TF-FRAP reveals number of cells in a subpopulation in hMSCs determines its differentiation potential

Purpose (the aim of the study): hMSCs are of great interest in tissue engineering and regenerative medicine to treat cartilage and bone defects due to their multi-lineage differentiation potential. Decades of research has uncovered multiple factors including signaling pathways, biomaterials, microenvironments that steer the differentiation of hMSCs into a particular lineage. Although several different protocols are in place to differentiate hMSCs into chondrocytes, in vitro differentiations are still sub-optimal and donor variation poses an additional challenge in stem cell-based therapies. As such, there is a need employ quick tool to assess the differentiation potential of a donor. We used Transcription Factor – Fluorescence Recovery After Photobleaching (TF-FRAP) to map the dynamics of SOX9 and RUNX2 transcription factors and show that number of cells in a specific subpopulation in hMSCs may determine the differentiation potential of a donor.

Methods: hMSCs were isolated from bone marrow from patients with no known musculoskeletal diseases. Cells from three hMSC donors were grown on poly-l-lysine or fibronectin coated coverslips and differentiated to chondrogenic and osteogenic lineages. The differentiation potential was assessed using GAG, ALP and Oli Red O staining for chondro-, osteo- and adipogenic differentiation, respectively. The hMSCs transiently expressed SOX9-mGFP or eGFP-RUNX2. FRAP measurements were done at days 0 (undifferentiated cells). FRAP was performed in at least 40 cells per condition. Unsupervised hierarchical clustering was done to identify subpopulations with distinct SOX9 and RUNX2 dynamics. To correlate protein mobility and DNA binding with the protein activity, we quantified the expression levels of SOX9, RUNX2, PPARγ, ALPL and COL2A1 target genes (ACAN and COL2A1) and RUNX2 target genes (MMP13 and COL10) levels by RT-qPCR. Immunostaining was performed to visualize the DNA localization patterns of endogenous SOX9 and RUNX2.

Results: Hierarchical clustering revealed at least four subpopulations with a distinct SOX9 (Clusters A-D) /RUNX2 (clusters a-d) dynamics in undifferentiated hMSCs (Table 1 and 2, respectively). The number of cells per subpopulation varied among the donors. Using SOX9 dynamics revealed that a donor with a high number of cells in cluster A, i.e. higher SOX9-DNA binding and longer DNA residence time, showed higher chondrogenic differentiation ability. Similarly, using RUNX2 dynamics showed that a donor with a high number of cells in cluster a, i.e. higher RUNX2-DNA binding and longer DNA residence time, had higher osteogenic differentiation capacity. Interestingly, a donor with a equal amounts of cells per subpopulation showed a moderate differentiation potential towards all three lineages. Interestingly, some of the cell subpopulations did not differentiate at all, indicating an inability to respond to the differentiation stimuli. Moreover, these subpopulations had distinct nuclear localization patterns of SOX9/RUNX2. Cells with higher SOX9/RUNX2-DNA binding had discreet foci and the cells with lower SOX9/RUNX2-DNA binding showed diffused patterns of these transcription factors.

Conclusions: The dynamics of the transcription factors SOX9 and RUNX2 is related to the differentiation and the differentiation potential of cells toward either the chondrogenic, osteogenic, or adipogenic lineage. In the heterogenic pool of hMSCs, RUNX2 and SOX9 activity is not same among sub-populations. In some populations the TFs are highly active, while in others their activity is lower. SOX9 and RUNX2 dynamics and subsequent cluster analysis can be used to predict the differentiation potential of the donors even before differentiation. This is specifically advantageous as compared to multiple week- long experiments to assess the differentiation potential of a donor.