Applications of whole genome sequencing to TB epidemiology

Scope of investigations covered by next generation molecular epidemiologyIdentifying …  risk factors that could not be identified by conventional or early-generation molecular biology l

National Institute of Infectious Disease

January 18, 2017

Next generation molecular epidemiology

epidemiology

Scope of investigations covered by next generation molecular epidemiology

Identifying …

 risk factors that could not be identified by conventional or early-generation molecular biology laboratory methods

 new or hidden transmission pathways

 direction of transmission of an infectious agent

 endogenous reactivation vs exogenous reinfection

 ecological niche from which clonal pathogenic strains are selected and

disseminate

 pathogen microbial population structures associated with a syndrome

 host commensal microbial population structures that determine

non-communicable disease outcomes

Methods used to genotypeM tuberculosis

Endonucleases used to cut DNA

5’ overhand

Blunt end

PvuII cutting site

PvuII site:

Restriction fragment length polymorphism (RFLP)

analysis of M tuberculosis DNA (Ferrazoli et al)

IS6110 Southern blot hybridization of RFLP analysis

(Ferrazoli et al)

Distribution of spoligotypes

(Sougakoff, Clin Microbiol Infect 2011; 17: 800–805)

Beijing , T, T1-Ghana , Haarlem , , LAM, S, Dehli/CAS and undetermined types

Distribution of spoligotypes in Brazil

Fortaleza (2007-8) Santos LR et al; 2013 115 33 12 5 5 45

Prison, RS (2007-08) Kuhleis D et al 2012 63 40 17.5 12.5 22 3 5

11 states (1996-2005) Gomes HM et al 2011 1991 46 12.2 18.6 23

Sao Paulo (2006-08) Mendes NH et al 2011 93 28 12 26 34

Minas Gerais (2004) Miranda SS et al 2011 114 55 7 10.5 5.3 22

Parana (2005-08) Noquti EN et al 2010 93 27 17 12 44

-A variation in the number of short, repeated segments found

in a specific locus

-Amplification of these loci (using primers specific to

flanking regions) will produce DNA fragments whose lengths vary between strains.

Variation in the Number of Tandem Repeats (VNTR)

• The Mycobacterium tuberculosis genome contains 41 loci with

direct tandem repeats of 50 – 70 bp

• The number of repeats per locus varies between strains

• Strains can be typed based on the number of repeats per locus

Mycobacterial Interspersed Repetitive Units

MIRU-VNTR for M tuberculosis rationale

24-loci MIRU VNTR as discriminating as IS6110-RFLP

12-loci MIRU VNTR

SNP analysis of M. tuberculosis strains (Gutacker et al, JID, 2006)

Beijing family strains

Role of Beijing clade in the transmission of MDRTB

• China (Heilongjiang Province: Wang J et al, J Clin Microbiol, 2011):

– Overall: 8%

– Beijing clade: 27%

• Russian Federation and former Soviet Union countries

(Niemann S et al, J Clin Microbiol, 2010; Afanas’ev MV et al, Eur J Clin Microbiol Infect Dis, 2011):

– Overall (retreatment cases): ~40%

– Beijing clade: 70-78%

Epidemiologic studies using IS6110-based strain-typing analysis

• Differentiating recent from past transmission

• Confirming an outbreak in institutional or community settings

• Identifying an outbreak in what appears to be sporadic cases of TB

• Identifying risk factors in community-based settings

• Identifying virulence factors associated with M tuberculosis

• Tracking geographic spread of clones of M tuberculosis

• Managing a clinical case of TB

• Detecting drug-resistant strains of M tuberculosis

• Evaluating laboratory contamination

• Evaluating effect of an intervention (e.g., BCG, INH prophylaxis)

Differentiating new cases of TB arising from recent infection vs remote past infection

• Recent transmission TB:

– Rapidly progressive or primary TB (CDC: within 2 yrs of new transmission)

• Remote past transmission TB:

– Reactivation TB

• Assumptions:

– Identical IS6110 pattern observed in 2 or more persons (called cluster) with

newly-diagnosed TB represent cases due to recent exogenous infection

– Isolates with unique RFLP pattern (noncluster pattern) in a community

represent isolates from a case of reactivation of endogenous infection.

• May not apply in rural areas with highly stable populations (Braden et al, JID, 1997)

• Low-risk countries: IS6110-based test may underestimate the

proportion of recent infection in young people and overestimate in older people with TB (Vinnycky et al, EID, 2003)

• High-risk countries: IS6110-based test may underestimate the

proportion of disease due to recent infection (Vinnycky et al, EID, 2003)

Proportion (%) of M tuberculosis isolates belonging to cluster

IS6110 RFLP groups (low-incidence regions)

Los Angeles, 1994-96 Texas, 1994-95 Colorado,1992-94 Arkansas,1992-93 Amsterdam,1992-95

Proportion (%) of M tuberculosis isolates belonging to cluster IS6110 RFLP

groups (high-incidence regions)

2000-10

Rio de Janeiro, 1996

Campinas, 1996-99

Methods used to genotype M tuberculosis

Are these genotyping tests able to truly

differentiate recent vs past transmission?

Problems with assumption based on genotyping tests

 More recent studies show the same genotypes of M

tuberculosis in the same geographic sites (>20 yrs!)

 What is the definition of “recent transmission”

 What is the threshold of SNPs needed to claim recent transmission vs past transmission?

Discussion for presentation TB

Mycobacteriumtuberculosis outbreaks: a retrospective observational study Lancet

ID, 2013.

Tuberculosis Outbreak in London: A Retrospective Observational Study PLoS

Medicine, 2016.

 What type of public health intervention can be devised based on WGS data?

 What are the disadvantages of WGS over other genotyping tests to study

epidemiology of TB?

 What criteria can be used to infer recent transmission TB using WGS data?

 What other types of information are needed to demonstrate recent

transmission TB in addition to WGS data?