antibiotic resistance; Genomic clusters; Mycobacterium tuberculosis; Public health; TB Diagnostics; Whole-genome sequencing.