Establishment of NOD-Pdcd1-/- mice as an efficient animal model of type I diabetes

Abstract

Mice deficient in programmed cell death 1 (PD-1, Pdcd1), an immunoinhibitory receptor belonging to the CD28/cytotoxic T lymphocyte-associated antigen-4 family, spontaneously develop lupus-like autoimmune disease and autoimmune dilated cardiomyopathy on C57BL/6 and BALB/c backgrounds, respectively. However, how PD-1 deficiency induces different forms of autoimmune diseases on these two strains was unknown. Here, we report that PD-1 deficiency specifically accelerates the onset and frequency of type I diabetes in NOD (nonobese diabetic) mice, with strong T helper 1 polarization of T cells infiltrating into islets. These results suggest that PD-1 deficiency accelerates autoimmune predisposition of the background strain, leading to the induction of different forms of autoimmune diseases depending on the genetic background of the strain. Using NOD-Pdcd1-/- mice as an efficient animal model of type I diabetes, we screened diabetes-susceptible loci by genetic linkage analysis. The diabetic incidence of NOD-Pdcd1-/- mice was controlled by five genetic loci, including three known recessive loci [Idd (insulin-dependent diabetes) 1, Idd17, and Idd20] and two previously unidentified dominant loci [Iddp (Idd under PD-1 deficiency) 1 and Iddp2].

Fig. 1.

The early onset and complete penetrance of type I diabetes in NOD-Pdcd1^-/- mice. (a) Incidence of type I diabetes in female NOD-Pdcd1^-/- (filled circles), NOD-Pdcd1^+/- (triangles), and NOD WT (open circles) mice. (b) Incidence of type I diabetes in male mice. Symbols are the same as in a. (c) Severity of insulitis in female NOD-Pdcd1^-/-, NOD-Pdcd1^+/-, and NOD WT mice was evaluated at 4 and 6 weeks of age. Insulitis was scored as follows: presence of periinsulitis (grade 1), moderate insulitis (grade 2), and severe insulitis (grade 3), and the percentage of islets with each grade of insulitis was calculated. Black, red, and pink bars represent the percentages of islets with grade 3, grade 2, and grade 1 insulitis, respectively. White bars represent the percentage of islets without inflammation. The data are the mean of at least five mice for each genotype. On average, 203 islets were analyzed from 12 independent sections for each mouse. (d) Severity of insulitis in female NOD-Pdcd1^-/-, NOD-Pdcd1^+/-, and NOD WT mice at 6 weeks. Insulitis was scored as in c, and the mean score of individual mice is plotted. Bars represent mean value. Symbols are the same as in a. (e) Hematoxylin/eosin staining of representative islets, parotid glands, and submandibular glands of NOD WT (Upper) and NOD-Pdcd^-/- (Lower) mice at 6 weeks.

Fig. 2.

Invasion of CD8⁺ T cells in islets and expression of PD-L1 on beta cells. (a) CD4⁺ (red) and CD8⁺ (blue) T cells and PD-L1⁺ (green) cells were detected in the islets of grade 0, 1, 2, and 3 insulitis of NOD WT (Upper) and NOD-Pdcd1^-/- (Lower) mice. (b) PD-L1 expression on B cells (Left) and CD11c⁺ DCs (Right) in prediabetic NOD WT mice. Red signal represents IgM⁺ B cells (Left) and CD11c⁺ DCs (Right). Green signal represents PD-L1⁺ cells. (c) PD-L1 is expressed on beta cells (Left) and glucagon-producing alpha cells (Right) in prediabetic NOD WT mice. Red signal represents insulin⁺ beta cells (Left) and glucagon⁺ alpha cells (Right). Green signal represents PD-L1⁺ cells. (d and e) CD8/CD4 ratio of islet infiltrates. (d) Representative FACS profiles of islet infiltrates are shown for prediabetic NOD WT and NOD-Pdcd1^-/- mice. (e) Mean CD8/CD4 ratios are shown. Error bars represent standard error.

Fig. 3.

Th1 response is augmented in the islets of NOD-Pdcd1^-/- mice. (a and b) Islet-infiltrating T cells were collected and examined for the production of IFN-γ and IL-4. (a) Representative FACS profiles of NOD WT (Left) and NOD-Pdcd1^-/- (Right) mice are shown. (b) Mean percentage of IFN-γ-producing cells (Left) and IL-4-producing cells (Right) are shown. White bars represent NOD WT mice and black bars represent NOD-Pdcd1^-/- mice. Error bars represent standard error. (c) Production of IFN-γ (Left) and IL-4 (Right) by splenocytes (spl) and pancreatic lymph node cells (pLN) was examined as above. Error bars represent standard error.

Fig. 4.

Genetic association of Idd1, Idd17, and Idd20 with diabetic incidence of NOD-Pdcd1^-/- mice. (a) Diabetic incidence of BC1 backcross progenies (NOD-Pdcd1^-/- × C57BL/6-Pdcd1^-/-)F1 × NOD-Pdcd1^-/-. Filled circles, open circles, and triangles represent cumulative diabetic incidence of NOD-Pdcd1^-/- and C57BL/6-Pdcd1^-/- mice and their BC1 progenies, respectively. (b-d) Logarithm of odds (lod) score plots for individual chromosomes containing the putative locus controlling diabetic incidence in BC1 progenies. The length of each chromosome is adjusted to the same size, although the centimorgan length of each chromosome varies. Genetic maps were constructed by using mapmanager qtx. Arrows indicate the positions of microsatellite markers used for genotyping. The lod score is given on the y axis. Dashed lines indicate the suggestive lod score of 1.9 and the significant lod score of 3.3 (31).

Fig. 5.

Identification of Iddp1 and Iddp2. (a) Diabetic incidence of (NOD-Pdcd1^-/- × C57BL/6-Pdcd1^-/-)F2 progenies. Closed circles, open circles, and triangles represent cumulative diabetic incidence of NOD-Pdcd1^-/- and C57BL/6-Pdcd1^-/- mice and their F2 progenies, respectively. (b and c) lod score plots for individual chromosomes containing the putative dominant locus controlling diabetic incidence among F2 progenies. Dashed lines indicate the suggestive lod score of 2.0 and the significant lod score of 3.4 (31). Representation of chromosomes, lod score, and genetic markers are as described in the legend of Fig. 4.