TL;DR
This study reveals that traditional methods for tracking tuberculosis (TB) transmission can overestimate recent outbreaks in high-burden settings. By using whole-genome sequencing (WGS), researchers found that a cluster identified by older methods had actually been circulating for decades, highlighting the need for WGS to accurately understand TB epidemiology and inform control strategies.
What Problem Does This Paper Address?
Tracking the spread of infectious diseases like tuberculosis is crucial for effective public health interventions. Traditional methods, such as MIRU-VNTR typing, are often used to group similar bacterial strains and infer transmission events, forming "clusters." However, the utility and interpretation of these clusters in high-burden settings, where TB is endemic and transmission is continuous, remain unclear. Specifically, it's uncertain whether these clusters represent recent outbreaks or long-standing endemic strains. This paper aims to clarify this by applying whole-genome sequencing (WGS) to a large TB cluster in Lima, Peru, a high-prevalence setting, and comparing its findings to traditional methods.
Methods
The researchers performed whole-genome sequencing on 61 isolates from a larger cluster of 148 strains identified by MIRU-VNTR typing. They then compared transmission link inference using both epidemiological data and genomic data, including controversial variants. A time-calibrated phylogenetic tree was generated to estimate the emergence dates of the cluster and drug resistance acquisition events. To validate their findings, they analyzed genomic data from a TB outbreak in London and a larger dataset of public Mycobacterium tuberculosis (Mtb) genomes. They also investigated bacterial factors potentially contributing to transmission fitness.
Datasets
The study utilized:
- 61 Mycobacterium tuberculosis isolates from a large MIRU-VNTR cluster in Lima, Peru.
- Genomic data from 148 strains initially identified by MIRU-VNTR typing.
- Genomic data from 325 TB cases from an outbreak in London for validation.
- A dataset of 12,032 public Mtb genomes to identify bacterial factors under positive selection.
Key Insights and Ideas
- MIRU-VNTR clusters can mask long-term endemicity: What appears as a recent outbreak via MIRU-VNTR typing can, with WGS, be revealed as a lineage that has been circulating for many decades. This challenges the assumption that clusters directly equate to recent transmission events.
- Genomic regions matter: Excluding certain genomic regions (like PE/PPE genes) from WGS analysis can lead to an underestimation of genetic diversity and potentially misinterpret transmission links. Including these regions, along with indels, provides a more granular and accurate picture.
- Drug resistance evolution is complex: Resistance to drugs like rifampicin and pyrazinamide can emerge multiple times independently within a single cluster, suggesting different evolutionary pathways and acquisition events.
- Host and bacterial factors intertwine: Transmission success is not solely due to bacterial evolution; host factors (like age and substance use) also play a significant role, interacting with bacterial traits to influence spread.
Key Results
- The 57 most genetically homogeneous isolates within the cluster diverged around 1968 (95% HPD: 1945-1985), indicating decades of circulation rather than a recent outbreak.
- Rifampicin resistance emerged at least three times independently within the cluster.
- The study identified five cluster-defining single nucleotide polymorphisms (SNPs), including one in the esxV gene, that may contribute to transmissibility.
- In a London outbreak dataset, including PE/PPE variants and indels reduced the number of inferred transmission links by 11%, suggesting these regions are important for accurate inference.
- Five patients were identified as having a particularly high "propensity to propagate" (PTP), linked to being younger, male, smear-positive, and having a history of substance use.
Limitations and Future Work
- Sample size: While the study analyzed 61 isolates, the authors acknowledge that sequencing a subset of isolates might have led to missed links and an underestimation of diversity.
- Contact tracing limitations: Contact tracing was primarily limited to households, potentially missing community-based transmission events.
- Dating estimates: The dating estimates rely on previously reported molecular clock rates, which can have inherent uncertainties.
The authors suggest that WGS, especially when incorporating PE/PPE regions and indels, offers significantly higher resolution for understanding TB transmission in high-burden settings. Future work could focus on integrating WGS data more seamlessly with epidemiological investigations and exploring the identified bacterial factors for potential intervention targets.
Summary updated