Bacteria resistant to multiple antibiotics are a growing threat for global public health and the economy. Multidrug-resistant tuberculosis (MDR-TB) and extensively drug-resistant TB (XDR-TB) are particular problems in many parts of the world including South Africa. Yet, little is known on the bacterial factors driving the global epidemics of MDR/XDR-TB. In the past, de novo acquisition of drug resistance determinants during patient treatment was considered the main driver of drug resistance in TB because the fitness costs associated with resistance were thought to make drug-resistant bacteria less transmissible. However, recent experimental and epidemiological data show that the fitness of drug-resistant Mycobacterium tuberculosis (Mtb) is heterogeneous and that a large proportion of MDR-TB is in fact transmitted. Preventing ongoing transmission of drug-resistant TB (DR-TB) is fundamental to controlling the epidemic. Understanding the interplay between drivers of de novo emergence and transmission of DR-TB, and the impact of treatment is key to developing new treatment regimens and ultimately improving control strategies.
Aim, Hypotheses and Objectives
We aim to use next-generation whole genome sequencing (WGS) to study the evolution and epidemiology of DR-TB in Khayelitsha, a high HIV, TB and DR-TB burden setting in South Africa. We have access to a detailed patient-level clinical database, maintained as part of an ongoing community-based DR-TB treatment programme initiated in Khayelitsha in 2008. We propose to link these data to a biobank of stored Mtb strains held at Stellenbosch University to test the following Hypotheses:
1. The transmission success of drug-resistant Mtb varies as a function of the specific combination of drug resistance-conferring mutations, compensatory mutations and strain genetic background;
2. The transmission success of drug-resistant Mtb further depends on the HIV status of the host;
3. The within-host evolution of Mtb populations (incl. resistance acquisition) differs in HIV-coinfected and HIV-negative TB patients;
4. The within-host evolution of Mtb populations also depends on the efficacy of the treatment regimen administered.
We will test these hypotheses by addressing the following objectives:
1. Use WGS to assess drug resistance evolution and transmission of DR-TB over time (10-years from 2008-2017) and in relation to level of drug resistance, HIV status and geographical location.
2. Assess ‘within-patient’ Mtb evolution among drug-resistant strains during treatment of individual patients, focussing on resistance acquisition for existing and new TB drugs, and contrasting HIV-negative and HIV-infected individuals.
3. To enhance capacity for bacterial population genomic analyses and genomic epidemiology in South Africa.
Secondary objectives include:
1. To assess the utility of WGS for rapid determination of drug resistance profiles and the potential for use of this data to inform individualisation of treatment.
2. To assess innate resistance to new and repurposed anti-TB drugs.
The study will be both retrospective, using clinical data and stored Mtb isolates from 2008-2015 and prospective, utilising data from the Khayelitsha programme from 2016-2017. Mtb isolates will be sub-cultured and bacterial DNA extracted in South Africa and sent to Switzerland for WGS. Data analyses will be performed in close collaboration between the teams in South Africa and Switzerland; regular visits in both directions will support these exchanges.
WGS will be conducted on an estimated 1,900 isolates from patients across the 10 years of the Khayelitsha cohort, using the Illumina technology. Sequence reads will be analysed using the bioinformatics pipeline developed at the Swiss Tropical and Public Health Institute (Swiss TPH) and concurrently with the bioinformatics pipeline utilised by Stellenbosch University (SUN). Phylogenomic and population genomic methods will be used to study the bacterial population structure and diversity of Mtb isolates across and within individual patients. To assess bacterial fitness in terms of between-patient transmissibility, we will collaborate with ETH Zurich to apply phylodynamic models to the sequence data.
A sub-sample of patient isolates will be selected to undergo quantitative phenotypic drug susceptibility testing to a range of key first-line and second-line anti-TB drugs. This will include the new drugs bedaquiline and delamanid, and the repurposed drug linezolid. This data will be used to assess phenotype/genotype relationships for existing, new and repurposed drugs, in addition to assessing the utility of using WGS-derived resistance profiles to individualise second-line treatment.
Training and capacity building
Capacity building is a key objective of this proposal. There is capacity for two South African PhD students within this study; one has already been identified. Both students will benefit from regular visits to Switzerland for training, and from visits from Swiss investigators to South Africa. Technology transfer will be facilitated through regular visits between Swiss and South African researchers and through the conduct of two 3-day training workshops to be held in Cape Town. These workshops will bring together researchers in the field of Mtb population genomics, genomic epidemiology and bioinformatics, and will be scheduled to coincide with visits from Swiss researchers. These workshops will be open to researchers from beyond the scope of the study, including other researchers from the University of Cape Town (UCT), SUN and the University of the Western Cape.
Through describing the role of Mtb strain diversity and evolution in the transmission of DR-TB in a high DR-TB setting, this study has the potential to dramatically improve our understanding of DR-TB epidemiology and the impact of current control strategies. In addition, as Khayelitsha is one of the pilot sites for expansion of access to new TB drugs, this study provides opportunity to study the impact of the use of these new drugs on resistance evolution. The data generated through this study also provides an opportunity to assess the feasibility and impact of using WGS for routine, rapid determination of TB drug resistance, in order to optimise and individualise second-line treatment. Finally, this study offers unique opportunities to build research capacity in bioinformatics, population genomics and genomic epidemiology of TB in South Africa. The study proposes expansion of an existing collaboration between researchers at UCT, SUN and Swiss TPH, and aims to bring a much wider circle of researchers in Cape Town together to share research and collaborate further. We aim to train and support two promising doctoral students and through them build capacity in genomic epidemiology and bioinformatics in Cape Town.