Single-cell RNA sequencing and spatial transcriptomics reveal the heterogeneity and intercellular communication of cancer-associated fibroblasts in gastric cancer

Abstract

Background: Gastric cancer is a highly aggressive malignancy characterized by a complex tumor microenvironment (TME). Cancer-associated fibroblasts (CAFs), which are a key component of the TME, exhibit significant heterogeneity and play crucial roles in tumor progression. Therefore, a comprehensive understanding of CAFs is essential for developing novel therapeutic strategies for gastric cancer.

Methods: This study investigates the characteristics and functional information of CAF subtypes and explores the intercellular communication between CAFs and malignant epithelial cells (ECs) in gastric cancer by analyzing single-cell sequencing data from 24 gastric cancer samples. CellChat was employed to map intercellular communication, and Seurat was used to integrate single-cell sequencing data with spatial transcriptome data to reconstruct a comprehensive single-cell spatial map. The spatial relationship between apCAFs and cancer cells was analyzed using multicolor immunohistochemistry.

Results: Cells were categorized into nine distinct categories, revealing a positive correlation between the proportions of epithelial cells (ECs) and fibroblasts. Furthermore, six fibroblast subpopulations were identified: inflammatory (iCAFs), pericytes, matrix (mCAFs), antigen-presenting (apCAFs), smooth muscle cells (SMCs), and proliferative CAFs (pCAFs). Each of these subpopulations was linked to various biological processes and immune responses. Malignant ECs exhibited heightened intercellular communication, particularly with CAF subpopulations, through specific ligand-receptor interactions. High-density regions of CAF subpopulations displayed spatial exclusivity, with pericytes serving as a source for iCAFs, mCAFs, and apCAFs. Notably, malignant ECs and apCAFs showed increased interactions, with certain ligand-receptor pairs potentially impacting the prognosis of gastric cancer. Multiplex immunohistochemistry (mIHC) confirmed the close spatial proximity of apCAFs to cancer cells in gastric cancer.

Conclusion: Our study provided a comprehensive characterization of CAF heterogeneity in gastric cancer and revealed the intricate intercellular networks within the TME. The identified CAF subpopulations and their interactions with malignant cells could serve as potential therapeutic targets.

Keywords: Cancer-associated fibroblasts; Cell–cell interactions; Gastric cancer; Single-cell transcriptomics; Tumor microenvironment.

Fig. 1

Single-cell transcriptomics revealed distinct cellular compositions in gastric cancer samples. A UMAP diagram shows the source of cells. B Cell grouping results. C Cell annotation results. D Marker gene expression. E Proportion of cells in each sample. F InferCNV results. G Proportion of malignant cells in each sample

Fig. 2

Subclustering and annotation of fibroblasts. A UMAP diagram shows the origin of fibroblasts. B Cell grouping results of fibroblast subpopulations. C Annotation results of fibroblast subpopulation cells. D The proportion of fibroblast subpopulation cells in each sample. E The expression of marker genes in each subpopulation of fibroblasts. F Heatmap showing pathway activity scored by AUCell in each CAF subtype. G Scatter plot showing the regulon specificity score (RSS) of transcription factors in each fibroblast subtype, with the top five transcription factors highlighted. H Trajectory analysis results for each CAF subtype

Fig. 3

Cell–cell interactions. The heat map shows the number (A) and intensity (B) of interactions between cells. The vertical axis represents cells that send signals, and the horizontal axis represents cells that receive signals. The darker the color, the stronger the intensity. The heat map shows the pathways involved in cell incoming signals (C) and outgoing signals (D). The horizontal axis represents the cell type, and the vertical axis represents the signal pathway. The darker the color, the stronger the intensity of the incoming (outgoing) pathway. E Ligand-receptor pairs of epithelial cells to CAFs (efferent signals from epithelial cells, incoming signals from CAFs). F Ligand-receptor pairs from CAFs subpopulations to epithelial cells (efferent signals from CAFs, incoming signals from epithelial cells) cell incoming signal)

Fig. 4

Spatial annotation and trajectory analysis results of gastric cancer tissue sections. A–C, Spatial annotation results of gastric cancer tissue sections. D–F, Spatial distribution characteristics of CAFs subpopulations. G–I, Spatial distribution characteristics of pseudo-timeline analysis of CAFs subpopulations. J Results of pseudo-time analysis of CAFs subpopulations in spatial slices

Fig. 5

Effects of CAFs on the tumor microenvironment through paracrine signaling. A Heat map shows the interaction between cells in the spatial slice of gastric cancer2 sample. B Key ligand-receptor pairs of malignant cells and apCAFs, the abscissa is the corresponding slice information. C Integrated ranking of ligand-receptor interactions (LRIs) from malignant epithelial cells to apCAFs in 9 tissue sections. The results of GO (D), KEGG (E), and Hallmark (F) enrichment analysis. G The overall survival rate of gastric cancer patients in the low and high risk groups

Fig. 6

Multiple immunofluorescence revealed the spatial proximity of apCAFs (double-positive for CD74 and PDGFRα) and tumor cells (Pan-CK)