TL;DR
This study provides the first comprehensive single-cell analysis of transcriptional changes in the aging human lung, revealing that aging impacts different cell types heterogeneously. The researchers developed a "LungAge" signature that can predict physiological aging and identified immune and progenitor cells as particularly prone to senescence.
What Problem Does This Paper Address?
Aging is a natural process characterized by a decline in organ function and increased susceptibility to chronic diseases. The lung, a complex organ with numerous cell types, is particularly vulnerable to age-related decline and diseases. While previous studies have begun to map transcriptional changes in the lung, many haven't adequately accounted for the specific impact of aging itself, or how age-related changes contribute to disease. Understanding how different lung cell types are affected by aging at a molecular level is crucial for developing targeted therapeutic strategies. This paper aims to fill this gap by providing a systemic, single-cell resolution analysis of healthy aging lungs.
Methods
The researchers employed single-cell RNA sequencing (scRNA-seq) to analyze lung tissue from young and aged healthy human donors. Key techniques and approaches include:
- scRNA-seq Data Generation and Processing: New samples were generated, and publicly available datasets (Human Lung Cell Atlas) were integrated. Standard scRNA-seq processing pipelines using Seurat were applied, including normalization (SCTransform), batch effect correction (Harmony), and cell type annotation using known marker genes.
- Differential Gene Expression (DEG) Analysis: To account for batch effects and pseudo-replication, a "Pseudobulk" approach was used, followed by batch correction with Combat-seq. DEGs between young and aged groups were identified for each cell population.
- Pathway and Module Score Analysis: Ingenuity Pathway Analysis (IPA) was used to identify enriched canonical pathways. Module scores were calculated to infer gene set activities, including those related to aging and senescence.
- Random Forest (RF) Classifier: An RF classifier was developed and trained using the identified "LungAge" gene signatures to predict the aging state of lung cells. This involved stratified splitting of data and cross-validation for robust evaluation.
- Senescence Score Calculation: Two established gene sets for senescence were used to calculate senescence scores for individual cells and cell populations.
Datasets
The study utilized a combination of newly generated and publicly available scRNA-seq datasets:
- Newly generated data: 10 new healthy lung samples.
- Publicly available datasets: Data from the Human Lung Cell Atlas (HLCA), specifically including datasets from Adams et al. (2020), Habermann et al. (2020), and Reyfman et al. (2019), as well as samples from previous studies by the authors (Morse et al., 2019; Valenzi et al., 2019).
- Validation datasets: Additional datasets from Adams et al. (2020) and Habermann et al. (2020) were used for validating the developed LungAge classifier.
The study focused on comparing young (≤23 years old) and aged (≥55 years old) individuals, excluding samples with smoking history or other lung pathologies.
Key Insights and Ideas
- Heterogeneous Aging: Aging doesn't affect all lung cells equally. Some cell populations show significant transcriptional changes, while others remain relatively stable. This highlights the complexity of lung aging.
- Monocytes and Macrophages as Key Players: Monocytes and alveolar macrophages were identified as the most transcriptionally dynamic cell types during aging, with changes strongly linked to inflammation and immune response regulation. This suggests their central role in "inflammaging."
- The "LungAge" Signature: The development of a "LungAge" score, based on consistently differentially expressed genes, provides a powerful tool to predict physiological lung aging and identify signatures of age-related lung diseases.
- Immune and Progenitor Cells Prone to Senescence: Immune cells and progenitor cells were found to be most biased towards senescence in aged lungs, a finding consistent with their roles in tissue repair and immune function.
- CTCF as a Central Regulator: The transcription factor CTCF emerged as a key regulator, influencing both upregulated and downregulated gene signatures associated with aging, suggesting its role in chromatin organization and transcriptional regulation during the aging process.
Key Results
- 22 Cell Types Identified: The analysis identified 22 distinct cell populations in the human lung, including stromal, endothelial, epithelial, and immune lineages.
- Monocyte-like and FABP4 Macrophages Most Affected: Monocyte-like macrophages showed the highest number of differentially expressed genes (DEGs) (705), followed by FABP4 macrophages (498), indicating significant transcriptional shifts in these immune cells.
- LungAge Classifier Performance: The developed LungAge classifier achieved a high Area Under the Receiver Operating Characteristic (AUROC) of 0.87 for predicting aging state, with downregulated genes being slightly more predictive (AUROC=0.85) than upregulated genes (AUROC=0.76).
- Top Predictive Genes: Key genes predicting aging included those involved in DNA damage repair (YBX1), inflammation (S100A8, S100A9), and fatty acid metabolism.
- Senescence Scores Higher in Aged Lungs: Senescence scores, calculated using established gene sets, were significantly higher in aged lung samples compared to young controls across multiple cell types.
Limitations and Future Work
- Sample Size and Representation: The study acknowledges that the sample size may limit the representation of certain cellular populations and could impact the statistical power of DEG analysis. The use of pseudobulk methods was employed to mitigate this.
- Environmental Exposure: The influence of environmental exposures (e.g., pollution) on DNA methylation age and transcriptional changes in lung cells was noted as a critical consideration for future studies.
- ETV2 Function: While ETV2 was identified as differentially expressed in aged lung epithelium, its specific function in this context remains to be reported.
- Future Directions: The authors suggest that the LungAge signature can serve as a valuable tool for predicting physiological aging and identifying molecular pathways linking age and chronic lung diseases. Further research could explore the therapeutic potential of targeting the identified transcription factors.
Summary updated