Latest News

RNA Biosignatures: Overcoming Hurdles In Diagnosing Childhood Bacterial Infections

By Dana Barberio

September 14, 2016 | Bacterial and viral infections are difficult to distinguish clinically, and culturing techniques to diagnose bacterial infections—blood cultures, urine cultures, and sometimes a lumbar puncture to obtain cerebral spinal fluids—are slow, invasive, and not as specific as they could be. Infants and children are particularly vulnerable as bacterial infections such as bacteremia, meningitis, and urinary tract infections can be missed or antibiotics unnecessarily prescribed.

Preliminary studies published in the August 2016 edition of the Journal of the American Medical Association by two groups discuss genetic techniques that have the potential to make diagnosis of bacterial infections faster, more sensitive, more specific, and less invasive. In both studies, researchers experimentally defined diagnostic patterns of RNA expression, or RNA biosignatures, to distinguish bacterial from viral infections. These novel applications make use of the fact that bacterial infections induce an immune response in their host that create the unique patterns of RNA biosignatures.

In an interview with Diagnostics World, Nathan Kuppermann of UC Davis, one of the three principal investigators of the first JAMA study (doi:10.1001/jama.2016.9207) and one of the principal investigators of the Pediatric Emergency Care Applied Research Network (PECARN), elaborated on some of the drawbacks of standard techniques for culturing patient samples to diagnose infections.

“Blood and spinal fluid samples typically require 24 to 48 hours for the cultures to grow, and in the meantime the patient frequently is hospitalized and given antibiotics,” he said. In addition, there’s a risk that normal skin flora can contaminate blood cultures and cause false positives if skin isn’t cleaned properly before a blood draw. And maybe most importantly, if enough blood isn’t drawn, labs may not be able to grow bacteria that are truly in the blood, giving a false negative result. In contrast to culturing, Kuppermann says that the RNA biosignature technique can be rapid, and doesn’t require much blood.

The study was a large multicenter effort through PECARN and led by Kuppermann and two other principal investigators: Prashant Mahajan of Wayne State University and Octavio Ramilo of Nationwide Children’s Hospital. 279 infants younger than 60 days old were randomly chosen from a group of 1883 children with fever admitted to emergency rooms over a 2-year period. The chosen cases had RNA biosignature analysis in addition to having standard cultures performed on their blood samples. The patients included those with bacteremia, urinary tract infections, and healthy infants. The analysis was done by Illumina microarray analysis of blood leukocytes.

Using the microarray-based RNA biosignature technique, the researchers identified 66 classifier genes that distinguished the presence or absence of bacterial infections in the test set with 89% specificity and 87% sensitivity. Ten classifier genes successfully distinguished bacteremia from non-bacterial infections in the test set with 95% specificity and 94% sensitivity. Before this study, it was not known if infants have a sufficiently strong immune response for this type of evaluation of their immune response to be an option.

How do their results compare to the decades-long technique of taking cultures for diagnosis? According to Kuppermann: “Many times the culture and the biosignature agreed. In the cases when they didn’t agree, when we looked at lab screening tests such as white blood cell count and other markers of infection, it seemed to suggest that the biosignature was more frequently correct.”

This technique has the potential to be applied to other types of infections as well. For a disease such as pneumonia, using RNA biosignatures has a potentially large benefit. “It’s hard to get a sample from your lungs,” Kuppermann says. “Adults can cough up sputum, but in infants, it is extremely difficult to obtain this sample, making it hard to make a microbiologic-based diagnosis. So the RNA biosignatures may be an ideal test for lung infections.”

In order to get this new test into clinical settings, the PECARN researchers plan to test many more samples, examine the stability of the biosignatures regardless of illness stage, and test a new platform that is based on a PCR technique.

In a similar multi-center study, also published in the August 2016 JAMA (doi:10.1001/jama.2016.11236) researchers also identified RNA biosignatures which distinguished viral and bacterial infections. 240 febrile children with an average age of 19 months in hospitals in the U.S., United Kingdom, Spain and the Netherlands, were recruited over a four-year period and placed in a discovery or validation group. All infections were categorized as bacterial, viral, or undetermined using bacterial cultures. The researchers developed a 38-transcript RNA expression signature based on the infections that were definitively bacterial or viral, and then assessed performance using the validation group.

The researchers then identified a 2-transcript RNA biosignature which successfully distinguished all 23 patients classified with bacterial infection as determined by culture techniques. The 2-transcript RNA biosignature also distinguished 27 of 28 viral infections (96% specificity) that had been classified as viral by culture analysis. Additional analysis of children with meningococcal disease identified bacterial infections with 92% sensitivity and 96% specificity; analysis of inflammatory diseases identified bacterial infections with 90% sensitivity and 96% specificity. As next steps, the authors plan to investigate the clinical accuracy and utility in diverse sets of patients.

The use of RNA biosignatures would potentially reduce unnecessary antibiotic use in clinical practice. The statistics on antibiotic use in infants is daunting, as noted in an accompanying JAMA editorial (doi:10.1001/jama.2016.11137) by Editor-in-Chief Howard Bauchner, MD. The vast majority of infants younger than 30 days presenting with fever to an emergency room are subsequently sent to the hospital and given antibiotics. This is despite the fact that “in this age group, approximately 5% to 8% of infants presenting with fever to an emergency department have serious bacterial infection,” noted Bauchner.

Researchers hope RNA biosignatures will more quickly identify these children, while saving others from unnecessary antibiotics.