Islet cell replacement and transplantation immunology in a mouse strain with inducible diabetes

Generation, genetics, and immunological phenotypes of B6 RIP-DTR mice

RIP-DTR mouse strains were initially generated and maintained on mixed B6/CBA genetic backgrounds for studies of beta cell regeneration7,10,11. To make an inbred strain of RIP-DTR mice for transplantation studies, we repeatedly backcrossed to B6.SJL-Ptprca Pepcb/BoyJ mice, commonly referred to as Pep Boy or B6 CD45.1 (hereafter, B6 CD45.1), to generate an inbred line with a defined histocompatibility genotype. After backcrossing, we analyzed B6 CD45.1 RIP-DTR mice (abbreviated B6 RIP-DTR) using single nucleotide polymorphism (SNP) genome scanning analysis (Supplementary Table S1) to measure relative C57BL/6J identity. After four rounds of backcrossing to B6 CD45.1 and subsequent sibling matings to maintain homozygosity for RIP-DTR (N4F4 breeding scheme; “Methods”), we achieved an average 96% C57BL/6J identity at 120 single nucleotide polymorphisms (SNPs) analyzed across 20 chromosomes from 12 mice. It is not possible to obtain 100% C57BL/6J identity throughout the genome, due to the presence of multiple polymorphisms between the B6 CD45.1 mice used for backcrossing and C57BL/6J strain, most notably on chromosome 112. We achieved 98% C57BL/6J identity, if chromosome 1 is excluded from analysis. B6 RIP-DTR have multiple strain characteristics similar to the commonly used C57BL/6J mouse strain. B6 RIP-DTR mice have lifespans of at least one year, and average litter size of 7 ± 2 (n = 9). At 10 weeks of age, average male weight is 27.3 g ± 2.3 (n = 13), and average female weight is 19.7 g ± 1.5 (n = 13), similar to C57BL/6J mice13. In sum, our genotyping indicated successful generation of inbred B6 RIP-DTR mice.

B6 RIP-DTR mice exhibit 100% C57BL/6J identity on chromosome 17 (Supplementary Table S1). Chromosome 17 contains all class I and class II major histocompatibility complex (MHC) genes responsible for antigen presentation and distinguishing self from non-self14, which are inherited together through linkage15. To phenotype the MHC haplotypes of B6 RIP-DTR mice, we used flow cytometry to analyze peripheral blood lymphocytes (Supplementary Fig. S1, Supplementary Table S2). The MHC Class II gene I-A is expressed on antigen-presenting cells, and the I-Ab allele is expressed by C57BL/6J mice, while the I-Ak allele is expressed by CBA/J mice. I-Ab only was expressed by blood cells in both C57BL/6J and B6 RIP-DTR mice. I-Ek, another MHC Class II gene expressed by CBA/J blood cells, was not detectable in C57BL/6J nor B6 RIP-DTR mice. The MHC Class I gene H-2K is expressed ubiquitously, and the H-2Kb allele is characteristic of C57BL/6J mice while H-2Kk is characteristic of CBA/J mice. Flow cytometry revealed that only H-2Kb was expressed by blood cells in both C57BL/6J and B6 RIP-DTR mice. Thus, molecular phenotyping confirmed that B6 RIP-DTR mice express MHC Class I and Class II gene products characteristic of C57BL/6J mice with the MHC haplotype H2b.

RIP-DTR and CD45.1 in B6 RIP-DTR mice

We verified the presence of the RIP-DTR transgene in B6 RIP-DTR mice using PCR (Supplementary Fig. S2; “Methods”). Since this transgene was inserted in the Hprt locus of the X chromosome, female RIP-DTR heterozygotes remain normoglycemic despite 50% beta cell loss after diphtheria toxin administration, reflecting random X chromosome inactivation in β cells7. Therefore, we sought to produce female B6 RIP-DTR homozygous for RIP-DTR, and males hemizygous for RIP-DTR. To achieve this, we developed a comparative qPCR method to distinguish female heterozygotes from homozygotes (Supplementary Fig. S2). Homozygous females (2 copies of RIP-DTR) result in a one cycle threshold (CT) difference from heterozygous females with one copy of RIP-DTR. Thus, our methods reliably measured RIP-DTR copy number, and generated breeder pairs that ensure production of B6 RIP-DTR female mice homozygous for RIP-DTR and males hemizygous for RIP-DTR.

To facilitate studies in B6 RIP-DTR mice receiving allogeneic hematopoietic cell transplantation, we sought to make immune cells in B6 RIP-DTR mice readily distinguishable from allogeneic donors. Most available wild-type mouse strains express the CD45.2 epitope on immune cells, encoded by the Ptprcb allele. Thus, our breeding strategy aimed to produce B6 RIP-DTR homozygous for the mutant Ptprca allele, which encodes CD45.1, an isoform of the CD45 cell surface protein readily distinguishable from CD45.2 using flow cytometry4. To achieve this goal, we backcrossed RIP-DTR mice with B6 CD45.1 mice. We generated genotyping tools (“Methods”) to track and confirm homozygosity of the Ptprca allele. Specifically, dual endpoint qPCR was used to distinguish the Ptprca allele from Ptprcb in the initial backcrosses, and in all subsequent generations of B6 RIP-DTR CD45.1 mice (Supplementary Fig. S2). Ptprca genotyping was confirmed by flow cytometry analysis on peripheral blood showing CD45.1 expression only (Supplementary Fig. S3). To demonstrate additional advantages of this novel strain for studies that require differentiation of host and donor hematopoietic cells, we established mixed hematopoietic chimeras using B6 RIP-DTR host mice and BALB/c bone marrow donors (Fig. 1a, “Methods”). In stable mixed chimeric B6 RIP-DTR mice, host CD45.1+ immune cells, including CD19+ B cells, CD3+ T cells, and CD11b+ myeloid cells, were readily distinguished from CD45.2+ donor cells in peripheral blood by flow cytometry (Fig. 1b). Thus, our approaches generated and validated uses of inbred B6 RIP-DTR mice homozygous for the CD45.1 cell surface marker.

Figure 1
figure 1

Phenotyping of CD45 in B6 RIP-DTR mice with BALB/c mixed chimerism. (a) Schematic showing generation of mixed chimerism in B6 RIP-DTR mice (CD45.1) using bone marrow (BM) from BALB/c CD45.2 donors. (b) Representative flow analysis of peripheral blood from a B6 RIP-DTR mixed chimera at 8 weeks after hematopoietic cell transplant with 1.5 × 106 CD45.2 BALB/c donor hematopoietic cells. Live single cells are gated on CD19 to distinguish B cells, CD3 to distinguish T cells, or CD11b to distinguish myeloid cells, which are subsequently gated on CD45.1 and CD45.2 to distinguish host and donor cells.

Fully penetrant diabetes in RIP-DTR mice after diphtheria toxin injection

To measure the efficiency of diabetes induction in B6 RIP-DTR mice, we administered a single dose of diphtheria toxin (DT) intraperitoneally (i.p.) to males (n = 16) and females (n = 13) between 8 and 24 weeks of age (Fig. 2; “Methods”). Blood glucose was measured before and after injection until mice were overtly hyperglycemic (> 250 mg/dL). 100% of male mice were hyperglycemic approximately 3 days after DT injection, regardless of age (Fig. 2a). Likewise, 100% of female mice were also hyperglycemic by approximately 4 days after DT injection, regardless of age (Fig. 2b). Within 24–48 h after the onset of hyperglycemia, the health of diabetic mice deteriorated, reflected by rapid weight loss, unless provided exogenous insulin. The observed sex difference between male and females for time to diabetes onset could reflect differences in sex hormones, where estrogen confers mild protection from hyperglycemia16,17. For subsequent transplantation studies, we were able to consistently predict the timing of hyperglycemia onset for both sexes. Additionally, histology of the pancreas from RIP-DTR mice at 2 weeks and 4 months after single DT injection shows effective β cell ablation, with little to no insulin producing cells remaining (Fig. 2c). This is consistent with prior reports of extremely low β cell regeneration rate after DT-dependent ablation in RIP-DTR mice with a mixed CBA/B6 genetic background7. By contrast, histology of the pancreas from control B6 CD45.1 mice injected with DT revealed normal-appearing islets that included intact β cells and other islet endocrine cells (Fig. 2d). Thus, our studies revealed that the RIP-DTR transgene was functional in B6 RIP-DTR mice, and resulted in rapid and reliable β cell loss, and conditional diabetes induction.

Figure 2
figure 2

Diabetes in RIP-DTR mice after DT administration. (a) Non-fasting blood glucose and percentage of starting weight in RIP-DTR mice after single dose administration of DT (i.p.) in n = 16 hemizygous males and (b) n = 13 homozygous females. DT was injected on day 0 after baseline non-fasting blood glucose was recorded. Mice were between 8 and 24 weeks of age at time of injection, and no exogenous insulin was administered. (c) Representative histology of pancreas taken at 2 weeks and 4 months from B6 RIP-DTR or (d) B6 CD45.1 mice given single dose of DT. After confirming hyperglycemia on two consecutive days, we maintained the health of diabetic B6 RIP-DTR mice up to the 2-week timepoint by providing 40 U/kg exogeneous insulin daily. To maintain diabetic B6 RIP-DTR mice to 4 months, insulin pellets (“Methods”) were administered subcutaneously after confirmation of hyperglycemia on two consecutive days. INS Insulin, SST somatostatin.

Islet transplantation in B6 RIP-DTR mice

To validate the use of B6 RIP-DTR mice for studies of immune tolerance, we transplanted diabetic mice with wild-type H2b C57BL/6 islets, H2k CBA/J islets, or MHC-mismatched H2d BALB/c islets, and subsequently monitored blood glucose (Fig. 3a; Supplementary Table S2). Six B6 RIP-DTR mice, at 10–24 weeks of age, were injected with DT, and after hyperglycemia onset were transplanted in the renal sub-capsular space with islets from C57BL/6 donors (Fig. 3b). By 3 days after transplantation, all recipients became euglycemic without supplemental insulin. To confirm the functional status of the transplanted islets, grafts were removed from three mice at 10–12 days, and histology showed intact insulin+ β cells without CD3+ T cell infiltration and little CD45+ immune cell infiltrate (Fig. 4a, Supplementary Fig. S4). The remaining three mice remained euglycemic for at least 56 weeks following islet transplantation (Fig. 3b). Reversion to hyperglycemia invariably followed nephrectomy of the islet graft-bearing kidney at 56 weeks, confirming long-term tolerance of the C57BL/6 islet graft as expected. Furthermore, histology of islet grafts after 56 weeks showed intact insulin producing β cells without detectable CD3+ T cell infiltration and little CD45+ immune cell infiltrate (Fig. 4b; Supplementary Fig. S4). Pancreatic histology in B6 RIP-DTR mice at 2 weeks or 1-year after DT injection showed persistent β cell ablation, in contrast to healthy pancreas from DT-injected B6 CD45.1 mice (Fig. 4e–g). In summary, C57BL/6 islets are functional after transplantation and therefore durably tolerated as congenic in the novel B6 RIP-DTR mouse strain.

Figure 3
figure 3

Islet transplantation in diabetic B6 RIP-DTR mice. (a) Schematic of congenic and allogeneic islet transplantation in diabetic B6 RIP-DTR mice. (b) Male RIP-DTR mice (n = 6; age 10–24 weeks) were injected with DT on day 0 and transplanted with C57BL/6J islets on day 4 after confirmation of hyperglycemia. Mice (n = 6) were monitored for about 2 weeks post-transplant. Mice (n = 3) were monitored for approximately another 1 year, until nephrectomy was performed at 56 weeks post-transplant to remove the islet graft. (c) Male RIP-DTR mice (n = 3; age 24 weeks) were injected with DT on day 0 and transplanted with allogeneic CBA/J islets on day 4 after confirmation of hyperglycemia. Mice were monitored for repeat hyperglycemia for up to 14 days. (d) Female and male RIP-DTR mice (n = 5; 4F, 1M, age 10–24 weeks) were injected with DT on day 0 if female and day 1 if male and transplanted with BALB/c (allogeneic) islets on day 5 after confirmation of hyperglycemia. Mice were monitored for repeat hyperglycemia for up to 14 days. In all cohorts, a single dose of exogenous insulin (40 U/kg) was administered on the morning of islet transplantation following confirmation of hyperglycemia. No further exogenous insulin treatment was provided.

Figure 4
figure 4

Histological assessment of transplanted and pancreatic islets. (a) Representative histology of C57BL/6J islet graft at 2 weeks and (b) 1-year post-transplant into B6 RIP-DTR mice. (c) Representative histology of CBA/J islet graft at approximately 2 weeks post-transplant into B6 RIP-DTR mice. (d) Representative histology of BALB/c islet graft at approximately 2 weeks post-transplant into B6 RIP-DTR mice. (e) Representative histology of B6 RIP-DTR pancreas from animals transplanted with C57BL/6J islets at 2 weeks and (f) 1 year after single DT injection. No exogenous daily insulin was necessary due to the function of transplanted islets. (g) Representative histology of B6 CD45.1 pancreas (wild type) at 2 weeks after single DT injection. INS Insulin, CD3 CD3+ T cells, GCG glucagon.

To verify that B6 RIP-DTR mice are no longer immunologically compatible with the prior mixed B6/CBA background, we transplanted CBA/J islets into three B6 RIP-DTR mice after administration of DT and confirmation of hyperglycemia (Fig. 3c). After transplantation and reversion to euglycemia, none of these mice remained euglycemic beyond two weeks, consistent with the timeline of adaptive immunological rejection. As expected, similar outcomes were observed after transplanting third-party MHC-mismatched BALB/c islets into diabetic B6 RIP-DTR (Fig. 3d). Histologic studies for both groups revealed heavy infiltration of the islet graft site with CD3+ and CD45+ immune cells, with few or no remaining insulin+ β cells (Fig. 4c,d; Supplementary Fig. S4). Thus, our breeding scheme successfully produced H2b B6 RIP-DTR mice that are no longer compatible with H2k donors and normally reject MHC-mismatched grafts.