Cell-Cell Communication and Gene Regulation of Stem Cell-Parenchymal Cell Fusion

Introduction: Mesenchymal stromal cells (MSCs) can fuse with parenchymal cells in various organs, including the brain, liver, small intestine, and heart. This fascinating process of cell fusion is critical for development, regeneration, and understanding cancer progression. Elucidating the mechanisms of cell fusion is thus paramount for differentiating between beneficial and detrimental fusion events. This knowledge holds promise for therapeutic strategies that promote controlled cell fusion for tissue repair and regeneration, while preventing pathological fusion. Here, we revisit previously published data (Freeman et al. Sci Rep 2016) on the fusion of mouse HL-1 cardiomyocytes (mHL1cm) and mouse MSCs (mMSC). We aim to gain deeper insights into cell-cell communication and the underlying gene networks orchestrating this specific fusion process.

Materials and

Methods: The published data (Figure 1A) is accessible from the Gene Expression Omnibus (GEO) database, (GSE69926). To identify distinguished groups of cells after fusion, we first attempted UMAP (Uniform Manifold Approximation and Projection). Using CellCall (Zhang et al., Nucleic Acids Res 2021), we analyzed the intercellular communication pathways between all possible permutations of sender-receiver pairs. Using SCENIC (Single-Cell rEgulatory Network Inference and Clustering, Aibar et al. Nat Meth 2017), we attempted to understand the gene regulatory network associated with the distinguished group of cells. Using hdWGCNA (high dimensional Weighted Gene Co-expression Network Analysis, Morabito and Miyoshi et al., Nat Genet 2021), we identified unique hub genes associated with gene regulation.

Results, Conclusions, and Discussions: Of 68 samples analyzed, we identified 20 samples of mHL1cm, 20 of mMSC, 9 of mHL1-like fusion and 19 of mMSC-like fusion. Using CellCall, we found multiple pairs of intercellular communication pathways between mHL1cm, mMSC and fusion products with some exceptions (Figure 1B). CellCall analysis also predicted that unique pathways are associated with specific type of communication between all possible permutation of cell types (Figure 1C), For example, intercellular communication of mMSC and mHL1cm triggers pathways related to tight junction molecules, endocrine resistance, and cellular senescence, while reducing Wnt signaling pathways. CellCall analysis further identified unique cell-cell communication and the associated transcription factor (TF, Figure 1D). Using UMAP, we demonstrated that fusion products can be clustered with either type of parental cells (Figure 2A). The results of SCENIC analysis, visualized in Figure 2B, demonstrate that high RSS (regulon specificity score) regulons associated with each unique group of cells are clearly separated. In addition to intercellular communication and gene regulatory network, we also analyzed gene co-expression network using hdWGCNA. We identified 37 module eigengenes (MEs) and visualized three hub genes and five other genes in a graph (Figure 3A). Our analysis revealed that each ME exhibits relative exclusivity to its corresponding parental and fusion cell types. This suggests minimal overlap in ME expression across these groups (Figure 3B). Next, we will analyze unique TFs and their target genes associated with regeneration or tumorigenesis. This analysis will focus on TFs and targets that become active as the fusion cells acquire phenotypic and functional properties following accidental cell fusion.

Acknowledgements (Optional): This research was funded in part by the NSF CAREER award (JPJ)