Polymyositis in Kooiker dogs is associated with a 39 kb deletion upstream of the canine IL21/IL2 locus

Abstract

Recently we characterized polymyositis in the Dutch Kooiker dog. The familial occurrence of the disease were suggestive of an inherited cause. Here we report the results of our molecular genetic investigation. A genome-wide association study of 33 cases and 106 controls indicated the involvement of a region on chromosome CFA19 (p = 4.7*10-10). Haplotype analysis indicated that the cases shared a 2.9 Mb region in the homozygous or the heterozygous state. Next Generation Sequencing of genomic DNA implicated a deletion of a 39 kb DNA fragment, located 10 kb upstream of the neighbouring interleukin genes IL21 and IL2. The frequency of the deletion allele was 0.81 in the available cases and 0.25 in a random sample of the Kooiker dog breed. Leukocytes of affected, untreated dogs that were homozygous for the deletion overexpress IL21 and IL2 upon stimulation with mitogens. We suggest that elements located 10-49 kb upstream of the IL21/IL2 locus play an important role in the regulation of the canine genes and that deletion of these elements is a risk factor for polymyositis in Kooiker dogs. Postulating causality, the penetrance of the disease phenotype was estimated at 10-20% for homozygous dogs and 0.5-2% for dogs that were heterozygous for the deletion. Our results suggest that distant variants upstream of IL21 could also be important for human autoimmune diseases that have been found to be associated with the IL21/IL2 chromosome region.

Fig 1. Genome-wide association of polymyositis in Kooiker dogs.

Twenty-eight affected dogs and 106 controls were genotyped at more than 170,000 SNP positions. Differences between the groups were evaluated with Genabel software and the -log₁₀(p) value is plotted against the chromosomal position. A) Manhattan plot summarizing results from all chromosomes. The red line marks the significance threshold after Bonferroni correction. B) Results from chromosome 19. The position of the IL21/IL2 locus is indicated.

Fig 2. Integrative genomics view of next generation sequence data of the region upstream of IL21 in Kooiker dogs.

The DNA of the polymyositis affected dog did not generate data from a 39 kb region. The same was observed in the data of the other 7 affected dogs that were analyzed by NGS. Control 1 shows average coverage in the region, while the coverage of control 2 is approximately half of that. A PCR assay, which distinguishes the reference and the deletion allele confirmed the heterozygosity of this control. The position on CFA19 corresponds to CanFam3.1. The corresponding positions of the borders of the deletion on the reference genome UU_Cfam_GSD_1.0 are 18,414,090 and 18,453,526.

Fig 3. Three primer assay for genotyping of the 39 kb deletion variant upstream of IL21.

A) The positions of the PCR primers I, II and III are indicated relative to the 39 kb deletion. The combination I–II generates a product of 662 bp from the normal variant. The combination I–III produces a fragment of 242 bp from the deletion variant only. B) Result of the assay for heterozygotes (lanes 1 and 2), a dog homozygous for the deletion (lane 3), and a homozygous normal dog (lane 4). C) Genotype distribution of the 39 kb deletion (del) in 102 IM cases and 112 random Kooiker dogs.

Fig 4. Segregation of the 39 kb deletion and polymyositis in a pedigree of Kooiker dogs.

The birth year of the first-born founder of the pedigree was 1992 and DNA of all dogs was available. W: reference allele; Δ: 39 kb deletion allele. The frequency of the deletion allele was 0.24 in the 29 founders of this pedigree and 0.85 in the 27 affected dogs.

Fig 5. Stimulation of IL21 and IL2 expression in PBMCs from PM affected dogs.

Blood samples for leukocyte isolation were taken from two untreated Kooiker dogs with PM. The dogs PM1 and PM2 were homozygous for the 39 kb deletion. The cells were cultured with and without stimulating mitogen, RNA was isolated and RT-qPCR performed. Plotted are the Ct-values of the unstimulated cells minus the Ct-values of the stimulated cells from the same dogs. The control cells were left over from a routine diagnostic procedure. The control dog did not have the 39 kb deletion.