Diabetes Protection and Restoration of Thymocyte Apoptosis in NOD Idd6 Congenic Strains

NOD and C57Bl/6 (B6) inbred mouse strains were maintained in our animal facilities through brother-sister mating. In our NOD colony, the incidence of diabetes at 35 weeks of age reaches 86% in females and 60% in males. Diabetes was screened weekly by measuring glycosuria from the age of 10 weeks using colorimetric test strips. Reciprocal NOD and B6 Idd6 congenic strains were constructed by the introgression of chromosomal segments telomeric to D6Mit291, representing the Idd6 locus. To initiate the breeding of the Idd6 congenic strains, F2(B6×NOD) mice were genotyped and screened for either B6 or NOD homozygozity at markers D6Mit198, D6Mit57, D6Mit14, and D6Mit15. Two F2 males were selected as founders to establish the NOD.B6-Idd6 and the B6.NOD-Idd6 congenic strains. Introgression of the founder chromosomal segment was obtained by repeated backcrossing. At each backcross generation, the offspring was genotyped and selected for heterozygozity from markers D6Mit198 to D6Mit304. The NOD.B6(A)-Idd6 (N.Idd6) and the B6.NOD-Idd6 (B.Idd6) strains were backcrossed for 11 generations. At generation 8 of the NOD.B6(A)-Idd6 strain, a recombinant was isolated that was B6 homozygous from marker D6Mit259 to D6Mit200 and heterozygous from marker D6Mit340 to D6Mit304. Further backcrossing of this recombinant for 12 generations gave rise to the NOD.B6(B)-Idd6 strain (N.Idd6-15). To achieve homozygozity over the congenic region, the strains were bred through brother-sister mating. Thus, we intercrossed the N.Idd6 strain for seven generations, the N.Idd6-15 strain for nine generations, and the B.Idd6 strain for seven generations. Therefore, the mice analyzed in this report were N.Idd6 (N11F7), N.Idd6-15 (N12F9), and B.Idd6 (N11F7). To control for residual genomic contamination, congenic strains were genotyped with a set of markers mapping to the main Idd loci detected in crosses between the NOD and the B6 strain.

To measure the frequency of diabetes in the N.Idd6 strain, littermates that were NOD homozygous for the Idd6 region, NOD.NOD(A)-Idd6 (N.N.Idd6), were used as controls.

The protocol for induction of apoptosis in thymocytes by in vivo treatment with Dxm has been described in detail (26). In brief, 8-week-old female mice were injected intraperitoneally with 0.2 mg Dxm (Merck Sharp Dohme). Thymuses were collected and the thymocytes analyzed at 0, 8, 12, 18, and 48 h after treatment.

Thymocytes were surface-stained with anti-mouse CD8 and CD4 antibodies (Becton Dickinson). For detection of apoptotic cells, the terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick-end labeling (TUNEL) reaction was performed as described by Colucci et al. (22). Briefly, the thymocytes were fixed with 4% paraformaldehyde for 30 min at room temperature, permeabilized with 0.1% saponin for 3 min on ice, and stained with flourescein-labeled dUTP in the presence of TdT for 1 h at 37°C. Cells were analyzed by flow cytometry on a FACscan (Becton Dickinson). For detection of apoptotic cells by propidium iodide (PI) staining, thymocytes were fixed in 70% ethanol for 2 h on ice and stained overnight with PI (50 μg/ml) (Sigma) in PBS with 0.1% sodium citrate and 0.1% Triton X-100 at 4°C. Apoptotic cells were identified by hypodiploid content using flow cytometry on a FACscan.

Genomic DNA was extracted from tails according to standard techniques. The genotypes of each locus were determined by DNA marker amplification using primers purchased from Research Genetics. The genetic markers used distinguish the parental strains (B6 and NOD) by length polymorphisms in simple sequence repeats. Amplification products were analyzed in ethidium bromide-stained 4% agarose gels (3% Nusieve GTG agarose [FMC] and 1% type II A agarose [Sigma]).

Linkage of marker loci to apoptosis resistance was evaluated by a χ² test in contingency tables (2 d.f.). Marker orders and recombination fractions were calculated by three-point analysis using Mapmaker/EXP 3.0 software (beta release 3.0b) (27). Analysis of quantitative trait loci was performed using Mapmaker QTL 1.1 software (28). This program calculates logarithm of odds (LOD) scores over intervals between linked markers, provides estimates of the percent of phenotypic variation explained by a given quantitative trait locus (QTL), and compares recessive, dominant, and additive models of gene action.

The log-rank test was used to compare the cumulative incidence of type 1 diabetes between mouse strains. ANOVA tests (SSPS 11.0 for Windows) were performed to determine statistical significance of the phenotypes observed.

Our earlier study reported that resistance of NOD DP thymocytes to depletion by Dxm, scored 48 h after treatment, maps to the Idd6 locus (25). It is possible, therefore, that the Idd6 effect on type 1 diabetes may be exerted at the level of the thymic selection (25). To further define this resistance phenotype, we performed a time-course experiment where we measured DP thymocyte depletion and thymocyte apoptosis induction after dexamethazone treatment in NOD and B6 females. The TUNEL technique was used to directly quantify the proportion of apoptotic cells. Throughout the time-course experiment, the level of apoptotic thymocytes in NOD mice reflected the kinetics of the DP depletion and was significantly lower than in B6 mice (Fig. 1). The percentage of TUNEL-positive thymocytes reached a peak at 12 h after treatment in both NOD (12.9 ± 3.6) and B6 (21.5 ± 3.3) females, coinciding with a maximal difference between the two strains (Fig. 1). The proportion of remaining DP thymocytes at 6, 12, 18, 24, and 48 h after Dxm treatment was measured by staining to detect expression of the surface markers CD4 and CD8. The depletion of DP cells reached a maximum difference between NOD and B6 mice at 12 h after treatment, coinciding with the peak of apoptosis induction in NOD and B6 mice (data not shown).

To further investigate the genetic control of resistance to Dxm-induced apoptosis in NOD thymocytes, we analyzed a cohort of 72 F2 (B6×NOD) females 12 h after Dxm treatment. For each mouse, the proportion of DP thymocytes was measured. In addition, the proportion of TUNEL-positive thymocytes in 42 animals was also measured. The DP depletion trait and the apoptosis resistance trait displayed normal phenotype distributions (data not shown). For both traits, the phenotypic variance in the F2 population was larger than in the parental strains, indicating that the genetic factors controlling the trait were segregating in this cross (data not shown). An estimation of the nongenetic variance (29) ascribes 53% of the DP depletion trait and 54% of the apoptosis resistance trait variance observed in the F2 progeny to genetic factors.

To test for the involvement of the Idd6 locus in the Dxm apoptosis resistance trait, we analyzed the cosegregation of the DP depletion trait in the F2 progeny with markers comapping with the main Idd loci detected in crosses between NOD and B6. Using the χ² test of association, linkage was found between the DP depletion trait and markers linked to the Idd6 locus on chromosome 6. The association reached the highest significance values (P = 9 × 10⁻⁵) at markers D6Mit14 and D6Mit15 (data not shown), previously reported to be strongly linked to the Idd6 locus (30).

To verify that markers linked to the Idd6 locus quantitatively control the DP depletion trait in the F2 progeny, we scanned chromosome 6 using Mapmaker/QTL software (29). The LOD score values peak at markers linked to the Idd6 locus, with a maximum value of LOD 6.5 at markers D6Mit14 and D6Mit15 (Fig. 2). The QTL analysis preliminarily predicted that the locus explains 34% of the phenotypic variation in the F2 progeny, and that the NOD allele is likely to control the trait in a dominant fashion. In agreement with these results, the apoptosis resistance trait reached LOD scores of 3.5 at markers D6Mit14 and D6Mit15 (data not shown).

To further define the region of linkage of the apoptosis resistance phenotype, and to evaluate the contribution of Idd6 to type 1 diabetes pathogenesis, we constructed three congenic strains over the Idd6 region as detailed in research design and methods. The introgressed chromosomal regions are illustrated in Table 1. We analyzed the N.Idd6 strain at the N11F7 generation, the N.Idd6 strain at the N12F9 generation, and the B.Idd6 strain at the N11F7 generation.

We next studied how the type 1 diabetes development was influenced by the Idd6 region in our N.Idd6 congenic strain, which carried a 8-cM B6-derived region on distal chromosome 6. Thus, the diabetes frequency was compared in N.Idd6 congenic males and females that were either B6 or NOD homozygous for the Idd6 congenic chromosomal interval. The frequency of diabetes in female N.N.Idd6 littermate controls reached 93%, whereas in N.Idd6 females, it was significantly lower (75%) (P = 0.008) (Fig. 3A). The protective effect of the Idd6 region to type 1 diabetes was even more pronounced in males, in which the frequency of diabetes was 60% in N.N.Idd6 males compared to 30% in N.Idd6 (P = 0.003) (Fig. 3B). This demonstrates that the most distal 8-cM region on chromosome 6, carried by the N.Idd6 congenic strain, includes the Idd6 locus and protects from type 1 diabetes.

NOD, B6, N.Idd6-15, and B.Idd6 females were analyzed 12 h after Dxm treatment (Fig. 4). Apoptosis induction was scored as the proportion of TUNEL-positive thymocytes. The proportion of TUNEL-positive apoptotic cells observed in the NOD strain (19.9% ± 7.7) was significantly different from that of the B6 strain (35.1% ± 5.9) (P < 0.000), whereas the apoptosis resistance phenotype was largely reversed in N.Idd6-15 (30.3% ± 6.2) (P = 0.170), thus confirming that the apoptosis resistance in thymocytes is restored by the most distal region of the Idd6 locus. Conversely, the degree of TUNEL-positive cells was comparable in the B6.N-Idd6 strain (21.4% ± 6.1) and the NOD strain (P = 0.943) (Fig. 4A). Apoptosis detection using PI staining confirmed these observations (Fig. 4B). Collectively, these data demonstrate that a 3-cM chromosomal segment telomeric to marker D6Mit200 is sufficient to mediate the restoration of the Dxm-induced apoptosis phenotype in NOD thymocytes. We conclude that the NOD allele and the B6 allele of the Idd6 locus control 90% and 70%, respectively, of the phenotypic difference observed between NOD and B6 mice.

In this study, we have localized the Idd6 locus to an 8-cM region on distal chromosome 6 by analyzing a newly constructed N.Idd6 congenic strain. The B6-derived chromosomal region spans from marker D6Mit259 to D6Mit304 and partly protects from type 1 diabetes (18% reduced incidence in females and 50% reduced incidence in males). We conclude that Idd6 is located in a chromosomal region defined by markers D6Mit259 to D6Mit304 and controls the restoration of the apoptosis induction phenotype and partial protection from type 1 diabetes.

In agreement with our results, it was recently demonstrated that the Idd6 region from both C57BL/6 and C3H/HeJ strains partly protects from type 1 diabetes (6,31). Carnaud et al. (31) constructed an Idd6 congenic NOD strain carrying a 10-cM B6-derived chromosomal region that included the Idd6 region and the natural killer (NK) complex (31). Furthermore, a set of Idd6 congenic strains were analyzed by Rogner et al. (6), indicating that the 4-cM distal part of chromosome 6 defined by markers D6Mit57 and D6Mit304 accounts for the disease protection observed in those strains. Furthermore, two additional Idd loci on chromosome 6 were reported, namely a protective locus, Idd19, and a susceptibility locus, Idd20.

Herein, we report that N.Idd6-15 mice, carrying a B6-derived region (3 cM) spanning from marker D6Mit340 to D6Mit304, exhibit a restored thymocyte apoptosis response to Dxm. The level of apoptosis induction in N.Idd6-15 mice is similar to the high level observed in B6 mice, whereas B.Idd6 mice display a resistance phenotype similar to that detected in NOD mice. The analysis of the Idd6 congenic strains presented here indicates that a 3-cM region on the distal part of the Idd6 region controls the apoptosis resistance observed in NOD immature thymocytes.

The dominant action of this apoptosis resistance trait contrasts with the behavior of the Idd6 locus believed to be recessive. It should be noted, however, that the Idd6 locus was described in the context of a genetic backcross in which all Idd resistance alleles appear only in the heterozygous state (30). In such a situation, it is difficult to distinguish between a recessive mode of action and a dominant mode with reduced penetrance. On the other hand, we have here used an F2 intercross, allowing us to analyze the B6 resistance allele in both the homozygous and heterozygous state, concluding that in the heterozygous state the NOD allele was dominant with reduced penetrance.

Based on the results from diabetes protection in our NOD-Idd6 congenic stains, and those described by Carnaud et al. (31) and Rogner et al. (6), we conclude that a 4-cM region on distal chromosome 6 controls apoptosis resistance and accounts for the type 1 diabetes protection mediated by the Idd6 locus. We hypothesize that a genetic component within the Idd6 locus contributes to type 1 diabetes and mediates resistance to thymocyte apoptosis in the NOD mouse by altering the thymic selection process. However, it is still possible that the genetic factor representing the Idd6 locus segregates from the gene(s) controlling the apoptosis resistance trait. Further analysis of Idd6 subcongenic strains will address this question.

By comparing the diabetes protection in our congenic strain with the congenic strains described by Rogner et al. (6), we can exclude the NK complex as a candidate for Idd6. However, the NK locus cannot be excluded as a candidate for the Idd19 locus. It remains to be clarified if the NK complex of the NOD mouse can exhibit a diabetes protective effect.

Possible functions for glucocorticoids (of which Dxm is the most potent) in T-cell development have been suggested (32,33). Thus, independent signaling through the TCR or through the glucocorticoid receptor has been reported to induce apoptosis, whereas the simultaneous stimulation through both receptors may result in reduction of cell death (32). In a model of dual antagonism, it was proposed that glucocorticoid receptor (GR) signaling was required to transform death-inducing signals through the TCR of developing thymocytes into positive selection and, conversely, that activation of the glucocorticoid receptor could modulate TCR-mediated cell death (34). Interpreting our data in the frame of this model, it could be argued that the observed weak signaling through the glucocorticoid pathway should indirectly affect TCR-based selection. The role of glucocorticoids in the thymus remains controversial (35,36), however, and while it was reported that thymic T-cell development and selection proceeds normally in the absence of GR signaling (37), a defect in central selection, in the Iaβ^g7-independent negative deletion of semimature thymocytes, was recently described in the NOD mouse (23), implicating the role of thymic selection in the development of type 1 diabetes in the NOD mouse.

T-cell apoptosis is a crucial event in tolerance induction, during peripheral immunoregulation and during thymic deletion of self-reactive cells (38). There are distinct thymocyte apoptosis pathways described (39–41). Disruption of these apoptotic pathways can prevent the deletion of autoreactive cells (42) and impair the resolution/control of ongoing inflammatory responses (43,44). In fact, MRL mice with defective Fas receptors (MRL/lpr) or Fas ligands (MRL/gld) develop a lupus-like syndrome (45). Moreover, Ctla4-deficient mice develop a lethal lymphoproliferative syndrome and multiorgan inflammation resulting in death at 4 weeks of age (46). It has also been suggested that autoimmune manifestations in the NOD mouse are related to mechanisms controlling lymphocyte apoptosis, and dysfunctions in apoptosis induction of NOD lymphocytes have been reported (22,26,47).

We have previously reported that the low rate of proliferation in NOD immature thymocytes upon complete DP thymocyte depletion maps to the Idd6 region (48). This observation, taken together with the data presented in this article, suggests that Idd6 is involved in the control of both apoptosis and proliferation. We envisage two possible explanations for the involvement of Idd6 in the development of type 1 diabetes. First, Idd6 could represent a gene controlling the survival/apoptosis decisions during T-cell development. Signaling pathways leading to apoptosis and cell proliferation are frequently tightly coupled, and proto-oncoproteins, such as c-Myc and Ras, control both cell proliferation and pathways leading to apoptosis (49,50). Therefore, a single genetic modification could lead to alterations in both apoptosis and proliferations in immature thymocytes. Second, Idd6 could constitute a gene controlling proliferation of immature thymocytes. Such slower dynamics of thymocyte differentiation in the NOD mouse could potentially lead to an accumulation of cells resistant to Dxm-induced apoptosis. NOD mature CD4 T-cells have previously been reported to be resistant to activation-induced cell death due to the interaction with NOD antigen-presenting cells (APCs) (51). Therefore, it remains to be solved whether the apoptosis resistance reported here is due to a thymocyte autonomous effect or if it is mediated through APCs.

This work was supported by grants from the Swedish Research Council and the Portuguese Foundation for Science and Technology.

We wish to acknowledge Hans Andersson for help with statistical analysis, Céu Conceição for technical assistance, and Dr. Mario Penha-Gonçalves for advice and discussions. The experiments reported here comply with the national laws and rules for animal experimentation.