SMAD6 variants in craniosynostosis: genotype and phenotype evaluation

Abstract

Purpose: Enrichment of heterozygous missense and truncating SMAD6 variants was previously reported in nonsyndromic sagittal and metopic synostosis, and interaction of SMAD6 variants with a common polymorphism nearBMP2 (rs1884302) was proposed to contribute to inconsistent penetrance. We determined the occurrence of SMAD6 variants in all types of craniosynostosis, evaluated the impact of different missense variants on SMAD6 function, and tested independently whether rs1884302 genotype significantly modifies the phenotype.

Methods: We performed resequencing of SMAD6 in 795 unsolved patients with any type of craniosynostosis and genotyped rs1884302 in SMAD6-positive individuals and relatives. We examined the inhibitory activity and stability of SMAD6 missense variants.

Results: We found 18 (2.3%) different rare damaging SMAD6 variants, with the highest prevalence in metopic synostosis (5.8%) and an 18.3-fold enrichment of loss-of-function variants comparedwith gnomAD data (P < 10^-7). Combined with eight additional variants, ≥20/26 were transmitted from an unaffected parent but rs1884302 genotype did not predict phenotype.

Conclusion: Pathogenic SMAD6 variants substantially increase the risk of both nonsyndromic and syndromic presentations of craniosynostosis, especially metopic synostosis. Functional analysis is important to evaluate missense variants. Genotyping of rs1884302 is not clinically useful. Mechanisms to explain the remarkable diversity of phenotypes associated with SMAD6 variants remain obscure.

Keywords: BMP2; digenic inheritance; metopic synostosis; protein instability; two-locus.

Fig. 1. Human SMAD6 gene and protein showing variants identified in craniosynostosis (CRS).

Top, SMAD6 comprises four exons; positions of variants affecting translation initiation or splicing are indicated. Middle, cartoon of encoded protein showing conserved domains (MH1 and MH2, including highly conserved L3 region) and PY and PLDLS motifs. Colored shading indicates the position of the MH1 (transparent purple) and MH2 (transparent orange) domains according to Uniprot, Pfam, and CDD resources, with darker shading denoting overlapping domain assignments. Novel or rare (AF_max < 0.000045) variants identified in CRS patients that are also predicted damaging (loss-of-function [LoF], plus missense variants with DS²⁵ ≥ 4) are indicated above the cartoon, whereas below in gray are additional missense variants predicted to have lower pathogenicity (DS ≤ 3 and/or AF_max > 0.000045); negative and positive controls used in the functional assays are colored green and blue, respectively. Frameshifts and stop-gain variants are shown with filled and empty arrowheads, respectively; § = de novo variants; * = novel/rare damaging variants found in addition to the CRS cohort screen; # = AF_max ≥ 0.000045 in gnomAD. Bottom, conservation profiles of inhibitory SMADs SMAD6 (black), SMAD7 (gray), and all SMAD members (SMAD1–8) combined (purple line). AF allele frequency, DS deleterious score.

Fig. 2. Preoperative clinical presentations of craniosynostosis (CRS) in association with pathogenic heterozygous SMAD6 variants.

(a) Subject 8260 aged 4 months with metopic synostosis, the most frequently associated CRS phenotype, showing hypotelorism (front view) and trigonocephaly (top view). Newly described clinical presentations include sagittal and bicoronal synostosis (b, subject 3711 aged 11 months, note narrow, saddle-shaped skull with frontal bossing) and right unicoronal synostosis (c, subject 4370 aged 10 months, note facial asymmetry and recessed brow on right).

Fig. 3. Functional analysis of SMAD6 variants.

(a) Luciferase assay. The cartoon at top shows a simplified representation of the BMP signaling pathway, indicating in red the components transfected into C2C12 cells to perform the assay. Firefly luciferase activity of the BRE-luc transcriptional reporter induced by constitutively active BMPR1A (BMPR1A c.a.) was used to monitor the inhibitory effects of SMAD6 variants on BMP signaling, similar to previously described. Below, graphs represent means ± SEM from three independent experiments. Data were normalized (using Renilla levels), relativized to the wild type (WT) and analyzed by one-way analysis of variance (ANOVA) with Dunnett’s multiple-comparisons test; ^∗P ≤ 0.05, ^∗∗P ≤ 0.01, and ^∗∗∗P ≤ 0.001. Color-coding of SMAD6 variants follows the same scheme as in Fig. 1. (b) Analysis of SMAD6 protein stability. Above are representative examples of western blots (using aliquots of protein extracts from the luciferase assays), showing SMAD6 protein levels (detected with anti-FLAG) compared with anti-GAPDH loading control. Control missense variants (negative, green bar, p.A325T; positive, blue bar, p.C484F) were selected as previously described. Data were normalized (using GAPDH) and relativized to the WT. The bars represent means ± SEM from three independent experiments, analyzed and visualized as in (a). (c) SMAD6 average protein levels of 21 different missense variants (from b) plotted against respective deleterious score (DS).