Intronic variant in POU1F1 associated with canine pituitary dwarfism

Abstract

The anterior pituitary gland secretes several endocrine hormones, essential for growth, reproduction and other basic physiological functions. Abnormal development or function of the pituitary gland leads to isolated or combined pituitary hormone deficiency (CPHD). At least 30 genes have been associated with human CPHD, including many transcription factors, such as POU1F1. CPHD occurs spontaneously also in mice and dogs. Two affected breeds have been reported in dogs: German Shepherds with a splice defect in the LHX3 gene and Karelian Bear Dogs (KBD) with an unknown genetic cause. We obtained samples from five KBDs presenting dwarfism and abnormal coats. A combined analysis of genome-wide association and next-generation sequencing mapped the disease to a region in chromosome 31 and identified a homozygous intronic variant in the fourth exon of the POU1F1 gene in the affected dogs. The identified variant, c.605-3C>A, resided in the splice region and was predicted to affect splicing. The variant's screening in three new prospective cases, related breeds, and ~ 8000 dogs from 207 breeds indicated complete segregation in KBDs with a carrier frequency of 8%, and high breed-specificity as carriers were found at a low frequency only in Lapponian Herders, a related breed. Our study establishes a novel canine model for CPHD with a candidate POU1F1 defect.

Fig. 1

Affected KBDs. a An affected female KBD (case 1) at the age of 9 months. b An affected female dog (case 1) and an unaffected male littermate at 10 months. c An adult affected female KBD (case 2) showing severe hair loss. d An affected KBD puppy from the Swedish litter at the age of 6 months was pictured with a dog from another breed. Photos published with permission from the dog owners under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/)

Fig. 2

A pedigree drawn around the affected dogs suggests an autosomal recessive condition. The case numbers are denoted as #1, #2 etc. (case 1, case 2). Genotypes for the POU1F1 c.605-3C>A are marked in the pedigree. Cases 1–5 were used in the genome-wide analysis, whereas cases 6–8 were recognised later in the variant validation phase of the study

Fig. 3

Genome-wide association study in five cases and 139 controls. SNP correlation structures (r2) in images b and d were calculated with plink. In images e and g, only ten randomly selected controls are shown for better visualisation. a Manhattan plot. Two genome-wide significant loci were revealed on chromosomes 9 and 31. b A locus plot of chr9:30–38 Mb. c A quantile–quantile plot of raw (UNADJ) and genomic control corrected (GC) p values. Lambda was 1.88 before genomic control correction. d A locus plot of chr31:0–10 Mb. e A genotype plot of chr9:30–38 Mb. One of the top SNPs at chr9:34197651 is indicated with a black triangle. All five cases are above the horizontal line, and ten randomly selected controls below the line. f An MDS plot with two dimensions. Cases are denoted in red and controls in black. g Genotype plot of chr31:0–10 Mb. The candidate variant in POU1F1 is indicated with a red triangle and the top SNP at chr31:4769865 with a black triangle. All five cases are above the horizontal line, and ten randomly selected controls below the line