Francesca Tomasi received her B.A. from the University of Chicago and is now a microbiologist.
Three years ago, a 2-year-old having recently returned from a trip to India arrived at the Johns Hopkins Children’s Center with persistent fever and malaise. A slew of tests yielded no conclusive diagnosis, until a chest x-ray showed a lung abnormality. Eventually, she was diagnosed with extensively drug-resistant tuberculosis (XDR-TB). This was going to be a difficult case: only a handful of TB cases in young children appear in medical literature, giving from their lungs like adults can, and sputum samples are the main source for diagnostic testing of tuberculosis. Also, ask your toddler to swallow multiple large pills over several months and see how that goes.
Figuring out the child had XDR-TB took about 12 weeks, and the diagnosis was just the start. The physicians at Johns Hopkins dealt first-hand with the limitations of current diagnostic tools, which are time consuming and lack in sensitivity (that is, there are a lot of false negatives). Tracking drug resistance over the disease’s course was another daunting task, and a lack of truly reliable disease markers made it nearly impossible to monitor the child’s response to drug treatment. Monitoring drug treatment was especially important for this pediatric case, since any antibiotics used had to be altered to the physiological and metabolic needs of a two-year-old. This included mushing up the drugs and strategically placing them in her meals.
Enter Dr. Sanjay Jain, Johns Hopkins Children’s Center pediatrician and tuberculosis expert. His team of researchers has been working on real-time imaging techniques to track tuberculosis in patients, methods including CT imaging and PET scanning to track bacterial behavior. Having successfully imaged tuberculosis in mice, it was time to bring the bench to the bedside. Using a child-friendly modification of CT imaging with minimal radiation, clinicians repeatedly performed CT scans of their patient and tracked the progress of her illness over several weeks. A lack of reliable biomarkers for pediatric tuberculosis, Jain explained, prompted his team to turn to this emerging method of live infection imaging. As CT scans over time revealed lower bacterial counts, the patient showed physiological signs of improvement, showing an elegant – and rare – temporal parallel between diagnostic data and physiological response. Now, three years later, the child’s tuberculosis has cleared and she is being monitored to make sure that her infection does not relapse.
This case and its success provide a precedent to the implications of infectious disease imaging, an emerging field with incredible potential to radically improve treatment of all kinds of infectious diseases. Sanjay Jain’s team is one of the pioneering groups working on imaging of infectious diseases (IOI), and has recently developed a way to detect and monitor infections in real-time caused by Gram-negative bacteria. This class of bacteria is physically and physiologically distinct from the acid-fast group to which M. tuberculosis belongs, and increasingly drug-resistant forms of Gram-negative pathogens are responsible for many serious hospital and community-acquired infections.
Jain’s imaging system identifies the source of an infection and allows for real-time tracking of bacterial proliferation, allowing physicians to quickly judge the efficacy of antibiotic therapies in a patient. Because it uses existing technologies and compounds already known to be safe, translation of this work from animal studies to humans is in the near future. The technique works by exploiting a unique metabolic feature of Gram-negative pathogens: these bacteria, which include E. coli, Salmonella, and Klebsiella species, readily consume sorbitol, a compound often found in artificial sweeteners. On the other hand, other microbes and human cells do not absorb sorbitol. Thus, Jain’s team generated radio-labeled sorbitol and incorporated it into a pre-existing PET imaging tracer, which provides a light tag on specific targets. In this case, the researchers’ hypothesis was that Gram-negative bacteria would take up the modified sorbitol and glow on a PET scanner image. Indeed, when Jain’s research team injected mice with live E. coli on one thigh and dead E. coli in the other, the sorbitol-based radio tracer localized to the thigh with live bacteria, and only one thigh lit up on the PET scan. This procedure is pivotal because it discerns between bacteria-induced inflammation and non-pathogenic inflammation, such as something caused by an immune response to a separate trigger.
Being able to rapidly identify the class of organism responsible for a patient’s infection will lead to more appropriate antibiotic choices, and faster. Being able to pinpoint the epicenter of an infection will allow a physician to directly target the source with a specific drug and avoid using broad-spectrum stepping stones that promote drug resistance by challenging all sorts of bacteria living in a person’s body. Furthermore, tracking the light signal of a PET tracer is a simple way to track the efficacy of a therapy: if a signal weakens over time, the drugs are working. If it strengthens or remains constant, a different course of action is due. Rapid feedback against a dynamic source of illness is key to winning the fight. In the end, the patient gets better, the healthcare system spends less money on unnecessary drug treatment, and the bad bugs lose.