Genetic Reconstitution of Autoimmune Type 1 Diabetes With Two Major Susceptibility Genes in the Rat

Tester Moriyama (TM) rats were obtained from Dr. T. Serikawa at the Institute of Laboratory Animals, Kyoto University. Both TM and KDP rats (19) were maintained under specific pathogen-free (SPF) conditions with a 12-h light/dark cycle. A commercial diet (CE-2; CLEA Japan, Tokyo, Japan) and water were available ad libitum at the Institute for Experimental Animals, Kobe University School of Medicine. The TM and KDP rats were crossed to produce the F1 progeny. Following this, six successive backcrosses (N7) to the TM rats were performed. An intercross between the N7 animals was conducted to obtain the congenic strain (N7F1) carrying the mutated Cblb allele of the KDP rat on the TM genetic background. This congenic strain was maintained at the same facility. All animal experiments were approved by the Committee on Animal Experimentation, Kobe University School of Medicine (permission no. P-031207) and carried out in accordance with the Guidelines for Animal Experimentation at Kobe University. The KDP rats are available from Japan SLC (Shizuoka, Japan). The congenic strain (TM.KDP-Cblb) has been deposited (ref. no. 0354) and is available from the National Bio-Resource Project of Rat in Japan (www.anim.med.kyoto-u.ac.jp/nbr/).

Phenotyping was completed using a previously described protocol (16). Diabetes was defined as glycosuria positivity and blood glucose levels ≥250 mg/dl under ad libitum dietary conditions. Data were obtained from five generations (N7F1 to N7F5) of the congenic strain and one generation of the KDP strain.

Simple sequence-length polymorphism markers used in this study have been described elsewhere (RATMAP, available at http://ratmap.gen.gu.se/; Rat Genome Database, available at http://rgd.mcw.edu/; 16,17,19). Genotyping was performed as described in prior investigations (16,17,19).

Histological analysis was performed as described previously (7,16), with some modifications. Briefly, tissues were fixed in 10% formalin, and paraffin sections were stained with hematoxylin and eosin. Serial sections were viewed via light microscopy by an examiner blind to the experimental conditions of the animals. Each animal was rated on the degree of insulitis, which ranged from none to severe (none [0%], slight [>0 and ≤5%], mild [>5 and ≤20%], moderate [>20 and ≤70%], and severe [>70%]) based mainly on the percentage of moderately and severely infiltrated islets. The term “slight insulitis” refers to at least one infiltrated islet across the sections and a percentage of infiltrated islets of ≤5%. Animals were also rated on the degree of thyroiditis, which ranged from none to severe (none [0%], slight to mild [>0% and ≤20%], and moderate to severe [>20%]) based on the percentage of infiltrated regions in thyroid sections. Data were obtained from five generations (N7F1 to N7F5) of the congenic strain and one generation of the KDP strain, except that the degree of insulitis at 90 days of age was evaluated in the N7F9 generation of the congenic strain.

Differences in the incidence of diabetes and in the degree of insulitis or thyroiditis were assessed using χ² tests. Differences in body weights were assessed using two-tailed Student’s t tests. Nominal P values are listed for all analyses.

By crossing the TM and KDP rats followed by six successive backcrosses with the speed congenic approach (20), we successfully produced a congenic strain carrying the mutated Cblb allele of the KDP rat on the TM genetic background harboring the RT1^u haplotype, TM.KDP-Cblb. Genetic profiling revealed that the ∼14-cM (∼15-Mb) region between D11Rat68 and D11Mgh4 on chromosome 11 harboring the mutated Cblb allele of the KDP rat had been introgressed into the TM genetic background and that the KDP genome was not detected on other chromosomes (Table 1). Since the Cblb homozygous mutants have poor reproductive ability (17,19), we have been maintaining the congenic strain that has the Cblb region in the heterozygous state.

We compared the incidence of diabetes in the congenic strain with that in KDP rats under the same SPF condition (Fig. 1A). At 210 days of age, 26% (9 of 35) of the Cblb homozygous mutants of the congenic strain developed diabetes, while none of the heterozygous (0 of 22) or wild-type (0 of 14) animals developed the disease (TM.KDP-Cblb homozygotes vs. heterozygotes, χ² [df = 1] = 6.72, P = 0.0095; TM.KDP-Cblb homozygotes vs. wild-types, χ² [df = 1] = 4.41, P = 0.0357). In contrast, all (18 of 18) of the Cblb homozygous mutants of KDP rats developed diabetes at 210 days of age (KDP homozygotes vs. TM.KDP-Cblb homozygotes, χ² [df = 1] = 26.3, P < 0.0001). In addition to the low incidence, the onset of diabetes was delayed in the congenic strain. KDP rats developed diabetes as early as 60 days of age (96.3 ± 8.8 days [means ± SE], n = 18), while the congenic strain developed the disease at ∼120 days of age (156.3 ± 9.4 days, n = 9). Despite the low incidence and the delay of onset, it is noteworthy that type 1 diabetes was reconstituted on the non-KDP genetic background with the two major susceptibility genes (RT1^u haplotype and Cblb mutation).

We also compared body weight and changes in blood glucose level in the congenic and KDP strains (Fig. 1B–D). At 210 days of age, both females and males in the congenic strain exhibited higher body weight than KDP rats (KDP vs. congenic: female 230 ± 4.5 vs. 246.5 ± 3.7 g, P = 0.014; male 377.8 ± 5.7 vs. 450.5 ± 10.2 g, P < 0.0001). In contrast, there was no difference in changes in blood glucose levels between the strains, except for age of onset of diabetes: blood glucose levels elevated rapidly from ∼100 to ∼400 mg/dl within a week in both the congenic and KDP strains.

We then investigated the degree of insulitis shortly after the onset of diabetes in diabetic animals and at 210 days of age in nondiabetic animals (Fig. 2). In the congenic strain, the Cblb homozygous mutants (n = 35) showed slight to severe insulitis, while heterozygous (n = 22) and wild-type (n = 14) animals exhibited none or only slight insulitis (TM.KDP-Cblb homozygotes vs. heterozygotes, χ² [df = 4] = 43.7, P < 0.0001; TM.KDP-Cblb homozygotes vs. wild-types, χ² [df = 4] = 37.2, P < 0.0001). In contrast, all of the Cblb homozygous mutants (n = 18) of KDP rats showed severe insulitis, which is consistent with 100% incidence of diabetes (KDP homozygotes vs. TM.KDP-Cblb homozygotes, χ² [df = 3] = 26.2, P < 0.0001). Although the degree of insulitis was relatively mild in the congenic strain compared with that in KDP rats, these results indicate that insulitis was reconstituted on the non-KDP genetic background with the two major susceptibility genes (RT1^u haplotype and Cblb mutation).

To clarify the onset of insulitis in the congenic strain, we further investigated the degree of insulitis at 90 days of age (Fig. 2). At that age, KDP rats showed 70% incidence of diabetes, whereas none of the animals in the congenic strain developed diabetes. The Cblb homozygous mutants (n = 10) of the congenic strain showed none to moderate insulitis, which is milder than at 210 days of age. These results indicate that the onset and progression of insulitis were delayed in the congenic strain, thus resulting in the delayed onset of diabetes.

We also investigated the degree of thyroiditis of the congenic and KDP strains at the same ages as those for insulitis (Fig. 3). Regarding the congenic strain, 23% (8 of 35) and 20% (7 of 35) of the Cblb homozygous mutants showed slight to mild and moderate to severe thyroiditis, respectively, while heterozygous (n = 22) and wild-type (n = 14) animals exhibited none or only slight thyroiditis (TM.KDP-Cblb homozygotes vs. heterozygotes, χ² [df = 2] = 10.0, P = 0.0067; TM.KDP-Cblb homozygotes vs. wild-types, χ² [df = 2] = 8.65, P = 0.0133). In KDP rats, 11% (2 of 18) of the Cblb homozygous mutants showed moderate to severe thyroiditis. In both strains, mild to severe thyroiditis was observed in either the nondiabetic animals or those with delayed onset of diabetes. In contrast to the results for insulitis and diabetes, the congenic strain showed a higher incidence of thyroiditis than that seen in KDP rats (KDP homozygotes vs. TM.KDP-Cblb homozygotes, χ² [df = 2] = 6.43, P = 0.0401). Because thyroiditis is one of the autoimmune symptoms observed in KDP rats (17), these results indicate that the autoimmune phenotype was reconstituted on the non-KDP genetic background with the two major susceptibility genes (RT1^u haplotype and Cblb mutation).

In the present study, we successfully reconstituted autoimmune type 1 diabetes with two major susceptibility genes, the RT1^u haplotype and Cblb mutation, on a non-KDP genetic background using the congenic approach. Genetic analysis of type 1 diabetes in the MHC-matched backcrosses, (TM × KDP)F1 × KDP and (LETO [Long-Evans Tokushima Otsuka] × KDP)F1 × KDP, revealed that the genetic predisposition to diabetes is accounted for by a recessively acting locus, Iddm/kdp1 (16). This raised the possibility that introgression of Iddm/kdp1 (Cblb) onto the TM or LETO genetic backgrounds might confer sufficient genetic susceptibility for type 1 diabetes. Because the LETO strain was derived from the same outbred colony as LETL (the original strain of KDP) (6), a considerable part of the genetic background may be common among LETO, LETL, and KDP rats. Therefore, we selected the TM strain as the recipient strain for backcrossing. To confirm the susceptibility loci and identify the responsible genes in animal models of type 1 diabetes, many congenic strains have been produced in NOD mice and BB rats. These studies have shown that most congenic strains carrying susceptibility allele(s) on nondiabetic genetic backgrounds do not develop diabetes (12–14). In addition, most congenic strains carrying resistant allele(s) on diabetes-prone genetic backgrounds suppressed the development of the disease. The BB congenic strains carrying segments of chromosome 6 or 18 of the spontaneously hypertensive rat on the BB/OK background showed a reduction in diabetes frequency compared with BB/OK rats (86% vs. 14 or 34%, respectively) (21). Moreover, NOD congenic strains carrying resistant alleles (Idd3, Idd5, Idd10, or Idd18) suppressed the development of both insulitis and diabetes (22). In contrast, the NOD congenic strain carrying B6-derived alleles on chromosome 13 accelerated diabetes, suggesting that B6 mice harbor more diabetogenic alleles than NOD mice for this locus (23). These findings indicate that restricted combinations of susceptibility alleles are not sufficient for the development of type 1 diabetes on nondiabetic genetic backgrounds. Therefore, to our knowledge, the present study is the first demonstration of genetic reconstitution of autoimmune type 1 diabetes on a nondiabetic genetic background with a combination of susceptibility genes.

The KDP rats showed 100% incidence of diabetes in this study. However, as reported previously (19), the incidence of diabetes in each generation fluctuates from 70 to 100%. In addition, the present KDP colony at Japan SLC exhibited ∼80% incidence of diabetes (N. Masui, H. Asai, S. Yokose, personal communication). Therefore, the incidence of diabetes in the present KDP colony is similar to that of the previous reports (16,19). The degree of insulitis in the congenic strain supports the previous findings that homozygosity for the KDP allele (Cblb mutation) at the Iddm/kdp1 (Cblb) locus is strongly associated with the development of moderate to severe insulitis and that Iddm/kdp1 (Cblb) acts in a recessive manner (16). However, in addition to the observation in (TM × KDP)F1 animals (16), slight insulitis was observed in 68% (15 of 22) of Cblb heterozygous and 43% (6 of 14) of wild-type animals in the congenic strain (Fig. 2). These data indicate that slight insulitis may occur in the KDP, TM, or (TM × KDP)F1 genetic backgrounds, irrespective of Cblb mutation, and that the Cblb mutation acts in a semidominant manner in the development of slight insulitis.

The congenic strain showed a higher incidence of thyroiditis than KDP rats. Mild to severe thyroiditis was rarely found in animals with early onset of diabetes but was found in those with delayed onset of diabetes in both the KDP and congenic strains. In addition, mild to severe thyroiditis was found in nondiabetic congenic rats. These results suggest that the onset and progression of thyroiditis are delayed compared with those of insulitis and are independent of the development of insulitis. Similar to the case of insulitis, homozygosity for the Cblb mutation is necessary for the development of mild to severe thyroiditis. However, it is not clear whether RT1^u is involved in the development of thyroiditis. The study of BB rats suggested that the susceptibility of RT1 haplotypes to thyroiditis is distinct from that of insulitis and that RT1^a confers greater susceptibility to thyroiditis than RT1^u (24).

Although type 1 diabetes was reconstituted in the congenic strain, the incidence of diabetes was low and progression of the disease was delayed compared with KDP rats. Since breeding and phenotyping of both the congenic and KDP strains were performed in the same room at the same facility, environmental factors are likely to be similar for both strains, except for the microenvironment, such as cages or mothers. Thus, the primary cause of the difference in incidence of diabetes between the two strains is most likely due to genetic rather than environmental factors. These genetic factors (modifiers) may well interact with environmental factors to promote or suppress the development of insulitis and type 1 diabetes. There are diabetes-promoting modifiers in the KDP genetic background, whereas there are diabetes-suppressing modifiers in the TM genetic background. Differences in the phenotypes of Cblb-deficient mice also support the involvement of modifiers: Bachmaier et al. (25) did, whereas Chiang et al. (26) did not, observe spontaneous development of several autoimmune phenotypes in mice. Based on genetic analyses of backcrosses between the KDP rat and three strains (TM, LETO, and BN rats), we previously identified the two recessively acting genes, the RT1^u haplotype and Cblb mutation, and found that most of the genetic predisposition to diabetes in these crosses can be accounted for by these two major genes (16,17). However, in these analyses using backcrosses, dominantly acting genes could not be identified. The results in this study, together with the previous genetic studies, strongly suggest that there are dominantly acting diabetes-promoting modifiers in the KDP genetic background. The two recessively acting genes, the RT1^u haplotype and Cblb mutation, are necessary for the development of type 1 diabetes, while dominantly acting diabetes-promoting modifiers are not essential but influence the development of the disease. Although both KDP and TM rats have the same RT1^u haplotype, it is possible that subtle differences in coding or regulatory sequences of RT1.B^u and RT1.D^u between KDP and TM rats may influence the incidence and age of onset of diabetes. In addition, accumulating evidence in both humans and NOD mice (27–29) indicates that genes outside the class II MHC may contribute to susceptibility to type 1 diabetes.

Genetic variations in the human CBLB gene do not seem to be a major cause of type 1 diabetes (30–32). However, human orthologues of diabetes-promoting modifiers of the KDP rat may be involved in the development of the disease. The responsible genes for these modifiers and their physiological roles remain to be investigated. The congenic strain established in this study should be useful for mapping and identifying such modifiers and for the elucidation of mechanisms underlying the development of autoimmune diseases.

This study was supported by Grants-in-Aid for Scientific Research (15700313) and Specially Promoted Research (15002002) from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

We thank the late Professor K. Komeda (Tokyo Medical University) and Drs. T. Serikawa (Kyoto University) and K. Yasuda (International Medical Center of Japan) for their helpful advice during the course of this study. We also thank A.C. Kentner (University of Ottawa) and T. Hirata for critical reading of the manuscript and technical assistance, respectively.

Part of this study was conducted at the Department of Cellular and Molecular Medicine, Graduate School of Medicine, Chiba University, with which N.Y., H.-Y.W., and S.S. were formerly affiliated.