The Rat Diabetes Susceptibility Locus Iddm4 and at Least One Additional Gene Are Required for Autoimmune Diabetes Induced by Viral Infection

BBDR/Wor, WF.Iddm4, and WF.ART2a rats (all RT1^u/u, ART2^a) were obtained from colonies maintained by us. WF.Iddm4 congenic rats were generated by repetitive (BBDR/Wor × WF) × WF backcrosses using a marker-assisted selection protocol as described (10). They were studied at the N6 generation. WF.ART2a congenic rats were also developed by us and differ from ordinary WF animals in that they express the “a” rather than the “b” allotype of the ART2 T-cell alloantigen on chromosome 1 (10). For simplicity, we refer to them here as WF rats. An F2 intercross was bred from (BBDR × WF)F1 hybrids. Animals were housed in viral antibody–free conditions, confirmed monthly to be serologically free of rat pathogens (10), and maintained in accordance with recommendations in the Guide for the Care and Use of Laboratory Animals (National Academy of Sciences, 1996).

Genomic DNA was prepared as described (10) (online appendix Fig. 3 [available from http://diabetes.diabetesjournals.org]). Microsatellite markers were placed evenly throughout the 20 autosomes. The source of primers and positions of markers on the genetic map are given in online appendix Fig. 3. Primers were end labeled using ³²Pγ-ATP, used in a PCR reaction, and resolved by polyacrylamide gel electrophoresis as described (6). The position of markers on the genetic map was established by inspection of the dataset and conventional calculation methods to establish meiotic map distances, which are expressed in centimorgans (cM) or megabases (Mb) according to the rat genome sequence, June 2003 build (available at http://genome.ucsc.edu).

Linkage of diabetes with segregation of BBDR-origin alleles was evaluated by composite interval mapping (CIM) using model 6 of the Zmapqtl program in Windows QTL Cartographer v1.30 (available at statgen.ncsu.edu/qtlcart/cartographer.html). CIM combines classical interval mapping with multiple regression analysis, allowing for more precise quantitative trait loci (QTL) localization than classical interval mapping (12).

KRV-UMass was propagated in normal rat kidney cells grown in Dulbecco’s minimal essential medium. Poly I:C (Sigma, St. Louis, MO) was dissolved in Dulbecco’s PBS, sterile filtered, and stored at −20°C until used. Contaminating endotoxin concentration was <50 units/mg (Charles River Endosafe, Charleston, SC). In studies of KRV alone, rats of either sex 22–28 days old were injected intraperitoneally with 10⁷ PFU in a volume of 1 ml. In other experiments, rats 21–25 days of age of either sex were injected intraperitoneally with poly I:C (1 μg/g body wt on 3 consecutive days) and either not treated further or injected on the following day with KRV. Pretreatment with poly I:C was used because it increases the frequency of diabetes in KRV-treated BBDR rats from ∼30 to 100% (8). Animals were screened three times weekly for glycosuria (Tes-Tape; Eli Lilly, Indianapolis, IN). Diabetes was diagnosed on the basis of a plasma glucose concentration >11.1 mmol/l (OneTouch Ultra Glucometer; LifeScan, Milpitas, CA). For study of insulitis, pancreata were removed, fixed in formalin, and stained with hematoxylin and eosin. Insulitis was graded by a qualified pathologist on a scale of increasing intensity from 0 to 4+ as described (10).

We first confirmed (8) that a significant fraction (41%) of parental BBDR rats become diabetic after infection with KRV alone, that none become diabetic in response to a 3-day course of poly I:C alone, and that 100% become diabetic in response to KRV after poly I:C (Table 1). We also confirmed (13) that WF rats resist diabetes induction in response to either KRV or KRV plus poly I:C (Table 1).

We next tested N6 generation WF.Iddm4 congenic rats (10) for disease susceptibility. To our surprise, we observed WF.Iddm4^d rats to be uniformly resistant to diabetes in response to KRV infection either alone or after poly I:C, despite maintaining the expected degree of susceptibility to diabetes after treatment with poly I:C and Treg depletion (Table 1).

To determine whether Iddm4 acts only in the presence of additional BBDR-origin genes, we generated (BBDR × WF)F1 progeny. We observed that all F1 rats were resistant to KRV alone, but 38% were susceptible to treatment with KRV plus poly I:C. To identify susceptibility genes, we then generated (BBDR × WF)F2 progeny and treated them with KRV plus poly I:C. Diabetes occurred in 59 of 190 animals (31%) in this segregating population and affected both males and females (Table 1). Looking first at the Iddm4 interval, we observed that diabetes occurred almost exclusively in animals with at least one BBDR-origin allele of Iddm4 (58 of 59 diabetic animals, Table 2). The presence of the BBDR allele of Iddm4 was 98% sensitive but only 31% specific in predicting susceptibility to diabetes, implying that at least one gene of BBDR origin is required for the expression of diabetes in response to infection. We therefore performed a genome-wide scan on this F2 cohort. The remaining genome was assessed for linkage using 144 markers on the 20 autosomes (online appendix Fig. 3).

Composite interval analysis of the linkage data is shown in Fig. 1. Iddm4, as expected, showed strong linkage to diabetes (logarithm of odds [LOD] >6.0 at ∼36 cM on chromosome 4). A second locus on chromosome 17 was linked to diabetes in the F2 population with an LOD score of 3.7. This locus has been designated Iddm20 by the curators of the Rat Genome Database (available at www.rgd.mcw.edu).

To determine the mode of inheritance and the interaction between these two loci, we analyzed the dataset using a life-table analysis. As shown in Fig. 2, the highest likelihood of diabetes onset occurs in animals in which the BBDR-origin allele of Iddm4 is homozygous or heterozygous and Iddm20 is homozygous.

The genome-wide scan positioned the Iddm20 locus in a 1-LOD interval bounded by D17Rat61 (at 19.7 Mb) and D17Rat115 (at 27.7 Mb) (available at http://genome.ucsc.edu). To identify candidate genes within this interval, we constructed a preliminary map using the databases at UCSC and Ensembl (available at http://genome.ucsc.edu and http://www.ensembl.org). The genes in the Iddm20 region have their human orthologs on human chromosomes 5, 6, and 9 and on mouse chromosome 13. Of interest is a confirmed mouse diabetes QTL (Idd14) in this homologous region (14,15). Candidate genes in the Iddm20 interval are listed in online appendix Table 4.

Histologic analysis of islets revealed nearly complete concordance of insulitis scores with diabetes phenotype. Among the 59 diabetic F2 animals, the mean insulitis score was 3.7 with 49 scored 4+ or end-stage insulitis. In contrast, among 127 nondiabetic rats with technically satisfactory specimens, the mean insulitis score was 0.3 with 111 (87%) being entirely normal and 8 of the remaining 16 exhibiting only 1+ insulitis. Exocrine pancreatitis was absent.

These data establish linkage of rat genotype to a form of environmental perturbation—infection—that is potentially important in the pathogenesis of autoimmunity. They confirm that Iddm4 is an exceptionally strong non-MHC determinant of susceptibility to autoimmune diabetes in the rat (6,9–11). In previous studies of Iddm4, diabetes was induced by chronic treatment with poly I:C plus Treg depletion. The present data now extend the role of Iddm4 in diabetes pathogenesis to virus-induced disease expression. They also illuminate the complexity of environmental interaction with genetic susceptibility. The diabetogenic potential of Iddm4 is readily discernable in congenic rats treated with poly I:C and Treg depletion but is far less apparent in animals treated with KRV plus poly I:C unless additional BBDR genes are present. We have discovered at least one of these genes, designated Iddm20, on chromosome 17.

The Iddm20 interval (online appendix Table 3) contains at least one gene of particular interest: Syk. This gene is involved in T-cell receptor–dependent signaling pathways and interacts with Cblb (16). This could be important because Cblb is a known rat diabetes susceptibility gene (17). Loss of function mutations in Cblb lead to activation of autoreactive diabetogenic T-cells in the absence of full costimulation (17).

Iddm20, like Iddm4, appears to act as a genetic dominant with incomplete penetrance. There is clear disease-promoting activity in the Iddm20 heterozygote. In the poly I:C plus Treg system, ∼70% of both WF.Iddm4^d/w heterozygotes and WF.Iddm4^d/d homozygotes become diabetic (10). In the KRV plus poly I:C model, homozygosity for diabetogenic alleles at the Iddm20 locus increases the penetrance of diabetes in Iddm4^d/d rats from 25% (in Iddm20^w/w animals) to 56%, and it increases penetrance in Iddm4^d/w rats from 11 to 59%. We therefore regard Iddm20 as a modifier of the Iddm4 locus.

The mechanisms by which poly I:C and KRV infection act to induce diabetes in genetically susceptible rats are not yet known. We speculate that Iddm4, Iddm20, or both define a strain-specific response of BBDR rats to KRV infection. KRV is known to infect lymphocytes in the pancreatic lymph nodes (but not the islets) of BBDR rats (18). More recent studies have revealed that KRV also causes a decrease in splenic CD4⁺CD25⁺ Treg cells in both BBDR and normal WF rats (8). In adult LEW rats, KRV-UMass infection is associated with several potentially important effects on both CD4⁺ and CD8⁺ T-cell populations (19) but whether allelic variations in Iddm4 or Iddm20 regulate those responses is not yet known. By itself, KRV infection typically induces diabetes in ∼30–40% of BBDR/Wor rats (8). Pretreatment poly I:C, at a dose that is itself incapable of inducing disease, dramatically increases the penetrance of diabetes (8). The mechanism of this synergy is not clear but is likely to relate to innate immunity because poly I:C is a ligand of toll-like receptor 3 and a potent inducer of type I interferon production by various cells (20) and interleukin-1 production by monocytes (21). It also activates natural killer cells (22) and B-cells (23). In rats, it has been shown that interferon production in response to poly I:C varies substantially in different inbred strains (24), an effect that is presumably genetically determined. It will be of interest to determine whether Iddm4 and/or Iddm20 is a determinant of the magnitude of the immune response to poly I:C.

This study was supported in part by grants DK49106 (to D.L.G., J.P.M., and E.P.B.), DK 36024 (to D.L.G.), DK25306 (to J.P.M.), and Center Grant DK32520 from the National Institutes of Health. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.

We thank Dr. Michael Appel for scoring histology specimens, Michael Bates and Deborah Mullen for technical assistance, and Dennis Guberski for logistic support.