Abstract
Despite the exciting reports regarding the ability of human mesenchymal stem cells (MSC) to differentiate into different cells of different organs, the mechanism by which this process occurs remains controversial. Several possible explanations have been put forth as an alternative to the existence of a true differentiation mechanism. We previously showed that MSC, at a single cell level, are able to differentiate into cells of different germ cell layers. In the present study, we investigated whether transfer of mitochondria or membrane-derived vesicles between cells and/or cell fusion participate in the events that lead to the change of phenotype of MSC upon transplantation (Tx). To this end, 54 sheep fetuses (55–60 gestational days) were Tx intra-peritoneally with Stro-1+,CD45−, Gly-A- MSC labeled prior to Tx with either CFSE, that irreversibly couples to both intracellular and cell-surface proteins, or DiD that efficiently labels all cell membranes and intracellular organelles, such as mitochondria. Evaluation of the recipients’ different organs started at 20h post-Tx and continued at 25,30,40,60 and 120h. MSC reached the liver at 25h post-Tx (0.033%±0.0) with maximal engraftment at 40h (0.13%±0.02). MSC were first detected in the lung (0.028%±0.0) and brain (0.034%±0.0) at 30h and 40h respectively. In the brain, engraftment peaked at 60 hours post-Tx (0.08%±0.0) and in the lung at 120h (0.09%±0.01). Normalization of the number of engrafted cells per tissue mass and number of Tx cells revealed that 26% of the Tx MSC reached the lung; 2% the liver; and 3% the brain. Since the decreasing number of CFSE+ and DiD+ cells detected after 120h could be due to cell division, Ki67 staining was performed and revealed that 85–95% of the engrafted cells proliferated upon lodging in the organs, and divided throughout the evaluation period. To determine MSC differentiative timeline, confocal microscopy was performed to assess whether CFSE+ or DiD+ cells expressed tissue-specific markers (MSC were negative for these markers prior to transplant) within the engrafted organs. In the liver at 25h post-Tx, all CFSE+ or DiD+ cells co-expressed alpha-fetoprotein, demonstrating the rapid switch from an MSC to a fetal hepatocyte-like phenotype. In the lung, co-localization of pro-surfactant protein and CFSE/DiD was first detected at 30h post-Tx, but cells remained negative for Caveolin1; a phenotype that is consistent with differentiation to a type II epithelial cell, but not to a more mature type I. In the brain, MSC expressed Tau promptly, but synaptophysin expression was not detected until 120h. In situ hybridization on serial sections using either a human- or sheep-specific probe, with simultaneous visualization of CFSE+ or DiD+ cells allowed us to show that no membrane or mitochondrial transfer had occurred, since none of the sheep cells contained CFSE or DiD, and all of the dye+ cells hybridized only to the human probe. Furthermore, this combined methodology enabled us to determine that differentiation to all of the different cell types had occurred in the absence of cell fusion. In conclusion, MSC engraft multiple tissues rapidly, undergo proliferation, and give rise to tissue-specific cell types in the absence of cellular fusion or the transfer of mitochondria or membrane vesicles.
Efficient generation of human hepatocytes by the intrahepatic delivery of clonal human mesenchymal stem cells in fetal sheep