Clinical and public health utility of Mycobacterium tuberculosis whole genome sequencing

Abstract

The World Health Organization (WHO) estimates that around 10 million people develop tuberculosis (TB) every year, with 1.5 million deaths attributed to TB in 2019 (World Health Organization, 2020). The majority of the disease burden occurs in low-income countries, where access to diagnostics and tailored treatment remains problematic. The current COVID-19 pandemic further threatens to impact global TB control by diverting resources, reducing notifications and hence significantly increasing deaths attributable to TB (World Health Organization, 2020).

Whole genome sequencing (WGS) is becoming increasingly accessible, and has particular value in the diagnosis and management of TB disease (Cabibbe et al., 2018; Meehan et al., 2019). Not only does it have the potential to give more rapid and complete information on drug-resistance, but the high discriminatory power it offers allows detection of clusters and transmission pathways, as well as likely contamination events, mixed infections and to differentiate between re-infection and relapse with much greater confidence than previous typing methods.

The basic biology of Mycobacterium tuberculosis, with its lack of horizontal gene transfer, simplifies bioinformatics analyses and makes it an ideal organism to establish ‘proof of principle’ for the clinical and public health value of WGS. It has been adopted as an integral part of the public health strategy in several low-burden, high-income countries with the benefits previously described (Gurjav et al., 2016; Shea et al., 2017; Olaru et al., 2018). Routine WGS results will also facilitate timely cross-jurisdictional communication and data-sharing in the management of outbreaks that cross national or international borders. Initially the costs required for sequencing infrastructure, maintenance and expertise were high, but this has been considerably reduced over the last two decades. Big challenges remain however in the availability and funding of adequately trained bioinformaticians, while the cost and sophistication of the technology, as well as the large number of tuberculosis patients, limits its use in most high TB incidence countries with limited resources. The potential benefits and challenges of using WGS in these different settings are summarised in Table 1.

Table 1. Benefits and challenges of WGS in TB diagnostics.

WGS considerations	All settings	Specific priorities/challenges in high TB incidence settings
Benefits	Public health	High TB incidence setting
	Drug resistance surveillance Outbreak/cluster recognition Identifying unknown cases in the transmission chain Data-sharing/cross-jurisdictional communication	Routine WGS in every single patient not feasible: Focus on drug resistance surveillance to inform optimal programmatic treatment options Detailed transmission tracking in particular ‘hot spot’ areas Monitor for resistance against newer drugs, such as bedaquiline, in MDR- TB patients Monitor laboratory contamination events to identify potential ‘systemic issues’
	Individual patient
	Drug resistance prediction Differentiating relapse from re-infection Identifying mixed infections Identifying laboratory contamination events
Challenges	Costs – infrastructure, maintenance Expertise Quality control/assurance Current requirement for culture Comprehensive genotype-phenotype databases Data handling & security Data handling & security Data interoperability

TB − tuberculosis; WGS − whole genome sequencing; MDR-TB − multidrug resistant TB.

Drug-resistant TB poses a major global public health threat, since it is associated with greater morbidity and mortality, and potentially drug resistance amplification and increased transmission if management is sub-optimal. It requires prolonged treatment with expensive, difficult to access and potentially more toxic second-line drugs. In 2019 there was an estimated half a million new cases of rifampicin-resistant TB; 78% of which had multidrug-resistant TB (MDR-TB) (World Health Organization., 2020). Globally, 3.3% of new cases and 17.7% of previously treated cases had drug-resistant TB, indicating active transmission of these strains within communities (World Health Organization., 2020). Large outbreaks of drug resistant strains that have gone undetected for decades when smear microscopy was the only available diagnostic tool have been well characterised with the advent of WGS (Bainomugisa et al., 2018). In 2016 it was estimated that 6.2% of MDR-TB cases had additional drug-resistance causing extensively drug-resistant TB (World Health Organization., 2017), which is even more difficult to treat.

The most commonly used microbiological methods for detecting antimicrobial resistance involve phenotypic drug susceptibility testing (DST) using liquid broth culture, mostly using the Mycobacteria Growth Indicator Tube (MGIT) system. While this has been considered the gold-standard in TB antimicrobial resistance diagnostics, it still takes a long time to obtain results as M. tuberculosis is a slow growing organism. Moreover, the expertise and infrastructure required is often in excess of what is available. With access to routine WGS a complete resistome may be available within days of the initial positive culture. While several studies have evaluated the correlation between certain resistance polymorphisms and phenotypic resistance (Rodwell et al., 2014; Walker et al., 2015; Miotto et al., 2017; CRyPTIC Consortium, 2018), to date this has been most complete and useful for first-line drugs. Continued research in this area will aid in improving resistance prediction and creating comprehensive genome-based drug susceptibility profiling against first- and second-line drugs.

In addition, several studies have shown accurate genomic DST prediction when sequencing directly from patient specimens, using enhanced methods for human nucleic acid depletion and target amplification, which has the potential to dramatically reduce turn-around times and the requirement for expensive biosafety facilities to first grow these drug-resistant TB strains (Brown et al., 2015; Votintseva et al., 2017; Doyle et al., 2018). WGS may also allow the detection of novel drug resistance mechanisms when performed in conjunction with phenotypic DST, and help to determine whether resistance is due to de novo acquisition or clonal spread, as well as inform decisions about which resistance mutations should be included on more accessible extended molecular panels (Cohen et al., 2019). Ultimately, accurate and prompt diagnosis of MDR-TB allows more effective TB control through more timely initiation of appropriate treatment of individual patients and institution of the necessary public health measures to limit its spread.

The superior resolution of WGS compared to conventional typing methods, such as mycobacterial interspersed repetitive units-variable number tandem repeats (MIRU-VNTR), provides several key advantages. Firstly, it has given us deeper insight into the transmission dynamics of TB within communities, facilitating the detection of clusters and probable transmission pathways with much greater confidence; especially in strains that are poorly differentiated by MIRU-VNTR such as the Beijing lineage (Walker et al., 2013; Stucki et al., 2016). The implementation of routine WGS will allow detection of unsuspected transmission events, as well as highly contagious patients in order to guide improved public health containment measures. Secondly, the greater discriminatory power of WGS may allow for the distinction between cases of endogenous relapse from exogenous reinfection (Bryant et al., 2013; Witney et al., 2017), which can provide invaluable information when evaluating treatment adherence and the efficacy of current drug regimens. Thirdly, WGS has improved the ability to detect mixed infections with several different strains of M. tuberculosis (Tarashi et al., 2017; Wyllie et al., 2018). Mixed infections may involve strains with different susceptibility patterns, which has important implications for treatment selection and therapeutic outcomes. And lastly, the resolution offered by WGS has enabled the detection of probable laboratory contamination events with much greater precision than previous typing methods (Min et al., 2019; Wu et al., 2019; Kizny Gordon et al., 2021). This has major consequences for individual patients by preventing unnecessary exposure to prolonged courses of toxic drugs, and also has value in highlighting possible failures in institutional infection control or laboratory processes.

Despite the challenges of 2020, global travel and migration are likely to return to pre-COVID rates in the future. While cluster detection and contact tracing have been less of a priority than diagnosis and treatment in high burden countries due to constrained resources, with access to WGS these public health measures may be better targeted and more feasible. Although it remains problematic, collecting genomic data in a standardised format will allow easier data exchange and comparison across jurisdictions, facilitating early and accurate outbreak detection or even recognition of laboratory contamination events (Kizny Gordon et al., 2021; Meehan et al., 2019; Walker et al., 2018). This has important implications for an improved understanding of M. tuberculosis transmission dynamics providing better guidance for local and ultimate global TB control. Although developed countries are ‘leading the way’ in integrating routine WGS into clinical and public health practice, it is important to consider strategies that would extend the benefits of routine WGS to high TB incidence settings as well.

The development of effective techniques for culture-independent sequencing would dramatically change the landscape of TB diagnostics, and targeted next generation sequencing focusing on panels of genes or genetic regions may be an alternative solution when sequencing directly from specimens (Cabibbe et al., 2018; Cohen et al., 2019). The low abundance of target M. tuberculosis genomes in respiratory microbiomes from patients without advance pulmonary disease presents a major challenge for the evaluation and translation of ‘deep’ sequencing into diagnostic practice. Perhaps WGS greatest utility in high burden settings will be in antimicrobial resistance surveillance to inform appropriate empiric treatment regimens when resistance prevalence is low and to guide tailored therapy when resistance is more prevalent, as well as guiding broader public health measures. We are experiencing a ‘genomics revolution’ in microbiology with major potential benefits for TB patients and TB control programs, but we need to continue to invest in and advance these technologies and ensure that high TB incidence countries also share in the benefits as the technology becomes more mature and cost efficient.