Gene-expression patterns in whole blood identify subjects at risk for recurrent tuberculosis.

The majority of patients with tuberculosis who comply with appropriate treatment are cured. However, approximately 5% subsequently have a repeat disease episode, usually within 2 years of successful combination therapy. Presently, there is no way of predicting which patients will experience a relapse. We identified 10 subjects who had previously experienced recurrent tuberculosis and carefully matched them to cured subjects who had had only 1 episode of tuberculosis, to patients with active tuberculosis, and to latently infected healthy subjects. We compared their ex vivo whole-blood gene-expression profiles by use of DNA array technology and confirmed the results by use of quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR). The 4 clinical tuberculosis groups exhibited distinct patterns of gene expression. The gene-transcript profiles of the patients with recurrent tuberculosis were more similar to those of the patients with active tuberculosis than to those of the cured or latently infected subjects. Discriminant analysis of a training data set showed that 9 genes were sufficient to classify the subjects. We confirmed that measurement of the expression of these genes by qRT-PCR can accurately discriminate between subjects in a test set of samples. A simple test based on gene-expression patterns may be used as a biomarker of cure while identifying patients who are at risk for relapse. This would facilitate the introduction of new tuberculosis drugs.