The current recommended treatment duration for pulmonary tuberculosis (TB) is six months. Despite this recommendation, approximately 5% of TB patients will fail treatment and an additional 5% will relapse. While failure and relapse patients might well have been cured with longer or more intensive therapy, many patients can clearly be treated successfully in four months or less. Identifying markers that could assign patients to shorter or longer therapy could transform TB treatment, improving outcomes, increasing adherence and decreasing costs. This proposal will identify biomarkers that predict response to four and six month TB treatment. We found that the number of M. tuberculosis genomes (genome load) present in sputum at the start of treatment, and the rate of genome load decline with treatment, is predictive of treatment outcome. We have also shown that even in subjects with fully drug-susceptible M. tuberculosis (as defined by standard breakpoints), considerable heterogeneity among the minimal inhibitory concentrations (MICs) of each isolate remains. Importantly, we found that MIC correlated very strongly with treatment outcome, even controlling for clinical characteristics. Although our novel biomarkers appear to be highly prognostic, host drug absorption- distribution (pharmacokinetics, PK) is also likely to be of prognostic value, especially as PK interacts with MIC at the patient level. We will test, improve and validate each of our novel treatment response markers using samples obtained from TB patients enrolled into the joint Tuberculosis Trials Consortium and AIDS Clinical Trial Group study 31/A5349, treated for four or six months and followed 18 months for relapse. We will create the first comprehensive predictive model of TB treatment response and cure incorporating genome load, pre- treatment MIC and host PK, and identify genomic and transcriptional markers that can replace bacterial MIC with a rapid pre-treatment test. This will be done in the following three aims: 1) Test and refine molecular measures of (a) initial bacterial load and (b) change in bacterial load (as measured by the slope) in response to therapy, as predictors of successful four and six month treatment for drug-susceptible TB. 2) Investigate the contribution of bacterial sub-breakpoint MICs as a predictor of successful four and six month treatment for drug-susceptible TB; and identify simplified markers (new critical MIC values, and mRNA or mutations in the bacterial chromosome) with the potential to be developed into a predictive test. 3) Develop and test a highly predictive algorithm that identifies TB patients who will be cured with short course (i.e. ? 4 months) TB treatment, in a model that includes measurable clinical characteristics, host pharmacokinetics (PK), and the measures refined and validated individually in aims 1 and 2.

Public Health Relevance

BENEFIT This study will help us understand and predict why tuberculosis must be usually be treated for at least six months to be cured. This will help us design generally faster ways to treat tuberculosis that are less expensive and safer.

Agency
National Institute of Health (NIH)
Institute
National Institute of Allergy and Infectious Diseases (NIAID)
Type
Research Project (R01)
Project #
5R01AI126788-02
Application #
9302263
Study Section
Special Emphasis Panel (ZRG1-IDM-S (02)M)
Program Officer
Lacourciere, Karen A
Project Start
2016-06-22
Project End
2021-05-31
Budget Start
2017-06-01
Budget End
2018-05-31
Support Year
2
Fiscal Year
2017
Total Cost
$728,068
Indirect Cost
$229,436
Name
Rutgers University
Department
Internal Medicine/Medicine
Type
Schools of Medicine
DUNS #
078795851
City
Newark
State
NJ
Country
United States
Zip Code
07103
Colangeli, Roberto; Jedrey, Hannah; Kim, Soyeon et al. (2018) Bacterial Factors That Predict Relapse after Tuberculosis Therapy. N Engl J Med 379:823-833