Abstract The intestinal epithelium normally acts as a barrier and coordinates immune responses. In Ulcerative colitis (UC), this barrier is compromised, and the epithelial-immune communication becomes dysregulated, leading to persistent inflammation and impaired repair. Understanding how crosstalk changes over time or under different exposures is critical for identifying new therapeutic targets and the optimal window for intervention. While advances in single-cell and spatial transcriptomic technologies allow us to measure gene activity in individual cells and map their location in tissue, most existing cell-cell communication inference tools fail to infer how cellular crosstalk effects change dynamically as disease progresses. To address this gap, we developed DynamicCC, a cell-cell communication inference method that tracks signaling shifts over time and across conditions (Figure 1). Unlike conventional approaches, DynamicCC uses the full distributions of ligand-receptor expression (both mean & variance) to capture systematic and dynamic changes in signaling. It also integrates spatial information of the cells and enables the detection of ordered shifts in signaling as inflammation develops, peaks, and recovers. We applied DynamicCC to a high-resolution MERFISH dataset of dextran sodium sulfate (DSS)-induced colitis in mice. The dataset included ∼1.35 million cells across three stages: healthy (day 0), early inflammation (day 3), and peak inflammation (day 9). Using this approach, we uncovered dynamic signaling between epithelial, fibroblast, and immune compartments that conventional approaches overlooked. For example, signaling between Inflammation Associated Epithelial (IAE) and Fibroblast (F) via Cxcl12-Itga5 spiked sharply at day 3 before declining. This transient interaction triggered fibroblasts to secrete TGF-β, IL-6, and other cytokines, which further recruited immune cells (e.g., neutrophils, regulatory T cells) and reinforced a pro-inflammatory microenvironment. Other interactions showed distinct temporal patterns. Macrophage-T cell signaling (e.g., Col1a2-Cd44, Tgfb1-Tgfbr2) and Fibroblast-T cell crosstalk (e.g., Col1a2-Itga1, Cxcl12-Itga4) shifted over time, with some mediators strengthened progressively from day 0 to day 9 and others followed U-shaped dynamics, with early suppression followed by late reactivation. These temporal shifts converged on transcriptional programs regulated by Activator Protein 1 (AP-1) and Nuclear Factor (NF-κB), consistent with stage-specific increases in VEGFA, FOS, and ITGB5/6, which modulate inflammatory cytokine production. Overall, the results demonstrate that UC progression is not defined by static signaling events. Instead, DynamicCC captures the evolving communication patterns that orchestrate fibroblast remodeling and immune regulation, providing critical insights for therapeutic discovery. Figure 1:This figure illustrates the process of analyzing dynamic cell-cell communication over time. It begins with (i) a group of interacting cells (Epithelia, T-cell, Fibroblast, Macrophage cells) that are communicating through secreted ligands (triangles) binding to their respective receptors (Y-shaped structures). This communication is then measured at multiple time points (ii), t = 1, t = 2, and t = 3, using single-cell RNA sequencing or spatial transcriptomics to quantify the expression of ligands and receptors. Finally, the developed computational method, DynamicCC, processes this time-series data to generate a plot (iii) that shows how the communication strength between specific signaling between two cell types changes over time.
Egbon et al. (2026) studied this question.