Human Immune System Development and Rejection of Human Islet Allografts in Spontaneously Diabetic NOD-Rag1^null IL2rγ^null Ins2^Akita Mice

NOD-Rag1^null IL2rγ^null (NRG) (14), NOD-Rag1^null Prf1^null Ins2^Akita (10), and NOD-scid IL2rγ^null (NSG) (15) mice have previously been described. To generate NRG-Akita mice, NOD.Cg-Rag1^tm1Mom Prf1^null Ins2^Akita mice were crossed with the NOD-Rag1^null IL2rγ^null strain. Additional crosses fixed the wild-type Prf1 allele and the IL2rγ^null mutation to homozygosity while maintaining the Ins2^Akita mutation in the heterozygous state. Mice were housed in a specific pathogen-free facility in microisolator cages (15). All animal use was in accordance with the guidelines of the IACUCs of the University of Massachusetts and The Jackson Laboratory.

The phenotypes of murine cells were determined using four color analyses (15). Anti-human FoxP3 (eBioscience, San Diego, CA), anti-human CD123 and BDCA2 (Miltenyi Biotec, Bergisch Gladbach, Germany), anti-human CD11c (Biolegend, San Diego, CA), and anti-human CD235a (Coulter Immunotech, Miami, FL) were obtained as indicated. All other antibodies were purchased from BD Biosciences (San Jose, CA).

Bone marrow, spleen, and thymus cell suspensions and heparinized blood were incubated with anti-mouse FcR11b to block Fc binding (14,19). Cells were prepared for flow cytometry, and at least 50,000 events were acquired on BD Biosciences LSRII or FACSCalibur instruments (BD Biosciences) (14,15,20). For rare populations, we acquired at least 500 specific events. Data analysis was performed with FlowJo (Tree Star, Ashland, OR) software.

Handpicked BALB/c islets (21) were also dissociated into single-cell suspensions (22). Dissociated islet cell suspensions were treated with Cytofix/Cytoperm (BD Biosciences) for intracellular staining using a biotinylated anti-insulin antibody and developed with Streptavidin-allophycocyanin (SA-APC). Dissociated preparations contained a mean ± SD of 62 ± 19% (n = 10) insulin-positive cells. Islets or dissociated cells were transplanted intrapancreatically or subrenally into STZ diabetic NSG mice (150 mg/kg) or unmanipulated diabetic NRG-Akita recipients.

Human islets were obtained from Juvenile Diabetes Research Foundation Islet Isolation Centers or the National Institutes of Health Integrated Islet Distribution Program. Human islets (IEQs) (4,000) were transplanted subrenally into diabetic NSG or NRG-Akita mice. Nonfasting blood glucose levels were determined twice weekly. Loss of graft function was determined by two consecutive blood glucose values >250 mg/dl.

NRG mice or progeny of NRG × NRG-Akita matings (∼50% heterozygous for Ins2^Akita) 1–3 days old were irradiated with 400 cGy and engrafted with umbilical cord blood. HSCs were provided by the University of Massachusetts Memorial Umbilical Cord Blood Donation Program under insitutional review board approval (14,20). Mice were analyzed phenotypically 12–16 weeks later and/or transplanted with human islet allografts.

Pancreata and islet transplants were stained with hemotoxlin-eosin (H-E) and for insulin, glucagon, and human CD45 (16). Unilateral nephrectomy of the graft-bearing kidney was performed on selected cohorts to confirm recurrent hyperglycemia. In some cases, islet allograft survival was inferred with histological study.

Most data are presented as means ± SD, and flow cytometry data are presented as means ± SEM. Parametric data were compared by one-way ANOVA with Bonferroni posttests to compare individual pairwise groupings. Nonparametric data were compared by a Kruskal-Wallis with Dunns posttest to compare individual pairwise groupings. Significant differences were assumed for P values <0.05. All statistical analyses were performed using GraphPad Prism software (version 4.0c; GraphPad, San Diego, CA).

To develop an immunodeficient strain of mice bearing the Ins2^Akita mutation that will support the development of a human immune system following engraftment with HSC, we generated NRG-Akita mice. Because mice homozygous for Ins2^Akita develop a rapid hyperglycemia and die shortly after weaning (10), all studies were performed using mice heterozygous for the Ins2^Akita allele.

Flow cytometry analysis of NRG-Akita splenocytes confirmed the lack of mature T, B, and natural killer (NK) cells (supplemental Table 1, available in the online appendix [http://diabetes.diabetesjournals.org/cgi/content/full/db10-0323/DC1]). Although cells with an NK phenotype (DX5⁺/CD122⁺) were observed, these cells are not mature NK cells (15). Populations of granulocytes and macrophages in NRG and NRG-Akita mice were similar (supplemental Table 1), documenting that NRG-Akita mice are deficient in adaptive immunity and NK cells.

All male and female NRG-Akita mice developed hyperglycemia between 3 and 5 weeks of age and remained diabetic, viable, and healthy in the absence of exogenous insulin (Fig. 1). In contrast, NRG (Ins2^+/+) littermates remained euglycemic throughout their life span.

Normal islet architecture and insulin and glucagon staining were observed in the pancreata of NRG mice at all ages (Fig. 1). At ∼3 weeks of age, the pancreata of NRG-Akita mice were only slightly disorganized and strong insulin staining was observed. By ∼32 weeks of age in hyperglycemic NRG-Akita mice, islet architecture was disorganized and condensed. Some insulin-positive cells were present but scattered throughout the islet. No inflammatory infiltrates were observed.

Subrenal transplantation of 20 mouse islets/g body wt into hyperglycemic NRG-Akita mice restored normoglycemia (Fig. 2). Islet graft recipients remained normoglycemic to the end of observation periods or, in selected cases, reverted to hyperglycemia following removal of the graft-bearing kidney (data not shown).

Transplantation of 4,000 IEQ human islets into the subrenal capsular space routinely restored euglycemia in NRG-Akita mice (Fig. 2). These islets remained functional throughout the observation periods—in some cases over 300 days. Graft survival was confirmed by histology or by removal of the graft-bearing kidney and reversion to hyperglycemia (data not shown).

Intrapancreatic transplantation of 20 islets/g body wt into STZ diabetic NSG mice failed to restore normoglycemia, whereas 40 islets/g body wt restored normoglycemia in two of seven recipients (supplemental Fig. 1). Using dissociated islet cells, more than 5 × 10⁵ insulin-positive β-cells was required for restoration of normoglycemia following subrenal transplantation in diabetic NSG mice (supplemental Fig. 1).

In diabetic NRG-Akita mice, intrapancreatic transplantation of 40 mouse islets/g body wt restored normoglycemia in three of four NRG-Akita mice, whereas 20 islets/g body wt restored normoglycemia in only one of three NRG-Akita recipients (supplemental Fig. 2). This is in contrast to restoration of normoglycemia in diabetic NRG-Akita recipients following subrenal transplantation of 20 mouse islets/g body wt (Fig. 2).

Intrapancreatic transplantation of dissociated mouse islet cells containing 1–4 × 10⁶ insulin-positive β-cells restored euglycemia in only one of five diabetic NRG-Akita recipients (supplemental Fig. 2). Only two of six hyperglycemic NRG-Akita mice became euglycemic after subrenal transplantation with 1–4 × 10⁶ dissociated mouse islet cells (supplemental Fig. 2). These data document 1) that dissociated mouse islets can restore normoglycemia in diabetic NSG and NRG-Akita mice and 2) that fewer insulin-positive cells are required for restoration of normoglycemia following subrenal compared with intrapancreatic transplantation.

We have shown that NRG mice engraft with human HSC at levels similar to those observed in NSG mice (14), but the effect of chronic hyperglycemia on the development of a human immune system is unknown.

NRG-Akita and NRG littermates were injected with human HSCs (14,20). Twelve weeks later, hyperglycemia was confirmed in NRG-Akita mice (data not shown). Similar percentages of splenic human CD45⁺ cells and all human cell subsets were observed in hyperglycemic NRG-Akita acompared with euglycemic NRG mice (Table 1). Comparable levels of human CD45⁺ and human cell subsets were also observed in the bone marrow, blood, and thymus of hyperglycemic NRG-Akita and euglycemic NRG mice (supplemental Tables 2–4). These data document that the human immune systems that develop in hyperglycemic NRG-Akita mice appear phenotypically comparable with those developing in normoglycemic NRG mice.

Eight of nine human islet recipients of 4,000 IEQ restored normoglycemia (<250 mg/dl) (Fig. 3) in non–HSC-engrafted diabetic NRG-Akita mice. In contrast, eight of 13 human islet allografts were rejected in HSC-engrafted NRG-Akita mice (Fig. 3). Insulin staining in the absence of mononuclear cell infiltration was detected in human islet allografts in non–HSC-engrafted NRG-Akita mice (Fig. 3). In the eight HSC-engrafted NRG-Akita human islet recipients that reverted to hyperglycemia, human CD45⁺ cell infiltration into the graft site and loss of β-cells were observed. In the five HSC-engrafted NRG-Akita mice that did not revert to hyperglycemia, insulin-positive cells and human CD45⁺ cell islet infiltration were observed (Fig. 3).

We have previously described the development of spontaneously diabetic NOD-Rag1^null Prf1^null Ins2^Akita mice (10), but NOD-Rag1^null Prf1^null mice engraft poorly with human immune systems (23). To address this limitation, we generated NRG-Akita mice that spontaneously develop hyperglycemia. Euglycemia can be restored following transplantation with mouse or human islets or with dissociated mouse islet cells. Even in the presence of a state of chronic hyperglycemia, a human immune system develops following engraftment with human HSCs that is capable of rejecting human islet allografts in some of the transplanted mice.

Unmanipulated NRG-Akita mice maintain a phenotype consistent with a severely immunodeficient mouse bearing IL2rγ and Ins2^Akita mutated genes. Both male and female mice develop hyperglycemia between 3 and 5 weeks of age, and islets progressively lose insulin-positive cells with age. However, small numbers of insulin-positive cells remain throughout life, perhaps explaining why administration of exogenous insulin is not required for long-term survival.

We confirmed that transplantation of mouse or human islets restores euglycemia in diabetic NRG-Akita and NSG mice. Higher numbers of islets were required following intrapancreatic implantation than were needed following subrenal transplantation to restore euglycemia. Our data also suggest that single-cell suspensions of insulin-producing cells may be less functional than comparable numbers contained within an islet structure.

We confirmed that a human immune system develops in diabetic NRG-Akita mice engrafted with human HSCs, and this immune system is phenotypically comparable with that generated in euglycemic NRG mice. However, only ∼60% of human immune system–engrafted NRG-Akita mice rejected human islet allografts. In separate studies, HSC-engrafted NSG mice readily reject human skin allografts (M.A.B., unpublished observations), but it is known that skin allografts prompt a more robust alloimmune response than do islet allografts (24,25). Alternatively, the human immune system may be impaired as a result of the hyperglycemic environment—or we may not have allowed sufficient time for graft rejection.

In summary, we have developed a new model of spontaneous hyperglycemia based on immunodeficient NOD mice bearing mutations in the IL2rγ gene and the Ins2^Akita gene. NRG-Akita mice can be engrafted with a human immune system, providing a recipient environment similar to that which will be encountered when transplanting allogeneic β-cells in the clinic. These mice provide a novel model system suitable for validation of the function of stem cell–derived human β-cells in the absence or presence of an alloreactive human immune system.

This work was supported by National Institutes of Health grants (AI46629, DK53006, HL077642, DK66636, DK69603, DK68854, and DK72473), the VA Research Service, an institutional Diabetes Endocrinology Research Center (DERC) grant (DK32520), Cancer Center Core Grant CA34196, the Vanderbilt Diabetes Research and Training Center (DK20593), the Beta Cell Biology Consortium (DK72473), an institutional Center for AIDS Research grant (AI042845), and grants from the Juvenile Diabetes Foundation, International, and the Helmsley Foundation. D.L.G. is a member of the UMass DERC (DK32520). Human islets were obtained from the Juvenile Diabetes Research Foundation Islet Isolation Centers or the Islet Cell Resource Consortium (now termed the Integrated Islet Distribution Program).

No potential conflicts of interest relevant to this article were reported.

M.A.B. researched data, wrote the manuscript, and reviewed and edited the manuscript. R.B. reviewed and edited the manuscript. P.D. researched data, contributed to discussion, and reviewed and edited the manuscript. J.L. researched data and contributed to discussion. J.L. reviewed and edited manuscript. A.C. researched data. C.Y. reviewed and edited the manuscript. M.H. reviewed and edited the manuscript. L.B. researched data and wrote the manuscript. B.G. researched data and wrote the manuscript. O.F. researched data. A.C.P. contributed to discussion and wrote the manuscript. D.L.G. researched data, contributed to discussion, wrote the manuscript, and reviewed and edited the manuscript. L.D.S. researched data, contributed to discussion, wrote the manuscript, and reviewed and edited manuscript.

We thank Linda Paquin, University of Massachusetts Medical School, Allison Ingalls, The Jackson Laboratory, and Greg Poffenberger, Vanderbilt University Medical Center, for their technical assistance, and Anoop Kavirayani, The Jackson Laboratory, for critical reading of the manuscript. We also thank the medical staff of the University of Massachusetts Memorial Umbilical Cord Blood Donation Program.