Background: Streptozotocin (STZ) induced diabetes in RAG-/- mice has become a standard in vivo model for testing human islet function. However the model is limited by significant STZ off target affects and poor engraftment of human cells. Immunocompromised NRG (NOD.Cg-Rag1tm1Mom Il2rgtm1Wjl/SzJ) mice have improved human cell engraftment rates and radiation tolerability compared to other strains. With development of therapies that involve the co-transplantation of accessory cells such as endothelial progenitor cells (EPCs) or mesenchymal stem cells (MSCs) there is a need to demonstrate stable function of grafts in an animal model.  
Aim: To develop a stable NRG model for co-transplanting human islets and EPCs, which can be humanised to host allogeneic immune cells.
Methods: To optimise diabetes induction, NRG male mice (8-12 weeks) were treated with varying doses of STZ (150-170mg/kg), where diabetes was defined by 3 consecutive blood glucose readings (BGL)  of  > 16.6mmol/L. Diabetic mice were transplanted with either  human islets alone (n=4, 2000IEQ)  or  co-transplanted with accessory vasculogenic human EPCs (n=4) under the kidney capsule of  diabetic mice. BGLs were monitored for 28 days post-transplant for reversal of diabetes. Graft function was assessed by intraperitoneal glucose tolerance test (IPGTT; 2g/kg glucose) on cured mice displaying two consecutive BGL<11.1mM. NRG engraftment of human CD45⁺ cells was initially determined in non-transplanted mice, where allogeneic PBMNCs were injected IP and monitored for 28 days.  Flow cytometry was used to determine the level of engraftment in blood and spleen. In separate islet transplant experiments the rejection model was verified where islet only transplanted mice (2000IEQ) were injected (IP) with an optimised dose of 10x10⁶ allogeneic PBMNCs 15 days post islet transplant, and monitored for daily BGL and the reversion to hyperglycaemia.
Results: 150mg/kg STZ dose induced a low diabetes rate (40%) compared to 160 or 170mg/kg, which both showed a 100% diabetes rate within 4 days.  Subsequent experiments used 160mg/kg dose to induce diabetes. Mice transplanted with islets alone showed a 75% cure rate by day 28 (n=4). Co-transplantation of EPC did not affect the capacity of the islets to reach normoglycemia (75% cure rate, n=4). Additionally mice transplanted with either islets alone or co-transplanted with EPC displayed normal glucose clearance when challenged with a bolus dose of glucose following a 4 hr fast. Mice injected with 10x10⁶ showed a high rate of engraftment of 88%. We saw no significant weight loss or GVHD development over the study period. Allogeneic islets were rejected (BGL>16.6mmol/L) by 10x10⁶  PBMNC in 21 days (±SEM 4.5days, n=3).
Conclusion: NRG mice have good tolerance for STZ diabetes induction and serve as a stable pre-clinical humanised islet transplant model, which is able to tolerate the presence of EPCs without affecting the cure rate. In addition high engraftment of human immune cells and a lack of GVHD symptoms in our 4 week study period will enable this model to be useful in testing novel immuno-therapies and therapies that use accessory cells.