Molecular Signatures of Tuberculosis-Diabetes Interaction: Transcriptomics Study

Post Date: 
Clinical Sites: 

Type 2 diabetes is an acquired host factor that increases the risk for and severity of pulmonary tuberculosis. A report published in 2007 estimated that, in India, diabetes accounted for 14.8% (range 7.1% to 23.8%) of pulmonary tuberculosis and 20.2% (8.3% to 41.9%) of smear-positive cases. Emerging data from Brazil and India suggest that the spectrum of host metabolic disorders adversely effecting tuberculosis defense may include dyslipidemia and dysglycemia below the diagnostic cutoff for diabetes classification. The global burden of tuberuclosis attributable to the full spectrum of host metabolic disorders may therefore be substantially greater than current estimates. The ~3 fold greater individual tuberculosis risk with diabetes is one tenth the risk impact of HIV but there are ~10 times more people living with diabetes than with HIV making the population-attributable risk similar for these two conditions. Diabetes prevalence is rising at the fastest rate in countries with high tuberculosis burden including India and Brazil. New knowledge is urgently needed to identify modifiable factors influencing tuberculosis risk and disease outcomes in people living with diabetes, and to inform the rational development of strategies to address this significant barrier to global tuberculosis control.

Whole blood transcriptomic profiling revealed a signature of active tuberculosis disease dominated by interferon signaling predominantly in neutrophils. Quantitative metrics of global transcriptional perturbation were shown to correlate with radiographic severity of tuberculosis at baseline and to regress to the healthy control background with effective anti-tuberculous treatment (ATT). This original observation was supported by over ten additional blood transcriptome studies in several different countries. None of the tuberculosis blood transcriptome studies published through 2016 included participants with diabetes or used included samples from tuberculosis patients living in the Indian subcontinent. We recently published a preliminary study using Illumina HT-2 microarrays to measure whole blood gene expression in tuberculosis patients with diabetes (TBDM) and tuberculosis patients without DM (TB) at baseline (<7 days ATT), comparing to healthy controls (HC) and to diabetic controls (DM) without active tuberculosis disease. In that study, the differences in gene expression between the TBDM and TB primarily reflected non-immune pathways linked to the microvascular and macrovascular complications of diabetes. On the other hand, tuberculosis disease appeared to the main driver of the immune pathway transcriptional response with no qualitative separation of TBDM vs TB by hierarchical clustering or quantitative separation by the molecular degree of perturbation (MDP).

Our published transcriptomic analysis had several limitations. It was conducted at a single site, there was an unexpected degree of heterogeneity within all four subgroups at the level of MDP, and it used a microarray technology that limits comparability with different transcriptomic datasets and which has now gone out of production. A multi-site gene expression profiling study of TBDM vs TB patients is warranted to verify (or challenge) the results from our preliminary study, to identify a transcriptional signature of TBDM, and to create a high quality dataset using RNA sequencing (RNA-seq) which is more sensitive and specific than microarray.

The overall objective of this study is to identify transcriptomic signatures unique to TBDM and to characterize its change in response to ATT. In doing so, we hope to generate strong preliminary data for a programmatic grant application seeking funding for a comprehensive systems biology investigation of the TBDM interaction in India.

Specific Aims: 

  1. Compare baseline whole blood RNA-seq transcriptional signatures between newly diagnosed drug-sensitive PTB patients with and without DM in Pune and Chennai, India and Salvador, Brazil.
  2. Characterize and compare longitudinal change in transcriptional signatures in response to ATT between newly diagnosed drug-sensitive PTB patients with and without DM in Chennai, India.

University of Massachusetts Medical School, Worcester, MA

Instituto Brasileiro para a Investigacao da Tuberculose, Salvador, Brazil