Evaluation of Whole Genome Sequencing-Based Predictions of Antimicrobial Resistance to TB First Line Agents: A Lesson from 5 Years of Data.

Phenotypic susceptibility testing of the Mycobacterium tuberculosis complex (MTBC) isolate requires culture growth, which can delay rapid detection of resistant cases. Whole genome sequencing (WGS) and data analysis pipelines can assist in predicting resistance to antimicrobials used in the treatment of tuberculosis (TB). This study compared phenotypic susceptibility testing results and WGS-based predictions of antimicrobial resistance (AMR) to four first-line antimicrobials-isoniazid, rifampin, ethambutol, and pyrazinamide-for MTBC isolates tested between the years 2018-2022. For this 5-year retrospective analysis, the WGS sensitivity for predicting resistance for isoniazid, rifampin, ethambutol, and pyrazinamide using Mykrobe was 86.7%, 100.0%, 100.0%, and 47.8%, respectively, and the specificity was 99.4%, 99.5%, 98.7%, and 99.9%, respectively. The predictive values improved slightly using Mykrobe corrections applied using TB Profiler, i.e., the WGS sensitivity for isoniazid, rifampin, ethambutol, and pyrazinamide was 92.31%, 100%, 100%, and 57.78%, respectively, and the specificity was 99.63%. 99.45%, 98.93%, and 99.93%, respectively. The utilization of WGS-based testing addresses concerns regarding test turnaround time and enables analysis for MTBC member identification, antimicrobial resistance prediction, detection of mixed cultures, and strain genotyping, all through a single laboratory test. WGS enables rapid resistance detection compared to traditional phenotypic susceptibility testing methods using the WHO TB mutation catalog, providing an insight into lesser-known mutations, which should be added to prediction databases as high-confidence mutations are recognized. The WGS-based methods can support TB elimination efforts in Canada and globally by ensuring the early start of appropriate treatment, rapidly limiting the spread of TB outbreaks.

A real-time PCR assay for rapid identification of inducible and acquired clarithromycin resistance in Mycobacterium abscessus.

BACKGROUND: Mycobacterium abscessus is a rapidly growing mycobacteria involved in severe infections of the lung, skin, or soft tissue. Macrolides such as clarithromycin are the recommended first line drugs for treatment of M. abscessus infections. However, M. abscessus has dual mechanisms of resistance to macrolides, making treatment by macrolides difficult. A functional erm(41) gene confers for inducible resistance while acquired mutations on the 23S rRNA rrl gene confer for constitutive resistance. METHODS: We have developed a real-time PCR assay to detect both inducible and acquired resistance to clarithromycin, and compared the results to traditional erm(41) and rrl sequencing and phenotypic susceptibility testing using Sensititre™ plates. RESULTS: Of the total 126 M. abscessus isolates tested, truncated erm(41) was found in 23/126 (18.3%) of the samples, 27/126 (21.4%) had a T28C mutation in erm(41), and 2/126 (1.6%) had an acquired A2058C mutation in rrl. The phenotypic results correlated with the expected sequencing results in 121/126 samples (96%). Phenotypic testing compared to real-time PCR resolved 2 of these discrepancies by showing the existence of both erm(41) alleles in the isolates that sequencing missed. One culture was found to be mixed with two M. abscessus subsp. as per hsp65 sequencing and 2 isolates had discordance between molecular and phenotypic results. It was presumed that 3 isolates showed discrepancy between sequencing and real-time PCR, but one culture was mixed and other 2 detected both alleles by real-time PCR leading to 100% concordance when compared to sequencing. CONCLUSION: In conclusion, real-time PCR is more accurate for detection of both acquired and induced clarithromycin resistance, specifically when mixed genic profiles are present in a sample.

A multilocus sequence typing scheme for Mycobacterium abscessus complex (MAB-multilocus sequence typing) using whole-genome sequencing data.

Background: Mycobacterium abscessus is a rapid growing nontuberculous mycobacteria (NTM) and a clinically significant pathogen capable of causing variable infections in humans that are difficult to treat and may require months of therapy/surgical interventions. Like other NTMs, M. abscessus can be associated with outbreaks leading to complex investigations and treatment of affected cases. Typing schemes for bacterial pathogens provide numerous applications; including identifying chain of transmission and tracking genomic evolution, are lacking or limited for many NTMs including M. abscessus. Methods: We extended the existing scheme from PubMLST using whole-genome data for M. abscessus by extracting data for 15 genetic regions within the M. abscessus genome. A total of 168 whole genomes and 11 gene sequences were used to build this scheme (MAB-multilocus sequence typing [MLST]). Results: All seven genes from the PubMLST scheme, namely argH, cya, gnd, murC, pta, purH, and rpoB, were expanded by 10, 14, 13, 10, 13, 10, and 9 alleles, respectively. Another eight novel genes were added including hsp 65, erm(41), arr, rrs, rrl, gyrA, gyrB, and recA with 16, 16, 25, 7, 32, 35, 29, and 15 alleles, respectively, with 85 unique sequence types identified among all isolates. Conclusion: MAB-MLST can provide identification of M. abscessus complex to the subspecies level based on three genes and can provide antimicrobial resistance susceptibility prediction based on results from seven genes. MAB-MLST generated a total of 85 STs, resulting in subtyping of 90 additional isolates that could not be genotyped using PubMLST and yielding results comparable to whole-genome sequencing (WGS). Implementation of a Galaxy-based data analysis tool, MAB-MLST, that simplifies the WGS data and yet maintains a high discriminatory index that can aid in deciphering an outbreak has vast applicability for routine diagnostics.