Microscope image from the lung of someone who died with pneumonia, with different types of immune cells showing up as different colors.

Research

Analyses of Human Lungs Reveal Seven Subphenotypes of Pneumonia

Sub-phenotypes of pneumonia suggest a foundation for developing and testing therapies for patients.

Pneumonia is responsible for a tremendous burden of disease worldwide. In the U.S., it is a leading cause of death due to infection, especially for those of advanced age. For survivors, pneumonia’s lingering effects such as reduced lung function, scarring and new or worsened respiratory issues like asthma or COPD, may accelerate unhealthy aging.  While pneumonia is fundamentally a disease of the lung tissue characterized by inflammation and alveolar damage, medical science has historically relied on symptoms, imaging (X-rays), and microbiological cultures (microbe-directed) to classify the disease, rather than analyzing the specific cellular damage and structural changes in the lungs (histopathology) to create personalized treatment subgroups (subphenotyping).

In a new study, researchers have identified seven different forms of pneumonia. This is the first systematic examination of pulmonary histopathology during pneumonia, resulting in a new framework for understanding pneumonia heterogeneity based on cellular resolution of lung biology.

“While pneumonia is a pulmonary pathophysiology, pneumonia patients have not been subphenotyped based on biological processes within the lungs. Focusing on the lungs, where the pneumonia process is occurring, we were hoping to determine whether distinct sub-phenotypes of pneumonia would emerge based on variations in local inflammation and damage revealed by pulmonary histopathology,” explains corresponding author Joseph P. Mizgerd, ScD, the Jerome S. Brody, MD, Professor of Pulmonary Medicine and director of the Pulmonary Center at the school.

The researchers used various microscope modalities to examine the lungs of several hundred people who died with pneumonia. After scoring each lung for 20 different types of histopathology, machine learning algorithms clustered these hundreds of subjects into seven distinct groups. The groups had discrete patterns of aberrations in their lungs, which was associated with distinctly different microbes and immune cells. Similar groupings were also observed in lung samples from experimental models, suggesting similar biology applies across different mammalian species and providing tools for deciphering how these pathologies develop and can be prevented or treated.

According to the researchers, the next steps involve finding biomarkers that report these subphenotypes, defining mechanisms that drive them and testing whether host-directed therapies have distinct outcomes for different subgroups. “Creating a better understanding of the multiple distinct biological processes that damage the lungs during pneumonia will guide innovative approaches to treating and possibly preventing pneumonia,” adds Mizgerd.

These findings appear online in the American Journal of Respiratory and Critical Care Medicine.