A division-of-labor mode contributes to the cardioprotective potential of mesenchymal stem/stromal cells in heart failure post myocardial infarction

Abstract

Background: Treatment of heart failure post myocardial infarction (post-MI HF) with mesenchymal stem/stromal cells (MSCs) holds great promise. Nevertheless, 2-dimensional (2D) GMP-grade MSCs from different labs and donor sources have different therapeutic efficacy and still in a low yield. Therefore, it is crucial to increase the production and find novel ways to assess the therapeutic efficacy of MSCs.

Materials and methods: hUC-MSCs were cultured in 3-dimensional (3D) expansion system for obtaining enough cells for clinical use, named as 3D MSCs. A post-MI HF mouse model was employed to conduct in vivo and in vitro experiments. Single-cell and bulk RNA-seq analyses were performed on 3D MSCs. A total of 125 combination algorithms were leveraged to screen for core ligand genes. Shinyapp and shinycell workflows were used for deploying web-server.

Result: 3D GMP-grade MSCs can significantly and stably reduce the extent of post-MI HF. To understand the stable potential cardioprotective mechanism, scRNA-seq revealed the heterogeneity and division-of-labor mode of 3D MSCs at the cellular level. Specifically, scissor phenotypic analysis identified a reported wound-healing CD142⁺ MSCs subpopulation that is also associated with cardiac protection ability and CD142^- MSCs that is in proliferative state, contributing to the cardioprotective function and self-renewal, respectively. Differential expression analysis was conducted on CD142⁺ MSCs and CD142^- MSCs and the differentially expressed ligand-related model was achieved by employing 125 combination algorithms. The present study developed a machine learning predictive model based on 13 ligands. Further analysis using CellChat demonstrated that CD142⁺ MSCs have a stronger secretion capacity compared to CD142^- MSCs and Flow cytometry sorting of the CD142⁺ MSCs and qRT-PCR validation confirmed the significant upregulation of these 13 ligand factors in CD142⁺ MSCs.

Conclusion: Clinical GMP-grade 3D MSCs could serve as a stable cardioprotective cell product. Using scissor analysis on scRNA-seq data, we have clarified the potential functional and proliferative subpopulation, which cooperatively contributed to self-renewal and functional maintenance for 3D MSCs, named as "division of labor" mode of MSCs. Moreover, a ligand model was robustly developed for predicting the secretory efficacy of MSCs. A user-friendly web-server and a predictive model were constructed and available (https://wangxc.shinyapps.io/3D_MSCs/).

Keywords: CD142; MSCs; division-of-labor; predictive model; scRNA-seq.