Pathogens Unveiled

When patients come in with respiratory complaints (like coughs, sore throats, or runny noses), physicians go through a typical diagnostic process to figure out the source of the problem. Most of the time, symptoms like these are due to infections – but that’s not always the case. For example, allergies can present identically to a viral infection. Congestive heart failure can sometimes be hard to distinguish from pneumonia. Other inflammatory conditions can also adopt a very similar appearance to infection. So even the most basic question – does the patient have an infection at all? – can be quite difficult to answer. But once the presence of a pathogen is established, things get even more complicated when that typical question arises: is it a bacterium or a virus?

Viral infections, for the most part, don’t require medical intervention. Bacterial infections will sometimes resolve without intervention, too, but most physicians prefer to prescribe antibiotics if a bacterial pathogen can be confirmed. That “if” is key, though; in the interests of patient health and antibiotic stewardship, it’s important to prescribe drugs only where warranted – and that’s why distinguishing between these various causes of illness is so important. It ensures we get the right treatments to the right patients, and gives us the ability to offer not just a diagnosis, but a prognosis as well. Failing to give antibiotics to patients with bacterial infections could result in their condition declining, rather than improving. On the other hand, giving antibiotics to someone who doesn’t need them exposes that individual to the ever-growing list of drug side effects and also increases the risk of selecting for antibiotic-resistant pathogens. Distinguishing bacterial from viral infections, then, has serious implications for the future health of both the individual patient and the general population.

Studying signatures

The human genome contains an estimated 20,000 protein-coding genes, representing the blueprint for everything our bodies need to grow and survive. Some of those genes are active all the time, while others are activated only under certain conditions. Out of that very large number, my colleagues and I have identified 130 genes that are active in a limited set of situations – which include viral infections, bacterial infections, and non-infectious illnesses. But not all of those genes are active in every case. Some are turned on only when a patient has a viral infection, whereas others are activated only in the case of a bacterial infection, and still others only during non-infectious illnesses. The test we’re developing works by looking at how active each of these 130 genes is in a patient with respiratory tract symptoms. We then compare that patient’s gene activity pattern to the known patterns we describe in our recent paper (1). How well the patterns match determines the patient’s diagnosis.

The specific genes in the signatures were selected to maximize our ability to distinguish between the different illness groups – not chosen because of any known role in the viral or bacterial response. That’s one of the reasons we think it worked so well. We didn’t bias the process by restricting the genes only to those we thought might work; instead, we let mathematical models make those determinations for us. When the genes in the signatures had been selected, though, we did note that most of them have known roles in the immune response. For example, the bacterial classifier includes genes involved in processes like cell cycle regulation, cell growth, and differentiation, while the viral classifier includes ones involved in interferon response, T cell signaling, and RNA processing.

Developing diagnostics

In its current form, the new test is most suitable for research purposes. Why? It requires an entire vial of blood, which then has to undergo a fair amount of preparation. In total, it takes a full day’s work in the hands of an experienced laboratory technician – a requirement that we recognize is untenable for routine clinical use. That’s why we’re working with diagnostics developers to generate a test that would use no more than a few drops of blood, require minimal or even no pre-processing, and return results in an hour or less. Just load the blood sample onto the test cartridge and let it run! Although we haven’t finalized a test like that yet, we’re making exciting progress toward that goal.

The main steps we need to take now are to put the assay on a testing platform that can be used at the point of care, and to continue working to show that this paradigm – the host response – can be used in all populations, including infants, the elderly, and all ethnic groups. At the same time, we are expanding the test’s repertoire to include not only viral and bacterial pathogens, but also fungal infections. Ultimately, we’d like the test to address both patients in their general practitioners’ offices and the challenges of critically ill patients on hospital wards and in intensive care.

Knowing when to prescribe antibiotics is a major challenge in patient care, particularly when managing respiratory infections. Today, without accurate information on the cause of infection, most doctors prescribe antibiotics to ensure they are treating the most dangerous potential cause of infection. But many – perhaps most – respiratory infections are caused by viruses, which means that these drugs are being significantly overprescribed. Our technology is aimed at providing information to help doctors make the best possible decisions regarding which patients truly need antibiotics and which do not. If we can reduce the overuse of unnecessary drugs, then we might see a corresponding reduction in pathogens’ resistance. The ideal scenario, should this test ultimately be approved for broad use, is that a patient with a respiratory issue would go to the doctor’s office, have a simple blood test administered, and receive results by the time they meet with their physician. Hopefully, that’s what we’ll see in the not-too-distant future!

Ephraim Tsalik is Assistant Professor of Medicine, specializing in infectious diseases, at the Duke University School of Medicine, Durham, USA.