INTRODUCTION

Implants of human pluripotent stem cell (hPSC)-derived pancreatic endoderm (PE) can form a functional beta cell mass (FBM) in immune-compromised mice which can protect against streptozotocin-induced hyperglycemia (1). This formation also occurs in device-encapsulated implants (2,3,4). Device-encapsulated PE cells are under clinical investigation in type 1 diabetes patients using ViaCyte’s PEC-01 cells.

We investigated devices from mice that had received a subcutaneous implant of PEC-01 encapsulated in a closed device (3,5). The number of beta cells isolated from devices correlated with the glucose-inducible plasma human (hu)-C-peptide levels. At least 20 weeks were needed to establish a FBM that achieved glucose control (5). Analysis of devices also identified other (non)-endocrine cells of donor origin within the inner membranes which formed a barrier to recipient cells and tissue. Local tissue reactivity was minimal in immune-compromised mice, thus allowing development and maintenance of a FBM (3, 5). This was not the case when these device-encapsulated implants were placed in Rowett nude rats, known to exhibit a stronger innate reactivity, and therefore relevant in a translational path to man. The fibrotic reactivity around the membranes is held responsible for an impaired survival of the cells. This observation led ViaCyte to modify the membranes so that capillaries can enter the inner chamber of their cell therapy product (PEC-Direct); this modification allowed beta cell formation in nude rats (VC, unpublished) and in clinical trials (6,7,8). 

AIM

In the present study we examined devices from Rowett nude rat recipients of PEC-Direct* to investigate whether their FBM correlates with the density and/or localization of ingrown endothelial cells and other recipient cells.  

METHODS

Sentinel devices of PEC-Direct were implanted in the subcutis of normoglycemic nude rats. Formation of FBM was followed by measuring basal and glucose-stimulated plasma hu-C-peptide levels over 50 weeks. At the end of follow-up, devices were retrieved and analyzed for their composition in cells of donor and recipient origin, their respective identity, density and localization. Recipient endothelial cells were identified by CD34-staining, myofibroblasts by α-SMA-staining.

RESULTS
At post transplant week 50, 5/10 rats presented glucose-induced hu-C-peptide levels >6 ng/ml as well as signs of glucose control by human beta cells. The cells in the inner chamber of the device were >60% of donor origin, of which a quarter was insulin-positive (INS+). Clustering of INS+cells led to areas with >40% INS+cells; these areas presented a high density of CD34+cells (10% on average) forming capillaries in the vicinity of the clusters. Areas with <1% INS+ cells contained only 1% CD34+cells but exhibited a high density of ingrown α-SMA+ cells; these areas were predominant in devices from rats with lower hu-C-peptide levels.

CONCLUSION

This study confirms that ingrowth of recipient endothelial cells in devices containing hPSC-PE cells can protect their capacity to form a FBM in a subcutaneous site with innate reactivity. It demonstrates their preferential localization near beta cell clusters reaching a density as in pancreatic islets. Ingrowth of recipient fibroblasts can interfere with this process.