A106-12 A Single-Nucleus RNA-seq Atlas of Childhood Interstitial Lung Disease: Building a Foundational Resource for Pediatric Rare Lung Disorders

Rationale Childhood interstitial and diffuse lung diseases (chILD) comprise a group of rare disorders with high morbidity and mortality, limited therapeutic options, and poorly understood molecular mechanisms. Because of the rarity and heterogeneity of chILD, comprehensive molecular atlases are lacking, limiting discovery and translational progress. To address this gap, we developed the first single-nucleus RNA sequencing (snRNA-seq) atlas of chILD, integrating diverse genetic and histopathologic forms to enable systematic study of its cellular and molecular landscape. Methods With informed consent, we collected 37 lung tissue samples (biopsy, explant, or autopsy) from 32 infants and children encompassing 16 defined chILD diagnoses, five control lungs, and one chILD ‘unknown’ diagnosis. Samples were obtained through Boston Children’s Hospital institutional biobanks and national collaborations, including the BRINDL biorepository, Washington University in St. Louis, Johns Hopkins University, Children’s Hospital Colorado, Children’s Hospital of Pittsburg, and University of North Carolina. Nuclei were isolated and sequenced using 10x Genomics technology. Raw data were processed into Seurat objects, and metadata were curated for each sample, including diagnosis, genetic variant, and tissue source. Rigorous quality control included filters on unique molecular identifier counts per cell and mitochondrial RNA percentage per cell. PCA was conducted, a UMAP was generated to visualize potential batch effects, and Harmony was used to correct for batch effects. Cell type annotation was performed with Seurat’s label transfer function using two reference snRNA-seq atlases across multiple age groups and disease contexts to ensure robust classification. Cells with annotations that were inconsistent between the label transfer annotations were manually annotated and checked for biological discrepancies. Results The atlas consists of a unified atlas of approximately 341,046 high-quality nuclei representing epithelial, stromal, endothelial, and immune compartments across chILD and unaffected lung samples. Samples included individuals with variants in ABCA3, SFTPB, SFTPC, NOD2, FLNA, STAT1, and others. The inclusion of both unaffected and disease lung tissues sampled from a diverse set of chILD subtypes allows comparative analyses that can reveal convergent and divergent cellular responses to different pathogenic mechanisms. Conclusions This collaborative snRNA-seq atlas represents the largest pediatric lung disease single-nucleus resource to date and a foundational tool for the study of chILD. By providing an integrated, well-annotated dataset across diverse chILD disorders, this atlas is intended to accelerate mechanistic insights, identify candidate biomarkers of diagnosis, and inform potential therapeutic target discovery and drug repurposing for chILD. This abstract is funded by: Chan Zuckerberg Initiative, NIH U54HL165443, NIH U01HL148861