The type 1 diabetes candidate gene Dexi does not affect disease risk in the nonobese diabetic mouse model

Abstract

Genome-wide association studies have implicated more than 50 genomic regions in type 1 diabetes (T1D). A T1D region at chromosome 16p13.13 includes the candidate genes CLEC16A and DEXI. Conclusive evidence as to which gene is causal for the disease association of this region is missing. We previously reported that Clec16a deficiency modified immune reactivity and protected against autoimmunity in the nonobese diabetic (NOD) mouse model for T1D. However, the diabetes-associated SNPs at 16p13.13 were described to also impact on DEXI expression and others have argued that DEXI is the causal gene in this disease locus. To help resolve whether DEXI affects disease, we generated Dexi knockout (KO) NOD mice. We found that Dexi deficiency had no effect on the frequency of diabetes. To test for possible interactions between Dexi and Clec16a, we intercrossed Dexi KO and Clec16a knockdown (KD) NOD mice. Dexi KO did not modify the disease protection afforded by Clec16a KD. We conclude that Dexi plays no role in autoimmune diabetes in the NOD model. Our data provide strongly suggestive evidence that CLEC16A, not DEXI, is causal for the T1D association of variants in the 16p13.13 region.

FIGURE 1

Generation of Dexi KO NOD mice. (A) Schematic representation of the region targeted by CRISPR-Cas9 genome editing in the Dexi genomic (top) region and of the mutant allele #1 (bottom). Only the first 8bp of exon 1 remain, followed by a 3bp insertion and a 544bp deletion at the start of intron 1-2. (B) Inheritance pattern of the two mutant Dexi alleles (#1 and #2) present in the founder male NOD mouse. The proportion of wild-type and mutant alleles inherited from the founder in the F1 progeny (total 41 mice, of which 10 carried allele #1 and 9 carried allele #2) is shown. (C) Quantitation of Dexi mRNA in the spleen, thymus and pancreatic islets of WT and Dexi KO mice by quantitative PCR. n=4 mice per group, data show individual values and mean +/− SEM and are representative of at least three similar experiments. *** P < 0.001 (two-tailed t-test). (D) Detection of Dexi protein by western blotting following immune-precipitation with anti-Dexi antibody. Data are shown for WT, Clec16a KD, Dexi KO and Clec16a KD/Dexi KO mice and are representative for two similar experiments. (E-G) Quantitation of Clec16a (E), Socs1 (F) and Ciita (G) mRNA by quantitative PCR in spleen, thymus and pancreatic islets. n=2-4 mice per group. Data show individual values, mean +/− SEM and are representative of at least two similar experiments. (H) Interferon beta expression in bone marrow-derived macrophages from WT and Dexi KO mice stimulated with poly I:C. Data show individual values, mean +/− SEM and are representative of two similar experiments.

FIGURE 2

Dexi KO does not modify the frequency of diabetes in NOD mice. (A) Cyclophosphamide-accelerated diabetes was measured in groups of WT (n=20), Dexi KO (n=18), Clec16a KD (n=19) and Clec16a KD/Dexi KO (n=17) male NOD mice injected with cyclophosphamide at age 9-10 weeks. (B) Spontaneous diabetes was measured in groups of WT (n=54), Dexi KO (n=47), Clec16a KD (n=39) and Clec16a KD/Dexi KO (n=50) female NOD mice. Differences between groups were measured using the Log-rank test. * P < 0.05, ** P < 0.01, *** P < 0.001. (C) WT and Dexi KO mice (9 weeks old) were injected with glucose intraperitoneally to test their glucose tolerance. Data show mean +/− SEM values from 6 mice per group. (D) Histological analysis was performed on 10 week-old WT and Dexi KO mice to quantify the degree of insulitis. Data show the proportion of WT islets (n=402) and Dexi KO islets (n=354) with no infiltration, moderate or severe insulitis from 3 mice per group. Fisher’s exact test P=0.1 comparing the proportion of infiltrated islets in WT and Dexi KO mice.