NEW YORK – Circulating cell-free RNA (cfRNA) in blood can potentially be harnessed as a host-response biomarker for tuberculosis, according to a recent study led by Cornell University researchers.
Published in a Nature Communications paper last month, the proof-of-concept study suggests possible utility of plasma cfRNA for TB diagnosis and treatment. However, more work is needed to further validate these biomarkers in terms of their clinical effectiveness and practicability.
"We have been interested in looking at circulating RNA in the setting of infection and immunity," said Iwijn De Vlaminck, a biomedical engineering professor at Cornell and the corresponding author of the study. "The goal of this study was to develop a host signature which can inform [TB] disease severity and disease states."
Often secreted as a by-product of cell death in tissues or organs, plasma cfRNA can potentially reflect the dynamics and state of a host’s immune system during infection, according to De Vlaminck.
To explore a possible link between plasma cfRNA and TB, De Vlaminck and his team profiled 251 blood samples from three countries: Uganda, Vietnam, and the Philippines. These samples were collected from adult patients with coughing symptoms that lasted at least two weeks who were tested for TB at local outpatient clinics. Of the samples analyzed in the study, 142 were microbiologically confirmed to have TB, and 37 were from individuals living with HIV.
The researchers extracted cfRNA from patients' plasma, followed by reverse transcription of the RNA and next-generation sequencing. Since cfRNA in blood tends to be degraded, meaning many molecules no longer have a polyA tail available, De Vlaminck said, random primers were deployed during the reverse transcription step. Because random priming is also sensitive to DNA, one of the critical steps in the experiment was to deplete DNA molecules in the samples, he added.
After sequencing, the researchers deployed various computational strategies to help discern potential TB biomarkers. In the end, they proposed a six-gene signature that showed differential expression in TB cases versus controls. These genes are guanylate binding protein 5 (GBP5), BCL2/adenovirus E1B interaction protein 3-like (BNIP3L), Kruppel-like factor 6 (KLF6), dysferlin (DYSF), LIM and SH3 protein 1 (LASP1), and poly(rC)-binding protein 1 (PCBP1).
In a test cohort that contained 98 samples, the signature discriminated between TB positive and negative groups with 91.8 percent accuracy, 97.1 percent sensitivity, and 85.2 percent specificity. The researchers further tested the biomarkers in a validation cohort of 60 samples, where the signature discriminated between TB positive and negative groups with 88.3 percent accuracy, 97.1 percent sensitivity, and 76.0 percent specificity.
Compared with traditional TB tests, a potential advantage of the blood-based cfRNA assay is that it alleviates the need for sputum samples, which are not always available from all patients, De Vlaminck noted. In addition, while tests directly targeting Mycobacterium tuberculosis can accurately identify the pathogen, he said, host response assays can glean information on disease progression in the host and a patient’s immune status, potentially informing prognosis and treatment.
On the flip side, De Vlaminck acknowledged that the host-response signature is an indirect way to detect TB and might not be disease-specific, meaning a positive result might be caused by inflammation in the body other than TB. Therefore, future studies will need to confirm the clinical utility of the markers.
"By all means, the test is not perfect," De Vlaminck said, adding that there are other limitations of the study, which was just a proof of principle. For instance, he noted that the RNA-seq workflow used can be biased toward longer RNA molecules, leaving out information about smaller RNAs. Additionally, the authors noted they did not investigate patients with extrapulmonary TB, latent TB infection, or other forms of TB.
Tony Hu, a molecular biologist at Tulane University who was not involved in the study, said he appreciates this type of work. "If there are new tools for us to better understand how a host responds to pathogen infections, then we probably can better organize the treatment strategy for certain patients," he said.
According to Hu, current gold-standard TB tests, which are typically sputum-based, can only diagnose about 70 percent of cases. Therefore, the field needs to develop diagnostic assays using alternative sample types. Moreover, because tuberculosis is a chronic disease, learning about an infected host’s immune response can help inform their diagnosis and treatment, he added.
Hu said an important consideration for host-response assays is specificity, making further evaluation of the biomarkers’ performance necessary. Additionally, they should be tested in more populations to account for host genetic heterogeneity, he noted.
Besides accuracy and specificity, researchers also need to take into account the technical practicality of an assay, given TB tests are mostly performed in low- and middle-income countries with limited resources. De Vlaminck said the workflow described in the paper is "likely not" optimal for routine clinical implementation, given that next-generation sequencing can still be expensive to carry out, though its cost has been declining in recent years.
However, one potential upside for RNA biomarkers is that they may be translated into a PCR assay, lowering the technical barrier to adoption. With that in mind, De Vlaminck said his team has filed a patent for the biomarkers described in the study and is interested in working with a commercial partner to develop a PCR-based test using the host-response cfRNA signature once it is further validated.
Other host-response TB tests are already commercially available and approved by the US Food and Drug Administration. These include blood tests developed by Qiagen and by Oxford Immunotec that leverage interferon-gamma release assay (IGRA) technology for detecting latent TB infection.
Meanwhile, researchers have also been working on TB assays using non-sputum samples. Hu’s team, for instance, has been working to develop blood-based tests targeting host antigen proteins against the TB pathogen as well as extracellular vesicles secreted from M. tuberculosis-infected macrophages. There are also efforts by other research teams underway to achieve TB diagnosis using a mouth swab PCR test.
Encouraged by the results from the present study, De Vlaminck said his team plans to further validate the biomarkers in a much larger cohort. Another study is ongoing to investigate the cfRNA signature in pediatric patients.
"If you could develop high-quality [cfRNA] signatures, then there is an opportunity to potentially use those signatures to monitor disease severity, treatments, and perhaps to triage patients for the right treatment," De Vlaminck said.