Search

We took advantage of the proliferative and permissive environment of the developing preimmune fetus to develop a noninjury large animal model in sheep, in which the transplantation of defined populations of human hematopoietic stem cells resulted in the establishment of human hematopoiesis and led to the formation of significant numbers of long-lasting, functional human liver cells, with some animals exhibiting levels as high as 20% of donor (human) hepatocytes 11 months after transplantation. A direct correlation was found between hepatocyte activity and phenotype of transplanted cells, cell dose administered, source of cells used on a cell-per-cell basis (bone marrow, cord blood, mobilized peripheral blood), and time after transplantation. Human hepatocytes generated in this model retained functional properties of normal hepatocytes, constituted hepatic functional units with the presence of human endothelial and biliary duct cells, and secreted human albumin that was detected in circulation. Transplanting populations of hematopoietic stem cells can efficiently generate significant numbers of functional hepatic cells in this noninjury large animal model and thus could be a means of ameliorating or curing genetic diseases in which a deficiency of liver cells or their products threatens the life of the fetus or newborn.

We have previously reported the isolation of Stro-1+, CD45-, GlyA-human MSC from human bone marrow (BM), liver (LV), and brain (BR). Despite their similar morphology and phenotype, these cells, upon in vivo transplantation into the pre-immune fetal sheep model, seemed to differentiate more efficiently into cells of their tissue of origin. For instance, all 3 sources of MSC gave rise to blood, with average levels of human CD45+ cells in chimeric animals of 8.7% for LVMSC, 3.3% for BRMSC, and 4.5%for BMMSC. Human neural cells were also detected in the brain of transplanted animals with the highest levels obtained with BRMSC (1.33%) and lowest with LVMSC (0.4%). However, BRMSC generated very few human hepatocytes (2–10 hepatocytes/section), while both LVMSC and BMMSC gave rise to much higher levels of hepatocytes. This led us to hypothesize that despite the similar phenotype, the proteome makeup of MSC from different tissues may be distinct. To test this hypothesis, we examined the commonalities and differences in the protein repertoire of cultured MSC from these 3 different sources. The eighty most prominent protein spots from the 2-D gel of each cell population were removed and identified by MOLDI-TOF/TOF mass spectroscopy after Trypsin digestion. Our results demonstrate that there are a large number of common proteins between the LV, BM and BR MSC (> 75%). Besides the cytoskeletal and associative proteins such as Vimentin, β-actin, cofilin, Tropomyosin and Lamin A/C, the cells contained a large quantity of proteins for protein processing such as Cyclophilin A, Endoplasmic reticulum protein ERp29, Disulfide-isomerase ER60 and Glucose-regulated proteins 75 and 78, which may be important for the broad differentiative potential of MSC. Despite these similarities, forty percent of the identified proteins showed significantly different expression levels between the three types of cells (> 300%), the majority of which were not proteins previously shown to be BM-, Liver or Brain-specific. Interestingly, however, LV-derived cells contained a 5-fold greater concentration of Albumin than BM or Brain cells, which may explain their greater tendency to form liver tissue in vivo. Likewise, since BMMSC were enriched in smooth muscle associated proteins, potentially they may constitute a better source to be used for muscle repair. Thus, the definition of the different MSC proteomes can lead not only to the discover of critical differences involved in differentiation/plasticit into different cell fates, but may also help us to identify distinct cell populations that are ideally suited for cell therapies for specific organs.

The ability of bone marrow-derived Hematopoietic and Mesenchymal Stem Cells (MSC) to generate hepatocytes in a large non-injury sheep model of human stem cell transplantation has been demonstrated. Although in our sheep model the liver is still a hematopoietic organ at the time of transplantation, we hypothesized that the human BM-derived MSCs migrated preferentially to the BM, thus leaving few MSCs available to lodge in the liver and give rise to hepatocytes. Thus, in the present studies, we compared the ability and efficiency of clonally-derived BM MSC populations to give rise to liver cells upon either intra-hepatic (IH) or intra-peritoneal (IP) transplantation into fetal sheep. Five MSC clones were established from human adult BM by single cell deposition of Stro-1+ Gly-A-CD45- cells and expanded until sufficient cells were obtained. Phenotypic characterization demonstrated that all clones were positive for CD90, CD13, and CD29, and negative for hematopoietic markers such as CD34, CD133, and CD45. Five different clonal populations were used and each clone was transplanted either IP (n=7) or IH (n=6) into 55-60 days old fetal sheep recipients at 5x105cells/fetus. Evaluation of the recipients at 56-70 days post-transplant showed that clonally-derived MSC were not only able to generate hepatocytes, but they were also able to generate hematopoietic cells detectable both in the BM and in the PB. IP transplanted animals had a significantly higher level of human CD45+ in their BM (CD45=2.7±0.4%) compared to the IH transplanted animals (CD45=0.25±0.05%), demonstrating that IP injection favors the homing of these cells to the BM while upon IH injection, the human blood cells generated remain in circulation (2.6% CD45+). When we evaluated the presence of hematopoietic cells in the livers of these animals, the IP transplanted animals contained almost 2% CD45+ cells within their livers, while IH transplanted animals had levels of only 0.5-0.65%. When we looked at the generation of liver cells in both groups of animals by immunohistochemistry using monoclonal antibodies specific for human hepatocytes and albumin, we observed that IH injection was much more efficient at generating human hepatocytes than IP injection. The levels of human hepatocytes in most of the IH transplanted animals (10-15% positive cells) were roughly 5X higher than their IP counterparts (2.5-3%). Furthermore, human hepatocytes produced following IH injection were widely distributed throughout the liver parenchyma, while IP injection resulted in a periportal pattern of human hepatocytic activity. We also confirmed the functionality of the generated hepatocytes by detecting human albumin in circulation in these animals by ELISA. In conclusion, these studies show that in a non-injury model the number of available donor cells reaching the liver can be controlled by altering the route of administration, and that this availability determines both the levels of donor-derived hepatocytes generated and the pattern of distribution of these hepatocytes throughout the recipient liver.