A New Look at COPD Heterogeneity
—A recent study used multilayer omic networks to help identify molecularly distinct groups of patients with COPD in an effort to explain why patients with similar presentations might have different and distinct clinical courses.
A unique new study inches us closer to an understanding of the conundrum that has long baffled clinicians caring for patients with chronic obstructive lung disease (COPD): Why do patients with similar clinical presentations have dramatically different clinical courses?1
To address this perplexing heterogeneity, a multinational team of investigators undertook a novel multilayer network analysis of lung tissue from patients with COPD.2
“Of note, we found that patients with similar airflow limitation severity can have molecularly distinct small airways and immune response patterns, indicating that different endotypes can lead to similar clinical presentation,” the investigators of this recent study wrote of their findings.2
Why are these findings important?
In the accompanying editorial, Matthew R. Moll, MD, MPH, from Brigham and Women’s Hospital in Boston, and John McDonough, PhD, of McMaster University in Ontario, Canada, contend that this biologic heterogeneity among those with similar disease severity has thwarted development of targeted therapies. “Thus, there has been increasing interest in identifying COPD endotypes, or subgroups of individuals with shared pathobiology, that can be leveraged for personalized therapeutic intervention,” they wrote.1
In this groundbreaking study, Dr. Moll told ֱ, “Olvera and colleagues utilized lung tissue from 135 former smokers with COPD and constructed a multilayer network integrating mRNA, mi-RNA, and DNA methylation data.”
The study didn’t end here. Dr. Moll explained: “The authors then constructed a machine learning model to identify these endotypes in 2 independent cohorts, and validated their findings in these independent cohorts (including one with spatial transcriptomics).”
Dr. Moll described 3 novel features of the study. The team:
- “For the first time, integrated 3 types of omics data into a network from a large sample size of COPD lung tissue.”
- “Demonstrated that amongst those with severe disease, there can be heterogeneity with respect to biological processes, and this finding may explain the variable clinical courses of severe COPD patients.”
- “Built a machine learning model that allowed them to validate their findings in additional cohorts, a step that is critical but rarely performed via this machine learning-based approach.”
“To our knowledge, this is the first study to use an unbiased, integrative, multilayer network approach to investigate the molecular heterogeneity of lung tissue in COPD, and to correlate it with clinical characteristics,” wrote the authors in their paper.2
Clinically similar but molecularly divergent communities identified
The study’s main findings, according to the authors, were the following2:
- Five molecularly defined communities (labeled C#1 to C#5) of patients with COPD were identified by integrating 3 omic layers: miRNA, methylome, and transcriptome
- Each patient community was associated with different clinical features of COPD, such as severity of airflow obstruction
- Notably, 2 clinically similar but molecularly different communities of patients with the most severe disease (C#3 and C#4) demonstrated divergent biological responses
Dr. Moll elaborated on this most provocative finding: “The investigators found that amongst those with severe and very severe disease, there were 2 subgroups with similar clinical characteristics in terms of emphysema and diffusing capacity of the lungs for carbon monoxide (DLCO), but with divergent biological processes—ie, distinct endotypes. Specifically, these endotypes demonstrated opposite regulation of B and T cell signatures and secretory and ciliated cells.”
What does this mean to clinical practice?
“The key finding as a clinician,” said Dr. Moll, “is to recognize that severe COPD patients have heterogeneity in terms of biological processes, which likely explains variable clinical courses and responses to therapies in this population. This means that we cannot assume the same treatments will be appropriate for all patients within a GOLD grade. This emphasizes the importance of considering how ancillary information might change management, such as the measurement of eosinophils and characterization of comorbidities such as allergic rhinitis, asthma, bronchiectasis, and others. More biomarkers of disease subtypes and endotypes are likely to become part of clinical practice in the future.”
While these findings clearly set the stage for a future of more personalized management of COPD, Dr. Moll pointed out that, “The findings do not instantly change clinical practice except to emphasize the importance of recognizing heterogeneity within severe and very severe COPD patients.”
What is needed before precision medicine can take hold? “There remain many important limitations to the work, including linking the identified lung endotypes to important clinical outcomes (eg, exacerbations, lung function decline) and a method to identify endotypes without obtaining lung tissue. The study is an important step forward but is an early example of how omics has the potential to change our approach to chronic lung diseases.”
Published:
References