High-resolution mapping of the embryonic lung using single-cell and spatial transcriptomics

Introduction: The mature mammalian lung contains ~50 different cell types that are located in precise positions along the airway epithelial tree. It remains unclear how these precise spatial patterns arise during development. High-throughput single-cell RNA-sequencing (scRNA-seq) has been used to identify new cell types and cell states in the embryonic lung, but at the expense of spatial resolution. Commercial platforms have been used to conduct spatial transcriptomic analysis of the embryonic mouse and human lung, but their limited resolution and sensitivity have prevented spatial mapping at the single-cell level. Here, we optimized a microfluidics-enabled, deterministic barcoding-based sequencing (DBiT-seq) workflow to achieve high-resolution and highly sensitive spatial mapping of tissue sections at close-to-single-cell level. Our data identify new cell types that appear to be involved in spatial patterning of adjacent tissues.

Materials and

Methods: To spatially map cell types and cell states during early development of the embryonic mouse lung, we acquired lungs on embryonic days E11.5, E12.5, and E13.5 using C57BL/6J inbred mice. We optimized a microfluidics-based DBiT-seq workflow to obtain unbiased, high-resolution (down to 10 m), highly sensitive spatial mapping of transcripts throughout all tissues of the lung at these stages of development. We then integrated published scRNA-seq data with spatial transcriptomics to spatially map known cell types and uncover novel clusters and tissue-tissue interactions. We also used orthogonal methods, including in-situ hybridization, immunostaining, and EdU analysis, to further validate the accuracy of our spatial transcriptomic mapping.

Results, Conclusions, and Discussions: We took advantage of the spatial information provided by our data to map different cell types in the major tissues of the embryonic lung. The embryonic lung epithelium differentiates into Sox2-expressing airway epithelial cells located proximally, and Sox9-expressing alveolar epithelial cells located distally. We mapped the epithelial subclusters, carried out pseudotime analysis, and used our spatial barcoding to map the differentiation trajectories at each stage of lung development. This spatial pseudotime analysis revealed that the distal epithelium is in a more progenitor-like state, while the proximal epithelium is in a more differentiated state; these findings were further supported by RNA velocity analysis.
We used a similar approach to explore the notoriously heterogeneous mesenchyme of the embryonic lung. We uncovered a novel mesenchymal cluster showing high expression of Wnt2 and strong enrichment in hallmarks of proliferation, suggesting these cells are mesenchymal progenitors. These cells are located at the periphery of the lung, corroborating previous findings in the human embryonic lung suggesting that most immature progenitors are located peripherally. Spatial pseudotime and RNA velocity analyses suggest that these progenitors differentiate into immature smooth muscle, which is enriched in markers of both proliferation and smooth muscle contraction as well as actin filaments, before they become differentiated smooth muscle.
Finally, we identified a novel mesenchymal cluster that is spatially localized around the trachea and the medial boundary of the lung lobes. Differential expression and gene set enrichment analysis showed that this new mesenchymal cluster expresses high levels of genes that encode for extracellular matrix proteins, suggesting a potential mechanical role for these cells. We are currently using transgenic and tissue-specific knockout mice to define the role of these cells in embryonic lung development.
In summary, we spatially mapped the early embryonic mouse lung at the single-cell level. Our spatial analysis confirmed epithelial differentiation along the distal-proximal axis, unveiled the localization of mesenchymal progenitors at the lung periphery, and identified a novel mesenchymal cluster potentially involved in mechanical functions. These findings underscore the potency of integrating single-cell RNA-seq and spatial transcriptomics in developmental biology.

Acknowledgements (Optional): This work is funded by an NIH Director’s Pioneer Award HD111539 and PBI2 Distinguished Postdoc Fellowship