Enhanced Spatial Transcriptomics Analysis of Mouse Lung Tissues Reveals Cell-Specific Gene Expression Changes Associated with Pulmonary Hypertension

Abstract

Spatial transcriptomics technologies have emerged as powerful tools for understanding cellular identity and function within the natural spatial context of tissues. Traditional transcriptomics techniques, such as bulk and single-cell RNA sequencing, lose this spatial information, which is critical for addressing many biological questions. Here, we present a protocol for high-resolution spatial transcriptomics using fixed frozen mouse lung sections mounted on 10X Genomics Xenium slides. This method integrates multiplexed fluorescent in situ hybridization (FISH) with high-throughput imaging to reveal the spatial distribution of mRNA molecules in lung tissue sections, allowing detailed analysis of gene expression changes in a mouse model of pulmonary hypertension (PH). We compared two tissue preparation methods, fixed frozen and fresh frozen, for compatibility with the Xenium platform. Our fixed frozen approach, utilizing a free-floating technique to mount thin lung sections onto Xenium slides at room temperature, preserved tissue integrity and maximized the imaging area, resulting in high-fidelity spatial transcriptomics data. Using a predesigned 379-gene mouse panel, we identified 40 major lung cell types. We detected key cellular changes in PH, including an increase in arterial endothelial cells (AECs) and fibroblasts, alongside a reduction in capillary endothelial cells (CAP1 and CAP2). Through differential gene expression analysis, we observed markers of endothelial-to-mesenchymal transition and fibroblast activation in PH lungs. High-resolution spatial mapping further confirmed increased arterialization in the distal microvasculature. These findings underscore the utility of spatial transcriptomics in preserving the native tissue architecture and enhancing our understanding of cellular heterogeneity in disease. Our protocol provides a reliable method for integrating spatial and transcriptomic data using fixed frozen lung tissues, offering significant potential for future studies in complex diseases such as PH.

Keywords: Arterialization; Endothelial cells; Fixed frozen tissue; Mesenchymal transition; Pulmonary hypertension; Spatial transcriptomics; Xenium platform.

Figure 1.

Protocol outline for fixed frozen mouse lung thin section mounting on a Xenium slide. (A) Protocol flowchart. Control (WT) and pulmonary hypertension (PH, Egln1^Tie2cre, KO) models were generated and validated. Mouse lungs were collected after intracardial perfusion with 4% PFA. Thin cryosections were mounted on a Xenium slide and run on a Xenium analyzer with prebuild gene panels, followed by bioinformatics analyses. (B) Photomicrographs illustrate steps of mouse lung OCT embedding, thin cryosectioning and mounting on the Xenium slide. Tissue sections were further stained with HE to visualize tissue cytoarchitectures. (C) In comparison, the 10X Genomics validated fresh frozen tissue cryosectioning and mounting leads to tissue folding and insufficient utilization of the imaging area.

Figure 2.

Initial analyses and clustering of lung samples. (A) Overview of tissue section layout on a Xenium slide. Five sections (two control and three Egln1^Tie2cre/KO) were mounted in the imaging area. (B) Each tissue section was delivered as a separate dataset, which was plotted on both spatial coordinates (left) and UMAP embeddings (right). (C) Selected top gene markers from the 23 unsupervised clusters, with expression levels denoted by heatmap. The data was generated from 5 Xenium spatial data. (D) Heatmap of selected gene expressions further organized by sample ID and unsupervised Clusters. The data was generated from 5 Xenium spatial data.

Figure 3.

Cell staining based segmentation and spatial distribution of selected transcripts. (A) Overview of the 5 stained mouse lung thin tissue sections mounted on a Xenium slide. (B) Spatial plotting of detected cell clusters in one tissue section (Ctrl #2). (C) Enlarged view of the boxed area in (B). (D) Enlarged view of boxed area in (C), with detected cell types and boundaries overlaid on the stained tissue image. (E) The same field of view in (D), with detected transcript locations overlaid onto H & E-stained bright field image. Cell type annotation and their top gene markers were listed to the right. (F) Bright field view of a remodeling pulmonary artery from KO mice. (G) Fluorescent stain of the same region as in (F). (H) Cell segmentation and overlaid selected transcripts. (I) Enlarged view of boxed area in (H). (J) Bright field view of distal alveolars from WT mice. (K) Fluorescent stain of the same region as in (J). (L) Cell segmentation and overlaid selected transcripts of the same region in (J,K).

Figure 4.

Customized analysis of Xenium dataset using Scanpy/Squidpy pipeline. (A) Spatial plot of four tissue samples with annotated lung cell types. (B) UMAP embedding of all the cells integrated after unsupervised clustering (n = 4). (C) Annotated cell types from the same UMAP space as in (B) (n = 4). (D) Sanky plot showing cell type transition from the 28 unsupervised clusters to the selected top 20 most abundant, annotated cell type clusters. Cells with less than 100 were excluded for Sanky analysis. (E) Heatmap revealed enriched top marker genes for each of the annotated cell types (n = 4). (F) Histogram distribution on the total number of 300,848 cells assigned to the predicted 40 annotated cell types (n = 4).

Figure 5.

Clustering and cell type annotation using the Seurat pipeline. (A) UMAP embedding of individual dataset after cell annotation. (B) Spatial plot of WT and KO samples with annotated lung cell types. (C) Cell proportion analysis showed the increase of AF1, AF2, SCMF, AEC, EPC and a reduction of CAP1, CAP2, Treg, CD4 and CD8 T cells in KO mice (WT n = 2, KO n = 2). (D) Statistical analysis of the cell proportional change in KO mice. (WT n = 2, KO n = 2). (E,F) Differentially expressed genes (DEGs) on lung cell types between WT and KO mice. (WT n = 2, KO n = 2).

Figure 6.

Spatial plot showing the change of EC subpopulation in PH mice. (A) Spatial plots showing the arterial ECs (AEC) and endothelial progenitor cells (EPC) were increased in KO mice (n = 2). (B) Quantification of AECs and EPCs cell proportions in (A). (C) Spatial plots showing the general capillary ECs (CAP1) and aerocytes (CAP2) were reduced in KO mice (n = 2). (D) Quantification data of AECs and EPCs in (C). (E) Spatial and Violin plots showing the increased AEC marker Sox17, decreased CAP1 (gCap) marker Plvap, and decreased CAP2 (aCap) marker Prx in the KO mice (n = 2). *** p < 0.001, **** p < 0.0001. (F) Immunostaining and RNASCOPE analysis validated the reduction of CAP2 (Ednrb⁺/CD31⁺) in KO mice (WT n = 4, KO n = 3). * p < 0.05.