Identification of novel genetic risk loci in Maltese dogs with necrotizing meningoencephalitis and evidence of a shared genetic risk across toy dog breeds

Abstract

Necrotizing meningoencephalitis (NME) affects toy and small breed dogs causing progressive, often fatal, inflammation and necrosis in the brain. Genetic risk loci for NME previously were identified in pug dogs, particularly associated with the dog leukocyte antigen (DLA) class II complex on chromosome 12, but have not been investigated in other susceptible breeds. We sought to evaluate Maltese and Chihuahua dogs, in addition to pug dogs, to identify novel or shared genetic risk factors for NME development. Genome-wide association testing of single nucleotide polymorphisms (SNPs) in Maltese dogs with NME identified 2 regions of genome-wide significance on chromosomes 4 (chr4:74522353T>A, p = 8.1×10-7) and 15 (chr15:53338796A>G, p = 1.5×10-7). Haplotype analysis and fine-mapping suggests that ILR7 and FBXW7, respectively, both important for regulation of immune system function, could be the underlying associated genes. Further evaluation of these regions and the previously identified DLA II locus across all three breeds, revealed an enrichment of nominal significant SNPs associated with chromosome 15 in pug dogs and DLA II in Maltese and Chihuahua dogs. Meta-analysis confirmed effect sizes the same direction in all three breeds for both the chromosome 15 and DLA II loci (p = 8.6×10-11 and p = 2.5×10-7, respectively). This suggests a shared genetic background exists between all breeds and confers susceptibility to NME, but effect sizes might be different among breeds. In conclusion, we identified the first genetic risk factors for NME development in the Maltese, chromosome 4 and chromosome 15, and provide evidence for a shared genetic risk between breeds associated with chromosome 15 and DLA II. Last, DLA II and IL7R both have been implicated in human inflammatory diseases of the central nervous system such as multiple sclerosis, suggesting that similar pharmacotherapeutic targets across species should be investigated.

Figure 1. Manhattan plot of genome wide association analysis in Maltese dogs with necrotizing meningoencephalitis.

The raw –log₁₀ p-values for each SNP as determined by Fisher’s exact tests are plotted (y axis) against the chromosome position (x axis). The horizontal gray line represents the threshold for significant association after Bonferroni correction. Regions that reach genome wide significance are indicated in black.

Figure 2. Fine mapping of genome wide significant regions on (A) chromosomes 4 and (B) 15 in Maltese dogs by haplotype analysis.

The raw –log₁₀ p-values for each haplotype test are plotted (y axis) against the chromosome position (x axis). The colored lines represent haplotypes formed by a 5-SNP sliding window across the region, and the black lines represent haplotypes formed by the underlying LD structure calculated by the 4-gamete rule. In the bottom, the LD between the SNPs in the region is shown (dark red = high D’; blue = low D’). The haploblocks based on the 4-gamete rule are indicated in black.

Figure 3. Forest plots of four SNPs in the dog leukocyte antigen II region on chromosome 12 across three toy breeds with necrotizing meningoencephalitis.

(a) BICF2P22942: p = 1.57×10⁻⁷, OR = 0.18 (0.09–0.35), (b) BICF2P738783: p = 4.11×10⁻⁸, OR = 0.18 (0.10–0.34), (c) BICF2P178662: p = 1.11×10⁻⁹, OR = 9.48 (4.19–21.40), and (d) BICF2P608380: p = 8.6×10⁻¹¹, OR = 11.08 (4.72–26.01). The 95% confidence interval for each study is represented by a horizontal line, and the point estimate is given by a square, the height of which is inversely proportional to the standard error of the estimate after each study. The summary OR is indicated by a diamond with horizontal limits as the confidence limits and height inversely proportional to its standard error. The meta-analyses for all these SNPs were significant and effect sizes were in the same direction.