Extracellular vesicle bioactivity and potential for clinical development are determined by mesenchymal stromal cell clonal subtype

BACKGROUND: Mesenchymal stromal cells (MSCs) have been used in numerous clinical trials but very few reach phase 3 or market authorisation. Progress is often hampered by the use of non-clonal, heterogeneous and uncharacterised MSC cultures and lack of mechanistic understanding. There is limited evidence of MSC engraftment in vivo and disease resolution may be the result of the paracrine effects of the MSC secretome, rather than the cells per se. Extracellular vesicles (EVs) are key components of the MSC secretome and there is growing interest in the use of EVs as cell-free therapies. However, like MSCs, heterogeneity can exist within any therapeutic EV pool. Here we used immortalised clonal MSC lines, termed Y201 and Y202, to examine how MSC phenotype influences EV character and function. METHODS: EVs were isolated by ultracentrifugation and characterised by nano-sizing, ultrastructural morphometric analysis, western blotting, mass spectrometry and miRNA screening. Bioactivity was determined by phosphorylation of ERK1/2, proliferation and T cell polarisation assays and using two in vivo models of inflammatory disease. RESULTS: EVs from Y201 and Y202 MSCs were morphologically similar, however, Y201 EVs were more abundant in EV biomarkers versus Y202 EVs, with an enhanced miRNA and proteomic content. Computational analysis of the Y201 EV proteome identified significant enrichment in matrix-associated proteins, predicted to contribute to an elaborate EV corona particularly abundant in RGD-containing proteins fibronectin and MFG-E8, which was confirmed by western blotting. Y201 EVs, but not Y202 EVs, significantly increased the proliferation of articular chondrocytes in a dose-dependent manner, and the proliferative effect of Y201 EVs was mediated at least in part via an RGD (integrin)-FAK-ERK1/2 axis. Both Y201 and Y202 EV subsets significantly reduced proliferative index scores of activated T cells. However, only Y201 EVs, not Y202 EVs, suppressed disease activity compared to controls in different in vivo models of inflammatory peritonitis and arthritis. CONCLUSIONS: EVs released by closely related MSC subtypes within the same heterogeneous population differ significantly in terms of cargo abundance, bioactivity, and pre-clinical in vivo efficacy. Analysis of defined EV subsets will aid mechanistic understanding and prioritisation for EV therapeutics.