Skip to main content
Premium Trial:

Request an Annual Quote

Study Finds Blood-based Expression Signature for Childhood Tuberculosis

NEW YORK (GenomeWeb) – An international team has uncovered a gene expression signature in host blood that appears to have potential for more accurately detecting and diagnosing tuberculosis infections in children.

"It has taken seven years and the combined efforts of clinicians and scientists in the UK, Africa and Singapore to identify this gene signature of childhood TB," Imperial College London tropical medicine researcher Michael Levin said in a statement.

"What we now need is collaboration from biotechnology and industrial partners to turn these findings into a simple, rapid and affordable test for TB that can be used in hospitals worldwide," Levin added.

In a study published online last night in the New England Journal of Medicine, Levin and other researchers involved in the ILULU Consortium and the KIDS TB Study Group described the array-based approach they used to profile gene expression patterns in blood samples from hundreds of African children with potential TB infections.

From that data, they narrowed in on a set of 51 genes showing distinct expression profiles in the blood of children with TB infections compared to their TB-free counterparts or children with latent TB infections.

Follow-up experiments suggest the blood levels of genes in the TB-associated set can be used to detect the disease with almost 83 percent sensitivity and nearly 84 percent specificity in TB cases that were verifiable by independent, culture-based methods.

The sensitivity and specificity was somewhat lower in instances where children had probable or possible TB infections that could not be culture confirmed, the study authors noted, though it still caught as many as 80 percent or more apparent TB cases.

Unlike adult TB cases, which are typically diagnosed using microscopic and culture-based methods, the Mycobacterium tuberculosis culprit behind the condition does not always turn up in such tests of children with TB, researchers noted. Such diagnoses are further complicated by the fact that TB shares symptoms with other conditions that may manifest themselves in childhood.

"The symptoms of TB in children are common to many other childhood diseases, and the standard tests used on adults are not effective in children," Levin said.

"Although the disease is treatable," he noted, "thousands of children still die each year due to late diagnosis and many more are left with damage to their brain, bones and lungs."

In an effort to find a more reliable means of identifying authentic TB cases in children, Levin and colleagues used Illumina Human HT-12 v.4 expression BeadChip arrays to assess gene expression levels in blood samples from 157 South African children and 189 Malawian children with possible TB infections.

Within that discovery cohort, the researchers saw a set of 51 genes with differential expression in infections caused by culture-confirmed TB compared to other types of disease, while expression profiles for 42 genes differed in the blood of those with active and latent TB infections.

The team subsequently verified that 51-gene signature using blood samples from another 157 Kenyan children with or without culture-confirmed TB. The latter group included children who were TB-free as well as children with TB infections that were deemed highly probable, probable, or possible based on other clinical features.

In that group, the signature could pick up culture-confirmed TB cases with 82.9 percent sensitivity and 83.6 percent specificity. The sensitivity fell to between 62.5 and 82.3 percent for children with "highly probable" TB infections, the researchers reported, and between 42.1 to 80.8 percent in those with "probable" TB infections.

For children with "possible" infections, they found that the blood-based expression profiles proved useful for finding an estimated 35.3 to 79.6 percent of cases.

The team noted that that sensitivity of the signature was higher than that achieved when it tested samples from the Kenyan cohort with Cepheid's Xpert MTB/RIF, a DNA-based, real-time PCR assay for detecting TB. In that group, the Xpert test reportedly had roughly 54 percent sensitivity (ranging from 37.1 to 68.6 percent) and 100 percent specificity.

The sensitivity of that test dipped in the probable or possible TB cases that could not be verified by culture-based methods, the researchers reported, though the specificity remained at 100 percent.