Can we finally reliably distinguish between viral and bacterial pneumonia, or identify the specific cause? Promising results from a nanoparticle-based test

Unpredictable pneumonias and their diagnosis

According to the development team, the main challenge lies in the fact that a wide range of pathogens is involved in the etiology of pneumonias, and identification is successful in only about half of the patients, even with the most extensive and advanced testing. Distinguishing between viral and bacterial pneumonia is also crucial for diagnosis, especially due to treatment modalities. Administering antibiotics to patients with viral pneumonia represents an ineffective therapy modality that risks developing bacterial resistance. Scientists at MIT have therefore developed a diagnostic tool with which they were able to determine the etiology of pneumonia within 2 hours in a study using mouse models, based on the detection of nanoparticles in urine.

Testing based on the detection of specific proteases

Given the large number of pathogens causing pneumonia, experts decided to focus on a “simpler” aspect, namely the specific features of immune defense. Each pathogen triggers a different immune response, characterized by the production of specific proteases. Certain forms of proteases are more active in either viral or bacterial infections. The human genome codes for more than 500 proteases, many of which are produced by immune system cells, such as T lymphocytes, neutrophils, and NK cells.

From available genetic data, scientists were able to identify 39 proteases that likely respond differently to various types of infectious agents. They then created 20 nanoparticle sensors that interact with specific proteases. The sensors consist of nanoparticles coated with peptides that can only be cleaved by a specific protease. Each peptide is labeled with a reporter molecule that is released when the peptide is cleaved by the activated protease. The reporter molecules are excreted in the urine, where their presence is subsequently determined using mass spectrometry. This test thus makes it possible to determine which proteases are most active in pulmonary tissue.

Researchers tested the sensors on 5 different models of pneumonia, specifically diseases caused by Streptococcus pneumoniae, Klebsiella pneumoniae, Haemophilus influenzae, influenza viruses, and mouse pneumonia viruses. They subjected the urine test results to machine learning analysis and were able to distinguish cases of pneumonia from healthy controls, as well as differentiate between bacterial and viral etiologies using the mentioned specific sensors. They also managed to distinguish individual pathogens, but with much less accuracy than when differentiating bacterial and viral etiologies.

What's next for research and practical application?

The sensors were administrated to mice via direct tracheal injection. In testing this diagnostic method on human patients, scientists are developing a new version applied using a nebulizer or inhaler. Also in development is a breath test, where the result could be read directly from exhaled air rather than urine, potentially further speeding up the diagnostic process.

Researchers have identified 5 sensors that could eventually be used to create diagnostic strips. Their vision includes the possibility of home testing for the etiology of pneumonia, similar to pregnancy tests. Another challenge is improving the method to distinguish between individual pathogens with high accuracy, not only differentiating viral and bacterial etiologies.

(holi)

Sources:
1. Trafton A. Nanoparticle sensor can distinguish between viral and bacterial pneumonia. MIT News, 2022 Jun 13. Available at: https://news.mit.edu/2022/sensor-viral-bacterial-pneumonia-0613
2. Anahtar M., Chan L. W., Ko H. et al. Host protease activity classifies pneumonia etiology. PNAS 2022 Jun 21; 119 (25): e2121778119, doi: 10.1073/pnas.2121778119.