Genotype-phenotype associations in Sotos syndrome: an analysis of 266 individuals with NSD1 aberrations

Abstract

We identified 266 individuals with intragenic NSD1 mutations or 5q35 microdeletions encompassing NSD1 (referred to as "NSD1-positive individuals"), through analyses of 530 subjects with diverse phenotypes. Truncating NSD1 mutations occurred throughout the gene, but pathogenic missense mutations occurred only in functional domains (P < 2 x 10(-16)). Sotos syndrome was clinically diagnosed in 99% of NSD1-positive individuals, independent of the molecular analyses, indicating that NSD1 aberrations are essentially specific to this condition. Furthermore, our data suggest that 93% of patients who have been clinically diagnosed with Sotos syndrome have identifiable NSD1 abnormalities, of which 83% are intragenic mutations and 10% are 5q35 microdeletions. We reviewed the clinical phenotypes of 239 NSD1-positive individuals. Facial dysmorphism, learning disability, and childhood overgrowth were present in 90% of the individuals. However, both the height and head circumference of 10% of the individuals were within the normal range, indicating that overgrowth is not obligatory for the diagnosis of Sotos syndrome. A broad spectrum of associated clinical features was also present, the occurrence of which was largely independent of genotype, since individuals with identical mutations had different phenotypes. We compared the phenotypes of patients with intragenic NSD1 mutations with those of patients with 5q35 microdeletions. Patients with microdeletions had less-prominent overgrowth (P = .0003) and more-severe learning disability (P = 3 x 10(-9)) than patients with mutations. However, all features present in patients with microdeletions were also observed in patients with mutations, and there was no correlation between deletion size and the clinical phenotype, suggesting that the deletion of additional genes in patients with 5q35 microdeletions has little specific effect on phenotype. We identified only 13 familial cases. The reasons for the low vertical transmission rate are unclear, although familial cases were more likely than nonfamilial cases (P = .005) to carry missense mutations, suggesting that the underlying NSD1 mutational mechanism in Sotos syndrome may influence reproductive fitness.

Figure 1

Facial features and NSD1 mutations in patients with Sotos syndrome. A, Typical facial features of patients with different NSD1 abnormalities: 5q35 microdeletion, partial NSD1 deletion, truncating mutation, and missense mutation. B, Number of truncating and missense mutations identified in different NSD1 domains, demonstrating that missense mutations occur only within functional domains. The number of truncating mutations is at the top of each section of NSD1, and the number of missense mutations is shown at the bottom. Functional domains are shown as shaded boxes.C, Position of missense mutations (red) in PHD and C5HCH domains, showing bias toward mutations at consensus cysteine and histidine residues (black).

Figure 2

Cumulative frequency distribution graphs for patients with NSD1 mutations or 5q35 microdeletions. A, Height distribution. B, Head circumference distribution. The distributions are significantly different for height (Wilcoxon rank-sum test) but not for head circumference.

Figure 3

Bar graph showing the frequency and severity of learning disability in patients with NSD1 mutations (blackened) and in those with 5q35 microdeletions (hatched). The actual number of patients is given above each bar. The distributions are significantly different (Wilcoxon rank-sum test).