Type 1 Diabetes: Cellular, Molecular & Clinical Immunology

Chapter 3 - Animal Models of Type 1 Diabetes: Genetics and Immunological Function
George S. Eisenbarth modified from second web edition Julie Lang and Don Bellgrau

Updated 7/08, slides updated 3/08 Click to download Powerpoint slide set
Underlined references are clickable links and will take you to the reference on the PubMed website.

Introduction
NOD Mouse

Genetics of the NOD mouse

Idd1
Idd2-5
Idd6-10
Idd11
“Autoimmune” Polymorphisms

Immunology of the NOD Mouse

Regulatory Cells

Antigenic Targets of the NOD Mouse

Autoantigen Gene Knockouts

BioBreeding Diabetes-Prone (BB-DP) Rat Model

Derivation of BB Rats
Genetics of BB Rats
Immune Dysfunction in the BB-DP Rat
Disease Prevention in the BB-DP Rat

Long Evans Tokushima Lean (LETL) Rat Model (Komeda Diabetes-prone)
LEW1.AR1 Rat Model
TCR (T Cell Receptor) transgenic Mouse Models of Type 1 Diabetes 
Transgenic NeoSelf-Antigens in Beta Cells
Immunization/Immunoregulatory Induced Models of Diabetes

Islet Peptides
Co-stimulation

“Humanized” Mice

Human Histocompatibility Alleles

HLA-A2
DQ8

Human Lymphocyte Transplantation

Conclusion

Introduction
The past decade has seen remarkable advances in understanding the genetics and pathophysiology of spontaneous animal models of immune mediated diabetes (Type 1A) and the creation of new animal models (1), either with genetic manipulation of autoantigens (2;3) or with application of immunologic insights (e.g., toll receptor activation; peptide immunization, mutations of regulatory pathways such as AIRE (4) or foxP3 (5)), or a combination of genetic and environmental manipulation (6-17). “Humanized” mouse strains are being created and studied and promise additional insights relevant to type 1A diabetes of man (18-20).  In addition at the Jackson laboratory a repository of type 1 diabetes related mouse models has been created which is an important resource for the long-term preservation of the multiple strains that have been created and a means to facilitate sharing of strains by different research groups.
Spontaneous type 1 diabetes-susceptible models include the non-obese diabetic (NOD) mouse, the BioBreeding Diabetes-Prone (BB-DP) rat, the Komeda Diabetes-Prone (KDP) sub-line of the Long-Evans Tokushima Lean rat, and the Lew.1AR1/Ztm rat. Multiple experimentally-induced models of type 1 diabetes are available including: 1) T cell receptor (TCR) transgenic (Tg) mice with the T cell receptors of naturally occurring diabetogenic clones 2) Neo-antigen (Ag) expression under the control of the rat insulin promoter (RIP) to establish neo-self antigen pancreatic expression that can be the target of autoimmunity, and 3) RIP-driven expression of costimulatory molecules on beta cells enhancing diabetogenic stimuli. Mice with knockouts of putative islet autoantigens have allowed direct testing of the pathogenic significance of specific target molecules. Strains of mice with mutations of genes associated with type 1 diabetes in man (FoxP3 and AIRE) are being studied in detail even though such strains usually do not develop diabetes, but rather novel autoimmune phenotypes (21-23).
A major conclusion from these models is that type 1 diabetes is not the result of a single defined pathway but can be the result of numerous distinct mechanisms (24). Nonetheless, some generalities do exist including the polygenic nature of the disease, the involvement of T cells in the destruction of pancreatic beta cells, and the incomplete penetrance of the disease implying that environmental or “stochastic” factors influence disease susceptibility. These models provide useful tools for studying the autoimmune process as well as treatments for the prevention of type 1 diabetes. There has been some pessimism (25), appropriately expressed (8), that therapies effective in animal models such as the NOD mouse may not be directly relevant to man. No doubt the usual animal models are inbred (thus not diallelic at any locus) and finely “balanced” in terms of progression to diabetes while raised in pathogen-free environments. Thus, it is likely that many more therapies will be effective in the animal models compared to man (Entelos has developed computer model of NOD mice and catalogued therapeutics (26). Nevertheless, the initial experimental success of anti-CD3 monoclonal antibody therapy in man to delay loss of insulin secretion is a direct result of studies in the NOD mouse (27), and large trials such as the Diabetes Prevention Trial (DPT) have provided preliminary data that oral insulin (also first studied in the NOD mouse) in subjects with elevated insulin autoantibodies may delay disease progression in the subset of individuals expressing high levels of insulin autoantibodies (28). We suspect that informed optimism relative to the utility of animal models is in order (25), and that the more robust the therapy in the animal model (e.g., ability to abrogate insulitis, ability to prevent diabetes in models engineered to be more highly penetrant [e.g., insulin 2 gene knockout NOD mice (29)]), reversal of hyperglycemia, and the closer the human studies mimic the animal studies (e.g., dose and timing), the more relevant the study will be to human type 1 diabetes.
The Nonobese Diabetic Mouse
There are now more than 5,000 articles listed in PubMed on the NOD mouse making any comprehensive review of the model a difficult task with the certainty of failure to cite all important contributions and a certainty that not all the different pathways being pursued will be covered.  Given the ability to electronically access primary articles through the internet with services such as PubMed and Google my purpose of a review on such a large topic is to provide an overall organization, a particular viewpoint (with caveats), and to attempt to highlight many important “relevant” observations.  The general hypothesis my laboratory group is pursuing is that NOD mice 1. have relatively mild defects of immune tolerance (particularly in comparison to foxP3 or AIRE mutant mice) and thus the polygenic nature of the disease and the small influence of most loci; 2. that the disease results because of an immune response that once directed at a specific beta cell peptide is essentially normal, though the targeting leads to disease and there is spreading of autoimmunity, with implication that all mechanisms available to the normal immune system for destruction are operative; 3. that there is a primary autoantigenic epitope presented by the class II I-Ag7 molecule and for the NOD mouse it is the insulin B:9-23 peptide recognized primarily by a conserved alpha chain of the T cell receptor (30) and 4. the killing of islet beta cells is asynchronous with different islets destroyed over time until enough destruction has occurred for hyperglycemia to develop.  These “biases” are stated, as to some extent they influence “commissions” and “omissions” of this review that many investigators and hopefully myself with time will correct

Figure 1. Derivation of the spontaneous animal model of type 1A diabetes, the NOD mouse.

The NOD mouse is the most-studied animal model of type 1A diabetes (25;26;31-46). The NOD inbred mouse line was established at the Shionogi Research Laboratories in Japan through sib breeding of a hyperglycemic female mouse from the CTS sub line (47) (Figure 1). The line that actually became the NOD mouse strain was originally a control line for the modestly hyperglycemic line that became the NON strain (Non-obese Non-diabetic). NOD mice show islet infiltration by lymphocytes around 5-7 weeks of age followed by the spontaneous development of overt diabetes in approximately 70% of the females and 40% of the males by 30 weeks of age (48;49). Similar to humans, NOD mice usually (but not always) express anti-insulin autoantibodies in their serum prior to hyperglycemia (50-52) (Figure 2).  Of note NOR mice which do not usually develop diabetes follow the same pattern of expression of insulin autoantibodies.  NOR mice are a complex congenic inbred strain with approximately 80% NOD genome and following depletion of CD4+CD25+ T cells NOR splenocytes can mediate diabetes in NOD-scid mice (53).

Figure 2. Insulin autoantibodies and glucose of individual NOD mice followed from 4 weeks of age.

Although all NOD mice develop insulitis, this is not always followed by diabetes. Diabetes incidence varies among established NOD colonies, although in all cases, unlike in humans, gender influences disease incidence with higher disease frequencies in female mice than males (49). Type 1 diabetes incidence is also affected by environmental conditions with higher disease frequencies in sterile conditions (54) and lower disease penetrance with infectious agents (55;56). The disease can be prevented in numerous ways (>134) (25) including immunological or genetic manipulations of NOD mice (57).

Genetics of the NOD Mouse

Figure 3. Summary of NOD (idd) and human (IDDM) loci.

Diabetes susceptibility in the NOD mouse results from the interaction of multiple genes (58), with the strongest predisposing effect deriving from genes within the major histocompatibility complex (MHC) (59-61). Multiple studies suggest that the effect of the MHC (59) is due to the combination of the unique sequence of the I-Ag7 allele with its beta chain aspartic acid at position 57 and proline at position 56 (62) (the I-Aalpha chain sequence of NOD is identical to BALB/c), a lack of expression of I-E (63) due to a common mouse mutation (also present in C57BL/6 mice) and specific class I alleles (64), with additional polymorphisms of other genes within the MHC (e.g. idd16 (65;66). A great deal of effort has been involved in mapping the diabetes-associated alleles in both NOD mice and humans. In both cases, numerous loci have been identified (>20), although to date, very few predisposing genes are identified for the NOD mouse (MHC genes (59), potentially CTLA-4 (67), IL2 (58;68) and beta-2 microglobulin important exceptions (61;69;70)). Genetic analysis has involved mapping of the mouse genome usually with crossing of loci from nonautoimmune-prone strains onto the NOD background as well as analysis of recombinant and related strains. Loci (identified as Iddn in the mouse and IDDMn in the human) with significant LOD scores for association have been analyzed for candidate genes. Multiple congenic strains containing Idd locus (n) from disease-resistant strains have been generated and analyzed for disease incidence, insulitis, and more recently immunological phenotypes (e.g., insulin autoantibodies (71)) associated with diabetes in the NOD mouse (70;72;73). A general finding is that several identified regions contain multiple-disease associated loci which alone rarely confer protection but in multitudes of two or more can nearly completely block disease (70). Positional cloning of numerous Idd loci is underway with potential for success of identifying “causative” polymorphisms likely to be dependent upon the number of genes within a given locus. For example a 1.52Mb region of locus Idd5.2 contains 45 genes, while the locus Idd5.1 (2.1-Mb region) contains four genes, including CTLA-4 (58), with evidence for influence of CTLA-4 polymorphisms influencing disease in man and mouse.
A list of many of the identified Idd loci, their positions on mouse chromosomes, identified phenotypes, and candidate genes is presented in Table 3.1. These loci can either confer susceptibility or resistance to disease development and even some alleles of the NOD mouse confer resistance rather than susceptibility. Given the identification of these loci, it is possible to rapidly introduce loci from other strains onto the NOD background with what has been termed speed congenics. Usually within five generations of backcrossing, essentially all NOD diabetogenic loci can be fixed with the locus of interest introduced (74). A major disadvantage of the NOD mouse is the lack of embryonic stem cell lines to directly create gene knockouts on the NOD background, a lack that is partially alleviated with the availability of “mixed” NOD strain embryonic cell lines (75).

Table 3.1. IAA= Insulin Autoantibodies, bold genes likely candidate.

Idd1 has been mapped to the MHC locus on mouse chromosome 17 (49;59). Although the class II allele within the MHC does not completely explain the association of disease (66;76) to this locus (other non-MHC class II genes within this locus contribute to disease susceptibility including class I alleles (77;78)), the MHC class II locus is the most studied genes within this region (70). The I-Ag7 molecule behaves as a recessive allele, normally required in the homozygous state for diabetes development in the NOD model (59). Of note, both human DQ and mouse IA diabetes-associated sequences (DQB1*0302 and IAg7) carry a serine at position 57 of their b-chain instead of the aspartic acid residue conserved on other mouse strains and on other DQB1 alleles conferring low diabetes risk in humans (60;79). These findings have suggested that this particular residue may influence diabetes susceptibility, but it is likely that other amino acids within DQ or IA as well as other MHC genes may be important both in humans and mice. In fact, population studies (80) and experiments in transgenic mice (63;81) have failed to show position 57 alone as a susceptibility factor and in addition support a significant role for alleles at the DRB1 locus in humans and for the corresponding IE molecule in the mice. The NOD mouse lacks surface expression of the I-E molecule and its expression as a transgene prevents diabetes (63;81;82). These data must be interpreted with caution since control experiments with transgenic expression of the diabetogenic IAg7 molecule also protect NOD mice from diabetes (83). The Idd1 locus may act as a gene complex with at least two susceptibility loci, I-A and I-E (81) as well as class I and class III loci (66;76). It must be noted that, as with any one identified Idd locus, the combination of the I-A and I-E NOD alleles is not sufficient for diabetes development as CTS mice share the NOD class II alleles but are disease-resistant (84;85). In addition, the requirement for IAg7 homozygosity is not absolute as models of mice heterozygous for IAg7 have been found to be susceptible to type 1 diabetes (86).
Recent crystallography studies of the IAg7 molecule showed this MHC protein is structurally stable. In comparison to other class II molecules, this diabetes-associated allele contains an altered peptide binding groove which allows more promiscuous binding of numerous peptides (87;88), potentially related to earlier studies of weak peptide binding (89). Furthermore, diabetes-associated DQ alleles in humans were also found to have similarly altered peptide-binding grooves in crystallography studies (31). One possible role of weak presentation of self-peptides by MHC in autoimmunity could be inefficient thymic deletion. Support for this concept comes from limiting dilution studies which showed a high frequency of autoreactive T cells in IAg7 mice (90). In general though it appears that I-Ag7 functions normally in peptide presentation and has high affinity for a subset of peptides.  The hypothesis that there is some overarching abnormality of the class II molecules associated with autoimmunity is difficult to reconcile with the observation that some MHC haplotypes protect from one autoimmune disease, while enhancing another, such as human multiple sclerosis and type 1 diabetes (e.g. DRB1*1501, DQB1*0602) (91).
Investigations into the role of MHC on thymic tolerance processes provide evidence for both a lack of thymic deletion in IAg7 homozygous mice and positive selection of regulatory T cells by protective MHC alleles. The CD4+ anti-islet 4.1 TCR transgenic is positively selected by IAg7 but negatively selected with coexpression of other MHC class II alleles (92). In addition, Ridgway and coworkers showed that dosing the IAg7 allele correlates with the degree of insulitis and autoreactive repertoire (93). On the other hand, BDC2.5 TCR CD4+ anti-islet transgenic mice show selection of regulatory subsets by protective alleles (94). Thus, the current data suggest altered MHC alleles result in inefficient thymic deletion and altered positive selection in the NOD mouse. However, there is also evidence for a role of the IAg7 molecule in the peripheral activation of autoreactive T cells (95). A mutated form of the IAg7 molecule replacing the His and Ser at positions 56 and 57 with the more common Pro and Asp residues generates a transgenic MHC molecule known as BALBg7PD. This MHC transgenic mouse mediates positive selection of BDC2.5 TCR transgenic cells, however, these mature T cells are unable to mount an anti-islet response with BALBg7PD antigen-presenting cells (APCs) whereas responses with NOD WT APCs are adequate (96). In conclusion, the Idd1 MHC locus may contribute to type 1 diabetes both in altering thymic selection processes and facilitating activation of autoreactive T cells in the periphery. Several groups including our own have provided evidence that insulin may a primary autoantigen for the NOD mouse and in particular insulin peptide B:9-23 (2;3;97-99).  If this insulin peptide is indeed essential for the development of diabetes of the NOD mouse, the manner in which the B:9-23 peptide binds to I-Ag7 may be a crucial determinant of disease, with report of two binding registers (99).  With the lack of a crystal structure, and clear molecular definition of the B:9-23 register for pathogenic T cells utilizing a dominant T cell receptor alpha chain (Kobabyashi et al in press) recognizing the peptide further evaluation will be important.
Idd2 has been assigned to mouse chromosome 9 and linked to the T lymphocyte marker thy1, although a significant association with diabetes has not been found in all studies (100).
Idd3 Progress in gene identification and contribution to disease has been made with the Idd3 locus. NOD mice congenic for the Idd3 region of chromosome 3 from C57BL/6 mice show a low incidence of diabetes (25% compared to 80%) (101). The Idd3 locus contains the IL2 gene which has a unique glycosylation form in the NOD (102), however, the protein appears to have normal function and the glycosylation difference has been genetically ruled out as contributing to diabetes (68). Identified  idd3 candidate genes in what is now approximately a 650Kb congenic region include Tenr, IL2, IL21, and Fgf2, and Cetn4,  (77) (70) and two genes of unknown function (58;103). A combination of idd1 and idd3 introgressed onto the C57BL/6 strain is not sufficient for the induction of diabetes with C57 background genes (104).  Engineered haploinsufficiency of IL2 similar to low  expression of IL2 of the NOD mouse associated with idd3 (105) locus results in reduced CD4+CD25+ regulatory T cells and enhanced diabetes (68).
Idd4 maps to chromosome 11 and may influence the frequency and severity of insulitis and progression to diabetes (49;54). In particular, Idd4 homozygosity determines the age of onset of diabetes and Idd2, Idd3, and Idd4 together may accelerate progression to overt diabetes (106). A recent report indicates that NOD mice with deleted 12/15 lipoxygenase involved in the production of proinflammatory fatty acids increase the development of diabetes (107). In addition constituitive phosphorylation of Stat5 of NOD mice is associated with idd4 (108).
The Idd5 locus is located on chromosome 1 and alone confers 50% protection from diabetes and in combination with Idd3 provides nearly complete protection from infiltration of the pancreas, thyroid, and salivary glands (109). Therefore, these genes are likely to influence tolerance processes in the animal (110). The synergistic effect follows a model of additivity rather than multiple epistasis (111). Two loci have been identified within the Idd5 region which results in delayed onset of disease. Idd5.1 overlaps with the IDDM12 locus in humans (candidate gene is CD152 or CTLA-4) (109) and which has been narrowed to 2.1-Mb containing CTLA-4, ICOS, Als2cr19, and Nrp2, with CTLA-4 being the primary candidate (67;112) but higher levels on T cells of ICOS on NOD T cells (113). Idd5 F2 mice show a resistance to gamma-induced apoptosis in the NOD and NOR strains while T cells from C57BL/6 and DBA/2 strains show a “high-apoptosis” phenotype (114). The CTLA-4 candidate gene is intriguing because CTLA-4-/- mice are also resistant to apoptosis (115). The NOD idd5 locus mediates a bone marrow cell derived defect in negative selection of t cells (116).  Microarray data have implicated CD55 (decay accelerating factor) and acyl coenzyme A dehydrogenase expression (117) associated with idd5.
The Idd6 locus contains the NK cell cluster; NOD.Idd6 (containing NK1.1) congenic mice have been shown to have reduced disease incidence (118). The importance of NK cell function in reducing disease in NOD mice has been a story of great interest (118-120), although the complete localization to Idd6 is still unknown. F2 intercrosses between NOD and C57BL/6 mice showed an association between the Idd6 locus on chromosome 6 and decreased proliferation of immature thymocytes in NOD mouse (121), although the contribution of this phenotype to disease is still unknown. Additional phenotypes associated with idd6 loci include downregulation of expression of Toll like receptor 1 and decreased expression of the Lmp gene (91;122).
Little information is available on Idd7 and Idd8 except that these loci appear to be protective in the homozygous state (106). A recent study indicates that the IDD7 locus influences thymic deletion of specific CD8 autoreactive T cells such as AI4 (123).
The Idd9 locus on chromosome 4 is now known to contain 3 distinct loci. NOD.B10 Idd9.1/9.2/9.3 triple congenic mice are almost completely protected from diabetes yet still show insulitis, suggesting these loci are involved in regulating autoreactive T cells (101). A change in cytokine production from IFNg to IL4 by infiltrating cells in triple congenic mice compared to wild-type NOD supports this interpretation. The presence of insulitis and salivary gland infiltrates in the Idd9 congenic mice suggests that tolerance defects still exist. Genes of the TNFR superfamily with polymorphisms between NOD and B10 mice include candidate genes CD30 and TNFR2 for Idd9.2 (124) and CD137 for Idd9.3 (101). Idd11 has been localized to one of the three Idd9 loci.  It is reported that in Idd9 mice autoreactive T cell accumulate in the pancreatic lymph node (123).
Idd10 and Idd18 are closely linked on mouse chromosome 3. NOD.B610/18 congenic mice show reduced incidence of diabetes (50% vs. 80%) and include candidate genes Csfm, CD53, KCNA3, Nras, and Rap1a (125). Recent studies evaluating the IIS mouse (126) and with congenic mapping indicates that Idd10 is not Fcgr1 (77) and CD101 that differs in sequence from NOD for IIS and B6 is a candidate. Adding another locus from chromosome 3 to create NOD.B6 3/10/18 congenic mice results in almost complete loss of diabetes as well as insulitis (127). The few animals that do develop diabetes have a delayed-onset.
Idd13, like idd5, is involved in the regulation of T cell progression from benign to destructive insulitis. It contains at least two identified loci: idd13a and idd13b, which contains the candidate gene b2-microglobulin (73). The studies of Serreze and coworkers have firmly established a polymorphism of b2-microglobulin as influencing development of NOD diabetes, and it is one of the few established “genes” outside of the MHC (69). The NOD allele is a standard a isoform, and the b isoform differs from the a by one amino acid (alanine instead of aspartic acid at amino acid 85. The b allele does not suppress diabetes in the presence of the a allele, but the b allele cannot restore diabetes development by transgenesis in mice lacking beta-2 microglobulin but the a isoform does (69).  NOR derived idd13 locus increases invariant NKT cells (128).
"Monogenic Autoimmune Mutations"
In 1926, Schmidt described a patient with Addison’s disease and thyroiditis (129), and eventually multiple clinical syndromes consisting of multiple autoimmune disorders were recognized (11). A subset of these syndromes develops from single gene mutations; many such single gene mutations now have animal models. In particular, the Autoimmune Polyendocrine Syndrome Type 1 (APS-I) results from a mutation of the AIRE gene and approximately 18% of patients with this syndrome eventually develop type 1 diabetes (usually in addition to Addison’s disease, mucocutaneous candidiasis, and hypoparathyroidism) (11). When this mutation (deletion) is bred onto mouse strains, lymphocytic infiltrates occur, but no diabetes (23;130;131). Of note, however, the AIRE mutation appears to have a major role in expression of “peripheral antigens” within the thymus (22). Hanahan coined the term “peripheral” antigens (132-134), as molecules expressed at low levels within the thymus (e.g., rat-insulin-promoter-driven expression in his studies) and it is clear that such expression has a major influence on autoimmunity to the respective molecules. There is one Italian family reported with an autosomal dominant form of APS-I (135) and Anderson and coworkers have introduced this mutation into NOD mice (4).  An autosomal dominant autoimmune syndrome develops that differs from the phenotype of NOD mice with both AIRE genes knocked out.  In particular the autosomal dominant disease does not result in pancreatitits (4).  The dominant negative mutation appears to act by recruiting wild type AIRE away from active sites of transcription and does down regulate expression of peripheral antigens within the thymus.  The AIRE knockout on NOD mice did not accelerate the development of diabetes and combined myd88 knockouts (eliminating major pathways of toll like receptor signaling) or raising mice in a germ free environment did not influence disease (136).
The IPEX (Immune dysregulation, Polyendocrinopathy, Enteropathy, X-linked) syndrome is another particularly informative autoimmune syndrome, with mutations of the  Foxp3 gene, and the homologous gene scurfin in mouse. This mutation results in loss of a major subset of regulatory T lymphocytes (CD4+CD25+), and overwhelming autoimmunity in man, such that children often die as neonates, and immune-mediated diabetes can occur in the first days of life (21;137). In the scurfy mouse, the disease can be cured with partial T cell chimerism, and the same appears to be true of man with normal T lymphocytes able to regulate the abnormal immune system in a dominant fashion (138).  Mice with the foxP3 (scurfy) mutation die from overwhelming autoimmunity, but T cell receptor transgenics combined with this mutation, allow studies of specific autoimmunity, and accelerated development of diabetes (139).  Depending upon the specific T cell receptor diabetes can be accelerated even in  rag-/- mice expressing a single anti-insulin B:9-23 T cell receptor (Jasinski et al unpublished).
The more common autoimmune disorders, similar to the NOD mouse, are polygenic in origin. Human and rat studies of type 1 diabetes, multiple sclerosis, collagen-induced arthritis and systemic lupus erythematosus (SLE) have been mapped to the same chromosomal sites, suggesting a common genetic basis (72;140;141). However, given the number of  loci, chance overlap is to be predicted. The recently discovered autoimmune polymorphism of the PTPN22 gene (LYP gene of man) does contribute to multiple autoimmune diseases (type 1 diabetes, rheumatoid arthritis, Graves’ disease, lupus erythematosus (142-144)). The NOD mouse itself develops multiple autoimmune manifestations in addition to type 1 diabetes including thyroiditis, lymphocyte infiltration of the salivary and lachrymal glands (61), and can develop polyneuropathy (145) if B7.2 deficient.  NOD mice have been reported to target “Schwann” like cells surrounding islets (146), and has been shown to be susceptible to both experimental autoimmune encephalitis (EAE - an animal model for multiple sclerosis) (147) and experimental autoimmune prostatitis (148). Therefore certain Idd genes or combinations thereof may alter the immune system such that tolerance processes are defective. The form of autoimmunity may then depend on other genetic (e.g., MHC and antigens) or environmental factors. One recent study shows that autoimmunity in the NOD switches from pancreas-specific to destruction of the peripheral nervous system by altering the costimulatory molecules present on dendritic cells (145) and NOD.c3c4 mice with multiple B6 and B10 protective alleles from chromosome 3 and 4 bred onto NOD do not develop diabetes, but develop autoimmune biliary disease (149). Of note, breeding only one of the two chromosomal regions onto NOD produces mice with biliary autoantibodies without liver infiltration (149). In a similar manner, NOD congenics can express anti-insulin autoantibodies with rare progression to diabetes (71). Thus disease targeting is influenced both by MHC alleles and by other polymorphisms in an autoimmune “background”.
Immunology of the NOD mouse
Beta cell destruction in the NOD requires CD4 T cells, CD8 T cells (150-152) and B cells for its spontaneous occurrence (153), and maternal transfer of autoantibodies has also been found to have a role (154). The NOD mouse allows studies of the progression of the disease including the timing of insulitis and the nature of insulitis, i.e., cytokine production during the course of the disease. Prior to developing sufficient beta cell damage to produce hyperglycemia, a period of insulitis is observed followed by eventual development of hyperglycemia. There remains a debate among investigators whether the beta-cell loss is a gradual or acute process, occurring suddenly once regulation of autoreactive T cells wanes, but we suspect the term “benign” insulitis is a misnomer. More than 90% of untreated female NOD mice lose beta cell mass with time, such that by 52 weeks beta cell mass is 1/5th to 1/10th of normal even for the great majority of female mice that have not progressed to diabetes ( (155) and Gianani et al ADA abstract 2008). There is considerable endocrine reserve and mice even with 1/5th of normal beta cell mass can remain non-diabetic. As illustrated in Figure 4, even 8-9 week old NOD mice have lost beta cell mass. Of note, the histology of the NOD pancreas suggests asynchronous destruction of islet beta cells. Within the same mouse, one can find normal islets with beta cells, islets with insulitis and beta cells, and islets where all the beta cells (insulin producing cells) have been destroyed, and only non-beta cells (e.g., alpha cells producing glucagon) remain. Additionally during the phase of active destruction there is evidence of beta cell replication (155;156) that eventually cannot keep pace with beta cell destruction.

Figure 4. Quantitation of beta cell mass, comparing NOD-SCID mice that lack T cells, to loss of beta cell mass in NOD with age.

There is now clear evidence that mice can regenerate beta cell mass following acute beta cell destruction but it is not clear if humans have such an ability (36;157).  Anecdotal reports of beta cell proliferation (158) followed by analysis of multiple individuals (159) suggest that replication of human beta cells is infrequent.  A subset of patients with type 1 diabetes retain C-peptide secretion long-term, but the great majority have very low c-peptide levels (160).  The mechanism related to retention of some beta cell mass in man and in the NOD mouse model is currently unknown.  Herold and coworkers analyzing NOD mice prior to the development of diabetes and then during therapy with both anti-CD3 and regulatory T lymphocytes have concluded that inflammation increases beta cell replication, that most of the recovery of beta cells following therapy is a result of regranulation of degranulated beta cells and following therapy beta cell replication is reduced (161) with lower percentage of Ki67+ beta cells post recovery (figure below).

It is likely that in man, similar to the NOD mouse, that a significant and important beta cell mass is present at onset of hyperglycemia.
In NOD mice, expression of insulin autoantibodies (IAA) is usually noted prior to onset of hyperglycemia (162). In addition, the NOR mice, a strain closely related to the NOD with the IAg7 MHC haplotype as well as other NOD congenic strains (29;71), develop IAA but with limited or no progression to type 1 diabetes (71). The presence of insulin autoantibodies correlates primarily with insulitis and not simply progression to hyperglycemia. Genetic studies using congenic NOD strains show the expression of IAA antibodies is controlled by multiple loci. (71) In addition, the amount of IAA was found to associate with degrees of insulitis more than disease incidence (71).  There is evidence of additional autoantibodies reacting with islets cell antigens of NOD mice, but in workshops it has been difficult to confirm the presence of such autoantibodies with highly specific fluid phase radioassays (163;164).  It is likely that B-lymphocytes (though insufficient by themselves (165) can not only contribute to diabetes through the production of autoantibodies (e.g. evidence that “transplacental” passage of autoantibodies of NOD mice important for disease (166)) but have additional roles, particularly as antigen presenting cells and for the maturation of CD8 T cells (154;167).  Treatment of with a “humanized” CD20 NOD mice by an anti-CD20 monoclonal in clinical use prevents the bulk of diabetes (168).
It is likely that antigen presentation for the activation of pathogenic T lymphocytes occurs in the pancreatic lymph node with evidence provided by the BDC 2.5 transgenic T cells, where labeling of cells with the dye CFSE indicates initial proliferation in such lymph nodes (169;170), the finding by Fathman and coworkers that pathogenic cells in the pancreatic lymph node are CD4high (171), and removal of pancreatic lymph nodes at 3 weeks but not 10 weeks prevents diabetes (172). The role of various cell populations in the beta cell damage have been addressed using knockout animals, T cell clones, and adoptive transfer models of the NOD mouse and are discussed in the following chapter. To summarize numerous findings, spontaneous disease requires CD4 and CD8 T cells as well as B cells, which are thought to be involved in antigen presentation. Mathis and colleagues have implicated NK specific transcripts and proportion of NK cells in development of destructive islet autoimmunity (173).
The confirmation that lymphocytes are required for the beta cell destruction in type 1 diabetes led to numerous studies on the immune system in NOD mice compared to nonautoimmune-prone strains. A deficiency in the ability of NOD APCs to mount equivalent T cell responses has been reported (174;175). These NOD APCs have been found to have low CD86 expression (176), and recent studies have reported phenotypic and functional defects in bone-marrow-derived dendritic cells (DCs) in response to GM-CSF (177). Wong et al. have found a deficiency in priming of NOD T cell responses to both endogenous and exogenous antigens (178). Similarly, defects in TCR-mediated signaling resulting in inferior NOD T cell responses have been linked to alterations in p21ras signaling (179). Neonatal CD28 costimulation has been recently found to restore this signaling as well as protect NOD mice from diabetes, suggesting a relationship between this T cell hyporesponsiveness and disease onset (180). These abnormalities in T cell function may relate “lymphopenia” of the NOD mouse with the hypothesis that homeostatic expansion may contribute to autoimmunity (181).
Bellgrau and coworkers found a similar T cell hyporesponsiveness among numerous strains of autoimmune-prone mice susceptible to EAE, SLE, rheumatoid arthritis, and autoimmune hemolytic anemia (182). The prevention of diabetes resulting from administration of immune adjuvants (183;184) or DNA vaccinations to self-antigens (185) may be related to reversing this hyporesponsive immunity observed in the NOD. Further work is required to clarify the mechanism of protection and possible role of a hyporeactive immune response in autoimmunity.
One possible mechanism by which a hyporeactive T cell response may contribute to autoimmunity is through defective negative selection in the thymus. Evidence for poor central tolerance in the NOD includes direct demonstration of inferior thymic deletion in NOD compared to BALB/c mice upon administration of anti-TCR antibodies, increased frequency of autoreactive T cell responses to pancreatic antigens (90;186-189), and a recent demonstration that epitope spreading in NOD follows a hierarchy from highest affinity TCR:self-antigen interactions to lower affinities suggesting the spreading hierarchy is determined by the extent of negative selection (190). The concept that with disease progression lower affinity TCR reactions are favored is challenged by the studies of Santamaria and coworkers who observe affinity maturation of the TCR utilized by cells targeting the molecule IGRP (islet-specific glucose-6-phosphatase catalytic subunit-related protein), which is the native molecule targeted by NRP (NOD Related Peptide) CD8 lymphocytes which form the majority of T cells infiltrating islets (191-193). Other thymic defects in the NOD include abnormal corticomedullary environments (194), a deficiency in the number of TCR+ CD4-, CD8- populations in NOD mice, and low proliferation of immature thymocytes (121). Whether these phenotypes have relevance to autoimmunity remains to be determined, although the deficiency in TCR+ DN thymocytes does not appear to be related as it is seen in a number of congenic Idd mouse strains which are not diabetic (195).

Regulatory Cells

Table 3.2. Summary of Regulatory Cells Influencing Diabetes.

It is well established that autoreactive T cells can be controlled by regulatory cells (Table 3.2). There has been a recent tremendous expansion in knowledge concerning regulatory T lymphocytes (9;196), and in particular, their role in the NOD mouse and the ability to generate large numbers of CD4+CD25+ regulatory T cells (197;198) and prevent diabetes (196). Evidence for the role of regulatory T cells in delaying or preventing overt diabetes are numerous including prevention of adoptive transfer of diabetes by mixing splenocytes (199) from pre-diabetic NOD mice with the disease-inducing splenocytes from diabetic donors (200), induction of regulatory cells dependent upon TGFbeta with anti-CD3 therapy in recent onset NOD mice (201), and the large literature concerning mucosal tolerance with bystander suppression, including ability of oral insulin to delay diabetes of NOD mice (202). It is clear that these regulatory T cells are heterogeneous in nature, and distinct populations have been elucidated (43;197;203-206).
Related to the initial studies of thymectomy-induced autoimmunity (207;208), the remarkable human autoimmune phenotype of mutations of the Foxp3 gene (IPEX syndrome-see chapter 8) associated with neonatal immune mediated diabetes (137), the ability of Foxp3 expression to generate CD4+CD25+ regulatory T cells (209) and suppress NOD diabetes (210), and the elegant models inducing deficiencies of these regulatory T cells (207;211), the CD4+CD25+ T cells have become perhaps the best characterized regulatory T cell of the NOD mouse (9;196;212;213). NOD mice genetically deficient for CD28 expression succumb to diabetes with faster kinetics than wild type NOD mice (214). This accelerated disease incidence was found to be associated with a loss of CD4+CD25+ regulatory T cells  (100). Of note, as recently published by Bluestone and coworkers, 107 in vitro expanded FoxP3 expressing Tregs (using the BDC2.5 T cell receptor transgenic (recognizes islet membrane antigen) as proof of principal) reversed new-onset diabetes in 60% of NOD mice. BDC2.5 Treg cells have been demonstrated in the pancreatic islets and may act at this site to suppress disease given the ability of ICOS blockade to accelerate development of diabetes (215). Similar regulatory T cells have been induced by immunization with the B:9-23 insulin peptide as well as a B24-C36 proinsulin peptide (216;217).
A critical role of NK T cells in protection from diabetes has recently been demonstrated in NOD mice. A deficiency in number and function (IL4 secretion) of CD1-restricted NK T cells has been observed for many years (218;219), however, the contribution of this phenotype to diabetes development was unclear. Several recent studies demonstrate the protection offered by increased numbers or function of CD1-restricted NK T cells (118-120).  NK T cells show a restricted TCR repertoire with a predominance of Va14/Ja281 chains. These cells are specific for lipid antigens presented in the CD1 MHC molecule. Upon activation, which can be achieved with alpha-galactosylceramide (aGalCer), NK T cells produce large amounts of cytokines, most notably IL-4 and IL-10. The Idd6 locus contains the NKR-P1 gene cluster, which includes the NK1.1 gene. The NK1.1 gene is absent in NOD mice, making NK T cells more difficult to analyze. Carnaud and coworkers demonstrated a contribution of the Idd6 locus to disease incidence as NOD.NK1.1 congenic mice have reduced diabetes incidence which correlated with improved NK cell function (118). The contribution of NK T cell defects to diabetes incidence was also demonstrated in NOD CD1 KO mice which showed an exacerbated disease course   (50). In addition, increasing the numbers of NK T cells in the NOD by transgenic expression of the Va14Ja281 receptor resulted in reduced disease (119). The protection appears to be related to increased IL4 production observed in the pancreas as IL12 or anti-IL4 antibodies were shown to abolish this protection (119). It appears that increasing numbers of NK T cells is not necessary for protection as activation of the existing NK T cells with aGalCer alone protects diabetes in CD1-sufficient NOD   (50-52) (120) but not in CD1-deficient NOD mice   (52). This CD1-restricted NK T activating reagent also enhances the survival of transplanted islets (120).
NOD mice injected with complete Freund’s adjuvant are dramatically protected from the development of diabetes. Lee and coworkers have recently reported that this protection is abrogated when NK asialo-GM1 cells are depleted and restoring cells expressing CD3-DX5+ restored protection (220).
Harrison and coworkers have studied the administration of insulin by various routes as a means of protecting NOD mice. They find that administration of insulin by the naso-respiratory route protects from diabetes with the induction of CD8+ alpha (TCR gamma delta T regulatory cells that can prevent NOD diabetes). These cells cannot be generated in day 3 thymectomized mice, but are restored with administration of the above gamma delta T cells (217).
Homann and coworkers have utilized anti-CD40 ligand blockade and the RIP-LCMV diabetes model to generate cells with the unusual set of markers of both NK cells and dendritic cells, with the finding that these cells can block the development of diabetes (221).
Antigenic Targets
A fundamental question relative to the immunopathogenesis of type 1A diabetes (immune- mediated diabetes) is whether there are primary islet autoantigens. It is clear that multiple islet molecules are the target of autoimmunity in man and animal models (222;223). In particular, there are multiple T cell clones (224) whose target antigens are currently unknown (BDC2.5) and a list of well-characterized beta cell “specific” targets for T cell clones (e.g., insulin, IGRP, proinsulin) as well as target molecules such as heat shock proteins, GAD, and recently described dystrophia myotonica kinase (DMK) where the antigen is either minimally or not expressed in mouse islets (GAD) (225) or widely expressed in multiple tissues (226;227).
In terms of islet autoantibodies of NOD mice, two autoantibody workshops suggest that only insulin autoantibodies can be specifically detected using sensitive radioassays (164;228). Despite the large number of islet autoantigens, there is increasing evidence that insulin, and in particular, a specific epitope of insulin (the Wegmann insulin B chain amino acids 9-23) may be an essential target in the NOD mouse model (29;216;229-231). Both cloned CD4 and CD8 T cells reacting with an overlapping insulin B chain peptide can mediate adoptive transfer of type 1 diabetes (178;222). MacLaren and coworkers reported that subcutaneous administration of insulin and insulin B chain prevents the development of diabetes in NOD mice, and Weiner and coworkers reported that oral insulin decreased the development of diabetes (232-235). Wegmann and coworkers isolated T lymphocytes directly from islets of NOD mice. The T lymphocytes were stimulated with islets as “antigen” and following the development of a series of clones, the majority was found to react with insulin (222;229). Of the clones reacting with insulin, more than 90% reacted with insulin peptide B:9-23. These clones were notable for utilizing a conserved AV13S3 and AJ53 segment, with variation in the junctional region and no apparent conservation of the Vb chain (236;237). Despite utilization of this dominant a-chain motif, one of the clones studied recognized insulin peptide B:9-16 and another four B:13-23 (238). Administration of the B:9-23 peptide either intranasally without adjuvant or with a single injection in incomplete Freud’s adjuvant protects the majority of NOD mice from progression to diabetes (239). Wong and coworkers have isolated a NOD CD8 clone reacting with insulin peptide B:15-23 and reported a very high frequency of B:15-23 tetramer-positive cells within islets of NOD mice, though the exact percentage remains controversial (178;240). Follow-up studies suggest that though present early in lesions the percentage of CD8 anti-B:15-23 T cell clones is more limited with a major population of IGRP (islet-specific glucose-6-phosphatase catalytic subunit-related protein)-reactive CD8 T cells (223).
Autoantigen Gene Knockouts
A significant number of genes for potential autoantigens of the NOD mouse have been “knocked out.” To date, only knockouts of insulin genes have influenced development of diabetes (Table 3.3). Mice have two insulin genes, and since both insulins are present in islets and both insulins are metabolically active, it is possible to knock out either gene and not develop metabolic forms of diabetes. The insulin 2 gene is the proinsulin gene that is expressed within the thymus. The proinsulin 1 gene is a retroposon, with almost no expression within the thymus. Single insulin 2 gene knockouts (produced by J. Jami) were bred onto NOD mice by our group (29) and Boitard and coworkers (98;241). The insulin 2 gene knockout greatly accelerates the development of diabetes and increases the levels of insulin autoantibodies. The insulin 1 gene knockout prevents approximately 90% of the development of diabetes of female NOD mice but does not alter the expression of insulin autoantibodies and the majority of mice with the insulin 2 knockout have insulitis, consistent with the subset progressing to overt diabetes. Jaeckel and coworkers using a technology to induce tolerance similar to that which failed to alter progression to diabetes with GAD65 transgene (225), have recently reported that with an insulin 2 preproinsulin transgene with invariant chain promoter, progression to diabetes is markedly decreased. They concluded that insulin is a “key” autoantigen of the NOD mouse but not “essential” (230). Of note, they only utilized an insulin 2 transgene. Insulin 2 differs from insulin 1 for two amino acids (as well as multiple additional polymorphisms in the leader and connecting peptide sequence).

Figure 5. Life table analysis of progression to diabetes of NOD mice with insulin 1 or insulin 2 gene knocked out, for male and female mice. Homozygous knockout mice are represented by the triangles. Heterozygous knockout by the squares, and wild type at the insulin gene by circles. Knocking out insulin 2 accelerates diabetes whereas insulin 1 knockout prevents the majority of progression to diabetes.

One of the amino acids that differ between insulin 1 and insulin 2 is position 9 of the B:9-23 peptide (serine for insulin 2, proline for insulin 1), and we have in vivo evidence that the immune response to the two peptides can be dramatically different (242). A double insulin gene knockout NOD mouse (insulin 1 and insulin 2) with a mutant preproinsulin transgene to prevent metabolic diabetes has neither insulitis nor anti-insulin autoantibodies (Nakayama et al., In review). The dramatic effect of insulin gene knockouts is consistent with multiple studies indicating induction of widespread insulin gene expression prevents diabetes, including expression in bone marrow-derived cells (243;244).
Insulin as the Primary Autoantigen of the NOD Mouse
Multiple islet molecules are the target of autoantibodies or T cells in man and animal models (37;224;245), with the islet Zinc transporter (ZnNt8) a recent addition (246).  There is not universal agreement if any of the target molecules are essential for immune mediated beta cell destruction, though studies from the laboratories of Kay (3;247), Boitard (98;241), Jaeckel (230) our group and others, strongly implicate immune responses to insulin as a central component of type 1 diabetes of the NOD mouse. 
Our studies both in man and mouse have concentrated on insulin (29).  An insulin 2 gene knockout greatly accelerates the development of diabetes and increases the levels of insulin autoantibodies.  The insulin 1 gene knockout prevents approximately 90% of the development of diabetes of female NOD mice but does not alter the expression of insulin autoantibodies and the majority of mice with the insulin 1 knockout alone have insulitis, consistent with a subset progressing to overt diabetes. Kay and coworkers utilizing a transgene driving proinsulin expression with an I-E promoter completely prevent diabetes and of interest abrogate development of IGRP CD8 reactive T cells, and even prevent diabetes in a TCR transgenic targeting IGRP (3;247).  Of note a similar I-E promoter transgene but driving IGRP expression has no effect on progression to diabetes (247), and thus they concluded that immune responses to proinsulin are “upstream” of IGRP T cell targeting and crucial for diabetes, and lack of immune response to insulin even prevented diabetes in mice with an anti-IGRP T cell receptor transgene (3).

We have analyzed NOD mice lacking both insulin 1 and 2 genes, with multiple transgenic founders with either a native insulin sequence or a sequence with a single amino acid altered (B16:A replacing B16:Y).  These mice fail to develop diabetes and development of insulin autoantibodies and insulitis are markedly decreased (2).  Restoring the native B:9-23 sequence with an islet transplant (but not bone marrow transplant) or peptide immunization, or a native proinsulin transgene, restores anti-insulin autoimmunity and generates CD4 T cells able to cause diabetes (97).  Unanue and coworkers have defined two different registers of binding of the B:9-23 peptide to I-Ag7 and multiple investigators have utilized the B:9-23 peptide for prevention of diabetes (248-250).
Insulin is a major antigen recognized by islet infiltrating T cells of NOD mice, (222) a proinsulin transgene prevents disease in NOD mice; (243) insulin is within the islets β-cell specific (e.g. not expressed in islet α-cells); a polymorphism of the insulin gene which influences thymic insulin messenger RNA is associated with diabetes protection. (251;252) Both cloned CD4 and CD8 T cells reacting with an overlapping insulin B chain peptide can mediate adoptive transfer of type 1 diabetes. (178;222).
Cogent arguments can be and have been made for the importance of other target autoantigens including GAD, IA-2, IA-2beta, heat shock proteins, targets of fascinating cloned T cells such as BDC2.5 and the IGRP reactive clones described by Santamaria and Serreze (223), and the recently described ZnT8 autoantigen (246).  Baekkeskov has produced a GAD65 knockout and breeding this gene onto NOD mice does not effect diabetes development. (253) Similar studies of knocking out potential islet target molecules such as IA-2 and IA-2 beta similarly did not influence progression to diabetes (254;255).  Lack of immune response to GAD65 did not influence progression to diabetes (225;256) but there is interesting evidence that anti-GAD responses can be protective (256).  Vignali and coworkers have produced a series of retrogenics with T cell receptors targeting islet molecules, and though anti-GAD T Cell Receptors did not induce disease, anti-insulin (BDC12-4.1) induced delayed diabetes (257) (Note 12-4.4 sequence studied in that manuscript was different from our diabetogenic BDC12-4.4 retrogenics), and there are a series of additional important islet antigens awaiting discovery including those such as the 10.1 T cell receptor that markedly induced diabetes (257).

Table 3.3. Autoantigen Knockouts.

This suggests that presence of GAD65 is irrelevant to development of diabetes of NOD mice but does not address a potential role of GAD67 and the ability of GAD peptide immunization to prevent diabetes. The lack of effect of the GAD65 knockout is concordant with studies inducing tolerance to GAD and finding no influence on progression to diabetes (225). Nevertheless studies where anti-GAD T cell receptor transgenic mice inhibit rather than accelerate development of diabetes (256). For the NOD mouse there is a consensus that immune recognition of GAD is more related to protection than beta cell destruction (258).
Studies knocking out potential islet target molecules such as IA-2 and IA-2 beta similarly did not influence progression to diabetes (254;255) and Kay and coworkers have demonstrated that the immune response to IGRP is “downstream” of the immune response to insulin (247). Yoon and coworkers produced five anti-sense GAD transgenic lines (259) with the transgene bred onto the NOD background. Follow-up of these lines indicates that of the two lines with any diabetes suppression, one develops diabetes spontaneously and the other after cyclophosphamide induction (oral communication). Further study of these transgenic lines will be of particular interest relative to the mechanism of disease alteration. Cohen and coworkers have studied in detail the ubiquitous HSP60 molecule and peptide p277-responsive T lymphocytes of NOD mice (260-263). The administration of the p277 peptide has been reported to prevent diabetes, but to have no effect in a study by Atkinson and coworkers (264). The evidence that HSP60/p277 is an islet “autoantigen” is relatively weak, with no demonstration in workshops of specific reactivity and with an alternative hypothesis for its effects in NOD mice related to activation of the innate immune system (227).
BioBreeding Diabetes-Prone (BB-DP) Rat
The first inbred animal model for spontaneous, type 1 diabetes evolved from the discovery of diabetic animals in an outbred colony of Wistar rats maintained at the Bio Breeding Research Laboratories in Ottawa, Canada. Selective breeding led to the first paper on the Bio Breeding (BB) rat in 1975 wherein the clinical description of insulin-dependent diabetes was documented (265). Prevention of diabetes with antiserum to rat lymphocytes suggested an autoimmune pathogenesis for disease (266). Further support for this hypothesis was provided by studies describing the linkage of the disease with the MHC (267). The disease is now known to have a polygenic basis for susceptibility (268), and abnormalities in T cell function have been described (269-272). In comparison to the NOD animal model for type 1 diabetes, the BB model has certain advantages including higher and more uniform disease penetrance and no gender bias as diabetes is observed in >90% of both male and female BB-DP rats when maintained in pathogen-free conditions (273;274). The time between insulitis and overt diabetes in the BB-DP rat is much shorter than in the NOD mouse, as diabetes develops between 2-4 weeks after insulitis in the rat model. However, the BB-DP model requires a unique lymphopenia (lyp) locus for spontaneous disease which is not present in the NOD mouse or the vast majority of humans with type 1 diabetes (269). This lyp locus is required in the homozygous state for disease and results in a profound loss of peripheral T cells including a >85% reduction in CD8+ T cells and a 50-80% loss of CD4+ T cells. In this section, we will focus on the genetics and immune status of the BB rat, with a special emphasis on the relationship between T cell lymphopenia, T cell function, and disease susceptibility.
Derivation of BB Rats
All BB rat colonies worldwide were derived from a progenitor stock in Ottawa. As there was considerable genetic heterogeneity in the original Ottawa stock, inbreeding at other locations has predictably led to genetically distinct but related, inbred BB rat lines. In a study of 11 different BB rat colonies describing over 20 different BB sub-lines, inbred lines could be classified as belonging to one of four groups distinguished by eight protein markers from at least three linkage groups (275). Selected for sensitivity or resistance to diabetes, BB rats can be further subdivided into diabetes-prone (BB-DP) or diabetes-resistant (BB-DR) (276). As described later, the BB-DR sub-line progeny are designated by their ability to develop diabetes following experimental manipulations that do not induce diabetes in other animals (277).
Genetics of BB Rats
It has long been known that diabetes of the BB rat depends upon polymorphisms of the RT1U major histocompatibility complex as well as a “peculiar” severe T cell lymphopenia inherited in an autosomal recessive manner (lyp [lymphopenia] gene in BB rats that is unrelated to the Lymphoid Tyrosine Phosphatase gene (LYP) associated with human type 1 diabetes (142)) (278;279). Lymphopenia results from mutation of one of the Immune Associated Nucleotide Related genes (IAN4).  The frameshift mutation of IAN4 is associated with increased apoptosis, mitochondrial dysfunction (280), and potentially heightened T cell receptor signaling (281).  In the mouse IAN family genes are predominantly expressed in lymphocytes and expression is upregulated with thymocyte differentiation (282).

Figure 6. The lymphopenia gene (IDDM1 of rat) of BB-DP rats (Ian gene).

Essentially all rat models of type 1 diabetes have the permissive RT1u MHC haplotype (283). Linkage to the MHC for BB rats was first defined by Colle and coworkers in an analysis of diabetes incidence among F2 animals from a cross of BB-DP with Lewis strain rats (267). Lewis rats are not prone to diabetes and are incompatible with the BB-DP at the MHC. All diabetic F2 animals were shown to be homozygous for the BB-DP MHC which is the susceptible RT1u haplotype.
Multiple BB rat disease-associated loci (iddmn) have been reported (284-286). The lyp locus was mapped to a syntenic region of rat chromosome 4, is designated iddm1, is responsible for the severe T cell loss when present in the homozygous state (269), and has been identified as a frameshift mutation of the IAN4 gene (287;288).
Although the contributions of the iddm1 (lyp) and idddm2 (MHC) loci to disease are quantitatively large, they are not, in and of themselves, sufficient for disease. Fixing the RT1u and lyp alleles when breeding the BB-DP rat to diabetes-resistant ACI (278) and PVG (289) strains did not result in diabetes. Therefore, other loci are clearly involved in rendering the animal susceptible to autoimmunity. Reports link iddm4 with diabetes susceptibility in crosses between disease-resistant Wistar-Furth (WF) and BB-DR animals which do not normally develop spontaneous disease (284). The BB-DR animals are genetically similar to the BB-DP rat yet are resistant to spontaneous type 1 diabetes due to the lack of the iddm1 locus (lymphopenia gene). Thus the BB-DR animals have normal peripheral T cell numbers. However, autoimmune-susceptibility genes are clearly present in the BB-DR as type 1 diabetes can be induced in BB-DR animals with immune manipulations if administered in a short time window of 25-35 days of age. These manipulations include viral infections (290), poly IC injections (291) and depletion of RT6+ T cells along with an environmental trigger (292) and thymectomy (293).  One of the most fascinating BB-DR models is the induction of diabetes with the Kilham rat virus (277;290;294-296).  The ability of this virus to induce diabetes in BB-DR rats was discovered following the spontaneous infection of BB-DR colony with the virus.  The virus acts without infection of islet beta cells, with evidence that induction of disease follows TLR9 mediated induction of a series of cytokines and in particular IL12p40.  Chloroquine therapy decreases the development of diabetes (277). Genetic loci influencing induced diabetes of the BB-DR and related strains have been defined (286;297) with TCR beta (stop codon in WF rats) as one candidate antigen for IDDM4 (298).
Backcrosses of (WFxDR-BB)F1 to WF rats resulted in approximately one-half of the pups susceptible to type 1 diabetes induction. Furthermore, this susceptibility mapped strongly to the iddm4 locus on chromosome 4 (284). This 2.8-cM region on rat Chromosome (Chr) 4 locus contains several major autoimmunity loci including aia2, aia3, and cia3, and it has been assigned to a 2.8cM region proximal to Lyp/Ian4l1 (299). These data, as well as studies in the BB/OK strain (285;300), crosses between BB-DP and non-autoimmune-prone rats (301), and RT1u congenic strains (283) support the contribution of non-MHC loci in general susceptibility to autoimmunity. Two other loci found to be associated with type 1 diabetes in the BB-DP rat, including the diabetes-susceptible iddm3 and diabetes-resistant iddm5 loci (285;300).
Immune Dysfunction in the BB-DP Rat and Its Relationship to Lymphopenia
Lymphopenia was first defined in 1981 by Jackson and colleagues (302). An early important characteristic distinguishing lymphopenic from nonlymphopenic animals was the absence of the RT6+ subset of peripheral T cells (303;304). The under-representation of the CD45R+ T cell subset isoform was also documented (305). The absence of RT6+CD45R+ T cells correlated well with the overall reduction in the numbers of peripheral T cells found in the BB-DP. Since thymocytes are both RT6- (303) and 98% CD45R- (306), a plausible role of the lymphopenia gene is to retard T cell maturation at a stage prior to the expression of these peripheral T cell antigens. The RT6+ population is known to contain regulatory function since 1) depletion of the RT6+ population along with poly I:C injection induces type 1 diabetes in BB-DR rats (307), 2) diabetes can be induced by adoptive transfer of BB-DR lymphocytes into athymic nude WAG rats only if the BB-DR lymphocytes are pretreated with anti-RT6 antibodies while cotransfer of RT6+ T cells prevents diabetes (308) and 3) injection of RT6+ cells from BB-DR rats into BB-DP lymphopenic rats prevents their spontaneous development of type 1 diabetes (309). Both CD4+CD25+ and CD4 T cells that express neither CD25 nor Foxp3 have regulatory function (310;311).
Studies into the immune function of lyp T cells from BB-DP rats report contradictory results of poor in vitro responses and hyper-activated phenotypes. Early studies report weak proliferative (312) and cytotoxic responses (271) to alloantigen as well as altered CD4-mediated signaling (313). Given the reduced numbers of T cells in the lyp animals, especially in the CD8 compartment, the differences in responses in these bulk read-out in vitro assays may be due to differences in functional T cells present. It is now known that many of the peripheral T cells in the BB-DP are undergoing apoptosis and therefore may not be functional, although present, in these assays. More recent studies suggest that lyp T cells show increased signs of activation ex vivo, including increased expression of CD25 (314) and OX40 activation markers (315), increased INFg production (316), and increased mRNA of T cell signaling adapter protein vav (317). The lyp cells also show increased signs of spontaneous proliferation ex vivo with 1) 2X number of cycling cells as determined by increased DNA content (318), 2) a 90% reduction in levels of cell-cycle inhibitor p27kip, increased levels of PCNA and phosphorylated Rb (319), and 3) incorporation of BrdU into >90% of lyp T cells compared to ~30% of normal T cells in a 13-day period (320). Collectively these data suggest that lyp T cells go through DNA synthesis prior to their imminent demise although no studies have investigated progression through mitosis or cell division. The resultant loss of T cells, as opposed to an accumulation, suggests this proliferative state is non-productive.
In addition to T cell abnormalities associated with the lyp gene, studies of non-T cell subsets of the BB-DP immune system suggest thymic and APC defects similar to the NOD mouse model. A lower number of splenic DCs with decreased expression of MHC class II and costimulatory ligand CD80 (321) as well as  decreased clustering which improves stimulating capacity of DCs is reported (322). In addition, defects in the NK cell population (323) as well as thymic B cells (324) have been observed. The data support an overwhelming defect in peripheral T cell regulation as a mechanism for disease and reports of enhanced proliferative responses to superantigens following in vivo administration (272) support the conclusion of defects in peripheral tolerance in lyp rats. Few studies on thymic tolerance have been reported in the BB-DP model although alterations in thymic medullary (269) and cortical (325) architecture have been associated with thymic tolerance defects were noted.
A remarkable phenotype afforded by the lyp mutation is the severely shortened lifespan of the lyp T cells. Although thymic development appears normal in lyp animals with an almost normal distribution of double-positive and CD4+ and CD8+ single-positive subsets which vary depending on background strain (289), peripheral T cell numbers are drastically reduced. The CD8+ T cell subset is nearly absent in the spleen and lymph nodes of lyp animals with a few cells expressing lower levels of CD8 (326), suggesting a faster death following thymic selection than CD4+ T cells (327;328). Tracking the export of T cells from the thymus with fluorescent-labeling, Zadeh and coworkers measured the lifespan of CD4+ lyp T cells to be less than one week (318). In addition, only few labeled thymocytes expressed RT6, which is normally upregulated 1-2 weeks following thymic export. The majority of the labeled thymocytes expressed the Thy1 Antigen which is characteristic of recent thymic emigrants (RTEs) in the rat (318). The thymus of the lyp rat also showed reduced thymic export of RTEs, data which is consistent with increased death of thymocytes in adult thymic organ cultures (ATOCs) from lyp rats (329). Further data supporting an overwhelmingly shortened lifespan of lyp T cells is the rapid loss of T cells from lymph nodes (LNs) and spleen following thymectomy in lyp rats compared to a fairly stable T cell pool in thymectomized normal rats (330;331).
The efficient clearance of apoptotic cells by phagocytosis has made detection of apoptosis difficult in vivo. Death by apoptosis in lyp T cells is rapid in vitro (289;320) with a majority (~80%) of the cells dying in overnight cultures compared to negligible death (~20%) observed in cultured T cells from nonlymphopenic rats. Advances in detection of apoptotic cells allowed identification of these dying cells in vivo through TUNEL and FITC-Annexin V staining assays (332). Further characterization identified the clearance of the apoptotic T cells in lyp rats in the liver (323). In mice, dying apoptotic CD8+ T cells are cleared in the liver whereas CD4+ T cells remain in the LN even when apoptotic (333). The dying cells were difficult to define as T cells due to their detectable but low expression of TCR, CD8, CD4, B220 and HSA proteins on their surface (323).
The vast reduction in RT6+ cells along with a preponderance of Thy1+ cells suggests lyp impacts T cell development at the RTE (Recent Thymic Emigrant) stage. The effects of the lyp mutation appear to be cell autonomous and restricted to bone marrow cells as demonstrated through lyp à normal bone marrow chimeras (334). In addition, transplants of normal thymus into lyp recipients do not reverse the lymphopenia, pointing to the developing T cells themselves harboring the genetic disturbance (335). It is noteworthy that the T cell loss, activation, and death phenotype observed in the BB-DP rat is also observed in lyp congenic rats from diabetes-resistant strains (Fisher-lyp and PVG-lyp) which show T cell lymphopenia but not diabetes (289;319).
The reduced number and early death of the T cells and their necessary role in type 1 diabetes induction present a paradox in the BB-DP model. Although a lack of regulation in the BB-DP rat is well-documented, less is known about the properties of the effector T cells responsible for the beta cell destruction. There is disagreement whether the short-lived RTEs are responsible for disease, as suggested by disease prevention with adult thymectomy (331). However, the thymectomy is required during a distinct time period for protection – when performed at 8 weeks of age animals still develop diabetes even though the vast majority of RTEs) still undergo rapid death. Combined with newer studies showing antigen activation can rescue lyp T cells from death both in vivo (320) and in vitro (289), another model suggests that higher affinity autoreactive T cells are activated by autoantigen, survive, and serve the effector functions necessary for immune clearance. Isolation and characterization of the effector T cells responsible for beta cell destruction are still necessary to differentiate among these models.
Disease Prevention in the BB-DP Rat
As in the NOD mouse model, numerous treatments have been shown to prevent diabetes in the BB-DP rat model. Early immunosuppression to prevent effector T cell function, including T cell depletion with antibodies (336-338), thymectomy (330;331;339) or FK506 (340;341) prevents diabetes in the BB-DP rat. Similar to the effects in the NOD model, complete Freund’s adjuvant (342) and viral infections (343;344) prevent disease in the BB-DP model as well. Reversing the defect in the regulatory population also prevents diabetes. This reversal has been shown to be successful in disease prevention with the addition of RT6+ cells (345-348) or with human TNFalpha, a cytokine which influences regulatory properties of T cells (349) . Although alterations in peripheral tolerance are well-documented in the BB-DP, little is known about alterations in thymic tolerance in this model. However a report that in thymic organ culture of BB-DP thymocytes, coculture with islets, but not thyrocytes, prevents diabetes and reduces insulitis upon adoptive transfer (350). These data suggest diabetes can be reversed with increased thymic tolerance. The non-MHC and non-lyp loci may indeed influence the repertoire of autoreactive T cells and distinguish lyp animals susceptible or resistant to autoimmunity.
BB rats raised in a germ-free environment develop diabetes.  Nevertheless recent studies indicate that treatment of diabetes-prone BB rats with antibiotics which change their GI flora (with correlation with diabetes outcome), especially when combined with a casein free diet are protected from the development of diabetes (351).
Long Evans Tokushima Lean (LETL) Rat Model (Komeda Diabetes-prone)
An additional rat model of spontaneous development of type 1 diabetes was described in 1991 (352). Inbreeding of a rat showing signs of diabetes in 1983 resulted in the LETL strain maintained at the Tokushima Research Institute. These rats have the advantage of spontaneous disease incidence without a gender bias or a requirement for lymphopenia. Like the NOD mouse, infiltration of the pancreas, salivary, and lachrymal glands is observed (352). The disease shows a polygenic mode of inheritance (353) including a large contribution of the MHC RT-1u haplotype shared with the BB-DP rat (352). Selection of the Komeda Diabetes-Prone rat substrain improved disease penetrance, with 100% insulitis and 70% diabetes by 120 days of age (354). A genome-wide scan of this strain identified a novel non-MHC associated disease locus on rat chromosome 11 which was shown to be essential for development of insulitis (354).  Mutated Cblb alleles when introduced on a non-KDP RT1 (u) rat resulted in a low incidence of diabetes and thyroidits, suggesting additional modifier genes influence penetrance (355).  An additional Komeda Non-Diabetic (KND) substrain was established which, although genetically very similar to the KDP strain, does not develop type 1 diabetes (356). This strain serves as an important control in both genetic and immunological studies of the KDP rat. This rat model thus most closely resembles human type 1 diabetes in its spontaneous onset, lack of gender bias, lack of lymphopenia, and association with MHC class II. The non-MHC mutation segregating as a recessive is a nonsense mutation of the Cblb gene (Casitas B-Lineage Lymphoma b gene) which is important for T cell regulation (357). Cbl-b is a ubiquitin ligase important for CD28 co-stimulation during T cell activation. Analysis by two groups for polymorphisms of the human Cbl-b gene failed to find an association with human type 1 diabetes (358;359).  It does not appear that Cbl-b locus contributes to development of type 1 diabetes of man though there is one report of an interacting with CTLA-4 (358-360).

Figure 7. The Cblb mutation of the Komeda diabetic rat.

LEW.1AR1/Ztm- iddm rat
In 2001 Lenzen and coworkers described a new rat model of insulin deficient diabetes in which approximately 20% of the animals (no sex bias) developed diabetes characterized by lymphocytic infiltration of the pancreas from rats that were not lymphopenic (361;362). Similar to other rat strains with immune-mediated diabetes, these animals had RT1 haplotype with u alleles (RT1.Aa B/Du Cu (361) and is a recombinant RT1 haplotype (RT1r2). In addition to islets, larger pancreatic ducts show mononuclear infiltrates. The islet infiltrates are predominantly CD8 T cells (>50%) with a relatively small percentage of CD4 cells (4%), and many ED1-positive macrophages (362). At early stages of infiltration macrophages predominate (363).  Of note non-diabetic rats lacked islet infiltration, with none of the non-diabetic animals having infiltrates! Other organs, including endocrine organs did not have infiltrates. In summary, the LEW1.AR1/Ztm-iddm rat is a novel model of islet destruction in the presence of islet infiltrates that differs from other models in that age-matched animals lacking diabetes did not have infiltrates. This suggests an acute disease process and one with dramatic heterogeneity between rats, which may be due to an environmental factor, though at present, there is no evidence for such.
TCR Tg (transgenic) Mouse Models of Type 1 Diabetes
A number of T cell clones have been isolated from the spleens of diabetic NOD mice (BDC2.5, BDC6.9) (364), and the pancreata of pre-diabetic (222;365-367) NOD or RIP-B7.1 NOD transgenic (368) mice and islet-transplanted diabetic (369) NOD mice, and from islets of NOD mice (BDC12-4.1) (222). Both CD4 and CD8 clones have been generated (370) which have varying abilities to induce or suppress insulitis and diabetes (371). The TCR (T Cell Receptor) of many of these clones have been utilized to produce TCR transgenic (Tg) mice on various backgrounds (257). A summary of TCR transgenic diabetic mice is listed in Table 3.4.  A major advance has been the development of the technology to produce retrogenics mice (257;372-375) (and Kobayashi et al PNAS in press) that has rapidly been applied to diabetogenic T cell receptors.

Table 3.4. T Cell Receptor transgenic NOD Mice.

The BDC2.5 TCR transgenic mouse was generated from the BDC2.5 CD4 T cell clone specific for pancreatic islet antigen (Ag) (377). The BDC2.5 TCR transgenic mouse develops a uniform insulitis at 3 weeks of age followed by a low incidence of overt diabetes. Surprisingly, BDC2.5 transgenic mice on the C57BL/6.IAg7 background showed a higher incidence of diabetes than on the NOD background. The difference in disease incidence was mapped to the Idd7 disease-resistance interval in the NOD (61). The BDC2.5 TCR transgene bred onto the NOD.scid background resulted in a much accelerated onset of type 1 diabetes, suggesting regulatory cells prevent diabetes induction in the normal NOD setting. Of note, diabetes is restored in the absence of NKT cells with deletion of the CD1 locus (378). Another  CD4+ clone, BDC6.9, isolated in Kathryn Haskin’s laboratory has been made into a TCR transgenic mouse. Unlike the BDC2.5 transgenic mice, BDC6.9 transgenic mice develop early diabetes by the 3rd and 4th generation backcross to the NOD background (162). The advantage to this clone is the ability to manipulate the presence and absence of self-Ag, as the stimulatory Ag is found in pancreas from NOD and SWR mice, but not other common strains of mice. BDC2.5 T cells have been utilized to create potent regulatory T cells that are able to suppress the spontaneous disease of the NOD mouse (197;379).
Daniel and Wegmann isolated a series of T cell clones from islet infiltrates and found that the majority reacted with insulin and the great majority of insulin-reactive (>90%) CD4 clones reacted with the insulin peptide B:9-23 (B chain, amino acids 9-23). A transgenic T cell receptor mouse has been established with the 12-4.1 T cell receptor anti-B:9-23 clone. This T cell recognizes insulin peptide sequences 13-23, using the common alpha chain motif (Valpha 13.3 (TRAV5D-4*04) with Jalpha 53) (380). As a retrogenic the T cell receptor of the 12-4.4 clone is also able to induce insulin autoantibodies and diabetes (Kobayashi et al PNAS in press).  The 12-4.1 and 12-4.4 alpha chains have the same TRAV5D-4*04 T cell receptor segment, related Jalpha segments, but different alpha chain N region and different Vbeta segments.  The ability of multiple T cell receptors with the related alpha chains that target the insulin B:9-23 peptide to induce insulin autoantibodies and diabetes has led to the hypothesis that such targeting is primarily driven by the alpha chains (30).
Both Class II-restricted CD4+ and class-I restricted CD8+ T cell clones isolated in the laboratory of  Santamaria have been developed into TCR transgenic mice with either beta chains alone or in combination with alpha chains. A majority of CD8+ T cell clones isolated from the pancreas of pre-diabetic NOD mice show the same TCR Va17/Ja42 with N region sequence M-R-D/E, which has been used to generate 8.3 TCRab transgenic mice. The Vb8.1 CD8+ TCR transgenic mice, transgenic for the TCR beta chain only, and 8.3 CD8+ transgenic mice show an accelerated incidence of disease with a reduced period of benign insulitis (365). In the 8.1 beta-chain only transgenic mice, the islet-reactive T cells were found to have the same alpha chain as the original clone, suggesting a selection of this clone in diabetogenesis (365). The target antigen of the 8.3 clone is the molecule IGRP (islet-specific glucose-6-phosphatase catalytic subunit-related protein) and an excellent tetramer reagent (using a mimotope of the native peptide NRP-V7) allows staining of reactive T cells from peripheral blood of NOD mice, with numbers of T cells correlating with disease progression (192). Islet-reactive CD8+ T cells increase their avidity binding to the 8.3-reactive peptide presented in Kd as shown through tetramer analysis (193).  Kay and coworkers have recently found that the 8.3 transgenic induction of diabetes is dependent upon an immune response to insulin by blocking such a response with a proinsulin transgene (3).
The NY4.1 transgenic mouse is from a CD4 clone. This TCR transgenic mouse has the same disease incidence on RAG2 KO or normal NOD background, confirming the ability of some CD4+ cells to induce diabetes in the absence of CD8+ T cells (365). However, breeding the Vb8.3 CD8+ TCRab transgenic mice onto the RAG2 KO background results in greatly reduced disease onset and demonstrates a role for CD4 cells in the efficient recruitment of CD8+ T cells to islets (366). The NY4.1 as a retrogenic induces diabetes (257). The transition from benign insulitis to overt diabetes appears to be a loss in the ability to regulate autoreactive cells. Epitope spreading and affinity maturation have been shown to be important in this process.
Previous reports demonstrated that a CD8+ T cell clone (G9C8) isolated from the pancreas of RIP-B7.1 transgenic mice did not require CD4 cells for disease transfer (368). In a more recent study a TCR transgenic mouse (NOD.AI4) was generated from a CD8 clone which was isolated from an unmanipulated NOD pancreas. The transgenic T cells were capable of inducing diabetes free of CD4 help or transgenic costimulation, supporting the idea that high-affinity CD8 T cells are independently capable of beta cell destruction (367). Of interest the clone apparently requires both Kd and Db to be present for cytotoxicity (78). However, another CD8+ islet-reactive TCR transgenic, the 9.3 transgenic, showed accelerated diabetes compared to normal NOD mice but no diabetes on the NOD.scid background (381). The unusual finding in this transgenic system was the necessity for two dissimilar receptors on the cell surface for disease incidence as identified by co-staining with antibodies specific for the transgene and a mix of other Vbeta-specific Antibodies. The authors conclude that the lowered expression of the pancreatic-specific receptor due to coexpression of other TCRs allowed escape of the transgenic receptors from negative selection. Adoptive transfer studies showed these cells did require CD4 help for diabetes induction (381).
Vignali and coworkers have introduced a powerful new technique for the study of T cell receptor expression that they term “retrogenics” in distinction to transgenics. The use retroviral- mediated stem cell gene transfer of T cell receptor sequences and bone marrow transplantation of the transduced bone marrow. For the production of retrogenic mice a retrovirus is used to introduce genes coding for T Cell receptor chains into bone marrow cells that are then transplanted into immunodeficient mice (usually SCID or rag-/-).  Within weeks one can observe the development of insulin autoantibodies, insulitis and diabetes.  The figure below illustrates the general methodology for production of transgenic mice where one can introduce a series of Valpha chains of anti-islet T cell receptors (Nakayama et al unpublished) that differ in different segments or complete alpha a beta chain T cell receptors. If the donor marrow is from a Calpha knockout mouse, the only alpha chain produced in the retrogenic mouse will be the introduced chain.  If the donor mouse is a SCID and complete alpha and beta TCR chains are introduced, only a single T cell receptor will be produced.

Vignali and coworkers studied two T cell receptor positive controls and showed that BDC2.5 and 4.1 retrogenic bone marrow induced diabetes in NOD-scid mice. In contrast T cell receptors for GAD206-220 and GAD524-538, though reactive cells were present in vivo, failed to influence the development of diabetes of NOD mice (375). We believe this is consistent with the lack of GAD within islets of mice and the demonstration that tolerance to GAD did not influence development of diabetes (225). More generally the retrogenic technique provides a powerful tool to rapidly study many T cell receptors in vivo (257) .  Studies of multiple T cell receptors targeting different islet autoantigens indicate that only a subset lead to insulitis and less produce diabetes.  In particular anti-GAD T cell receptors are not able to induce insulitis or disease, T cell receptors targeting the insulin B:9-23 peptide can cause disease, and a set of T cell receptors are more diabetogenic that react with currently unknown islet autoantigens.  It is likely that production of retrogenics will accelerate the study of T cell receptors and importantly segments of T cell receptors to determine targeting and pathogenicity.   For instance DiLorenzo and coworkers have utilized retrogenics to produce large numbers of CD8 T cells with specific T cell receptors to study in vitro (372). A disadvantage of the retrogenics technique is that the introduced T cell receptors are naturally not heritable, and for each experiment new mice need to be transduced.
Transgenic Expression of NeoSelf-Antigens in Pancreas and Diabetes Incidence
A limitation for many of the studies using TCRs from pancreatic-reactive clones is the lack of knowledge of the antigen. This is a limitation that is rapidly being addressed, now with anti-insulin TCR transgenics (BDC12-4.1 TCR), IGRP (islet-specific glucose-6-phosphatase catalytic subunit-related protein) transgenics (8.3 TCR transgenics) (191), and AI4 transgenics (?DMK target autoantigen) (226). Nevertheless, studies with defined antigens introduced into the pancreas by transgenesis on the RIP promoter, combined with T cell receptor transgenics have provided an elegant system for evaluating tolerance to islet expressed molecules. A list of models is present in Table 3.5. A important caveat, not recognized in the very initial studies of Rat Insulin Promoter driven Neo-Antigen expression is the expression of the transgene neoSelf-Antigen in lymphoid tissues such as the thymus. Thus autoimmunity can be influenced not only by antigen in the target tissue but dramatically by central tolerance mechanisms. The initial studies by Miller and coworkers with RIP induced expression of a “foreign” class I molecule, elegantly demonstrate the importance of minute expression within the thymus of neo-Self-Antigen, with tolerance dependent upon thymic expression of the antigen and not expression within islets (382).
Transgenic expression of viral LCMV under the RIP promoter is associated with ignorance of the T cells to the LCMV antigen unless prompted by LCMV systemic infection which leads to diabetes (383;384). It is hypothesized that this ignorance is due to poor cross-presentation of antigens such as gp33 by CD4 T cells (385). The diabetes in this model always requires CD8+ T cells whereas CD4+ T cells are only required in some lines (386-388). When the RIP-LCMV is crossed with a TCR transgenic mouse specific for LCMV, the animal still requires systemic LCMV infection for diabetes induction, although in this model the disease onset is accelerated compared to the non TCR transgenic model (384). In the LCMV transgenic model, the CD8+ T cells do not require CD4+ help as suggested by disease incidence with LCMV infection in class II-deficient (387) and CD40L-deficient animals   (120). As with other systems, crossing the RIP-LCMV with RIP-B7.1 transgenic mice, which provides costimulation directly on islet cells, results in spontaneous diabetes (389;390). Recent work with a defined LCMV protein (NP) crossed with a RIP-IL4 transgenic mouse showed protection from diabetes following LCMV-infection (391). The IL4 was shown to inhibit the acquisition of cytolytic activity yet promote expansion and survival of CD8+ T cells through an increased expression of B7.2 on DCs (391).
Another neo self-Antigen model targets expression of flu hemagglutinin (HA) antigen to pancreatic beta cells using the RIP promoter. In one set of models HA expressed with Igk promoter results in expression in thymic medullary and cortical epithelial cells (as wells a B lymphocytes)
and thymuses induce CD25+ regulatory T cells (392). Unlike the LCMV protein, the Ins-HA mice appear tolerant to the HA Antigen following flu immunization if administered in adulthood (389;393). A CD8+ TCR transgenic specific for HA (CL4) can adoptively transfer diabetes in Ins-HA recipients if the transgenic T cells are pre-activated in vitro (394). In addition, CL4xIns-HA double transgenic mice are not tolerant and get spontaneous diabetes by 8 days of age with few T cells in the islet infiltrate (393). Further work on the CL4 TCR showed this TCR possessed an unusually high affinity to HA/Kd.

Table 3.5. Neo-Self Antigen Models to Type 1 Diabetes.

This high affinity is evident in the activation of the TCR in the absence of costimulation or with soluble monomeric complexes (394). Like this CD8+ TCR transgenic mouse, tolerance was also not established to the HA Antigen in mice doubly transgenic for the Ins-HA transgenic and a CD4+ anti-HA TCR transgenic (395). Other HA-specific CD4+ TCR transgenic mice crossed with Ins-HA developed varying degrees of diabetes depending on the background strain (395;396). Recently, the Ins-HA model has been used in studies of peripheral tolerance of CD8+ T cells in various mouse studies. Defects in peripheral tolerance in NOD mice compared to BALB/c and B10.D2 were confirmed by showing higher avidity binding of NOD T cells to Kd-HA tetramers (189).
Similar to the Ins-HA model, RIP-OVA transgenic mice can become diabetic when transferred with pre-activated anti-OVA CD8+ T cells (397). In addition, early diabetes is observed in OT-1 (TCR transgenic anti-OVA)xRIP-OVA double transgenic mice, suggesting a lack of tolerance and ignorance in this model (398). Another model places the class I Kb molecule under control of the RIP-promoter (382). Crossing the RIP-Kb mice with a CD8+ TCR anti-Kb transgenic mouse shows tolerance to this pancreatic islet, which is associated with both thymic deletion of high-affinity clones by residual thymic expression and functional unresponsiveness of remaining clones. This tolerance can be broken, resulting in diabetes induction, with RIP expression of IL2, a cytokine which has been shown to reverse T cell anergy (399).
Transgenic expression of the exogenous hen egg lysozyme (HEL) protein by the RIP promoter, together with an anti-HEL TCR transgenic, shows thymic deletion and peripheral anergy of transgenic cells on the C57BL/6 background. However when bred onto the NOD background, maintaining the necessary H-2k MHC, the HEL model showed defects in both thymic deletion and T cell anergy induction leading to overt diabetes in the NOD RIP-HEL/TCR double-transgenic mice. This model clearly demonstrates the influence of NOD non-MHC genes on tolerance processes (400).
As mentioned above, transgenic mice expressing the costimulatory molecule B7.1 on pancreatic islet cells (RIP-B7.1) have been useful in studying “spontaneous” diabetes by crossing this mouse with models which otherwise do not show disease incidence. This mouse on the C57BL/6 background does not get diabetes but can become diabetic when crossed with RIP-LCMV, RIP-TNFa, or RIP-IE or with mice expressing human HLA molecules associated with type 1 diabetes. The usefulness of the RIP-B7.1 model is to address the sensitivity of disease induction allowing the systems to be stressed to determine if diabetes can occur. RIP-B7.1 transgenic expression on the NOD background results in diabetes incidence as early as the first backcross although IAg7 homozygosity is required (394). The largest effect of B7.1 expression is on the CD8 cells, thereby negating the necessity of CD4+ help although the presence of CD4+ cells accelerates the disease progression by aiding in migration of cells to the pancreas. Unlike spontaneous NOD diabetes, B cells are not required for disease in the RIP-B7.1 model (401).
Immunization/Immunoregulatory Induced Models of Diabetes
For many experimental autoimmune disorders it is possible to induce pathology by immunizing with the presumed target autoantigen, creating disorders such as Experimental Autoimmune Encephalitis (EAE) (113). Until recently it has not been possible to induce immune mediated diabetes in mice or rats with specific islet autoantigens. Immunization with a heat shock protein was associated with transient hyperglycemia (402) and this molecule, despite its ubiquitous expression is reported to be both an autoantigen and activator of the innate immune system (227).
There is actually an “old” literature where insulin immunization of cows and rabbits induced “insulitis” and even diabetes in rabbits (403-408) but with the “difficulty” of the bovine model, and apparently problems in reproducibly inducing diabetes in the rabbit model, (as available insulin became purer), this area of investigation was not pursued. It was certainly recognized that most patients treated with insulin (again becoming less with improved insulin purity) produced insulin antibodies and immunoreactivity to insulin was studied in mouse models (409) and man (410). It is somewhat surprising that it was not possible to induce diabetes with proinsulin/insulin/or insulin peptides. Recent studies indicate that insulin and its peptides can induce anti-islet autoimmunity (insulin autoantibodies and insulitis), but in normal mouse strains (e.g., BALB/c, C57BL/6) this is not sufficient to engender diabetes, while in mice carrying a transgene inducing B7.1 (costimulatory molecule) expression on islets, diabetes is routinely induced following insulin/ insulin peptide B:9-23 immunization given the appropriate MHC alleles (216;411-414) .

Figure 8. Induction of insulin autoantibodies following subcutaneous injection of insulin peptide B:9-23 in incomplete Freund's adjuvant.

Mice with I-Ad (e.g., BALB/c mice) or I-Ag7 respond to subcutaneous immunization with insulin peptide B:9-23 (e.g.,, NOD mice) with the rapid production of insulin autoantibodies (Figure 8) (411-413). NOD mice respond in a similar fashion, but because they already produce insulin autoantibodies, one observes increased levels rather than de novo production of autoantibodies. In NOD mice the expression of insulin autoantibodies persist for months, rather than being transient following immunization with B:9-23 peptide. These are actual autoantibodies in that they can be blocked with “mouse” insulin, and the antibodies react with intact insulin and cannot be absorbed with the immunizing B:9-23 peptide. Of note the response is MHC restricted as shown by analysis of congenic strains and no other peptide of proinsulin to date has been found to induce the autoantibodies.
When NOD mice are immunized with the insulin peptide B:9-23 they also produce antibodies reacting with the peptide itself, that depending on the chronicity of administration can induce anaphylaxis (415;416). Though insulin autoantibodies are induced by insulin peptide B:9-23 immunization of BALB/c mice, insulitis was not observed. To produce insulitis, immunization of BALB/c mice was combined with two weeks of administration of subcutaneous poly-IC. Poly-IC is a toll 3 receptor activator and mimic of double stranded RNA. It induces circulating levels of interferon alpha, and acts through induction of interferon alpha to induce diabetes (417).   Despite the development of both insulin autoantibodies and insulitis, the BALB/c mice do not progress to diabetes. In mice expressing B7.1 in islet beta cells, diabetes can be induced with either poly-IC or insulin B:9-23 peptide alone, as well as with both together (412). Thus tolerance to insulin can be readily broken, and there is a dramatic restriction in terms of the peptides able to induce disease. We suspect that normal BALB/c mice are already sensitized to the insulin B:9-23 peptide prior to immunization with this peptide given the rapidity of the induction of insulin autoantibodies reacting with a conformational epitope of insulin. For the C57Bl6 mouse that does not respond to insulin peptide B:9-23 (I-Ab) disease can be induced in B7.1 islet transgenic mice by immunization with intact insulin, presumably by reactivity to a different epitope of insulin (414).  Induction of cytokines can also accelerate the development of diabetes in NOD mice depending upon the stage at which for instance NOD mice with established insulitis when infected with rotavirus (418).
Co-stimulation
As mentioned previously, the expression of the B7.1 molecule on islets enhances the diabetogenicity of a series of stimuli that by themselves are insufficient to induce diabetes. This includes insulin peptide B:9-23 (BALB/c mice) (412), poly-IC (C57BL/6 mice) (419), LCMV-GP peptide (420), LCMV-GP gp33 and preproinsulin encoding plasmid (421), human DR4 transgene (422), human DQ8 transgene (422), transgenic islet expression of IL2 (423), and transgene induced expression of TNF alpha on beta cells (424). It was hypothesized the B7-H1 would be a negative regulator of autoimmunity, but transgene induced expression of B7-H1 on islets also induced diabetes (425).
"Humanized" Mice
Another approach undertaken in the past decade in the study of type 1 diabetes is the transgenic expression in mice of human genes.  There are two general approaches:  Introduction of human genes into mice and 2. Transfer of human cells into immunodeficient mice (18).  Both approaches represent important technical advances (19;426;427).
The class II HLA molecule DQB1 chain with serine, alanine or valine at P57 of the b chain, similar to the NOD MHC class II haplotype, confers the greatest risk for type 1 diabetes in humans. The DQ8 (DQA1*0301 plus DQB1*0302 (alanine at P57)) haplotype has been expressed in transgenic mice (428). The DQ8 transgenic mice do not develop diabetes even when backcrossed to the NOD background. However, when these mice are crossed with the RIP-B7.1 transgenic mice, 80% of the mice develop insulitis and type 1 diabetes whereas control human DQ6 (diabetes-resistant haplotype)xRIP-B7.1 double-transgenic mice do not develop insulitis or diabetes. The DQ8xRIP-B7.1 transgenic mice have been used to map approximately 10 different T cell epitopes in human GAD65, one of which is also a dominant T cell epitope in NOD mice restricted to IAg7 (429;430). transgenic expression of the human disease-associated DR4 allele in mice showed surprisingly that this allele, in combination with the highly susceptible DQ8 allele, actually protected mice from disease. Whereas 25% of DR4/RIP-B7 mice developed diabetes compared to >80% of DQ8/RIP-B7 mice, the triple transgenic DR4/DQ8/RIP-B7 had only a 23% disease incidence, showing a dominant protective effect for the DR4 allele   (137). The diabetes incidence afforded by the immunostimulatory RIP-B7.1 transgene enables studies of human autoimmunity genes in mice with the important caveat that the pattern of disease can be very unusual as found by Lipes and coworkers with pancreatitis depending on the strain in the absence of murine class II genes (431). RIP-B7/DRB1*0404 HLA transgenic mice spontaneously develop T cells reacting with GAD65 and Glial fibrillary acid protein (GFAP) and approximately 20% develop diabetes at a mean age of 40 weeks (432). A human T cell receptor transgenic mouse (anti GAD65/67 (555-567) develops insulitis on a human DR4 rag-/- background (433).
The introduction of human DQ8 onto the NOD background has resulted in a series of autoimmune disorders. In particular a disorder resembling dermatitis herpetiformis is induced followed immunization with gluten (434), a disorder resembling polychondritis following immunization with type II collagen (435).
Marron and coworkers demonstrated that transgenic expression of the common human class I allele, HLA-A2.1, accelerated the development of diabetes in NOD mice. This effect was not simply a result of expression of any human class I allele, as HLA-B27 (associated with ankylosing spondylitis in man) if anything decreased development of diabetes (436). The A2.1 mouse islets have been used to study a model of human lymphocytes in NOD-Rag1 (null) Prf1 (null) mice rejecting these islets. Transplanting both into the spleen enhanced graft and lymphocyte interaction. When lymphocytes came from A2.1 positive donors the grafts survived, while from HA-A2 negative donors, the grafts were rejected (437).  Humanized HLA-A2.1 mice (further refined to eliminate murine class I antigens) have been utilized to identify peptides recognized by CD8 t cells of potential relevance to human type 1 diabetes (426;438;439).
Human Lymphocyte Transplantation
It is likely that many additional human molecules will be introduced into murine strains to “humanize” the immune system.  An approach to utilize transplantation/transfer of human cells directly into mice allows the rapid introduction of multiple genes and cell types (19).  These studies have been greatly enhanced with the development of IL2 common gamma chain knockout mice on the NOD-scid background (440). This combines severe adaptive immunodeficiency with abnormalities of NK T cells and allows engraftment of both peripheral blood cells and stem cells and the study of alloreactivity with transplant rejection (18).
Conclusion
The multiple animal models of type 1A (immune mediated) diabetes amply demonstrate that there are many pathways that can result in immune mediated (autoimmune) destruction of islet beta cells, and that the spontaneous animal models each represent a different pathway. One can generalize with the following observations:

1. T cell destruction of islet beta cells can be entrained by the targeting of single islet molecular determinants by T cells with a single T cell receptor (and obviously also with multiple receptors). In general however, both CD4 and CD8 T cells collaborate in causing beta cell destruction. It is likely that each T cell receptor (e.g., clone or transgenic) will follow its own rules, depending on the avidity and other qualities of the receptor), in terms of whether ignorance, tolerance or beta cell destruction results.
2. There is often T cell receptor biased chain utilization of the T cells targeting a specific peptide autoantigen.
3. In the spontaneous animal models multiple islet antigens are the targets of autoimmunity.
4. It is likely in some common models, such as the NOD mouse, a single determinant may be essential for disease, such as the insulin sequence B:9-23.
5. The finding that a given antigen is the target of autoreactive T lymphocytes, and even that T cells targeting that antigen can transfer disease, does not mean that the targeted molecule is important for pathogenesis of a spontaneous disorder, and the best current test for importance of a target epitope, is gene knockout or mutation. If disease occurs at the same rate in the absence of the molecule, it cannot be essential, though may contribute in a minor way to the disorder.
6. Multiple genes synergistically contribute to autoimmunity and the MHC is often essential in determining the organ targeted.
7. Both pathogenic T cells and regulatory T cells (and other regulatory cells) modulate the development of disease.

The tremendous increase in knowledge relative to the animal models will hopefully translate into informed trials for the prevention of type 1A diabetes in man and long-term success of islet transplantation. There is an important opportunity to attempt to develop assays and paradigms in the animal models that are directly applicable to man. In particular in man one must sample the peripheral blood for disease prediction, and every effort should be made in the “simpler” animal models to develop assays sampling blood that can be applied to man.

Reference List - links to PubMed available in Reference List.

Chapter 3 Powerpoint slide set - Updated 3/08

For comments, corrections or to contribute teaching slides, please contact Dr. Eisenbarth at: george.eisenbarth@ucdenver.edu

Back to Type 1 Diabetes Table of Contents