Molecular Characterization of Non-tuberculous Mycobacteria Among Symptomatic Tuberculosis-negative Patients in Kenya

Abstract

Introduction: Non-Tuberculous Mycobacteria (NTM) are transmitted to humans through inhalation of dust or water vapor containing NTM, leading to NTM pulmonary diseases (NTM-PD), often misdiagnosed as tuberculosis (TB) due to similar symptoms. In Kenya, studies on NTM distribution have only been conducted in Western Kenya but the results are not adequate for a countrywide extrapolation and elucidation of NTM evolutionary diversity. Treatment of NTM-PD depends on the species, with macrolides and aminoglycosides for Rapid-Growing Mycobacteria (RGM) and macrolides with rifampicin for Slow-Growing Mycobacteria (SGM). Treatment failures are due to mutations in drug target regions: rrl for macrolides, rrs for aminoglycosides, and rpoB for rifampicin. While some NTM species are pathogenic and cause disease, others exist as harmless lung commensals; their ability to cause NTM-PD depends on the presence of mycobacterial virulence factors. We conducted this study to identify the most prevalent NTM species in Kenya, characterize the mutation patterns of resistance genes for macrolides, aminoglycosides, and rifampicin, and describe the diversity of virulence genes in the NTM. Methods: The study was conducted at the National Tuberculosis Reference Laboratory from February to November 2020, analyzing 166 sputum samples that showed growth in mycobacterial culture but tested negative for Mycobacterium tuberculosis using the Capilia rapid diagnostic test. Polymerase Chain Reaction (PCR) targeting the hsp65 gene was performed on these samples. Samples showing a 441 bp band in gel electrophoresis underwent sequencing with an ABI 3730XL analyzer, and species-level identification of NTM was achieved using NCBI BLAST. Subsequently, all identified NTM species underwent PCR and sequencing targeting the rrl gene, with additional rrs and rpoB sequencing for RGM and SGM respectively. Multiple sequence alignment for these genes against their wild-type sequences was conducted using Geneious software. Whole genome sequencing was performed using Oxford Nanopore Technology (ONT) on 12 species, each representing the dominant species within the phylogenetic clades. The sequences were analyzed with bioinformatic tools including Guppy, Trimmomatic, Unicycler, Prokka, and FreeBayes to identify virulence genes. Results: Out of 166 samples analyzed, 122 (73.5%) tested positive for NTM. The majority of these samples originated from the Coastal region (41%) and Nairobi (26%). Majority of participants fell within the 30-39 age range, with a median age of 39 years (IQR 20-50). Males comprised 73% of the NTM cases, and 31% were for those undergoing TB treatment follow-up. We identified 43 different NTM species, with the most prevalent being Mycobacterium avium complex species (31%), followed by M. fortuitum complex species (20%), and M. abscessus complex species (14%). The SGM comprised 71% of the isolates, while RGM accounted for 29%. Of the 122 NTM, 10.4% showed mutations in the rrl gene, with 58.3% of these being RGM and 41.7% SGM. iii Additionally, 11.1% of RGM had rrs mutations, and 14.7% of SGM had mutations in the rpoB gene. Whole genome sequencing revealed NTM genome sizes ranging from 4.6Mbp for M. terrae to 6.5Mbp for M. gordonae. Notably, M. gordonae had the highest (81%) number of virulence genes, with representation from all secretion systems, while M. chelonae had the lowest (38.1%). All NTM genomes contained ESX-3 genes, but ESX-1 genes were less prevalent, and PE/PPE protein genes were less common, except in M. alsense and M. abscessus bollettii. Conclusion: This study offers the first comprehensive characterization of NTM diversity in Kenya, revealing a high prevalence of NTM among symptomatic TB-negative patients. The majority of isolates were SGM, with Mycobacterium avium complex (MAC) species being the most prevalent. A significant proportion of NTM exhibited mutations in drug target genes for at least one antibiotic used in macrolide-based therapy. Genomic analysis indicated that while all NTM genomes contained ESX-3 genes, ESX-1 genes were the least common within the ESX transport system.