PLoS ONE (Jan 2022)

Prediction of drug resistance by Sanger sequencing of Mycobacterium tuberculosis complex strains isolated from multidrug resistant tuberculosis suspect patients in Ethiopia.

  • Eyob Abera Mesfin,
  • Matthias Merker,
  • Dereje Beyene,
  • Abreham Tesfaye,
  • Yassir Adam Shuaib,
  • Desalegn Addise,
  • Belay Tessema,
  • Stefan Niemann

DOI
https://doi.org/10.1371/journal.pone.0271508
Journal volume & issue
Vol. 17, no. 8
p. e0271508

Abstract

Read online

BackgroundEthiopia is one of the high multidrug-resistant tuberculosis (MDR-TB) burden countries. However, phenotypic drug susceptibility testing can take several weeks due to the slow growth of Mycobacterium tuberculosis complex (MTBC) strains. In this study, we assessed the performance of a Sanger sequencing approach to predict resistance against five anti-tuberculosis drugs and the pattern of resistance mediating mutations.MethodsWe enrolled 226 MTBC culture-positive MDR-TB suspects and collected sputum specimens and socio-demographic and TB related data from each suspect between June 2015 and December 2016 in Addis Ababa, Ethiopia. Phenotypic drug susceptibility testing (pDST) for rifampicin, isoniazid, pyrazinamide, ethambutol, and streptomycin using BACTEC MGIT 960 was compared with the results of a Sanger sequencing analysis of seven resistance determining regions in the genes rpoB, katG, fabG-inhA, pncA, embB, rpsL, and rrs.ResultDNA isolation for Sanger sequencing was successfully extracted from 92.5% (209/226) of the MTBC positive cultures, and the remaining 7.5% (17/226) strains were excluded from the final analysis. Based on pDST results, drug resistance proportions were as follows: isoniazid: 109/209 (52.2%), streptomycin: 93/209 (44.5%), rifampicin: 88/209 (42.1%), ethambutol: 74/209 (35.4%), and pyrazinamide: 69/209 (33.0%). Resistance against isoniazid was mainly mediated by the mutation katG S315T (97/209, 46.4%) and resistance against rifampicin by rpoB S531L (58/209, 27.8%). The dominating resistance-conferring mutations for ethambutol, streptomycin, and pyrazinamide affected codon 306 in embB (48/209, 21.1%), codon 88 in rpsL (43/209, 20.6%), and codon 65 in pncA (19/209, 9.1%), respectively. We observed a high agreement between phenotypic and genotypic DST, such as 89.9% (at 95% confidence interval [CI], 84.2%-95.8%) for isoniazid, 95.5% (95% CI, 91.2%-99.8%) for rifampicin, 98.6% (95% CI, 95.9-100%) for ethambutol, 91.3% (95% CI, 84.6-98.1%) for pyrazinamide and 57.0% (95% CI, 46.9%-67.1%) for streptomycin.ConclusionWe detected canonical mutations implicated in resistance to rifampicin, isoniazid, pyrazinamide, ethambutol, and streptomycin. High agreement with phenotypic DST results for all drugs renders Sanger sequencing promising to be performed as a complementary measure to routine phenotypic DST in Ethiopia. Sanger sequencing directly from sputum may accelerate accurate clinical decision-making in the future.