My 33+ year career at The Jackson Laboratory has concentrated on the interaction between genetic and environmental factors that predispose inbred mouse strains to development of both forms of diabetes: autoimmune and non-autoimmune forms of Type I insulin-dependent diabetes (T1D), and Type 2 diabetes, associated with insulin resistance and obesity. With collaborators, I have also contributed an analysis of the genetic basis for differential inbred strain susceptibility to development of inflammatory bowel disease. At this writing, I am in the process of closing my active research career but will continue to be involved with diabetes resources development and management at the Laboratory. The studies described below done under the aegis of The Type 1 Diabetes Resource at The Jackson Laboratory emphasize the importance of thoroughly characterizing genetically manipulated mouse models used in diabetes research.
The Type 1 Diabetes Resource (TIDR) imports, or, as necessary, develops genetically-modified stocks of NOD mice important for diabetes research. One such NOD transgenic stock strongly expressing enhanced green fluorescent protein under control of the mouse insulin 1 promoter (commonly designated MIP-GFP, formally designated NOD.Tg(Ins1-EGFP/GH1)12Hara) was imported from Dr. Manami Hara at the University of Chicago. We found that homozygous transgene expression produced a developmental lethal with only small numbers of homozygous mice surviving to wean. All survivors developed severe diabetes within weeks of weaning that was not associated with insulitis; rather, beta cell-depleted islets indicated a defect in beta cell survival. Hemizygous mice were born in normal numbers and were diabetes-free post weaning. Indeed, development of clinical diabetes was almost completely suppressed in a cohort of MIP-GFP hemizygous females followed to 30 weeks of age. Whereas standard NOD/ShiLtJ males are typically more diabetes-resistant than are females, MIP-GFP hemizygous males showed a higher frequency of diabetes than hemizygous females. Our evidence indicates that hemizygous MIP-GFP expression impaired beta cell glucose responsiveness in an age- and male sex-dependent fashion. Plasma insulin of non-diabetic hemizygous males were marginally lower at 8 weeks, but markedly (4-fold) lower by 30 weeks Insulitis, when present at 30 weeks, varied considerably among individuals. Some pancreata (2/22 males and 3/17 females) showed widespread, invasive insulitis typical of standard NOD/Lt mice, and these mice may potentially have developed insulitis-driven T1D had they been aged for a longer period. However, pancreata from other individuals (13/22 males, 9/17 females) were completely free of intra-islet insulitis, showing only perivascular-periductular infiltrates, while the islets themselves exhibited hyperplastic/hypertrophic size increases. A third category included a combination of insulitic and insulitis-free islets (7/22 males, 5/17 females). What was most notable in the two latter classes, regardless of the presence or absence of insulitis, was the appearance of extensive peri-insular and intra-islet fibroconnective material, a phenotype not seen in standard NOD/ShiLtJ mice at any stage of the insulitic process. A final peculiarity of the stock was the finding that enriched populations of two distinct islet-reactive CD8+ T cells (AI4 and NY8.3 specificities) adoptively transferred clinical diabetes into 600R-irradiated MIP-GFP recipients at a much lower frequency than observed in standard NOD controls. Hence, high beta cell expression of the enhanced GFP clearly deviated from NOD model characteristics.
Because only one NOD-MIP-GFP line was available for study, it could not be established whether the transgene-mediated effects were due to copy number-induced toxicity or integration site mutagenesis that affected immune recognition of beta cells by down-regulating their function. However, analysis of differential diabetes protection produced in another important set of NOD transgenic stocks indicated that copy number and level of expression was the critical factor. The ability to generate beta cell-specific "knockout" mice on the NOD background required an NOD stock constitutively expressing the Cre recombinase if a Cre-loxP system is employed. To generate such a stock, a RIP2-Cre construct was microinjected directly into NOD/ShiLtJ zygotes. Four NOD-RIP-Cre-expressing lines were established, only two of which could be bred to homozygosity (Line 5 and Line 9). A significant diabetes-suppressive effect was observed in transgene homozygous Line 5 mice of both sexes that was not observed in Line 9. Line 5 was distinguished from Line 9 homozygotes by a non-mosaic beta cell expression pattern in the former that was not observed in the latter. This major difference in Cre expression was confirmed by qPCR comparison of whole pancreatic mRNA. The diabetes-suppressive effect of homozygous Line 5 transgene expression was alleviated in the hemizygous mice, whose expression was half that of homozygotes. Thus, expression level rather than site of integration likely accounts for the diabetes suppression.
With a one-year innovative grant from the Juvenile Diabetes Research Foundation, and in collaboration with Dr. Michael Grunze and Mr. Marcel Müller of the Institute for Molecular Biophysics, we have used infrared spectroscopy to indirectly measure blood glucose fluctuations in the tail of a mouse via heat emanating from the tail. We have developed a prototype instrument that establishes proof of principle. However, considerably more technical improvements will be required to overcome problems in reproducibility and to allow faster measurements.