C29-25 Cellular Hallmarks of Lung Aging Revealed by Single-Cell and Spatial Transcriptomics

Rationale

Aging is a major risk factor for respiratory morbidity and mortality, yet the cellular and molecular mechanisms underlying human lung aging remain poorly defined. Prior studies have been limited by bulk tissue analyses, small cohorts, or lack of spatial context, obscuring cell type-specific and intercellular aging processes. We sought to define the cellular composition, transcriptional programs, and spatial interactions that characterize normal human lung aging at single-cell resolution and to identify biomarkers of lung biological age.

Methods

We developed an integrative computational framework leveraging single-cell RNA sequencing (scRNA-seq) data from 184 normal human lung parenchyma samples (ages 15-80 years) from the Human Lung Cell Atlas, complemented by single-cell spatial transcriptomics from 70 lung parenchyma samples profiled using the Xenium platform. Bulk RNA-seq data from GTEx and single-cell RNA/TCR sequencing from peripheral blood were used for validation and cross-tissue comparison. We analyzed age-associated changes in cell composition, senescence and proliferation markers, de novo gene expression modules, ligand-receptor interactions, and spatial cellular neighborhoods. Functional validation of mitochondrial dysfunction was performed using metabolic assays in human monocyte-derived macrophages. Finally, we trained a machine learning model to predict lung biological age using single-cell-derived transcriptional modules.

Results

Across cohorts, aging was associated with reduced abundance of alveolar epithelial and macrophage cell subsets, alongside increased T, NK, mast, and stromal cell populations. Senescence markers increased with age in a highly cell type-specific manner, while proliferation declined broadly, most prominently in alveolar type II (AT2) cells and macrophages. Epithelial analyses revealed accumulation of a KRT8⁺ pre-alveolar transitional (PATS-like) state and reduced autocrine WNT signaling, consistent with impaired alveolar regeneration. Myeloid cells exhibited marked mitochondrial dysfunction and increased interferon-responsive and inflammatory programs, including CXCL9/10 expression. T cells expanded with age and displayed heightened interferon-γ signaling, cytotoxicity, activation, and exhaustion. Spatial transcriptomics identified an age enriched CXCL9⁺ myeloid niche associated with increased T cell recruitment and activation. A machine learning-based lung aging clock outperformed bulk RNA data-based models and highlighted interferon and oxidative phosphorylation modules as key predictors of biological age.

Conclusions

Human lung aging is characterized by coordinated, multicellular alterations involving impaired alveolar regeneration, mitochondrial dysfunction, chronic inflammation, and aberrant immune crosstalk. Integrating single-cell and spatial transcriptomics enables robust identification of aging mechanisms and biomarkers, providing a foundation for improved risk stratification and targeted interventions in age-related lung diseases.

This abstract is funded by: NIH grant R01 AG089078-01A1