New avenues for systematically inferring cell-cell communication: through single-cell transcriptomics data

Abstract

For multicellular organisms, cell-cell communication is essential to numerous biological processes. Drawing upon the latest development of single-cell RNA-sequencing (scRNA-seq), high-resolution transcriptomic data have deepened our understanding of cellular phenotype heterogeneity and composition of complex tissues, which enables systematic cell-cell communication studies at a single-cell level. We first summarize a common workflow of cell-cell communication study using scRNA-seq data, which often includes data preparation, construction of communication networks, and result validation. Two common strategies taken to uncover cell-cell communications are reviewed, e.g., physically vicinal structure-based and ligand-receptor interaction-based one. To conclude, challenges and current applications of cell-cell communication studies at a single-cell resolution are discussed in details and future perspectives are proposed.

Keywords: cell-cell communication; chemical signal-dependent communication; ligand-receptor interaction; network biology; physical contact-dependent communication; single-cell RNA sequencing.

Figure 1

General procedures of cell-cell communication studies using scRNA-seq techniques. Human, mouse, or C. elegans samples were first dissected followed by scRNA-seq analysis to obtain the single-cell transcriptomic data, prepared for construction of cell-cell communicating network. Physical contact-dependent and chemical signal-dependent communicating networks constitute the two kinds of the cell-cell communicating network. In combination with statistical analysis, a cell-gene matrix generated from scRNA-seq protocols and a cell-cell connection matrix generated from the cell-cell communicating network are integrated to investigate cell-cell communications at a single-cell resolution. For experimental validation of inferred cell-cell communications, histological sections are evaluated to verify the physical contact-dependent cell-cell communications based on physically vicinal structure of cells, while perturbation experiments under inhibiting conditions are applied to verify the chemical signal-dependent cell-cell communications based on ligand-receptor interactions

Figure 2

Two strategies used to investigate cell-cell communications at a single-cell resolution. (A) Physically vicinal structure-based strategy according to physical contact-dependent communication. Physically vicinal cellular structures (doublets, triplets, etc.) are obtained by microdissection or microfluidics followed by processing scRNA-seq protocols. After annotation of cell types of physically vicinal cellular structures, the cell-cell communicating network is constructed combined with the cell-cell connection matrix for inference of physical contact-dependent cell-cell communication. (B) Ligand-receptor interaction-based strategy according to chemical signal-dependent communication. A ligand-receptor matrix is obtained from known ligand-receptor interactions and a cell-gene matrix is generated from scRNA-seq protocols. The matrices are integrated to construct the cell-cell connection matrix and the cell-cell communicating network using. The chemical signal-dependent cell-cell communication can be further inferred based on the constructed cell-cell communicating network

Figure 3

Current applications of cell-cell communications at a single-cell resolution. Cell-cell communication studies using scRNA-seq techniques can be applied to elucidate in-depth mechanisms underlying physiological processes (e.g., embryogenesis, homeostasis, and organogenesis), disease pathogenesis and progression (cancers, liver diseases and inflammation), or pharmacological research for efficacy and resistance

Figure 4

Challenges and opportunities of investigating cell-cell communication at a single-cell resolution. (A) Spatial reconstruction of single-cell transcriptomes from single-cell transcriptomic data without spatial location will shed light on the integration of physical contact-dependent and chemical signal-dependent cell-cell communications. (B) Incorporation of network topology and features will help infer cell-cell communications. (C) Recent advances in spatial transcriptomics at a single-cell resolution will facilitate the identification of single-cell intercellular communications in situ. (D) Establishing the comprehensive molecular view of the cell by multimodal profiling in the future will definitely benefit the inference of cell-cell communicating modes