Overcoming the Challenges of Pyrazinamide Susceptibility Testing in Clinical Mycobacterium tuberculosis Isolates

Evaluating the efficacy of whole genome sequencing in predicting susceptibility profiles for first-line antituberculosis drugs

OBJECTIVES

This study aimed to examine the efficacy of whole genome sequencing (WGS) in accurately predicting susceptibility profiles, potentially eliminating the need for conventional phenotypic drug susceptibility testing (pDST) for first-line antituberculosis drugs in routine tuberculosis diagnosis.

METHODS

Over the period of 2017 to 2020, 1114 Mycobacterium tuberculosis complex isolates were collected with drug susceptibility testing conducted using the MGIT960 system and WGS performed for predicting drug resistance profiles. In addition, we implemented a new algorithm with an updated WGS workflow, omitting pan-susceptible strains from pDST.

RESULTS

Results showed that out of 1075 analysed isolates, WGS-based genotypic sensitivity predictions for isoniazid, rifampicin, ethambutol, and pyrazinamide were 100% (95% CI, 99.6-100%), 100% (95% CI, 99.62-100%), 99.8% (95% CI, 99.26-99.94%), and 100% (95% CI, 99.63-100%), respectively. In contrast, the WGS-based genotypic resistance prediction, was 98.85% (95% CI, 93.77-99.79%) for isoniazid, 94.74% (95% CI, 82.71-98.54%) for rifampicin, 86.96% (95% CI, 67.87-95.46%) for ethambutol, and 75.7% (95% CI, 59.9-86.63%) for pyrazinamide. Moreover, WGS enabled the implementation of a new testing algorithm that made it unnecessary to perform pDST in 954 of all 1075 samples (88.7%) and in 890 of 901 pan-susceptible samples (98.8%).

DISCUSSION

Integrating WGS into tuberculosis management offers significant potential to replace phenotypic drug susceptibility testing, especially for problematic drugs like pyrazinamide and ethambutol, potentially improving treatment outcomes.

Applying whole genome sequencing to predict phenotypic drug resistance in Mycobacterium tuberculosis: leveraging 20 years of nationwide data from Denmark

Infection with Mycobacterium tuberculosis remains one of the biggest causes of death from a single microorganism worldwide, and the continuous emergence of drug resistance aggravates our ability to cure the disease. New improved resistance detection methods are needed to provide adequate treatment, such as whole genome sequencing (WGS), which has been used increasingly to identify resistance-conferring mutations over the last decade. The steadily increasing knowledge of resistance-conferring mutations increases our ability to predict resistance based on genomic data alone. This study evaluates the performance of WGS to predict M. tuberculosis complex resistance. It compares WGS predictions with the phenotypic (culture-based) drug susceptibility results based on 20 years of nationwide Danish data. Analyzing 6,230 WGS-sequenced samples, the sensitivities for isoniazid, rifampicin, ethambutol, and pyrazinamide were 82.5% [78.0%-86.5%, 95% confidence interval (CI)], 97.3% (90.6%-99.7%, 95% CI), 58.0% (43.2%-71.8%, 95% CI), and 60.5% (49.0%-71.2%, 95% CI), respectively, and specificities were 99.8% (99.7%-99.9%, 95% CI), 99.8% (99.7%-99.9%, 95% CI), 99.4% (99.2%-99.6%, 95% CI), and 99.9% (99.7%-99.9%, 95% CI), respectively. A broader range of both sensitivities and specificities was observed for second-line drugs. The results conform with previously reported values and indicate that WGS is reliable for routine resistance detection in resource-rich tuberculosis low-incidence and low-resistance settings such as Denmark.

Genomic and phenotypic characterization of Mycobacterium tuberculosis' closest-related non-tuberculous mycobacteria

Four species of non-tuberculous mycobacteria (NTM) rated as biosafety level 1 or 2 (BSL-1/BSL-2) organisms and showing higher genomic similarity with Mycobacterium tuberculosis (Mtb) than previous comparator species Mycobacterium kansasii and Mycobacterium marinum were subjected to genomic and phenotypic characterization. These species named Mycobacterium decipiens, Mycobacterium lacus, Mycobacterium riyadhense, and Mycobacterium shinjukuense might represent "missing links" between low-virulent mycobacterial opportunists and the highly virulent obligate pathogen Mtb. We confirmed that M. decipiens is the closest NTM species to Mtb currently known and found that it has an optimal growth temperature of 32°C-35°C and not 37°C. M. decipiens showed resistance to rifampicin, isoniazid, and ethambutol, whereas M. lacus and M. riyadhense showed resistance to isoniazid and ethambutol. M. shinjukuense was sensitive to all three first-line TB drugs, and all four species were sensitive to bedaquiline, a third-generation anti-TB drug. Our results suggest these four NTM may be useful models for the identification and study of new anti-TB molecules, facilitated by their culture under non-BSL-3 conditions as compared to Mtb. M. riyadhense was the most virulent of the four species in cellular and mouse infection models. M. decipiens also multiplied in THP-1 cells at 35°C but was growth impaired at 37°C. Genomic comparisons showed that the espACD locus, essential for the secretion of ESX-1 proteins in Mtb, was present only in M. decipiens, which was able to secrete ESAT-6 and CFP-10, whereas secretion of these antigens varied in the other species, making the four species interesting examples for studying ESX-1 secretion mechanisms.IMPORTANCEIn this work, we investigated recently identified opportunistic mycobacterial pathogens that are genomically more closely related to Mycobacterium tuberculosis (Mtb) than previously used comparator species Mycobacterium kansasii and Mycobacterium marinum. We confirmed that Mycobacterium decipiens is the currently closest known species to the tubercle bacilli, represented by Mycobacterium canettii and Mtb strains. Surprisingly, the reference strain of Mycobacterium riyadhense (DSM 45176), which was purchased as a biosafety level 1 (BSL-1)-rated organism, was the most virulent of the four species in the tested cellular and mouse infection models, suggesting that a BSL-2 rating might be more appropriate for this strain than the current BSL-1 rating. Our work establishes the four NTM species as interesting study models to obtain new insights into the evolutionary mechanisms and phenotypic particularities of mycobacterial pathogens that likely have also impacted the evolution of the key pathogen Mtb.

Proficiency of phenotypic drug susceptibility testing for Mycobacterium tuberculosis in China, 2008-2021

To analyze the results of proficiency testing for anti-tuberculosis drug susceptibility testing (DST) in China. Number of laboratory participating the proficiency testing performed DST, and the sensitivity, specificity, reproducibility, and accordance rate were calculated from data of 13 rounds proficiency testing results for DST from 2008 to 2021. A total of 30 and 20 strains of Mycobacterium tuberculosis with known susceptibility results were sent to each laboratory in 2008 to 2019, 2020 and 2021, respectively. The number of participating laboratories ranged from 30 in 2009 to 546 in 2021. L-J DST was the predominant method. The specificity presented relatively higher than sensitivity. Improvement of specificity were observed for all drugs through the years, while sensitivity did not show improvement for amikacin and capreomycin. Accordance rate of pyrazinamide and kanamycin and reproducibility of capreomycin and pyrazinamide were not significantly improved through the years. Most of the participating laboratories significantly improved the quality of their DST through the consecutive rounds of proficiency testing except for second-line injectable drugs and pyrazinamide. The results highlight the importance of developing novel and/or improving existing methods for phenotypic DST for certain drugs.

Emerging insight of whole genome sequencing coupled with protein structure prediction into the pyrazinamide-resistance signature of Mycobacterium tuberculosis

Pyrazinamide (PZA) is considered to be a pivotal drug to shorten the treatment of both drug-susceptible and drug-resistant tuberculosis, but its use is challenged by the reliability of drug-susceptibility testing (DST). PZA resistance in Mycobacterium tuberculosis (MTB) is relevant to the amino acid substitution of pyrazinamidase that is responsible for the conversion of PZA to active pyrazinoic acid (POA). The single nucleotide variants (SNVs) within ribosomal protein S1 (rpsA) or aspartate decarboxylase (panD), the binding targets of POA, has been reported to drive the PZA-resistance signature of MTB. In this study, whole genome sequencing (WGS) was used to identify SNVs within the pncA, rpsA and panD genes in 100 clinical MTB isolates associated with DST results for PZA. The potential influence of high-confidence, interim-confidence or emerging variants on the interplay between target genes and PZA or POA was simulated computationally, and predicted with a protein structure modelling approach. The DST results showed weak agreement with the identification of high-confidence variants within the pncA gene (Cohen's kappa coefficient=0.58), the analytic results of WGS coupled with protein structure modelling on pncA mutants (Cohen's kappa coefficient=0.524) or related genes (Cohen's kappa coefficient=0.504). Taken together, these results suggest the practicable application of a genotypic-coupled bioinformatic approach to manage PZA-containing regimens for patients with MTB.

Co-resistance to isoniazid and second-line anti-tuberculosis drugs in isoniazid-resistant tuberculosis at a tertiary care hospital in Thailand

Isoniazid-resistant tuberculosis (Hr-TB) is an important drug-resistant tuberculosis (TB). In addition to rifampicin, resistance to other medications for Hr-TB can impact the course of treatment; however, there are currently limited data in the literature. In this study, the drug susceptibility profiles of Hr-TB treatment and resistance-conferring mutations were investigated for Hr-TB clinical isolates from Thailand. Phenotypic drug susceptibility testing (pDST) and genotypic drug susceptibility testing (gDST) were retrospectively and prospectively investigated using the Mycobacterium Growth Indicator Tube (MGIT), the broth microdilution (BMD) method, and whole-genome sequencing (WGS)-based gDST. The prevalence of Hr-TB cases was 11.2% among patients with TB. Most Hr-TB cases (89.5%) were newly diagnosed patients with TB. In the pDST analysis, approximately 55.6% (60/108) of the tested Hr-TB clinical isolates exhibited high-level isoniazid resistance. In addition, the Hr-TB clinical isolates presented co-resistance to ethambutol (3/161, 1.9%), levofloxacin (2/96, 2.1%), and pyrazinamide (24/118, 20.3%). In 56 Hr-TB clinical isolates, WGS-based gDST predicted resistance to isoniazid [katG S315T (48.2%) and fabG1 c-15t (26.8%)], rifampicin [rpoB L430P and rpoB L452P (5.4%)], and fluoroquinolones [gyrA D94G (1.8%)], but no mutation for ethambutol was detected. The categorical agreement for the detection of resistance to isoniazid, rifampicin, ethambutol, and levofloxacin between WGS-based gDST and the MGIT or the BMD method ranged from 80.4% to 98.2% or 82.1% to 100%, respectively. pDST and gDST demonstrated a low co-resistance rate between isoniazid and second-line TB drugs in Hr-TB clinical isolates.

IMPORTANCE

The prevalence of isoniazid-resistant tuberculosis (Hr-TB) is the highest among other types of drug-resistant tuberculosis. Currently, the World Health Organization (WHO) guidelines recommend the treatment of Hr-TB with rifampicin, ethambutol, pyrazinamide, and levofloxacin for 6 months. The susceptibility profiles of Hr-TB clinical isolates, especially when they are co-resistant to second-line drugs, are critical in the selection of the appropriate treatment regimen to prevent treatment failure. This study highlights the susceptibility profiles of the WHO-recommended treatment regimen in Hr-TB clinical isolates from a tertiary care hospital in Thailand and the concordance and importance of using the phenotypic drug susceptibility testing or genotypic drug susceptibility testing for accurate and comprehensive interpretation of results.

Investigation of genomic mutations and their association with phenotypic resistance to new and repurposed drugs in Mycobacterium tuberculosis complex clinical isolates

BACKGROUND

WGS has the potential to detect resistance-associated mutations and guide treatment of MDR TB. However, the knowledge base to confidently interpret mutations associated with the new and repurposed drugs is sparse, and phenotypic drug susceptibility testing is required to detect resistance.

METHODS

We screened 900 Mycobacterium tuberculosis complex genomes from Ireland, a low TB incidence country, for mutations in 13 candidate genes and assessed their association with phenotypic resistance to bedaquiline, clofazimine, linezolid, delamanid and pretomanid.

RESULTS

We identified a large diversity of mutations in the candidate genes of 195 clinical isolates, with very few isolates associated with phenotypic resistance to bedaquiline (n = 4), delamanid (n = 4) and pretomanid (n = 2). We identified bedaquiline resistance among two drug-susceptible TB isolates that harboured mutations in Rv0678. Bedaquiline resistance was also identified in two MDR-TB isolates harbouring Met146Thr in Rv0678, which dated back to 2007, prior to the introduction of bedaquiline. High-level delamanid resistance was observed in two isolates with deletions in ddn, which were also resistant to pretomanid. Delamanid resistance was detected in two further isolates that harboured mutations in fbiA, but did not show cross-resistance to pretomanid. All isolates were susceptible to linezolid and clofazimine, and no mutations found were associated with resistance.

CONCLUSIONS

More studies that correlate genotypic and phenotypic drug susceptibility data are needed to increase the knowledge base of mutations associated with resistance, in particular for pretomanid. Overall, this study contributes to the development of future mutation catalogues for M. tuberculosis complex isolates.

Implementation of national whole-genome sequencing of Mycobacterium tuberculosis, National Public Health Laboratory, Singapore, 2019-2022

The National Tuberculosis Programme (NTBP) monitors the occurrence and spread of tuberculosis (TB) and multidrug-resistant TB (MDR-TB) in Singapore. Since 2020, whole-genome sequencing (WGS) of Mycobacterium tuberculosis isolates has been performed at the National Public Health Laboratory (NPHL) for genomic surveillance, replacing spoligotyping and mycobacterial interspersed repetitive unit-variable number tandem repeats analysis (MIRU-VNTR). Four thousand three hundred and seven samples were sequenced from 2014 to January 2023, initially as research projects and later developed into a comprehensive public health surveillance programme. Currently, all newly diagnosed culture-positive cases of TB in Singapore are prospectively sent for WGS, which is used to perform lineage classification, predict drug resistance profiles and infer genetic relationships between TB isolates. This paper describes NPHL's operational and technical experiences with implementing the WGS service in an urban TB-endemic setting, focusing on cluster detection and genomic drug susceptibility testing (DST). Cluster detection: WGS has been used to guide contact tracing by detecting clusters and discovering unknown transmission networks. Examples have been clusters in a daycare centre, housing apartment blocks and a horse-racing betting centre. Genomic DST: genomic DST prediction (gDST) identifies mutations in core genes known to be associated with TB drug resistance catalogued in the TBProfiler drug resistance mutation database. Mutations are reported with confidence scores according to a standardized approach referencing NPHL's internal gDST confidence database, which is adapted from the World Health Organization (WHO) TB drug mutation catalogue. Phenotypic-genomic concordance was observed for the first-line drugs ranging from 2959/2998 (98.7 %) (ethambutol) to 2983/2996 (99.6 %) (rifampicin). Aspects of internal database management, reporting standards and caveats in results interpretation are discussed.

Global status of phenotypic pyrazinamide resistance in Mycobacterium tuberculosis clinical isolates: an updated systematic review and meta-analysis

Pyrazinamide (PZA) is an essential first-line tuberculosis drug for its unique mechanism of action active against multidrug-resistant-TB (MDR-TB). Thus, the aim of updated meta-analysis was to estimate the PZA weighted pooled resistance (WPR) rate in M. tuberculosis isolates based on publication date and WHO regions. We systematically searched the related reports in PubMed, Scopus, and Embase (from January 2015 to July 2022). Statistical analyses were performed using STATA software. The 115 final reports in the analysis investigated phenotypic PZA resistance data. The WPR of PZA was 57% (95% CI 48-65%) in MDR-TB cases. According to the WHO regions, the higher WPRs of PZA were reported in the Western Pacific (32%; 95% CI 18-46%), South East Asian region (37%; 95% CI 31-43%), and the Eastern Mediterranean (78%; 95% CI 54-95%) among any-TB patients, high risk of MDR-TB patients, and MDR-TB patients, respectively. A negligible increase in the rate of PZA resistance were showed in MDR-TB cases (55% to 58%). The rate of PZA resistance has been rising in recent years among MDR-TB cases, underlines the essential for both standard and novel drug regimens development.

Mycobacterium tuberculosis complex whole-genome sequencing in New York State: Implementation of a reduced phenotypic drug susceptibility testing algorithm

Whole-genome sequencing (WGS) can predict drug resistance and antimicrobial susceptibility in Mycobacterium tuberculosis complex (MTBC) and has shown promise in partially replacing culture-based phenotypic drug susceptibility testing (pDST). We performed a two-year side by side study comparing the prediction of drug resistance and antimicrobial susceptibility by WGS molecular DST (mDST) to pDST to determine resistance at the critical concentration by Mycobacterial Growth Indicator Tube (MGIT) and agar proportion testing. Negative predictive values of WGS results were consistently high for the first-line drugs: rifampin (99.9%), isoniazid (99.0%), pyrazinamide (98.5%), and ethambutol (99.8%); the rates of resistance to these drugs, among strains in our population, are 2.9%, 10.4%, 46.3%, and 2.3%, respectively. WGS results were available an average 8 days earlier than first-line MGIT pDST. Based on these findings, we implemented a new testing algorithm with an updated WGS workflow in which strains predicted pan-susceptible were no longer tested by pDST. This algorithm was applied to 1177 isolates between October 2018 and September 2020, eliminating pDST for 66.6% of samples and reducing pDST for an additional 22.0%. This algorithm change resulted in faster turnaround times and decreased cost while maintaining comprehensive antimicrobial susceptibility profiles of all culture-positive MTBC cases in New York.

pncA Large Deletion is the Characteristic of Pyrazinamide-Resistant Mycobacterium tuberculosis belonging to the East Asian Lineage

BACKGROUND

Pyrazinamide (PZA) is often used as an add-on agent in the treatment of multidrug-resistant tuberculosis, regardless of phenotypic drug susceptibility testing (pDST) results. However, evaluating the effectiveness of PZA is challenging because of its low pH activity, which can result in unreliable pDST results. This study aimed to investigate the genomic characteristics associated with PZA resistance that can be used to develop genotypic DST.

MATERIALS AND METHODS

A publicly available whole genome sequencing (WGS) dataset of 10,725 Mycobacterium tuberculosis complex genomes (3,326 phenotypically PZA-resistant and 7,399 phenotypically PZA-susceptible isolates) were analyzed.

RESULTS

In total, 2,934 pncA non-silent mutations were identified in 2,880 isolates (26.9%). Detected mutations were found throughout the entire coding region of pncA in a scattered pattern, of which the most frequent mutation was p.Q10P (n = 278), followed by p.H57D (n = 167) and c.-11A>G (n = 122). The sensitivity and specificity of the group 1 or 2 mutations reported by the World Health Organization (WHO) mutational catalogue were 73.0% and 98.9%, respectively. We further identified 18 novel pncA mutations that were significantly associated with phenotypically PZA-resistant. In addition to these mutations, we identified 102 large deletions in the pncA gene, and all but two isolates were phenotypically resistant to PZA isolates. Notably, pncA deletions were mutually exclusive to pncA mutations, and more than half of the isolates with pncA large deletions belonged to the East Asian lineage (67.6%). The sensitivity, specificity, positive predictive value, and negative predictive value of the pooled variants (group 1 or 2 mutations, novel resistance-associated mutations, and large deletions of the pncA gene) were 79.0%, 98.9%, 97.0%, and 91.3%, respectively. The area under the curve (AUC) value for the pooled variants was significantly higher than the AUC value for the group 1 or 2 mutations (P <0.001), indicating that the pooled variants have a better discriminative ability for predicting PZA resistance.

CONCLUSION

Using WGS, we found that the pncA mutations are scattered without specific mutational hotspots, and large deletions associated with PZA resistance are more common in the East Asian lineage of M. tuberculosis isolates. Our data also demonstrated the reliability of group 1 or 2 mutations presented in the WHO mutation catalogue and the need for further investigation on group 3 mutations, contributing to the evaluation of the current knowledge base on mutations associated with the PZA-resistant M. tuberculosis complex.

Synthesis, cytotoxicity, and biomacromolecule binding: Three isomers of nitrosylruthenium complexes with bidentate bioactive molecules as co-ligands

Three isomeric nitrosylruthenium complexes [RuNO(Qn)(PZA)Cl] (P1, P2, and P3) with bioactive small molecules 8-hydroxyquinoline (Qn) and pyrazinamide (PZA) as co-ligands were synthesized, and their crystal structures were determined using X-ray diffraction technique. The cellular toxicity of the isomeric complexes was compared to understand the effects of the geometries on the biological activity of the complexes. Both the complexes and the human serum albumin (HSA) complex adducts affected the extent of proliferation of HeLa cells (IC₅₀: 0.77-1.45 μM). P2 showed prominent activity-induced cell apoptosis and arrested cell cycles at the G1 phase. The binding constants (K_b) of the complex with calf thymus DNA (CT-DNA) and HSA were quantitatively evaluated using fluorescence spectroscopy in the range of 0.17-1.56 × 10⁴ M^-1 and 0.88-3.21 × 10⁵ M^-1, respectively. The average binding site (n) number was close to 1. Moreover, the structure of HSA and the P2 complex adduct solved at the resolution of 2.48 Å revealed that one PZA-coordinated nitrosylruthenium complex bound at the subdomain I of HSA via a noncoordinative bond. HSA could serve as a potential nano-delivery system. This study provides a framework for the rational design of metal-based drugs.

Analysis of Factors Influencing Multidrug-Resistant Tuberculosis and Validation of Whole-Genome Sequencing in Children with Drug-Resistant Tuberculosis

Objective

Pediatric tuberculosis (TB) is one of the top ten causes of death in children. Our study was to analyze influencing factors of multidrug-resistant tuberculosis (MDR-TB) and validation of whole-genome sequencing (WGS) used in children with drug-resistant TB (DR-TB).

Methods

All Mycobacterium tuberculosis (Mtb) strains were isolated from patients aged below 18 years old of Children’s Hospital of Chongqing Medical University, China. A total of 208 Mtb isolates were tested for eight anti-TB drugs with phenotypic drug susceptibility test (DST) and for genetic prediction of the susceptible profile with WGS. The patients corresponding to each strain were grouped according to drug resistance and genotype. Influencing factors of MDR-TB and DR-TB were analyzed.

Results

According to the phenotypic DST and WGS, 82.2% of Mtb strains were susceptible to all eight drugs, and 6.3% were MDR-TB. Using the phenotypic DSTs as the gold standard, the kappa value of WGS to predict isoniazid, rifampin, ethambutol, rifapentine, prothionamide, levofloxacin, moxifloxacin and amikacin was 0.84, 0.89, 0.59, 0.86, 0.89, 0.82, 0.88 and 1.00, respectively. There was significant difference in the distribution of severe TB, diagnosis, treatment and outcome between MDR and drug-susceptible group (P<0.05). The distribution of severe TB and treatment between DR and drug-susceptible group was statistically different (P<0.05). The results of binary logistic regression showed that Calmette–Guérin bacillus (BCG) vaccine is the protective factor for MDR-TB (OR=0.19), and MDR-TB is the risk factor for PTB and EPTB (OR=17.98).

Conclusion

The BCG vaccine is a protective factor for MDR-TB, and MDR-TB might not be confined to pulmonary infection, spreading to extrapulmonary organs in children. MDR-TB had more severe cases and a lower recovery rate than drug-susceptible TB. WGS could provide an accurate prediction of drug susceptibility test results for anti-TB drugs, which are needed for the diagnosis and precise treatment of TB in children.

Activity of Pyrazinamide against Mycobacterium tuberculosis at Neutral pH in PZA-S1 Minimal Medium

Susceptibility testing of tuberculosis (TB) drugs on Mycobacterium tuberculosis is essential for the rapid detection of strains resistant to the drugs, providing the patient with effective treatment, and preventing the spread of drug-resistant TB strains. Pyrazinamide (PZA) is one of the first-line agents used for the treatment of TB. However, current phenotypic PZA susceptibility testing is unreliable due to its performance in acidic pH conditions. The aims of this study were to develop minimal media to determine the activity of PZA at a neutral pH at 37 °C to avoid problems caused by an acidic pH, which is currently used in PZA susceptibility tests, and to identify PZA-resistant M. tuberculosis in media with reproducibility and accuracy. Different minimal media were used to determine the activity of PZA using the broth microdilution method with M. tuberculosis H37Ra as the reference strain. The PZA-S1 minimal medium was proposed as the most suitable medium. PZA inhibited the growth of M. tuberculosis in PZA-S1 at a neutral pH of 6.8, which is the optimal pH for M. tuberculosis growth. Moreover, PZA showed activity at a neutral pH on a PZA-S1 agar plate when using the disk diffusion method. PZA-resistant M. tuberculosis could be identified at a neutral pH in PZA-S1 minimal medium. This study establishes valuable information regarding the testing of PZA's susceptibility in relation to M. tuberculosis at a neutral pH of 6.8 with reliability and accuracy in clinical settings.