Molecular characterization of mutations associated with resistance among 72 multidrug-resistant strains of Mycobacterium tuberculosis by whole genome sequencing / 中国热带医学

ABSTRACT

@#Abstract Objective　To understand the characteristics of mutations associated with resistance among 72 multidrug-resistant tuberculosis (MDR-TB) strains using whole genome sequencing (WGS) and to evaluate the performance of WGS for predicting MDR-TB drug resistance. Methods　The clinical strains isolated from patients who visited the outpatient department of Tianjin Center for Tuberculosis Control from January to September in 2020 were collected. Identification tests using p-nitrobenzoic acid (PNB) medium were performed. Drug susceptibility tests (proportion method) on L-J medium were performed. After excluding duplicate strains, 72 MDR-TB strains were selected for WGS. Data were analyzed by using online databases and the phenotypic drug susceptibility test results were compared with resistance profiles predicted by WGS. Results　All of 72 MDR-TB strains belonged to linage 2, and there was no significant difference in rate of pre-extensive drug-resistant tuberculosis (pre-XDR-TB) between modern type and ancestral type (χ2=0.287, P=0.592). A total of 81 mutation types were found from resistance-related genes for 12 anti-tuberculosis drugs, and the common mutation types in different drug-resistant strains were streptomycin (SM) rpsL Lys43Arg; isoniazid (INH) katG Ser315Thr; rifampicin (RIF) rpoB Ser450Leu; ethambutol (EMB) embB Met306Val; ofloxacin (OFX), levofloxacin (LFX), moxifloxacin (MFX) gyrA Asp94Gly; kanamycin (KAM), capreomycin (CAP), amikacin (AMK) rrs 1401a>g; para-aminosalicylic acid (PAS) folC Ile43Thr. Nine mutation types were found in 9 prothionamide (PTO)-resistant strains, one type for each strain. The sensitivity and specificity of WGS for predicting resistance to different drugs were SM 98.15% and 88.89%, INH 90.28% and -, RIF 98.62% and -, EMB 79.49% and75.76%, OFX 97.30% and 85.71%, KAM 85.71% and 98.46%, PAS 27.27% and 95.08%, PTO 81.82% and 60.66%, CAP 60.00% and 98.51%, LFX 97.22% and 83.33%, MFX 97.30% and 85.71%, AMK100.00% and 100.00%, respectively. Conclusion　WGS is a rapid and promising method which has high consistency with the phenotypic drug sensitivity test. Therefore, it has good application prospects in predicting drug resistance in MDR-TB.