Successful treatment of type 1 diabetes must overcome two problems: the first is the ability to engraft or expand an adequate islet cell mass and the second is the attenuation of likely immune reactivity either due to histocompatibility discrepancies (via allogeneic transplantation) or autoimmunity should autologous replacement methods be found more effective. With the continual refinement of safe and effective methods of immunosuppression, it appears unlikely that immune reactivity will be the major impediment to successful treatment (36
Human fetal pMSC were chosen as we expected them to generate islet activity due in part to fetal SCs increased expansion and differentiation potential over their adult counterparts (7
) and the fact that they exhibit similar telomerase activity to ESC but pose a much with lower risk of malignant transformation (26
). We isolated a population of pMSC from human fetal pancreas using established procedures that rely on the MSC adherence to plastic, resistance to trypsinization during passaging, expression of specific cell surface markers, and multi-lineage differentiation. Our three-step culture system was initiated with a heterogeneous pancreatic cell population that consisted of endothelial cells, mesenchymal cells and a small proportion of hemopoietic SC that were lost following culture of the adherent population in restrictive medium.
The putative pMSC obtained after culture exhibited morphologic and phenotypic characteristics similar to MSC derived from bone marrow (Figures -; 23
). Previous studies have shown that fetal MSC express HLA class I but not HLA class II antigens (37
). Accordingly, HLA class II expression in the pMSC population was lost after culture. Transcription factor profiles presented in evaluated gene expression pertinent to the pancreas or to MSC. Pancreatic MSC are reported to express both nestin and Pdx1 markers (22
). Our MSC population expressed nestin but not Pax6 or Pdx1 (early pancreatic development marker) suggesting that this population is not committed to the pancreatic lineage (). Our pMSC population did not express the ESC marker, nanog, either (37
). The presence of the hepatocyte growth factor receptor (c-Met; ) was noted in both the pancreatic tissue and the pMSC population. c-Met is reported to be expressed in early pancreatic progenitors, β-cells, HSC and BM derived MSC (41
). The pMSCs demonstrated multi-lineage differentiation into osteocytes, adipocytes and chondrocytes (). These results lead us to believe that our three-step culture system yielded a relatively pure, but not clonal, pMSC population that might be suitable for transplantation.
We then transplanted the human pMSC population into fetal sheep during the transplant receptivity / tolerance phase of gestation. The engraftment frequency (chimeric index) was 79% with functional engraftment at 50% (62.5% of chimeric sheep). This was seen without manipulations designed to improve graft expression (10
). Serum C-peptide concentrations in some sheep approached basal levels seen in humans 1.86-4.49 ng/mL (46
). In situ
staining for human insulin was noted up to 27 months following transplantation (Figures , ). This supports the capability of our non-injury xenograft model to allow/promote engraftment and differentiation of human pMSC into human insulin secreting structures that remain functional years after transplantation. This same SC population was also noted to engraft in fetal liver, express human albumin and differentiate into the hematopoietic lineage in some animals (data not shown).
While it is well established that MSC differentiate into multiple lineages in vitro
, an alternate explanation namely cell fusion between donor and recipient cells has been proposed to explain multi-lineage differentiation in vivo
. A series of bone marrow transplant studies using different gene markers identified fusion of donor MSC with multiple lineages in vivo
rather than differentiation as the mechanism underlying multi-lineage donor expression (47
). Our and others formal investigation of this issue using xenograft systems (including human MSC transplantation in utero
) failed to demonstrate significant cell fusion to explain multi-lineage differentiation, including hepatocyte differentiation (48
). It should be noted that neither study specifically investigated chimeric pancreas. While the anti-human insulin antibody is human specific (Millipore product information
), we cannot confirm that these cell clusters are uniformly human and contain requisite islet cell components. Studies are ongoing in our laboratory to investigate these possibilities. In addition, while we demonstrate human insulin in situ
and in the circulation, we did not determine if circulating insulin is derived from the pancreatic circulation. This leaves open the possibility that ectopic production of some or all of the observed human insulin is occurring in these sheep.
Our preliminary observations support the possibility that human pMSC will engraft and differentiate into functioning islets. Reports demonstrating pluripotency of stem/progenitor cells, including differentiation into islet phenotypes in vitro
and in vivo
are supportive of this conclusion (41
). Further investigation is required to prove that the full complement of cell phenotypes and requisite human endocrine islet function is possible using our technique.
In summary, we isolated a mesenchymal stem cell population from human fetal pancreas. Following transplantation of these pMSC in utero, evidence of pancreatic endocrine engraftment, differentiation and islet cell function is presented. These findings support further feasibility studies on mesenchymal stem cells’ potential to differentiate into functioning human islets in vivo.